A near-field symbiotic radio receiver detection method
By selecting the receiving antenna in the visible area of the target signal under near-field conditions and performing local null forcing and interference cancellation, the problems of large computational load and noise interference in full array detection are solved, thereby improving the effectiveness and efficiency of signal detection.
Patent Information
- Application Number
- CN202611302900.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-26
- Publication Date
- 2026-11-10
AI Technical Summary
Under near-field conditions, existing full-array detection methods involve large computational loads, and the antenna in the non-visible area introduces noise and interference, resulting in a decrease in signal-to-interference-plus-noise ratio and poor bit error rate performance. Furthermore, performing unnecessary zero-fork operations on weakly correlated users increases computational complexity.
By acquiring the visible area information of the target signal, selecting an effective receiving antenna, performing local null-forcing and serial interference cancellation, reducing the computational dimension, and dynamically adjusting the detection order according to significant interference relationships, prioritizing the detection of the main transmitted signal, and locally eliminating interference.
It improves the signal-to-interference-plus-noise ratio and bit error rate performance, reduces computational complexity, and maintains the robustness and efficiency of detection.
Smart Images

Figure CN123074609A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of wireless communication technology, specifically relating to a near-field symbiotic radio reception detection method. Background Technology
[0002] Symbiotic radio (SR) achieves coordinated main transmission and backscatter transmission without additional spectrum usage by having a low-power backscatter device (BD) multiplex the radio frequency carrier and spectrum resources of the primary transmitter (PT) to transmit its own information. The receiver simultaneously receives the direct signal from the PT and the backscattered signal formed by multiple BDs loading their own symbols, and needs to perform joint detection of both types of signals.
[0003] To achieve higher array gain and support a large number of low-power devices, a large-scale antenna array can be configured at the receiver. As the array aperture increases, the Fraunhofer distance increases with the square of the aperture, and the PT or BD (Pressure Point) is easily located in the near-field region with the array. At this point, different array elements have different propagation distances from the signal source, and the signal exhibits spherical wave propagation with amplitude and phase varying with the element position. The traditional far-field plane wave and spatially stationary channel assumptions no longer apply. The specific Fraunhofer distance is given by the following formula:
[0004]
[0005] in, Indicates Fraunhofer distance, Indicates the aperture of the receiving array. This represents the carrier wavelength. Due to near-field distance-dependent amplitude and phase variations, as well as factors such as obstruction, user location, and propagation environment, the effective energy of the same transmitter is often concentrated only on a portion of the array's antennas, resulting in spatial non-stationarity (SNS) and visibility region (VR).
[0006] Existing full array detection methods typically do not distinguish the visible range of the target signal on the array, directly using the entire array. Each receiving antenna performs either Maximum Ratio Combining (MRC) or Zero Forcing (ZF). Antennas outside the target VR provide very little effective energy to the target signal, yet still introduce receive noise, channel estimation errors, and other user interference. Therefore, full-array processing not only has a high computational dimensionality but can also reduce the effective signal-to-interference-plus-noise ratio and bit error rate performance of the target signal by accumulating invalid components on non-VR antennas.
[0007] Even if the ZF is confined to the target VR, if the interference subspace is still constructed uniformly for all other users, unnecessary zero-forcing will be performed on signals that only slightly overlap with or are separated from the target VR. This increases the scale of matrix operations and may project the desired signal into a subspace unrelated to the actual significant interference, resulting in the loss of useful components. Traditional successive interference cancellation (SIC), if it reconstructs and subtracts the detected signal across the entire array, will also introduce additional cancellation errors at array elements where the signal is not visible. Therefore, how to utilize the VR spatial locality of the near-field SNS channel while reducing the antenna processing dimension and the dimension of interfering users participating in zero-forcing is a problem that near-field multi-user SR receivers urgently need to solve. Summary of the Invention
[0008] This invention aims to overcome the shortcomings of existing full-array receiver detection methods under near-field SNS conditions, such as high computational complexity, accumulation of non-VR noise and interference, and unnecessary zero-forking for weakly correlated users. It provides a visible-area sensing receiver detection method for near-field co-occurring radio. This method first selects an effective receiving antenna based on the VR corresponding to the target signal, then filters out significant interference based on the overlap relationship between different VRs, and performs partial zero-forking and local serial interference cancellation within the target VR, thus balancing detection performance and computational complexity.
