TW201023543A - Method and apparatus for calibration for beamforming of multi-input-multi-output orthogonal frequency division multiplexing transceivers - Google Patents

Abstract

An embodiment of the present invention includes a calibration system employed in a multi-input-multi-output (MIMO) system for beamforming and receiving a plurality of streams. The system includes a first calibration circuit responsive to inphase (I) and quadrature (Q) pairs of stream and operative to calibrate each I and Q pair and a second calibration circuit responsive to the calibrated I and Q pairs for all streams, wherein the first and second calibration circuits perform calibration in the time domain.

Description

201023543 mt. Description of the Invention: [Technical Field] The present invention relates to a wireless communication transmitter and receiver, and more particularly to one of Orthogonal Frequency Division Multiplexing (OFDM) technologies. Multi-Input-Multi-Output (ΜΙΜΟ) transmitter and receiver to achieve antenna beamforming methods and devices. ❸ [Prior Art] Personal devices, such as computers, mobile phones, personal digital assistants and similar devices, have been widely welcomed in recent years. As technology advances, the size of personal devices continues to shrink and is more portable. Today's various portable wireless devices communicate with each other, making it easier for users to use and making the transfer of data, sound and audio signals easier. In order to achieve this goal, the construction of mobile networks and φ portable wireless devices are indispensable. For wireless networks applied to personal devices, a data machine suitable for the industry standard IEEE 802.11n is expected to be widely used. That is to say, an array formed by a plurality of antennas is placed inside or around the personal device, and a radio frequency (RF) semiconductor device transmits through the antenna array and an analog-to-digital converter (analog-to-digital converter) , ADC) Receive signals and data. Analog digital converters are typically built into personal devices to convert received signals to the fundamental frequency. Then, a baseband processor processes and decodes the received signal to obtain the original data, and the 5 201023543 horizontal data may be a file that is accessed by another person. Or similar equipment transmits via remote wireless transmission. To achieve this goal, the antenna array (called a multi-input multi-output system because it is composed of multiple antennas) to - the receiving and transmitting keys between specified locations Rate of data or information, single-ended reception, and connection. The most == industry standard _G2.11n in order to synchronize parallel strings has been incorporated into the advanced antenna technology to increase the total processing power, and to improve the quality of the connection by transmitting and receiving RF tfl numbers "smartly". The fresh IEEE_n-draft (10) towel specifies two basic beamform 2 methods: ExpHdt and Implid. In the explicit wave formation method, the 'receive H measurement—one of the channel status information (channel state), and the channel status information is returned to the light transmitter. The transmitter can use the channel status information to calculate an optimal "path" or "direction" to transmit packets for a particular user, and this use of channel status information = method is sometimes referred to as "county formation." The material explicit (4) silking method directly measures the channel when 'f wants to close the transmission channel state Wei, causing the problem of the permanent resident time-consuming, which makes the total processing capacity of the network decrease. In the inner implicit beamforming method, the beam state information does not need to be transmitted back to the transport H(10) on the basis of a packet-by-packet, but depends on the channel-reciprocal property (ChannelRedpiOdty). The channel's mutual f assumption—the upper nuclear channel is essentially the same as the 201023543 streaming channel in the transpose operation (Transp〇se〇perati〇n), so that the receiver can use the measured channel state information and form The beam mode sends the packet of the channel status information back to the transmitter. In this way, the explicit channel status information does not need to be transmitted on the line, thereby eliminating the problem of network resident time. In the intrinsic implicit beamforming mode, the lower stream channel requires a correction procedure between the transmitter and the receiver to confirm that the reciprocal nature of the channel is established. The calibration procedure requires the use of a complex coordinate (c〇mplexcoordinati〇n) between an access point (Ac) and a user (a large number of channel status information is periodically exchanged here). An access point is a device that transmits and receives signals via wireless communication between devices. However, the recent proposed method for implementing the corrected implicit beamforming has problems that are difficult to handle and which are inefficient. For example, in the current standard, each sub-carrier (SUb-carrier, typically 56 or 118) is required to perform a fast Fourier transform 'and transfer the result of fast Fourier transform from one device to another' This reduces network processing power and efficiency, and complicates the problem. Therefore, even if an implicit beamforming method is proposed, the problem of inter-chassis consumption and complicated correction exchange cannot be avoided. As defined by the standard EE8〇2 UnD2 ’, in the implicit beamforming method, a positive exchange is required between a pair of transmitters and receivers, such as the transfer of the sneakers. Since the correction of the system - the time-consuming process of performing the correction step F will reduce the overall processing efficiency of the network, and no, how the status of the channel changes, the correction steps must be repeated, causing a large amount of data to be transmitted into the network. in. 7 201023543 In addition to the fact that the implicit beamforming method requires complex correction switching, the issue of complexity and internal operational capability becomes more important. In this case, minor differences in the implementation process or data arrangement may create a risk that the solutions provided by different vendors are not compatible. In addition, the calibration exchange must actually be implemented via hardware/firmware. 10 The correction as described in the standard IEEE 8 〇 2.11 nD 2.0 cannot compensate for the non-linearity of the front end attenuation of the non-linear components such as the in-phase component and the quadrature component. As mentioned earlier, in order to improve the correction procedure of the implicit beamforming method, it is necessary to develop a wireless transceiver for receiving and transmitting signals in accordance with the standard IEEE8〇2.l In DZ0 development. SUMMARY OF THE INVENTION Briefly, the present invention-embodiment includes a correction for beamforming and receiving a plurality of beams, and a miscellaneous positive system-recording multiple output lines. The correction system includes a -first-correction circuit corresponding to the -beam component and the quadrature component for correcting the in-phase component and the quadrature component of each beam; and a second correction circuit corresponding to all beam-corrected in-phase a component and an orthogonal component; wherein the first correction circuit and the second correction circuit are used to perform correction in the time domain. [Embodiment] 8 201023543 An embodiment of the present invention provides a device (e.g., a receiver) and method for correcting beamforming of a multiple input multiple output system to perform a two-stage calibration process. In the first stage, a transmitter and receiver perform a gain correction and a group delay calculation on the in-phase and quadrature transmit and receive signals. The second phase performs a phase correction and a second group delay correction on the results of the first phase to correct the signals across the transmitter and receiver. The 'stage-of-phase correction circuit and the second stage correction circuit perform correction in the time domain. Referring to FIG. 1, FIG. 1 is a general-purpose multiple input multiple output wireless transceiver beamforming system (Universal Multi-Input-Multi-Output) according to an embodiment of the present invention.

Schematic of TransceiverBeamformingSystem) 10. The system 1A includes a receiver circuit 12, a transmitter circuit 14, a channel processor 16, a user profile storage unit I8, and a Multiple Access Contrd' MAC unit 20. The receiver circuit 丨2 and the transmitter circuit 14 are sometimes referred to as wireless transceivers, and φ is used to transmit and receive information in the form of packets. As shown in FIG. 1, the receiver circuit 12 includes a receiver RF (Radio Frequency, RF) module 22, a receiver calibration module 24, and a Fast Fourier Transform (FFT). The device 26, a frequency equalizer 'FEQ' channel estimation circuit 28, a decoder 3A, and a Cyclic Redundancy Code (CRC) detector 32. The transmitter circuit 14 includes a series of transmitter RF modules 34, a 9 201023543 transmitter calibration unit 36, a reverse speed (ίην (10) fine f〇u thin τ surface f_) converter 38, - space Mapping matrix unit 4q, - star map material (ConsteU know Mappe 〇 42, ― listen code _ 41. The map mapper phantom receives the encoded request from the data edit 41, and maps the encoded tilt to the symbol Star constellation point (Symbol ConstelIati〇nP〇ints). 第 As shown in Fig. 1, the receiver circuit u is further transmitted through the estimation circuit 28 at the wheel = (4) transmitter circuit 14 and through the spatial mapping matrix unit (4) connected to ^ ^ The file processor 16 is further switched to the cyclic redundancy check J 32, and the user profile storage unit ls is additionally connected to the multiple access control unit 2 to receive the multiple access control unit. 2〇The output signal. - Generally speaking, the system 1G in the figure is used to perform orthogonal frequency division multiplexing for the architecture of _ multiple input multiple output (mUlti-input_multi_〇呻ut, MIM〇). rth-nalFrequencyDivi—the key feature of the calendar is to perform corrections on the time domain to avoid In particular, the present invention eliminates the need to transmit correction information on the network, and it is necessary to perform correction on the basis, and Yi Zheng = #简子钟 - transmission. Solution, household tL1 picture / reception Each component in the circuit 12 performs a different function to receive the signal. The module 22 is used to convert the RF signal to the private sector for correcting beamforming. Fast Fourier transform over the earth, group change execution - shouting – 201023543 is used to decode the channel, and the checker % is used to perform a redundancy error. The channel estimation performed by circuit 28 contains parameters for each-minimum orthogonal multi-join work (or - parameter_卜/付Body 'computation

