CN117787625A - Task allocation method in manned-unmanned aerial vehicle formation scene - Google Patents

Task allocation method in manned-unmanned aerial vehicle formation scene Download PDF

Info

Publication number
CN117787625A
CN117787625A CN202311827315.9A CN202311827315A CN117787625A CN 117787625 A CN117787625 A CN 117787625A CN 202311827315 A CN202311827315 A CN 202311827315A CN 117787625 A CN117787625 A CN 117787625A
Authority
CN
China
Prior art keywords
unmanned aerial
radius
threat
target
aerial vehicle
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311827315.9A
Other languages
Chinese (zh)
Inventor
李晖
于彦泽
刘卓
戴睿
吴洵
孟多南
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Avic Research Institute Yangzhou Science And Technology Innovation Center
Original Assignee
Avic Research Institute Yangzhou Science And Technology Innovation Center
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Avic Research Institute Yangzhou Science And Technology Innovation Center filed Critical Avic Research Institute Yangzhou Science And Technology Innovation Center
Priority to CN202311827315.9A priority Critical patent/CN117787625A/en
Publication of CN117787625A publication Critical patent/CN117787625A/en
Pending legal-status Critical Current

Links

Landscapes

  • Aiming, Guidance, Guns With A Light Source, Armor, Camouflage, And Targets (AREA)

Abstract

The invention relates to a task allocation method in a manned-unmanned aerial vehicle formation scene, which comprises the following steps: receiving data acquired by a sensor, and constructing a random variable and a derivative variable based on the acquired data; based on the random variable and the derivative variable, calculating threat probability of the opponent target to the my helicopter by adopting a dynamic Bayesian network; and carrying out task allocation by adopting an improved Markov decision according to the threat probability. The invention carries out threat assessment based on a dynamic Bayesian network, adopts a Markov decision process to carry out task allocation, can complete danger assessment under a given scene and generate a UAV task list to assist the decision of a pilot, makes a decision similar to the decision of the pilot, reduces the threat degree of an enemy target and improves the viability of a helicopter.

