Collaborative Planning Parallel System Of UAVs Based On Immune Clone And Adaptive Ant Colony

In the modern war,with the continuous progress of air defense technology and air defense system,unmanned aerial vehicle(UAV)has become an effective weapon in the battlefield environment due to its strong mobility,good concealment and low cost.As the key of military war,the path planning and task assignment of UAVs have attracted more and more attention.On one hand,UAV sometimes can't make a comprehensive and accurate judgment quickly when facing with the complex and changeable battlefield environment,so UAV has to change the battle plan adaptively.On the other hand,as the air defense technology becomes more and more advanced and the air defense system becomes more and more perfect,the operational requirement for UAV becomes more and more strict.The concept of the parallel system applied to collaborative planning of UAVs becomes a new research hotspot.In this thesis,specific solutions for path planning and task assignment of UAVs as well as its application in parallel system are proposed as follows:(1)The collaborative path planning of UAVs should not only consider the threat in the surrounding environment and the constraint of UAV 's own performance,but also meet the requirement of time collaboration between UAVs.In order to reduce the planning difficulty of the problem,a two-layer structure is usually adopted for collaborative path planning,that is,the path planning layer and the collaborative planning layer.In this thesis,the immune clone algorithm is applied to path planning layer by establishing two objective function which include threat cost and fuel consumption cost.In the process of initialization,guidance factor and enlightenment factor are introduced to generate initial antibodies of high quality and accelerate the convergence rate.The feasibility of antibody can be improved by considering the relationship between track nodes in immune gene manipulation.And then some potential infeasible antibodies are modified to increase the diversity of antibodies in the process of antibody modification.After obtaining the candidate tracks of UAV in the track planning layer,the collaborative function between different tracks is designed in the collaborative planning layer,and the final collaborative scheme is selected according to the principle of minimum cost and time collaboration.(2)The task assignment for multiple UAVs should not only assign tasks to each UAV,but also consider the matching between the UAV teams and the target groups.In order to reduce the difficulty of problem,task assignment is usually divided into three sub-problems which include target clustering,cluster allocation and target assignment.Firstly,k-means clustering is used to cluster the targets in the battlefield to obtain different target groups.And then the hungary algorithm is used to match the UAV teams with different target groups.Finally,the adaptive ant colony algorithm with increase revenue matrix is used to allocate the targets within each UAV team.In the adaptive ant colony algorithm with revenue matrix proposed in this thesis,the increase revenue matrix is firstly constructed in the initialization process to optimize the update of transition probability and accelerate the convergence rate.And then 2-opt optimization is added to expand the search scope in each iteration.Finally,the maximum and minimum pheromone concentration and adaptive factor are set to adjust the volatility coefficient automatically so as to balance the global search ability and convergence speed of the algorithm.(3)To build the parallel simulation system of UAVs path planning,MFC interface is established at first.Artificial simulation system is building by Matlab and actual combat system is building by electronic map,and then connect them by information interacting.The artificial simulation system is established by collecting data from the actual combat system.The optimal route planning scheme is obtained by calculating experiment which is transmitted to the actual combat system for operation and dynamic demonstration in real time.At the same time,the parameters of the route planning algorithm and model parameters are constantly adjusted,and the scheme deduction and auxiliary decision-making are carried out again,so as to achieve the effect of parallel execution.