A radar communication dual-function system-oriented active and passive cooperative perception fusion method

Abstract

The application discloses a kind of active and passive cooperative perception fusion methods for radar communication dual-function system, including 1, embedding special perception symbol in the signal sent by base station;2, power distribution is solved using optimization problem;3, passive perception signal is received by receiving access point, direct path interference is handled using whitening filter, so that it becomes white noise;4, receiving access point uses generalized likelihood ratio test detector to judge whether target exists, then receiving access point and base station transmit binary decision result and target detection probability to fusion center;5, fusion center aggregates by voting the information received, judges whether target exists.The application can improve the accuracy and efficiency of target detection, positioning and environment information acquisition and reconstruction, and may significantly improve the performance of wireless communication perception network.

Description

Technical Field

[0001] This invention belongs to the field of radar power allocation, and in particular relates to an active and passive cooperative sensing fusion method for radar communication dual-function systems. Background Technology

[0002] Communication systems are evolving from 5G to 6G, presenting unique challenges for networks aiming for global coverage, green intelligence, interconnected sensing, and converged communication. To fully realize the potential of 6G, it is necessary to acquire environmental sensing information, achieve seamless information interaction and sharing, and intelligently control information processing. Therefore, Dual-Function Radar Communication (DFRC) has become a key technology for 6G, with researchers focusing on integrating communication systems and radar sensors. However, some problems exist in DFRC research. Firstly, the fading and interference characteristics of wireless communication systems lead to sensing uncertainties, meaning the data received by a single radar may not be perfect. Secondly, DFRC research primarily focuses on waveform design and signal processing, rarely considering multi-static sensing capabilities, resulting in a waste of sensing resources. Summary of the Invention

[0003] Purpose of the invention: The purpose of this invention is to provide an active-passive cooperative sensing fusion method for dual-function radar communication systems, which integrates active and passive sensing in the DFRC system to compensate for the waste of sensing resources.

[0004] Technical Solution: This invention provides an active-passive cooperative sensing fusion method for a dual-function radar communication system. The dual-function radar communication system includes a dual-function radar communication base station (BS) equipped with a fusion center (FC) and receiving access points (RAPs). The BS performs active sensing while communicating with users, while the RAPs perform passive sensing and are subject to direct path interference from the signals transmitted by the BS. The RAPs and the BS are connected via a backhaul link, but the link capacity is limited and does not support the transmission of large amounts of data. The method includes the following steps:

[0005] Step 1: Embed dedicated sensing symbols into the signals transmitted by the dual-function radar communication base station BS and pre-encode them using a zero-forcing radar beamformer;

[0006] Step 2: Model the optimization problem as maximizing the average target detection probability while ensuring that the total transmission power does not exceed the budget of the dual-function radar communication base station (BS) and that the user's signal-to-interference-plus-noise ratio (SINR) meets the minimum requirements. Solve this problem using a power allocation algorithm based on ergonomics.

[0007] Step 3: The receiving access point RAPs receives the passive sensing signal from the dual-function radar communication base station BS to the target and then to the receiving access point RAPs. The direct path interference in the passive sensing signal is processed by the whitening filter and turned into white noise.

[0008] Step 4: The receiving access point RAPs uses the generalized likelihood ratio test to determine whether the GLRT detector exists. The receiving access point RAPs and the dual-function radar communication base station BS transmit the judgment result and target detection probability to the fusion center FC.

[0009] Step 5: The Fusion Center (FC) aggregates the received information through voting to determine whether the target exists.

[0010] Further, step 1 specifically involves: using the dedicated sensing symbol s0, precoding is performed using a zero-forcing radar beamformer, and the precoding vector is...

[0011]

[0012] Where H = [h1, h2, ..., h K ] H Let θ be the communication channel matrix, a(θ) be the transmission steering vector, θ be the azimuth angle of the target relative to BS, and the operator |||| represents the 2-norm operation.