In Fig. 1, the signals transmitted by the respective components in the receiver circuit 14 are mapped to a known star map, and the spatial mapping of the output of the 40 pairs of mappers 42 is performed. ^ 41 Binary Convolutional Encoder), (Puncturer) ; Interleaved operation. Inverse fast Fu Li misfolded 38 to map the space map to the single (four) ^ out through the reverse fast Lai Li #赖从亲到至_. The transmission ^ calibration touch / $ is often corrected for the inverse fast Fourier turn _8 green line correction, and the transmitter shoots and transmits the base frequency signal to the radio frequency, so that the converted RF signal can be in the network between the devices. It is transmitted, and a short description will be given later. ,

Channel processor 16 is used to determine the optimal beamforming parameters for each sub-carrier. The user setting dump unit 18 hides the beamforming information of the 1G per-manager, the beamforming information recording beam, and stores it in the form of a fine-grained format to specifically indicate the transmission between the H antenna and the receiver antenna. The best gain and the best phase = direction. The spatial mapping matrix unit 40 receives the user setting information from the unit 18, and for each of the single _ (four) sub-news. The township heavy access control element 2 is used to identify the destination of the package and the Wei's destination. Thus, the 'multiple access (4) unit 2G can perform the authorization of the transmission line, and the authorization of the address, transmission station and orthogonal phase component update of the user multiple access control of 11 201023543. Modules 24 and 36 are used to correct axis attenuation and alignment error

Wgnment). In system ω, the RF attenuation and the alignment error are caused by the receiver analog circuit and the transmitter analog circuit, respectively. β 议 细 实施 实施 实施 实施 实施 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一 一In Fig. 1A, the decoder 3〇_ is in a frame check sequence (F_2 of the cyclic redundancy code circuit 43. The circuit 43 in the processor 16' processing state 16 contains a dagger ~ yuan 45, two _ The circuit = 44 and - the transmitter beam forming matrix single input terminal, and the -45 (trans) (output) terminal of the unit 45 is secreted to the one of the unit 45 mapping circuit 46. The circuit 49 _ is completely independent of the regional memory. The same applies to the application. The area memory mapping circuit 46 of the upper two t/Z is available for selection 46. The circuit 49 includes different area memory mapping circuits.

A matrix is included, the vertical direction of the matrix and the horizontal direction of the matrix is a stream, such as the u-th vertical direction is a sub-carrier and the water 20 outputs a signal to the circuit 49 q, q2° multiple access control unit shot matrix unit 4〇 and Wei 49 The next number secret #. The road read ^ to the space map 12 201023543 The verification performed on the circuit 43 can prevent the channel information containing the error packet from entering the storage circuit 49', thereby reducing the amount of work of the unit π to increase the performance of the network. In operation, 'circuit 43 verifies the signal received by the fast Fourier transformer % (or ^(8) in the signal is Μ orthogonal (four) k shout. If the silk certificate is successful, circuit 43 10 uses the interesting redundancy code check to verify all received orthogonalities. The accuracy of the frequency division multiplexing packet. If any-verification fails, the signal received by Lin will not pass through circuit 43. Circuits 46 and 49 together include user-defined storage unit 18 in Figure 1. 卩 电路 电路 43 Verification has been passed, still need to pass the multiple access control unit such as - verification to check whether the side belongs to this network. In order to achieve this goal ~ take the butterfly single 70 2G control - 彡 heavy wire _ table _ the identity of the two User's identity. If the user belongs to this network, the system 10 continues the process and stores the wire to form the f-message; otherwise, the conversion 1126 receives the no longer processed. According to the transmission or connection (four) action is the collision, multiple Access control early 70 2G provides the sender and the cake to her to the circuit 46. Specially operated as G, the storage beam forming ability becomes so due to each of the scales makes __ very large, the network must be Lose., 峨娜(四)时性The inspections of all the districts have passed. These inspections can verify the corresponding information, and the inspection method is based on the above-mentioned signal inspection and cycle inspection. For the continuation, the received seal (4) is passed. All ages, but the package itself is not to be _ he _. scale, township heavy noodle unit 2 (four) work on f to confirm the package is exhausted W峨 (four), Qian Chi Na. For ς ❹ and 36-42), to properly store Beamforming data to be transmitted. More specifically, for each packet 'multiple access control list = 20, the f path 46 - the sender's button (or the domain user's rhyme) is provided, and then the flow as shown in Fig. 1B is executed. Figure 1B is a flow chart showing the steps performed by circuit 46 in Figure iA. After the channel processor 16 enables the beamforming calculation, the beamforming information is stored in the -temporary position (memory circuit 49) while the circuit 43 checks the validity of the corresponding packet in step 5A. Then, in step 5G2, by checking whether the user belongs to the network, the multiple access control unit 2 determines whether the received packet is valid. If the verification of step 502 fails, then step 5 () 4 is performed. The settings made by processor π and step 500 in step 5() 4 will be reset. Conversely, if the verification of the step is passed, then step 506 is performed. In the step, the user who sent the packet is assigned to a temporary location (which is considered part of the user identification). Taking the table of Figure 18 as an example, the user in step 506 is assigned to a temporary position in the straight line titled "TEMp". Next step 508 will determine if the user is a new user. If 201023543 is used as a new user, then step training is performed to determine the location of the user who is the most ineffective (10) should be e). In the 丨8_example, the most invalid user location is the line 4 "丨". In addition, the user who has failed is the user who has not sent the packet to the beam-shaped wire 1G recently. On the other hand, if the user is not a new user, then the step is made to treat the user's previous position as the next temporary position. In a typical _ sub-user, the single-user continuously transmits multiple packets, causing the user to beam-and-control (four) jobs to be stored in two locations of the note (as described in step 512). - After steps 510 and 512, the flow proceeds to step 514. In step 514, the 'per-user's-failure count is updated to be linked to a most temporary location, i.e., the recent user's failure count is zero, while the other users' expiration counts are increasing. There are two ways to perform the above steps: the first method is to increase the touch count according to the increase of the failure count of a system clock; thus - the 'failure count' gradually increases and stops after a preset time. Increase (because the preset time is still). The second method is to increment the failure count every time a valid new packet is received, and reset the failure count after the failure count reaches a preset threshold. In the second method, the failure determination is a measure of the relative activity between the users. Regardless of which incremental count increment method is used, the user's time-out will cause the information stored by the user to become invalid, and the user must be recalculated through a new packet before restarting the beam control for a particular user. Information stored. 2 is a schematic diagram of an application embodiment of system 10 in FIG. In the second 15 201023543 diagram, a wireless network 5 includes wireless devices %, S8, 62, 68, and 8〇, which communicate with each other via the wireless access method. Access point % - example base station. The wireless devices 52, 58, 62, 68, and 80 each include a corrected wireless transceiver module.

