Networking radar collaborative resource management algorithm based on random opportunity constraint programming

Abstract

The invention belongs to the technical field of radar system resource management and optimization, and particularly relates to a networking radar collaborative resource management algorithm based on random opportunity constraint programming. The algorithm comprises the following steps of: 1, determining a target priority by adopting a fuzzy reasoning algorithm; step 2, intelligent grouping of the radar and the target is realized by adopting a clustering algorithm; 3, taking the Bayesian Cramer-Rao bound as a measurement standard of target tracking performance, and constructing a random opportunity constraint programming model; and step 4, embedding a stochastic simulation algorithm into a genetic algorithm to form a hybrid intelligent algorithm to solve the stochastic opportunity constraint programming model, and obtaining a power optimization distribution result. According to the method, the problem of resource allocation during multi-target tracking can be solved, intelligent grouping of the radar and the targets is realized through fuzzy reasoning and clustering methods, and resource allocation based on chance constraint planning can ensure the robustness of an algorithm and can also ensure that a model accords with reality.

Description

Technical Field

[0001] The present application belongs to the field of radar system resource management and optimization technology, and in particular to a collaborative resource management algorithm for networked radars based on random chance constrained programming. Background Art

[0002] In the process of tracking multiple targets by networked radar, due to differences in various factors such as target location, category, and movement form, the priority of each target relative to different radar nodes should be different. Therefore, in many works on radar node allocation, target priority is an important reference. However, most of the literature uses closed-form solutions to calculate priority, which is just a simple assumption about the mapping relationship between target priority and influencing factors. In fact, due to the different dimensions of influencing factors and the coupling between some factors, target priority is difficult to accurately quantify through the above simple mapping relationship and is a typical fuzzy variable.

[0003] In theory, in order for the radar to obtain better detection performance, each transmitting beam of the radar should maximize the radar system resources it occupies. However, when the radar system is located on a mobile carrier platform and the system resources are limited, for multi-task and multi-functional applications, it is only necessary to pre-set the radar's detection performance for a certain target, and minimize the resources consumed by the radar system while meeting the radar detection performance, so as to complete more tasks. Radar system power resource management is an important part of radar system resource management, and power resource management is mainly reflected in the management of beam power size. As the degree of radar digitization continues to increase, the radar system can control the transmit power of each array element separately to adjust the transmit power of the synthetic beam, thereby saving more power resources to complete more tasks.

[0004] Traditional resource allocation models are generally deterministic models, but due to the uncertainty of the radar system and the target environment, the target's RCS, the attenuation factor of signal transmission, and the system noise are all uncertain. Under uncertain conditions, building the resource allocation model as a deterministic model (both the cost function and the constraint function are deterministic) can neither guarantee the robustness of the algorithm nor the established model is realistic.

[0005] Therefore, it is desired to have a technical solution to overcome or at least alleviate at least one of the above-mentioned defects of the prior art. Summary of the invention

[0006] The purpose of this application is to provide a networked radar collaborative resource management algorithm based on random chance constrained programming to solve at least one problem existing in the prior art.

[0007] The technical solution of this application is:

[0008] A collaborative resource management algorithm for networked radars based on random chance constrained programming, including:

[0009] Step 1: Use fuzzy reasoning algorithm to determine the target priority;

[0010] Step 2: Use clustering algorithm to realize intelligent grouping of radar and target;

[0011] Step 3: Using the Bayesian Cramer Rao bound as the target tracking performance criterion, a random chance constrained programming model is constructed;

[0012] Step 4: embed the random simulation algorithm into the genetic algorithm to form a hybrid intelligent algorithm to solve the random chance constraint programming model to obtain the power optimization allocation result.

[0013] In at least one embodiment of the present application, in step 1, a fuzzy inference algorithm is used to determine the target priority, including:

[0014] S11, obtaining input changes, inputting the input changes into the fuzzification module for fuzzification processing, and converting them into fuzzy sets on a given domain;

[0015] S12, activating the fuzzy rules in the fuzzy rule base, and performing fuzzy logic reasoning on the fuzzy set to obtain a fuzzy logic reasoning result;

[0016] S13, defuzzifying the fuzzy logic reasoning result to obtain the target priority.

[0017] In at least one embodiment of the present application, the input change includes tracking scene, target class attribute, and motion form.

[0018] In at least one embodiment of the present application, in S11, the fuzzification module uses a triangle fuzzification method to perform fuzzy processing on the input change.

[0019] In at least one embodiment of the present application, in step 2, a clustering algorithm is used to implement intelligent grouping of radars and targets, including:

[0020] S21, assuming that at time k, the fuzzy logic reasoning results of each target for all radar nodes Form an N×Q matrix W k ;

[0021] S22, let the matrix W temp,k =W k ;

[0022] S23, from the matrix W temp,k Select the one with the highest value Assign radar i′ to target q′, matrix W temp,kAll data in the i′th row and q′th column are set to zero, and the matrix W k All data in row i′ are set to zero;

[0023] S24, repeat step S23 until the number of radars assigned to each target is M, and finally obtain the radar node clustering situation in, Represents the radar node clustering for target q.

[0024] In at least one embodiment of the present application, in step 3, a random chance constrained programming model is constructed using the Bayesian Cramer-Rao bound as a measure of target tracking performance, including:

[0025] S31. Using the Bayesian Cramer Rao bound as a measure of target tracking performance, the Bayesian Cramer Rao bound is:

[0026]

[0027] in, represents the target state vector at time k, represents the radar node clustering for target q, represents the radar transmission power, represents the Bayesian Cramé-Rao matrix;

[0028] S32. According to the radar node clustering results, the transmit power of each radar node is optimally allocated to obtain a random chance constrained programming model:

[0029]

[0030] in, represents the probability measure, η represents the target tracking error threshold, α represents the chance constraint confidence level, represents the set of RCS values ​​of target q in different radar directions, P max represents the upper limit of radar transmission power, P min Indicates the lower bound of radar transmit power.

[0031] In at least one embodiment of the present application, in step 4, a random simulation algorithm is embedded into a genetic algorithm to form a hybrid intelligent algorithm to solve the random chance constraint programming model to obtain a power optimization allocation result, including:

[0032] S41, initialize the population, set the number of evolutions to G max , the population size is popsize, the mutation probability is Pm, and the crossover probability is Pc;

[0033] S42, verifying whether the initial chromosome satisfies the constraints by a random simulation algorithm, and updating the chromosomes that do not satisfy the constraints until the constraints are satisfied;

[0034] S43, according to the mutation probability Pm, the chromosome is updated through the mutation operation, and the new chromosome is verified through the random simulation algorithm whether it satisfies the constraint. If it does not satisfy the constraint, the mutation operation is repeated until the constraint is satisfied;

[0035] S44, according to the crossover probability Pc, the chromosome is updated through a crossover operation, and a random simulation algorithm is used to verify whether the new chromosome satisfies the constraints. If the constraints are not satisfied, the crossover operation is repeated until the constraints are satisfied;

[0036] S45, calculating the objective function value of each chromosome in the population to obtain the corresponding fitness function;

[0037] S46. Based on the fitness function, a roulette algorithm is used to select popsize chromosomes to form a new population;

[0038] S47, repeat steps S42 to S46 until the cycle is G times,

[0039] max returns the optimal chromosome as the power optimization allocation result

[0040] In at least one embodiment of the present application, in S42, S43, and S44, verifying whether the chromosome satisfies the constraints by a random simulation algorithm includes:

[0041] S401, set S′=0;

[0042] S402: Randomly generate a set of sample vectors from a set of RCS values ​​of target q in different radar directions

[0043] S403, if Then let S′=S′+1;

[0044] S404, repeat step S402 to step S403 for S times;

[0045] S405, let Pr simulation =S′ / S, if Pr simulation ≥a, then the decision variable If the constraints are met, the probability can reach the confidence level, otherwise it cannot.

