A signal detection method and system based on the phase difference vector of array signals

Abstract

The present invention provides a signal detection method and system based on the phase difference vector of array signals. The method constructs a phase difference vector by solving the phase differences of the received signals between different array elements of a receiving array, and uses the mean value of the correlation coefficients of the phase difference vectors of multiple segments of data before and after time as a detection parameter to reduce the influence of environmental interference. Compared with existing algorithms, the proposed scheme of the present invention has basically the same computational complexity and better detection effect, that is, it ensures and improves the advantages of array signal processing without increasing the computational complexity, and has good detection performance under low signal-to-noise ratio conditions.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and more specifically, to a signal detection method and system based on an array signal phase difference vector. Background Art

[0002] Signal detection technology is an essential primary step in non-cooperative communications such as electronic reconnaissance and electronic countermeasures. Signal detection technology avoids analyzing invalid data by detecting the presence of a target signal, so as to reduce the workload of the system, and is widely applied in directions such as communications, radar, seismology, speech processing, biomedicine, and image processing. According to the number and location distribution of receiving devices, detection technologies can be divided into single-antenna detection algorithms and multi-antenna detection algorithms.

[0003] Relatively classical single-antenna detection algorithms include the likelihood ratio test (LRT), the matched filter (MF), the energy detection (ED), and the cyclostationary detection (CSD). The LRT algorithm constructs a likelihood ratio statistic of the received signal as a detection feature, and determines the presence or absence of the target signal by comparing it with a test threshold; the MF algorithm filters out interference signals by constructing a suitable filter to highlight the target signal, so prior information such as the frequency and waveform of the target signal needs to be known before detection; the ED algorithm does not require prior information of the target signal, compares the total energy of the received signal in a determined frequency band with a decision threshold to obtain a detection result, and the decision threshold of this method changes with the ambient noise, and the variance of the surrounding ambient noise needs to be accurately estimated; the CSD algorithm uses the cyclostationarity of the target signal for detection, and the computational complexity is relatively high, making it difficult to meet the real-time requirements.

[0004] With the development of technologies and detection devices, array signal processing has become a mainstream research direction in the field of statistical signal processing, and multi-antenna detection algorithms have emerged as the times require. Since the covariance matrix can measure the difference between signals and noise, the eigenvalue detection algorithm has been proposed. According to different covariance matrix eigenvalue functions, the eigenvalue detection algorithm can be divided into the scaled largest eigenvalue (SLE), the arithmetic-to-geometric mean (AGM), and the John detection. Existing multi-antenna signal detection algorithms are mainly based on analyzing the noise characteristics, but there are still great difficulties in the theoretical derivation and algorithm implementation of the signal covariance structure characteristics.

[0005] A Chinese patent with a publication date of October 18, 2019 and a publication number of CN110348402A discloses a signal detection method combining the expected likelihood of characteristic frequencies, belonging to the technical field of signal detection. First, the present invention establishes a data covariance matrix from the sampled data of the signal and obtains the Fourier transform Wm(θk) of the eigenvectors of the data covariance matrix; then defines the frequency where the peak point of the spectrum of the obtained Wm(θk) is located as the characteristic frequency, calculates the expected likelihood statistic combining the characteristic frequency, and calculates the eigenvector detection threshold; then calculates the threshold of the expected likelihood statistic combining the characteristic frequency using a numerical method; finally, obtains the detection result from the new binary hypothesis testing formula. This patent needs to analyze the covariance matrix of the received array signal, and its detection performance is not good in the case of low signal-to-noise ratio. Summary of the Invention

[0006] The primary object of the present invention is to provide a signal detection method based on the phase difference vector of array signals, which does not need to analyze the covariance matrix of the received array signal and has better detection performance in the case of low signal-to-noise ratio.

[0007] A further object of the present invention is to provide a signal detection system based on the phase difference vector of array signals.

[0008] To solve the above technical problems, the technical solution of the present invention is as follows:

[0009] A signal detection method based on the phase difference vector of array signals, comprising the following steps:

[0010] S1: Generate Gaussian white noise signals of multiple array elements;

[0011] S2: Construct noise signal phase difference vectors at multiple continuous times according to the Gaussian white noise signals, and solve the mean value of the correlation coefficients of the noise signal phase difference vectors according to the noise signal phase difference vectors at the multiple continuous times;

[0012] S3: Repeat step S2 to calculate the mean values of the correlation coefficients of multiple noise signal phase difference vectors. The mean values of the correlation coefficients of the noise signal phase difference vectors follow a normal distribution, and obtain the mean value and standard deviation of the mean values of the correlation coefficients of the noise signal phase difference vectors;

[0013] S4: Set a decision threshold according to the mean value and standard deviation of the mean values of the correlation coefficients of the noise signal phase difference vectors obtained in S3;

[0014] S5: Input the signals received by the array elements;

[0015] S6: Construct received signal phase difference vectors at multiple continuous times according to the signals received by the array elements, and solve the mean value of the correlation coefficients of the received signal phase difference vectors according to the received signal phase difference vectors at the multiple continuous times;

[0016] S7: Compare the mean value of the correlation coefficient of the received signal phase difference vector with the decision threshold set in step S4;

[0017] S8: Output the decision result.