[0013] Furthermore, the optimization problem described in step 2 is as follows:

[0014]

[0015] in, These are the target detection probabilities of BS and RAPs, where R is the number of RAPs. It is a non-central chi-square cumulative distribution function with 2 degrees of freedom, ρ r The non-centrality parameters for BS and RAPs can be given by the generalized likelihood ratio test detector used, ξ. r These are the thresholds for the BS and RAP detectors. It is the SINR of the i-th user. It is a normalized communication transmission precoding vector, where λ is the regularization parameter, and p = [p0, p1, ..., p]. K [Power allocation for sensor symbols and communication symbols] Let P be the variance of additive white Gaussian noise (AWGN), K be the number of users, Γ be the required SINR threshold, and P be the variance of the additive white Gaussian noise (AWGN). T This is the power budget for BS.

[0016] Furthermore, in step 2, the power allocation algorithm based on traversal algorithm specifically includes:

[0017] Step 2.1: When the power p0 assigned to the sensor symbol gradually increases and satisfies the BS power budget, the joint detection probability... Gradually increase, initialization: p′0 = P T ,Δp,p sum =p′0, where p′0 is the initialization parameter set, Δp is the step size, and p sum As a condition for judgment;

[0018] Step 2.2: Set p0 = p′0 - Δp;

[0019] Step 2.3: Use the CVX toolkit to solve the convex optimization problem min||p||1s.t.γ k ≥Γ, where ||||1 is a norm 1;

[0020] Step 2.4, set p sum =||p||1'p′0=p0;

[0021] Step 2.5, when p sum ≥P T Proceed to step 2.2; otherwise, output p.

[0022] Furthermore, step 3 specifically involves:

[0023] When a target is present, the passive sensing signal received by the r-th RAP is:

[0024]

[0025] Where, n′ r [l] represents the AWGN matrix, α r It is the combined sensing channel gain. It is the turning vector of the receiving antenna. Let X = WS be the azimuth angle of the target relative to the r-th RAP, and let WS be the DFRC signal matrix. S = [s0, s1, ..., s K ], s i Let i = 1, 2, ..., K be the communication symbols of the i-th user, and G be the communication symbols of the i-th user. r This represents the targetless channel between the BS and the r-th RAP in the absence of a target. The signal is obtained by passing through a matched filter and vectorizing:

[0026]

[0027] in W = WW H ,vec() represents the vectorization operator, ε r It follows a zero-mean, complex Gaussian distribution and has a block covariance matrix of:

[0028]

[0029] in, The whitening filter is obtained as Depend on Obtained by disassembling Cherosky.

[0030] Furthermore, step 4 specifically involves:

[0031] The binary assumptions of the GLRT detector after using the whitening filter are:

[0032]

[0033] in The corresponding GLRT detector is then given by the following formula:

[0034]

[0035] and The joint probability density function of the received signals is used to calculate the test statistic, which is:

[0036]

[0037] in, Then, a binary decision is made based on (8).

[0038] Furthermore, step 5 specifically involves:

[0039] FC makes the following binary assumptions

[0040]

[0041] Where D r D represents the binary reasoning results of BS and the i-th RAP, respectively. r =0 indicates no target, D r =1 indicates the existence of a target, and n represents the voting threshold, which is given by the following formula:

[0042]

[0043] in,

[0044]

[0045] P FAr Let represent the false alarm probabilities of BS and RAPs.

[0046] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: Compared with the prior art, the present invention integrates active and passive sensing in the DFRC system, improving the accuracy and efficiency of target detection. Its sensing performance is significantly higher than that of the DFRC system that only uses passive sensing. Furthermore, the binary decision result voting aggregation scheme solves the problem of huge overhead caused by directly sending sensing signals to the fusion center, and may significantly improve the performance of wireless communication sensing networks. Attached Figure Description

[0047] Figure 1 This is a model diagram of the DFRC system of the present invention.

[0048] Figure 2 This is a schematic diagram of the method flow of the present invention.

