Perception method, perception system, electronic equipment and related device

Abstract

The embodiment of the invention provides a sensing method, a sensing system, electronic equipment and a related device, and relates to the technical field of terminals. The sensing method comprises the steps that a sensing initiating end sends a first signal to a sensing response end, so that the sensing response end performs sensing measurement based on a data symbol for communication in the first signal, and a sensing measurement result is obtained. The perception initiating end can obtain the perception measurement result transmitted by the perception responding end. The sensing measurement result comprises the distance and the speed of each target in the at least one target between the sensing initiating end and the sensing responding end. According to the sensing method, environment sensing can be realized without special pilot signals, and the probability of occurrence of reduced communication efficiency caused by the fact that the special pilot signals occupy communication resources is reduced.

Description

Technical Field

[0001] The present application relates to the field of terminal technology, and in particular to a sensing method, a sensing system, an electronic device and related devices. Background Art

[0002] Some communication systems have both communication and perception functions, which expands the application scope of communication systems.

[0003] In some implementations, a transmitter in a communication system can send wireless signals to a receiver. While enabling communication between the transmitter and receiver, the transmitter can also sense the environment between the transmitter and receiver based on the wireless signals. The transmitter can then make decisions based on the sensed environmental information.

[0004] However, in some implementations, when communication and perception are performed simultaneously, there may be a problem of low communication efficiency between the transmitter and the receiver. Summary of the Invention

[0005] The embodiments of the present application provide a sensing method, a sensing system, an electronic device, and related devices, which are applied to the field of terminal technology and can reduce the probability of low communication efficiency between a transmitting end and a receiving end.

[0006] In a first aspect, embodiments of the present application provide a sensing method, comprising: receiving a first signal indicating a data symbol of a communication; and transmitting a sensing measurement result obtained based on a measurement of the data symbol.

[0007] For example, the first signal in this embodiment may be Figure 1 Orthogonal Frequency Division Multiplexing (OFDM) signal in the embodiment shown. It can also be Figure 2-Figure 9 The first signal in the embodiment shown. The perception measurement results in this embodiment may include Figure 1 The distance and speed of the target in the embodiment shown can also be Figure 2 The sensory measurement results in the embodiment shown can also be Figure 6 The updated target parameters of each target in at least one target output by the perception algorithm module in the embodiment shown can also be Figure 8-Figure 9 Perceptual measurements or system outputs in the illustrated embodiment.

[0008] The specific implementation principle of this embodiment can be found in Figure 1 The specific implementation principle of the embodiment shown can also be found in Figure 2 The specific implementation principle of the embodiment shown can also be found in Figure 3-Figure 8 The specific implementation principle of the embodiment shown can also be found in Figure 9 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0009] Thus, taking the execution subject of this embodiment as the perception responding end as an example, the perception method provided by the embodiment of the present application, the perception responding end can receive the first signal of the data symbol sent by the perception initiating end for indicating the communication. The perception responding end can obtain the perception measurement result based on the data symbol measurement in the first signal, and send the perception measurement result to the perception initiating end. The perception measurement result, for example, the distance and speed of each target in at least one target between the perception initiating end and the perception responding end. The perception initiating end can obtain the perception measurement result sent by the perception responding end, and realize the perception of the environment between the perception initiating end and the perception responding end. Compared with the method of performing environmental perception through a dedicated pilot signal, in the perception method of the embodiment of the present application, the perception responding end can obtain the perception measurement result based on the data symbol measurement, and there is no need for the perception initiating end to send a dedicated pilot signal, which can reduce the occupation of communication resources by the dedicated pilot signal, thereby reducing the probability of communication efficiency being reduced due to the occupation of communication resources by the dedicated pilot signal. It can also save communication resources and improve the resource utilization of data symbols. Among them, the perception initiating end can be Figures 1-9 The sensing initiator or initiator in the embodiment shown. The sensing responder can be Figures 1-9 The sensing responder or responder in the illustrated embodiment.

[0010] In a possible implementation, before receiving the first signal, the method further includes: sending first indication information, where the first indication information is used to indicate that data symbols are used to perform perception measurement.

[0011] For example, the first indication information may be Figure 3 The first indication information in the embodiment shown. The specific implementation principle of the embodiment of this application can be found in Figure 3 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0012] In this way, the data symbols in the first signal are used for perception measurement to achieve environmental perception without the need for a dedicated pilot signal, thereby reducing the probability of communication resources being occupied by the dedicated pilot signal, resulting in reduced communication efficiency.

[0013] In a possible implementation manner, before receiving the first signal, the method further includes: sending first capability information, where the first capability information is used to indicate support for a data symbol-based perception measurement function.

[0014] Exemplarily, the first capability information may be Figure 3 The first capability information in the embodiment shown. The specific implementation principle of the embodiment of this application can be found in Figure 3 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0015] In this way, before receiving the first signal, capability negotiation between the perception responding end and the perception initiating end is achieved, which facilitates the perception responding end to use data symbols to perform perception measurement and achieve environmental perception.

[0016] In one possible implementation, before sending the perception measurement result, the method further includes: inputting the first signal into a first model, so that the first model outputs the perception measurement result. The first model is configured to perform perception processing on a data symbol indicated by the first signal to obtain the perception measurement result. The data symbol is obtained by demodulating the first signal, or is obtained by updating a demodulation result of the demodulated first signal.

[0017] For example, the first model can be Figures 4-10 The neural network model or feedback loop system in the embodiment shown. The specific implementation principle of this embodiment can be found in Figure 4-Figure 8 The specific implementation principle of the embodiment shown can also be found in Figure 9 The specific implementation principle of the illustrated embodiment.

[0018] In this way, the first model performs perception processing on the data symbols indicated by the first signal to obtain perception measurement results, enabling environmental perception and communication without the need for dedicated pilot signals. The data symbols used for perception processing are obtained by demodulating the first signal, which can improve resource utilization. The data symbols used for perception processing are obtained by updating the demodulation results of the demodulated first signal. By updating the data symbols to obtain more accurate data symbols, the measurement accuracy of the perception measurement results measured based on the data symbols can be improved.

[0019] In one possible implementation, the sensing measurement result includes the speed and distance of each target object in at least one target object, where the target object is an object between a sensing initiator and a sensing responder. Sensing and processing data symbols in a first signal to obtain the sensing measurement result includes calculating a channel frequency response matrix of a channel corresponding to the first signal based on the data symbols and the first signal. Calculating and processing the channel frequency response matrix to obtain the distance and speed of each target object in the at least one target object.

[0020] For example, the target object may be Figures 1-10 In the embodiment shown, the target object or target, the target object can be simply referred to as the target. The distance can be Figures 1-10 The target distance or distance in the embodiment shown. The speed can be Figures 1-10 The target speed or speed in the embodiment shown. The specific implementation principle of this embodiment can be found in Figure 5-Figure 6 The specific implementation principle of the illustrated embodiment.

[0021] In this way, perception measurement is performed based on the data symbols to obtain a perception measurement result.

[0022] In one possible implementation, the first model includes at least one layer of neural network. The channel frequency response matrix satisfies the following formula:

[0023]

[0024] in, is the channel frequency response matrix calculated by the k-th neural network in at least one layer of neural network or the channel matrix calculated by the k-th neural network, for The elements in is the matrix representation of the first signal, is a matrix The element in row n and column l in is the matrix representation of the data symbols demodulated or updated by the k-th layer neural network, is a matrix The element in the nth row and lth column, 1≤n≤N, 1≤l≤L, k is the sequence number of the neural network layer in at least one neural network, n is the row sequence number and is an integer, l is the column sequence number and is an integer, N is the number of matrix rows, L is the number of matrix columns, k is a natural number, N is an integer, and L is an integer.

[0025] In this way, the channel frequency response matrix is ​​determined based on the first signal and the data symbols in the first signal. The channel frequency response matrix is ​​used for perceptual measurement to achieve determination of perceptual measurement results based on the data symbols.

[0026] In one possible implementation, calculating and processing a channel frequency response matrix to obtain the distance and velocity of each target object in the at least one target object includes: calculating the channel frequency response matrix using a spectral estimation algorithm to obtain the time delay and Doppler shift corresponding to the distance and velocity of each target object. The distance and velocity of each target object are determined based on the time delay and Doppler shift corresponding to the distance and velocity of each target object. Each target object is each target object in the at least one target object.

[0027] For example, the spectrum estimation algorithm may be a 2D MUSIC algorithm. The specific implementation principle of this embodiment may refer to the specific implementation principle of the embodiment shown in S603-S604, or the specific implementation principle of the embodiment shown in S603-S605.

[0028] In this way, perception measurement is performed based on the data symbols to obtain a perception measurement result.

[0029] In a possible implementation, the first signal is modulated by orthogonal frequency division multiplexing modulation. A spectrum estimation algorithm is used to calculate the channel frequency response matrix to obtain the time delay and Doppler frequency shift corresponding to the distance and speed of each target object, including: using a vectorization operator to convert the channel frequency response matrix Vectorization, get a one-dimensional long vector . For long vectors Calculate the covariance matrix to get the covariance matrix R. Perform eigenvalue decomposition and subspace separation on the covariance matrix R to get the noise subspace Based on the noise subspace The constructed power spectrum is used for peak search to obtain the time delay and Doppler shift corresponding to the distance and speed of each target object.

[0030] For example, the specific implementation principle of this embodiment may refer to the specific implementation principle of S603, or may refer to the specific implementation principles of S6031-S6037.

[0031] In this way, the distance and speed of each target object can be determined based on the time delay and Doppler shift corresponding to the distance and speed of each target object.

[0032] In a possible implementation, the time delay and Doppler shift corresponding to at least one target object include the time delay and Doppler shift corresponding to each peak in I peaks. Performing a peak search on the constructed power spectrum to obtain a time delay and Doppler shift corresponding to the distance and velocity of each target object includes: performing a peak search on the power spectrum to obtain a plurality of peaks; determining I peaks from the plurality of peaks that are greater than or equal to a first parameter; the coordinate position of each of the I peaks in the power spectrum includes the time delay and Doppler shift corresponding to each peak; the first parameter is determined based on the noise power of the first signal; I is the number of the at least one target object, and I is a positive integer.

[0033] In this way, the first parameter is used to perform threshold judgment on the multiple peaks obtained by the peak search. It is possible to determine the peak generated by the target from the multiple peaks searched, that is, to determine the I peaks corresponding to I targets between the initiator and the responder from the multiple peaks searched. In the case where the side lobe interferes with the peak generated by the target, it is possible to distinguish the true target peak from the side lobe (or noise). The true target peak is a peak greater than or equal to the first parameter. The side lobe (or noise) corresponds to a peak less than the first parameter. The true target peak is, for example, any one of the I peaks. This can improve the accuracy of the perception measurement results. The first parameter can be Figure 6 and Figures 8-10 Multi-target detection threshold in an embodiment.

[0034] In one possible implementation, the long vector Satisfies the formula:

[0035]

[0036] The covariance matrix R satisfies the formula:

[0037]

[0038] Power spectrum Satisfies the formula:

[0039]

[0040] Where T represents transpose, express The transpose of for The first element of for The second column element of for The Lth column element of for The conjugate transpose of is an operator, and Indicates the calculation of covariance, for The conjugate transpose of for The conjugate transpose of is the two-dimensional joint steering vector, is the power spectrum The independent variable represents the delay, is the power spectrum The independent variable represents the Doppler frequency shift. This makes it easier to search for peak values ​​based on the power spectrum and thus obtain the target's distance and speed.

[0041] In one possible implementation, the two-dimensional joint steering vector Satisfies the formula:

[0042]

[0043] in, represents the Kronecker product, which can also represent the tensor product or outer product. The one-dimensional distance-oriented vector constructed for the k-th layer of the neural network, , is the one-dimensional velocity guidance vector constructed for the k-th layer of the neural network, , e is a natural constant or transcendental number, j is the imaginary unit, is the carrier spacing of OFDM, for The independent variable is The corresponding delay, for The independent variable is The corresponding Doppler shift is, is the period of an OFDM symbol or the length of time occupied by an OFDM symbol, express This makes it easier to construct the power spectrum.

[0044] In a possible implementation, the distance of the i-th target object among the at least one target object satisfies the formula:

[0045]

[0046] The velocity of the i-th target object among at least one target object satisfies the formula:

[0047]

[0048] in, is the distance of the i-th target object extracted by the k-th layer neural network, The time delay corresponding to the distance of the i-th target object calculated by the k-th layer neural network, is the speed of the i-th target object extracted by the k-th layer neural network, is the Doppler frequency shift corresponding to the velocity of the i-th target object calculated by the k-th layer neural network, c is the propagation speed of electromagnetic waves or the speed of light, is the preset carrier frequency.

[0049] In this way, the distance of the target is determined based on the target corresponding time delay, and the speed of the target is determined based on the target corresponding Doppler frequency shift.

[0050] In one possible implementation, the perception measurement result includes an updated distance of each target object in at least one target object and an updated speed of each target object in at least one target object. The method further includes: performing a fusion update on the distance of the i-th target object and the speed of the i-th target object to obtain an updated distance of the i-th target object and an updated speed of the i-th target object. The updated distance of the i-th target object and the updated speed of the i-th target object satisfy the formula:

[0051]

[0052] in, , , is the updated distance of the i-th target object obtained by the k-th layer neural network, is the updated speed of the i-th target object obtained by the k-th layer neural network, is the updated distance of the i-th target object obtained by the k-1-th layer neural network, is the updated speed of the i-th target object obtained by the k-1-th layer neural network, is the second parameter of the k-th layer neural network, which is determined based on the noise power of the first signal. [ ] represents a matrix.

[0053] In this way, since the second parameter is generated based on the noise power of the first signal, using the second parameter to fuse and update the extracted target parameter can further reduce the impact of noise such as side lobes on the target parameter, improve the accuracy of the updated target parameter, and thus improve the perception accuracy. Among them, the second parameter can be Figure 6 and Figures 8-10 Parameter fusion coefficients in the embodiment: Target parameters, such as the distance and speed of the target, or, for example, the distance of the i-th target object extracted by the k-th layer neural network and the speed of the i-th target object extracted by the k-th layer neural network.

[0054] In one possible implementation, the data symbols include initial data symbols, which are obtained by demodulating the first signal in the following manner: performing channel estimation based on the pilot symbols in the first signal to obtain initial channel information, and demodulating the first signal using the initial channel information to obtain initial data symbols.

[0055] For example, the initial channel information may be the initial channel information in the embodiment shown in S501, or the initial channel matrix in the embodiment shown in S501. For the specific implementation principle of demodulating the first signal using the initial channel information to obtain the initial data symbol, reference may be made to the specific implementation principle of S503-S504.

[0056] This facilitates perceptual measurements using data symbols.

[0057] In one possible implementation, the first model includes a K+1-layer neural network, where K is an integer greater than or equal to 1. The data symbols also include data symbols updated by the K+1-layer neural network. The data symbols updated by the K+1-layer neural network are obtained by updating the demodulation result after demodulating the first signal in the following manner: using the third parameter of the K+1-layer neural network, the channel matrix updated by the K-layer neural network, and the noise vector updated by the K-layer neural network, the data symbols updated by the K-layer neural network are updated to obtain data symbols updated by the K+1-layer neural network. The third parameter is determined based on the noise power of the first signal, and the data symbols updated by the 1st layer neural network are initial data symbols.

[0058] For example, the third parameter may be Figure 5 The signal update coefficients in the embodiment shown. The noise vector can be Figure 5 The residual or residual vector in the embodiment shown. The specific implementation principle of this embodiment can be found in the specific implementation principle of S504, which will not be repeated here.

[0059] In this way, the remaining data symbols in the first signal except for the demodulated data symbols can be demodulated more accurately, thereby obtaining more accurate data symbols.

[0060] In one possible implementation, the data symbol updated by the K+1th layer neural network is Satisfies the formula:

[0061]

[0062] in, , , is the symbol of the first signal after being transformed into the symbol domain, For the The data symbols updated by the layer neural network, Can be the first The third parameter of the layer neural network, is the channel matrix Middle Path The corresponding amplitude component, For the The channel matrix updated by the layer neural network, is the residual vector or noise vector updated by the K-th layer neural network, is the channel matrix Middle Path The corresponding phase component, for The transpose of for The transpose of is the path number of the path for transmitting the first signal.

[0063] In this way, the iterative algorithm based on gradient descent is used to solve , it is possible to more accurately demodulate the first signal except for the demodulated data symbols (such as ) to obtain more accurate data symbols.

[0064] In one possible implementation, the residual vector updated by the K-th layer neural network satisfies the formula:

[0065]

[0066] in, is the residual vector or noise vector updated by the K-th layer neural network, It can be the residual vector or noise vector updated for the K-1th layer neural network, It can be the fourth parameter of the K-th layer neural network, and the fourth parameter is determined based on the noise power of the first signal. is the channel matrix Middle Path The corresponding amplitude component, is the channel matrix updated by the K-1th layer neural network, is the channel matrix Middle Path The corresponding phase component.

[0067] Exemplarily, the fourth parameter may be Figure 5 The residual update coefficient or noise update parameter in the illustrated embodiment can be referred to the specific implementation principle of S505 for the specific implementation principle of this embodiment, which will not be described in detail here.

[0068] Thus, by introducing the momentum term Updating the residual can accelerate convergence and suppress oscillation, improve noise estimation accuracy, and reduce the probability of large noise estimation errors leading to reduced accuracy of acquired data symbols. This can further improve data demodulation accuracy.

[0069] In one possible implementation, Channel matrix updated by layer neural network Satisfies the formula:

[0070]

[0071] in, , is the fifth parameter of the K-th layer neural network, and the fifth parameter is determined based on the noise power of the first signal. is the updated channel matrix of the K-1th layer neural network, is a communication and environment perception channel matrix reconstructed by the K-th layer neural network based on the Doppler frequency shift of each target object in at least one target object, the time delay of each target object, the receiving angle of each target object, and the transmitting angle of each target object, the Doppler frequency shift of the i-th target object in at least one target object is calculated based on the velocity of the i-th target object, the time delay of the i-th target object in at least one target object is calculated based on the distance of the i-th target object, and i is a positive integer.

[0072] Exemplarily, the fifth parameter may be Figure 7 The channel update coefficient in the embodiment shown. The specific implementation principle of this embodiment can be found in Figure 7 The specific implementation principle of the illustrated embodiment or S705 will not be described in detail here.

