Design And Implementation Of A Mobile Robot Autonomous Navigation System In Indoor Environments

The current demand for fully autonomous robot applications from indoor scenes is increasing.Representative products include household sweeping robots,indoor distribution robots working in environments such as hotels,shopping malls,restaurants,and hospitals,and special-purpose ones for inspection and security within the plant.The purpose of this paper is to design and implement an autonomous navigation system for mobile robots in an indoor environment,and develop a fully autonomous mobile robot that integrates a robot chassis,perception system and navigation system.Real-time acquisition and processing of sensor information can be achieved.Through autonomous exploration,reliable construction of indoor maps can be completed,which provides support for autonomous path planning of mobile robots.In this paper,an indoor robot autonomous navigation system under the framework of the Robot Operation System(ROS)is designed and implemented.The system is mainly divided into two parts:autonomous environment exploration and path planning,which mainly solves the problem of exploration and path planning in the indoor environment.This paper implements environment exploration based on Partially Observable Markov Decision Process(POMDP).The information entropy is used as the reward function,and the observation model and state transition function are contructed.On this basis,the Monte Carlo Tree Search method is utilized to approximate the optimal solution of POMDP.Simultaneously,the Rao-Blackwellized particle filter algorithm is applied to perform simultaneous localization and mapping(SLAM).For the problem of indoor robot path planning,this paper proposes an improved global path planning algorithm based on A*algorithm,which is based on the environment map constructed by the POMDP algorithm.Moreover,the dynamic window method for local path planning is improved,which makes the path closer to the one obtained by the global path planning.In order to achieve the function of multi-point navigation of the robot and complete specific target tasks,the autonomous cruise function of the mobile robot is added.The cruise route setting is completed by collecting coordinate information in the environment map in advance,thereby realizing the navigation function according to the set sequence waypoints.Finally,the functional test of the autonomous navigation system designed and implemented is carried out.In the performance test of the mobile robot environment exploration algorithm,the environment exploration algorithm designed in this paper is compared with the boundary exploration method that only considers the exploration benefits.It is found that the proposed algorithm improves the positioning accuracy compared with the existing boundary exploration method.At the same time,in the process of exploration,the robot successfully constructs a reliable environment map.In the functional test of path planning,each subsystem can operate correctly and steadily,and the design goals are achieved.