Description

Task allocation method in manned-unmanned aerial vehicle formation scene
Technical Field
The invention relates to the technical field of unmanned aerial vehicle combat, in particular to a task allocation method in a manned-unmanned aerial vehicle formation scene.
Background
Generally, unmanned aerial vehicles (Unmanned Aerial Vehicle, UAV) require remote control by ground personnel, and space splitting of UAV and ground personnel cannot fully exploit the advantages of UAV due to communication and collaboration problems. In order to solve the problem, in combat exercises, strategies for controlling UAVs by a manned flight platform (such as a helicopter) are proposed, namely, the manned flight platform (such as the helicopter) is cooperatively matched with an unmanned aerial vehicle, so that ground command and navigation can be effectively avoided. Based on situation awareness, the manned platform performs threat assessment and situation prediction, and distributes certain tasks to UAVs for execution. In a Manned-and-Unmanned-Team (M & UMT), the effectiveness of a task is a first issue to be considered, and during task execution, helicopter drivers often decide when, where and how to release UAVs based on the current real-time status. Thus, the main influencing factor of this decision-making process is the risk assessment of the current helicopter state. But pilots may not be able to release UAVs in a timely manner due to their heavy tasks or lack of efficient situational awareness.
Disclosure of Invention
The invention aims to provide a task allocation method in a manned-unmanned aerial vehicle formation scene, which is used for improving timeliness of task allocation.
In order to achieve the above object, the present invention provides the following technical solutions:
a task allocation method in a manned-unmanned aerial vehicle formation scene comprises the following steps:
receiving data acquired by a sensor, and constructing a random variable and a derivative variable based on the acquired data;
based on the random variable and the derivative variable, calculating threat probability of the opponent target to the my helicopter by adopting a dynamic Bayesian network;
and carrying out task allocation by adopting an improved Markov decision according to the threat probability.
In one embodiment, the random variables include target type, weapon radius, investigation radius, distance, direction, terrain, speed, within weapon radius, within investigation radius, membership, unmanned plane investigation.
The derived variables include capability assessment, risk, mobility, intent, threat.
The adoption of the random variable and the derivative variable is convenient for acquisition and has good effect of evaluating the threat of the opposite party.
As an embodiment, a trapezoidal function f is used T (x) And f R (x) The function processes the random variable and the derivative variable into discrete values, and the random variable is divided into f at two ends of a value interval R (x) Function determination:f T (x) The function is: />
Wherein lambda, sigma, alpha, beta, gamma,As a parameter limiting the value of the current and next discrete values, x represents the threat of the counterpart target.
As an implementation manner, the threat probability of the ith target of the counterpart to the reference point w of the helicopter on the flight path is calculated by the following formula:
is the ith at the t momentTarget state of individual partner, ->For the ith counterpart target state at time t-1, pa(s) is s parent node, n=5,>is->In the case->Is a probability of (2).
As an embodiment, the modified markov decision is a five-tuple, expressed as < S, a, T, p, R >, where S is the state set, S e S, and s= (R, N) A ,T),s=(R,N A T), R is defined by a probability distribution function f R (x) The discrete value of the detection radius is obtained, T is the value obtained by the probability distribution function f T (x) Discrete values of the resulting capacity assessment, N A For the number of unmanned aerial vehicles, A is a possible action set of the helicopter in a state S epsilon S, p is threat probability, and r is an objective function value obtained by the helicopter executing action a in a state of t-1 to a state of t moment.
Compared with the prior art, the method and the system have the advantages that threat probability evaluation is carried out based on the dynamic Bayesian network, the Markov decision process is adopted for task allocation, the risk evaluation can be completed under a given scene, the UAV task list is generated to assist the pilot in decision making, the decision making similar to the human pilot is made, the threat degree of an enemy target is reduced, and the survivability of the helicopter is improved.
Other advantages of the present invention will be apparent from the description of the embodiments.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments or the description of the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a task allocation method in a manned-unmanned aerial vehicle formation scenario in an embodiment of the present invention.
FIG. 2 is a graph of variables and relationships between each other for assessing threat to a helicopter by a counterpart target in an embodiment.
FIG. 3 f R (x) The value of the function is shown schematically.
FIG. 4 f T (x) The value of the function is shown schematically.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The following describes in detail the implementation of the present invention in connection with specific embodiments:
in a manned-unmanned aerial vehicle formation scenario, a manned platform (such as a helicopter) and a plurality of unmanned aerial vehicles are generally included, and the unmanned aerial vehicles are controlled by the helicopter, and precisely a control system configured in the helicopter is used for task allocation, i.e. determining when to release the unmanned aerial vehicles, which unmanned aerial vehicles to release, and the like.
Referring to fig. 1, the task allocation method in the manned-unmanned aerial vehicle formation scene provided in the embodiment includes the following steps:
s10, receiving data acquired by the sensor, and constructing a random variable based on the acquired data.
In this embodiment, the sensors are mounted on each unmanned aerial vehicle, and the data collected by the sensors are related information of the target of the other party, including distance, direction, topography and speed, and based on these information, random variables for evaluating threat of the target of the other party to the my helicopter are constructed, where the random variables include target type, weapon radius, investigation radius, distance, direction, topography, speed, within weapon radius, within investigation radius, membership and unmanned aerial vehicle investigation degree.
The type of target is determined by the image acquired by the camera, and the type of target comprises, for example, an individual hand-held gun, an individual portable air defense weapon, as shown in the first row in table 1 below.
Distance refers to the distance between the counterpart target and the my helicopter. The method of the present invention is applicable to the scene of two-party combat, such as exercise, so that one party of the unmanned aerial vehicle formation is the my party, the other party is the other party, and the targets of the other party can be various types, as listed in the first row in table 1 below.
