Intelligent identification and emergency countering system for low-altitude aircraft

Abstract

A low-altitude aircraft intelligent identification and emergency countering system relates to the technical field of low-altitude aircraft safety management and comprises a detection unit, an identification unit, a command decision platform, a scheduling unit and a countering unit. The detection unit is used for detecting all low-altitude aircrafts in a prevention and control area and obtaining flight data of all the low-altitude aircrafts; the identification unit is used for receiving the flight data and analyzing and judging the threat level of the low-altitude aircraft based on a fusion algorithm; the command decision platform is used for receiving the threat level of each low-altitude aircraft and sending a disposal instruction to the countering unit; the scheduling unit is used for giving a specific countering equipment scheduling instruction through an intelligent optimization scheduling algorithm according to the disposal instruction; the countering unit controls different countering devices to accurately strike low-altitude aircrafts with corresponding threat levels according to the scheduling instruction; according to the method, the malicious unmanned aerial vehicle can be accurately identified and can be countered in time.

Description

Technical Field

[0001] The present invention relates to the technical field of low-altitude aircraft safety management, and in particular to a low-altitude aircraft intelligent identification and emergency countermeasure system. Background Art

[0002] At present, the commonly used detection methods for low-altitude aircraft such as drones mainly include radar, radio spectrum, photoelectricity and sound waves. These detection methods have their own characteristics and are applicable to different scenarios. These detection sensors all have a certain detection probability and false alarm probability. Using a single detection method will inevitably result in target omissions and false alarms. Target omissions will bring security loopholes to the control area, while false alarms will cause unnecessary psychological panic among on-duty personnel and may bring additional police handling costs.

[0003] In addition, after discovering and identifying malicious aircraft, appropriate countermeasures need to be taken as quickly as possible to respond to and eliminate threats, avoid or minimize possible losses in various aspects. The current low-altitude aircraft countermeasure technology is also increasingly mature, mainly including navigation suppression interference, navigation deception interference, aircraft command interference, projectile interception, capture net interception, high-power microwave damage, laser damage and other disposal methods. However, the equipment cost and strike cost-effectiveness of various low-altitude aircraft countermeasure equipment vary greatly, the importance of different points in the control area is not the same, and the deployment location and quantity configuration of different low-altitude aircraft countermeasure equipment will also be different.

[0004] The Chinese patent with publication number CN112598046B discloses a method for identifying target tactical intentions in multi-aircraft coordinated air combat. It can conduct research on multi-aircraft coordinated tactical reasoning and establish a target tactical intention recognition model in multi-aircraft coordinated air combat. Specifically, by constructing an SVM classifier that can handle different linearity and correlation features, the model performance of the traditional tactical intention recognition method is improved. In addition, combined with the parameters used in the traditional method, a variety of collaborative tactical feature extraction methods are proposed to extract higher-level features from the preliminary features. In addition, the decision tree model is trained through air combat simulation samples, and then all the above features are input into the trained model to effectively identify the collaborative tactical intention, thereby effectively meeting actual combat needs; However, this method is designed to identify hostile drones during wartime, and it will not work for malicious drones disguised as normal drones. Therefore, we propose a system that can accurately identify malicious drones and take timely countermeasures against them. Summary of the invention

[0005] The purpose of the present invention is to provide a low-altitude aircraft intelligent identification and emergency countermeasure system, which can accurately identify malicious drones and take timely countermeasures against them.

[0006] The present invention is achieved through the following technical solutions: A low-altitude aircraft intelligent identification and emergency countermeasure system, including a detection unit, an identification unit, a command decision platform, a dispatch unit and a countermeasure unit; The detection unit is used to detect all low-altitude aircraft in the control area and obtain flight data of all low-altitude aircraft; The identification unit is used to receive the flight data obtained by the detection unit, and analyze and determine the threat level of all low-altitude aircraft based on the fusion algorithm; The command decision platform is used to receive the threat level of each low-altitude aircraft, and send a disposal instruction to the countermeasure unit according to the importance of the control area and the threat level of the low-altitude aircraft; The dispatch unit is used to give specific dispatch instructions for countermeasure equipment according to the disposal instructions and the deployment information of the city-wide countermeasure equipment through an intelligent optimization dispatch algorithm; The countermeasure unit controls different countermeasure devices to carry out precise strikes on low-altitude aircraft of corresponding threat levels according to the dispatch instructions.