[0009] The technical solution of this invention is as follows:
[0010] A near-field co-occurring radio receiver detection method is applied to a multi-user co-occurring radio system comprising a main transmitter (PT), K backscattering devices (BDs), and a large-aperture receiving array with N antennas, where K is the number of BDs and N is the number of receiving antennas. The method includes the following steps:
[0011] The receiving array receives baseband signals, which include direct signals transmitted by the PT and K-channel BD backscattered equivalent signals;
[0012] The visible region (VR) information of each signal on the receiving array is obtained, the VR information indicating the subset of antennas with significant energy for each signal in the array;
[0013] Based on the VR information, an antenna selection matrix is constructed for each signal to be detected. The receive vector and local channel within the target VR are extracted from the full array receive vector, reducing the receiver processing dimension from N to the cardinality of the target VR. , For the first The set of VR antenna indices for the road signal;
[0014] Based on the receiver vector and local channel within the target VR, linear detection is performed to obtain an equivalent signal estimate;
[0015] Based on the spatial overlap relationship between the VR signals of each signal, a set of significant interferences to the target signal is selected;
[0016] Within the target VR, an interference subspace is constructed only for signals in the significant interference set, and partial zero-forcing is performed to obtain a partial zero-forcing receiver-combining vector;
[0017] Prioritize the detection of the PT main transmit signal and use the detected PT symbols to recover each BD information symbol;
[0018] After each signal detection is completed, the signal components are reconstructed and eliminated only within the corresponding VR of that signal. The significant interference relationships of the remaining signals are updated, and the detection and local elimination are repeated until all signals are detected.
[0019] Furthermore, the baseband signal reception vector is represented as:
[0020] ,
[0021] in, For the receive vector, For complex fields, This refers to the PT's transmit power. This is a direct channel from the PT to the receiving array. For BD reflectance, For PT to the scalar channels of BD For the first One BD to the receiver array channel, and PT and the 1st The information symbol of a BD, for 3D noise vector;
[0022] In near-field large-aperture arrays, the channel exhibits range-dependent amplitude and phase variations, with the same signal having significant energy only on a subset of array elements, resulting in spatial nonstationarity (SNS). The SNS channel of each BD is represented as follows:
[0023] ,
[0024] in, For the near-field channel without VR mask applied, For the corresponding VR indicator vector, In the The values are 1 for visible elements of the BD array and 0 or a smaller weight for invisible elements. This represents element-wise product.
[0025] Furthermore, specific methods for obtaining the visible area information of each signal on the receiving array include:
[0026] The VR indication vector is obtained in advance through channel measurement, energy detection, signal detection, statistical estimation, or machine learning; or
[0027] Assuming ideal VR prior conditions, VR information is considered to be precisely known; or
[0028] VR information with a certain recognition error is used as input to verify the robustness of the receiving detection.
[0029] Furthermore, the specific method for extracting the receive vector and local channel within the target VR from the full array receive vector is as follows:
[0030] The estimated equivalent channel matrix of the entire array is as follows: ,in, for Dimensional estimation of the equivalent channel matrix, For PT direct signal dimensional estimation of channel vector, For the first The equivalent backscattered signal corresponding to each BD The channel vector is estimated in dimensionality. Wherein, With respect to the aforementioned PT to the first scalar channels of BD They have different meanings.
[0031] Constructing the antenna selection matrix Through formula The receiver vector within the target VR is obtained, and the target channel and related interference channels are limited to the same antenna selection relationship. Inside, through Obtain the locally estimated equivalent channel matrix.
[0032] Furthermore, the specific method for performing linear detection based on the receive vector and local channel within the target VR is as follows:
[0033] Using VR maximum ratio merging (VR-MRC), the merge vector is:
[0034] ,
[0035] Among them, superscript This indicates the conjugate transpose. yes Middle and the first The channel column corresponding to the signal, corresponding to the first The estimated channel vector of the path signal. for 3D VR-MRC merged vector;
[0036] Alternatively, VR-ZF (VR-ZF) can be used, with the following projection matrix:
[0037] ,
[0038] in, It is the identity matrix. for 3D VR-ZF projection matrix, To remove the target column 3D interference channel matrix, This represents finding the inverse of a matrix; the result is... 3D VR-ZF merged vector :
[0039] ;
[0040] Obtain the The equivalent signal of the path is estimated as follows:
[0041] ,
[0042] For detector type, For the first Road sign quantities are merged and output.