[0018] Preferably, in step S2, a plurality of noise signal phase difference vectors at continuous times are constructed according to the Gaussian white noise signal, specifically:

[0019] Each array element intercepts a signal with a data volume length of L in the Gaussian white noise signal N(t) = [n1(t), n2(t), …, n i (t)] T Among them, taking the signal intercepted by the first array element as the standard, calculate the phase difference φ N-m1 between the signal intercepted by the m-th array element and the first array element, and construct a phase difference vector ψ N = [φ N-21 ,..., φ N-m1 ,..., φ N-i1 T , where i is the number of array elements, i is a positive integer not less than 4, n1(t), n2(t), n i (t) are the Gaussian white noise signal of the first array element, the Gaussian white noise signal of the second array element, and the Gaussian white noise signal of the i-th array element, respectively.

[0020] Preferably, in step S2, calculating the phase difference φ N-m1 between the signal intercepted by the m-th array element and the first array element, specifically:

[0021] The multiple array elements are arranged as a uniform circular array with a radius of R. When there is only Gaussian white noise signal, the received signal is:

[0022] X(t) = N(t)

[0023] Taking the first array element as the reference point, the frequency-domain Fourier transform of the Gaussian white noise signal received by the first array element is expressed as:

[0024]

[0025] Let the time delay of the Gaussian white noise signal received by the second array element relative to the first array element be τ, then the frequency-domain Fourier transform of the Gaussian white noise signal received by the second array element is expressed as:

[0026]

[0027] Multiply X N1 (ω) and X N2 (ω) conjugately to obtain its conjugate product Y N21 (ω): ​

[0028] Y N21 (ω) = |X N1 (ω)| 2 e -jωτ

[0029] The phase difference of the received noise signal between the first array element and the second array element at the highest frequency point ω0 is φ N-21 :

[0030]

[0031] wherein, the phase difference between array elements is within [-π, π];

[0032] Perform the above steps on other array elements to obtain the noise signal phase difference vector:

[0033] ψ N = [φ N-21 ,..., φ N-m1 ,…, φ N-i1 T .

[0034] Preferably, in step S2, the mean value of the correlation coefficient of the noise signal phase difference vector is solved according to the noise signal phase difference vectors at the plurality of continuous times, specifically:

[0035] Intercept M segments of signals with a continuous data volume length of L without overlap in time, and construct a set of vector data:

[0036] [ψ N1 , ψ N2 ,…, ψ NM

[0037] In the formula, ψ N1 , ψ N2 and ψ NM are respectively the noise signal phase difference vector constructed from the signal intercepted in the first time period, the noise signal phase difference vector constructed from the signal intercepted in the second time period, and the noise signal phase difference vector constructed from the signal intercepted in the Mth time period;

[0038] The correlation coefficient J coef between the front and rear noise signal phase difference vectors is calculated as:

[0039]

[0040] In the formula, ψ t1 = [φ 21 , φ 31 ,..., φ i1 T refers to the first noise signal phase difference vector participating in the calculation, and ψ t2 = [φ'​​​21 , φ' 31 ,..., φ' i1 T It refers to the second noise signal phase difference vector participating in the calculation;

[0041] The phase difference vector ψ of the received noise data N1 and ψ N2 The correlation coefficient is:

[0042]

[0043] Similarly, construct a set of correlation coefficients:

[0044] [J N-21 , J N-32 ,…, J N-M(M-1)

[0045] Then the mean value J of the correlation coefficients of the noise signal phase difference vectors N-mean is:

[0046]

[0047] Preferably, in step S3, obtaining the mean value and standard deviation of the correlation coefficients of the noise signal phase difference vectors specifically is:

[0048] Calculate the mean values of the correlation coefficients of the noise signal phase difference vectors in Num different time periods. At this time, the mean values of the correlation coefficients of the noise signal phase difference vectors are random and follow a normal distribution, and its probability density function satisfies:

[0049]

[0050] In the formula, μ is the mean value of the correlation coefficients of the noise signal phase difference vectors, and σ is the standard deviation of the correlation coefficients of the noise signal phase difference vectors.

[0051] Preferably, in step S4, setting the decision threshold specifically is:

[0052] th = μ + 3σ

[0053] In the formula, th is the decision threshold.

