Unmanned aerial vehicle cluster threat assessment method
By identifying the spatial and directional similarity of the drone swarm, combining membership and hierarchical analysis methods, using Bayesian network to process data, the problems of large amount of calculation and unreasonable selection of threat elements in the drone swarm threat assessment method are solved, and the problem of unreasonable selection of threat elements is achieved Efficient and reliable drone swarm threat assessment.
Patent Information
- Application Number
- CN202410972282.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-11-05
AI Technical Summary
The existing drone swarm threat assessment method is large in calculation and inconsistent with the actual situation when large-scale targets are attacked, the selection of threat factors is unreasonable, and the accuracy and credibility of the threat assessment algorithm using a single method are poor.
By determining the spatial similarity and directional similarity between drone individuals, threat factor indicators are analyzed using a combination of membership and hierarchy methods, and data is processed using Bayesian network to determine the quantitative and qualitative threats of drone clusters. and weighted fusion to obtain the comprehensive threat value.
It realizes efficient threat assessment of drone clusters, reduces the amount of calculation and improves real-time performance, and has high credibility in the threat assessment results, which can fully reflect the threat level of the group target. It is suitable for small-scale drone clusters and stand-alone independent combat scenarios .
Smart Images

Figure CN118898029A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of threat assessment of drone swarms, and in particular to a drone cluster threat assessment method. Background Art
[0002] With the rapid development of cutting-edge technologies such as drones, artificial intelligence, communication networks, and big data, drone swarms will become an important asymmetric means in the future, and anti-drone swarm-related technologies have become a research hotspot. Among them, threat assessment of enemy incoming drone swarms is a key link in situation judgment and program decision-making.
[0003] Threat assessment is a quantitative description of the threat level of enemy forces to us, and is a comprehensive measurement of the attribute values of enemy targets. Threat assessment generally includes three aspects: threat factor extraction, threat factor quantification, and threat level determination. Threat factor extraction refers to the extraction of indicators that have a greater impact on the threat level of the target, such as height, speed, etc. Threat level calculation is to use a certain algorithm to achieve a quantitative description of the target threat level, and finally make a qualitative or quantitative threat level determination.
[0004] In the field of threat assessment research on drone swarms, current research results have the following problems:
[0005] 1. The threat assessment is conducted on individual drone targets without considering the cluster combat characteristics of drone swarms. When large-scale targets attack, the calculation amount is large and does not conform to combat reality.
[0006] 2. Improper selection of threat factors, such as payload, stealth capability, and endurance. These threat factors cannot be obtained through commonly used radar, optoelectronic, radio and other detection equipment in actual battlefield environments. Therefore, the threat assessment method calculated based on this has no practical application value.
[0007] 3. In the selection of threat assessment algorithms, a single method is often used to process threat factors, and the accuracy and credibility of threat assessment results are poor. Summary of the invention
[0008] The purpose of the present invention is to provide a drone swarm threat assessment method to solve the technical problems existing in the prior art.
[0009] The present invention is achieved through the following technical solutions:
[0010] A drone swarm threat assessment method, comprising:
[0011] Determine the spatial similarity and directional similarity between individual drones, and determine a drone cluster based on the spatial similarity and directional similarity;
[0012] Obtain threat factor indicators corresponding to the drone cluster, and analyze the threat factor indicators by combining membership degree and hierarchical analysis method to determine the quantitative threat degree corresponding to the drone cluster;
[0013] Determine the drone cluster data, process the drone cluster data using a Bayesian network to obtain probabilities corresponding to different threat levels, and determine the qualitative threat degree corresponding to the drone cluster according to the probabilities corresponding to the different threat levels;
[0014] According to the quantitative threat degree and qualitative threat degree corresponding to the drone cluster, the weighted fusion method is used to determine the comprehensive threat value of the group target and complete the threat assessment of the drone cluster.