[0049] Figure 3 The graph showing the relationship between the average detection probability and the number of RAPs provided in this embodiment of the invention. Detailed Implementation

[0050] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0051] The technical problem to be solved by the embodiments of the present invention is to provide an active and passive cooperative sensing fusion method for radar communication dual-function systems. This method integrates active and passive sensing in the DFRC system to remedy the waste of sensing resources.

[0052] like Figure 1 As shown, the MIMO DFRC BS described in this example is equipped with a fusion center (FC) that communicates with users while performing active sensing; the receiving access points (RAPs) perform passive sensing; the RAPs and the BS are connected via a backhaul link, but the link capacity is limited and does not support the transmission of large amounts of data.

[0053] like Figure 2 The flowchart shown is a method for active and passive cooperative sensing fusion in a radar-communication dual-function system, provided by an embodiment of the present invention. The method includes the following steps:

[0054] Step 1: Embed dedicated sensing symbols into the signals transmitted by the BS and pre-encode them using a zero-forcing radar beamformer;

[0055] Step 2: Model the optimization problem as maximizing the average target detection probability while ensuring that the total transmission power does not exceed the BS budget and that the user's signal-to-interference-plus-noise ratio (SINR) meets the minimum requirements. Solve this problem using a power allocation algorithm based on ergonomics.

[0056] Step 3: RAPs receive passive sensing signals and use whitening filters to process direct path interference (DPI) into white noise;

[0057] Step 4: RAPs use the generalized likelihood ratio test (GLRT) detector to determine whether the target exists. Then, RAPs and BS transmit the binary decision result and the target detection probability to FC.

[0058] Step 5: The FC aggregates the received information through voting to determine whether the target exists.

[0059] In this embodiment, the dedicated sensing symbol in step one is s0, which is pre-coded using a zero-forcing radar beamformer, and the pre-coding vector is...

[0060]

[0061] Where H = [h1, h2, ..., h K ] H Let θ be the communication channel matrix, a(θ) be the transmission steering vector, θ be the azimuth angle of the target relative to BS, and the operator |||| represents the 2-norm operation.

[0062] In this embodiment, the optimization problem in step two is as follows:

[0063]

[0064] in, These are the target detection probabilities of BS and RAPs, where R is the number of RAPs. It is a non-central chi-square cumulative distribution function with 2 degrees of freedom, ρ r The non-centrality parameters of BS and RAPs can be given by the generalized likelihood ratio test detector used, ξ. r These are the thresholds for the BS and RAP detectors. It is the SINR of the i-th user. It is a normalized communication transmission precoding vector, where λ is the regularization parameter, and p = [p0, p1, ..., p]. K [Power allocation for sensor symbols and communication symbols] Let P be the variance of additive white Gaussian noise (AWGN), K be the number of users, Γ be the required SINR threshold, and P be the variance of the additive white Gaussian noise (AWGN). T This is the power budget for BS.

[0065] In the embodiment, as p0 gradually increases and meets the BS power budget, It increases gradually. Therefore, the power allocation algorithm based on the traversal algorithm in step two is as follows.

[0066] (1) Initialization: p′0=PT ,Δp,p sum =p′0;

[0067] (2) Set p0 = p′0 - Δp;

[0068] (3) Using the CVX toolkit to solve the convex optimization problem min||p||1s.t.γ k ≥Γ;

[0069] (4) Set p sum =||p||1, p′0 = p0;

[0070] (5) When p sum ≥P T Go to (2), otherwise output p;

[0071] In the embodiment, in step three, when a target is present, the passive sensing signal received by the r-th RAP is:

[0072]

[0073] Where, n′ r [l] represents the AWGN matrix, α r It is the combined sensing channel gain. It is the turning vector of the receiving antenna. Let X = WS be the azimuth angle of the target relative to the r-th RAP, and let WS be the DFRC signal matrix. S = [s0, s1, ..., s K ], s i Let i = 1, 2, ..., K be the communication symbols of the i-th user, and G be the communication symbols of the i-th user. r This represents the targetless channel between the BS and the r-th RAP in the absence of a target. The signal is obtained by passing through a matched filter and vectorizing.

