Multi-goal Path Planning For A Mobile Robot Indoor Environment

In recent years,with the development of intelligent mobile robots,research related to the mobile robots gradually become the focus.When people use mobile robots to solve practical problems,they often encounter situations where the robot needs to complete corresponding tasks between multiple goals,such as searching tasks,inspection tasks,handling tasks,etc.The previous path planning algorithms,such as graph-based algorithms and cluster optimization algorithms,can solve multi-goal path planning problems,but the completeness of these algorithms are limited by the discretization of the state space of mobile robots,meanwhile,these algorithms have poor performance in high-dimensional state space.In order to solve the limitations of traditional algorithms when facing multi-goal path planning problems,this thesis proposes two multi-goal path planning methods based on RRT*algorithm.The first method uses the idea of multi-tree extension and improves the related functions in the basic RRT*algorithm to suit multi-tree expansion.At the same time,this thesis gives the expansion strategy,connection strategy and important parameters used in multi-tree expansion with RRT*.The second method uses the idea of hierarchical planning to divide the entire multi-goal path planning task into two parts:exploration layer and planning layer.The exploration layer explore the free space continuously with multiple RRT*trees,and the planning layer uses the information of the random tree in the exploration layer to create a distance exploration square matrix,meanwhile,this thesis proposes a method that can work out a shortest path for multi-goal path planning problem based on the Prim minimum spanning tree.After the analysis of the simulation results and the experimental results in indoor environment with ROS platform,the proposed two methods can work out the traversal path for multiple goals.Moreover,the mobile robot can navigate this path to accomplish multi-goal path planning task,which indicates that the proposed two methods can apply to the practical application.