Propagation angle estimation in multipath communication systems
A multilayer perceptron feedforward neural network transforms channel estimates into frequency-dependent spirographs to adjust phase and amplitude, addressing the challenge of unpredictable multipath interference in Wi-Fi systems and enhancing signal strength.
Patent Information
- Application Number
- JP2022502597
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2019-09-27
- Filing Date
- 2020-06-11
- Publication Date
- 2024-12-04
- Estimated Expiration
- 2040-06-11
AI Technical Summary
Existing methods for channel estimation and beamforming in frequency-diversity multi-antenna Wi-Fi communication systems with multipath interference fail to accurately determine propagation angles and relative channel delays due to unpredictable amplitude and phase variations, leading to insufficient signal strength maximization.
Employ a multilayer perceptron feedforward neural network (FFNN) to estimate multipath reflections and generate transmit correction coefficients by transforming channel estimates or beamforming feedback into frequency-dependent H and V spirographs in the I/Q plane, using in-phase and quadrature components to adjust phase and amplitude for improved signal transmission.
Enhances received signal strength by accurately determining propagation angles and compensating for multipath interference, thereby maximizing signal-to-noise ratio in multipath environments.
Smart Images

Figure 0007594577000002 
Figure 0007594577000003 
Figure 0007594577000004
Abstract
Description
[Technical field]
[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application is an international application of U.S. Patent Application No. 16 / 585,400, filed September 27, 2019, and claims priority to U.S. Provisional Patent Application No. 62 / 874,079, filed July 15, 2019, both of which are incorporated herein by reference.
[0002] Field This disclosure relates to maximizing the received signal-to-noise ratio in Wi-Fi communication systems using channel estimation and beamforming. In particular, this disclosure is directed to maximizing the received signal-to-noise ratio in a frequency diversity multi-antenna communication system in a multipath signal environment by mapping the in-phase and quadrature components of a matrix representation of the communication channel to angles of arrival using a trained feedforward neural network. [Background technology]
[0003] In a single-frequency, multi-antenna Wi-Fi communication system without multipath, a given null data packet (NDP) x i The communication channel can be estimated by transmitting the signal. In a system with m transmit antennas and n receive antennas, the channel effect can be expressed as Y = HX, where H is the m-by-n channel matrix, X is the transmit NDP represented by a diagonal matrix with n rows and m columns, and Y is the receive information represented by a matrix with m rows and m columns. Each element of the H matrix represents the attenuation and phase delay of each of the m-by-n channels (i.e., for each transmit / receive antenna pair). In this case, the channel matrix H is expressed as H = YX -1The receiving station can then send the H matrix to the transmitting station, where the amplitude and phase correction factors can be calculated as an n-by-m matrix V, which is the inverse of H, normalized with respect to the selected reference channel. The normalized V matrix can be used to predistort the transmitted signal. That is, if V is the normalized inverse of H, then the received signal Y can be calculated as Y=VHX=(H -1 H) = IX = X, where I is the identity matrix. Alternatively, the receiving station can calculate the correction coefficients and send them as explicit beamforming feedback to the transmitting station.
[0004] However, in frequency diversity (e.g., OFDMA) based multi-antenna Wi-Fi communication systems with multipath, signals combine at the receiver and manifest themselves as unpredictable variations in amplitude and phase over frequency in the H- and V-matrix representations of the communication channel. The traditional approach to this unpredictability is to average the amplitude and phase components over frequency under the implicit assumption that no multipath signals exist. This approach has proven inadequate for determining accurate Angle of Propagation (AoP) and relative channel delay and attenuation in the presence of multipath interference. Summary of the Invention [Means for solving the problem]
[0005] This disclosure describes example systems and methods for determining the line-of-sight propagation angle between transmitters and receivers in a multi-channel, multi-carrier communication system in the presence of multi-path interference, and for determining the transmit amplitude and phase corrections required to compensate for channel conditions and multi-path interference to maximize received signal strength.
[0006] In one example, a transceiver is configured to receive a frequency-dependent channel estimate or beamforming feedback in a multi-carrier, multi-antenna communication system, and a multi-layer perceptron feed-forward neural network (FFNN) component coupled to the transceiver is configured to estimate parameters of multipath reflections using a representation of the channel estimate or beamforming feedback and generate transmit correction factors for the transceiver.
[0007] In one example, the system also includes a processor coupled to the transceiver and the FFNN and configured to extract frequency-dependent n-by-m H-matrix amplitude and phase data for a plurality of carrier frequencies from the channel estimates or frequency-dependent n-by-m V-matrix amplitude and phase data for a plurality of carrier frequencies from the beamforming feedback, where n is a number of transmit antennas in the communication system, m is a number of receive antennas in the communication system, and the product of m and n is a number of channels in the communication system.
[0008] In one example, a memory coupled to the processor and the FFNN is configured to store weights, biases and activation functions for the FFNN and to store estimated parameters of multipath reflections, where the processor is further configured to adjust the phase and amplitude of the transmission channel based on the estimated parameters of multipath reflections.
[0009] In one example, the processor is further configured to convert the H matrix data into a frequency-dependent H spirograph in the I / Q plane based on the in-phase (I) and quadrature (Q) components of the H matrix data, or convert the V matrix data into a frequency-dependent V spirograph in the I / Q plane based on the in-phase (I) and quadrature (Q) components of the V matrix data.