[0046] The invention has at least the following beneficial technical effects:

[0047] The networked radar collaborative resource management algorithm based on random chance-constrained programming of the present application can solve the resource allocation problem during multi-target tracking, and realize intelligent grouping of radars and targets through fuzzy reasoning and clustering methods. The resource allocation based on chance-constrained processing (CCP) can ensure the robustness of the algorithm and the conformity of the model to reality. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flowchart of a collaborative resource management algorithm for networked radars based on random chance constrained programming according to an implementation mode of the present application;

[0049] Figure 2 is a flow chart of a fuzzy reasoning algorithm of one embodiment of the present application;

[0050] Figure 3 It is a schematic diagram of a radar node clustering algorithm based on fuzzy target priority according to an embodiment of the present application;

[0051] Figure 4 is a schematic diagram of radar node and target distribution according to an embodiment of the present application;

[0052] Figure 5 This is a schematic diagram of the priorities of targets relative to different radar nodes in one embodiment of the present application;

[0053] Figure 6 It is a schematic diagram of radar node clustering and power allocation in one embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical scheme and advantages of the implementation of this application clearer, the technical scheme in the embodiment of this application will be described in more detail below in conjunction with the drawings in the embodiment of this application. In the drawings, the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The described embodiments are part of the embodiments of this application, not all of them. The embodiments described below with reference to the drawings are exemplary and are intended to be used to explain this application, and should not be construed as limitations on this application. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. The embodiments of this application are described in detail below in conjunction with the drawings.

[0055] In the description of the present application, it should be understood that the terms "center", "longitudinal", "lateral", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the scope of protection of the present application.

[0056] The following is combined with Figures 1 to 6 This application is described in further detail.

[0057] This application provides a collaborative resource management algorithm for networked radars based on random chance constrained programming, such as Figure 1 As shown, the following steps are included:

[0058] Step 1: Use fuzzy reasoning algorithm to determine the target priority;

[0059] Step 2: Use clustering algorithm to realize intelligent grouping of radar and target;

[0060] Step 3: Using the Bayesian Cramer Rao bound as the target tracking performance criterion, a random chance constrained programming model is constructed;

[0061] Step 4: Embed the random simulation algorithm into the genetic algorithm to form a hybrid intelligent algorithm to solve the random chance constraint programming model and obtain the power optimization allocation result.

[0062] The present application discloses a collaborative resource management algorithm for networked radars based on random chance constrained programming. First, in step one, a fuzzy reasoning algorithm is used to determine the target priority. Usually, system behavior decisions need to consider the combined impact of multiple factors, but it is difficult to find an accurate mapping relationship between these factors and decision results. Fuzzy logic reasoning is similar to the reasoning and decision-making process of the human brain. Based on fuzzy logic rules, it can achieve complex mapping of multiple input spaces to one output space. Therefore, fuzzy logic reasoning is used in the application to calculate the fuzzy priority of the target relative to the radar. Its main process is as follows: Figure 2 As shown. The fuzzy reasoning algorithm is used to determine the target priority, including:

[0063] S11, obtaining input changes, inputting the input changes into the fuzzification module for fuzzification processing, and converting them into fuzzy sets on a given domain;

[0064] S12, activating the fuzzy rules in the fuzzy rule base, performing fuzzy logic reasoning on the fuzzy set to obtain a fuzzy logic reasoning result;

[0065] S13, defuzzifying the fuzzy logic reasoning results to obtain the target priority.

[0066] In this embodiment, the input changes include tracking scenes, target attributes, and motion forms. In S11, the fuzzification module uses a triangle fuzzification method to fuzzify the input changes. Preferably, three representative target priority influencing factors are considered: target speed, distance from the radar node, and target threat level.

[0067] In the networked radar collaborative resource management algorithm based on random chance constrained programming of the present application, in step 2, a radar node clustering algorithm based on fuzzy target priority is used to realize intelligent grouping of radars and targets. According to the target priority calculated in step 1, the radar node clustering algorithm based on fuzzy target priority includes the following process:

[0068] S21, assuming that at time k, the fuzzy logic reasoning results of each target for all radar nodes Form an N×Q matrix W k ;

[0069] S22, let the matrix W temp,k =W k ;

[0070] S23, from the matrix W temp,k Select the one with the highest value Assign radar i′ to target q′, matrix W temp,k All data in the i′th row and q′th column are set to zero, and the matrix W k All data in row i′ are set to zero;

[0071] S24, repeat step S23 until the number of radars assigned to each target is M, and finally obtain the radar node clustering situation in, Represents the radar node clustering for target q.

[0072] For ease of understanding, Figure 3 A schematic diagram of radar node clustering algorithm based on fuzzy target priority is given. In the figure, radar numbers are represented as R1, R2, …, R10; and target numbers are represented as T1, T2 and T3.

[0073] The networked radar collaborative resource management algorithm based on random chance constrained programming of the present application optimizes the allocation of the transmit power of each radar node after obtaining the clustering situation of the radar node for each target, aiming to minimize the radar transmit power consumption on the basis of meeting the target tracking accuracy.

[0074] In step 3, the Bayesian Cramer-Rao bound is used as the target tracking performance criterion to construct a random chance constrained programming model, including:

[0075] S31. Using the Bayesian Cramer Rao bound as a measure of target tracking performance, the Bayesian Cramer Rao bound is:

[0076]

[0077] in, represents the target state vector at time k, represents the radar node clustering for target q, represents the radar transmission power, represents the Bayesian Cramé-Rao matrix;

[0078] S32. According to the radar node clustering results, the transmit power of each radar node is optimally allocated to obtain a random chance constrained programming model:

[0079]

[0080] in, represents the probability measure, η represents the target tracking error threshold, α represents the chance constraint confidence level, represents the set of RCS values ​​of target q in different radar directions, P max represents the upper limit of radar transmission power, P min Indicates the lower bound of radar transmit power.

[0081] In this embodiment, step S31 can be specifically divided into the following processes:

[0082] In nonlinear filtering problems, This provides a lower bound on the target state estimation error. Assume represents the target q state vector The unbiased estimate of , then the BCRLB inequality can be expressed as:

[0083]

[0084] In the formula, Indicates about and The mathematical expectation of Represents the Bayesian Information Matrix (BIM), which can be expressed as:

[0085]

[0086] In the formula, is the prior information matrix, and its expression is:

[0087]

[0088] In the formula, Represents the data information matrix, and the expression is:

[0089]

[0090] In the formula, represents the Jacobian matrix.

[0091] Therefore, the BIM matrix can be expressed as:

[0092]

[0093] Where, the measurement noise covariance matrix is is the transmit power function.

[0094] Therefore, the BCRLB matrix can be expressed as:

[0095]

[0096] In the formula, the vector Represents the transmission power distribution of all radar nodes tracking target q.

[0097] The Bayesian Craméro matrix is the inverse matrix of the Bayesian information matrix J(·),

[0098] is the prior information matrix, is the data information matrix.

[0099] because Each element on the diagonal of the matrix gives a lower bound for the estimation error of the corresponding target state component. Therefore, the following expression is given as a function to measure the target tracking accuracy:

[0100]

[0101] In this embodiment, step S32 can be specifically divided into the following processes:

[0102] Get radar node clustering results After that, the transmit power of each radar node is optimally allocated. The purpose is to minimize the total power consumption of unmanned cluster radar tracking multiple targets on the basis of meeting the multi-target tracking accuracy. The following radar power optimization allocation model is established:

[0103]

[0104] Where, the first constraint indicates that the tracking accuracy of each target must be better than the target tracking error threshold η; the second constraint gives the upper bound P of each radar’s transmit power max With the lower bound P min .

[0105] In addition to the decision variables In addition, the target tracking accuracy is also affected by the random variable target In the radar resource allocation problem, if the randomness of the target RCS is ignored, the algorithm output will be unreliable. In order to cope with the impact of random RCS and enable the algorithm to output stable and reliable resource allocation results, the radar power optimization allocation model is transformed into the following random chance constrained programming model:

[0106]

[0107] In the formula, represents the probability measure, which can be understood as the probability that the inequality in the curly brackets holds; α represents the confidence level of the chance constraint; represents the set of RCS values ​​of target q in different radar directions. Compared with the radar power optimization allocation model, the random chance constrained programming model allows random variables in its constraints, but it is necessary to ensure that the model is Under random changes, the probability that the output power allocation results meet the constraints is always greater than the confidence level α.

[0108] The networked radar collaborative resource management algorithm based on random chance constrained programming of the present application, finally, due to the existence of chance constraints, the random chance constrained programming model is a non-convex and nonlinear optimization problem, which is difficult to solve directly using convex optimization methods. Therefore, in this embodiment, the random simulation is embedded in the genetic algorithm to form a hybrid intelligent algorithm to solve it directly. In step 4, the random simulation algorithm is embedded in the genetic algorithm to form a hybrid intelligent algorithm to solve the random chance constrained programming model, and the power optimization allocation result is obtained, including:

[0109] S41, initialize the population, set the number of evolutions to G max , the population size is popsize, the mutation probability is Pm, and the crossover probability is Pc;

[0110] S42, verifying whether the initial chromosome satisfies the constraints by a random simulation algorithm, and updating the chromosomes that do not satisfy the constraints until the constraints are satisfied;

[0111] S43, according to the mutation probability Pm, the chromosome is updated through the mutation operation, and the new chromosome is verified through the random simulation algorithm whether it satisfies the constraint. If it does not satisfy the constraint, the mutation operation is repeated until the constraint is satisfied;

[0112] S44, according to the crossover probability Pc, the chromosome is updated through a crossover operation, and a random simulation algorithm is used to verify whether the new chromosome satisfies the constraints. If the constraints are not satisfied, the crossover operation is repeated until the constraints are satisfied;

[0113] S45, calculating the objective function value of each chromosome in the population to obtain the corresponding fitness function;

[0114] S46. Based on the fitness function, a roulette algorithm is used to select popsize chromosomes to form a new population;

[0115] S47, repeat steps S42 to S46 until loop G max times, returns the optimal chromosome as the power optimization allocation result

[0116] Random simulation can obtain the probability that the decision variables meet the random chance constraints through a large number of numerical simulation calculations. Taking the target q as an example, in S42, S43, and S44, the random simulation algorithm is used to verify whether the chromosome meets the constraints. The specific steps are as follows:

[0117] S401, set S′=0;

[0118] S402: Randomly generate a set of sample vectors from a set of RCS values ​​of target q in different radar directions

[0119] S403, if Then let S′=S′+1;

[0120] S404, repeat step S402 to step S403 for S times;

[0121] S405, let Pr simulation =S′ / S, if Pr simulation ≥a, then the decision variable If the constraints are met, the probability can reach the confidence level, otherwise it cannot.

[0122] After the above solution process, we can get the radar node clustering of the target tracking at each moment. And the distribution of transmission power Since different targets are dispersed in space, the problem of multi-radar tracking of multiple targets in this application can be split into several multi-radar tracking of single targets. The target state estimation can be obtained by using UKF.

[0123] In a specific implementation of the present application, the parameters are configured as follows: In order to verify the effectiveness and robustness of the radar node and power collaborative management algorithm, the following simulation experiment is designed. Assume that there are 3 targets moving in a straight line at a uniform speed in the tracking scenario. Taking into account the limited scale of the task, a small-scale cluster is used for simulation, assuming that an unmanned cluster containing 12 radar nodes performs collaborative tracking of these targets. The system always assigns M = 3 radar nodes to each target. The radar node position information is given in Table 1. The radar waveform parameters are given in Table 2. The target initial position and motion information are given in Table 3. Assume that the tracking error threshold η = 100; the lower limit of the radar transmit power is P min =0.05P total ; The upper limit is P max =0.8P total ; The confidence level of random chance constraint is α = 0.9. The simulation experiment has 40 frames of data.

[0124] Table 1 Radar node location information

[0125]

[0126] Table 2 Radar waveform parameters

[0127]

[0128] Table 3 Target initial state information

[0129]

[0130] The distribution of radar nodes and targets is shown in Figure 4 Given.

[0131] Due to differences in position and movement path, each target has different priorities relative to different radars. As the target moves, its priority will change accordingly. Figure 5 The priority changes of each target relative to different radar nodes during the tracking process are given. Among them, red indicates a higher priority, and blue indicates a lower priority. It can be found that the target closer to the radar node has a higher priority. Figure 6The radar node clustering and power allocation (the ratio of actual transmit power to the upper limit of transmit power) output by the radar node and power cooperative management algorithm are given. According to the target priority, the system tends to select radar nodes closer to the target to track the target. As the target position changes, the radar node and power allocation of each target also change accordingly. Since the RCS of the target is related to many factors such as the observation angle, aircraft attitude, polarization mode, etc., the power resource constraint planning model of the random chance constraint programming model is adopted to enable the radar to complete as many tasks as possible under the premise of meeting the pre-set tracking error. It can be seen from Table 4 that under the premise of meeting the pre-set tracking error, as the confidence level decreases, the radar can save more transmit power, thereby completing more tasks. The confidence level is set within an acceptable risk range.

[0132] Table 4 Comparison of total power consumption of target tracking for different optimization models

[0133]

[0134] The networked radar collaborative resource management algorithm based on random chance constrained programming of the present application adopts fuzzy reasoning algorithm to calculate target priority by integrating multi-dimensional composite factors of the target, designs radar node clustering algorithm based on target priority to intelligently group radars and targets, represents the randomly fluctuating target RCS with random variables, establishes a random chance constrained programming model, embeds the random simulation algorithm into the genetic algorithm to form a hybrid intelligent optimization algorithm to solve the power resource allocation result that satisfies the chance constraint. The present application combines fuzzy logic reasoning and clustering algorithm to select the optimal radar combination to achieve target tracking; at the same time, considering the information uncertainty brought by the target RCS, introduces the random chance constrained programming model, controls the relationship between risk and confidence level, obtains a more robust power allocation scheme, and also provides a reliable basis for radar system resource allocation decision-making.

[0135] The networked radar collaborative resource management algorithm based on random chance constrained programming of the present application adopts a random chance constrained programming model, which can not only ensure the robustness of the algorithm, but also better handle the uncertainty of target measurement information, making the target model closer to reality. The model takes into account that the decision made may not meet the constraints in some extreme cases. The principle adopted by the model is: it is allowed that the planning scheme does not fully meet the constraints, but the probability that the constraints are established is not less than a given confidence level. By setting the confidence level, the risk of system default can be effectively regulated, and at the same time, resources can be greatly saved because the extreme cases where the constraints are met at a very small confidence level are abandoned.

[0136] The networked radar collaborative resource management algorithm based on random chance constraint programming of the present application, the random chance constraint programming model can complete as many tasks as possible under the condition of ensuring that the preset tracking accuracy is met; at the same time, the randomness of the target RCS is considered, and the power resource allocation algorithm of the random chance constraint programming model is used to deal with the problem, which can control the relationship between risk and confidence level and is more robust; at the same time, the present invention adopts a fuzzy logic reasoning algorithm to calculate the target priority, which can integrate the multi-dimensional complex factors of the target; the planning schemes under different confidence levels also provide a reliable basis for making resource allocation decisions.

[0137] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A collaborative resource management algorithm for networked radars based on random chance constrained programming, characterized in that: include: Step 1: Use fuzzy reasoning algorithm to determine the target priority; Step 2: Use clustering algorithm to realize intelligent grouping of radar and target; Step 3: Using the Bayesian Cramer Rao bound as the target tracking performance criterion, a random chance constrained programming model is constructed; Step 4: embed the random simulation algorithm into the genetic algorithm to form a hybrid intelligent algorithm to solve the random chance constraint programming model to obtain the power optimization allocation result.

2. The networked radar collaborative resource management algorithm based on random chance constrained programming according to claim 1 is characterized in that: In step 1, a fuzzy reasoning algorithm is used to determine the target priority, including: S11, obtaining input changes, inputting the input changes into the fuzzification module for fuzzification processing, and converting them into fuzzy sets on a given domain; S12, activating the fuzzy rules in the fuzzy rule base, and performing fuzzy logic reasoning on the fuzzy set to obtain a fuzzy logic reasoning result; S13, defuzzifying the fuzzy logic reasoning result to obtain the target priority.

3. The networked radar collaborative resource management algorithm based on random chance constrained programming according to claim 2 is characterized in that: The input changes include tracking scenes, target attributes, and motion forms.

4. The networked radar collaborative resource management algorithm based on random chance constrained programming according to claim 3 is characterized in that: In S11, the fuzzification module uses a triangle fuzzification method to perform fuzzification processing on the input change.

5. The networked radar collaborative resource management algorithm based on random chance constrained programming according to claim 4 is characterized in that: In step 2, a clustering algorithm is used to realize intelligent grouping of radars and targets, including: S21, assuming that at time k, the fuzzy logic reasoning results of each target for all radar nodes Form an N×Q matrix W k ; S22, let the matrix W temp,k =W k ; S23, from the matrix W temp,k Select the one with the highest value Assign radar i′ to target q′, matrix W temp,k All data in the i′th row and q′th column are set to zero, and the matrix W k All data in row i′ are set to zero; S24, repeat step S23 until the number of radars assigned to each target is M, and finally obtain the radar node clustering situation in, Represents the radar node clustering for target q.

6. The networked radar collaborative resource management algorithm based on random chance constrained programming according to claim 5 is characterized in that: In step 3, the Bayesian Cramer-Rao bound is used as the target tracking performance criterion to construct a random chance constrained programming model, including: S31. Using the Bayesian Cramer Rao bound as a measure of target tracking performance, the Bayesian Cramer Rao bound is: in, represents the target state vector at time k, represents the radar node clustering for target q, represents the radar transmission power, represents the Bayesian Cramé-Rao matrix; S32. According to the radar node clustering results, the transmit power of each radar node is optimally allocated to obtain a random chance constrained programming model: in, represents the probability measure, η represents the target tracking error threshold, α represents the chance constraint confidence level, represents the set of RCS values ​​of target q in different radar directions, P max represents the upper limit of radar transmission power, P min Indicates the lower bound of radar transmit power.

7. The networked radar collaborative resource management algorithm based on random chance constrained programming according to claim 6 is characterized in that: In step 4, the random simulation algorithm is embedded into the genetic algorithm to form a hybrid intelligent algorithm to solve the random chance constraint programming model to obtain the power optimization allocation result, including: S41, initialize the population, set the number of evolutions to G max , the population size is popsize, the mutation probability is Pm, and the crossover probability is Pc; S42, verifying whether the initial chromosome satisfies the constraints by a random simulation algorithm, and updating the chromosomes that do not satisfy the constraints until the constraints are satisfied; S43, according to the mutation probability Pm, the chromosome is updated through the mutation operation, and the new chromosome is verified through the random simulation algorithm whether it satisfies the constraint. If it does not satisfy the constraint, the mutation operation is repeated until the constraint is satisfied; S44, according to the crossover probability Pc, the chromosome is updated through a crossover operation, and a random simulation algorithm is used to verify whether the new chromosome satisfies the constraints. If the constraints are not satisfied, the crossover operation is repeated until the constraints are satisfied; S45, calculating the objective function value of each chromosome in the population to obtain the corresponding fitness function; S46. Based on the fitness function, a roulette algorithm is used to select popsize chromosomes to form a new population; S47, repeat steps S42 to S46 until the cycle is G times, max returns the optimal chromosome as the power optimization allocation result 8. The networked radar collaborative resource management algorithm based on random chance constrained programming according to claim 7 is characterized in that: In S42, S43, and S44, a random simulation algorithm is used to verify whether the chromosome satisfies the constraints, including: S401, set S′=0; S402: Randomly generate a set of sample vectors from a set of RCS values ​​of target q in different radar directions S403, if Then let S′=S′+1; S404, repeat step S402 to step S403 for S times; S405, let Pr simulation =S′ / S, if Pr simulation ≥a, then the decision variable If the constraints are met, the probability can reach the confidence level, otherwise it cannot.

Images