An aircraft endgame system

Abstract

The application belongs to the technical field of aircraft terminal countermeasure system design, and particularly relates to an aircraft terminal countermeasure system, which simultaneously covers four links of observation, judgment, decision and execution in an aircraft countermeasure task, and an integrated architecture solves the problem of poor coupling of existing terminal countermeasure sensing and decision technology, shortens the response time of the countermeasure system to external air countermeasure situation changes, and compared with traditional terminal countermeasure technology, is realized based on a deep learning framework, has strong feature extraction and nonlinear mapping capability, and has good generalization capability after training due to a good neural network structure, and is more adaptable to the countermeasure system under complex air countermeasure situation conditions, and greatly improves the survival capability of the aircraft under complex situation conditions.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of aircraft terminal countermeasure system design, and particularly relates to an aircraft terminal countermeasure system. BACKGROUND

[0002] In modern air combat, the attack and defense confrontation between aircraft and attacking objects is becoming more and more intense, and the continuous improvement of the maneuverability and intelligent level of attacking objects poses a serious threat to the survival of aircraft. In order to improve the survivability of aircraft, it is necessary to enhance the terminal state perception, type identification and countermeasure strategy generation capabilities of the aircraft.

[0003] Terminal perception mainly relies on multi-sensor track fusion technology, which fuses the measurement data of attacking objects by various sensors to achieve accurate estimation of the state of attacking objects. The multi-sensor track fusion problem was first proposed by Singer in the 1970s, marking the beginning of the research on multi-sensor track fusion algorithm. Since the domestic track fusion technology started late, the performance and engineering of the track fusion algorithm are still in a catching-up position, especially the rapidity and accuracy of fusion required by the terminal countermeasure task are more stringent, and the existing technology is difficult to meet.

[0004] The attacking object type identification technology, as an important part of the terminal threat perception of the aircraft, the accuracy of its identification result will directly affect the success rate of the countermeasure task. The attacking object type identification technology makes full use of various properties of attacking objects obtained by various sensors and extracts feature information to infer the type of attacking objects. Foreign countries have been researching this technology for many years and have accumulated a lot of experience, but due to technical blockade, it is difficult to find complete and effective methods in public documents. Domestic scholars regard the attacking object type identification as a dynamic uncertainty reasoning process, and construct the credibility of evidence based on entropy gain description and integrate it into the Bayesian network reasoning process, and have achieved certain results, but under complex situation (interference, multiple targets), the recognition accuracy decreases seriously, which is difficult to meet the requirements of aircraft terminal countermeasure.

[0005] The aircraft air combat decision-making technology is a key link of the terminal countermeasure task, and has experienced five main stages of expert maneuver logic, automatic rule generation, rule evolution, machine learning and demonstration verification in the past few decades, from traditional expert system based on human experience to new paradigm characterized by machine intelligence self-evolution. At present, foreign countries have engineering experience of air combat decision-making systems such as AML, Paradyne, Alpha air combat. Domestic scholars focus more on theoretical research, with too many restrictions, and there is a big gap in engineering experience compared with foreign countries.

[0006] The emergence of omnidirectional attacking projectiles makes the conditions for projectile locking launch more easily met, and with the continuous improvement of the intelligence level and maneuverability of the projectiles, the existing perception, identification and decision-making technologies are fragmented, and it is increasingly difficult to meet the aircraft terminal countermeasure requirements in terms of perception speed, identification accuracy and maneuvering and decoy composite decision-making.

[0007] The present application is proposed in view of the above technical defects.

[0008] It should be noted that the disclosure of the above background art is only used to assist in understanding the inventive concept and technical solutions of the present application, and it does not necessarily belong to the prior art of the present patent application. In the absence of explicit evidence that the above content has been disclosed on the filing date of the present application, the above background art should not be used to evaluate the novelty and inventiveness of the present application. SUMMARY

[0009] The purpose of the present application is to provide an aircraft terminal countermeasure system to overcome or alleviate at least one aspect of the known technical defects.

