Radar interference effect evaluation method, device and equipment based on small sample learning

Abstract

The invention relates to the technical field of radars, and provides a radar interference effect evaluation method, device and equipment based on small sample learning. And determining a plurality of different target type interference parameters corresponding to each preset type of radar waveform and a plurality of interference evaluation indexes in the plurality of preset type interference parameters. And determining a small sample task training set according to the plurality of preset type radar waveforms, the plurality of different target type interference parameters corresponding to each preset type radar waveform and the plurality of interference evaluation indexes. And training is carried out according to the small sample task training set and a Reptile meta-learning algorithm, and a radar interference effect evaluation model is obtained. The radar interference effect evaluation model is trained through the small sample task training set, so that the method can better adapt to scenes with scarce data or small samples during radar confrontation reconnaissance, the problems of weak generalization ability and poor robustness of the radar interference effect evaluation model are avoided, and a relatively accurate interference effect evaluation result is obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of radar, in particular to a radar jamming effect evaluation method, device and equipment based on small sample learning. BACKGROUND

[0002] In the field of radar countermeasure reconnaissance technology, radar jamming effect evaluation is crucial. Especially for active protection systems (APS), accurately predicting the impact of jamming methods on radar perception performance is of great significance for electronic countermeasure strategy formulation. In the prior art, Liu Songtao et al. published a review paper "Review of Jamming Effect Evaluation Technology" (Journal of CAEIT, 2020, 15(4):306-316) which discloses a classification method and research progress of jamming effect evaluation, especially focusing on wartime online jamming effect evaluation. Specifically, current jamming effect evaluation is mostly concentrated on offline evaluation in peacetime, mainly relying on theoretical derivation, simulation modeling and expert experience to analyze the effectiveness of jamming patterns. This kind of method is usually based on complete target information acquisition, stable tactical scene setting and sufficient sample data support, and has the problems of long evaluation process cycle and insufficient flexibility. The paper introduces the introduction of machine learning algorithms such as support vector machine and neural network for jamming effect evaluation in recent years, which improves the evaluation efficiency to a certain extent. However, in the prior art, due to the high dependence on a large amount of prior knowledge and high-quality sample input, when facing the situation of dynamic change of radar parameters, complex and variable jamming conditions and limited data acquisition in real battlefield, the model performance will be significantly reduced, and it is difficult to meet the actual demand of real-time and accurate evaluation of jamming effect. Especially in the condition of data scarcity or small sample, the existing technology has the problems of weak generalization ability and poor robustness, and cannot obtain accurate jamming effect evaluation results. SUMMARY

[0003] Therefore, it is necessary to provide a radar jamming effect evaluation method, device and equipment based on small sample learning to solve the above technical problems.

[0004] In a first aspect, the embodiments of the present application provide a radar jamming effect evaluation method based on small sample learning, comprising: For different preset type radar waveforms, determine a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms and a plurality of jamming evaluation indexes in a plurality of preset type jamming parameters, the jamming evaluation index is used to indicate the jamming performance of the target type jamming to the preset type radar signal; According to a plurality of preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and a plurality of jamming evaluation indexes, a small sample task training set is determined; According to the small sample task training set and a Reptile meta-learning algorithm, training is performed to obtain a radar jamming effect evaluation model.

[0005] In an embodiment, for different preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms and a plurality of jamming evaluation indexes are determined from a plurality of preset type jamming parameters, and the method further comprises: According to artificial experience, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms are obtained from a plurality of preset type jamming parameters; According to a preset evaluation formula, a jamming evaluation index of each of the preset type jamming on the preset type radar signal is calculated.

[0006] In an embodiment, the preset type jamming is a suppression type jamming or a deception type jamming, the preset evaluation formula is a power criterion evaluation formula or an information entropy evaluation formula, the suppression type jamming corresponds to the power criterion evaluation formula one by one, the deception type jamming corresponds to the information entropy evaluation formula one by one, and the calculation of the jamming evaluation index of each of the preset type jamming on the preset type radar signal according to the preset evaluation formula comprises: When it is determined that the preset type jamming is the suppression type jamming, the jamming evaluation index is calculated according to the power criterion evaluation formula; Or, when it is determined that the preset type jamming is the deception type jamming, the jamming evaluation index is calculated according to the information entropy evaluation formula.

