NZ772974A - Access node for end-to-end beamforming communications system - Google Patents

Abstract

A method of communication at an access node of a communication system comprising a plurality of access nodes at geographically distributed locations providing a communication service to user terminals distributed over multiple return user beam coverage areas via an end-to-end relay comprising multiple return receive/transmit signal paths, comprises receiving a return downlink signal comprising a composite return signal and a relay beacon signal from the end-to-end relay. The composite return signal comprises return uplink signals from a plurality of the user terminals relayed by the end-to-end relay. The method further comprises demodulating the relay beacon signal to obtain receive timing information; multiplexing the composite return signal with the receive timing information to obtain a multiplexed composite return signal; and sending the multiplexed composite return signal to a return beamformer for determining a return beam signal by applying return beam weights to the multiplexed composite return signal and at least one other multiplexed composite return signal, wherein the return beam weights account for wireless signal uplink paths up to the end-to-end relay, the multiple return receive/transmit signal paths through the end-to-end relay, and wireless signal downlink paths down from the end-to-end relay.

Description

ACCESS NODE FOR END-TO-END BEAMFORMING COMMUNICATIONS SYSTEM TECHNICAL FIELD The disclosed systems, methods, and apparatuses relate to end-to-end rming in a system using an end-to-end relay.

BACKGROUND Wireless communication systems, such as ite communication systems, provide a means by which data, including audio, video, and various other sorts of data, may be communicated from one location to another. Information originates at a first station, such as a first ground-based station, and is transmitted to a wireless relay, such as a communication satellite. Information received by the wireless relay is retransmitted to a second station, such as a second ground-based station. In some wireless relay communication systems, either the first or second station (or both) are mounted on a craft, such as an aircraft, watercraft, or landcraft.

Information may be transmitted in just one direction (e.g., from a first ground-based n to a second ground-based station only) or may be itted in both directions (e.g., also from the second ground-based station to the first ground-based n).

In a wireless relay communication system in which the wireless relay is a ite, the satellite may be a tionary satellite, in which case the satellite’s orbit is synchronized to the on of the Earth, keeping the coverage area of the satellite essentially stationary with respect to the Earth. In other cases, the satellite is in an orbit about the Earth that causes the coverage area of the satellite to move over the surface of the Earth as the satellite traverses its orbital path.

The signals that are ed to or from a first station may be directed by using an antenna that is shaped to focus the signal into a narrow beam. Such antennas lly have a paraboloid shaped reflector to focus the beam.

In some cases, a beam may be formed onically by adjusting the gain and phase (or time delay) of signals that are transmitted, received, or both from several elements of a phased array antenna. By properly selecting the relative phase and gain transmitted and/or received by each element of a phased array antenna, the beam may be directed. In most cases, all of the energy being transmitted from a ground-based n is intended to be received by one wireless relay. Similarly, information received by the second station is lly received from one wireless relay at a time. Therefore, it is typical that a transmit beam that is formed to transmit information to the wireless relay (whether by use of electronic beamforming or by use of an a with a shaped reflector) is relatively narrow to allow as much of the transmitted energy as possible to be directed to the wireless relay. Likewise, a receive beam that is formed to receive information from the wireless relay is typically narrow to gather energy from the direction of the wireless relay with minimal interference from other sources.

In many cases of interest, the signals that are itted from the ss relay to the first and second stations are not directed to a single station. Rather, the wireless relay is able to transmit s over a relatively large geographic area. For example, in one satellite communication system, a satellite may service the entire continental United States. In such a case, the satellite is said to have a satellite coverage area that includes the entire continental United States. Nonetheless, in order to increase the amount of data that may be transmitted h a satellite, the energy transmitted by the satellite is focused into beams. The beams may be directed to geographic areas on the Earth.

BRIEF PTION OF THE S The drawings are provided for purposes of illustration only and merely depict examples.

These drawings are provided to facilitate the reader’s understanding of the disclosed method and apparatus. They do not limit the breadth, scope, or applicability of the claimed invention. For clarity and ease of illustration, these drawings are not necessarily made to scale. is an illustration of an example of a ite communication system. is a diagram g an example pattern of beams that covers the continental United States. is an illustration of an example of the forward link of a satellite communication system in which the satellite has a phased array feed per beam on—board beamforming capability. is an illustration of an example of the d link of a satellite communication system having ground—based rming. is an illustration of an example end—to—end rming system. is an illustration of example signal paths for signals in the return direction. is an illustration of example signal paths in the return direction from a user terminal. is a simplified illustration of an example end—to—end return channel matrix model. is an illustration of example signal paths in the forward direction. is an illustration of example signal paths in the forward direction to a user terminal located within a user beam coverage area. is a simplified illustration of an example end—to—end forward channel matrix model. is an illustration of an example end—to—end relay satellite supporting forward and return data. is an illustration of an example of an uplink frequency range being divided into two portions. is an ration of an example end—to—end relay being time multiplexed between forward data and return data. is a block diagram of components of an example end—to—end relay implemented as a satellite. is a block diagram of an example transponder including a phase r. is a graph of example signal th patterns of several antenna elements. is an ration of example 3 dB signal strength contours for several antenna elements. is an illustration of example overlapping signal strength patterns of l antenna elements.

A — 20E is an illustration of example overlapping 3 dB signal strength contours for l antenna ts. is an illustration of an example enumeration of 16 antenna elements and their overlapping 3 dB signal strength rs. is a table showing example mappings of receive antenna elements to transmit antenna elements h 16 transponders. is an illustration of a section of a paraboloid antenna reflector and an array of elements centered at the focal point of the parabola. is an illustration of a cross—section of a paraboloid antenna reflector and an array of elements placed away from the focal point of the parabola. is an illustration of an example relay ge area (shown with single cross— hatching) and the area (shown with double cross—hatching) defined by the points within the relay coverage area that are also contained within six antenna element coverage areas.

W0 2016/209332 is an illustration of an e relay antenna pattern in which all of the points within a relay coverage area are also ned within at least four antenna element coverage areas. is an ration of an e bution of access nodes (ANs) and user beam coverage areas. is an example graph of normalized forward and return link capacity as a function of the number of ANS deployed. is a block diagram of an example ground segment 502 for an end—to—end beamforming system. is a block diagram of an example forward/retum beamformer. is a block diagram of an example forward beamformer comprising multiple return lice beamformers with time—domain de—multiplexing and lexing. is an illustration of a simplified example ground segment showing the operation of a forward time-slice beamformer. is a block diagram of an example return beamformer comprising multiple return time-slice beamformers with time-domain de-multiplexing and multiplexing. is an illustration of a fied example ground segment showing the operation of a return beamformer employing time-domain multiplexing. is a block diagram of an example multi—band forward/return beamformer that employs sub—band de—multiplexing and multiplexing. and is an illustration of example timing alignment for the forward link. is a block diagram of an example AN. is a block diagram of part of an example of an AN. is a block diagram of an example AN 515 in which multiple frequency sub— bands are processed tely. is an illustration of an example —end rming system for enabling distinct user—link and feeder—link coverage areas. is an illustration of an example model of signal paths for signals carrying return data on the end—to—end return link. is an illustration of an example model of signal paths for signals carrying forward data on the end—to—end forward link.

FIGS. 44A and 44B are an illustration of an e forward signal path and return signal path, respectively. is an illustration of an example of an end—to—end relay visible Earth coverage area. is an illustration of an e of an end—to—end relay North American coverage area.

FIGS. 47A and 47B are block diagrams of an example forward signal path and return signal path, respectively, each having selective tion of multiple user—link antenna subsystems.

FIGS. 48A and 48B are an illustration of an example of an end—to—end relay coverage area that includes multiple, ively activated user coverage areas. is a block diagram of an example forward signal path having selective tion of multiple user-link a subsystems and multiple -link antenna subsystem.

Reference ators (e.g., 100) are used herein to refer to aspects of the drawings.

Similar or like aspects are typically shown using like numbers. A group of similar or like ts may be referred to collectively by a single reference designator (e.g., 200), while individual elements of the group may be referred to by the reference designator with an appended letter (e.g., 200a, 200b).

The figures are not ed to be exhaustive or to limit the claimed invention to the precise form disclosed. The disclosed method and apparatus may be practiced with modification and alteration, and that the invention is d only by the claims and the equivalents f.

DETAILED DESCRIPTION [005 8] This detailed description is organized as follows. First, an introduction to wireless relay communication systems using satellite communication and beamforming are described. Second, end—to—end beamforming is described generally and at the system level using satellite end—to—end beamforming as an example, although application of end—to—end beamforming is not limited to satellite communications. Third, ion of forward and return data is described in context of end—to—end beamforming. Fourth, end—to—end relays and their antennas are described using a communication satellite as an example. Next, ground networks to form the end—to—end beams are described, including related aspects, such as delay equalization, feeder—link impairment removal, and beam weight computation. Finally, end—to—end beamforming with distinct user—link and feeder—link coverage areas is bed, as well as systems with multiple ge areas.

Satellite Communication is an illustration of an example of a hub and spoke satellite communication system 100. The satellite serves as an example of a wireless relay. Though many es are described throughout this disclosure in context of a satellite or satellite communication , such examples are not intended to be limited to satellite; any other suitable wireless relay may be used and operate in a similar fashion. The system 100 ses a ground—based Earth station 101, a ication satellite 103, and an Earth transmission source, such as a user terminal 105. A satellite coverage area may be broadly defined as that area from which, and/or to which, either an Earth transmission source, or an Earth receiver, such as a ground—based Earth station or a user terminal, can communicate through the satellite. In some systems, the coverage area for each link (e.g., forward uplink ge area, forward nk coverage area, return uplink coverage area, and return downlink ge area) can be different. The forward uplink coverage area and return uplink coverage area are collectively referred to as the uplink satellite coverage area. Similarly, the forward downlink coverage area and the return downlink coverage area are collectively referred to as the downlink satellite coverage area. While the satellite coverage area is only active for a satellite that is in service (e.g., in a service orbit), the satellite can be considered as having (e.g., can be designed to have) a satellite antenna pattern that is independent of the relative location of the satellite with respect to the Earth. That is, the satellite antenna pattern is a pattern of bution of energy transmitted from an antenna of a ite (either transmitted from or ed by the antenna of the satellite). The satellite antenna n illuminates mits to, or receives from) a particular satellite coverage area when the satellite is in a service orbit. The satellite ge area is defined by the satellite antenna pattern, an orbital position and attitude for which the satellite is designed, and a given antenna gain threshold. In general, the intersection of an antenna pattern (at a particular ive antenna gain, e. g. 3 dB, 4 dB, 6 dB 10 dB from peak gain) with a particular physical region of interest (e.g., an area on or near the earth surface) defines the coverage area for the antenna. as can be designed to provide a particular antenna n (and/or coverage area) and such antenna patterns can be determined computationally (e.g., by analysis or simulation) and/or measured experimentally (e.g., on an antenna test range or in actual use).

While only one user terminal 105 is shown in the figure for the sake of simplicity, there are typically many user terminals 105 in the system. The satellite communication system 100 operates as a point to multi—point system. That is, the Earth station 101 within the satellite coverage area can send information to, and receive information from, any of the user terminals 105 within the satellite coverage area. However, the user terminals 105 only communicate with the Earth station 101. The Earth station 101 receives forward data from a communication network 107, modulates the data using a feeder link modem 109 and transmits the data to the satellite 103 on a forward feeder uplink 111. The satellite 103 relays this forward data to user terminals 105 on the forward user downlink imes called a d service downlink) 113.

In some cases, the forward direction communication from the Earth station 101 is intended for several of the user terminals 105 (e.g., information is multicast to the user terminals 105). In some cases, the forward communication from the Earth station 101 is intended for only one user terminal 105 (e.g., unicast to a particular user terminal 105). The user terminals 105 transmit return data to the ite 103 on a return user uplink (sometimes called a return service uplink) 115. The ite 103 relays the return data to the Earth station 101 on a return feeder downlink 117. A feeder-link modem 109 demodulates the return data, which is forwarded to the communication network 107. This -link capability is generally shared by a number of user terminals 105. is a diagram g an example of one ration of beam coverage areas of a satellite to service the continental United States. Seventy beams are shown in the example configuration. A first beam 201 covers approximately two thirds of the state of Washington. A second beam 203 adjacent to the first beam 201 covers an area immediately to the east of the first beam 201. A third beam 205 approximately covers Oregon to the south of the first beam 201. A fourth beam 207 covers an area y southeast of the first beam 201. Typically, there is some overlap n adjacent beams. In some cases, a multi—color (e.g., two, three or four— color re—use pattern) is used. In an example of a four—color pattern, the beams 201, 203, 205, 207 are individually allocated a unique combination of frequency (e.g., a frequency range or ranges or one or more channels) and/or antenna polarization (e.g., in some cases an antenna may be configured to transmit signals with a hand circular polarization (RHCP) or a left—hand circular polarization ; other zation techniques are available). Accordingly, there may be relatively little mutual interference between signals itted on different beams 201, 203, 205, 207. These combinations of frequency and antenna polarization may then be re—used in the repeating non—overlapping “four—color” re—use pattern. In some situations, a desired communication capacity may be achieved by using a single color. In some cases, time g among beams and/or other erence mitigation techniques can be used.

Within some limits, focusing beams into smaller areas and thus increasing the number of beams, increases the data capacity of the satellite by allowing greater opportunity for frequency re—use. However, sing the number of beams can increase the xity of the system, and in many cases, the complexity of the satellite. xity in the design of a satellite typically results in larger size, more weight, and r power consumption. Satellites are expensive to launch into orbit. The cost of launching a satellite is determined in part by the weight and size of the satellite. In addition, there are absolute limits on the weight and size of a satellite if the satellite is to be launched using presently available rocket technology. This leads to tradeoffs between features that may be designed into a satellite. Furthermore, the amount of power that may be provided to ents of a satellite is limited. Therefore, weight, size, and power consumption are parameters to be considered in the design of a ite.

Throughout this disclosure, the term receive antenna t refers to a physical transducer that converts an electro—magnetic signal to an electrical signal, and the term transmit antenna element refers to a physical transducer that launches an electro-magnetic signal when excited by an electrical signal. The antenna element can include a horn, septum polarized horn (e.g., which may function as two combined elements with different polarizations), multi-port multi-band horn (e.g., dual-band 20 GHz/30 GHz with dual polarization LHCP/RHCP), - backed slot, inverted—F, slotted waveguide, Vivaldi, Helical, loop, patch, or any other configuration of antenna element or combination of interconnected sub-elements. An antenna element has a corresponding a pattern, which describes how the antenna gain varies as a function of direction (or angle). An antenna element also has a coverage area which ponds to an area (e. g., a portion of the Earth surface) or volume (e.g., a portion of the Earth surface plus airspace above the surface) over which the antenna element provides a desired level of gain (e.g., within 3 dB, 6 dB, 10 dB, or other value relative to a peak gain of the antenna t).

The coverage area of the antenna element may be modified by various structures such as a reflector, frequency selective surface, lens, radome, and the like. Some satellites, including those described herein, can have several transponders, each able to ndently receive and transmit signals. Each transponder is coupled to antenna elements (e.g., a receive element and a transmit element) to form a receive/transmit signal path that has a ent radiation pattern na n) from the other receive/transmit signal paths to create unique beams that may be allocated to different beam coverage areas. It is common for a single receive/transmit signal path to be shared across multiple beams using input and/or output multiplexers. In both cases, the number of aneous beams that may be formed is generally limited by the number of receive/transmit signal paths that are deployed on the satellite.

Beamforming Beamforming for a communication link may be performed by adjusting the signal phase (or time delay), and mes signal amplitude, of signals transmitted and/or received by multiple elements of one or more antenna arrays with overlapping coverage areas. In some cases, some or all antenna elements are arranged as an array of constituent receive and/or transmit elements that cooperate to enable end—to—end beamforming, as described below. For transmissions (from transmit elements of the one or more antenna arrays), the relative phases, and sometimes amplitudes, of the transmitted signals are adjusted, so that the energy transmitted by transmit antenna elements will constructively superpose at a desired on. This phase/amplitude adjustment is commonly referred to as “applying beam weights” to the transmitted signals. For reception (by receive ts of the one or more antenna ), the relative phases, and mes amplitudes, of the received signals are adjusted (i.e., the same or ent beam weights are applied) so that the energy received from a desired location by receive antenna ts will constructively ose at those receive antenna elements. In some cases, the beamformer computes the desired antenna element beam s. The term beamforming may refer in some cases to the application of the beam weights. Adaptive beamformers include the function of dynamically computing the beam s. Computing the beam weights may require direct or indirect discovery of the communication channel characteristics. The processes of beam weight computation and beam weight application may be performed in the same or different system ts.

The antenna beams may be steered, selectively formed, and/or otherwise reconfigured by applying different beam weights. For example, the number of active beams, coverage area of beams, size of beams, relative gain of beams, and other parameters may be varied over time.

Such versatility is desirable in n situations. rming antennas can generally form relatively narrow beams. Narrow beams may allow the signals itted on one beam to be guished from signals transmitted on the other beams (e.g., to avoid interference).

Accordingly, narrow beams can allow frequency and polarization to be re—used to a greater extent than when larger beams are formed. For example, beams that are narrowly formed can service two discontiguous coverage areas that are non—overlapping. Each beam can use both a right hand polarization and a left hand polarization. Greater reuse can increase the amount of data transmitted and/or received. 2016/026815 Some satellites use rd beamforming (OBBF) to electronically steer an array of antenna elements. is an illustration of a satellite system 300 in which the satellite 302 has phased array multi—feed per beam (MFPB) on—board beamforming capability. In this example, the beam weights are computed at a ground based computation center and then transmitted to the satellite or pre— stored in the satellite for application (not shown). The forward link is shown in although this architecture may be used for forward links, return links, or both forward and return links. Beamforming may be employed on the user link, the feeder link, or both. The illustrated d link is the signal path from one of a plurality of gateways (GWs) 304 to one or more of a plurality of user terminals within one or more spot beam coverage areas 306. The satellite 302 has a receive antenna array 307, a transmit a array 309, a down—converter (D/C) and gain module 311, a receive beamformer 313, and a transmit beamformer 315. The satellite 302 can form beams on both the feeder link 308 and the user link 310. Each of the L elements of the receive array 307 receives K signals from the K GWs 304. For each of the K feeder link beams that are to be created (e.g., one beam per GW 304), a different beam weight is applied (e.g., a phase/amplitude adjustment is made) by the receive beamformer 313 to each signal received by each of the L receive antenna array elements (of e antenna array 307).

Accordingly, for K beams to be formed using a receive a array 307 having L e antenna elements, K different beam weight vectors of length L are applied to the L signals received by the L receive antenna array elements. The receive beamformer 313 within the ite 302 adjusts the phase/amplitude of the s received by the L receive antenna array elements to create K receive beam signals. Each of the K e beams are focused to receive a signal from one GW 304. ingly, the receive beamformer 313 outputs K receive beam signals to the D/C and gain module 311. One such receive beam signal is formed for the signal received from each transmitting GW 304.

The D/C and gain module 311 down—converts each of the K receive beam signals and adjusts the gain appropriately. K signals are output from the BIG and gain module 311 and coupled to the transmit beamformer 315. The it beamformer 315 applies a vector of L weights to each of the K signals for a total of L x K transmit beam weights to form K beams on the user downlink 310.

In some cases, significant processing capability may be needed within the satellite to control the phase and gain of each antenna element that is used to form the beams. Such processing power increases the complexity of the satellite. In some cases, satellites may e with ground—based beamforming (GBBF) to reduce the complexity of the satellite while still providing the advantage of electronically forming narrow beams. is an illustration of one example of a ite communication system 400 having forward GBBF. GBBF is performed on the forward user link 317 via an L element array similar to that described above. The phases/amplitudes of the signals transmitted on the user link 317 are weighted such that beams are formed. The feeder link 319 uses a Single Feed per Beam (SFPB) scheme in which each receive and it antenna element of an antenna 324 is dedicated to one feeder link beam.

Prior to transmission from a GW or GWs 304, for each of the K forward feeder link beams, a transmit beamformer 321 applies a respective one of K beam weight vectors, each of length L, to each of K signals to be itted. Determining the K vectors of L weights and applying them to the signals enables K forward beams to be formed on the ground for the forward user downlink 317. On the feeder uplink 319, each of the L different signals is lexed into a frequency division multiplexed (FDM) signal by a multiplexer 323 (or the like). Each FDM signal is transmitted by the GWs 304 to one of the receive antenna elements in the antenna 324 on the feeder link 319. An FDM receiver 325 on the satellite 327 receives the s from the antenna 324. An analog to digital converter (A/D) 326 converts the received analog signals to digital signals. A digital channel processor 328 demultiplexes the FDM signals, each of which was appropriately weighted by the rmer 321 for transmission h one of the L ts of an array of transmit antenna elements of a transmit antenna 329. The digital channel processor 328 outputs the signals to a digital to analog converter (D/A) 331 to be converted back to analog form. The analog s of the D/A 331 are up—converted and amplified by an up—converter (U/C) and gain stage 330 and transmitted by the associated element of the transmit antenna 329. A complimentary process occurs in reverse for the return beams. Note that in this type of system the FDM feeder link requires L times as much bandwidth as the user beams making it impractical for systems with wide data bandwidths or systems that have a large number of elements L.

End-to-end Beamforming Systems The end—to—end beamforming systems described herein form end—to—end beams through an end—to—end relay. An end—to—end beamforming system can connect user terminals with data s/sinks. In contrast to the beamforming systems discussed above, in an end—to—end beamforming , beam weights are computed at a central processing system (CPS) and end— to—end beam weights are applied within the ground network r than at a satellite). The s within the end—to—end beams are transmitted and received at an array of access nodes (ANs), which may be satellite access node (SANs). As described above, any suitable type of end—to—end relays can be used in an end—to—end beamforming system, and ent types of ANs may be used to communicate with different types of end—to—end relays. The term “central” refers to the fact that the CPS is accessible to the ANs that are involved in signal transmission and/or reception, and does not refer to a particular geographic on at which the CPS resides. A beamformer within a CPS computes one set of end—to—end beam weights that accounts for: (l) the wireless signal uplink paths up to the end—to—end relay; (2) the receive/transmit signal paths h the end—to—end relay; and (3) the wireless signal nk paths down from the end-to— end relay. The beam weights can be represented mathematically as a matrix. As discussed above, OBBF and GBBF satellite s have beam weight vector dimensions set by the number of a elements on the satellite. In contrast, end—to—end beam weight vectors have dimensions set by the number of ANs, not the number of elements on the —end relay. In general, the number of ANs is not the same as the number of antenna elements on the —end relay. r, the formed end—to—end beams are not terminated at either transmit or receive antenna ts of the end—to—end relay. , the formed end—to—end beams are effectively relayed, since the end-to—end beams have uplink signal paths, relay signal paths (via a satellite or other suitable end-to-end relay), and downlink signal paths.

Because the end-to-end beamforming takes into t both the user link and the feeder link (as well as the end-to-end relay) only a single set of beam weights is needed to form the desired end—to—end user beams in a particular direction (e.g., forward user beams or return user beams). Thus, one set of end—to-end forward beam weights (hereafter referred to simply as forward beam weights) results in the signals itted from the ANs, through the forward uplink, through the end—to—end relay, and through the forward downlink to combine to form the end-to-end forward user beams (hereafter referred to as forward user beams). Conversely, signals transmitted from return users through the return uplink, through the end—to—end relay, and the return downlink have end—to—end return beam weights (hereafter referred to as return beam weights) applied to form the end—to-end return user beams (hereafter ed to as return user beams). Under some conditions, it may be very difficult or impossible to distinguish between the characteristics of the uplink and the downlink. Accordingly, formed feeder link beams, formed user beam directivity, and individual uplink and downlink carrier to interference ratio (C/I) may no longer have their traditional role in the system design, while concepts of uplink and downlink signal—to—noise ratio (Es/No) and end—to—end C/I may still be relevant. is an illustration of an example end—to—end beamforming system 500. The system 500 includes: a ground t 502; an end—to—end relay 503; and a plurality of user terminals 517. The ground segment 502 comprises M ANs 515, spread geographically over an AN WO 09332 coverage area. ANs 515 and user terminals 517 can be collectively referred to as Earth receivers, Earth transmitters, or Earth transceivers, depending upon the particular functionality at issue, since they are located on, or near, the Earth and both it and receive signals. In some cases, user terminals 517 and/or ANs 515 can be located in aircraft, watercraft or mounted on landcraft, etc. In some cases, the user terminals 517 can be phically distributed. The ANs 515 can be geographically distributed. The ANs 515 e signals through a distribution network 518 to a CPS 505 within the ground segment 502. The CPS 505 is connected to a data source (not shown), such as, for example, the internet, a video headend or other such entity.

User terminals 517 may be grouped with other nearby user terminals 517 (e.g., as illustrated by user terminals 517a and 517b). In some cases, such groups of user terminals 517 are serviced by the same user beam and so reside within the same geographic forward and/or return user beam coverage area 519. A user terminal 517 is within a user beam if the user terminal 517 is within the coverage area serviced by that user beam. While only one such user beam coverage area 519 is shown in to have more than one user terminal 517, in some cases, a user beam coverage area 519 can have any suitable number of user terminals 517.

Furthermore, the depiction in is not intended to te the relative size of different user beam coverage areas 519. That is, the user beam coverage areas 519 may all be approximately the same size. Alternatively, the user beam ge areas 519 may be of varying sizes, with some user beam coverage areas 519 much larger than . In some cases, the number of ANs 515 is not equal to the number of user beam coverage areas 519.