[0007] Furthermore, the detection unit includes radar detection equipment, radio spectrum detection equipment, photoelectric detection equipment and sound wave detection equipment, an Internet of Things network between the devices in the detection unit, and a detection equipment deployment information database.

[0008] Furthermore, the fusion algorithm includes a recurrent neural network model, an information extraction module, a Bayesian network reasoning model and a confidence weight calculation model.

[0009] Furthermore, the method for analyzing and determining the threat level of all low-altitude aircraft based on the fusion algorithm is specifically as follows: Screen and remove outliers from flight data to obtain initial data; After the initial data is calculated by the recurrent neural network model, the first threat level value of each low-altitude aircraft is obtained; The initial data is passed through the information extraction module to obtain high-level information about each low-altitude aircraft; After the high-order information is calculated by the Bayesian network inference model, the second threat level value of each low-altitude aircraft is obtained; After the first threat level value and the second threat level value are calculated by the confidence weight calculation model, the threat level of each low-altitude aircraft is determined.

[0010] Furthermore, the calculation steps of the confidence weight calculation model include: The calculation steps of the confidence weight calculation model include: The entropy method is used to calculate and obtain the weight value WA1 of the first threat level value and the weight value WA2 of the second threat level value respectively; The analytic hierarchy process is used to calculate and obtain the weight value WB1 of the first threat level value and the weight value WB2 of the second threat level value respectively; The confidence weight value W1 of the first threat level value and the confidence weight value W2 of the second threat level value are calculated using the following formula; ;

[0011] According to TL1, TL2, W1 and W2, the threat level value TL is calculated, and the threat level The calculation formula is: .

[0012] Furthermore, the specific calculation steps of the entropy method include: Performing non-negative processing on the first threat level value and the second threat level value; Normalizing the first threat level value and the second threat level value data after non-negative processing; Based on the normalized first threat level value and second threat level value, according to the entropy value formula: , calculate the entropy value corresponding to the first threat level value and the second threat level value, in, is the Boltzmann constant, Normalized first threat level value and second threat level value; Based on multiple entropy values, according to the weight calculation formula: ; Calculate the weight values ​​corresponding to the first threat level value and the second threat level value; is the value of the first threat level and the value of the second threat level. The weight value of the data.

[0013] Furthermore, the specific calculation steps of the analytic hierarchy process include: Construct a hierarchical structure including the goal layer, the criterion layer and the solution layer; Construct a judgment matrix based on the criteria layer; The judgment matrix is ​​calculated using the sum method to obtain the weight vector ; Calculate the relative importance of all factors at a certain level in the hierarchy to the target layer, and calculate the scores of each solution in the solution layer under different criteria in the criterion layer, build a score matrix for each criterion, and finally calculate the weight of each solution for each criterion based on each score matrix. ; according to and , calculate the weight value WB1 of the first threat level value and the weight value WB2 of the second threat level value.

[0014] Furthermore, the intelligent optimization scheduling algorithm adopts an ant colony algorithm with multi-factor balance.

[0015] Furthermore, the ant colony algorithm for multi-factor balance adds a crossover operator that determines the global convergence of the algorithm and a mutation operator that avoids local convergence of the algorithm and ensures the diversity of the solution set based on the ant colony algorithm.

[0016] Furthermore, the countermeasure unit includes suppression jammers, decoy jammers, command jammers, projectile interceptors, net interceptors, microwave destroyers and laser destroyers, an Internet of Things network between the devices in the countermeasure unit, and a countermeasure equipment deployment information database.

