CN-120822739-B - Unmanned aerial vehicle cluster comprehensive task complexity index system construction and dynamic evaluation method

CN120822739BCN 120822739 BCN120822739 BCN 120822739BCN-120822739-B

Abstract

The invention relates to the technical field of intelligent unmanned aerial vehicle manufacturing, and discloses a method for constructing and dynamically evaluating an unmanned aerial vehicle cluster comprehensive task complexity index system, which comprises the steps of providing perfect target state parameters and task-equipment matching functions; the unmanned aerial vehicle cluster task allocation is converted into a multi-objective optimization problem, the task allocation under the multi-objective optimization condition is carried out by using an improved NSGA-II algorithm, the index composition of the task target dimension and the task allocation dimension is comprehensively considered, a task complexity evaluation index system is established, and multi-dimensional task complexity comprehensive evaluation is carried out. The invention can enable personnel to quickly and accurately know the change of the task situation and make relatively correct judgment according to the change. The invention considers the target state and the dynamic change of the task and can provide guidance for the task execution of the unmanned aerial vehicle.

Inventors

ZHANG YU
DU LINLIN
XIAO GANG
Ma Qiongmin
HU JIANWEI
MA JINGJING
WANG YING

Assignees

军事科学院系统工程研究院系统总体研究所

Dates

Publication Date: 20260508
Application Date: 20250627

Claims (4)

1. The unmanned aerial vehicle cluster comprehensive task complexity index system construction and dynamic evaluation method is characterized by comprising the following steps of: Step one, determining target point target state parameters, wherein the target point target state parameters comprise target point space distribution density, target point shape and target point dynamic change frequency, and the determining the target point target state parameters comprises the following steps: construction of target points FIG. calculating each Thiessen polygon area Calculating the ratio of standard deviation to mean value of each Thiessen polygon area to obtain the spatial distribution density of the target point , wherein, Is the standard deviation of the area of the lens, Is the mean value; characterizing the shape of the target point by the target point target contour, carrying out parameterization coding on the target point target contour, wherein the contour shape is characterized in that , As the total number of contour points, Represent the first The complex coordinates of the individual contour points, Is the Fourier coefficient order; Collecting the moving track of a target point Calculating the maximum Index number : When (when) At the time, the target point dynamically changes frequency When (1) At the time, the target point dynamically changes frequency ; Step two, constructing a matching degree function of a task type matrix and equipment performance parameters, and calculating task-equipment matching degree to prove whether the used equipment is suitable for executing a target task; Thirdly, converting unmanned aerial vehicle cluster task allocation into a multi-target optimization problem, taking the target point target state parameter as a task target, and taking task-equipment matching degree as constraint of the multi-target optimization problem for improvement The algorithm performs task allocation under the multi-objective optimization condition; According to the condition of triggering the reassignment, the reassignment of the task is carried out, wherein the condition of triggering the reassignment comprises a hard threshold trigger and a soft threshold trigger, the hard threshold trigger is forced reassignment, the standby unmanned aerial vehicle is called to fill a gap of the task, and the deviation rate of the hard threshold trigger is as follows: wherein, the plan progress is the ratio of the current time to the total time limit of the task, if the deviation rate is more than 15%, the task is judged to be lagged; the soft threshold trigger is elastic redistribution, task priority and unmanned aerial vehicle formation are dynamically adjusted, and threat amplification is as follows: Threat augmentation = In the formula, For an average threat level of the past 5 minutes, If the threat increases by more than 20% and lasts for more than 30 seconds, judging the current instantaneous threat degree as burst risk; the objective function of task reassignment is: in the formula, The income of the task; Is a comprehensive risk; Cost for path collision; Is a weight coefficient; step four, based on the index composition of the task target dimension and the task allocation dimension, a task complexity assessment index system is established, wherein the task complexity assessment index system comprises a target layer, a criterion layer and an index layer, the target layer is comprehensive task complexity, the criterion layer is divided into the task target dimension and the task allocation dimension, the index of the task target dimension comprises target point space distribution density and target point dynamic change frequency, the index of the task allocation dimension comprises cross-domain task switching cost, resource conflict degree and cooperative constraint intensity, and the multi-dimensional task complexity assessment comprises the following steps: normalizing the indexes of the task target dimension and the task allocation dimension to the [0,1] interval by using a range method: wherein, C' is a normalized value, C is a current index value; And Respectively the minimum value and the maximum value of the normalized index; Designing membership functions of all indexes of the task target dimension, and calculating to obtain the membership of the task target dimension by adopting a trapezoid membership function: ; designing a membership function of each index of the task allocation dimension, and calculating by using a Gaussian membership function to obtain the membership of the task allocation dimension: ; Wherein mu is the central value of each index of the task allocation dimension, and sigma is the standard deviation; defuzzification is carried out by adopting a gravity center method, fuzzy output is converted into an accurate value, and comprehensive task complexity is calculated: in the formula, Assigning membership degrees of the task target dimension and the task dimension; And assigning weights of dimension membership degrees to the target dimension and the task.
2. The method of claim 1, wherein in step two, the task type matrix comprises a scout task Throwing task Evaluation task Wherein the scout task The equipment performance parameters of (a) include the detection distance Accuracy of identification And simultaneously tracking the number of targets Throwing task The equipment performance parameters of (a) include throwing height Hit accuracy Number of throws Number of targets to be thrown guided simultaneously Assessment task The matching degree function of the construction task type matrix and the equipment performance parameters comprises the following steps: Quantifying task requirements; Normalizing equipment performance parameters and mission requirements, wherein the forward equipment performance parameters The normalized formula of (2) is: , Negative equipment performance parameters The normalized formula of (2) is: , in the formula, Is provided with Is the first of (2) The value of the performance parameter of the item, 、 Respectively the first in the cluster Maximum and minimum values of the item performance parameters; calculating task-equipment matching degree according to task type , wherein, Is the weight of the j-th performance parameter.
3. The method of claim 1, wherein translating the unmanned cluster task allocation into a multi-objective optimization problem in step three comprises: task allocation is built into a mathematical model, and decision variables are defined as allocation matrixes In which, in the process, Representing unmanned aerial vehicle Whether or not to execute a task , For the total number of unmanned aerial vehicles, Is the total number of tasks; The optimization objective function of task allocation is: in the formula, Representing the maximum task benefit; Representing minimizing the overall risk; representing a minimum energy cost; Is a task Is a basic benefit of (1); Representing unmanned aerial vehicle Executing tasks Matching degree of (3); Threat degree of the task; Representing unmanned aerial vehicle For tasks Is a risk resistance of (1); Representation unmanned aerial vehicle Executing tasks Energy consumption of (2); the constraint conditions include: at most three unmanned aerial vehicles are distributed to each task: ; Task load Not exceeding unmanned plane Is of the bearing capacity of (a) : ; Time consuming task j The time window is within: ; Communication topology map To maintain strong communication, algebraic communication degree ; Task-equipment matching degree 。
4. The method of claim 1, wherein the modification in step three The task allocation under the multi-objective optimization condition by the algorithm comprises the following steps: For the following Unmanned aerial vehicle frame, in total Group initialization is carried out on each task scene data, two-dimensional matrix coding is adopted between the unmanned aerial vehicle and the task, and each individual represents an allocation scheme Repairing individuals who do not meet the constraint; non-dominant ordering, defining dominant relationship, and solving Dominance of If and only if: dividing the population into a plurality of leading edge classes by fast non-dominant ordering; And (3) calculating the crowding degree, namely calculating the crowding distance for individuals in the same front layer, and maintaining the diversity of solution sets: in the formula, Representing individuals Is the first of (2) Target values; 、 Is the first in the current leading edge layer Maximum and minimum values of the individual targets; randomly selecting from a population Individual, reserve the optimal person, for the selected parent individual Partitioning and crossing according to task types; And performing mutation operation.