The end—to—end relay 503 relays signals wirelessly between the user terminals 517 and a number of network access nodes, such as the ANs 515 shown in The end—to—end relay 503 has a ity of signal paths. For example, each signal path can include at least one receive antenna element, at least one transmit antenna element, and at least one transponder (as is discussed in detail below). In some cases, the ity of receive antenna elements are arranged to receive signals reflected by a receive reflector to form a receive antenna array. In some cases, the plurality of transmit antenna elements is arranged to transmit signals and thus to form a transmit antenna array.

In some cases, the end—to—end relay 503 is provided on a satellite. In other cases, the end— to—end relay 503 is ed on an aircraft, blimp, tower, underwater structure or any other suitable structure or vehicle in which an —end relay 503 can reside. In some cases, the system uses different frequency ranges (in the same or different ncy bands) for the uplinks and downlinks. In some cases, the feeder links and user links are in different frequency ranges.

In some cases, the end—to—end relay 503 acts as a passive or active reflector.

As described herein, various features of the end—to—end relay 503 enable end—to—end beamforming. One feature is that the end—to—end relay 503 includes le transponders that, in the context of end—to—end beamforming systems, induce multipath between the ANs 515 and the user terminals 517. r feature is that the antennas (e.g., one or more antenna subsystems) of the end—to—end relay 503 contribute to end—to—end beamforming, so that d and/or return user beams are formed when properly beam—weighted signals are communicated through the multipath induced by the end—to—end relay 503. For example, during forward communications, each of multiple onders receives a tive superposed composite of (beam ed) forward uplink signals 521 from le (e.g., all) of the ANs 515 (referred to herein as composite input forward signals), and the transponders output corresponding composite signals (referred to herein as forward nk signals). Each of the forward downlink signals can be a unique composite of the beam-weighted d uplink signals 521, which, when transmitted by the transmit antenna elements of the end—to—end relay 503, superpose to form the user beams 519 in desired locations (e.g., recovery locations within forward user beams, in this case). Return end-to-end beamforming is similarly enabled. Thus, the end-to-end relay 503 can cause multiple superpositions to occur, thereby enabling end-to-end beamforming over induced multipath channels.

Return Data is an illustration of an example model of signal paths for signals carrying return data on the —end return link. Return data is the data that flows from user terminals 517 to the ANs 515. Signals in flow from right to left. The signals originate with user terminals 517. The user terminals 517 transmit return uplink signals 525 (which have return user data s) up to the end—to—end relay 503. Return uplink signals 525 from user terminals 517 in K user beam coverage areas 519 are received by an array of L receive/transmit signal paths 1702.

In some cases, an uplink ge area for the end—to—end relay 503 is defined by that set of points from which all of the L receive antenna elements 406 can receive signals. In other cases, the relay coverage area is defined by that set of points from which a subset (e.g., a desired number more than 1, but less than all) of the L receive antenna elements 406 can receive signals.

Similarly, in some cases, the downlink coverage area is defined by the set of points to which all of the L transmit antenna elements 409 can reliably send signals. In other cases, the downlink coverage area for the end—to—end relay 503 is defined as that set of points to which a subset of the transmit antenna elements 409 can ly send s. In some cases, the size of the subset of either receive antenna elements 406 or transmit antenna elements 409 is at least four. In other WO 09332 cases, the size of the subset is 6, 10, 20, 100, or any other number that provides the desired system performance.

For the sake of simplicity, some examples are described and/or illustrated as all L e antenna elements 406 receiving signals from all points in the uplink coverage area and/or all L transmit antenna elements 409 transmitting to all points in the downlink coverage area. Such descriptions are not intended to require that all L elements receive and/or transmit signals at a significant signal level. For example, in some cases, a subset of the L receive antenna ts 406 es an uplink signal (e.g., a return uplink signal 525 from a user terminal 517, or a forward uplink signal 521 from an AN 515), such that the subset of receive antenna elements 406 receives the uplink signal at a signal level that is close to a peak ed signal level of the uplink signal (e.g., not substantially less than the signal level corresponding to the uplink signal having the highest signal level); others of the L e antenna elements 406 that are not in the subset receive the uplink signal at an appreciably lower level (e.g., far below the peak received signal level of the uplink ). In some cases, the uplink signal received by each receive antenna element of a subset is at a signal level within 10 dB of a maximum signal level received by any of the receive antenna elements 406. In some cases, the subset includes at least 10% of the receive antenna elements 406. In some cases, the subset includes at least 10 receive a elements 406.

Similarly, on the transmit side, a subset of the L transmit antenna elements 409 transmits a downlink signal to an Earth receiver (e.g., a return downlink signal 527 to an AN 515, or a forward downlink signal 522 to a user terminal 517), such that the subset of transmit antenna elements 409 transmits the downlink signal to the er with a received signal level that is close to a peak transmitted signal level of the downlink signal (e.g., not substantially less than the signal level corresponding to the downlink signal having the highest received signal level); others of the L transmit antenna ts 409 that are not in the subset transmit the downlink signal such that it is received at an appreciably lower level (e.g., far below the peak transmitted signal level of the downlink ). In some cases, the signal level is within 3 dB of a signal level corresponding to a peak gain of the transmit antenna element 409. In other cases, the signal level is within 6 dB of the signal level corresponding to a peak gain of the transmit antenna element 409. In yet other cases, the signal level is within 10 dB of the signal level corresponding to a peak gain of the transmit antenna element 409.

In some cases, the signal received by each receive a element 406 originates at the same source (e.g., one of the user terminals 517) due to overlap in the receive antenna pattern of each receive antenna element. However, in some cases, there may be points within the end—to- end relay coverage area at which a user terminal is located and from which not all of the receive antenna ts can receive the signal. In some such cases, there may be a icant number of receive a elements that do not (or cannot) receive the signal from user terminals that are within the end—to—end relay coverage area. However, as described herein, inducing multipath by the —end relay 503 can rely on receiving the signal by at least two receive elements.

As shown in and sed in r detail below, in some cases, a receive/transmit signal path 1702 comprises a receive antenna element 406, a transponder 410, and a transmit a element 409. In such cases, the return uplink signals 525 are received by each of a plurality of onders 410 via a respective receive antenna element 406. The output of each receive/transmit signal path 1702 is a return downlink signal 527 corresponding to a respective ite of received return uplink signals. The return downlink signal is created by the receive/transmit signal path 1702. The return downlink signal 527 is transmitted to the array ofM ANs 515. In some cases, the ANs 515 are placed at geographically distributed locations (e.g., reception or recovery locations) throughout the end-to-end relay coverage area. In some cases, each transponder 410 couples a respective one of the receive antenna elements 406 with a respective one of the transmit antenna elements 409. Accordingly, there are L different ways for a signal to get from a user terminal 517 located in a user beam coverage area 519 to a particular AN 515. This creates L paths between a user terminal 517 and an AN 515. The L paths n one user terminal 517 and one AN 515 are referred to collectively as an end-to-end return multipath channel 1908 (see . Accordingly receiving the return uplink signal 525 from a transmission location within a user beam coverage area 519, h the L transponders 410, creates L return nk signals 527, each itted from one of the transponders 410 (i.e., h L collocated communication paths). Each end—to—end return multipath channel 1908 is associated with a vector in the uplink radiation matrix Ar, the payload matrix E, and a vector in downlink radiation matrix Ct. Note that due to antenna element coverage ns, in some cases, some of the L paths may have relatively little energy (e.g., 6 dB, 10 dB, 20 dB, 30 dB, or any other suitable power ratio less than other paths). A superposition 1706 of return downlink 527 signal is received at each of the ANs 515 (e.g., at M geographically distributed reception or recovery locations). Each return downlink signal 527 comprises a superposition of a plurality of the transmitted return downlink signals 527, resulting in a respective composite return signal.

The respective composite return signals are coupled to the return beamformer 531 (see FIGS. 5 and 29). illustrates an example end—to—end return link 523 from one user terminal 517 located within a user beam coverage area 519 to the ANs 515. The return uplink signal 525 WO 09332 transmitted from the user terminal 517 is received by the array of L receive antenna elements 406 on the end—to—end relay 503 (e.g., or received by a subset of the L e antenna elements 406).

Ar is the L x K return uplink radiation . The values of the return uplink radiation matrix model the signal path from a reference location in the user beam coverage area 519 to the end—to—end relay receive antenna elements 406. For example, Arm is the value of one element of the return uplink radiation matrix (1'. e. the amplitude and phase of the path) from a reference location in the 1St user beam coverage area 519 to the Lm e antenna element. In some cases, all of the values in the return uplink radiation matrix Ar may be non—zero (e.g., there is a significant signal path from the reference location to each of the receive antenna elements of the receive antenna array).

E (dimension L x L) is the payload matrix and provides the model (amplitude and phase) of the paths from the receive antenna elements 406 to the transmit antenna elements 409. A "payload" of an end—to—end relay 503, as used herein, generally includes the set of components of the -end relay 503 that affect, and/or are ed by, signal communications as they are received by, relayed through, and itted from the end-to-end relay 503. For example, an end-to-end relay payload can include antenna elements, reflectors, transponders, etc.; but the end-to-end relay can further include batteries, solar cells, sensors, and/or other components not considered herein as part of the payload (since they do not affect signals when operating normally). Consideration of the set of components as a d can enable mathematically modeling the overall impact of the end-to—end relay as a single payload matrix E). The predominant path from each e antenna element 406 to each corresponding transmit antenna element 409 is modeled by the value that lies on the al of the payload matrix E.

Assuming there is no crosstalk between receive/transmit signal paths, the off—diagonal values of the d matrix are zero. In some cases, the crosstalk may not be zero. Isolating the signal paths from each other will minimize crosstalk. In some cases, since the alk is negligible, the payload matrix E can be estimated by a diagonal matrix. In some cases, the off—diagonal values (or any other suitable ) of the payload matrix can be treated as zero, even where there is some signal impact corresponding to those values, to reduce mathematical complexity and/or for other reasons.

Ct is the M x L return downlink radiation matrix. The values of the return downlink radiation matrix model the signal paths from the transmit antenna elements 409 to the ANs 515.

For example, Ct3,2 is the value of the return downlink ion matrix (e.g., the gain and phase of the path) from the second transmit antenna element 40% to the third AN 515C. In some cases, 2016/026815 all of the values of the downlink radiation matrix Ct may be non—zero. In some cases, some of the values of the downlink ion matrix Ct are essentially zero (e.g., the antenna pattern established by a corresponding transmit a elements 409 of the transmit antenna array is such that the transmit a element 409 does not transmit useful signals to some of the ANS 515).

As can be seen in the —end return multipath channel from a user terminal 517 in a particular user beam ge area 519 to a particular AN 515 is the sum of the L different paths. The end—to—end return multipath channel has multipath induced by the L unique paths h the transponders 410 in the end—to—end relay. As with many multipath channels, the paths' amplitudes and phases can add up favorably (constructively) to produce a large end— to—end channel gain or unfavorably (destructively) to produce a low end—to—end channel gain.

When the number of different paths, L, between a user terminal and an AN is large, the end-to— end channel gain can have a Rayleigh distribution of the amplitude. With such a distribution, it is not uncommon to see some end—to—end channel gains from a particular user terminal 517 to a particular AN 515 that are 20 dB or more below the average level of the channel gain from a user terminal 517 to an AN 515. This -end beamforming system intentionally s a multipath environment for the end-to-end path from any user terminal to any AN. is a simplified illustration of an example model of all the end-to-end return multipath channels from user beam ge areas 519 to ANs 515. There are M x K such end— to—end return multipath channels in the end—to—end return link (i.e., M from each of the K user beam coverage areas 519). Channels 1908 connect user terminals in one user beam coverage area 519 to one AN 515 over L different receive/transmit signal paths 1702, each path going through a different one of the L receive/transmit signal paths (and associated transponders) of the relay. While this effect is referred to as “multipat ” herein, this multipath differs from conventional multipath (e.g., in a mobile radio or multiple—input multiple—output (MIMO) system), as the multiple paths herein are intentionally induced (and, as described herein, affected) by the L receive/transmit signal paths. Each of the M x K end—to—end return multipath channels that originate from a user terminal 517 within a particular user beam coverage area 519 can be modeled by an end—to—end return multipath channel. Each such end—to—end return multipath channel is from a reference (or ry) location within the user beam coverage area 519 to one of the ANs 515.

Each of the M x K end—to—end return multipath channels 1908 may be individually modeled to compute a ponding element of an M x K return l matrix Hret. The return channel matrix Hret has K vectors, each having dimensionality equal to M, such that each vector models the end—to—end return channel gains for multipath communications between a reference location in one of a respective K user beam coverage areas and the M ANs 515. Each end—to—end return multipath channel couples one of the M ANS 515 with a reference location within one of K return user beams via L transponders 410 (see . In some cases, only a subset of the L transponders 410 on the end—to—end relay 503 is used to create the end—to—end return multipath channel (e.g., only a subset is considered to be in the signal path by contributing significant energy to the end—to—end return multipath channel). In some cases, the number of user beams K is greater than the number of transponders L that is in the signal path of the —end return multipath channel. Furthermore, in some cases, the number of ANs M is greater than the number of transponders L that is in the signal path of the end—to—end return multipath channel 1908. In an example, the element 2 of the return channel matrix Hret is associated with the channel from a reference on in the second user beam ge area 1903 to the fourth AN 1901. The matrix Hret models the end—to—end channel as the product of the matrices Ct x E x Ar (see . Each element in Hret models the end—to—end gain of one —end return multipath channel 1908. Due to the multipath nature of the channel, the channel can be t to a deep fade. Return user beams may be formed by the CPS 505. The CPS 505 computes return beam weights based on the model of these M x K signal paths and forms the return user beams by applying the return beam weights to the plurality of composite return signals, each weight being computed for each end-to-end return multipath channel that couples the user terminals 517 in one user beam coverage area with one of the plurality of ANs 515. In some cases, the return beam weights are computed before receiving the composite return signal. There is one end—to—end return link from each of the K user beam coverage areas 519 to the M ANS 515. The weighting (i.e., the complex relative phase/amplitude) of each of the s received by the M ANs 515 allows those signals to be combined to form a return user beam using the rming lity of the CPS 505 within the ground segment 502. The computation of the beam weight matrix is used to determine how to weight each end-to—end return multipath channel 1908, to form the plurality of return user beams, as described in more detail below. User beams are not formed by directly adjusting the relative phase and amplitude of the signals transmitted by one end—to—end relay antenna element with respect to the phase and ude of the signals transmitted by the other end—to—end relay a elements. Rather, user beams are formed by ng the weights ated with the M x K channel matrix to the M AN signals.

It is the plurality of ANs that provide the receive path diversity, single transmitter (user terminal) to multiple receivers (ANs), to enable the successful transmission of information from any user terminal in the presence of the ionally induced multipath channel.

WO 09332 Forward Data is an illustration of an example model of signal paths for signals carrying forward data on the end—to—end forward link 501. Forward data is the data that flows from ANS 515 to user als 517. Signals in this figure flow from right to left. The signals originate with M ANs 515, which are located in the footprint of the end—to—end relay 503. There are K user beam coverage areas 519. Signals from each AN 515 are relayed by L receive/transmit signal paths 2001.

The receive/transmit signal paths 2001 transmit a relayed signal to user terminals 517 in user beam coverage areas 519. Accordingly, there may be L different ways for a signal to get from a particular AN 515 to a user terminal 517 located in a user beam coverage area 519. This creates L paths between each AN 515 and each user terminal 517. Note that due to antenna element coverage patterns, some of the L paths may have less energy than other paths. illustrates an example end—to—end forward link 501 that couples a ity of access nodes at geographically buted locations with a user terminal 517 in a user beam (e.g., located at a recovery location within a user beam coverage area 519) via an end-to-end relay 503. In some cases, the forward data signal is received at a beamformer prior to generating forward uplink s. A plurality of forward uplink signals is generated at the beamformer and communicated to the plurality of ANs 515. For example, each AN 515 receives a unique (beam weighted) forward uplink signal generated according to beam s corresponding to that AN 515. Each AN 515 has an output that transmits a forward uplink signal via one ofM uplinks.

Each forward uplink signal comprises a forward data signal associated with the forward user beam. The forward data signal is “associated with” the forward user beam, since it is intended to be received by user terminals 517 serviced by the user beam. In some cases, the forward data signal comprises two or more user data streams. The user data streams can be multiplexed er by time—division or ncy—division multiplexing, etc. In some cases, each user data stream is for transmission to one or more of a plurality of user als within the same forward user beam.

As is sed in greater detail below, each forward uplink signal is transmitted in a time—synchronized manner by its respective transmitting AN 515. The forward uplink signals 521 transmitted from the ANs 515 are received by a plurality of transponders 410 on the end—to— end relay 503 via receive antenna elements 406 on the end—to-end relay 503. The superposition 550 of the forward uplink s 521 received from geographically distributed locations creates a composite input d signal 545. Each transponder 410 rently receives a composite 2016/026815 input forward signal 545. However, each transponder 410 will receive the signals with slightly different timing due to the differences in the location of the receive antenna element 406 associated with each transponder 401.

Cr is the L x M forward uplink radiation matrix. The values of the forward uplink radiation matrix model the signal path (amplitude and phase) from the ANs 515 to the receive antenna elements 406. E is the L x L payload matrix and provides the model of the transponder signal paths from the receive antenna elements 406 to the it antenna elements 409. The direct path gain from each e antenna element 406 h a corresponding one of a ity of transponders to each corresponding transmit antenna t 409 is d by the diagonal values of the payload matrix. As noted above with respect to the return link, assuming there is no cross—talk between antenna elements, the off—diagonal elements of the payload matrix are zero. In some cases, the crosstalk may not be zero. Isolating the signal paths from each other will minimize crosstalk. In this example, each of the transponders 410 couples a respective one of the receive a elements 406 with a respective one of the transmit antenna ts 409.

Accordingly, a forward downlink signal 522 output from each of the transponders 410 is transmitted by each of the plurality of transponders 410 (see via the transmit antenna elements 409, such that the forward downlink signals 522 form a forward user beam (by uctively and destructively superposing in desired geographic recovery locations to form the beam). In some cases, a ity of user beams is formed, each corresponding to a geographic user beam coverage area 519 that services a respective set of user terminals 517 within the user beam coverage area 519. The path from the first transmit antenna t 409a (see ) to a reference (or recovery) location in the first user beam coverage area 519 is given in the At11 value of the forward downlink radiation matrix. As noted with regard to the return link, this end—to—end rming system intentionally induces a multipath environment for the end—to—end path from any AN 515 to any user terminal 517. In some cases, a subset of the transmit antenna elements 409 transmits forward downlink signals 522 with significant energy to a user terminal 517. The user terminal 517 (or, more generally, a reference or recovery location in the user beam coverage area 519 for receiving and/or ry) receives the plurality of forward downlink signals 522 and recovers at least a portion of the d data signal from the received plurality of forward downlink signals 522. The transmitted forward downlink signals 522 may be received by the user terminal 517 at a signal level that is within 10 dB of a maximum signal level from any of the other signals transmitted by the transmit antenna elements 409 within the subset. In some cases, the subset of transmit antenna elements includes at least 10% of the plurality of transmit antenna elements t in the end—to—end relay 503. In 2016/026815 some cases, the subset of transmit antenna elements include at least 10 transmit antenna elements, less of how many transmit antenna elements 409 are t in the end—to—end relay 503. In one case, receiving the plurality of forward downlink signals ses receiving a superposition 551 of the plurality of forward downlink signals. is a simplified illustration of a model of all the end—to—end forward multipath channels 2208 from the M ANs 515 to the K user beam coverage areas 519. As shown in , there is an end—to—end forward multipath channel 2208 that couples each AN 515 to each user beam coverage area 519. Each channel 2208 from one AN 515 to one user beam ge area 519 has multipath induced as a result of L unique paths from the AN 515 through the plurality of transponders to the user beam ge area 519. As such, the K X M multipath channels 2208 may be individually modeled and the model of each serves as an element of a K x M forward channel matrix wad. The forward channel matrix Hde has M vectors, each having dimensionality equal to K, such that each vector models the end—to—end forward gains for multipath ications between a respective one of the M ANs 515 and reference (or recovery) locations in K forward user beam coverage areas. Each end-to-end d multipath channel couples one of the M ANs 515 with user terminals 517 serviced by one of K forward user beams via L transponders 410 (see ). In some cases, only a subset of the L transponders 410 on the end-to-end relay 503 are used to create the end-to-end forward multipath channel (i.e., are in the signal path of the end-to-end forward multipath channel). In some cases, the number of user beams K is greater than the number of transponders L that are in the signal path of the end—to—end d multipath channel. Furthermore, in some cases, the number of ANs M is greater than the number of transponders L that are in the signal path of the end-to-end forward multipath channel.

Hde may represent the end—to—end forward link as the t of matrices At X E x Cr. Each element in wad is the end—to—end d gain due to the multipath nature of the path and can be subject to a deep fade. An appropriate beam weight may be computed for each of the plurality of end—to—end forward multipath channels 2208 by the CPS 505 within the ground t 502 to form forward user beams from the set ofM ANs 515 to each user beam coverage area 519. The plurality of ANs 515 provide transmit path diversity, by using multiple transmitters (ANs) to a single receiver (user terminal), to enable the successful transmission of information to any user terminal 517 in the presence of the intentionally induced multipath channel.

Combined Forward and Return Data illustrates an example end—to—end relay supporting both forward and return ications. In some cases, the same end—to—end relay signal paths (e.g., set of receive a ts, transponders, and transmit antenna elements) may be used for both the end—to— end forward link 501 and the end—to—end return link 523. Some other cases include forward link transponders and return link transponders, which may or may not share receive and it antenna elements. In some cases, the system 1200 has a plurality of ANs and user terminals that are located in the same general geographic region 1208 (which may be, for e, a particular state, an entire y, a region, an entire visible area, or any other le phic region 1208). A single end—to—end relay1202 (disposed on a satellite or any other le end—to—end relay) receives forward uplink s 521 from ANs and transmits forward downlink signals 522 to user terminals. At ate times, or on alternate frequencies, the end—to—end relay1202 also receives return uplink signals 525 from the user terminals and transmits return downlink signals 527 to the ANs. In some cases, the end—to—end relay 1202 is shared between forward and return data using techniques such as time domain duplexing, frequency domain duplexing, and the like. In some cases, time domain duplexing between forward and return data uses the same frequency range: forward data is transmitted during ent (non-overlapping) time als than those used for transmitting return data. In some cases, with frequency domain duplexing, different frequencies are used for d data and return data, thereby permitting concurrent, non-interfering transmission of forward and return data. is an illustration of an uplink frequency range being divided into two portions.

The lower—frequency (left) portion of the range is allocated to the forward uplink and the upper— frequency (right) portion of the range is allocated to the return uplink. The uplink range may be divided into multiple portions of either forward or return data. is an illustration of the forward data and return data being time division multiplexed. A data frame period is shown in which forward data is transported during the first time interval of the frame, while return data is transported during the last time interval of the frame. The end—to—end relay receives from one or more access nodes during a first (forward) receive time interval and from one or more user terminals during a second (return) receive time interval that doesn’t overlap the first receive time interval. The end—to—end relay transmits to one or more user terminals during a first (forward) transmit time interval and to one or more access nodes during a second n) transmit time interval that doesn’t overlap the first receive time interval. The data frame may be repeated or may change dynamically. The frame may be divided into multiple (e.g., non—contiguous) portions for forward and return data.

-End Beamforming Satellites In some cases, the end—to—end relay 503 is ented on a satellite, so that the satellite is used to relay the signals from the ANs (which can be referred to as satellite access nodes (SANs) in such cases) to the user terminals and vice versa. In some cases, the satellite is in geostationary orbit. An example satellite operating as an end—to—end relay has an array of receive antenna elements, an array of transmit antenna elements, and a number of transponders that connect the e antenna elements to the transmit antenna elements. The arrays have a large number of antenna elements with overlapping antenna t coverage areas, similar to traditional single link phased array antennas. It is the overlapping antenna t coverage areas on both the transmit antenna elements and receive antenna elements that create the multipath environment previously described. In some cases, the antenna patterns established by the corresponding antenna elements, and those that result in the overlapping antenna element coverage areas (e.g., pping ent beam antenna patterns), are identical. For the es of this disclosure, the term “identical” means that they follow essentially the same distribution of power over a given set of points in space, taking the antenna element as the point of reference for locating the points in space. It is very difficult to be perfectly identical.

Therefore, patterns that have relatively small ions from one n to another are within the scope of “identical” patterns. In other cases, receive component beam antenna patterns may not be identical, and in fact may be significantly different. Such antenna patterns may yet result in overlapping antenna element coverage areas, however, those resulting coverage areas will not be identical.

Antenna types include, but are not limited to, array fed reflectors, al arrays, direct radiating arrays and other forms of antenna arrays. Each antenna can be a system including additional optical components to aid in the receipt and/or transmission of signals, such as one or more reflectors. In some cases, a satellite includes components that assist in system timing alignment and beamforming calibration. is a diagram of an example satellite 1502 that can be used as an end—to—end relay 503. In some cases, the satellite 1502 has an array fed reflector transmit antenna 401 and an array fed reflector receive antenna 402. The receive a 402 comprises a receive reflector (not shown) and an array of receive antenna elements 406. The e antenna elements 406 are illuminated by the receive reflector. The transmit antenna 401 comprises a it reflector (not shown) and an array of transmit a elements 409. The transmit antenna elements 409 are arranged to illuminate the transmit reflector. In some cases, the same reflector is used for both e and transmit. In some cases, one port of the antenna element is used for receiving and r port for transmission. Some as have the ability to distinguish between s of different polarizations. For example, an antenna element can include four waveguide ports for right—hand circular polarization (RHCP) receive, left—hand circular polarization (LHCP) receive, RHCP transmit, and LHCP transmit, respectively. In some cases, dual polarizations may be used to increase capacity of the system; in other cases, single polarization may be used to reduce interference (e.g., with other systems using a different polarization).

