Radar networking active and passive joint tracking scheduling method and device under main lobe interference

Abstract

The invention relates to an active and passive joint tracking scheduling method and device for radar networking under main lobe jamming, and the method comprises the steps: designing an evaluation function with a minimum tracking error as a target through constructing a radar active measurement model and a passive positioning model, and obtaining a radar active measurement model and a passive positioning model; and a reinforcement learning agent of a near-end strategy optimization algorithm is used for carrying out radar active and passive working mode distribution, and the optimal residence time is further solved, so that the tracking performance and the system anti-interference capability of the radar networking cooperative detection system in a strong interference environment are improved. Therefore, the problems that the optimal performance of the system cannot be achieved in the prior art, the tracking performance of the radar networking cooperative detection system under the main lobe interference is easily reduced, and multi-radar task allocation under the main lobe suppression interference environment is difficult to reasonably plan are solved.

Description

Technical Field

[0001] The present application relates to the technical field of radar anti-interference and distributed simulation systems, and in particular to a radar network active-passive joint tracking and scheduling method and device under main lobe interference. Background Art

[0002] With the continuous development of radar technology, modern radars have rich resource scheduling capabilities, can flexibly change working modes and waveform parameters, and complete tasks such as search, tracking, and anti-interference. Electronic interference has become an important threat to radars. It limits the detection efficiency of radars through "soft kill" means, and even uses main lobe interference to make radar detection ineffective. Usually, the impact of electronic interference can be reduced by increasing the power aperture product of the radar and strengthening signal anti-interference processing; however, the physical scale of a single radar node is currently close to its limit, and it is necessary to achieve multi-radar task allocation and collaborative processing through networking collaborative technology to maximize the use of radar detection resources and improve the system's ability to resist main lobe interference.

[0003] In recent years, multi-radar collaborative resource scheduling has received extensive attention and in-depth research. The main purpose is to achieve the best overall performance through resource management based on multi-radar collaborative processing, including improving target tracking accuracy, increasing target tracking capacity, and reducing beam interception probability. With the development of artificial intelligence technology, deep learning methods have also been applied to radar resource management tasks such as radar waveform generation, time resource management, and detection task scheduling. Specifically, the existing technology can use the deep deterministic policy gradient (DDPG) reinforcement learning model to achieve multi-radar single target tracking, and the long-term tracking accuracy is better than the result of each step of optimization solution; in addition, the existing technology can also use deep reinforcement learning to adjust the waveform spectrum parameters in a crowded spectrum environment where radar and communication coexist to improve detection performance.

[0004] Due to the high attention paid to the technology of countering mainlobe interference, there have been many studies in recent years on achieving anti-interference through radar resource scheduling, including the use of deep neural networks for interference perception, strategy learning, game confrontation and other methods. The existing technology also considers the allocation of time resources of each node of phased array radar in a suppressed interference environment to improve target tracking accuracy and capacity; in addition, the existing technology can also deceive the jammer's jamming strategy by reasonably allocating the transmit power of multiple radars, thereby improving the target tracking performance of the system, or can jointly optimize the radar frequency and transmit power for dynamically changing interference strategies, so as to achieve detection performance optimization while avoiding interference as much as possible; at the same time, the existing technology can also model radars and jammers as multi-agents, and achieve the Nash equilibrium of interference and anti-interference strategies through mutual game.

[0005] However, the radars that report the interference pointing lines are often selected based on manual experience, which cannot achieve the optimal performance of the system. This can easily lead to a decrease in the tracking performance of the radar network collaborative detection system under main lobe interference, and it is difficult to reasonably plan the multi-radar task allocation under the main lobe suppression interference environment, which needs to be solved urgently. Summary of the invention

[0006] The present application provides a method and device for active and passive joint tracking scheduling of radar networking under main lobe interference, so as to solve the problems that the prior art cannot achieve the optimal performance of the system, easily leads to the degradation of tracking performance of the radar networking collaborative detection system under main lobe interference, and it is difficult to reasonably plan the multi-radar task allocation in the main lobe suppression interference environment.

[0007] The first aspect of the present application provides a radar network active-passive joint tracking and scheduling method under main lobe interference, comprising the following steps: determining the working mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario, and constructing an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode; determining the working mode vector of each phased array radar and the residence time variable of each phased array radar for the target to be measured according to the active measurement model and the passive positioning model, so as to establish a joint tracking and scheduling evaluation function through the working mode vector and the residence time variable; constructing a reinforcement learning agent corresponding to the networked detection system, and training the reinforcement learning agent through a preset proximal strategy optimization algorithm, so as to determine the target working mode of each phased array radar by using the trained reinforcement learning agent, and solving the joint tracking and scheduling evaluation function through a preset convex optimization solution strategy, so as to obtain the residence time of each phased array radar for the target to be measured.

[0008] Optionally, in one embodiment of the present application, the operating mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario is determined, and an active measurement model or a passive positioning model corresponding to the operating mode is constructed, wherein the operating mode includes an active detection mode and a passive detection mode, including: when the phased array radar is in the active detection mode, obtaining radial distance and angle information of the phased array radar to the target to be measured, and constructing the active measurement model according to the radial distance and the angle information; when the phased array radar is in the passive detection mode, obtaining the interference source pointing line information of the phased array radar, and constructing the passive positioning model according to the interference source pointing line information.

[0009] Optionally, in one embodiment of the present application, the joint tracking scheduling evaluation function is established through the working mode vector and the dwell time variable, including: determining the state transfer matrix and state transfer noise of the target to be measured in the kth detection cycle, and estimating the state vector of the target to be measured in the kth detection cycle according to the state transfer matrix, state transfer noise and the measurement result of the target to be measured in the k-1th detection cycle, wherein k is a positive integer; calculating the prior information Fisher information matrix of the target to be measured based on the covariance matrix of the state transfer noise, the state transfer matrix and the state vector in the k-1th detection cycle; calculating the radar measurement Jacobian matrix of each phased array radar on the target to be measured in the kth detection cycle, so as to estimate the state vector of the target to be measured in the kth detection cycle based on the state transfer matrix, state transfer noise and the measurement result of the target to be measured in the k-1th detection cycle. The radar measures the Jacobian matrix, and calculates the active tracking Fisher information matrix of the target to be measured; calculates the measurement noise of each phased array radar and the covariance matrix corresponding to the measurement noise, and calculates the passive tracking Fisher information matrix of the target to be measured based on the passive positioning model and the covariance matrix corresponding to the measurement noise; calculates the Bayesian information matrix of the target to be measured in the kth detection cycle based on the prior information Fisher information matrix, the passive tracking Fisher information matrix and the active tracking Fisher information matrix; calculates the Bayesian Cramer-Rao bound of the target to be measured according to the Bayesian information matrix in the kth detection cycle, so as to establish the joint tracking scheduling evaluation function through the Bayesian Cramer-Rao bound, the working mode vector and the dwell time variable.

