JP2024516594A - A method for time synchronization and localization in mesh networks - Google Patents

Abstract

The method includes scheduling transmission of a first synchronization signal by a first node and scheduling transmission of a second synchronization signal by a second node. The method also includes receiving from the first node a first phase reference associated with the first synchronization signal after transmission of the first synchronization signal, and receiving from the second node a first arrival phase of the first synchronization signal at the second node. The method further includes receiving from the second node a second phase reference associated with the second synchronization signal after transmission of the second synchronization signal, and receiving from the first node a second arrival phase of the second synchronization signal at the first node. The method further includes calculating a propagation delay between the first node and the second node based on the first phase reference, the second phase reference, the first arrival phase, and the second arrival phase.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This application is a continuation of U.S. Patent Application No. 17/338,543, filed June 3, 2021, which claims the benefit of U.S. Provisional Patent Application No. 63/177,805, filed April 21, 2021, each of which is incorporated by reference in its entirety.

The present invention relates generally to the field of networking and digital communications, and more particularly to a novel and useful method of time synchronization and location determination in the field of networking and digital communications.

Figure 1A is a flow chart of the method, Figure 1B is a flow chart of a first variant of the method, Figure 1C is a flow chart of a second variant of the method, and Figure 1D is a flow chart of a third variant of the method. FIG. 2 is a schematic diagram of a mesh network. Figure 3A is a schematic diagram of a node pair in a mesh network, Figure 3B is a schematic diagram of a node in a mesh network, and Figure 3C is a schematic diagram of a node's self-reception hardware. 4A and 4B are schematic diagrams of a synchronous slot structure; 5A and 5B are flow diagrams of the transmission and reception of a synchronization signal and a self-reception signal, respectively. 6A, 6B, and 6C are flow diagrams of a second method, a variation of the second method, and a variation of the second method, respectively.

The following description of the embodiments of the present invention is not intended to limit the invention to these embodiments, but rather to enable those skilled in the art to make and use the invention. The variations, configurations, implementations, exemplary implementations, and examples described herein are optional and are not limited to the variations, configurations, implementations, exemplary implementations, and examples they describe. The invention described herein may include any and all combinations of these variations, configurations, implementations, exemplary implementations, and examples.

1. Methods for Characterizing Time Bias and/or Propagation Delay As shown in FIG. 1A, a method S100 for characterizing a time bias and propagation delay between a pair of nodes includes transmitting a first synchronization signal at a first node of the pair of nodes during a first synchronization slot at a first node at a first time according to a first clock of the first node in block S110A, combining back the first synchronization signal to generate a first self-received signal in block S120A, receiving the first self-received signal in block S130A, calculating a time of arrival of the first self-received signal according to the first clock in block S140A, receiving a second synchronization signal from a second node in block S150A, and calculating a time of arrival of the second synchronization signal according to the first clock in block S160A. Method S100 also includes transmitting a second synchronization signal at the second node during the first synchronization slot at a second time according to a second clock of the second node in block S110B, combining back the second synchronization signal to generate a second self-received signal in block S120B, receiving the second self-received signal in block S130B, calculating a time of arrival of the second self-received signal according to the second clock in block S140B, receiving the first synchronization signal from the first node in block S150B, and calculating a time of arrival of the first synchronization signal according to the second clock in block S160B. Method S100 further includes calculating at the second node in block S170 a time bias and a propagation delay between the node pair based on the time of arrival of the first self-received signal, the time of arrival of the second synchronization signal, the time of arrival of the second self-received signal, and the time of arrival of the first synchronization signal.

As shown in FIG. 1B, a first variation of method S100 includes transmitting a coordination signal to a second node of the node pair in block S102, the coordination signal indicating a second time for transmitting a second synchronization signal, transmitting the first synchronization signal at the first time according to a first clock of the first node in block S110, combining back the first synchronization signal to generate a first self-received signal in block S120, receiving the first self-received signal in block S130, calculating the arrival time of the first self-received signal in block S140 according to the first clock, receiving a second synchronization signal in block S150, the second synchronization signal being transmitted by the second node at the second time according to a second clock of the second node, and combining back the first self-received signal in block S160 according to the first clock. Calculating the arrival time of the synchronization signal; receiving the arrival time of a second self-received signal from the second node in block S142, where the second self-received signal is coupled back by the second node upon transmission of the second synchronization signal, and the arrival time of the second self-received signal is calculated according to a second clock; receiving the arrival time of a first synchronization signal from the second node in block S162, where the arrival time of the first synchronization signal is calculated by the second node according to the second clock; and calculating a first time bias between the first clock and the second clock and a first propagation delay between the first node and the second node based on the arrival time of the first self-received signal, the arrival time of the first synchronization signal, the arrival time of the second self-received signal, and the arrival time of the second synchronization signal in block S170.

1C, a second variation of method S100 includes scheduling transmission of a first synchronization signal at a first time by the first node in block S104 and scheduling transmission of a second synchronization signal at a second time by the second node in block S106. The second variation of method S100 also includes receiving from the first node, after transmission of the first synchronization signal at the first time by the first clock of the first node, a time of arrival of a first self-received signal by the first clock in block S144, the first self-received signal being a return combination of the first synchronization signal, and receiving a time of arrival of the first synchronization by the second clock of the second node in block S164. The second variation of method S100 further includes receiving from the second node, after transmission of the second synchronization signal by the second node at the second time by the second clock, a time of arrival of a second self-received signal by the second clock in block S146, the second self-received signal being a return combination of the second synchronization signal, and receiving, in block S166, a time of arrival of the second synchronization signal by the first clock. The second variation of method S100 also includes, in block S170, calculating a first time bias between the first clock and the second clock and a first propagation delay between the first node and the second node based on the arrival time of the first self-received signal, the arrival time of the first synchronization signal, the arrival time of the second self-received signal, and the arrival time of the second synchronization signal.

A third variation of method S100 includes predicting, in block S180, a magnitude of drift in the time bias between the first node and the second node between the first synchronization slot and the second synchronization slot based on the predictive drift model, and setting, in block S190, a duration of the second synchronization slot equal to the sum of the maximum time bias uncertainty, the maximum accumulated drift magnitude over the duration of the first synchronization signal, the propagation delay, and the duration of the first synchronization signal.

1D, a fourth variation of method S100 includes scheduling transmission of a first synchronization signal by a first node during a first synchronization slot in block S104, the first synchronization signal being characterized by a first set of carrier frequencies, and scheduling transmission of a second synchronization signal by a second node during a second synchronization slot in block S106, the second synchronization signal being characterized by a second set of carrier frequencies. This variation of method S100 also includes, after transmission of the first synchronization signal by the first node during the first synchronization slot, receiving from the first node a phase reference for each carrier frequency of the first set of carrier frequencies in block S147, and receiving from the second node an arrival phase for each carrier frequency of the first set of carrier frequencies based on the first synchronization signal received at the second node in block S167. This variation of method S100 further includes receiving a phase reference for each carrier frequency of the second set of carrier frequencies from the second node in block S148 after transmission of the second synchronization signal by the second node at the second time, and receiving an arrival phase for each carrier frequency of the second set of carrier frequencies from the first node in block S168. This variation of method S100 further includes calculating a propagation delay between the first node and the second node in block S172 based on the phase reference for each carrier frequency of the first set of carrier frequencies, the arrival phase for each carrier frequency of the first set of carrier frequencies, the phase reference for each carrier frequency of the second set of carrier frequencies, and the arrival phase for each carrier frequency of the second set of carrier frequencies.

1.1 Applications In general, method S100 is performed by a pair of node devices (hereinafter "nodes") in a network and/or by a remote server to synchronize the clocks of the node pair with high precision, e.g., within 1 nanosecond. Method S100 can calculate the relative time bias between the clocks, e.g., within 1 nanosecond or sub-nanosecond precision, without precise calibration of either node's hardware and without prior information about the signal propagation delay or physical distance between the two nodes (or even without the requirement that the nodes be stationary), provided that each node uses standard electronic clocking technology, such as a crystal oscillator clock. Furthermore, method S100 can utilize a frequency bandwidth of less than 5 megahertz (e.g., 500-100 kHz). This small bandwidth of method S100 allows transmission in lower frequency bands with large propagation range and/or penetration. Examples of such bands include frequency bands for unlicensed use, cellular communications, and/or public safety applications (e.g., the 902-928 MHz ISM band, sub-GHz cellular bands, or the Citizens Radio Service (CBRS) at 26.965-27.405 MHz). Thus, method S100 enables applications such as, but not limited to, precise location of RF emitting devices, remote sensing of public safety infrastructure, and time-synchronized distributed antenna systems that may further enable improved time-based data transfer protocols.

More specifically, the method S100 for synchronizing time between two nodes can be classified as a two-way ranging and synchronization protocol. However, the method S100 is distinguished from other two-way ranging protocols in that each node transmits independently of the other node in absolute time (according to each node's clock) rather than transmitting in response to receiving a signal from the leader node of the pair. Furthermore, each node is configured to provide a local reference copy of the transmitted signal within the node's transceiver hardware to implement a scheme that provides a "time of departure" for each transmitted synchronization signal (through back-coupling and/or reflection of the transmitted synchronization signal). The arrival time of the reflected and/or back-coupled version of this transmitted synchronization signal (hereinafter the "self-received signal") includes the transmission and reception chain delays incurred by the signal transmitted by another node, thereby allowing the arrival times of various signals to be directly compared without precise characterization of these delays and without assuming that the response times of the individual nodes are deterministic and symmetrical.

The blocks of method S100 may be performed by a system that may include pairs of node devices (i.e., nodes) in a mesh network, leader and follower nodes in the mesh network, and/or a remote server cooperating with pairs of nodes in the mesh network. Each node in the mesh network may include radio and baseband processing hardware, such as antennas, transceiver hardware, FPGA/DSP, clocks, and self-reception hardware (e.g., directional couplers, RF power splitters, combiners, circulators), as described further below. However, method S100 may also be performed between nodes in a wired network. The nodes in the mesh network are mutually connected to the Internet or a local area network such that the initial time bias between any pair of nodes in the mesh network is initially bounded by the Network Time Protocol (hereinafter "NTP") or any other network time synchronization protocol. This time bias may range from tens of milliseconds to a few microseconds in many state-of-the-art networks.

Nodes in a mesh network communicate with each other on a common frequency band or bands shared among them using time division multiple access (hereinafter "TDMA") and/or code division multiple access (hereinafter "CDMA") to minimize frequency bandwidth usage. Thus, the nodes may perform the blocks of method S100 during one or more synchronization slots in a TDMA frame structure. In one embodiment, the synchronization slot duration and frame length are dynamically adjustable. In particular, method S100 may leverage a smaller initial time bias between pairs of nodes to reduce the synchronization slot duration and further reduce the uncertainty in the time bias calculation. Additionally, the frame structure may include a data transfer frame that allows nodes to communicate with each other or with a remote server via the Internet Protocol Suite to perform the blocks of method S100. For example, the nodes may transfer a set of arrival times to a remote server for further processing by method S100.

Assuming initial coarse clock synchronization (e.g., 1 to 10 ms) between the node pairs, each node of the node pair transmits a synchronization signal to the other node at the beginning of a synchronization slot according to each node's clock. However, due to a relative time bias between each node, which may not be constant or generally known a priori, these synchronization signals are transmitted at times offset by the time bias between the two nodes. The synchronization signal may be a frequency-, amplitude-, or phase-modulated pseudorandom code, or a combination of multiple codes (e.g., on multiple center carrier frequencies). Upon receiving a synchronization signal from the other node of the node pair, each node calculates the time of arrival (hereinafter "TOA"), such as by using the magnitude of the autocorrelation peak associated with one or more codes, the time offset, and the carrier phase.

When each node transmits a synchronization signal, it generates a self-received signal as shown in FIG. 3A by utilizing self-receiving hardware (shown in FIG. 3C) that back-couples and/or internally reflects a portion of the synchronization signal. The self-receiving hardware reflects and/or back-couples an attenuated copy of the outgoing synchronization signal, which is then received by the same node that transmitted the synchronization signal. Each node then calculates a TOA for the self-received signal after the self-received signal has propagated through the same receive chain as the incoming synchronization signal that is received at the antenna of that node. Thus, the TOA of the self-received signal serves as the time of departure of the synchronization signal offset by the receiver delay time of the transmitting node. Because method S100 involves generating a self-received signal, the TOA of an incoming synchronization signal from another node can be directly compared to the time of departure of the originating node without requiring precise measurement and/or calibration of the receiver delay.

Each node of the pair then receives the synchronization signal from the other node and calculates the TOA for the synchronization signal. Thus, by the end of the synchronization slot, each node has recorded the TOA for its own received signal and the TOA for the synchronization signal of the other node. These TOAs are then transmitted to one of the nodes or to a remote server, which solves two simultaneous equations for two unknowns: the relative time bias (including the receiver delay) between the node pair and the propagation delay between the node pair. This simultaneous equation can be solved based on the reciprocity theorem of electromagnetics. The relative time bias between the clocks of the two nodes can then be calculated and tracked by the leader node or a remote server cooperating with the mesh network. Alternatively, the relative time bias between the nodes can then be reported to both nodes, so that one node (i.e., the follower node) can synchronize its clock to match the other (i.e., the leader node).

While the first synchronization between the two node clocks by method S100 may be very accurate, even greater accuracy and reduced TDMA overhead is achieved by slot refinement, where the synchronization slot duration is reduced based on the smaller time bias between the two clocks. Thus, with each successive synchronization slot for a node pair, the synchronization slot duration is reduced, thereby reducing error sources such as cumulative jitter and/or environmental frequency drift that may occur during the synchronization process. For example, a typical crystal oscillator may be expected to have a frequency error of 6 ppm, corresponding to a drift of 6 nanoseconds during a synchronization slot of 1 millisecond duration. If the synchronization slot duration is reduced to 100 microseconds, the expected drift is reduced to less than 1 nanosecond.

The system can also improve the data transfer rate of a node by implementing a synchronization slot refinement technique to increase the period within a frame for data transmission. Additionally, a shorter synchronization slot duration allows method S100 to be performed at a higher repetition rate, thereby improving synchronization in changing environmental conditions.

After synchronization is achieved between a pair of nodes in the mesh network, the common time can be distributed throughout the mesh network by placing other nodes in the mesh network in successive synchronization slots and performing method S100 pair by pair.

In some implementations, method S100 leverages iterative synchronization between two nodes and pairwise synchronization between multiple nodes in the network in combination with environmental data recorded from an internal measurement unit (hereinafter "IMU") or temperature sensor in each node to create a predictive drift model for each node in the mesh network. The predictive drift model can characterize the time drift of the node as a function of the measured environmental input data over time. Method S100 can also include triggering a synchronization slot of the next frame or modifying the length of the synchronization slot based on the output of the predictive drift model. Alternatively, method S100 can include adjusting the clock of the node between synchronization slots.