The example satellite 1502 also comprises a plurality of transponders 410. A transponder 410 connects the output from one receive antenna element 406 to the input of a transmit antenna element 409. In some cases, the transponder 410 amplifies the received signal.

Each receive antenna t outputs a unique received signal. In some cases, a subset of receive antenna elements 406 receive a signal from an Earth transmitter, such as either a user terminal 517 in the case of a return link signal or an AN 515 in the case of a forward link .

In some of these cases, the gain of each receive antenna element in the subset for the received signal is within a relatively small range. In some cases, the range is 3 dB. In other cases, the range is 6 dB. In yet other cases, the range is 10 dB. Accordingly, the satellite will e a signal at each of a plurality of e antenna elements 406 of the satellite, the communication signal originating from an Earth transmitter, such that a subset of the receive antenna elements 406 receives the communication signal at a signal level that is not substantially less than a signal level corresponding to a peak gain of the receive antenna element 406.

] In some cases, at least 10 transponders 410 are provided within the satellite 1502. In another case, at least 100 transponders 410 are provided in the satellite 1502. In yet another case, the number of transponders per polarity may be in the range of 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024 or numbers in—between or greater. In some cases, the transponder 410 includes a low noise amplifier (LNA) 412, a frequency converter and associated filters 414 and a power amplifier (PA) 420. In some cases in which the uplink frequency and downlink frequency are the same, the transponder does not include a frequency converter. In other cases, the plurality of receive antenna elements operate at a first frequency. Each receive a element 406 is associated with one transponder 410. The receive antenna element 406 is coupled to the input of the LNA 412. Accordingly, the LNA independently amplifies the unique ed signal provided by the e antenna element associated with the transponder 410. In some cases, the output of the LNA 412 is coupled to the frequency converter 414. The ncy converter 414 converts the amplified signal to a second frequency.

] The output of the transponder is coupled to an associated one of the transmit antenna elements. In these es, there is a one to one relationship n a transponder 410, an associated receive a element 406, and an associated transmit a element 409, such that the output of each receive antenna element 406 is connected to the input of one and only one transponder and the output of that transponder is connected to the input of one and only one transmit antenna t. is an illustration of an example transponder 410. The transponder 410 can be an example of a transponder of an —end relay 503, as described above (e.g., the satellite 1502 of ). In this example, the transponder includes a phase shifter 418 in addition to the low noise ier (LNA) 412, frequency converter and associated s 414, and power amplifier (PA) of transponder 410. As illustrated in , the example transponder 410 can also be coupled with a phase shift controller 427. For e, the phase shift controller 427 can be coupled (directly or indirectly) with each of some or all of the transponders of an end—to—end relay 503, so that the phase shift controller 427 can individually set the phases for each transponder. The phase shifters may be helpful for calibration, for example, as discussed below.

Antennas To create the multipath environment, antenna element coverage areas can overlap with antenna t coverage areas of at least one other antenna element of the same polarity, frequency, and type mit or e, respectively). In some cases, a plurality of receive component beam antenna patterns, operable at the same receive polarization and e frequency (e.g., having at least a portion of the receive frequency in common), overlap with one another. For example, in some cases, at least 25% of the receive component beam antenna ns, operable at the same receive polarization and receive frequency (e.g., having at least a portion of the receive frequency in common), overlap with at least five other receive ent beam a patterns of the e antenna elements. Similarly, in some cases, at least 25% of the transmit component beam antenna patterns, operable at the same transmit polarization and it frequency (e.g., having at least a portion of the transmit frequency in common), p with at least five other transmit component beam antenna patterns. The amount of overlap will vary from system to system. In some cases, at least one of the receive antenna elements 406 has component beam antenna patterns that overlap with the antenna patterns of other receive antenna elements 406 operable at the same receive frequency (e.g., having at least a portion of the receive frequency in common) and same receive polarization. Therefore, at least some of the plurality of receive antenna elements are capable of receiving the same s from the same source. Similarly, at least one of the transmit antenna elements 409 has a component beam antenna pattern that overlaps with the antenna patterns of other transmit antenna elements 409 operable at the same transmit frequency (e.g., having at least a portion of the transmit frequency in common) and it zation. Therefore, at least some of the plurality of transmit antenna elements are capable of transmitting signals having the same frequency at the same polarization to the same receiver. In some cases, overlapping component beam antenna patterns may have gains that differ by less than 3 dB (or any other suitable value) over a common geographic area. The antenna elements, Whether receive or transmit, may have a broad component beam antenna pattern, and thus a relatively broad antenna element coverage area. In some cases, signals transmitted by an Earth transmitter, such as a user terminal 517 or access node 515, are received by all of the receive antenna elements 406 of the —end relay (e.g., satellite). In some cases, a subset of the elements 406 receives the signals from an Earth transmitter. In some cases, the subset includes at least 50% of the receive antenna elements. In other cases, the subset includes at least 75% of the receive antenna elements. In still other cases, the subset includes at least 90% (e.g., up to and including all) of the receive antenna elements.

Different subsets of the e antenna elements 406 may receive signals from different Earth transmitters. Similarly, in some cases, a subset of the elements 409 transmits s that may be ed by a user terminal 517. In some cases, the subset includes at least 50% of the transmit a elements. In other cases, the subset includes at least 75% of the it antenna elements. In still other cases, the subset includes at least 90% (e.g., up to and including all) of the transmit antenna elements. Different subsets of the elements 409 may transmit signals that are ed by different user terminals. Furthermore, user terminals may be within several formed user beam coverage areas 519. For the purpose of this disclosure, an antenna pattern is a n of distribution of energy itted to, or received from, an antenna. In some cases, the energy may be directly radiated from/to the antenna element. In other cases, the energy from one or more transmit a elements may be reflected by one or more reflectors that shape the antenna element pattern. Similarly, a receive element may receive energy directly, or after the energy has ed off one or more reflectors. In some cases, as can be made up of several elements, each having a component beam a pattern that establishes a corresponding antenna element coverage area. Similarly, all or a subset of receive and it antenna elements that receive and transmit signals to ANs 515 may overlap, such that a plurality of receive antenna elements receives signals from the same AN 515 and/or a plurality of it antenna elements transmits signals to the same AN 515. ] is an illustration of component beam antenna patterns produced by several antenna elements (either receive antenna elements 406, or transmit antenna elements 409) that intersect at the 3 dB points. The component beam antenna pattern 1301 of a first antenna element has peak component beam antenna gain along the boresight 1303. The component beam antenna pattern 1301 is shown to attenuate about 3 dB before it intersects with the component beam antenna pattern 1305. Since each pair of two adjacent component beam antenna patterns p about the 3 dB line 1307 for only a relatively small portion of the component beam antenna pattern, the antenna elements that produce these component beam antenna patterns are considered not to be overlapping. ] shows idealized 3 dB antenna contours 3901, 3902, 3903 of several elements 406, 409 with the peak gain designated with the letter ‘x’. The contours 3901, 3902, 3903 are referred to herein as “idealized” because the contours are shown as circular for the sake of simplicity. r, the contours 3901, 3902, 3903 need not be circular. Each contour indicates the place at which the transmitted or received signal is 3 dB below the peak level. e the contour, the signal is more than 3 dB below the peak. Inside the contour, the signal is less than 3 dB below the peak (i.e., within 3 dB of the peak). In a system in which the ge area of a receive component beam antenna pattern is all points for which the receive ent beam antenna gain is within 3 dB of peak receive component beam antenna gain, the area inside the contour is referred to as the antenna element coverage area. The 3 dB antenna contour for each element 406, 409 is not pping. That is, only a relatively small portion of the area inside the 3 dB antenna contour 3901 overlaps with the area that is inside the adjacent 3 dB antenna patterns 3902, 3903. is an illustration of the antenna ns 1411, 1413, 1415 of several antenna elements (either receive antenna elements 406 or transmit antenna elements 409). In contrast to the component beam antenna patterns of , the component beam antenna patterns shown in ect 1417 above the 3 dB line 1307.

A h E illustrate 3 dB antenna contours for several antenna elements 406, 409 with the beam center point (peak gain) designated with the letter ‘x’. A shows the particular antenna contour 1411 of a first a element 406. B shows the 3 dB antenna contours 1411, 1413 for two particular ts 406. C shows the 3 dB antenna contours for three elements 406. D shows the 3 dB antenna contours for four antenna elements 406. E shows the 3 dB a contours for an array of 16 antenna elements 406. The 3 dB antenna rs are shown to overlap 1418 (e.g., 16 such 3 dB antenna contours are shown). The antenna elements in either the receive or transmit antenna may be arranged in any of several different configurations. For example, if ts have a generally circular feed horn, the elements may be arranged in a honeycomb configuration to tightly pack the elements in a small amount of space. In some cases, the antenna ts are aligned in horizontal rows and vertical s. is an example illustration of relative positions of receive antenna 3 dB antenna contours associated with receive antenna elements 406. The element 406 beam centers are numbered 1 — 16, with element 4064 identified by the number ‘4’ to the upper left of the beam center indicator ‘X’. In some cases, there may be many more than 16 receive antenna elements 406. However, for the sake of simplicity, only 16 are shown in . A corresponding array of transmit antenna elements 409 and their associated 3 dB antenna contours will look similar to . Therefore, for the sake of simplicity, only the array of receive antenna elements 406 are shown. The area 2101 in the center is where all of the antenna element coverage areas overlap.

In some cases, at least one point within the relay coverage area (e.g., ite coverage area) falls within the 3 dB antenna contour of the component beams of several antenna elements 406. In one such case, at least one point is within the 3 dB antenna contour of at least 100 different antenna elements 406. In another case, at least 10% of the relay coverage area lies within the 3 dB a contours of at least 30 ent a elements. In another case, at least 20% of the relay ge area lies within the 3 dB antenna contours of at least 20 different a elements. In another case, at least 30% of the relay coverage area lies within the 3 dB a contours of at least 10 different a elements. In another case, at least 40% of the relay coverage area lies within the 3 dB antenna contours of at least eight different antenna elements. In another case, at least 50% of the relay coverage area lies within the 3 dB antenna contours of at least four different antenna elements. However, in some cases, more than one of these relationships may be true.

] In some cases, the end-to-end relay has a relay coverage area (e.g., satellite coverage area) in which at least 25% of the points in the uplink relay coverage area are within (e.g., span) overlapping coverage areas of at least six receive antenna elements 406. In some cases, 25% of the points within the uplink relay coverage area are within (e.g., span) overlapping ge areas of at least four receive antenna elements 406. In some cases, the end—to—end relay has a coverage area in which at least 25% of the points in the downlink relay coverage area are within (e.g., span) overlapping coverage areas of at least six transmit a elements 409. In some cases, 25% of the points within the downlink relay coverage area are within (e.g., span) overlapping ge areas of at least four transmit antenna elements 409.

In some cases, the e antenna 402 may be pointed roughly at the same coverage area as the transmit antenna 401, so that some receive antenna element coverage areas may lly correspond to particular transmit antenna element coverage areas. In these cases, the 2016/026815 receive antenna elements 406 may be mapped to their corresponding transmit antenna ts 409 via the onders 410, yielding similar transmit and receive antenna element coverage areas for each receive/transmit signal path. In some cases, however, it may be advantageous to map receive antenna elements 406 to transmit antenna elements 409 that do not correspond to the same component beam coverage area. Accordingly, the mapping of the ts 406 of the receive a 402 to the elements 409 of the it antenna 401 may be randomly (or otherwise) permuted. Such permutation includes the case that results in the receive antenna elements 406 not being mapped to the transmit antenna elements 409 in the same relative location within the array or that have the same coverage area. For example, each receive antenna element 406 within the receive antenna element array may be associated with the same onder 410 as the transmit a element 409 d in the mirror location of the transmit antenna element array. Any other permutation can be used to map the receive antenna elements 406 to the transmit a elements 409 ing to a permutation (e.g., pair each receive antenna element 406 with the same transponder to which an associated transmit antenna element 409 is coupled in accordance with a particular permutation of the receive antenna element 406 and the transmit antenna element 409). is a table 4200 showing example mappings of receive antenna elements 406 to transmit antenna elements 409 through 16 transponders 410. Each transponder 410 has an input that is ively coupled to an associated receive antenna element 406 and an output that is exclusively coupled to an associated transmit antenna element 409 (e.g., there is a one to one relationship between each receive antenna element 406, one transponder 410 and one transmit antenna element 409). In some cases, other receive antenna elements, transponders and transmit antenna elements may be present on the end—to—end relay (e.g., ite) that are not configured in a one to one relationship (and do not operate as a part of the end-to—end beamforming system).

The first column 4202 of the table 4200 identifies a transponder 410. The second column 4204 identifies a receive antenna t 406 to which the transponder 410 of the first column is coupled. The third column 4206 of the table 4200 identifies an associated transmit antenna element 409 to which the output of the transponder 410 is coupled. Each receive a element 406 is coupled to the input of the transponder 410 identified in the same row of the table 4200. Similarly, each transmit antenna t 409 is coupled to the output of the transponder 410 identified in the same row of the table 4200. The third column of the table 4200 shows an example of direct g in which each receive antenna element 406 of the receive a array is coupled to the same transponder 410 as a transmit antenna element 409 in the same relative on within the transmit antenna array. The fourth column 4208 of table 4200 shows an example of interleaved mapping in which the first receive antenna element 406 is coupled to the first transponder 410 and to the tenth transmit antenna element 409. The second receive antenna element 406 is coupled to the second transponder 410 and to the ninth transmit antenna element 409, and so on. Some cases have other permutations, including a random mapping in which the particular pairing of the receive antenna element 406 and the transmit element 409 with a transponder 410 are randomly selected.

The direct mapping, which attempts to keep the transmit and receive antenna element coverage areas as similar as possible for each e/transmit signal path, generally yields the highest total capacity of the system. Random and interleaved ations lly produce slightly less capacity but provide a more robust system in the face of AN outages, fiber outages in the terrestrial network, or loss of receive/transmit signal paths due to electronic failure on the end—to—end relay (e.g., in one or more transponders). Random and interleaved permutations allow lower cost non-redundant ANs to be used. Random and interleaved permutations also provide less variation between the capacity in the best performing beam and the ty in the worst performing beam. Random and interleaved permutations may also be more useful to lly operate the system with just a fraction of the ANs ing in only a fraction of the total capacity being available but no loss in coverage area. An example of this is an incremental rollout of ANs, where the system was lly operated with only 50% of the ANs deployed.

This may provide less than the full capacity, while still ng operation over the entire coverage area. As the demand increases, more ANs can be deployed to increase the ty until the full capacity is achieved with all the ANs active. In some cases, a change in the composition of the ANs results in a re-calculation of the beam weights. A change in composition may include changing the number or characteristics of one or more ANs. This may require a re— estimation of the end—to—end forward and/or return gains.

] In some cases, the a is an array—fed reflector antenna with a paraboloid reflector.

In other cases, the reflector does not have a loid shape. An array of receive antenna elements 406 may be arranged to receive signals reflected by the or. Similarly, an array of transmit antenna elements 409 may be arranged to form an array for nating the reflector.

One way to provide elements with overlapping component beam antenna patterns is to have the elements 406, 409 defocused (unfocused) as a consequence of the focal plane of the reflector being behind (or in front of) the array of elements 406, 409 (i.e., the receive antenna array being located outside the focal plane of the receive reflector). is an ration of a cross—section of a center—fed paraboloid reflector 1521.

A focal point 1523 lies on a focal plane 1525 that is normal to the central axis 1527 of the reflector 1521. Received signals that strike the reflector 1521 el to the central axis 1527 are focused onto the focal point 1523. Likewise, signals that are transmitted from an antenna element located at the focal point and that strike the reflector 1521 will be reflected in a focused beam from the reflector 1521 parallel to the central axis 1527. Such an arrangement is often used in Single Feed per Beam systems to maximize the ivity of each beam and minimize overlap with beams formed by adjacent feeds. is an ration of another paraboloid reflector 1621. By locating antenna elements 1629 (either receive antenna elements or it antenna elements 406, 409, 3416, 3419, 3426, 3429,) outside the focal plane (e.g., in front of the focal plane 1625 of the reflector 1621), the path of transmitted signals 1631 that strike the reflector 1621 will not be parallel to one another as they reflect off the or 1621, resulting in a wider beam width than in the focused case. In some cases, ors that have shapes other than paraboloids are used. Such reflectors may also result in defocusing the antenna. The end—to—end beamforming system may use this type of defocused antenna to create overlap in the ge area of adjacent antenna elements and thus provide a large number of useful receive/transmit paths for given beam locations in the relay coverage area.

In one case, a relay coverage area is established, in which 25% of the points within the relay coverage area are within the a element coverage areas of at least six component beam antenna patterns when the end-to-end relay is deployed (e.g., an end—to—end satellite relay is in a e orbit). Alternatively, 25% of the points within the relay coverage area are within the antenna element coverage areas of at least four receive antenna ts. is an illustration of an example relay coverage area (for an end—to—end satellite relay, also referred to as satellite coverage area) 3201 (shown with single cross—hatching) and the area 3203 (shown with double hatching) defined by the points within the relay coverage area 3201 that are also contained within six antenna element coverage areas 3205, 3207, 3209, 3211, 3213, 3215.

The coverage area 3201 and the antenna element coverage areas 3205 , 3207, 3209, 3211, 3213, 3215 may be either receive antenna element coverage areas or transmit antenna element coverage areas and may be associated with only the forward link or only the return link. The size of the antenna element coverage areas 3205 3215 is determined by the , 3207, 3209, 3211, 3213, desired performance to be provided by the system. A system that is more nt of errors may have antenna element coverage areas that are larger than a system that is less tolerant. In some cases, each antenna t coverage area 3205, 3207, 3209, 3211, 3213, 3215 is all points for which the component beam antenna gain is within 10 dB of the peak component beam antenna gain for the antenna element establishing the component beam antenna pattern. In other cases, each antenna element coverage area 3205, 3207, 3209, 3211, 3213, 3215 is all points for which the component beam antenna gain is within 6 dB of peak component beam antenna gain. In still other cases, each antenna element coverage area 3205, 3207, 3209, 3211, 3213, 3215 is all points for which the ent beam antenna gain is within 3 dB of peak component beam a gain. Even when an end—to—end relay has not yet been deployed (e.g., an end—to—end satellite relay is not in a service orbit, the end—to—end relay still has component beam antenna ns that conform to the above definition. That is, antenna element coverage areas corresponding to an end—to—end relay in orbit can be calculated from the component beam antenna patterns even when the end—to—end relay is not in a service orbit. The —end relay may include additional antenna elements that do not contribute to beamforming and thus may not have the recited characteristics. is an illustration of an end—to—end relay (e.g., satellite) antenna pattern 3300 in which all of the points within a relay coverage area 3301 (e.g. satellite coverage area) are also contained within at least four antenna element coverage areas 3303, 3305, 3307, 3309. Other antenna elements may exist on the end-to-end relay and can have antenna element coverage areas 3311 that contain less than all of the points within the relay coverage area 3301.

The system may operate in any suitable spectrum. For e, an end-to-end beamforming system may e in the C, L, S, X, V, Ka, Ku, or other suitable band or bands.

In some such systems, the receive means operates in the C, L, S, X, V, Ka, Ku, or other suitable band or bands. In some cases, the forward uplink and the return uplink may operate in the same frequency range (e.g., in ty of 30 GHZ); and the return downlink and the forward downlink may operate in a erlapping frequency range (e.g., in the vicinity of 20 GHz). The end—to— end system may use any suitable dth (e.g., 500 MHz, 1 GHZ, 2 GHZ, 3.5 GHz, etc.). In some cases, the forward and return links use the same transponders.

To assist in system timing alignment, path lengths among the L transponders are set to match signal path time delays in some cases, for example through appropriate cable length selection. The end—to—end relay (e.g., satellite) in some cases has a relay beacon generator 426 (e.g. satellite ) within a calibration support module 424 (see ). The beacon generator 426 generates a relay beacon signal. The end—to—end relay broadcasts the relay beacon signal to further aid in system timing alignment as well as support feeder link ation. In some cases, the relay beacon signal is a pseudo—random (known as PN) sequence, such as a PN direct sequence spread spectrum signal that runs at a high chip rate (e.g., 100, 200, 400, or 800 n chips per second (Mcps), or any other suitable value). In some cases, a linearly polarized relay (e. g., satellite) beacon, receivable by both RHCP and LHCP antennas, is broadcast over a wide coverage area by an antenna, such as an antenna horn (not shown) or coupled into one or more of the transponders 410 for transmission through the associated transmit antenna element 409. In an e system, beams are formed in multiple 500 MHZ bandwidth channels over the Ka band, and a 400 Mcps PN code is filtered or pulse—shaped to fit within a 500 MHZ bandwidth channel. When multiple channels are used, the same PN code may be transmitted in each of the channels. The system may employ one beacon for each channel, or one beacon for two or more channels.

Since there may be a large number of receive/transmit signal paths in an end—to—end relay, redundancy of individual receive/transmit signal paths may not be required. Upon failure of a receive/transmit signal path, the system may still perform very close to its previous performance level, although modification of beamforming coefficients may be used to t for the loss.

Ground Networks The ground network of an example end-to-end beamforming system contains a number of phically distributed Access Node (AN) Earth ns pointed at a common end-to-end relay. Looking first at the d link, a l Processing System (CPS) computes beam s for transmission of user data and interfaces to the ANs through a distribution network. The CPS also interfaces to the sources of data being provided to the user terminals.

The distribution k may be implemented in various ways, for e using a fiber optic cable infrastructure. Timing n the CPS and SANs may be deterministic (e.g., using circuit—switched channels) or non—deterministic (e. g., using a packet-switched network). In some cases, the CPS is implemented at a single site, for example using custom application specific integrated circuits (ASICs) to handle signal processing. In some cases, the CPS is implemented in a distributed manner, for example using cloud computing techniques.

Returning to the example of the CPS 505 may include a plurality of feeder link modems 507. For the forward link, the feeder link modems 507 each receive forward user data streams 509 from various data sources, such as the internet, a video headend (not shown), etc.

The received forward user data streams 509 are modulated by the modems 507 into K forward beam signals 511. In some cases, K may be in the range of l, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024 or numbers ween or greater. Each of the K d beam signals carries forward user data s to be transmitted on one of K forward user beams. Accordingly, if K = 400, then there are 400 forward beam s 511, each to be transmitted over an associated one of 400 forward user beams to a d user beam coverage area 519. The K forward beam signals 511 are coupled to a forward beamformer.

If M ANs 515 are present in the ground segment 502, then the output of the forward beamformer is M access node—specific forward signals 516, each comprising weighted forward beam signals corresponding to some or all of the K forward beam signals 511. The d beamformer may generate the M access node—specific forward signals 516 based on a matrix product of the K x M forward beam weight matrix with the K forward data signals. A distribution network 518 distributes each of the M access node—specific forward signals to a corresponding one of the M ANS 515. Each AN 515 its a forward uplink signal 521 comprising a respective access node—specific d signal 516. Each AN 515 transmits its respective forward uplink signal 521 for relay to one or more (e.g., up to and ing all) of the forward user beam coverage areas via one or more (e.g., up to and including all) of the forward receive/transmit signal paths of the end—to—end relay. Transponders 410, 411 within the end—to-end relay 503 receive a composite input forward signal comprising a superposition 550 of forward uplink signals 521 itted by a plurality (e.g., up to and including all) of the ANs 515. Each transponder (e.g., each receive/transmit signal path through the relay) relays the ite input d signal as a respective forward downlink signal to the user terminals 517 over the forward downlink. is an illustration of an example distribution of ANs 515. Each of the smaller numbered circles represents the on of an AN 515. Each of the larger circles indicates a user beam coverage area 519. In some cases, the ANs 515 are spaced approximately evenly over the coverage area of the end—to—end relay 503. In other cases, the ANs 515 may be distributed unevenly over the entire coverage area. In yet other cases, the ANs 515 may be distributed evenly or unevenly over one or more sub—regions of the relay coverage area. lly, system performance is best when the ANs 515 are uniformly distributed over the entire coverage area.

However, erations may dictate compromises in the AN placement. For example, an AN 515 may be placed based on the amount of interference, rain, or other environmental conditions, cost of real estate, access to the distiibution network, etc. For e, for a satellite—based end— to—end relay system that is sensitive to rain, more of the ANs 515 may be placed in areas that are less likely to experience rain—induced fading (e.g., the western United States). As r example, ANs 515 may be placed more densely in high rain regions (e.g., the southeastern United States) to provide some diversity gain to counteract the effects of rain fading. ANs 515 may be located along fiber routes to reduce distribution costs associated with the ANs 515.