[0017] The technical solution of the present invention has at least the following advantages and beneficial effects: The present invention discloses a low-altitude aircraft intelligent identification and emergency countermeasure system, which can fuse and process target data detected by radar, optoelectronics, acoustics and spectrum sensors, solve the problem of target underreporting and false alarms easily generated by using a single or a single sensor, and improve the comprehensiveness and accuracy of urban low-altitude safety control; Using recurrent neural networks, we build a behavior model for malicious low-altitude aircraft, classify target data detected by radar, optoelectronics, acoustics, and spectrum sensors, and identify threat aircraft with malicious behavior trends from detected low-altitude aircraft, making urban low-altitude safety protection and law enforcement more accurate and efficient. Using Bayesian network reasoning, the threat level of malicious low-altitude aircraft is judged and ranked, providing data support for the selection of the optimal threat target disposal method and the determination of the disposal sequence, and laying the foundation for the realization of the urban low-altitude safety graded protection mechanism; Using the ant colony algorithm with multi-factor balanced optimization, various low-altitude aircraft countermeasure equipment and resources deployed in the prevention and control network are intelligently dispatched to achieve target destruction or threat elimination with the highest efficiency and lowest cost, minimizing the human and economic costs of urban low-altitude security protection; There are no special requirements or restrictions on the specific types and quantities of low-altitude aircraft detection equipment and low-altitude aircraft counter-equipment, and the system has good openness, compatibility and scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A structural schematic diagram of the present invention; Figure 2 Flow chart of the identification unit of the present invention; Figure 3 The target threat classification process of the recurrent neural network of the present invention; Figure 4 The recurrent neural network model of the present invention; Figure 5 The Bayesian network reasoning model of the present invention; Figure 6 An example of calculating weight vectors using the analytic hierarchy process of the present invention; Figure 7 Schematic diagram of the input and output of the scheduling unit of the present invention; Figure 8 Flowchart of the improved ant colony algorithm of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0022] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0023] Example 1 As attached Figure 1 A low-altitude aircraft intelligent identification and emergency countermeasure system shown includes a detection unit, an identification unit, a command decision platform, a dispatch unit and a countermeasure unit; The detection unit is used to detect all low-altitude aircraft in the control area and obtain flight data of all low-altitude aircraft; In particular, the detection unit includes radar detection equipment, radio spectrum detection equipment, photoelectric detection equipment and acoustic wave detection equipment, Internet of Things networking between the devices in the detection unit, and a detection equipment deployment information database; The detection unit mainly realizes real-time detection and tracking of various low-altitude aircraft flying in the control area, and uses various types of sensors to output the real-time three-dimensional position, speed, image, sound wave parameters (frequency / amplitude) and other information of each aircraft.

[0024] The identification unit is used to receive the flight data obtained by the detection unit, and analyze and determine the threat level of all low-altitude aircraft based on the fusion algorithm; The hardware of the identification unit is a set of rack-mounted high-performance servers, equipped with high-performance CPUs, multiple GPUs, large amounts of memory and fast storage, a network bandwidth of at least 10Gbps, a high-reliability power supply, and a strong cooling system. At the same time, software and system support such as data backup, cluster management, automated deployment, and status monitoring are considered. The overall process of the identification unit for identifying malicious aircraft is as follows: Figure 2 As shown. The threat level of aircraft is predefined as 0-9, with a total of 10 levels. Threat level 0 represents an aircraft with no malicious intent and performing tasks normally. The larger the number, the higher the threat level. The malicious aircraft identification process mainly includes three steps: recurrent neural network classification, Bayesian network reasoning, and comprehensive weight calculation. Both recurrent neural network classification and Bayesian network reasoning can derive a target threat level value. The comprehensive weight calculation step calculates different weights for the results obtained in the above two steps and derives the final threat level values ​​of each target. The aircraft data such as distance, altitude, and speed obtained by the detection equipment enter the malicious aircraft identification unit and first screen and eliminate abnormal values. The initial data after screening enters the recurrent neural network classification module and the feature matching and high-order information extraction module respectively. The initial data is classified by the recurrent neural network to obtain the target threat level value. After the initial data enters the high-order information extraction module, high-order information such as target type, whether the target carries weapons, the number of targets, the target's anti-damage capability, and the target's maneuverability are obtained. These high-order target information are then input into the Bayesian network reasoning to obtain the threat level value of each target; The identification unit analyzes the real-time status information of low-altitude aircraft to identify targets that are performing malicious acts or have a tendency to engage in malicious acts, and assesses their threat level based on information such as the type, size parameters, structural material, flight status, and type of weapons carried by the low-altitude aircraft.

[0025] The command decision platform is used to receive the threat level of each low-altitude aircraft, and send a disposal instruction to the countermeasure unit according to the importance of the control area and the threat level of the low-altitude aircraft; The command and decision-making platform mainly includes data receiving and processing workstations, network servers, global situation display projection and display equipment, and duty personnel operation desks. The data receiving and processing workstation is responsible for collecting real-time data such as alarm information, malicious aircraft parameter information, and video streams. The network server is used for internal and external communications, including telephone systems, intercom systems, wired / wireless network interconnection systems, etc. The global situation display projection and display equipment is the core display equipment of the command center, which is used to display key information such as maps, monitoring images, security situation of the control area, data reports, etc., to provide intuitive visual display for duty operators and help command decision-makers quickly obtain key information. The duty personnel operation desk is the operator's workstation for performing duty tasks, equipped with a touch screen, microphone and various control buttons for signal switching, information retrieval and command and dispatch.

[0026] The dispatch unit is used to give specific dispatch instructions for countermeasure equipment according to the disposal instructions and the deployment information of the city-wide countermeasure equipment through an intelligent optimization dispatch algorithm; The dispatch unit selects the most appropriate disposal means and equipment based on the threat level, location, speed and other status information of the target to be disposed, with the highest reliability, highest execution efficiency and minimum disposal cost as the optimization goal. The low-altitude aircraft countermeasure subsystem automatically selects the appropriate time, location and configuration of appropriate damage parameters based on the dispatch instructions received from the countermeasure equipment, maximizes the effectiveness of the countermeasure equipment and achieves the expected disposal goal of malicious low-altitude aircraft. The countermeasure unit controls different countermeasure devices to carry out precise strikes on low-altitude aircraft of corresponding threat levels according to the dispatch instructions.

[0027] In particular, the countermeasure unit includes suppression jammers, decoy jammers, command jammers, projectile interceptors, net interceptors, microwave destroyers and laser destroyers, an Internet of Things network between the devices in the countermeasure unit, and a countermeasure equipment deployment information database.

[0028] Example 2 The fusion algorithm includes a recurrent neural network model, an information extraction module, a Bayesian network reasoning model and a confidence weight calculation model.

[0029] In addition, the method for analyzing and determining the threat level of all low-altitude aircraft based on the fusion algorithm is specifically as follows: Screen and remove outliers from flight data to obtain initial data; After the initial data is calculated by the recurrent neural network model, the first threat level value TL1 of each low-altitude aircraft is obtained; The initial data is passed through the information extraction module to obtain high-level information about each low-altitude aircraft; After the high-order information is calculated by the Bayesian network inference model, the second threat level value TL2 of each low-altitude aircraft is obtained; After the first threat level value TL1 and the second threat level value TL2 are calculated by the confidence weight calculation model, the threat level TL of each low-altitude aircraft is determined.

[0030] Among them, the classification process of the recurrent neural network is as follows Figure 3 The working process is divided into two stages: model training and optimization and model practical application. In the model training and optimization stage, the neural network model is established and the parameters are determined. In the model practical application stage, the trained neural network model is used to classify and distinguish the input real-time target data. The neural network model designed by the present invention is shown in FIG. Figure 4 As shown in Figure 1, it is divided into input layer, hidden layer and output layer. The input layer consists of 9-order vectors The input data represents the aircraft's altitude, speed, distance, heading, acceleration, noise decibels, length, width, and height. The hidden layer consists of The present invention adopts a rectangular network structure. , that is, the hidden layer has 6 layers and 18 orders. The output layer consists of 10-order vectors Indicates, representing 0-9, a total of 10 levels of threat. The present invention uses a recurrent neural network with hidden states. The network has a "memory unit" that can store information from previous time steps and use it for calculations at the current time step. In this recurrent neural network, the same parameters are used at different time steps. When processing each time step of the sequence, the calculation method is consistent, only the input data is different. At each time step, the hidden state Based on the current input and the hidden state of the previous time step To calculate, the update formula is ,in , is the weight matrix, is the bias term, is the activation function. The output layer of the recurrent neural network outputs the first threat level value TL1.

[0031] The information extraction module is designed to process, predict and match the feature library of continuous initial data such as aircraft altitude, speed, distance, heading, acceleration, noise decibel number, length, width, height and image, and extract high-level target-related information such as target type, number of targets, target maneuverability, whether weapons are carried, type of weapons carried, possible attack points, and anti-damage capability.

[0032] Bayesian network inference model such as Figure 5 As shown, a Bayesian network is constructed with eight threat factors, namely, target type, number of targets, target mobility, whether to carry weapons, type of weapons carried, damage resistance, intended attack point, and distance from the attack point. The parameters of each node in the network are mainly obtained by summarizing expert experience. The present invention uses a clique tree propagation algorithm for reasoning, including constructing a clique tree structure, initializing the probability potential function of clique tree nodes, performing message transmission, and calculating the probability of target variables. Among them, the probability potential function is a non-negative function, which is related to the joint probability distribution of variables within the clique. For a clique containing variables and The maximum group of According to the Bayesian network and The conditional probability table is initialized, The inference result of the model is the second threat level value TL2.

