Research On Swarm Formation Of Quadrotors Based On Local Behavioral Rules

Multiple quadrotors cooperation can improve the robustness of system and efficiency of mission.With the increase of the number of UAVs,it is difficult for centralized control nodes to obtain the information of each UAV in real time.Therefore,each UAV must realize the coordination of non-centralized dispatching groups through the interaction of information and behaviors among individuals with limited neighborhoods.In addition,the swarm formation must have the ability to form different formations under different mission scenarios.This thesis focuses on the method of swarm formation of multiple quadrotors based on the acquisition of local information and distributed individual behavior rules.The first work of this thesis is the transformation between acceleration control and the attitude control,for simplifying the algorithm complexity to control swarm formation.The local behavior rules of swarm formation are proposed with reference to the motion model of the biological cluster.By constructing artificial potential field,velocity and position consistency,the language rules are transformed into the specific control law of formation,and achieve uniform coverage of the moving target.On the aspect of formation reconfiguration,by changing the behavior rules,the relevant control law is designed to achieve the effective reconstruction from uniform coverage to surrounding formation.Finally multi-UAVs simulation experiment is conducted to verify the effectiveness of algorithm and the time efficiency of the swarm formation.From the point of security,analyze the relationship between the parameter of behavior rules,the distribution of initial position and collision-free aggregation are analyzed in order to provide a security check for the algorithm's application on multi-UAVs system.The results indicate that the method based on local behavior rules can effectively form the uniform coverage formation,and achieve reconfiguration to surrounding formation by changing the local behavior rules.Multi-UAVs simulation and related conclusion can provide a test to determine the parameter of neighborhoods and the distribution of initial position,improving the safety of algorithm deployment on multiple quadrotors system.