[0010] Optionally, in one embodiment of the present application, the construction of the reinforcement learning agent corresponding to the networked detection system includes: determining the stable state of the reinforcement learning agent corresponding to the mth step in the Markov decision process based on a preset Markov decision strategy, wherein the stable state includes the radar position of the phased array radar, the state vector of the target to be measured, the principal element of the prior information matrix of the target to be measured, the passive measurement error of the interfering target in the target to be measured, and the unique hot vector of the phased array radar corresponding to the mth step, wherein m is a positive integer; determining the dynamic state of the reinforcement learning agent through the radar coding of the phased array radar, the working mode and the prediction error of the mth step, and establishing the state space of the reinforcement learning agent according to the stable state and the dynamic state; determining the action space of the reinforcement learning agent according to the working mode of the phased array radar; and constructing the reward function of the reinforcement learning agent based on the prediction error of the mth step, the prediction error of the m-1th step and preset scaling hyperparameters.

[0011] Optionally, in one embodiment of the present application, the training of the reinforcement learning agent through a preset proximal policy optimization algorithm includes: constructing the preset proximal policy optimization algorithm through a preset fully-linked network, and randomly generating a target training environment to train the preset proximal policy optimization algorithm using the target training environment, wherein the target training environment includes a randomly generated initial state and interference intensity of the target to be measured within the coverage range of the phased array radar; training the reinforcement learning agent based on the target training environment and the trained preset proximal policy optimization algorithm.

[0012] Optionally, in one embodiment of the present application, the mathematical expression of the joint tracking scheduling evaluation function is:

[0013]

[0014] Among them, ∈ k is the working mode vector of each phased array radar in the kth detection cycle; τ q,k is the dwell time variable of each phased array radar on the target q in the kth detection cycle; M represents the number of phased array radars.

[0015] The second aspect of the present application provides a radar network active-passive joint tracking and scheduling device under main lobe interference, including: a modeling module, used to determine the working mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario, and construct an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode; an establishment module, used to determine the working mode vector of each phased array radar and the residence time variable of each phased array radar for the target to be measured according to the active measurement model and the passive positioning model, so as to establish a joint tracking scheduling evaluation function through the working mode vector and the residence time variable; a solution module, used to construct a reinforcement learning agent corresponding to the networked detection system, and train the reinforcement learning agent through a preset proximal strategy optimization algorithm, so as to use the trained reinforcement learning agent to determine the target working mode of each phased array radar, and solve the joint tracking scheduling evaluation function through a preset convex optimization solution strategy to obtain the residence time of each phased array radar for the target to be measured.

[0016] Optionally, in one embodiment of the present application, the modeling module includes: an active detection unit, used to obtain the radial distance and angle information of the phased array radar to the target to be measured when the phased array radar is in the active detection mode, and construct the active measurement model according to the radial distance and the angle information; a passive detection unit, used to obtain the interference source pointing line information of the phased array radar when the phased array radar is in the passive detection mode, and construct the passive positioning model according to the interference source pointing line information.

[0017] Optionally, in one embodiment of the present application, the establishment module includes: a first determination unit, used to determine the state transfer matrix and state transfer noise of the target to be measured in the kth detection cycle, and estimate the state vector of the target to be measured in the kth detection cycle according to the state transfer matrix, state transfer noise and the measurement result of the target to be measured in the k-1th detection cycle, wherein k is a positive integer; a first calculation unit, used to calculate the prior information Fisher information matrix of the target to be measured based on the covariance matrix of the state transfer noise, the state transfer matrix and the state vector in the k-1th detection cycle; a second calculation unit, used to calculate the radar measurement Jacobian matrix of each phased array radar on the target to be measured in the kth detection cycle, so as to calculate the radar measurement Jacobian matrix based on the radar measurement Jacobian matrix. an active tracking Fisher information matrix of the target to be measured; a third calculation unit, used to calculate the measurement noise of each phased array radar and the covariance matrix corresponding to the measurement noise, and calculate the passive tracking Fisher information matrix of the target to be measured based on the passive positioning model and the covariance matrix corresponding to the measurement noise; a fourth calculation unit, used to calculate the Bayesian information matrix of the target to be measured in the kth detection cycle based on the prior information Fisher information matrix, the passive tracking Fisher information matrix and the active tracking Fisher information matrix; a fifth calculation unit, used to calculate the Bayesian Cramer-Rao bound of the target to be measured according to the Bayesian information matrix in the kth detection cycle, so as to establish the joint tracking scheduling evaluation function through the Bayesian Cramer-Rao bound, the working mode vector and the residence time variable.

[0018] Optionally, in one embodiment of the present application, the solution module includes: a second determination unit, used to determine the stable state of the reinforcement learning agent corresponding to the mth step in the Markov decision process based on a preset Markov decision strategy, wherein the stable state includes the radar position of the phased array radar, the state vector of the target to be measured, the principal element of the prior information matrix of the target to be measured, the passive measurement error of the interfering target in the target to be measured, and the unique hot vector of the phased array radar corresponding to the mth step, wherein m is a positive integer; a third determination unit, used to determine the dynamic state of the reinforcement learning agent through the radar coding of the phased array radar, the working mode and the prediction error of the mth step, and establish the state space of the reinforcement learning agent according to the stable state and the dynamic state; a fourth determination unit, used to determine the action space of the reinforcement learning agent according to the working mode of the phased array radar; a construction unit, used to construct the reward function of the reinforcement learning agent based on the prediction error of the mth step, the prediction error of the m-1th step and a preset scaling hyperparameter.

[0019] Optionally, in one embodiment of the present application, the solution module also includes: a generation unit, used to construct the preset proximal strategy optimization algorithm through a preset fully-linked network, and randomly generate a target training environment to train the preset proximal strategy optimization algorithm using the target training environment, wherein the target training environment includes a randomly generated initial state and interference intensity of the target to be measured within the coverage range of the phased array radar; a training unit, used to train the reinforcement learning agent based on the target training environment and the trained preset proximal strategy optimization algorithm.

[0020] Optionally, in one embodiment of the present application, the mathematical expression of the joint tracking scheduling evaluation function is:

[0021]

[0022] Among them, ∈ k is the working mode vector of each phased array radar in the kth detection cycle; τ q,k is the dwell time variable of each phased array radar on the target q in the kth detection cycle; M represents the number of phased array radars.