Weapon radius and reconnaissance radius refer to the target weapon radius of the partner, for example the reconnaissance radius of the partner radar. When the target type variable is determined, the weapon/investigation radius of the target can be further determined, and then the specific discrete value of the helicopter in the weapon/investigation radius can be determined according to the distance (as shown in the following table 1).
The direction refers to the direction of movement of the counterpart target relative to the my helicopter. The speed refers to the current moving speed of the counterpart target. Topography describes the effect of the surrounding environment on the movement of a target, such as a flat or obstructed area, determined by the average inclination. Membership refers to a relationship of a target to a helicopter, such as neutral, opponent, or unknown. The unmanned aerial vehicle detection degree refers to whether the unmanned aerial vehicle detects a target, and detection is 1, and detection is not 0.
S20, constructing derivative variables for evaluating threat of the counterpart target to the helicopter according to the random variables, wherein the derivative variables comprise capability evaluation, risk, mobility, intention and threat.
As shown in fig. 2, fig. 2 shows the relationship of random variables to derivative variables and to each other. The capability assessment describes the capability of the counterpart target, i.e. essentially the threat level of the counterpart target to the my helicopter, the target type and weapon radius, investigation radius, distance are the main variables that pose a threat to the my drone, wherein the target type is as shown in table 1 below in the first line. By way of example, the capability assessment includes, for example, three discrete values, low, medium, and high. If the opponent target is a 'single-soldier hand-held machine gun', and the my helicopter is positioned inside the weapon radius and outside the investigation radius of the opponent, the capability assessment value of the opponent target is low; if the opponent target is a vehicle-mounted portable air defense weapon, and the my helicopter is outside the radius of the opponent weapon and is within the opponent investigation radius, the capability assessment value of the opponent target is the center; if the counterpart target is a "SAM missile air defense system" and the my helicopter is within the weapon radius and the reconnaissance radius of the counterpart, the capability assessment value of the counterpart target is high.
The movement capability describes the movement condition of the counterpart target relative to the my helicopter, and the movement capability of the counterpart target is influenced by the movement direction, the terrain and the speed of the my helicopter, and can take the values of inconsequential, attention and important attention.
The risk is determined by the type of target and the evaluation of the capability, i.e. the target of the opposite party is determined to be a low risk target, a medium risk target or a high risk target.
The intention describes the size of the intention of the opponent's target to attack the my helicopter. By way of example, attack intents can be categorized into low intents, medium intents, high intents, the intents being determined by capability assessment, mobility and membership.
In this embodiment, threat variables of the counterpart target to the helicopter are described in a hierarchical manner, e.g., five levels of none, low, medium, high, and very high, resulting from a combination of threat (three levels) and intent (three levels) variables of the counterpart target.
The variable is adopted in the embodiment, so that the variable can be obtained easily, and the effect of evaluating the striking capability of the weapon is good.
S30, processing the constructed random variable and the derivative variable into discrete values.
The data collected by the sensor are continuous data, the continuous data are processed into discrete data, and the discrete data are more beneficial to the calculation of membership function values. In the present embodiment, by means of a trapezoidal function f T (x) And f R (x) The function defines a probability distribution of discrete values, i.e. the value of each random variable is defined by f T (x) And f R (x) And (5) determining a function. The two functions function to giveAnd the probability distribution of the random variable is obtained so as to avoid jump of threat level caused by hard discretization and improve the accuracy of threat assessment. The random variable is divided into two ends of the value interval by probability f R (x) Function determination:
wherein lambda andas parameters limiting the values of the current and next discrete values, x represents the threat of the counterpart target, i.e. the value of the threat variable shown in fig. 2, also the value of the penultimate threat variable in table 1, is calculated, f R (x) The value of the function is shown in fig. 3.
f T (x) The function is:
wherein, alpha, beta, gamma,Is a parameter of a trapezoidal function, aims to smooth the discretization process, does not generate violent jump on the calculation of the target threat level, and f T (x) The function is shown in fig. 4.
For the example in the simulation test, the discrete values obtained by using the above-described respective random variables with the formulas (2) and (3) are shown in the following table 1. For example, the weapon radius, the investigation radius variable have discrete values of {1500,5000, infinity, and {3000,7500, ≡ }, the discrete value of the distance variable is 500,2000,20000, and infinity, and then the relationship between the my helicopter and the weapon/investigation radius of the counterpart target can be obtained, with the values {0,1,2,4, ≡ }.
S40, determining threat probability of the opponent target to the my helicopter by adopting a dynamic Bayesian network according to the discrete value.
Bayesian networkThe states of the nodes in (a) are represented by s, s= (R, N) A T), R is defined by a probability distribution function f R (x) The discrete value of the detection radius is obtained, T is the value obtained by the probability distribution function f T (x) The discrete value of the obtained capability assessment; n (N) A Is the number of unmanned aerial vehicles.
In order to calculate the threat degree on the flight path of the helicopter, the threat probability of the opponent target in the task area to W reference points with equal distances on the route is estimated, and the average threat degree on the flight path is obtainedIs that
Wherein N is E For the number of opponent targets, W is a reference point set on the flight path of the helicopter, p (i, W) is a threat probability value of the ith opponent target to the reference point W, and W epsilon W. p (i, w) is calculated from formula (1):
wherein,for the ith opponent target state at time t, < >>For the i-th counterpart target state at time t-1, pa(s) is the parent node of s, N is the number of nodes of the dynamic bayesian network (Dynamic Bayesian Networks, DBN), n=5 in this embodiment; />Is->In the case->Is a probability of (2). p (·) represents probability.
