Research On SLAM And Path Planning Method For Mobile Robot

Automatic transportation is the basis of realizing automation and intelligence in modern logistics warehouse.Compared with the conveyor belt and the traditional AGV,the autonomous mobile robot has the ability to explore and decision-making in the face of the unknown environment,which has the advantages of flexible and reconfigurable environment.In this thesis,the simultaneous localization and mapping building(SLAM)and path planning of autonomous mobile robot were studied.,the main contents include:Firstly,the method of simultaneous localization and mapping has been studied.Established the system model of the mobile robot and the mathematical model of simultaneous localization and mapping,and analyzed the key problems of the system.Then,the EKF-SLAM and FAST-SLAM algorithms were analyzed and simulated respectively.Simulation results show that both algorithms can achieve better simulation results,but the FAST-SLAM algorithm has less error and more suitable to the actual environment.Secondly,Proposed an improved algorithm of social spider algorithm.In order to solve the local optimization problem of path planning,the social spider algorithm was improved.Using the standard function to evaluate the function of the improved social spider algorithm,the social spider algorithm and the particle swarm optimization algorithm,The results show that the improved spider social algorithm has strong searching ability,and the optimal results are close to the optimal value.Then,with the help of simulation experiment in MATLAB simulation environment,simulation results show that the improved algorithm of gregarious spider is feasible in the global path planning;its convergence is better than other intelligent optimization algorithms.At last,research on the simultaneous localization and map building method of autonomous mobile robot based on stereo vision.Design the autonomous mobile robot prototype based on RGB-D sensor in the environment of Robot Operating System(ROS),then using the depth image instead of laser radar to provide information on the 2D laser.On the basis of matching the RGB and Depth information of features,splicing the point cloud,the localization and map building under laboratory conditions at the same time have been realized.Prototype test results show that:the mobile robot system based on stereo vision can achieve the indoor localization and mapping.