Research Of Mobile Robot Trajectory Planning And Obstacle Avoidance Algorithm In Dynamic Environment

With the rapid development of robotics,new challenges are presented to the research of path planning,navigation and positioning of mobile robots.For the problem of robot path planning,the requirements for robot planning algorithm are increasing due to its increasingly complex application scenarios.Single planning algorithm can no longer meet the needs of robots in the real working environment.In response to this problem,in this thesis,the QBot 2 mobile platform is taken as the research object,focusing on the global trajectory planning of mobile robots,the local obstacle avoidance planning of multiple mobile robots,and the hybrid trajectory planning of mobile robots in complex environments.Firstly,the non-holonomic constraints of QBot 2 mobile platform are analyzed,and the kinematics model of QBot 2 mobile platform is established.Based on the problem of global trajectory planning for mobile robots,this thesis analyzes the basic theory and deficiency of rapidly exploring random tree(RRT)algorithm.A hierarchical random sampling weak random RRT algorithm is proposed,in addition,the algorithm also combined with a fast amplitude limiting feedback controller,and transformed the path into a pose trajectory conforming to the constraints of the kinematics model of the mobile robot,so as to realize the global trajectory planning of the mobile robot.The simulation results show that the hierarchical random sampling weak random RRT algorithm can ensure high search efficiency in the narrow environment,the pose trajectory generated by the fast amplitude limiting feedback controller satisfies the kinematics model constraints of the mobile robot,and the trajectory parameters strictly conform to the physical model constraints.Secondly,aiming at the dynamic planning problem of mobile robot,the basic theory of optimal reciprocal collosion avoidance algorithm and its limitation and deficiency in application scenarios are analyzed,a real-time-constant speed-ORCA algorithm is proposed.The algorithm takes into account the equivalent physical model of mobile robot,combined with trajectory calculation strategy,and improves the calculation method of ORCA algorithm to avoid collision velocity set.The simulation and experimental results show that the real-time-constant speed-ORCA algorithm can overcome the "deadlock" problem and other constraints.Moreover,the local trajectory planning of multiple mobile robots can be realized in the dynamic environment.Finally,this thesis focuses on the trajectory planning of mobile robot in dynamic environment.The basic principle of dynamic window approach and its shortcomings in robot navigation are analyzed,and a dynamic window approach based on multi-level virtual target is proposed.This algorithm can recover the trajectory of mobile robot when it deviates from the global trajectory.At the same time,a hybrid trajectory planning algorithm for multiple mobile robots in dynamic environment is designed.The algorithm can realize global trajectory planning,real-time obstacle avoidance planning and trajectory recovery for multiple mobile robots in dynamic environment.Simulation and experiments show that the proposed hybrid planning algorithm can ensure the mobile robot to reach the target area safely and without collision in the dynamic environment,and has high reliability.