The number of ANS 515, M, is a selectable parameter that can be ed based upon l criteria. Fewer ANS can result in a simpler, lower cost ground segment, and lower operational costs for the bution network. More ANs can result in larger system ty. shows a simulation of the normalized forward and return link ty as a function of the number of ANs ed in an example system. Normalized capacity is the capacity with M ANS divided by the capacity obtained with the largest number of ANS in the simulation. The capacity increases as the number of ANS increases, but it does not increase without bound. Both forward link and return link capacities approach an asymptotic limit as the number of ANS is sed. This simulation was performed with L = 517 transmit and receive antenna elements and with the ANs distributed uniformly over the coverage area, but this asymptotic behavior of the ty can be seen with other values for L and other AN l distributions. Curves like those shown in can be helpful in selection of the number of ANS, M, to be deployed and in understanding how the system ty can be phased in as ANS are incrementally ed, as discussed previously. is a block diagram of an example ground segment 502 for an end—to—end beamforming system. may illustrate, for example, ground segment 502 of The ground segment 502 comprises CPS 505, distribution network 518, and ANS 515. CPS 505 ses beam signal interface 524, forward/return beamformer 513, distribution interface 536, and beam weight generator 910.

For the forward link, beam signal interface 524 obtains forward beam signals (FBS) 511 associated with each of the forward user beams. Beam signal interface 524 may include forward beam data multiplexer 526 and forward beam data stream modulator 528. Forward beam data multiplexer 526 may receive forward user data streams 509 comprising forward data for transmission to user terminals 517. d user data s 509 may comprise, for example, data packets (e.g., TCP packets, UDP packets, etc.) for transmission to the user terminals 5 17 via the end—to—end beamforming system 500 of Forward beam data multiplexer 526 groups (e.g., lexes) the forward user data streams 509 according to their respective forward user beam coverage areas to obtain forward beam data streams 532. Forward beam data multiplexer 526 may use, for example, time—domain multiplexing, frequency—domain multiplexing, or a combination of multiplexing techniques to generate forward beam data streams 532. Forward beam data stream modulator 528 may modulate the forward beam data streams 532 according to one or more modulation schemes (e.g., mapping data bits to modulation symbols) to create the forward beam signals 511, which are passed to the forward/return beamformer 513. In some cases, the modulator 528 may frequency multiplex 2016/026815 multiple modulated signals to create a carrier beam signal 511. Beam signal interface 524 may, for example, implement the onality of feeder link modems 507 discussed with reference to d/retum beamformer 513 may include forward beamformer 529 and return beamformer 531. Beam weight generator 910 generates an M x K forward beam weight matrix 918. Techniques for generating the M x K forward beam weight matrix 918 are discussed in more detail below. Forward beamformer 529 may include a matrix multiplier that calculates M access—node ic forward signals 516. For example, this ation can be based on a matrix product of the M x K forward beam weight matrix 918 and a vector of the K forward beam signals 511. In some examples, each of the K forward beam signals 511 may be associated with one of F d frequency sub—bands. In this case, the forward beamformer 529 may generate samples for the M access—node specific forward signals 516 for each of the F forward frequency sub—bands (e.g., ively implementing the matrix product operation for each of the F sub— bands for respective subsets of the K forward beam s 511. Distribution interface 536 distributes (e.g., via distribution network 518) the M access node—specific forward signals 516 to the respective ANs 515.

For the return link, the distribution interface 536 obtains composite return signals 907 from ANs 515 (e.g., via distribution network 518). Each return data signal from user terminals 517 may be included in multiple (e.g., up to and including all) of the composite return signals 907. Beam weight generator 910 generates a K x M return beam weight matrix 937. ques for generating the K x M return beam weight matrix 937 are discussed in more detail below. Return beamformer 531 calculates K return beam signals 915 for the K return user beam coverage areas. For example, this calculation can be based on a matrix product of the return beam weight matrix 937 and a vector of the respective composite return signals 907. Beam signal interface 524 may include return beam signal demodulator 552 and return beam data de— lexer 554. Return beam signal demodulator 552 may demodulate each of the return beam signals to obtain K return beam data streams 534 associated with the K return user beam ge areas. Return beam data de—multiplexer 554 may tiplex each of the K return beam data streams 534 into respective return user data streams 535 associated with the return data signals transmitted from user terminals 517. In some examples, each of the return user beams may be ated with one of R return frequency sub—bands. In this case, the return beamformer 531 may generate respective subsets of the return beam signals 915 associated with each of the R return frequency sub—bands (e.g., effectively implementing the matrix product operation for each of the R return frequency sub—bands to generate respective subsets of the return beam signals 915). is a block diagram of an example forward/return beamformer 513. The forward/return beamformer 513 comprises a d beamformer 529, a forward timing module 945, a return beamformer 531, and a timing module 947. The forward timing module 945 associates each of the M access pecific forward signals 516 with a time stamp (e.g., multiplexes the time stamp with the access node—specific forward signal in a multiplexed access node—specific forward signal) that indicates when the signal is desired to arrive at the end—to—end relay. In this way, the data of the K forward beam signals 511 that is split in a splitting module 904 within the forward rmer 529 may be transmitted at the appropriate time by each of the ANs 515. The timing module 947 aligns the receive signals based on time stamps. Samples of the M AN composite return signals (CRS) 907 are associated with time stamps indicating when the particular samples were transmitted from the end—to—end relay. Timing considerations and generation of the time stamps are discussed in greater detail below.

The d beamformer 529 has a data input 925, a beam weights input 920 and an access node output 923. The d beamformer 529 applies the values of an M x K beam weight matrix to each of the K forward data signals 511 to generate M access node specific forward signals 521, each having K ed forward beam signals. The forward beamformer 529 may e a splitting module 904 and M forward weighting and g modules 533.

The splitting module 904 splits (e.g., duplicates) each of the K forward beam signals 511 into M groups 906 of K d beam signals, one group 906 for each of the M forward weighting and summing s 533. Accordingly, each forward weighting and summing module 533 es all K forward data signals 511.

A forward beam weight generator 917 generates an M x K forward beam weight matrix 918. In some cases, the forward beam weight matrix 918 is generated based on a channel matrix in which the elements are estimates of end—to—end forward gains for each of the K x M end-to—end forward multipath ls to form a forward channel matrix, as discussed further below. Estimates of the end—to—end forward gain are made in a channel estimator module 919. In some cases, the channel estimator has a channel data store 921 that stores data related to s parameters of the end—to—end multipath channels, as is discussed in r detail below. The channel estimator 919 outputs an estimated end—to—end gain signal to allow the forward beam weight generator 917 to generate the d beam weight matrix 918. Each of the weighting and summing modules 533 are coupled to receive respective vectors of beamforming weights of the forward beam weight matrix 918 (only one such connection is show in for simplicity). The first weighting and summing module 533 applies a weight equal to the value of the 1,1 element of the M x K forward beam weight matrix 918 to the first of the K forward beam signals 511 (discussed in more detail below). A weight equal to the value of the 1,2 element of the M x K d beam weight matrix 918 is applied to the second of the K forward beam signals 511. The other weights of the matrix are applied in like fashion, on through the Km forward beam signal 511, which is weighted with the value equal to the 1,K t of the M x K forward beam weight matrix 918. Each of the K weighted forward beam signals 903 are then summed and output from the first weighting and g module 533 as an access node— ic forward signal 516. The access node—specific forward signal 516 output by the first ing and summing module 533 is then coupled to the timing module 945. The timing module 945 outputs the access node—specific forward signal 516 to the first AN 515 through a bution network 518 (see . Similarly, each of the other weighting and summing modules 533 receive the K forward beam signals 511, and weight and sum the K forward beam signals 511. The outputs from each of the M weighting and summing modules 533 are coupled through the distribution network 518 to the associated M ANS 515 so that the output from the mm weighting and summing module is coupled to the mm AN 515. In some cases, jitter and uneven delay through the distribution network, as well as some other timing considerations, are handled by the timing module 945 by associating a time stamp with the data. Details of an e timing technique are provided below with regard to FIGs. 36 and 37.

] As a consequence of the beam weights applied by the forward beamformers 529 at the ground segment 502, the signals that are itted from the ANs 515 through the end—to—end relay 503 form user beams. The size and location of the beams that are able to be formed may be a function of the number of ANs 515 that are ed, the number and antenna patterns of relay antenna elements that the signal passes through, the location of the end—to—end relay 503, and/or the geographic spacing of the ANs 515.

Referring now to the —end return link 523 shown in a user terminal 517 within one of the user beam coverage areas 519 transmits signals up to the end—to—end relay 503.

The signals are then relayed down to the ground segment 502. The signals are received by ANs 515.

Referring once again to , M return downlink signals 527 are received by the M ANs 515 and are coupled, as ite return signals 907, from the M ANs 515 through the distribution network 518 and received in an access node input 931 of the return beamformer 531. Timing module 947 aligns the composite return signals from the M ANs 515 to each other and outputs the time—aligned signals to the return beamformer 531. A return beam weight generator 935 generates the return beam weights as a K x M return beam weight matrix 937 based on information stored in a l data store 941 within a channel estimator 943.The return beamformer 531 has a beam weights input 939 through which the return beamformer 531 receives the return beam weight matrix 937. Each of the M AN composite return signals 907 is coupled to an associated one ofM splitter and weighting s 539 within the return beamformer 531. Each splitter and weighting module 539 splits the time—aligned signal into K copies 909. The splitter and weighting modules 539 weight each of the K copies 909 using the k, m t of the K x M return beam weight matrix 937. Further details regarding the K x M return beam weight matrix are provided below. Each set of K ed composite return signals 911 is then coupled to a combining module 913. In some cases, the combining module 913 combines the km weighted composite return signal 911 output from each splitter and weighting module 539. The return beamformer 531 has a return data signal output 933 that outputs K return beam signals 915, each having the samples associated with one of the K return user beams 519 (e.g., the s received through each of the M ANS). Each of the K return beam signals 915 may have samples from one or more user terminals 517. The K combined and aligned, beamformed return beam signals 915 are coupled to the feeder link modems 507 (see .

Note that the return timing adjustment may be performed after the splitting and weighting.

Similarly, for the forward link, the forward timing adjustment may be med before the beamforming.

As discussed above, d beamformer 529 may perform matrix product operations on input samples of K forward beam signals 511 to ate M access node—specific forward signal 516 in real—time. As the beam bandwidth increases (e.g., to support shorter symbol duration) and/or K and M become large, the matrix product operation becomes ationally intensive and may exceed the capabilities of a single computing node (e.g., a single computing server, etc). The operations of return beamformer 531 are similarly computationally intensive. Various approaches may be used to partition computing resources of multiple computing nodes in the forward/retum beamformer 513. In one example, the forward beamformer 529 of may be ioned into separate weighting and summing modules 533 for each of the M ANs 515, which may be distributed into different computing nodes.

Generally, the considerations for implementations include cost, power consumption, scalability relative to K, M, and bandwidth, system availability (e.g., due to node failure, eta), eability, and system y. The example above is per row (or column). Vice versa is possible. Other manners of grouping the matrix operations may be considered (e.g., split into four with [1,1 to K/2,M/2], [. . .], computed individually and summed up).

In some cases, the forward/return beamformer 513 may include a time—domain multiplexing architecture for processing of beam weighting operations by time—slice rmers. is a block diagram of an example forward beamformer 529 comprising multiple forward time—slice beamformers with time—domain de—multiplexing and multiplexing.

The forward beamformer 529 es a forward beam signal de—multiplexer 3002, N forward lice beamformers 3006, and a forward access node signal multiplexer 3010.

] Forward beam signal de—multiplexer 3002 receives forward beam s 511 and de—multiplexes the K forward beam s 511 into forward time slice inputs 3004 for input to the N forward time—slice beamformers 3006. For example, the forward beam signal de— multiplexer 3002 sends a first omain subset of s for the K forward beam s 511 to a first forward time—slice beamformer 3006, which generates samples associated with the M access node—specific forward signals corresponding to the first time—domain subset of samples. The forward time-slice beamformer 3006 s the samples associated with the M access node—specific forward signals for the first time—domain subset of samples Via its forward time slice output 3008 to the forward access node signal multiplexer 3010. The forward time— slice beamformer 3006 may output the samples associated with each of the M access node- specific forward signals with synchronization timing information (e.g., the corresponding time- slice index, etc.) used by the access nodes to cause (e.g., by pre-correcting) the respective access node-specific forward signals to be synchronized when received by the end-to-end relay. The forward access node signal lexer 3010 lexes time-domain subsets of samples for the M access node—specific forward signals received via the N forward time slice outputs 3008 to generate the M access pecific d signals 516. Each of the forward time—slice beamformers 3006 may include a data buffer, a beam matrix buffer, and beam weight processor implementing the matrix product operation. That is, each of the forward time—slice beamformers 3006 may implement computations mathematically equivalent to the splitting module 904 and d weighting and summing modules 533 shown for forward beamformer 529 of during processing of the samples of one time slice—index. Updating of the beam weight matrix may be performed incrementally. For example, the beam weight matrix buffers for forward lice beamformers may be updated during idle time in a rotation of time—slice indices I through the N forward time—slice beamformers 3006. Alternatively, each d time—slice beamformer may have two buffers that can be used in a ping—pong configuration (e.g., one can be updated while the other is being used). In some cases, multiple buffers can be used to store beam weights corresponding to multiple user beam patterns (e.g., multiple user coverage areas).

Beam weight buffers and data buffers for forward time—slice beamformers 3006 may be implemented as any type of memory or storage including dynamic or static random access memory (RAM). Beam weight processing may be implemented in an application specific integrated circuit (ASIC) and/or a field programmable gate array (FPGA), and may include one or more processing cores (e.g., in a cloud computing environment). Additionally or atively, the beam weight buffer, data buffer, and beam weight processor may be integrated within one component. illustrates a simplified example ground segment showing the operation of a forward time—slice beamformer 529. In the example of , d beamformer 529 receives four forward beam signals (e.g., K24), generates access node—specific forward signals for five ANs (e.g., M25), and has three forward time—slice rmers (e.g., N23). The forward beam signals are denoted by FBk2t, where k is the forward beam signal index and t is the time— slice index (e. g., corresponding to a time—domain subset of samples). The forward beam signal tiplexer 3002 receives four time—domain subsets of samples of the forward beam signals associated with four forward user beams and de—multiplexes each forward beam signal so that one forward time slice input 3004 includes, for a particular time—slice index I, the omain subsets of s from each of the forward beam signals 511. For example, time-domain subsets can be a single sample, a contiguous block of s, or a discontiguous (e.g., eaved) block of samples as described below. The forward time-slice beamformers 3006 generate (e.g., based on the forward beam signals 511 and forward beam weight matrix 918) each of the M access-node specific forward signals for the time-slice index I, denoted by Aszt.

For e, the time—domain subsets of s FB1:0, FB220, FB320, and FB4:0 for time— slice index 120 are input to the first forward time—slice beam former TSBF[1] 3006, which generates corresponding samples of access node—specific forward signals AFle, AF2:0, AF320, AF420, and AF5:0 at a forward time slice output 3008. For subsequent time—slice index values 121, 2, the time—domain subsets of samples of forward beam signals 511 are tiplexed by the forward beam signal de—multiplexer 3002 for input to second and third forward time— slice beamformers 3006, which generate access node—specific forward signals associated with the corresponding time—slice indices 5 at forward time slice outputs 3008. also shows that at time—slice index value I23, the first forward time—slice beamformer generates access node— specific forward signals associated with the corresponding time—slice index 3. The matrix product operation performed by each d time—slice beamformer 3006 for one time— slice index value I may take longer than the real time of the omain subset of samples (e.g., the number of samples S multiplied by the sample rate ts). However, each forward time—slice rmer 3006 may only process one omain subset of samples every N time—slice indices t. d access node signal multiplexer 3010 receives forward time slice outputs 3030 from each of the d time—slice beamformers 3006 and multiplexes the time—domain subsets of samples to generate the M access node—specific forward signals 516 for distribution to respective ANs. is a block m of an example return beamformer 531 comprising multiple return time—slice beamformers with time—domain tiplexing and multiplexing.

The return beamformer 531 includes a return ite signal de—multiplexer 3012, N return time—slice beamformers 3016, and a return beam signal multiplexer 3020. Return composite signal de—multiplexer 3012 receives M composite return signals 907 (e.g., from M ANs) and de— multiplexes the M composite return signals 907 into return time slice inputs 3014 for input to the N return time—slice beamformers 3016. Each of the return time—slice beamformers 3016 output the samples ated with the K return beam signals 915 for ponding time—domain subsets of samples via respective return time slice outputs 3018 to the return beam signal multiplexer 3020. The return beam signal multiplexer 3020 multiplexes the time—domain subsets of samples for the K return beam signals received via the N return time slice outputs 3018 to generate the K return beam signals 915. Each of the return time-slice beamformers 3016 may e a data buffer, a beam matrix buffer, and beam weight processor implementing the matrix product operation. That is, each of the return time-slice rmers 3016 may implement computations mathematically equivalent to the splitter and weighting s 539 and combining module 913 shown for return rmer 531 of during processing of the samples of one time slice—index. As sed above with the forward lice beamformers, updating of the beam weight matrix may be performed incrementally using a ping—pong beam weight buffer configuration (e.g., one can be d while the other is being used). In some cases, multiple buffers can be used to store beam weights corresponding to multiple user beam patterns (e. g., multiple user coverage areas). Beam weight buffers and data buffers for return time-slice beamformers 3016 may be implemented as any type of memory or storage including dynamic or static random access memory (RAM). Beam weight processing may be implemented in an application specific integrated circuit (ASIC) and/or a field programmable gate array (FPGA), and may include one or more processing cores. Additionally or alternatively, the beam weight buffer, data buffer, and beam weight processor may be integrated within one component. illustrates a simplified example ground segment showing the operation of a return beamformer 531 employing time—domain multiplexing. In the example of , return beamformer 531 receives five composite return signals (e.g., M25), generates return beam signals for four return user beams (e.g., K25), and has three time—slice beamformers (e.g., N23).

The composite return signals are denoted by Rszt, where m is the AN index and t is the time- slice index (e. g., corresponding to a time—domain subset of samples). The return composite signal de—multiplexer 3012 receives four time—domain subsets of s of the composite return signals from five ANs and de—multiplexes each composite return signal so that one return time slice input 3014 includes, for a particular time—slice index I, the ponding time—domain subsets of samples from each of the composite return signals 907. For example, time—domain subsets can be a single sample, a contiguous block of samples, or a discontiguous (e.g., interleaved) block of samples as described below. The return time—slice beamformers 3016 generate (e.g., based on the composite return signals 907 and return beam weight matrix 937) each of the K return beam signals for the time—slice index I, denoted by RBkzt. For example, the omain subsets of samples RC1:0, RC2:0, RC3:0, RC4:0, and RC5:0 for lice index t=0 are input to a first return time—slice beam former 3016, which generates ponding samples of return beam signals RB1:0, RB2:0, RB3:0, and RB420 at a return time slice output 3018. For subsequent time—slice index values t=l, 2, the time—domain subsets of samples of composite return signals 907 are tiplexed by the return composite signal de—multiplexer 3012 for input to a second and a third return time—slice beamformer 3016, respectively, which generate samples for the return beam signals ated with the corresponding time-slice indices t at return time slice outputs 3018. also shows that at time-slice index value 1:3, the first return time-slice rmer generates samples of return beam signals associated with the corresponding time-slice index 3. The matrix product operation performed by each return lice beamformer 3016 for one time-slice index value I may take longer than the real time of the time—domain subset of samples (e.g., the number of samples S multiplied by the sample rate ts). However, each return lice beamformer 3016 may only process one time—domain subset of samples every N lice indices t. Return beam signal multiplexer 3020 receives return time slice outputs 3018 from each of the return time—slice beamformers 3016 and multiplexes the omain subsets of samples to generate the K return beam signals 915.

Although FIGs. 31—34 illustrate the same number N of forward time—slice beamformers 3006 as return time—slice beamformers 3016, some implementations may have more or fewer forward lice beamformers 3006 than return time—slice beamformers 3016.

In some examples, forward beamformer 529 and/or return beamformer 531 may have spare capacity for robustness to node failure. For example, if each forward time—slice beamformer 3006 takes fins to process one set of samples for a time—slice index I having a real—time time— slice duration ID, where tFTszN-tD, the forward rmer 529 may have N+E forward time- slice beamformers 3006. In some examples, each of the N+E forward time—slice beamformers 3006 are used in operation, with each forward time—slice beamformer 3006 having an effective extra capacity of E/N. If one forward time—slice beamformer 3006 fails, the operations may be shifted to another forward time—slice beamformer 3006 (e.g., by adjusting how time—domain samples (or groups of samples) are routed through the time—domain de—multiplexing and multiplexing). Thus, forward beamformer 529 may be tolerant of up to E forward time— slice beamformers 3006 g before system performance is impacted. In addition, extra capacity allows for system maintenance and upgrading of time—slice beamformers while the system is operating. For example, upgrading of time—slice beamformers may be performed incrementally because the system is tolerant of different performance between time— slice beamformers. The data samples associated with a time—slice index I may be interleaved. For example, a first time— slice index to may be associated with samples 0, P, 2P, . . . (S—l)*P, while a second time—slice index t1 may be associated with samples 1, P+l, 2P+l . . . (S—l)*P+l, etc., where S is the number of samples in each set of samples, and P is the interleaving on. The interleaving may also make the system more robust to time—slice beamformer failures, e each time— slice beamformer block of samples are separated in time such that errors due to a missing block would be buted in time, similarly to the advantage from interleaving in forward error tion. In fact, the buted errors caused by time-slice beamformer failure may cause s similar to noise and not result in any errors to user data, especially if forward error coding is employed. Although examples where N = 3 have been illustrated, other values of N may be used, and N need not have any particular relationship to K or M.

As discussed above, forward beamformer 529 and return beamformer 531 illustrated in FIGs. 31 and 33, respectively, may perform time—domain de—multiplexing and lexing for time-slice beamforming for one channel or ncy sub—band. Multiple sub—bands may be processed independently using an additional sub—band mux switching layer. is a block diagram of an example multi-band forward/retum beamformer 5 13 that employs sub—band de—multiplexing and multiplexing. The band forward/retum beamformer 5 13 may support F forward sub—bands and R return sub—bands.

Multi—band forward/return beamformer 5 13 includes F forward sub—band beamformers 3026, R return sub—band beamformers 3036, and a sub—band multiplexer/de— multiplexer 3030. For example, the d beam signals 511 may be split up into F d nds. Each of the F d sub—bands may be associated with a subset of the K forward user beam coverage areas. That is, the K forward user beam coverage areas may include multiple subsets of forward user beam coverage areas associated with different (e.g., different frequency and/or polarization, etc.) frequency nds, where the forward user beam coverage areas within each of the subsets may be non—overlapping (e.g., at 3 dB signal contours, etc.). Thus, each of the forward sub—band beamformer inputs 3024 may include a subset K1 of the forward beam signals 511. Each of the F forward beamformers 3026 may include the functionality of forward beamformer 529, generating forward sub—band beamformer outputs 3028 that comprise the M access node—specific forward signals associated with the subset of the forward beam signals 511 (e.g., a matrix product of the K1 forward beam signals with an M x K1 d beam weight matrix). Thus, each of the ANs 515 may receive multiple access node—specific forward signals associated with different frequency sub—bands (e.g., for each of the F forward sub— bands). The ANs may combine (e.g., sum) the signals in different sub—bands in the forward uplink signals, as sed in more detail below. Similarly, ANS 515 may generate multiple ite return signals 907 for R different return sub—bands. Each of the R return sub—bands may be associated with a subset of the K return user beam coverage areas. That is, the K return user beam coverage areas may include multiple subsets of return user beam ge areas associated with different frequency sub—bands, where the return user beam coverage areas within each of the subsets may be non—overlapping (e.g., at 3 dB signal contours, etc.). The nd multiplexer/de-multiplexer 3030 may split the composite return signals 907 into the R return sub-band beamformer inputs 3034.Each of the return sub-band beamformers 3036 may then generate a return sub-band beamformer output 3038, which may include the return beam signals 915 for a subset of the return user beams (e. g., to the feeder link modems 507 or return beam signal demodulator, etc.) In some examples, the band forward/return beamformer 513 may support multiple polarizations (e.g., right—hand circular polarization , left—hand circular zation (LHCP), etc), which in some cases may effectively double the number of sub-bands.

In some cases, time—slice multiplexing and de—multiplexing for forward beamformer 529 and return beamformer 531 (e.g., beam signal de—multiplexer 3002, forward access node signal lexer 3010, return ite signal de—multiplexer 3012, return beam signal multiplexer 3020) and sub—band multiplexing/de—multiplexing (sub—band multiplexer/de— lexer 3030) may be med by packet switching (e.g., Ethernet switching, etc). In some cases, the time—slice and sub-band switching may be performed in the same switching nodes, or in a different order. For example, a fabric switching architecture may be used where each switch fabric node may be coupled with a subset of the ANs 515, forward lice beamformers 3006, return time—slice rmers 3016, or feeder link modems 507. A fabric ing architecture may allow, for example, any AN to connect (e.g., via switches and/or a switch fabric interconnect) to any forward time—slice beamformer or return time—slice beamformer in a low—latency, hierarchically flat architecture. In one example, a system WO 09332 2016/026815 supporting K S 600, M S 600, and a 500 MHz bandwidth (e.g., per sub—band) with fourteen sub—bands for the forward or return links may be ented by a cially available interconnect switch platform with 2048 lOGigE ports.