[0073] In this way, the fifth parameter can be used to control the weight balance between the channel matrix updated by the K-1th layer neural network and the channel reconstruction of the Kth layer neural network, making the updated channel matrix more accurate. In addition, the channel matrix reconstructed by the Kth layer neural network It can be reconstructed based on the perception results measured by the K-th layer neural network by introducing the channel update coefficient of the K-th layer neural network , it is possible to control the contribution of the perception results of the Kth layer neural network and the K-1th layer neural network to channel reconstruction, further improving the accuracy of the updated channel matrix. Perception results, such as perception measurement results or updated target parameters.

[0074] In one possible implementation, the third parameter of the k-th layer neural network is included in the hyperparameter set of the k-th layer neural network. , Satisfies the formula:

[0075]

[0076] Where k is the sequence number of the neural network layer in the K+1 layer neural network, 1≤k≤K+1, represents the activation function, 、 、 and are all model parameters of the k-th layer neural network, 、 、 and All are preset in the first model. 、 、 and All are obtained by training the first model. is the noise power of the first signal, It also includes the first parameter, second parameter, fourth parameter and fifth parameter of the k-th layer neural network, the first parameter is used to judge the threshold of the searched peak, the second parameter is used to update the fusion of the distance and speed of the target object, the fourth parameter is used to update the noise vector or to update the residual vector, and the fifth parameter is used to update the channel matrix.

[0077] For example, the specific implementation principle of this embodiment can be found in Figure 8 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0078] In this way, the first parameter, the second parameter, the third parameter, the fourth parameter and the fifth parameter are determined based on the noise power of the first signal, so that the first model adopts these parameters to improve the perception measurement accuracy.

[0079] In a possible implementation, the perception measurement result is carried in a media access control layer frame payload or a first management frame, and the first management frame is used to transmit the perception measurement result.

[0080] In this way, the perception measurement result can be transmitted through the medium access control layer frame or the first management frame.

[0081] In a second aspect, embodiments of the present application provide a sensing method, comprising: transmitting a first signal, the first signal being used to indicate a data symbol of a communication, and receiving a sensing measurement result, the sensing measurement result being obtained based on a measurement of the data symbol.

[0082] In this way, taking the execution subject of this embodiment as the perception initiator as an example, the perception method provided by the embodiment of the present application, the perception initiator can send a first signal for indicating the data symbol of communication to the perception responder, so that the perception responder obtains the perception measurement result based on the data symbol measurement in the first signal. The perception initiator can obtain the perception measurement result sent by the perception responder, and realize the perception of the environment between the perception initiator and the perception responder. Compared with the method of performing environmental perception through a dedicated pilot signal, in the perception method of the embodiment of the present application, the perception responder can obtain the perception measurement result based on the data symbol measurement, without the need for the perception initiator to send a dedicated pilot signal, which can reduce the occupation of communication resources by the dedicated pilot signal, and thus reduce the probability of communication efficiency being reduced due to the occupation of communication resources by the dedicated pilot signal. It can also save communication resources and improve the resource utilization of data symbols. Among them, the perception initiator can be Figures 1-9 The sensing initiator or initiator in the embodiment shown. The sensing responder can be Figures 1-9 The sensing responder or responder in the illustrated embodiment.

[0083] In a possible implementation, before sending the first signal, the method further includes: receiving first indication information, where the first indication information is used to indicate that data symbols are used to perform perception measurement.

[0084] In this way, the data symbols in the first signal are used for perception measurement to achieve environmental perception without the need for a dedicated pilot signal, thereby reducing the probability of communication resources being occupied by the dedicated pilot signal, resulting in reduced communication efficiency.

[0085] In a possible implementation manner, before sending the first signal, the method further includes: receiving first capability information, where the first capability information is used to indicate support for a data symbol-based perception measurement function.

[0086] In this way, before sending the first signal, capability negotiation between the perception responding end and the perception initiating end is achieved, which facilitates the perception responding end to use data symbols to perform perception measurement and realize environmental perception.

[0087] In a third aspect, an embodiment of the present application provides a perception system, comprising: a perception initiator and a perception responder, the perception responder being used to execute the method described in the first aspect or any possible implementation of the first aspect, and the perception initiator being used to execute the method described in the second aspect or any possible implementation of the second aspect.

[0088] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising: a signal demodulation module, a perception algorithm module, a channel reconstruction module, and a noise perception module. The electronic device is configured to receive a first signal, the first signal being used to indicate a data symbol for communication. The signal demodulation module is configured to demodulate the first signal to obtain the data symbol indicated by the first signal, or to update the demodulation result after demodulating the first signal to obtain the data symbol in the first signal. The perception algorithm module is configured to perform perception processing on the data symbols in the first signal to obtain the distance and velocity of each target object in at least one target object. The channel reconstruction module is configured to reconstruct a communication and environment perception channel matrix based on the distance and velocity of each target object in the at least one target object, and to update the reconstructed communication and environment perception channel matrix. The updated communication and environment perception channel matrix is ​​used to update the demodulation result after demodulating the first signal. The electronic device is further configured to transmit a perception measurement result, the perception measurement result including the distance and velocity of each target object in at least one target object between a perception initiator and a perception responder, the perception measurement result being obtained based on data symbol measurements.

[0089] For example, the signal demodulation module may be Figures 4-10 The signal demodulation module in the embodiment shown. The perception algorithm module can be Figures 4-10 The sensing algorithm module in the embodiment shown. The channel reconstruction module can be Figures 4-10The channel reconstruction module in the embodiment shown. The noise perception module can be Figures 4-10 The noise perception module in the illustrated embodiment.

[0090] In this way, a sensing measurement result is obtained based on the measurement of the data symbols in the first signal, thereby achieving perception of the environment between the sensing initiator and the sensing responder. Compared to the method of performing environmental perception through dedicated pilot signals, the sensing method in the embodiment of the present application obtains a sensing measurement result based on the measurement of data symbols, eliminating the need for dedicated pilot signals. This can reduce the occupation of communication resources by dedicated pilot signals, thereby reducing the probability of communication resources being occupied by dedicated pilot signals, resulting in reduced communication efficiency. This can also save communication resources and improve the resource utilization of data symbols.

[0091] In a fifth aspect, embodiments of the present application provide a sensing device, which may be an electronic device or a chip or chip system within an electronic device. The sensing device may include a display unit and a processing unit. When the sensing device is an electronic device, the display unit may be a display screen. The display unit is configured to perform the display step so that the electronic device implements a sensing method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. When the sensing device is an electronic device, the processing unit may be a processor. The sensing device may also include a storage unit, which may be a memory. The storage unit is configured to store instructions, and the processing unit executes the instructions stored in the storage unit so that the electronic device implements a sensing method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. When the sensing device is a chip or chip system within an electronic device, the processing unit may be a processor. The processing unit executes the instructions stored in the storage unit so that the electronic device implements a sensing method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. The storage unit may be a storage unit within the chip (eg, a register, a cache, etc.), or a storage unit within the electronic device that is located outside the chip (eg, a read-only memory, a random access memory, etc.).

[0092] In a sixth aspect, an embodiment of the present application provides an electronic device comprising a processor and a memory, the memory being used to store computer-executable instructions, and the processor being used to run the computer-executable instructions stored in the memory to execute the method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect.

[0093] In the seventh aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program or instruction is stored. When the computer program or instruction is run on a computer, the computer executes the method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect.

[0094] In an eighth aspect, an embodiment of the present application provides a computer program product comprising a computer program, which, when the computer program is run, enables the computer to execute the method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect.

[0095] In a ninth aspect, an embodiment of the present application provides a chip or chip system, comprising at least one processor and a communication interface, wherein the communication interface and the at least one processor are interconnected via a line, and the at least one processor is configured to run a computer program or instruction to execute the method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. The communication interface in the chip may be an input / output interface, a pin, or a circuit, etc.

[0096] In one possible implementation, the chip or chip system described above in the embodiments of the present application further includes at least one memory, wherein instructions are stored in the at least one memory. The memory may be a storage unit within the chip, such as a register or cache, or a storage unit of the chip (such as a read-only memory or random access memory).

[0097] It should be understood that the fifth to ninth aspects of the embodiments of the present application correspond to the technical solutions of the first or second aspects of the embodiments of the present application, and the beneficial effects achieved by each aspect and the corresponding feasible implementation methods are similar and will not be repeated. BRIEF DESCRIPTION OF THE DRAWINGS

[0098] Figure 1 A schematic diagram of a perception system provided in an embodiment of the present application;

[0099] Figure 2 A schematic diagram of a flow chart of a sensing method provided in an embodiment of the present application;

[0100] Figure 3 Another flowchart of the sensing method provided in the embodiment of the present application;

[0101] Figure 4 A schematic diagram of a feedback loop structure provided in an embodiment of the present application;

[0102] Figure 5A schematic diagram of a flow chart of a signal demodulation module provided in an embodiment of the present application;

[0103] Figure 6 A schematic diagram of a flow chart of the perception algorithm module provided in an embodiment of the present application;

[0104] Figure 7 A schematic diagram of a flow chart of a channel reconstruction module provided in an embodiment of the present application;

[0105] Figure 8 A schematic diagram of a cascade connection of the noise perception module provided in an embodiment of the present application;

[0106] Figure 9 A schematic diagram of the structure of a single-layer neural network in the neural network model provided in an embodiment of the present application;

[0107] Figure 10 A schematic diagram of the training process of the neural network model provided in the embodiment of the present application. DETAILED DESCRIPTION

[0108] To facilitate a clear description of the technical solutions of the embodiments of the present application, some of the terms and technologies involved in the embodiments of the present application are briefly introduced below:

[0109] 1. Physical layer protocol data unit (PLCP protocol data unit, PPDU)

[0110] PPDU can be understood as the data unit of the physical layer (PHY) in wireless communication.

[0111] According to the data encapsulation format defined by the physical layer convergence protocol (PLCP), the PPDU may include a preamble, a PLCP header, a PLCP service data unit (PSDU), and a trailer.

[0112] The preamble can be used for synchronization and coarse channel estimation. Synchronization can include time and frequency synchronization. The data payload can carry media access control layer (MAC) frames, such as data packets. The trailer can include a frame check sequence (FCS). The FCS can be used to detect errors during frame transmission.

[0113] 2. Orthogonal frequency division multiplexing (OFDM)

[0114] OFDM can be understood as a digital multicarrier modulation technique. By dividing a data stream into multiple lower-rate sub-signals and transmitting them using orthogonal subcarriers, it improves spectrum efficiency and mitigates multipath interference. OFDM can be used in wireless communication systems, such as wireless fidelity (WiFi), long-term evolution (LTE), terrestrial digital video broadcasting (DVB-T), fifth-generation (5G) systems, new radio (NR), and future evolutionary communication systems. It achieves efficient spectrum utilization and robustness to channel fading.

[0115] For example, in actual transmission of the PPDU, the PPDU may be modulated into an OFDM signal for transmission.

[0116] 3. Null Data Packet Announcement (NDPA), Null Data Packet (NDP), and Trigger Frames

[0117] NDPA can be understood as a control frame that does not carry actual data and can be used to notify that an NDP is about to be sent.

[0118] NDPs are special data packets that do not contain data payloads, typically used for channel measurement and other control purposes. Data payloads, such as data symbols, can also be used to send dedicated pilot signals, which can be used for sensing.

[0119] Trigger frames can be understood as control frames used to coordinate device communications in a multi-user environment.

[0120] 4. Other terms

[0121] In the embodiments of this application, terms such as "first" and "second" are used to distinguish between identical or similar items with substantially the same functions and effects. For example, the terms "first chip" and "second chip" are used solely to distinguish between different chips and do not define their order. Those skilled in the art will understand that terms such as "first" and "second" do not define the quantity or execution order, and do not necessarily define differences.

[0122] It should be noted that in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0123] In the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: the existence of A alone, the existence of A and B at the same time, and the existence of B alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, a--c, bc, or abc, where a, b, c can be single or plural.

[0124] 5. Electronic devices

[0125] The electronic devices of the embodiments of the present application may include handheld devices, vehicle-mounted devices, etc. with communication functions. For example, some electronic devices include: mobile phones, tablet computers, PDAs, laptop computers, mobile internet devices (MIDs), virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to wireless modems, vehicle-mounted devices, terminal devices in 5G networks, or terminal devices in future evolved public land mobile networks (PLMNs), etc. The embodiments of the present application are not limited to this.

[0126] As an example and not a limitation, in the embodiments of the present application, the electronic device may also be a wearable device. Wearable devices may also be referred to as wearable smart devices, which are a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing, and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not only hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are fully functional, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0127] In addition, in the embodiments of the present application, the electronic device can also be a terminal device in the Internet of Things (IoT) system. IoT is an important part of the future development of information technology. Its main technical feature is to connect objects to the network through communication technology, thereby realizing an intelligent network that interconnects people and machines and things.

[0128] The electronic devices in the embodiments of the present application may also be referred to as: terminal equipment, user equipment (UE), mobile station (MS), mobile terminal (MT), access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication equipment, user agent or user device, etc.

[0129] In the embodiments of the present application, electronic devices or network devices include a hardware layer, an operating system layer running on top of the hardware layer, and an application layer running on top of the operating system layer. The hardware layer includes hardware such as a central processing unit (CPU), a memory management unit (MMU), and memory (also known as main memory). The operating system can be any one or more computer operating systems that implement business processing through processes, such as the Linux operating system, Unix operating system, Android operating system, iOS operating system, or Windows operating system. The application layer includes applications such as browsers, address books, word processing software, and instant messaging software.

[0130] Some communication systems have both communication and perception capabilities. This perception function can be applied to scenarios such as indoor positioning, security systems, and health monitoring, expanding the application scope of communication systems. A communication system can include a transmitter and a receiver. The transmitter and receiver can communicate and / or achieve perception using wireless signals.

[0131] With the widespread deployment and application of indoor WiFi technology, there's a growing demand for integrated sensing and communication (ISAC) that leverages existing WiFi infrastructure for both indoor communication and environmental awareness. In many indoor environments, 5G base station coverage may be absent, or 5G NR signal perception may be poor due to factors like penetration loss. Leveraging ubiquitous WiFi signals for sensing can improve this performance.

[0132] In a scenario where WiFi signals are used for perception, the transmitter may be a WiFi access point (AP) device, the receiver may be a user equipment (UE), and the wireless signal transmitted between the transmitter and the receiver may be a WiFi signal.

[0133] In some possible implementations, environmental perception is achieved by using dedicated pilot signals or dedicated sensing sequences. For example, using dedicated pilot signals as an example, some transmitting ends can communicate and sense by sending WiFi signals containing dedicated pilot signals for sensing to receiving ends. Dedicated pilot signals can be sent in specific sensing time slots. To improve sensing performance, a larger number of dedicated pilot signals may be included in the transmitted signal. A larger number of dedicated pilot signals will occupy more time and frequency resources of the transmitted signal. The transmitted signal can be understood as the WiFi signal sent by the transmitting end.

[0134] In this way, valuable communication resources are occupied by a large number of dedicated pilot signals, which may affect the data transmission between the transmitter and the receiver, and further reduce the communication efficiency between the transmitter and the receiver, resulting in low communication efficiency between the transmitter and the receiver.

[0135] In view of this, an embodiment of the present application provides a sensing method. A sensing initiator may send a first signal to a sensing responder. Correspondingly, the sensing responder may receive the first signal. The first signal may indicate a data symbol in a communication. The sensing responder may obtain a sensing measurement result based on data symbol measurements. The sensing responder may send the sensing measurement result to the sensing initiator. Correspondingly, the sensing initiator may receive the sensing measurement result. For example, if the transmitted signal is the first signal, obtaining the sensing measurement result based on data symbol measurements in the transmitted signal can achieve environmental perception. Because data symbols are used for sensing, dedicated pilot signals in the transmitted signal can be eliminated. This reduces communication resources occupied by dedicated pilot signals and, in turn, improves communication efficiency between a transmitting end (e.g., a sensing initiator) and a receiving end (e.g., a sensing responder). The sensing initiator may be a transmitter of a wireless signal, such as a Wi-Fi signal. The sensing responder may be a receiver of wireless signals. The first signal may be a Wi-Fi signal.

[0136] Figure 1 A schematic diagram of a perception system provided in an embodiment of the present application is shown.

[0137] In the embodiment of the present application, the perception system also belongs to the communication system.

[0138] like Figure 1As shown, the perception system may include a WiFi access point (WiFi AP) and a user equipment (UE). The WiFi AP can use multiple antennas to simultaneously transmit and receive signals. The UE can also use multiple antennas to simultaneously transmit and receive signals. In one possible implementation, the WiFi AP can be the perception initiator, and the UE can be the perception responder. In another possible implementation, the UE can be the perception initiator, and the WiFi AP can be the perception responder.

[0139] Taking a WiFi AP as a sensing initiator and a UE as a sensing responder as an example, the WiFi AP can send an orthogonal frequency division multiplexing (OFDM) signal to the UE, where the OFDM signal includes data symbols used for communication.

[0140] Correspondingly, the UE can receive an OFDM signal from the WiFi AP and use the data symbols in the OFDM signal to measure the distance and speed of a target object between the WiFi AP and the UE.

[0141] In the embodiments of the present application, the target object may be referred to as a target. For example, the target object between a WiFi AP and a UE may be referred to as a target between the WiFi AP and the UE.

[0142] The target between WiFi AP and UE is such as Figure 1 Target 1, Target 2, or Target 3 as shown in the figure.

[0143] The UE may send a sensing measurement result to the WiFi AP. The sensing measurement result may include distance 1 and speed 1 of target 1, distance 2 and speed 2 of target 2, and distance 3 and speed 3 of target 3.

[0144] In this way, the WiFi AP sends a signal containing data symbols to the UE, enabling the UE to measure the distance and speed of each of multiple targets in the environment between the WiFi AP and the UE based on the data symbols. The WiFi AP then obtains the distance and speed of each of the multiple targets measured by the UE. This enables the WiFi AP to perceive the environment between the WiFi AP and the UE. The UE measures the distance and speed of the target based on the data symbols, eliminating the need for the WiFi AP to send dedicated pilot signals. This reduces the communication resource usage of dedicated pilot signals, conserves communication resources, and improves the resource utilization of data symbols.

[0145] Figure 2 A flow chart of the perception method provided in an embodiment of the present application is shown.

[0146] like Figure 2As shown, the perception method provided in the embodiment of the present application may include S201-S202.