[0043] Furthermore, the specific method for screening the significant interference set of the target signal is as follows:
[0044] According to the formula Calculate the first The signal of the path relative to the first Normalized VR overlap of road target signals, setting VR overlap threshold. , will satisfy And the signal index that has not yet been eliminated Classified as Significant interference set of road target signals , This is a set of signal indexes that includes PT and all BD.
[0045] Furthermore, it also includes:
[0046] Each equivalent signal to be detected is taken as a signal node, and each target VR is taken as a VR node. When the first... The signal belongs to the first Significant interference set of road target signals At that time, in the Road signal node and the first Edges are established between VR nodes of the road target to form a bipartite graph representing significant interference relationships;
[0047] The bipartite graph is used to manage the set of significant interferences, determine the BD detection order, and update the remaining interference relationships after local serial interference elimination.
[0048] Furthermore, the partial zero-fork performed only on signals within the significant interference set within the target VR includes:
[0049] Define the set of indices of signals that have not yet been detected and eliminated. , for the Road signal, only China still belongs to And not for Its own signal, as a residual significant interference, has the following quantity: ;
[0050] when At that time, construct the remaining significant interference channel matrix. Calculate the partial zero-forced projection matrix The subscript I represents the set of interfering channels, and a PZF receive combining vector is generated:
[0051] ,
[0052] in, for Dimensional PZF-SIC merged vector;
[0053] when At that time, the zero-force step is skipped, and the process degenerates to performing maximum ratio merging (VR-MRC) within the target VR.
[0054] Furthermore, the specific method for prioritizing the detection of the PT main transmit signal and recovering each BD information symbol using the detected PT symbols is as follows:
[0055] Initialize the untested set First of all, let And in Internal detection PT main information symbol:
[0056] ,
[0057] in, The output is a scalar combination of the PT equivalent signal. A collection of constellations;
[0058] For BD equivalent signal The BD signal is recovered using the decision result of the PT main information symbol:
[0059] ,
[0060] in, For the first The decision result for each BD symbol. Candidate symbols for constellations;
[0061] The specific methods of repeated detection and local elimination are as follows:
[0062] From the undetected BDs, the BD corresponding to the target VR with the smallest remaining interference degree is selected as the next signal to be detected:
[0063] ,
[0064] in, The index that has the fewest significant residual interferences in the currently undetected BD;
[0065] After the k-th signal is detected, the signal component is reconstructed and eliminated only on the antenna corresponding to its VR.
[0066] ,
[0067] in, for 3D VR mask vector, For the first Equivalent signal of road sign quantity This indicates an assignment update; Represents the set difference. For the updated version The number of remaining significant interferences in the road signal; from Delete the signal path, and delete the corresponding signal node and its associated edge in the bipartite graph. Update the remaining interference degree of the remaining target VR nodes until all signals have been detected.
[0068] A computer-readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the method of the present invention.
[0069] The beneficial effects of this invention are as follows:
[0070] 1) By utilizing the spatial locality of VR in the near-field SNS channel, only antennas where the target signal is visible are selected for detection, thus avoiding the accumulation of noise, channel estimation errors, and multi-user interference on the VR external antennas into target detection, which can improve the effective signal-to-interference-plus-noise ratio and bit error rate performance.
[0071] 2) The receiving and processing dimension of each signal is changed from the entire array. Limited to target VR This reduces the scale of operations such as merging, channel matrix construction, and matrix inversion.
[0072] 3) Within the target VR, partial zero-forcing is performed only on interferences with significant VR overlap that have not yet been eliminated, further reducing the number of users involved in zero-forcing and avoiding the loss of useful signals caused by unnecessary projection of weakly correlated or spatially separated signals; it can adaptively degenerate to VR-MRC when there is no significant interference.