[0010] In one example, the FFNN is configured to receive H Spirograph in-phase and quadrature-phase data at each carrier frequency as an input vector and output an effective propagation angle, an effective channel attenuation for each channel, and an effective channel delay for each channel based on characteristics of the H Spirograph; alternatively, the FFNN is configured to receive V Spirograph in-phase and quadrature-phase data at each carrier frequency as an input vector and output an effective propagation angle, and a relative delay and relative attenuation for each channel relative to a selected reference channel based on characteristics of the V Spirograph.
[0011] As an example, a method includes, at a transceiver in a multi-carrier, multi-antenna communication system, receiving one of a channel estimate and a beamforming estimate for a plurality of carrier frequencies, estimating parameters of multipath reflections using a representation of the channel estimate or the beamforming feedback using a multi-layer perceptron feedforward neural network (FFNN), and adjusting a phase and amplitude of a channel transmission based on the estimated parameters of the multipath reflections.
[0012] In one example, the method also includes extracting, in the processor, frequency-dependent n-by-m H-matrix amplitude and phase data for a plurality of carrier frequencies from the channel estimates or extracting frequency-dependent n-by-m V-matrix amplitude and phase data for a plurality of carrier frequencies from the beamforming feedback, where n is a number of transmit antennas in the communication system, m is a number of receive antennas in the communication system, and the product of m and n is a number of channels in the communication system; storing, in a memory coupled to the processor and the FFNN, weights, biases and activation functions for the FFNN and storing estimated parameters of multipath reflections.
[0013] In one example, the method also includes, in the processor, converting the H matrix data into a frequency-dependent H spirograph in the I / Q plane based on in-phase (I) and quadrature (Q) components of the H matrix data, or converting the V matrix data into a frequency-dependent V spirograph in the I / Q plane based on in-phase (I) and quadrature (Q) components of the V matrix data.
[0014] In one embodiment, the method also includes receiving, in the FFNN, H Spirograph in-phase and quadrature-phase data at each carrier frequency as an input vector and outputting an effective propagation angle, an effective channel attenuation, and an effective channel delay for each channel based on characteristics of the H Spirograph; or receiving, in the FFNN, V Spirograph in-phase and quadrature-phase data at each frequency as an input vector and outputting an effective propagation angle and a relative delay and relative attenuation for each channel relative to a selected reference channel based on characteristics of the V Spirograph.
[0015] For a more complete understanding of the various examples, reference is now made to the following detailed description taken in conjunction with the accompanying drawings, in which like identifiers correspond to like components and in which: [Brief description of the drawings]
[0016] [Figure 1] 1 is a block diagram illustrating an example multi-frequency, multi-antenna communication system according to the present disclosure. [Diagram 2] 1 illustrates an exemplary communication system without multipath interference in accordance with the present disclosure. [Diagram 3] 3 illustrates the communication system of FIG. 2 using channel estimation or beamforming feedback according to the present disclosure. [Figure 4] 1 illustrates an exemplary communication system having one multipath signal per channel in accordance with the present disclosure. [Diagram 5] 1 illustrates an exemplary communication system having two multipath signals per channel. [Figure 6]1 is a graphical representation of a frequency-dependent channel estimate in a multipath environment in accordance with the present disclosure. [Figure 7] 1 is a block diagram illustrating an example system for determining a propagation angle in a multipath signal environment according to the present disclosure. [Figure 8] 1 is a graphical representation of frequency-dependent channel estimation in a multipath signal environment in accordance with the present disclosure. [Figure 9] 1 is a graphical representation of frequency-dependent beamforming feedback estimates in a multipath signal environment in accordance with the present disclosure. [Figure 10] 1 is a graphical representation of a frequency-dependent channel estimate in a second-order multipath signal environment. [Figure 11] 1 is a graphical representation of frequency-dependent beamforming feedback estimates in a second-order multipath signal environment. [Figure 12] 5 is a flowchart illustrating an example method for training a neural network to estimate a propagation angle in a multipath signal environment in accordance with the present disclosure. [Figure 13] 5 is a flowchart illustrating an exemplary method for determining a propagation angle in a multipath signal environment according to the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0017] This disclosure describes example systems and methods for determining channel estimates, beamforming feedback estimates, and line-of-sight propagation angle estimates between transceivers in a multi-frequency, multi-antenna communication system in the presence of multipath interference.
[0018] 1 illustrates an exemplary multi-frequency, multi-antenna communication system 100 that includes two transceivers: transceiver 101 and transceiver 102. As illustrated in FIG. 1, transceiver 101 operating in a transmit mode includes transceiver 101, which is designated by TX1 to TX2. m The transceiver 102 has m transmit antennas, RX1 to RX nThe system 100 may be, for example, but not limited to, a Wi-Fi® communication system or a cellular phone system.
[0019] The channel conditions between any pair of antennas in the system 100 can be characterized by a frequency-dependent h parameter that defines the attenuation and phase delay between the antenna pair. For example, the channel between transmit antenna TX1 and receive antenna RX1 can be characterized by h 11 (f), where f is the frequency of a subband carrier, such as an OFDM (orthogonal frequency division multiplexing) carrier in a Wi-Fi or cellular system. If symbol stream x1 is being transmitted from antenna TX1, then antenna RX1 will receive the product (h 11 )(x1). This relationship is expressed by the following equation (1):
number
[0020] Because the H matrix is initially unknown, the signals received at the transceiver 102 may combine in unpredictable ways, either through constructive or destructive interference. However, if the transmitted symbol vector X is a training vector known to the transceiver 102, then the transceiver 102 can calculate the H matrix as H=YX -1 (3) can be estimated as follows, where X -1 is the inverse of X.