Generally, method S100 is described herein as including detection of the TOA of the synchronization signal and the self-received signal at the nodes in the mesh network. However, method S100 may also include detection of the phase of arrival (hereinafter "POA") of the synchronization signal and the self-received signal to calculate time bias and propagation delay between pairs of nodes in the mesh network. Additionally, method S100 may also include transmission and reception of a frequency hopping spread spectrum signal as the synchronization signal, thereby enabling precise measurement of the relative time delay of the synchronization signal in transmission between pairs of nodes compared to the self-received signal.

1.2 Example The method S100 for synchronizing time between two node devices can be applied pairwise to a mesh network of nodes to improve the accuracy and flexibility of the location protocol executed by those nodes. Additionally, the method S100 can increase the data transfer rate of a time-based multiple access protocol (e.g., by decreasing the buffer duration between slots of the protocol).

In one application of method S100, a mesh network of nodes can perform method S100 pairwise to synchronize their clocks across the mesh network. Each node of the mesh network can then individually calculate the TOA of a signal from an RF source not included in the mesh network. By comparing the TOA between nodes in the mesh network and performing multilateration, the location of the RF source can be calculated with an accuracy limited by the temporal synchronization between the nodes. For example, with temporal synchronization within 1 nanosecond, the location of the RF source can be calculated to within 30 centimeters relative to a node in the mesh network (assuming a phase velocity corresponding to the speed of light in a vacuum, 299792458 meters per second). If one node in the mesh network acts as an anchor node and is calibrated with accurate global position information, the absolute location of the RF source can also be found.

In one application of mesh network based localization, nodes are distributed throughout an area, in buildings, or on cars driving on urban roads. These nodes execute the blocks of method S100 to periodically synchronize their clocks and then collect TOA data for other RF sources in the mesh area. A remote server cooperating with the nodes can then convert these TOA data into three-dimensional geospatial locations of the RF sources to improve location services, asset tracking, and object detection. Thus, the remote server or reader node can calculate the relative positions of the nodes in the mesh network based on the propagation delays between the nodes, receive the arrival times of signals from the transmitting device from several nodes in the mesh network, and calculate a relative position estimate for the transmitting device via multilateration that accounts for the relative time bias between each node in the mesh network.

In another example, TDMA-based protocols may be limited to slot durations of tens of milliseconds. Thus, the blocks of method S100 may be performed to synchronize clocks of devices in a wired or wireless network and more accurately calculate time biases to allow for reduced dead time and inter-slot interference while improving bit error rate (BER) and bandwidth utilization by reducing inefficiencies in multiple access protocols.

1.3 Telecommunications Deployment Generally, the method S100 can be performed by a telecommunications network, such as a cellular network implementing the LTE, 4G, 5G, and 5G NR standards proposed by the Third Generation Partnership Project (hereinafter "3GPP"). For example, the method S100 can be performed by a node in a cellular network, such as a 5G radio node (hereinafter "gNB") and/or enhanced 4G eNodeBs (hereinafter "ng-eNB"), instead of a partial timing support (hereinafter "APTS") or full timing support (hereinafter "FTS") time distribution and/or synchronization method. For example, a node in a telecommunications network can perform the method S100 to enable over-the-air synchronization (hereinafter "OAS") to reduce networking overhead and increase the accuracy of inter-node time synchronization.

1.4 Node Hardware As mentioned above, method S100 is performed by a node in the network or through cooperation with a node in the network. Examples of wireless node hardware are shown in Figures 3B and 3C. In general, a node includes transmit and receiver components, FPGAs and/or DSPs configured to generate and process signals, a clock, and self-receiving signal hardware.

In one embodiment, the nodes transmit information wirelessly and therefore include RF transceiver hardware such as the super-heterodyne radio architecture and Rx/Tx antennas shown in FIG. 3B. Alternatively, the nodes may include a zero-IF architecture (i.e., direct conversion receiver). Each node includes a "receive chain" and a "transmit chain." The receive chain includes a pipeline of hardware components that process signals received from the Rx port. The transmit chain includes a pipeline of hardware components that process transmit signals generated by the FPGA or DSP and send them to the Tx port. The receive chain and transmit chain impart a "receive chain delay" and a "transmit chain delay," respectively. "Receive chain delay" and "transmit chain delay" refer to the amount of time that elapses as a signal traverses the receive chain or transmit chain, respectively.

In an alternative embodiment, the nodes may communicate over a wired network. In this embodiment, the nodes may include I/O ports and/or appropriate interface converters to communicate over any wired medium (e.g., Ethernet/twisted pair, coax, or fiber optic) instead of antennas. These interface converters may also be used to provide the self-received signals necessary to measure the propagation delay.

In one embodiment, the node includes an existing transceiver infrastructure, such as a cell tower, a mobile phone, or any other RF transceiver device adapted to perform method S100. A cell tower or other existing transceiver can be adapted to perform method S100 with just a software update. In an alternative embodiment, the node can include hardware optimized to improve the implementation of method S100, which may include an impedance matching network at the antenna interface between passive coupling devices (e.g., coupling the transmit and receive chains to the antenna).

In a wireless node embodiment, each node's FPGA or DSP is configured to generate a complex digital signal and output the generated signal to a DAC. The complex components of the digital signal represent the in-phase and quadrature (i.e., I/Q) portions of the analog signal generated by the DAC. Additionally, the node's FPGA or DSP receives the digital signal from the node's antenna via an ADC, and then timestamps the received synchronization signal according to the instantaneous value of its own clock and the TOA calculation process described further below.

Each node includes a clock, such as a crystal oscillator clock or an atomic clock, that performs timekeeping functions at the node and is also used for sampling, digital synthesis, and processing. Method S100 can be performed to synchronize the clocks of multiple nodes in the network. In particular, method S100 can effectively synchronize crystal oscillator clocks that meet the fundamental frequency stability, phase noise, and frequency requirements of wireless and/or wired communications. In one example, the clock is a crystal oscillator with an AT cut and a clock frequency of 10 megahertz (MHz). However, a node can include a crystal oscillator of any frequency or cut, provided that the above constraints are met.

Each node may include specific self-reception hardware that back couples and/or reflects the self-received signal in block S120. In general, when a node transmits a synchronization signal, the self-receive hardware reflects and/or couples an attenuated repetition of that synchronization signal back to the receive port of the transmitting node. By processing the transmitted synchronization signal through the receive chain and calculating the TOA for the self-received signal, each node can timestamp the transmitted signal delayed by the receive chain delay. When the node later receives a synchronization signal from another node, the TOA of that synchronization signal will also be subject to the same receive chain delay. Because both the TOA of the self-received signal and the TOA of the synchronization signal include the receive chain delay, they can be directly compared without precise calibration of the receive chain hardware.

In one embodiment, the self-reception hardware includes an impedance-mismatched directional coupler as shown in FIG. 3C. The self-reception hardware may also include a variable impedance circuit controlled by the node to adjust the gain of the reflected Tx signal to the Rx port to generate the self-reception signal. Depending on the specific hardware implementation of the node, similar impedance matching may be applied to a circulator, power splitter, or any other transmission line device, as described further below. In one embodiment, the node may include standard antenna interface hardware with an unintentionally mismatched antenna impedance (e.g., characterized by a voltage standing wave ratio greater than 1), which may function as self-reception hardware by reflecting a synchronization signal at the antenna interface. Thus, each node need not include dedicated self-reception hardware to perform method S100.

The self-reception hardware includes an impedance matching network between the antenna and the passive coupling device to define a tuned reflection coefficient for the signal incident on the interface between the node's passive coupling device and the antenna. The impedance matching network transforms the impedance of the antenna to an impedance that achieves a precise reflection coefficient between the passive coupling device and the antenna. The reflection coefficient is selected so that the reflected power of the self-reflected signal is greater than the noise floor of the ADC and less than the saturation voltage of the ADC.

In another embodiment, the matching network can maintain the reflection coefficient for a wide range of signal frequencies. For example, the self-receiving hardware can include a switchable wideband matching network to improve matching of the reflection coefficient over a wide range of transmission frequencies. In this example, the self-receiving hardware can actively switch between multiple impedance matching networks depending on the frequency of the signal being transmitted by the node.

In yet another embodiment, the self-receiving hardware can include a frequency multiplexer connected to multiple impedance matching networks such that each separate impedance matching network sees the signal at a frequency that produces a precise reflection coefficient.

In a further embodiment, the system 100 may include an adaptive impedance matching network that can adjust its own impedance and therefore the reflection coefficient between the passive coupling device and the antenna. In this case, the node may use the adaptive impedance matching network to adjust the reflection coefficient, thereby ensuring reception of the self-reflected signal despite changes in gain in amplifiers of the RF transmit and/or receive chains and/or changes in noise levels in the ADC due to signal interference in one or more frequency bands of interest.

In one embodiment, the self-receiving hardware includes a directional coupler as a passive coupling device. The directional coupler includes four ports, two for each of the coupled transmission lines. In this embodiment, the transmit port and the antenna port are located on the same transmission line in the directional coupler, but the receive port is located on the coupled port opposite the antenna, so that the coupled power is received from the antenna and the reflected power is received from the interface between the directional coupler and the antenna. Alternatively, the self-receiving hardware can include a directional coupler with the receive port and the antenna port located on the same transmission line. As a result, the receive port receives direct power from the antenna port and also receives reflected power from the antenna port. However, in this alternative embodiment, the antenna port receives the signal from the transmit port at a lower power due to the coupling between the two transmission lines. Thus, a higher power at the transmit port will result in the same power at the antenna port. However, in an embodiment where the receive port and the antenna port are coupled to the same transmission line, the self-receiving hardware can improve the sensitivity of the transceiver to signals received at the antenna.

In an alternative embodiment, the self-receiving hardware includes a power divider as a passive coupling device. The power divider divides the power from the input port between the two output ports. In this embodiment, the power input at the antenna port is divided between the transmit and receive ports, so any power reflected at the interface between the antenna and the power divider is also divided between the transmit and receive ports.

In yet another embodiment, the self-receiving hardware includes a circulator as a passive coupling device. The circulator can directly couple the transmit port to the antenna port when the antenna port is directly coupled to the receive port. In this case, any reflections generated at the interface of the antenna port are coupled back to the receive port.

In one embodiment, each node may include a software defined radio architecture that performs the functions of any of the hardware elements described above.

However, self-receiving hardware can include any software or hardware system for sending a Tx signal to a node's Rx port.

1.5 Overall Time Synchronization Prior to execution of method S100, the clocks of each node are roughly synchronized using another time synchronization protocol. For example, node pairs may be connected to the Internet and communicate with one or more NTP servers. In one example, node pairs may synchronize their clocks with NTP to within 1 to 10 milliseconds. Alternatively, in another example, prior to execution of the subsequent blocks of method S100, node pairs may communicate with a Global Navigation Satellite System (hereinafter "GNSS") time synchronization server and synchronize to within 1 microsecond.

In this way, method S100 leverages existing time synchronization protocols to enable nodes to communicate using the same synchronization slot in the TDMA described below.

Thus, the remote server may coarsely synchronize the clock of the first node with the clock of the second node, schedule transmission of the first synchronization signal within a first synchronization slot characterized by a first synchronization slot duration of the first frame, schedule transmission of the second synchronization signal within a second synchronization slot characterized by the first synchronization slot duration of the first frame, configure the second node to receive the first synchronization signal during the first synchronization slot of the first frame, and configure the first node to receive the second synchronization signal during the second synchronization slot of the first frame.

1.6 Slot and Frame Definition In general, in method S100, communication between node pairs is performed according to a TDMA structure of slots and frames, an example of which is shown in Figures 4A and 4B. The TDMA structure may include one or more slots for both time synchronization via method S100 and time synchronization via data transfer. In one embodiment shown in Figure 4A, each TDMA frame includes a synchronization slot for synchronizing each unique node pair in the mesh network. For example, a mesh network including three nodes _n1 , _n2 , and _n3 includes a first synchronization slot for synchronizing _n1 and _n2 , a second synchronization slot for _n1 and _n3 , and a third synchronization slot for _n2 and _n3 . Alternatively, as shown in Figure 4B, the TDMA frame may include a synchronization slot for each node in the mesh network to transmit a synchronization signal to all other nodes in the mesh network. For example, a mesh network including three nodes _n1 , _n2 , and _n3 includes a first synchronization slot for _n1 to transmit a synchronization signal to _n2 and _n3 , a second synchronization slot for _n2 to transmit a synchronization signal to _n1 and _n3 , and a third synchronization slot for _n3 to transmit a synchronization signal to _n1 and _n2 .

In one embodiment, each synchronization slot is further divided into two subsequent sub-slots to reduce synchronization overhead (e.g., when the initial synchronization performed by NTP is sufficiently low). For example, in synchronizing _n1 and _n2 , _n1 transmits during the first sub-slot and _n2 receives the transmission from _n1 . Then, in the second sub-slot, _n2 transmits and _n1 receives.

Each TDMA frame may also include a header indicating the presence and order of synchronization slots and data transfer slots for each node in the mesh network. In one embodiment, each TDMA frame includes a set of synchronization slots. Alternatively, each TDMA frame may or may not include a set of synchronization slots depending on the header of that TDMA frame. In yet another embodiment, a TDMA frame may include multiple sets of synchronization slots. In one embodiment, the header may include a coordination signal transmitted by a remote server or reader node coordinating with the TDMA protocol to communicate a specific transmission time of the synchronization signal to nodes in the mesh network.

The duration of a synchronization slot and the buffer time between two synchronization slots corresponding to different node pairs may vary depending on the implementation and factors such as the currently known time bias between the nodes and the associated uncertainties. Method S100 may include adjusting the synchronization slot duration based on the output of a predictive drift model, which is described further below. Additionally, the synchronization slot duration may have a lower bound equal to the sum of the uncertainty of the time bias between the node pairs, the propagation time between the node pairs, the uncertainty of the propagation time between the node pairs, and the duration of the synchronization signal, of which the duration of the synchronization signal is typically the most significant factor. In one implementation, the slot and the buffer between the synchronization slots are both set to 1 millisecond. Additionally or alternatively, method S100 may also include adjusting the synchronization slot duration as a function of the signal-to-noise ratio of the previously received synchronization signal to provide additional measurement acquisition time to determine the TOA of signals having a lower signal-to-noise ratio.

Furthermore, the duration of each TDMA frame is subject to a set of practical constraints. The TDMA frame duration may have an upper limit defined by the expected drift of the least stable node clock of a node pair compared to the desired accuracy of the clock synchronization process. For example, if it is known that one node clock typically drifts at most 1 nanosecond per second due to accumulated jitter and frequency drift, and the desired accuracy of the clock synchronization process is 1 nanosecond, then the TDMA frame duration (or the time between synchronization slots) is limited to 1 second, so that the time bias between nodes cannot exceed 1 nanosecond.

Furthermore, the TDMA frame duration has a lower bound based on the expected initial clock offset between nodes, plus the total duration of the synchronization slots, the total duration of any data transfer slots, and the total duration of any time buffers between slots. The frame duration must be long enough for each node to receive the initial synchronization signal from every other node in the network, while allowing for a large initial time bias. Thus, when nodes are initially synchronized using NTP, the frame duration should be at least tens of milliseconds to allow for the common time biases between NTP-synchronized clocks.