Delay Equalization In some cases, differences in the propagation delays on each of the paths between the end—to—end relay 503 and the CPS 505 are insignificant. For example, on the return link, when the same signal (e.g., data to or from a particular user) is received by multiple ANs 515, each ce of the signal may arrive at the CPS essentially aligned with each other instance of the signal. Likewise, when the same signal is transmitted to a user terminal 517 through several ANs 515, each ce of the signal may arrive at the user terminal 517 essentially aligned with each other instance of the signal. In other words, signals may be phase and time aligned with sufficient precision that signals will coherently combine, such that the path delays and beamforming effects are small relative to the transmitted symbol rate. As an illustrative example, if the difference in path delays is 10 microseconds, the beamforming bandwidth can be on the order of tens of kHz and one can use a narrow dth signal, say le ksps with a small possible ation in performance. The 10 ksps signaling rate has a symbol duration of 100 microseconds and the 10 microsecond delay spread is only one tenth of the symbol duration.

In these cases, for the purposes of the system analysis, it may be assumed that signals received by the end—to—end relay at one instant will be relayed and transmitted at essentially the same time, as described earlier.

In other cases, there may be a significant difference in the propagation delay relative to the signaling interval (transmitted symbol duration) of the signals transmitted from the transmit a elements 409 to the ANs 515. The path that the signals take from each AN 515 through the distribution network 518 may contain significant delay ions. In these cases, delay equalization may be employed to match the path delays.

For end—to—end return link signals received through the distribution network 518 by the CPS 505, signals may be time aligned by using a relay beacon signal transmitted from the end— to—end relay, for example a PN beacon as bed earlier. Each AN 515 may time stamp the composite return signal using the relay beacon signal as a reference. ore, different ANs 515 may receive the same signal at different times, but the received signals in each AN 515 may be time stamped to allow the CPS 505 to time align them. The CPS 505 may buffer the signals so that beamforming is done by combining signals that have the same time stamp.

WO 09332 Returning to FIGS. 33 and 34, delay zation for the return link may be performed by de—multiplexing the composite return signals to the return time— slice beamformers 3016. For example, each AN may split up the composite return signal into sets of samples associated with time—slice indices t, which may include interleaved samples of the composite return signal. The time—slice s I may be determined based on the relay beacon signal. The ANs may send the subsets of samples multiplexed with the corresponding time—slice indices t (e.g., as a multiplexed composite return signal) to the return beamformer 531, which may serve as synchronization timing information on the return link. The subsets of samples from each AN may be de—multiplexed (e.g., Via switching) and one return time—slice beamformer 3016 may receive the subsets of samples from each AN for a time—slice index I (for one of multiple sub— bands, in some cases). By performing the matrix t of the return beam weight matrix and the subsets of samples from each of the M composite return signals associated with the time— slice index t, return time—slice beamformer 3016 may align the signals relayed by the end—to—end relay at the same time for ng the return beam weight matrix.

For the forward link, the beamformer 513 within the CPS 505 may generate a time stamp that indicates when each access node-specific d signal transmitted by the ANs 515 is desired to arrive at the -end relay 503. Each AN 515 may it an access node beacon signal 2530, for example a loopback PN signal. Each such signal may be -back and transmitted back to the ANs 515 by the end-to-end relay 503. The ANs 515 may receive both the relay beacon signal and the relayed (looped—back) access node beacon signals from any or all of the ANs. The received timing of the access node beacon signal relative to receive timing of the relay beacon signal indicates when the access node beacon signal arrived at the end—to—end relay. Adjusting the timing of the access node beacon signal such that, after relay by the end—to— end relay, it arrives at the AN at the same time as the relay beacon signal s at the AN, forces the access node beacon signal to arrive at the —end relay synchronized with the relay beacon. Having all ANs perform this function enables all access node beacon signals to arrive at the end—to—end relay synchronized with the relay beacon. The final step in the process is to have each AN transmit its access node—specific forward signals onized with its access node beacon . This can be done using timestamps as described subsequently.

Alternatively, the CPS may manage delay equalization by sending the respective access node— specific forward signals offset by the respective time—domain offsets to the ANs (e.g., where the timing Via the distribution network is deterministic). [0015 8] is an illustration of PN sequences used to align the timing of the system. The horizontal axis of the figure represents time. An AN1 PN sequence 2301 of chips 2303 is transmitted in the access node beacon signal from the first AN. The relative time of arrival of this sequence at the end—to—end relay is depicted by the PN sequence 2305. There is a time shift of PN sequence 2305 with respect to AN1 PN ce 2301, due to the propagation delay from the AN to the end—to—end relay. A relay PN beacon ce 2307 is generated within, and transmitted from, the end—to—end relay in a relay beacon signal. A PN chip of the relay PN beacon sequence 2307 at time To 2315 is aligned with a PN chip 2316 of the AN1 PN received signal 2305 at time T0. The PN chip 2316 of the AN1 PN received signal 2305 is aligned with the PN chip 2315 of the relay PN beacon 2307 when the AN1 transmit timing is adjusted by the proper amount. The PN sequence 2305 is looped back from the end—to—end relay and the PN sequence 2317 is received at AN1. A PN sequence 2319 transmitted from the end—to—end relay in the relay PN beacon is received at AN1. Note that the PN sequences 2317, 2319 are aligned at AN1 indicating that they were d at the end—to—end relay. shows an example of an ANZ that has not properly adjusted the timing of the PN sequence ted in the AN2. Notice that the PN sequence 2311 generated by the ANz is received at the end—to—end relay shown as ce 2309 with an offset by an amount dt from the relay PN beacon PN sequence 2307. This offset is due to an error of the timing used to generate the sequence in the ANg. Also, note that the arrival of the AN2 PN sequence 2321 at AN; is offset from the arrival of the relay PN beacon PN sequence at AN; 2323 by the same amount dz. The signal processing in AN2 will observe this error and may make a correction to the transmit timing by adjusting the timing by an amount dt to align the PN sequences 2321, 2323.

In FIGs. 36 and 37 the same PN chip rate has been used for the relay PN beacon and all of the AN (loopback) PN signals for ease of illustration of the concept. The same timing concepts can be applied with different PN chip rates. ing to FIGs. 31 and 32, the time— slice indices I may be used for onizing the access node—specific forward signals received from each of the ANs at the end—to-end relay. For example, the time—slice indices I may be multiplexed with the access node—specific forward signals 516. Each AN may transmit samples of the access node—specific forward signals with a particular time-slice index I aligned with corresponding timing information in the PN ce of chips transmitted in the respective access node beacon signals. Because the respective access node beacon signals have been adjusted to compensate for the tive path delays and phase shifts between the ANs and the end—to—end relay, the samples associated with the lice index I will arrive at the end—to—end relay with timing synchronized and phase aligned tly relative to each other.

In cases where ANs receive their own access node beacon signals, it is possible to loop back the access node beacon signals using the same end—to—end relay communication hardware that is also carrying the forward direction communication data. In these cases, the relative gains and/or phases of the transponders in the end—to—end relay can be adjusted as subsequently described. is a block diagram of an example AN 515. AN 515 comprises receiver 4002, receive timing and phase adjuster 4024, relay beacon signal demodulator 2511, lexer 4004, network interface 4006, controller 2523, tiplexer 4060, transmit timing and phase compensator 4020, and transmitter 4012. Network interface 4006 may be connected to, for example, CPS 505 via k port 4008.

On the return link, receiver 4002 receives a return downlink signal 527. The return downlink signal 527 may include, for example, a composite of return uplink signals relayed by the end—to—end relay (e.g., via multiple receive/transmit signal paths, etc.) and the relay beacon signal. Receiver 4002 may perform, for example, down—conversion and sampling. Relay beacon signal demodulator 2511 may demodulate the relay beacon signal in the digitized composite return signal 907 to obtain relay timing information 2520. For example, relay beacon signal demodulator 2511 may perform demodulation to recover the chip timing associated with the relay PN code and generate time stamps corresponding to the transmission time from the end-to- end relay for samples of the digitized composite return signal 527. Multiplexer 4004 may multiplex the relay timing information 2520 with the samples of the digitized ite return signal (e.g., to form a multiplexed composite return signal) to be sent to the CPS 505 (e.g., Via network interface 4006). Multiplexing the relay timing ation 2520 may include ting subsets of samples corresponding to time—slice indices t for sending to the CPS 505. For e, multiplexer 4004 may output subsets of samples associated with each time slice index t for input to the return time— slice rming architecture described above with reference to FIGs. 33, 34, and 35. Multiplexer 4004 may include an interleaver 4044 for interleaving samples for each subset of samples, in some cases.

On the forward link, network ace 4006 may obtain AN input signal 4014 (e.g., via network port 4008). De—multiplexer 4060 may de—multiplex AN input signal 4014 to obtain access node—specific forward signal 516 and forward signal transmit timing information 4016 indicating transmission timing for the access node—specific forward signal 516. For example, the access node—specific forward signal 516 may comprise the forward signal transmit timing ation (e.g., multiplexed with data samples, etc). In one example, the access node—specific forward signal 516 comprises sets of samples (e.g., in data packets), where each set of s is associated with a time—slice index I. For example, each set of samples may be samples of the access node—specific d signal 516 generated according to the forward time—slice beamforming architecture discussed above with reference to FIGS. 31, 32 and 35. De— multiplexer 4060 may include a de—interleaver 4050 for erleaving samples associated with time—slice indices t.

Transmit timing and phase compensator 4020 may receive and buffer access node— specific forward signal 516 and output d uplink signal s 4022 for transmission by the transmitter 4012 at an appropriate time as forward uplink signal 521. The transmitter 4012 may perform digital—to—analog conversion and up—conversion to output the forward uplink signal 521. Forward uplink signal samples 4022 may include the access node—specific forward signal 516 and an access node beacon signal 2530 (e.g., loopback PN signal), which may include transmit timing information (e.g., PN code chip timing information, frame timing information, etc). Transmit timing and phase compensator 4020 may multiplex the access node—specific forward signal 516 with the access node beacon signal 2530 such that the forward signal transmit timing and phase information 4016 is synchronized to corresponding transmit timing and phase information in the access node beacon signal 2530.

In some examples, generation of the access node beacon signal 2530 is performed y at the AN 515 (e.g., in access node beacon signal tor 2529). Alternatively, generation of the access node beacon signal 2530 may be performed in a separate component (e.g., CPS 505) and sent to the AN 515 (e.g., via network interface 4006). As discussed above, the access node beacon signal 2530 may be used to compensate the forward uplink signal 521 for path differences and phase shifts between the AN and the end—to-end relay. For example, the access node beacon signal 2530 may be transmitted in the forward uplink signal 521 and relayed by the end—to—end relay to be received back at receiver 4002. The controller 2523 may compare d transmit timing and phase information 4026 obtained (e.g., by demodulation, etc.) from the relayed access node beacon signal with e timing and phase information 4028 obtained (e.g., by demodulation, etc.) from the relay beacon signal. The controller 2523 may generate a timing and phase ment 2524 for input to the it timing and phase compensator 4020 to adjust the access node beacon signal 2530 to compensate for the path delay and phase shifts.

For example, the access node beacon signal 2530 may include a PN code and frame timing information (e.g., one or more bits of a frame , etc). The transmit timing and phase compensator 4020 may, for example, adjust the frame timing information for coarse compensation for the path delay (e. g., output frame timing information in the access node beacon signal such that the relayed access node beacon signal will have the relayed transmit frame timing information coarsely d with corresponding frame timing information in the relay beacon signal, changing which chip of the PN code is considered to be the LSB, eta).

Additionally or alternatively, the transmit timing and phase compensator 4020 may perform timing and phase adjustments to the forward uplink signal samples 4022 to compensate for timing or phase differences between the relayed transmit timing and phase information 4026 and receive timing and phase information 4028. For example, where the access node beacon signal 2530 is ted based on a local oscillator, timing or phase differences between the local oscillator and the received relay beacon signal may be corrected by timing and phase adjustments to the forward uplink signal samples 4022. In some es, demodulation of the access node beacon signal is performed locally at the AN 515 (e.g., in access node beacon signal demodulator 2519). Alternatively, demodulation of the access node beacon signal may be performed in a separate component (e.g., CPS 505) and the relayed transmit timing and phase information 4026 may be ed in other signaling (e.g., via network interface 4006). For example, deep fading may make ion and demodulation of the ANS own relayed access node beacon signal difficult without transmission at higher power than other signaling, which may reduce the power budget for communication signals. Thus, combining reception of the relayed access node beacon signal from multiple ANs 515 may increase the effective received power and demodulation accuracy for the relayed access node beacon signal. Thus, demodulation of the access node beacon signal from a single AN 515 may be performed using downlink s received at multiple ANs 515. Demodulation of the access node beacon signal may be performed at the CPS 505 based on the ite return signals 907, which may also include signal information for the access node beacon s from most or all ANs 515. If desired, end—to—end beamforming for the access node beacon s can be performed taking into account the access node beacon uplinks (e.g., Cr), relay loopback(e.g., E), and/or access node beacon downlinks (e.g., Ct).

Feeder Link Impairment l In addition to delay equalization of the signal paths to the end—to—end relay from all the ANs, the phase shifts induced by feeder links can be removed prior to beamforming. The phase shift of each of the links n the end—to—end relay and the M ANs will be different. The causes for different phase shifts for each link include, but are not limited to, the propagation path length, atmospheric ions such as scintillation, Doppler frequency shift, and different AN oscillator errors. These phase shifts are generally ent for each AN and are time varying (due to scintillation, Doppler shift, and difference in the AN oscillator errors). By removing dynamic feeder link impairments, the rate at which beam weights adapt may be slower than an alternative where the beam weights adapt fast enough to track the cs of the feeder link.

In the return direction, feeder downlink impairments to an AN are common to both the relay PN beacon and user data signals (e.g., return downlink s). In some cases, coherent demodulation of the relay PN beacon provides channel information that is used to remove most or all of these impairments from the return data signal. In some cases, the relay PN beacon signal is a known PN sequence that is continually transmitted and located in—band with the communications data. The lent (or effective) isotropically radiated power (EIRP) of this d PN signal is set such that the interference to the ications data is not larger than a maximum acceptable level. In some cases, a feeder link impairment removal process for the return link involves coherent demodulation and tracking of the received timing and phase of the relay PN beacon signal. For example, relay beacon signal demodulator 2511 may determine receive timing and phase adjustments 2512 to compensate for feeder link impairment based on comparing the relay PN beacon signal with a local nce signal (e.g., local oscillator or PLL). The recovered timing and phase differences are then removed from the return downlink signal (e.g., by receive timing and phase adjuster 4024), hence removing feeder link impairments from the ications signal (e.g., return downlink signals 527). After feeder link impairment removal, the return link signals from a beam will have a common frequency error at all ANs and thus be le for beamforming. The common frequency error may include, but is not limited to, contributions from the user terminal frequency error, user terminal uplink Doppler, end—to—end relay frequency translation frequency error and relay PN beacon frequency error.

In the forward direction, the access node beacon signal from each AN may be used to help remove feeder uplink ments. The feeder uplink impairments will be imposed upon the forward link ications data (e.g., the access node—specific signal) as well as the access node beacon signal. Coherent demodulation of the access node beacon signal may be used to r the timing and phase differences of the access node beacon signal (e.g., relative to the relay beacon signal). The recovered timing and phase differences are then removed from the transmitted access node beacon signal such that the access node beacon signal arrives in phase with the relay beacon signal.

In some cases, the forward feeder link removal process is a phase locked loop (PLL) with the path delay from the AN to the end—to—end relay and back within the loop structure. In some cases, the round—trip delay from the AN to the end—to—end relay and back to the AN can be significant. For example, a geosynchronous satellite functioning as an —end relay will generate round—trip delay of approximately 250 milliseconds (ms). To keep this loop stable in the presence of the large delay, a very low loop bandwidth can be used. For a 250 ms delay, the PLL closed loop bandwidth may typically be less than one Hz. In such cases, high—stability oscillators may be used on both the ite and the AN to maintain reliable phase lock, as indicated by block 2437 in (see below).

In some cases, the access node beacon signal is a burst signal that is only transmitted during calibration intervals. During the calibration al, communications data is not transmitted to eliminate this interference to the access node beacon signal. Since no communications data is transmitted during the calibration interval, the transmitted power of the access node beacon signal can be large, as compared to what would be required if it were broadcast during communication data. This is because there is no concern of causing interference with the communications data (the communications data is not t at this time).

This technique enables a strong signal—to—noise ratio (SNR) for the access node beacon signal when it is transmitted during the calibration interval. The frequency of occurrence of the calibration intervals is the reciprocal of the elapsed time between calibration intervals. Since each ation interval provides a sample of the phase to the PLL, this calibration frequency is the sample rate of this discrete time PLL. In some cases, the sample rate is high enough to support the closed loop bandwidth of the PLL with an insignificant amount of aliasing. The product of the calibration frequency (loop sample rate) and the calibration interval represents the fraction of time the -end relay cannot be used for ications data t additional interference from the l sounding probe signal. In some cases, values of less than 0.1 are used and in some cases, values of less than 0.01 are used. is a block diagram of an example AN transceiver 2409. The input 2408 to the AN transceiver 2409 receives -end return link signals received by the AN 515 (e.g., for one of a plurality of frequency sub—bands). The input 2408 is coupled to the input 2501 of a down converter (D/C) 2503. The output of the D/C 2503 is coupled to an analog to digital converter (A/D) 2509. The output of the A/D 2509 is coupled to an Rx time adjuster 25 15 and/or Rx phase adjuster 25 17. RX time adjuster 25 15 and RX phase adjuster 25 17 may illustrate aspects of the receive timing and phase adjuster 4024 of . The D/C 2503 is a quadrature down converter. Accordingly, the D/C 2503 outputs an in—phase and quadrature output to the A/D 2509. The received s may e communications signals (e.g., a composite of return uplink signals itted by user terminals), access node beacon signals (e.g., transmitted from the same AN and/or other ANs) and a relay beacon signal. The digital samples are coupled to a relay beacon signal lator 251 l. The relay beacon signal demodulator 25ll demodulates the relay beacon signal. In addition, the relay beacon signal demodulator 25ll generates a time l signal 2513 and a phase control signal 2514 to remove feeder link impairments based on the received relay beacon signal. Such impairments include Doppler, AN ncy error, scintillation effects, path length changes, etc. By performing nt demodulation of the relay beacon signal, a phase locked loop (PLL) may be used to correct for most or all of these errors.

By correcting for the errors in the relay beacon signal, corresponding errors in the ication signals and access node beacon signals on the feeder link are corrected as well (e.g., since such errors are common to the relay beacon signal, the access node beacon signals and the communications s). After feeder link impairment removal, the end—to—end return link communication signal from a user terminal 517 nominally have the same frequency error at each of the M ANs 515. That common error includes the user al frequency error, the user link Doppler, the end—to—end relay frequency translation error, and the relay beacon signal ncy error.

The digital samples, with feeder link impairments removed, are coupled to a multiplexer 2518, which may be an example of the multiplexer 4004 of . The lexer 2518 associates (e.g., time stamps) the s with the relay timing information 2520 from the relay beacon signal lator 2511. The output of the multiplexer 2518 is coupled to the output port 2410 of the AN transceiver 2409. The output port 2410 is coupled to the multiplexer 2413 and through the interface 2415 (see ) to the CPS 505. The CPS 505 can then use the time stamps associated with the received digital samples to align the digital samples received from each of the ANs 515. Additionally or alternatively, feeder link impairment removal may be performed at the CPS 505. For example, digital samples of the end— to—end return link signals with the embedded relay beacon signal may be sent from the AN 515 to the CPS 505, and the CPS 505 may use the synchronization timing information (e.g., embedded relay beacon signal) in each of the composite return signals to ine respective adjustments for the respective composite return signals to compensate for downlink channel impairment.

] An access node beacon signal 2530 may be generated locally by an access node beacon signal generator 2529. An access node beacon signal demodulator 2519 demodulates the access node beacon signal received by the AN 515 (e.g., after being relayed by the end—to—end relay and received at input 2408). The relay beacon signal demodulator 2511 provides a received relay timing and phase information signal 2521 to a controller 2523. The controller 2523 also receives a relayed transmit timing and phase information signal 2525 from the access node beacon signal demodulator 2519. The controller 2523 es the received relay timing and phase information with the relayed transmit timing and phase information and generates a coarse time adjust signal 2527. The coarse time adjust signal 2527 is coupled to the access node beacon signal generator 2529. The access node beacon signal generator 2529 generates the access node beacon signal 2530 with embedded transmit timing information to be itted from the AN 515 to the end—to—end relay 503. As noted in the discussion above, the ence between the relay timing and phase information (embedded in the relay beacon signal) and the transmit time and phase information (embedded in the access node beacon signal) is used to adjust the transmit timing and phase information to synchronize the relayed transmit timing and phase information with the received relay timing and phase information. Coarse time is adjusted by the signal 2527 to the access node beacon signal tor 2529 and fine time is adjusted by the signal 2540 to the TX time adjuster 2539. With the relayed transmit timing and phase information 2525 from the access node beacon signal demodulator 2519 synchronized with the ed relay timing and phase information 2521, the access node beacon signal generator 2529 generates timestamps 2531 that assist in the onization of the access node beacon signal 2530 and the access node-specific forward signal from the CPS 505 that is transmitted. That is, data samples from the CPS 505 are received on input port 2423 together with timestamps 2535 that indicate when the associated data samples is desired to arrive at the end-to-end relay 503. A buffer, time align and sum module 2537 buffers the data samples d from the CPS 505 and sums them with the samples from the access node beacon signal generator 2529 based on the timestamps 2535, 2531. PN samples and communication data s with identical times, as indicated by the time stamps, are summed together. In this example, the multiple beam signals (xk(n) * bk) are summed together in the CPS 505 and the access node-specific forward signal comprising a composite of the multiple beam signals is sent to the AN by the CPS 505.

When aligned properly by the ANs, the data samples arrive at the end—to—end relay 503 at the d time (e.g., at the same time that the same data samples from other ANs arrive). A transmit time adjuster 2539 performs fine time adjustments based on a fine time controller output signal 2540 from the time controller module 2523. A transmit phase er 2541 performs phase adjustments to the signal in response to a phase control signal 2542 generated by the access node beacon signal demodulator 2519. Transmit time adjuster 2539 and transmit phase adjuster 2541 may illustrate, for example, aspects of the transmit timing and phase compensator 4020 of .

] The output of the transmit phase er 2541 is d to the input of a digital to analog converter (D/A) 2543. The quadrature analog output from the D/A 2543 is coupled to an up—converter (U/C) 2545 to be transmitted by the HPA 2433 (see ) to the end—to—end relay 503. An amplitude control signal 2547 provided by the access node beacon signal demodulator 2519 provides amplitude feedback to the U/C 2545 to compensate for items such as uplink rain fades.

In some cases, the PN code used by each AN for the access node beacon signal 2530 is different from that used by every other AN. In some cases, the PN codes in the access node beacon signals are each different from the relay PN code used in the relay beacon . ingly, each AN 515 may be able to distinguish its own access node beacon signal from those of the other ANs 515. ANs 515 may distinguish their own access node beacon signals from the relay beacon signal.

] As was previously described, the end—to—end gain from any point in the coverage area to any other point in the area is a multipath channel with L different paths that can result in very deep fades for some point to point channels. The transmit diversity (forward link) and receive diversity (return link) are very effective in mitigating the deep fades and enable the communications system to work. However for the access node beacon signals, the transmit and receive diversity is not t. As a result, the point—to-point link of a ck signal, which is the transmission of the signal from an AN back to the same AN, can experience end-to-end gains that are much lower than the average. Values of 20 dB below the e can occur with a large number of e/transmit signal paths (L). These few low end-to-end gains result in lower SNR for those ANs and can make link closure a challenge. Accordingly, in some cases, higher gain antennas are used at the ANs. Alternatively, referring to the example transponder of , a phase adjuster 418 may be included in each of the receive/transmit signal paths. The phase er 418 may be individually adjusted by the phase shift ller 427 (for example, under control of a telemetry, tracking, and command (TT&C) link from an Earth—based control center). Adjusting the relative phases may be ive in increasing the end—to—end gains of the low-gain loopback paths. For example, an objective may be to choose phase shift settings to increase the value of the worst case loopback gain (gain from an AN back to itself). Note that the selection of phases generally does not change the distribution of the gains when evaluated for all points in the coverage area to all other points in the coverage area, but it can increase the gains of the low gain loopback paths.

To elaborate, consider the set of gains from each ofM ANs 515 to all of the other ANs 515. There are M gains, of which, only M of them are loopback paths. Consider two gain distributions, the first is the total distribution of all paths (M 2) which can be estimated by ing a histogram of all M 2paths. For ANs distributed evenly over the entire coverage area, this distribution may be representative of the distribution of the end—to—end gain from any point to any other point in the coverage area. The second distribution is the loopback gain distribution (loopback distribution) which can be estimated by compiling a histogram of just the M ck paths. In many cases, custom selection of the receive/transmit signal path phase settings (and ally gain settings) does not provide a significant change to the total distribution. This is especially the case with random or interleaved mappings of transmit to receive elements.

However, in most cases, the loopback distribution can be improved with custom selection (as opposed to random values) of the phase (and optionally gain) settings. This is because the set of loopback gains consist of M paths (as opposed to M 2total paths) and the number of degrees of freedom in the phase and gain adjustments is L. Often times L is on the same order as M which enables significant improvement in low loopback gains with custom phase selection. Another way of looking at this is that the custom phase selection is not necessarily eliminating low end— to—end gains, but rather moving them from the set of loopback gains (M members in the set) to the set of non—loopback gains (M 2—M members). For non—trivial values of M, the larger set is often much larger than the former.

An AN 515 may process one or more frequency nds. is a block diagram of an e AN 515 in which multiple frequency sub-bands are processed separately.