[0021] After estimating the H matrix for the system, transceiver 102 determines how to adjust its transmit gain and phase to compensate for channel attenuation and delay, and transmits the H matrix to transceiver 101 so that transceiver 101 can make the same adjustments to maximize the received signal strength of its transmission at transceiver 102.
[0022] Alternatively, since the channel is reciprocal, the transceiver 102 can simply send the necessary adjustments directly to the transceiver 101 as beamforming feedback, and the transceiver 101 can make the adjustments without having to manipulate the H matrix data. This matrix data is the inverse of the H matrix, called the V matrix, and can be used to directly adjust the gain and phase of the transceiver 101 transmission to compensate for the channel. That is, instead of sending X, the transceiver 101 can VX=H -1 X(4) This causes the transceiver 102 to Y=HVX=HH -1 X=X(5) Receive.
[0023] In this case, the transceiver 101 can apply the same corrections to normal data transmissions. To compensate for changing channel conditions, the above-described channel estimation and beamforming feedback operations can be repeated periodically. However, as mentioned above, this channel calibration process breaks down in the presence of multipath interference.
[0024] Figure 2 shows a 2x1 communication system 200 with two transmit antennas TX1 and TX2 and one receive antenna RX1, with no multipath interference and no correction from the H-matrix or V-matrix. In the example of Figure 2, the location of RX1 relative to the center C of the transmit baseline 210 results in a somewhat acute angle of propagation (AoP), causing line of sight (LOS) signals from TX1 (201) and TX2 (202) to combine out of phase and with different amplitudes (due to path length differences), resulting in destructive interference at RX1 as shown by combined signal 203. It should be noted that even if the AoP were orthogonal and path lengths were equal for signals 201 and 202, they could still combine out of phase due to phase mismatch in the analog front ends of the transmitters associated with TX1 and TX2.
[0025] Figure 3 illustrates the effect of correcting the transmit phase of TX2 (204) relative to TX1 (202) based on H-matrix or V-matrix data to create constructive interference between the TX1 and TX2 signals as shown by combined signal 205 in Figure 3. Additionally, the angle of propagation (AoP) can be determined from the known frequency of transmission and the amount of phase shift required to maximize signal 205.
[0026] As shown in FIG. 4, however, the presence of multipath interference can defeat the traditional H-matrix or V-matrix approach. FIG. 4 shows a communication system 300 similar to system 200, but with reflective surfaces 310 in the environment. In FIG. 4, LOS signal 201 is combined with reflected signal 301 to generate combined signal 303, and LOS signal 202 is combined with reflected signal to generate combined signal 304. In this scenario, receiver RX1 cannot distinguish between the LOS signal and the reflected signal, so assume that combined signal 303 is the LOS signal from TX2 and combined signal 304 is the LOS signal from TX1. These two signals combine at RX1 to generate combined signal 305 as shown in FIG. 4. Comparing FIG. 2 to FIG. 4, it can be seen that the apparent LOS signal 303 from TX2 in FIG. 4 is significantly different from the actual LOS signal 201 from TX2 in FIG. 2, and the apparent LOS signal 304 from TX1 in FIG. 4 is significantly different from the actual LOS signal 202 from TX1 in FIG. 2. As a result, the overall combined signal 305 at RX1 in FIG. 4 also differs significantly from the overall combined signal 203 in FIG.
[0027] Therefore, the H or V matrices calculated from the received signals in FIG. 4 will be significantly different from those calculated from the received signals in FIG. 2. Furthermore, in a multi-carrier system such as an OFDM system, the "apparent" H and V matrices change unpredictably. In this case, the level of unpredictability increases as the number of multipath signals increases. FIG. 5 shows an exemplary communication system 400 similar to system 300, but with an additional reflecting surface 410 that creates two additional multipath signals 401 and 402. These signals are combined with the aforementioned LOS and multipath signals to generate a signal 411 at RX1 that appears to be a LOS signal from TX2, and a signal 412 at RX1 that appears to be a LOS signal from TX1. These two signals are combined at RX1 to generate an overall received signal at RX1 that is different from the combined signal 203 in FIG. 2 or the combined signal 305 in FIG. 4.
[0028] FIG. 6 shows the H matrix for a communication system with two transmit and two receive antennas. 11 (f), h 12 (f), h 21 (f) and (h) 22 FIG. 6 is a graph of the variation with frequency of the H matrix parameters for the case of four composite parameters of |h xy Graph 602 in FIG. 6 plots the phase angle (ang [h xy 6 plots the variation of the frequency of the h parameter ((f)) with frequency. The horizontal axis of graphs 601 and 602 represents the normalized subcarrier frequency index relative to a reference center frequency of index 0. The vertical axis of graph 601 represents the normalized amplitude of the h parameter with respect to frequency relative to a reference amplitude of one of the h parameters at a selected frequency. The vertical axis of graph 602 represents the normalized phase of the h parameter with respect to frequency relative to a reference phase of one of the h parameters at a selected frequency. Similar magnitude and phase representations (not shown) can be made for the parameters of the V matrix, although the curves will be different due to the inverse nature of the V matrix relative to the H matrix.