[0033] In addition, the calculation steps of the confidence weight calculation model include: The entropy method is used to calculate the weight values ​​WA1 and WA2 of the first threat level value and the second threat level value respectively; The analytic hierarchy process is used to calculate the weight values ​​WB1 and WB2 of the first threat level value and the second threat level value respectively; The specific steps of calculating the weight using the entropy method are as follows: 1) Data preprocessing The data, namely the first threat level value and the second threat level value, are non-negatively processed to ensure that all data are non-negative. The threat levels output by the two models in the present invention are both within the range of 0-9. This step can be skipped if the processing flow is correct.

[0034] 2) Normalization The maximum and minimum index values ​​(threat levels) are counted, and the index value X, i.e., the non-negative first threat level value and the second threat level value, is normalized. In the present invention, it is hoped that the smaller the index value, the better, and the formula is used. .

[0035] 3) Calculate entropy Define the entropy value formula , is the Boltzmann constant, is the normalized index value (threat degree).

[0036] 4) Calculate weights The weight of each indicator is calculated based on the entropy value. The weight calculation formula is: ,in It is The weight values ​​of the indicators, j are 1 and 2 respectively, corresponding to the weight values ​​WA1 and WA2 of the first threat level value and the second threat level value in the present invention.

[0037] The specific steps of the hierarchical analysis method described in the present invention are as follows: 1) Build a hierarchy The decision problem (target threat level) in the present invention is decomposed into three levels: target layer, criterion layer, and solution layer. The target layer is at the highest layer, the criterion layer is at the middle layer, and the solution layer is at the bottom layer. T "two threat level selection weights, that is, the weight values ​​of the first threat level value and the second threat level value" is the target layer; R1 "the number of detectors that detect the target at the same time", R2 "whether a clear target image is obtained", R3 "whether the radar continuously detects the target point trace", R4 "whether there are multiple targets at a close distance" and R5 "whether the target speed exceeds the threshold" are the criterion layer; P1 (2 / 8), P2 (3 / 7), ..., P7 (8 / 2) are the solution layer, and the (a / b) in the brackets represents the weights of the two threat levels, such as P1 (2 / 8) means that the weights of the two threat levels are 0.2 and 0.8.

[0038] 2) Construct a judgment matrix Constructing a judgment matrix is ​​to compare each factor with each other and determine the weight of each criterion layer to the target layer.

[0039] For the criterion layer, we can construct a judgment matrix A:

[0040] Among them, m is the number of criteria levels, n is the number of scheme levels, To be the quasi-measurement element in the quasi-measurement layer, the elements in A satisfy the following conditions:

[0041]

[0042]

[0043] As shown in Table 1, an example of a judgment matrix constructed by implementing the present invention is shown.

[0044] 3) Calculate the weight vector The present invention adopts the sum method to calculate the weight, that is, firstly standardize each column of the judgment matrix, then sum the standardized elements by row, and finally standardize the sum result to obtain the weight vector , m is the number of criterion levels. Figure 6 As shown, an example of implementing this process on Table 1; Table 1 Example of AHP judgment matrix

[0045] 4) Total ranking of levels The total hierarchical ranking is to calculate the relative importance of all factors at a certain level to the target level, calculate the "score" of each plan under different criteria, and construct a score matrix for each criterion.

[0046] As shown in Table 2, an example of implementing this process is shown; Table 2 Example of AHP hierarchical analysis to obtain the total ranking score matrix of weighted levels

[0047] And according to each scoring matrix, calculate the weight of each scheme P1, P2, ..., P7 for each criterion R1, R2, ..., R5 .

[0048] 5) Synthetic weight Get the weight values ​​of each plan P1, P2, ..., P7 , is the weight of each criterion in the above step 4), is the weight vector obtained in the above step 3). According to these scheme weight values, the weight values ​​of the first threat level value and the second threat level value are calculated.

[0049] Assuming that the weight values ​​of the schemes P1, P2, ..., P7 are calculated as 0.1, 0.2, 0.1, 0.1, 0.1, 0.2, 0.2 respectively, the weight values ​​WB1 and WB2 of the first threat value and the second threat value are calculated as follows: WB1=0.2×0.1+0.3×0.2+0.4×0.1+0.5×0.1+0.6×0.1+0.7×0.2+0.8×0.2=0.53 WB2=0.8×0.1+0.7×0.2+0.6×0.1+0.5×0.1+0.4×0.1+0.3×0.2+0.2×0.2=0.47.