[0073] 4) By combining priority detection of the main transmitted signal, dynamic determination of the backscatter signal order based on the number of remaining significant interferences, and execution of local SIC only within the corresponding VR, the set of subsequent interferences can be continuously reduced, and good detection robustness can be maintained even when there are certain errors in the VR information. Attached Figure Description
[0074] Figure 1 This is a schematic diagram of the multi-user near-field co-occurrence radio system of the present invention and the visible area formed by each signal on a large aperture receiving array;
[0075] Figure 2 This is a comparison chart of the main transmission bit error rate performance of the full array detection, visible region-based detection, and PZF-SIC detection based on visible region overlap relationship used in this invention under different received signal-to-noise ratios.
[0076] Figure 3 This is a comparison chart of the average bit error rate performance of multi-backscattering devices under different received signal-to-noise ratios for the full array detection, visible region-based detection, and PZF-SIC detection based on visible region overlap relationships used in this invention.
[0077] Figure 4 A schematic diagram showing the visible regions and spatial overlap of the main transmitted signal and multiple backscattered equivalent signals on the receiving array;
[0078] Figure 5 This is a schematic diagram of the PZF-SIC dynamic detection, local elimination, and interference relationship update process based on the visible region overlap relationship. Detailed Implementation
[0079] The present invention will now be described in detail with reference to the accompanying drawings and simulation examples.
[0080] The method of the present invention can be logically summarized into the following steps:
[0081] Step 1: Construct a near-field multi-user co-occurring radio receiving system and a spatial non-stationary channel model; the main transmitter sends a main signal, each backscattering device uses the main signal and loads its own backscattering symbol, and the receiving array simultaneously receives the main transmitted direct signal and multiple backscattered signals.
[0082] Step 2: Obtain the visible area information and corresponding channel information of the main transmitted signal and each backscattered signal on the receiving array; the visible area information can be obtained in advance by channel measurement, signal detection, statistical estimation, machine learning or other methods, and the present invention does not limit the specific acquisition method; the channel information can be obtained by least squares estimation or other existing channel estimation methods.
[0083] Step 3: Let For the first VR antenna index set for road signals, Given its base, construct the antenna selection matrix. , in order to from the whole Extract the receive vector, target channel, and related interference channels within the target VR from the antenna array.
[0084] Step 4: Perform linear detection based on the receive vector and local channel within the target VR; the linear detection can employ VR-MRC, VR-ZF, or other receive combining methods constructed on a selected subset of antennas, reducing the processing dimension from... Reduce to .
[0085] Step 5: Calculate the VR overlap degree of different signals to be detected, and identify the signals whose VR overlap degree reaches the preset or adjustable threshold and have not been eliminated as significant interference signals of the target signal, forming a significant interference set.
[0086] Step 6: Within the target VR, construct the interference channel matrix, the interference subspace orthogonal projection matrix, and the partial zero-forced receiver combining vector based only on the set of significant interferences; when there are no significant interferences that have not yet been eliminated within the target VR, the partial zero-forced detection degenerates into maximum ratio combining within the target VR.
[0087] Step 7: First, detect the main transmitted symbol within the VR corresponding to the main transmitted signal; then, among the undetected backscattered signals, prioritize the backscattered signals with fewer significant interferences to detect, so as to recover the corresponding backscattered symbol using the detected main transmitted symbol.
[0088] Step 8: After each signal detection is completed, reconstruct and eliminate its signal components only in the VR corresponding to that signal, delete the signal from the undetected set, update the significant interference relationship and the number of significant interferences among the remaining signals, and repeat the detection, local elimination and update until all target signals are detected.
[0089] Example:
[0090] This embodiment considers, for example Figure 1 The multi-user SR system shown includes a single-antenna PT. A single antenna BD and a configuration A large-aperture receiver array with multiple antennas. PT transmits main information symbols. , No. Each BD backscatters the incident PT signal by adjusting its reflection coefficient, and loads its own information symbols onto the reflected signal. The receiver receives the PT direct signal and The equivalent backscattered signal from the BD path is jointly received.
[0091] The following is a detailed description of this example, combining the steps outlined above:
[0092] In step 1, a near-field non-uniform spherical wave channel capable of reflecting spherical wave propagation is established. (The text abruptly ends here.) Taking a channel from a BD to a receiving array as an example, it includes both line-of-sight and non-line-of-sight components:
[0093]
[0094] in, For the near-field channel without VR mask applied, and These are its LoS and NLoS components, respectively, and all three have the same dimension. ; For component weights.