[0015] In a possible implementation, determining the spatial similarity and the directional similarity between individual drones, and determining a drone cluster based on the spatial similarity and the directional similarity includes:
[0016] Select an unvisited drone individual as a core point, and add the core point to the drone cluster;
[0017] Based on the core point, determine the spatial similarity and directional similarity between the core point and other individual drones that are not part of the drone cluster, and determine the overall similarity based on the spatial similarity and directional similarity;
[0018] Determine whether the overall similarity between other drone individuals that have not joined the drone cluster and the core point is greater than a preset similarity threshold C0. If so, determine the drone individual as a target reachable point and add the target reachable point to the drone cluster. Otherwise, do not operate on the drone individual.
[0019] Determine whether each individual drone in the drone cluster has been used as a core point. If so, the drone cluster is obtained. Otherwise, a new unvisited drone individual is selected from the drone cluster as the core point, and return to the step of determining the overall similarity.
[0020] In a possible implementation, the spatial similarity between the core point and other individual drones that are not part of the drone cluster is determined as:
[0021]
[0022] Among them, d(f i ,f j) represents the spatial similarity between the first drone individual and the second drone individual, the first drone individual represents the core point, the second drone individual represents other drone individuals that have not joined the drone cluster, d0 represents the preset distance threshold, k is the coefficient, exp represents the exponential function with the natural constant e as the base, dis(f i ,f j ) represents the spatial distance between the first drone individual and the second drone individual, (x i ,y i ,z i ) represents the spatial position data of the first drone individual, (x j ,y j ,z j ) represents the spatial position data of the second drone individual.
[0023] In a possible implementation, the direction similarity between the core point and other individual drones that are not part of the drone cluster is determined as:
[0024] h(f i ,f j )=cos(θ i -θ j )
[0025] Among them, h(f i ,f j ) represents the direction similarity between the first drone individual and the second drone individual, cos represents the cosine function, θ i Represents the heading angle data corresponding to the first drone individual, θ j Indicates the heading angle data corresponding to the second drone individual.
[0026] In a possible implementation, the overall similarity is determined according to the spatial similarity and the directional similarity as:
[0027] c(f i ,f j )=ω d ×d(f i ,f j )+ω h ×h(f i ,f j )
[0028] Among them, c(f i ,f j ) represents the overall similarity between the first drone individual and the second drone individual, ω d represents the first weighting coefficient, ω h represents the second weighting coefficient, d(f i ,f j) represents the spatial similarity between the first drone individual and the second drone individual, h(f i ,f j ) represents the direction similarity between the first drone individual and the second drone individual.
[0029] In a possible implementation, the threat factor index corresponding to the drone cluster is obtained, and the threat factor index is analyzed by combining the degree of membership and the analytic hierarchy process to determine the quantitative threat level corresponding to the drone cluster, including:
[0030] Obtain threat factor indicators corresponding to the drone cluster; wherein the threat factor indicators include average speed, average height, average distance, average route shortcut and number of individuals; the average speed, average height, average distance and average route shortcut are obtained by averaging the data corresponding to all drone individuals in the drone cluster;
[0031] According to the threat factor indicators corresponding to the drone cluster, the speed membership, altitude membership, distance membership, route shortcut membership and individual number membership are determined respectively;
[0032] The analytic hierarchy process is used to analyze the indicator weights corresponding to each threat factor indicator, and the quantitative threat degree corresponding to the drone cluster is determined based on the indicator weights, speed membership, altitude membership, distance membership, route shortcut membership and individual number membership.