On the end-to-end return link 523 (see , the AN 515 receives the return downlink signals 527 from the end-to-end relay 503 h a low noise amplifier (LNA) 2401. The amplified s are coupled from the LNA 2401 to a power divider 2403. The power divider 2403 splits the signal into multiple output signals. Each signal is output on one of the output ports 2405, 2407 of the power divider 2403. One of the output ports 2407 may be provided as a test port.

The other ports 2405 are coupled to an input 2408 of a corresponding one of multiple AN transceivers 2409 (only one . The AN transceivers 2409 process the signals received on corresponding sub—bands. The AN eiver 2409 performs several functions, discussed in detail above. The outputs 2410 of the AN eivers 2409 are coupled to input ports 2411 of a sub-band multiplexer 2413. The outputs are combined in the sub—band multiplexer 2413 and output to a distribution network interface 2415. The ace 2415 provides an interface for data from/to AN 515 to/from the CPS 505 over the distribution network (see . sing frequency sub—bands may be ageous in reducing mance requirements on the RF components used to implement the end—to—end relay and AN. For example, by splitting up 3.5 GHZ of bandwidth (e.g., as may be used in a Ka—band system) into seven sub—bands, each sub— band is only 500 MHZ wide. That is, each of the access node—specific forward signals may e le sub—signals associated with the different sub—bands (e.g., associated with different subsets of the forward user beam coverage areas), and the AN transceivers 2409 may upconvert the sub—signals to different carrier frequencies. This bandwidth splitting may allow 2016/026815 for lower tolerance components to be used since amplitude and phase ions between different sub—bands may be compensated by separate beamforming weights, calibration, etc. for the different sub—bands. Of course, other systems may use a different number of sub—bands and/or test ports. Some cases may use a single sub—band and may not include all the ents shown here (e.g., omitting power divider 2403 and mm; 2413).

On the end—to—end forward link 501, data is received from the CPS 505 by the interface 2415. The received data is d to an input 2417 of a sub—band de—multiplexer 2419. The sub—band de—multiplexer 2419 splits the data into multiple data signals. The data signals are coupled from output ports 2421 of the sub—band de—multiplexer 2419 to input ports 2423 of the AN transceivers 2409. Output ports 2425 of the AN transceivers 2409 are coupled to input ports 2427 of the summer module 2429. The summer module 2429 sums the signals output from the seven AN transceivers 2409. An output port 2431 of the summer module 2429 couples the output of the summer module 2429 to the input port 2433 of a high power amplifier (HPA) 2435. The output of the HPA 2435 is coupled to an antenna (not shown) that transmits the signals output to the end—to—end relay 503. In some cases, an stable oscillator 2437 is coupled to the AN transceivers 2409 to provide a stable reference ncy source.

Beam weight Computation Returning to which is an example description of signals on the return link, a mathematical model of the end—to—end return link may be used to describe the link as: y = Bretht E(Ar X + 11111 ) + ndl EQ' 1 = Bret[Hret X + Ct Enlll + ndl] where, X is the K X 1 column vector of the transmitted signal. In some cases, the magnitude squared of every element in X is defined to be unity (equal it power). In some cases, this may not always be the case. y is the K X 1 column vector of the received signal after beamforming.

Ar is the L X K return uplink radiation matrix. The t alk ns the gain and phase of the path from a reference location located in beam K to the [m (the letter “el”) receive antenna element 406 in the RX array. In some cases, the values of the return uplink radiation matriX are stored in the channel data store 941 (see ).

E is the L x L payload matrix. The element eij defines the gain and phase of the signal from the jth antenna element 406 in the e array to an im antenna element 409 in the transmit array. In some cases, aside from incidental crosstalk between the paths (resulting from the finite isolation of the electronics), the E matrix is a diagonal matrix. The matrix E can be normalized such that the sum of the magnitude squared of all elements in the matrix is L. In some cases, the values of the d matrix are stored in the channel data store 941 (see ).

Ct is the M x L return downlink radiation matrix. The element cm, contains the gain and phase of the path from 1th (the letter “el”) antenna element in the Tx array to an mth AN 515 from among the M ANs 515. In some cases, the values of the return downlink radiation matrix are stored in the channel data store 941 (see ).

Hret is the M x K return l matrix, which is equal to the t Ct x E x Ar. nut is an L x 1 noise vector of complex an noise. The covariance of the uplink noise is E|nuln§1| 2 205m. IL is the L x L ty matrix. 02 is noise variance. aglis enced on the uplink, while Uglis experienced on the downlink. ndl is an M x 1 noise vector of complex Gaussian noise. The covariance of the downlink noise is E|nd1n§1| = 205111“. 1M is the M x M identity matrix.

Bret is the K x M matrix of end-to—end return link beam weights. es are generally described above (e.g., with reference to FIGS. 6 — 11) in a manner that assumes certain rities between forward and return end—to—end multipath channels. For example, the forward and return channel matrices are described above with reference generally to M, K, E, and other models. However, such descriptions are intended only to simplify the description for added clarity, and are not intended to limit examples only to cases with identical urations in the forward and return directions. For example, in some cases, the same transponders are used for both forward and return traffic, and the payload matrix E can be the same for both d and return end—to—end beamforming (and ponding beam weight computations), accordingly. In other cases, different transponders are used for forward and return traffic, and a different forward payload matrix (Efwd) and a return payload matrix (Eret) can be used to model the corresponding end—to—end multipath channels and to compute corresponding beam weights. Similarly, in some cases, the same M ANs 515 and K user terminals 517 are considered part of both the forward and return end—to—end multipath channels.

In other cases, M and K can refer to different subsets of ANs 515 and/or user terminals 517, and/or different s of ANs 515 and/or user terminals 517, in the forward and return directions.

Beam s may be computed in many ways to satisfy system requirements. In some cases, they are computed after deployment of the end—to—end relay. In some cases, the payload matrix E is measured before deployment. In some cases, beam weights are computed with the objective to increase the signal to erence plus noise (SINR) of each beam and can be computed as follows: Bret = (R‘IH)H EQ‘ 2’ 3 R = zaglIM + zagctEEflcf‘ + HHH where R is the covariance of the received signal and ( * )H is the conjugate transpose (Hermetian) operator.

The k, m element of the K x M return beam weight matrix Bret provides the weights to form the beam to the mth AN 515 from a user terminal in the kth user beam. Accordingly, in some cases, each of the return beam weights used to form return user beams are computed by estimating end-to-end return gains (i.e., elements of the channel matrix Hret) for each of the end-to-end multipath ls (e.g., each of the end-to-end return multipath channels).

EQ. 2 holds true where R is the covariance of the received signal as ed in EQ. 3.

Therefore, when all of the matrices of EQ. 1, 2 and 3 are known, the beam weights used to form -end beams may be ly determined.

This set of beam weights reduces the mean d error between x and y. It also increases the —end signal to noise plus interference ratio (SINR) for each of the K end—to— end return link signals 525 (originating from each of the K beams).

The first term 205111“ in EQ. 3 is the covariance of the downlink noise (which is uncorrelated). The second term ZO'EICtEEHCEi in EQ. 3 is the covariance of the uplink noise (which is correlated at the ANs). The third term HHH in EQ. 3 is the covariance of the signal.

Setting the variance of the uplink noise to zero and ignoring the last term ( HHH) results a set of weights that increases the signal to downlink noise ratio by phase-aligning the ed s on each of the M ANs 515. Setting the downlink noise variance to zero and ignoring the 3rd term results in a set of weights that increases the uplink SINR. Setting both the uplink and downlink noise variances to zero results in a de—correlating receiver that increases the carrier to interference (C/l) ratio.

In some cases, the beam weights are normalized to make the sum of the magnitude squared of any row of Bret sum to unity.

In some cases, the solution to EQ. 2 is determined by a priori knowledge of the matrices Ar, Ct, and E as well as the variances of the noise s 11.11 and Hal. dge of the element values of the matrices can be obtained during measurements made during the manufacturing and testing of relevant components of the end—to—end relay. This may work well for systems where one does not expect the values in the matrices to change significantly during system operation. However, for some s, ally ones operating in higher frequency bands, such expectations may not be present. In such cases, the matrices Ar, Ct, and E may be estimated subsequent to the deployment of a craft (such as a satellite) on which the —end relay is disposed.

In some cases where a priori information is not used to set the weights, the solution to EQ. 2 may be determined by estimating the values of R and H. In some cases, designated user als 517 in the center of each user beam coverage area 519 transmit known signals X during calibration periods. The vector received at an AN 515 is: u=Hx+CtEnul+ndl EQ.4 In an example, the CPS 505 estimates the values of R and H based on the following relationships: fi = 2 uuH EQ. 5 H = [15111521 13K] EQ 6 13K 2 Z ui}: EQ.:I I: is an estimate of the covariance matrix R, II is an estimate of channel matrix H and [3k is an estimate of the ation vector, 3?}: is the conjugate of the km component of the transmitted vector with the frequency error introduced by the uplink transmission. In some cases, no return communication data is transmitted during the calibration period. That is, only calibration signals that are known to the ANs are itted on the end—to—end return link during the calibration period in order to allow the value of pk to be determined from the received vector u using the equation above. This, in turn allows the value of II to be determined.

Both the covariance matrix te Ii and the channel matrix estimate II are determined based on the signals received during the calibration period.

In some cases, the CPS 505 can estimate the covariance matrix I? while communication WO 09332 data is present (e. g., even when X is unknown). This may be seen from the fact that B is determined based only on the received signal u. Nonetheless, the value of H is ted based on s ed during a calibration period during which only calibration signals are transmitted on the return link.

In some cases, estimates of both the channel matrix H and the covariance matrix B are made while communication data is being transmitted on the return link. In this case, the covariance matrix B is estimated as noted above. However, the value of x is determined by demodulating the received signal. Once the value of X is known, the channel matrix may be estimated as noted above in EQ. 6 and EQ. 7.

The signal and interference components of the signal after beamforming are ned in the vector Bret H x. The signal and interference powers for each of the beams are contained in the K x K matrix Bret H. The power in the km diagonal element of Bret H is the desired signal power from beam k. The sum of the magnitude squared of all elements in row k except the diagonal element is the interference power in beam k. Hence the C/I for beam k is: EQ. 8 where Skj are the elements of Bret H. The uplink noise is contained in the vector Bret Ct Enul, which has a K x K covariance matrix of Zaleret Ct E EH CtH BretH . The km diagonal element of the covariance matrix contains the uplink noise power in beam k. The uplink signal to noise ratio for beam k is then computed as: EQ. 9 111 t kk where tkk is the km diagonal element of the uplink covariance matrix. The downlink n01se is contained in the vector Bret ndl, which has a covariance of 20dZIK by v1rtue of the normalized beam weights. Hence the downlink signal to noise ratio is: s Is kk I2 = EQ. 10 N 2 (11 k 2 6 The end—to—end SINR is the combination of EQ. 8 — 10: ‘1 _1 44 C s s SINsz (—j + — + — EQ.11 I k Nul k Nd] k The above equations describe how to calculate the end—to—end SINR given the d matrix E. The payload matrix may be constructed by intelligent choice of the gain and phases of each of the elements of E. The gain and phase of the diagonal elements of E that optimize some utility metric (which is lly a function of the K beam SINR’s as computed above) may be selected and implemented by setting the phase r 418 in each of the L onders 411. Candidate utility functions include, but are not limited to, sum of SH\IRk (total SINR), sum of Log(1+SINRk) (proportional to total throughput) or total power in the channel matrix, H. In some cases, the improvement in the y function by customizing the gains and phases is very small and insignificant. This is sometimes the case when random or interleaved mappings of antenna elements are used. In some cases, the utility function can be improved by a non-trivial amount by custom ion of the receive/transmit signal gain and phase.

Returning to a mathematical model of the end-to-end forward link 501 may be used to describe the link 501 as: y= AtE[Cr Bfwd X+nul ]+mdl EQ. 12 =wad Bfwd X+AEnu l +Ildl where, X is the K x 1 column vector of the transmitted signal. The magnitude squared of every t in X is defined to be unity (equal signal power). In some cases, unequal transmit power may be achieved by selection of the forward beam weights. y is the K x 1 column vector of the received signal.

Cr is the L x M forward uplink radiation matrix. The element clm contains the gain and phase of the path 2002 from mm AN 515 to the [”1 (letter “el”) receive a element 406 of the Rx array of antenna on the end—to—end relay 503. In some cases, the values of the forward uplink radiation matrix are stored in the channel data store 921 (see ).

E is the L x L d matrix. The element 6, s the gain and phase of the signal from jth receive array antenna element to the im antenna element of the transmit array. Aside from incidental crosstalk between the paths (resulting from the finite isolation of the electronics), the E matrix is a diagonal matrix. In some cases, the matrix E is normalized such that the sum of the magnitude d of all elements in the matrix is L. In some cases, the values of the payload matrix are stored in the channel data store 921 (see ).

At is the K x L forward downlink radiation matrix. The element aid contains the gain and phase of the path from antenna element L (letter “el”) in the Tx array of the end—to—end relay 503 to a reference location in user beam k. In some cases, the values of the forward nk radiation matrix are stored in the channel data store 921 (see ). wad is the K x M forward channel matrix, which is equal to the t AtECr. nul is an L x 1 noise vector of x an noise. The covariance of the uplink noise is: E[nulnul ]: 2CSulILaH 2 where IL is the L x L identity matrix. ndl is an K x 1 noise vector of complex Gaussian noise. The covariance of the downlink noise is: Eindlnclil] = 203111? where IK is the K x K identity matrix.

Bfwd is the M x K beam weight matrix of —end forward link beam weights.

The beam weights for user beam k are the elements in column k of Bfwd. Unlike the return link, the C/I for beam k is not determined by the beam weights for beam k. The beam weights for beam k determine the uplink signal to noise ratio (SNR) and the downlink SNR, as well as the carrier (C) power in the C/I. However, the interference power in beam k is determined by the beam weights for all of the other beams, except for beam k. In some cases, the beam weight for beam k is selected to se the SNR. Such beam weights also increase the CH for beam k, since C is increased. However, interference may be generated to the other beams.

Thus, unlike in the case of the return link, optimal beam weights are not computed on a beam— by—beam basis (independent of the other beams).

In some cases, beam weights (including the radiation and payload matrices used to compute them) are determined after deployment of the end—to—end relay. In some cases, the payload matrix E is measured before deployment. In some cases, one can compute a set of beam s by using the interference created in the other beams by beam k and counting it as the interference in beam k. gh this approach may not compute optimum beam weights, it may be used to fy weight computation. This allows a set of weights to be determined for each beam independent of all other beams. The resulting forward beam weights are then computed similar to the return beam weights: Bfwd: HH R? where, EQ. 13 R = ZafllK + 2021AtEEH AtfI +H HH 11 EQ 14 The first term 20(111 K in EQ. 14 is the covariance of the downlink noise (uncorrelated). The 2 H H second term 20-ulAt EE At is the ance of the uplink noise (which is correlated at the ANs). The third term HHILI is the covariance of the signal. g the variance of the uplink noise to zero and ignoring the last term (HHH) results in a set of weights that increases the signal to nk noise ratio by phase aligning the received signals at the M ANs 515. Setting the downlink noise variance to zero and ignoring the 3rd term results in a set of weights that increases the uplink SNR. Setting both the uplink and downlink noise variances to zero results in a relating receiver that increases the C/I ratio. For the forward link, the downlink noise and interference generally dominate. Therefore, these terms are lly useful in the beam weight computation. In some cases, the second term in EQ. 14 (the uplink noise) is insignificant compared to the first term (the downlink noise). In such cases, the second term can be ignored in co—variance calculations, further simplifying the calculation while still yielding a set of beam s that ses the end—to—end SINR.

As with the return link, the beam weights may be normalized. For transmitter beam weights with equal power allocated to all K forward link signals, each column of Bfwd may be scaled such that the sum of the magnitude squared of the elements in any column will sum to unity. Equal power sharing will give each of the signals the same fraction of total AN power (total power from all ANs ted to signal Xk). In some cases, for forward links, an unequal power sharing between forward link signals is implemented. Accordingly, in some cases, some beam signals get more than an equal share of total AN power. This may be used to equalize the SINR in all beams or give more important beams larger SINR’s than lesser important beams. To create the beam weights for unequal power sharing, the M x K equal power beam weight matrix, Bfwd, is post lied by a K x K diagonal matrix, P, thus the new Bfwd = Bfwd P. Let P = diag (J17) then the squared valued of the km diagonal element represents the power allocated to user signal WO 09332 xk. The power sharing matrix P is normalized such that the sum or the square of the diagonal elements equals K (the agonal elements are zero).

In some cases, the solution to EQ. 13 is determined by a priori knowledge of the matrices At, Cr, and E, as well as the variances of the noise vectors 11.11 and ndl. In some cases, knowledge of the matrices can be obtained during measurements made during the manufacturing and g of relevant components of the end—to—end relay. This can work well for systems where one does not expect the values in the matrices to change significantly from what was measured during system operation. However, for some systems, especially ones operating in higher frequency bands, this may not be the case.

In some cases where a priori information is not used to set the weights, the values of R and H for the forward link can be estimated to determine the solution to EQ. 13. In some cases, ANS transmit a channel sounding probe during calibration s. The channel sounding probes can be many different types of signals. In one case, different, orthogonal and known PN sequences are transmitted by each AN. The channel sounding probes may be pre—corrected in time, frequency, and/or phase to remove the feeder link impairments (as discussed further below). All communication data may be turned off during the calibration interval to reduce the erence to the channel sounding . In some cases, the channel ng probes can be the same signals as those used for feeder link impairment removal.

During the calibration interval, a terminal in the center of each beam may be designated to receive and process the channel sounding . The Kx] vector, u, of received signals during the calibration period is u=H X + At E nu1+ ndl where X is the Mx] vector of transmitted channel sounding probes. In some cases, each designated terminal first s the incidental frequency error (resulting from Doppler shift and terminal oscillator error), and then correlates the resulting signal with each of the M known, orthogonal PN sequences. The results of these correlations are M x numbers (amplitude and phase) for each terminal and these results are itted back to the CPS via the return link. The M complex numbers calculated by the terminal in the center of the kth beam can be used to form the kth row of the estimate of the channel matrix, II. By using the ements from all of K designated terminals, an estimate of the entire channel matrix is obtained. In many cases, it is useful to combine the measurement from multiple calibration intervals to improve the estimate of the channel matrix.

Once the estimate of the channel matrix is determined, an estimate of the ance , R, can be determined from EQ. 14 using a value of 0 for the second term. This may be a very te estimate of the covariance matrix if the uplink noise (the second term in EQ. 14) is negligible relative to the downlink noise (the first term in EQ. 14). The forward link beam weights may then be computed by using the estimates of the channel matrix and covariance matrix in EQ. 13. ingly, in some cases, the computation of beam weights comprises estimating end—to—end forward gains (i.e., the values of the elements of the channel matrix wad) for each of the end—to—end forward multipath channels between an AN 515 and a reference location in a user beam coverage area. In other cases, computation of beam weights comprises estimating end—to—end forward gains for K x M end—to—end forward multipath channels from M ANs 515 to nce locations located within K user beam coverage areas.

The signal and interference components of the signal after rming are contained in the vector H Bfwd X (product of H, Bfwd, X). The signal and interference powers for each of the beams are contained in the K x K matrix H Bfwd. The power in the km al element of H Bfwd is the desired signal power intended for beam k. The sum of the magnitude squared of all elements in row k except the diagonal element is the interference power in beam k. Hence the C/I for beam k is: (g) lskkl z EQ. 15 where Skj are the elements of H B fwd. The uplink noise is contained in the vector AtE nu], 2 H H which has a K x K covariance matrix of 20,41At EE Att . The k’h al element of the covariance matrix ns the uplink noise power in beam k. The uplink signal to noise ratio for beam k is then computed as: S lekkl EQ. 16 ul k tkk where tkk is the km diagonal element of the uplink covariance matrix. The downlink noise is contained in the vector ndl, which has a covariance of 2 0-d1 IK . Hence the downlink signal to noise ratio is: s Is 12 2% EQ.17 Ndl k 26dl ] The end—to—end SINR is the combination of EQ. 15— EQ. 17: _1 —1 —1_ SINsz [2] EQ. 18 k +[L]Nul k +[L]Ndl k The above equations describe how to calculate the end—to—end SINR given the payload matrix E. The payload matrix may be constructed by intelligent choice of the gain and phases of each of the elements of E. The gain and phase of the al elements of E that optimize some utility metric (which is generally a function of the K beam SINR’s as computed above) may be selected and implemented by g the phase shifter 418 in each of the L transponders 411.

Candidate utility functions include, but are not limited to, sum of SINRk (total SINR), sum of Log(1+SINRk) (proportional to total throughput) or total power in the channel matrix, H. In some cases, the improvement in the y function by customizing the gains and phases is very small and ificant. This is sometimes the case when random or interleaved mappings of a elements are used. In some cases, the utility function can be improved by a non—trivial amount by custom selection of the receive/transmit signal gain and phase.

Distinct Coverage Areas Some es described above assume that the end-to-end relay 503 is designed to service a single coverage area shared by both the user terminals 517 and the ANs 515. For example, some cases describe a satellite having an antenna system that nates a satellite coverage area, and both the ANs and the user terminals are geographically distributed hout the satellite coverage area (e.g., as in ). The number of beams that can be formed in the satellite ge area, and the sizes (beam coverage areas) of those beams can be affected by aspects of the antenna system design, such as number and arrangement of antenna elements, reflector size, etc. For example, realizing a very large capacity can involve deploying a large number of ANs (e.g., hundreds) with sufficient g between the ANs to allow for end-to—end beamforming. For example, as noted above with reference to , increasing the number of ANs can se system capacity, although with diminishing returns as the number increases. When one antenna system supports both the user terminals and the ANs, achieving such a deployment with sufficient spacing between ANs can force a very wide geographical distribution of the ANs (e.g., across the entire satellite coverage area, as in ). Practically, achieving such a distribution may involve g ANs in undesirable locations, such as in areas with poor access to a high—speed network (e.g., a poor fiber infrastructure back to the CPS 505, one or more oceans, etc.), multiple legal jurisdictions, in ive and/or highly populated areas, etc. Accordingly, AN placement often involves various tradeoffs.

Some examples of the end—to—end relay 503 are designed with multiple antenna systems, thereby enabling te servicing of two or more distinct coverage areas from a single end—to—end relay 503. As bed below, the end—to—end relay 503 can include at least a first antenna system that services an AN ge area, and at least a second antenna system that services a user coverage area. Because the user al and AN coverage areas are serviced by different antenna systems, each antenna system can be designed to meet different design parameters, and each coverage area can be at least partially distinct (e.g., in geography, in beam size and/or density, in frequency band, etc.). For example, using such a multi—antenna system approach can enable user terminals distributed over a relatively large area geographic area (e.g., the entire United States) to be serviced by a large number of ANs distributed over a relatively small geographic area (e.g., a portion of the Eastern United States). For e, the AN coverage area can be a fraction (e.g., less than one half, less than one quarter, less than one fifth, less than one tenth) of the user coverage area in physical area. ] is an illustration of an example end-to-end beamforming system 3400. The system 3400 is an -end beamforming system that includes: a plurality of geographically distributed access nodes (ANs) 515; an end-to-end relay 3403); and a plurality of user terminals 517. The end-to-end relay 3403 can be an example of end-to-end relay 503 described herein.

The ANs 515 are geographically buted in an AN coverage area 3450, the user terminals 517 are geographically distributed in a user coverage area 3460. The AN coverage area 3450 and the user coverage area 3460 are both within a coverage area of the end—to—end relay 3403, but the AN coverage area 3450 is distinct from the user coverage area 3460. In other words, the AN area is not the same as the user coverage area, but rather has a substantial rivial) area (e.g., more than one—quarter, one—half, etc. of the AN coverage area) that does not overlap with the user coverage area. For example, in some cases, at least half of the user coverage area does not overlap the AN coverage area. As described above (e.g., in , the ANs 515 can provide signals through a distribution network 518 to a CPS 505 within a ground segment 502, and the CPS 505 can be connected to a data source.

The end—to—end relay 3403 includes a separate —link antenna subsystem 3410 and user—link antenna subsystem 3420. Each of the feeder—link antenna subsystem 3410 and the user—link antenna subsystem 3420 is capable of supporting end—to—end beamforming. For e, as bed below, each antenna subsystem can have its own array(s) of cooperating antenna elements, its own reflector(s), etc. The —link antenna subsystem 3410 can e an array of ating feeder—link tuent receive elements 3416 and an array of ating feeder—link constituent transmit elements 3419. The user—link antenna subsystem 3420 can include an array of cooperating user—link constituent e elements 3426 and an array of cooperating user—link constituent transmit ts he tuent elements are "cooperating" in the sense that the array of such constituent elements has characteristics making its respective antenna subsystem suitable for use in a beamforming . For example, a given user—link constituent receive element 3426 can receive a superposed composite of return uplink signals from multiple (e.g., all) user beam coverage areas 519 in a manner that contributes to forming of return user beams. A given user—link constituent it element 3429 can transmit a forward downlink signal in a manner that superposes with corresponding transmissions from other user—link constituent transmit elements 3429 to form some or all forward user beams. A given feeder—link constituent receive element 3416 can e a superposed ite of forward uplink signals from multiple (e.g., all) ANs 515 in a manner that contributes to forming of forward user beams (e.g., by inducing multipath at the end—to—end relay 3403). A given feeder-link constituent transmit element 3419 can transmit a return downlink signal in a manner that superposes with ponding transmissions from other feeder-link constituent transmit elements 3419 to contribute to forming of some or all return user beams (e.g., by enabling the ANs 515 to receive ite return s that can be beam- weighted to form the return user beams).