That is, the system 10 in Fig. 1. A wireless transceiver refers to a transmitter and a receiver circuit. B In the second fiscal 'wireless devices' 52, 58, 62, 68, and 8 (), the corrected wireless transceiver module is included. For example, device 52 includes a calibrated wireless transceiver module 56, device 62 includes a calibrated wireless transceiver module, device % includes a corrected wireless transceiver H 〇 6 〇, and device 8G includes the corrected no The squad group 82, and the device 68 includes the calibrated wireless transceiver module 72. In addition, wireless devices 52, 58, 62, 68, and 8 can be used to transmit and receive information via antennas coupled to wireless devices 52, 58, 62, 68, and 80. For example, the device 58 transmits and receives information through one of the antennas 54 coupled thereto, and the device 62 transmits and receives information through the coupled antenna 64. The device 58 transmits and receives information through the coupled antenna 98. The device 80 transmits and receives information through one of the coupled antennas 84, and the device 68 transmits and receives information through the coupled antenna 7 . Access point 74 can include a calibrated wireless transceiver module 76 and transmit and receive information via one of the coupled antennas 78. The access point 74 (wireless communication) can also be interfaced to a network interface 86 to access a Wide Area Network (WAN) or a regional network (L〇calAreaNetw〇rk) , LAN) Receive 16 201023543 Input signal. In addition, in FIG. 2, as described in the prior art, each of the antennas 54, 64, 7〇78, 84, and 98 may be an antenna array or a plurality of antenna arrays for simultaneous transmission and Receive data. As shown in Figure 2,? The antennas are taken care of by different wireless devices, thus creating an environment with multiple inputs and multiple outputs. Devices 52, 62, 58, 68, and 8 are all coupled to an area correction cell that can be beamed to access point 74. Figure 2 shows the various types of 职 的 本 本 本 本 本 本 本 , , , , , , , , , , , , , , , , , , , , , , , , 本 本 本 本 本 本 笔记 笔记 笔记 笔记 笔记Share photos via webcam, wireless elity WiFi f-reading, watch videos or get movie source code through high-quality digital TV (Digltal TV 'HDTV) and video server , the voice stream function of network broadcast programs, and so on. ❹ S ® is a detailed schematic diagram of each component in Figure 1. The circuit 12 is used to group the RF Λ1〇〇. N receiver RF module off = secret group of N RF signals 1〇0, used to receive RF. Frequency module 102 fixed gain correction early ^ 1〇8 group is lightly connected N-healing device; n gain correction mm phase correction unit m consists of group consumption party%^heart 70 group of groups °N group delay adjustment unit 111 ^成^ group, _ connected to positive The phase correction corrects the N receiver baseband signals 142 of the single pair of forks. 201023543 Each module or unit in the group 'in circuit 12 and circuit 14' is reduced to a module or unit in another group. For example, the module 1〇4 is formed and includes a receiver RF module of the group of receiving café modules 1〇2, which is lightly connected to the group of the gain correction unit (10). In the group, the gain correction unit 1〇6, the benefit correction unit (10) is secreted in the same phase as the orthogonal correction unit, and the in-phase quadrature phase correction unit 11G. The Orthogonal Her Correction Unit_ group-delay adjustment unit 114' consisting of the group delay adjustment unit 116 is used to generate the (4) baseband signal 丨. She, in circuit 12, the units or modules in the group are also in the unit or module. In the group formed by the N gain correction units 1〇6, each gain correction is used to correct one of the in-phase component paths and the quadrature component paths of the received signal. This correction function is usually done by correcting the offset first, and the correction bias is achieved by measuring the power per path (using the following formula ig and equation (1) ϋ and the _ component path and the orthogonal component diameter The - or both - multipliers are used to equalize the gods of the mechanical and quadrature component signals, as shown in Equation 12, Equation 13, and Figure. In Figure 3, the circuit U is used to process one. The group formed by the transmitter baseband signal (10) includes a group formed by the unit 124, and the group of the in-phase quadrature phase correction unit 122 is formed by the same phase orthogonal The gain corrects the group formed by the early 120 and the group formed by the μ transmitter RF module 。 8. The above groups are connected in a sequence as shown in Fig. 3. 18 201023543 Groups The block delay adjustment unit 126' is the Γ phase correction unit 128 in the group formed by the in-phase quadrature phase correction unit 122. She corrects the unit 128 _ _ orthogonal gain correction 7L early One of the groups formed by 120 is a gain correction unit 13 〇. Gain is more The positive unit 130 engages with the transmitter face group (3) in the group of transmission (four) quilts 118. Similarly, in circuit 14, the remaining units or modules in the group are purely connected to the corresponding ones. Το or a module. The group formed by the transmitter squaring group 118 is used to generate the corrected N radio frequency signals. The group formed by the module 104 is substantially equal to the chic 22 in the first figure, and the transmitter shoots The group formed by the module m is substantially equal to the module 3 core gain correction unit 108, the in-phase quadrature phase correction unit m, and the group delay adjustment unit w in the first figure. 24, Off-phase quadrature gain correction single (four) g, in-phase © quadrature phase correction unit 122 and group delay adjustment list are included in module 36 in figure i. In-phase quadrature phase correction unit 112 and in-phase quadrature phase correction unit The '_positive residual correction unit' in the group formed by 122 can be used to adjust the phase to ensure that each in-phase component channel and the corresponding quadrature component channel are substantially orthogonal (ie, in-phase knives and positive enthalpy components) Between 90 degrees); in addition, it can be used to equalize each The total phase between the transmitter and the receiver. 19 201023543 In the group formed by the group delay adjustment unit 114 and the group delay adjustment unit 124, each group__unit can be used to face or correct the circuit layout. Limiting 'and each of the receive path and the transmit path separately, eliminating the mismatch caused by the path delay (the path delay of each-in-phase component channel and the corresponding orthogonal component channel is not the same). Group delay adjustment unit The group formed by 1M is used to solve the problem that the delay of the ridges in the receiver channel is not equal; at the same time, the group delay adjustment unit uses the formed group to correct the unequal group delay caused by the crossover transmitter circuit. Problem: Beamforming requires precise correction steps to eliminate or reduce the unintended effects of analog multi-output wireless transceiver analog circuits. For example, irregular antenna traces, variations in power amplifiers, mismatches in RF circuits across receiver channels, and transmitter channels can reduce system performance. If the transmitter does not correct (4) defects, these defects will cause the beam to be transmitted to a defect that deviates from the intended receiver, making the system less efficient. Similarly, a receiver RF defect can cause beam mismatch into the baseband processor and reduce the benefits of beamforming. The embodiment in Fig. 3 is to reduce the problems caused by the defects of the woman. As shown in Figure 3, each correction module requires a parameter set, and she is pre-set in a correction mode. In the correction mode _, - the external device is installed - the _ circuit η has been calibrated. Calculate the gain, phase, and group_delay parameters for the parent-receive chain between the received signal strings by the corrected reference signal, as shown in Figure 4, the single 20 201023543 兀162 Show. Here, a receiving signal chain includes a radio frequency circuit coupled to a pair of analog digital bit conversion g (Anaiogu^gitai Converter, ADC), which is coupled to the base frequency digital compensation circuits 108, 112, and 114. Fig. 4 is a schematic view of a correction system 15G according to an embodiment of the present invention. The calibration system 150 权 3 is a positive mode device 152 and an external calibration device 154. In Fig. 4, the external board positive unit 154 is coupled to the correction mode unit 152 for providing N sets of correction signals 13 to 156 to the correction mode unit 152 (N is an integer). In a multiple input multiple output environment, the correction mode device 152 is calibrated and beamed (4), which is the circuit 12, 14 of the embodiment of the invention. In the external correction device 154, a corrected signal source unit 174 is used to generate the correction signal string 156. The correction mode device 152 includes a group formed by the N receiving RF modules 160 and a compensation parameter calculation unit 162. Receive the shooting group _ wire connection positive news ❿ 156. The compensation parameter calculation unit 162 knots N gains, her and group delay parameters. The calibration mode device 152 further includes a group formed by the N transmission RF modules 168 for receiving the output signals of the correction signal string generator 164 and generating an uncorrected group of the transmission RF signals 158 to One of the external correction units 补偿 54 compensates the parameter calculation unit 172. In order to transmit the instantaneous use of the compensation 166, the compensation calculation unit 172 generates a transmission gain and a subtraction parameter (compensation parameter). The transmission compensation module 166 includes transmission compensation units 12, η and . 21 201023543 and Transmitter Combination or Reception The term "wireless transceiver" as used herein is a circuit of a receiver and transmitter. In the correction of the _, the school Qianlin unit m touch-touch signal string is assembled to the group 16G of the RF module. In the foregoing manner, the device 152 is sequentially

❿ The received signal string (N corrected signal strings (9)) calculates the gain, phase and group delay parameters for each receive chain. A portion of the transfer analog circuit (e.g., circuit 14 of FIG. 3) is a group of m transmit RF modules 168. The transmission of the RF module 168, the string generation request - the reference job (the reference signal can be the same as the correction signal string generated by the correction signal source 174), and the transmission of the transmission module (5) to the external %c positive Device 154. As shown in FIG. 4, by offsetting the compensation parameter calculation 172 to the complement 166, the transmitter compensation parameters can be calculated, stored, and returned to the device 152 at the device (9). The unitary mode device (5) may be a laboratory device, such as a signal generator for generating a correction signal:. The transmitter compensation parameters can be taken and counted by the signal analyzer. The calibration die set 152 can be a mated (transparent line) a pre-corrected scale area network card on an uncorrected vibration. The transmit correction module "pre-corrects" the signal to eliminate the effects of the transmitter RF module. 22 201023543 The procedure for calculating the correction signal string and parameters is as follows: · = Calculation (performed by unit (10)): Calculate the average power per connection (10). The leaf-normal normalization constant (N_aliZingCGnstants) is such that the received ones are equal in all the _component/positive filament pairs. 2. Executed by unit 112): She corrects the splitting of the parameters into two. The relative phase between the first phase and the inphase component/orthogonal component is determined by averaging the outer product value between the received component _ and the quadrature component signal (using a suitable wideband reference signal). In the second phase, the relative phase between each receiving path is calculated by using one or more sine wave reference signals (10) on each receive path to be relatively low, and measuring the _ bits between the received turns. 3. Finally, calculate the group delay between the in-phase component and the quadrature component (performed by cell ιΐ6) and the group delay between each-receiving channel. The channel referred to herein may be any one of channels i to N or any of _. A square wire suitable for measuring the orthogonal frequency division multiplexing subtraction group is in reference to a number of solutions (preferably covering the entire Weiwei belt). For the paste, as specified in the standard IEEE802.11I1, the _ subcarrier (numbered from -28 to 28) 'reference signal string is composed of 8 reference tones (T〇ne) under the Hz bandwidth. The sub-vectors are separated by 'sub-carriers numbered -28, -2, _14, -7, 7, 14, 21 and 28 in the set of sub-carriers, respectively. The phase of each subcarrier is measured 23 201023543 篁 and stored. The linear phase (or slope) of the frequencies in order can be used to measure the group delay on the esoteric channel. The subchannel with the largest group delay determines the amount of compensation required for other channels. In other words, other channels will be delayed to match the subchannel with the largest group delay. The delay of each subchannel is changed from the phase slope to the fraction of the sampling delay. In this way, in the time domain, time delay compensation can be achieved by using a resampler. © Wireless Induction (including (4) and Ship II) The predecessor correction can make the next beam control accurate, and the process of Chengzheng ensures that the reciprocity of the channel is essentially established. The mutual interface of the channel refers to the mathematical model of the upper stream channel, such as the mathematical model of the lower stream channel.