[0033] In a possible implementation, determining the speed membership, the altitude membership, the distance membership, the route shortcut membership, and the individual number membership respectively includes:
[0034] Determine the speed membership as:
[0035]
[0036] Among them, F(v) represents the speed membership, e represents the natural constant, k1 represents the first coefficient, and v represents the average speed;
[0037] Determine the height membership as:
[0038]
[0039] Wherein, F(h) represents the height membership, k2 represents the second coefficient, h represents the average height, and h0 represents the preset height threshold;
[0040] Determine the distance membership as:
[0041]
[0042] Wherein, F(d) represents the distance membership, k3 represents the third coefficient, d represents the average distance, d1 represents the first preset distance threshold, d2 represents the second preset distance threshold, and d1 represents the distance <d2;
[0043] Determine the route shortcut membership as:
[0044]
[0045] Among them, F(p) represents the route shortcut membership, k4 represents the fourth coefficient, and p represents the average route shortcut;
[0046] Determine the individual quantity membership as:
[0047]
[0048] Among them, F(n) represents the individual number membership, k5 represents the fifth coefficient, and n represents the number of individuals.
[0049] In a possible implementation, according to the indicator weight, speed membership, altitude membership, distance membership, route shortcut membership and individual number membership, the quantitative threat level corresponding to the drone cluster is determined as follows:
[0050] F 定量 =ω1F(v)+ω2F(h)+ω3F(d)+ω4F(p)+ω5F(n)
[0051] Among them, F 定量 It represents the quantitative threat degree corresponding to the drone cluster, ω1 represents the speed index weight, ω2 represents the height index weight, ω3 represents the distance index weight, ω4 represents the route shortcut index weight, and ω5 represents the individual number index weight.
[0052] In a possible implementation, determining drone cluster data, processing the drone cluster data using a Bayesian network to obtain probabilities corresponding to different threat levels, and determining a qualitative threat degree corresponding to the drone cluster according to the probabilities corresponding to the different threat levels, including:
[0053] The target type, combat intent and combat capability are used as implicit nodes of the Bayesian network;
[0054] The average distance, average height, average speed, average route shortcut, average electromagnetic radiation, number of individuals in a group, and density of individuals in a group are used as evidence nodes, i.e., as input information of a Bayesian network; wherein the conditional probability of the Bayesian network is a preset value;
[0055] According to the Bayesian network, obtaining probabilities corresponding to different threat levels;
[0056] The probabilities corresponding to different threat levels are weighted and summed to obtain the qualitative threat level F corresponding to the drone cluster. 定性 .
[0057] In a possible implementation, according to the quantitative threat level and qualitative threat level corresponding to the drone cluster, a weighted fusion method is used to determine the group target comprehensive threat value:
[0058] F=ω×F 定量 +(1-ω)×F 定性
[0059] Among them, F represents the comprehensive threat value of the group target, ω represents the quantitative threat weight, and F 定量 Indicates the quantitative threat level, F 定性 Indicates the qualitative threat level.
[0060] The invention provides a method for assessing the threat of unmanned aerial vehicle swarms. The method identifies and divides multiple unmanned aerial vehicles into swarms and takes swarm targets as threat assessment objects, thereby reducing the amount of calculation and achieving higher real-time performance. The method does not require specifying the number of swarms and the scale of individuals in the swarm, and single drones can also form a swarm, thereby being compatible with target threat assessment of small-scale unmanned aerial vehicle swarms or even single drones in independent combat scenarios. The threat factors can comprehensively reflect the threat level of swarm targets, so that the obtained threat assessment results have high credibility. The threat factors are quantitatively processed by constructing a membership function to obtain a quantitative threat assessment value, and the threat factors are qualitatively inferred by constructing a Bayesian network to obtain a qualitative threat assessment value, and the two are merged to obtain a final threat level assessment value, so that the threat assessment has high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:
[0062] Figure 1 A flowchart of a drone swarm threat assessment method provided in an embodiment of the present invention.
[0063] Figure 2 An example diagram of a Bayesian network provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0064] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0065] like Figure 1 As shown, an embodiment of the present invention provides a method for assessing the threat of a drone cluster, which may include:
[0066] S1. UAV cluster target recognition.
[0067] According to the characteristics of unmanned swarm combat, drone swarms often achieve improved combat effectiveness through individual quantity advantages. There is consistency of action between individuals in the unmanned swarm. Therefore, the degree of similarity of drone individuals is used as the basis for judging the characteristics of the target group. The individual similarity is defined as the weighted fusion value of spatial similarity and directional similarity, and then the density clustering method is used for cluster division.
