Vehicle Path Planning Method For Unstructured Operation Area In Opencast Mines

Integrating advanced sensor technology,high-precision positioning technology,decision-making technology,and modern control technology,autonomous driving technology effectively improves vehicle operation efficiency,improves driving safety,and reduces the driving burden on drivers.Hence,it becomes the trend of automotive technology development.The opencast mines are the excellent scenario for applying autonomous driving technology due to their low transportation speed,and low interference from pedestrians and other vehicles.In this paper,the path planning method for autonomous driving mining vehicles in common unstructured operation areas of opencast mines is studied.The main research contents are as follows:(1)Improved Hybrid A* path search method based on equal-step hierarchical expansion.The unstructured operation area in opencast mines is an open environment and lies numerous irregular obstacles.Aiming at the problem that the naive Hybrid A* algorithm has low path search efficiency and poor collision safety in this scenario,this paper proposes an equal-step hierarchical expansion method and constructs a new motion primitive for the expansion of the search tree,which can greatly reduce the expansion times and consequently improve the efficiency of path search.Based on the proposed motion primitive,the corresponding path search policy is designed,and an additional collision penalty term is introduced into the cost function to improve the collision safety of the searching path.In addition,a real-time heuristic value calculation method based on improved A* is designed to solve the problems of the large and non-closed-boundary searching areas,which ensures the rationality of heuristic value and calculation efficiency.(2)The initial path optimization method based on quadratic programming.Aiming at the problem of unnecessary turning and discontinuous curvature of the planned path,a path optimization method based on quadratic programming is proposed.In the objective function,the smoothness of the path,the uniformity of the waypoints,and the closeness to the initial path are considered.In the constraints,the endpoint-fixed constraint,the box position constraint of each waypoint,and the curvature maximum constraint are included.As for the box position constraint,a fast generation method is designed.In addition,a mathematical method for curvature calculation is proposed to express the maximum curvature constraints.Then these constraints are linearized and slacked to ensure solving efficiency and feasibility.In summary,this method can efficiently smooth the planned path and also enhance path safety.(3)Path interpolation method based on the improved cubic polynomial curve.To make the waypoints in the final path uniform,and avoid distortion in the interpolation process,an improved path interpolation method based on the cubic polynomial curve is proposed.Firstly,the heading angle of each waypoint is calculated using the coordinate of two adjacent waypoints.Then,the cubic polynomial curve interpolation is performed between the two adjacent points.In this step,the coordinate system is rotated to avoid solving failure due to infinite derivatives at the endpoints,and consequently,the calculation is simplified and the efficiency is improved.(4)Simulation test using MATLAB and field test using ROS(Robotics Operation System).To verify the effectiveness of the proposed path search method,path optimization method,and path interpolation method,simulation experiments are conducted in MATLAB environment,and field tests are conducted for the algorithm programmed by C++and executed in ROS.The test results show that the proposed method can plan a smooth,safe,efficient,and kinematics-constraint-satisfaction path for the autonomous driving mining vehicle,and be effectively applied to the unstructured operation area in opencast mines.