In one embodiment, during initial synchronization of a set of nodes in a mesh network, the set of nodes is configured (e.g., by a remote server) to broadcast a cooperation signal at a predetermined time according to each node's clock. If initial coarse time synchronization is performed between the nodes, the system can designate a leader node as the first node to transmit the cooperation signal to the other nodes. Each node in the mesh network can then designate itself as the leader node by acknowledging receipt of the leader node's confirmation signal. In this manner, a first node of a node pair can receive confirmation of receipt of the cooperation signal from a second node of the node pair, and can designate the first node as the leader node in response to receiving confirmation of receipt of the cooperation signal before the first node receives a second cooperation signal from another node in the mesh network.

In an alternative embodiment, the TOA of the coordination signal can be calculated at each node and transmitted to a remote server, which can then determine a leader node in the mesh network based on the measured propagation delay and corresponding distance between the nodes or link quality metrics such as a received signal strength indicator (RSSI). In yet another alternative embodiment, a remote server designates a leader node or coordinates the TDMA slots and method S100 for a set of nodes in the mesh network.

When communicating within a TDMA synchronization slot, each node of a node pair performing method S100 may transmit a synchronization signal during the synchronization slot and receive a synchronization signal and a self-reception signal from the other node. Thus, in one embodiment, the nodes may perform a time division duplex (hereinafter "TDD") or half-duplex scheme to communicate within the synchronization slot. In this embodiment, the synchronization signals transmitted by each node of the node pair are separated by a predetermined delay time (e.g., about half the synchronization slot duration) such that a first node of the node pair may transmit a first synchronization signal during a first segment of the synchronization slot, receive a self-reception signal corresponding to the first synchronization signal, and then receive a second synchronization signal (transmitted from a second node of the node pair) during a second segment of the synchronization slot. In this embodiment, the node performing method S100 may establish a half-duplex communication link between the two nodes by transmitting a coordination signal to the other node in the mesh network that includes timing information such as the duration and timing of the slot segments.

In one embodiment, the system may implement a full-duplex (i.e., frequency division duplex, hereafter "FDD") communication scheme in which each node of a node pair may simultaneously transmit and receive a synchronization signal (e.g., each synchronization signal is transmitted over multiple frequencies). Additionally, each node of a node pair may be configured to simultaneously receive both its own received signal coupled back to its receiver and the synchronization signal from the second node of the node pair.

Thus, in one embodiment, the leader node of the node pair can transmit a coordination signal to the second node indicating a second time point at which to transmit a second synchronization signal that is equal to the first time point at which to transmit the first synchronization signal, and establish a full-duplex communication link with the second node.

However, in applications where larger bandwidths are available, nodes may also communicate using frequency division multiple access or any other channel access method.

Furthermore, the coordination signal and/or the synchronization signal header may specify a set of carrier frequencies characterizing each synchronization signal and the timing of each frequency hop in the synchronization signal transmitted between nodes of the mesh network (based on coarse time synchronization between the nodes) by communicating a frequency hopping spread spectrum (hereinafter "FHSS") scheme for one or more subsequent synchronization signals.

1.7 Characterization of Time Bias and Propagation Delay As shown in FIG. 1A, each node of a node pair can execute blocks S110, S120, S130, S140, S150, and S160 to transmit a synchronization signal to each other at a predetermined transmission time evaluated according to each node's clock, combine the synchronization signal back through the node's self-reception hardware to form a self-received signal, and record the TOA and/or POA for the self-received signal and the synchronization signal received from the other node of the node pair. Thus, each node of the node pair calculates two TOAs according to the node's clock, or calculates a set of POAs corresponding to each carrier frequency of the synchronization signal. Each node can then transmit the TOA (or POA) of the self-received signal and the TOA (or POA) of the synchronization signal to one of the nodes of the node pair (e.g., during a data transfer slot in the frame) or to a remote server in order to calculate the time bias between the node pair and the propagation delay between the node pair in block S170.

The variations of method S100 shown in FIG. 1B and FIG. 1C include additional blocks for coordinating the node pairs performing method S100. In a first variation shown in FIG. 1B, the leader node of the node pair performs blocks S102, S142, S162, and S170 in addition to the above-mentioned blocks of method S100 to communicate the transmission time to the follower node, receive the TOA of the self-received signal calculated by the follower node, receive the TOA of the synchronization signal calculated by the follower node, and calculate the time bias and propagation delay between the leader node and the follower node at the leader node. Thus, the leader node can communicate the transmission time to the follower node so that the leader node can receive the synchronization signal transmitted by the follower node during the synchronization slot of the TDMA communication method.

In a second variation of method S100 shown in FIG. 1C, a remote server or other computing device in communication with the node pair performs blocks S104, S106, S144, S146, S164, S166, and S170 of method S100 to schedule transmission times for each node of the node pair, receive TOAs from each node for the synchronization signal received by each node of the node pair and the self-received signal received at each node of the node pair, and calculate time biases and propagation delays between the node pairs.

Thus, in any of the above variations, various entities included in the system may cooperate to execute a time synchronization protocol between pairs of nodes in the mesh network to calculate time biases and propagation delays between pairs of nodes. In one implementation, the time bias is tracked by a remote server and is expected when managing a TDMA communication protocol, performing multilateration of transmitting RF devices, or executing any other protocol that utilizes a time-synchronized mesh network. Alternatively, one of the nodes may act as a leader node to track time biases and propagation times between nodes in the mesh network.

To synchronize the clocks of a node pair with each other, each node of the node pair may execute blocks S110, S120, S130, S140, S150, and S160 simultaneously or sequentially. For example, a first node of the node pair may execute a first instance of these blocks of method S100 at S110A, S120A, S130A, S140A, S150A, and S160A, and a second node of the node pair may execute a second instance of these blocks at S110B, S120B, S130B, S140B, S150B, and S160B. For clarity, any description of blocks S110, S120, S130, S140, S150, or S160 applies to blocks S110A, S120A, S130A, S140A, S150A, or S160A, respectively, executed by the first node, or blocks S110B, S120B, S130B, S140B, S150B, or S160B, respectively, executed by the second node.

Depending on the relative time bias between the two nodes, the first node may execute blocks S110A, S120A, S130A, and S140A at any time relative to the second node executing blocks S110B, S120B, S130B, and S140B (assuming both nodes are executing during a synchronization slot). However, the first node executes blocks S150A and S160A to receive and calculate the TOA for synchronization sent by the second node after the second node executes blocks S110B, S120B, S130B, and S140B. Similarly, the second node executes blocks S150B and S160B after the first node executes blocks S110A, S120A, S130A, and S140A.

1.7.1 Transmission Time In blocks S110A and S110B, each node of a node pair transmits a synchronization signal at a predetermined or coordinated transmission time within the synchronization slot. In an embodiment in which the nodes perform TDD, the first transmission time in block S110A may be offset by a transmission interval from the second transmission time in block S110B. Thus, each node of a node pair transmits a synchronization signal at a predetermined transmission time according to each node's own clock. For example, if the first transmission time of the first node is set to 1:00:00, the first node transmits a synchronization at 1:00:00 according to its own clock. Thus, from a third-party perspective, each node transmits a synchronization signal offset by a time bias between the two nodes and also offset by a transmission interval between the transmission times of each node.

Because each node directly measures its own reception delay, rather than relying on a deterministic delay that is consistent between different node hardware units, as is the case with symmetric two-sided bidirectional ranging protocols that use a poll-response scheme, a system performing method S100 does not require precise calibration of each node's receive chain delay, transmit chain delay, or intervening processing delay to accurately calculate the time bias and propagation time between node pairs. Furthermore, these delays need not be fixed and can vary between measurements, in which case the system can use the measured self-reception delay to remove any systematic delay offset due to variable receive chain delay, transmit chain delay, or intervening processing delay.

1.7.2 Synchronization Signals In general, nodes in a mesh network can communicate (with each other or with a central server) to establish synchronization signals corresponding to each synchronization slot of the synchronization protocol. More specifically, the synchronization signal for each synchronization slot is predetermined according to the synchronization protocol. Thus, each node can access and/or generate a template signal that is cross-correlated (i.e., cross-correlated via a bit matched filter or an I/Q matched filter) with the synchronization signal received from the other node.

In one embodiment, the system generates a deterministic and unique synchronization signal for each synchronization slot according to a predetermined pattern. Additionally, the synchronization signal may include information designating the node pairs with which the synchronization signals are exchanged (for both self-received signals and for reception of synchronization signals transmitted between node pairs). In another embodiment, the agreement signal is static across a synchronization slot.

In general, the synchronization signal may comprise a carrier signal modulated (e.g., by quadrature modulation) by a complex-valued baseband signal (e.g., a phase shift keying signal, a frequency shift keying signal) generated from a known code sequence to improve the signal-to-noise ratio of the signal when it is received by the nodes in the mesh network. More specifically, the system comprises a carrier signal characterized by a carrier frequency f _c modulated by a baseband signal that may comprise a code sequence of transmitter chip period T′ _c . However, the system may generate a synchronization signal characterized by any code sequence, provided that the code sequence is known to both nodes of a node pair executing a synchronization protocol.

Furthermore, the system can receive and calculate synchronous TOAs and/or POAs transmitted over multiple simultaneous or consecutive carrier frequencies (e.g., multi-carrier or FHSS signals). More specifically, nodes in the mesh network can receive FHSS synchronization signals that contain multiple narrowband frequency components in order to extract the POA for each carrier frequency and thus calculate precise time biases and propagation times between pairs of nodes.

1.7.2.1 Baseband Signals In general, the system generates a synchronization signal that includes a baseband signal that further includes a code sequence. More specifically, the system can receive a signal that includes a pseudorandom binary code sequence, such as a maximum length code sequence (hereinafter "MLS"), a Gold code sequence, a Kasami code sequence, a Barker code sequence, or any other binary code sequence. In one embodiment, the system can receive and time stamp a signal that includes a code sequence that includes a constant amplitude, zero autocorrelation waveform to enhance the sharpness of peaks generated in the magnitude response of the cross-correlation function. For example, the system can generate a synchronization signal that includes a modulated MLS or FSK modulated code sequence.

The system can calculate the TOA and POA of signals within shorter signal durations for signals with the desired autocorrelation characteristics (e.g., high autocorrelation at zero delay and low autocorrelation elsewhere). The system can be of the following form:

のベースバンド信号ｓ（ｔ）を有する信号を受信することができ、ここで、Ｔ’_ｃは、送信機チップ周期であり、ｂ_ｎ（例えば｛－１，１｝内）は、既知のコードシーケンス

A signal having a baseband signal s(t) of T′ c is received, where T′ _c is the transmitter chip period and b _n (eg, in {−1,1}) is a known code sequence.

であり、ｇ’（ｔ）は、チップ波形（方形パルスまたはその他の任意のパルス形状）である。したがって、信号の持続時間は、ＮＴ’_ｃである。

and g'(t) is the chip waveform (rectangular pulse or any other pulse shape). The duration of the signal is therefore _NT'c .

In one embodiment, the system can generate a synchronization signal that includes a fixed preamble (e.g., a bit sequence that does not change between synchronization signals), a variable sync word, and a data payload. In one example, the system can generate a synchronization signal that includes the code sequence b _n as a combination of the fixed preamble and the variable sync word. In another example, the system can generate a synchronization signal that includes the code sequence b _n only within the variable sync word. In yet another example, the system can generate a synchronization signal that includes the code sequence b _n within a predetermined portion of the data payload.

In another embodiment, the system can calculate (e.g., by brute force simulation) a code sequence (within a larger data-carrying portion of the synchronization signal) that maximizes an autocorrelation peak ratio, defined as the ratio of the magnitude of the largest peak to the magnitude of the second largest peak in the autocorrelation function of the signal. More specifically, the system can generate the synchronization signal such that the autocorrelation of the synchronization signal is characterized by an autocorrelation peak ratio that is greater than a threshold autocorrelation peak ratio.

Alternatively, the system may generate a synchronization signal based on a continuous and/or complex-valued baseband signal, as opposed to a binary code sequence. For example, the system may generate a synchronization signal that includes a Zadoff-Chu sequence.

1.7.2.2 Carrier Signal In general, the system can generate a synchronization signal that includes a carrier signal modulated by a code sequence. More specifically, the system can receive a signal that includes a carrier signal characterized by a carrier frequency generated based on a local oscillator of the transmitting node. Thus, each node can transmit a synchronization signal to the other node of a node pair via a given operating frequency.

In one embodiment, the system may generate a synchronization signal characterized by a set of carrier signals. In one example, the system may generate a synchronization signal that defines a multi-carrier signal, such as an OFDM signal, over multiple carrier frequencies. In an alternative example, the system may generate a synchronization signal that defines an FHSS signal that hops between multiple carrier frequencies. Thus, in this embodiment, the system identifies, for each carrier frequency of the synchronization signal, the POA of the synchronization signal received from the receiving node of the node pair relative to a phase reference (e.g., the POA of the self-received signal) calculated from the self-received signal received from the transmitting node.

Additionally or alternatively, the system can refine the calculation of the TOA of the signal based on a phase value corresponding to a cross-reference peak of a matched filter of a synchronous waveform with the measurement signal. Thus, the system can use a synchronous signal characterized by any type of modulation, such as amplitude modulation, frequency modulation, or phase modulation.

In one embodiment, for nodes in a mesh network to receive a synchronization signal via standard RF transceiver techniques, a local node can generate a synchronization signal from the above-mentioned baseband signal by mixing the baseband signal with a local oscillator tone at the transmit carrier frequency (i.e., upconverting the baseband signal). Upon receiving this upconverted signal, a remote node can downconvert the receive passband to a receive baseband signal by filtering the receive passband signal and mixing it with a local oscillator tone at the receive carrier frequency (which may be different from the transmit carrier frequency as in frequency division multiplexing).

1.7.3 TOA and POA Detection In general, the system (e.g., a receiving node or a server receiving a sample of a synchronization signal or a self-received signal) can extract the TOA and/or POA based on the received synchronization signal or the self-received signal by implementing a matched filter, such as a bit matched filter or an I/Q matched filter, based on a template of the synchronization signal. In this way, the system can calculate the cross-correlation between the template signal of the synchronization signal and the received baseband sample. The system can then identify a peak in the cross-reference to obtain the TOA of the synchronization signal at the receiving node. The system can also extract the phase of the synchronization signal at the receiving node (or the phase of the self-received signal at the receiving node) based on the phase response of the matched filter output.

1.7.4 Synchronization Protocol In general, each node generates a synchronization signal (e.g., based on a known synchronization signal template of a synchronization slot), transmits the synchronization signal to the other node of the node pair simultaneously or consecutively, receives a self-received signal based on the transmitted synchronization signal, and receives a synchronization signal from the other node of the node pair. For ease of explanation, the nodes of the node pair are referred to as _n1 and _n2 . However, _n1 and _n2 can refer to any two-node pair in the mesh network. The method S100 includes calculating a relative time bias and propagation delay τ between _n1 and _n2 , represented as ( _b1 - _b2 ). Based on the reciprocity theorem of electromagnetics, the propagation delay of a signal transmitted from _n1 to _n2 is equal to the propagation delay of a signal transmitted from _n2 to _n1 ( _τ1,2 = _τ2,1 ≡ τ).

