Distributed communication signal detection method, apparatus and system based on generalized likelihood ratio

Abstract

Embodiments of the application disclose a kind of distributed communication signal detection method, device and system based on generalized likelihood ratio, wherein the method is applied to fusion center, comprising the following steps: S1: receiving the observation data transmitted by multiple receiving nodes; S2: under the assumption that radiation source does not exist and radiation source exists, respectively construct the joint conditional probability density function of all observation data of all receiving nodes within the observation time; S3: adopt NP criterion to construct the expression of generalized likelihood ratio test statistic; S4: according to maximum likelihood criterion, estimate the covariance matrix in the joint conditional probability density function under the assumption of two kinds; S5: the estimated value of the covariance matrix under the assumption of two kinds is substituted into the expression of generalized likelihood ratio test statistic, and compared with the decision threshold set according to false alarm probability, to complete distributed communication signal detection. The application can improve the detection probability of communication signal under low signal-to-noise ratio, and has good noise resistance when there is noise fluctuation in different receiving nodes.

Description

Technical Field

[0001] This invention relates to distributed communication signal detection technology, specifically to a distributed communication signal detection method, apparatus, and system based on generalized likelihood ratio. Background Technology

[0002] Two factors make signal detection challenging in practical applications. First, the signal-to-noise ratio (SNR) at each receiving node can be very low. With the increasing number of wireless communication devices, the growing complexity and variability of the electromagnetic environment, the deteriorating channel transmission conditions, and the increasing background noise and interference, the signal at the receiving end is often interfered with by strong background noise, sometimes even drowning out the useful signal, resulting in a very low received SNR. Second, noise can vary with time and location, leading to noise uncertainty. Under these conditions, false detections or missed detections may occur, and erroneous detection results will inevitably have a serious impact on modulation scheme identification and signal processing.

[0003] Weak signal detection is a crucial application area in the military field. For example, in battlefield environments, individual military equipment needs to accurately identify enemy communication signals to enable rapid decision-making during combat. Military aircraft and satellites also require weak signal detection capabilities to accurately identify and avoid threats during flight. With the continuous development of modern military technology, military equipment and tactics require more radio spectrum to support various combat missions; therefore, accurate detection of weak signals is becoming increasingly important in the military field. Military-grade weak signal detection systems need to possess efficient signal detection and localization capabilities and be able to correctly identify weak signals in high-noise and interference environments. In military communications, the increasing use of non-standard systems places higher demands on communication signal detection. Communication modes represented by spread spectrum communication, ultra-wideband communication, and frequency hopping communication have low power spectral density and good concealment. In the detection of non-cooperative communication signals without prior information, their presence is often difficult to determine.

[0004] Therefore, it is necessary to propose an effective blind detection algorithm for non-cooperative communication signals. Against this backdrop, researching signal detection algorithms that are easy to implement, have high accuracy, and meet practical application requirements in low signal-to-noise ratio (SNR) environments, continuously lowering the lower limit of detectable SNR, further improving signal detection performance under low SNR conditions, and exhibiting good noise immunity when there are noise fluctuations at different receiving nodes, is of great significance. Summary of the Invention

[0005] The purpose of this invention is to propose a distributed communication signal detection method, apparatus, and system based on generalized likelihood ratio.

[0006] According to a first aspect of the present invention, a distributed communication signal detection method based on generalized likelihood ratio is provided, the method being applied to a fusion center, comprising the following steps:

[0007] S1: Receive observation data transmitted by multiple distributed receiving nodes;

[0008] S2: Based on the pre-built channel model and received signal model, under the two assumptions of no radiation source and the presence of radiation source, the joint conditional probability density function of all observation data of all receiving nodes during the observation time is constructed respectively.

[0009] S3: Based on the joint conditional probability density function under the two assumptions, the NP criterion is used to construct the generalized likelihood ratio test statistic expression, where the unknown variable in the generalized likelihood ratio test statistic expression is the covariance matrix in the joint conditional probability density function under the two assumptions.

[0010] S4: Estimate the covariance matrix in the joint conditional probability density function under the two assumptions according to the maximum likelihood criterion;

[0011] S5: Substitute the estimated values ​​of the covariance matrix under the two assumptions into the expression of the generalized likelihood ratio test statistic, and compare the obtained generalized likelihood ratio test statistic with the decision threshold set according to the false alarm probability. When the generalized likelihood ratio test statistic is greater than the decision threshold, it is determined that there is a radiation source in the detection frequency band; otherwise, it is determined that there is no radiation source in the detection frequency band.

[0012] According to a second aspect of the present invention, a distributed communication signal detection device based on generalized likelihood ratio is provided, the device being applied in a fusion center, and comprising the following modules:

[0013] The receiving module is used to receive observation data transmitted by multiple receiving nodes that are distributed in a dispersed manner.

