Research On Mapping And Path-planning Method Of Mobile Robot

The mobile robot is a strategic product that integrates mechanical manufacturing technology,computer technology,electromechanical integration technology,electronic information technology,multi-sensor data fusion,and artificial intelligence.In recent years,mobile robots have emerged in restaurants,hotels,stores,warehouse management,and personal homes.Environmental construction and path planning are key technologies for mobile robot navigation.They are responsible for mapping the environment where the mobile robot is located as a map,providing environmental support for navigation,and planning a safe collision-free path from the current position of the mobile robot to the target position in the environment.In order to improve the autonomous navigation performance of mobile robots,this topic researches the simultaneous localization and mapping(SLAM)and path planning technology of mobile robots.The main contents include the following aspects:Firstly,designing the upper navigation system of the mobile robot and its communication with the underlying hardware platform for the existing hardware platform so that the mobile robot can realize the autonomous navigation function;the laser radar point cloud is optimized to make it closer to the real environment.And designed a humancomputer interaction client to achieve remote control of the mobile robot.Secondly,we study the method of mapping,and propose a mapping method based on constructing submap.First,a heuristic search and localization algorithm is proposed to speed up the scan match.Second,the probability grid is updated by creating a look-up table,and constructing submap.Then,the optimal historical subgraph is found through the fast filtering algorithm,which provides a reference for the construction of new submap.Combined with a heuristic search and location algorithm,the pose relationship between subgraphs is determined,and the map creation is completed.This method reduces the error accumulation and improves the accuracy and consistency of the map with the environment.Thirdly,we study the path planning of mobile robots and proposes a method of using the reinforcement learning method to modify the global path.This method combines the Q-learning algorithm with the cost map to modify the global path.The cost information of the cost map is used to create the state space,action space,reward function and value function.The mobile robots are trained in complex scenes to obtain a Q value table that can be used in mobile robots.Finally,by using mobile robot platform design experiment,the feasibility of this paper's mapping algorithm and path planning algorithm is verified.