When calculating the TOA, such as in blocks S140 and S160, the node may cross-correlate the received synchronization signal or the self-received signal with the template signal to determine a timestamp corresponding to the peak value of the autocorrelation function. In one embodiment, each node performs a digital autocorrelation between the received signal and the template signal. Alternatively, each node performs an analog conversion of the digital synchronization signal and an analog autocorrelation between the template signal, the latter allowing for the values of the samples in addition to the period between samples. Additionally, the node may refine the calculated TOA based on phase information it extracts from the synchronization signal. In one embodiment, the node may transmit each synchronization signal over multiple frequency bands, as described further below, to increase the number of carrier phase measurements obtained by the receiving node, thereby increasing the accuracy of the TOA calculation.

In block S110A, at a local time T ₁ (i.e., the first transmission time), n ₁ generates a baseband synchronization signal in an FPGA or DSP. In one embodiment, n ₁ upconverts the baseband synchronization signal to a carrier frequency synchronization signal for transmission via an antenna, and experiences a transmission chain delay t ₁ by propagating the carrier synchronization signal along n ₁ 's transmission chain. The analog carrier signal then interacts with the self-reception hardware, which reflects or otherwise couples the synchronization signal back to n ₁ 's Rx port as shown in block S120A. The self-received signal (i.e., the reflected or otherwise coupled back synchronization signal) is a replica of the synchronization signal preserving phase and group delay, although the power of the two signals may differ. In one embodiment, the power of the self-received signal is adjusted through a return coupling mechanism and/or an attenuator so that the voltage level of the self-received signal does not saturate n ₁ 's ADC.

At block S130A, _n1 receives its own received signal as it transmits the synchronization signal. As the self-received signal propagates through _n1 's receive chain, it experiences a receive chain delay _r1 . At block S140A, _n1 calculates the local TOA, _S1,1 , of the self-received signal for that node, which is related to the unknown parameter of interest by the following equation:
_S1,1 = _T1 + _t1 + _r1
Thus, S _1,1 represents the TOA of the self-received signal from n ₁ according to n ₁ 's clock.

As shown in FIG. 5A and blocks S110B, S120B, S130B, and S140B, _n2 transmits a second synchronization signal and calculates the TOA of the self-received signal generated by _n2 by performing steps equivalent to those performed by _n1 in blocks S110A, S120A, S130A, and S140A.
_S2,2 = _T2 + _t2 + _r2
where _T2 is the local time at _n2 (i.e., the second transmission time), _t2 is the transmit chain delay of _n2 , and _r2 is the receive chain delay of _n2 . In an embodiment where the transmit times of each node are offset by a transmission interval Δ, _T2 = _T1 + Δ. Thus, _S2,2 represents the measured TOA of the self-received signal from _n2 according to the clock of _n2 .

5B and block S150A, _n1 receives a synchronization signal from _n2 that propagates through _n1 's receive chain and experiences a receive chain delay _r1 . Then, in block S160A, _n1 calculates the local TOA, _S1,2, of the received synchronization signal from _n2 , which is represented by the following equation:
_S1,2 = _T2 - _b2 + _t2 + τ + _r1 + _b1
Thus, S _1,2 represents the measured TOA of the synchronized signal from n ₂ by the clock of n ₁ .

In blocks S150B and S160B, _n2 receives the synchronization signal from _n1 by performing steps equivalent to those performed by _n1 in blocks S150A and S160A, and calculates the local TOA, _S2,1, of the synchronization signal (at _n2 ), which can be expressed as:
_S2,1 = _T1 - _b1 + _t1 + τ + _r2 + _b2

Thus, _S2,1 represents the measured TOA of the synchronization signal from _n1 by the clock of _n2 .

In block S170, one of the nodes (e.g., the leader node of the node pair), either _n1 or _n2 , or another computing device, such as a remote server or separate leader node, collects the TOAs, _S1,1 , _S1,2 , _S2,2 , and _S2,1 , and calculates τ and _b1 - _b2 + _r1 - _r2 , or the relative time bias plus the difference in receiver chain delays. By subtracting the TOA of the self-received signal and the TOA of the corresponding synchronization signal received by the other node, the following equation is derived:
_S1,2 - _S2,2 = τ + _r1 + _b1 - _r2 - _b2
S2,1 - _S1,1 = τ + _r2 + _b2 - _r1 - _b1
Therefore, this system is

として計算することができ、ｂ_１－ｂ_２＋ｒ_１－ｒ_２は、

and b ₁ −b ₂ +r ₁ −r ₂ is

として計算される。

It is calculated as:

Although the value of the pure relative time bias b ₁ -b ₂ cannot be calculated without the value of r ₁ -r ₂ , adjusting the clock of one node of a node pair by b ₁ -b ₂ +r ₁ -r ₂ ensures that any signal received simultaneously by both nodes receives the same timestamp at each node, even though the instantaneous points in time at each clock differ by r ₁ -r ₂ . Alternatively, if r ₁ -r ₂ ≈ 0, then

である。

It is.

After calculating τ and b ₁ -b ₂ +r ₁ -r ₂ , method S100 may include synchronizing time between n ₁ and n ₂ by adding b ₁ -b ₂ +r ₁ -r ₂ to the clock of n ₁ or n ₂ to compensate for the initial time bias. In an embodiment with multiple nodes in the mesh network, one node is designated as the "leader node" and the other nodes are designated as "follower nodes." Thus, in this embodiment, method S100 includes adjusting the clocks of the follower nodes to match the leader node. Alternatively, a leader node or a remote server in the system may track the relative time bias of each pair of nodes in the mesh network to compensate for the calculated time bias when executing processes that rely on precise time synchronization between nodes in the mesh network, such as time-based communication protocols or multilateration of other RF devices.

In one embodiment, each node in the mesh network can subtract these variables to calculate the time bias and propagation delay when the time bias measurements between various pairwise combinations of nodes are overdefined. For example, in a network including nodes _n1 , _n2 , and _n3 , the time bias between _n1 and _n2 can be calculated as follows:

Additionally, method S100 may also include indirectly calculating the time bias between nodes that do not have a direct communication line (e.g., due to a communication obstruction between the two nodes, etc.) For example, in a network including nodes _n1 , _n2 , and _n3 , the time bias between _n1 and _n3 may be calculated as follows:

したがって、ｎ_２およびｎ_３は、方法Ｓ１００の別の反復を実行することができ、このシステムは、第１の時間バイアス（ｂ_１－ｂ_２）と第２の時間バイアス（ｂ_２－ｂ_３）の和に基づいてｎ_１の第１のクロックとｎ_３の第３のクロックの間の時間バイアスを計算することができる。

Thus, n ₂ and n ₃ can perform another iteration of method S100, and the system can calculate a time bias between the first clock of n ₁ and the third clock of n ₃ based on the sum of the first time bias (b ₁ -b ₂ ) and the second time bias (b ₂ -b ₃ ).

Additionally, method S100 may include calculating an instantaneous uncertainty δ(b ₁ −b ₂ ) in the relative time bias between the nodes and an instantaneous uncertainty δτ in the propagation delay. Sources of uncertainty may include propagated uncertainties based on the width of the peaks in the autocorrelation function of each TOA calculation, as well as any expected phase noise that may occur in the clocks of each node between synchronization slots.

1.7.4.1 Phase-Based Synchronization Protocol As shown in Figure 1D, the system can implement (through communication with a node pair) a phase-based variant of the synchronization protocol in which each node detects the synchronization signal at _n1 and _n2 and the POA of its own received signal at _n1 and _n2 (based on a template signal that matches the synchronization signal) in blocks S167, S168, S147, and S148, respectively. Thus, the system can measure the relative phase delay in block S172 between the POA of the synchronization signal received at the other node of the node pair and the local phase reference provided by the POA of the own received signal, and use this phase information to calculate the time bias and propagation delay between _n1 and _n2 . Furthermore, the system can detect a set of POAs for each synchronization signal and self-received signal, including a first self-received signal received at n ₁ , a first synchronization signal received at n ₂ , a second self-received signal received at n ₂ , and a second synchronization signal received at n ₁ , by performing a phase-based variation of the synchronization protocol for each carrier frequency of the synchronization signal. Upon receiving the first self-received signal, n ₁ detects the carrier phase of the first self-received signal, which can be expressed as:
φ _1,1,m = (2πf _c,m T ₁ + φ _1,Tx,m + φ _1,Rx,m ) mod 2π
where f _c,m represents the carrier frequency, φ _1,Tx,m represents the frequency dependent phase offset of the n ₁ transmit chain, and φ _1,Rx,m represents the frequency dependent phase offset induced by the n ₁ receive chain. Thus, for a set of carrier frequencies f _c,1 to f _c,M , the system can generate a set of phase references (f _c,1 ,φ _1,1,1 ) to (f _c,M ,φ _1,1,M ) for the self-received signal received at n _1. Similarly, n ₂ detects the carrier phase of the second self-received signal, which can be expressed as:
_φ2,2,m = (2πfc _{, mT2} + _φ2,Tx,m + _φ2,Rx,m ) mod 2π
where f _c,m represents the carrier frequency, φ _2,Tx,m represents the frequency dependent phase offset induced by the n ₂ transmit chains, and φ 2,Rx,m represents the frequency dependent phase offset induced by the n ₂ receive chains. Thus, for a set of carrier frequencies f _c,1 through f _c,M , the system can generate a set of phase references (f _c,1 , φ _2,2,1 ) through (f _c,M , φ _2,2,M ) for the self-received signal received at n ₂ .

The first set of phase/frequency points (f _c,1 , φ _1,1,1 ) through (f _c,M , φ _1,1,M ) and the second set of phase/frequency points (f _c,1 , φ _2,2,1 ) through (f _c,M , φ _2,2,M ) represent phase references for the phase-dependent frequency offset that the synchronization signal experiences as it interacts with the hardware of nodes n ₁ and n ₂ .

Furthermore, _n1 can detect the carrier phase of the synchronization signal it receives from _n2 , which can be expressed as:
φ _1,2,m = (2πf _c,m (T ₂ -b ₂ +b ₁ +τ) + φ _2,Tx,m + φ _1,Rx,m ) mod 2π
Thus, for a set of carrier frequencies f _c,1 to f _c,M , the system can generate a _first set of POAs (f _c,1 , φ _1,2,1 ) to (f _c,M , φ _1,2,M ) for the synchronization signal received at n 1.

Similarly, _n2 can detect the carrier phase of the synchronization signal it receives from _n1 , which can be expressed as:
φ _2,1,m = (2πf _c,m (T ₁ -b ₁ +b ₂ +τ) + φ _1,Tx,m + φ _2,Rx,m ) mod 2π
Thus, for a set of carrier frequencies f _c,1 to f _c,M , the system can generate a _second set of POAs (f _c,1 , φ _2,1,1 ) to (f _c,M , φ _2,1,M ) for the synchronization signal received at n 2.

To calculate τ for node pairs _n1 and _n2 , the system may calculate a first set of phase/frequency points (f _c,m ,φ _2,1,m -φ 1,1,m ) by calculating the phase difference between the phase reference of each carrier frequency of the first synchronization signal and the POA of each carrier frequency of the first synchronization signal. Similarly, the system may calculate a second set of phase/frequency points (f _c,m ,φ _1,2,m -φ _{2,2 ,m} ) by calculating the phase difference between the phase reference of each carrier frequency of the second synchronization signal and the POA of each carrier frequency of the second synchronization signal. The relationship between phase and frequency for the first and second sets of phase/frequency points may be expressed as follows:
φ _2,1,m - φ _1,1,m
= (2πf _c,m (T ₁ -b ₁ +b ₂ +τ) + φ _1,Tx,m + φ _2,Rx,m - 2πf _c,m T ₁ -φ _1,Tx,m -φ _1,Rx,m ) mod 2π
= (φ _{2, Rx, m} + 2π _{f c, m} (τ - b ₁ + b ₂ ) - φ _{1, Rx, m} ) mod 2π
φ _1,2,m - φ _2,2,m
= (2πf _c,m (T ₂ -b ₂ +b ₁ +τ) + φ _2,Tx,m + φ _1,Rx,m - 2πf _c,m T ₂ -φ _2,Tx,m -φ _2,Rx,m ) mod 2π
= (φ _{1, Rx, m} + 2π _{f c, m} (τ - b ₂ + b ₁ ) - φ _{2, Rx, m} ) mod 2π
Thus, the system can sum ( _φ2,1,m - _φ1,1,m ) and ( _φ1,2,m - _φ2,2,m ) for each carrier frequency fc _,m to generate a set of total phase/frequency points (( _φ2,1,m - _φ1,1,m ) + ( _φ1,2,m - _φ2,2,m ), fc _,m ). The system can then calculate the value of τ by the following formula:

したがって、合計位相／周波数点の線形回帰（例えば位相／周波数点の各２πラップアラウンドについての周期的線形回帰）（（φ_{２，１，１}－φ_{１，１，１}）＋（φ_{１，２，１}－φ_{２，２，１}），ｆ_ｃ，１）から（（φ_{２，１，Ｍ}－φ_{１，１，Ｍ}）＋（φ_{１，２，Ｍ}－φ_{２，２，Ｍ}），ｆ_ｃ，Ｍ）を計算することにより、このシステムは、上述の同期プロトコルによるサブサンプル確度でτの値を計算することができる。さらに詳細には、このシステムは、τを計算するために、概算で４πτに等しい周期的線形回帰の傾きを抽出することができる。

Thus, by calculating a linear regression of the total phase/frequency points (e.g., a periodic linear regression for each 2π wraparound of the phase/frequency points) (( _φ2,1,1 - _φ1,1,1 ) + ( _φ1,2,1 - _φ2,2,1 ), _fc,1 ) to (( _φ2,1,M - _φ1,1,M ) + ( _φ1,2,M - _φ2,2,M ), fc _,M ), the system can calculate a value of τ with sub-sample accuracy according to the synchronization protocol described above. More specifically, the system can extract the slope of the periodic linear regression, which is approximately equal to 4πτ, in order to calculate τ.

Thus, in one embodiment, a remote server or reader node executing the blocks of method S100 may receive from _n1 the phase of the first self-received signal by _n1 after transmission of the synchronization signal by _{n1 at T1 by its clock, receive from n2 the phase of} the synchronization signal from _n1 by _n2 after transmission of the synchronization signal by _n2 at _T2 by its clock, receive _{from n2 the} phase of the second self-received signal by _n2 's clock, and receive _{from n1} the phase of the synchronization signal from _n2 by its clock. The remote server or reader node may then refine the first propagation delay based on the phase of the first self-received signal, the phase of the first synchronization signal, the phase of the second self-received signal, and the phase of the second synchronization signal.