[0050] The confidence weight values ​​W1 and W2 of the first threat level value and the second threat level value are calculated using the following formula; ; ; According to TL1, TL2, W1 and W2, calculate the threat level value TL, then the threat level The calculation formula is: .

[0051] The final TL is the threat level from 0 to 9.

[0052] Example 3 The intelligent optimization scheduling algorithm adopts an ant colony algorithm for multi-factor balance.

[0053] In addition, the ant colony algorithm for multi-factor balance adds a crossover operator that determines the global convergence of the algorithm and a mutation operator that avoids local convergence of the algorithm and ensures the diversity of the solution set on the basis of the ant colony algorithm.

[0054] The hardware of the dispatch unit is a set of rack-mounted high-performance servers, and the hardware configuration requirements are the same as those of the identification unit. Figure 7 As shown. The scheduling unit takes the threat target disposal instructions (including the target information to be disposed) issued by the command decision platform and the system deployment countermeasure equipment information stored in the countermeasure equipment deployment information database as input, runs the multi-factor balanced ant colony algorithm module, and obtains the optimal threat target disposal plan. The countermeasure equipment intelligent scheduling unit extracts the scheduling instructions for each countermeasure equipment that performs the threat target disposal task from the threat target disposal plan and issues them immediately. After receiving the scheduling instructions, each countermeasure equipment performs the disposal task to eliminate the threat of malicious aircraft to the control point.

[0055] The ant colony algorithm module process of the multi-factor balance described in the present invention is as follows: Figure 8 As shown. Based on the traditional ant colony algorithm, a crossover operator that determines the global convergence of the algorithm and a mutation operator that avoids local convergence of the algorithm and ensures the diversity of the solution set are added to ensure the strong robustness and execution efficiency of the algorithm. , disposal costs and disposal time The objective function of the combination To optimize the target, the objective function designed by the present invention is expressed as , , , is a weight factor, and the values ​​in the present invention are 0.6, 0.3 and 0.1 respectively, that is, the disposal success rate is set as the primary optimization target, followed by the disposal cost, and the least is the disposal time. The present invention uses an improved ant colony optimization algorithm to solve the scheduling of each disposal device when the objective function value is minimum, that is, the above objective function is used as the adaptability function of the improved ant colony algorithm, and the path of ants walking is used to represent the scheduling strategy of each disposal device in the low-altitude aircraft countermeasure subsystem. All paths of the entire ant colony constitute the disposal plan of the low-altitude aircraft countermeasure subsystem for threatening targets.

[0056] The intelligent scheduling method for countermeasure equipment based on ant colony algorithm described in the present invention uses a directed graph to represent the solution space of the intelligent scheduling scheme of countermeasure equipment, and the countermeasure equipment scheduling scheme finally outputted is a set of paths in the directed graph. The nodes in the directed graph are represented by "task-equipment pairs", which means that a certain countermeasure equipment is used to complete a certain threat target disposal task. The path in the directed graph represents the order in which the ants (a search agent in the algorithm) move between the nodes, that is, the order of task allocation. For example, an ant starts from the "threat target 1 disposal task-countermeasure equipment 1" node and arrives at the "threat target 2 disposal task-countermeasure equipment 3" node. This path represents a task allocation method, that is, countermeasure equipment 1 performs the threat target 1 disposal task and countermeasure equipment 3 performs the threat target 2 disposal task. The path set embodies a series of continuous scheduling decisions. At each step, the ant decides to move to the next task-equipment pair based on the pheromone concentration and heuristic information. In the described threat target handling scenario with multiple tasks and multiple devices, the path taken by the ants during the search process may include allocation decisions of multiple tasks on different devices, and the quality of this path will be evaluated by the above objective function (maximizing the handling success rate, minimizing the handling cost, and minimizing the handling time).