[0014] The probability density function construction module is used to construct the joint conditional probability density function of all observation data of all receiving nodes within the observation time, based on the pre-built channel model and received signal model, under the two assumptions of no radiation source and the presence of radiation source.

[0015] The test statistic construction module is used to construct a generalized likelihood ratio test statistic expression based on the joint conditional probability density function under the two constructed hypotheses, using the NP criterion. The unknown variable in the generalized likelihood ratio test statistic expression is the covariance matrix in the joint conditional probability density function under the two hypotheses.

[0016] The covariance matrix estimation module is used to estimate the covariance matrix in the joint conditional probability density function under the two assumptions according to the maximum likelihood criterion.

[0017] The comparison and decision module is used to substitute the estimated values ​​of the covariance matrix under the two assumptions into the expression of the generalized likelihood ratio test statistic, and compare the obtained generalized likelihood ratio test statistic with the decision threshold set according to the false alarm probability. When the generalized likelihood ratio test statistic is greater than the decision threshold, it is determined that there is a radiation source in the detection frequency band; otherwise, it is determined that there is no radiation source in the detection frequency band.

[0018] According to a third aspect of the present invention, a distributed communication signal detection system based on generalized likelihood ratio is provided, comprising multiple distributed receiving nodes and a fusion center.

[0019] The multiple receiving nodes collect communication signals from the surrounding environment in real time to obtain observation data, and transmit the collected observation data to the fusion center.

[0020] The fusion center includes a memory and a processor. The fusion center stores the observation data transmitted by the multiple receiving nodes in the memory. The memory also stores a computer program, which is loaded and executed by the processor to implement the aforementioned distributed communication signal detection method based on generalized likelihood ratio.

[0021] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing one or more computer programs, which, when executed by a processor, implement the aforementioned distributed communication signal detection method based on generalized likelihood ratio.

[0022] The beneficial effects of the various embodiments of the present invention are as follows:

[0023] The present invention provides a distributed communication signal detection method, apparatus, and system based on generalized likelihood ratio. It assumes that the channel parameters from the radiation source to each receiving node remain unchanged during the observation period. The signal generated by the radiation source reaches each receiving node after channel fading. Each receiving node transmits the received data to a fusion center. Based on a pre-constructed channel model and received signal model, the fusion center constructs a joint conditional probability density function for all observation data from all receiving nodes during the observation period, under two assumptions: the absence of a radiation source and the presence of a radiation source. Then, it uses the NP criterion to construct a generalized likelihood ratio (GLR) test statistic expression. Next, it estimates the unknown variables in the GLR test statistic expression according to the maximum likelihood (ML) criterion, obtaining estimated values ​​of the covariance matrix in the joint conditional probability density function under the two assumptions. By substituting these estimated values ​​into the GLR test statistic expression and comparing them with a decision threshold set according to the false alarm probability, distributed communication signal detection is completed. Because the present invention employs distributed detection across multiple nodes, it improves the detection probability of communication signals under low signal-to-noise ratios compared to a single receiving node. Theoretically, even under low signal-to-noise ratios, it can achieve excellent detection performance when there are a sufficient number of receiving nodes. Furthermore, by utilizing the correlation between data from multiple receiving nodes, the present invention estimates the unknown variables in the GLR test statistic constructed using the NP criterion according to the maximum likelihood (ML) criterion, thereby obtaining the estimated value of the covariance matrix in the joint conditional probability density function under the two hypotheses. Therefore, when the noise of the receiving node fluctuates, the method of the present invention has good noise resistance. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those skilled in the art can obtain other drawings based on these drawings. In the drawings:

[0025] Figure 1 The diagram illustrates a flowchart of a distributed communication signal detection method based on generalized likelihood ratio provided by an embodiment of the present invention.

[0026] Figure 2 A schematic diagram of the structure of the distributed communication signal detection device based on generalized likelihood ratio provided in an embodiment of the present invention is shown.

[0027] Figure 3 A schematic diagram of the architecture of a distributed communication signal detection system based on generalized likelihood ratio provided by an embodiment of the present invention is shown.

[0028] Figure 4This illustrates the impact of different modulation methods on the detection probability of a distributed communication signal detection radiation source;

[0029] Figure 5 The performance comparison of distributed communication signal (BPSK modulation) detection versus single-node communication signal detection is shown, as well as the impact of the number of nodes on the detection probability.

[0030] Figure 6 The effect of noise fluctuations at the receiving node on the detection probability is shown in the distributed communication signal (BPSK modulation) detection.

[0031] Figure 7 The impact of asynchronous reception by each node on detection performance in distributed communication signal (BPSK modulation) detection is illustrated. Detailed Implementation

[0032] Embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. These embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Although exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein.

[0033] Figure 1 This diagram illustrates a flowchart of a distributed communication signal detection method based on generalized likelihood ratio, provided by an embodiment of the present invention. This method is applied in a fusion center. The method of this embodiment includes:

[0034] S1: Receives observation data transmitted from multiple distributed receiving nodes.

[0035] In actual testing, multiple receiving nodes, which are distributed in a decentralized manner, collect communication signals from the surrounding environment in real time to obtain raw observation data, and transmit this raw observation data to the fusion center in real time. The fusion center receives the observation data transmitted by multiple receiving nodes.

[0036] S2: Based on the pre-constructed channel model and received signal model, under the two assumptions of no radiation source and the presence of a radiation source, the joint conditional probability density function of all observation data of all receiving nodes during the observation time is constructed respectively.

[0037] The fusion center constructs joint conditional probability density functions for all observation data received from all receiving nodes within the observation time, based on the pre-built channel model and received signal model, under two assumptions: the absence of a radiation source and the presence of a radiation source.

[0038] The fusion center pre-constructs the channel model and the received signal model using the following method:

[0039] In distributed communication signal detection, assuming there is a radiation source, there are N receiving nodes distributed in a dispersed manner receiving the radiation source signal from different directions, resulting in N spatial diversity channels.

[0040] Assuming the radiation source signal is a narrowband signal, and when both the transmitter and receiver have only one antenna each, and the transmitter and receiver are relatively stationary, the channel model from the transmitter to the l-th receiving node is as follows: It is an unknown complex number that follows a Rice distribution; the generated radiation source signal is attenuated by the channel and superimposed with Gaussian white noise at the receiving node. The superimposed white noise has a mean of 0 and a variance of . The complex Gaussian distribution;

[0041] Radiation source detection can be formulated as a hypothesis testing problem, i.e., there are two hypotheses: This indicates that the radiation source does not exist. Indicating the presence of a radiation source, under both assumptions, the signal received by the l-th receiving node during the observation time can be expressed as:

[0042]

[0043] In the formula, l = 1,...,N, t = 1,...,T, where T is the number of samples collected by each receiving node during the observation time; y l (t) The observation data received by the l-th receiving node at time t; n l (t) represents the received noise of the l-th receiving node during the observation time, which follows a mean of 0 and a variance of . The noise follows a complex Gaussian distribution, and the received noise at different receiving nodes is statistically independent of each other; h l Let be the channel parameters from the radiation source to the l-th receiving node. These are unknown complex parameters that follow a Rice distribution and are used to describe the channel propagation effect; s(t) represents the radiation source signal.

[0044] Based on the pre-constructed channel model and received signal model, in and Under the two assumptions, at a certain time (time t), the joint conditional probability density function of the signals received by the N receiving nodes is as follows:

[0045]

[0046] Therefore, in and Under the assumption that the fusion center constructs the joint conditional probability density function for all observation data from N receiving nodes within the observation time, the following are given:

[0047]

[0048] In the formula, express Assume the joint conditional probability density function value of all observation data from all receiving nodes within the observation time. express Assume the joint conditional probability density function value of all observation data from all receiving nodes within the observation time. U is Assume that the covariance matrix V of all observation data from all receiving nodes within the observation time is... Let det(·) represent the covariance matrix of all observation data from all receiving nodes within the observation time; tr(·) represents the determinant of the matrix and tr(·) represents the trace of the matrix.

[0049] S3: Based on the joint conditional probability density function under the two assumptions, the NP criterion is used to construct the generalized likelihood ratio test statistic expression, where the unknown variable in the generalized likelihood ratio test statistic expression is the covariance matrix in the joint conditional probability density function under the two assumptions.

[0050] In signal statistical detection theory, commonly used detector design methods are mainly based on Bayesian criteria, derived Bayesian criteria, minimax criteria, and Neyman-Pearson (NP) criteria. In distributed communication signal detection, the probability of target occurrence is often unknown a priori, nor can the cost of different decisions be predetermined. Therefore, the NP criterion is adopted for distributed communication signal detector design. The principle of the NP criterion is to maximize the detection probability while constraining the false alarm probability to a certain extent; the corresponding optimal detector is the Likelihood Radio (LR) detector.

[0051] In practice, it's impossible to know the likelihood function precisely because it contains one or more uncertain parameters. For example, the power of noise. and channel parameter h l When the unknown parameters exist, the hypothesis test is called a composite hypothesis test.

[0052] The joint conditional probability density function constructed in step S2 above is a likelihood function, in which the covariance matrices U and V are unknown variables. Therefore, this invention considers the Generalized Likelihood Radio (GLR) detector. For the joint conditional probability density function under the two hypotheses, the GLR test statistic expression is constructed using the NP criterion as follows:

[0053]

[0054] By taking the logarithm, the above expression for the GLR test statistic can be simply expressed as:

[0055]

[0056] In the formula ξ GLR This represents the generalized likelihood ratio test statistic; This represents the covariance matrix of all observations from all receiving nodes within the observation time. S is the known variable in the GLR test statistic expression, and U and V are the unknown variables in the GLR test statistic expression.

[0057] S4: Estimate the covariance matrix in the joint conditional probability density function under the two assumptions according to the maximum likelihood criterion.

[0058] To solve for the unknown variables U and V in the expression of the GLR test statistic, consider Assume the following distributed signal reception model:

[0059] y(t)=Hs(t)+n(t)

[0060] In the formula, s(t) indicates the existence of one signal source, and the mean of s(t) is generally zero; Let represent the channel parameters; n(t) represent the N-dimensional noise vector, which follows a complex Gaussian distribution with zero mean, i.e., n(t) ~ CN(0,Σ), where the covariance matrix Σ is a diagonal matrix; y(t) represent the received signal vector, which follows a Gaussian distribution with zero mean, and its covariance matrix is ​​R. R = HH H +Σ, meaning the covariance matrix of the received signal y(t) is a rank-1 matrix plus a diagonal matrix. in

[0061] The maximum likelihood estimation problem is transformed into the following optimization problem:

[0062]

[0063] stR = HH H +Σ

[0064]

[0065] In the formula, R is the covariance matrix to be solved, and the received signal is... t represents time, t = 1, 2, ..., T. Let represent the power of the received noise at each receiving node, det(·) represent the determinant of the matrix, and tr(·) represent the trace of the matrix; Assuming H = 0; in Assuming, This represents the channel parameters from the radiation source to each receiving node, where N represents the number of receiving nodes;

[0066] The above optimization problem does not have a closed-form solution, but a local maximum can be found using an alternating optimization method. This invention uses an alternating optimization method to find the local maximum of the above optimization problem, obtaining the covariance matrix R to be solved; in Assuming that the R value is an estimate of V. exist Assuming that the value of R when H=0 is an estimate of U.

[0067] Specifically, the present invention uses the following iterative algorithm to solve the above optimization problem and find the local maximum value of the optimization problem:

[0068] S41: k = 0, random initialization

[0069] S42: k = k + 1;

[0070] S43: According to renew

[0071] S44: According to renew

[0072] S45: judge If convergence has not occurred, return to step S42; if convergence has occurred, proceed to the next step.

[0073] S46: Output

[0074] The convergence condition for the above iterative algorithm is:

[0075]

[0076] In the formula, ε represents the convergence accuracy. Arranged in descending order eigenvalues, Q is any orthogonal matrix.

[0077] exist Under the assumption that the unknown parameter V is an ML estimate This is what the above algorithm yielded. Value, that is

[0078] exist Under the assumption that the ML estimate of the unknown parameter U is obtained by substituting H=0 into the above algorithm, The estimated value, i.e.

[0079] S5: Substitute the estimated values ​​of the covariance matrix under the two assumptions into the expression of the generalized likelihood ratio test statistic, and compare the obtained generalized likelihood ratio test statistic with the decision threshold set according to the false alarm probability. When the generalized likelihood ratio test statistic is greater than the decision threshold, it is determined that there is a radiation source in the detection frequency band; otherwise, it is determined that there is no radiation source in the detection frequency band.

[0080] In signal source detection algorithms, a test statistic ξ is constructed by processing the sampled data, and a decision is made based on a predetermined decision threshold η. The decision criterion can be described as follows:

[0081]

[0082] When the test statistic is less than the decision threshold, it is judged as... (i.e., no signal source is detected within the frequency band); otherwise, it is judged as... (That is, assuming that a signal source exists in the detection frequency band). It can be seen that the detection performance of a signal source is generally evaluated by two indicators: detection probability and false alarm probability. The false alarm probability refers to the fact that, since noise always exists, when the amplitude of the noise signal exceeds the detection threshold, the radar (or other detection system) will mistakenly identify the target. This error is called "false alarm", and its probability of occurrence is called the false alarm probability.

[0083] The detection probability and false alarm probability are defined as follows:

[0084]

[0085]

[0086] As can be seen from the above two equations, the detection probability and the false alarm probability are a pair of contradictory evaluation parameters. A detection algorithm with good detection performance should have a high detection probability and a low false alarm probability.

[0087] Based on the preceding steps, the generalized likelihood ratio (GLR) detector constructed in this invention is as follows:

[0088]

[0089] In the formula η GLR η represents the detection threshold of the generalized likelihood ratio detector. GLR Configure this setting when the false alarm probability is specified.

[0090] The GLR detector can be considered a suboptimal option, as it will store the unknown variables in the likelihood function (the joint conditional probability density function constructed in the S2 step). and Assuming that the covariance matrices U and V of the data received by each node are used for maximum likelihood estimation (MLE), then the estimator ( and Substitute the likelihood ratio (the generalized likelihood ratio test statistic constructed in step S3) into the sample. In the expression.

[0091] That is, to Assume the estimated value of V The estimated value of U is assumed to be... Substituting the test statistic ξ of the generalized likelihood ratio detector GLR The expression, if ξ GLR >η GLR Then it is determined to be That is, if a radiation source exists within the detection frequency band, otherwise it is determined to be... This means there is no radiation source within the detection frequency band. This completes the distributed communication signal detection of this invention.

[0092] and Figure 1 The distributed communication signal detection method based on generalized likelihood ratio shown belongs to the same technical concept. The embodiments of the present invention also provide a distributed communication signal detection device based on generalized likelihood ratio. Figure 2 This diagram illustrates the structure of a distributed communication signal detection device based on generalized likelihood ratio provided in an embodiment of the present invention. The device is applied in a fusion center, such as... Figure 3 As shown, it includes the following modules:

[0093] The receiving module 21 is used to receive observation data transmitted by multiple receiving nodes that are distributed in a dispersed manner.

[0094] The probability density function construction module 22 is used to construct the joint conditional probability density function of all observation data of all receiving nodes within the observation time, based on the pre-constructed channel model and received signal model, under the two assumptions of no radiation source and the presence of radiation source.

[0095] The test statistic construction module 23 is used to construct a generalized likelihood ratio test statistic expression based on the joint conditional probability density function under the two constructed assumptions and using the NP criterion. The unknown variable in the generalized likelihood ratio test statistic expression is the covariance matrix in the joint conditional probability density function under the two assumptions.

[0096] The covariance matrix estimation module 24 is used to estimate the covariance matrix in the joint conditional probability density function under the two assumptions according to the maximum likelihood criterion.

[0097] The comparison and decision module 25 is used to substitute the estimated value of the covariance matrix under the two assumptions into the expression of the generalized likelihood ratio test statistic, and compare the obtained generalized likelihood ratio test statistic with the decision threshold set according to the false alarm probability. When the generalized likelihood ratio test statistic is greater than the decision threshold, it is determined that there is a radiation source in the detection frequency band; otherwise, it is determined that there is no radiation source in the detection frequency band.

[0098] Figure 2 The implementation process of each module in the device shown can be found in the aforementioned method embodiments, and will not be repeated here.

[0099] and Figure 1 The distributed communication signal detection method based on generalized likelihood ratio shown belongs to the same technical concept. The embodiments of the present invention also provide a distributed communication signal detection system based on generalized likelihood ratio. Figure 3 This diagram illustrates the architecture of a distributed communication signal detection system based on generalized likelihood ratio, as provided in an embodiment of the present invention. Figure 3 As shown, it includes a fusion center 30 and multiple receiving nodes (31, 32, 33, ...) arranged in a distributed manner.

[0100] The multiple receiving nodes (31, 32, 33, ...) respectively collect communication signals from the surrounding environment in real time to obtain observation data, and transmit the collected observation data to the fusion center 30;

[0101] The fusion center 30 includes a memory and a processor. The fusion center stores the observation data transmitted by multiple receiving nodes (31, 32, 33, ...) in the memory. The memory also stores a computer program, which is loaded and executed by the processor to implement the aforementioned distributed communication signal detection method based on generalized likelihood ratio.

[0102] Figure 3 The implementation process of the fusion center 30 in the system shown can be found in the aforementioned method embodiments, and will not be repeated here.

[0103] The present invention also proposes a computer-readable storage medium that stores one or more computer programs, which, when executed by a processor, implement the aforementioned distributed communication signal detection method based on generalized likelihood ratio, which will not be elaborated further here.

[0104] In summary, the distributed communication signal detection method, apparatus, and system based on generalized likelihood ratio provided by this invention, due to its multi-node distributed detection, improves the detection probability of communication signals under low signal-to-noise ratios compared to a single receiving node. Theoretically, even under low signal-to-noise ratios, good detection performance can be achieved when there are enough receiving nodes. Furthermore, this invention utilizes the correlation between data from multiple receiving nodes to estimate the unknown variables in the GLR test statistic constructed using the maximum likelihood (ML) criterion, obtaining an estimate of the covariance matrix in the joint conditional probability density function under the two hypotheses. Therefore, when the noise of the receiving node fluctuates, the method of this invention exhibits good noise resistance.

[0105] The GLR detection method provided by this invention will be further illustrated below through simulation experiments.

[0106] First, a simulated radiation source generates communication signals. The communication signal transmission and reception process includes baseband signal generation, shaping and filtering, modulation, and the modulated signal passing through different spatial diversity channels to reach the receiving node. After demodulation and filtering, the received data is obtained at each node. Modulation includes the generation of various modulation signals such as BPSK modulation and QPSK modulation.

[0107] Taking Quadrature Phase Shift Keying (QPSK) modulation as an example, QPSK modulation represents the waveform changes of the baseband signal by varying the phase of the carrier signal. In QPSK modulation, in the bit sequence to be transmitted, every two consecutive bits are grouped together to form a quaternary symbol, or a two-bit symbol. The four states of the two-bit symbol are represented by four different phases of the carrier (k = 1, 2, 3, 4). This correspondence is called phase logic. The transmitting end first generates a bit stream, which is then converted from serial to parallel and divided into two paths (I-path and Q-path). These paths are then shaped and filtered. Odd-numbered bits are fed into the I-path and paralleled with cos(ω). o Multiply by t, enter Q path with even numbers and combine with sin(ω) o Multiply the I signal by the Q signal, and then subtract the Q signal from the I signal to obtain the QPSK modulated signal.

[0108] BPSK modulated signals represent the waveform changes of baseband signals by varying the phase of the carrier signal. m(t) represents the baseband signal with amplitudes of -1 and 1. Therefore, a BPSK signal can be represented as:

[0109]

[0110] In the following simulation experiments, the communication signal modulation method is BPSK modulation, the carrier frequency is 1.42e9Hz, the symbol rate is 20kHz, the observation time per node is 0.5s, the sampling rate is 120kHz, the low-pass filter has a passband attenuation factor of 4 and a stopband attenuation factor of 20. The Monte Carlo simulation was performed 600 times. The accuracy of the unknown parameters in the maximum likelihood estimation (MLE) is ε = 10. -6 To compare detection performance, the fusion center used detection methods including single-node energy detection (Single-node ED), multi-node energy detection (Multi-node ED), and the multi-node generalized likelihood ratio detection (Multi-node GLR) of this invention, with the number of nodes set to 3, 5, and 8, respectively.

[0111] First, the impact of different modulation methods on the detection performance of the radiation source signal is analyzed. Then, the detection performance of the above detection methods is compared. The impact of the number of nodes on the detection performance and the impact of noise fluctuations of the receiving nodes on the detection performance are analyzed. Finally, the impact of asynchronous reception on the detection performance is also analyzed.

[0112] Simulation Scenario 1: All nodes are assumed to have the same signal-to-noise ratio, with no time delay error. The number of nodes is set to 8. The radiation source signal uses BPSK / PAM4 / CPFSK modulation. The GLR detection method of this invention is used to observe the impact of different modulation schemes on the detection probability. Simulation results are as follows: Figure 4 As shown, Figure 4 The effect of different modulation methods on the detection probability of a distributed communication signal detection radiation source is shown.

[0113] from Figure 4 The simulation results show that different modulation methods used in the radiation source signal have some impact on the detection performance, but it is not significant.

[0114] Simulation Scenario 2: Based on the simulation parameters above, we set the signal-to-noise ratio of all receiving nodes to be consistent, with no time delay error and no fluctuation in noise energy at each receiving node, to analyze the impact of the number of nodes on detection performance. The simulation results are as follows: Figure 5 As shown, Figure 5 The performance of distributed communication signal (BPSK modulation) detection compared to single-node communication signal detection is shown, as well as the impact of the number of nodes on the detection probability.

[0115] from Figure 5The comparison results show that: 1) In a distributed detection system, changes in the signal-to-noise ratio will affect the detection performance, and the detection probability will increase with the increase of the signal-to-noise ratio; 2) The detection performance of multi-node systems is better than that of single-node systems; 3) The detection performance is related to the number of nodes, and the more nodes there are, the better the detection performance; 4) Under ideal noise conditions, the performance difference between the multi-node energy detection method and the GLR detection method of this invention is not significant.

[0116] Simulation Scenario 3: All nodes have a consistent signal-to-noise ratio and no time delay error. The noise energy of the receiving nodes fluctuates. The impact of this noise fluctuation on the detection probability is observed. To describe the noise fluctuation of each receiving node, σ is used. 2 σ represents the average power level of the noise at the receiving node. 2 =10 -b / 10 , b∈[-1,1], b follows a uniform distribution within the interval. Simulation results are as follows: Figure 6 As shown, Figure 6 The effect of noise fluctuations at the receiving node on the detection probability is shown in the distributed communication signal (BPSK modulation) detection.

[0117] from Figure 6 The simulation results show that in a distributed detection system, noise fluctuations at each node have a significant impact on detection performance. When the noise levels of each node differ, the overall noise level of the system becomes unbalanced, affecting the signal-to-noise ratio (SNR). Conversely, when the noise fluctuations at each node are excessive, the overall noise level of the system becomes very high, impacting detection performance. (Comparison) Figure 6 and Figure 5 The simulation results of (Simulation Scenario 2) show that the GLR detection method of the present invention has good anti-noise performance when the actual noise fluctuates.

[0118] Simulation Scenario 4: All nodes have a consistent signal-to-noise ratio, but there is a time delay error. The noise energy of each receiving node fluctuates. The impact of asynchronous reception at each node on detection performance is observed. To describe the different delays in the radiation source signal reaching each receiving node, the arrival delays are randomly generated. k follows a uniform distribution within this interval. To describe the fluctuation of noise at each receiving node, σ is used. 2 σ represents the average power level of the noise at the receiving node. 2 =10 -b / 10 , b∈[-1,1], b follows a uniform distribution within the interval. Simulation results are as follows: Figure 7 As shown, Figure 7 The impact of asynchronous reception by each node on detection performance in distributed communication signal (BPSK modulation) detection is illustrated.

[0119] from Figure 7As can be seen from the simulation diagram, in a distributed detection system, asynchronous data received by each node will impact detection performance. This is because when the data from each node is asynchronous, the overall data of the system loses consistency, and the system cannot accurately calculate correlations. This will affect the system's detection accuracy. Therefore, to improve the performance of a distributed detection system, it is usually necessary to take measures to control data synchronization, such as implementing a precise time synchronization mechanism or using data synchronization algorithms. Figure 7 and Figure 6 The simulation results of (Simulation Scenario 3) show that when the received data is asynchronous, the detection performance will decrease when the data of each node is not processed by time synchronization.

[0120] Finally, it should be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0121] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for distributed communication signal detection based on generalized likelihood ratio, characterized in that, The method includes the following steps: S1: Receive observation data transmitted by multiple distributed receiving nodes; S2: Based on the pre-built channel model and received signal model, under the two assumptions of no radiation source and the presence of radiation source, the joint conditional probability density function of all observation data of all receiving nodes during the observation time is constructed respectively. S3: based on the constructed joint conditional probability density function under the two hypotheses, adopting criterion constructs a generalized likelihood ratio test statistic expression, the unknown variable in the generalized likelihood ratio test statistic expression is the covariance matrix in the joint conditional probability density function under the two hypotheses; S4: Estimate the covariance matrix in the joint conditional probability density function under the two assumptions according to the maximum likelihood criterion; S5: Substitute the estimated values ​​of the covariance matrix under the two assumptions into the expression of the generalized likelihood ratio test statistic, and compare the obtained generalized likelihood ratio test statistic with the decision threshold set according to the false alarm probability. When the generalized likelihood ratio test statistic is greater than the decision threshold, it is determined that there is a radiation source in the detection frequency band; otherwise, it is determined that there is no radiation source in the detection frequency band.

2. The method of claim 1, wherein, The channel model and the received signal model are constructed in advance using the following method: In distributed communication signal detection, assuming that there is a radiation source, there are several distributedly arranged receiving nodes receiving the radiation source signal from different directions to obtain a number of spatial diversity channels; Assuming the radiation source signal is a narrowband signal, and when the transmitter and receiver each have only one antenna, and the transmitter and receiver are relatively stationary, the signal from the transmitter to the... The channel model for each receiving node is as follows: The signal is an unknown complex number that follows a Rice distribution; the generated radiation source signal is attenuated by the channel and superimposed with Gaussian white noise at the receiving node. The superimposed white noise has a mean of 0 and a variance of . The complex Gaussian distribution; The radiation source detection formulation is cast as a hypothesis testing problem, i.e., there are two hypotheses: H0: The radiation source is not present, H1: The radiation source is present, then the signal received by the first receiving node in the observation time is given by: ; In the formula, , , This represents the number of samples collected by each receiving node during the observation period. No. Each receiving node The observation data received at any time; For the first The received noise of each receiving node during the observation time follows a mean of 0 and a variance of . The noise follows a complex Gaussian distribution, and the received noise at different receiving nodes is statistically independent of each other; From the radiation source to the first The channel parameters of each receiving node are unknown complex parameters that follow a Rice distribution and are used to describe the channel propagation effect. This indicates the signal from the radiation source.

3. The method of claim 2, wherein, Step S2 specifically includes: In the above and Assuming the above, the joint conditional probability density function of all observation data of the K receiving nodes within the observation time is respectively: ​ ; In the formula, express Assume the joint conditional probability density function value of all observation data from all receiving nodes within the observation time. express Assume the joint conditional probability density function value of all observation data from all receiving nodes within the observation time. ; for Assume the covariance matrix of all observation data from all receiving nodes within the observation time. for Assume the covariance matrix of all observation data from all receiving nodes within the observation time. express The determinant, Represents the trace of a matrix.

4. The method of claim 3, wherein, Step S3 specifically includes: For the joint conditional probability density function under the two hypotheses, we use The generalized likelihood ratio test statistic is constructed as follows: ; By taking the logarithm, the expression for the generalized likelihood ratio test statistic above can be simply expressed as: ; wherein denotes a generalized likelihood ratio test statistic; denotes the covariance matrix of all observation data of all receiving nodes in the observation time, is a known variable; and is an unknown variable in the expression of the generalized likelihood ratio test statistic.

5. The method of claim 4, wherein, Step S4 specifically includes: Consider Assume the following distributed received signal model: ; In the formula, Indicates existence One signal source, The mean is zero; Indicates channel parameters; express The noise vector follows a complex Gaussian distribution with a mean of zero, i.e. The covariance matrix It is a diagonal matrix; The received signal vector follows a Gaussian distribution with zero mean, and its covariance matrix is: , , That is, receiving signals The covariance matrix is ​​a rank-1 matrix plus a diagonal matrix. ,in ; The maximum likelihood estimation problem is transformed into the following optimization problem: ; In the formula, Given the covariance matrix to be solved, and the received signal... , , Indicates time, , , … These represent the power of the noise received by each receiving node. Represents the determinant of a matrix. Represents the trace of a matrix; in Assuming, ;exist Assuming, , representing the channel parameters from the radiation source to each receiving node. Indicates the number of receiving nodes; By using the alternating optimization method, the local maxima of the above optimization problem are found, and the covariance matrix to be solved is obtained. ;exist Assuming, Value is The estimated value ;exist Assuming, Below Value is The estimated value .

6. The method of claim 5, wherein, The local maximum of the optimization problem is found by an alternating optimization method, and a covariance matrix to be solved is obtained Specifically comprising: The following iterative algorithm is used to solve the above optimization problem and find its local maximum: S41: Random initialization , ; S42： ； S43: In accordance with update ; S44: In accordance with update ; S45: , determine whether the convergence is achieved, if not, return to S42, if yes, go to next step; S46: Output ; The convergence condition for the above iterative algorithm is: ; In the formula, represents the convergence precision, , , are arranged in descending order characteristic values of the matrix, , is an arbitrary orthogonal matrix.

7. The method of claim 5, wherein, Step S5 specifically includes: The generalized likelihood ratio detector is constructed as follows: ; In the formula denotes the detection threshold of the generalized likelihood ratio detector, which is set in the case where the false alarm probability is specified; Will Assuming The estimated value , Assuming The estimated value Substituting the test statistic of the generalized likelihood ratio detector The expression, if Then it is determined to be That is, if a radiation source exists within the detection frequency band, otherwise it is determined to be... This means that there is no radiation source within the detection frequency band.

8. A generalized likelihood ratio based distributed communication signal detection apparatus, comprising: The device is used in the fusion center and includes the following modules: The receiving module is used to receive observation data transmitted by multiple receiving nodes that are distributed in a dispersed manner. The probability density function construction module is used to construct the joint conditional probability density function of all observation data of all receiving nodes within the observation time, based on the pre-built channel model and received signal model, under the two assumptions of no radiation source and the presence of radiation source. a test statistic construction module, configured to construct, based on the constructed joint conditional probability density function under the two hypotheses, a criterion to construct a generalized likelihood ratio test statistic expression, wherein an unknown variable in the generalized likelihood ratio test statistic expression is a covariance matrix in the joint conditional probability density function under the two hypotheses; The covariance matrix estimation module is used to estimate the covariance matrix in the joint conditional probability density function under the two assumptions according to the maximum likelihood criterion. The comparison and decision module is used to substitute the estimated value of the covariance matrix under the two assumptions into the expression of the generalized likelihood ratio test statistic, and compare the obtained generalized likelihood ratio test statistic with the decision threshold set according to the false alarm probability. When the generalized likelihood ratio test statistic is greater than the decision threshold, it is determined that there is a radiation source in the detection frequency band. Conversely, if the test result is negative, it is determined that there is no radiation source within the detection frequency band.

9. A distributed communication signal detection system based on generalized likelihood ratio, characterized by, This includes a fusion center and multiple distributed receiving nodes. The multiple receiving nodes collect communication signals from the surrounding environment in real time to obtain observation data, and transmit the collected observation data to the fusion center. The fusion center includes a memory and a processor. The fusion center stores observation data transmitted by multiple receiving nodes in the memory. The memory also stores a computer program, which is loaded and executed by the processor to implement the distributed communication signal detection method based on generalized likelihood ratio as described in any one of claims 1 to 7.

10. A computer readable storage medium storing one or more computer programs which, when executed by a processor, implement the generalized likelihood ratio based distributed communication signal detection method of any one of claims 1-7.

Images