[0023] The third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the active and passive joint tracking and scheduling method for radar networking under main lobe interference as described in the above embodiment.

[0024] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned active and passive joint tracking and scheduling method for radar networking under main lobe interference.

[0025] The fifth aspect of the present application provides a computer program product, including a computer program, which is executed to implement the above-mentioned active and passive joint tracking and scheduling method for radar networking under main lobe interference.

[0026] Therefore, the embodiments of the present application have the following beneficial effects:

[0027] The embodiment of the present application can determine the working mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario, and construct an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode; determine the working mode vector of each phased array radar and the dwell time variable of each phased array radar on the target to be measured according to the active measurement model and the passive positioning model, so as to establish a joint tracking scheduling evaluation function through the working mode vector and the dwell time variable; construct a reinforcement learning agent corresponding to the networked detection system, and train the reinforcement learning agent through a preset proximal strategy optimization algorithm, so as to use the trained reinforcement learning agent to determine the target working mode of each phased array radar, and solve the joint tracking scheduling evaluation function through a preset convex optimization solution strategy to obtain the dwell time of each phased array radar on the target to be measured, so as to effectively improve the tracking performance of the radar radar networked collaborative detection system in a strong interference environment and improve the system's anti-interference ability. Thus, the problems that the prior art cannot achieve the optimal performance of the system, easily lead to a decrease in the tracking performance of the radar networked collaborative detection system under main lobe interference, and it is difficult to reasonably plan the multi-radar task allocation under the main lobe suppression interference environment are solved.

[0028] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0030] Figure 1 A flowchart of a radar network active and passive joint tracking scheduling method under main lobe interference provided according to an embodiment of the present application;

[0031] Figure 2 A schematic diagram of a network detection system model provided for an embodiment of the present application;

[0032] Figure 3A schematic diagram of radar position and target activity range provided for one embodiment of the present application;

[0033] Figure 4 (a) is a schematic diagram of a tracking error curve provided for an embodiment of the present application.

[0034] Figure 4 (b) is a schematic diagram of resource allocation results provided for an embodiment of the present application.

[0035] Figure 5 This is an example diagram of a radar network active and passive joint tracking and scheduling device under main lobe interference according to an embodiment of the present application;

[0036] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0037] Among them, 10-radar network active and passive joint tracking and scheduling device under main lobe interference; 100-modeling module, 200-establishment module, 300-solution module; 601-memory, 602-processor, 603-communication interface. DETAILED DESCRIPTION

[0038] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0039] The following describes the active and passive joint tracking and scheduling method and device for radar networking under main lobe interference in an embodiment of the present application with reference to the accompanying drawings. In response to the problems mentioned in the above background technology, the present application provides a radar network active and passive joint tracking and scheduling method under main lobe interference. In this method, the working mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario is determined, and an active measurement model or a passive positioning model corresponding to the working mode is constructed, wherein the working mode includes an active detection mode and a passive detection mode; the working mode vector of each phased array radar and the residence time variable of each phased array radar on the target to be measured are determined according to the active measurement model and the passive positioning model, so as to establish a joint tracking scheduling evaluation function through the working mode vector and the residence time variable; a reinforcement learning agent corresponding to the networked detection system is constructed, and the reinforcement learning agent is trained through a preset proximal strategy optimization algorithm, so as to use the trained reinforcement learning agent to determine the target working mode of each phased array radar, and the joint tracking scheduling evaluation function is solved through a preset convex optimization solution strategy to obtain the residence time of each phased array radar on the target to be measured, thereby effectively improving the tracking performance of the radar network collaborative detection system in a strong interference environment and improving the system's anti-interference capability. This solves the problems that the existing technology cannot achieve the optimal performance of the system, easily leads to the degradation of tracking performance of the radar network collaborative detection system under main lobe interference, and is difficult to reasonably plan the multi-radar task allocation under the main lobe suppression interference environment.

[0040] Specifically, Figure 1 A flowchart of a radar network active and passive joint tracking scheduling method under main lobe interference provided in an embodiment of the present application.

[0041] like Figure 1 As shown, the radar network active and passive joint tracking scheduling method under main lobe interference includes the following steps:

[0042] In step S101, the working mode of each phased array radar in the networked detection system in the main lobe interference penetration scenario is determined, and an active measurement model or a passive positioning model corresponding to the working mode is constructed, wherein the working mode includes an active detection mode and a passive detection mode.

[0043] It should be noted that the networked detection system in the embodiment of the present application is composed of M radars, and the deployment position of each radar is represented by (x i ,y i ),i=1,2,...,M, the detection cycle of the network detection system is T0. In the kth detection cycle, the system needs to track Q spatially dispersed targets, including one self-defense interference target that releases noise suppression interference. At this time, the position and speed of the target are expressed as (x q,k ,y q,k )and

[0044] In addition, the networked detection system consists of a central node responsible for assigning working modes and corresponding parameters to each radar, such as Figure 2 As shown, the central node (i.e., fusion center) includes the following two fusion engines:

[0045] 1. Interference source cross-location engine, used to fuse interference pointing lines;

[0046] 2. Radar track fusion engine, used to fuse tracking track information.

[0047] In an embodiment of the present application, after the fusion of multi-radar information, the target tracking process can be completed, and the position of the target in the next detection cycle can be predicted based on this, and the working mode and corresponding parameters are assigned to the corresponding radar, where the working mode includes active detection and passive detection mode, and the working parameter is the residence time of the active detection radar on the target.

[0048] Assuming that the fusion center has obtained the target quantity Q and the identification information of the interference target in the search phase, the state vector of the target q to be tracked in the kth detection cycle in the embodiment of the present application is defined as:

[0049]

[0050] The target's motion state transfer is described by the uniform linear motion equation, as shown below:

[0051]

[0052] Among them, F q is the state transfer matrix, is the noise of state transition.

[0053] Therefore, the embodiment of the present application considers a networked detection system composed of multiple multifunctional phased array radars in a two-dimensional space, and collaboratively tracks multiple targets including jammers; wherein each radar node in the networked detection system collaborates under the control of a fusion center in a timing cycle, and the fusion center can allocate detection tasks and parameters at the beginning of each cycle based on the detection results of each radar in the previous cycle, thereby achieving the purpose of maximizing tracking performance.

