Research On Robot Motion Planning Based On RoboCup Standard Platform League

This paper takes the Robot World Cup Standard Platform League(Robo Cup SPL)competition as the background,the humanoid robot NAO as the experimental platform,and studies the problems related to the robot motion planning involved in the actual soccer match,mainly around the robot Global path planning and trajectory tracking control,local path planning,and site information fusion are carried out.The specific research work is as follows:Firstly,in order to solve the problem of motion planning in special rules such as friendly free kicks and penalty kicks,a global path based on an improved heuristic Informed Rapidly-exploring random tree*(Informed-RRT*)is proposed The planning algorithm is optimized for the problems of low initial convergence efficiency and poor performance in the narrow environment of the original algorithm: the initial state estimation strategy is used to adjust the convergence time of the algorithm;the algorithm is improved by introducing an optimized multi-random tree model The exploration efficiency of the feasible space;the real-time performance of the algorithm is improved by incorporating a hybrid caching mechanism in the sampling stage.As a supplement,this paper uses a trajectory tracking algorithm based on two types of error feedback to achieve effective tracking of the global path.Experimental results show that the scheme in this paper is feasible.Secondly,considering that the task environment always changes dynamically in the competition in the non-special penalty stage,this paper combines local path planning with global path planning to ensure that the robot can meet the dynamic obstacle avoidance requirements in the field.To this end,the problems such as velocity curve oscillation and local optimality existing in the classical dynamic window method are improved: Dynamic calculation of reference points is used to replace the fixed terminal attitude,which alleviates the jitter phenomenon when the robot approaches the global control point.By introducing the radiation value calculation theory,the local optimal tendency of the algorithm is effectively reduced,and a minimum value detection method is proposed.By establishing and solving the multi-objective optimization problem,the adaptive weight coefficient is realized,so that the robot can change its behavior mode in a specific stage and improve the success rate of planning.By introducing attitude correction,the reachability of the global control points is improved,and the moving trajectory of the robot is optimized.Finally,considering that the solution of motion planning problem must be based on accurate environmental information,in order to solve the problem that a single robot cannot accurately obtain field information due to its limited sensing ability,the team information fusion method adopted by robots in actual football matches is studied,and the overall verification of the planning scheme in this paper is conducted based on the entity NAO robot.In order to fully explain the effectiveness and progress of the work of this paper.