[0147] S201: A sensing initiator may send a first signal to a sensing responder. Correspondingly, the sensing responder may receive the first signal from the sensing initiator. The first signal may be used to indicate a data symbol of communication.

[0148] Exemplarily, the sensing initiator and the sensing responder can both use multiple antennas to simultaneously transmit and receive signals. The first signal can be a WiFi signal. The first signal can be an OFDM signal modulated based on the PPDU.

[0149] The sensing initiator can send an OFDM signal to the sensing responder. Correspondingly, the sensing responder can receive the OFDM signal from the sensing initiator. The OFDM signal includes data symbols used for communication.

[0150] The first signal may be referred to as a received signal, and the first signal includes a data symbol. In a possible implementation, the first signal may also be understood as a data signal received by the sensing response terminal.

[0151] S202: The sensing responding end may send the sensing measurement result to the sensing initiating end. Correspondingly, the sensing initiating end may receive the sensing measurement result from the sensing responding end.

[0152] The perception measurement result is obtained based on data symbol measurement. The data symbol may be an initial data symbol obtained by demodulating the first signal, or an updated data symbol obtained by updating a demodulation result after demodulating the first signal.

[0153] Exemplarily, the OFDM signal may also include pilot symbols. Upon receiving the OFDM signal from the initiator, the responder may perform channel estimation based on the pilot symbols to obtain initial channel information. The initial channel information may be used to perform data demodulation on the OFDM signal to obtain initial data symbols.

[0154] In the embodiments of the present application, the sensing initiator may be referred to as the initiator. The sensing responder may be referred to as the responder. The sensing initiator may be a sensing transmitter. The sensing responder may be a sensing receiver. Data symbols may also be referred to as data signals, communication signals, or communication symbols.

[0155] Exemplarily, based on the data symbols and the OFDM signal, the responding end may calculate and obtain a channel frequency response (CFR) matrix.

[0156] The responding end may use a spectrum estimation algorithm to calculate and process the channel frequency response matrix to obtain the distance and speed of each target in at least one target between the initiating end and the responding end.

[0157] The spectrum estimation algorithm may include at least one of a multiple signal classification (MUSIC) algorithm, an estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithm, and / or a root multiple signal classification (Root-MUSIC) algorithm. The MUSIC algorithm may include a one-dimensional (1D) MUSIC algorithm and / or a two-dimensional (2D) MUSIC algorithm.

[0158] The responding end may send the sensing measurement result to the sensing end. Correspondingly, the initiating end may receive the sensing measurement result from the responding end. The sensing measurement result may include the distance and speed of each target in at least one target between the initiating end and the responding end.

[0159] Optionally, the perception measurement result is carried in a MAC frame payload or a first management frame, wherein the first management frame can be used to transmit the perception measurement result.

[0160] like Figure 2 As shown, in the sensing method provided in an embodiment of the present application, a sensing initiator sends a PPDU containing data symbols to a sensing responder, so that the sensing responder measures the distance and speed of each of at least one target between the sensing initiator and the sensing responder based on the data symbols (e.g., initial data symbols or updated data symbols). The sensing initiator can obtain the distance and speed of each of the at least one target measured by the sensing responder, thereby achieving perception of the environment between the sensing initiator and the sensing responder. Compared to the method of performing environmental perception using dedicated pilot signals, the sensing method in an embodiment of the present application uses the sensing responder to measure the distance and speed of the target based on the data symbols, eliminating the need for the sensing initiator to send a dedicated pilot signal. This reduces the occupation of communication resources by dedicated pilot signals, thereby reducing the probability of communication efficiency being reduced due to dedicated pilot signals occupying communication resources. This method can also improve the utilization of data symbols, thereby improving communication resource utilization.

[0161] The following combination Figure 3-10 , the perception method provided in the embodiment of the present application is described.

[0162] The perception measurement process can include the following four stages.

[0163] Phase 1: Sensing capabilities exchange

[0164] The sensory capability exchange phase (or phase 1) can be called the association phase or the pre-association phase.

[0165] During the sensing capability exchange phase, the initiator and responder can inform each other of their supported sensing capabilities. Sensing capabilities, also known as sensing functions, can include supported bandwidth, antenna configuration, and / or reporting format.

[0166] Phase 2: Sensing measurement session

[0167] During the perception measurement session phase (or phase 2), the initiator and responder can negotiate a set of parameters related to this measurement. This set of parameters can include role assignment, report type, and / or measurement times. The initiator can assign a measurement session ID to this set of parameters.

[0168] Phase 3: Sensing Measurement Exchange

[0169] During the sensing measurement exchange phase (Phase 3), based on the parameters negotiated in the previous step (e.g., Phase 2), the responder collects channel information through one or more measurement rounds and reports the measurement results to the peer (e.g., the initiator). Channel information may include channel state information (CSI), delay, and / or amplitude. Measurement results may also include channel information.

[0170] In some possible implementations, based on air interface preamble packets sent by the initiator via one or more measurement rounds, the responder may collect channel information and target perception measurement information. Each measurement round is marked with a measurement session identifier. The air interface preamble packet may include, for example, an NDPA, an NDP, and a trigger frame. The NDP may include a dedicated pilot signal. The measurement results may include channel information and target perception measurement information. The target perception measurement information may include the target's distance. The target perception measurement information is measured based on the dedicated pilot signal.

[0171] Phase 4: Sensing measurement session termination

[0172] After the measurement is completed, the initiator may send a session termination message to the responder, or the responder may send a session termination message to the initiator.

[0173] The initiator and responder can release the resources reserved for the session (such as the perception measurement session). The measurement session identifier allocated during the perception measurement session phase can be reused in the next perception measurement session.

[0174] Figure 3 Another flow chart of the perception method provided in an embodiment of the present application is shown.

[0175] like Figure 3 As shown, the perception method provided in the embodiment of the present application may further include: S301, S201-S202.

[0176] S301: The perception responding end may send first capability information to the perception initiating end. Correspondingly, the perception initiating end may receive the first capability information from the perception responding end.

[0177] The first capability information can be understood as capability information of the perception responder. The first capability information can be used to indicate support for data symbol-based perception measurement functions. Support for data symbol-based perception measurement functions can also be understood as support for the perception method provided in the embodiments of this application. The first capability information can also include functions such as supported bandwidth, antenna configuration, and / or reporting format.

[0178] For example, before S201, for example, in the sensing capability exchange phase, the sensing responding end may send the first capability information to the sensing initiating end. Correspondingly, the sensing initiating end may receive the first capability information from the sensing responding end.

[0179] The sensing initiator may send the second capability information to the sensing responder, and correspondingly, the sensing responder may receive the second capability information from the sensing initiator.

[0180] The second capability information can be understood as capability information of the sensing initiator. The second capability information can be used to indicate support for data symbol-based sensing measurement functions. The second capability information can also include supported bandwidth, antenna configuration, and / or reporting format.

[0181] S201: A sensing initiator may send a first signal to a sensing responder. Correspondingly, the sensing responder may receive the first signal from the sensing initiator, wherein the first signal includes data symbols used for communication.

[0182] S202: The sensing responding end may send a sensing measurement result to the sensing initiating end. Correspondingly, the sensing initiating end may receive the sensing measurement result from the sensing responding end. The sensing measurement result is obtained based on data symbol measurement.

[0183] For example, during the sensing measurement exchange phase, the sensing initiator may send a first signal to the sensing responder, and correspondingly, the sensing responder may receive the first signal from the sensing initiator.

[0184] The sensing responding end may measure and obtain the distance and speed of each target in at least one target between the initiating end and the responding end based on the data symbols in the first signal.

[0185] The sensing responding end may send a sensing measurement result including a distance and a speed of each target in at least one target to the sensing initiating end.

[0186] The specific implementation principles and technical effects of S201-S202 can be found in Figure 2 The specific implementation principles and technical effects of S201-S202.

[0187] In this way, during the perception measurement exchange phase, the perception initiator can send a first signal (e.g., a PPDU) to the perception responder. The perception initiator does not need to send an NDP to the perception responder. The perception responder can collect channel information based on the pilot portion (e.g., pilot symbols) in the PPDU to demodulate the data of the first signal. The target's range and speed can be measured based on the data symbols, achieving environmental perception without relying on the perception initiator to send an NDP.

[0188] As shown in S301, S201-S202, in the perception method provided by the embodiment of the present application, the perception initiator and the perception responder can inform each other of their respective capability information during the perception capability exchange phase. The capability information of the perception initiator and the capability information of the perception responder both include support for data symbol-based perception measurement functions or support for the perception method provided by the embodiment of the present application, so that the perception initiator and the perception responder can use the perception method provided by the embodiment of the present application to perceive the environment during the perception measurement exchange phase. For example, the perception responder can collect channel information based on the pilot part (such as pilot symbols) in the PPDU to achieve data demodulation of the first signal. The distance and speed of the target are obtained based on the data symbol measurement. Environmental perception can be achieved without relying on the perception initiator to send NDP.

[0189] Alternatively, as Figure 3 As shown, the perception method provided in the embodiment of the present application may also include: S301-S302, S201-S202.

[0190] Among them, the specific implementation principles and technical effects of each step in S301, S201-S202, and S303 can be found in the specific implementation principles and technical effects of the corresponding steps in the above embodiments, which will not be repeated here.

[0191] S302: The sensing responding end may send first indication information to the sensing initiating end. Correspondingly, the sensing initiating end may receive the first indication information from the sensing responding end.

[0192] The first indication information may be used to indicate the use of data symbols for perception measurement, and may also be understood as the use of a perception measurement mode in the perception method provided in an embodiment of the present application.

[0193] For example, before S201, for example, in the perception measurement session phase, the perception responding end may send first indication information to the perception initiating end. Correspondingly, the perception initiating end may receive the first indication information from the perception responding end.

[0194] Exemplarily, during the perception measurement session phase, the parameter set related to this measurement negotiated by the initiating end and the responding end may include the perception measurement mode of the perception method provided in the embodiment of the present application.

[0195] In this way, the perception response end uses data symbols to perform perception measurement in the perception measurement exchange phase to measure the distance and speed of the target and realize environmental perception.

[0196] As shown in S301-S302 and S201-S202, in the perception method provided by the embodiment of the present application, the perception initiator and the perception responder inform each other of their respective capability information during the perception capability exchange phase, including support for the data symbol-based perception measurement function or support for the perception method provided by the embodiment of the present application. The parameter set negotiated by the initiator and the responder during the perception measurement session phase may include the perception measurement mode of the perception method provided by the embodiment of the present application. This allows the perception responder to use the perception method provided by the embodiment of the present application to perform perception measurement during the perception measurement exchange phase to measure the distance and speed of the target, achieve environmental perception without relying on dedicated pilot signals, and improve resource utilization such as data symbols.

[0197] Alternatively, as Figure 3 As shown, the perception method provided in the embodiment of the present application may also include: S301-S302, S201-S202, S303.

[0198] Among them, the specific implementation principles and technical effects of each step in S301-S302 and S201-S202 can be found in the specific implementation principles and technical effects of the corresponding steps in the above embodiments, which will not be repeated here.

[0199] S303: The perception initiator and the perception responder may release resources.

[0200] Exemplarily, the perception initiator may send a session termination message to the perception responder, and correspondingly, the perception responder may receive the session termination message from the perception initiator.

[0201] Alternatively, the perception responder may send a session termination message to the perception initiator, and correspondingly, the perception initiator may receive a session termination message from the perception responder.

[0202] The perception initiator and the perception responder may release resources reserved for the perception measurement session.

[0203] As shown in S301-S302, S201-S202, and S303, the perception method provided by the embodiment of the present application, the perception initiator and the perception responder, in the perception capability exchange phase, inform each other of their respective capability information, including support for data symbol-based perception measurement capabilities or support for the perception method provided by the embodiment of the present application, and the parameter set negotiated by the initiator and the responder in the perception measurement session phase may include a mode for using data symbols for perception measurement. This allows the perception responder to perform perception measurements based on data symbols during the perception measurement exchange phase to measure the distance and speed of the target, achieve environmental perception, and achieve environmental perception without relying on dedicated pilot signals. When the perception measurement is completed, the perception initiator and the perception responder can release resources so that the released resources can be used for the next communication and / or environmental perception, thereby improving resource utilization.

[0204] The following combination Figures 4-10 , explain S201-S202.

[0205] In one possible implementation of the embodiments of the present application, a neural network model may be deployed in the sensing response end. The neural network model may include at least one neural network layer. Each neural network layer may include a feedback loop. The sensing response end may input the first signal received from the sensing initiator end into the neural network model. The neural network model may output the distance and speed of each of the at least one target.

[0206] Figure 4 A schematic diagram of a feedback loop structure provided in an embodiment of the present application is shown.

[0207] like Figure 4 As shown, the feedback loop may include: a signal demodulation module, a sensing algorithm module, a channel reconstruction module and a noise perception module.

[0208] The signal demodulation module receives signals. Its output is fed into the perception algorithm module. The perception algorithm module also feeds into the channel reconstruction module. The channel reconstruction module's output is fed back to the signal demodulation module. The noise perception module generates hyperparameters for each layer of the neural network based on the estimated noise power or signal-to-noise ratio (SNR). These hyperparameters influence the operations of the signal demodulation module, the perception algorithm module, and the channel reconstruction module.

[0209] Still taking the first signal as an OFDM signal as an example, the modules of the feedback loop in the neural network model are introduced in detail below.

[0210] Figure 5 A flow chart of a signal demodulation module provided in an embodiment of the present application is shown.

[0211] In the perception method provided in the embodiment of the present application, the signal demodulation module may execute the following S501-S506.

[0212] S501, the responding end can receive the first signal from the initiating end , the channel matrix estimated in the previous iteration , the data signal estimated in the previous iteration and / or the residual vector from the previous iteration , input signal demodulation module. Correspondingly, the signal demodulation module can receive 、 、 and / or .

[0213] Among them, the data signal estimated in the previous iteration can be understood as the data signal estimated in the previous iteration. The previous iteration, for example, the k-1th iteration. The channel matrix estimated in the previous iteration can be understood as the channel matrix output by the k-1th layer neural network. The data signal estimated in the previous iteration can be understood as the data signal output by the k-1th layer neural network. The residual vector of the previous iteration can be understood as the residual vector output by the k-1th layer neural network. k is a natural number. k can be understood as the sequence number of the neural network layer in the neural network model. For example, k can be any value of 1, 2, ..., K or K+1. K is a positive integer. For example, K can be 4, 5, 6, or 7. K can also be other values ​​besides 4, 5, 6, and 7.

[0214] It can be an OFDM signal. The symbol after transformation into the symbol domain can be expressed as . Satisfies the formula:

[0215]

[0216] in, It can be a data symbol sent by the initiator. The channel matrix Middle Path The corresponding amplitude component. The channel matrix Middle Path The corresponding phase component. It can be noise. Can be the number of paths. p It can be a path number. For example, p It can be 1, 2, ..., or Any value in .

[0217] For example, Can satisfy the formula:

[0218]

[0219] in, is the channel matrix updated or output by the k-th layer neural network, The channel matrix Middle Path The corresponding amplitude component. The channel matrix Middle Path The corresponding phase component. The data symbol that can be updated or output by the k-th layer neural network. The noise vector or residual vector that can be updated or output by the k-th layer neural network.

[0220] The number of paths can be understood as the number of transmission paths for OFDM signals from the initiator to the responder. Figure 1 In the perception system shown, the number of paths It can be 4. Figure 1 The transmission path of the OFDM signal in the sensing system shown may include Figure 1 Path 1, Path 2, Path 3, and Path 4 are shown.

[0221] For example, when k=1, the responding end may receive the first signal from the initiating end. Input signal demodulation module. The signal demodulation module can perform channel estimation based on the pilot symbol in the first signal to obtain initial channel information. The initial channel information may include the estimated initial channel matrix . So that the signal demodulation module adopts the initial channel matrix Data demodulation is performed on the first signal to obtain initial data symbols, which may also be referred to as initial data signals.

[0222] For example, the responder can receive the first signal from the initiator. In the case of , the responding end can start the neural network model and input the first signal into the signal demodulation module. When the neural network model is started, the signal demodulation module is also started. After starting, the signal demodulation module can perform channel estimation based on the pilot symbols in the first signal to obtain the estimated initial channel matrix . After obtaining the initial channel matrix In this case, the signal demodulation module may execute S502.

[0223] In this way, the signal demodulation module can start the signal demodulation module based on the response end and determine that the signal demodulation module needs to perform its function in the first layer of the neural network.

[0224] In the case of k≠1, or in the output of the channel reconstruction module In the case of , the output of the channel reconstruction module , and the signal demodulation module output and , input signal demodulation module. So that the signal demodulation module can update the data signal and residual vector output by the k-1 layer neural network.

[0225] in, is the channel matrix output by the k-1th layer neural network, is the data signal output by the k-1th layer neural network, is the residual vector output by the k-1th layer neural network.

[0226] k≠1, for example, k>1.

[0227] For example, when the signal demodulation module receives 、 、 and In this case, the signal demodulation module can skip S502 and S503 and execute S504.

[0228] Optionally, the responder can 、 、 and / or Pack the data into data packets and input the data packets into the signal demodulation module. Correspondingly, the signal demodulation module can receive the data packets.

[0229] The signal demodulation module checks that the received data packet only contains In this case, the signal demodulation module may execute S502.

[0230] The signal demodulation module detects that the received data packet contains 、 、 and In this case, the signal demodulation module may execute S504.

[0231] S502: The signal demodulation module may be initialized to obtain initialization parameters.

[0232] Among them, the initialization parameters can include and . , . The symbol of the first signal after transformation is obtained by transforming the first signal into the symbol domain. For the specific implementation of transforming the first signal into the symbol domain, please refer to the subsequent S503.

[0233] Exemplarily, when k=1, the signal demodulation module may be initialized to obtain initialization parameters.

[0234] In the case where k>1, the signal demodulation module may not be initialized. That is, in the case where k>1 and the signal demodulation module receives the information input in S501, the signal demodulation module may skip S502 and S503 and execute S504.

[0235] Exemplarily, when the signal demodulation module completes initialization, the signal demodulation module can maintain the first flag bit of the signal demodulation module as a first value. The first value can indicate that the signal demodulation module has completed initialization. When the signal demodulation module is powered off or receives a session termination message transmitted by the responding end, the signal demodulation module can maintain the first flag bit of the signal demodulation module as a second value or as empty. The value of the first flag bit is the second value or the first flag bit is empty, which may indicate that the signal demodulation module has not completed initialization. Maintaining the first flag bit of the signal demodulation module is empty, for example, clearing the value of the first flag bit.