The example end-to-end relay 3403 includes a plurality of forward-link transponders 3430 and a plurality of return—link onders 3440. The transponders can be any le type of bent-pipe signal path between the antenna subsystems. Each forward—link transponder 3430 couples a respective one of the feeder-link tuent receive elements 3416 with a respective one of the user—link constituent transmit ts 3429. Each return-link transponder 3440 couples a respective one of the user-link constituent receive ts 3426 with a respective one of the feeder—link tuent transmit elements 3419. For example, some examples are described as having a one—to—one correspondence between each user—link constituent receive element 3426 and a respective feeder—link constituent transmit element 3419 (or vice versa), or that each user—link constituent receive element 3426 is coupled with "one and only one" feeder— link constituent transmit element 3419 (or vice versa), or the like. In some such cases, one side of each transponder is coupled with a single receive element, and the other side of the transponder is coupled with a single transmit element. In other such cases, one or both sides of a transponder can be selectively coupled (e.g., by a switch, or other means, as described below) with one of multiple elements. For example, the end—to—end relay 3403 can include one feeder— link a subsystem 3410 and two user—link antenna subsystems 3420; and each transponder can be coupled, on one side, to a single feeder—link element, and selectively coupled, on the other side, either to a single ink element of the first user—link antenna subsystem 3420 or to a single user—link element of the second user—link antenna subsystem 3420. In such ively coupled cases, each side of each transponder can still be considered at any given time (e.g., for a particular —related transaction) as being coupled with "one and only one" element, or the like.

For forward communications, transmissions from the ANs 515 can be received (via feeder uplinks 521) by the feeder—link constituent receive elements 3416, relayed by the forward—link transponders 3430 to the user—link constituent it ts 3429, and transmitted by the user—link tuent transmit elements 3429 to user terminals 517 in the user coverage area 3460 .For return ications, transmissions from the user terminals 517 can be received by user—link constituent receive elements, relayed by the retum—link transponders 3440 to the feeder—link constituent transmit elements 3419, and itted by the feeder-link constituent transmit elements 3419 to ANs 515 in the AN coverage area 3450 (via feeder downlink signals 527). The full signal path from an AN 515 to a user terminal 517 via the end- to-end relay 3403 is referred to as the end-to-end forward link 501; and the full signal path from a user terminal 517 to an AN 515 via the -end relay 3403 is referred to as the end-to-end return link 523. As described herein, the end-to-end forward link 501 and the end-to-end return link 523 can each include multiple ath ls for forward and return communications.

In some cases, each of the plurality of geographically distributed access nodes (e.g., the ANs 515) has an end—to—end beam-weighted forward uplink signal output. The end—to—end relay (e. g., the end—to—end relay 3403) comprises an array of cooperating feeder—link tuent receive elements 3416 in wireless ication with the distributed access nodes, an array of cooperating user—link constituent it elements 3419 in wireless communication with the plurality of user terminals 517, and a plurality of forward—link transponders 3430. The forward— link transponders 3430 are "bent—pipe" (or non—processing) transponders, so that each transponder outputs a signal that corresponds to the signal it receives with little or no processing.

For example, each forward—link transponder 3430 can amplify and/or frequency translate its received signal, but does not perform more complex processing (e.g., there is no demodulation and/or modulation, no on—board beamforming, etc.). In some cases, each forward—link transponder 3430 accepts an input at a first frequency band (e.g., 30 GHz LHCP) and outputs at a second frequency band (e.g., 20 GHZ RHCP), and each return—link transponder 3440 s an input at the first frequency band (e.g., 30 GHz RHCP) and outputs at the second ncy band (e.g., 20 GHz LHCP). Any suitable combination of frequency and/or polarization can be used, and the user—link and feeder—link can use the same or different frequency ranges. Each forward—link transponder 3430 is coupled between a respective one of the feeder—link tuent receive elements 3416 and a respective one of the user—link constituent it elements 3419 (e.g., with a —one correspondence). The forward—link transponders 3430 convert superpositions of a plurality of beam—weighted forward uplink signals via the feeder—link constituent receive elements 3416 into forward downlink s (e.g., composite input forward signals). Transmission of the forward downlink signals by the user—link constituent transmit elements 3429 contributes to forming a forward user beam ing at least some of the plurality of user terminals 517. As described , the forward uplink signals can be end—to— end beam—weighted and onized (e.g., phase—synchronized, and, if d, time— synchronized) prior to transmission from the ANs 515, which can enable the desired superposition of those signals at the feeder—link constituent receive ts 3416.

The transmission contributes to forming the forward user beam in the sense that the beamforming is end—to-end, as described herein; the beamforming is a result of multiple steps, including ing and applying appropriate weights to the forward uplink s prior to transmission to the relay from the ANs 515, inducing multipath by the multiple forward-link transponder 3430 of the end-to-end relay 3403, and transmitting the forward downlink signals with a user-link array antenna. Still, for the sake of simplicity, some descriptions can refer to the forward beam as being formed by superposition of the transmitted forward downlink signals.

In some cases, each of the plurality of user terminals 517 is in wireless communication with the array of cooperating user—link constituent transmit elements 3429 to receive a composite (e.g., a superposition) of the transmitted forward downlink signals.

In some cases, the end-to-end relay 3403 further includes an array of user—link constituent e elements 3426 in wireless communication with the user terminals 517, an array of cooperating feeder—link constituent it elements 3419 in wireless communication with the buted ANs 515, and a plurality of return—link transponders 3440. The retum—link transponders 3440 can be similar or identical to the forward—link transponder 3430 (e.g., bent— pipe transponders), except that each is d between a respective one of the user—link constituent receive elements 3426 and a respective one of the feeder—link constituent transmit elements 3419. Receipt of return uplink signals via the array of ating user—link tuent receive element 3426 forms return downlink signals in the link transponders 3440. In some cases, each return downlink signal is a respective superposition of return uplink signals received by a user—link tuent receive element 3426 from multiple user terminals 517 (e.g., from multiple user beam coverage areas 519). In some such cases, each of the 2016/026815 plurality of user terminals is in wireless communication with the array of cooperating user—link constituent receive elements 3426 to transmit a respective return uplink signal to multiple of the user—link constituent receive elements 3426.

In some cases, the return downlink signals are transmitted by the feeder—link constituent transmit element 3419 to the geographically distributed ANs 515. As described herein, each AN 515 can receive a superposed composite of the return downlink signals transmitted from the feeder—link constituent transmit elements 3419 (i.e., which pond to the return downlink signals). The received return downlink signals red to as composite receive signals) can be coupled to a return beamformer, which can combine, synchronize, beam weight, and perform any other suitable processing. For example, the return rmer can weight the received superpositions 1706 of the return downlink signals (i.e., apply return beam weights to the composite return signals) prior to combining the signals. The return beamformer can also synchronize the composite return signals prior to combining the signals to account at least for respective path delay differences between the end-to—end relay 3403 and the ANs 515.

In some cases, the synchronizing can be according to a ed beacon signal (received by one or more, or all, of the ANs 515).

Because of the end-to-end nature of the beamforming, proper application of return beam weights by the return beamformer enables formation of the return user beams, even though the return rmer may be coupled to the feeder—link side of the end-to—end multipath channels, and the user beams may be formed at the ink side of the end—to—end multipath channels. Accordingly, the return beamformer can be referred to as contributing to the g of the return user beams (a number of other s of the system 3400 also contribute to the end-to-end return beamforming, such as the inducement of multipath by the retum—link transponders 3440 of the end—to—end relay 3403). Still, the return beamformer can be referred to as forming the return user beams for the sake of simplicity.

] In some cases, the end-to—end relay 3403 further includes a feeder—link antenna subsystem 3410 to illuminate an access node coverage area (AN coverage area 3450) within which the ity of distributed access nodes is located. The feeder—link antenna subsystem 3410 ses the array of cooperating feeder—link constituent receive elements 3416. In some cases, the —end relay 3403 also includes a user—link antenna subsystem 3420 to illuminate a user coverage area 3460 within which the plurality of user terminals 517 is geographically distributed (e.g., in a plurality of user beam coverage areas 519). The user—link antenna tem 3420 comprises the array of cooperating user—link constituent transmit elements 3429. In some cases, the user—link antenna subsystem 3420 includes a user—link receive array and a user—link transmit array (e.g., separate, half—duplex arrays of ating user—link constituent elements). The user—link receive array and the user—link transmit array can be spatially interleaved (e.g., to point to a same reflector), spatially separated (e.g., to point at receive and transmit reflectors, respectively), or arranged in any other le manner. In other cases, the user—link antenna subsystem 3420 includes full—duplex ts (e.g., each user—link constituent it element 3429 shares ing structure with a respective ink constituent receive element 3426). Similarly, in some cases, the feeder—link antenna subsystem 3410 includes a feeder—link receive array and a —link transmit array, which may be spatially related in any suitable manner and may directly radiate, point to a single reflector, point to separate transmit and receive ors, etc. In other cases, the feeder—link antenna subsystem 3410 includes full—duplex elements. The feeder—link antenna subsystem 3410 and the user—link antenna subsystem 3420 can have the same or different aperture sizes. In some cases, the feeder—link antenna subsystem 3410 and the user—link antenna subsystem 3420 operate in a same frequency band (e.g., Ka band, etc.). In some cases, the feeder—link a subsystem 3410 and the user-link antenna subsystem 3420 operate in different frequency bands (e.g., feeder-link uses V band, the user-link uses Ka band, etc.).

In examples, such as those illustrated by , the AN ge area 3450 is distinct from the user coverage area 3460. The AN coverage area 3450 can be a single, contiguous ge area, or multiple disjoint coverage areas. Similarly (and independently of whether the AN coverage area is single or multiple), the user coverage area 3460 can be a single, contiguous coverage area, or multiple disjoint coverage areas. In some cases, the AN coverage area 3450 is a subset of the user coverage area 3460. In some cases, at least half of the user coverage area 3460 does not overlap the AN coverage area 3450. As described below, in some cases, the —link antenna subsystem 3410 further comprises one or more feeder—link reflectors, and the user—link antenna subsystem 3420 further comprises one or more user—link reflectors. In some cases, the feeder—link reflector is significantly larger (e.g., at least twice the physical area, at least five times, ten times, fifty times, eighty times, etc.) than the user—link reflector. In some cases, the feeder—link reflector is approximately the same physical area (e.g., within 5%, 10%, 25%) as the user—link reflector.

In some cases, the system 3400 operates in context of ground network ons, as described with reference to For example, the end—to—end relay 3403 communicates with ANs 515, which communicate with a CPS 505 via a distribution k 518. In some cases, the CPS 505 es a forward beamformer 529 and/or a return beamformer 531, for example, as described with nce to . As described above, the forward beamformer 529 can participate in g forward end—to—end beams by applying computed d beam weights (e.g., supplied by a forward beam weight generator 918) to forward—link signals; and the return beamformer 531 can participate in forming return end—to—end beams by applying ed return beam weights (e.g., supplied by a return beam weight generator 935) to return—link signals. As described above, the end—to—end forward beam weights and/or the set of end—to—end return beam weights can be computed according to ted end—to—end gains for end—to—end multipath channels, each —end multipath channel communicatively coupling a respective one of the distributed ANs 515 with a respective location in the user coverage area (e.g., a user terminal 517 or any suitable nce location) via a respective plurality of the forward—link bent—pipe onders 3430 and/or via a respective plurality of the retum—link ipe transponders 3440. In some cases, though not shown, the end—to—end relay 3403 includes a beacon signal transmitter. The beacon signal transmitter can be implemented as described above with reference to the beacon signal generator and calibration support module 424 of . In some cases, the generated beacon signal can be used so that the plurality of distributed ANS 515 is in time-synchronized wireless communication with the end-to-end relay 3403 (e. g., with the ity of feeder-link constituent receive elements 3416 according to the beacon signal).

In some cases, the system 3400 includes a system for forming a plurality of forward user beams using end-to-end beamforming. Such cases include means for transmitting a plurality of forward uplink signals from a plurality of geographically distributed locations, wherein the plurality of forward uplink s is formed from a weighted combination of a plurality of user beam signals, and wherein each user beam signal corresponds to one and only one user beam. For example, the plurality of geographically distributed locations can include a plurality of ANs 515, and the means for transmitting the plurality of forward uplink signals can include some or all of a forward rmer 529, a distribution network 518, and the geographically distributed ANs 515 (in communication with the end-to—end relay 3403). Such cases can also include means for relaying the plurality of forward uplink signals to form a plurality of d downlink signals. Each forward nk signal is created by amplifying a unique superposition of the plurality of forward uplink signals, and the ity of forward downlink signals superpose to form the plurality of user beams, wherein each user beam signal is dominant within the ponding user beam. For example, the means for relaying the plurality of forward uplink signals to form the plurality of forward downlink signals can e the end—to—end relay 3403 (in communication with one or more user terminals in user beam coverage areas 519) with its collocated ity of signal paths, which can include d—link transponders 3430 and retum—link transponders 3440.

Some such cases include first means for receiving a first osition of the plurality of forward downlink s and recovering a first one of the plurality of user beam signals. Such first means can include a user terminal 517 (e.g., including a user terminal antenna, and a modem or other components for recovering user beam signals from the forward nk signals). Some such cases also include second means (e.g., including a second user terminal 517) for receiving a second superposition of the plurality of forward downlink signals and recovering a second one of the plurality of user beam signals. For example, the first means for receiving is located within a first user beam, and the second means for receiving is located within a second user beam. is an illustration of an example model of signal paths for signals carrying return data on the end—to—end return link 523. The example model can operate similarly to the model described with reference to FIGS. 6 — 8, except that the end—to—end relay 3403 includes return—link signal paths 3502 dedicated for return—link ications. Each —link signal path 3502 can include a retum—link transponder 3440 coupled between a user—link constituent receive element 3426 and a feeder-link constituent transmit element 3419. Signals originating with user terminals 517 in K user beam ge areas 519 are transmitted (as return uplink signals 525) up to the end-to-end relay 3403, ed by an array of L return-link signal paths 3502, communicated through L return-link transponders 3440 to L corresponding feeder-link tuent transmit elements 3419, and transmitted by each of the L -link constituent transmit ts 3419 to some or all of the M ANs 515 (similar to what is shown in . In this way, the multiple return—link signal paths 3502 (e.g., the return-link transponders 3440) induce multipath in the return—link communications. For example, the output of each retum—link signal path 3502 is a return downlink signal 527 corresponding to a respective one of the received return uplink signals 525 (e.g., corresponding to a received ite of the return uplink signals 525 transmitted from multiple of the user beam coverage areas 519) and is itted over the return downlink signals 527 to the M ANs 515 (e.g., geographically distributed over an AN coverage area 3450). As described above, there are L (or up to L) different ways for a signal to get from a user terminal 517 located in a user beam coverage area 519 to a particular AN 515. The end—to—end relay 3403 thereby creates L paths n a user terminal 517 and an AN 515, referred to collectively as an end—to—end return multipath channel 1908 (e.g., similar to .

The end—to—end return multipath channels can be modeled in the same manner bed above. For example, Ar is the L x K return uplink radiation matrix, Ct is the M x L return downlink radiation matrix, and Eret is the L x L return payload matrix for the paths from WO 09332 the user—link constituent receive elements 3426 to the feeder—link tuent transmit elements 3419. As described above, the end—to—end return multipath channel from a user al 517 in a particular user beam coverage area 519 to a particular AN 515 is the net effect of the L different signal paths induced by L unique retum—link signal paths 3502 through the end—to—end relay 3403. With K user beam coverage areas 519 and M ANS 515, there can be M x K induced end—to—end return ath channels in the end—to—end return link 523 (via the end—to—end relay 3403), and each can be dually modeled to e a corresponding element of an M x K return channel matrix Hret (Ct x Eret x Ar). As noted above (e.g., with reference to FIGS. 6 — 8), not all ANs 515, user beam coverage areas 519, and/or retum—link transponders 3440 have to participate in the end—to—end return multipath channels. In some cases, the number of user beams K is greater than the number of transponders L in the signal path of the end—to—end return multipath channel; and/or the number of ANs M is greater than the number of transponders L in the signal path of the end—to—end return multipath channel. As described with reference to the CPS 505 can enable forming of return user beams by applying return beam weights to the ed downlink return signals 527 (the received signals, after reception by the AN are referred to as composite return signals 907, as explained further below). The return beam weights can be computed based on the model of the M x K signal paths for each end-to-end return multipath channel that couples the user terminals 517 in one user beam ge area 519 with one of the plurality of ANs 515. is an ration of an example model of signal paths for signals carrying forward data on the —end d link 501. The example model can operate rly to the model bed with reference to FIGS. 9 — 11, except that the end—to—end relay 3403 includes forward—link signal paths 3602 dedicated for forward—link ications. Each forward—link signal path 3602 can include a forward—link transponder 3430 coupled between a feeder—link constituent receive element 3416 and a user—link constituent transmit element 3429.

As described above, each forward uplink signal 521 is beam weighted (e.g., at a forward beamformer 515 in the CPS 505 of the ground segment 502) prior to transmission from an AN 515. Each AN 515 receives a unique forward uplink signal 521 and transmits the unique forward uplink signal 521 via one ofM uplinks (e.g., in a time—synchronized manner). The forward uplink signals 521 are received from geographically distributed locations (e.g., from the ANs 515) by some or all of the forward—link transponders 3430 in a superposed manner that creates composite input forward signals 545. Each forward—link transponder 3430 concurrently receives a composite input forward signal 545, though with slightly different timing due to differences in the locations of each receiving feeder—link constituent receive element 3416 associated with each forward—link transponder 3430. For example, even though each feeder—link constituent receive element 3416 can e a composite of the same plurality of forward uplink s 521, the received composite input forward signals 545 can be slightly different. The composite input forward signals 545 are received by L forward—link onders 3430 via respective —link constituent receive elements 3416, communicated h the L forward— link transponders 3430 to L corresponding user—link constituent transmit elements 3429, and transmitted by the L user—link constituent transmit elements 3429 to one or more of the K user beam coverage areas 519 (e.g., as forward nk signal 522, each corresponding to a respective one of the received composite input forward signals 521). In this way, the multiple forward—link signal paths 3602 (e.g., forward—link transponders 3430) induce multipath in the d—link communications. As described above, there are L different ways for a signal to get from an AN 515 to a ular user terminal 517 in a user beam coverage area 519. The end—to— end relay 3403 thereby induces multiple (e.g., up to L) signal paths between one AN 515 and one user terminal 517 (or one user beam ge area 519), which may be referred to collectively as an end—to—end forward multipath channel 2208 (e.g., similar to ).

The end-to-end forward multipath channels 2208 can be modeled in the same manner described above. For example, Cr is the L x M forward uplink radiation matrix, At is the K x L forward downlink radiation matrix, and Efwd is the L x L forward payload matrix for the paths from the feeder-link constituent receive elements 3416 to the user-link constituent it elements 3429. In some cases, the forward payload matrix Efwd and return payload matrix Eret may be different to reflect ences n the forward—link signal paths 3602 and the return—link signal paths 3502. As described above, the end-to-end forward multipath channel from a particular AN 515 to a user terminal 517 in a particular user beam ge area 519 is the net effect of the L different signal paths induced by L unique forward—link signal paths 3602 through the end—to—end relay 3403. With K user beam coverage areas 519 and M ANs 515, there can be M x K induced end—to-end forward ath channels in the end—to—end forward link 501, and each can be individually modeled to compute a corresponding element of an M x K forward channel matrix Hde (At x Efwd x Cr). As noted with reference to the return direction, not all ANs 515, user beam ge areas 519, and/or forward—link transponders 3430 have to participate in the end—to—end forward multipath ls. In some cases, the number of user beams K is greater than the number of transponders L in the signal path of the end—to—end forward multipath channel; and/or the number of ANs M is greater than the number of transponders L in the signal path of the end—to—end forward multipath channel. As described with reference to an appropriate beam weight may be computed for each of the plurality of end—to—end forward multipath ls by the CPS 505 to form the forward user beams.

Using multiple transmitters (ANS 515) to a single receiver (user terminal 517) can provide transmit path diversity to enable the successful transmission of information to any user terminal 517 in the presence of the intentionally d ath channel.

FIGS. 41 — 43 describe end—to—end relays 3403 implemented with separate forward— link transponders 3430 and return—link onders 3440. FIGS. 44A and 44B show an illustration of an e forward signal path 3700 (like the forward signal path 3602 of ) and return signal path 3750 (like the return signal path 3502 of ), respectively. As described above, the forward signal path 3700 includes a forward—link transponder 3430 coupled between a —link constituent receive element 3416 and a user—link constituent transmit element 3429. The return signal path 3750 es a return—link transponder 3440 coupled between a user—link constituent receive element 3426 and a feeder—link constituent transmit element 3419. In some cases, each forward—link transponder 3430 and each return—link transponder 3440 is a cross-pole transponder. For example, the forward—link transponder 3430 receives a d uplink signal at an uplink frequency with left-hand circular polarization (LHCP) and outputs a forward downlink signal at a downlink frequency with right-hand circular zation (RHCP); and each retum-link onder 3440 receives a return uplink signal at the uplink ncy with right-hand circular zation (RHCP) and outputs a return downlink signal at the downlink frequency with left-hand ar polarization (LHCP). One such case (i.e., following the polarizations described in the preceding example) is illustrated by ing only the solid lines of FIGS. 44A and 44B, and another such case (i.e., following opposite polarizations from those described in the preceding example) is illustrated by following only the dashed lines of FIGS. 44A and 44B. In other cases, some or all transponders can provide a dual—pole signal path pair. For example, following both the solid and dashed lines of FIGS. 44A and 44B, the forward—link transponders 3430 and the retum—link transponders 3440 can receive forward uplink signals at the same or different uplink ncy with both polarizations (LHCP and RHCP) and can both output forward downlink signals at the same or different downlink frequency with both polarizations (RHCP and LHCP). For example, such cases can enable multiple systems to operate in parallel using any suitable type of interference mitigation techniques (e.g., using time division, frequency on, etc.). In some cases, the end—to—end relay 3403 es a large number of transponders, such as 512 forward—link transponders 3430 and 512 retum—link transponders 3440 (e.g., 1,024 transponders total). Other implementations can include smaller numbers of transponders, such as 10, or any other suitable number. In some cases, the antenna elements are implemented as full—duplex structures, so that WO 09332 each receive antenna element shares structure with a respective transmit antenna element. For example, each illustrated antenna element can be implemented as two of four waveguide ports of a radiating structure adapted for both transmission and receipt of signals. In some cases, only the feeder—link ts, or only the user—link elements, are full duplex. Other implementations can use different types of polarization. For example, in some implementations, the transponders can be coupled between a receive antenna element and transmit antenna element of the same polarity.

Both the example forward—link transponder 3430 and return—link transponder 3440 can include low noise amplifiers (LNAs) 3705, frequency converters and ated filters 3710, l amplifiers 3715, phase shifters 3720, power amplifiers 3725 (e.g., traveling wave tube amplifiers ), solid state power amplifiers (SSPAs), etc.) and harmonic s 3730. In dual—pole implementations, as shown, each pole has its own signal path with its own set of transponder components. Some implementations can have more or fewer components. For example, the ncy converters and associated s 3710 can be useful in cases where the uplink and downlink frequencies are different. As one example, each d—link transponder 3430 can accept an input at a first frequency band and can output at a second frequency band; and each return-link transponder 3440 can accept an input at the first frequency band and can output at the second frequency band.

In some cases, multiple sub-bands are used (e.g., seven 500 MHz sub—bands, as described above For e, in some cases, transponders can be provided that operate over the same sub—bands as used in a multiple sub—band implementation of the ground network, effectively to enable multiple independent and parallel end—to—end beamforming systems through a single —end relay (each end-to-end beamforming system operating in a different sub— band). In other cases, a wide bandwidth end—to—end beamforming system may use multiple sub— bands in the ground network, but pass one or more (or all) sub—bands through wideband transponders (e.g., g 7 sub—bands, each 500 MHz wide, through a 3.5 GHz bandwidth transponders). In such cases, each transponder can include le frequency ters and associated filters 3710, and/or other components, dedicated to handling one or more of the sub— bands. ). The use of multiple frequency sub—bands may allow relaxed requirements on the amplitude and phase response of the transponder, as the ground network may separately determine beam weights used in each of the sub—bands, ively calibrating out passband amplitude and phase variation of the transponders. For example, with te forward and return transponders, and using 7 sub—bands, a total 14 different beam weights may be used for each beam, i.e. 7 sub—bands * 2 directions (forward and return). In some cases, each transponder path includes only a LNA 3705, a l amplifier 3715, and a power amplifier 3725. Some implementations of the end—to—end relay 3403 include phase shift controllers and/or other controllers that can individually set the phases and/or other characteristics of each transponder as bed above.

The antenna ts can transmit and/or receive s in any suitable manner. In some cases, the end—to—end relay 3403 has one or more array fed reflectors. For example, the feeder—link antenna subsystem 3410 can have a feeder—link reflector for both transmit and receive, or a separate feeder—link transmit reflector and feeder—link receive reflector. Similarly, the user—link antenna subsystem 3420 can have a user—link reflector for both transmit and receive, or a separate user—link transmit reflector and user—link receive reflector. In one example case, the feeder—link antenna subsystem 3410 comprises an array of radiating ures, and each radiating structure includes a feeder—link constituent receive element 3416 and a feeder—link constituent transmit element 3419. In such a case, the feeder—link antenna tem 3410 can also include a feeder—link reflector that illuminates the —link constituent receive elements 3416 and is illuminated by the feeder—link constituent transmit elements 3419. In some cases, the reflector is implemented as le reflectors, which may be of ent shapes, sizes, orientations, etc. In other cases, the feeder-link antenna subsystem 3410 and/or the user-link antenna subsystem 3420 is implemented without reflectors, for example, as a direct radiating array.