[0054] Preferably, in step S6, constructing the received signal phase difference vectors at multiple continuous times according to the signals received by the array elements specifically is:

[0055] The radius of the multi-element uniform circular array is R. Let the detection array receive a narrowband far-field signal s(t) with an azimuth angle θ and an elevation angle , and the signal gain of each array element is 1. Then the received signal is:

[0056] X(t) = AS(t) + N(t)​​

[0057] Among them, A = A(θ) = [a1(θ), a2(θ), …, a i (θ)] T is the steering vector, and the received signal of the m-th array element is a m (θ)·s m (t) + n m (t). The data volume of length L is intercepted for the received data of each array element to facilitate the subsequent construction of the received signal phase difference vector;

[0058] Taking the first array element as the reference point, the frequency-domain Fourier transform of its received signal can be expressed as:

[0059]

[0060] Let the time delay of the received signal of the second array element relative to the reference point be τ, then the frequency-domain Fourier transform of the received signal of the second array element is:

[0061]

[0062] Multiply X S1 (ω) and the conjugate of X S2 (ω) to obtain its conjugate product Y S21 (ω):

[0063] Y S21 (ω) = |X S1 (ω)| 2 e -jωτ

[0064] The phase difference of the received signal between the first array element and the second array element at the highest frequency point ω0 is φS-21:

[0065]

[0066] Among them, the phase difference between array elements is within [-π, π];

[0067] Perform the above steps for other array elements to obtain the received signal phase difference vector:

[0068] ψ S = [φ S-21 ,..., φ S-m1 ,..., φ S-i1 T .

[0069] Preferably, in step S6, solving the mean value of the received signal phase difference vector correlation coefficient according to the received signal phase difference vectors at the multiple consecutive times is specifically:

[0070] ​Intercept signals with a length of L and continuous and non-overlapping data volume for M periods of time, and construct a set of vector data:

[0071] [ψ S1 , ψ S2 , …, ψ SM

[0072] In the formula, ψ S1 , ψ S2 and ψ SM are the received signal phase difference vectors constructed from the signals intercepted in the first period of time, the received signal phase difference vectors constructed from the signals intercepted in the second period of time, and the received signal phase difference vectors constructed from the signals intercepted in the Mth period of time, respectively;

[0073] The correlation coefficient between the received signal data phase difference vectors ψ S1 and ψ S2 is:

[0074]

[0075] Similarly, construct a set of correlation coefficients:

[0076] [J S-21 , J S-32 , …, J S-M(M-1)

[0077] Then the average value J S-mean of the received signal phase difference vector correlation coefficient is:

[0078]

[0079] Preferably, step S8 outputs a judgment result, specifically:

[0080] If J S-mean > th, the judgment result flag = 1, indicating the existence of the target signal;

[0081] If J S-mean ≤ th, the judgment result flag = 0, indicating the non-existence of the target signal.

[0082] A signal detection system based on the phase difference vector of array signals includes:

[0083] A noise generation module, which is used to generate Gaussian white noise signals of multiple array elements;

[0084] A first calculation module, which is used to construct phase difference vectors of noise signals at multiple continuous times according to the Gaussian white noise signals, and solve the average value of the correlation coefficients of the phase difference vectors of the noise signals;

[0085] ​​An acquisition module, which is used to acquire the mean and standard deviation of the mean of the correlation coefficients of the noise signal phase difference vectors, and the mean of the correlation coefficients of the noise signal phase difference vectors follows a normal distribution;

[0086] A threshold setting module, which is used to set a decision threshold according to the mean and standard deviation of the mean of the correlation coefficients of the noise signal phase difference vectors acquired by the acquisition module;

[0087] A receiving module, which is used to input the signals received by the array elements;

[0088] A second calculation module, which is used to construct the received signal phase difference vectors at multiple consecutive times according to the received signals of the array elements, and solve the mean of the correlation coefficients of the received signal phase difference vectors according to the received signal phase difference vectors at the multiple consecutive times;

[0089] A comparison module, which is used to compare the mean of the correlation coefficients of the received signal phase difference vectors with the decision threshold set in step S4;

[0090] An output module, which is used to output the decision result.

[0091] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0092] Compared with the existing single-antenna energy detection (ED) algorithm and multi-antenna minimum description length detection method (Minimum Description Length Detection, MDLD), the solution proposed by the present invention has a similar computational complexity to the ED algorithm, and the detection effect is better than the ED algorithm and the MDLD algorithm. That is, on the premise of not increasing the computational complexity, the advantages of array signal processing are ensured and improved, and the detection performance is good under low signal-to-noise ratio. Description of the Drawings

[0093] Figure 1 It is a schematic flowchart of the method of the present invention.

[0094] Figure 2 It is a schematic diagram of the structure of a multi-element uniform disk.

[0095] Figure 3 It is a schematic diagram of the detection accuracy of the method proposed by the present invention, the single-antenna ED algorithm, and the multi-antenna MDLD algorithm at different signal-to-noise ratios when the detected target signal is an FM signal.

[0096] Figure 4 It is a schematic diagram of the detection accuracy of the method proposed by the present invention, the single-antenna ED algorithm, and the multi-antenna MDLD algorithm at different signal-to-noise ratios when the detected target signal is a 4ASK signal.