[0068] The present invention firstly performs target group feature recognition and takes each group of aircraft as the object of threat assessment, which conforms to the actual anti-unmanned cluster combat and optimizes the computing efficiency when large-scale targets attack.
[0069] S1.1. Calculation of spatial similarity between drones.
[0070] The relative spatial distance between two target drone individuals is defined as spatial similarity. Assume that drone individual f i The spatial position data is (x i ,y i ,z i ), then any two drone individuals f i and f j The spatial distance dis(f i ,f j ):
[0071]
[0072] Calculate the spatial similarity d(f i ,f j ):
[0073]
[0074] S1.2. Calculation of direction similarity between drones.
[0075] The difference in maneuvering direction between two target individuals is defined as direction similarity. Assume that the drone individual f i The heading data is θ i , then any two drone individuals f i and f j The directional similarity h(f i ,f j ):
[0076] h(f i ,f j )=cos(θi -θ j )
[0077] S1.3. Cluster division.
[0078] The density clustering method is used to divide the clusters and define two drone individuals f i and f j The similarity between i ,f j ) is its spatial similarity d(f i ,f j ) and direction similarity h(f i ,f j ) weighted fusion value, the weights are respectively d ,ω h :
[0079] c(f i ,f j )=ω d ×d(f i ,f j )+ω h ×h(f i ,f j )
[0080] When dividing clusters, specify the similarity threshold as C0. The detailed steps are as follows:
[0081] ① Select the core point: select an unvisited drone individual as the core point;
[0082] ②Mark the core point: Mark the core point and find all its reachable points. The reachable points are defined as other drone individuals within the specified similarity threshold C0.
[0083] ③ Form a new cluster: add these reachable points to the cluster where the core point is located to form a new cluster.
[0084] ④ Expand the cluster: Repeat steps 2 and 3 for each point in the newly formed cluster until the cluster can no longer be expanded.
[0085] ⑤ Repeat steps ① to ④ until all drone target individuals have been visited.
[0086] S2. Group target threat assessment.
[0087] The threat factor indicators selected for the quantitative threat degree calculation of group targets include: speed, height, distance, route shortcuts, and number of individuals. Since the spatial and motion characteristics of individual drones in the group targets are similar, the speed, height, distance, and route shortcuts can all be calculated based on the average value of the data of individual drones in the cluster.
[0088] Among them, distance refers to the distance between the UAV target and our key protection target. The smaller the distance, the higher the threat to the protection target. The so-called route shortcut refers to the vertical distance from the projection of the ray in the direction of movement of the UAV group on the water surface to our key protection target. The smaller the route shortcut, the closer the incoming UAV group flies to the protection target, and the higher the threat to the protection target.
[0089] The present invention selects speed, altitude, distance, route shortcut, number of individuals and electromagnetic radiation, which can be directly or indirectly obtained by radar, optoelectronics and radio detection equipment, as threat factor indicators for each cluster target, so as to solve the problem that the threat factor selection is not in line with reality while being able to fully reflect the threat of the target cluster.
[0090] S2.1. Calculation of quantitative threat value of group targets.
[0091] The quantitative threat assessment of group targets uses membership function to calculate the quantitative threat index, and then uses AHP method to calculate the quantitative threat value.
[0092] S2.1.1. Quantitative processing of quantitative threat indicators.
[0093] For each quantitative threat indicator, its membership function is defined as follows:
[0094] ① The speed membership function is as follows, where the parameter k1 = 5 × 10 -3 , speed in m / s:
[0095]
[0096] ② The height membership function is as follows, where the parameter k2 = 2 × 10 -8 , h0=100, height in m:
[0097]
[0098] ③ The distance membership function is as follows, where the parameter k3 = 5 × 10 -2 ,d1=30,d2=300,distance unit is km:
[0099]
[0100] ④ The route shortcut membership function is as follows, where the parameter k4 = 5 × 10 -3 , the unit of route shortcut is km:
[0101]
[0102] ⑤ The individual number membership function is as follows, where the parameter k5 = 5 × 10 -3 :
[0103]
[0104] S2.1.2. Quantitative threat calculation.