[0029] 7 illustrates an example system 700 for estimating a propagation angle in a multipath signal environment according to the present disclosure. The system 700 includes a transceiver 701 configured to receive a channel estimate or beamforming feedback in a multi-carrier, multi-antenna communication system and output a digitized channel estimate or beamforming feedback. The system 700 also includes a processor 702 coupled to the transceiver 701 by a communication bus 703 and configured to receive the digitized channel estimate or beamforming feedback from the transceiver 701. The processor is also configured to extract frequency-dependent n-by-m H-matrix amplitude and phase data (such as the amplitude data 601 and phase data 602) for multiple carrier frequencies from the channel estimate or frequency-dependent n-by-m V-matrix amplitude and phase data for multiple carrier frequencies from the beamforming feedback, where n is the number of transmit antennas in the communication system, m is the number of receive antennas in the communication system, and the product of m and n is the number of channels in the communication system.
[0030] The processor 702 may be one or more general-purpose processing devices, such as a microprocessor, a central processing unit, etc. More specifically, the processing device may be a Complex Instruction Set Computing (CISC) microprocessor, a Reduced Instruction Set Computer (RISC) microprocessor, a Very Long Instruction Word (VLIW) microprocessor, or a processor implementing other instruction sets or a combination of instruction sets. The processing device may also be one or more special-purpose processing devices, such as an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), a network processor, etc.
[0031] The system 700 also includes a multi-layer perceptron feed-forward neural network (FFNN) 704 coupled to the transceiver 701 and the processor 702 by a communication bus 703, the FFNN 704 configured to estimate parameters of multipath reflections using a channel estimate or a representation of the beamforming feedback and generate transmit correction factors for the transceiver 701.
[0032] The system 700 also includes a memory 705 coupled to the transceiver 701, the processor 702, and the FFNN 704 by a communication bus 703, which stores weights, biases, and activation functions for the FFNN 704, and stores estimated parameters of multipath reflections. The memory 705 may be a volatile memory or a non-volatile memory, or may include volatile and non-volatile memory. The non-volatile memory may be a Read-Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), or a flash memory. Also, the volatile memory may be a Random Access Memory (RAM) used as an external cache. Many forms of Random Access Memory (RAM) can be used, including, by way of non-limiting illustrative example, Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDR SDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), Synchronous Link Dynamic Random Access Memory (SLDRAM), and Direct Rambus® DRAM (DRDRAM).
[0033] In one example, the processor 702, the communication bus 703, the FFNN 704, and the memory 705 may be implemented as a system on a chip (SOC) 706. In another example, the transceiver 701, the processor 702, the communication bus 703, the FFNN, and the memory 705 may comprise an IEEE 802 compatible Wi-Fi® access point.
[0034] In one example, the processor 702 is further configured to adjust the phase and amplitude of a transmission channel of the transceiver based on the estimated parameters of the multipath reflections. In one example, the processor is further configured to convert the digitized H matrix amplitude and phase data into a frequency-dependent H matrix-based Spirograph (H Spirograph) in the I / Q plane based on the in-phase (I) and quadrature (Q) components of the H matrix data, or convert the V matrix data into a frequency-dependent V matrix-based Spirograph (V Spirograph) in the I / Q plane based on the in-phase (I) and quadrature (Q) components of the V matrix data.
[0035] The H-matrix transformation is illustrated in Figure 8, where a spirograph based on the H-matrix (H-spirograph) for an exemplary 2x2 communication system (such as communication system 100 with m=2, n=2) with one multipath reflection per channel (as in system 300) is plotted in the I / Q plane in graph 800. The I component for each h parameter is |h xy |cosφ xy where |h xy | is the parameter h xy is the size of φ xy is the parameter h xy (Here, the frequency index is omitted.) The Q component of each h parameter is |h xy |sinφ xy is given by:
[0036] In the case of one multipath reflection shown in Figure 8, the H spirographs are located at coordinates (I,Q) xyCenter, radius R xy , and the arc span θ xy By modeling and simulating many different single multipath scenarios, we have found that the center of the arc (I,Q) xy It can be shown that the location of is a function of the analog chain mismatch and the line of sight (LOS) angle of propagation (AoP) in the system's transceiver front-end. Also, the radius of the arc R xy It can also be shown that is a function of the magnitude of the multipath reflections for each channel relative to the magnitude of the LOS signal for that channel. One implication of the relationship between the radius of the H-spirograph and the strength of the multipath reflections is that in the absence of multipath reflections, the H-spirograph is frequency independent and represents a single point (I,Q) xy The point is to degenerate into
[0037] Also, the span of the arc θ xy is a function of the phase of the multipath reflections relative to the LOS signal, which can also be shown to depend on the propagation angle of the multipath reflections relative to the LOS signal. Finally, for a single multipath, the arcs are congruent, i.e., they can be made to overlap each other under one rotation and one translation in the I / Q plane.