[0095] make For the receiving antenna index, the first of the line-of-sight components The coefficients of each array element are expressed using near-field spherical waves as follows:
[0096]
[0097] in, and The first The receiving antenna and the first The location of BD For transmission distance, It is the Euclidean norm. Pi It is a natural constant. It is the imaginary unit.
[0098] To describe spatial nonstationarity, the near-field channel is multiplied element-wise by the VR indicator vector:
[0099]
[0100] in, For the SNS channel after applying VR mask, The VR indicator vector has a dimension of 1. ; For the first A set of VR antenna indices for each BD. This represents element-wise product. For those belonging to... antenna, For those that do not belong to antenna, Alternatively, a smaller weight can be used to represent the weakly visible state. The PT direct link can be represented similarly as follows:
[0101]
[0102] in, , and These represent the PT direct channel with VR mask applied, the PT near-field direct channel without mask applied, and their VR indication vectors, respectively, each with dimension [missing information]. .
[0103] The baseband vector received by the receiving array at one symbol time is:
[0104]
[0105] in, For complex numbers; For the receive vector, for 3D noise vector; This refers to the PT's transmit power. For BD reflectance, For PT to the scalar channels of BD The symbol is PT. For the first A BD symbol, The corresponding multiplicative coupling equivalent signal; the summation range is... .
[0106] For ease of consistent description, the received signal is written in matrix form:
[0107]
[0108]
[0109]
[0110] in, The equivalent channel matrix for the entire array. The SR equivalent signal vector, superscript Indicates transpose; cursive set This is the set of all signal indices.
[0111] In step 2, it is assumed that the receiver has obtained the PT and the VR information corresponding to each BD, and performs subsequent channel estimation and reception detection accordingly. Ideally, the VR information can be considered as precisely known to characterize the ideal VR prior conditions; in practical systems, the VR information can be obtained through energy detection, signal measurement, statistical estimation, machine learning, or other existing methods. This invention does not limit the specific acquisition method of the VR information. For channel information, the receiver uses the least squares method to estimate the direct link of the PT and the backscatter equivalent link corresponding to each BD, and constructs the full array equivalent channel matrix from the obtained channel estimation results. Furthermore, to account for actual VR acquisition errors, subsequent simulation examples further use VR information with a certain recognition error, such as a VR indication with approximately 90% accuracy, to verify the robustness of the reception detection method to imperfect VR information. Let the estimated full array equivalent channel matrix be:
[0112]
[0113] in, for Dimensional estimation of the equivalent channel matrix, and PT and the 1st Road BD The channel vector is estimated using a dimensional method; the cap symbol indicates the estimator.
[0114] In steps 3 and 4, let For the first VR antenna index set for road signals, Represent subset relations; define the antenna selection matrix. .pass Extract the receive vector and local channel within the target VR:
[0115]
[0116]
[0117]
[0118] in, , and These are the received vectors within the target VR, and the... The path estimation channel vector and the local estimation equivalent channel matrix have dimensions of [dimensions to be filled in]. , and .
[0119] One basic implementation uses VR-MRC, whose receive merging vector is:
[0120]
[0121] Among them, superscript Indicates conjugate transpose; for Dimensional VR-MRC merged vector.
[0122] Another basic implementation uses VR-ZF. Let To remove the target column If the interference channel matrix is 3D, then its orthogonal projection matrix is expressed as:
[0123]
[0124]
[0125] in, It is the identity matrix. for 3D VR-ZF projection matrix, This represents finding the inverse of a matrix. for Dimensional VR-ZF merged vector. After VR linear merging, the th... The equivalent signal of the path is estimated as follows:
[0126]
[0127] VR-MRC and VR-ZF are two implementations of the fundamental technical concept of "selecting an effective subset of antennas based on the target VR and performing detection on the selected subset of antennas." For detector type, For the first Road sign quantities are merged and output.