[0105] The AHP method (hierarchical analysis method) is used to determine the weight of each quantitative indicator. By comparing two indicators, a comparison matrix of speed, height, distance, route shortcuts and individual quantity indicators is constructed. The normalized eigenvector in the comparison matrix is calculated to obtain the weight of each indicator. An example of the comparison matrix is shown in Table 1:
[0106] Table 1 Comparison matrix of threat factor indicators
[0107] speed high distance Route shortcuts Number of individuals speed 1 2 1 / 2 4 1 / 2 high 1 / 2 1 1 / 3 3 1 / 4 distance 2 3 1 7 2 Route shortcuts 1 / 4 1 / 3 1 / 7 1 1 / 5 Number of individuals 2 4 1 / 2 5 1
[0108] Calculate the normalized eigenvector [ω1, ω2, ω3, ω4, ω5] of the above comparison matrix, and denote the weight of each indicator, and then obtain the quantitative threat degree calculation formula as follows:
[0109] F 定量 =ω1F(v)+ω2F(h)+ω3F(d)+ω4F(p)+ω5F(n)
[0110] S2.2. Calculation of qualitative threat value of group targets.
[0111] The qualitative threat degree calculation of group targets is based on the Bayesian network reasoning method. The selected indicators include: target type, combat intention, and combat capability, which are used as implicit nodes of the Bayesian network and are obtained through reasoning from the input information. At the same time, drone cluster data that can be obtained by various detection methods are selected as evidence nodes, representing the input information of the Bayesian network, namely distance, altitude, speed, route shortcuts, electromagnetic radiation, number of group individuals, and density of group individuals. These input node data are divided into qualitative descriptions according to needs. For example, altitude data greater than 7000 meters is treated as high altitude, 1000 to 7000 meters is treated as medium altitude, and less than 1000 meters is treated as low altitude.
[0112] like Figure 2 As shown, the embodiment of the present invention provides an example of constructing a Bayesian network to perform probability prediction on the threat level.
[0113] The node status values are shown in Table 2:
[0114] Table 2 Bayesian network node state value table
[0115]
[0116] The conditional probability table of the Bayesian network uses empirical values. According to Bayesian reasoning, the probabilities of high, medium, and low threat levels are P and P respectively. H , P M, P L , then the qualitative threat value can be calculated according to the following formula:
[0117] F 定性 =0.9×P H +0.5×P M +0.1×P L
[0118] S2.3. Calculation of comprehensive threat value of group targets.
[0119] The comprehensive threat value of the drone swarm target is calculated by weighted fusion of its quantitative threat degree and qualitative threat degree, as shown in the following formula:
[0120] F=ω×F 定量 +(1-ω)×F 定性
[0121] Based on the characteristics of drone swarm combat, the present invention first identifies and divides multiple drones into swarms. Under the current development trend of miniaturization and large-scale combat of drone swarms, when facing a large number of drone swarm targets, the present invention uses group targets as threat assessment objects, which has smaller calculation amount and higher real-time performance. At the same time, the method uses density clustering method to define the swarm according to density, without specifying the number of swarms and the individual scale in the swarm, and a single machine can also form a swarm. Therefore, the method is compatible with target threat assessment of small-scale drone swarms and even single-machine independent combat scenarios, and has universality.
[0122] The threat factors selected by the present invention include: distance, altitude, speed, route shortcuts, electromagnetic radiation, number of group individuals, and group individual density. These indicators cover the motion information, behavior information, and group structure information of the group targets. These threat factors can comprehensively reflect the threat level of the group targets, so that the obtained threat assessment results are highly credible. Moreover, these indicator data can be directly or indirectly obtained through commonly used radar, optoelectronics, radio and other detection equipment in the battlefield environment, and have practical application value.