[0038] The V-matrix transformation is illustrated in FIG. 9, where a spirograph based on the V-matrix (V-spirograph) for an exemplary 2×2 communication system (such as communication system 100 with m=2, n=2) with one multipath reflection per channel (as in system 300) is plotted in the I / Q plane in graph 900. Graph 900 of FIG. 9 differs from graph 800 of FIG. 8 in at least one respect. While the h parameter represents an estimate of the absolute channel condition, the v parameter represents an estimate of the relative change in the transmitted signal required to compensate for the channel condition. Thus, the v parameter can be normalized to a reference channel. In the example graph 900 of FIG. 9, v 22The channel associated with v is selected as the reference channel, and the other three channels are 22 As a result, the v22 trace in graph 900 is a single point (I,Q) that is frequency independent. 22 It appears as:
[0039] For the other three v parameters in graph 900, the V spirograph for the I component is |v xy |cosφ xy where |v xy | is |v 22 Parameter v for the size of | xy is the relative size of φ xy v 22 Parameter v for the phase of xy Similarly, the Q component for each v parameter is |v xy |sinφ xy is given by:
[0040] In the case of one multipath reflection shown in Figure 9, the V spirographs (excluding the reference channel) are each represented by coordinates (I,Q) xy Center, radius R xy , and an arc span θ having non-uniformly spaced frequency points xy By modeling and simulating many different single multipath scenarios, it can be shown that the location of the center of the arc, the radius of the arc, and the span of the arc are determined by the analog chain mismatch in the transceiver front-end, the line-of-sight (LOS) propagation angle of the channel, the magnitude of the multipath reflections relative to the LOS signal, the phase of the multipath reflections relative to the LOS signal, and the propagation angle of the multipath reflections relative to the LOS signal.
[0041] In the absence of multipath interference, all V Spirographs are frequency independent and correspond to a single point (I,Q), in a similar fashion to the H Spirograph. xyAlso, for a chosen reference channel, the V-spirographs for all channels with one multipath reflection are similar under translation, rotation, and radial scaling.
[0042] FIG. 10 is a graph 1000 in the I / Q plane showing the H spirograph as a function of frequency for the case where there are two multipath reflections on each channel (similar to system 400 of FIG. 5). By modeling and simulating many two-path reflection scenarios, the h parameter in the I / Q plane, h 11 (f), h 12 (f), h 21 (f), and h 22 It can be shown that (f) are congruent compound hypocycloids that can be superimposed under one translation and one rotation in the I / Q plane.
[0043] 11 is a graph 1100 in the I / Q plane showing the V spirograph as a function of frequency for the case where there are two multipath reflections on each channel (similar to system 400 of FIG. 5). By modeling and simulating many two-multipath reflection scenarios, it can be shown that the V parameters in the I / Q plane for the reference V spirograph are similar under rescaling, rotation, and resampling.
[0044] It will be appreciated that the multipath scenarios illustrated in Figures 8 through 11 are under-constrained and are not subject to direct analytical solution, however, the same modeling and simulations used to characterize the H-spirograph and V-spirograph in the preceding description can be used in system 700 as training data for training a multi-layer perceptron feed-forward neural network 704 to estimate multipath reflection parameters, determine propagation angles, and apply corrections to the transmitted signal to maximize received signal strength.
[0045] 12 is a flow chart illustrating an example method 1200 for training a multi-layer perceptron feed-forward neural network (e.g., FFNN 704) to estimate parameters of propagation angles and multipath reflections using representations of channel estimates and beamforming feedback in a multi-frequency, multi-antenna communication system. The method 1200 begins at operation 1202 with preparing training data including simulations of channel estimates and beamforming feedback in a multi-channel, multi-carrier communication system for multiple multipath scenarios. The method 1200 continues to operation 1204 with training a multi-layer perceptron feed-forward neural network using in-phase and quadrature-phase representations of the channel estimates (e.g., H-spirographs) and beamforming feedback (e.g., V-spirographs) to estimate parameters of multipath reflections. The method 1200 ends at operation 1206 with storing the weights, biases, and activation functions of the trained feed-forward neural network in memory.
[0046] In one example, operation 1204 of training a multi-layer perceptron feedforward neural network (FFNN) includes generating a Spirograph feature set from in-phase and quadrature representations of channel estimates and beamforming feedback for each of a plurality of multipath scenarios for each channel at each carrier frequency, applying the feature sets in batches to the FFNN in a forward pass to generate a loss function with a representation of the training data, and propagating the loss function in a backward pass through the neural network to update weights and biases of the neural network to minimize the loss function.
[0047] 13 is a flow chart illustrating an example method 1300 according to the present disclosure for determining a propagation angle in a multipath signal environment. The method 1300 begins with operation 1302, where a transceiver (e.g., transceiver 701) in a multicarrier, multiantenna communication system (e.g., system 700) receives one of a channel estimate (e.g., H matrix) and a beamforming estimate (e.g., V matrix) for a plurality of carrier frequencies. The method 1300 continues with operation 1304, where a multi-layer perceptron feed-forward neural network (FFNN) estimates parameters of multipath reflections using a representation of the channel estimate (e.g., H spirograph) or the beamforming feedback (e.g., V spirograph). The method 1300 continues with operation 1306, where a phase and amplitude of a channel transmission is adjusted based on the estimated parameters of the multipath reflections.