[0128] The PT main information symbol is detected using the following minimum distance criterion:
[0129]
[0130] Among them, the floral collection A collection of constellations. This indicates a minimum distance search. Here, we denote the modulus of a complex number. The decision result is for the PT symbol. Because of the... The backscattering equivalent current corresponding to each BD is: The BD symbol needs to be used Decision after removing the PT main information symbol:
[0131]
[0132] in, For the first The decision result for each BD symbol. It is a candidate symbol for the constellation.
[0133] In step 5, according to Figure 4 The spatial overlap relationship of different signals VR is shown, and the normalized VR overlap degree is calculated. And based on the threshold Form a significant interference set Significant overlap indicates that another signal has a sufficiently large visible range within the target VR, thus requiring active suppression; weakly overlapping or non-overlapping signals are not included in this partial zero-forcing.
[0134] like Figure 4 As shown, the PT main transmit signal and three BD equivalent signals , and Different waveforms (VRs) are formed on the array, with some VRs significantly overlapping and others having only minor overlap or being separate from each other. Therefore, uniformly performing ZF on all other signals within each target VR is neither necessary nor efficient, and would increase computation and lose the desired signal energy.
[0135] In step 6, let the flower body assembly... This is the set of indices for signals that have not yet been detected and eliminated. ; Define the first The number of remaining significant interferences in the road signal is:
[0136]
[0137] in, The remaining significant interference quantity, and These represent the difference and intersection of sets, respectively. for The remaining significant interference channel matrix is given by , where the subscript I denotes the set of interference channels. The partially zero-forked projection matrix is:
[0138]
[0139] in, for The partial zero-forced projection matrix. The corresponding PZF receiver merging vector is:
[0140]
[0141] in, for Dimensional PZF-SIC merged vector. When When performing partial zero-force; season It degenerates into VR-MRC.
[0142] In steps 7 and 8, initialization is performed. First of all, let And in Internal detection of PT main information symbols. After PT detection is completed, the signal with fewer significant interferences in the remaining BD is selected as the next signal to be detected:
[0143]
[0144] in, This is the index of the element with the fewest significant residual interferences in the currently undetected BD.
[0145] For the After the path signal is decided, the path signal component is reconstructed and eliminated only on the antenna corresponding to its VR.
[0146]
[0147] in, for 3D VR mask vector, For the first Equivalent signal of road sign quantity This indicates an assignment update.
[0148] Figure 5 A bipartite graph implementation of the above dynamic process is presented. Figure 5 (a) in the text indicates that first in VR- Internal main transmission signal Perform PZF detection and local elimination; Figure 5 (b) in the text indicates Since there is no significant residual interference within the corresponding VR, VR-MRC detection and local elimination are adopted. Figure 5 (in) ) indicates in VR- The inner pairs still have significant overlap Detection after PZF inhibition ; Figure 5 (d) in the text indicates that only one remains. VR-MRC was used to complete the detection. The remaining edge relations and local degrees were updated at each stage until all BD detections were completed.
[0149] Complexity analysis:
[0150] For the road signal, Indicates asymptotic complexity; the complexities of full array MRC and VR-MRC are respectively and Full array ZF and VR-ZF are respectively and The complexity of PZF-SIC is... related:
[0151]
[0152] in, To detect the first The PZF-SIC complexity of the path signal. Therefore, this invention first reduces the antenna dimension from Reduce to Then, the zero-forced interference dimension is changed from Reduce to .when At that time, the complexity is For example, in , , and At that time, the magnitudes of full array ZF, VR-ZF, and PZF-SIC were 450261, 97461, and 21000, respectively; these values are only examples to illustrate the complexity.
[0153] Simulation example:
[0154] This embodiment uses the simulation conditions described in the paper to verify the receiver detection performance. Unless otherwise specified, the number of receiver array antennas... BD quantity Each link employs a near-field SNS channel, and both PT and BD use QPSK modulation. Benchmark schemes include full-array MRC and full-array ZF without utilizing VR information, and VR-MRC and VR-ZF that only limit linear detection to the target VR; a further scheme is the PZF-SIC of this invention. To examine robustness to imperfect VR input, a VR indication with approximately 90% accuracy is set as a control condition; this accuracy is only a simulation input parameter and does not limit the VR acquisition method.
