Research On UAV Cluster Control Technology In Restricted Environment

With the development of science and technology and the increasing demand of application,UAV has shown a blowout development in recent years.Due to the limitations of traditional single machine operation,UAV cluster has attracted more and more attention.Researchers began to explore how to make UAV cluster complete specific tasks as efficiently and autonomously as natural biological groups.Aiming at this problem,based on the position difference vector between UAVs in the cluster,this paper designs the cohesion speed,repulsion speed and alignment speed of UAVs to form the total expected speed,and makes UAVs track the expected speed to achieve the purpose of UAV cluster control.Among them,the cluster cohesion algorithm is designed by using local information interaction,and the proportional differential control is introduced to shorten the cohesion time and reduce the probability of collision.Finally,the effectiveness and reliability of the algorithm are verified by numerical simulation and real flight experiments.The main contributions and innovative work of this paper are summarized as follows.(1).Research on cluster rules in restricted environment.Constrained by the local communication conditions,the information obtained by the individual cluster is limited.At the same time,the movement space of clusters is also limited by physical boundaries and obstacles.In these restricted environments,based on Reynolds rule,this paper designs cluster rules of cohesion,exclusion and alignment,and analyzes the speed decomposition method and parameters in detail.In the cluster rule,cohesion component becomes the dominant factor when UAV is far away;when UAV is near,but not in the cohesive range,there are velocity exclusion term,velocity alignment term and velocity cohesion term in individual velocity expectation;when UAV is near and within the cohesive boundary,velocity exclusion and velocity alignment are different The role of cohesion is neglected.(2).The simulation design of cluster rules.The simulation design of cluster rules is divided into three stages: the first stage establishes the mathematical model of four rotor UAV for experiment,and designs the controller to track the desired speed,and completes the speed control of four rotors through numerical simulation;the second stage is based on the robot SIM framework of vicsk team in Hungary,and realizes the algorithms of speed rejection,speed alignment and velocity cohesion in the restricted environment,and sets up the algorithm The controller is designed to reduce the clustering time and the probability of collision between individuals.Finally,the effectiveness of the algorithm is verified by the numerical simulation of multiple four rotor UAVs.In the third stage,the algorithm is transplanted to the gazebo simulation,and the rationality and effectiveness of the algorithm and parameters are verified before the flight.(3).Four rotor flight experiment.In order to verify the rationality of the cluster rule,the flight experiments of several quadrotor UAVs are designed.In the experiments,the procedures and steps of the quadrotor decision-making layer are analyzed and designed.The flight results show that the velocity decomposition method,velocity exclusion,alignment,boundary exclusion and cohesion algorithm are reasonable and effective.