Research On Autonomous Navigation System Of Mobile Robot In Indoor Environment With Multiple Obstacles

Autonomous navigation in the working environment is the premise for mobile robots to complete other complex tasks.Autonomous navigation includes simultaneous localization and mapping(SLAM)and path planning.With the increase of the number of obstacles in the indoor environment,the common mobile robot autonomous navigation system will suffer from poor real-time performance,low accuracy of localization and mapping,and low success rate of path planning.Therefore,the research of autonomous navigation system applicable to indoor multi-obstacle environment has very important theoretical significance and application value.First of all,in this thesis,the research status of mobile robot navigation subject at home and abroad is elaborated,the common sensors and related models in autonomous navigation system are discussed,the SLAM technology and path planning technology involved are analyzed.On this basis,the overall scheme of mobile robot autonomous navigation system suitable for indoor multi-obstacle environment is designed.Then,the state estimation and the data association in the SLAM operation of mobile robot are studied deeply in this thesis.Fast SLAM algorithm is selected to perform SLAM state estimation in the indoor multi-obstacle environment because of its low computational complexity and high estimation accuracy.At the same time,the joint compatibility branch and bound(JCBB)algorithm is improved by setting the adaptive local data association region and using the adaptive Gaussian mixture clustering grouping method,so that the algorithm can get high association accuracy with less association time when performing SLAM data association in the indoor multi-obstacle environment.The experimental results show that,in the multi-obstacle environment,the SLAM operation of mobile robot based on Fast SLAM algorithm and the improved JCBB algorithm has high localization and mapping accuracy,high execution efficiency.Subsequently,the artificial potential field(APF)algorithm is improved in this thesis,in order to solve the goal non-reachable problem and the falling into local minimum point problem,which often occur when mobile robot using this algorithm to perform path planning operation in the indoor multi-obstacle environment.The goal non-reachable problem is solved by introducing the distance between the robot and the goal processed by arc tangent function into the repulsive field function.And the fallinginto local minimum point problem is also solved by breaking the balance of forces with the compulsive disturbing force constructed on the basis of attractive force.The experimental results show that,in the multi-obstacle environment,the improved APF algorithm has a higher success rate of path planning,and the quality of the planned path is also better.In the end,based on the algorithms researched in the subject,the mobile robot autonomous navigation system designed in this thesis is implemented with the software platform——robot operating system(ROS),and the hardware platform built by relevant equipment.Besides,the SLAM experiment and path planning experiment are carried out in the real indoor multi-obstacle environment.The experimental results confirm the effectiveness of the autonomous navigation system designed in this thesis.