Research On Cooperative Formation Of Multi-quadrotor UAVs Based On The Theory Of Consistency

In today’s world,the application of UAVs has become a hot research object in both military and civilian fields.Faced with diverse tasks and complex environments,a single UAV is usually unable to meet the needs,while multiple UAVs rely on effective coordination strategies and the different functions of the UAV itself can effectively solve many difficult problems such as unmanned combat,target search,and communication relay.The coordinated formation of multiple UAVs is an important part of mission execution.This thesis takes quadrotor UAVs as the research object,and designs and implements several hot topics in the coordinated formation of quadrotor UAVs.First of all,in order to maintain the formation of UAV swarm,the distributed information exchange strategy is adopted,and the formation control method of the leader-follow method is integrated into the information consistency theory,the communication topology is described with graph theory,thereby the distributed consistency formation is designed,the control algorithm enables the formation of UAVs to maintain state consistency during the flight.In the formation change problem,the KM algorithm is extended to the path planning of the formation UAVs,and the task allocation thinking is used to find the best flight path for the formation change,therefore reducing the flight distance of the entire formation system,it effectively reduces the time required in the formation change process,and reduces the flight energy consumption,which has high practical application significance.Secondly,a collaborative formation avoidance algorithm is designed.During the formation flight,UAVs may collide between aircraft or collisions with obstacles in the airspace.Therefore,effective obstacle avoidance algorithms are needed to solve these problems.In this thesis,the consensus theory and the artificial potential field method are used to jointly design obstacle avoidance algorithms.The traditional artificial potential field method is prone to fall into the local minimum and cause the target to be unreachable.Therefore,this thesis improves the traditional method by segmenting the potential field function and adding the loop force to the calculation of the resultant force to solve the problem.In the obstacle avoidance algorithm,the output obtained by the improved artificial potential field function is used as the desired state value,so that it can adjust the system input and correct the state error,and finally the formation achieves the same state under the premise of successful obstacle avoidance.Finally,a software-in-the-loop simulation environment is built to systematically verify the effectiveness of the proposed algorithm in a simulated three-dimensional environment.Among them,a simulation platform with the ROS robot operating system as the distributed communication carrier,PX4 as the UAV flight control and Gazebo as the three-dimensional simulation simulator was constructed.Under this platform,the problems of UAV positioning and obstacle recognition faced in 3D simulation are solved.The VIO and GPS data fusion method is used to solve the inaccurate positioning of UAVs under special circumstances.The application of the Yolo V4 algorithm takes into account the speed and accuracy of obstacle recognition,lays the foundation for later implementation on the embedded platform.Using Python code to write ROS nodes in a three-dimensional environment,the entire formation system was implemented,and the data was collected using MATLAB tools for intuitive flight trajectory display,which showed that the entire system and the designed algorithm are correct and effective.