[0057] The crossover operator described in the present invention is used to enhance the search ability of the algorithm, and to generate new and more potential solutions by combining excellent genes (good task-device allocation fragments) in different individuals (ant paths). It helps to jump out of the local optimal solution, accelerate the convergence speed of the algorithm, and enable the algorithm to find the optimal or approximately optimal device scheduling solution in a wider solution space. The present invention uses a sequential coding method to represent the path of the ant. Assuming that there are n tasks and m devices, a vector of length n is used to represent the path of the ant, and each element i in the vector represents the device number assigned to the i-th task. Assuming n=5, m=3, the encoding is [2,1,3,2,1] to indicate that task 1 is assigned to device 2, task 2 is assigned to device 1, task 3 is assigned to device 3, task 4 is assigned to device 2, and task 5 is assigned to device 1. The ant colony algorithm described in the present invention determines which ants (individuals) are selected from the ant colony for crossover operation according to the quality of the solution found by the ants. Individuals with better quality are preferred for crossover, so that high-quality offspring are more likely to be produced. The crossover operator of the present invention selects multiple crossover points, divides the parent individuals into multiple parts, and then exchanges these parts according to certain rules to generate offspring. For example, two crossover points k1=2 and k2=4 are selected, and for the parent individuals and, they are first divided into and, and then the middle parts are exchanged to obtain the offspring and. This method can more flexibly combine the good parts of the parent individuals, and can generate more diverse task-equipment allocation combinations in equipment scheduling, increasing the probability of finding a better scheduling solution.

[0058] The mutation operator of the present invention introduces new genes (new elements in the scheduling scheme) to avoid the algorithm from converging to the local optimal solution too early and increase the diversity of the population (ant colony). By performing a small probability mutation operation on individuals (scheduling schemes represented by ants), some under-discovered areas can be explored in the solution space, so that it is possible to find a better device scheduling strategy. The mutation probability is a key parameter that determines the frequency of mutation operations. If the mutation probability is too high, the algorithm will become similar to a random search and lose the advantages of the ant colony algorithm based on pheromone and heuristic information search; if the mutation probability is too low, it is difficult to effectively introduce new solutions and the role of the mutation operator cannot be fully utilized. The mutation probability of the present invention has a value range of 0.01-0.1. In the mutation operation of the present invention, a task is randomly selected and reallocated from the currently assigned device to another device. In the current device scheduling scheme, a threat target T1 is currently assigned to the countermeasure device M1 for disposal. Through the mutation operation, the threat target T1 is reallocated to the countermeasure device M2. Assuming that the processing time of threat target T1 in countermeasure device M1 is t1, and the processing time in countermeasure device M2 is t2, this variation may change the total processing time of the entire scheduling plan.

[0059] The fitness function described in this method is a function used to measure the quality of each ant in the search process. The fitness function is constructed based on the objective function, and the value of the objective function directly or indirectly determines the value of the fitness function. The present invention takes minimizing the objective function as the optimization goal and directly uses the objective function as the fitness function. This method is simple and intuitive, and can directly reflect the degree of closeness between the scheduling scheme and the optimal solution.

[0060] In the improved ant colony algorithm described in the present invention, ants release pheromones on the searched paths, but the pheromones previously left behind will also gradually evaporate over time. is the volatility factor of the pheromone. When the ant traverses all nodes, it updates the pheromone in the graph. The update formula is:

[0061]

[0062]

[0063] In the formula, represents the concentration of pheromone released by an ant k on a path (i, j), It represents the total concentration of pheromones released by the entire ant colony on a certain path, Q is the total amount of pheromones, is the sum of the lengths of the paths traveled by an ant k in this cycle.

[0064] In the improved ant colony algorithm described in the present invention, the probability calculation formula for an ant k to move from node i to node j at a certain time t during the search process is:

[0065] in, represents the total amount of pheromone on the path (i, j) at a certain time t, represents the expected degree of the ant's movement from node i to node j, The distance between node i and node j Inversely proportional, is the set of all nodes that ant k can reach at a certain time t.

[0066] In the improved ant colony algorithm described in the present invention, the state transition rule is used to search for paths, so that it is more inclined to choose those shorter paths with high pheromone concentrations. Assuming that an ant k is at node i at this time, the state transition rule defines the selection strategy for the next access node j. The state transition rule in the present invention is defined as:

[0067] In the formula, is a random number in the interval (0,1) generated by the program. are the configured algorithm parameters, and the other variable definitions are the same as above. When , the ants will select the next node to visit based on the shortest path and the highest pheromone concentration to develop a new area for search; on the contrary, when When , the optimal solution will be searched according to the probability calculation formula. The algorithm adjusts the parameters To configure the probability of using the two state transition rules, determine whether the algorithm continues to search near the current optimal path or jumps to a new area to search.