^(Xrec (Hds) T (Formula 1) where is the mathematical model of the upper stream channel, 5 is the mathematical model of the lower stream channel, © is the reciprocity constant (possibly any food value), 〇 『 The difference between the upper stream channel and the lower stream channel means that the information in the channel is transmitted in different directions. For details of the mountain bar, refer to the standard 1EEE802.lln draft 2.0. If the two ends of the line are discussed above The method is corrected and redness is established. The beam control parameters are calculated for use by each subcarrier of an orthogonal frequency division multiplex symbol. The calculation process is divided into two phases. The first phase is used to Each sub-carrier is evaluated by the channel estimation. The complex number of the filament-MxN is estimated. M is the number of the matrix column 'equal to the number of antennas; N is the health of the matrix row, equal to the number of data strings ( One 24 201023543

Generally speaking, N<M). In the second stage, the channel matrix of the estimation result is disassembled by a numerical method such as QR decomposition or singUiar vaiueDecompositipn (SVD). The disassembled matrix provides control information for the transmitter to "pre-equalize" the data packet. For a single wave packet, the sfl number of the receiving end has the highest signal-to-noise ratio (Signal t〇N〇ise Rati〇, SNR), or for example, 'when transmitting multiple streams' the signal at the receiving end has effectively "decoupled" the data stream. ^ When using singular value disassembly, the resolution of a known channel // can be written as: Η^ϋΣν* (Equation 2) When QR is used, the disassembly of a known channel can be written as: h=qr ( Equation 3) // represents the channel matrix, the Diagonal Singular Matrix, the matrix 0, ί/ and K are UnitaryMatrices, and the 酉 matrix means that the eigenvalues are all "1" Matrix. Therefore, through the Hermitian transposition operation, the face matrix can perfectly achieve the inverse operation. In other words, tA/ = /, 2*0=/...etc. It is important that these matrices, which are disassembled from the channel matrix of the estimation results, can be used to beam control each subcarrier. Through this implicit beam steering method and these disassembled matrices, Figure 17 provides some alternative beam steering implementations. In Fig. 17, the circled matrix is the conjugate value of the original matrix. These common vehicle values are required when implementing implicit beam steering. 25 201023543 The beam steering matrix is calculated (according to the receiving_packet), stored and Na (the packet transmitted for beam control), so that the f Medium ACcess Contrd' MAC layer layer module can be finally controlled by a one-way access control As shown in Figure t. When a packet is received, the channel estimation and disassembly (using QR disassembly, singular value disassembly, etc.) begins. If the packet is found to have a valid signal, the cyclic redundancy check is continued. The identification packet needs to be passed through - the effective and unique user midway access control address, so that a channel status information (❿福 stateinf_at^, CSI) can be stored in the appropriate location (for later retrieval). The midway access control layer module calculates the user address and maps the user address to one of the physical memory locations. In addition, the mid-way access control layer touches whether or not the user is authorized to transmit beam control. The foregoing embodiments of the present invention are used to reduce time consumption and eliminate complex correction beta exchange procedures, thereby improving the overall processing performance of the network. However, regardless of how the channel conditions change, the total processing power of the network is affected by the repeated time consumption process, and the repeated time consumption process requires a large amount of data to be transferred between the networks. In addition, the present invention solves the problem of complexity and internal operational capability by eliminating the problem of different suppliers' positions (small differences in the layout of the actual axis), and at the same time, realizing the calibration exchange process. Λ ^ ___compensation (4) Front-end attenuation, for example, the in-phase component of the stream and the quadrature component are non-orthogonal. In contrast, in the embodiment of the present invention, 26 201023543 compensates for nonlinear front-end attenuation, and beamforming in the embodiment of the present invention can improve network performance. For the typical channel condition (5) EE model B: small home/office environment), the performance of a multi-input multi-output wireless transceiver (including a correction circuit including both ends) with implicit beamforming can be improved 4~ 7 (jB. Through the beam control method or "smart antenna", the embodiment of the invention can optimize the performance of the antenna array. In this case, by optimizing the antenna performance, the four antennas can be reduced to three by three. The three antennas can be relied on two " and so on. Therefore, the aforementioned beam control system can be used to reduce the complexity (cost) of the multi-input and multi-output wireless local area network. High Power Consumption The advantages of the embodiments of the present invention include material implementation, no complicated correction, and low network resident time (because wireless channel transceivers do not need to transmit channel state information), and no interoperability is required. Problem (because this solution is essentially for one side) and can improve performance by 4 to 7 dB. ° In addition, the added performance through beamforming can be used to remove one or more receivers/passers. Wei Dao, Qi Qi low power and cost of future products. Other details of the calibration method and device of the monthly embodiment can be referred to the drawings and corresponding programs. Here, the origin of these equations will be introduced. 27 201023543 ... mathematical symbol 0, Xia and Su are used to indicate the behavior of a wireless device (Workstation A), as shown in Figure 5. Similarly, the 'mathematical symbol gift, the necessary financial system is used to gamma B. Teng represents the channel at the receiver end. In addition, because of the data (or signal) string, the aforementioned mathematical symbols represent the matrix. Ideally, the matrix is deleted as a transposed matrix, and the matrix is fine. The matrix is a transposed matrix (the following equations assume that these two conditions hold). The matrix and the milk matrix correction matrix, the matrix and the value of 尬 are determined during the calibration exchange to ensure the reciprocity of the channel. Standard ΙΕΕΕ 802.11 Draft 2.0 of n re-specifies the correction of implicit beamforming. The attenuation of the analog/RF portion of the matrix, muscle T, and the wireless transceivers, or defects, must be corrected or corrected. In the orthogonal frequency division multiplexing spectrum (including 56 or 114 subcarriers according to the channel bandwidth), the correction matrix 4 and the lamp are generally implemented at the fundamental frequency. According to the positive EE8〇2, the correction The matrix is determined during a "positive exchange", whereby the measured channel status information is transmitted between workstation A and workstation B. Because the channel status information in the upper stream channel and the lower stream channel can be Therefore, the condition of channel reciprocity can be used to calculate the correction matrix: BrxHabAtxKa= {ArxHbaB txKb) t (Formula 4) where 'State' represents the receiver attenuation matrix matrix of station B, indicating the channel from station A to station B , indicating the transmitter attenuation matrix of the workstation, especially the correction matrix of the workstation A, the representation of the receiver attenuation matrix of the workstation A, / / seems to represent the channel from the workstation B to the workstation a, Wan 7 ^ represents the workstation B Transmitter attenuation moment 28 201023543 array, and the correction matrix representing station B. As indicated by IEEE 802.ilη, the reciprocity of the channel implies the following relationship: 1 TxBTjix inverse Kb=A-}txATrx (Equation 5) where "]" indicates the inverse operation, so it is known as the inverse matrix.资讯 The information about the attenuation of each sub-carrier must be broadcast. Problems arising from the standard ffiEE 802.11 ri, such as a large amount of resident time consuming, implementation complexity, and possible coffee operation capability, can be solved by the embodiments of the present invention. 5 is a schematic diagram of a wireless device (a communication model between a workstation a and a workstation (1) according to Embodiment 2 of the present invention. In FIG. 5, a correction process for compensating for attenuation is performed at a receiver end and a transmitter end of the wireless transceiver. The transmitter circuit is stenciled in the workstation A, and is turned into a single 兀 200, and the receiving pirate circuit is modularized into the unit 2 (10) in the workstation B. The transmitter circuit of the workstation B is modularized into the unit 2 〇 6 The connection of the workstation A is modularized into a unit. In terms of frequency, as shown in the fifth riding, on the workstation a, the single 兀 (10) surface attenuates the scale tear, the unit _ surface rhyme; at the same time, = workstation B single The compensation surface is turned to Na, and the unit gives the compensation attenuation moment ^ Hua. The new correction beamforming method can be regarded as a - regional solution 5 by

V, especially V, V 29 201023543 where the heart, the meta, and the v table report constants are defined during the execution of the calibration procedure (similar to the method discussed in the equation below). In addition to this, the corrections performed at each of the receiver and transmitter channels operate in the time domain; conversely, the correction method performed on each of the individual subcarriers operates in the frequency domain. In the embodiment of the present invention, the correction module can be corrected in the wireless transceiver through a method of built-in verification (4) and face-slice-external symmetry. In this way, it is not necessary to estimate the channel information in the air so that the correction process without attenuation can be realized. As in the 51st, the lower stream channel is the signal string on the upper side (from station a to station B). In other words, represented in the form of a matrix, the lower stream channel can be expressed as:

