Path Planning And Driving Behavior Decision Of Intelligent Trackless Rubber-Tyred Vehicle In Confined Space Of Deep Well

Coal production is high-risk and hard.Using intelligent equipment or robots to participate in or replace workers in coal mining work can not only improve coal mining safety but also improve coal production efficiency.As an important auxiliary transportation equipment for mining,trackless rubber-tyred vehicles are one of the main objects of intelligent research and implementation.Therefore,the differences and reasons between underground unmanned driving and ground unmanned driving are analyzed by referring to the relevant methods of ground robots and unmanned vehicles through on-site research.Based on the characteristics of deep confined space underground and the working requirements of underground trackless rubber-tyred vehicles,the intelligent decision-making methods of intelligent trackless rubber tire vehicles are studied from three aspects,i.e.,working environment modeling,obstacle avoidance path planning,and driving behavior decision-making.The specific work is summarized as follows:1.According to the working requirements of the coal mine environment and underground auxiliary transportation equipment and the research status of ground unmanned vehicles,the key technologies of intelligent trackless rubber-tyred vehicles are analyzed and the unmanned driving system architecture is designed.Firstly,the unmanned driving system architecture for intelligent transformation of trackless rubber-tyred vehicles is analyzed and discussed from three main aspects,i.e.,vehicle environmental perception,intelligent decision-making,and autonomous operation control.Secondly,based on the work requirements of trackless rubber tire vehicles and the main differences between them and the key technologies of ground unmanned vehicles,the key technologies of the intelligent decision-making system were designed from work environment modeling,obstacle avoidance planning,and driving behavior decision-making.The overall architecture and intelligent decision-making process of the intelligent trackless rubber-tyred vehicle decision-making system were built.2.Considering the obstacles threat from multiple planes,e.g.,the roadway floor,roof,and side walls,work environment modeling method based on risk grid and dimension reduction is proposed.Firstly,the basic grid map uses Lidar perception information to project obstacle information from multiple planes onto a two-dimensional grid map for storage,while the risk grid map combines the perception information of binocular cameras to assign grid risk information.Secondly,in order to remove redundant information and retain basic features available for path planning,the dimension redundant framework is used to reduce the map dimension.Thirdly,to compensate for the shortcomings of this dimensionality reduction framework,a reverse automaton is proposed to solve the problem of some key points being unable to participate in path planning after dimensionality reduction,and a Bezier curve smoothing method based on slope interpolation is proposed to solve the collision problem of smooth paths caused by fewer planned path points after dimensionality reduction.The feasibility verification of environment modeling and map dimensionality reduction framework using A* algorithm shows that the algorithm can plan safe paths faster.3.In order to solve the problem of the high uncertainty of the underground environment,which makes it difficult to achieve optimal path planning,an obstacle avoidance path planning method based on risk-grid particle swarm optimization for the trackless rubber-tyred vehicle in confined spaces of the deep well is proposed.Firstly,number the grids of the grid map to transform the path planning problem into a permutation and combination problem.Secondly,the velocity/position iteration formula of the particle swarm optimization algorithm has been improved,with the velocity term representing one path.The inertia weight and cognitive factor are compared with the replacement rate to replace the path points in the path,achieving population diversity.The position term represents the fitness of the current path.Then,a fitness function is designed with the goal of path length cost and path risk cost,and risk factors are used to balance the planning proportions of these two costs.Compared with four advanced path planning methods based on evolutionary computation,the results show that the proposed method has higher efficiency and progressiveness.4.In order to avoid the impact of sudden and dynamic obstacles on safe driving of vehicles,a driving behavior decision-making method based on multiple driving modes and safety assessment is proposed.Firstly,to solve the problems of low underground illumination and excessive dust and mist,which make it difficult for sensors to fully perform,a hierarchical multi-source perception data management fusion method is proposed.The multi-source perception data is layered and prioritized according to the warning layer,target layer,and road condition layer,and then three driving modes of trackless rubber wheeled vehicles are proposed.Based on the depth of the obstacle field,the current driving mode of the vehicle is determined.Secondly,two safety evaluation standards are introduced to improve the driving behavior decision-making performance of intelligent trackless rubber-tyred vehicles.Combining the obstacle avoidance planning method,path smoothing method,and driving behavior decision-making method in this dissertation,two sets of simulation experiments were designed.The simulation results showed that the vehicle can make accurate driving behavior decisions on static obstacles,lateral motion dynamic obstacles,and longitudinal motion dynamic obstacles,and achieve driving behavior decisions i.e.,turning,changing lanes,and returning to the initial path,etc.