Research On Path Planning Method For Tracked Unmanned Vehicles

With the rapid development of artificial intelligence and computer technology,unmanned ground vehicles(UGVs)have gradually become one of the important areas of artificial intelligence applications.Driverless technology is the core of UGVs.It is a comprehensive frontier science and technology which integrates computer,mathematics,control,information,surveying and mapping and other disciplines.UGVs are mainly divided into wheeled vehicles and tracked vehicles.At present,more research is on wheeled vehicles,such as Baidu driverless car and Google driverless vehicle.Wheeled unmanned vehicles are mainly used in civilian areas such as life and enterprises,while tracked unmanned vehicles are relatively rare,mainly used in military,fire,earthquake relief and other fields.The UGV technology mainly acquires environmental obstacle information through a large number of sensors,and then independently performs path planning to make motion decisions,so that the vehicle can reach the target position safely and quickly,thereby realizing the motion planning navigation control of the ground unmanned vehicle.Due to the characteristics and needs of its own structure,crawler vehicles are not the same as wheeled vehicles.Therefore,the path planning problem of tracked unmanned vehicles needs to be solved urgently to realize the practical application of tracked unmanned vehicles in military,industrial,firefighting and disaster relief fields.This paper will focus on the research of tracked unmanned vehicle.The main contents include the following aspects:(1)A vehicle global path planning method with known environmental information is proposed.Considering the mobile robot vehicle as a mass point,the grid decomposition method is used to model the environment of the map information of the mobile robot vehicle,and the mathematical model of the path planning problem is constructed without considering the motion state of the vehicle.Then,the traditional ant colony algorithm is improved to improve the accuracy and efficiency of path search.The global path planning method based on improved ant colony algorithm with the shortest path as the optimization goal is proposed.Finally,the global path planning method of mobile robot based on improved ant colony algorithm is simulated by MATLAB,and the feasibility and effectiveness of the method are verified.(2)A vehicle local path planning method with unknown environmental information is proposed.Considering the mobile robot vehicle as a mass point,without considering the motion state of the vehicle,firstly,the theory of rolling window algorithm for local path planning is outlined and mathematically modeled.Then,based on the principle of improved ant colony algorithm and A*algorithm,a local path planning method based on improved ant colony-A*hybrid algorithm with the shortest path as the optimization goal is proposed.The method improves the convergence speed and search efficiency of the algorithm.Finally,the MATLAB is used to simulate the local path planning method of mobile robot based on improved ant colony-A*hybrid algorithm,and the feasibility and real-time performance of the method are verified.(3)A path planning method based on tracked vehicle center steering is proposed.The tracked vehicle is regarded as a rigid body.For the differential steering principle of the tracked vehicle,the motion state of the tracked vehicle is analyzed,and the relationship between the track speed difference between the two sides and the steering of the vehicle is studied,and then the kinematics of the tracked vehicle is performed.Then,considering the motion state of the tracked vehicle,the prediction calculation method of the tracked vehicle steering time is proposed.Finally,the shortest travel time is used as the optimization target.Based on the proposed classical path planning method with the shortest path as the optimization goal,a path planning method based on the track steering of unmanned vehicles is proposed,which lays a theoretical foundation for the navigation control of crawler-type unmanned vehicles.(4)Experiments were carried out on the proposed path planning method for crawler-type unmanned vehicles.Firstly,the architecture of the experimental platform G1-RAS crawler robot car is summarized.The accuracy of the tracked vehicle steering time prediction calculation method is verified on the platform.Then the experimental research on the path planning method of the crawler unmanned vehicle is carried out.Taking the typical scene of the ordinary map and the tunnel map scene as the experimental environment,the feasibility and effectiveness of the global path planning method of the crawler unmanned vehicle in the known environment and the local path planning method of the crawler unmanned vehicle in the unknown environment are verified respectively.Experiments show that the path planning method based on the tracked unmanned vehicle center steering with the shortest travel time as the optimization goal can ensure the minimum total travel time and steering times under the premise of the shortest path.