[0010] The technical solution of the present application is:

[0011] An aircraft terminal countermeasure system, comprising a situation information input module, a multi-sensor attacking projectile target track fusion module, a multi-sensor attacking projectile type identification module, a terminal countermeasure decision module and a terminal countermeasure strategy output module;

[0012] The situation information input module comprises a data analysis module and a first network communication module, wherein the data analysis module analyzes airborne sensor information from the aircraft data bus, and analyzes aircraft state indication information; the first network communication module realizes communication between the situation information input module and the multi-sensor attacking projectile target track fusion module, sends the airborne sensor information analyzed by the data analysis module, and realizes communication between the situation information input module and the terminal countermeasure decision module, sends the aircraft state indication information analyzed by the data analysis module;

[0013] The multi-sensor striking object target track fusion module comprises an LSTM neural network module, a data alignment module and a second network communication module, wherein the LSTM neural network module takes the historical track features and current track information of the striking object detected by each sensor as input and takes the fusion striking object track information at a certain future time as output; the data alignment module uses the internal extrapolation and external interpolation method to unify the "time base" of the local track data of the striking object collected by different sensors and performs coordinate system transformation; the second network communication module realizes communication between the multi-sensor striking object track fusion module and the multi-sensor striking object type identification module, sends the fused striking object track information, realizes communication between the multi-sensor striking object track fusion module and the terminal countermeasure decision module, and sends the fused striking object track information;

[0014] The multi-sensor striking object type identification module comprises a GoogLeNet neural network module, a data preprocessing module and a third network communication module, wherein the GoogLeNet neural network module adopts a GoogLeNet neural network, performs convolution operation on the multi-sensor striking object measurement data through a plurality of convolution kernels with different sizes to form feature mapping, selects the maximum value in a small neighborhood for activation output after subsampling, can effectively reduce the dimension of the feature vector, and can effectively associate the striking object motion feature, the radar emitter signal feature and the striking object type after training; the data preprocessing module has data normalization and whitening functions, places the processed data near the mean value, and accelerates the training process of the GoogLeNet neural network; the third network communication module realizes communication between the multi-sensor striking object track fusion module and the multi-sensor striking object type identification module, receives the fused striking object track information, realizes communication between the multi-sensor striking object type identification module and the terminal countermeasure decision module, and sends the striking object type identification result;

[0015] The terminal countermeasure decision module comprises a countermeasure decision model module, a situation threat assessment module and a fourth network communication module. The countermeasure decision model module takes the fused strike object target track information, the strike object type identification result, the situation threat assessment result and the aircraft state indication information as inputs, performs inference through forward propagation of a policy network based on a deep reinforcement learning algorithm, and outputs a countermeasure strategy. The situation threat assessment module calculates the situation threat degree of the strike object relative to the aircraft according to the information of the movement of the strike object and the aircraft, considers the performance parameters of the strike object, and sends the situation threat assessment result to the countermeasure decision model module. The fourth network communication module realizes communication between the multi-sensor strike object target track fusion module and the terminal countermeasure decision module, receives the fused strike object target track information sent by the multi-sensor strike object target track fusion module, realizes communication between the multi-sensor strike object type identification module and the terminal countermeasure decision module, receives the strike object type identification result sent by the multi-sensor strike object type identification module, and realizes communication between the terminal countermeasure decision module and the terminal countermeasure strategy output module, and sends the countermeasure strategy.

[0016] The terminal countermeasure strategy output module comprises a display module, a control instruction calculation module and a fifth network communication module. The display module is used to display the aircraft evasion maneuver and the decoy bomb release point or the interceptor launch point position. The control instruction calculation module parses the countermeasure strategy into the control instruction of the aircraft. The fifth network communication module realizes communication between the terminal countermeasure decision module and the terminal countermeasure strategy output module, and receives the countermeasure strategy generated by the terminal countermeasure decision module.

[0017] According to at least one embodiment of the present application, in the aircraft terminal countermeasure system, the airborne sensor information comprises radar, DAS and ESM sensor information.

[0018] The aircraft state indication information comprises IAS / TAS, HSI and ADI state indication information.

[0019] According to at least one embodiment of the present application, in the aircraft terminal countermeasure system, the training process of the LSTM neural network module is that the loss is obtained by comparing the output of the LSTM neural network module with the actual strike object track information, the network parameters are constantly adjusted to minimize the loss through the SGD optimizer, and the mapping relationship between the multi-sensor strike object target historical track and future track feature data is established.