Alternatively, the system may calculate the time bias between n1 and n2 by generating a set of subtracted phase/frequency points based on a first set of phase/frequency points (fc _,m , _φ2,1,m - _φ1,1,m ) and a second set of phase/frequency points (fc _,m , _{φ1,2 ,m} - _{φ2,2 ,m} ). More specifically, the system may calculate the time bias based on the following formula:
( _φ2,1,m - _φ1,1,m ) - ( _φ1,2,m - _φ2,2,m )
= (φ _{2, Rx, m} + 2πf _{c, m} (τ-b ₁ + b ₂ ) - φ _{1, Rx, m} )
−(φ _1,Rx,m + 2πf _c,m (τ-b ₂ +b ₁ )−φ _2,Rx,m ) mod 2π
( _φ2,1,m - _φ1,1,m ) - ( _φ1,2,m - _φ2,2,m ) = 2φ2 _{,Rx,m -} 2φ1 _,Rx,m + 4πf _c,m ( _b2 - _b1 ) mod 2π

Therefore, by calculating the linear regression of the subtracted phase/frequency points (f _c,m , (φ _2,1,m -φ _1,1,m ) - (φ _1,2,m -φ _2,2,m )) and extracting the slope of the linear regression, the system can calculate b ₂ -b ₁ with sub-sample accuracy.

1.8 Clock Adjustment After the system executes the synchronization protocol and/or phase refinement for a node pair, the system may shift the clock of one node of each node pair by the calculated time bias to synchronize the clocks of the node pairs. Alternatively, as shown in FIG. 2, the system may maintain a time bias of each node relative to a leader node or other time standard (e.g., a clock of a remote server) and correct the scheduling time sent to and/or timestamps received at each node based on the node's most recently calculated time bias relative to a leader node or time standard, such as by subtracting the node's time bias from any timestamps calculated by the node. Additionally, the system may predict the current time bias for each node based on the node's most recently calculated time bias and a predictive drift model (discussed further below).

1.9 Refining Synchronization Slots At block S180, the method S100 includes calculating the duration of subsequent synchronization slots between previously synchronized nodes. When a node pair executes the method S100, their clocks may be aligned to within 1 to 10 nanoseconds. Thus, as the synchronization between the two nodes improves, the synchronization slots may be made successively smaller.

The method S100 may also include calculating the duration of the subsequent synchronization slot based on a combination (e.g., uncertainty propagation sum) of the time bias uncertainty, the expected drift of the relative time bias between synchronization slots (calculated by a predictive drift model described further below), the propagation delay, the propagation delay uncertainty, the greater of the first node's transmit chain delay or the second node's transmit chain delay, the greater of the first node's receive chain delay or the second node's receive chain delay, the duration of the synchronization signal, and/or the expected change in propagation delay due to relative movement between the node pair.

In one embodiment, the synchronization slot is further extended by a time buffer to ensure that each node receives the complete synchronization signal within the synchronization slot. When the synchronization slot duration is set to be equal to the above sum, the synchronization slot duration is long enough to allow a synchronization signal generated by one node to be completely received by a second node in applications that experience unexpected drift or error in any of the above quantities.

In one example, the receive chain delay and the transmit chain delay may be negligible compared to the synchronization signal duration. Thus, method S100 may include calculating the synchronization slot duration by adding a time buffer of a duration close to the maximum sum of the normal receive chain delay and the normal transmit chain delay.

In one embodiment, after a synchronization failure, such as due to an unexpected drift in the time bias between nodes or a large increase in the propagation delay, the method S100 can include inserting a synchronization slot into the subsequent TDMA frame and increasing the synchronization slot duration to provide additional time for transmitting and receiving the synchronization signal. The method S100 can increase the synchronization slot duration incrementally (e.g., in 10 microsecond increments) until a synchronization process is received, or can include extending the synchronization slot duration to a significantly longer duration to increase the likelihood of receiving the signal.

In this manner, the remote server or reader node executing the blocks of method S100 can decrease the synchronous slot duration of the subsequent synchronous slot corresponding to the node pair such that the updated synchronous slot duration is less than the initial synchronous slot duration but greater than the first uncertainty of the first time bias between the clocks of the node pair plus the synchronous signal duration. The remote server or reader node can then schedule transmission times for each node of the node pair within the updated synchronous slot duration.

Furthermore, the system can schedule the transmission time of the synchronization signal for a node pair based on a recently calculated time bias between the node pairs. In this way, the system can account for the relative time bias between the nodes and ensure that both nodes are transmitting during the next synchronization slot even if the duration of the synchronization slot is decreasing.

1.10 Predictive Drift Model In one variation of method S100, the leader node or remote server executing block S170 may also execute a predictive drift model based on characterization of continuous time bias and propagation delay between nodes in the mesh network. In general, the predictive drift model characterizes the drift of a node's clock (e.g., a crystal oscillator clock) relative to the clocks or time standards (e.g., UTC) of other nodes in the network as a function of environmental factors such as temperature, humidity, movement, and vibration. More specifically, method S100 may include recording environmental data at each node and observing the relative drift calculated over multiple previous synchronization slots to train a predictive drift model of the drift of a particular node in the mesh network. The predictive drift model outputs a predicted drift of a node in the mesh network relative to the time at the leader node or remote server based on the time elapsed since the node's most recent synchronization slot and recent environmental data collected at the node.

The system can maintain predictive drift models that characterize the drift of each pair of nodes in the mesh network relative to each other. Alternatively, the leader node and/or a remote server can compress the pairwise predictive drift models and instead maintain one predictive drift model for each node in the mesh network relative to the clock of the leader node in the mesh network or a time standard accessed at a remote server.

In one embodiment, method S100 includes periodically adjusting the node's clock time between synchronization cycles according to the output of the predictive drift model (e.g., updating the node's clock value every millisecond according to the node's predictive drift). Alternatively, the system may calculate an estimate of the relative time bias between synchronization cycles for a node based on the node's predictive drift model.

Additionally, method S100 may include adjusting the duration of the subsequent synchronization slot according to the output of the predictive drift model. Additionally, method S100 may include triggering the inclusion of a synchronization slot in the TDMA frame in response to the predicted drift value being greater than a predefined drift threshold (e.g., triggering the inclusion of a synchronization slot in the TDMA frame when the predictive drift model predicts a drift between the follower node's clock and the leader node's clock of greater than 1 nanosecond).

The predictive drift model may be a combination of a set of physical models representing relevant environmental parameters. In one embodiment, the predictive drift model includes a temperature model of drift as a function of temperature and crystal cut. The temperature model may include a set of suitable models, each relating drift (in parts per million) to temperature for a corresponding crystal cut. In one embodiment, the method S100 may include sorting the crystal oscillator of each node according to observed temperature drift over multiple synchronization slots to select a polynomial approximation of the temperature drift associated with the particular node as a function of time.

The predictive drift models may also include predictive models for temperature hysteresis, ambient pressure, humidity, electric and/or magnetic field strength, crystal oscillator drive level, and/or crystal oscillator reference voltage.

In one embodiment, the predictive drift model can output a confidence interval for the predicted drift based on the randomly distributed phase noise and systematic drift of the crystal oscillator. The predictive drift model can calculate the distribution of the crystal oscillator's phase noise based on factors such as reference source noise, power supply noise, vibration induced noise, and/or acceleration induced noise.

In this manner, the system may perform successive iterations of method S100 for a node in the mesh network and record a time series of relative time bias between the node and a reference time, which may be a time maintained at a leader node or a remote server. At the same time, the system may record a time series of environmental data from the node. The system may then correlate the time series of environmental data with the time series of time bias to predict drift in the node's time bias based on changes in the environmental data. In this manner, the system may calculate a temperature correlation between the temperature at a particular node and the drift rate exhibited by the node based on a time series of temperature data and a time series of the node's time bias. Additionally or alternatively, the system may also calculate a movement correlation between the movement of the node (e.g., acceleration measured by an IMU at the node) and the drift rate exhibited by the node based on a time series of the node's movement data and a time series of the time bias.

In yet another embodiment, the system can detect node movement by recording a time series of propagation delays between one node in the mesh network and another node in the mesh network. Thus, the system can calculate a first relative position of the node based on a first propagation delay calculated during a first iteration of method S100, calculate a second relative position of the node based on a second propagation delay calculated during a second iteration of method S100, and calculate a movement correlation of drift of the node's clock based on the first relative position, the second relative position, the first time bias, and the second time bias.

In one embodiment, the system can calculate a time bias uncertainty for a recently calculated time bias of a first node relative to a second node, predict a magnitude of drift in the time bias at the node based on a predictive drift model, predict a change in propagation delay between the first node and the second node based on motion data from the first node and the second node, and set a synchronization slot duration equal to the sum of the time bias uncertainty, the drift magnitude, the propagation delay, and the change in propagation delay. Thus, the system can ensure reception of subsequent synchronization signals transmitted between node pairs by accounting for relative drift between node pairs in a mesh network.

1.11 Triggering Resynchronization One variation of method S100 includes triggering the inclusion of one or more synchronization slots in a subsequent TDMA frame based on the output of a predictive drift model or detection of an environmental change. In one embodiment, method S100 includes triggering the inclusion of a synchronization slot when a predictive drift model outputs a predicted drift magnitude greater than a threshold drift value (e.g., drift greater than 10 nanoseconds since the last synchronization slot). Alternatively, method S100 may include triggering the inclusion of a synchronization slot based on observed temperature changes (e.g., via a digital thermometer at the node) and/or acceleration or vibration data (e.g., via IMU data collected at the node). Additionally, method S100 may include triggering the inclusion of a synchronization slot based on movement of one node relative to another node. Method S100 may include detecting movement via an IMU at the node, a measurement of Doppler shift on an incoming signal from the node, or multilateration between nodes.

In one example, after an initial iteration of method S100, the system can measure a temperature at a first node of a node pair, calculate a time bias drift based on the temperature and a temperature drift model, and schedule a synchronization slot in the next TDMA frame based on the time bias drift.

Generally, when method S100 triggers the inclusion of a synchronization slot, the synchronization slot is included in the subsequent TDMA frame and indicated in the header of the TDMA frame.

2. Method of locating a transmitter As shown in FIG. 6A, a method S200 of locating a device through a network includes, at each node of each unique node pair in the network, transmitting an outbound synchronization signal at block S210, generating a self-received signal based on the outbound synchronization signal at block S220, detecting the self-received signal at one of the self-received time of arrival (hereinafter "TOA") pairs at block S222, and detecting an inbound synchronization signal transmitted from a partner node of the unique node pair at one of the synchronous TOA pairs at block S212. The method S200 also includes, for each unique node pair in the network, calculating a pairwise time offset between the unique node pair from a set of pairwise time offsets based on the pair of self-received TOAs and the pair of synchronous TOAs at block S230, and calculating a pairwise distance between the unique node pair from a set of pairwise distances at block S232. Method S200 further includes, for each node in the network, calculating a relative position of the node in the network with respect to one node in the network based on the set of pairwise distances, block S240, and calculating a time bias of the node in the network with respect to one node in the network based on the set of pairwise time offsets, block S242. Method S200 also includes, at each node in the network, detecting a positioning signal transmitted from the device at the positioning TOA, block S250, and calculating, for each node in the network, a position of the device with respect to the network based on the positioning signal detected at the node, the node's time bias, and the node's relative position, block S260.

As shown in FIG. 6B, a first variant of method S200 includes, at each node in a network characterized by a position relative to the network and a clock synchronized with respect to the network clock, receiving a carrier on the multiplexed channel in block S252, demodulating the carrier to detect an identification signal transmitted from the device in block S254, detecting in the carrier a set of location signals resulting from multipath propagation of the carrier in block S256, and, for each location signal in the set of location signals, calculating a location TOA of that location signal in a set of location TOAs in block S258. This first variation also includes, for each node in the network, selecting a line-of-sight (hereinafter "LOS") TOA from the node's location TOA, in block S262, calculating a set of time difference of arrival (hereinafter "TDOA") based on the LOS TOA for each node, in block S264, and calculating a device location based on the set of TDOA, in block S266.

6C, a second variation of method S200 includes, at each node in the network, transmitting an outbound synchronization signal in block S210, generating a self-received signal based on the outbound synchronization signal in block S220, detecting the self-received signal at the self-received TOA in block S222, and for each inbound synchronization signal of the set of inbound synchronization signals received from other nodes in the network in block S212, detecting the synchronization TOA of the inbound synchronization signal in the set of synchronization TOAs. This second variation also includes, for each node in the network, calculating a time bias of the node relative to a node in the network in block S242 based on the self-received TOA and the set of synchronization TOAs, and calculating pairwise distances between the node and each other node in the network in the set of pairwise distances in block S232. This second variation further includes calculating, in block S240, a relative position of the nodes in the network based on the set of pairwise distances; detecting, in block S250, a positioning signal transmitted from the device at each node in the network; calculating, in block S268, for each node in the network, a TDOA for that node based on the positioning signal detected at each node; and calculating, in block S270, a position of the device relative to the network based on the relative positions of the nodes in the network, the TDOA of each node in the network, and the time bias of each node in the network.

2.1 Application In general, method S200 is performed by a system including a network (e.g., a mesh network) of nodes and/or remote servers to estimate the location of a radio frequency (hereinafter "RF") transmitting device relative to the node network, such as to within 30 centimeters. The system can estimate (or "locate") the location of an RF transmitting device within the RF range of the node network without prior information regarding the node's location, without requiring the nodes in the network to be stationary, without precise calibration of the nodes (e.g., time calibration, gain calibration, and/or frequency calibration), and with each node including a standard electronic clock (e.g., a crystal oscillator clock). Each node in the network can include networking hardware such as antennas, transceiver hardware, FPGA/DPS, clocks, and self-receiving signal generators (e.g., impedance-mismatched directional couplers, RF power splitters, combiners, circulators, etc.), as described further below.

The system can execute the blocks of method S200 to locate RF transmitting devices detectable by nodes in the network (e.g., at least three nodes for two-dimensional localization, at least four nodes for three-dimensional localization). More specifically, the system can identify and locate third party devices, such as RFID-enabled devices, ZIGBEE-enabled devices, BLUETOOTH-enabled devices, WIFI-enabled devices, and/or LTE-enabled devices, without requiring modification of the transmission protocol implemented by the device. Additionally or alternatively, the system can locate specially designed devices (e.g., active tags) configured to transmit specific location signals to nodes in the network.

To obtain sub-meter location accuracy, the system performs a calibration and synchronization process to obtain nanosecond-level clock synchronization between the nodes in the network by identifying the time offset between each unique pair of nodes in the network. The system can then calculate the relative time bias of the nodes in the network and compensate for the calculated time bias. By performing a calibration and synchronization process, the system can also calculate the propagation delay and therefore the distance between each unique pair of nodes in the network. Once the system calculates the distance between the nodes, it can establish the relative position of each node in the network, provided there are enough nodes in total (e.g., at least five nodes to obtain a three-dimensional relative position). In this way, if the global position of one of the nodes is known, the system can determine the global position of all other nodes in the network.

The system can then leverage the calculated location information and time synchronization of each node in the network to precisely locate any RF transmitting device within the RF range of the network. The system can uniquely identify an RF transmission from a device (e.g., an RF transmission in the form of a carrier wave) as it is received at each node in the network. In one implementation, each node can precisely calculate the TOA of the RF transmission as it propagates to each node. The system can then compare the TOA from each node in the network to calculate the TDOA of the received signal between each node. Alternatively, the node can transmit a portion of the received signal to a remote server and/or one of the nodes in the network (e.g., a leader node) and determine the TDOA of the received signal by cross-correlation of the received signals. The system can then perform multilateration calculations to estimate the location of the device.