[0097] Figure 5 This is a schematic diagram of the system of the present invention. Detailed implementation manners

[0098] The accompanying drawings are only for illustrative purposes and should not be construed as limitations on this patent.

[0099] For better illustration of this embodiment, some components in the accompanying drawings are omitted, enlarged or reduced, which do not represent the dimensions of the actual product.

[0100] For those skilled in the art, it is understandable that some well-known structures and their descriptions in the accompanying drawings may be omitted.

[0101] The technical solution of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0102] Embodiment 1

[0103] This embodiment provides a signal detection method based on the phase difference vector of array signals. As Figure 1 shown, it includes the following steps:

[0104] S1: Generate Gaussian white noise signals of multiple array elements;

[0105] S2: Construct phase difference vectors of noise signals at multiple continuous times according to the Gaussian white noise signals, and solve the mean value of the correlation coefficients of the phase difference vectors of the noise signals according to the phase difference vectors of the noise signals at the multiple continuous times;

[0106] S3: Repeat step S2 to calculate the mean values of the correlation coefficients of multiple phase difference vectors of the noise signals. The mean values of the correlation coefficients of the phase difference vectors of the noise signals follow a normal distribution, and obtain the mean value and standard deviation of the mean values of the correlation coefficients of the phase difference vectors of the noise signals;

[0107] S4: Set a decision threshold according to the mean value and standard deviation of the mean values of the correlation coefficients of the phase difference vectors of the noise signals obtained in S3;

[0108] S5: Input the signals received by the array elements;

[0109] S6: Construct phase difference vectors of the received signals at multiple continuous times according to the signals received by the array elements, and solve the mean value of the correlation coefficients of the phase difference vectors of the received signals according to the phase difference vectors of the received signals at the multiple continuous times;

[0110] S7: Compare the mean value of the correlation coefficients of the phase difference vectors of the received signals with the decision threshold set in step S4;

[0111] S8: Output the decision result.

[0112] The above method is applicable to a detection array with the number of array elements greater than or equal to 4. For the convenience of description, this embodiment takes an 8-element detection array as an example:

[0113] In step S2, multiple noise signal phase difference vectors at multiple continuous times are constructed according to the Gaussian white noise signal, specifically as follows:

[0114] Each array element intercepts a segment of the Gaussian white noise signal N(t) = [n1(t), n2(t), …, n8(t)] T with a signal length of L. In the formula, taking the signal intercepted by the first array element as the standard, the phase difference φ N-m1 between the signal intercepted by the m-th array element and the signal intercepted by the first array element is calculated, and a phase difference vector ψ N = [φ N-21 , φ N-31 , …, φ N-81 T is constructed, where n1(t), n2(t), and n8(t) are the Gaussian white noise signals of the first array element, the second array element, and the eighth array element, respectively.

[0115] In step S2, the phase difference φ N-m1 between the signal intercepted by the m-th array element and the signal intercepted by the first array element is calculated, specifically as follows:

[0116] As Figure 2 shown, the radius of the eight-element uniform circular array is R. When only receiving the Gaussian white noise signal, the received signal is:

[0117] X(t) = N(t)

[0118] Taking the first array element as the reference point, the frequency-domain Fourier transform of the Gaussian white noise signal received by the first array element is expressed as:

[0119]

[0120] Let the time delay of the Gaussian white noise signal received by the second array element relative to the first array element be τ, then the frequency-domain Fourier transform of the Gaussian white noise signal received by the second array element is expressed as:

[0121]

[0122] Multiply X N1 (ω) and the conjugate of X N2 (ω) to obtain their conjugate product Y N21 (ω):

[0123] Y N21 (ω) = |X N1 (ω)| 2 e -jωτ

[0124] ​The phase difference of the received noise signal between the first array element and the second array element at the highest frequency point ω0 is φ N-21 :

[0125]

[0126] wherein, the phase difference between array elements is within [-π, π];

[0127] Perform the above steps on other array elements to obtain the noise signal phase difference vector:

[0128] ψ N =[φ N-21 , φ N-31 , φ N-41 , φ N-51 , φ N-61 , φ N-71 , φ N-81 T .

[0129] In step S2, solve the mean value of the correlation coefficient of the noise signal phase difference vector according to the noise signal phase difference vectors at the multiple consecutive times, specifically:

[0130] Intercept M segments of signals with a continuous and non-overlapping data volume length of L, and construct a set of vector data:

[0131] [ψ N1 , ψ N2 , …, ψ NM

[0132] In the formula, ψ N1 , ψ N2 and ψ NM are respectively the noise signal phase difference vector constructed from the signal intercepted in the first period of time, the noise signal phase difference vector constructed from the signal intercepted in the second period of time, and the noise signal phase difference vector constructed from the signal intercepted in the Mth period of time;

[0133] The correlation coefficient J coef of the two adjacent noise signal phase difference vectors is calculated as:

[0134]

[0135] In the formula, ψ t1 =[φ 21 , φ 31 , φ 41 , φ 51 , φ 61 , φ 71 , φ 81 T refers to the first noise signal phase difference vector participating in the calculation, and ψ t2 =[φ'​​​21 , φ' 31 , φ' 41 , φ' 51 , φ' 61 , φ' 71 , φ' 81 T It refers to the second phase difference vector of the noise signal participating in the calculation;

[0136] The phase difference vector ψ of the received noise data N1 and ψ N2 The correlation coefficient is:

[0137]

[0138] Similarly, construct a set of correlation coefficients:

[0139] [J S-21 , J S-32 , …, J S-M(M-1)

[0140] Then the mean J of the correlation coefficients of the noise signal phase difference vectors N-mean is:

[0141]

[0142] Step S3 to obtain the mean and standard deviation of the mean of the correlation coefficients of the noise signal phase difference vectors is specifically as follows:

[0143] Calculate the mean of the correlation coefficients of the noise signal phase difference vectors in Num different time periods. In the case of low signal-to-noise ratio, the phase difference vector of the received signal shows randomness, and the mean of the correlation coefficients of the noise signal phase difference vectors has randomness. At this time, the mean of the correlation coefficients of the noise signal phase difference vectors approaches zero and follows a normal distribution, and its probability density function satisfies:

[0144]

[0145] In the formula, μ is the mean of the mean of the correlation coefficients of the noise signal phase difference vectors, and σ is the standard deviation of the mean of the correlation coefficients of the noise signal phase difference vectors. According to the set statistical parameter Num, after Num times of statistical solutions, estimate the mean μ and standard deviation σ of the coefficient J N-mean .

[0146] In step S4, set the decision threshold. When the false alarm probability P f = 10 -3 , select the 3σ criterion:

[0147] th = μ + 3σ

[0148] In the formula, th is the decision threshold.

[0149] ​​In step S6, multiple received signal phase difference vectors at consecutive times are constructed based on the received signals of the array elements, specifically as follows:

[0150] As Figure 2 shown, the radius of the eight-element uniform circular array is R. Let the detection array receive a narrowband far-field signal s(t) with an azimuth angle θ and an elevation angle . If the signal gain of each array element is 1, the received signal is:

[0151] X(t) = AS(t) + N(t)

[0152] where A = A(θ) = [a1(θ), a2(θ), …, a i (θ)] T is the steering vector, and the received signal of the m-th array element is a m (θ)·s m (t) + n m (t). The data volume of length L is intercepted for the received data of each array element to facilitate the subsequent construction of the received signal phase difference vector;

[0153] Taking the first array element as the reference point, the frequency-domain Fourier transform of its received signal can be expressed as:

[0154]

[0155] Let the time delay of the received signal of the second array element relative to the reference point be τ. Then, the frequency-domain Fourier transform of the received signal of the second array element is:

[0156]

[0157] Multiply X S1 (ω) and X S2 (ω) conjugately to obtain their conjugate product Y S21 (ω):

[0158] Y S21 (ω) = |X S1 (ω)| 2 e -jωτ

[0159] The phase difference of the received signal between the first and the second array elements at the highest frequency point ω0 is φS-21:

[0160]

[0161] where the phase difference between array elements is within [-π, π];

[0162] Perform the above steps for other array elements to obtain the received signal phase difference vector:

[0163] ψS = [φ S-21 , φ S-31 , φ S-41 , φ S-51 , φ S-61 , φ S-71 , φ S-81 T .

[0164] In step S6, the mean value of the correlation coefficient of the received signal phase difference vector is solved according to the received signal phase difference vectors at the multiple consecutive times, specifically:

[0165] Intercept M segments of signals with a continuous data volume length of L and no overlap in time, and construct a set of vector data:

[0166] [ψ S1 , ψ S2 , …, ψ SM

[0167] In the formula, ψ S1 , ψ S2 and ψ SM are the received signal phase difference vectors constructed from the signals intercepted in the first time period, the received signal phase difference vectors constructed from the signals intercepted in the second time period, and the received signal phase difference vectors constructed from the signals intercepted in the Mth time period, respectively;

[0168] The correlation coefficient between the phase difference vectors ψ S1 and ψ S2 of the received signal data is:

[0169]

[0170] Similarly, construct a set of correlation coefficients:

[0171] [J S-21 , J S-32 , …, J S-M(M-1)

[0172] Then the mean value J S-mean of the correlation coefficient of the received signal phase difference vector is:

[0173]

[0174] Step S8 outputs the judgment result, specifically:

[0175] If J S-mean > th, the judgment result flag = 1, indicating that the target signal exists;

[0176] If J S-mean ≤ th, the judgment result flag = 0, indicating that the target signal does not exist. ​​​

[0177] In a specific implementation process, the simulation parameters are set as follows: The incoming signal is a single FM complex signal with a signal frequency of 10 MHz, an azimuth angle of the incoming signal of 120°, an elevation angle of 12°, the electromagnetic wave propagation speed of 3×10 8 m / s, a sampling frequency of 20 kHz, the number of array elements of 8, and the radius of the array element of 50 m; the noise type is complex Gaussian white noise, and the false alarm probability is 10 -3 , the data volume L = 512, the number of data segments M = 4, the Monte Carlo experiment is set 51,200 times, and the statistical parameter Num = 100,000 times.

[0178] Compare the detection performance of the new method proposed in this embodiment with that of the single-antenna ED algorithm and the multi-antenna MDLD algorithm. The ED algorithm is divided into the energy accumulation detection algorithm and the average energy detection algorithm, which are specifically as follows: According to Figure 3 As shown, when the target signal to be detected is an FM signal, the detection correct rate at different signal-to-noise ratios is different. Among them, the Energy curve corresponds to the energy accumulation detection algorithm; the Power curve corresponds to the average energy detection algorithm; the PDCD curve corresponds to the detection algorithm proposed in this embodiment (Phase Difference Correlation Detection, PDCD); the MDLD curve corresponds to the minimum description length detection algorithm. The detection correct rates of the four detection algorithms increase with the increase of the signal-to-noise ratio. The detection effect of the PDCD algorithm is better than that of the MDLD algorithm, and the detection effect of the MDLD algorithm is better than that of the two ED algorithms. Among them, the comparison between the MDLD algorithm and the ED algorithm verifies that array signal processing has a high gain on the signal detection result, and the comparison between the PDCD algorithm and the MDLD algorithm verifies that the new detection algorithm proposed in this embodiment improves the advantage of array signal processing. The detection correct rate of the PDCD algorithm starts to rise at SNR = -21 dB and approaches 100% at -13 dB, verifying that the algorithm has good detection performance in the case of low signal-to-noise ratio.

[0179] The above embodiment shows that in the case where the target detection signal is an analog signal FM and the SNR of the received signal is low, the method described in this embodiment can reduce the influence of environmental noise, achieve a good signal detection effect, and has a low computational complexity.

[0180] On the basis of the above embodiment, verify the detection performance of the digital signal 4ASK signal. Specifically, during the process of setting the simulation parameters, modify the incoming signal to a single 4ASK complex signal, the data volume L = 2,048, the number of data segments M = 4, and other parameters remain unchanged. Compare the detection performance of the new method proposed in this embodiment with that of the single-antenna ED algorithm and the multi-antenna MDLD algorithm. The results are as follows:

[0181] As Figure 4As shown in the figure, when the target signal to be detected is a 4ASK signal, the detection accuracy rates under different signal-to-noise ratios are different. Among them, the detection accuracy rates of the four detection algorithms increase with the increase of the signal-to-noise ratio. The detection effect of the PDCD algorithm is better than that of the MDLD algorithm, and the detection effect of the MDLD algorithm is better than that of the two ED algorithms. Among them, although the detection accuracy rate of the MDLD algorithm reaches 100% earlier than that of the PDCD algorithm, its sensitivity to the signal-to-noise ratio of the received signal is higher than that of the PDCD algorithm. The PDCD algorithm is more suitable for the harsh reception environment with a low signal-to-noise ratio. The detection accuracy rate of the PDCD algorithm starts to rise when SNR=-23dB and approaches 100% when SNR=-15dB, verifying that the algorithm has good detection performance in the case of a low signal-to-noise ratio.

[0182] The above embodiments show that in the case where the target detection signal is a digital modulation signal 4ASK and the SNR of the received signal is low, the method described in this embodiment can reduce the influence of environmental noise, achieve a good signal detection effect, and has a low computational complexity.

[0183] In summary, the method provided in this embodiment is mainly aimed at the detection scenario of a stationary emission source and a fixed detection array. Utilizing the characteristic that the direction of arrival of the wave from the emission source remains unchanged, the mean value of the phase difference vector correlation coefficients of multiple segments of data before and after time is used as the detection parameter. Compared with the MDLD algorithm, this method does not need to analyze the covariance matrix of the received array signal, avoids complex calculations, and at the same time maintains the high gain of array signal processing. The detection performance is far better than that of the ED algorithm using a single-antenna receiving device.

[0184] Embodiment 2

[0185] This embodiment provides a signal detection system based on the phase difference vector of array signals, as Figure 5 shown, including:

[0186] A noise generation module, which is used to generate Gaussian white noise signals of multiple array elements;

[0187] A first calculation module, which is used to construct phase difference vectors of noise signals at multiple consecutive times according to the Gaussian white noise signals, and solve the mean value of the phase difference vector correlation coefficients of the noise signals according to the phase difference vectors of the noise signals at the multiple consecutive times;

[0188] An acquisition module, which is used to acquire the mean value and standard deviation of the mean value of the phase difference vector correlation coefficients of the noise signals. The mean value of the phase difference vector correlation coefficients of the noise signals follows a normal distribution;

[0189] A threshold setting module, which is used to set a decision threshold according to the mean value and standard deviation of the mean value of the phase difference vector correlation coefficients of the noise signals acquired by the acquisition module;

[0190] A receiving module for receiving signals by input array elements;

[0191] A second calculation module for constructing phase difference vectors of received signals at multiple continuous times based on the received signals of the array elements and solving the mean value of the correlation coefficients of the phase difference vectors of the received signals;

[0192] A comparison module for comparing the mean value of the correlation coefficients of the phase difference vectors of the received signals with the decision threshold set in step S4;

[0193] An output module for outputting a decision result.

[0194] Identical or similar reference numerals correspond to identical or similar components;

[0195] The terms describing the positional relationship in the drawings are for illustrative purposes only and should not be construed as a limitation of this patent;

[0196] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention and are not intended to limit the embodiments of the present invention. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A signal detection method based on the phase difference vector of array signals, characterized in that, It includes the following steps: S1: Generate Gaussian white noise signals of multiple array elements; S2: Construct phase difference vectors of noise signals at multiple continuous times according to the Gaussian white noise signals, and solve the mean value of the correlation coefficients of the phase difference vectors of the noise signals according to the phase difference vectors of the noise signals at the multiple continuous times; S3: Repeat step S2 to calculate the mean values of the correlation coefficients of multiple phase difference vectors of noise signals. The mean values of the correlation coefficients of the phase difference vectors of the noise signals follow a normal distribution, and obtain the mean value and standard deviation of the mean values of the correlation coefficients of the phase difference vectors of the noise signals; S4: Set a decision threshold according to the mean value and standard deviation of the mean values of the correlation coefficients of the phase difference vectors of the noise signals obtained in S3; S5: Input the received signals of the array elements; S6: Construct phase difference vectors of the received signals at multiple continuous times according to the received signals of the array elements, and solve the mean value of the correlation coefficients of the phase difference vectors of the received signals according to the phase difference vectors of the received signals at the multiple continuous times; S7: Compare the mean value of the correlation coefficients of the phase difference vectors of the received signals with the decision threshold set in step S4; S8: Output the decision result; Step S2 is specifically as follows: Each array element intercepts a signal with a data volume length of L from the Gaussian white noise signal N(t) = [n1(t), n2(t), …, n i (t)] T In the formula, taking the signal intercepted by the first array element as the standard, the phase difference φ N-m1 between the signal intercepted by the m-th array element and the signal intercepted by the first array element is calculated, and the phase difference vector ψ N = [φ N-21 ,..., φ N-m1 ,..., φ N-i1 T In the formula, i is the number of array elements, i is a positive integer not less than 4, and n1(t), n2(t), n i (t) are the Gaussian white noise signals of the 1st array element, the 2nd array element, and the i-th array element respectively;​ The multiple array elements are arranged as a uniform circular array with a radius of R. When there is only Gaussian white noise signal, the received signal is: X(t) = N(t) Taking the first array element as the reference point, the frequency-domain Fourier transform of the Gaussian white noise signal received by the first array element is expressed as: Let the time delay of the Gaussian white noise signal received by the second array element relative to the first array element be τ, then the frequency-domain Fourier transform of the Gaussian white noise signal received by the second array element is expressed as: Multiply X N1 (ω) by the conjugate of X N2 (ω) to obtain its conjugate product Y N21 (ω): Y N21 (ω) = |X N1 (ω)| 2 e -jωτ The phase difference of the received noise signals between the first array element and the second array element at the highest frequency point ω0 is φ N-21 : Wherein, the phase difference between the array elements is within [-π, π]; Perform the above steps for other array elements to calculate the phase difference vector of the noise signal: ψ N = [φ N-21 ,..., φ N-m1 ,..., φ N-i1 T ​ Intercept M segments of signals with a data volume length of L that are continuous in time and have no overlap, and construct a set of vector data: [ψ N1 , ψ N2 , …, ψ NM ​ where ψ N1 , ψ N2 and ψ NM are the phase difference vectors of the noise signals constructed from the signals intercepted in the first period, the phase difference vectors of the noise signals constructed from the signals intercepted in the second period, and the phase difference vectors of the noise signals constructed from the signals intercepted in the Mth period, respectively; The correlation coefficient J of the phase difference vectors of the front and rear noise signals coef The calculation formula is as follows: where ψ t1 = [φ 21 , φ 31 ,..., φ i1 T represents the first phase difference vector of the noise signals participating in the calculation, and ψ t2 = [φ' 21 , φ' 31 ,..., φ' i1 T represents the second phase difference vector of the noise signals participating in the calculation;​​ The phase difference vector ψ of the received noise data N1 and ψ N2 The correlation coefficient of is as follows: Similarly, construct a set of correlation coefficients: [J N-21 ,J N-32 ,…,J N-M(M-1) ​ Then the mean value J of the correlation coefficient of the noise signal phase difference vectors N-mean is as follows: Step S6 is specifically as follows: The radius of the multi-element uniform circular array is R. Let the detection array receive a narrowband far-field signal s(t) with an azimuth angle θ and an elevation angle . If the signal gain of each element is 1, the received signal is as follows: X(t) = AS(t) + N(t) where \(A = A(\theta)=[a_1(\theta),a_2(\theta),\cdots,a i (\theta)] T is the steering vector, the received signal of the \(m\)-th array element is \(a m (\theta)\cdot s m (t)+n m (t)\). For the received data of each array element, a data volume with a length of \(L\) is intercepted to facilitate the subsequent construction of the received signal phase difference vector; Taking the first array element as the reference point, the frequency-domain Fourier transform of its received signal can be expressed as: Let the time delay of the received signal of the second array element relative to the reference point be τ, then the frequency-domain Fourier transform of the received signal of the second array element is: Multiply X S1 (ω) by the conjugate of X S2 (ω) to obtain its conjugate product Y S21 (ω): Y S21 (ω) = |X S1 (ω)| 2 e -jωτ The phase difference of the received signals between the first array element and the second array element at the highest frequency point ω0 is φ S-21 : Wherein, the phase difference between the array elements is within [-π, π]; Perform the above steps for other array elements to calculate the phase difference vector of the received signal: ψ S = [φ S-21 ,..., φ S-m1 ,..., φ S-i1 T ​ Intercept M segments of signals with a data volume length of L that are continuous in time and have no overlap, and construct a set of vector data: [ψ S1 ,ψ S2 ,…,ψ SM ​ where ψ S1 , ψ S2 and ψ SM are the received signal phase difference vectors constructed from the signals intercepted in the first time period, the received signal phase difference vectors constructed from the signals intercepted in the second time period, and the received signal phase difference vectors constructed from the signals intercepted in the Mth time period, respectively; Phase difference vector ψ of received signal data S1 and ψ S2 The correlation coefficient is: Similarly, construct a set of correlation coefficients: [J S-21 ,J S-32 ,…,J S-M(M-1) ​ Then the mean value J of the correlation coefficient of the received signal phase difference vector S-mean is as follows:

2. The signal detection method based on the array signal phase difference vector according to claim 1, wherein Step S3 obtains the mean value and standard deviation of the mean values of the correlation coefficients of the phase difference vectors of the noise signals specifically as follows: Calculate the mean values of the correlation coefficients of the phase difference vectors of the noise signals in Num different time periods. At this time, the mean values of the correlation coefficients of the phase difference vectors of the noise signals are random and follow a normal distribution, and its probability density function satisfies: In the formula, μ is the mean value of the mean values of the correlation coefficients of the phase difference vectors of the noise signals, and σ is the standard deviation of the mean values of the correlation coefficients of the phase difference vectors of the noise signals.

3. The signal detection method based on the array signal phase difference vector according to claim 2, characterized in that In step S4, setting the decision threshold is specifically as follows: th = μ + 3σ In the formula, th is the decision threshold.

4. The signal detection method based on the array signal phase difference vector according to claim 3, wherein Step S8 outputs the decision result, specifically as follows: If J S-mean > th, then the decision result flag = 1, indicating the existence of the target signal; If J S-mean ≤ th, then the decision result flag = 0, indicating that the target signal does not exist.

5. A signal detection system based on the phase difference vector of array signals, characterized in that, The signal detection system applies the signal detection method based on the array signal phase difference vector according to any one of claims 1 to 4, and includes: A noise generation module, which is used to generate Gaussian white noise signals of multiple array elements; A first calculation module, which is used to construct noise signal phase difference vectors at multiple continuous times according to the Gaussian white noise signals, and solve the mean value of the correlation coefficients of the noise signal phase difference vectors according to the noise signal phase difference vectors at the multiple continuous times; An acquisition module, which is used to acquire the mean value and standard deviation of the correlation coefficients of the noise signal phase difference vectors, and the mean value of the correlation coefficients of the noise signal phase difference vectors follows a normal distribution; A threshold setting module, which is used to set a decision threshold according to the mean value and standard deviation of the correlation coefficients of the noise signal phase difference vectors acquired by the acquisition module; A receiving module, which is used to input the received signals of the array elements; A second calculation module, which is used to construct received signal phase difference vectors at multiple continuous times according to the received signals of the array elements, and solve the mean value of the correlation coefficients of the received signal phase difference vectors according to the received signal phase difference vectors at the multiple continuous times; A comparison module, which is used to compare the mean value of the correlation coefficients of the received signal phase difference vectors with the decision threshold set in step S4; An output module, which is used to output the decision result.

Images