[0155] like Figure 2 and Figure 3 As shown, the horizontal axis of both graphs represents the received signal-to-noise ratio (SNR), and the vertical axis represents the bit error rate (BER). Among them, Figure 2 Give the main transmission bit error rate. Figure 3 The average bit error rate (BER) of multiple backscattering devices is given. Compared to full array processing, VR-MRC and VR-ZF have lower overall BERs at the same receive SNR, indicating that ignoring the antennas that mainly carry noise and interference outside the target VR can improve detection performance; the corresponding gain in this simulation is approximately 3 dB.
[0156] The bit error rate performance of PZF-SIC is close to that of VR-ZF, while only performing zero-forcing on significantly overlapping interference and avoiding cancellation on non-VR antennas, thus resulting in lower computational scale. Even with a VR indication accuracy of approximately 90%, the performance ranking of various VR sensing schemes remains largely unchanged and still outperforms the corresponding full-array benchmark, indicating that this invention is robust to a certain degree of VR input error. For BD transmission, since its symbol recovery depends on... The PT detection error will be added to the BD detection, so the average bit error rate of multiple BDs is usually higher than the main transmission bit error rate, which is consistent with the aforementioned SR multiplicative coupling model.
[0157] In summary, this invention uses the VR formed by the near-field SNS channel as the basis for spatial processing, forming a progressive reception mechanism from full array processing to VR subarray processing, and then to VR overlapping sensing PZF-SIC. This mechanism does not require a specific VR learning model, nor does it limit the channel estimation method, threshold value, array size, or bipartite graph data structure, and can achieve synergistic improvement in performance and complexity while maintaining the SR multiplicative signal model.
Claims
1. A near-field co-occurring radio receiver detection method, applied to a multi-user co-occurring radio system comprising a main transmitter (PT), K backscattering devices (BDs), and a large-aperture receiving array configured with N antennas, wherein K is the number of BDs and N is the number of receiving antennas; characterized in that, Includes the following steps: The receiving array receives baseband signals, which include direct signals transmitted by the PT and K-channel BD backscattered equivalent signals; The visible region (VR) information of each signal on the receiving array is obtained, the VR information indicating the subset of antennas with significant energy for each signal in the array; Based on the VR information, an antenna selection matrix is constructed for each signal to be detected. The receive vector and local channel within the target VR are extracted from the full array receive vector, reducing the receiver processing dimension from N to the cardinality of the target VR. , For the first The set of VR antenna indices for the road signal; Based on the receiver vector and local channel within the target VR, linear detection is performed to obtain an equivalent signal estimate. Based on the spatial overlap relationship between the VR signals of each signal, a set of significant interferences to the target signal is selected; Within the target VR, an interference subspace is constructed only for signals in the significant interference set, and partial zero-forcing is performed to obtain a partial zero-forcing receiver-combining vector; Prioritize the detection of the PT main transmit signal and use the detected PT symbols to recover each BD information symbol; After each signal detection is completed, the signal components are reconstructed and eliminated only within the corresponding VR of that signal. The significant interference relationships of the remaining signals are updated, and the detection and local elimination are repeated until all signals are detected.
2. The near-field co-occurrence radio reception detection method according to claim 1, characterized in that, The baseband signal reception vector is represented as: , in, For the receive vector, For complex fields, This refers to the PT's transmit power. This is a direct channel from the PT to the receiving array. For BD reflectance, For PT to the scalar channels of BD For the first One BD to the receiving array channel, and PT and the 1st The information symbol of a BD, for 3D noise vector; In near-field large-aperture arrays, the channel exhibits range-dependent amplitude and phase variations, with the same signal having significant energy only on a subset of array elements, resulting in spatial nonstationarity (SNS). The SNS channel of each BD is represented as follows: , in, For the near-field channel without VR mask applied, For the corresponding VR indicator vector, In the The values are 1 for visible elements of the BD array and 0 or a smaller weight for invisible elements. This represents element-wise product.
3. The near-field symbiotic radio reception detection method according to claim 1, characterized in that, Specific methods for obtaining the visible area information of each signal on the receiving array include: The VR indication vector is obtained in advance through channel measurement, energy detection, signal detection, statistical estimation, or machine learning; or Assuming ideal VR prior conditions, VR information is considered to be precisely known; or VR information with a certain recognition error is used as input to verify the robustness of the receiving detection.