[0123] The present invention combines quantitative processing of threat factors by constructing a membership function to obtain a quantitative threat assessment value, and qualitative reasoning of threat factors by constructing a Bayesian network to obtain a qualitative threat assessment value, and the two are integrated to obtain a final threat assessment value. This method combines quantitative assessment with qualitative assessment, reasonably reflects the subjective weight of factors, and gives full play to the advantages of Bayesian networks in the field of uncertainty and fuzzy indicator processing, and has high threat assessment accuracy.
[0124] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A drone swarm threat assessment method, characterized in that: include: Determine the spatial similarity and directional similarity between individual drones, and determine a drone cluster based on the spatial similarity and directional similarity; Obtain threat factor indicators corresponding to the drone cluster, and analyze the threat factor indicators by combining membership degree and hierarchical analysis method to determine the quantitative threat degree corresponding to the drone cluster; Determine the drone cluster data, process the drone cluster data using a Bayesian network to obtain probabilities corresponding to different threat levels, and determine the qualitative threat degree corresponding to the drone cluster according to the probabilities corresponding to the different threat levels; According to the quantitative threat degree and qualitative threat degree corresponding to the drone cluster, the weighted fusion method is used to determine the comprehensive threat value of the group target and complete the threat assessment of the drone cluster.
2. The drone swarm threat assessment method according to claim 1, characterized in that: Determining the spatial similarity and directional similarity between individual drones, and determining a drone cluster based on the spatial similarity and directional similarity, including: Select an unvisited drone individual as a core point, and add the core point to the drone cluster; Based on the core point, determine the spatial similarity and directional similarity between the core point and other individual drones that are not part of the drone cluster, and determine the overall similarity based on the spatial similarity and directional similarity; Determine whether the overall similarity between other drone individuals that have not joined the drone cluster and the core point is greater than a preset similarity threshold C0. If so, determine the drone individual as a target reachable point and add the target reachable point to the drone cluster. Otherwise, do not operate on the drone individual. Determine whether each individual drone in the drone cluster has been used as a core point. If so, the drone cluster is obtained. Otherwise, a new unvisited drone individual is selected from the drone cluster as the core point, and return to the step of determining the overall similarity.
3. The drone swarm threat assessment method according to claim 2, characterized in that: The spatial similarity between the core point and other drone individuals that are not part of the drone cluster is determined as: Among them, d(f i ,f j ) represents the spatial similarity between the first drone individual and the second drone individual, the first drone individual represents the core point, the second drone individual represents other drone individuals that have not joined the drone cluster, d0 represents the preset distance threshold, k is the coefficient, exp represents the exponential function with the natural constant e as the base, dis(f i ,f j ) represents the spatial distance between the first drone individual and the second drone individual, (x i ,y i ,z i ) represents the spatial position data of the first drone individual, (x j ,y j ,z j ) represents the spatial position data of the second drone individual.
4. The drone swarm threat assessment method according to claim 2, characterized in that: Determine the direction similarity between the core point and other drone individuals that are not part of the drone cluster: h(f i ,f j )=cos(θ i -θ j ) Among them, h(f i ,f j ) represents the direction similarity between the first drone individual and the second drone individual, cos represents the cosine function, θ i Represents the heading angle data corresponding to the first drone individual, θ j Indicates the heading angle data corresponding to the second drone individual.
5. The drone swarm threat assessment method according to claim 2, characterized in that: The overall similarity is determined based on the spatial similarity and the directional similarity: c(f i ,f j )=ω d ×d(f i ,f j )+ω h ×h(f i ,f j ) Among them, c(f i ,f j ) represents the overall similarity between the first drone individual and the second drone individual, ω d represents the first weighting coefficient, ω h represents the second weighting coefficient, d(f i ,f j ) represents the spatial similarity between the first drone individual and the second drone individual, h(f i ,f j ) represents the direction similarity between the first drone individual and the second drone individual.