Hds=KbrBrxHabA τχΚΑΤ (Formula 6Α) ❹ Similarly, the above-mentioned stream channel is the signal string on the lower side (arrow from workstation β to workstation). Expressed in the form of a matrix, the upper stream channel can be expressed as:

Hus=KarArxHBaBtxKBt (Equation 6B) If the correction _ defined by Equation 6 is used, the lower (four) channel and the upper stream channel are divided into

HDs=dBRaATHAB 30 201023543 (Formula 7Α) and ^us^bARbBTHBA (Formula 7B) Because the ideal channel satisfies (Xun 5) 7 The nature of substitutions, Equations 7A and 7B can be combined into: Moved over Ο

Hus ^AR^Bl

' HDS, T aBRaAT (Equation 8A) Re-arrangement 8A, can obtain hus=^i. (Hdj aBRaAT (Equation 8B) type reciprocity condition, where i

a REC

The reciprocity constant can be written as: (Equation 9) The calibration method and device of the embodiment of the present invention (in the time domain _ 々 々 、, * know +, do Le 仃 仃) can use a second-order ^-type description. The first stage contains corrections for each-stream and the second stage contains corrections for cross-streams. The above-mentioned stream includes a receiving stream and a transmitting stream, and the receiving stream (or receiver chain) refers to a hardware formed by a receiver RF circuit, an analog digital converter, and an associated fundamental frequency correction and chopper module. The transport stream (or transmitter key) is the base frequency; the fccJL/relying group, the analog analogy, the transmission circuit, and the associated transmission antenna. In the purchase, "streaming" and "chain" are essentially equal. For the transmission part of the process of the receiving part of the process of 31 201023543 calibration process, the description of this two-stage process is based on the same calibration process as the receiving part. Understanding Correction The model shown in Figure 5A is used to modularize the attenuation process of the stream through the 3rd and 2nd stage processes according to the RF/Analog attenuation mode correction process. 1000 scoops in a private circle' RF antenna and demixer circuit attenuation model ^ _ antenna 1002, RF antenna 1 〇 _ connected to the subtraction model 1004. Model u circuit clothing shows the fiber H math or Huan 102 with a mathematical table. In order to perform the aforementioned beamforming, the attenuation is divided into two I5 white segments so that the captured defects can be corrected and compensated. The picture is not the 'stage-stage circuit contains the antenna 臓 and the related RF circuit "display TF" 'The second stage circuit (that is, the module brain) contains the "split filter" ❹ and the other 5, the splitter stage will The RF signal is converted to the fundamental frequency. The RF/antenna circuit includes a time delay (or group delay) circuit surface, and each antenna time delay circuit 1006 is different and can be modularized into a production form, where Γ is a delay constant and 7b represents Ν The delay constant Γ of the first stream circuit in the stream, Ν is an integer and ω represents the frequency. The phase delay (or offset) circuit _ and the frequency ω are linear and can be modularized as e, where Ρ represents the phase and ❼ represents the first of the streams in the stream. In addition, 'each antenna material has its relative phase wide 32 201023543 As shown in Figure 5A, the splitter circuit is modularized to - non-orthogonal parameters (with Φ table), time U group) delay The difference and gain difference are the in-phase component and the quadrature component of the fundamental frequency signal. / is the gain of the in-phase component channel, # Orthogonal age channel gain. The wire-to-wire parameter 0 is the offset of the wire-wire orthogonal phase component and the positive component (90 degrees difference). The first stage of the correction/more domain is divided into the splitting effect. For each-receive H circuit, the removal step of the first paragraph must be performed. In order to improve the signal quality and the accuracy of the second-stage correction, the nonlinear effect of removing the splitting wave is very heavy. In Figure 5A, the nonlinear effect or attenuation is modularized into a circuit ι〇ι〇~ι〇 twenty two. The mixer 1012 mixes (or multiplies) the in-phase component of the circuit surface and the output of the circuit. - The mixer combines the quadrature components of the subcarriers with the output of the circuit and the output of the circuit 1008 is mixed (or multiplied). The output of the mixer is passed through a phase difference component/orthogonal component time delay difference circuit surface, and the time difference of the segment in-phase component/orthogonal component is generated. Circuitry 1018 provides an output to an orthogonal component gain offset circuit 1〇22 of the first stream circuit. The mixer 1 〇 12 supplies the output to a phase shift component gain circuit 1020, and /de represents a gain mismatch between the in-phase component and the quadrature component. The circuits 1〇2〇 and 1〇22 respectively provide the outputs to the analog-to-digital converters of the analog-to-digital conversion module 1024, as shown in Fig. 5A. Generally, the position of the analog digital conversion module 1024 is between the module 1〇2 and the unit 1〇6. Module 33 201023543 1024 provides an output to a demultiplexer in-phase component/orthogonal component nonlinearity correction circuit 1026' circuit 1〇26 is actually a combination of cell 106, cell 110, and cell 114. In FIG. 5A, the RF circuit 1〇4〇 having the group delay and the phase effect includes the antenna 1002, the time delay circuit 1〇〇6, and the phase offset circuit 1〇〇 circuit 〇4〇 providing the output to An RF demultiplexer circuit 1042, the circuit 1042 includes circuits 1〇12, 1016, 1018, 1020, and 1022. ❹ As shown in Figure 5B, the second phase of the calibration process involves measuring and correcting the phase and delay variations of the antenna circuit (or stream) to facilitate resolving the reciprocity conditions. In order to understand how the process of the second stage solves the reciprocity condition (as listed in Equations 4 to 9), the embodiment corresponding to Figure 5B is taken as an example, n is equal to three, and the receiver has n streams (each The second receiver circuit in the stream responsible for an RF signal is found to have the largest delay parameter η. Conceptually, other receiver circuits (〇 and 2) can be compensated for matching purposes. In particular, the RF signals 1 and 1 are respectively mixed by signals having 妁-Ρο and Pi port 2 phases to equalize the phase. Similarly, the RF signals 〇 and i are time-shifted by 71 - Τ〇 & Τι -η, respectively, to equalize the time delay, as shown in Figure 5B. It should be noted here that the antennas 1〇5〇, 1〇52 and 1〇54 in Fig. 5B are used to receive RF signals 0, 1 and 2, respectively. The received signal is processed via a stream correction circuit. Circuit 1056 is actually a model of a signal chain 354. More specifically, the three-string-to-stream includes a time delay (or group delay) circuit, a phase 34 201023543 offset circuit, and a - corrected _ mechanical/positive pair. As mentioned previously, the second of the three streams has the longest delay. Therefore, after the circuit deletion process, the outputs of the 'streaming' and the third stream are supplied to - she corrects the circuit 1058, and the circuit is actually a model of the valley circuit group 358. Circuit deletion

The processing includes mixing the output of the circuit capped stream circuit with the phase, and outputting the mixed signal to the -group correction circuit, and the circuit lion is used to correct the delay of the first stream and the third stream. Circuit 1060 is actually a model of a resampler mud. Circuitry 1060 transmits the output signal to a multiple input multiple output circuit to demodulate and decode the received signal. Demodulation and Decoding... After the two-stage calibration process string is obtained, the requirements of Equation 6 are satisfied. For example, if a two-stage calibration process is applied to workstation A in Figure 5, the resulting reciprocity condition can be expressed as: f^AB4RX=ciAR^~^Tl ω + ^1)/ ❽ 9Α) Let αΑΚ=^τλω+^ (Equation 9Β) where / is the unit matrix of 3x3 (iden%). The transmitter calibration process is transmitted through a similar method: where the ^ 5 'right positive reference signal is generated in the transmitter, the azole radio frequency circuit, the miscellaneous - the external county (such as the Nalu meter element 172) is connected ^ Shoot 35 201023543 = number = estimation mode and subtraction, after the material is found, the number of gamma receiver compensation modes, and (groups of modules 118 ~ 124). For example, assuming that the third-sided circuit has the longest delay, the reciprocity constant can be written as: e/(72w+^2) (Equation 9C) It should be noted here that for the transmitter and receiver circuits, the correction The process can be completed in the Station A and Workstation B areas. In this way, Equation 9 can be satisfied (each coefficient is represented by the shout of Equation 9B). For this reason, the transmitter RF and mixer circuits are equally important to the receiver RF and splitter circuits. The circuits of Figures 6 and 10 are embodiments for implementing a two-stage calibration process, but there are also variability, for example, by changing the order of circuit elements to produce different embodiments. FIG. 6 is a first stage calibration module 300 according to the first embodiment of the present invention. The module 300 includes a radio frequency correction reference signal source 25 (or a correction signal string 156) coupled to a receive state analog circuit 252. In sequence, circuit 252 is coupled to a first stage correction circuit 254. The receiver analog circuit 252 includes a radio frequency module 256 (or module 1〇2), an analog converter of the in-phase component 258, and an analog component of the quadrature component. The first stage correction circuit 254 includes The phase component of the jS circuit 266 is coupled to 36 201023543 to a gamma circuit 270 of a non-inverting component and an adder 274 of an in-phase component, and the adder 274 is coupled to a resampler 278 of an in-phase component. The first stage correction circuit 254 further includes an orthogonal component alpha circuit 264 coupled to an orthogonal component adder 276, and the adder 276 is coupled to an orthogonal component resampler 280. The first stage correction circuit 254 further includes a gain correction calculation unit 262 for receiving a digital in-phase component number 290 and a digital positive parent component signal 292. At the same time, the digital in-phase component signal 290 is also input to the 0 circuit 266, and the digital quadrature component signal 292 is also used as the input of the circuit 264. The first stage correction circuit 254 further includes an orthogonal correction calculation circuit 268' for receiving the output signals of the circuits 266 and 264, and generating a signal γ at the second output 298. Before the output of the output signal to the adders 274 and 276, the output circuit 298 is simultaneously operated at the output of the circuit 66 and the alpha circuit 264, the manner of operation being listed by equation. In the receiver analog circuit 252, the RF module 256 is configured to receive the output signal of the RF calibration source 250 and generate an analog component signal 294 and an analog component 296. The RF module 256 converts an RF signal (the output signal of the RF correction reference source 25A) to the base frequency. The analog to digital converter 258 can convert the analog in-phase component signal 294 into a digital form. Similarly, the analog to digital converter 26 turns the analog quadrature component signal 296 into a digital form. The outputs of the resamplers 278 and 280 are calibrated in-phase component quadrature component signal chains 1, which are supplied to a second stage correction circuit. The gain correction calculation unit 262 is similar to the gain correction unit 1〇6 of FIG. 3 for correcting the gain offset between the received fundamental frequency signal and the quadrature component. The signal series for correcting measurement and compensation is arranged in order, so that the quadrature component gain offset of the in-phase component can be compensated first. For gain offset measurements, face correction 250 injects a wideband RF signal as an input to RF module 256. The RF module '256 converts the injected signal to a fundamental frequency signal (or converts it into an equivalent intermediate frequency signal according to actual requirements), and the in-phase component and the quadrature component of the fundamental frequency signal are equal to the signals ® 294 and 296, respectively. Ideally, the power of signals 290 and 292 (converted from signals 294 and 296, respectively) is the same, but the power of signals 290 and 292 may be due to non-ideal effects such as analog-to-digital converters, RF amplifiers, and circuit gain mismatches. different. The data formed by the in-phase component of the average analog-to-digital converter and the sampling points of the quadrature component (the data string must be long enough), the power of signals 290 and 292 can be estimated:

lPOW ^ /1=1 (Equation 10) ® QP〇,=j±Ql (Equation 11) In Equations 10 and 11, '/pew is the estimated power level on the in-phase component signal 290, which is orthogonal. The estimated power level on the component signal 292 path.士/〗 indicates the sum of the in-phase components/continuous sums from n=\ N-1 to L, and the value of |; 2 represents the sum of the orthogonal components δ from the N-1 to L sampling points. In addition, L is an integer 'the value must be large enough to ensure that the accuracy of the power measurement is sufficient, k is the signal 290 and the signal 292 is measured during the measurement period 38 201023543. The gain adjustment value is calculated by comparing the measured power with a target power level p: (Expression 12) yQpow (Equation 13) 《 The function performed by the representation circuit 264 in Equation 12, the circuit 266 is represented in Equation 13 The function performed. As shown in Figure 6, the gain adjustment is achieved by the gain unit. / represents the square root operation. After gain offset correction is complete, unit 268 performs a non-orthogonal correction on the in-phase component and the quadrature component signal. The non-orthogonal correction is performed according to the following equation: ^ Λ = Ι ❹ (式 μ) Ideally, if the in-phase component signal 290 and the quadrature component signal 292 are orthogonal, the in-phase component signal 290 and the quadrature component signal 292 are The average inner product is zero. Any orthogonality will result in a two-branch correction and an unintended average inner product. The measured value of 7 is proportional to #正纽, which can be ashamed_the correlation between the component and the orthogonal component is as shown in Fig. 6. In Equation 13, the integer L, which is represented according to the correction period, must be large enough to ensure that the measurement of non-orthogonality is accurately produced. The non-orthogonal correction is obtained by multiplying the gain correction correct multiplier, circuit 266 and 264 39 201023543 output signals by one-half of the measured non-orthogonality constant g (used in multiplier circuits 270 and 272). . The output of the quadrature component multiplier circuit 272 is subtracted from the in-phase component signal by the subtractor 274. Similarly, the output of the in-phase component multiplier circuit 270 is subtracted from the quadrature component signal by a subtractor 276. After signals 290 and 292 have undergone gain equalization steps and correction steps for non-linear calculations, unit 282 measures and corrects the group delay between the in-phase component signals and the quadrature component signals. In one embodiment of the invention, the measurement and correction function of unit 282 is the use of a fast-rising group, which is a hardware unit commonly found in orthogonal frequency division multiplexing wireless transceivers. A test signal string is fed into signals 29〇 and 29f,

汛唬290 and 292 each include a reference signal on each of the fast Fourier transform tones. As shown in Fig. 7, the in-phase component of the fast Fourier transform and the quadrature component output phase are measured, recorded, and stored on each-tone. The relationship between the number of riding sub-carriers and the number of sub-carriers is proportional to the branch delay and the group delay. As shown in Figure 8 and the ninth riding, if the group delay does not match, the in-phase component will be different from the slanting that she is looking for. The face group _ thing turn pro δ θ slope ratio, such as the $ 不 0 brother 7 diagram is a schematic diagram of the group delay meter 1_ Τ 早 70 70 304 of the embodiment of the present invention.君 , 迟 早 70 70 304 includes - in-phase component fast Fourier transform circuit 3 positive parent component fast Fourier transform circuit 312. In Fig. 7, a positive mode 302 is used to generate an in-phase component signal of 3% and an orthogonal component into the parent. 201023543 ❹

The transfer surface is subtracted from the transfer surface Μ leaf conversion circuit 312. The circuitry of gain and quadrature correction module 302 is similar to circuit 254 of Figure 6 (except circuit 282, resampler 278, and others). The benefits and orthogonal correction module 302 perform the gain and orthogonal correction in the time domain. = NFFT is used to measure the frequency of the group delay between the phase component and the quadrature component path. The circuit and circuit crimes are used to perform the bribery fast Fourier transform. The calculation of the group t' group delay in the embodiment of Fig. 7 is performed in the frequency domain. In another alternative implementation, the calculation of the group is late in the time domain, as shown in the diagram of P. Figure 8 is the relationship between the phase of the in-phase component and the quadrature component signal. Or the %' phase heart and the respectively in-phase component path and the quadrature component path 2 group are determined to be positively reduced by k, k is the sub-collar, and the group delay is measured by the green parent. During the measurement period, the same signal in the time domain is processed by the benefit and positive _2 = 3 () 2, and is the output of the adder circuits 274 and 276 "© part of the calibration process". - Stage group delay correction Fig. 9 is the relationship between the in-phase component and the quadrature component signal phase difference y vs. phase, as shown in the following equation: 7 glaze represents the phase (Equation 15) and the x-axis represents the index k of the sub-carrier. After the quadrature component gain correction correction of the in-phase component of the first-stage correction process, there is still a phase shift that varies linearly with the number of sub-carriers, 41 201023543 phase, 2 group between the in-phase component and the quadrature component signal After the delay mismatches into the age-old age-added money face Nyula, the invention is corrected _ 'listening to the same phase of the age group's secret group, such as the following brief description. Group, 'and I thank you 贞 estimate#γ The filament is generated by the phase difference 1 and the 'family approximation for Fig. 9. One of the methods for generating a linear approximation is the least square curve ❹ ° & %aSt squ_ curve fitting method ) 'which can be used to reduce Measuring the effect of noise. Least square The curve fitting method assumes a linear model, as listed below: y^mx+b (Equation 16) For group delay compensation, the least square curve fitting method needs to determine the slope m. The simultaneous solutions of m and 6 are listed below: b = (κΤ^ΥκΓδθ ❿ (Equation 17) where δθ is the measured phase difference vector, as shown in Fig. $. Subcarrier index ^-%, -7%+1', .., under-1, and especially one His 2 matrix:

(Equation 18) 42 201023543 In the middle of the matrix (the illusion is stored in advance to facilitate subsequent calculations. The evaluation is proportional to the difference in group delay between the in-phase component and the quadrature component, and the estimate is The sampled wealth is converted into a fractional delay, and the fractional delay is applied to the extension of the short branch. In other words, the most dominant branch is delayed, causing the in-phase component and the group on the orthogonal component to delay the scale. The branching step is generally performed by a resampler. ❹ For example, a typical calculation process produces ... 5. subcarriers. Because it is a negative number, the delay in the in-phase component channel is greater than in the quadrature component channel. The delay is longer. Therefore, the partial extension is to be added to the orthogonal component channel. The problem with this step is how much delay is required to be added to the orthogonal component channel. For example, suppose there are 64 subcarriers. That is, at the outermost subcarrier 5χ32 = 16〇., 16〇. Conversion to a sampling point requires drift = _ delayed sampling points. In this case, when the resampler 278 of Fig. 6 is bypassed, re Sampler 278 provides The required delay is added to the sample delay by a standard number resampling circuit. The above calibration method uses an external signal source to correct the in-phase component/orthogonal component channel. In another embodiment of the invention, The loop correction can also achieve an equivalent compensation function without the need to pass an external signal source. The following is a description of the second phase of the calibration process through a related second stage correction circuit. The receiver and transmitter are streamed through After the calibration process of the first phase, the calibration process of the second phase will compensate or align the receiver and transmitter streams. Although the single stream 43 201023543 has been corrected, there is no guarantee that the calibration between the streams is completed. Therefore, the second phase of the positive process is still necessary. Specifically, her and group delay variation between streams must be corrected. One of the methods for correcting the phase between the streams and the group delay variation is shown in Figure 10. In Fig. 10, a second stage correction circuit 35 is coupled to a group 354 of corrected in-phase component orthogonal component signal chains, and group 354 is used. Receiving an output of a radio frequency 10 positive reference signal source 352. Each link of the group 3M towel is a corrected in-phase component orthogonal component chain. For example, the corrected in-phase component of the group 354 is the parent. Component Chain 1 is used to receive the signals generated by resampling!! 278 and 28(). Therefore, the output of the Yangshuo correction process is supplied to the second phase calibration process. The output of group 354 is provided to circuit 350' The circuit 350 includes a second-stage group delay and phase calculation circuit 3 and a group of - (four) paths. The circuit is used to receive the output signal of the group 354. Each of the packet circuits 358 is used to receive the group. The output of the group of orthogonal components of the in-phase component of the group 10 is transmitted to the -resampler group 362, which provides all of the corrected input multiple input signals. Circuit 366 performs phase correction between the in-phase component signal and the corresponding quadrature component signal and performs another group of delay operations. The phase correction performed here is similar to the phase correction performed by the circuit. After the in-phase component signal and the quadrature component signal are corrected, the fast Fourier transform unit 44 201023543 measures and corrects the _ group delay and phase difference of the N receiving/transport chain branches, and the method and the first stage of the measurement and correction The correction circuit is the same. After the receiver measurement and correction method is proposed, a similar measurement and correction method is applied to the transmitter. The calculation of the parental parameters 纟1, 纟2, ..., 〖N and the delay parameter is transmitted through circuit 366, as shown in Fig. 11. In Figure u, circuit 366 includes N fast Fourier transform circuits, such as circuits 370 and 37; The inputs to circuits 370 and 371 are provided by a group 354 of orthogonal component chains of the corrected syndiotactic components. When there is N-Shenjian, each fast Fourier transform circuit performs - N-point fast Fourier transform. The circuit and calculation used for the phase and delay parameters between the streams are the same as those for the in-phase correction. The measurement here also uses the fast Fourier transform circuit' but the output of the calibrated signal module is complex. The complex output of the fast Fourier transform module is converted to phase and is often used to calculate the slope history and the zero frequency (DC) wear point b (the linear estimator used in the past): mb -[κτκ)'κτδθ in this case Under the 'because the input is the plural', the phase of the fiscal origin does not have to pass through the origin. Under the n-segment, she recorded the actual number of tracks (in terms of the in-phase component and the positive component), and the phase at the origin of the figure would pass through the origin. At the origin, bl, b2, ..., 1^ the estimated phase needs to be compensated (or removed) between streams. This step of compensating (or removing) can be done through a phasor:

Wb2 45 201023543

WN As shown in Figure 10, the phase is implemented as a complex multiplier. For the receive chain of each second stage correction process, Fig. 12 is a phase diagram of the complex signal. For the second phase of the calibration process, Figure 13 is a plot of the phase difference between the different signal chains. The group delay between the stream or the signal chain is compensated in the same way as the in-phase component is orthogonal to the group delay. In other words, by comparing all measured slopes 1, 2, ..., mN (as listed in Equation 18), the longest delay branch can be found. The difference between the remaining branches (shorter delay) is also calculated. The difference between the remaining branches is converted to a fraction of a sample point and is achieved by a resampler, as shown in Figure 10, such that the group delay between each stream is the same. ♦ If the delay difference is found to exceed a full sample point, a complete sample point delay link delay is used to compensate for the delay difference. ♦ ♦ The fast Fourier transform circuit of circuit 366 is used to estimate the group delay but is not required to perform this correction mechanism. Specifically, circuit 366 is a non-essential circuit. Conversely, pure tones at discrete frequency points can be used to break into the phase measurement vector. This method can be implemented all in the time domain and has the advantage of being simple and easy to implement. Another alternative method is detailed below. • In another implementation, the two calibration stages (4) the sampling circuit can be combined into a 'resampler' combined 姨 sampling. Combined with all measured 46 201023543 delays to simplify the circuit. For example, the resampler π: and the resampler of circuit 254 can be combined into one resampler. The parameters of the first-stage and second-stage calibration procedures are estimated to be completed, and the overall structure of the calibration circuit is shown in Figure 14. In Fig. 14, a correction system side is applied in a multiple input multiple output system, which includes a group 402 of antennas, which is grouped in a group 404' of a first stage correction circuit to transmit the output money to The group is thorough. In sequence, the group transmits the output signal to the first P to the correction circuit group 4〇6, and the group 4〇6 is used to generate the corrected stream for the multi-input multi-output receiver signal processing. Unit - processing. Each of the groups in Figure 14 contains N devices, where N is the number of receivers. In addition, Μ denotes the number of transmitters or transmissions, and lowercase k denotes an index in the fast Fourier transform output. Group 404 contains circuitry 254 similar to that of Figure 6, and group _406 contains circuitry similar to circuit 35 of Figure 10. For the transmitter, the direction of the arrow in the figure will be reversed, and the above calibration process is done in a separate unit, which may be a radio frequency data collection system or a calibrated RF/baseband module. . As described in the interaction between modules 172 and 166 in Figure 4, the parameters are loaded from the remote processor for the correction of the traced filaments. In an exemplary application, the embodiments shown and discussed herein are applied to system commands using orthogonal 47 201023543. Various embodiments show a calibration process for beamforming, where beamforming refers in particular to implicit beamforming. Figure 15 shows the time domain weights used to receive the phase and group delays between the II subcarrier circuits. In the 15th, the time domain phase and group delay correction system 5 is coupled to an external RF signal source 501, and the external RF signal source 5〇1 provides an RF signal to the system 500 to correct the system 5〇〇. . The system 500 includes a group 5〇4 of RF/analog digit converter units, the group 504 is connected to a group 506 of an in-phase component orthogonal component correction module, and the group 5〇4, 5〇6 includes three Unit (or module), but the actual application is not limited to three. Modules 5〇6 are coupled to a parent inter-correlation circuit 508 and 510. More specifically, the first module and the first module in the group 506 are connected to the circuit 508, and the second module and the third module in the group 50 are coupled to the circuit 510. Circuit 508 outputs the signal to an accumulator circuit 512 which outputs the signal to an accumulator circuit 514. The circuit 512 is coupled to a look-up-table (LUT) circuit 516'. The circuit 514 is coupled to a look-up table circuit 518. The look-up table circuits 516 and 518 transmit the output signals to an estimated parameter circuit wo. System 500 of the present invention includes groups 504, 506, circuits 5〇8, 510, 512, 514, look-up table circuits 516, 518, and circuitry 520. System 500 operates in the time domain to determine group delay and phase compensation parameters in the time domain. To achieve this, signal source 501 provides an analog sine wave reference signal source and group 504 converts the sine wave reference signal source to digital form. As mentioned earlier, 48 201023543 7y((wr+wcV+(?0) =::=:==一& = γ_((«Ά)ι+0丨) ^ _ eJ{{K+^c)t+ 02) ββ, medium 4 sine wave reference frequency, We is any-long offset, Θ. And Θ4 the phase of the early morning signal. The output of the group is interactively paired in pairs, such that the electrical predicate can generate an interaction _ component and an orthogonal component subtraction for use by circuits 512 and 514. The cross-correlation signal generated by circuit 5〇8 can be written as: 5〇 X 51 = ^ SQ^

SlXS2~0i-02 = δθη circuit 512 accumulates the output I of circuit 508, which adds the output 2 of the circuit. The angle and phase of each-accumulated signal is proportional to the frequency of the reference signal provided by the source. Circuits 512 and 514 must accumulate the signal for a sufficient time to reduce any noise effects at the measurement angle. The look-up table circuits 516 and 518 are used to determine the real angle of the accumulated signal and the secret one. Similarly, a m (Coordinate Rotation Digital Computer > CORDIC) module (or a numerical arctangent function) is used to calculate the phase of the corresponding accumulated complex number. The angle is calculated at equal frequency (in the signal bandwidth), for example, for a quadrature frequency division multiplex signal with a bandwidth of 20 MHz, the angle is at frequency _6 MHz, 49 201023543 - dirty, duty It is calculated and stored in the form of a vector. Through the actual situation in Figure 15, she only has to go through two simple designs.