[0054] Optionally, in one embodiment of the present application, the operating mode of each phased array radar in the networked detection system in the main lobe interference penetration scenario is determined, and an active measurement model or a passive positioning model corresponding to the operating mode is constructed, wherein the operating mode includes an active detection mode and a passive detection mode, including: when the phased array radar is in the active detection mode, the radial distance and angle information of the phased array radar to the target to be measured is obtained, and an active measurement model is constructed based on the radial distance and angle information; when the phased array radar is in the passive detection mode, the interference source pointing line information of the phased array radar is obtained, and a passive positioning model is constructed based on the interference source pointing line information.

[0055] It should be noted that since each radar node in the networked detection system can quickly switch between active detection and passive detection modes during the detection cycle, the working mode of the i-th radar in the k-th detection cycle is represented as ∈ i,k :

[0056]

[0057] Specifically, when the radar works in the active detection mode, the radar can obtain target information such as the radial distance and angle information of the target to be detected; assuming that the i-th radar detects the target q, the radar measurement model is expressed as:

[0058]

[0059] in, is the measurement vector, which contains the distance R i,q,k and angle θ i,q,k Two elements, as shown below:

[0060]

[0061] Among them, w in formula (4) i,q,k represents the measurement error of active tracking, which is a zero-mean Gaussian distribution, and its covariance ∑ i,q,k Contains the Cramér-Rao Lower Bound (CRLB) for range and bearing estimates.

[0062]

[0063] Where c is the speed of light; B i,q,j is the radar waveform bandwidth for a specific phased array radar i; β i is a constant, i.e., the normalized angle measurement error of the radar; μ in the interference scenario i,q,k is the signal to interference plus noise ratio (SINR) measured by the radar this time, and:

[0064] μ i,q,k ∝τ i,q,k (7)

[0065] Among them, τ i,q,k is the dwell time of radar i tracking the waveform of target q.

[0066] In addition, when the radar works in passive detection mode, that is, ∈ i,k = 0, the radar can only obtain the angle information of the interference signal, that is, the information of the interference source pointing line; assuming that the i-th radar detects the interference target q, similar to the active measurement model, the passive measurement model is expressed as:

[0067]

[0068]

[0069] Since phased array radar can use multiple sub-arrays on the array to estimate the direction of the interference signal, the measurement error v i,q,k is the error of the linear array DOA estimation, which is zero-mean Gaussian noise. By modeling the passive radar as a K i Sub-arrays are arrays, and it is assumed that the interference signal is within the radar operating frequency band, that is, the array spacing is half of the signal wavelength. According to the uncorrelated zero-mean Gaussian source signal in the prior art, the variance of the measurement error can be calculated:

[0070]

[0071] Among them, formula (10) can be solved by block matrix method, α i,q,k is a constant term containing the front beam width, and γ i,q,k is the cumulative interference-to-noise ratio.

[0072] After obtaining the interference source direction-finding information of multiple radars, the fusion center can perform cross-location of interference sources based on this information. For the entire network detection system, the measurement equation can be expressed as:

[0073]

[0074] in, The measurement noise v q,k The covariance matrix of is as follows:

[0075]

[0076]

[0077] Wherein, formula (13) represents the variance of cross positioning, i.e., the variance of DOA estimation, when the radar is passively detecting. If the radar is actively detecting, no interference pointing line information is provided.

[0078] Therefore, the embodiments of the present application establish an active measurement model and a passive positioning model of the phased array radar under the lobe interference penetration scenario, thereby providing reliable technical support and basis for the implementation of active and passive joint tracking scheduling of radar networking under subsequent main lobe interference.

[0079] In step S102, the working mode vector of each phased array radar and the dwell time variable of each phased array radar on the target to be measured are determined according to the active measurement model and the passive positioning model, so as to establish a joint tracking scheduling evaluation function through the working mode vector and the dwell time variable.

[0080] It should be understood by those skilled in the art that, as the jammer interference intensity increases, the SINR of active detection of the radar decreases and the JNR of passive detection increases, which correspondingly reduces the active tracking accuracy of the jammer and improves the passive direction finding accuracy. At the same time, the measurement accuracy of other conventional targets is mainly maintained by active detection; therefore, in theory, the embodiments of the present application can reasonably allocate the working modes of each radar and schedule the working parameters, wherein the working parameters include the working mode vector and the dwell time variable, as shown in the following formula:

[0081] ∈ k =[∈ 1,k ,∈ 2,k ,...,∈ M,k ] T (14)

[0082] τ q,k =[τ 1,q,k ,τ 2,q,k ,...,τ M,q,k ] T (15)

[0083] Among them, ∈ k is the vector composed of all radar working modes, τ q,k is the dwell time of all radars on target q.

[0084] Therefore, the embodiments of the present application optimize the global tracking accuracy of all targets by reasonably allocating the working modes of each radar and scheduling the working parameters.

[0085] Optionally, in one embodiment of the present application, a joint tracking scheduling evaluation function is established through the working mode vector and the residence time variable, including: determining the state transfer matrix and state transfer noise of the target to be measured in the kth detection cycle, and estimating the state vector of the target to be measured in the kth detection cycle based on the state transfer matrix of the target to be measured, the state transfer noise and the measurement result of the target to be measured in the k-1th detection cycle, wherein k is a positive integer; calculating the prior information Fisher information matrix of the target to be measured based on the covariance matrix of the state transfer noise, the state transfer matrix and the state vector in the k-1th detection cycle; calculating the radar measurement Jacobian moment of each phased array radar on the target to be measured in the kth detection cycle. The invention discloses a phased array radar system, and a method for calculating an active tracking Fisher information matrix of the target to be measured based on the radar measurement Jacobian matrix; calculating the measurement noise of each phased array radar and the covariance matrix corresponding to the measurement noise, and calculating the passive tracking Fisher information matrix of the target to be measured based on the passive positioning model and the covariance matrix corresponding to the measurement noise; calculating the Bayesian information matrix of the target to be measured in the kth detection cycle based on the prior information Fisher information matrix, the passive tracking Fisher information matrix and the active tracking Fisher information matrix; calculating the Bayesian Cramer-Rao bound of the target to be measured according to the Bayesian information matrix in the kth detection cycle, and establishing a joint tracking scheduling evaluation function through the Bayesian Cramer-Rao bound, the working mode vector and the dwell time variable.

[0086] It can be understood that the Bayesian Cramé-Rao Bound (BCRLB) is one of the most classic and commonly used performance indicators in target tracking. It can give a statistical lower bound on the mean square error of state estimation and can be obtained by solving the inverse matrix of the Bayesian Information Matrix (BIM) of target tracking.

