Research On Swarm Confrontation Motion Control Based On Dynamic Grouping Strategy

In recent years,multi-agent systems(MAS)have become a research focus both domestically and internationally.UAV swarm,consisting of homogeneous or heterogeneous types of UAVs,is centered around inter-agent control strategies that facilitate collaboration to accomplish a wide range of complex tasks.UAVs possess typical agent characteristics and rely on their ability to exchange information with one another to coordinate their actions.Therefore,MAS theory can be applied to address the challenges of large-scale UAV swarm confrontation.This thesis focuses on the problem of UAV swarm confrontation in dynamic environments,and studies the autonomous control strategies for swarm motion in three typical confrontation scenarios: swarm tracking,swarm encirclement,and swarm-toswarm confrontation,corresponding to the three interaction modes of 1v1,Nv1,and Nv N between the enemy and the swarm.The main work includes the design and optimization of distributed dynamic grouping strategies,as well as the study of swarm confrontation motion planning methods in typical scenarios.The main work of this thesis is as follows:Firstly,this thesis presents the motion control problem of UAV swarm in adversarial environments,including the logical structure diagram of swarm-to-swarm control system and the UAV swarm motion model consisting of single swarm motion model and dualswarm interaction model.The control strategy consists of two parts: dynamic grouping and swarm motion planning.Dynamic grouping is a prerequisite for swarm motion planning,setting task goals for individual UAV in the swarm.In dynamic grouping,the distributed target assignment strategy and its optimization of classical algorithms are mainly introduced,while swarm motion planning considers the planning problems of several classic motion scenarios from multiple perspectives and directions.Secondly,the distributed dynamic grouping strategy for UAV swarms was studied,and the ECBBA algorithm(Extend consensus-based bundling algorithm,ECBBA)was proposed for the target assignment problem in dynamic UAV swarm adversarial scenarios.This algorithm optimizes the CBBA distributed allocation algorithm,making it applicable to the target assignment problem in swarm-swarm adversarial scenarios.Then,this thesis studied the group motion planning methods for UAV swarms,and designed control strategies for different task objectives such as single swarm motion,target tracking,swarm encirclement,and swarm-swarm confrontation.Specifically,for single swarm motion,a group motion algorithm based on central aggregation was designed to optimize the boid model.For swarm encirclement behavior,a comprehensive control strategy was implemented using a circular convergence controller,spacing layout controller,and collision avoidance controller to encircle dynamic targets.For swarmswarm confrontation,tactical behaviors were studied from the perspectives of greedy strategy,conservative strategy,and friendly support.Finally,simulation experiments were conducted to evaluate the dynamic grouping algorithm and the control laws for single-group motion and inter-group interaction discussed in the previous sections.The experimental results provide guidance and reference for real-world UAV swarm confrontation scenarios.