Path Planning Of Mobile Robot Based On Global Map

Mobile robots,as an emerging research field,assist human beings to accomplish many tasks in industry,aerospace,defense and transportation.It is necessary to study the path planning algorithms and optimization problems in a variety of environments to perform these tasks.At present,many algorithms have many defects in solving the path planning problem,such as low planning efficiency,non-optimal path,and poor dynamic obstacle avoidance performance.This dissertation proposes a hybrid path planning method,which finds a global optimal path by RRT algorithm,uses DWA and TEB for local optimization and dynamic obstacle avoidance,respectively.The feasibility and effectiveness of the algorithm are verified by simulation and experiments on the ROS-based mobile robot.The main achievements can be concluded as follows.(1)In order to solve the problems what long time,incomplete map and accumulative error,because of the change of illumination and view angle when mobile robots construct map with SLAM.The global map of indoor environments is constructed by grids method combining aerial vision and image processing.(2)An improved algorithm is proposed to overcome the deficiencies of slow convergence rate,dense sampling nodes,complex and tortuous path in previous studies on RRT method.Firstly,finds an initial feasible path quickly in the biased region,to reduce the execution time and accelerate the convergence rate.Futhermore,the node rejection is used to remove high cost and invalid nodes,to generate a more promising tree and reduce the memory requirements.Finally,all the vertices of the initial path are traversed to determine the cost between the start point and the vertices,and whether to cross the obstacle or not.Then the extra nodes are pruned to optimize the path.The results show that this method is superior to the RRT and RRT*method and converges to shorter paths with fewer nodes and execution time.(3)In the aspect of local path planning,the principle and workflow of DWA and TEB are analyzed.The kinematics model and constraint function of mobile robot are expounded.Aiming at the problems of oscillations,stops and collision in the narrow channels for traditional DWA algorithm,the obstacle expansion processing and state classification strategy are introduced to ensure the safety of the mobile robot.Finally,both the DWA based on the two-wheel differential mobile robot model and the TEB based on the Ackermann steering mobile robot model can avoid the temporary obstacles effectively in the simulation experiment.(4)A hybrid path planning algorithm can compensate for the deficiencies of global or local algorithm.Firstly,the improved RRT is used to find a global optimal path in the global static environment.Secondly,DWA and TEB can track and optimize the global path for different mobile robot models,ensuring that mobile robots avoid dynamic obstacles quickly.Finally,on the ROS-based mobile robot,the hybrid algorithm is used to complete the path planning task in the indoor environment,avoid temporary obstacles and reach the target point.Overall,the hybrid path planning algorithm given in this dissertation has desirable comprehensive properties for mobile robot path planning problem,and with fast search speed,good optimization effect and strong adaptability to environmental changes.