[0087] The BIM of the target q to be measured in the kth time period includes the Fisher Information Matrix (FIM) of the prior information and the measurement information. Since the network detection system in the embodiment of the present application has two different fusion engines, it can be used represents the FIM of active detection (i.e., active tracking Fisher information matrix), and is used The FIM (Fisher information matrix for passive tracking) for passive detection is shown in the following formula:

[0088]

[0089] in, is the prior information FIM (i.e., the prior information Fisher information matrix) of the target q to be measured; is the BIM of the target q to be measured; and It can be calculated by the following formula:

[0090]

[0091]

[0092] in, The state vector of the target q is the predicted state obtained by tracking estimation at the previous moment; G i,q,k is the Jacobian matrix of the radar measurement:

[0093]

[0094] From formula (6), we can see that and the parameter τ to be assigned i,q,k is a linear relationship, therefore, the embodiment of the present application can be rewritten as at this time It can be expressed as ∈ i,k With τ i,q,k The function is as follows:

[0095]

[0096] and Similarly, the embodiment of the present application can obtain the radar system measurement FIM during passive tracking, as shown in the following formula:

[0097]

[0098]

[0099] In summary, when obtaining the BIM of the k-1th time period (i.e., detection period) After that, the fusion center can Predict the tracking performance for the kth time period, where for∈ k and τ q,k The function of for∈ k Therefore, the embodiment of the present application can write the tracking BCRLB of the target q to be measured as a function of the parameters to be allocated:

[0100]

[0101] Among them, Λ is the matrix introduced to unify the dimensions of BCRLB:

[0102]

[0103] Therefore, the embodiment of the present application establishes a joint tracking scheduling evaluation function through the Bayesian Cramer-Rao bound, the working mode vector and the residence time variable, so as to achieve the minimization of the overall BCRLB by selecting the working mode of each radar and reasonably allocating the residence time based on the networked detection system in a noise interference environment.

[0104] Optionally, in one embodiment of the present application, the mathematical expression of the joint tracking scheduling evaluation function is:

[0105]

[0106]

[0107] Among them, ∈ k is the working mode vector of each phased array radar in the kth detection cycle; τ q,k is the dwell time variable of each phased array radar on the target q in the kth detection cycle; M represents the number of phased array radars.

[0108] It should be noted that the mathematical form of the joint tracking scheduling evaluation function for minimizing the overall BCRLB in the embodiment of the present application is expressed as:

[0109]

[0110] Among them, ∈ k is the working mode vector of each phased array radar in the kth detection cycle; τ q,k is the dwell time variable of each phased array radar on the target q in the kth detection cycle; M represents the number of phased array radars; T k =[τ 1,k ,τ 1,k ,...,τ Q,k ] is a matrix composed of the dwell time of all radars detecting targets.

[0111] Therefore, the constraints of the joint tracking scheduling evaluation function in the embodiment of the present application respectively constrain the number of radars in active detection mode, the beam dwell time, and the maximum dwell time that each radar can schedule under different working modes, so as to achieve the optimization goal of minimizing the global target tracking error through the joint tracking scheduling evaluation function.

[0112] In step S103, a reinforcement learning agent corresponding to the networked detection system is constructed, and the reinforcement learning agent is trained by a preset proximal strategy optimization algorithm, so as to use the trained reinforcement learning agent to determine the target working mode of each phased array radar, and solve the joint tracking scheduling evaluation function by a preset convex optimization solution strategy to obtain the residence time of each phased array radar on the target to be measured.

[0113] The embodiment of the present application constructs a radar active measurement model and a passive positioning model, designs an evaluation function with the goal of minimizing tracking error, and uses a reinforcement learning agent of a proximal strategy optimization algorithm to allocate radar active and passive working modes, and further solves the optimal residence time, so as to exert the system detection efficiency through multi-node collaboration to counter the main lobe interference threat. Compared with a fixed radar working mode allocation strategy, the embodiment of the present application can effectively improve the tracking performance and system anti-interference capability of the radar radar network collaborative detection system in a strong interference environment.

[0114] Optionally, in one embodiment of the present application, a reinforcement learning agent corresponding to the networked detection system is constructed, including: based on a preset Markov decision strategy, determining the stable state of the reinforcement learning agent corresponding to the mth step in the Markov decision process, wherein the stable state includes the radar position of the phased array radar, the state vector of the target to be measured, the principal element of the prior information matrix of the target to be measured, the passive measurement error of the interference target in the target to be measured, and the unique hot vector of the phased array radar corresponding to the mth step, wherein m is a positive integer; determining the dynamic state of the reinforcement learning agent through the radar coding, working mode and prediction error of the phased array radar, and establishing the state space of the reinforcement learning agent according to the stable state and the dynamic state; determining the action space of the reinforcement learning agent according to the working mode of the phased array radar; and constructing the reward function of the reinforcement learning agent based on the prediction error of the mth step, the prediction error of the m-1th step and the preset scaling hyperparameters.

[0115] It should be noted that the optimization model described by equation (25) (i.e., the joint tracking scheduling evaluation function) contains two optimization variables: radar working mode allocation ∈ k and the radar's dwell time T for target detection k , where ∈ k is a binary optimization variable, and its result is similar to T k are coupled with each other, so this problem is a non-convex optimization problem of hybrid programming.

[0116] In the optimization model of formula (25), given a fixed radar working mode distribution After that, the embodiment of the present application can solve T separately k ; Without loss of generality, in the embodiment of the present application, the set of interference targets passively tracked among the Q targets is Q p , the set of active tracking targets is Q a , the optimization problem of formula (25) is equivalent to:

[0117]

[0118] It can be understood that the optimization problem of formula (26) has been proved to be a convex optimization problem. Therefore, the embodiment of the present application can be solved by using a convex optimization toolkit, such as SciPy's sequential least squares programming and other tools; based on this, the embodiment of the present application constructs a two-step method to solve the original problem of formula (25), namely:

[0119] Step 1: Train a reinforcement learning agent to estimate the radar's operating mode ∈ k ;

[0120] Step 2: Solve for T k Convex optimization problem.

[0121] Specifically, the problem of formula (25) is described as a Markov decision process (MDP), and the MDP is designed to determine a radar working mode at each step. The state space, action space and reward function of the constructed reinforcement learning agent are as follows:

[0122] (1) State space: The state space is divided into two parts. One part is the stable state that does not change in one step, including the radar position, the target prediction state (i.e., the state vector of the target to be measured), the principal element of the target prior information matrix, and the predicted passive measurement error of the interference target. The other part is the dynamic state that changes with each step of the MDP, including the current radar code, each radar working mode, and the current prediction error. The state space of the mth step is as follows:

[0123]

[0124] In the formula, e m is the one-hot vector corresponding to the radar at the current step, the vector length is M, the mth element is 1, and the rest are 0; represents the covariance matrix of formula (12) calculated according to the predicted position of the target; and Represents the value of the optimization variable at the previous step.

[0125] (2) Action space: Since the embodiment of the present application makes a decision for only one radar working mode at a time, the action space contains only two elements, namely, radar working mode ∈ i,k The corresponding values ​​are:

[0126] A={0,1}(28)

[0127] (3) Reward function: The goal of the network detection system is to minimize the tracking error. Therefore, the decrease in the tracking error of each step compared to the previous step is used as the reward function, and the hyperparameter ζ is used to scale the reward function to a reasonable range:

[0128]

[0129] It can be understood that the optimization goal of formula (25) is to minimize the global target tracking error, and the constraints respectively constrain the number of radars in active detection mode, the beam dwell time, and the maximum dwell time that each radar can schedule under different working modes.

[0130] Optionally, in one embodiment of the present application, a reinforcement learning agent is trained through a preset proximal strategy optimization algorithm, including: constructing a preset proximal strategy optimization algorithm through a preset fully-linked network, and randomly generating a target training environment, so as to train the preset proximal strategy optimization algorithm using the target training environment, wherein the target training environment includes a randomly generated initial state and interference intensity of a target to be measured within the coverage range of a phased array radar; and training the reinforcement learning agent based on the target training environment and the trained preset proximal strategy optimization algorithm.

[0131] For the training of reinforcement learning, the embodiment of the present application can use PPO (Proximal Policy Optimization, reinforcement learning model) to train the intelligent agent, and the deep learning network uses a fully connected network; in addition, for the training of the intelligent agent, the embodiment of the present application can use a randomly generated environment with a distribution similar to the expected environment for sampling, and the random environment includes the initial state of randomly generated targets within the radar coverage range. And the interference intensity P q .

[0132] For the trained intelligent agent, the embodiment of the present application can adopt the combined optimization problem of the two-step method and the network detection system working mode and the residence time allocation algorithm to calculate formula (25), as shown in the following table.

[0133] Table 1

[0134]

[0135]

[0136] The following application analyzes and verifies the effectiveness of the iterative algorithm based on deep reinforcement learning and convex optimization solution involved in the active and passive joint tracking and scheduling method of radar networking under main lobe interference proposed in this application through simulation experiments and combined with the accompanying drawings.

[0137] 1. Parameter settings

[0138] In the simulation experiment, the specific embodiment of the present application can be set to have M = 7 radars and Q = 6 targets, where the first target is a jammer; the position of the radar is fixed, and the target penetrates from a certain range covered by the radar power at a random position and speed, and the distance from the radar to the position is about 30km to 100km, such as Figure 3 shown.

[0139] The working parameters of each radar are the same. The radar system parameters are shown in the table. The radar cross-sectional area of ​​the target is 1m 2 The target movement speed ranges from 100m / s to 10m / s, and the power spectrum density of the interference signal ranges from 10W / Mhz to 10kW / Mhz.

[0140] Table 2

[0141]

[0142] When performing simulation training for PPO learning, the initial position, speed, and interference power of the target are randomly generated each time a scene is generated, and the training hyperparameter settings are shown in the table.

[0143] Table 3

[0144]

[0145]

[0146] 2. Performance verification and analysis

[0147] In order to evaluate the performance of the iterative algorithm based on deep reinforcement learning and convex optimization solution of the present application, the embodiment of the present application can implement the iterative algorithm based on deep reinforcement learning and convex optimization solution through the PPO agent, and compare the performance with the following three radar working mode allocation strategies, as described below:

[0148] 1. Fixed allocation strategy: This strategy generally believes that a larger detection angle baseline can bring better interference cross-positioning performance. Figure 3 The scenario shown selects Radar 2 and Radar 3 as passive detection radars;

[0149] Second, random allocation strategy, which can randomly select the radar responsible for passive detection from all radars in the feasible domain;

[0150] 3. Exhaustive allocation strategy: This strategy needs to exhaust all possible allocation strategies. Obviously, this strategy can achieve the best performance. However, since the iteration of the convex optimization problem of formula (26) is time-consuming, this strategy needs to solve 2 M Subconvex optimization problems are difficult to meet the real-time requirements of scheduling radar resources for tracking.

[0151] Depend on Figure 4 (a) and Figure 4 (b) It can be seen that the performance of the algorithm shown in the embodiment of the present application can exceed the results of the fixed allocation strategy and the random allocation strategy, and is close to the result of the optimal exhaustive allocation strategy.

[0152] In summary, the embodiments of the present application solve the problems of radar working mode allocation and dwell time optimization by constructing an active measurement model and a passive positioning model of a phased array radar in a main lobe interference scenario, and designing a joint tracking scheduling objective function with the goal of minimizing tracking error. Therefore, the embodiments of the present application can improve the overall tracking performance in a strong main lobe interference environment in a multi-radar networked collaborative tracking task. By reasonably allocating the resources of each radar, a lower tracking error is achieved, which well overcomes the difficulty of the overall tracking performance degradation of the networked detection system when facing a main lobe interference environment.

[0153] According to the active and passive joint tracking and scheduling method of radar networking under main lobe interference proposed in the embodiment of the present application, the working mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario is determined, and an active measurement model or a passive positioning model corresponding to the working mode is constructed, wherein the working mode includes an active detection mode and a passive detection mode; the working mode vector of each phased array radar and the residence time variable of each phased array radar on the target to be measured are determined according to the active measurement model and the passive positioning model, so as to establish a joint tracking and scheduling evaluation function through the working mode vector and the residence time variable; a reinforcement learning agent corresponding to the networked detection system is constructed, and the reinforcement learning agent is trained through a preset proximal strategy optimization algorithm, so as to determine the target working mode of each phased array radar by using the trained reinforcement learning agent, and the joint tracking and scheduling evaluation function is solved through a preset convex optimization solution strategy to obtain the residence time of each phased array radar on the target to be measured, thereby effectively improving the tracking performance of the radar networked collaborative detection system in a strong interference environment and improving the system's anti-interference capability.

[0154] Secondly, the active and passive joint tracking and scheduling device for radar networking under main lobe interference proposed in an embodiment of the present application is described with reference to the accompanying drawings.

[0155] Figure 5 It is a block diagram of a radar network active and passive joint tracking and scheduling device under main lobe interference in an embodiment of the present application.

[0156] like Figure 5 As shown, the radar network active and passive joint tracking and scheduling device 10 under main lobe interference includes: a modeling module 100, an establishing module 200 and a solving module 300.

[0157] Among them, the modeling module 100 is used to determine the working mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario, and construct an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode.

[0158] Establishing module 200, for determining the working mode vector of each phased array radar and the dwell time variable of each phased array radar for the target to be measured according to the active measurement model and the passive positioning model, so as to establish a joint tracking scheduling evaluation function through the working mode vector and the dwell time variable.

[0159] The solution module 300 is used to construct a reinforcement learning agent corresponding to the networked detection system, and train the reinforcement learning agent through a preset proximal strategy optimization algorithm, so as to use the trained reinforcement learning agent to determine the target working mode of each phased array radar, and solve the joint tracking scheduling evaluation function through a preset convex optimization solution strategy to obtain the residence time of each phased array radar on the target to be measured.

[0160] Optionally, in one embodiment of the present application, the modeling module 100 includes: an active detection unit and a passive detection unit.

[0161] Among them, the active detection unit is used to obtain the radial distance and angle information of the phased array radar to the target to be measured when the phased array radar is in the active detection mode, and to build an active measurement model based on the radial distance and angle information.

[0162] The passive detection unit is used to obtain the interference source pointing line information of the phased array radar when the phased array radar is in the passive detection mode, and to build a passive positioning model according to the interference source pointing line information.

[0163] Optionally, in one embodiment of the present application, the establishing module 200 includes: a first determining unit, a first calculating unit, a second calculating unit, a third calculating unit, a fourth calculating unit and a fifth calculating unit.

[0164] Among them, the first determination unit is used to determine the state transfer matrix and state transfer noise of the target to be measured in the kth detection cycle, and estimate the state vector of the target to be measured in the kth detection cycle based on the state transfer matrix, state transfer noise and measurement results of the target to be measured in the k-1th detection cycle, wherein k is a positive integer.

[0165] The first calculation unit is used to calculate the prior information Fisher information matrix of the target to be detected based on the covariance matrix of the state transfer noise, the state transfer matrix and the state vector in the k-1th detection cycle.

[0166] The second calculation unit is used to calculate the radar measurement Jacobian matrix of the target to be detected by each phased array radar in the kth detection cycle, so as to calculate the active tracking Fisher information matrix of the target to be detected based on the radar measurement Jacobian matrix.

[0167] The third calculation unit is used to calculate the measurement noise of each phased array radar and the covariance matrix corresponding to the measurement noise, and calculate the passive tracking Fisher information matrix of the target to be measured based on the passive positioning model and the covariance matrix corresponding to the measurement noise.

[0168] The fourth calculation unit is used to calculate the Bayesian information matrix of the target to be detected in the kth detection cycle based on the prior information Fisher information matrix, the passive tracking Fisher information matrix and the active tracking Fisher information matrix.

[0169] The fifth calculation unit is used to calculate the Bayesian Cramer Rao bound of the target to be detected according to the Bayesian information matrix in the kth detection cycle, so as to establish a joint tracking scheduling evaluation function through the Bayesian Cramer Rao bound, the working mode vector and the residence time variable.

[0170] Optionally, in one embodiment of the present application, the solution module 300 includes: a second determination unit, a third determination unit, a fourth determination unit and a construction unit.

[0171] Among them, the second determination unit is used to determine the stable state of the reinforcement learning agent corresponding to the mth step in the Markov decision process based on a preset Markov decision strategy, wherein the stable state includes the radar position of the phased array radar, the state vector of the target to be measured, the principal element of the prior information matrix of the target to be measured, the passive measurement error of the interference target in the target to be measured, and the unique hot vector of the phased array radar corresponding to the mth step, wherein m is a positive integer.

[0172] The third determination unit is used to determine the dynamic state of the reinforcement learning agent through the radar coding, working mode and prediction error of the phased array radar, and establish the state space of the reinforcement learning agent according to the stable state and the dynamic state.

[0173] The fourth determination unit determines the action space of the reinforcement learning agent according to the working mode of the phased array radar.

[0174] A construction unit for constructing a reward function for a reinforcement learning agent based on the prediction error at the mth step, the prediction error at the m-1th step, and preset scaling hyperparameters.

[0175] Optionally, in one embodiment of the present application, the solution module 300 further includes: a generation unit and a training unit.

[0176] Among them, the generation unit is used to construct a preset proximal strategy optimization algorithm through a preset full-link network, and randomly generate a target training environment to train the preset proximal strategy optimization algorithm using the target training environment, wherein the target training environment includes a randomly generated initial state and interference intensity of a target to be measured within the coverage range of the phased array radar.

[0177] The training unit is used to train the reinforcement learning agent based on the target training environment and the preset proximal policy optimization algorithm after training.

[0178] Optionally, in one embodiment of the present application, the mathematical expression of the joint tracking scheduling evaluation function is:

[0179]

[0180] Among them, ∈ k is the working mode vector of each phased array radar in the kth detection cycle; τ q,k is the dwell time variable of each phased array radar on the target q in the kth detection cycle; M represents the number of phased array radars.

[0181] It should be noted that the above explanation of the embodiment of the active and passive joint tracking and scheduling method for radar networking under main lobe interference is also applicable to the active and passive joint tracking and scheduling device for radar networking under main lobe interference of this embodiment, and will not be repeated here.

[0182] According to the radar network active-passive joint tracking and scheduling device under main lobe interference proposed in the embodiment of the present application, it includes a modeling module 100, which is used to determine the working mode of each phased array radar in the networked detection system under the main lobe interference penetration scenario, and construct an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode; an establishment module 200, which is used to determine the working mode vector of each phased array radar and the residence time variable of each phased array radar on the target to be measured according to the active measurement model and the passive positioning model, so as to establish a joint tracking and scheduling evaluation function through the working mode vector and the residence time variable; a solution module 300, which is used to construct a reinforcement learning agent corresponding to the networked detection system, and train the reinforcement learning agent through a preset proximal strategy optimization algorithm, so as to determine the target working mode of each phased array radar by using the trained reinforcement learning agent, and solve the joint tracking and scheduling evaluation function through a preset convex optimization solution strategy, so as to obtain the residence time of each phased array radar on the target to be measured, thereby effectively improving the tracking performance of the radar networked collaborative detection system in a strong interference environment and improving the system's anti-interference capability.

[0183] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0184] A memory 601 , a processor 602 , and a computer program stored in the memory 601 and executable on the processor 602 .

[0185] When the processor 602 executes the program, the active and passive joint tracking scheduling method for radar networking under main lobe interference provided in the above embodiment is implemented.

[0186] Furthermore, the electronic device further comprises:

[0187] The communication interface 603 is used for communication between the memory 601 and the processor 602 .

[0188] The memory 601 is used to store computer programs that can be executed on the processor 602 .

[0189] The memory 601 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0190] If the memory 601, the processor 602 and the communication interface 603 are implemented independently, the communication interface 603, the memory 601 and the processor 602 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0191] Optionally, in a specific implementation, if the memory 601, the processor 602 and the communication interface 603 are integrated on a chip, the memory 601, the processor 602 and the communication interface 603 can communicate with each other through an internal interface.

[0192] The processor 602 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0193] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned active and passive joint tracking and scheduling method for radar networking under main lobe interference.

[0194] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned radar network active and passive joint tracking and scheduling method under main lobe interference.

[0195] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0196] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0197] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0198] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0199] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0200] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0201] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0202] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A radar network active and passive joint tracking scheduling method under main lobe interference, characterized in that: The following steps are involved: Determine the working mode of each phased array radar in the networked detection system in the main lobe interference penetration scenario, and construct an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode; Determine the working mode vector of each phased array radar and the dwell time variable of each phased array radar on the target to be measured according to the active measurement model and the passive positioning model, so as to establish a joint tracking scheduling evaluation function through the working mode vector and the dwell time variable; A reinforcement learning agent corresponding to the networked detection system is constructed, and the reinforcement learning agent is trained by a preset proximal strategy optimization algorithm, so as to use the trained reinforcement learning agent to determine the target working mode of each phased array radar, and solve the joint tracking scheduling evaluation function by a preset convex optimization solution strategy to obtain the residence time of each phased array radar on the target to be measured.

2. The method according to claim 1, characterized in that The determining of the working mode of each phased array radar in the networked detection system in the main lobe interference penetration scenario, and constructing an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode, includes: When the phased array radar is in the active detection mode, obtaining radial distance and angle information of the phased array radar to the target to be measured, and constructing the active measurement model according to the radial distance and the angle information; When the phased array radar is in the passive detection mode, the interference source pointing line information of the phased array radar is obtained, and the passive positioning model is constructed according to the interference source pointing line information.

3. The method according to claim 2, characterized in that The establishing of a joint tracking scheduling evaluation function through the working mode vector and the residence time variable comprises: Determine the state transfer matrix and state transfer noise of the target to be measured in the kth detection cycle, and estimate the state vector of the target to be measured in the kth detection cycle according to the state transfer matrix, state transfer noise and the measurement result of the target to be measured in the k-1th detection cycle, where k is a positive integer; Calculating a priori information Fisher information matrix of the target to be detected based on the covariance matrix of the state transition noise, the state transition matrix and the state vector in the k-1th detection cycle; Calculating a radar measurement Jacobian matrix of each phased array radar for the target to be detected in the kth detection cycle, so as to calculate an active tracking Fisher information matrix of the target to be detected based on the radar measurement Jacobian matrix; Calculating the measurement noise of each phased array radar and the covariance matrix corresponding to the measurement noise, and calculating the passive tracking Fisher information matrix of the target to be measured based on the passive positioning model and the covariance matrix corresponding to the measurement noise; Based on the prior information Fisher information matrix, the passive tracking Fisher information matrix and the active tracking Fisher information matrix, calculating the Bayesian information matrix of the target to be detected in the k-th detection cycle; The Bayesian Cramer-Rao bound of the target to be detected is calculated according to the Bayesian information matrix in the k-th detection cycle, so as to establish the joint tracking scheduling evaluation function through the Bayesian Cramer-Rao bound, the working mode vector and the residence time variable.

4. The method according to claim 3, characterized in that The construction of the reinforcement learning agent corresponding to the network detection system includes: Based on a preset Markov decision strategy, determine the stable state of the reinforcement learning agent corresponding to the mth step in the Markov decision process, wherein the stable state includes the radar position of the phased array radar, the state vector of the target to be measured, the principal element of the prior information matrix of the target to be measured, the passive measurement error of the interference target in the target to be measured, and the unique hot vector of the phased array radar corresponding to the mth step, wherein m is a positive integer; Determining the dynamic state of the reinforcement learning agent through the radar code of the phased array radar, the working mode and the prediction error of the m-th step, and establishing the state space of the reinforcement learning agent according to the stable state and the dynamic state; Determining the action space of the reinforcement learning agent according to the working mode of the phased array radar; Based on the prediction error of the m-th step, the prediction error of the m-1-th step and the preset scaling hyperparameters, a reward function of the reinforcement learning agent is constructed.

5. The method according to claim 4, characterized in that The step of training the reinforcement learning agent by using a preset proximal strategy optimization algorithm includes: The preset proximal strategy optimization algorithm is constructed through a preset full-link network, and a target training environment is randomly generated to train the preset proximal strategy optimization algorithm using the target training environment, wherein the target training environment includes a randomly generated initial state and interference intensity of the target to be measured within the coverage range of the phased array radar; The reinforcement learning agent is trained based on the target training environment and the pre-set proximal strategy optimization algorithm after training.

6. The method according to claim 1, characterized in that The mathematical expression of the joint tracking scheduling evaluation function is: Among them, ∈ k is the working mode vector of each phased array radar in the kth detection cycle; τ q,k is the dwell time variable of each phased array radar on the target q in the kth detection cycle; M represents the number of phased array radars.

7. A radar network active and passive joint tracking and scheduling device under main lobe interference, characterized in that: include: A modeling module is used to determine the working mode of each phased array radar in the networked detection system in the main lobe interference penetration scenario, and to construct an active measurement model or a passive positioning model corresponding to the working mode, wherein the working mode includes an active detection mode and a passive detection mode; An establishment module is used to determine the working mode vector of each phased array radar and the residence time variable of each phased array radar for the target to be measured according to the active measurement model and the passive positioning model, so as to establish a joint tracking scheduling evaluation function through the working mode vector and the residence time variable; A solution module is used to construct a reinforcement learning agent corresponding to the networked detection system, and train the reinforcement learning agent through a preset proximal strategy optimization algorithm, so as to use the trained reinforcement learning agent to determine the target working mode of each phased array radar, and solve the joint tracking scheduling evaluation function through a preset convex optimization solution strategy to obtain the residence time of each phased array radar on the target to be measured.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the active and passive joint tracking and scheduling method for radar networking under main lobe interference as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the active and passive joint tracking and scheduling method for radar networking under main lobe interference as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed to implement the active and passive joint tracking scheduling method for radar networking under main lobe interference as described in any one of claims 1-6.

Images