Research On UAV Cluster Control Algorithm Under Network Attack Environment

With the increasing scale of UAV formation and the improvement of Ad-Hoc network performance,the security problem becomes more and more prominent in UAV cluster task planning.To solve this problem,this thesis mainly selects several typical attack methods in the network environment,namely forgery attack,tampering attack and flooding attack.Aiming at the three attack methods,this thesis proposes a forgery attack detection method based on differential constraint,combined with ant colony algorithm based on octree,which is used to complete the path planning in forgery attack environment;and proposes a multi attack detection algorithm based on information redundancy gradient constraint,combined with ant colony algorithm based on particle swarm optimization,which is used to complete the path planning in multi attack environment.The main contributions of this thesis are as follows:1.When forgery attack occurs in the network environment,UAV cluster uses the position information sent by forged nodes to plan the path,which leads to the collision between UAV clusters.To solve this problem,this thesis proposes an ant colony algorithm based on octree(Oc-ACO)to optimize the UAV formation defense avoidance model in network environment.Firstly,the UAV’s real-time attack model is simplified and the UAV’s detection efficiency is improved.The experimental results show that the Oc-ACO algorithm can complete the path planning quickly in the case of forgery attack.2.When there are tampering attacks and flooding attacks in the network environment,the attacked UAV will move irregularly and may collide with the UAV cluster.To solve this problem,a fast ant colony algorithm based on particle swarm optimization(PSO-MACO)is proposed to optimize the UAV formation defense model.Firstly,A multi attack detection algorithm with information redundancy gradient constraint is proposed to enable UAV cluster to detect tampering attack and flooding attack in real time.Secondly,the attacked UAVs are transformed into dynamic obstacles,and the UAV formation defense model is established.Finally,a fast ant colony algorithm based on particle swarm optimization is proposed.The experimental results show that the PSO-MACO algorithm meets the requirements of real-time path planning of UAV formation system in.3.The hardware in the loop simulation platform is built.The platform consists of UAV autonomous cooperation framework,communication framework and display platform.This platform uses service-oriented architecture,encapsulates path planning,task allocation,sensor data processing,waypoint management and other functions as services,and uses asynchronous communication message queue to connect them.UAV cluster system can realize the simulation experiments of ad hoc network attack,distributed control and route planning on this platform.By testing the hardware in the loop simulation platform,the feasibility of each function of the platform is verified;and the two algorithms are verified on the hardware in the loop simulation platform,and the experimental results are consistent with the theoretical simulation results of the algorithm on MATLAB.