Research Of Autonomous Navigation Control Methods For AGV Based On Turtlebot

With the rapid development of technology and the increase of labor cost,the level of intelligence in the process of production and transportation is increasing.AGV(Automated Guided Vehicle)is a kind of intelligent transportation vehicle,which integrates environmental perception,planning decision and intelligent control,and navigation is the core technology of AGV vehicle,which determines the level of intelligence.In this paper,the autonomous navigation control methods for AGV are studied to realize obstacle avoidance and line tracking based on the Turtlebot,and a hybrid navigation control method is proposed.Firstly,the depth image obtained by Kinect and processed to laser data as the input of SLAM algorithm.Based on the Adaptive Monte Carlo localization and RBPF particle filter,the environment map is constructed for the study on obstacle avoidance and path planning.Secondly,according that global path planning algorithm A* can obtain the global optimal path and local path planning algorithm DWA(Dynamic Window Approach)can perform real-time obstacle avoidance,combining the advantages of two algorithms,based on A* and DWA algorithm,hybrid dynamic path planning is achieved in the grid map,laying the foundation for the realization of hybrid navigation.Thirdly,it is found that navigation line cannot be detected stability while realizing line tracking besed on machine vision.Thus a new method is proposed that the RGB image is converted to the HSV space for threshold segmentation,and then the Canny edge detection results are taken as logic,and then the Hough transform is used to fit the detection of the navigation trajectory.Experimental results show that the proposed method can improve the stability of navigation trajectory detection.Fourthly,based on the realization of obstacle avoidance and line tracking,a hybrid navigation control method is proposed.In the case of the obstacles that don't affect the line tracking,the navigation area is divided into two parts,and then the K-means algorithm is used to cluster the laser data,thus to determine whether to avoid obstacles.In the case of the situation that navigation line is lost after the obstacle avoidance,it is increased the function of automatic acquisition of navigation line to improve the hybrid navigation control algorithm.Finally,based on the Turtlebot,the map built by SLAM,the obstacle avoidance and path planning,the line tracking navigation,the line tracking and the obstacle avoidance hybrid navigation are carried out respectively to verify the reliability of the navigation algorithm and achieve automatic navigation of AGV vehicle in complex task.