The system can also resolve overlapping repetitions of RF transmissions caused by multipath propagation of the RF transmissions from the device. After resolving the individual multipath signals, the system can determine the TOA or TDOA of each multipath signal individually and then select the LOS TOA or TDOA that represents the LOS RF transmission between the device and each node. By resolving the multipath artifacts in the RF transmissions, the system can better locate the device in a complex propagation environment. The system can perform the above-mentioned multipath detection and rejection by utilizing frequency, time, phase, spatial, and/or orientation diversity that can be applied at the transmitting end (e.g., by specifying the device's RF transmissions through a custom protocol), at the receiving end (e.g., by specialized receiving and processing at each node), and/or at the back end (e.g., in post-processing at the reader node and/or remote server).

A system performing the blocks of method S200 can be deployed for a wide range of applications. In one example, the system is deployed locally in a warehouse, store, medical facility, and/or any other building. A network of nodes of the system can be deployed locally around the building to locate and track asset tracking RF tags (active tags), cell phones, BLUETOOTH, WIFI devices, and the like within the building. In another example, the system can be deployed as a telecommunications network, where each node in the network is a cellular site that can locate, transmit, and receive signals from cellular devices. In yet another example, the system can be deployed as a network of low earth orbit (hereinafter "LEO") satellites that can locate RF transmissions from a large volume of space at or above the surface of the Earth.

2.2 Telecommunications Deployment Generally, the method S200 may be performed by a telecommunications network, such as a cellular network implementing the LTE, 4G, 5G, and 5G NR standards proposed by the Third Generation Partnership Project (hereinafter "3GPP"). For example, the method S200 may be performed by a node in the cellular network, such as a 5G radio node (hereinafter "gNB") and/or enhanced 4G eNodeBs (hereinafter "ng-eNB"), as a component of a server-side location management function (hereinafter "LMF") and/or an access and mobility management function (hereinafter "AMF").

2.3 System A system performing method S200 can include a network (e.g., a mesh network) of nodes, remote servers, and/or active tags or controllable devices (i.e., non-third party devices). The nodes in the network are RF transceivers that perform the blocks of method S200 that involve transmitting or receiving signals between nodes or from devices. In general, blocks of method S200 that do not involve receiving or transmitting signals can be performed on nodes and/or remote servers that can perform the blocks of method S200 via an internet connection with the nodes (i.e., "in the cloud") to reduce the processing load on the nodes. Additionally, the system can include various devices configured to interact with the nodes in order to be located by signals. These "controllable devices" can include smartphones or other transmitters configured to transmit special location signals so that the system can more easily detect and locate the devices.

2.4 Nodes As shown in Figures 3A and 3B, the system includes a network of nodes. The network can include more than one node, but the more nodes included in the network, the higher the accuracy of locating both the nodes and devices within the RF range of the node network. In one embodiment, the network includes three nodes and can locate devices in two-dimensional space. Alternatively, the network includes four nodes and can locate devices in three-dimensional space. In yet another embodiment, the network includes five nodes and can locate each node based on the propagation time measured from each of the other four nodes in the network. Thus, the functionality of the system increases as more nodes are included in the system.

In general, a node includes transmit and receiver components, an FPGA or DSP configured to generate and process signals, a clock, and a self-receiving signal generator. The node transmits and receives information wirelessly and therefore includes RF transceiver hardware such as the superheterodyne radio architecture and Rx/Tx antennas shown in FIG. 3A. In this implementation, each node includes a "receive chain" and a "transmit chain." The receive chain includes a pipeline of hardware components that process signals received from the Rx port. The transmit chain includes a pipeline of hardware components that process the transmit signals generated by the FPGA or DSP and send them to the Tx port. The receive chain and transmit chain impart a "receive chain delay" and a "transmit chain delay," respectively. "Receive chain delay" and "transmit chain delay" refer to the amount of time that elapses as a signal traverses the receive chain or transmit chain, respectively.

In one embodiment, the nodes can also communicate with a remote server over a wired network. In this embodiment, the nodes can include I/O ports and/or appropriate interface converters for communicating over any wired medium (e.g., Ethernet/twisted pair, coax, or fiber optic).

In one embodiment, the node is integrated into an existing transceiver infrastructure, such as a cellular site, that is adapted to perform method S200. The cellular site/tower or other existing transceiver can be adapted to perform method S200 by updating the cellular site's software. In an alternative embodiment, the node can include hardware optimized to improve aspects of method S200.

In one embodiment, each node's FPGA or DSP is configured to generate a complex digital signal and output the generated signal to the DAC. The complex components of the digital signal represent the in-phase and quadrature (i.e., I/Q) portions of the analog signal generated by the DAC. Additionally, the node's FPGA or DSP receives a digital signal from the node's antenna via an ADC, and then timestamps the received synchronization signal according to the instantaneous value of the clock and the TOA calculation process described below.

Each node also includes a clock, such as a crystal oscillator clock or an atomic clock, that may be responsible for time management and time stamping functions at the node. The system may execute method S200 to synchronize the clocks of multiple nodes in the network to locate devices. In particular, the system may effectively synchronize crystal oscillator clocks that meet the fundamental frequency stability, phase noise, and frequency requirements of wireless communications. In one example, the clock is a crystal oscillator with an AT cut and a clock frequency of 10 megahertz (MHz). However, a node may include a crystal oscillator of any frequency or cut, provided that the above constraints are met.

Each node includes a self-received signal generator 110 that generates a self-received signal in block S220. Typically, the self-received signal generator sends an attenuated copy of the synchronization signal being transmitted to another node to the receive port of the transmitting node. By processing the transmitted synchronization signal through the receive chain and determining a TOA and/or phase reference for the self-received signal, each node can timestamp the transmitted signal delayed by the receive chain delay (i.e., the time delay the signal undergoes as it is processed by the receive chain). When the node later receives a synchronization signal from another node, the TOA or POA of the synchronization signal will also have the same receive chain delay. Because both the TOA and/or POA of the self-received signal and the TOA or POA of the synchronization signal include the receive chain delay, they can be directly compared without precise calibration of the receive chain hardware.

In one embodiment, the self-received signal generator is a directional coupler as shown in FIG. 3B. The self-received signal generator may also include a variable impedance circuit, which may be software controlled to vary the gain of the reflected Tx signal to the Rx port. Depending on the specific hardware implementation of the node, similar impedance matching may be applied to a circulator, power splitter, or any other transmission line device. However, the self-received signal generator may include any software or hardware system that sends a Tx signal to the Rx port.

In one embodiment, each node may include a software defined radio architecture that performs the functions of any of the hardware elements described above.

2.5 Devices The system performs the method S200 to locate devices. In general, the devices may include controllable devices and/or third party devices. Although the system can locate any category of controllable devices 106, the system may include certain optimizations that may improve the location of the controllable devices. More specifically, the controllable devices may include any devices manufactured and/or configured (e.g., via firmware or software) to transmit a specific RF signal (i.e., a location signal) that can be detected by nodes in the network. In one embodiment, the controllable devices include "active tags," which may be low-power RF transceivers configured for asset tracking applications. In another embodiment, the controllable devices may include smartphones or any other devices that run software or firmware applications that transmit location signals that are detected by nodes in the network. In yet another embodiment, the controllable devices may also include low-power transmitters that are not capable of receiving signals from the system and that are not capable of operating according to predefined custom transmission protocols.

Third party devices may include any device that generates RF transmissions of sufficiently high power to be received by the nodes in the network. In general, third party devices communicate wirelessly by running standardized wireless protocols such as BLUETOOTH, WIFI, LTE, 5G, and/or any other wireless protocol that can be detected by packet analysis techniques performed by the nodes. Although third party devices may not transmit location signals as defined by this system, any identifiable RF transmission may be considered a location signal of a third party device. For example, the system may detect periodic LTE and/or 5G synchronization sequences received from a particular device as a location signal for that device.

However, the system can locate any RF emitting device that can be detected by a sufficient number of nodes in the network.

2.6 Relative Location of Nodes in a Network Generally, at block S240, the system can calculate the relative location of each node in the network based on the pairwise distance between each unique pair of nodes in the network. Additionally, the system can calculate the global location of each node in the network given the global location of a reference node in the network and the locations of other nodes in the network relative to the reference node. More specifically, a system including at least four nodes can calculate the relative three-dimensional location of each node in the network based on the pairwise distance between each node in the network. Similarly, a system including four nodes can calculate the relative two-dimensional location of each node in the network based on the pairwise distance between each node. In one embodiment, the system includes more than five nodes to improve the accuracy of the relative location calculation through measurement redundancy/overlap determination.

The system can determine the relative position of each node in the network by defining the origin of a coordinate system at a first reference node in the network and defining the axes of the coordinate system through a second reference node. Alternatively, the system can define a coordinate system based on three reference nodes that form a two-dimensional plane. The system can then solve a set of self-consistent equations for pairwise distances in the network. These self-consistent equations are given by: for each unique node pair n _i and n _j in the node network,
(x _i -x _j ) ² + (y _i -y _j ) ² + (z _i -z _j ) ² = d _i,j ²
Since errors in the computation of pairwise distances between nodes and/or possibly in the overlap determination of the system of self-consistency equations make the system of self-consistency equations unsolvable deterministically, the system may solve these self-consistency equations by iterative or probabilistic methods. Additionally, the system may incorporate any additional location information to further constrain the system of self-consistency equations and improve the computation of the location of each node in the network.

However, the system can implement any mathematical technique to convert a set of pairwise distances between nodes in the network into the relative positions of each node in the network.

2.7 Locating a Device In general, the system performs a combination of blocks S250, S252, S254, S256, S258, S260, S262, S264, S268, and/or S270 to locate a device within the RF reception range of a sufficient number of nodes in the network. More specifically, the system may receive a signal in the form of a carrier wave in block S252, disambiguate signals from multiple devices in a crowded wireless environment, uniquely identify a device based on the disambiguated signals in block S254, calculate a set of TDOAs of signals from the device among nodes in the network, and perform TDOA multilateration in block S258 to locate the device. Additionally, the system may detect multipath artifacts in signals from the device and select a LOS signal from the set of multipath signals. Additionally, the system may perform statistical noise removal and TDOA bounding to improve the accuracy of the location determination.

2.7.1 Signal Disambiguation To disambiguate a particular signal from general RF noise and/or other interfering signals, the nodes implement one or more wireless protocols to detect devices also communicating via those protocols. The system may implement variations of standardized wireless protocols, such as various versions of IEEE 802.15.4, BLUETOOTH, WIFI, GSM, CDMA, LTE, and/or 5G protocols, or any other standard wireless communication protocol. Additionally, the system may implement custom wireless protocols specifically designed for device location via the node network. Both standardized and custom wireless protocols may implement one or more forms of multiple access structures, such as TDMA, frequency division multiple access (hereinafter "FDMA"), code division multiple access (hereinafter "CDMA"), or some hybrid or variation thereof. Thus, the system may also implement these multiple access techniques to disambiguate transmissions from different devices. When a node in the network accesses the multiplexed channel, the node records a sample of the carrier transmitted from the device through the node's receive chain and Rx port, so that the node can then demodulate the carrier and identify the signal transmitted from the device, block S254.

The system can locate third party devices and controllable devices such as active tags. Thus, the demodulated signal can include a specific location signal (e.g., a signal from a controllable device) that is specifically configured for the system's calculation of the TOA, TDOA, POA, or phase difference of arrival (hereinafter "PDOA"). Additionally, the controllable device can also be configured to transmit an identification signal in addition to the location signal so that the system can more easily identify the device and associate the device's calculated location with the device's previously calculated location. However, if the demodulated signal is received from a third party device that operates a standardized wireless protocol, the demodulated signal may not include a specific location signal for calculating the signal's TOA, TDOA, POA, or PDOA. Instead, the system can identify and locate the device based on a specifically identifiable and locatable transmission specified by that standardized wireless protocol.

2.7.2 Custom Protocols In one embodiment, the system locates the controllable devices via a custom wireless communication protocol. The custom wireless communication protocol may specify a multiple access method utilized to disambiguate signals from different controllable devices, specify the structure of the identification signal used to establish a unique identity of the controllable device, and specify the structure of the location signal transmitted from the controllable device such that the system can calculate the TOA or TDOA of the location signal at each node in the network. In one embodiment, the custom wireless protocol specifies a tunable wireless protocol such that the system can instruct the controllable device to change the channel (in the multiple access method) and/or the timing, duration, format, or structure of the device's location or identification signal.

2.7.2.1 Structure of the Localization Signal In general, the localization signal of the controllable device comprises a pseudorandom sequence that exhibits a higher peak autocorrelation when aligned compared to when unaligned. In one embodiment, the localization signal is a Zadoff-Chu sequence.

In implementations where the controllable devices are one-way transmitters and cannot receive signals from the system (e.g., certain implementations of active tags), the location signal can include a single static pseudo-random sequence pre-assigned to each controllable device. The controllable devices can then periodically transmit the pre-assigned sequence to allow nodes to locate the controllable devices.

In alternative embodiments in which the controllable device is configured to receive signals from the system, the system may transmit instructions to the controllable device to specify the location signals that the controllable device will transmit. Additionally, the system may communicate synchronization, timing, or channel information to the controllable device to cause the controllable device to transmit at times and on channels specified by the system, thereby enabling nodes in the network to consistently receive location signals from the device in a variety of wireless propagation environments.

In another embodiment, the controllable device is configured to transmit the positioning signal over multiple bands to mitigate the effects of multipath fading and improve the likelihood of detecting and calculating the TOA or TDOA of the LOS signal of each node in the network. For example, the controllable device may be configured to transmit the positioning signal over multiple carrier frequencies by frequency division multiplexing (hereinafter "FDM"). Additionally or alternatively, the controllable device may improve the time diversity of the positioning signal by transmitting the positioning signal in multiple time division multiplexing (hereinafter "TDM") slots.

In yet another embodiment, the controllable devices are configured to transmit a location signal having a structure similar to the synchronization signal described above with reference to block S100 of the method. In this embodiment, each controllable device may transmit a location signal implemented according to any variation of the synchronization signal described above.

2.7.3 Standardized Protocols In one embodiment, the system locates third party devices through standardized wireless communication protocols such as ZIGBEE, BLUETOOTH, WIFI, GSM, CDMA, and/or LTE. Although standardized wireless protocols may not explicitly specify location or identification signals in transmissions between communicating devices, standardized wireless protocols frequently require devices utilizing the protocol to identify the initial communication of the transmitting device and communicate on a specific multiplexed channel so that the devices can consistently transmit signals to the correct device. Additionally, many standardized wireless protocols utilize pseudo-random sequences in the protocol for a variety of purposes beyond location.

The system can therefore implement a TDMA protocol that detects these location identifiable sequences and can locate devices transmitting over standardized wireless protocols by intercepting transmissions that the device would make during normal operation. For example, LTE and 5G protocols specify a pseudorandom (Zadoff-Chu) sequence for the Physical Uplink Control Channel (PUCCH). Nodes in the network can be configured to detect this sequence and use it as a location signal for a particular LTE or 5G device. Similar methods can be applied to other periodically broadcast synchronization and control sequences in other standardized wireless protocols. Thus, the location signal can include standardized wireless protocol communications.