[0020] The present application has at least the following technical effects:

[0021] The application provides an aircraft terminal confrontation system, which covers observation, judgment, decision and execution of an aircraft confrontation task, and an integrated architecture solves the problem of poor coupling of existing terminal confrontation sensing and decision technology, shortens the response time of the confrontation system to external air confrontation situation changes, and has strong feature extraction and nonlinear mapping capacity based on a deep learning framework compared with traditional terminal confrontation technology, has good generalization capacity after training due to a good neural network structure, and is more adaptable to the confrontation system under a complex air confrontation situation, and greatly improves the survival ability of the aircraft under the complex situation. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 FIG. 1 is a schematic diagram of an aircraft terminal confrontation system provided by the application;

[0023] Figure 2 FIG. 1 is a schematic diagram of an aircraft terminal confrontation system provided by the application;

[0024] Figure 3 FIG. 1 is a schematic diagram of an aircraft terminal confrontation system provided by the application;

[0025] Figure 4 FIG. 1 is a schematic diagram of an aircraft terminal confrontation system provided by the application;

[0026] Figure 5 FIG. 1 is a schematic diagram of an aircraft terminal confrontation system provided by the application; DETAILED DESCRIPTION

[0027] In order to make the technical scheme of the application and its advantages clearer, the technical scheme of the application will be further clearly and completely described below with reference to the drawings. It should be understood that the specific embodiments described herein are only part of the embodiments of the application, and are only used to explain the application, but not to limit the application. It should be noted that, for the purpose of description, only parts related to the application are shown in the drawings, and other related parts can be referred to the general design. In the case of no conflict, the embodiments in the application and the technical features in the embodiments can be combined to obtain new embodiments.

[0028] In addition, unless otherwise defined, technical terms or scientific terms used in the description of the present application shall be understood as the common meaning understood by one of ordinary skill in the art to which the present application belongs. The words "upper", "lower", "left", "right", "center", "vertical", "horizontal", "inner", "outer" and the like used in the description of the present application only indicate relative directions or positional relationships, and are not intended to imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and the relative positional relationship may also change accordingly when the absolute position of the described object changes, therefore it cannot be understood as a limitation to the present application. The "first", "second", "third" and the like used in the description of the present application are only for the purpose of description, to distinguish different components, and cannot be understood as indicating or implying relative importance. The "one", "a" or "the" and the like used in the description of the present application should not be understood as an absolute limitation on the quantity, but should be understood as the presence of at least one. The "includes" or "contains" and the like used in the description of the present application means that the elements or objects appearing before the word are encompassed by the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0029] In addition, it should be further pointed out that, unless otherwise explicitly specified and limited, the "installation", "connection", "connection" and the like used in the description of the present application should be understood in a broad sense, for example, the connection can be fixed connection, or detachable connection, or integral connection; can be mechanical connection, or electrical connection; can be directly connected, or indirectly connected through intermediate medium, or the internal communication of two elements, one skilled in the art can understand its specific meaning in the present application according to the specific circumstances.

[0030] The following will be described in detail in combination with the accompanying drawings Figures 1 to 5 The present application will be further described in detail.

[0031] An aircraft terminal countermeasure system, as shown in Figure 1 It includes a situation information input module, a multi-sensor strike object target track fusion module, a multi-sensor strike object type identification module, a terminal countermeasure decision module and a terminal countermeasure strategy output module.

[0032] The situation information input module comprises a data analysis module and a first network communication module, wherein the data analysis module analyzes airborne sensor information including radar, DAS and ESM from an aircraft data bus, and analyzes aircraft state indication information including IAS / TAS, HSI and ADI; the first network communication module realizes communication between the situation information input module and the multi-sensor strike target track fusion module, transmits the airborne sensor information analyzed by the data analysis module, realizes communication between the situation information input module and the terminal countermeasure decision module, and transmits the aircraft state indication information analyzed by the data analysis module.

[0033] The multi-sensor strike target track fusion module comprises an LSTM neural network module, a data alignment module and a second network communication module, wherein the LSTM neural network module has strong data feature extraction capability and memory capability, the LSTM neural network module takes historical track features and current track information of a strike target detected by each sensor as model input, takes fusion strike track information at a future time as output, and trains the model by comparing the output of the LSTM neural network module with actual strike track information to obtain a loss, continuously adjusts network parameters by minimizing the loss through an SGD optimizer, and establishes a mapping relationship between historical track and future track feature data of a multi-sensor strike target; the data alignment module uses an internal extrapolation method to unify the "time base" of local track data of a strike target collected by different sensors and performs necessary coordinate system conversion, so as to reduce the influence of independent work of sensors, scanning period differences and different coordinate systems used by sensors on track fusion; the second network communication module realizes communication between the multi-sensor strike target track fusion module and the multi-sensor strike type identification module, transmits fusion strike target track information; and realizes communication between the multi-sensor strike target track fusion module and the terminal countermeasure decision module, and transmits fusion strike target track information.

[0034] The multi-sensor striking object type identification module comprises a GoogLeNet neural network module, a data preprocessing module and a third network communication module. The GoogLeNet neural network module adopts a GoogLeNet neural network, which is a kind of convolutional neural network. The multi-sensor striking object measurement data is subjected to convolution operation by a plurality of convolution kernels with different sizes to form feature mapping. After sub-sampling, the maximum value in a small neighborhood is selected to activate the output, thereby effectively reducing the dimension of the feature vector and having strong feature extraction capability. After proper training, the GoogLeNet neural network can effectively associate the striking object motion features, radar emitter signal features and striking object types. The data preprocessing module has data normalization and whitening functions, and places the processed data near the mean value to accelerate the training process of the GoogLeNet neural network. The third network communication module realizes communication between the multi-sensor striking object target track fusion module and the multi-sensor striking object type identification module, receives the fused striking object target track information, realizes communication between the multi-sensor striking object type identification module and the terminal countermeasure decision module, and sends the striking object type identification result.

[0035] The terminal countermeasure decision module comprises a countermeasure decision model module, a situation threat assessment module and a fourth network communication module. The countermeasure decision model module is based on a deep reinforcement learning algorithm and has continuous action decision capability in a large continuous state, which can cope with complex air countermeasure situations and does not require experts to write countermeasure rules. The countermeasure decision model module takes the fused striking object target track information, the striking object type identification result, the situation threat assessment result and the aircraft state indication information as input, performs forward propagation through a policy network based on the deep reinforcement learning algorithm to infer, and outputs a countermeasure strategy. The situation threat assessment module calculates the situation threat degree of the striking object relative to the aircraft according to the information of the motion of the striking object and the aircraft and considering the performance parameters of the striking object, and sends the situation threat assessment result to the countermeasure decision model module. The fourth network communication module realizes communication between the multi-sensor striking object target track fusion module and the terminal countermeasure decision module, receives the fused striking object target track information sent by the multi-sensor striking object target track fusion module, realizes communication between the multi-sensor striking object type identification module and the terminal countermeasure decision module, receives the striking object type identification result sent by the multi-sensor striking object type identification module, and realizes communication between the terminal countermeasure decision module and the terminal countermeasure strategy output module, and sends the countermeasure strategy.

[0036] The end confrontation strategy output module comprises a display module, a control instruction calculation module and a fifth network communication module, wherein the display module is used to display the carrier evasion maneuver and the position of the decoy bomb release point or the intercept object launch point; the control instruction calculation module parses the confrontation strategy into the control instructions of the aircraft, such as the throttle, rudder, etc.; and the fifth network communication module realizes the communication between the end confrontation decision module and the end confrontation strategy output module, and receives the confrontation strategy generated by the end confrontation decision module.

[0037] The specific working process of the aircraft end confrontation system is as shown in Figure 2

[0038] Step 1: The situation information input module is connected with the aircraft sensor bus

[0039] The situation information input module is connected with the aircraft sensor bus to initialize the data analysis module to correctly read and analyze the information of each airborne sensor and the carrier state indication information, and to initialize each network communication module to enable each network communication module to normally transmit data.

[0040] Step 2: Multi-sensor strike object target track fusion

[0041] The local track data of the strike object target obtained by different sensors is preprocessed through time alignment and coordinate system transformation, and then input into the trained LSTM neural network module to obtain the fused strike object target track information through forward inference, and the fused strike object target track information is sent to the multi-sensor strike object type identification module and the end confrontation decision module through the second network communication module.

[0042] Step 3: Multi-sensor strike object type identification

[0043] The multi-sensor strike object type identification module extracts the target motion features and radar radiation source signal features according to the fused strike object target track information, performs inference through forward propagation of the GoogLeNet neural network, outputs the results of strike object target type identification and confidence, and sends the strike object type identification results to the end confrontation decision module through the third network communication module.

[0044] Step 4: Evaluate the threat situation and generate confrontation decisions

[0045] ​The situation threat assessment module calculates the situation threat and the threat degree relative to the carrier according to the motion information of the striking object target and the carrier, considers the target performance parameters, obtains a situation threat assessment result, and sends the situation threat assessment result to the terminal countermeasure decision module through the fourth network communication module. The countermeasure decision model module based on the deep reinforcement learning algorithm formulates a composite countermeasure strategy formed by different terminal countermeasure means combinations according to the carrier state, the fused striking object target track information, the striking object type identification result, and the situation threat assessment result, and comprehensively configures the terminal countermeasure means.

[0046] Step 5: Display and execute the countermeasure strategy

[0047] The fifth network communication module sends the countermeasure strategy obtained by the terminal countermeasure decision module to the display module and the carrier control instruction calculation module, and the display module displays the carrier evasion maneuver and the position of the decoy bomb release point or the interceptor launch point.

[0048] In the aircraft terminal countermeasure system disclosed in the above embodiment, the situation information input module can be developed using an object-oriented method, wherein the data analysis module parses the data on the data bus according to the corresponding protocol, thereby completing the reading of information of each sensor.

[0049] In the aircraft terminal countermeasure system disclosed in the above embodiment, the multi-sensor striking object track fusion module can be developed using Python language based on the PyTorch framework, and the LSTM neural network module is deployed in ONNX format after training, so that the model can run independently of the PyTorch framework. The training process is as shown in Figure 3 .

[0050] In the aircraft terminal countermeasure system disclosed in the above embodiment, the multi-sensor striking object type identification module can be developed using Python language based on the PyTorch framework, and the GoogLeNet neural network module is deployed in ONNX format after training, so that the model can run independently of the PyTorch framework. The training process is as shown in Figure 4 .

[0051] In the aircraft terminal countermeasure system disclosed in the above embodiment, the terminal countermeasure decision module can realize countermeasure decision in a continuous action space based on the Actor-Critic framework. The terminal countermeasure decision model can include four neural networks, namely, a target evaluation network, a current evaluation network, a target policy network, and a current policy network. The training architecture of the terminal countermeasure decision model is as shown in Figure 5 .

[0052] In the aircraft terminal confrontation system disclosed in the above embodiment, the terminal confrontation strategy output module can be developed based on the Qt framework using an object-oriented method, the display module displays the evasive maneuvers of the host aircraft and the positions of the decoy bomb release points or the launch points of the interceptors, and the control instruction calculation module parses the confrontation strategy into control instructions of the aircraft, such as the throttle, the rudder, etc.

[0053] In the aircraft terminal confrontation system disclosed in the above embodiment, each network communication module can realize communication between the situation information input module, the multi-sensor strike object target track fusion module, the multi-sensor strike object type identification module, the terminal confrontation decision module, and the terminal confrontation strategy output module based on the UDP protocol.

[0054] In the aircraft terminal confrontation system disclosed in the above embodiment, under the current development level of strike technology, the use conditions of the strike objects are expanded, and it is increasingly easy to meet the launch conditions of the strike objects in air confrontation. In the background of omnidirectional attack strike objects canceling the launch direction restrictions of the strike objects, the confrontation strategy can be intelligently and efficiently generated to improve the survivability of the aircraft.

[0055] In the aircraft terminal confrontation system disclosed in the above embodiment, through the strategy output module, the confrontation strategy can be displayed on the display, and at the same time, the confrontation strategy can be parsed into control instructions that can be executed by the aircraft through the control instruction calculation module, which are automatically executed by the aircraft. The pilot can intuitively see the confrontation strategy being executed, which increases the pilot's trust in the terminal confrontation system and improves the automation level of the terminal confrontation under the premise of safety.

[0056] In the aircraft terminal confrontation system disclosed in the above embodiment, the terminal confrontation decision model is constructed based on a deep reinforcement learning algorithm, which greatly improves the adaptability of the terminal confrontation system to complex situation scenarios, and at the same time, does not require artificial writing of expert confrontation rules. When making decisions, only the current strategy network is needed to input the situation information, and the optimal confrontation strategy can be obtained through forward reasoning.

[0057] In the aircraft terminal confrontation system disclosed in the above embodiment, the four links of observation, judgment, decision, and execution in the terminal confrontation task are covered, and the integrated architecture reduces the communication time between modules and shortens the response time of the terminal confrontation system to external air confrontation situation changes.

[0058] In addition, those skilled in the art should also be able to realize that each module of the aircraft terminal confrontation system disclosed in the embodiments of the present application can be realized by electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the embodiments of the present application are generally described in terms of functions. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can choose different methods to implement the described functions for each specific application and its actual constraints, but such implementation should not be considered beyond the scope of the present application.

[0059] So far, the technical solution of the present application has been described in conjunction with the preferred embodiments shown in the drawings. Those skilled in the art should understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application. The technical solutions after such changes or replacements will all fall within the protection scope of the present application.

Claims

1. An aircraft endgame countermeasures system, characterized by, The system comprises a situation information input module, a multi-sensor striking object target track fusion module, a multi-sensor striking object type identification module, an endgame confrontation decision module and an endgame confrontation strategy output module. The situation information input module comprises a data analysis module and a first network communication module, wherein the data analysis module analyzes airborne sensor information and aircraft state indication information from an aircraft data bus; the first network communication module realizes communication between the situation information input module and the multi-sensor striking object target track fusion module, sends the analyzed airborne sensor information, and realizes communication between the situation information input module and the endgame confrontation decision module, sends the analyzed aircraft state indication information; The multi-sensor striking object target track fusion module comprises an LSTM neural network module, a data alignment module and a second network communication module, wherein the LSTM neural network module takes the historical track features and current track information of the striking object detected by each sensor as input, and takes the fusion striking object track information at a future time as output; the data alignment module uses the internal extrapolation and external interpolation method to unify the "time base" of the local track data of the striking object collected by different sensors and performs coordinate system transformation; the second network communication module realizes communication between the multi-sensor striking object target track fusion module and the multi-sensor striking object type identification module, sends the fused striking object track information, realizes communication between the multi-sensor striking object target track fusion module and the endgame confrontation decision module, and sends the fused striking object track information; The multi-sensor striking object type identification module comprises a GoogLeNet neural network module, a data preprocessing module and a third network communication module, wherein the GoogLeNet neural network module uses the GoogLeNet neural network, performs convolution operation on the multi-sensor striking object measurement data through a plurality of convolution kernels with different sizes to form feature mapping, selects the maximum value in a small neighborhood for activation output after subsampling, which can effectively reduce the dimension of the feature vector, and after training, can effectively associate the striking object motion features, radar emitter signal features and striking object types; the data preprocessing module has data normalization and whitening functions, places the processed data near the mean value, and accelerates the training process of the GoogLeNet neural network; the third network communication module realizes communication between the multi-sensor striking object target track fusion module and the multi-sensor striking object type identification module, receives the fused striking object track information, and realizes communication between the multi-sensor striking object type identification module and the endgame confrontation decision module, and sends the striking object type identification result; The terminal countermeasure decision module comprises a countermeasure decision model module, a situation threat assessment module and a fourth network communication module. The countermeasure decision model module takes the fused strike object track information, the strike type identification result, the situation threat assessment result and the aircraft state indication information as inputs, performs inference through forward propagation of a policy network based on a deep reinforcement learning algorithm, and outputs a countermeasure strategy. The situation threat assessment module calculates the situation threat degree of the strike object relative to the aircraft according to the information of the movement of the strike object and the aircraft, considers the performance parameters of the strike object, and sends the situation threat assessment result to the countermeasure decision model module. The fourth network communication module realizes communication between the multi-sensor strike object track fusion module and the terminal countermeasure decision module, receives the fused strike object track information sent by the multi-sensor strike object track fusion module, realizes communication between the multi-sensor strike type identification module and the terminal countermeasure decision module, receives the strike type identification result sent by the multi-sensor strike type identification module, and realizes communication between the terminal countermeasure decision module and the terminal countermeasure strategy output module, and sends the countermeasure strategy. The terminal countermeasure strategy output module comprises a display module, a control instruction calculation module and a fifth network communication module. The display module is used to display the aircraft evasive maneuver and the position of the decoy bomb release point or the interceptor launch point. The control instruction calculation module parses the countermeasure strategy into aircraft control instructions. The fifth network communication module realizes communication between the terminal countermeasure decision module and the terminal countermeasure strategy output module, and receives the countermeasure strategy generated by the terminal countermeasure decision module.

2. The aircraft terminal countermeasure system according to claim 1, wherein The airborne sensor information comprises radar, DAS and ESM sensor information. The aircraft state indication information comprises IAS / TAS, HSI and ADI state indication information.

3. The aircraft terminal countermeasure system according to claim 1, wherein The training process of the LSTM neural network module is to obtain a loss by comparing the output of the LSTM neural network module with the actual strike track information, to constantly adjust the network parameters by minimizing the loss through an SGD optimizer, and to establish a mapping relationship between the multi-sensor strike object historical track and future track feature data.

Images