[0048] In one example, the method 1300 also includes, in the processor, extracting frequency-dependent n-by-m H-matrix amplitude and phase data for multiple carrier frequencies from the channel estimates or extracting frequency-dependent n-by-m V-matrix amplitude and phase data for multiple carrier frequencies from the beamforming feedback, where n is a number of transmit antennas in the communication system, m is a number of receive antennas in the communication system, and the product of m and n is a number of channels in the communication system; storing, in a memory coupled to the processor and the FFNN, weights, biases and activation functions for the FFNN, and storing estimated parameters of multipath reflections.
[0049] In one example, the method 1300 also includes, in the processor, converting the H matrix data into a frequency-dependent H spirograph in the I / Q plane based on the in-phase (I) and quadrature (Q) components of the H matrix data, or converting the V matrix data into a frequency-dependent V spirograph in the I / Q plane based on the in-phase (I) and quadrature (Q) components of the V matrix data.
[0050] In one example, the method 1300 also includes receiving, in the FFNN, in-phase and quadrature-phase data of an H spirograph at each carrier frequency as an input vector, and outputting an effective propagation angle, an effective channel attenuation, and an effective channel delay for each channel based on characteristics of the H spirograph, or receiving, in the FFNN, in-phase and quadrature-phase data of a V spirograph at each frequency as an input vector, and outputting an effective propagation angle and a relative delay and relative attenuation for each channel relative to a selected reference channel based on characteristics of the V spirograph.
[0051] In the preceding description, numerous specific details have been mentioned, such as examples of specific systems, components, methods, and the like, to provide a thorough understanding of some examples in the present disclosure. However, it will be apparent to one of ordinary skill in the art that at least some examples of the present disclosure can be practiced without these specific details. In other instances, well-known components or methods have not been described in detail, but have been presented in simple block diagram form, in order to avoid unnecessarily obscuring the present disclosure. Thus, the specific details mentioned are merely exemplary. It is contemplated that certain examples may differ from these example details and still be within the scope of the present disclosure.
[0052] References throughout this specification to "one example" or "an example" mean that a particular feature, structure, or characteristic described in connection with the example is included in at least one example. Thus, the appearances of the phrase "in one example" or "in an example" in various places throughout this specification are not necessarily all referring to the same example.
[0053] Although the operations of the methods herein have been shown and described in a particular order, the order of operations of each method may be changed such that certain operations are performed in the reverse order, or certain operations are performed at least in part concurrently with other operations. Instructions or sub-operations of separate operations may be performed intermittently or alternately.
[0054] The above description of examples shown for the present invention, including those described in the Abstract, is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Specific implementations of the present invention, and examples thereof, are described herein for illustrative purposes, but various equivalent modifications are possible within the scope of the present invention, as those skilled in the relevant art will recognize. As used herein, the words "example" or "exemplary" are intended to serve as an example, instance, or illustration. Any aspect or design described herein as "example" or "exemplary" is not necessarily to be construed as preferred or advantageous over other aspects or designs. Rather, use of the word "example" or "exemplary" is intended to illustrate a concept in a specific manner. As used in this application, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or". That is, unless otherwise specified or clear from the context, "X includes A or B" is intended to mean any of the natural inclusive permutations. That is, if X includes A, X includes B, or X includes both A and B, then "X includes A or B" will be satisfied in any of the foregoing cases. Moreover, the articles "a" and "an" as used in this application and the appended claims should be construed generally to mean "one or more" unless otherwise specified or clearly directed from the context to the singular form.
Claims
1. 1. A system comprising: A transceiver configured to receive frequency dependent channel estimates or beamforming feedback in a multi-carrier, multi-antenna communication system; a multi-layer perceptron feed-forward neural network (FFNN) component coupled to the transceiver and configured to estimate parameters of multipath reflections using the channel estimates or a representation of the beamforming feedback and generate transmit correction factors for the transceiver; a processor coupled to the transceiver and the FFNN; a memory coupled to the processor and the FFNN; Equipped with the processor is configured to extract from the channel estimates frequency-dependent n-by-m H-matrix amplitude and phase data for a plurality of carrier frequencies, or extract from the beamforming feedback frequency-dependent n-by-m V-matrix amplitude and phase data for the plurality of carrier frequencies, where n is a number of transmit antennas in the multi-carrier, multi-antenna communication system, m is a number of receive antennas in the multi-carrier, multi-antenna communication system, and the product of m and n is a number of channels in the multi-carrier, multi-antenna communication system; The memory stores weights, biases and activation functions for the FFNN and stores estimated parameters of the multipath reflections; The processor is further configured to adjust a phase and an amplitude of a transmission channel based on the estimated parameters of the multipath reflections; the processor is further configured to convert the n-by-m H-matrix magnitude and phase data into a frequency-dependent H-spirograph in an I / Q plane based on in-phase (I) and quadrature (Q) components of the n-by-m H-matrix magnitude and phase data. system.
2. A system, comprising: A transceiver configured to receive frequency dependent channel estimates or beamforming feedback in a multi-carrier, multi-antenna communication system; a multi-layer perceptron feed-forward neural network (FFNN) component coupled to the transceiver and configured to estimate parameters of multipath reflections using the channel estimates or a representation of the beamforming feedback and generate transmit correction factors for the transceiver; a processor coupled to the transceiver and the FFNN; a memory coupled to the processor and the FFNN; Equipped with the processor is configured to extract from the channel estimates frequency-dependent n-by-m H-matrix amplitude and phase data for a plurality of carrier frequencies, or extract from the beamforming feedback frequency-dependent n-by-m V-matrix amplitude and phase data for the plurality of carrier frequencies, where n is a number of transmit antennas in the multi-carrier, multi-antenna communication system, m is a number of receive antennas in the multi-carrier, multi-antenna communication system, and the product of m and n is a number of channels in the multi-carrier, multi-antenna communication system; The memory stores weights, biases and activation functions for the FFNN and stores estimated parameters of the multipath reflections; The processor is further configured to adjust a phase and an amplitude of a transmission channel based on the estimated parameters of the multipath reflections; the processor is further configured to convert the n-by-m V-matrix magnitude and phase data into a frequency-dependent V-spirograph in an I / Q plane based on in-phase (I) and quadrature (Q) components of the n-by-m V-matrix magnitude and phase data. system.
3. the processor, the FFNN, the memory and the communication bus comprise a system on chip (SOC); 3. The system according to claim 1 or 2.
4. the transceiver, the processor, the FFNN and the memory comprise an IEEE 802 compatible Wi-Fi access point; 3. The system according to claim 1 or 2.
5. The FFNN is configured to receive as an input vector the in-phase and quadrature-phase data of the H Spirograph at each carrier frequency, and output an effective propagation angle, an effective channel attenuation for each channel, and an effective channel delay for each channel based on characteristics of the H Spirograph. The system of claim 1 .
6. The FFNN is configured to receive as input vectors in-phase and quadrature-phase data of a V-spirograph at each carrier frequency, and output an effective propagation angle and a relative delay and relative attenuation of each channel with respect to a selected reference channel based on characteristics of the V-spirograph. The system of claim 2.
7. The H-spirograph for a channel without multipath reflections has a single point in the I / Q plane with a location determined by analog chain mismatch in the transceiver front-end and the line-of-sight (LOS) propagation angle of the channel. The system of claim 5.
8. The H-spirograph for a channel with one multipath reflection has a circular arc in the I / Q plane having a center, a radius, and an arc span with uniformly spaced carrier frequency points; the location of the center of the arc is determined by analog chain mismatch and line of sight (LOS) propagation angle in the front end of the transceiver; the radius of the arc is determined by the magnitude of the multipath reflections relative to the LOS signal; the span of the arc is determined by the phase of the multipath reflections relative to the LOS signal and the propagation angle of the multipath reflections relative to the LOS signal. The system of claim 5.
9. The H-spirographs for all channels with one multipath reflection are congruent under rotation and translation. The system of claim 8.
10. The V-spirograph for a channel without multipath reflections has a single point in the I / Q plane with a location determined by the analog chain mismatch and line of sight (LOS) propagation angle at the front end of the transceiver. The system of claim 6.
11. A V-spirograph for a channel having one multipath reflection has a circular arc in the I / Q plane having a center, a radius, and an arc span with non-uniformly spaced carrier frequencies; the location of the center of the arc, the radius of the arc and the span of the arc are determined by analog chain mismatch in the front end of the transceiver, the line of sight (LOS) propagation angle of the channel, the magnitude of the multipath reflections relative to the LOS signal, the phase of the multipath reflections relative to the LOS signal, and the propagation angle of the multipath reflections relative to the LOS signal; The system of claim 6.
12. The V-spirographs for all channels with one multipath reflection relative to a selected reference channel are similar under translational, rotational and radial scaling. The system of claim 11.
13. The H-spirograph for a channel having two or more multipath reflections has a complex hypocycloid in the I / Q plane. The system of claim 5.
14. The H-spirographs for all channels with two multipath reflections are congruent under one translation and one rotation. The system of claim 13.
15. the V-spirograph for a channel with two multipath reflections has an asymmetric trace in the I / Q plane, and the V-spirographs for all channels with two multipath reflections relative to a selected reference channel are similar under scaling, rotation and resampling; The system of claim 6.
16. 1. A method comprising: receiving one of a channel estimate and a beamforming estimate for a plurality of carrier frequencies at a transceiver in a multi-carrier, multi-antenna communication system; estimating parameters of multipath reflections using said channel estimates or a representation of beamforming feedback by a multi-layer perceptron feedforward neural network (FFNN); adjusting the phase and amplitude of the channel transmission based on the estimated parameters of the multipath reflections; Including, The method comprises: extracting, in a processor, from the channel estimates, frequency-dependent n-by-m H-matrix amplitude and phase data for a plurality of carrier frequencies; or extracting frequency-dependent n-by-m-column V-matrix amplitude and phase data for the plurality of carrier frequencies from the beamforming feedback. Further comprising: n is the number of transmit antennas in the multi-carrier, multi-antenna communication system, m is the number of receive antennas in the multi-carrier, multi-antenna communication system, and the product of m and n is the number of channels in the multi-carrier, multi-antenna communication system; a memory coupled to the processor and to the FFNN, storing weights, biases and activation functions for the FFNN and storing estimated parameters of the multipath reflections; The method further includes, in the processor, converting the n-by-m-column H-matrix amplitude and phase data into a frequency-dependent H-spirograph in an I / Q plane based on in-phase (I) and quadrature-phase (Q) components of the n-by-m-column H-matrix amplitude and phase data. method.