S50, task allocation is carried out by adopting a Markov decision according to the threat probability.
In this embodiment, a markov decision (Markov Decision Process, MDP) is adopted for task allocation, and an improvement is made on the basis of the task allocation, and the four-element MDP is added by one element T, which is described as follows:
<S,A,T,p,r> (4)
wherein S is a state set, s∈s, and s= (R, N) A T), R is a investigation radius, T is a capability assessment; n (N) A Is the number of unmanned aerial vehicles. A is a possible action set of the helicopter in a state S epsilon S, p is threat probability, for example, p (S '|s, a) represents probability that the helicopter reaches a future state S epsilon S in a given state S and action a, r is an objective function value obtained by the helicopter executing the action a in a t-1 moment state to reach a t moment state, for example, r (S, a) is an objective function value obtained by the helicopter executing the action a in the state S to reach the future state S'.
The actions in action set a include:
(1) Re-planning the path: re-planning the helicopter path to avoid dangerous points and reduce the threat degree of weapons of the other side;
(2) Detecting a target x of the other party: releasing an unmanned aerial vehicle for detecting the target x of the other party;
(3) No action is taken: and flying according to a set route without other actions.
Helicopter motion a epsilon A, vectorRepresenting a ranking for the drone and the other party target, i.e. a i ={0,1},i∈{1,2,…,N E }. Then there is a= (0, …, 0) indicating no action and a= (0, …, 1) indicating a re-planned path. Since the intermediate actions indicate the ordering of how the drones are released, the number of "1" s in a should be less than or equal to N U ,N U For the number of available unmanned aerial vehicles, there are N U ≤N A . For example, if there are two opponent targets, there are three possible "investigation opponent target x" actions: releasing the unmanned aerial vehicle investigation target 1, releasing the unmanned aerial vehicle investigation target 2, and releasing the unmanned aerial vehicle investigation targets 1 and 2. Thus, the number of possible actions |a| is:
without losing generality, the unmanned aerial vehicle investigation degree of each unmanned aerial vehicle is set to be irrelevant to the number of available unmanned aerial vehicles and the threat degree on the helicopter airlines, and then the unmanned aerial vehicle has
Wherein N is U And N' U The number of available unmanned aerial vehicles in states s and s ' respectively, t and t ' are discrete values of the risk of the opposite target in states s and s ' to the helicopter respectively, r i And r' i The discrete values of the reconnaissance degree of the unmanned aerial vehicle to the target i in the states S and S' are respectively, a is the action under the state S, and S epsilon S.
The objective function of the whole system can be expressed as a size ofEach element in the matrix records an objective function that performs action a in state s. />The number of combinations of distances from each counterparty to the unmanned aerial vehicle is given, where |d| is the cardinal number of distances from counterparty to the unmanned aerial vehicle.
How to utilize UAVs is embodied in an objective function that considers the scout value of the scout target, the UAV to target distance, and how to reprogram the helicopter path to reduce the threat, then the objective function under action a is
r=a T (r rec v r +r uav v d +r plan v p +r thr v thr ) (8)
Wherein T is a transposed symbol, r rec >0, is the objective function drop value of the target detection radius of the counterpart, r uav <0, r for releasing penalty of UAVs plan <0, penalty for re-planning helicopter path, r thr >0, the objective function value for the re-planned path, and r rec 、r uav 、r planVector v r And v d Respectively representing the investigation radius of the opponent target and the distance of the opponent target from the helicopter (the last element in the two vectors is 0), v p = (0, …,0, 1) is a constant vector (i.e. the last element in the vector is 1), v thr = (0, …,0, 1), which is the difference between the current path and the re-planned path (r if the threat of re-planning the path increases thr <0; the last element in the vector is 1), and v r 、v d 、v p 、/>
In order to verify the feasibility of the task allocation method, 6 scenes are designed for simulation tests. In the test, the discrete values employed for the DBN are shown in table 1, while the objective function values for the MDP are shown in table 2.
Table 1: discrete values in DBN (all are dimensionless values)
In table 1, the opponent's target ability assessment, mobility, risk and intent are all discretized into three values, while the distance values are discretized into four values, the threat is classified into five classes, these variables are qualitative determinations, not specific values.
Table 2: objective function value used by MDP
The function values shown in table 2 are superior values obtained through a plurality of continuous simulations, and can be practically applied based on the values.
2 trials were performed for each scene and 12 test results are given in table 3. For example, column 3, the resulting set of threat levels is {0,0,0,3,9}, and the analysis results in a relatively concentrated threat level. However, a more "relaxed" distribution may sometimes occur, such as column 10, with a set of threat level discrete values of {1,5,2,3,1} for each threat level.
TABLE 3 threat level discrete values from 12 tests
The helicopter will act at each reference point on the flight path, and the number of actions taken against the threat in 6 scenarios is shown in table 4. The accuracy of the threat assessment and the accuracy of the advice taken are shown in table 5.
Table 4: action statistics in 6 scenarios
Action taken Scene 1 Scene 2 Scene 3 Scene 4 Scene 5 Scene 6
No action 0 0 1 2 0 0
Re-planning paths 7 11 9 0 1 4
Investigation of partner target 1 9 8 7 10 6 11
Investigation of partner target 2 9 10 8 11
Investigation of partner target 3 11 8
Note that: each column of data includes a condition of re-planning the path while probing for the counterpart target for threat.
Table 5: accuracy of threat assessment and accuracy statistics of advice taken in 6 scenarios
Accuracy of Scene 1 Scene 2 Scene 3 Scene 4 Scene 5 Scene 6 Mean value of
Quasi-threat assessmentDegree of certainty 75% 74% 74% 56% 82% 80% 73.67%
Accuracy of taking advice 100% 95% 85% 91% 100% 84% 92.50%
According to the test results, aiming at the combined investigation task of the unmanned aerial vehicle and the unmanned aerial vehicle formation of the helicopter, the invention adopts a dynamic Bayesian network and a Markov decision, and the risk assessment is completed under the given 6 scenes, the suggestion of a UAV task list is given to assist the decision of a pilot, the decision of the threat can be made, and an approximate reasonable decision is made.
The above embodiments are merely specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any equivalent modifications, substitutions and improvements will readily occur to those skilled in the art within the scope of the present invention, and these modifications, substitutions and improvements are intended to be included in the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.

