Autonomous Underground Vehicle Video Capture,Localization And Obstacle Avoidance Based On Industrial Internet

As the main energy source for many countries,coal has been universally emphasized.In recent twenty years,with the development of information network technology and automation control technology,the autonomous driving of underground vehicle based on the industrial internet will become the most important key part of underground unmanned technology.Considering the unmanned underground vehicle as research object,this thesis focuses on the real-time video capture,mapping,self-localization and real-time obstacle avoidance technologies in the local area of coal mine which is covered by the underground industrial internet.The main contributions are as follow:(1)The network architecture design of underground industrial internet based on the optical fiber ring network and wireless Mesh network is analyzed.The advatanges of optical fiber ring network and wireless Mesh network is also analyzed.Based on this network architecture design,the entire industrial internet can be divided into 4 layers.By using the optical fiber ring network as the backbone transmission network and wireless Mesh network as local area network,this kind of network architecture design can increase the coverage rate of industrial internet in the local aera of coal mine like fully mechanical coal mining area.(2)An underground vehicle video capture and driving control platform based on the Android system and coal mine industrial internet is proposed and designed.Through the platform,we can realize vehicle's surrounding environment video capture and driving control in the local area of coal mine.An experimental vehicle is assembled and various functions of platform are verified based on an Android device and this experimental vehicle.(3)According to the special environment of coal mine,a new method of vehicle localization in the underground area is proposed.The landmarks which contain barcode are deployed on the both sides of local area tunnel of coal mine.The landmarks can be localized and recognized by the on-board cameras.By combining three-dimensional positioning of space model,we realized the self-localization of underground vehicle and proved its feasibility in the actual real tunnel environment experiment.At the same time,the map of the tunnel in the experimental environment is drawn by using the on-board rotate radar and ROS system,(4)Based on the hierarchical fuzzy control algorithm,a real-time obstacle avoidance control method is proposed.By using the Mamdani fuzzy control method,we designed the specific fuzzy rules and proposed hierarchical fuzzy controller.We verified the obstacle avoidance method through the Matlab simulation and actual vehicle obstacle avoidance test and proved this method can effectively reduce the number of judgment during the obstacle avoidance process,thus reducing theamount of calculation and improve the accuracy of system.Based on the network architecture of underground industrial internet,the environment video capture,self-localization,environmental map building and real-time obstacle avoidance technologies for underground vehicle are theoretical studied and pratical verified in this thesis.Research results show that above work has important vaule to the realization of autonomous driving and monitoring for underground vehicle in the local area.