The coupler (ie, electricity (4) 8 and 51G) can be directly calculated. Underneath, in the frequency domain approach, the cup's relative phase needs to pass through a three-sided fast Fourier transform circuit. The accumulated signal secret 01 and the state 2 are plotted against the frequency. As shown in Figure 16, the slope of the secret 1 curve is _ and the group delay between the 〇 赖 and the 丨峨 chain is the same as '1'. The phase of the signal ' is proportional to the phase shift of the fresh scale & Similarly, the measurement result of the signal machine 2 can also find the corresponding phase and group delay. According to the foregoing method, the phase and group delay parameters can also be estimated by the estimator: "θ / . vi . I mmT' - ~ 1 *__T*__ b (Equation 20) where κ = + 2 1 + 6 1_ ( Equation 21) The number of akisaki_6, .2, +2, and +6 indicates the frequency point at which the angle is calculated, as previously described. The above is only a preferred embodiment of the present invention, and is applied according to the present invention. The equal changes and modifications made in the scope of patents should belong to the circumstance of the present invention. [Simple description of the diagram] 50 201023543 Into the multi-round beamforming wireless transceiver diagram is the invention of the multi-input and multi-input (four) beam formation The wireless transceiver diagram is the flow of steps performed by the area memory system circuit in Figure 1A.

Schematic diagram of the first diagram Figure 1A. 1B diagram. For the first embodiment of the present invention, FIG. 2 is a schematic diagram of an exemplary application of the system in FIG. Figure 3 is a detailed schematic diagram of the components of Figure 1. 4 is a schematic diagram of a calibration system according to an embodiment of the present invention. Figure 5 is a schematic diagram of a module of a second wireless communication device. Figures 5A and 5B are schematic views of the attenuation module.

FIG. 6 is a schematic diagram of a first stage correction module according to an embodiment of the present invention. FIG. 7 is a group delay calculation unit according to an embodiment of the present invention. Fig. 8 is a diagram showing the relationship between the number of sub-carriers corresponding to the group delay and Fig. 9 is a diagram showing the relationship between the in-phase component signal and the quadrature component signal. The phase difference corresponds to the number of sub-carriers. FIG. 10 is a schematic diagram of a second-stage correction circuit. Figure 11 is a schematic diagram of the delay and phase calculation circuit of the 1G towel set through the fast Fourier transform. The first step is to set up the relationship between the number of sub-carriers and the number of orthogonal components. The phase difference is shown in Fig. 13 as a relationship between the quiet phase component of the second-stage correction process and the number of orthogonal components corresponding to the number of sub-carriers. 51 201023543 Figure 14 is a calibration system according to an embodiment of the present invention. Figure 15 shows the delay in the time domain correction and the (4) sub-domain group delay. Figure 16 is the cumulative signal versus frequency diagram. Necessary Implicit Beamforming Implementations Figure 17 is a partial non-disassembly and disassembly matrix of an embodiment of the present invention. Figure 18 is an example of the contents stored in the area memory mapping circuit.

Channel point, Ατχ, KBR, KBT, jj, V, Η, Q, r matrix phase [Main component symbol description]

Hab, Ηβα

b], bN, bQl, bu Kar, KAT, KBR, KBT 0, δ~, outer 2, Θ, ,, I, ~ a, β, Y2 k

M〇i ' mn N STAA, STAB TEMP 10 12, 202, 204 Gain sub-carrier number slope data string number station header multi-input multi-round wireless transceiver beamforming system receiver 52 201023543 14,200,206 16 408 18 49 20 © 22 , 102 24 26 28 30 32 34 ' 132 ❿ 36 38 40 41 42 43 44 45 Transmitter channel processor multi-input multi-output receiver signal processor unit user profile storage unit sub-carrier matrix storage circuit multiple Access control unit receiver RF module receiver correction module fast Fourier converter frequency domain equalizer channel estimation circuit decoder cyclic redundancy code checker transmitter RF module transmitter correction module inverse fast Fourier converter space Mapping matrix unit encoder star map mapper frame check sequence cyclic redundancy code circuit channel estimator transmitter beamforming matrix unit 53 201023543 e 46 50 52, 58, 62, 68, 74, 80 54, 64, 70 , 78, 84, 98, 352 56, 60, 66, 72, 76, 82 86 88 100, 104, 108, 112, 116, 118 158, 160, 168, 354, 358, 354 506 Memory mapping circuit wireless network wireless device, 1002, 1050, 1052, 1054 antenna corrected wireless transceiver module network interface processor network wide area network or regional network connection, 120, 122, 124 , 140, 142, 156, 358, 362, 402, 404, 406, 504, 106, 130

110, 128 114, 126 150, 400 152 154 162 , 172 164 166 Group in-phase quadrature gain correction unit in-phase quadrature phase correction unit group delay adjustment unit correction system correction mode device external correction device compensation parameter calculation unit correction signal string Generator Transmit Compensation Module Correction Signal Source Unit 54 174 201023543 250 RF Correction Reference Signal Source 252 Receiver Analog Circuit 254 First Stage Correction Circuit 256 RF Module 258, 260 Analog Digit Converter 262 Gain Correction Calculation Unit 264, 266, 270, 272 multiplier circuit φ 274, 276 adder 278, 280, 364 resampler 282 group delay calculation unit 290 digital in-phase component signal 292 digital quadrature component signal 294 analog in-phase component signal 296 analog quadrature component signal φ 300 first stage correction module 302 correction module 306 in-phase component signal 308 orthogonal component signal 310, 312, 370, 371 fast Fourier transform circuit 350 second phase correction circuit 356, 372 corrected in-phase component orthogonal component signal chain 360 0 circuit 55 201023543

366 second stage group delay and phase calculation circuit 368 signal 500 time domain phase and group delay correction system 501 external RF signal source 508 ' 510 cross-correlation circuit 512, 514 accumulator circuit 516, 518 comparison table circuit 520 estimation parameter circuit 1000 RF antenna and demultiplexer circuit attenuation model 1004 Splitter circuit attenuation model 1006 Group delay delay circuit 1008 Phase delay circuit 1010, 1014 Circuit 1012, 1016 Mixer 1018 In-phase component / quadrature component time delay difference circuit 1020 In phase Component Gain Offset Circuit 1022 Quadrature Component Gain Shift Circuit 1024 Analog Digital Converter Module 1026 Splitter In-Phase/Orthogonal Nonlinearity Correction Circuit 56 201023543 1040 RF Circuit 1042 RF Splitter Circuit 1056 Stream Correction Circuit 1058 Phase Correction Circuit 1060 Group Correction Circuit ❹ 57

Claims (1)

201023543 VII. Patent application scope: 1. A calibration system for a Universal Multi-Input-Multi-Output Beamforming System, comprising: - a first-correction circuit corresponding to - string The in-phase (1叩hase) age and quadrature components are used to correct the in-phase component and the positive component of each stream; a second correction circuit corresponding to all the stream-corrected in-phase components and Orthogonal component; wherein 'the first correction circuit and the second correction circuit are used to perform correction in the time domain. 2. ❹ 3. The correction secret described in claim 1 'there is a correction correction circuit that includes ^ gain correction unit' per gain correction unit for receiving - streaming pairs, and correcting the pair with the stream A correlation gain offset. For example, in the correction system of claim 2, the first-correction f material contains N phase corrections as early as 1, and each-she corrects the unit in the first (four) benefiting element in a pair of red gain red units, which is to make up for her, so that each - The in-phase component of the core and the quadrature component are essentially the overall phase between the J-receiver channels. And a calibration system as claimed in claim 3, wherein the N phase correction units are used to average the received in-phase components and according to a suitable wideband reference signal. The orthogonal component is 'product' to determine the relative phase between the in-phase component and the quadrature component of each stream. 5. The calibration system of claim 4, wherein the N gain correction units are further configured to use one or more sinusoidal reference signals in each of the multiple paths, and to measure a phase between the received signals, The relative phase between the lines per _receiver path. 6. The calibration system of claim 3, wherein the second correction circuit further comprises a solid delay adjustment unit ' pure to the N her correction units, the base corrected signal is generated. The correction system of claim 6, wherein the second correction circuit is further configured to perform a time domain to frequency domain conversion using a K-speed Fourier transform (FastF〇urierTransf〇rm, Fpp). The calibration system of claim 6, wherein the second correction circuit additionally performs group delay adjustment on the domain. ^The correction system described in Item 1, Qian Feng (4) a plurality of packets into one of the orthogonal frequency division multiplexing (OrthogonalF: the group frequency frequency 59 9. 201023543 Multiplexing, OFDM) signal, and confirm the received signal It is an orthogonal frequency division multiplexing signal and is received correctly. 10. The calibration system of claim 1, further comprising a multiple access control (Multiple Access Contro L. MAC) unit for confirming the identity of the user through the calibration system. Eight, the pattern: