Radar radiation source weak signal detection method based on unmanned aerial vehicle group cooperation

Abstract

The invention discloses a radar radiation source weak signal detection method based on unmanned aerial vehicle group cooperation, and the method comprises the steps: firstly building a single unmanned aerial vehicle detection model based on energy detection, considering the Rayleigh fading characteristics of a channel, building a Rayleigh channel model, and analyzing the false alarm probability and the detection probability; and then providing a method for cooperatively detecting a radar radiation source by multiple unmanned aerial vehicles based on a local detection + information fusion architecture, establishing a global detection performance evaluation model, analyzing the performance of a detection system, and realizing signal detection. According to the model and the method provided by the invention, the detection probability of a radiation source target by unmanned aerial vehicle group collaboration in a low signal-to-noise ratio scene is effectively improved, the technical framework of the field of multi-unmanned aerial vehicle collaboration reconnaissance is enriched theoretically, and an effective technical means is provided in practical application; the method is used for coping with reconnaissance challenges in a low signal-to-noise ratio environment, and has a wide application prospect for improving the reliability and accuracy of a reconnaissance system.

Description

Technical Field

[0001] The present invention belongs to the technical field of radar target detection, and in particular relates to a method for detecting weak signals of radar radiation sources based on the collaboration of a group of unmanned aerial vehicles. Background Art

[0002] In the field of civilian surveillance, UAV swarms play an increasingly important role. They greatly enhance the depth and breadth of intelligence collection by providing real-time, highly maneuverable aerial surveillance. However, the execution of surveillance tasks is often constrained by low signal-to-noise ratio (SNR) environments, where target signals are often masked by background noise, resulting in reduced detection performance. This challenge has prompted researchers to explore more efficient collaborative reconnaissance strategies to improve reconnaissance effectiveness in complex signal environments.

[0003] Multi-UAV collaborative reconnaissance is a method to improve reconnaissance efficiency through the joint action of multiple UAVs. Compared with a single UAV, a multi-UAV system has significant advantages in target detection, data collection and information processing. First, multiple UAVs can observe targets from different angles and distances at the same time, increasing the diversity and reliability of information. Secondly, through information sharing and collaborative processing, the false alarm rate of individual UAVs can be effectively reduced and the detection capability of the overall system can be improved. In addition, the multi-UAV system has strong anti-interference ability and flexibility, and can adapt to dynamically changing environments.

[0004] In low signal-to-noise ratio environments, target signals are often drowned out by background noise, and existing signal detection methods have difficulty effectively identifying targets. In this case, weak signal detection of radar emitters becomes a major problem. The Rayleigh fading characteristics of the channel further exacerbate this challenge, making signal propagation and reception more unstable. In addition, during the flight of a drone, the channel conditions will also change due to changes in position and speed, which places higher requirements on the robustness of the detection algorithm.

[0005] Energy detection is a commonly used signal detection method. It is widely used under low signal-to-noise ratio conditions due to its simplicity and high efficiency. This method determines the presence of a target by detecting whether the signal energy exceeds a set threshold. However, single energy detection is susceptible to noise in a low signal-to-noise ratio environment. For this reason, information fusion technology is introduced to form a more accurate global decision by integrating the detection results of multiple drones. Information fusion can significantly improve the detection probability and reduce the false alarm rate. It is one of the key technologies to improve the collaborative monitoring task of multiple drones.

[0006] Currently, researchers are actively exploring new models and algorithms for multi-UAV collaborative monitoring to meet the challenges of reconnaissance in complex environments. Key research directions include: optimizing UAV collaboration strategies, enhancing the intelligence and adaptability of information fusion algorithms, developing detection technologies that adapt to time-varying channels, and using advanced simulation and numerical methods to improve the accuracy and efficiency of algorithms. At the same time, with the introduction of artificial intelligence and machine learning technologies, the intelligence level of multi-UAV collaborative reconnaissance systems will be further improved, providing stronger technical support for radar emitter signal detection.

[0007] This multi-level and multi-angle technological progress not only enriches the theoretical framework of UAV cooperative reconnaissance, but also provides a solid technical foundation for practical applications, and has broad application prospects. In the future, multi-UAV cooperative reconnaissance technology will continue to play an important role in complex and changing environments, and promote the development of reconnaissance technology in a more efficient and intelligent direction. Summary of the invention

[0008] In order to solve the above technical problems, the present invention proposes a weak signal detection method for radar radiation source based on UAV swarm collaboration to cope with the reconnaissance challenges in low signal-to-noise ratio environments and improve the reliability and accuracy of the reconnaissance system.

[0009] The technical solution adopted by the present invention is: a method for detecting weak signals of radar radiation sources based on the collaboration of drone groups, the steps are as follows:

[0010] S1. Establish a single UAV detection probability model based on energy detection and output local judgment results;

[0011] S2. Establish a Rayleigh channel model and use Monte Carlo experiments to observe and analyze the impact of time-varying channels and time-invariant channels on detection performance;

[0012] S3, based on the channel model established in step S2, construct a local detection performance evaluation model;

[0013] S4. Based on the local detection results of step S3, a method for collaborative detection of weak signals of multiple UAV radar radiation sources with a local detection + decision fusion architecture is proposed;

[0014] S5. Based on step S4, a global detection performance evaluation model is established to analyze the performance of the detection system and realize signal detection.

[0015] Furthermore, the step S1 is specifically as follows:

[0016] Suppose a network consisting of N drones, the number of sampling points in a detection interval is L, and the observation value x received by the i-th drone is i (k) The expression is as follows:

[0017]

[0018] Where k = 1, 2, ..., L, n i (k) means that both the real and imaginary parts have a mean of 0 and a variance of complex Gaussian white noise, a(k) represents the target signal, h i (k) represents the transmission channel gain, i=1,2,...,N.

[0019] When using energy detector, the energy statistics E at the i-th UAV is i The expression is as follows:

[0020]

[0021] Let T represent the decision threshold, u i Represents local decision making, and the decision rule expression of each drone is as follows:

[0022]

[0023] When the number of sampling points L in the detection interval is large enough, according to the central limit theorem, the energy statistic E in formula (2) i Under the assumptions H0 and H1, Gaussian distribution is used to represent the target signal energy E a The expression is as follows:

[0024]

[0025] The corresponding expectation and variance expressions are as follows:

[0026]

[0027] Among them, e0 and d0 represent the corresponding expectation and variance under the assumption H0, e1 and d1 represent the corresponding expectation and variance under the assumption H1, and η i It represents the instantaneous signal-to-noise ratio of the signal received by the i-th UAV. The specific expression is as follows:

[0028]

[0029] in, represents the average signal-to-noise ratio.

[0030] Then the false alarm probability P of the i-th UAV at a given threshold T is f,i , detection probability P d,i The expressions are as follows:

[0031]

[0032] Among them, P(·) represents probability, Q(·) represents the standard Q function, and the expression is as follows:

[0033]

[0034] Among them, x represents the lower limit of integration, and m represents the integral variable. Then consider the transmission channel gain h i (k) is time-invariant within the perception interval, i.e. h i (k) = h i ,k=1,2,...,L, the expression is as follows:

[0035]

[0036] Finally, the detection probability model expression of a single UAV based on energy detection is as follows:

[0037]

[0038] in,

[0039]

[0040] Among them, P f Represents a given false alarm probability value.

[0041] Furthermore, the step S2 is specifically as follows:

[0042] The Rayleigh channel model is used to simulate the process of transmitting the decision results of the drone to the decision center. It is assumed that the signals received by each drone are independent of each other. The probability density model expression of the Rayleigh channel is as follows:

[0043]

[0044] Where r represents the signal amplitude.

[0045] Then, the Monte Carlo method is used to perform numerical simulation to estimate the average detection probability under the influence of time-varying channels. Assume z i represents the result transmitted from the i-th reconnaissance aircraft to the decision center, v i It represents the decoding result of the fusion center for the transmission content of the i-th drone, and the definition expression is as follows:

[0046]

[0047] Among them, z i =y i +w i ,y i =g i u i , w i represents zero mean and variance is Gaussian variable, gi represents the channel gain between the decision center and each drone, and its probability density function f(g i ) is expressed as follows:

[0048]

[0049] Furthermore, the step S3 is specifically as follows:

[0050] The evaluation indicators of the local detection performance evaluation model include: local detection probability and local false alarm probability.

[0051] First, set the conditional probability density function expression as follows:

[0052]

[0053] Among them, the definition expression of U(·) is as follows:

[0054]

[0055] Among them, t represents the positive part.

[0056] Using signal detection theory to deduce, the expected and variance expressions of the detection results under various conditions are as follows:

[0057] E[v i |H j ]=P(v i =1|H j )×1+P(v i =0|H j )×0=P(u i ≥0|H j ) (18)

[0058]

[0059] Where j = 0, 1. From equation (19), we can see that the received local decision result v i Depends on the local false alarm probability and the detection probability, which are expressed as P F,i =f(u i ≥0|H0),P D,i =f(u i ≥0|H1), then u i The distribution expression of is as follows:

[0060]

[0061] Then it is deduced that when setting H j The equivalent probability expressions of false alarm probability and detection probability are as follows:

[0062]

[0063] Since f(z i |y i ,u i )=f(z i |y i ), then we can get:

[0064] f(z i |u i )=∫f(z i |y i ,u i )f(y i |u i )dy i (twenty three)

[0065] Then, given the local judgment u i = +1, f(z i |y i ) and f(z i |u i ) is substituted into equation (23) to obtain the received observation value z i The conditional probability density of is expressed as follows:

[0066]

[0067] make Substituting variables and using the integral properties of the Q function, we get the following expression:

[0068]

[0069] in, Q(-x)=1-Q(x), both sides of equation (25) are i Integrate and get the following expression:

[0070]

[0071] Similarly, when u i = -1, we can deduce that:

[0072]

[0073] P(z i ≥0|u i = +1) and P(z i ≥0|u i =-1) expression into P F,i and P D,i In the above equation, the local false alarm probability and detection probability are calculated and expressed as follows:

[0074]

[0075] Among them, ρ i Represents the signal-to-noise ratio of the transmission channel between each UAV and the decision-making center.

[0076] Furthermore, the step S4 is specifically as follows:

[0077] S41, based on the local detection result of step S3, evaluating and optimizing the detection performance of the drone;

[0078] In local detection, the false alarm and detection probability expressions of the i-th drone at a given threshold T are as follows:

[0079]

[0080] Then the relationship between the detection probability and the instantaneous signal-to-noise ratio is d,i (η i ) is expressed as follows:

[0081]

[0082] Among them, Q -1 (·) represents the inverse function of the standard Q function, η i Represents the instantaneous signal-to-noise ratio of the signal received by the drone.

[0083] S42, using the Bayesian global decision method to perform decision fusion and obtain a global detection result;

[0084] In decision fusion, the receiving channels from each UAV to the decision center are set to be independent and free from interference, and the delay in information transmission is ignored. i After that, according to the local decision result, decode to get v i , v i Independent of each other.

[0085] Then, the decision center adopts the Bayesian global decision method. The Bayesian global decision rule determines the global detection result by calculating the posterior probability. The posterior probability represents the probability of the target existing under the condition of the detection results of each drone. The expression is as follows:

[0086]

[0087] Among them, P(H1) represents the prior probability, P(v1,v2,...,v N |H1) represents the joint probability of the detection results of each drone under the condition that the target exists, P(v1,v2,...,v N ) represents the marginal probability.

[0088] Finally, the global decision rule expression is as follows:

[0089]

[0090] Since P(H1) and P(H0) are constants, the simplified expression is as follows:

[0091]

[0092] Assume that there are N1 1s in the local decision result obtained by decoding, and 0≤N1≤N, then

[0093]

[0094] Furthermore, the step S5 is specifically as follows:

[0095] The global detection performance evaluation model includes: a global false alarm probability evaluation model and a detection probability evaluation model.

[0096] Then the global false alarm probability evaluation model Q of the decision center is f And the detection probability evaluation model Q D The expressions are as follows:

[0097]

[0098] Among them, Ω represents the decision threshold of the decision center.

[0099] Finally, based on the global detection performance evaluation model, signal detection is achieved to improve the target detection probability.

[0100] Beneficial effects of the present invention: The method of the present invention first establishes a single UAV detection model based on energy detection, considers the Rayleigh fading characteristics of the channel, establishes a Rayleigh channel model and analyzes the false alarm probability and detection probability, and then proposes a method for collaborative detection of radar radiation sources by multiple UAVs based on a local detection + information fusion architecture, establishes a global detection performance evaluation model, analyzes the performance of the detection system, and realizes signal detection. The model and method proposed by the method of the present invention effectively improve the detection probability of the UAV group collaboratively detecting the radiation source target in the low signal-to-noise ratio scenario, which not only enriches the technical framework in the field of multi-UAV collaborative reconnaissance in theory, but also provides an effective technical means in practical applications to cope with the reconnaissance challenges in low signal-to-noise ratio environments, and has broad application prospects for improving the reliability and accuracy of the reconnaissance system. BRIEF DESCRIPTION OF THE DRAWINGS

[0101] Figure 1 The present invention is a flow chart of a method for detecting weak signals of radar radiation sources based on collaboration of a swarm of unmanned aerial vehicles.

[0102] Figure 2 1 is a flow chart of signal processing in an embodiment of the present invention.

[0103] Figure 3 Graph showing the relationship between the global detection probability and the number of drones in an embodiment of the present invention.

[0104] Figure 4 1 is a comparison diagram of detection probabilities under time-varying channels and time-invariant channels in an embodiment of the present invention. DETAILED DESCRIPTION

[0105] The method of the present invention is further described below with reference to the accompanying drawings and embodiments.

[0106] like Figure 1 As shown, a flow chart of a method for detecting weak signals of radar radiation sources based on UAV group collaboration of the present invention is shown, and the specific steps are as follows:

[0107] S1. Establish a single UAV detection probability model based on energy detection and output local judgment results;

[0108] S2. Establish a Rayleigh channel model and use Monte Carlo experiments to observe and analyze the impact of time-varying channels and time-invariant channels on detection performance;

[0109] S3, based on the channel model established in step S2, construct a local detection performance evaluation model;

[0110] S4. Based on the local detection results of step S3, a method for collaborative detection of weak signals of multiple UAV radar radiation sources with a local detection + decision fusion architecture is proposed;

[0111] S5. Based on step S4, a global detection performance evaluation model is established to analyze the performance of the detection system and realize signal detection.

[0112] In this embodiment, the step S1 is specifically as follows:

[0113] Suppose a network consisting of N drones, the number of sampling points in a detection interval is L, and the observation value x received by the i-th drone is i (k) The expression is as follows:

[0114]

[0115] Where k = 1, 2, ..., L, n i (k) means that both the real and imaginary parts have a mean of 0 and a variance of σ i 2 complex Gaussian white noise, a(k) represents the target signal, h i (k) represents the transmission channel gain, i=1,2,...,N.

[0116] When using energy detector, the energy statistics E at the i-th UAV is i The expression is as follows:

[0117]

[0118] Let T represent the decision threshold, u i Represents local decision making, and the decision rule expression of each drone is as follows:

[0119]

[0120] When the number of sampling points L in the detection interval is large enough, according to the central limit theorem, the energy statistic E in formula (2) i Under the assumptions H0 and H1, Gaussian distribution is used to represent the target signal energy E a The expression is as follows:

[0121]

[0122] The corresponding expectation and variance expressions are as follows:

[0123]

[0124] Among them, e0 and d0 represent the corresponding expectation and variance under the assumption H0, e1 and d1 represent the corresponding expectation and variance under the assumption H1, and η i It represents the instantaneous signal-to-noise ratio of the signal received by the i-th UAV. The specific expression is as follows:

[0125]

[0126] in, represents the average signal-to-noise ratio.

[0127] Then the false alarm probability P of the i-th UAV at a given threshold T is f,i , detection probability P d,i The expressions are as follows:

[0128]

[0129] Among them, P(·) represents probability, Q(·) represents the standard Q function, and the expression is as follows:

[0130]

[0131] Among them, x represents the lower limit of integration, and m represents the integral variable. Then consider the transmission channel gain h i (k) is time-invariant within the perception interval, i.e. h i (k) = h i ,k=1,2,...,L, the expression is as follows:

[0132]

[0133] Finally, the detection probability model expression of a single UAV based on energy detection is as follows:

[0134]

[0135] in,

[0136]

[0137] Among them, P f Represents a given false alarm probability value.

[0138] In this embodiment, step S2 is specifically as follows:

[0139] When the decision result of the drone is transmitted to the decision center, there will be noise and attenuation. The Rayleigh channel model is used to simulate this process. Assuming that the signals received by each drone are independent of each other, the probability density model expression of the Rayleigh channel is as follows:

[0140]

[0141] Where r represents the signal amplitude.

[0142] When faced with time-varying channel conditions, due to the dynamic changes in the channel state, it becomes extremely challenging to obtain an analytical solution for the detection probability based on the existing probability density function. In view of this, this embodiment uses the Monte Carlo method for numerical simulation to estimate the average detection probability under the influence of the time-varying channel. This method simulates the random changes of channel states through a large number of random sampling, and then statistically analyzes the detection performance under different signal-to-noise ratio conditions, providing an effective numerical solution for radar emitter signal detection in time-varying channels.

[0143] Considering that the decision coding received by the decision center from each drone may have bit errors, the decoding error w is defined as i To quantify this phenomenon. i represents the result transmitted from the i-th reconnaissance aircraft to the decision center, v i It represents the decoding result of the fusion center for the transmission content of the i-th drone, and the definition expression is as follows:

[0144]

[0145] Among them, z i =y i +w i ,y i =g i ui , w i represents zero mean and variance is Gaussian variable, g i represents the channel gain between the decision center and each drone, and its probability density function f(g i ) is expressed as follows:

[0146]

[0147] In this embodiment, step S3 is specifically as follows:

[0148] The local detection performance evaluation model is described by the local detection probability and the local false alarm probability. The expressions of the two probabilities are derived below.

[0149] First, set the conditional probability density function expression as follows:

[0150]

[0151] Among them, the definition expression of U(·) is as follows:

[0152]

[0153] Among them, t represents the positive part.

[0154] Using signal detection theory to deduce, the expected and variance expressions of the detection results under various conditions are as follows:

[0155] E[v i |H j ]=P(v i =1|H j )×1+P(v i =0|H j )×0=P(u i ≥0|H j ) (18)

[0156]

[0157] Among them, j = 0, 1, it can be seen from formula (19) that the received local decision result v i It depends on the local false alarm probability and the detection probability. For convenience, the local false alarm probability and the detection probability are expressed as P F,i =f(u i ≥0|H0),P D,i =f(u i ≥0|H1),u i The distribution expression of is as follows:

[0158]

[0159] Then it can be deduced that when setting H j The equivalent probability expressions of false alarm probability and detection probability are as follows:

[0160]

[0161]

[0162] Since f(z i |y i ,u i )=f(z i |y i ), then we can get:

[0163] f(z i |u i )=∫f(z i |y i ,u i )f(y i |u i )dy i (twenty three)

[0164] Then, given the local judgment u i = +1, f(z i |y i ) and f(z i |u i ) is substituted into equation (23) to obtain the received observation value z i The conditional probability density of is expressed as follows:

[0165]

[0166] make Substituting variables and using the integral properties of the Q function, we get the following expression:

[0167]

[0168] in, Q(-x)=1-Q(x), both sides of equation (25) are i Integrate and get the following expression:

[0169]

[0170] Similarly, when u i = -1, we can deduce that:

[0171]

[0172] P(z i ≥0|ui = +1) and P(z i ≥0|u i =-1) expression into P F,i and P D,i In the above equation, the local false alarm probability and detection probability are calculated and expressed as follows:

[0173]

[0174] Among them, ρ i Represents the signal-to-noise ratio of the transmission channel between each UAV and the decision-making center.

[0175] In this embodiment, step S4 is specifically as follows:

[0176] S41, based on the local detection result of step S3, evaluating and optimizing the detection performance of the drone;

[0177] In local detection, the false alarm and detection probability expressions of the i-th drone at a given threshold T are as follows:

[0178]

[0179] In the problem of radar radiation source target detection of UAVs, the instantaneous signal-to-noise ratio is a key parameter that directly affects the distribution characteristics of the detection statistics. Therefore, the detection threshold T can be analytically expressed, so that the relationship between the missed detection probability and the expected false alarm probability can be derived under the given false alarm probability constraint. Specifically, the detection threshold can be determined by setting the false alarm probability, and then the missed detection probability and detection probability under the threshold can be calculated based on the signal detection theory. This process provides an analytical method for the target detection problem, which enables the detection performance of UAVs to be evaluated and optimized while meeting the specific false alarm probability requirements. The relationship between the detection probability and the instantaneous signal-to-noise ratio P d,i (η i ) is expressed as follows:

[0180]

[0181] Among them, Q -1 (·) represents the inverse function of the standard Q function, η i Represents the instantaneous signal-to-noise ratio of the signal received by the drone.

[0182] S42, using the Bayesian global decision method to perform decision fusion and obtain a global detection result;

[0183] like Figure 2As shown in the signal processing flow chart, in decision fusion, the receiving channels from each UAV to the decision center are set to be independent and free from interference. At the same time, the delay in the information transmission process is ignored. The decision center receives the decision results of each UAV. i After that, according to the local decision result, decode to get v i , v i Based on this, the decision center adopts the Bayesian global decision method. The Bayesian global decision rule determines the global detection result by calculating the posterior probability. The posterior probability represents the probability of the target existing under the condition of the detection results of each drone. The expression is as follows:

[0184]

[0185] Among them, P(H1) represents the prior probability, P(v1,v2,...,v N |H1) represents the joint probability of the detection results of each drone under the condition that the target exists, P(v1,v2,...,v N ) represents the marginal probability.

[0186] Finally, the global decision rule expression is as follows:

[0187]

[0188] Since P(H1) and P(H0) are constants, the simplified expression is as follows:

[0189]

[0190] Assume that there are N1 1s in the local decision result obtained by decoding, and 0≤N1≤N, then

[0191]

[0192] In this embodiment, step S5 is specifically as follows:

[0193] The global detection performance evaluation model includes: a global false alarm probability evaluation model and a detection probability evaluation model.

[0194] Then the global false alarm probability evaluation model Q of the decision center is f And the detection probability evaluation model Q D The expressions are as follows:

[0195]

[0196] Among them, Ω represents the decision threshold of the decision center.

[0197] Finally, based on the global detection performance evaluation model, signal detection is achieved to improve the target detection probability.

[0198] This embodiment further performs simulation verification and analysis, as follows:

[0199] Any parameters L, N, η0 and γ0 set in the simulation of this embodiment are applicable to the derived local or global detection probability, so some classic parameters are used to draw the detection probability and false alarm probability curves to observe their impact on the system detection performance. Without loss of generality, consider fixing the position of each drone, and the number of samples in one observation time is L. The channel adopts the Rayleigh channel model. For all simulation cases, the channel vector can be modeled as a complex Gaussian random vector with zero mean. In addition, assume that the average signal-to-noise ratio of all drone receiving channels is the same, that is, η0 = -8dB, and set the channel error probability to 0.1, then γ0 = 5.5dB.

[0200] Figure 3 The results reveal the positive impact of the increase in the number of drones on the global detection probability under different instantaneous signal-to-noise ratio conditions. The results show that the increase in the number of drones significantly improves the accuracy of detection, especially in high signal-to-noise ratio environments, where the increase in detection probability is more significant.

[0201] Figure 4 The detection probability of drone swarms under time-varying channels is estimated by Monte Carlo simulation. The results show that Monte Carlo simulation can provide accurate detection probability estimation for drone swarms under complex channel conditions, providing important data support for decision-making in actual reconnaissance missions. The effectiveness of this method provides a feasible solution to the reconnaissance problem under time-varying channels.

[0202] In summary, the method of the present invention establishes a single UAV reconnaissance model based on energy detection, analyzes in detail the false alarm probability, detection probability and missed detection probability under Rayleigh fading channel, and explores the influence of instantaneous signal-to-noise ratio on detection performance, and further develops a multi-UAV collaborative detection strategy. By calculating the local and global false alarm probability and detection probability, it is revealed that the increase in the number of UAVs significantly improves the overall detection performance of the UAV group. The effectiveness of the model and method proposed in this embodiment is verified by the above simulation. The results show that with the increase in the number of UAVs, the global detection probability is significantly improved. At the same time, the influence of time-varying channels and time-invariant channels on the detection probability is simulated and analyzed, and the detection probability under time-varying channels is successfully estimated using the Monte Carlo method. The model and method proposed by the present invention effectively improve the detection probability of the UAV group to the radiation source target in the low signal-to-noise ratio scenario.

[0203] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A method for detecting weak signals of radar radiation sources based on UAV swarm collaboration, the steps are as follows: S1. Establish a single UAV detection probability model based on energy detection and output local judgment results; S2. Establish a Rayleigh channel model and use Monte Carlo experiments to observe and analyze the impact of time-varying channels and time-invariant channels on detection performance; S3, based on the channel model established in step S2, construct a local detection performance evaluation model; S4. Based on the local detection results of step S3, a method for collaborative detection of weak signals of multiple UAV radar radiation sources with a local detection + decision fusion architecture is proposed; S5. Based on step S4, a global detection performance evaluation model is established to analyze the performance of the detection system and realize signal detection.

2. According to the method for detecting weak signals of radar radiation sources based on UAV swarm collaboration according to claim 1, it is characterized in that: The step S1 is specifically as follows: Suppose a network consisting of N drones, the number of sampling points in a detection interval is L, and the observation value x received by the i-th drone is i (k) The expression is as follows: Where k = 1, 2, ..., L, n i (k) means that both the real and imaginary parts have a mean of 0 and a variance of complex Gaussian white noise, a(k) represents the target signal, h i (k) represents the transmission channel gain, i = 1, 2, ..., N; When using energy detector, the energy statistics E at the i-th UAV is i The expression is as follows: Let T represent the decision threshold, u i Represents local decision making, and the decision rule expression of each drone is as follows: When the number of sampling points L in the detection interval is large enough, according to the central limit theorem, the energy statistic E in formula (2) i Under the assumptions H0 and H1, Gaussian distribution is used to represent the target signal energy E a The expression is as follows: The corresponding expectation and variance expressions are as follows: Among them, e0 and d0 represent the corresponding expectation and variance under the assumption H0, e1 and d1 represent the corresponding expectation and variance under the assumption H1, and η i It represents the instantaneous signal-to-noise ratio of the signal received by the i-th UAV. The specific expression is as follows: in, represents the average signal-to-noise ratio; Then the false alarm probability P of the i-th UAV at a given threshold T is f,i , detection probability P d,i The expressions are as follows: Among them, P(·) represents probability, Q(·) represents the standard Q function, and the expression is as follows: Among them, x represents the lower limit of integration, m represents the integral variable; then consider the transmission channel gain h i (k) is time-invariant within the perception interval, i.e. h i (k) = h i ,k=1,2,…,L, the expression is as follows: Finally, the detection probability model expression of a single UAV based on energy detection is as follows: in, Among them, P f Represents a given false alarm probability value.

3. The method for detecting weak signals of radar radiation sources based on UAV swarm collaboration according to claim 1 is characterized in that: The step S2 is specifically as follows: The Rayleigh channel model is used to simulate the process of transmitting the decision results of the drone to the decision center. It is assumed that the signals received by each drone are independent of each other. The probability density model expression of the Rayleigh channel is as follows: Where r represents the signal amplitude; Then, the Monte Carlo method is used to perform numerical simulation to estimate the average detection probability under the influence of time-varying channels. Assume z i represents the result transmitted from the i-th reconnaissance aircraft to the decision center, v i It represents the decoding result of the fusion center for the transmission content of the i-th drone, and the definition expression is as follows: Among them, z i =y i +w i ,y i =g i u i , w i represents zero mean and variance is Gaussian variable, g i represents the channel gain between the decision center and each drone, and its probability density function f(g i ) is expressed as follows:

4. The method for detecting weak signals of radar radiation sources based on UAV swarm collaboration according to claim 1 is characterized in that: The step S3 is specifically as follows: The evaluation indicators of the local detection performance evaluation model include: local detection probability and local false alarm probability; First, set the conditional probability density function expression as follows: Among them, the definition expression of U(·) is as follows: Among them, t represents the positive part; Using signal detection theory to deduce, the expected and variance expressions of the detection results under various conditions are as follows: E[v i |H j ]=P(v i =1|H j )×1+P(v i =0|H j )×0=P(u i ≥0|H j ) (18) Where j = 0, 1. From equation (19), we can see that the received local decision result v i Depends on the local false alarm probability and the detection probability, which are expressed as P F,i =f(u i ≥0|H0),P D,i =f(u i ≥0|H1), then u i The distribution expression of is as follows: Then it is deduced that when setting H j The equivalent probability expressions of false alarm probability and detection probability are as follows: Since f(z i |y i ,u i )=f(z i |y i ), then we can get: f(z i |u i )=∫f(z i |y i ,u i )f(y i |u i )of i (23) Then, given the local judgment u i = +1, f(z i |y i ) and f(z i |u i ) is substituted into equation (23) to obtain the received observation value z i The conditional probability density of is expressed as follows: make Substituting variables and using the integral properties of the Q function, we get the following expression: in, Q(-x)=1-Q(x), both sides of equation (25) are i Integrate and get the following expression: Similarly, when u i = -1, we can deduce that: P(z i ≥0|u i = +1) and P(z i ≥0|u i =-1) expression into P F,i and P D,i In the above equation, the local false alarm probability and detection probability are calculated and expressed as follows: Among them, ρ i Represents the signal-to-noise ratio of the transmission channel between each UAV and the decision-making center.

5. The method for detecting weak signals of radar radiation sources based on UAV swarm collaboration according to claim 1 is characterized in that: The step S4 is specifically as follows: S41, based on the local detection result of step S3, evaluating and optimizing the detection performance of the drone; In local detection, the false alarm and detection probability expressions of the i-th drone at a given threshold T are as follows: Then the relationship between the detection probability and the instantaneous signal-to-noise ratio is d,i (η i ) is expressed as follows: Among them, Q -1 (·) represents the inverse function of the standard Q function, η i Indicates the instantaneous signal-to-noise ratio of the signal received by the drone; S42, using the Bayesian global decision method to perform decision fusion and obtain a global detection result; In decision fusion, the receiving channels from each UAV to the decision center are set to be independent and free from interference, and the delay in information transmission is ignored. i After that, according to the local decision result, decode to get v i , v i are independent of each other; Then, the decision center adopts the Bayesian global decision method. The Bayesian global decision rule determines the global detection result by calculating the posterior probability. The posterior probability represents the probability of the target existing under the condition of the detection results of each drone. The expression is as follows: Among them, P(H1) represents the prior probability, P(v1,v2,…,v N |H1) represents the joint probability of the detection results of each drone under the condition that the target exists, P(v1,v2,…,v N ) represents the marginal probability; Finally, the global decision rule expression is as follows: Since P(H1) and P(H0) are constants, the simplified expression is as follows: Assume that there are N1 1s in the local decision result obtained by decoding, and 0≤N1≤N, then 6. The method for detecting weak signals of radar radiation sources based on UAV swarm collaboration according to claim 1 is characterized in that: The step S5 is specifically as follows: The global detection performance evaluation model includes: a global false alarm probability evaluation model and a detection probability evaluation model; Then the global false alarm probability evaluation model Q of the decision center is f And the detection probability evaluation model Q D The expressions are as follows: Among them, Ω represents the decision threshold of the decision center; Finally, based on the global detection performance evaluation model, signal detection is achieved to improve the target detection probability.

Images