[0007] In an embodiment, the small sample task training set comprises a plurality of small sample task training subsets, and the determination of the small sample task training set according to a plurality of preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and a plurality of jamming evaluation indexes comprises: According to each of the preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and a plurality of jamming evaluation indexes, a small sample task training subset corresponding to each of the preset type radar waveforms is determined; According to a plurality of small sample task training subsets, a small sample task training set is determined.

[0008] In an embodiment, the determination of the small sample task training subset corresponding to each of the preset type radar waveforms according to each of the preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and a plurality of jamming evaluation indexes comprises: determine a set of small sample task training samples for each preset type of radar waveform, a target type jamming parameter corresponding to each preset type of radar waveform, and a jamming evaluation index; According to the multiple sets of small sample task training samples corresponding to each preset type of radar waveform, determine a small sample task training subset corresponding to each preset type of radar waveform.

[0009] In one embodiment, after determining the small sample task training set according to the multiple small sample task training subsets, the method further comprises: one-hot encoding processing is performed on the small sample task training set.

[0010] In one embodiment, the set of small sample task training samples corresponds to one jamming task, and the training according to the small sample task training set and the Reptile meta-learning algorithm to obtain the radar jamming effect evaluation model comprises: constructing a radar jamming effect evaluation model for the radar jamming problem; performing gradient descent training according to the target task specific formula and the small sample task training set to obtain a target task specific parameter corresponding to each jamming task; based on multiple target task specific parameters, obtaining the trained radar jamming effect evaluation model.

[0011] In one embodiment, the radar jamming effect evaluation model can be defined by the following expression:

[0012] wherein, represents the task specific parameter corresponding to the initial radar jamming effect evaluation model, represents the meta-learning rate, and N represents the number of jamming tasks; The target task specific formula can be defined by the following expression:

[0013] wherein, represents the mean square error loss, represents the learning rate, represents the task specific parameter at the tth iteration.

[0014] In a second aspect, an embodiment of the present application provides a radar jamming effect evaluation device based on small sample learning, comprising: An obtaining module is configured to determine, for different preset types of radar waveforms, multiple different target type jamming parameters corresponding to each of the preset types of radar waveforms and multiple jamming evaluation indexes in multiple preset type jamming parameters, wherein the jamming evaluation index is used to indicate the jamming performance of the target type jamming on the preset type of radar signal. The training set determination module is configured to determine a small sample task training set according to a plurality of preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and a plurality of jamming evaluation indexes. The training module is configured to train the radar jamming effect evaluation model according to the small sample task training set and a Reptile meta-learning algorithm.

[0015] In a third aspect, an electronic device is provided, which includes a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the first aspect when executing the computer program.

[0016] Compared with the prior art, the technical scheme provided by the embodiments of the present application has the following advantages: The radar jamming effect evaluation method based on small sample learning provided by the embodiments of the present application determines, for different preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms from a plurality of preset type jamming parameters, and a plurality of jamming evaluation indexes for indicating the jamming performance of the target type jamming on the preset type radar signal. The small sample task training set is determined according to the plurality of preset type radar waveforms, the plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and the plurality of jamming evaluation indexes. The radar jamming effect evaluation model is obtained by training the radar jamming effect evaluation model according to the small sample task training set and the Reptile meta-learning algorithm. In this way, the radar jamming effect evaluation model can better adapt to the scene of data scarcity or small sample during radar counter reconnaissance by training the radar jamming effect evaluation model based on the small sample task training set, avoiding the problems of weak generalization ability and poor robustness of the radar jamming effect evaluation model, and thus obtaining a more accurate jamming effect evaluation result. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings, which are incorporated into and form a part of the specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the application.

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, for those skilled in the art, other drawings can also be obtained from these drawings without any creative effort.

[0019] Figure 1 A flowchart of a radar jamming effect evaluation method based on small sample learning provided by the embodiments of the present application is shown in the figure. Figure 2 A structural diagram of a radar jamming effect evaluation device based on small sample learning provided by the embodiments of the present application is shown in the figure. Detailed Implementation

[0020] To better understand the above-mentioned objectives, features, and advantages of the present invention, the solutions of the present invention will be further described below. It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other.

[0021] Many specific details are set forth in the following description in order to provide a full understanding of the invention, but the invention may also be practiced in other ways different from those described herein; obviously, the embodiments in the specification are only some embodiments of the invention, and not all embodiments.

[0022] In the field of radar countermeasures and reconnaissance technology, the assessment of radar jamming effects is crucial. Especially for Active Protection Systems (APS), accurately predicting the impact of jamming methods on radar sensing performance is of great significance for formulating electronic countermeasures strategies. In existing technologies, Liu Songtao et al., in their review paper "A Review of Electronic Countermeasures Jamming Effect Evaluation Technology" (Journal of CAEIT, 2020, 15(4):306-316), disclosed a classification method and research progress for jamming effect evaluation, with a particular focus on wartime online jamming effect evaluation. Specifically, current jamming effect evaluations mostly focus on peacetime offline evaluations, relying mainly on theoretical derivation, simulation modeling, and expert experience to analyze the effectiveness of jamming patterns. These methods are usually based on complete target information acquisition, stable tactical scenario settings, and sufficient sample data support, resulting in a long evaluation process and insufficient flexibility. The paper introduces machine learning algorithms such as support vector machines and neural networks for jamming effect evaluation in recent years, which has improved evaluation efficiency to some extent. However, existing technologies heavily rely on extensive prior knowledge and high-quality sample input. When faced with dynamic changes in radar parameters, complex and varied interference conditions, and limited data acquisition in real-world battlefields, model performance significantly deteriorates, making it difficult to meet the practical needs of real-time and accurate evaluation of interference effects. Especially under conditions of scarce data or small sample sizes, existing technologies suffer from weak generalization ability and poor robustness, failing to obtain accurate interference effect evaluation results.

[0023] Therefore, the present application provides a radar jamming effect evaluation method based on small sample learning. For different preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each preset type radar waveform and a plurality of jamming evaluation indexes indicating the jamming performance of the target type jamming on the preset type radar signal are determined among a plurality of preset type jamming parameters. According to the plurality of preset type radar waveforms, the plurality of different target type jamming parameters corresponding to each preset type radar waveform, and the plurality of jamming evaluation indexes, a small sample task training set is determined. The radar jamming effect evaluation model is obtained by training according to the small sample task training set and the Reptile meta-learning algorithm. In this way, by training the radar jamming effect evaluation model through the small sample task training set, the radar jamming effect evaluation model can better adapt to the scene of data scarcity or small sample in radar counter reconnaissance, avoid the problem of weak generalization ability and poor robustness of the radar jamming effect evaluation model, and thus obtain a more accurate jamming effect evaluation result.

[0024] In one embodiment, as shown in Figure 1 Figure 1 A flowchart of a radar jamming effect evaluation method based on small sample learning provided by the embodiment of the present application is shown, which specifically includes the following steps: S10: For different preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each preset type radar waveform and a plurality of jamming evaluation indexes are determined among a plurality of preset type jamming parameters.

[0025] The preset type radar waveform refers to the type of the opponent's radar signal detected in radar counter reconnaissance. The preset type radar waveform includes, but is not limited to, angle frequency modulation continuous wave, sawtooth frequency modulation continuous wave, sinusoidal frequency modulation continuous wave, frequency modulation interrupted continuous wave, single frequency continuous wave + up-sweep frequency, pseudo code phase modulation continuous wave, pseudo code phase modulation quasi continuous wave, pulse Doppler, double frequency continuous wave, multi-frequency continuous wave, triangular frequency modulation continuous wave (search), sawtooth frequency modulation continuous wave (search), and pseudo code phase modulation continuous wave (search), but the present application is not limited thereto, and the person skilled in the art can set it according to the actual situation.

[0026] The preset type jamming parameter includes preset type jamming and control parameters during jamming operation. The preset type jamming refers to the jamming made for different preset type radar waveforms. The preset type jamming includes suppression type jamming and deception type jamming. The suppression type jamming includes noise frequency modulation jamming, noise amplitude modulation jamming, and noise phase modulation jamming. The deception type jamming includes frequency shift jamming, trailing false target jamming, multi-false target jamming, and leading + trailing false target jamming. The control parameters, for example, the parameters for noise amplitude modulation jamming include jamming power, jamming frequency offset, and amplitude modulation speed factor, but the present application is not limited thereto, and the person skilled in the art can set it according to the actual situation.​

[0027] The interference evaluation index is used to indicate the interference performance of the target type jamming on the preset type radar signal, and the interference evaluation index includes a power criterion evaluation index and an information entropy evaluation index. The power criterion evaluation index can be determined by a power criterion evaluation formula, and the information entropy evaluation index can be determined by an information entropy evaluation formula.

[0028] Optionally, the power criterion evaluation formula can be defined by the following expression:

[0029] represents the minimum interference power of a receiver input end receiving interference when the active protection system is used for interference operation, represents the target echo signal power of an input end in the active protection system.

[0030] The information entropy evaluation formula can be defined by the following expression:

[0031] represents the probability of the monitored radar signal.

[0032] It should be noted that for the power criterion evaluation index, the greater the power criterion evaluation index, the better the radar jamming effect, and vice versa, the radar jamming effect is not good. For the information entropy evaluation index, the greater, the better the radar jamming effect, and vice versa, the radar jamming effect is not good.

[0033] Specifically, for different preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each preset type radar waveform and a plurality of interference evaluation indexes indicating the interference performance of the target type jamming on the preset type radar signal are determined in a plurality of preset type jamming parameters.

[0034] Optionally, based on the above embodiment, in some embodiments of the present application, one implementation of S10 can be: S101: According to artificial experience, a plurality of different target type jamming parameters corresponding to each preset type radar waveform are obtained in a plurality of preset type jamming parameters.

[0035] Specifically, for different preset type radar waveforms, a plurality of different target type jamming corresponding to the preset type radar waveform are determined in a plurality of preset type jamming according to artificial experience, and the corresponding control parameters are obtained, that is, a plurality of different target type jamming parameters corresponding to each preset type radar waveform are obtained.

[0036] ​​S102: calculate the interference evaluation index of each preset type of interference on the preset type of radar signal according to a preset evaluation formula.

[0037] Optionally, based on the above-mentioned embodiment, in some embodiments of the present application, when the preset type of interference is the suppression type of interference, one implementation of S102 can be: calculating the interference evaluation index according to the power criterion evaluation formula.

[0038] Optionally, based on the above-mentioned embodiment, in some embodiments of the present application, when the preset type of interference is the deception type of interference, one implementation of S102 can be: calculating the interference evaluation index according to the information entropy evaluation formula.

[0039] Specifically, for the currently monitored preset type of radar waveform, when it is determined that the corresponding preset type of interference is the suppression type of interference, the interference evaluation index is calculated according to the power criterion evaluation formula, or when it is determined that the corresponding preset type of interference is the deception type of interference, the interference evaluation index is calculated according to the information entropy evaluation formula.

[0040] S11: determining a small sample task training set according to the multiple preset types of radar waveforms, the multiple different target type interference parameters corresponding to each preset type of radar waveform, and the multiple interference evaluation indexes.

[0041] The small sample task training set refers to a data training set constructed for training a radar interference effect evaluation model. Unlike a traditional task training set, the small sample task training set has a much smaller amount of training data than the traditional task training set. Training the radar interference effect evaluation model based on the small sample task training set can better adapt to the scenario of data scarcity or small sample in radar counter reconnaissance, avoid the problems of weak generalization ability and poor robustness of the radar interference effect evaluation model, and thus obtain a more accurate interference effect evaluation result.

[0042] Specifically, after obtaining the different preset types of radar waveforms, the multiple different target type interference parameters corresponding to each different preset type of radar waveform, and the multiple interference evaluation indexes, the small sample task training set is determined according to the multiple preset types of radar waveforms, the multiple different target type interference parameters corresponding to each preset type of radar waveform, and the multiple interference evaluation indexes.

[0043] Optionally, based on the above-mentioned embodiment, in some embodiments of the present application, one implementation of S11 can be: S111: Determine a small sample task training subset corresponding to each preset type radar waveform according to each preset type radar waveform, a plurality of different target type jamming parameters corresponding to each preset type radar waveform, and a plurality of jamming evaluation indexes.

[0044] Specifically, for each preset type radar waveform, each preset type radar waveform, a plurality of different target type jamming parameters corresponding to each preset type radar waveform, and a jamming evaluation index corresponding to each different target type jamming parameter for the preset type radar waveform are determined to determine a small sample task training subset corresponding to each preset type radar waveform.

[0045] It should be noted that one small sample task training subset corresponds to each different preset type radar waveform.

[0046] For example, based on the above embodiment, for the current preset type radar waveform, there are 13 types, and 13 small sample task training subsets corresponding to the 13 preset type radar waveforms are constructed, that is, 13 small sample task training subsets are constructed. For each small sample task training subset corresponding to each preset type radar waveform, the preset type radar waveform, a plurality of target type jamming parameters for performing jamming operation on the preset type radar waveform determined according to artificial experience, and a plurality of target type jamming parameters corresponding to the preset type radar waveform are included.

[0047] Optionally, based on the above embodiment, in some embodiments of the present application, one implementation of S111 can be: S1111: Determine each preset type radar waveform, one target type jamming parameter corresponding to each preset type radar waveform, and a jamming evaluation index as a group of small sample task training samples.

[0048] Specifically, for each preset type radar waveform, a plurality of target type jamming parameters for performing jamming operation can be determined according to artificial experience, and a jamming evaluation index of a plurality of target type jamming parameters for performing jamming operation on the preset type radar waveform is calculated. Each preset type radar waveform, one target type jamming parameter corresponding to each preset type radar waveform, and a jamming evaluation index corresponding to the target type jamming parameter and the preset type radar waveform are taken as a group of small sample task training samples.

[0049] S1112: Determine a small sample task training subset corresponding to each preset type radar waveform according to a plurality of groups of small sample task training samples corresponding to each preset type radar waveform.

[0050] Specifically, after obtaining the multiple groups of small sample task training samples corresponding to each preset type of radar waveform, a small sample task training subset corresponding to each preset type of radar waveform can be constructed according to the multiple groups of small sample task training samples corresponding to each preset type of radar waveform.

[0051] S112: Determine a small sample task training set according to the multiple small sample task training subsets.

[0052] Specifically, after obtaining the small sample task training subsets corresponding to each different preset type of radar waveform, the multiple small sample task training subsets are combined to form a small sample task training set.

[0053] It should be noted that each small sample task training set can be represented as: , wherein, represents the preset type of radar waveform and the target type of interference parameter, represents an interference evaluation index.

[0054] It should be noted that, in order to be able to train to obtain the interference effect evaluation model according to the small sample task training set, the preset type of radar waveform and the target type of interference parameter are processed by one-hot encoding. One-hot encoding is an encoding technology for converting a classification variable into a binary vector, and its core principle is to independently encode N states through an N-bit state register, and each state corresponds to a binary vector with only one valid bit.

[0055] For example, based on the above embodiment, for 13 preset types of radar waveforms, when the current preset type of radar waveform is sawtooth frequency modulation continuous wave, the corresponding target type of interference parameter is the rear false target jamming, then the sawtooth frequency modulation continuous wave is set to 1, and the other 12 preset types of radar waveforms are 0, and the rear false target jamming is 1, and the others are 0. Further, the processed one-hot encoding is converted into a binary vector, such as: but not limited to this, the present application does not specifically limit it, and persons skilled in the art can set it according to the actual situation.

[0056] S12: Train to obtain a radar interference effect evaluation model according to the small sample task training set and the Reptile meta-learning algorithm.

[0057] The Reptile meta-learning algorithm is a meta-learning algorithm proposed by OpenAI, which realizes efficient small sample learning through first-order optimization and solves the calculation bottleneck of traditional meta-learning (such as MAML). Its core idea is to guide the model into a high-performance parameter region through parameter initialization, rather than directly optimizing the final parameters. It realizes the optimization of the parameter space through multi-task parameter averaging, avoiding the dependence on the second-order derivative.

[0058] Specifically, after obtaining the small sample task training set, the radar jamming effect evaluation model is obtained by training according to the small sample task training set and the Reptile meta-learning algorithm.

[0059] Optionally, based on the above embodiment, in some embodiments of the present application, one interference task corresponds to a group of small sample task training samples, based on which, one implementation of S12 can be: S121: constructing a radar jamming effect evaluation model for the radar jamming problem.

[0060] Specifically, for the radar jamming problem, the radar jamming effect evaluation model is constructed, Optionally, based on the above embodiment, in some embodiments of the present application, the radar jamming effect evaluation model can be defined by the following expression:

[0061] wherein, represents the task-specific parameter corresponding to the initial radar jamming effect evaluation model, represents the meta-learning rate, and N represents the number of interference tasks.

[0062] S122: performing gradient descent training according to the target task-specific formula and the small sample task training set to obtain the target task-specific parameter corresponding to each interference task.

[0063] wherein, the target task-specific formula is used to calculate the target task-specific parameter corresponding to each iteration in the process of iterative training based on the Reptile meta-learning algorithm and the gradient descent algorithm. The target task-specific formula can be defined by the following expression:

[0064] wherein, represents the mean square error loss, represents the learning rate, represents the task-specific parameter at the tth iteration. , represents the meta-update step.

[0065] S123: obtaining the trained radar jamming effect evaluation model based on the plurality of target task-specific parameters.

[0066] Specifically, according to the target task-specific formula and the small sample task training set, the gradient descent training is performed, and the target task-specific parameter corresponding to each interference task is obtained at the end of the gradient descent training. After obtaining the plurality of target task-specific parameters, the plurality of target task-specific parameters are substituted into the radar jamming effect evaluation model to obtain the trained radar jamming effect evaluation model.

[0067] Thus, the radar jamming effect evaluation method based on small sample learning provided by the embodiment determines, for different preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms and a plurality of jamming evaluation indexes for indicating the jamming performance of the target type jamming on the preset type radar signal in a plurality of preset type jamming parameters. The small sample task training set is determined according to the plurality of preset type radar waveforms, the plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and the plurality of jamming evaluation indexes. The radar jamming effect evaluation model is obtained by training according to the small sample task training set and the Reptile meta-learning algorithm. In this way, the radar jamming effect evaluation model is trained by using the small sample task training set, which can better adapt to the scene of data scarcity or small sample in radar counter-reconnaissance, avoid the problems of weak generalization ability and poor robustness of the radar jamming effect evaluation model, and thus obtain a more accurate jamming effect evaluation result.

[0068] It should be understood that, although Figure 1 the steps in the flowchart of the method are shown in order according to the arrows, these steps are not necessarily executed in order according to the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, Figure 1 at least part of the steps of the method can include a plurality of sub-steps or a plurality of stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be alternately or alternately executed with other steps or sub-steps or stages of other steps.

[0069] In one embodiment, as shown in Figure 2 , a radar jamming effect evaluation device based on small sample learning is provided, which includes an acquisition module 10, a training set determination module 11, and a training module 12.

[0070] The acquisition module 10 is configured to determine, for different preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms and a plurality of jamming evaluation indexes in a plurality of preset type jamming parameters, the jamming evaluation indexes being used to indicate the jamming performance of the target type jamming on the preset type radar signal. The training set determination module 11 is configured to determine a small sample task training set according to a plurality of preset type radar waveforms, a plurality of different target type jamming parameters corresponding to each of the preset type radar waveforms, and a plurality of jamming evaluation indexes. The training module 12 is configured to train according to the small sample task training set and the Reptile meta-learning algorithm, and obtain the radar jamming effect evaluation model.

[0071] In this way, the radar jamming effect evaluation model is trained by the small sample task training set, which can better adapt to the scene of data scarcity or small sample in radar counter-reconnaissance, avoid the problems of weak generalization ability and poor robustness of the radar jamming effect evaluation model, and thus obtain a more accurate jamming effect evaluation result.

[0072] The specific limitations of the radar jamming effect evaluation device based on small sample learning can be referred to the limitations of the radar jamming effect evaluation method based on small sample learning in the above, which will not be repeated here. Each module in the server can be realized by software, hardware and their combination in whole or in part. Each module can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory in the computer device in software form, so as to call and execute the operation corresponding to each module by the processor.

[0073] The embodiment of the present application provides an electronic device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the radar jamming effect evaluation method based on small sample learning provided by the embodiment of the present application can be realized, for example, the processor can realize the radar jamming effect evaluation method based on small sample learning provided by the embodiment of the present application when executing the computer program. Figure 1 The technical solutions of any of the method embodiments are similar in implementation principle and technical effects, which will not be repeated here.

[0074] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing relevant hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiment methods. Any reference to memory, database or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static random access memory (SRAM) and dynamic random access memory (DRAM).

[0075] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0076] The above embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be pointed out that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A radar jamming effect evaluation method based on few-sample learning, characterized in that, include: For different preset type radar waveforms, multiple different target type interference parameters and multiple interference evaluation indicators are determined from multiple preset type interference parameters for each preset type radar waveform. The interference evaluation indicators are used to indicate the interference performance of target type interference on preset type radar signals. Based on multiple preset types of radar waveforms, multiple different target type interference parameters corresponding to each preset type of radar waveform, and multiple interference evaluation indicators, a small sample task training set is determined. A radar jamming effect evaluation model is obtained by training a small sample task training set and the Reptile meta-learning algorithm.

2. The method according to claim 1, characterized in that, The method of determining multiple different target type interference parameters and multiple interference evaluation indicators corresponding to each preset type radar waveform from multiple preset type interference parameters also includes: Based on human experience, multiple different target type interference parameters corresponding to each preset type radar waveform are obtained from multiple preset type interference parameters. According to the preset evaluation formula, the interference evaluation index of each preset type of interference on the preset type of radar signal is calculated.

3. The method according to claim 2, characterized in that, The preset type of interference is either suppression-type interference or deception-type interference. The preset evaluation formula is either a power criterion evaluation formula or an information entropy evaluation formula. The suppression-type interference corresponds one-to-one with the power criterion evaluation formula, and the deception-type interference corresponds one-to-one with the information entropy evaluation formula. The step of calculating the interference evaluation index of each preset type of interference on the preset type of radar signal according to the preset evaluation formula includes: When the preset type of interference is determined to be a suppression type of interference, the interference evaluation index is calculated according to the power criterion evaluation formula. Alternatively, when the preset type of interference is determined to be deception type interference, the interference evaluation index is calculated according to the information entropy evaluation formula.

4. The method according to claim 1, characterized in that, The small sample task training set includes multiple small sample task training subsets. The process of determining the small sample task training set based on multiple preset radar waveform types, multiple different target type interference parameters corresponding to each preset radar waveform type, and multiple interference evaluation indicators includes: Based on each preset type of radar waveform, multiple different target type interference parameters corresponding to each preset type of radar waveform, and multiple interference evaluation indicators, a small sample task training subset corresponding to each preset type of radar waveform is determined. The few-shot task training set is determined based on multiple few-shot task training subsets.

5. The method according to claim 4, characterized in that, The step of determining a small sample task training subset corresponding to each preset type of radar waveform based on each preset type of radar waveform, multiple different target type interference parameters corresponding to each preset type of radar waveform, and multiple interference evaluation indicators includes: Each preset type of radar waveform, a target type interference parameter corresponding to each preset type of radar waveform, and interference evaluation index are determined as a set of small sample task training samples; Based on multiple sets of small sample task training samples corresponding to each preset type of radar waveform, determine the small sample task training subset corresponding to each preset type of radar waveform.

6. The method according to claim 5, characterized in that, After determining the few-shot task training set based on multiple few-shot task training subsets, the process further includes: The small sample task training set is processed by one-hot encoding.

7. The method according to claim 5, characterized in that, The set of small sample training samples corresponds to one interference task. The step of training based on the small sample training set and the Reptile meta-learning algorithm to obtain a radar interference effect evaluation model includes: To address the radar jamming problem, a radar jamming effect evaluation model is constructed. Gradient descent training is performed based on the target task-specific formula and a small sample task training set to obtain the target task-specific parameters corresponding to each of the interference tasks. Based on specific parameters of multiple target tasks, a trained radar jamming effect evaluation model is obtained.

8. The method according to claim 7, characterized in that, The radar jamming effect evaluation model can be defined by the following expression: in, This represents the mission-specific parameters corresponding to the initial radar jamming effect evaluation model. represents the meta-learning rate, and N represents the number of perturbation tasks; The target task-specific formula can be defined by the following expression: in, This represents the mean squared error loss. Indicates the learning rate. This represents the task-specific parameter at the t-th iteration.

9. A radar jamming effect evaluation device based on few-sample learning, characterized in that, include: The acquisition module is used to determine, for different preset type radar waveforms, multiple different target type interference parameters and multiple interference evaluation indicators corresponding to each preset type radar waveform from multiple preset type interference parameters. The interference evaluation indicators are used to indicate the interference performance of the target type interference on the preset type radar signal. The training set determination module is used to determine a small sample task training set based on multiple preset types of radar waveforms, multiple different target type interference parameters corresponding to each preset type of radar waveform, and multiple interference evaluation indicators. The training module is used to train a radar jamming effect evaluation model based on a small sample task training set and the Reptile meta-learning algorithm.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the radar jamming effect evaluation method based on few-shot learning as described in any one of claims 1 to 8.

Images