4. The near-field co-occurrence radio reception detection method according to claim 1, characterized in that, The specific method for extracting the receiver vector and local channel within the target VR from the full array receiver vector is as follows: The estimated equivalent channel matrix of the entire array is as follows: ,in, for Dimensional estimation of the equivalent channel matrix, For PT direct signal dimensional estimation of channel vector, For the first The equivalent backscattered signal corresponding to each BD Dimensional estimation of channel vector; Constructing the antenna selection matrix Through formula The receiver vector within the target VR is obtained, and the target channel and related interference channels are limited using the same antenna selection relationship. Inside, through Obtain the locally estimated equivalent channel matrix.
5. The near-field co-occurrence radio reception detection method according to claim 4, characterized in that, The specific method for performing linear detection based on the receiver vector and local channel within the target VR is as follows: Using VR maximum ratio merging (VR-MRC), the merge vector is: , Among them, superscript This indicates the conjugate transpose. yes The Middle The element corresponding to the element The estimated channel vector of the path signal. for 3D VR-MRC merged vector; Alternatively, VR-ZF (VR-ZF) can be used, with the following projection matrix: , in, It is the identity matrix. for 3D VR-ZF projection matrix, To remove the target column 3D interference channel matrix, This represents finding the inverse of a matrix; the result is... 3D VR-ZF merged vector : ; Obtain the The equivalent signal of the path is estimated as follows: , For detector type, For the first Road sign quantities are merged and output.
6. The near-field co-occurrence radio reception detection method according to claim 5, characterized in that, The specific method for filtering the set of significant interferences to the target signal is as follows: According to the formula Calculate the first The signal of the path relative to the first Normalized VR overlap of road target signals, setting VR overlap threshold. , will satisfy And the signal index that has not yet been eliminated Classified as Significant interference set of road target signals , This is a set of signal indexes that includes PT and all BD.
7. The near-field co-occurrence radio reception detection method according to claim 6, characterized in that, Also includes: Each equivalent signal to be detected is taken as a signal node, and each target VR is taken as a VR node. When the first... The signal belongs to the first Significant interference set of road target signals At that time, in the Road signal node and the first Edges are established between VR nodes of the road target to form a bipartite graph representing significant interference relationships; The bipartite graph is used to manage the set of significant interferences, determine the BD detection order, and update the remaining interference relationships after local serial interference elimination.
8. The near-field co-occurrence radio reception detection method according to claim 7, characterized in that, The partial zero-fork performed within the target VR on signals only from the set of significant interferences includes: Define the set of indices of signals that have not yet been detected and eliminated. , for the Road signal, only China still belongs to And not for Its own signal, as a residual significant interference, has the following quantity: ; when At that time, construct the remaining significant interference channel matrix. Calculate the partial zero-forced projection matrix The subscript I represents the set of interfering channels, and a PZF receive combining vector is generated: , in, for Dimensional PZF-SIC merged vector; when At that time, the zero-force step is skipped, and the process degenerates to performing maximum ratio merging (VR-MRC) within the target VR.
9. The near-field co-occurrence radio reception detection method according to claim 8, characterized in that, The specific method for prioritizing the detection of the PT main transmit signal and recovering each BD information symbol using the detected PT symbols is as follows: Initialize the untested set First of all, let And in Internal detection PT main information symbol: , in, The output is a scalar combination of the PT equivalent signal. A collection of constellations; For BD equivalent signal The BD signal is recovered using the decision result of the PT main information symbol: , in, For the first The decision result for each BD symbol. Candidate symbols for constellations; The specific methods of repeated detection and local elimination are as follows: From the undetected BDs, the BD corresponding to the target VR with the smallest remaining interference degree is selected as the next signal to be detected: , in, The index that has the fewest significant residual interferences in the currently undetected BD; After the k-th signal is detected, the signal component is reconstructed and eliminated only on the antenna corresponding to its VR. , in, for 3D VR mask vector, For the first Equivalent signal of road sign quantity Indicates assignment and update; from Delete the signal path, and delete the corresponding signal node and its associated edge in the bipartite graph. Update the remaining interference degree of the remaining target VR nodes until all signals have been detected.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method of any one of claims 1 to 9.