6. The drone swarm threat assessment method according to claim 1, characterized in that: Obtain the threat factor indicators corresponding to the drone cluster, and analyze the threat factor indicators by combining membership degree and hierarchical analysis method to determine the quantitative threat degree corresponding to the drone cluster, including: Obtain threat factor indicators corresponding to the drone cluster; wherein the threat factor indicators include average speed, average height, average distance, average route shortcut and number of individuals; the average speed, average height, average distance and average route shortcut are obtained by averaging the data corresponding to all drone individuals in the drone cluster; According to the threat factor indicators corresponding to the drone cluster, the speed membership, altitude membership, distance membership, route shortcut membership and individual number membership are determined respectively; The analytic hierarchy process is used to analyze the indicator weights corresponding to each threat factor indicator, and the quantitative threat degree corresponding to the drone cluster is determined based on the indicator weights, speed membership, altitude membership, distance membership, route shortcut membership and individual number membership.
7. The drone swarm threat assessment method according to claim 6, characterized in that: Determine the speed membership, altitude membership, distance membership, route shortcut membership and individual number membership respectively, including: Determine the speed membership as: Among them, F(v) represents the speed membership, e represents the natural constant, k1 represents the first coefficient, and v represents the average speed; Determine the height membership as: Wherein, F(h) represents the height membership, k2 represents the second coefficient, h represents the average height, and h0 represents the preset height threshold; Determine the distance membership as: Wherein, F(d) represents the distance membership, k3 represents the third coefficient, d represents the average distance, d1 represents the first preset distance threshold, d2 represents the second preset distance threshold, and d1 represents the distance <d2; Determine the route shortcut membership as: Among them, F(p) represents the route shortcut membership, k4 represents the fourth coefficient, and p represents the average route shortcut; Determine the individual quantity membership as: Among them, F(n) represents the individual number membership, k5 represents the fifth coefficient, and n represents the number of individuals.
8. The drone swarm threat assessment method according to claim 6, characterized in that: According to the index weights, speed membership, altitude membership, distance membership, route shortcut membership and individual number membership, the quantitative threat level corresponding to the drone cluster is determined as: F 定量 =ω1F(v0+ω2F(h)+ω3F(d0+ω4F(p)+ω5Fωn) Among them, F 定量 It represents the quantitative threat degree corresponding to the drone cluster, ω1 represents the speed index weight, ω2 represents the height index weight, ω3 represents the distance index weight, ω4 represents the route shortcut index weight, and ω5 represents the individual number index weight.
9. The drone swarm threat assessment method according to claim 1, characterized in that: Determine the drone cluster data, process the drone cluster data using a Bayesian network, obtain the probabilities corresponding to different threat levels, and determine the qualitative threat degree corresponding to the drone cluster according to the probabilities corresponding to the different threat levels, including: The target type, combat intent and combat capability are used as implicit nodes of the Bayesian network; The average distance, average height, average speed, average route shortcut, average electromagnetic radiation, number of individuals in a group, and density of individuals in a group are used as evidence nodes, i.e., as input information of a Bayesian network; wherein the conditional probability of the Bayesian network is a preset value; According to the Bayesian network, obtaining probabilities corresponding to different threat levels; The probabilities corresponding to different threat levels are weighted and summed to obtain the qualitative threat level F corresponding to the drone cluster. 定性 .
10. The drone swarm threat assessment method according to claim 1, characterized in that: According to the quantitative threat level and qualitative threat level corresponding to the drone cluster, the weighted fusion method is used to determine the comprehensive threat value of the group target: F=ω×F 定量 +(1-ω)×F 定性 Among them, F represents the comprehensive threat value of the group target, ω represents the quantitative threat weight, and F 定量 Indicates the quantitative threat level, F 定性 Indicates the qualitative threat level.
Citation Information
Cited By
Unmanned aerial vehicle swarm detection method and system based on multi-layer early warning
CN121028067A
A multi-layer early warning-based unmanned aerial vehicle swarm detection method and system
CN121028067B