Description

Unmanned aerial vehicle cluster comprehensive task complexity index system construction and dynamic evaluation method Technical Field The invention relates to the technical field of intelligent unmanned aerial vehicle manufacturing, in particular to intelligent unmanned system task planning, and provides a method for constructing and dynamically evaluating an unmanned aerial vehicle cluster comprehensive task complexity index system. Background The unmanned aerial vehicle cluster is widely applied to various task occasions, and when various tasks are executed, the influence factors such as task distribution, redistribution, task risks and the like need to be considered, and no unmanned aerial vehicle performance is required to be provided. In the face of complex and changeable environmental situations, if it is required to correctly judge whether the unmanned aerial vehicle cluster can execute the task, modeling and quantifying the task which changes dynamically are necessary, and establishing a corresponding index system. Most of the current researches are based on static or semi-dynamic environment modeling, and cannot effectively cope with real-time changing scenes such as target movement, newly added threat and the like. Firstly, the conventional task allocation model (such as TSP and VRP) is difficult to adjust the task allocation policy in real time under the dynamic environment due to high computational complexity, which results in reduced task execution efficiency and limitation in the multi-task dynamic scene. When the number of unmanned aerial vehicles exceeds 4, the problems of communication interference and channel overlapping can cause the collision rate to be remarkably increased in a dynamic environment. Secondly, research on quantification of coupling relation between task types and equipment performance is limited, and the existing task allocation model is not fully combined with matching relation between unmanned aerial vehicle isomerism and task types. The task allocation of the heterogeneous unmanned aerial vehicle cluster needs to consider the differences of load capacity, sensor performance and the like, but the existing MILP, CMTAP and other models are difficult to quantify the coupling relations. In addition, the response speed of a redistribution algorithm under sudden events such as equipment faults, threat changes and the like is insufficient, the calculation time of a traditional heuristic algorithm such as genetic algorithm, particle swarm optimization and the like is exponentially increased during dynamic redistribution, and the real-time requirement is difficult to meet. Finally, the traditional index system ignores the influence of morphological characteristics and equipment on the countermeasure situation of the target, the traditional evaluation index has single dimensions of multi-focusing time, energy consumption and the like, does not integrate multiple factors of environment, task and the like, and has blank in the aspect of a quantitative evaluation system of multi-dimensional task model and dynamic risk linkage. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a method for constructing and dynamically evaluating an unmanned aerial vehicle cluster comprehensive task complexity index system, which comprises the following steps: Step one, determining target state parameters of a target point, wherein the target state parameters of the target point comprise target point space distribution density, target point shape and target point dynamic change frequency; Step two, constructing a matching degree function of a task type matrix and equipment performance parameters, and calculating task-equipment matching degree to prove whether the used equipment is suitable for executing a target task; Thirdly, converting unmanned aerial vehicle cluster task allocation into a multi-objective optimization problem, taking the target point target state parameter as a task target, taking task-equipment matching degree as constraint of the multi-objective optimization problem, and performing task allocation under the multi-objective optimization condition by using an improved NSGA-II algorithm; And step four, based on the index composition of the task target dimension and the task allocation dimension, establishing a task complexity evaluation index system, and carrying out multi-dimensional task complexity comprehensive evaluation. Further, in the first step, the determining the target state parameter of the target point includes: Constructing a Voronoi diagram of the target point, calculating the area A i of each Thiessen polygon, and calculating the ratio of the standard deviation to the mean value of the area of each Thiessen polygon to obtain the spatial distribution density of the target point Wherein σ A is the area standard deviation, and μ A is the mean; characterizing the shape of the target point by the target point target contour, carrying ou