As noted above, ting the feeder—link antenna subsystem 3410 and the user—link antenna subsystem 3420 can enable servicing of one or more AN coverage areas 3450 that are ct from one or more user coverage areas 3460. For example, the feeder—link antenna subsystem 3410 can be implemented with a reflector having an appreciably larger physical area than the or of the user coverage area 3460. The larger reflector can permit a large number of ANs 515 to be geographically distributed in an appreciably smaller AN ge area 3450, such as in a small subset of the user coverage area 3460. Some examples are shown in FIGS. 45 and 46. shows an example of an —end relay 3403 (e.g., a satellite) visible Earth coverage area 3800. In the example end—to—end relay 3403, the feeder—link antenna subsystem 3410 includes a 18—meter feeder—link reflector, and the user—link a subsystem 3420 includes a 2—meter user—link reflector (e.g., the feeder—link reflector area is about eighty times larger than the ink reflector area). Each antenna subsystem also includes an array of 512 ating constituent receive/transmit elements. For example, the example end—to—end relay 3403 can include 512 forward—link transponders 3430 (e.g., forming 512 forward signal WO 09332 paths 3700 as shown in A) and 512 return—link transponders 3440 (e.g., forming 512 return signal paths 3750 as shown in B). The user coverage area 3460 includes 625 user beam coverage areas 519. The small shaded region in the eastern United States is an AN coverage area 3450 having 597 ANs 515 distributed therein. The AN coverage area 3450 is a small subset of the large user coverage area 3460, but still es large capacity with a large number of ANs 515. Such a relatively dense AN coverage area 3450 is referred to herein as a "AN farm." shows an example of an end—to—end relay 3403 (e.g., a satellite) Continental United States (CONUS) coverage area 3900. The example end—to—end relay 3403 is similar to the example shown in , except that the user—link antenna subsystem 3420 includes a 5— meter user—link reflector (e.g., the feeder—link reflector is about four times larger than the user— link reflector). The user ge area 3460 includes 523 user beam coverage areas 519. The AN ge area 3450 (the AN farm) is the same as that of : a region in the eastern United States having 597 ANs 515 distributed therein, which is a small subset of the user ge area 3460.

Multiple Coverage Areas In the example end-to-end relays 3403 described above, the user-link antenna subsystem 3420 is described as a single antenna subsystem (e.g., with a single user—link reflector), and the feeder—link antenna subsystem 3410 is described as a single antenna subsystem (e.g., with a single feeder-link reflector). In some cases, the user—link antenna subsystem 3420 can include one or more antenna subsystems (e.g., two or more sub—arrays of constituent antenna elements) associated with one or more ink reflectors, and the feeder— link antenna subsystem 3410 can include one or more antenna subsystems associated with one or more —link reflectors. For example, some end—to—end relays 3403 can have a user—link a subsystem 3420 that includes a first set of ink constituent receive/transmit elements associated with a first ink reflector (e.g., each element is arranged to illuminate, and/or be illuminated by, the first user—link reflector) and a second set of user—link constituent receive/transmit elements associated with a second ink reflector. In some cases, the two user—link reflectors are approximately the same physical area (e.g., within 5%, 10%, 25%, 50%) of each other. In some cases, one user—link reflector is significantly larger (e.g., at least twice the physical area) than the other. Each set of the user—link constituent receive/transmit ts, and its associated user—link reflector, can illuminate a corresponding, distinct user coverage area.

For example, the multiple user coverage areas can be non—overlapping, partially overlapping, fully overlapping (e.g., a smaller user coverage could be contained within a larger user coverage area), etc. In some cases, the multiple user ge areas can be active (illuminated) at the same time. Other cases, as described below, can enable ive activation of the ent portions of user—link constituent e/transmit elements, thereby activating different user coverage areas at different times. Switching between multiple coverage areas may be coordinated with the CPS. For example, beamforming calibration, beam weight calculation and beam weight application may occur in two parallel rmers, one for each of two ent coverage areas. The usage of appropriate weights in the beamformers can be timed to pond to the operation of the end—to—end relay. Beam weights might in some cases change during a time slice if time—slice beamformers are employed.

FIGS. 47A and 47B show an example forward signal path 4000 and return signal path 4050, respectively, each having selective activation of multiple user—link antenna subsystems 3420. Each signal path has a onder coupled between constituent antenna elements. Turning first to A, the forward-link transponder 3430 is similar to the one described with reference to A, except that the output side of the forward-link transponder 3430 is selectively coupled to one of two user-link constituent transmit elements 3429, each part of a separate user-link antenna subsystem 3420 (e.g., each part of a separate array of cooperating user-link constituent transmit elements 3429). As described above, the d-link transponder 3430 can include some or all of LNAs 3705, frequency converters and associated filters 3710, channel amplifiers 3715, phase rs 3720, power amplifiers 3725, and harmonic filters 3730.

The forward—link transponder 3430 of A further includes switches 4010 rd—link es) that selectively couple the transponder either to a first user—link constituent transmit element 3429a (of a first user—link antenna tem 3420) via a first set of power amplifiers 3725 and harmonic filters 3730, or to a second user—link constituent transmit element 3429b (of a second user—link a subsystem 3420) via a second set of power amplifiers 3725 and harmonic filters 3730. For example, in a first switch mode, the forward— link transponder 3430 effectively forms a signal path between a feeder—link constituent receive element 3416 and a first user—link constituent transmit element 3429; and in a second switch mode, the forward—link onder 3430 effectively forms a signal path between the same feeder—link constituent receive element 3416 and a second user—link constituent transmit element 3429. The es 4010 can be ented using any suitable switching means, such as an electromechanical switch, a relay, a transistor, etc. Though shown as switches 4010, other implementations can use any other suitable means for selectively coupling the input of the forward—link transponder 3430 to multiple outputs. For example, the power amplifiers 3725 can be used as switches (e.g., ing high gain when "on," and zero gain (or loss) when "off").

Turning to B, the retum—link transponder 3440 functionally mirrors the forward—link transponder 3430 of A. Rather than selectively coupling the output side of the transponder, as in the forward—link case of A, the input side of the return—link transponder 3440 of B is selectively coupled to one of two user—link constituent receive elements 3426. Again, each ink constituent receive element 3426 can be part of a separate user—link antenna subsystem 3420 (e.g., each part of a separate array of cooperating user—link constituent receive elements 3426). As described above (e.g., in B), the return—link transponder 3440 can e some or all of LNAs 3705, frequency converters and associated filters 3710, l amplifiers 3715, phase shifters 3720, power amplifiers 3725, and ic filters 3730.

The return—link transponder 3440 of B further includes switches 4010 (return—link switches) that selectively couple the transponder either to a first user—link constituent receive element 3426 (of a first user-link antenna subsystem 3420) via a first set of LNAs 3705, or to a second user-link constituent receive element 3426 (of a second user-link antenna tem 3420) via a second set of LNAs 3705. For example, in a first switch mode, the return-link transponder 3440 effectively forms a signal path between a first user-link constituent receive element 3426 and a feeder—link constituent transmit element 3419; and in a second switch mode, the return—link transponder 3440 ively forms a signal path between a second user—link constituent receive element 3426 and the same feeder—link constituent transmit element 3419. The switches 4010 can be implemented using any suitable switching means, such as an electromechanical switch, a relay, a transistor, etc. Though shown as switches 4010, other implementations can use any other suitable means for selectively coupling the input of the d—link transponder 3430 to multiple outputs. For example, the amplifiers 3705 can be used as switches (e.g., providing high gain when "on," and zero gain (or loss) when "of ").

Examples of the end—to—end relay 3403 can e a switch controller 4070 to selectively switch some or all of the switches 4010 (or other suitable selective coupling means) ing to a switching schedule. For example, the switching le can be stored in a storage device on—board the end—to—end relay 3403. In some cases, the switching schedule ively selects which user—link antenna subsystem 3420 to activate (e.g., which set of user beams to illuminate) in each of a ity of time intervals (e.g., ots). In some cases, the switching allocates equal time to the multiple user—link a subsystems 3420 (e.g., each of two tems is activated for about half the time). In other cases, the switching can be used to WO 09332 e capacity—sharing goals. For example, one user—link antenna subsystem 3420 can be associated with higher—demand users and can be allocated a r portion of time in the schedule, while another user—link antenna subsystem 3420 can be associated with lower—demand users and can be allocated a smaller portion of time in the schedule.

FIGS. 48A and 48B show an e of end—to—end relay 3403 coverage areas 4100 and 4150 that include multiple, selectively activated user coverage areas 3460a, 3460b. The example end—to—end relay 503 is similar to the ones in FIGS. 38 and 39 except for different antenna subsystems. In this example, the user—link antenna subsystem 3420 includes two 9— meter user—link reflectors, and the transponders are configured to selectively activate only half of the user beams at any given time (e.g., the transponders are implemented as in FIGS. 47A and 47B). For example, during a first time al, as shown in A, the user ge area 3460a es five hundred ninety active user beam coverage areas 519. The active user beam coverage areas 519 effectively cover the western half of the United States. The AN coverage area 3450 (the AN farm) is the same as that of FIGS. 38 and 39: a region in the eastern United States having 597 ANs 515 distributed therein. During the first time interval, the AN coverage area 3450 does not p with the active user coverage area 3460. During a second time interval, as shown in B, the user coverage area 3460b includes another five hundred ninety active user beam coverage areas 519. The active user beam coverage areas 519 in the second time interval effectively cover the n half of the United States. The AN coverage area 3450 does not change. However, during the second time interval, the AN coverage area 3450 is fully overlapped by (is a subset of) the active user coverage area 3460. Capacity may be y allocated to various regions (e.g., between eastern and western user coverage areas) by dynamically adjusting the ratio of time allocated to the corresponding user—link antenna sub— systems.

While the previous example illustrates two similarly sized user coverage areas, other numbers of user ge areas can be provided (e.g., three or more) and can be of differing sizes (e.g., earth coverage, continental US. only, US. only, al only, etc.). In cases with multiple user coverage areas 3460, the user coverage areas 3460 can have any suitable geographic relationship. In some cases, first and second user coverage areas 3460 partially overlap (e.g., as shown in FIGS. 48A and 48B). In other cases, a second user coverage area 3460 can be a subset of a first user coverage area 3460 (e.g., as shown in FIGS. 45 and 46). In other cases, the first and second user coverage areas do not overlap (e.g., are disjoint).

FIGS. 47A — 47B describe signal path selection on the user—link side. r, some cases alternatively or onally include signal path switching on the feeder—link side.

WO 09332 shows an example forward signal path 4200 having ive tion of le user— link antenna subsystems 3420 and multiple feeder—link antenna subsystems 3410. The signal path has a forward—link transponder 3430 d between constituent antenna elements. As described above, the forward—link transponder 3430 can include some or all of LNAs 3705, frequency converters and associated filters 3710, channel amplifiers 3715, phase rs 3720, power amplifiers 3725, and harmonic filters 3730. The input side of the forward—link onder 3430 is selectively coupled to one of two feeder—link constituent receive elements 3416 (e.g., using switches 4010a and 4010b, or any other suitable path selection means). Each feeder—link constituent receive element 3416 can be part of a separate feeder—link a subsystem 3410 (e.g., each part of a separate array of cooperating feeder—link constituent receive elements 3416). The output side of the forward—link transponder 3430 is selectively coupled to one of two user—link constituent transmit elements 3429 (e.g., using switches 4010c and 4010d, or any other suitable path selection . Each user—link constituent transmit element 3429 can be part of a separate user-link antenna subsystem 3420 (e.g., each part of a separate array of cooperating user-link constituent transmit elements 3429). One or more switching llers (not shown) can be included in the end-to-end relay 3403 for selecting between some or all of the four possible signal paths enabled by the forward-link transponder 3430. The transponders of FIGS. 47A, 47B, and 49 are intended only to illustrate a few of many possible cases. Further, some cases can include path selection n more than two user-link antenna subsystems 3420 and/or more than two feeder—link a subsystem 3410. Similarly, onal path selection can be included in cases where the user—link antenna subsystem 3420 and/or the feeder- link constituent receive element 3416 has separate transmit and e reflectors, or the like.

] In a similar manner, multiple AN coverage areas may also be provided. As one example, it can be desirable for traffic of particular geographic regions to terminate in their respective regions. For example, an end—to—end relay 3403 with or without paired transponders like those rated in can service a first AN coverage area 3450 and a first user coverage area 3460, both in North America, and a second AN coverage area 3450 and a second user coverage area 3460, both in South America. Using path selection (e.g., switching) in the transponders, a single end—to—end relay 3403 (e. g., a single satellite) can service traffic associated with the North American user coverage area 3460 using ANs 515 in the North American AN ge area 3450, and e traffic associated with the South American user coverage area 3460 using ANs 515 in the South American AN coverage area 3450. Capacity may be flexibly allocated to various regions (e.g., between North and South American user coverage areas) by dynamically adjusting the ratio of time allocated to the corresponding antenna sub—systems.

In general, features of the end—to—end relay 3403 described in enable servicing of at least one user beam coverage area 5 19 distinct from at least one AN ge area 3450. In some cases, the distinct coverage area servicing can enable use of AN farms to provide high capacity to a large user coverage area 3460. FIGS. 45, 46, 48A, and 48B show various examples of such AN farm implementations. Deploying large s of ANs 515 in a relatively small geographic area can provide a number of features. For example, it can be easier to ensure that more (or even all) of the ANs 5 15 are ed closer to a peed network (e.g., in a region with good fiber connectivity back to the CPS 505), within borders of a single country or region, on land, etc., with less deviation from an ideal AN 5 15 bution.

Implementing distinct coverage area servicing with path selection (e.g., as in FIGS. 47A — 47B can provide additional features. For example, as described above, a single AN farm (and a single end—to—end relay 3403) can be used to selectively e multiple user coverage areas 3460. Similarly, a single end-to—end relay 3403 can be used to distinguish and service c by region.

] In some cases, the distinct coverage area servicing with path selection can enable s interference management and/or capacity management features. For example, turning back to FIGS. 48A and 48B, four categories of ications links can be considered: forward-link communications from the AN farm to the western active user coverage area 3460 of A ("Link A"); forward—link communications from the AN farm to the eastern active user coverage area 3460 of B ("Link B"); return—link communications from the n active user coverage area 3460 of A to the AN farm ("Link C"); and return—link communications from the eastern active user ge area 3460 of B to the AN farm ("Link D"). In a first time interval, the eastern user coverage area 3460 of B is active, so that communications are over Link B and Link D. Because there is full overlap between the AN coverage area 3450 and the eastern user coverage area 3460, Links B and D potentially interfere.

Accordingly, during the first time interval, Link B can be allocated a first portion of the bandwidth (e.g., 2 GHZ), and Link D can be allocated a second portion of the bandwidth (e.g., 1.5 GHZ). In a first second interval, the western user coverage area 3460 of A is active, so that communications are over Link A and Link C. Because there is no overlap between the AN coverage area 3450 and the western user ge area 3460, Link A and Link C can use the full bandwidth (e.g., 3.5 GHz) of the end—to—end relay 3403 during the second time interval. For example, during the first time interval, the forward uplink signals can be received using a first frequency range, and the return uplink signals can be received using a second frequency range different from the first frequency range; and during the second time interval, the forward uplink signals and the return uplink signals can be received using a same frequency range (e.g., the first, second, or other frequency range). In some cases, there can be frequency reuse during both the first and second time intervals, with other interference mitigation techniques used during the first time interval. In some cases, the path selection timing can be selected to compensate for such a difference in bandwidth allocation during different time als. For example, the first time interval can be longer than the second time interval, so that Links B and D are allocated less bandwidth for more time to at least lly compensate for allocating Links A and C more bandwidth for a shorter time.

] In some cases, first return uplink signals are received during the first time interval by the plurality of cooperating user—link constituent receive elements 3426a from a first portion of the plurality of user terminals 517 phically distributed over some or all of a first user ge area (e.g., the eastern user coverage area 3460), and second return uplink signals are received during the second time interval by the plurality of cooperating user—link constituent receive elements 3426b from a second n of the plurality of user terminals 517 geographically distributed over some or all of a second user coverage area (e.g., the western user coverage area 3460). When the AN coverage area 3450 (the AN farm) is a subset of the first user coverage area (e.g., as rated in FIGS. 48A and 48B), the AN timing can be calibrated with the —end relay 3403 during the first timeframe (e.g., when there is overlap between the user coverage area 3460 and the AN coverage area 3450).

] As described above, some cases can include determining a respective relative timing adjustment for each of the plurality of ANs 515, such that associated transmissions from the plurality of ANs 515 reach the end-to-end relay 3403 in synchrony (e.g., with sufficiently coordinated timing relative to the symbol duration, which is typically a fraction of the symbol duration such as 10%, 5%, 2% or other suitable value). In such cases, the forward uplink signals are transmitted by the plurality of ANs 515 according to the respective relative timing adjustments. In some such cases, a synchronization beacon signal (e.g., a PN signal generated by a beacon signal generator, as described above) is received by at least some of the plurality of ANs 515 from the end—to—end relay 3403, and the tive relative timing adjustments are determined ing to the synchronization beacon signal. In other such cases, some or all of the ANs 515 can receive loopback transmissions from the end—to—end relay 3403, and the respective relative timing adjustments are determined according to the ck issions.

The various approaches to calibrating the ANs 515 can depend on the ability of the ANs 515 to communicate with the end—to—end relay 3403. Accordingly, some cases can calibrate the ANs 515 only during time intervals during which appropriate coverage areas are illuminated. For example, loopback transmissions can only be used in time als during which there is some overlap between the AN coverage area 3450 and the user ge area 3460 (e.g., the ANs 515 communicate over a loopback beam which can use both a feeder—link antenna subsystem 3410 and a user—link antenna subsystem 3420 of the end—to—end relay 3403). In some cases, proper calibration can further rely on some overlap between the feeder downlink frequency range and the user downlink frequency range.

Conclusion Although the disclosed method and tus is bed above in terms of various examples, cases and implementations, it will be understood that the particular features, s, and functionality described in one or more of the individual examples can be applied to other examples. Thus, the breadth and scope of the claimed invention is not to be limited by any of the examples provided above but is rather defined by the claims.

Terms and phrases used in this document, and variations thereof, unless ise expressly stated, are to be construed as open ended as opposed to limiting. As examples of the foregoing: the term “including” is used to mean “including, without limitation” or the like; the term le” is used to provide examples of ces of the item in discussion, not an exhaustive or limiting list thereof; the terms (L S, 5, “ a or “an” mean “at least one, one or more” or the like.

Throughout the specification, the term “couple” or “coupled” is used to refer broadly to either physical or electrical (including wireless) connection between ents. In some cases, a first ent may be coupled to a second component through an intermediate third component disposed between the first and second component. For example, components may be coupled through direct tions, impedance matching networks, amplifiers, attenuators, filters, direct current blocks, alternating current , etc.

A group of items linked with the conjunction “and” means that not each and every one of those items is ed to be present in the grouping, but rather includes all or any subset of all unless expressly stated otherwise. Similarly, a group of items linked with the conjunction “or” does not require mutual exclusivity among that group, but rather includes all or any subset of all unless expressly stated otherwise. Furthermore, although items, elements, or components of the disclosed method and apparatus may be described or claimed in the singular, the plural is contemplated to be within the scope thereof unless limitation to the singular is explicitly stated.

The presence of ning words and phrases such as “one or more,:9 6‘at least,” or other like phrases in some instances does not mean that the narrower case is intended or required in instances Where such broadening phrases may be absent.

While nce signs may be included in the , these are provided for the sole function of making the claims easier to understand, and the inclusion (or omission) of reference signs is not to be seen as limiting the extent of the matter protected by the claims.

Claims (30)

1. A method of communication at an access node of a ication system comprising a plurality of access nodes at geographically distributed locations ing a communication service to user als distributed over multiple return user beam coverage areas via an end-toend relay comprising multiple return receive/transmit signal paths, the method comprising: receiving a composite return signal and a relay beacon signal from the end-to-end relay, the composite return signal being a superposition of return downlink signals each comprising a respective superposition of return uplink signals from a plurality of the user terminals relayed by the end-to-end relay; lating the relay beacon signal to obtain receive timing information; multiplexing the composite return signal with the receive timing information to obtain a multiplexed composite return signal; and sending the multiplexed composite return signal to a return beamformer for determining a return beam signal by applying return beam weights to the multiplexed composite return signal and at least one other multiplexed composite return signal, wherein the return beam weights t for wireless signal uplink paths up to the end-to-end relay, the le return receive/transmit signal paths through the end-to-end relay, and wireless signal downlink paths down from the end-to-end relay.

2. The method of claim 1, wherein the demodulating comprises: determining a e timing adjustment and a receive phase adjustment to compensate for downlink l impairment based at least in part on a comparison of the relay beacon signal to a local reference

3. The method of claim 2, further comprising: adjusting timing of the composite return signal based on the receive timing adjustment.

4. The method of claim 2 or 3, further comprising: adjusting a phase of the composite return signal based on the receive phase adjustment.

5. The method of any one of the preceding claims, wherein the multiplexed composite return signal comprises a ity of time-domain subsets of samples.

6. The method of claim 5, further comprising: interleaving samples of the multiplexed composite return signal in the plurality of time-domain subsets of samples.

7. The method of any one of the preceding claims, wherein the relay beacon signal comprises a pseudo-noise (PN) code.

8. The method of any one of the preceding claims, wherein the relay beacon signal comprises frame timing information.

9. The method of any one of the preceding claims, wherein the relay beacon signal comprises a plurality of signals at a ity of sub-bands.

10. The method of any one of the preceding claims, wherein the ication system has a first number of the access nodes and the end-to-end relay has a second number of the return receive/transmit signal paths, and the first number is greater than the second .

11. The method of any one of the preceding claims, wherein the composite return signal comprises a plurality of s in different frequency sub-bands.

12. The method of claim 11, wherein the multiple return user beam coverage areas comprise a plurality of subsets of return user beam coverage areas associated with the different frequency sub-bands.

13. The method of claim 12, n each of the plurality of subsets of return user beam coverage areas comprises a ity of non-overlapping return user beam coverage areas.

14. The method of any one of claims 11 to 13, wherein a dth for each of the different frequency sub-bands is greater than or equal to 500 MegaHertz.

15. An access node for communication in a communication system comprising a plurality of access nodes at geographically distributed locations providing a communication service to user terminals buted over multiple return user beam coverage areas via an end-to-end relay comprising multiple return e/transmit signal paths, comprising: a er configured to receive a composite return signal and a relay beacon signal from the end-to-end relay, the composite return signal being a superposition of return downlink signals each comprising a respective superposition of return uplink signals from a plurality of the user terminals relayed by the end-to-end relay; a relay beacon signal demodulator configured to demodulate the relay beacon signal to obtain relay timing information; a multiplexer configured to multiplex the composite return signal with the relay timing information to obtain a multiplexed composite return signal; and a signal interface configured to send the multiplexed composite return signal to a return beamformer for determining a return beam signal by applying return beam weights to the multiplexed composite return signal and at least one other lexed composite return signal, n the return beam weights account for wireless signal uplink paths up to the end-to-end relay, the multiple return receive/transmit signal paths through the end-to-end relay, and wireless signal downlink paths down from the end-to-end relay.

16. The access node of claim 15, wherein the relay beacon signal demodulator compares the relay beacon signal with a local reference signal to obtain a receive timing adjustment and a receive phase ment to compensate for downlink channel impairment.

17. The access node of claim 16, further comprising: a timing adjuster that adjusts timing of the composite return signal based on the receive timing adjustment.

18. The access node of claim 16 or 17, further comprising: a phase adjuster that adjusts a phase of composite return signal based on the e phase ment.

19. The access node of any one of claims 15 to 18, wherein the multiplexed composite return signal comprises a plurality of time-domain subsets of samples.

20. The access node of claim 19, r comprising: an interleaver that interleaves samples of the multiplexed composite return signal in the plurality of time-domain subsets of samples.

21. The access node of any one of claims 15 to 20, wherein the relay beacon signal comprises a pseudo-noise (PN) code.

22. The access node of any one of claims 15 to 21, wherein the relay beacon signal ses frame timing information.

23. The access node of any one of claims 15 to 22, wherein the relay beacon signal comprises a plurality of signals at a plurality of nds.

24. The access node of any one of claims 15 to 23, wherein the composite return signal comprises a circularly polarized signal and the relay beacon signal comprises a linearly polarized signal.

25. The access node of any one of claims 15 to 24, wherein the communication system has a first number of the access nodes and the end-to-end relay has a second number of the return receive/transmit signal paths, wherein the first number is different than the second number.

26. The access node of claim 25, wherein the first number is greater than the second .

27. The access node of any one of claims 15 to 26, wherein the end-to-end relay is a satellite.

28. The access node of any one of claims 15 to 27, wherein the composite return signal comprises a plurality of signals in different frequency sub-bands.

29. The access node of claim 28, wherein the multiple return user beam coverage areas comprise a ity of subsets of return user beam coverage areas associated with the different frequency sub-bands.

30. The access node of claim 29, n each of the plurality of s of return user beam coverage areas comprises a plurality of non-overlapping return user beam coverage areas.