[0074]

[0075] in W = WW H ,vec() represents the vectorization operator, ε r It is a zero-mean, complex Gaussian distribution with a block covariance matrix of...

[0076]

[0077] in, Therefore, the whitening filter is obtained as follows: Depend on Obtained by disassembling Cherosky.

[0078] In the embodiment, the binary assumption of the GLRT detector after the whitening filter is:

[0079]

[0080] in The corresponding GLRT detector is then given by the following formula.

[0081]

[0082] and It is the joint probability density function of the received signal. After calculation, the test statistic is obtained as follows:

[0083]

[0084] in, Then, a binary decision is made based on (8).

[0085] In the embodiment, in step five, FC makes the following binary assumptions.

[0086]

[0087] Where D r D represents the binary reasoning results of BS and the i-th RAP, respectively. r =0 indicates no target, D r =1 indicates the existence of a target, and n represents the voting threshold. n can be given by the following formula.

[0088]

[0089] in,

[0090]

[0091] P FAr Let represent the false alarm probabilities of BS and RAPs.

[0092] In this embodiment, we prepared 10,000 samples for simulation. The DFRC system parameter configuration is as follows:

[0093] Number of BS transmitting antennas 16 Number of BS receiving antennas 20 Number of RAP receiving antennas 20 L 30 Number of users 9 Base station coverage 500m Path loss model (dB) 128.1 + 37.6 log10(w), where w (km) is the distance between the user and the base station. Base station budget 1W False alarm probability 0.00001 (used to calculate the detector threshold)

[0094] This example is a special case of an embodiment of the present invention, but it can be extended to other similar situations.

[0095] Figure 3The figure shows the relationship between the average detection probability and the number of RAPs. We observed the change in average detection probability when the combined sensing channel gain variance was -37 / -36 / -35 dB (-37 dB, -36 dB, and -35 dB in the example). When the combined sensing channel gain variance was greater than or equal to -36 dB, the average detection probability increased with the increase of RAPs. However, the growth rate was much greater when the combined sensing channel gain variance was -35 dB than when it was -36 dB. Furthermore, we found that when the combined sensing channel gain variance was -37 dB, the average detection probability hardly increased with the increase of RAPs, always tending towards 0. This is because the binary inference result of FC is obtained through voting aggregation in a scenario with limited backhaul capacity, which focuses on the quality of the sensed signal of each individual RAP. Therefore, as the combined sensing channel gain variance increases, the detection probability of a single RAP will increase. And as the number of RAPs increases, the probability of misclassification after voting aggregation will further decrease.