[0236] When the signal demodulation module receives the information or data packet input in S501 , the signal demodulation module may read the value of the first flag bit.

[0237] When the value of the first flag is read to be the second value or the first flag is empty, the signal demodulation module may execute S502.

[0238] When the value of the first flag bit read is the first value, the signal demodulation module may skip S502 and S503 and execute S504.

[0239] S503, the signal demodulation module can Transform to the symbolic domain.

[0240] For example, the signal demodulation module can demodulate the first signal Perform nonlinear processing to transform the first signal into the symbol domain and obtain the symbol after the first signal is transformed .

[0241] For example, the signal demodulation module can perform preprocessing, fast Fourier transform (FFT), channel equalization and signal threshold determination on the first signal to obtain the transformed symbol of the first signal. .

[0242] For example, preprocessing can include symbol synchronization, carrier synchronization, and cyclic prefix removal. Symbol synchronization and carrier synchronization can achieve correct symbol delimitation and frequency correction. Removing the cyclic prefix (CP) is an example of removing the cyclic prefix (CP) used to combat multipath effects in OFDM signals.

[0243] The signal demodulation module can use a bandpass filter to demodulate the received signal of the RF front end (such as the first signal ) is filtered to remove noise and obtain the first denoised signal.

[0244] The signal demodulation module can pre-process the first signal after de-noising to obtain a signal after removing the cyclic prefix.

[0245] The signal demodulation module can perform fast Fourier transform on the signal after removing the cyclic prefix to obtain multiple frequency domain subcarriers.

[0246] The signal demodulation module can use the initial channel matrix estimated by the signal demodulation module based on the pilot symbol in the first signal in S501 , each subcarrier in the multiple frequency domain subcarriers is equalized to compensate for the amplitude and phase distortion of the channel, and obtain multiple equalized frequency domain subcarriers.

[0247] The signal demodulation module can set a decision threshold or threshold value for each subcarrier according to the modulation mode carried in the first signal.

[0248] The signal demodulation module can perform threshold judgment on each equalized frequency domain subcarrier signal to obtain the symbol after the first signal transformation. . Among them, the symbol after the first signal transformation Including the symbol sequence of each subcarrier. The symbol sequence of each subcarrier is determined by the signal demodulation module comparing each equalized frequency domain subcarrier signal with the corresponding threshold value of the subcarrier.

[0249] In the case that k>1, the signal demodulation module may not execute S503. For example, the signal demodulation module may skip S503 and execute S504.

[0250] Optionally, when the signal demodulation module completes S503, the signal demodulation module may maintain the second flag bit of the signal demodulation module as a third value. The third value may indicate that the signal demodulation module has completed data demodulation of the first signal. When the signal demodulation module is powered off or receives a session termination message transmitted by the responding end, the signal demodulation module may maintain the second flag bit of the signal demodulation module as a fourth value or as empty. The fourth value of the second flag bit or the empty second flag bit may indicate that the signal demodulation module has not completed data demodulation of the first signal.

[0251] When the signal demodulation module completes S502 , the signal demodulation module may read the value of the second flag bit.

[0252] When the value of the second flag is read as the fourth value or the second flag is empty, the signal demodulation module may execute S503.

[0253] Optionally, when the signal demodulation module completes S501, for example, when the signal demodulation module receives 、 、 and / or In this case, the signal demodulation module can read the value of the second flag bit.

[0254] When the value of the second flag bit read is the third value, the signal demodulation module may skip S502 and S503 and execute S504.

[0255] S504: The signal demodulation module can update the data symbol to obtain the updated data symbol. .

[0256] Satisfies the formula:

[0257]

[0258] in, The data symbol that can be updated or output by the k-th layer neural network. The data symbol that can be updated or output by the k-1th layer neural network. The signal update coefficients of the k-th layer neural network can be obtained. The channel matrix Middle Path The corresponding amplitude component. The channel matrix Middle Path The corresponding phase component. T represents transpose. For example, for The transpose of . for The signal update coefficient can be called the signal update parameter. The residual or noise that can be used to update the k-1th layer of the neural network.

[0259] In the case of k=1, = ,Right now is the initial channel matrix , Satisfies the formula:

[0260]

[0261] As in S504 As shown in the formula satisfied, by introducing the gradient descent step size The signal demodulation module uses an iterative algorithm based on gradient descent to solve , in order to achieve the optimization goal.

[0262] Among them, For example, the optimization goal can be: . express The square of the Frobenius norm (or F-norm) of . Can express smallest .

[0263] In the case of k>1, It can be expressed as The remaining data symbols outside. Due to the channel matrix It is the channel information reconstructed by the neural network model based on the target parameters, and the target parameters are based on the updated signal (such as ) is determined. It is possible to use the reconstructed channel information and some demodulated data symbols (such as ), and more accurately demodulate the remaining data symbols. More accurate data symbols can be obtained through signal updating.

[0264] In this way, the signal demodulation module uses an iterative algorithm based on gradient descent to solve , it is possible to more accurately demodulate the first signal except for the demodulated data symbols (such as ) to obtain more accurate data symbols. The demodulated data symbols, such as or .

[0265] S505: The signal demodulation module can perform residual update to obtain the updated residual. .

[0266] Satisfies the formula:

[0267]

[0268] in, The residual can be updated for the k-th layer of the neural network. The residual can be equivalent to noise, or the residual can represent noise. Or the residual can be understood as noise. The residual or noise that can be used to update the k-1th layer of the neural network. The residual update coefficient of the k-th layer neural network can be called the noise update parameter.

[0269] As shown in formula S505, by introducing the momentum term Updating the residual can accelerate convergence and suppress oscillation, improve noise estimation accuracy, and reduce the probability of large noise estimation errors leading to reduced accuracy of acquired data symbols. This can further improve data demodulation accuracy.

[0270] The signal update coefficient of the k-th layer neural network The signal update coefficient and the residual update coefficient of the k-th layer neural network can be obtained from the noise perception module. The specific implementation principle of the noise perception module to generate the signal update coefficient and the residual update coefficient can be found in Figure 8 The embodiments shown are for ease of understanding. Figure 8 The illustrated embodiment is described below.

[0271] S506, the signal demodulation module can output data symbols and the residual vector .

[0272] For example, when k=1, the signal demodulation module may output the signal demodulation module for the first signal Data symbols obtained by data demodulation The signal demodulation module can also output the residual vector obtained by the signal demodulation module through residual update .in, .

[0273] In the case of k>1, the signal demodulation module can output the updated data symbol and the updated residual vector .in, It can be obtained by executing S504 by the signal demodulation module. It can be obtained by executing S505 by the signal demodulation module.

[0274] like Figure 5 As shown, in the perception method provided in the embodiment of the present application, when k=1, for example, in the first layer of the neural network, the signal demodulation module can execute S501-S506, and the signal demodulation module can perform channel estimation based on the pilot symbol in the first signal to obtain the estimated initial channel matrix The signal demodulation module can also be initialized to obtain initialization parameters, such as and , and adopt , initialization parameters and signal update coefficients of the first layer neural network are used to demodulate the first signal to output data symbols . In order to facilitate the perception algorithm module based on the data symbols obtained by demodulation The signal demodulation module can also be based on 、 Perform residual update with the residual update coefficient of the first layer neural network to obtain the updated residual , so that the signal demodulation module in the second layer of the neural network adopts the updated residual The remaining data symbols are demodulated to update the data symbols, thereby improving the perception measurement accuracy.

[0275] In the case of k>1, or in the k-th layer neural network with k>1, the signal demodulation module can perform S501, S504, S505 and S506. By adopting the signal update coefficient of the k-th layer neural network Data symbols output by the k-1 layer neural network Perform an iterative algorithm based on gradient descent to solve , in addition to more accurately demodulating the first signal The remaining data symbols other than the original data are used to update the data symbols. This allows the perception algorithm module to perform perception measurement based on the updated data symbols, thereby improving the accuracy of the perception measurement. The signal demodulation module can also use the residual update coefficient of the k-th layer neural network to perform residual update on the residual of the k-1th layer neural network to obtain the updated residual. , so that the next neural network layer (or the next neural network layer) of the k-th neural network adopts the updated residual The remaining data symbols are demodulated to update the data symbols, thereby improving the accuracy of the perception measurement.

[0276] Figure 6 A flow chart of the perception algorithm module provided in an embodiment of the present application is shown.

[0277] In the perception method provided in the embodiment of the present application, the perception algorithm module can execute the following S601-S608.

[0278] S601, the responding end can convert the first signal into a symbol , the data signal output by the signal demodulation module , the target parameters estimated by the previous neural network layer are input into the perception algorithm module.

[0279] Among them, the target parameters estimated by the previous neural network layer can include and . The distance to the i-th target is estimated or calculated for the k-1-th layer of the neural network. The speed of the i-th target estimated or calculated by the k-1-th neural network layer. i can be understood as the target's sequence number. i is an integer. For example, if I is the number of targets between the initiator and the responder, 1 ≤ i ≤ I, and I is an integer.

[0280] S602: The perception algorithm module may calculate a channel frequency response (CFR) of the channel.

[0281] For example, the sensing algorithm module can be based on the symbol after the first signal is transformed and data symbols output by the signal demodulation module Perform channel CFR calculation to obtain the channel frequency response matrix . Channel frequency response matrix It can be called the channel matrix .

[0282] Channel Matrix Satisfies the formula:

[0283]

[0284] in, is the channel matrix estimated by the k-th layer neural network or the channel matrix of the k-th iteration, for The elements in is the symbol after the first signal is transformed The matrix representation of is a matrix The element in row n and column l in Data symbol The matrix representation of is a matrix The element in the nth row and lth column of the matrix, 1≤n≤N, 1≤l≤L. n is the row number and is an integer. l is the column number and is an integer. N is the number of rows in the matrix. L is the number of columns in the matrix. k is a natural number, N is an integer, and L is an integer. For example, assuming that both N and L are greater than 3, n can be 1, 2, ..., N - 1, or any value in N. l can be 1, 2, ..., L - 1, or any value in L.

[0285] S603, the perception algorithm module can use the spectrum estimation algorithm to calculate the channel matrix Calculation is performed to obtain the time delay and Doppler shift corresponding to the target parameters.

[0286] In this way, it is convenient for the perception algorithm module to determine target parameters based on time delay and Doppler frequency shift, wherein the target parameters may include the target distance and target speed of the target.

[0287] In the embodiment of the present application, the target distance may be referred to as distance, and the target speed may be referred to as speed.

[0288] For example, the perception algorithm module can use a 2D MUSIC algorithm to calculate the channel matrix Calculation is performed to obtain the time delay and Doppler shift corresponding to the target parameters.

[0289] For example, the perception algorithm module uses a 2D MUSIC algorithm to calculate the channel matrix The calculation process may include S6031-S6037.

[0290] S6031, the perception algorithm module can use the vec() operator to convert the channel matrix Vectorization is used to realize the channel matrix Convert to a one-dimensional long vector Among them, the vec() operator can be called a vectorization operator, which is used to convert a matrix into a column vector. For example, it can be used to stack the columns of a matrix into a column vector in column order.

[0291] Satisfies the formula:

[0292]

[0293] Where T represents transpose, for example, Represents the transpose of *. for The lth column element of . For example, for The first element of the column. for The second element of the column. for The Lth column element of . Taking L as 5 as an example, satisfy:

[0294]

[0295] S6032, the perception algorithm module can construct a two-dimensional joint guidance vector .

[0296] 2D Joint Steering Vector Satisfies the formula:

[0297]

[0298] in, represents the Kronecker product, which can also represent the tensor product or outer product. A one-dimensional distance-oriented vector constructed for the k-th layer of the neural network. The one-dimensional velocity steering vector constructed for the k-th layer of the neural network.

[0299] Satisfies the formula:

[0300]

[0301] Satisfies the formula:

[0302]

[0303] Where, e is a natural constant or transcendental number, and j is the imaginary unit. It can be the carrier spacing of OFDM. for The independent variable. It can be understood as The corresponding delay. for The independent variable. It can be understood as The corresponding Doppler shift is, It is the period of an Orthogonal Frequency Division Multiplexing (OFDM) symbol or the length of time occupied by an OFDM symbol. express The transpose of .

[0304] S6033, the perception algorithm module can Calculate the covariance matrix. The calculated covariance matrix R satisfies the formula:

[0305]

[0306] Where H represents the conjugate transpose, for example, for The conjugate transpose of . is an operator, and Indicates the calculation of covariance.

[0307] Optionally, S6032 and S6033 may be executed concurrently. Optionally, the perception algorithm module may execute S6033-S6035 before S6032, or the order of S6032-S6035 may be replaced with: S6033-S6035, S6032.

[0308] S6034, the perception algorithm module can perform eigenvalue decomposition on the covariance matrix R to obtain the eigenvector matrix E and the diagonal matrix with eigenvalues ​​on the diagonal The eigenvector matrix E and the diagonal matrix Satisfies the formula:

[0309]

[0310] S6035. The perception algorithm module can divide the eigenvectors of the eigenvector matrix E into two orthogonal subspaces according to the size of the eigenvalues. The two orthogonal subspaces include the signal subspace and noise subspace .

[0311] Among them, the signal subspace Is caused by The signal subspace is composed of the eigenvectors corresponding to the larger eigenvalues. The eigenvectors in represent the main components of the signal. The largest eigenvalues ​​can be understood as the first M eigenvalues ​​in the eigenvalue sequence obtained by arranging the eigenvalues ​​of the eigenvector matrix E in descending order.

[0312] Noise subspace is made up of the remaining The eigenvectors corresponding to the smaller eigenvalues ​​are spanned. These eigenvalues ​​should be 0 or smaller. The rest The smaller eigenvalues ​​can be understood as the eigenvalues ​​other than the first M eigenvalues ​​in the eigenvalue sequence, where M is an integer.

[0313] S6036, the perception algorithm module can construct the 2D MUSIC power spectrum of the k-th layer neural network. Satisfies the formula:

[0314]

[0315] in, for The conjugate transpose of . for The conjugate transpose of .

[0316] S6037. The perception algorithm module can perform peak search on the 2D MUSIC power spectrum to obtain the position coordinates of each peak among I peaks.

[0317] Among them, the coordinate position of the i-th peak among the I peaks can be ( , ). The delay corresponding to the i-th peak estimated by the k-th layer neural network can be expressed as: The Doppler frequency shift corresponding to the i-th peak estimated by the k-th layer neural network can be expressed as:

[0318] For example, the perception algorithm module can Corresponding The two-dimensional grid search peak is composed of the k-th layer neural network multi-target detection threshold , search to obtain I peaks and the position coordinates of each peak in the I peaks. Among them, each peak in the I peaks is greater than or equal to the multi-target detection threshold of the k-th layer neural network .

[0319] For example, the perception algorithm module can Corresponding The perception algorithm module can use a multi-target detection threshold to perform a threshold judgment on each of the multiple peaks to determine I peak from the multiple peaks.

[0320] Among them, threshold judgment, for example, judging whether the peak value is greater than or equal to the multi-target detection threshold of the k-th layer neural network The multi-target detection threshold is such as the multi-target detection threshold of the k-th layer neural network. .

[0321] In the case where there are multiple targets between the initiator and the responder, the side lobes generated by some of the multiple targets will interfere with the peaks generated by another part of the multiple targets. In the embodiment of the present application, a multi-target detection threshold is used to perform threshold judgment on the searched peaks, so that the peak generated by the target can be determined from the multiple searched peaks, that is, I peaks corresponding to I targets between the initiator and the responder are determined from the multiple searched peaks. In the case where the side lobes interfere with the peaks generated by the targets, the true target peaks and the side lobes (or noise) can be distinguished. Among them, the true target peak can be greater than or equal to the multi-target detection threshold. The side lobes (or noise) can be less than the multi-target detection threshold. The true target peak is, for example, any one of the I peaks.

[0322] Optionally, the perception algorithm module may adopt a multi-target detection threshold to perform a threshold judgment on each peak value among the multiple peak values, which may include S60371-S60373.

[0323] S60371. The perception algorithm module can perform a threshold judgment on a peak value among multiple peak values ​​that has not been subjected to a threshold judgment.

[0324] Exemplarily, the perception algorithm module may determine whether there is a peak among the multiple peaks for which a threshold determination has not been performed.

[0325] When there is a peak among the multiple peaks that has not been threshold-judged, the perception algorithm module may perform a threshold judgment on the peak that has not been threshold-judged. For example, the perception algorithm module may perform a threshold judgment on the peak that has not been marked among the multiple peaks.

[0326] The unmarked peak may be a peak that is not marked with the first or second marker. The first marker may indicate that the peak satisfies the first condition. The second marker may indicate that the peak does not satisfy the first condition. The first condition may be, for example, that the peak is greater than or equal to the multi-target detection threshold. The peak that does not satisfy the first condition may be, for example, that the peak is less than the multi-target detection threshold.

[0327] If there is no peak for which threshold determination has not been performed among the multiple peaks, the perception algorithm module may execute S605.

[0328] There is no peak in the plurality of peaks that has not been subjected to threshold determination, for example, each of the plurality of peaks is marked with an identifier, and the identifier of the peak mark may be a first identifier or a second identifier.

[0329] S60372: When the threshold determination is completed for the peak, the perception algorithm module may mark the peak as a first identifier or a second identifier based on the threshold determination result of the peak. The threshold determination result may include: the peak meets the first condition, or the peak does not meet the first condition.

[0330] S60373: If the peak is marked with the first identifier, the perception algorithm module may execute S604 for the peak. After executing S604 for the spectrum peak pair, the perception algorithm module may continue to execute S60371.

[0331] In the case where the peak is marked with a second identifier, the perception algorithm module may continue to execute S60371.

[0332] In this way, multi-target cyclic detection can be achieved.

[0333] In another possible implementation, the perception algorithm module may perform a threshold determination on each peak in the multiple peak pairs. Each time a threshold determination is completed for a peak, the perception algorithm module may mark the peak as the first identifier or the second identifier based on the threshold determination result for the peak. After completing the threshold determination for all of the multiple peaks, the perception algorithm module may execute S604 for each peak in the multiple peaks marked with the first identifier.

[0334] S604, the perception algorithm module can be used to Calculate and get the distance of the i-th target calculated by the k-th layer neural network The sensing algorithm module can detect Doppler frequency shift Calculate and get the speed of the i-th target calculated by the k-th layer neural network .

[0335] in, Satisfies the formula:

[0336]

[0337] Satisfies the formula:

[0338]

[0339] Where c is the propagation speed of electromagnetic waves or the speed of light. is the preset carrier frequency. It can represent the length of the path corresponding to the i-th target calculated by the k-th layer neural network. The path corresponding to the i-th target can be understood as the path where the i-th target is located. Figure 1 , with the i-th target as Figure 1 Taking target 1 as an example, the path corresponding to the i-th target may be path 4.

[0340] In this way, the perception algorithm module can determine the distance based on the time delay and the speed based on the Doppler frequency shift, thereby realizing the extraction of target parameters of each target in at least one target (such as I targets).

[0341] S605. The perception algorithm module may perform fusion update on the target parameters to obtain updated target parameters.

[0342] Exemplarily, when target parameters of each target in at least one target are obtained, the perception algorithm module may perform a fusion update on the target parameters of each target to obtain updated target parameters of each target.

[0343] The target parameters of the i-th target are still calculated using the k-th layer neural network For example, the perception algorithm module can calculate the target parameters of the i-th target Perform fusion update to obtain the updated target parameters of the i-th target .

[0344] Among them, the distance of the i-th target after the k-th layer neural network is updated and speed Satisfies the following formula:

[0345]

[0346] Or, the distance to the i-th target after the k-th layer neural network is updated and speed Satisfies the following formula:

[0347]

[0348]

[0349] in, , In the case of k=1, , . is the parameter fusion coefficient of the k-th layer neural network. The parameter fusion coefficient can be called the perception result update coefficient or perception result update parameter. It can be called the initial target parameter.

[0350] Since the parameter fusion coefficient is generated based on the noise power of the first signal, using the parameter fusion coefficient to fuse and update the target parameters extracted by the neural network can further reduce the impact of noise such as side lobes on the target parameters and improve the accuracy of the updated target parameters.

[0351] The multi-target detection threshold of the k-th layer neural network and the parameter fusion coefficient of the k-th layer neural network can both be obtained from the noise perception module. The specific implementation principle of the noise perception module to generate the multi-target detection threshold and parameter fusion coefficient can be found in the subsequent Figure 8 The specific implementation principle of the illustrated embodiment.

[0352] S606. The perception algorithm module may output updated target parameters of each target in at least one target.

[0353] Exemplarily, the perception algorithm module may input the output target parameters into the channel reconstruction module. Correspondingly, the channel reconstruction module may receive the target parameters from the perception algorithm module. For example, the perception algorithm module may transmit updated target parameters for each of the at least one target to the channel reconstruction module. The updated target parameters include an updated target distance and an updated target speed.

[0354] like Figure 6 As shown, the perception algorithm module can calculate a channel frequency response matrix based on the data symbols output by the signal demodulation module and the symbols after the first signal transformation. A 2D MUSIC algorithm and threshold determination are performed on the channel frequency response matrix to determine the peak corresponding to each of at least one target between the initiator and the responder from the multiple peaks found. The peaks then determine the time delay and Doppler shift corresponding to each target. Target parameters for each target are then calculated based on the time delay and Doppler shift corresponding to each target. Perception measurements are performed using the data symbols output (or demodulated) by the signal demodulation module as a perception reference, increasing the signal resources available for perception and improving the estimation accuracy and resolution of target parameters.

[0355] Figure 7 A flow chart of a channel reconstruction module provided in an embodiment of the present application is shown.

[0356] In the perception method provided in the embodiment of the present application, the channel reconstruction module may execute the following S701-S706.

[0357] S701. Obtain updated target parameters of each target in at least one target input by the perception algorithm module to the channel reconstruction module.

[0358] S702: The channel reconstruction module may calculate, based on the first signal, an angle of departure (AoD) and an angle of arrival (AoA) of each of the at least one target using an angle estimation algorithm.

[0359] The transmission angle is also called the departure angle, and the reception angle is also called the arrival angle.

[0360] The Angle of Departure (AoD) can be understood as the angle at which a signal (such as the first signal) leaves a transmitting antenna array. The transmitting antenna array is, for example, the antenna array at the initiator.

[0361] The angle of acceptance (AoA) can be understood as the angle at which a signal (such as the first signal) reaches a receiving antenna array. The receiving antenna array is, for example, the antenna array at the responding end.

[0362] Angle estimation algorithms, such as direction of arrival estimation (DOA estimation), AOA estimation, and / or angle estimation.

[0363] S703: The channel reconstruction module can calculate the time delay and Doppler frequency shift.

[0364] Exemplarily, the channel reconstruction module may calculate the time delay and Doppler frequency shift corresponding to each target based on the updated target parameters of each target in the at least one target.

[0365] For example, the target parameter of the i-th target output by the perception algorithm module is and For example, the channel reconstruction module can be based on the target parameters and , calculate the delay of the i-th target and Doppler shift .

[0366] in, Satisfies the formula:

[0367]

[0368] Satisfies the formula:

[0369]

[0370] It can represent the delay determined by the distance of the k-th layer neural network based on the update of the i-th target. It can represent the Doppler shift determined by the k-th layer neural network based on the updated velocity of the i-th target.

[0371] S704: The channel reconstruction module may reconstruct the ISAC channel matrix.

[0372] The MIMO-OFDM channel can be parameterized by parameters such as AoA, AoD, Doppler shift, and delay. Therefore, the channel reconstruction module can use the Doppler shift and delay of each target in at least one target, as well as the AoA and AoD of each target in at least one target, to reconstruct the channel model and obtain the reconstructed channel. .

[0373] The ISAC channel matrix reconstructed for the k-th layer of the neural network, or the channel matrix reconstructed (or reconstructed) for the k-th layer of the neural network.

[0374] S705: The channel reconstruction module may use hyperparameters to update the channel matrix.

[0375] For example, the channel reconstruction module can use the channel update coefficient of the k-th layer neural network , for the reconstructed channel matrix Update the channel matrix to obtain the updated channel matrix of the k-th layer neural network The channel update coefficient may be referred to as a channel update parameter.

[0376] Satisfies the formula:

[0377]

[0378] in, is the updated channel matrix of the k-1th layer neural network. .

[0379] In the case of k=1, .

[0380] The channel update coefficient can be used to control the weight balance between the channel matrix updated by the previous layer of neural network and the channel reconstruction of the current layer of neural network, making the updated channel matrix more accurate. In addition, since the channel matrix reconstructed by the k-th layer of neural network It is reconstructed based on the perception results measured by the k-th layer neural network, by introducing the channel update coefficient of the k-th layer neural network , can control the contribution of the perception results of the kth layer neural network and the k-1th layer neural network to channel reconstruction, further improving the accuracy of the updated channel matrix. Perception results, such as updated target parameters.

[0381] The channel update coefficient of the k-th layer neural network can be obtained from the noise perception module. The specific implementation principle of the noise perception module to generate the channel update coefficient can be found in the subsequent Figure 8 The specific implementation principle of the illustrated embodiment.

[0382] S706: The channel reconstruction module may output an updated ISAC channel matrix.

[0383] For example, the channel matrix The channel reconstruction module of the k-th layer neural network can output the channel matrix Alternatively, the channel reconstruction module of the k-th layer neural network can transmit the channel matrix to the signal demodulation module .

[0384] In this way, it is convenient for the signal demodulation module in the k+1 layer neural network to use the channel matrix updated by the k layer neural network Perform signal and residual updates to facilitate demodulation of the remaining data symbols.

[0385] Optionally, the channel reconstruction module outputs the updated channel matrix of the K+1th layer neural network In the case of k=K+1, the responding end may terminate the iterative calculation of the first signal. The responding end may send the perception measurement result to the initiating end. The perception measurement result may include the target parameter of each target in the at least one target output by the K+1 layer neural network. The perception measurement result sent by the responding end to the initiating end may include the target distance and target speed Alternatively, the sensing measurement results sent by the responding end to the initiating end may include .

[0386] like Figure 7 As shown, the channel reconstruction module can calculate the Doppler frequency shift and delay corresponding to the target parameters of at least one target output by the perception algorithm module. The channel reconstruction module can also calculate the angle of departure (AoD) and angle of arrival (AoA) of each target in at least one target based on the channel matrix estimated by the perception algorithm module. The channel reconstruction module uses the Doppler frequency shift and delay corresponding to the target parameters of each target, as well as the angle of departure (AoD) and angle of arrival (AoA) of each target, to reconstruct a more accurate channel matrix. The channel reconstruction module updates the reconstructed channel matrix using the channel update coefficient to balance the channel matrix updated by the previous neural network layer with the channel matrix reconstructed by the current neural network layer, resulting in a more accurate updated channel matrix.

[0387] Figure 8 A cascade schematic diagram of the noise perception module provided in an embodiment of the present application is shown.

[0388] In the perception method provided in the embodiment of the present application, the noise perception module may perform the following S801-S804.

[0389] S801. The noise perception module can perform global noise estimation.

[0390] The noise perception module may include a small 2-layer multilayer perceptron (MLP). The hidden layer width of the 2-layer MLP may be 8. The output layer of the 2-layer MLP may output a 5-dimensional vector.

[0391] Exemplarily, in the case of receiving the first signal from the initiating end, the responding end may input the first signal into the noise sensing module.

[0392] The noise perception module can estimate the noise power of the first signal , and then obtain the global noise estimate .

[0393] in, The first signal can be referred to as the received signal. Noise power The noise level can be calculated using a time domain method or a frequency domain method by calculating a section of pure noise in the received signal. For example, the pure noise in the received signal is the noise during the idle period of the communication between the initiator and the responder. Satisfies the formula:

[0394]

[0395] is the sampling value of the pure noise segment. is the number of sampling points. is the sampling point number. express The square of the absolute value of .

[0396] Optionally, the noise perception module may also be used to estimate a signal-to-noise ratio of the first signal.

[0397] S802. The noise perception module can generate hyperparameters of the neural network layer based on noise perception gating.

[0398] For example, in the k-th layer of the neural network, the noise perception module can estimate the global noise Input 2-layer MLP. Based on noise-aware gating , 2-layer MLP can generate a 5-dimensional vector of the k-th layer neural network .

[0399] in, It can be called the hyperparameter set of the k-th layer neural network. The hyperparameter set can include multiple hyperparameters, for example, .

[0400] Satisfies the formula:

[0401]

[0402] It can represent the activation function. , , and are all model parameters of the k-th layer neural network. , , and All of them are preset in the neural network model. , , and All of them are obtained by training the neural network model. It can be a column vector with 8 rows and 1 column (such as 8×1), and all elements in the column vector are real numbers. It can also be an 8×1 column vector, and all elements in the column vector are real numbers. It can be a matrix with 5 rows and 8 columns (such as 5×8), and all elements in the matrix are real numbers. It can be a 5×1 column vector, and all elements in the column vector are real numbers. In the k-th layer neural network, the 2-layer MLP is based on the global noise estimation of the input , and the preset model parameters of the k-th layer neural network, generate the 5-dimensional vector of the k-th layer neural network .

[0403] S803. The noise perception module can output hyperparameters.

[0404] For example, in the k-th layer of the neural network, the 2-layer MLP can output the hyperparameter set of the k-th layer of the neural network So that in the k-th layer of the neural network, the signal demodulation module adopts and Perform signal update and residual update, the perception algorithm module adopts and Perform threshold judgment and target parameter fusion update, and the channel reconstruction module adopts Update the channel matrix.

[0405] S804: In the k-th layer of the neural network, the noise perception module may input the hyperparameters of the k-th layer of the neural network into the signal demodulation module, the perception algorithm module, and the channel reconstruction module. The responding end may input the first signal and the data signal, channel matrix, residual vector, and target parameters output by the k-1-th layer of the neural network into the signal demodulation module, the perception algorithm module, and the channel reconstruction module in the k-th layer of the neural network.

[0406] For example, the noise perception module can update the signal coefficient and the residual update coefficients Input signal demodulation module, you can also set the multi-target detection threshold and parameter fusion coefficients Input perception algorithm module, you can also update the channel coefficient Input channel reconstruction module. The responding end can input the first signal and the data signal, channel matrix, and residual vector output by the k-1th layer neural network into the signal demodulation module in the kth layer neural network, input the target parameters output by the k-1th layer neural network, the data signal output by the signal demodulation module in the kth layer neural network, and the first signal into the perception algorithm module in the kth layer neural network, and input the target parameters output by the perception algorithm module in the kth layer neural network into the channel reconstruction module in the kth layer neural network.

[0407] Similarly, the data signal, channel matrix, residual vector and target parameters output by the k-th layer neural network can be input into the signal demodulation module, perception algorithm module and channel reconstruction module in the k+1-th layer neural network respectively.

[0408] Optionally, the MLP mechanism for generating hyperparameter sets in the noise perception module can be replaced by other context-based parameter generation mechanisms or other small neural network structures.

[0409] For example, when the neural network model is started, the responder can transmit the kth data packet to the signal demodulation module. When the responder transmits the kth data packet to the signal demodulation module, the responder can instruct the noise perception module to distribute the hyperparameters of the kth layer of the neural network. The noise perception module will and Input signal demodulation module, and Input perception algorithm module, Input channel reconstruction module.

[0410] For example, a neural network model includes three layers of neural network. When the responding end transmits the first data packet to the signal demodulation module, the responding end can instruct the noise perception module to distribute the hyperparameters of the first layer of neural network. The noise perception module can obtain the first signal from the responding end and calculate the hyperparameter set of each layer of neural network. and Input signal demodulation module, and Input perception algorithm module, Input channel reconstruction module.

[0411] When the responder transmits the second data packet to the signal demodulation module, the responder can instruct the noise perception module to distribute the hyperparameters of the second layer of the neural network. and Input signal demodulation module, and Input perception algorithm module, Input channel reconstruction module.

[0412] When the responder transmits the third data packet to the signal demodulation module, the responder can instruct the noise perception module to distribute the hyperparameters of the third layer of the neural network. and Input signal demodulation module, and Input perception algorithm module, Input channel reconstruction module.

[0413] Among them, the first data packet may include The second data packet may include 、 、 and The third data packet may include 、 、 and .

[0414] like Figure 8 In the illustrated embodiment, the noise perception module generates hyperparameter sets for each layer of the neural network based on the noise power corresponding to the first signal. This allows each layer of the neural network to use the corresponding hyperparameters to update the data symbols, residual vectors, target parameters, and channel matrices during iterative calculations of the first signal. This improves the accuracy of the data symbols, residual vectors, target parameters, and channel matrices output by each layer, thereby improving the measurement accuracy of the perception measurement performed by the responding end. The hyperparameters are generated based on the noise power of the received signal (e.g., the first signal). In this way, the noise perception module can generate hyperparameter sets corresponding to different signal-to-noise ratio environments. For example, when the responding end receives a second signal, the responding end can generate hyperparameter sets for each layer of the neural network corresponding to the different second signals based on the noise power of the second signal. This allows the responding end to achieve accurate perception measurements based on the second signal. The second signal is different from the first signal. The noise power of the second signal is different from the noise power of the first signal. The signal-to-noise ratio environment of the second signal is different from the signal-to-noise ratio environment of the first signal.

[0415] like Figure 5-Figure 8 In the embodiment shown, in the perception method provided by the embodiment of the present application, when the responding end receives the first signal, the noise perception module can generate a set of hyperparameters for each layer of the neural network in the neural network model based on the noise power of the first signal. The noise perception module can input the corresponding hyperparameters in the hyperparameter set of each layer of the neural network into the signal demodulation module, the perception algorithm module, and the channel reconstruction module.

[0416] In the first layer of the neural network, the signal demodulation module can use the pilot symbols in the first signal to perform channel estimation and obtain the initial channel matrix The signal demodulation module can be initialized to obtain initialization parameters. The signal demodulation module uses , initialization parameters and signal update coefficients of the first layer neural network are used to demodulate the first signal to output data symbols The signal demodulation module can also be based on 、 Perform residual update with the residual update coefficient of the first layer neural network to obtain the updated residual , so that the signal demodulation module in the second layer of the neural network adopts the updated residual The remaining data symbols are demodulated to update the data symbols, thereby improving the perception measurement accuracy.

[0417] At the output of the signal demodulation module and In the case of the signal demodulation module transmitting to the perception algorithm module and In the case of , the responding end or the signal demodulation module can also input the symbols after the first signal transformation into the perception algorithm module. The perception algorithm module can be based on The symbol after the first signal transformation Calculate the channel matrix The perception algorithm module can use the 2D MUSIC algorithm and the multi-target detection threshold of the first layer neural network. , for the channel matrix Perform calculations to obtain the perception results of the first layer of neural network. The perception results of the first layer of neural network include the distance of the i-th target and the speed of the i-th target The perception algorithm module can use the perception result update coefficient (or parameter fusion coefficient) of the first layer of neural network to update the perception result of the first layer of neural network to obtain the updated perception result of the first layer of neural network, so as to improve the accuracy of the perception result of the first layer of neural network. The updated perception result of the first layer of neural network can include the updated distance of the i-th target. and the updated velocity of the i-th target The updated perception result of the first layer of the neural network is called the perception measurement result of the first layer of the neural network. The perception measurement result of the first layer of the neural network is the target parameter output by the first layer of the neural network.

[0418] Output of the perception algorithm module and In the case of the perception algorithm module transmitting to the channel reconstruction module and In the case of the at least one target, the channel reconstruction module can calculate the transmission angle (AoD) and the reception angle (AoA) of each target. The channel reconstruction module can also calculate the delay corresponding to each target based on the updated target parameters of each target in the at least one target. and Doppler shift The channel reconstruction module can use the Doppler frequency shift and time delay of each target in at least one target, and the AoA and AoD of each target in at least one target to reconstruct the channel model and obtain the channel matrix reconstructed by the first layer of neural network The channel reconstruction module can use the channel update coefficient of the first layer of neural network to update the channel matrix , get the ISAC channel matrix after the first layer neural network update . is the channel matrix that can be output by the first layer of neural network. The channel matrix output by the first layer of neural network is fed back to the signal demodulation module of the next iteration (or the second layer of neural network) for more accurate data symbol recovery.