2.7.4 Multiple Protocol Stacks The system may locate devices transmitting according to any of multiple standardized wireless protocols. In one embodiment, the system does not implement a complete protocol stack for each standardized protocol included in the multiple protocol stack. In this embodiment, the system may implement portions of each standardized protocol for demodulating a physical waveform into a symbol or bit stream. Additionally or alternatively, the system may implement portions of the standardized protocol stack related to demultiplexing transmissions from multiple devices. For example, the system may demultiplex signals transmitted by devices transmitting according to standardized wireless protocols by using control/header frames or by exploiting time separation of the transmissions. After demultiplexing the received signals, the system may extract metadata such as protocol descriptions and transmission parameters to identify and/or locate the devices.

2.7.5 Device Identification and Tracking The system can also identify and track various located devices as they move relative to the network. In one embodiment, the system can associate a device with an identifier based on an identification signal transmitted by the device and associate signals subsequently received from the device with that identifier. If the device is running a custom wireless protocol, the protocol can specify a uniquely identifiable location or identification signal such that any transmission made by the device can be identified as having been transmitted from that device. Alternatively, if the devices are running a standardized wireless protocol, the system can implement a multiple access scheme based on each device transmitting identification information in the process of running the standardized wireless protocol.

In an alternative embodiment, the system can interface with a hub or computing device (e.g., a leader node in a BLUETOOTH piconet or an eNode-B cellular site in LTE or 5G) that cooperates with a standardized wireless protocol to identify the multiplexed channels on which various devices are transmitting. The system can then identify any transmissions received on the multiplexed channels at the expected time as originating from the device designated by the wireless hub or leader node.

In yet another embodiment, the nodes in the network themselves may run standardized wireless protocols and locate devices in communication with the nodes. For example, the nodes themselves may operate as eNode-B cellular sites and run LTE or 5G protocols. Thus, each node, in the process of operating the LTE or 5G protocols, will have information detailing the identity and channel information for each device in communication with the node. The system can utilize this information with other nodes to locate the devices.

Once the system identifies a device, it can assign it an internal identifier and associate any calculated or estimated location information for the device with that identifier, allowing the device's location and path to be tracked relative to nodes in the network.

2.7.5.1 Adaptive Protocol and Collision Detection In one embodiment, the system may also track and/or predict clock drift of a device that the system is locating relative to the synchronized clock of the node by identifying transmissions from the device having a device identifier. In general, the system may characterize the clock drift of the device based on the time difference between the location TOAs of subsequently received signals from the device relative to the boundaries of transmission slots defined by the wireless protocol, and modify the first standardized wireless protocol based on the clock drift of the device. Additionally, the system may estimate the expected clock drift of the device based on the location of the device across multiple subsequently received signals, and modify the first standardized wireless protocol based on the expected clock drift of the device.

In this manner, the system can predict when a device will no longer transmit within a TDMA slot specified by the wireless protocol based on the drift rate of the device's clock. In one implementation, the system can adjust the boundaries of the TDMA frame to accommodate drift in the device's clock. Or, if the device is a controllable device, the system can send a synchronization signal to realign the device's clock with the node's clock. In yet another implementation, the system can instruct the device to transmit over a different carrier frequency so that signals from multiple devices can be interpreted within the same TDMA slot.

However, the system may modify the TDMA frame structure in any manner that accommodates collisions and disambiguates colliding signals within one TDMA slot.

2.8 Calculation of TDOA Generally, at blocks S266 and S268, the system can calculate, for each node in the network, the TDOA of that node based on the positioning signal detected at each node. More specifically, the system can calculate the TDOA of the positioning signal either directly from the positioning signal itself (by compensating for the time bias previously calculated for each node) or by first calculating the accurate TOA at each node and subtracting the first TOA at the first node to receive the signal from each of the TOAs for the other nodes. Thus, for each node in the network, the system can adjust the positioning TOA at that node (i.e., the TOA of the positioning signal) of the set of positioning TOAs by the time bias of that node.

In embodiments in which the system calculates TDOA for nodes directly based on received positioning signals, each node may transmit the time-stamped positioning signals received at that node to a reader node or a remote server. The system may then pairwise cross-correlate the positioning signals to generate a set of TDOA for the positioning signals at each node in the system. In these embodiments, the system compensates for systematic offsets (e.g., time bias and frequency offset) between each pair of nodes when calculating the cross-correlation between the positioning signals received at each node.

In an alternative embodiment, the system calculates the TOA of the location signal at each node separately and then subtracts the calculated TOA values to determine a set of TDOA values. In calculating the TOA of the location signal, the receiving node can autocorrelate the received location signal with a template location signal to determine a timestamp corresponding to the peak value of the autocorrelation function. The location signal can include specially chosen sequences that have high autocorrelation values when the sequences are aligned and low autocorrelation values when they are not. In one embodiment, each node performs a digital autocorrelation between the received signal and the template signal. Alternatively, each node performs an analog autocorrelation between an analog conversion of the digital synchronization signal and the template signal, the latter allowing for the period between samples in addition to the values of the samples.

If the system calculates the TDOA of the positioning signals between each unique pair of nodes, the system can perform statistical techniques, such as averaging, to take advantage of the redundant TDOA to improve the TDOA estimate. For example, if there are three nodes _n1 , _n2 , and _n3 , the system can calculate the TDOA between _n1 and _n3 directly by cross-correlation of the positioning signals received at nodes _n1 and _n3 , or by subtracting the TOAs calculated at nodes _n1 and _n3 , or the system can calculate the TDOA between _n1 and _n2 and between _n2 and _n3 and sum them to get another value of the TDOA between _n1 and _n3 . The system can apply statistical techniques to the over-determined TDOA to improve the accuracy of the TDOA calculation.

2.8.1 Phase-Based Location In an implementation in which a device transmits location signals over multiple carrier frequencies, the system may also locate the transmitting device based on the POA of the location signal by recording the carrier phase offset of the received location signal for each transmitted carrier frequency in the FHSS location signal. In general, the system may calculate, at each node in the network, a set of carrier phase offsets for each carrier frequency of the location signal detected at that node to generate a set of carrier phase offsets for the location signal detected at that node, and calculate the device's location relative to the network based on, for each node in the network, the set of carrier phase offsets for the location signal detected at that node, the time bias of that node, and the relative position of that node. More specifically, the system may measure the PDOA of multiple frequencies to improve the location of the device.

2.9 Multipath Detection In blocks S252, S256, S258, S262, S264, and S266, the system may detect in a carrier a set of positioning signals resulting from multipath propagation of the carrier, for each positioning signal in the set of positioning signals, calculate a positioning TOA from a set of positioning TOAs for that positioning signal, for each node in the network, select a LOS TOA from that node's set of positioning TOAs, calculate a set of TDOAs based on the LOS TOA for each node, and calculate a location of the device based on the set of TDOAs.

The system calculates the TOA for the superimposed location signals in a manner similar to the calculation of the TOA for a single location signal. However, instead of performing digital autocorrelation on the demodulated signal, the system can apply a cross-correlation function of a template for an analog signal corresponding to the modulated location signal and an analog signal received at the node. In a multipath environment, the autocorrelation function can output multiple peaks, each corresponding to a TOA of a multipath component of the location signal. A common method for determining the LOS TOA from a set of multipath TOAs is to simply select the first TOA from the set of TOAs detected at the node as the LOS TOA. However, this method is susceptible to artifacts caused by wireless signal propagation, which can lead to inaccurate estimation of the LOS TOA. Thus, the system can utilize various techniques that exploit the location signals across multiple frequency bands to filter out the multipath signals and estimate the LOS TOA from the set of TOAs, exploiting the frequency, time, and/or spatial diversity of the location signals.

2.9.1 Frequency-Based Multipath Detection In one embodiment, the system receives a location signal transmitted over multiple frequency bands. The location signal thus includes a superposition of carriers of different frequencies, which may be frequency modulated versions of the same pseudorandom sequence. Because RF waves of different frequencies propagate differently in the physical environment, the timing of the set of location TOAs resulting from the set of multipath signals received at each node may vary depending on the frequency band in which the location signal transmitted. However, the LOS TOA is the same (within a threshold) across the frequency bands. Thus, the system can filter out location TOAs from the set of location TOAs that are not repeated across a threshold number of frequency bands. The system can determine that two location TOAs between frequency bands are sufficiently simultaneous to be considered "repeated" if they occur within a predetermined threshold period. Additionally or alternatively, the system can determine the LOS TOA by choosing the first location TOA among the set of all location TOAs across the frequency bands.

The system can thus access a frequency division multiplexed channel that includes a set of carriers, each characterized by a different carrier frequency, detect a set of location-specific TOAs for each carrier, compare the sets of location-specific TOAs for each carrier, remove non-repetitive location-specific TOAs within a predetermined TOA threshold to generate a set of remaining TOAs, and select the LOS TOA for each node from the set of remaining TOAs from that node.

2.9.2 MIMO Multipath Detection In one embodiment, the system can achieve spatial diversity by utilizing multiple input/multiple output (hereinafter "MIMO") multipath detection. In this embodiment, the nodes in the network are MIMO and include multiple antennas at different physical locations on the node. The displacement of these antennas relative to each other can be suitably small so that the TOAs of the position location signals are sufficiently similar between the antennas, but the multipath signals received at each antenna are significantly different. In this embodiment, each node records the incoming carriers containing the modulated position location signals at each antenna of the node. The system then calculates a set of position location TOAs resulting from the multipath environment between the device and each antenna. The system can then filter out position location TOAs that are not repeated between antennas in a manner similar to the process described above for frequency-based multipath detection. Alternatively or additionally, the system can select the first position location TOA as the LOS TOA across the set of position location TOAs from all antennas.

2.9.3 Time Domain Multipath Detection In one embodiment, the system can achieve time diversity by utilizing time domain multipath rejection. In this embodiment, the device can transmit positioning signals periodically with a predetermined time offset within a short period of time. As a result, each positioning signal in a sequence of positioning signals may experience different levels of interference in a multipath wireless environment. The system can then subtract the predetermined time offset between each positioning signal to align the positioning signals and calculate a set of positioning TOAs for each positioning signal. The system can then select the first detected positioning TOA from the time separated positioning signals as the LOS TOA.

2.10 TDOA Bounding After calculating a set of TDOAs at each node, the system may also apply TDOA bounding techniques to improve the likelihood of calculating accurate TDOAs between each pair of nodes in the network. The system may implement a bounding model to eliminate TDOAs that are inconsistent with prior information about the location of the nodes or the scale of the area in which the nodes are distributed. Depending on the implementation, the system may perform TDOA bounding before calculating the LOS TOAs or after calculating the LOS TOAs. For example, the system may calculate all possible TDOAs between two nodes (e.g., by calculating the difference between each TOA calculated at the first node and each TOA calculated at the second node) and then eliminate TDOAs outside the bounding function. Alternatively, the system may first select a LOS TOA for each node and then eliminate TDOAs calculated based on those LOS TOAs. In implementations involving multi-band locating signals, the system may apply TDOA bounding to either the TOAs or TDOAs calculated across the multiple bands. In one implementation, the system may exclude TDOAs that correspond to distances greater than the pairwise distance between the two nodes for which the TDOA was calculated. Additionally or alternatively, if the system is locating a device within a known region, the TDOA bounds may be reduced to reflect the maximum TDOA of transmissions emanating from that known region.

2.11 Multilateration After calculating the TDOA for each pair of nodes in the network, the system can perform multilateration to calculate the location of the device relative to the network for each node in the network based on the node's detected localization signal, the node's time bias, and the node's relative location, similar to blocks S260, S268, and S270. In one embodiment, the system can calculate several locations for the device and define an area in which the device may be located. Additionally or alternatively, the system can calculate the device's location with uncertainty in each dimension that indicates the device's likely location within a given confidence level. The system calculates the device's location in the same coordinate system where the relative location of each node in the network is known, so that the location calculated for the device is also relative to the node. In general, the system can calculate the device's two-dimensional relative location with three nodes (of known location) in the network, and the device's three-dimensional location with four nodes (of known location) in the network. With TDOA from more nodes (i.e., more than four), the system can perform least squares and/or linear or non-linear optimization to refine the device's location. Additionally or alternatively, a Kalman filter or other filtering function can be applied to the location estimates of a particular device over time.

The system performs multilateration by solving a system of simultaneous equations of the form:

ここで、ｔ_ｉ，ｊは、ノードｉとノードｊの間のＴＤＯＡであり、ｘ_ｉ、ｙ_ｉは、およびｚ_ｉは、ノードｉの座標であり、ｘ_ｊ、ｙ_ｊは、およびｚ_ｊは、ノードｊの座標であり、ｘ、ｙは、およびｚは、デバイスの座標である。

where t _i,j is the TDOA between node i and node j, x _i , y _i , and z _i are the coordinates of node i, x _j , y _j , and z _j are the coordinates of node j, and x, y, and z are the coordinates of the device.

2.12 Deployment The system can be deployed in any scenario with consistent speed wave propagation between multiple nodes, as well as in several wireless networking scenarios. Some examples of system deployments include local asset tracking deployments, telecommunications deployments, and global satellite deployments.

2.12.1 Local Asset Tracking Deployment The system can be deployed as an asset tracking system. In this embodiment, the system can include a network of nodes placed around a warehouse, assembly line/factory, hospital, school, office building, or any other facility. The nodes can be distributed around the facility to avoid interference and ensure that at least four nodes (if three-dimensional location tracking is desired) can receive location signals from devices distributed throughout the facility. The system can also include a set of active tags that can be attached to assets to be tracked and transmit location signals that are received by the nodes. Additionally, in a local deployment, other transmitting devices in or around the facility can also be tracked.

2.12.2 Telecommunications Deployment The system can be deployed in a telecommunications context where each node includes a cellular site and the network includes a cellular network. Method S200 can be implemented in existing or additionally installed cellular sites. By leveraging the telecommunications infrastructure of the cellular sites, the range of the system can be increased so that the system can locate cellular devices or other transmitters within the range of the cellular network. After calculating the location information of the cellular device, the system can transmit the location information of the cellular device to the cellular device through the cellular network. In this manner, the system can enable precise location services for cellular devices in the cellular network.

2.12.3 Global Satellite Deployment The system can be deployed as a global satellite network, with each node including a LEO satellite and the network including a geolocation system. Method S200 can be adapted to global position tracking applications by compensating for relativistic and atmospheric effects on positioning and synchronization signals. One advantage of the global satellite deployment of the system compared to other global navigation systems is that the satellites performing method S200 do not require continuous tracking and updates via an almanac. In this case, each satellite acting as a node can repeatedly self-locate and time synchronize before locating a transmitter on or around the Earth. Thus, maintenance costs are reduced compared to existing global navigation systems, since the system does not require continuous tracking or prior knowledge of orbital position/velocity. The global satellite deployment of the system allows transmitters to be located anywhere on the Earth, similar to GPS or other global navigation systems.