Claims (7)

1. The task allocation method in the manned-unmanned aerial vehicle formation scene is characterized by comprising the following steps of:
receiving data acquired by a sensor, and constructing a random variable and a derivative variable based on the acquired data;
based on the random variable and the derivative variable, calculating threat probability of the opponent target to the my helicopter by adopting a dynamic Bayesian network;
and carrying out task allocation by adopting an improved Markov decision according to the threat probability.
2. The method of task allocation in a unmanned aerial vehicle formation scenario of claim 1, wherein the random variables include target type, weapon radius, reconnaissance radius, distance, direction, terrain, speed, within weapon radius, within reconnaissance radius, membership, unmanned aerial vehicle reconnaissance.
3. The method for task allocation in a unmanned aerial vehicle formation scenario according to claim 2, wherein the derived variables include capability assessment, risk, mobility, intent, threat.
4. A method of task allocation in a unmanned aerial vehicle formation scenario according to claim 3, wherein the weapon radius is determined by the weapon radius and distance, the scout radius is determined by the scout radius and distance, the capability assessment is determined by the weapon radius and the scout radius, the mobility is determined by direction, terrain, speed, the intent is determined by the mobility, membership and capability assessment, the risk is determined by the target type and capability assessment, and the risk and intent determine the threat level.
5. The task allocation method in a unmanned aerial vehicle formation scenario according to claim 1, wherein a trapezoidal function f is used T (x) And f R (x) The function processes the random variable and the derivative variable into discrete values, and the random variable is divided into f at two ends of a value interval R (x) Function determination:
f T (x) The function is:
wherein lambda, sigma, alpha, beta, gamma, delta are all epsilonAs a parameter limiting the value of the current and next discrete values, x represents the threat of the counterpart target.
6. The task allocation method in a unmanned aerial vehicle formation scene according to claim 5, wherein the threat probability of the ith target of the opponent to the reference point w of the helicopter on the flight path is calculated by the following formula:
for the ith opponent target state at time t, < >>For the ith counterpart target state at time t-1, pa(s) is s parent node, n=5,>is->In the case->Is a probability of (2).
7. The method of task allocation in a unmanned aerial vehicle formation scenario of claim 6, wherein the modified markov decision is a quintuple, denoted < S, a, T, p, R >, where S is a set of states, S e S, and s= (R, N) A ,T),s=(R,N A T), R is defined by a probability distribution function f R (x) The discrete value of the detection radius is obtained, T is the value obtained by the probability distribution function f T (x) Discrete values of the resulting capacity assessment, N A For the number of unmanned aerial vehicles, A is a possible action set of the helicopter in a state S epsilon S, p is threat probability, and r is an objective function value obtained by the helicopter executing action a in a state of t-1 to a state of t moment.
CN202311827315.9A 2023-12-28 2023-12-28 Task allocation method in manned-unmanned aerial vehicle formation scene Pending CN117787625A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311827315.9A CN117787625A (en) 2023-12-28 2023-12-28 Task allocation method in manned-unmanned aerial vehicle formation scene