[0096] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for active and passive cooperative sensing fusion in a radar-communication dual-function system, characterized in that, The dual-function radar communication system includes a dual-function radar communication base station (BS) equipped with a fusion center (FC) and receiving access points (RAPs). The BS performs active sensing while communicating with users, while the RAPs perform passive sensing and are subject to direct path interference from the signals transmitted by the BS. The RAPs are connected to the BS via a backhaul link, including the following steps: Step 1: Embed dedicated sensing symbols into the signals transmitted by the dual-function radar communication base station BS and pre-encode them using a zero-forcing radar beamformer; Step 2: Model the optimization problem as maximizing the average target detection probability while ensuring that the total transmission power does not exceed the budget of the dual-function radar communication base station (BS) and that the user's signal-to-interference-plus-noise ratio (SINR) meets the minimum requirements. Solve this problem using a power allocation algorithm based on ergonomics. Step 3: The receiving access point RAPs receives the passive sensing signal from the dual-function radar communication base station BS to the target and then to the receiving access point RAPs. The direct path interference in the passive sensing signal is processed by the whitening filter and turned into white noise. Step 4: The receiving access point RAPs uses the generalized likelihood ratio test to determine whether the GLRT detector exists. The receiving access point RAPs and the dual-function radar communication base station BS transmit the judgment result and the average target detection probability to the fusion center FC. Step 5: The Fusion Center (FC) aggregates the received information through voting to determine whether the target exists. Step 1 specifically involves: using the dedicated sensing symbol s0, precoding is performed using a zero-forcing radar beamformer, and the precoding vector is... Where H = [h1, h2, ..., h K ] H Here, θ is the communication channel matrix, a(θ) is the transmission turning vector, θ is the azimuth angle of the target relative to BS, and the operator |||| represents the 2-norm operation. The optimization problem described in step 2 is as follows: in, This represents the average target detection probability. r = 0, 1, ..., R These are the target detection probabilities of BS and RAPs, where R is the number of RAPs. It is a non-central chi-square cumulative distribution function with 2 degrees of freedom, ρ r The non-centrality parameters for BS and RAPs can be given by the generalized likelihood ratio test detector used, ξ. r It is the threshold of the GLRT detector. It is the SINR of the i-th user. It is a normalized communication transmission precoding vector, where λ is the regularization parameter, and p = [p0, p1, ..., p]. K [Power allocation for sensor symbols and communication symbols] Let K be the variance of additive white Gaussian noise (AWGN), K be the number of users, Γ be the SINR threshold, and P be the variance of the additive white Gaussian noise (AWGN). T This is the power budget of BS; Step 2, specifically, includes the use of a power allocation algorithm based on traversal. Step 2.1: When the power p0 assigned to the sensor symbol gradually increases and satisfies the BS power budget, the average target detection probability... Gradually increase, initialization: p′0 = P T ,Δp,p sum =p′0, where p′0 is the initialization parameter set, Δp is the step size, and p sum As a condition for judgment; Step 2.2: Set p0 = p′0 - Δp; Step 2.3: Use the CVX toolkit to solve the convex optimization problem min||p||1 s.t.γ i ≥Γ, where || ||1 is a norm; Step 2.4, set p sum =||p||1, p′0 = p0; Step 2.5, when p sum ≥P T Proceed to step 2.2; otherwise, output p.

2. The active-passive cooperative sensing fusion method for a dual-function radar-communication system according to claim 1, characterized in that, Step 3 specifically involves: When a target is present, the passive sensing signal received by the r-th RAP is: Where, n′ r [l] represents the AWGN matrix, α r It is the combined sensing channel gain. It is the turning vector of the receiving antenna. Let X = WS be the azimuth angle of the target relative to the r-th RAP, and let WS be the DFRC signal matrix. S = [s0, s1, ..., s K ], s i Let i = 1, 2, ..., K be the communication symbols of the i-th user, and G be the communication symbols of the i-th user. r This represents the targetless channel between the BS and the r-th RAP in the absence of a target. The signal is obtained by passing through a matched filter and vectorizing: in vec() represents the vectorization operator, ε r It follows a zero-mean, complex Gaussian distribution and has a block covariance matrix of: in, The whitening filter is obtained as Depend on Obtained by disassembling Cherosky.

3. The active-passive cooperative sensing fusion method for a dual-function radar communication system according to claim 2, characterized in that, Step 4 is as follows: The binary assumptions of the GLRT detector after using the whitening filter are: in The corresponding GLRT detector is then given by the following formula: and The joint probability density function of the received signals is used to calculate the test statistic, which is: in, Then, a binary decision is made based on (8).

4. The active-passive cooperative sensing fusion method for a dual-function radar-communication system according to claim 1, characterized in that, Step 5 specifically involves: FC makes the following binary assumptions Where D r D represents the binary reasoning results of BS and the r-th RAP, respectively. r =0 indicates no target, D r =1 indicates the existence of a target, and n represents the voting threshold, which is given by the following formula: in, Let represent the false alarm probabilities of BS and RAPs.

Images