The systems and methods described herein may be implemented, at least in part, as a machine configured to receive a computer-readable medium having computer-readable instructions stored thereon. The instructions may be executed by a computer-executable component integrated with an application, an applet, a host, a server, a network, a website, a communication service, a communication interface, a hardware/firmware/software element of a user's computer or mobile device, a wristband, a smartphone, or any suitable combination thereof. Other systems and methods of embodiments may be implemented, at least in part, as a machine configured to receive a computer-readable medium having computer-readable instructions stored thereon. The instructions may be executed by a computer-executable component integrated with a computer-executable component integrated with the above-mentioned types of devices and networks. The computer-readable medium may be stored in any suitable computer-readable medium, such as RAM, ROM, flash memory, EEPROM, optical device (CD or DVD), hard drive, floppy drive, or any other device. The computer-executable component may be a processor, although (alternatively or in addition) any suitable dedicated hardware device may execute the instructions.

From the above detailed description and the drawings and claims, those skilled in the art will appreciate that modifications and variations can be made to the embodiments of the invention without departing from the scope of the invention as defined in the following claims.

Claims (20)

scheduling transmission of a first synchronization signal during a first synchronization slot by a first node;
scheduling transmission of a second synchronization signal during a second synchronization slot by a second node;
after transmission of the first synchronization signal in the first synchronization slot by the first node,
receiving from the first node a phase of a first local reference copy of the first synchronization signal;
receiving from the second node a phase of the first synchronization signal at the second node;
after transmission of the second synchronization signal in the second synchronization slot by the second node,
receiving from the second node a phase of a second local reference copy of the second synchronization signal;
receiving from the first node a phase of the second synchronization signal at the first node;
the phase of the first local reference copy;
the phase of the first synchronization signal at the second node;
calculating a propagation delay between the first node and the second node based on the phase of a second local reference copy and the phase of the second synchronization signal at the first node. the phase of the first local reference copy;
the phase of the first synchronization signal at the second node;
2. The method of claim 1, further comprising: calculating a time bias between the first node and the second node based on the phase of a second local reference copy and the phase of the second synchronization signal at the first node. Calculating the propagation delay between the first node and the second node includes:

where φ _1,1 represents the phase of the first local reference copy, φ _2,1 represents the phase of the first synchronization signal at the second node, φ _2,2 represents the phase of the second local reference copy, and φ _1,2 represents the phase of the second synchronization signal at the first node. receiving the phase of the first local reference copy includes receiving from the first node a phase of a first self-received signal associated with the first synchronization signal;
2. The method of claim 1, wherein receiving the phase of the second local reference copy comprises receiving, from the second node, a phase of a second self-received signal associated with the second synchronization signal. Scheduling transmission of the first synchronization signal includes scheduling transmission of a first set of synchronization signals during the first synchronization slot by the first node, the first set of synchronization signals being characterized by a first set of carrier frequencies;
Scheduling transmission of the second synchronization signal includes scheduling transmission of a second set of synchronization signals during the second synchronization slot by the second node, the second set of synchronization signals being characterized by a second set of carrier frequencies;
receiving the phases of the first local reference includes receiving from the first node a set of phases of a first set of local reference copies of the first set of synchronization signals;
receiving the phases of the first synchronization signal at the second node includes receiving a set of phases of the first set of synchronization signals at the second node from the second node;
receiving the phases of the second local reference includes receiving from the second node a set of phases of a second set of local reference copies of the second set of synchronization signals;
receiving the phases of the second set of synchronization signals at the first node includes receiving a set of phases of the second set of synchronization signals at the first node from the first node;
Calculating the propagation delay between the first node and the second node includes:
the set of phases of local reference copies of the first set of synchronization signals;
the set of phases of the first set of synchronization signals at the second node;
2. The method of claim 1, further comprising: calculating the propagation delay between the first node and the second node based on: the set of phases of a local reference copy of the second set of synchronization signals of the second set; and the set of phases of the second set of synchronization signals at the first node. scheduling transmission of a first synchronization signal during a first synchronization slot by a first node;
scheduling transmission of a second synchronization signal during a second synchronization slot by a second node;
after transmission of the first synchronization signal by the first node during the first synchronization slot,
receiving, from the first node, a first phase reference associated with the first synchronization signal;
receiving from the second node a first arrival phase of the first synchronization signal at the second node;
after transmission of the second synchronization signal by the second node during the second synchronization slot,
receiving, from the second node, a second phase reference associated with the second synchronization signal;
receiving from the first node a second arrival phase of the second synchronization signal at the first node;
calculating a propagation delay between the first node and the second node based on the first phase reference, the second phase reference, the first arrival phase, and the second arrival phase. Scheduling transmission of the first synchronization signal includes scheduling transmission of a first set of synchronization signals during the first synchronization slot by the first node, the first set of synchronization signals being characterized by a first set of carrier frequencies;
Scheduling transmission of the second synchronization signal includes scheduling transmission of a second set of synchronization signals during the second synchronization slot by the second node, the second set of synchronization signals being characterized by a second set of carrier frequencies;
receiving the first phase reference includes receiving, from the first node, a first set of phase references associated with the first set of synchronization signals;
receiving the first phase reference includes receiving, from the first node, a first set of phase references associated with the first set of synchronization signals;
receiving the second phase reference includes receiving, from the second node, a second set of phase references associated with the second set of synchronization signals;
receiving the first arrival phases includes receiving from the second node a first set of arrival phases of the first set of synchronization signals at the second node;
receiving the second arrival phases includes receiving from the first node a second set of arrival phases of the second set of synchronization signals at the first node;
7. The method of claim 6, wherein calculating the propagation delay between the first node and the second node comprises calculating the propagation delay between the first node and the second node based on the first set of phase references, the second set of phase references, the first set of arrival phases, and the second set of arrival phases. Calculating the propagation delay between the first node and the second node includes:
For each synchronization signal in the first set of synchronization signals:
a phase reference from the first set of phase references associated with the synchronization signal;
generating a first set of phase-versus-frequency points based on an arrival phase of the synchronization signal from the first set of arrival phases and a carrier frequency of the synchronization signal from the first set of carrier frequencies;
calculating a first phase delay of the first set of synchronization signals based on a first regression of the first set of phase versus frequency points;
For each synchronization signal in the second set of synchronization signals:
a phase reference from the second set of phase references associated with the synchronization signal;
generating a second set of phase-versus-frequency points based on an arrival phase of the synchronization signal from the second set of arrival phases and a carrier frequency of the synchronization signal from the second set of carrier frequencies;
calculating a second phase delay of the second set of synchronization signals based on a second regression of the second set of phase versus frequency points;
calculating the propagation delay between the first node and the second node based on the first phase delay and the second phase delay;
The method of claim 7, comprising: Scheduling transmission of the first synchronization signal includes scheduling transmission of the first synchronization signal during the first synchronization slot by the first node, the first synchronization signal including a first frequency shift keying modulation code sequence;
Scheduling transmission of the second synchronization signal includes scheduling transmission of the second synchronization signal by the second node during the second synchronization slot, the second synchronization signal including a second frequency shift keying modulation code sequence;
Receiving the first arriving phase of the first synchronization signal at the second node comprises:
calculating a first cross-correlation of the first synchronization signal and a first template signal received at the second node based on the first frequency shift keying modulation code sequence;
extracting the first arrival phase based on the first cross-correlation;
Receiving the second arrival phase of the second synchronization signal at the first node comprises:
calculating a second cross-correlation of the second synchronization signal and a second template signal received at the first node based on the second frequency shift keying modulation code sequence;
extracting the second arrival phase based on the second cross-correlation; and
The method of claim 6, comprising: Scheduling transmission of the first synchronization signal includes scheduling transmission of the first synchronization signal during the first synchronization slot by the first node, the first synchronization signal including a first modulated maximal length sequence;
Scheduling transmission of the second synchronization signal includes scheduling transmission of the second synchronization signal by the second node during the second synchronization slot, the second synchronization signal including a second modulated maximal length sequence;
Receiving the first arriving phase of the first synchronization signal at the second node comprises:
calculating a first cross-correlation of the first synchronization signal and a first template signal received at the second node based on the first modulated maximal length sequence;
extracting the first arrival phase based on the first cross-correlation;
Receiving the second arrival phase of the second synchronization signal at the first node comprises:
calculating a second cross-correlation of the second synchronization signal and a second template signal received at the first node based on the second modulated maximal length sequence;
extracting the second arrival phase based on the second cross-correlation; and
The method of claim 6, comprising: Scheduling transmission of the first synchronization signal includes scheduling transmission of the first synchronization signal during the first synchronization slot by the first node, the first synchronization signal including a first Zadoff-Chu sequence;
Scheduling transmission of the second synchronization signal includes scheduling transmission of the second synchronization signal during the second synchronization slot by the second node, the second synchronization signal including a second Zadoff-Chu sequence;
Receiving the first arriving phase of the first synchronization signal at the second node comprises:
calculating a first cross-correlation of the first synchronization signal and a first template signal received at the second node based on the first Zadoff-Chu sequence;
extracting the first arrival phase based on the first cross-correlation;
Receiving the second arrival phase of the second synchronization signal at the first node comprises:
calculating a second cross-correlation of the second synchronization signal and a second template signal received at the first node based on the second Zadoff-Chu sequence;
extracting the second arrival phase based on the second cross-correlation; and
The method of claim 6, comprising: Scheduling transmission of the first synchronization signal includes scheduling transmission of the first synchronization signal during the first synchronization slot by the first node, the first synchronization signal including a first frequency hopping spread spectrum signal;
Scheduling transmission of the second synchronization signal includes scheduling transmission of the second synchronization signal during the second synchronization slot by the second node, the second synchronization signal including a second frequency hopping spread spectrum signal;
Receiving the first arriving phase of the first synchronization signal at the second node comprises:
calculating a first cross-correlation of the first synchronization signal and a first template signal received at the second node based on the first frequency hopping spread spectrum signal;
extracting the first arrival phase based on the first cross-correlation;
Receiving the second arrival phase of the second synchronization signal at the first node comprises:
calculating a second cross-correlation of the second synchronization signal and a second template signal received at the first node based on the second frequency hopping spread spectrum signal;
extracting the second arrival phase based on the second cross-correlation; and
The method of claim 6, comprising: Scheduling transmission of the first synchronization signal includes scheduling transmission of the first synchronization signal during the first synchronization slot by the first node, the first synchronization signal comprising:
a first fixed preamble sequence;
a first variable sync word sequence;
configured to generate a first autocorrelation peak ratio greater than a threshold autocorrelation peak ratio;
Scheduling transmission of the second synchronization signal includes scheduling transmission of the second synchronization signal during the second synchronization slot by the second node, the second synchronization signal comprising:
a second fixed preamble sequence;
a second variable sync word sequence;
configured to generate a second autocorrelation peak ratio greater than the threshold autocorrelation peak ratio;
Receiving the first arrival phase of the first synchronization signal at the second node comprises:
calculating a first cross-correlation of the first synchronization signal and a first template signal received at the second node based on the first fixed preamble sequence and the first variable sync word sequence;
extracting the first arrival phase based on the first cross-correlation;
Receiving the second arrival phase of the second synchronization signal at the first node comprises:
calculating a second cross-correlation of the second synchronization signal and a second template signal received at the first node based on the second fixed preamble sequence and the second variable sync word sequence;
extracting the second arrival phase based on the second cross-correlation; and
The method of claim 6, comprising: scheduling transmission of a first synchronization signal during a first synchronization slot by a first node, the first synchronization signal being characterized by a first set of carrier frequencies;
scheduling transmission of a second synchronization signal during a second synchronization slot by a second node, the second synchronization signal being characterized by a second set of carrier frequencies;
after transmission of the first synchronization signal by the first node during the first synchronization slot,
receiving from the first node a phase reference for each carrier frequency of the first set of carrier frequencies;
receiving from the second node an arrival phase for each carrier frequency of the first set of carrier frequencies based on the first synchronization signal received at the second node;
after transmission of the second synchronization signal by the second node during the second synchronization slot,
receiving from the second node a phase reference for each carrier frequency of the second set of carrier frequencies;
receiving from the first node an arrival phase for each carrier frequency of the second set of carrier frequencies;
calculating a propagation delay between the first node and the second node based on the phase reference for each carrier frequency of the first set of carrier frequencies, the arrival phase for each carrier frequency of the first set of carrier frequencies, the phase reference for each carrier frequency of the second set of carrier frequencies, and the arrival phase for each carrier frequency of the second set of carrier frequencies. Calculating the propagation delay between the first node and the second node includes:
generating a first set of phase/frequency points based on the phase reference for each carrier frequency of the first set of carrier frequencies and the arrival phase for each carrier frequency of the first set of carrier frequencies;
generating a second set of phase/frequency points based on the phase reference for each carrier frequency of the second set of carrier frequencies and the arrival phase for each carrier frequency of the second set of carrier frequencies;
calculating the propagation delay between the first node and the second node based on the first set of phase/frequency points and the second set of phase/frequency points;
15. The method of claim 14, comprising: Calculating the propagation delay between the first node and the second node based on the first set of phase/frequency points and the second set of phase/frequency points comprises:
summing the first set of phase/frequency points with the second set of phase/frequency points to generate a set of total phase/frequency points;
calculating a linear regression of the first set of total phase/frequency points;
Extracting the slope of the linear regression; and
calculating the propagation delay between the first node and the second node based on the slope of the linear regression;
16. The method of claim 15, comprising: 15. The method of claim 14, further comprising: calculating a time bias between the first node and the second node based on the phase reference for each carrier frequency of the first set of carrier frequencies, the arrival phase for each carrier frequency of the first set of carrier frequencies, the phase reference for each carrier frequency of the second set of carrier frequencies, and the arrival phase for each carrier frequency of the second set of carrier frequencies. Calculating the time bias between the first node and the second node includes:
generating a first set of phase/frequency points based on the phase reference for each carrier frequency of the first set of carrier frequencies and the arrival phase for each carrier frequency of the first set of carrier frequencies;
generating a second set of phase/frequency points based on the phase reference for each carrier frequency of the second set of carrier frequencies and the arrival phase for each carrier frequency of the second set of carrier frequencies;
calculating the time bias between the first node and the second node based on the first set of phase/frequency points and the second set of phase/frequency points;
20. The method of claim 17, comprising: Calculating the time bias between the first node and the second node based on the first set of phase/frequency points and the second set of phase/frequency points comprises:
subtracting the first set of phase/frequency points from the second set of phase/frequency points to generate a set of subtracted phase/frequency points;
calculating a linear regression of the first set of subtracted phase/frequency points;
Extracting the slope of the linear regression; and
calculating the time bias between the first node and the second node based on the slope of the linear regression;
20. The method of claim 18, comprising: Scheduling transmission of the first synchronization signal includes scheduling transmission of the first synchronization signal based on a first frequency hopping spread spectrum scheme across the first set of carrier frequencies;
scheduling transmission of the second synchronization signal includes scheduling transmission of the second synchronization signal based on a second frequency hopping spread spectrum scheme across the second set of carrier frequencies.
The method of claim 14.

Images