[0419] In the k-th layer neural network where k>1, the responding end can take the first signal and the channel matrix output by the previous layer (such as the k-1th layer) of the neural network as the output. , the data symbol output by the previous neural network And the residual vector output by the previous neural network , input signal demodulation module. The signal demodulation module uses the signal update coefficient of the k-th layer neural network, 、 and Update the signal to obtain the data symbol that can be output by the k-th layer neural network The data signal is updated using the signal update coefficient to achieve the solution using an iterative algorithm based on gradient descent. , which can more accurately demodulate the remaining data symbols in the first signal. In this way, after K iterations, more accurate data symbols can be obtained, improving the data demodulation accuracy. The signal demodulation module also uses the residual update coefficients of the k-th layer neural network, 、 and Update the residual vector to obtain the residual vector that can be output by the k-th layer neural network Using the residual update coefficient to update the residual vector can accelerate convergence and achieve consistent oscillation, improve noise estimation accuracy, and reduce the probability of large noise estimation errors affecting the acquisition of accurate data symbols. This can further improve data demodulation accuracy. It is used in the next layer (such as the k+1th layer) of the neural network to demodulate the remaining data symbols to update the data symbols, thereby improving the perception measurement accuracy.

[0420] At the output of the signal demodulation module and In the case of the signal demodulation module transmitting to the perception algorithm module and In the case of The target parameters output by the previous neural network are input into the perception algorithm module. For example, the target parameters output by the previous neural network are: and The perception algorithm module can be based on The symbol after the first signal transformation , calculate the channel matrix . Calculate the channel matrix based on the updated data signal , to improve the channel matrix The accuracy of the subsequent channel matrix The perception algorithm module can use the 2D MUSIC algorithm and the multi-target detection threshold of the k-th layer neural network. , for the channel matrix Perform calculations to obtain the perception results of the k-th layer of neural network. The perception results of the k-th layer of neural network include the distance of the i-th target and the speed of the i-th target The perception algorithm module can use the perception result update coefficient of the k-th layer neural network to fuse and update the perception result updated by the k-1-th layer neural network with the perception result of the k-th layer neural network to obtain the updated perception result of the k-th layer neural network, which can improve the accuracy of the perception result of the k-th layer neural network. The updated perception result of the k-th layer neural network can include the updated distance of the i-th target. and the updated velocity of the i-th target The updated perception result of the k-th layer neural network is called the perception measurement result of the k-th layer neural network. The perception measurement result of the k-th layer neural network is the target parameter output by the k-th layer neural network.

[0421] Output of the perception algorithm module and In the case of the perception algorithm module transmitting to the channel reconstruction module and In the case of the at least one target, the channel reconstruction module can calculate the transmission angle (AoD) and the reception angle (AoA) of each target. The channel reconstruction module can also calculate the delay corresponding to each target based on the updated target parameters of each target in the at least one target. and Doppler shift The channel reconstruction module can use the Doppler shift of each target in at least one target and delay , and the AoA and AoD of each target in at least one target, reconstruct the channel model and obtain the channel matrix reconstructed by the k-th layer neural network The channel reconstruction module can use the channel update coefficient of the k-th layer neural network to update the channel matrix , to control the weight balance between the channel matrix updated by the k-1th layer neural network and the channel reconstruction of the kth layer neural network, and obtain the ISAC channel matrix after the kth layer neural network is updated . is the channel matrix that can be output by the k-th neural network. Compared to the accuracy of the channel matrix output by the k-1-th neural network, the k-th neural network can output a more accurate channel matrix. The channel matrix output by the k-th neural network is fed back to the signal demodulation module of the next iteration (or the k+1-th neural network) for more accurate data symbol recovery.

[0422] The perception method provided in the embodiments of the present application does not require dedicated pilot signals, reducing the probability of communication efficiency being reduced due to dedicated pilot signals occupying communication resources. The pilot symbols in the perception method provided in the embodiments of the present application only occupy a portion of the time and frequency resources of the transmitted signal, and the perception method provided in the embodiments of the present application fully utilizes the energy and information of the entire signal to achieve higher perception accuracy and resolution. This improves the signal's resistance to complex reflections and interference in the presence of multiple perception targets, achieving higher perception accuracy and higher target resolution. The noise perception module can generate hyperparameter sets for the neural network model based on different received signals. Different received signals can correspond to different channel environments or different signal-to-noise ratio (SNR) environments. This enables the noise perception module to adapt to dynamically changing channel environments (such as different SNRs) and generate hyperparameter sets corresponding to different SNR environments, accelerating the convergence of the neural network model iteration, improving neural network model performance, achieving real-time and accurate perception and communication, and achieving synergistic gains in communication and perception. For example, the perception method provided in the embodiment of the present application not only uses data signals (such as data symbols) to improve the estimation accuracy and resolution of target parameters (distance, speed), but more importantly, the target parameter information obtained by perception is fed back through the channel reconstruction module to optimize the communication demodulation process, forming a deep coupling and mutual enhancement of communication and perception, which can improve the overall perception accuracy and communication robustness of the communication system or perception system.

[0423] The neural network model in the embodiment of the present application can also be called a feedback loop system. The feedback loop is embedded in a layer-by-layer iterative neural network structure.

[0424] Figure 9A schematic diagram of the structure of a single-layer neural network in the neural network model provided in an embodiment of the present application is shown.

[0425] like Figure 9 As shown, the signal demodulation module can receive the system input of the feedback loop system. The signal demodulation module can output an updated communication signal. The communication signal is, for example, a data symbol. The perception algorithm module can receive the data symbol output by the signal demodulation module and the symbol after the first signal transformation. The perception algorithm module can output the target parameter. The channel reconstruction module can receive the target parameter output by the perception algorithm module and the channel matrix calculated by the perception algorithm module. The channel reconstruction module can output the updated ISAC channel matrix. The signal demodulation module can receive the ISAC channel matrix output by the channel reconstruction module to perform signal update and residual update, and output the updated communication signal. This cycle is iterated. After iterating K times, the feedback loop system can perform system output. The system output content may include: the data symbol updated or output by the K+1th layer neural network and the target parameter updated or output by the K+1th layer neural network.

[0426] The signal demodulation module in the feedback loop system is also based on the pilot symbol Y in the first signal p and the pilot symbol X known to the responding end p Perform channel estimation and obtain the initial channel matrix , and use the initial channel matrix Demodulate some data symbols in the first signal The feedback loop system uses hyperparameters to continuously iterate and improve the data demodulation accuracy, thereby continuously improving the perception measurement accuracy and obtaining more accurate data symbols and perception measurement results. The pilot symbol can also be called the pilot signal. The pilot symbol Y in the first signal p , which can be called a received pilot signal. The pilot symbol X known to the responding end p This may be referred to as sending a pilot signal.

[0427] Pilot symbol Y p Carried in the first signal . Pilot symbol X p It can be carried in the downlink control information (DCI) transmitted from the initiator to the responder. p It can be pre-negotiated between the initiator and the responder, or the pilot symbol X p It can be sent by the initiator to the responder during the sensing capability exchange phase. System input can be called initial input.

[0428] For example, Figure 9 As shown, the system input may include: receiving a data signal, that is, a first signal , receive pilot signal Y p , send pilot signal X p and / or initialization parameters (i.e. and ).

[0429] Initial channel estimation can be found in Figure 5 The initial channel matrix is ​​obtained in S501 The specific implementation principle is that the signal demodulation module can perform initial channel estimation based on the pilot symbol in the first signal to obtain the estimated initial channel matrix In one possible implementation, X p It can also be carried in the first signal.

[0430] like Figure 9 As shown, in the k-th neural network layer, the responding end can input the first signal, the channel matrix output by the previous neural network layer, the data symbols output by the previous neural network layer, and the residual vector output by the previous neural network layer into the signal demodulation module. The signal demodulation module can output updated data symbols. The updated data symbols can be referred to as updated communication signals.

[0431] The updated data symbols output by the signal demodulation module are input into the perception algorithm module. The responding end may also input the transformed symbols of the first signal and the target parameters output by the previous neural network layer into the perception algorithm module. The target parameters output by the previous neural network layer may be target parameters estimated by the previous neural network layer. The perception algorithm module may output the target distance and target speed of each of the at least one target.

[0432] The target distance and speed output by the perception algorithm module are input into the channel reconstruction module. The responder can also input the channel matrix output by the previous neural network layer into the channel reconstruction module. The channel reconstruction module can then output the reconstructed ISAC channel matrix.

[0433] In this way, the k-th layer of the neural network completes one iterative calculation of the first signal.

[0434] After the neural network model iterates K times, the updated data symbol of the K+1th layer neural network can be output , target distance after the K+1th layer neural network is updated and the target speed after the K+1th layer neural network is updated The updated data symbols may be referred to as updated communication demodulated signals.

[0435] Exemplarily, the responder may input the first signal, the initialization parameter, the initial channel matrix, and the target parameter output by the first layer of the neural network into the second layer of the neural network. The target parameter output by the first layer of the neural network may include: and . is the target distance of the i-th target output by the first layer of the neural network. is the target speed of the i-th target output by the first layer of the neural network.

[0436] Neural network models can be calculated layer by layer or iterated layer by layer.

[0437] Among them, in the layer-by-layer calculation or layer-by-layer iteration from the 2nd layer of neural network to the K+1th layer of neural network, the input of each layer of neural network is: the output from the previous layer (k-1th layer) of neural network, and the received signal .

[0438] Each layer of the neural network performs a cascade of signal demodulation, perception algorithm, and channel reconstruction to obtain the output of the current layer. The specific implementation principles and technical effects of the cascade of signal demodulation, perception algorithm, and channel reconstruction performed by each layer of the neural network can be found in Figure 5-Figure 7 The specific implementation principles and technical effects of the illustrated embodiments. Taking the current layer as the K+1th layer of the neural network, i.e., k=K+1, as an example, the output of the current layer may include: and . It can be represented as a matrix . It can be the channel matrix reconstructed for the K+1th layer of the neural network Updated. . It can represent the target parameters calculated based on the K+1 layer neural network The channel reconstruction function.

[0439] In each layer of the neural network, each module can use the hyperparameter set generated by the noise perception module for the current layer of the neural network to perform the update step. The update steps include: signal update, residual update, threshold judgment, fusion update and / or channel matrix update. Taking the current layer as the K+1 layer of the neural network as an example, the hyperparameter set generated by the noise perception module for the current layer of the neural network can be . .

[0440] The system input passes through the K+1 layer neural network, that is, after K iterations, the feedback loop system can output the output result of the K+1 layer neural network.

[0441] Figure 10 A schematic diagram of the training process of the neural network model provided in an embodiment of the present application is shown.

[0442] like Figure 10 As shown, the training process of the neural network model may include S1000-S1008.

[0443] S1000, data preparation.

[0444] For example, an electronic device may obtain a training data set. The training data set may include multiple sets of training data and true labels corresponding to each set of training data. The electronic device may be a responding end or an initiating end deployed with a neural network model, or a server deployed with a neural network model.

[0445] Different sets of training data correspond to different signal-to-noise ratios. A received signal in the training data (e.g., the first signal) can use a repeated training sequence to compensate for carrier frequency offset (CFO) and symbol timing offset. This allows the training data to have a smaller range of CFO and symbol timing offset. The training sequence can include a short training sequence (STS) and / or a long training sequence (LTS).

[0446] The training data set can be obtained by the responder from the server or input by the designer at the responder.

[0447] Take the training data set including S groups of training data and S groups of true labels as an example. The sth group of training data in the S groups of training data may include: the sth received signal , the sth signal-to-noise ratio (SNR ) and the sth group of pilot signals Y p and X p .

[0448] The sth group of real labels in the S groups of real labels may include: the sth real communication data , the sth true target distance and the sth true target velocity The true labels of the sth group are the true labels corresponding to the sth group of training data.

[0449] s is a natural number, 1≤s≤S, and S is a positive integer.

[0450] S1001. Initialization.

[0451] The electronic device can initialize model parameters. For example, the electronic device can initialize a neural network model, thereby initializing the network parameters and hyperparameter sets of each neural network layer in the neural network model.

[0452] Optionally, the neural network model may include 4-7 layers of neural networks. One layer of neural network may include 61 network parameters and 5 hyperparameters.

[0453] S1002, forward propagation.

[0454] The electronic device may input S sets of training data into the neural network model. The S sets of training data are forward propagated in the neural network model.

[0455] The forward propagation of the sth group of training data in the neural network model may include: the electronic device performing channel estimation based on the pilot signal in the sth group of training data to obtain an initial channel matrix corresponding to the sth group of training data.

[0456] S1003, network output.

[0457] Neural network model output: prediction results corresponding to each set of training data in the training data set. The prediction results may include: predicted communication data , predicted target distance and the predicted target speed Communication data such as data symbols.

[0458] For example, the prediction results corresponding to the sth group of training data may include: , the sth predicted target distance and the sth predicted target speed .

[0459] S1004. Calculate the loss function.

[0460] The electronics can calculate communication loss, sensing loss and total loss.

[0461] Total loss Satisfies the formula:

[0462]

[0463] The communication loss satisfies the formula:

[0464]

[0465] The sensing loss satisfies the formula:

[0466]

[0467] in, is the weight factor, is the weight factor. Communication data that can be expressed as predictions With real communication data The mean square error between . It can be understood as the mean square error of communication demodulation. is the normalized distance estimation error. is the normalized velocity estimation error. ,and .

[0468] S1005. Back propagation.

[0469] For example, the electronic device may employ a backpropagation algorithm to calculate the gradient of the loss function relative to the parameters of each neural network, thereby adjusting the model parameters of the neural network to minimize the loss function. The parameters of the neural network include, for example, the network parameters and hyperparameters of each layer of the neural network.

[0470] For example, the electronic device can start from the K+1th layer of the neural network and use the chain rule or chain derivation method to calculate the gradient of the loss function with respect to the hyperparameters of each layer of the neural network layer by layer. Layer by layer, for example, layer by layer to the first layer of the neural network.

[0471] S1006: Parameter update.

[0472] The electronic device can use the calculated gradient to update each parameter of the neural network through an optimization algorithm, such as the gradient descent algorithm.

[0473] S1007, training convergence.

[0474] The electronic device can determine whether the training has converged.

[0475] If the training does not converge, the electronic device may execute S1002.

[0476] When the training converges, the electronic device may execute S1008.

[0477] As model training progresses, the loss function's value gradually decreases and stabilizes, and the gradient decreases. A stable loss function and a gradient within the preset gradient range indicate training convergence. A non-stabilized loss function or a gradient outside the preset gradient range indicates training non-convergence.

[0478] S1008. Model evaluation.

[0479] The electronic device may perform communication performance evaluation and / or perception performance evaluation.

[0480] Communication performance evaluation may include evaluation of block error rate (BER) and evaluation of symbol error rate (SER).

[0481] Perception performance evaluation may include calculating speed accuracy and / or distance accuracy. Distance accuracy may be the ratio of the predicted target distance to the actual target distance. Speed ​​accuracy may be the ratio of the predicted target speed to the actual target speed.

[0482] Through model evaluation, electronic devices can obtain the block error rate, symbol error rate, speed accuracy, and / or distance accuracy of the trained neural network model, thereby expanding the application scope of the model based on its performance.

[0483] like Figure 10 In the illustrated embodiment, the training data set used by the neural network model provided in the embodiment of the present application covers diverse channel conditions, mobility characteristics, multiple SNR ranges, and environmental characteristics with small residual CFO and symbol timing offset. Using the training data set to train the neural network model can enhance the robustness of the model. The perception method provided in the embodiment of the present application uses the trained neural network model to iteratively calculate the first signal, which can improve communication performance and perception accuracy. One layer of neural network in the neural network model can include 66 model parameters (61 network parameters plus 5 hyperparameters). In the case where the neural network model includes 7 layers of neural networks, the total number of parameters of the neural network model does not exceed 500, so that the neural network model occupies less memory of the electronic device, thereby enabling the deployment of the neural network model on a variety of electronic devices.

[0484] An embodiment of the present application provides a sensing method, comprising: receiving a first signal indicating a data symbol of a communication; and sending a sensing measurement result obtained based on a measurement of the data symbol.

[0485] For example, the first signal in this embodiment may be Figure 1 Orthogonal Frequency Division Multiplexing (OFDM) signal in the embodiment shown. It can also be Figure 2-Figure 9 The first signal in the embodiment shown. The perception measurement results in this embodiment may include Figure 1 The distance and speed of the target in the embodiment shown can also be Figure 2 The sensory measurement results in the embodiment shown can also be Figure 6 The updated target parameters of each target in at least one target output by the perception algorithm module in the embodiment shown can also be Figure 8-Figure 9 Perceptual measurements or system outputs in the illustrated embodiment.

[0486] The specific implementation principle of this embodiment can be found in Figure 1 The specific implementation principle of the embodiment shown can also be found in Figure 2 The specific implementation principle of the embodiment shown can also be found in Figure 3-Figure 8The specific implementation principle of the embodiment shown can also be found in Figure 9 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0487] Thus, taking the execution subject of this embodiment as the perception responding end as an example, the perception method provided by the embodiment of the present application, the perception responding end can receive the first signal of the data symbol sent by the perception initiating end for indicating the communication. The perception responding end can obtain the perception measurement result based on the data symbol measurement in the first signal, and send the perception measurement result to the perception initiating end. The perception measurement result, for example, the distance and speed of each target in at least one target between the perception initiating end and the perception responding end. The perception initiating end can obtain the perception measurement result sent by the perception responding end, and realize the perception of the environment between the perception initiating end and the perception responding end. Compared with the method of performing environmental perception through a dedicated pilot signal, in the perception method of the embodiment of the present application, the perception responding end can obtain the perception measurement result based on the data symbol measurement, and there is no need for the perception initiating end to send a dedicated pilot signal, which can reduce the occupation of communication resources by the dedicated pilot signal, thereby reducing the probability of communication efficiency being reduced due to the occupation of communication resources by the dedicated pilot signal. It can also save communication resources and improve the resource utilization of data symbols. Among them, the perception initiating end can be Figures 1-9 The sensing initiator or initiator in the embodiment shown. The sensing responder can be Figures 1-9 The sensing responder or responder in the illustrated embodiment.

[0488] Optionally, before receiving the first signal, the method further includes: sending first indication information, where the first indication information is used to indicate that data symbols are used to perform perception measurement.

[0489] For example, the first indication information may be Figure 3 The first indication information in the embodiment shown. The specific implementation principle of the embodiment of this application can be found in Figure 3 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0490] In this way, the data symbols in the first signal are used for perception measurement to achieve environmental perception without the need for a dedicated pilot signal, thereby reducing the probability of communication resources being occupied by the dedicated pilot signal, resulting in reduced communication efficiency.

