Coordination Tasks Pre-allocation And Redistribution Studies In UAVs

With the development of military technology in various countries and the increasing improvement of the technological level in the aviation field,how to control a unmanned aerial vehicles(UAVs)team as a whole has become one of the hot spots in modern research to perform a certain military mission and to maximize the total operational utility.Cooperation of multiple UAVs as a team to accomplish tactical deployment will get better combat effectiveness and higher team performance.Due to the influence of mission environment and the UAV resources,the problem of multiple UAVs collaborative task allocation is a multi-objective optimization problem under multi-constrained conditions.This dissertation is dedicated to the pre-allocation and redistribution of multi-UAV cooperative missions,which takes the multi-UAV coordinated strike missions as the research background.The main work is as follows:1.A "integral" and "multi-level" task pre-distribution model are established.The "integral" task pre-distribution model is usually used in situations where the task scale is small and the solution is easy.It mainly considers constraints such as multi-UAV cooperation,combat radius,ammunition resources,etc.At the same time,it uses the target value and survival probability function,flight voyage function,and ammunition cost function as evaluation criteria.However,the "multi-level" task pre-distribution model is usually used for larger tasks and complicated solutions.It is usually divided into a first-level distribution model between the UAV formation and the target cluster and a two-level distribution model of the UAVs and the targets.The establishment of two kinds of task pre-distribution models lays a solid foundation for solving the multi-UAV collaborative task allocation scheme.2.The PSO_BAS algorithm is proposed.As a classical optimization algorithm,PSO has the characteristics of fast convergence and strong ability of global optimization.PSO(Particle swarm optimization)is a classic optimization algorithm,which has the characteristics of fast convergence and strong global optimization ability.However,when it is applied to the task assignment of multi-UAV cooperative tasks,there are some defects,for example,it is slow convergence in late stage and easy to fall into local minimum.In order to solve these problems,the local convergence ability of the BAS(Beetle Antennae Search)algorithm is used to search the global optimal solution of the PSO in this paper.At the same time,the acceptance principle is introduced to solve the problem of falling easily into local minimum,andthe inertia weight of the PSO is adjusted linearly to accelerate the convergence speed.So the improved algorithm is validated as valid by example simulation.3.The "integral" distribution model is solved.Taking the air-to-ground strike as the research background,this paper uses the PSO_BAS algorithm to solve the "integrated" task assignment model.Traditionally,the simulation experiments were performed with simple and complicated examples.Compared with the traditional PSO,the experimental results show that the PSO_BAS solves the problem of task distribution with faster convergence speed and higher precision,obtaining a satisfactory task pre-distribution scheme.4.A "multi-level" solution strategy is proposed.The complex task allocation is divided into two levels to solve separately by layering theory.In the first level,the(Fuzzy C-means)FCM algorithm is used to cluster the mission objectives,and the target cluster with the same number of UAV formations is obtained.Then complete the assignment between the UAV formation and the target cluster.After the first-level assignment scheme is determined,the second-level allocation utilizes a parallel solving strategy,and the PSO_BAS is used to complete the task allocation of the UAVs within the UAV formation and the targets in the target cluster.The resulting task allocation scheme is a "multi-level" one.Compared with the "integral" task pre-allocation solution time,it is verified that the "multi-level" solution strategy has better real-time performance for the "integral" solution strategy through experimental simulation when the task scale is large.5.A supervision-based sequential auction mechanism is proposed.Market auction mechanism is a commonly used distributed solution method.When it is used for multi-UAV mission redistribution problems,there are deficiencies such as the inability to auction multiple jobs at the same time and the inability to accomplish job migration when there are no auctioneers.In view of the above defects,the auction performance function and sequential auction redistribution algorithm are designed to overcome the above problems.For two unexpected situations,the task re-distribution scheme is obtained through simulation experiments,which verified the feasibility and effectiveness of the mechanism.