Active phased array system radar cognitive interference signal generation method, medium and equipment

Abstract

The invention discloses an active phased array system radar cognitive interference signal generation method, a medium and equipment, and relates to the technical field of radar electronic countermeasures, and the active phased array system radar cognitive interference signal generation method mainly comprises the steps: constructing a radar electronic countermeasure simulation model, obtaining a signal style state conversion table by using a TOPSIS sorting algorithm and the minimum interference-to-signal ratio data table; according to the radar electronic countermeasure simulation model, obtaining an interference pattern selection model by using the signal pattern state conversion table and a DDQN network; and performing optimal interference pattern decision by using the interference pattern selection model. By implementing the active phased array system radar cognitive interference signal generation method, the medium and the equipment provided by the invention, the accuracy and applicability of radar interference decision can be improved.

Description

Technical Field

[0001] The present invention relates to the field of radar electronic countermeasure technology, and more specifically, to a method, medium and device for generating cognitive jamming signals for an active phased array radar. Background Art

[0002] In the field of radar electronic countermeasures, airborne active phased array fire control radars are widely used, such as fighter jets and missile defense systems in the military field. At present, airborne active phased array fire control radars include typical signal styles such as STT, TAS, TWS, RWS, and VS. Cognitive jammers include typical DRFM jammers such as comb spectrum jammers, sample jammers, slice jammers, Doppler noise jammers, random false target jammers, and delayed superposition smart noise jammers. Traditional jamming decision methods may not be able to cope with complex working modes and flexible and changeable signal styles; in the current jamming decision methods based on reinforcement learning, there is no reasonable basis for the radar state transition relationship, and they all rely on prior knowledge to construct the transition relationship table, which lacks adaptation to the actual radar system. At the same time, the current DQN algorithm has the problem of overestimation of Q value in the decision-making process, which will affect the effect of the decision. In the face of the complex and changeable signal styles of airborne active phased array fire control radars, traditional experience-based jamming decisions are difficult to take effect. In addition, with the continuous development of radar technology and the expansion of application fields, the demand for more intelligent and adaptive jamming decision methods is also increasing. Summary of the invention

[0003] The object of the present invention is to provide a method for generating cognitive jamming signals for an active phased array radar, a storage medium, a computer device and a computer program product, which can improve the accuracy and applicability of radar jamming decisions.

[0004] The present invention provides a method for generating cognitive jamming signals for an active phased array radar, comprising the following steps: S1: constructing a radar electronic countermeasure simulation model, and obtaining a signal pattern state transition table by using a TOPSIS sorting algorithm and a minimum interference-to-signal ratio data table; S2: obtaining an interference pattern selection model by using a signal pattern state transition table and a DDQN network according to the radar electronic countermeasure simulation model; and S3: making an optimal interference pattern decision by using the interference pattern selection model.

[0005] Furthermore, step S1 specifically includes: S11: constructing a radar electronic countermeasure simulation model, the radar electronic countermeasure simulation model includes an airborne active phased array radar model and a cognitive jammer model; S12: obtaining an error measurement result according to the radar electronic countermeasure simulation model; S13: sorting the error measurement result using the TOPSIS sorting algorithm to obtain a sorting result; S14: correcting the sorting result using the minimum interference-to-signal ratio data table to obtain a corrected TOPSIS sorting table; S15: according to the corrected TOPSIS sorting table, obtaining a signal style state transition table using a conversion rule.

[0006] Furthermore, the above-mentioned airborne active phased array radar model includes an all-entry angle search AAS working mode, a high-entry angle search HAS working mode, a mixed-entry angle search CAM working mode, and a single-target tracking STT working mode; the airborne active phased array radar model also includes a radar signal style set, and the radar signal style set includes a single-target tracking STT radar signal style, a search and tracking TAS radar signal style, a search and tracking TWS radar signal style, a medium-repetition-frequency search and ranging RWS-1 radar signal style, a high-repetition-frequency search and ranging RWS-2 radar signal style, and a speed search VS radar signal style; the cognitive jammer model includes an interference signal style set; the interference signal style set includes a comb spectrum interference signal style, a sample interference signal style, a Doppler noise interference signal style, a random false target interference signal style, a delayed superposition smart noise interference signal style, and a slice interference signal style.

[0007] Furthermore, the above error measurement results include ranging error, speed measurement error, false alarm probability and detection probability.

[0008] Furthermore, step S13 specifically includes:

[0009] S131: According to the error measurement results, a decision matrix is ​​obtained, such as the formula:

[0010]

[0011] Among them, K is the decision matrix, k nm are the parameter values ​​corresponding to the nth evaluation index in the mth reconnaissance result;

[0012] S132: Normalize the decision matrix to obtain a normalized matrix, such as the formula:

[0013]

[0014] Among them, k ij is the i-th row and j-th column element of the decision matrix; k′ ij is the element of the i-th row and j-th column of the normalized matrix;

[0015] S133: Weight the normalized matrix to obtain a weighted matrix, such as the formula:

[0016] z ij =w j k ij ′(i=1,2,L,m; j=1,2,L,n),

[0017] W=[w1,w2,L,w n ] T ,

[0018] Among them, z ij is the i-th row and j-th column element of the weighted matrix; w j is the jth element of the weight vector, W is the weight vector;

[0019] S134: According to the weighted matrix, the optimal reconnaissance result and the worst reconnaissance result are obtained, such as the formula:

[0020]

[0021] Among them, I + is the optimal reconnaissance result, is the jth element of the optimal reconnaissance node, I - The worst reconnaissance result. is the jth element of the worst reconnaissance result, and max{·} means taking the maximum value;

[0022] S135: Calculate the distance between each reconnaissance result and the best reconnaissance result and the worst reconnaissance result respectively, and obtain the first distance and the second distance, as shown in the formula:

[0023]

[0024] in, is the first distance, is the second distance;

[0025] S136: According to the first distance and the second distance, the quality of each reconnaissance result is obtained, as shown in the formula:

[0026]

[0027] Among them, A i is the quality of the i-th reconnaissance result;

[0028] S137: Obtain ranking results based on the quality of each reconnaissance result.

[0029] Furthermore, step S2 specifically includes: S21: according to the radar electronic countermeasure simulation model, using the signal style state transition table and the DDQN network, obtaining an adversarial sample; S22: constructing and updating a memory library according to the adversarial sample; S23: using the memory library as a training sample to train the DDQN network to obtain an interference style selection model.

[0030] Further, step S21 specifically includes: S211: according to the radar electronic countermeasure simulation model, a state set and an action set are obtained; the state set includes a state, and the state is used to represent the signal style of the enemy's airborne active phased array fire control radar at a certain moment; the action set includes an action, and the action is used to represent the interference signal style selected by our jammer at that moment; S212: according to the state set and the action set, the Q value corresponding to each interference signal style is obtained by using the signal style state transition table and the DDQN network; S213: according to the Q value corresponding to each interference signal style, the first interference signal style is obtained by using the ε-greed strategy; S214: according to the first interference signal style, the change of the enemy radar signal style is obtained by using the radar electronic countermeasure simulation model and the signal style state transition table, and the first reward value is obtained according to the change of the enemy radar signal style; S215: according to the first reward value, the Q value is updated by using the DDQN network, and the next state is obtained by using the signal style state transition table; S216: the state, the first interference signal style, the first reward value and the next state are combined into a four-tuple sample to obtain an adversarial sample.

[0031] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned active phased array radar cognitive interference signal generation method are implemented.

[0032] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the above-mentioned active phased array radar cognitive interference signal generation method are implemented.

[0033] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned active phased array radar cognitive interference signal generation method.

[0034] The active phased array radar cognitive jamming signal generation method, storage medium, computer device and computer program product provided by the present invention have the following beneficial effects:

[0035] The present invention combines the simulation data obtained by the radar electronic countermeasure simulation model with the small sample measured data to form a fully effective signal style conversion table, and obtains the optimal interference decision-making plan through the DDQN algorithm; wherein the signal style conversion table is obtained by relying on the radar signal processing simulation system and the actual radar interference experiment; during the implementation of the signal style conversion table, the simulation system reads the radar signal parameters from the simulated radar data, uses the signal level simulation model for simulation, and adds different styles of interference therein, calculates the ranging error, speed measurement error, false alarm probability and detection probability and other indicators according to the results obtained by the PD / CFAR processing part of the system before and after the interference, and the interference effect can be sorted according to the TOPSIS sorting algorithm; wherein, the simulated radar system interference experiment is used to obtain the topsis sorting result of the target measurement error, and the actual interference experiment is used to obtain the minimum interference-signal ratio of different interference styles to different radar states, and uses The topsis sorting table is corrected by the minimum interference-signal ratio data table, and the signal style state transition is obtained according to the corrected topsis sorting table; then the adversarial memory library is generated according to the signal style transition table, and finally the optimal interference style is obtained through DDQN; wherein, the implementation of DDQN interference decision mainly relies on two neural networks to fit the Q value: MainNet is used to estimate the Q value, and TargetNet is used to generate the Q target value; the network parameters are updated by the gradient descent algorithm through the formula of the loss function; in order to improve the effect of interference training, the present invention adopts an adversarial memory library to store radar signal samples before and after interference, interference signal samples and their corresponding interference effect values; the interference effect value is determined by comparing the changes in the signals before and after interference; during the interference training process, samples in the memory library are randomly extracted to avoid the correlation between consecutive samples causing significant fluctuations in network fitting;

[0036] The present invention can cope with the challenge that traditional empirical jamming decisions are difficult to be effective in airborne active phased array fire control radars under complex working modes and changeable signal patterns; at the same time, the present invention also solves the problem of constructing a signal pattern conversion matrix for phased array radars, constructs a signal pattern conversion table through a radar signal processing simulation system and a real radar experiment, forms a countermeasure memory library for radars and jammers, and DDQN samples and trains the memory library and updates network parameters; the trained DDQN can decide on the optimal jamming solution;

[0037] The present invention constructs a signal style conversion table by utilizing error measurement results of a radar signal processing system and experimental data of an actual installed radar, thereby making up for the deficiencies in traditional methods and improving the accuracy and applicability of interference decisions. The present invention introduces a DDQN algorithm, and uses different value function networks for action selection and evaluation, and then uses different value function networks to implement action selection and action evaluation in the DDQN algorithm, thereby effectively eliminating the problem of over-estimation of the Q value, improving the accuracy and reliability of interference decisions, making the decision results more effective, and helping to improve the overall performance and robustness of radar electronic countermeasure systems, meeting the needs of scientific analysis and research, and having positive significance for the research on interference decision-making of complex airborne fire control radars. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0039] Figure 1 It is a flow chart of a method for generating cognitive jamming signals for active phased array radar provided by the present invention;

[0040] Figure 2 It is a flow chart of a cognitive jamming signal generation algorithm for an active phased array radar provided by the present invention;

[0041] Figure 3 It is a schematic diagram of the result of radar signal processing before interference provided by the present invention;

[0042] Figure 4 is a schematic diagram of radar signal processing after interference provided by the present invention;

[0043] Figure 5 It is a schematic diagram of the DQN algorithm model provided by the present invention;

[0044] Figure 6 It is a schematic diagram of the loss values ​​of the DQN and DDQN algorithms provided by the present invention;

[0045] Figure 7 It is a schematic diagram of the maximum Q value of the DQN and DDQN algorithms provided by the present invention;

[0046] Figure 8 It is a schematic diagram of the total reward value of the DQN and DDQN algorithms provided by the present invention;

[0047] Fig. 9 It is a structural block diagram of the computer device provided by the present invention. DETAILED DESCRIPTION

[0048] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0049] Figure 1 A schematic diagram of a method for generating cognitive jamming signals for an active phased array radar of this embodiment is shown. In this embodiment, the method for generating cognitive jamming signals for an active phased array radar comprises the following steps:

[0050] S1: Build a radar electronic countermeasure simulation model, use the TOPSIS sorting algorithm and the minimum interference-to-signal ratio data table to obtain the signal pattern state transition table;

[0051] In an exemplary embodiment, step S1 specifically includes:

[0052] S11: Build a radar electronic countermeasure simulation model, which includes an airborne active phased array radar model and a cognitive jammer model;

[0053] For an exemplary embodiment, the airborne active phased array radar model includes an all-entry angle search AAS working mode, a high-entry angle search HAS working mode, a mixed-entry angle search CAM working mode, and a single-target tracking STT working mode; the airborne active phased array radar model also includes a radar signal style set, and the radar signal style set includes a single-target tracking STT radar signal style, a search and tracking TAS radar signal style, a search and tracking TWS radar signal style, a medium-repetition-frequency search and ranging RWS-1 radar signal style, a high-repetition-frequency search and ranging RWS-2 radar signal style, and a speed search VS radar signal style; the cognitive jammer model includes an interference signal style set; the interference signal style set includes a comb spectrum interference signal style, a sample interference signal style, a Doppler noise interference signal style, a random false target interference signal style, a delayed superposition smart noise interference signal style, and a slice interference signal style;

[0054] As an embodiment, in step S11, a radar electronic countermeasure simulation model is constructed, including constructing an airborne active phased array radar model and a cognitive jammer model; the airborne active phased array radar has multiple working modes such as AAS, HAS, CAM, STT, etc., and the N working modes are recorded as S = {s1, s2, L, s N According to the interference effect on the radar system, it can be divided into two categories: suppression interference and deception interference. The cognitive jammer model constructed contains noise modulation interference, smart noise interference, comb spectrum interference, Doppler interference, etc. The M types of interference patterns are recorded as A = {a1, a2, L, a M}; The signal styles included in the airborne active phased array radar model are shown in Table 1;

[0055] Table 1: Airborne active phased array radar signal pattern set

[0056]

[0057] Among them, STT is single target tracking, TAS is search and tracking, TWS is search and track, RWS-1 is search and range measurement with medium repetition rate, RWS-2 is search and range measurement with high repetition rate, and VS is speed search. The jamming style set constructed by the cognitive jammer is shown in Table 2.

[0058] Table 2: Interference signal pattern set

[0059]

[0060] S12: Obtain error measurement results according to the radar electronic countermeasure simulation model;

[0061] In an exemplary embodiment, the error measurement results include ranging error, speed measurement error, false alarm probability and detection probability;

[0062] It should be noted that the radar electronic countermeasure simulation model includes the radar signal processing simulation system;

[0063] As an embodiment, in step S12, according to the radar electronic countermeasure simulation model, the radar signal pattern and the interference signal pattern are input; according to the radar signal processing simulation system, corresponding indicators such as ranging error, speed measurement error, false alarm probability and detection probability are obtained;

[0064] Specifically, the simulation program simulates the generation of radar signals and interference signals. The radar signal processing system imports the radar and interference signals into the signal processing process. The radar signal processing system includes conventional pulse compression, Doppler processing, constant false alarm rate processing and other processes, and obtains the final error result; the generated radar signal and interference signal are placed into the radar signal processing system. Figure 3 The results of radar signal processing before jamming are shown. Figure 3 (a) is the PD treatment result; Figure 3 (b) is the CFAR processing result; Table 3 shows the radar error measurement results without interference;

[0065] Table 3 Error measurement results without interference

[0066]

[0067] The radar signal processing results after adding random false target interference are as follows: Figure 4 As shown, Figure 4 (a) is the PD treatment result; Figure 4 (b) is the CFAR processing result.

[0068] Table 4 shows the error measurement results;

[0069] Table 4 Error measurement results without interference

[0070]

[0071] From the above, we can see that the system measures correctly without interference. Then select the interference style as dense false target interference. The interface is shown above. It can be seen from the Doppler response diagram that multiple false targets are generated. After CFAR processing, multiple targets are still identified, indicating that the system is successfully interfered.

[0072] All interference pattern sets are put into the radar signal processing system for loop traversal to obtain measurement results;

[0073] S13: sorting the error measurement results using the TOPSIS sorting algorithm to obtain a sorting result;

[0074] In an exemplary embodiment, step S13 specifically includes:

[0075] S131: According to the error measurement results, a decision matrix is ​​obtained, such as the formula:

[0076]

[0077] Among them, K is the decision matrix, k nm are the parameter values ​​corresponding to the nth evaluation index in the mth reconnaissance result;

[0078] S132: Normalize the decision matrix to obtain a normalized matrix, such as the formula:

[0079]

[0080] Among them, k ij is the i-th row and j-th column element of the decision matrix; k′ ij is the element of the i-th row and j-th column of the normalized matrix;

[0081] S133: Weight the normalized matrix to obtain a weighted matrix, such as the formula:

[0082] z ij =w j k ij ′ (i=1,2,L,m; j=1,2,L,n) (3)

[0083] W=[w1,w2,L,w n ] T ,

[0084] Among them, z ij is the i-th row and j-th column element of the weighted matrix; w j is the jth element of the weight vector, W is the weight vector;

[0085] S134: According to the weighted matrix, the optimal reconnaissance result and the worst reconnaissance result are obtained, such as the formula:

[0086]

[0087] Among them, I + is the optimal reconnaissance result, is the jth element of the optimal reconnaissance node, I - The worst reconnaissance result. is the jth element of the worst reconnaissance result, and max{·} means taking the maximum value;

[0088] S135: Calculate the distance between each reconnaissance result and the best reconnaissance result and the worst reconnaissance result respectively, and obtain the first distance and the second distance, as shown in the formula:

[0089]

[0090] in, is the first distance, is the second distance;

[0091] S136: According to the first distance and the second distance, the quality of each reconnaissance result is obtained, as shown in the formula:

[0092]

[0093] Among them, A i is the quality of the i-th reconnaissance result;

[0094] S137: Obtaining a ranking result according to the quality of each reconnaissance result;

[0095] It should be noted that in step S13, the TOPSIS sorting algorithm is used to sort the error measurement results, that is, a decision analysis is performed on multiple evaluation indicators to approximate the optimal reconnaissance results under multiple reconnaissance results; it can be seen from formula (10) that A i The size of is between 0 and 1; the quality of the best reconnaissance result is 1, and the quality of the worst reconnaissance result is 0; therefore, A i The larger the value is, the closer the reconnaissance result is to the positive ideal solution; after the word, the quality of all the reconnaissance results obtained in step 136 is comprehensively analyzed to obtain the evaluation result of the interference effect; in this embodiment, after the radar system simulation experiment, Table 5 shows the TOPSIS sorting results of the error measurement results, and VS is set to the final state without TOPSIS sorting; the sorting results are shown in Table 5;

[0096] Table 5: Topsis ranking results

[0097]

[0098] S14: using the minimum interference-to-signal ratio data table to correct the ranking results, and obtain a corrected TOPSIS ranking table;

[0099] As an embodiment, in step S14, a minimum interference-to-signal ratio data table is obtained according to a real-installation radar interference experiment, and then the TOPSIS ranking result is corrected; wherein, the minimum interference-to-signal ratio test result obtained through the real-installation radar experiment is shown in Table 6;

[0100] Table 6: Minimum interference-to-signal ratio test results (unit: dB)

[0101]

[0102] In this embodiment, the TOPSIS results are filtered out by the interference-to-signal ratio test results, and the TOPSIS results of the interference patterns with an interference-to-signal ratio of more than 20 dB are removed, and are not considered in the subsequent optimal interference pattern selection; the TOPSIS results after the filtering are shown in Table 7;

[0103] Table 7: TOPSIS table after deletion

[0104]

[0105] S15: according to the modified TOPSIS ranking table, using the conversion rule to obtain a signal pattern state conversion table;

[0106] As an embodiment, in step S15, the modified TOPSIS ranking table is converted into a signal pattern state conversion table, and the conversion rules are as follows:

[0107] An effective interference pattern to the current signal pattern will cause the current signal pattern to change to a relatively ineffective signal pattern. When the interference signal is effective for both RWS-1 and RWS-2, the signal pattern will change to a VS signal, and VS will be set to the termination state. Ineffective interference to the current signal pattern will cause the current signal pattern to remain the same or change to a signal pattern with a higher threat level.

[0108] In this embodiment, the signal pattern state transition table is shown in Table 7;

[0109] Table 7 Signal pattern state transition table

[0110]

[0111] It should be noted that effective interference of the current pattern will change the current signal pattern state to a relatively invalid signal pattern. If there are multiple relatively invalid signal patterns, they will be randomly selected. For example, the signal pattern RWS-2 in the above table is transferred to the signal pattern RWS-2 or TAS by sample interference. The signal pattern transfer has a certain degree of randomness, which is more in line with the characteristics of phased array radar.

[0112] S2: According to the radar electronic countermeasure simulation model, the interference style selection model is obtained by using the signal style state transition table and the DDQN network;

[0113] In an exemplary embodiment, step S2 specifically includes:

[0114] S21: According to the radar electronic countermeasure simulation model, the adversarial sample is obtained using the signal pattern state transition table and the DDQN network;

[0115] In an exemplary embodiment, step S21 specifically includes:

[0116] S211: according to the radar electronic countermeasure simulation model, a state set and an action set are obtained; the state set includes a state, and the state is used to represent the signal pattern of the enemy's airborne active phased array fire control radar at a certain moment; the action set includes an action, and the action is used to represent the interference signal pattern selected by our jammer at that moment;

[0117] S212: according to the state set and the action set, using the signal pattern state transition table and the DDQN network, obtain the Q value corresponding to each interference signal pattern;

[0118] S213: Obtaining a first interference signal pattern using an ε-greed strategy according to the Q values ​​corresponding to each interference signal pattern;

[0119] S214: according to the first interference signal pattern, using the radar electronic countermeasure simulation model and the signal pattern state transition table to obtain a change in the enemy radar signal pattern, and obtaining a first feedback value according to the change in the enemy radar signal pattern;

[0120] S215: Update the Q value using the DDQN network according to the first reward value, and obtain the next state using the signal pattern state transition table;

[0121] S216: The state, the first interference signal pattern, the first reward value, and the next state are combined into a four-tuple sample to obtain an adversarial sample;

[0122] As an embodiment, in step S21, a radar electronic countermeasure simulation model is used to continuously counter the radar signal pattern and the interference pattern to generate a countermeasure sample;

[0123] It should be noted that the traditional Q learning algorithm has certain limitations when the state and action space is large; the Q learning algorithm initializes the Q value and then updates it according to the state transition; however, when the state and action space is very large, the traditional Q learning algorithm is difficult to learn and update effectively; the model of the DQN algorithm is as follows Figure 5 As shown in the figure, in the DQN algorithm, a neural network is used to approximate the Q value function. The input state is used as the input of the neural network, and the output is the Q value of each possible action. By training the neural network, the DQN algorithm can learn the Q value relationship between the state and the action. The introduction of the DQN algorithm enables better processing and learning when the state and action space are large. By using a neural network to approximate the Q value function, the DQN algorithm can handle high-dimensional inputs and can be trained end-to-end through the back propagation algorithm. In the DQN algorithm, the network parameter θ is updated by setting the number of steps to make the Q function approach the optimal Q value, as shown in the following formula (11):

[0124] Q(s,a;θ i )≈Q(s,a)(1)

[0125] In the formula, θ i is the neural network parameter at the i-th training step.

[0126] There are two neural networks in DQN. The two networks have the same structure but different parameters. The current network (MainNet) is used to calculate the value function Q(s, a; θ) of the current state-action pair. i ), the output of the target network (TargetNet) is used to calculate the Q estimate The mathematical expression for calculating the Q estimate is as follows:

[0127]

[0128] Where: Represents the parameters of TargetNet;

[0129] The loss function is as follows:

[0130]

[0131] The parameters of MainNet are updated in each training, while the parameters of TargetNet are assigned to the parameters of MainNet after a certain number of steps;

[0132] In DQN, the maximization operation when calculating the Q value may cause the estimated value function to be higher than its true value, resulting in non-uniform "overestimation", which will affect the final decision; as an offline learning algorithm, DQN does not use the true action of the next interaction at each learning, but uses the action currently considered to have the highest value to update the target value function; however, the true strategy does not always choose the action with the largest Q value in a given state, so directly using the maximum Q value as the target value often results in a target value higher than the true value;

[0133] In the DDQN algorithm, the selection and evaluation of actions are achieved through different value function networks. First, the action with the largest Q value in the next state s' is calculated based on the current network. Then, this action is used to calculate the Target Q value, and the target network completes the selection of the optimal action. The calculation method of the target Q value in DDQN is as follows:

[0134]

[0135] Substitute formula (14) into the loss function calculation formula, and we get:

[0136] Loss(θ)=E[(TargetQ-Q(s,a;θ i )) 2 ] (5)

[0137] In an exemplary embodiment, the principle of the DDQN algorithm is applied to the interference signal pattern decision, and the following variables are defined: the state (s∈S) represents the signal pattern of the enemy's airborne active phased array fire control radar at a certain moment, and the set (S) represents all possible enemy signal patterns; the action (a∈A) represents the interference signal pattern selected by our jammer at that moment, and the set (A) is the set of interference signal patterns available to us; when our reconnaissance equipment determines the state (s) through radar signal pattern recognition, it is input into the DDQN network; the network calculates the Q value corresponding to each interference signal pattern through neural network fitting; according to the strategy ε-g reed (this strategy is used to balance the relationship between exploration and utilization to maximize the cumulative return), select a jamming signal style (a) to apply to the enemy radar; after the jamming, we evaluate the effect according to the change of the enemy radar signal style and obtain a reward (r∈R); since we are in a non-cooperative relationship with the enemy, we can evaluate the effectiveness of the jamming by observing the change of the signal style after the jamming, and thus give the corresponding reward value (r); the jammer uses the reward value to update the Q value to determine the jamming style to be adopted in the next state (s'), and the above four variables are combined into a four-tuple sample (s, a, r, s') and stored in the sample pool (D);

[0138] S22: Build and update memory based on adversarial examples;

[0139] As an embodiment, in step S22, an adversarial memory library is constructed according to the sample pool (D), and adversarial samples are stored in the memory library. When the sample volume exceeds the capacity of the memory library, new samples will replace old samples.

[0140] S23: Use the memory library as training samples to train the DDQN network and obtain the interference style selection model;

[0141] As an embodiment, in step S23, DDQN randomly extracts samples from the memory bank as training samples, and continuously updates network parameters during the training process; after reaching the total number of iterations, a trained interference pattern selection model is obtained; specifically, in the sample pool (D), a certain number of samples are randomly extracted for neural network training to update the network parameters; then, the interference signal pattern selection process is repeated until the termination state is reached;

[0142] It should be noted that since different interference styles produce different interference effects, the return value is defined as follows:

[0143]

[0144] Where: JSR represents the minimum interference-to-signal ratio, TL→min represents that the signal pattern threat level of the airborne active phased array fire control radar will be the lowest; TL↓ represents that the signal pattern is converted in the direction of reducing the threat level; TL=TL represents that there is no conversion between signal patterns; TL↑ represents that the signal pattern is converted in the direction of increasing the threat level;

[0145] S3: Use the interference pattern selection model to make the optimal interference pattern decision.

[0146] In an exemplary embodiment, a simulation experiment is conducted using the above-mentioned active phased array radar cognitive jamming signal generation method. Figure 2 The figure shows the flow chart of the cognitive jamming signal generation algorithm of the active phased array radar. The details are as follows:

[0147] In this embodiment, different threat levels are set for different signal styles, which are used as the basis for interference effect evaluation, where STT is the highest level and is set to 4, TAS is 3, TWS is set to 2, RWS-1 and RWS-2 are set to 1, and VS is set to 1. The signal style conversion table is shown in 3.2.1; the algorithm training parameters in the simulation experiment are shown in Table 8;

[0148] Table 8: Training parameters

[0149] parameter Value Discount Factor 0.9 Learning Rate 0.005 Memory capacity 500 Training start threshold 200 Sample size 64 Network replication frequency 5 Greed Factor [0.1-0.95]

[0150] Before the parameters of the two neural networks are exchanged, 200 steps are observed first; after that, the parameters of the estimated network are assigned to the target network every 5 steps; in the greedy selection, the initial value is set to 0.1 and gradually increased at a rate of 0.0005 per step until it reaches a termination value of 0.95; 64 groups of samples are randomly drawn from the memory bank for network training each time, and the network parameters are continuously updated through the back propagation algorithm to improve the estimation accuracy of the Q value;

[0151] Figure 6 The following are the loss graphs obtained from the training of the DQN and DDQN algorithms. Since the ε exploration value is low at the initial moment, randomly selecting interference patterns may cause dramatic fluctuations in the loss value. However, as the training progresses, the ε exploration value gradually increases, and the neural network's fitting of the Q value becomes more accurate. After about 300 steps, the loss values ​​of the two algorithms drop to close to 0. Although the ε exploration value is increased to 0.95 at the highest, random selection of interference signal patterns for exploration will still occur in the later stages of training, so the loss values ​​of the two algorithms will also fluctuate slightly in the later stages of training.

[0152] Record the maximum Q value of a state output by the current network during the training process, such as Figure 7 As shown; Figure 7 The maximum Q-value convergence of the two algorithms at RWS-M is shown. The DQN algorithm converges to around 100 after 500 iterations, while the DDQN algorithm converges to around 100 after 400 iterations. By comparing the maximum Q-value convergence of the two algorithms, it can be seen that DDQN converges faster than DQN and eliminates the problem of overestimation of some Q values.

[0153] The ultimate goal of the reinforcement learning algorithm is to maximize the total target reward. Therefore, the reward value generated by each interference style selection during the training process is added to obtain the total reward value, such as Figure 8 As shown in the figure, after 100 steps of training, the total reward value obtained by the DDQN algorithm is higher than that of the DQN algorithm, which shows that the DDQN algorithm is better than the DQN algorithm in interference style selection, thus verifying the effectiveness of this method.

[0154] This embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned active phased array radar cognitive jamming signal generation method are implemented. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory (Flash Memory), a hard disk (HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above-mentioned types of memory.

[0155] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the above-mentioned active phased array radar cognitive interference signal generation method are implemented.

[0156] like Fig. 9 As shown, the computer device 120 may include: at least one processor 121, such as a central processing unit (CPU), at least one communication interface 123, a memory 124, and at least one communication bus 122. Among them, the communication bus 122 is used to realize the connection and communication between these components. Among them, the communication interface 123 may include a display screen (Display), a keyboard (Keyboard), and the optional communication interface 123 may also include a standard wired interface and a wireless interface. The memory 124 may be a high-speed random access memory (Random Access Memory, RAM), or a non-volatile memory (non-volatile memory), such as at least one disk storage. The memory 124 may also be at least one storage device located away from the aforementioned processor 121. Among them, the memory 124 stores an application program, and the processor 121 calls the program code stored in the memory 124 to perform any of the above method steps. Among them, the communication bus 122 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The communication bus 122 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig. 9Only one line is used to represent it, but it does not mean that there is only one bus or one type of bus. Among them, the memory 124 may include a volatile memory (volatile memory), such as a random-access memory (random-access memory, RAM); the memory may also include a non-volatile memory (non-volatile memory), such as a flash memory (flash memory), a hard disk drive (hard disk drive, HDD) or a solid-state drive (solid-state drive, SSD); the memory 124 may also include a combination of the above-mentioned types of memory. Among them, the processor 121 may be a central processing unit (central processing unit, CPU), a network processor (network processor, NP) or a combination of CPU and NP. Among them, the processor 121 may further include a hardware chip. The above-mentioned hardware chip may be an application-specific integrated circuit (application-specific integrated circuit, ASIC), a programmable logic device (programmable logic device, PLD) or a combination thereof. The above-mentioned PLD may be a complex programmable logic device (complex programmable logic device, CPLD), a field-programmable gate array (field-programmable gate array, FPGA), a generic array logic (generic array logic, GAL) or any combination thereof. Optionally, the memory 124 is also used to store program instructions. The processor 121 can call the program instructions to implement the active phased array radar cognitive jamming signal generation method as in this embodiment.

[0157] This embodiment provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned active phased array radar cognitive interference signal generation method.

[0158] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A method for generating cognitive jamming signals for active phased array radar, characterized in that: The following steps are involved: S1: Build a radar electronic countermeasure simulation model, use the TOPSIS sorting algorithm and the minimum interference-to-signal ratio data table to obtain the signal pattern state transition table; S2: According to the radar electronic countermeasure simulation model, using the signal pattern state transition table and the DDQN network, obtain an interference pattern selection model; S3: Utilize the interference pattern selection model to make an optimal interference pattern decision.

2. The method for generating cognitive jamming signals for active phased array radar according to claim 1, characterized in that: Step S1 specifically includes: S11: constructing a radar electronic countermeasure simulation model, wherein the radar electronic countermeasure simulation model includes an airborne active phased array radar model and a cognitive jammer model; S12: Obtaining an error measurement result according to the radar electronic countermeasure simulation model; S13: sorting the error measurement results using a TOPSIS sorting algorithm to obtain a sorting result; S14: using the minimum interference-to-signal ratio data table to correct the ranking result, to obtain a corrected TOPSIS ranking table; S15: According to the modified TOPSIS ranking table, a signal pattern state transition table is obtained using a conversion rule.

3. The method for generating cognitive jamming signals for active phased array radar according to claim 2, characterized in that: The airborne active phased array radar model includes an all-entry angle search AAS working mode, a high-entry angle search HAS working mode, a mixed-entry angle search CAM working mode, and a single-target tracking STT working mode; The airborne active phased array radar model also includes a radar signal pattern set, which includes a single target tracking STT radar signal pattern, a search and tracking TAS radar signal pattern, a search and track TWS radar signal pattern, a medium repetition frequency search and ranging RWS-1 radar signal pattern, a high repetition frequency search and ranging RWS-2 radar signal pattern, and a speed search VS radar signal pattern; The cognitive jammer model includes a set of interference signal patterns; the interference signal pattern set includes a comb spectrum interference signal pattern, a sample interference signal pattern, a Doppler noise interference signal pattern, a random false target interference signal pattern, a delayed superposition smart noise interference signal pattern, and a slice interference signal pattern.

4. The method for generating cognitive jamming signals for active phased array radar according to claim 2, characterized in that: The error measurement results include distance measurement error, speed measurement error, false alarm probability and detection probability.

5. The method for generating cognitive jamming signals for active phased array radar according to claim 2, characterized in that: Step S13 specifically includes: S131: According to the error measurement result, a decision matrix is ​​obtained, such as formula: Among them, K is the decision matrix, k nm are the parameter values ​​corresponding to the nth evaluation index in the mth reconnaissance result; S132: Normalize the decision matrix to obtain a normalized matrix, such as formula: Among them, k ij is the i-th row and j-th column element of the decision matrix; k′ ij is the element of the i-th row and j-th column of the normalized matrix; S133: Weighting the normalized matrix to obtain a weighted matrix, such as the formula: z ij =w j k ij ′(i=1,2,L,m;j=1,2,L,n), W=[w1,w2,L,w n ] T , Among them, z ij is the i-th row and j-th column element of the weighted matrix; w j is the jth element of the weight vector, W is the weight vector; S134: According to the weighted matrix, the best reconnaissance result and the worst reconnaissance result are obtained, such as the formula: Among them, I + is the optimal reconnaissance result, is the jth element of the optimal reconnaissance node, I - The worst reconnaissance result. is the jth element of the worst reconnaissance result, and max{·} means taking the maximum value; S135: Calculate the distance between each reconnaissance result and the best reconnaissance result and the worst reconnaissance result respectively, and obtain the first distance and the second distance, as shown in the formula: in, is the first distance, is the second distance; S136: According to the first distance and the second distance, the quality of each reconnaissance result is obtained, such as the formula: Among them, A i is the quality of the i-th reconnaissance result; S137: Obtaining a ranking result according to the quality of each reconnaissance result.

6. The method for generating cognitive jamming signals for active phased array radar according to claim 1, characterized in that: Step S2 specifically includes: S21: According to the radar electronic countermeasure simulation model, using the signal pattern state transition table and the DDQN network, obtain a countermeasure sample; S22: construct and update a memory base according to the adversarial sample; S23: Using the memory library as a training sample, the DDQN network is trained to obtain an interference style selection model.

7. The method for generating cognitive jamming signals for active phased array radar according to claim 6, characterized in that: Step S21 specifically includes: S211: obtaining a state set and an action set according to the radar electronic countermeasure simulation model; the state set includes states, and the states are used to represent the signal pattern of the enemy's airborne active phased array fire control radar at a certain moment; the action set includes actions, and the actions are used to represent the interference signal pattern selected by our jammer at the moment; S212: according to the state set and the action set, using the signal pattern state transition table and the DDQN network, obtaining a Q value corresponding to each interference signal pattern; S213: Obtaining a first interference signal pattern using an ε-greed strategy according to the Q values ​​corresponding to the respective interference signal patterns; S214: according to the first interference signal pattern, using the radar electronic countermeasure simulation model and the signal pattern state transition table to obtain a change in the enemy radar signal pattern, and obtaining a first feedback value according to the change in the enemy radar signal pattern; S215: according to the first reward value, using the DDQN network to update the Q value, and using the signal pattern state transition table to obtain the next state; S216: The state, the first interference signal pattern, the first reward value and the next state are combined into a four-tuple sample to obtain an adversarial sample.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for generating cognitive jamming signals of an active phased array radar are implemented as described in any one of claims 1 to 7.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for generating cognitive interference signals for active phased array radar are implemented as described in any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method for generating cognitive jamming signals for active phased array radar described in any one of claims 1 to 7 are implemented.

Images