[0491] Optionally, before receiving the first signal, the method further includes: sending first capability information, where the first capability information is used to indicate support for a data symbol-based perception measurement function.

[0492] Exemplarily, the first capability information may be Figure 3 The first capability information in the embodiment shown. The specific implementation principle of the embodiment of this application can be found in Figure 3 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0493] In this way, before receiving the first signal, capability negotiation between the perception responding end and the perception initiating end is achieved, which facilitates the perception responding end to use data symbols to perform perception measurement and achieve environmental perception.

[0494] Optionally, before sending the perception measurement result, the method further includes: inputting the first signal into a first model, so that the first model outputs the perception measurement result. The first model is configured to perform perception processing on a data symbol indicated by the first signal to obtain the perception measurement result. The data symbol is obtained by demodulating the first signal, or is obtained by updating a demodulation result of demodulating the first signal.

[0495] For example, the first model can be Figures 4-10 The neural network model or feedback loop system in the embodiment shown. The specific implementation principle of this embodiment can be found in Figure 4-Figure 8 The specific implementation principle of the embodiment shown can also be found in Figure 9 The specific implementation principle of the illustrated embodiment.

[0496] In this way, the first model performs perception processing on the data symbols indicated by the first signal to obtain perception measurement results, enabling environmental perception and communication without the need for dedicated pilot signals. The data symbols used for perception processing are obtained by demodulating the first signal, which can improve resource utilization. The data symbols used for perception processing are obtained by updating the demodulation results of the demodulated first signal. By updating the data symbols to obtain more accurate data symbols, the measurement accuracy of the perception measurement results measured based on the data symbols can be improved.

[0497] Optionally, the sensing measurement result includes a speed and distance of each target object in at least one target object, where the target object is an object between the sensing initiator and the sensing responder. Sensing and processing data symbols in the first signal to obtain the sensing measurement result includes calculating a channel frequency response matrix of a channel corresponding to the first signal based on the data symbols and the first signal. Calculating and processing the channel frequency response matrix to obtain the distance and speed of each target object in the at least one target object.

[0498] For example, the target object may be Figures 1-10 In the embodiment shown, the target object or target, the target object can be simply referred to as the target. The distance can be Figures 1-10 The target distance or distance in the embodiment shown. The speed can be Figures 1-10 The target speed or speed in the embodiment shown. The specific implementation principle of this embodiment can be found in Figure 5-Figure 6 The specific implementation principle of the illustrated embodiment.

[0499] In this way, perception measurement is performed based on the data symbols to obtain a perception measurement result.

[0500] Optionally, the first model includes at least one layer of neural network. The channel frequency response matrix satisfies the following formula:

[0501]

[0502] in, is the channel frequency response matrix calculated by the k-th neural network in at least one layer of neural network or the channel matrix calculated by the k-th neural network, for The elements in is the matrix representation of the first signal, is a matrix The element in row n and column l in is the matrix representation of the data symbols demodulated or updated by the k-th layer neural network, is a matrix The element in the nth row and lth column, 1≤n≤N, 1≤l≤L, k is the sequence number of the neural network layer in at least one neural network, n is the row sequence number and is an integer, l is the column sequence number and is an integer, N is the number of matrix rows, L is the number of matrix columns, k is a natural number, N is an integer, and L is an integer.

[0503] In this way, the channel frequency response matrix is ​​determined based on the first signal and the data symbols in the first signal. The channel frequency response matrix is ​​used for perceptual measurement to achieve determination of perceptual measurement results based on the data symbols.

[0504] Optionally, calculating and processing the channel frequency response matrix to obtain the distance and velocity of each target object in the at least one target object includes: calculating the channel frequency response matrix using a spectral estimation algorithm to obtain a time delay and Doppler shift corresponding to the distance and velocity of each target object. Determining the distance and velocity of each target object based on the time delay and Doppler shift corresponding to the distance and velocity of each target object. Each target object is each target object in the at least one target object.

[0505] For example, the spectrum estimation algorithm may be a 2D MUSIC algorithm. The specific implementation principle of this embodiment may refer to the specific implementation principle of the embodiment shown in S603-S604, or the specific implementation principle of the embodiment shown in S603-S605.

[0506] In this way, perception measurement is performed based on the data symbols to obtain a perception measurement result.

[0507] Optionally, the first signal is modulated by orthogonal frequency division multiplexing modulation. A spectrum estimation algorithm is used to calculate the channel frequency response matrix to obtain the time delay and Doppler frequency shift corresponding to the distance and speed of each target object, including: using a vectorization operator to convert the channel frequency response matrix Vectorization, get a one-dimensional long vector . For long vectors Calculate the covariance matrix to get the covariance matrix R. Perform eigenvalue decomposition and subspace separation on the covariance matrix R to get the noise subspace Based on the noise subspace The constructed power spectrum is used for peak search to obtain the time delay and Doppler shift corresponding to the distance and speed of each target object.

[0508] For example, the specific implementation principle of this embodiment may refer to the specific implementation principle of S603, or may refer to the specific implementation principles of S6031-S6037.

[0509] In this way, the distance and speed of each target object can be determined based on the time delay and Doppler shift corresponding to the distance and speed of each target object.

[0510] Optionally, the time delay and Doppler shift corresponding to at least one target object include the time delay and Doppler shift corresponding to each peak in the I peaks. Performing a peak search on the constructed power spectrum to obtain a time delay and Doppler shift corresponding to the distance and velocity of each target object includes: performing a peak search on the power spectrum to obtain a plurality of peaks; determining I peaks from the plurality of peaks that are greater than or equal to a first parameter; the coordinate position of each of the I peaks in the power spectrum includes the time delay and Doppler shift corresponding to each peak; the first parameter is determined based on the noise power of the first signal; I is the number of the at least one target object, and I is a positive integer.

[0511] In this way, the first parameter is used to perform threshold judgment on the multiple peaks obtained by the peak search. It is possible to determine the peak generated by the target from the multiple peaks searched, that is, to determine the I peaks corresponding to I targets between the initiator and the responder from the multiple peaks searched. In the case where the side lobe interferes with the peak generated by the target, it is possible to distinguish the true target peak from the side lobe (or noise). The true target peak is a peak greater than or equal to the first parameter. The side lobe (or noise) corresponds to a peak less than the first parameter. The true target peak is, for example, any one of the I peaks. This can improve the accuracy of the perception measurement results. The first parameter can be Figure 6 and Figures 8-10 Multi-target detection threshold in an embodiment.

[0512] In one possible implementation, the long vector Satisfies the formula:

[0513]

[0514] The covariance matrix R satisfies the formula:

[0515]

[0516] Power spectrum Satisfies the formula:

[0517]

[0518] Where T represents transpose, express The transpose of for The first element of for The second column element of for The Lth column element of for The conjugate transpose of , for The conjugate transpose of for The conjugate transpose of is the two-dimensional joint steering vector, is the power spectrum The independent variable represents the delay, is the power spectrum is an independent variable and represents the Doppler shift.

[0519] Optionally, the two-dimensional joint steering vector Satisfies the formula:

[0520]

[0521] in, represents the Kronecker product, which can also represent the tensor product or outer product. The one-dimensional distance-oriented vector constructed for the k-th layer of the neural network, , is the one-dimensional velocity guidance vector constructed for the k-th layer of the neural network, , e is a natural constant or transcendental number, j is the imaginary unit, is the carrier spacing of OFDM, for The independent variable is The corresponding delay, for The independent variable is The corresponding Doppler shift is, is the period of an OFDM symbol or the length of time occupied by an OFDM symbol, express The transpose of .

[0522] Optionally, the distance of the i-th target object in at least one target object satisfies the formula:

[0523]

[0524] The velocity of the i-th target object among at least one target object satisfies the formula:

[0525]

[0526] in, is the distance of the i-th target object extracted by the k-th layer neural network, The time delay corresponding to the distance of the i-th target object calculated by the k-th layer neural network, is the speed of the i-th target object extracted by the k-th layer neural network, is the Doppler frequency shift corresponding to the velocity of the i-th target object calculated by the k-th layer neural network, c is the propagation speed of electromagnetic waves or the speed of light, is the preset carrier frequency.

[0527] In this way, the distance of the target is determined based on the target corresponding time delay, and the speed of the target is determined based on the target corresponding Doppler frequency shift.

[0528] Optionally, the perception measurement result includes an updated distance of each target object in at least one target object and an updated speed of each target object in at least one target object. Calculating and processing the channel frequency response matrix to obtain the distance and speed of each target object in at least one target object further includes: fusing and updating the distance of the i-th target object and the speed of the i-th target object to obtain an updated distance of the i-th target object and an updated speed of the i-th target object. The updated distance of the i-th target object and the updated speed of the i-th target object satisfy the formula:

[0529]

[0530] in, , , is the updated distance of the i-th target object obtained by the k-th layer neural network, is the updated speed of the i-th target object obtained by the k-th layer neural network, is the updated distance of the i-th target object obtained by the k-1-th layer neural network, is the updated speed of the i-th target object obtained by the k-1-th layer neural network, is the second parameter of the k-th layer neural network, which is determined based on the noise power of the first signal. [ ] represents a matrix.

[0531] In this way, since the second parameter is generated based on the noise power of the first signal, using the second parameter to fuse and update the extracted target parameter can further reduce the impact of noise such as side lobes on the target parameter, improve the accuracy of the updated target parameter, and thus improve the perception accuracy. Among them, the second parameter can be Figure 6 and Figures 8-10 Parameter fusion coefficients in the embodiment: Target parameters, such as the distance and speed of the target, or, for example, the distance of the i-th target object extracted by the k-th layer neural network and the speed of the i-th target object extracted by the k-th layer neural network.

[0532] Optionally, the data symbols include initial data symbols, which are obtained by demodulating the first signal in the following manner: performing channel estimation based on the pilot symbols in the first signal to obtain initial channel information, and demodulating the first signal using the initial channel information to obtain initial data symbols.

[0533] For example, the initial channel information may be the initial channel information in the embodiment shown in S501, or the initial channel matrix in the embodiment shown in S501. For the specific implementation principle of demodulating the first signal using the initial channel information to obtain the initial data symbol, reference may be made to the specific implementation principle of S503-S504.

[0534] This facilitates perceptual measurements using data symbols.

[0535] Optionally, the first model includes a K+1-layer neural network, where K is an integer greater than or equal to 1. The data symbols also include data symbols updated by the K+1-layer neural network. The data symbols updated by the K+1-layer neural network are obtained by updating the demodulation result after demodulating the first signal in the following manner: using the third parameter of the K+1-layer neural network, the channel matrix updated by the K-layer neural network, and the noise vector updated by the K-layer neural network, the data symbols updated by the K-layer neural network are updated to obtain data symbols updated by the K+1-layer neural network. The third parameter is determined based on the noise power of the first signal, and the data symbols updated by the 1st layer neural network are initial data symbols.

[0536] For example, the third parameter may be Figure 5 The signal update coefficients in the embodiment shown. The noise vector can be Figure 5 The residual or residual vector in the embodiment shown. The specific implementation principle of this embodiment can be found in the specific implementation principle of S504, which will not be repeated here.

[0537] In this way, the remaining data symbols in the first signal except for the demodulated data symbols can be demodulated more accurately, thereby obtaining more accurate data symbols.

[0538] Optionally, the data symbol updated by the K+1th layer neural network Satisfies the formula:

[0539]

[0540] in, , , is the symbol of the first signal after being transformed into the symbol domain, For the The data symbols updated by the layer neural network, Can be the first The third parameter of the layer neural network, is the channel matrix Middle Path The corresponding amplitude component, For the The channel matrix updated by the layer neural network, is the residual vector or noise vector updated by the K-th layer neural network, is the channel matrix Middle Path The corresponding phase component, for The transpose of for The transpose of is the path number of the path for transmitting the first signal.

[0541] In this way, the iterative algorithm based on gradient descent is used to solve , it is possible to more accurately demodulate the first signal except for the demodulated data symbols (such as ) to obtain more accurate data symbols.

[0542] Optionally, the residual vector updated by the K-th layer neural network satisfies the formula:

[0543]

[0544] in, is the residual vector or noise vector updated by the K-th layer neural network, It can be the residual vector or noise vector updated for the K-1th layer neural network, It can be the fourth parameter of the K-th layer neural network, and the fourth parameter is determined based on the noise power of the first signal. is the channel matrix Middle Path The corresponding amplitude component, is the channel matrix updated by the K-1th layer neural network, is the channel matrix Middle Path The corresponding phase component.

[0545] Exemplarily, the fourth parameter may be Figure 5 The residual update coefficient or noise update parameter in the illustrated embodiment can be referred to the specific implementation principle of S505 for the specific implementation principle of this embodiment, which will not be described in detail here.

[0546] Thus, by introducing the momentum term Updating the residual can accelerate convergence and suppress oscillation, improve noise estimation accuracy, and reduce the probability of large noise estimation errors leading to reduced accuracy of acquired data symbols. This can further improve data demodulation accuracy.

[0547] Optionally, Channel matrix updated by layer neural network Satisfies the formula:

[0548]

[0549] in, , is the fifth parameter of the K-th layer neural network, and the fifth parameter is determined based on the noise power of the first signal. is the updated channel matrix of the K-1th layer neural network, is a communication and environment perception channel matrix reconstructed by the K-th layer neural network based on the Doppler frequency shift of each target object in at least one target object, the time delay of each target object, the receiving angle of each target object, and the transmitting angle of each target object, the Doppler frequency shift of the i-th target object in at least one target object is calculated based on the velocity of the i-th target object, the time delay of the i-th target object in at least one target object is calculated based on the distance of the i-th target object, and i is a positive integer.

[0550] Exemplarily, the fifth parameter may be Figure 7 The channel update coefficient in the embodiment shown. The specific implementation principle of this embodiment can be found in Figure 7 The specific implementation principle of the illustrated embodiment or S705 will not be described in detail here.

[0551] In this way, the fifth parameter can be used to control the weight balance between the channel matrix updated by the K-1th layer neural network and the channel reconstruction of the Kth layer neural network, making the updated channel matrix more accurate. In addition, the channel matrix reconstructed by the Kth layer neural network It can be reconstructed based on the perception results measured by the K-th layer neural network by introducing the channel update coefficient of the K-th layer neural network , it is possible to control the contribution of the perception results of the Kth layer neural network and the K-1th layer neural network to channel reconstruction, further improving the accuracy of the updated channel matrix. Perception results, such as perception measurement results or updated target parameters.

[0552] Optionally, the third parameter of the k-th layer neural network is included in the hyperparameter set of the k-th layer neural network , Satisfies the formula:

[0553]

[0554] Where k is the sequence number of the neural network layer in the K+1 layer neural network, 1≤k≤K+1, represents the activation function, 、 、 and are all model parameters of the k-th layer neural network, 、 、 and All are preset in the first model. 、 、 and All are obtained by training the first model. is the noise power of the first signal, It also includes the first parameter, second parameter, fourth parameter and fifth parameter of the k-th layer neural network, the first parameter is used to judge the threshold of the searched peak, the second parameter is used to update the fusion of the distance and speed of the target object, the fourth parameter is used to update the noise vector or to update the residual vector, and the fifth parameter is used to update the channel matrix.

[0555] For example, the specific implementation principle of this embodiment can be found in Figure 8 The specific implementation principle of the illustrated embodiment will not be described in detail here.

[0556] In this way, the first parameter, the second parameter, the third parameter, the fourth parameter and the fifth parameter are determined based on the noise power of the first signal, so that the first model adopts these parameters to improve the perception measurement accuracy.

[0557] Optionally, the perception measurement result is carried in a media access control layer frame payload or a first management frame, and the first management frame is used to transmit the perception measurement result.

[0558] In this way, the perception measurement result can be transmitted through the medium access control layer frame or the first management frame.

[0559] An embodiment of the present application provides a sensing method, comprising: sending a first signal, the first signal being used to indicate a data symbol of communication, and receiving a sensing measurement result, the sensing measurement result being obtained based on a measurement of the data symbol.

[0560] In this way, taking the execution subject of this embodiment as the perception initiator as an example, the perception method provided by the embodiment of the present application, the perception initiator can send a first signal for indicating the data symbol of communication to the perception responder, so that the perception responder obtains the perception measurement result based on the data symbol measurement in the first signal. The perception initiator can obtain the perception measurement result sent by the perception responder, and realize the perception of the environment between the perception initiator and the perception responder. Compared with the method of performing environmental perception through a dedicated pilot signal, in the perception method of the embodiment of the present application, the perception responder can obtain the perception measurement result based on the data symbol measurement, without the need for the perception initiator to send a dedicated pilot signal, which can reduce the occupation of communication resources by the dedicated pilot signal, and thus reduce the probability of communication efficiency being reduced due to the occupation of communication resources by the dedicated pilot signal. It can also save communication resources and improve the resource utilization of data symbols. Among them, the perception initiator can be Figures 1-9 The sensing initiator or initiator in the embodiment shown. The sensing responder can be Figures 1-9 The sensing responder or responder in the illustrated embodiment.

[0561] Optionally, before sending the first signal, the method further includes: receiving first indication information, where the first indication information is used to indicate that data symbols are used to perform perception measurement.

[0562] In this way, the data symbols in the first signal are used for perception measurement to achieve environmental perception without the need for a dedicated pilot signal, thereby reducing the probability of communication resources being occupied by the dedicated pilot signal, resulting in reduced communication efficiency.

[0563] Optionally, before sending the first signal, the method further includes: receiving first capability information, where the first capability information is used to indicate support for a data symbol-based perception measurement function.

[0564] In this way, before sending the first signal, capability negotiation between the perception responding end and the perception initiating end is achieved, which facilitates the perception responding end to use data symbols to perform perception measurement and realize environmental perception.

[0565] An embodiment of the present application provides a perception system, including: a perception initiator and a perception responder, the perception responder is used to execute the method described in the first aspect or any possible implementation of the first aspect, and the perception initiator is used to execute the method described in the second aspect or any possible implementation of the second aspect.