[0068] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A low-altitude aircraft intelligent identification and emergency countermeasure system, characterized by: It includes detection unit, identification unit, command decision platform, dispatch unit and countermeasure unit; The detection unit is used to detect all low-altitude aircraft in the control area and obtain flight data of all low-altitude aircraft; The identification unit is used to receive the flight data obtained by the detection unit, and analyze and determine the threat level of all low-altitude aircraft based on the fusion algorithm; The command decision platform is used to receive the threat level of each low-altitude aircraft, and send a disposal instruction to the countermeasure unit according to the importance of the control area and the threat level of the low-altitude aircraft; The dispatch unit is used to give specific dispatch instructions for countermeasure equipment according to the disposal instructions and the deployment information of the city-wide countermeasure equipment through an intelligent optimization dispatch algorithm; The countermeasure unit controls different countermeasure devices to carry out precise strikes on low-altitude aircraft of corresponding threat levels according to the dispatch instructions.

2. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 1 is characterized by: The detection unit includes radar detection equipment, radio spectrum detection equipment, photoelectric detection equipment and sound wave detection equipment, an Internet of Things network between the devices in the detection unit, and a detection equipment deployment information database.

3. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 1 is characterized by: The fusion algorithm includes a recurrent neural network model, an information extraction module, a Bayesian network reasoning model and a confidence weight calculation model.

4. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 3 is characterized by: The method for analyzing and determining the threat level of all low-altitude aircraft based on the fusion algorithm is specifically as follows: Screen and remove outliers from flight data to obtain initial data; After the initial data is calculated by the recurrent neural network model, the first threat level value of each low-altitude aircraft is obtained; The initial data is passed through the information extraction module to obtain high-level information about each low-altitude aircraft; After the high-order information is calculated by the Bayesian network inference model, the second threat level value of each low-altitude aircraft is obtained; After the first threat level value and the second threat level value are calculated by the confidence weight calculation model, the threat level of each low-altitude aircraft is determined.

5. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 4 is characterized by: The calculation steps of the confidence weight calculation model include: The entropy method is used to calculate and obtain the weight value WA1 of the first threat level value and the weight value WA2 of the second threat level value respectively; The analytic hierarchy process is used to calculate and obtain the weight value WB1 of the first threat level value and the weight value WB2 of the second threat level value respectively; The confidence weight value W1 of the first threat level value and the confidence weight value W2 of the second threat level value are calculated using the following formula; ； ； According to TL1, TL2, W1 and W2, the threat level value TL is calculated, and the threat level The calculation formula is: 。 6. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 5 is characterized by: The specific calculation steps of the entropy method include: Performing non-negative processing on the first threat level value and the second threat level value; Normalizing the first threat level value and the second threat level value data after non-negative processing; Based on the normalized first threat level value and second threat level value, according to the entropy value formula: ； Calculate the entropy value corresponding to the first threat level value and the second threat level value, where: is the Boltzmann constant, Normalized first threat level value and second threat level value; Based on multiple entropy values, according to the weight calculation formula: ； Calculate the weight values ​​corresponding to the first threat level value and the second threat level value; is the value of the first threat level and the value of the second threat level. The weight value of the data, is the value of the first threat level and the value of the second threat level. The entropy value of data is m, and m is the number of data within the first threat level value or the second threat level value.

7. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 5 is characterized by: The specific calculation steps of the hierarchical analysis method include: Construct a hierarchical structure including the goal layer, the criterion layer and the solution layer; Construct a judgment matrix based on the criteria layer; The judgment matrix is ​​calculated using the sum method to obtain the weight vector ; Calculate the relative importance of all factors at a certain level in the hierarchy to the target layer, and calculate the scores of each solution in the solution layer under different criteria in the criterion layer, build a score matrix for each criterion, and finally calculate the weight of each solution for each criterion based on each score matrix. ; according to and , calculate the weight value WB1 of the first threat level value and the weight value WB2 of the second threat level value.

8. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 1 is characterized by: The intelligent optimization scheduling algorithm adopts an ant colony algorithm for multi-factor balance.

9. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 8 is characterized by: The multi-factor balancing ant colony algorithm is based on the ant colony algorithm, and adds a crossover operator that determines the global convergence of the algorithm and a mutation operator that avoids local convergence of the algorithm and ensures the diversity of the solution set.

10. The low-altitude aircraft intelligent identification and emergency countermeasure system according to claim 1 is characterized by: The countermeasure unit includes suppression jammers, decoy jammers, command jammers, projectile interceptors, net interceptors, microwave destroyers and laser destroyers, an Internet of Things network between the devices in the countermeasure unit, and a countermeasure equipment deployment information database.

Images