17. A method comprising: receiving one of a channel estimate and a beamforming estimate for a plurality of carrier frequencies at a transceiver in a multi-carrier, multi-antenna communication system; estimating parameters of multipath reflections using said channel estimates or a representation of beamforming feedback by a multi-layer perceptron feedforward neural network (FFNN); adjusting the phase and amplitude of the channel transmission based on the estimated parameters of the multipath reflections; Including, The method comprises: extracting, in a processor, from the channel estimates, frequency-dependent n-by-m H-matrix amplitude and phase data for a plurality of carrier frequencies; or extracting frequency-dependent n-by-m-column V-matrix amplitude and phase data for the plurality of carrier frequencies from the beamforming feedback. Further comprising: n is the number of transmit antennas in the multi-carrier, multi-antenna communication system, m is the number of receive antennas in the multi-carrier, multi-antenna communication system, and the product of m and n is the number of channels in the multi-carrier, multi-antenna communication system; a memory coupled to the processor and to the FFNN, storing weights, biases and activation functions for the FFNN and storing estimated parameters of the multipath reflections; The method further includes, in the processor, converting the n-by-m-column V matrix amplitude and phase data into a frequency-dependent V spirograph in an I / Q plane based on in-phase (I) and quadrature-phase (Q) components of the n-by-m-column V matrix amplitude and phase data. method.
18. The method further includes receiving, in the FFNN, in-phase data and quadrature-phase data of an H Spirograph at each carrier frequency as an input vector, and outputting an effective propagation angle, an effective channel attenuation, and an effective channel delay for each channel based on characteristics of the H Spirograph.
17. The method of claim 16.
19. The method further includes the step of receiving, in the FFNN, in-phase and quadrature-phase data of a V-spirograph at each frequency as an input vector, and outputting an effective propagation angle and a relative delay and a relative attenuation of each channel with respect to a selected reference channel based on characteristics of the V-spirograph.
20. The method of claim 17.
20. The H-spirograph for a channel without multipath reflections has a single point in the I / Q plane with a location determined by analog chain mismatch in the transceiver front-end and the line-of-sight (LOS) propagation angle of the channel.
17. The method of claim 16.
21. The H-spirograph for a channel with one multipath reflection has a circular arc in the I / Q plane having a center, a radius, and an arc span with uniformly spaced carrier frequency points; the location of the center of the arc is determined by analog chain mismatch and line of sight (LOS) propagation angle in the front end of the transceiver; the radius of the arc is determined by the strength of the multipath reflections relative to the LOS signal; the span of the arc is determined by the phase of the multipath reflections relative to the LOS signal and the propagation angle of the multipath reflections relative to the LOS signal.
17. The method of claim 16.
22. The H-spirographs for all channels with one multipath reflection are congruent under rotation and translation.
22. The method of claim 21.
23. The V-spirograph for a channel without multipath reflections has a single point in the I / Q plane with a location determined by the analog chain mismatch and line of sight (LOS) propagation angle at the front end of the transceiver.
20. The method of claim 17.
24. A V-spirograph for a channel having one multipath reflection has a circular arc in the I / Q plane having a center, a radius, and an arc span with non-uniformly spaced carrier frequencies; the location of the center of the arc, the radius of the arc, and the span of the arc are determined by analog chain mismatch in the front end of the transceiver, the line of sight (LOS) propagation angle of the channel, the strength of the multipath reflections relative to a LOS signal, the phase of the multipath reflections relative to the LOS signal, and the propagation angle of the multipath reflections relative to the LOS signal; 20. The method of claim 17.
25. The V-spirographs for all channels with one multipath reflection relative to a selected reference channel are similar under translational, rotational and radial scaling.
25. The method of claim 24.
26. The H-spirograph for a channel with two multipath reflections has a complex hypocycloid in the I / Q plane.
17. The method of claim 16.
27. The H-spirographs for all channels with two multipath reflections are congruent under one translation and one rotation.
27. The method of claim 26.
28. the V-spirograph for a channel with two multipath reflections has an asymmetric trace in the I / Q plane, and the V-spirographs for all channels with two multipath reflections relative to a selected reference channel are similar under scaling, rotation and resampling; 20. The method of claim 17.
29. 1. A method comprising: Providing training data including simulations of channel estimation and beamforming feedback in a multi-channel, multi-carrier communication system for a number of multipath scenarios; training a multi-layer perceptron feed-forward neural network (FFNN) with the in-phase and quadrature-phase representations of the channel estimates and beamforming feedback to estimate parameters of multipath reflections; storing the weights, biases and activation functions of the trained FFNN in memory; Including, The step of training the FFNN comprises: generating, for each channel at each carrier frequency, a set of Spirograph features from the in-phase and quadrature-phase representations of the channel estimates and beamforming feedback for each of the plurality of multipath scenarios; applying the Spirograph feature set in batches to the FFNN in a forward pass to generate a loss function having a representation of the training data; propagating the loss function in a backward pass through the multi-layer perceptron feedforward neural network to update weights and biases of the multi-layer perceptron feedforward neural network to minimize the loss function; Including, method.
Citation Information
Patent Citations
Method and system for indoor geolocation using an impulse response fingerprinting technique
US20070010956A1
Method and Apparatus for Directing an Antenna Beam based on a Location of a Communication Device
US20170195893A1