[0566] An embodiment of the present application provides an electronic device, comprising: a signal demodulation module, a perception algorithm module, a channel reconstruction module, and a noise perception module. The electronic device is configured to receive a first signal, the first signal being used to indicate a data symbol for communication. The signal demodulation module is configured to demodulate the first signal to obtain the data symbol indicated by the first signal, or to update the demodulation result after demodulating the first signal to obtain the data symbol in the first signal. The perception algorithm module is configured to perform perception processing on the data symbol in the first signal to obtain the distance and velocity of each target object in at least one target object. The channel reconstruction module is configured to reconstruct a communication and environment perception channel matrix based on the distance and velocity of each target object in the at least one target object, and to update the reconstructed communication and environment perception channel matrix. The updated communication and environment perception channel matrix is ​​used to update the demodulation result after demodulating the first signal. The electronic device is further configured to transmit a perception measurement result, the perception measurement result including the distance and velocity of each target object in at least one target object between a perception initiator and a perception responder, the perception measurement result being obtained based on data symbol measurements.

[0567] For example, the signal demodulation module may be Figures 4-10 The signal demodulation module in the embodiment shown. The perception algorithm module can be Figures 4-10 The sensing algorithm module in the embodiment shown. The channel reconstruction module can be Figures 4-10 The channel reconstruction module in the embodiment shown. The noise perception module can be Figures 4-10 The noise perception module in the illustrated embodiment.

[0568] In this way, a sensing measurement result is obtained based on the measurement of the data symbols in the first signal, thereby achieving perception of the environment between the sensing initiator and the sensing responder. Compared to the method of performing environmental perception through dedicated pilot signals, the sensing method in the embodiment of the present application obtains a sensing measurement result based on the measurement of data symbols, eliminating the need for dedicated pilot signals. This can reduce the occupation of communication resources by dedicated pilot signals, thereby reducing the probability of communication resources being occupied by dedicated pilot signals, resulting in reduced communication efficiency. This can also save communication resources and improve the resource utilization of data symbols.

[0569] An embodiment of the present application provides a sensing device, which may be an electronic device or a chip or chip system within an electronic device. The sensing device may include a display unit and a processing unit. When the sensing device is an electronic device, the display unit may be a display screen. The display unit is configured to perform the display step so that the electronic device implements a sensing method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. When the sensing device is an electronic device, the processing unit may be a processor. The sensing device may also include a storage unit, which may be a memory. The storage unit is configured to store instructions, and the processing unit executes the instructions stored in the storage unit so that the electronic device implements a sensing method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. When the sensing device is a chip or chip system within an electronic device, the processing unit may be a processor. The processing unit executes the instructions stored in the storage unit so that the electronic device implements a sensing method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. The storage unit may be a storage unit within the chip (eg, a register, a cache, etc.), or a storage unit within the electronic device that is located outside the chip (eg, a read-only memory, a random access memory, etc.).

[0570] An embodiment of the present application provides a chip or chip system, comprising at least one processor and a communication interface, wherein the communication interface and the at least one processor are interconnected via a line, and the at least one processor is configured to execute a computer program or instruction to perform the method described in the first aspect, any possible implementation of the first aspect, the second aspect, or any possible implementation of the second aspect. The communication interface in the chip may be an input / output interface, a pin, or a circuit.

[0571] Optionally, the chip or chip system described above in the embodiments of the present application further includes at least one memory, wherein instructions are stored in the at least one memory. The memory may be a storage unit within the chip, such as a register or cache, or a storage unit of the chip (such as a read-only memory or random access memory).

[0572] It should be noted that the module names involved in the embodiments of the present application can be defined as other names as long as the functions of each module can be achieved, and there is no specific restriction on the names of the modules.

[0573] The perception responding end in the embodiments of the present application may also be a WiFi chipset or module that integrates the technology of the perception method provided in the embodiments of the present application, or may be an application software or system such as indoor positioning, security monitoring, or smart home control system based on the technology of the perception method provided in the embodiments of the present application, or may be a WiFi AP device that supports the perception method provided in the embodiments of the present application, a WiFi station (STA) device that supports the perception method provided in the embodiments of the present application, or a UE that supports the perception receiving end of the perception method provided in the embodiments of the present application. The perception initiating end in the embodiments of the present application may also be a WiFi chipset or module that integrates the technology of the perception method provided in the embodiments of the present application, or may be an application software or system such as indoor positioning, security monitoring, or smart home control system based on the technology of the perception method provided in the embodiments of the present application, or may be a WiFi AP device that supports the perception method provided in the embodiments of the present application, a WiFi station device that supports the perception method provided in the embodiments of the present application, or a UE that supports the perception receiving end of the perception method provided in the embodiments of the present application.

[0574] The above describes the perception method of the embodiment of the present application. The following describes the device for performing the above method provided by the embodiment of the present application. It will be understood by those skilled in the art that the method and device can be combined and referenced with each other, and the relevant device provided by the embodiment of the present application can perform the steps in the above-mentioned list sorting method.

[0575] The sensing method provided in the embodiment of the present application can be applied to electronic devices with communication functions. The electronic devices include terminal devices. The specific device form of the terminal device can refer to the above related descriptions and will not be repeated here.

[0576] An embodiment of the present application provides a terminal device, which includes: a processor and a memory; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the terminal device executes the above method.

[0577] The present embodiment provides a chip. The chip includes a processor configured to invoke a computer program stored in a memory to execute the technical solution of the above embodiment. The implementation principles and technical effects are similar to those of the above-mentioned related embodiments and will not be further described here.

[0578] The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the above-mentioned method is implemented. The methods described in the above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. If implemented in software, the functions can be stored as one or more instructions or codes on a computer-readable medium or transmitted on a computer-readable medium. Computer-readable media can include computer storage media and communication media, and can also include any medium that can transfer a computer program from one place to another. The storage medium can be any target medium that can be accessed by a computer.

[0579] In one possible implementation, computer-readable media may include RAM, ROM, compact disc read-only memory (CD-ROM) or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium designed to carry or store the desired program code in the form of instructions or data structures and accessible by a computer. Furthermore, any connection is appropriately termed a computer-readable medium. For example, if software is transmitted from a website, server, or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. Disk and disc, as used herein, include laser discs, digital versatile discs (DVDs), floppy disks, and Blu-ray discs, where disks typically reproduce data magnetically, while discs reproduce data optically using lasers. Combinations of the above are also intended to be included within the scope of computer-readable media.

[0580] An embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed, the computer executes the above method.

[0581] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processing unit of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable device to produce a machine, so that the instructions executed by the processing unit of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0582] The above specific implementation methods further explain in detail the purpose, technical solutions and beneficial effects of the embodiments of the present application. It should be understood that the above are only specific implementation methods of the embodiments of the present application and are not intended to limit the scope of protection of the embodiments of the present application. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solutions of the embodiments of the present application should be included in the scope of protection of the embodiments of the present application.

Claims

1. A perception method, characterized in that: The method comprises: receiving a first signal, wherein the first signal is used to indicate a data symbol of communication; A perceptual measurement result is sent, where the perceptual measurement result is obtained based on the data symbol measurement.

2. The method according to claim 1, characterized in that Before receiving the first signal, the method further includes: First indication information is sent, where the first indication information is used to instruct to use data symbols to perform perception measurement.

3. The method according to claim 1 or 2, characterized in that Before receiving the first signal, the method further includes: First capability information is sent, where the first capability information is used to indicate support for a data symbol-based perception measurement function.

4. The method according to claim 1, wherein Before sending the perception measurement result, the method further includes: inputting the first signal into a first model so that the first model outputs the perception measurement result; The first model is used to perform perception processing on the data symbols indicated by the first signal to obtain a perception measurement result; the data symbols are obtained by demodulating the first signal, or are obtained by updating the demodulation result after demodulating the first signal.

5. The method according to claim 4, characterized in that The sensing measurement result includes the speed and distance of each target object in at least one target object, wherein the target object is an object between the sensing initiator and the sensing responder; The performing sensing processing on the data symbols in the first signal to obtain a sensing measurement result includes: Calculating a channel frequency response matrix of a channel corresponding to the first signal based on the data symbols and the first signal; The channel frequency response matrix is ​​calculated and processed to obtain the distance and speed of each target object in the at least one target object.

6. The method according to claim 5, characterized in that The first model includes at least one layer of neural network; The channel frequency response matrix satisfies the following formula: in, is the channel frequency response matrix calculated by the k-th layer neural network or the channel matrix calculated by the k-th layer neural network, for The elements in is the matrix representation of the first signal, is a matrix The element in row n and column l in is the matrix representation of the data symbols demodulated or updated by the k-th layer neural network, is a matrix The element in the nth row and lth column in , 1≤n≤N, 1≤l≤L, k is the sequence number of the neural network layer in the at least one neural network, n is the row sequence number and is an integer, l is the column sequence number and is an integer, N is the number of matrix rows, L is the number of matrix columns, k is a natural number, N is an integer, and L is an integer.

7. The method according to claim 6, characterized in that The calculating and processing the channel frequency response matrix to obtain the distance and speed of each target object in the at least one target object includes: Calculating the channel frequency response matrix using a spectrum estimation algorithm to obtain the time delay and Doppler frequency shift corresponding to the distance and speed of each target object; The distance and speed of each target object are determined based on the time delay and Doppler shift corresponding to the distance and speed of each target object.

8. The method according to claim 7, characterized in that The first signal is modulated by orthogonal frequency division multiplexing modulation; The using of a spectrum estimation algorithm to calculate the channel frequency response matrix to obtain the time delay and Doppler shift corresponding to the distance and speed of each target object includes: The channel frequency response matrix is ​​vectorized using a vector operator Vectorization, get a one-dimensional long vector ; For the long vector Perform covariance matrix calculation to obtain the covariance matrix R; Perform eigenvalue decomposition and subspace separation on the covariance matrix R to obtain the noise subspace ; Based on the noise subspace The constructed power spectrum is searched for peaks to obtain the time delay and Doppler shift corresponding to the distance and speed of each target object.

9. The method according to claim 8, characterized in that The time delay and Doppler shift corresponding to the at least one target object include the time delay and Doppler shift corresponding to each peak in the I peak; The pair is based on the noise subspace The constructed power spectrum is peak searched to obtain the time delay and Doppler shift corresponding to the distance and speed of each target object, including: Performing peak search on the power spectrum to obtain multiple peaks; I peaks greater than or equal to a first parameter are determined from the multiple peaks, the coordinate position of each peak in the power spectrum includes the time delay and Doppler frequency shift corresponding to each peak, the first parameter is determined based on the noise power of the first signal, I is the number of target objects of the at least one target object, and I is a positive integer.

10. The method according to claim 9, characterized in that The long vector Satisfies the formula: The covariance matrix R satisfies the formula: The power spectrum Satisfies the formula: Where T represents transpose, express The transpose of for The first column element of for The second column element of for The Lth column element of for The conjugate transpose of , for The conjugate transpose of for The conjugate transpose of is the two-dimensional joint steering vector, The power spectrum The independent variable represents the delay, The power spectrum is an independent variable and represents the Doppler shift.

11. The method according to claim 10, characterized in that The two-dimensional joint steering vector Satisfies the formula: in, represents the Kronecker product, also known as the tensor product or outer product, The one-dimensional distance-oriented vector constructed for the k-th layer of the neural network, , is the one-dimensional velocity guidance vector constructed for the k-th layer of the neural network, , e is a natural constant or transcendental number, j is the imaginary unit, is the carrier spacing of OFDM, for The independent variable is The corresponding delay, for The independent variable is The corresponding Doppler shift is, is the period of an OFDM symbol or the length of time occupied by an OFDM symbol, express The transpose of .

12. The method according to claim 7, characterized in that The distance of the i-th target object in the at least one target object satisfies the formula: The speed of the i-th target object in the at least one target object satisfies the formula: in, is the distance of the i-th target object extracted by the k-th layer neural network, The time delay corresponding to the distance of the i-th target object calculated by the k-th layer neural network, is the speed of the i-th target object extracted by the k-th layer neural network, is the Doppler frequency shift corresponding to the velocity of the i-th target object calculated by the k-th layer neural network, c is the propagation speed of electromagnetic waves or the speed of light, is the preset carrier frequency.

13. The method according to claim 12, characterized in that The perception measurement results include an updated distance of each target object in the at least one target object and an updated speed of each target object in the at least one target object; The calculating and processing the channel frequency response matrix to obtain the distance and speed of each target object in the at least one target object further includes: The distance of the i-th target object and the speed of the i-th target object are fused and updated to obtain an updated distance of the i-th target object and an updated speed of the i-th target object; the updated distance of the i-th target object and the updated speed of the i-th target object satisfy the formula: in, , , is the updated distance of the i-th target object obtained by the k-th layer neural network, is the updated speed of the i-th target object obtained by the k-th layer neural network, is the updated distance of the i-th target object obtained by the k-1-th layer neural network, is the updated speed of the i-th target object obtained by the k-1-th layer neural network, is the second parameter of the k-th layer neural network, which is determined based on the noise power of the first signal, and [ ] represents a matrix.

14. The method according to any one of claims 4 to 13, characterized in that The data symbols include initial data symbols, and the initial data symbols are obtained by demodulating the first signal in the following manner: Performing channel estimation based on pilot symbols in the first signal to obtain initial channel information; The first signal is demodulated using the initial channel information to obtain initial data symbols.

15. The method according to claim 14, characterized in that The first model includes a K+1 layer neural network, where K is an integer greater than or equal to 1; the data symbol also includes a data symbol updated by the K+1 layer neural network; the data symbol updated by the K+1 layer neural network is obtained by updating the demodulation result after demodulating the first signal in the following manner: Using the third parameter of the K+1th layer neural network, the channel matrix updated by the Kth layer neural network, and the noise vector updated by the Kth layer neural network, the data symbol updated by the Kth layer neural network is updated to obtain the data symbol updated by the K+1th layer neural network; The third parameter is determined based on the noise power of the first signal, and the data symbol updated by the first layer of the neural network is the initial data symbol.

16. The method according to claim 15, characterized in that The data symbols updated by the K+1th layer neural network Satisfies the formula: in, , , is the symbol of the first signal after being transformed into the symbol domain, For the The data symbols updated by the layer neural network, For the The third parameter of the layer neural network, is the channel matrix Middle Path The corresponding amplitude component, For the The channel matrix updated by the layer neural network, is the residual vector or noise vector updated by the K-th layer neural network, is the channel matrix Middle Path The corresponding phase component, for The transpose of for The transpose of is the path number of the path for transmitting the first signal.

17. The method according to claim 16, characterized in that The residual vector updated by the K-th layer neural network satisfies the formula: in, is the residual vector or noise vector updated by the K-th layer neural network, is the residual vector or noise vector updated by the K-1th layer neural network, is the fourth parameter of the K-th layer neural network, and the fourth parameter is determined based on the noise power of the first signal, is the channel matrix Middle Path The corresponding amplitude component, is the channel matrix updated by the K-1th layer neural network, is the channel matrix Middle Path The corresponding phase component.

18. The method according to claim 16, characterized in that The said Channel matrix updated by layer neural network Satisfies the formula: in, , is a fifth parameter of the K-th layer neural network, wherein the fifth parameter is determined based on the noise power of the first signal, is the updated channel matrix of the K-1th layer neural network, is a communication and environment perception channel matrix reconstructed by the K-th layer neural network based on the Doppler frequency shift of each target object in at least one target object, the time delay of each target object, the receiving angle of each target object, and the transmitting angle of each target object, the Doppler frequency shift of the i-th target object in at least one target object is calculated based on the velocity of the i-th target object, the time delay of the i-th target object in at least one target object is calculated based on the distance of the i-th target object, and i is a positive integer.

19. The method according to claim 15, characterized in that The third parameter of the k-th layer neural network is included in the hyperparameter set of the k-th layer neural network , Satisfies the formula: Wherein, k is the sequence number of the neural network layer in the K+1 layer neural network, 1≤k≤K+1, represents the activation function, 、 、 and are all model parameters of the k-th layer neural network, 、 、 and are all preset in the first model. 、 、 and are all obtained by training the first model. is the noise power of the first signal, It also includes the first parameter, second parameter, fourth parameter and fifth parameter of the k-th layer neural network, the first parameter is used to judge the threshold of the searched peak, the second parameter is used to update the fusion of the distance and speed of the target object, the fourth parameter is used to update the noise vector or to update the residual vector, and the fifth parameter is used to update the channel matrix.

20. The method according to claim 1, wherein The perception measurement result is carried in a media access control layer frame payload or a first management frame, and the first management frame is used to transmit the perception measurement result.

21. A sensing method, characterized in that: The method comprises: Sending a first signal; the first signal is used to indicate a data symbol for communication; A perceptual measurement result is received, where the perceptual measurement result is obtained based on the data symbol measurement.

22. The method according to claim 21, characterized in that Before sending the first signal, the method further includes: First indication information is received, where the first indication information is used to instruct to use data symbols to perform perception measurement.

23. The method according to claim 21 or 22, characterized in that Before sending the first signal, the method further includes: First capability information is received, where the first capability information is used to indicate support for a data symbol-based perception measurement function.

24. A perception system, characterized in that: include: A perception initiating end and a perception responding end, wherein the perception responding end is used to execute the method according to any one of claims 1 to 20, and the perception initiating end is used to execute the method according to any one of claims 21 to 23.

25. An electronic device, characterized in that: include: Signal demodulation module, perception algorithm module, channel reconstruction module and noise perception module; The electronic device is configured to receive a first signal, wherein the first signal is configured to indicate a data symbol for communication; The signal demodulation module is configured to demodulate the first signal to obtain data symbols indicated by the first signal, or to update a demodulation result after demodulating the first signal to obtain data symbols in the first signal; The sensing algorithm module is configured to sense and process the data symbols in the first signal to obtain the distance and speed of each target object in at least one target object; The channel reconstruction module is used to reconstruct a communication and environment perception channel matrix based on the distance and speed of each target object in the at least one target object, and update the reconstructed communication and environment perception channel matrix, where the updated communication and environment perception channel matrix is ​​used to update a demodulation result after demodulating the first signal; The electronic device is further configured to send a perception measurement result, where the perception measurement result includes a distance and a speed of each target object in at least one target object between a perception initiating end and a perception responding end, and the perception measurement result is obtained based on the data symbol measurement.

26. An electronic device, characterized in that: include: processor and memory; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the electronic device performs the method according to any one of claims 1 to 23.

27. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 23 is implemented.

28. A chip system, characterized in that: The system comprises at least one processor and a communication interface, wherein the communication interface and the at least one processor are interconnected via a line, and the at least one processor is used to run a computer program or instruction to execute the method according to any one of claims 1 to 23.

29. A computer program product, characterized in that The method comprises a computer program which, when being executed, causes a computer to execute the method according to any one of claims 1 to 23.

Images