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311827315.9A CN117787625A (en) 2023-12-28 2023-12-28 Task allocation method in manned-unmanned aerial vehicle formation scene

Publications (1)

Publication Number Publication Date
CN117787625A true CN117787625A (en) 2024-03-29

Family

ID=90385109

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311827315.9A Pending CN117787625A (en) 2023-12-28 2023-12-28 Task allocation method in manned-unmanned aerial vehicle formation scene

Country Status (1)

Country Link
CN (1) CN117787625A (en)

Similar Documents

Publication Publication Date Title
CN108731684B (en) Multi-unmanned aerial vehicle cooperative area monitoring airway planning method
CN110597264B (en) Unmanned aerial vehicle counter-braking system
CN111044052B (en) Unmanned aerial vehicle self-adaptive navigation system and method based on intelligent sensing
CN109960148B (en) Autonomy evaluation method and system for intelligent unmanned system
Pan et al. Maneuver decision for cooperative close-range air combat based on state predicted influence diagram
Duan et al. Multiple UCAVs cooperative air combat simulation platform based on PSO, ACO, and game theory
Fu et al. The overview for UAV air-combat decision method
Darrah et al. Using genetic algorithms for tasking teams of raven UAVs
Sun et al. Route evaluation for unmanned aerial vehicle based on type-2 fuzzy sets
Dantas et al. Engagement decision support for beyond visual range air combat
CN113139331A (en) Air-to-air missile situation perception and decision method based on Bayesian network
CN115759754A (en) Beyond-visual-range air combat simulation target threat assessment method based on dynamic game variable weight
Fan et al. Path planning for a reconnaissance UAV in uncertain environment
Peterson Dynamic grouping of cooperating vehicles using a receding horizon controller for ground target search and track missions
Karneeb et al. Distributed discrepancy detection for a goal reasoning agent in beyond-visual-range air combat
CN112396298A (en) Unmanned helicopter multi-machine cooperative task planning method
CN117787625A (en) Task allocation method in manned-unmanned aerial vehicle formation scene
Kim et al. Airborne multisensor management for multitarget tracking
CN116050515A (en) XGBoost-based parallel deduction multi-branch situation prediction method
McLemore et al. A model for geographically distributed combat interactions of swarming naval and air forces
Li et al. Improved heuristic algorithms for UAVs path planning in hazardous environment
CN115097861A (en) Multi-Unmanned Aerial Vehicle (UAV) capture strategy method based on CEL-MADDPG
Yuksek et al. Development of UCAV fleet autonomy by reinforcement learning in a wargame simulation environment
CN114330093A (en) Multi-platform collaborative intelligent confrontation decision-making method for aviation soldiers based on DQN
Liu et al. Research of UAV cooperative reconnaissance with self-organization path planning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination