Path Planning of Composite Trackless AGV Considering Map Preprocessing

With mounted robot arm, the compound trackless automatic guided vehicle (AGV) has strong flexibility and adaptability in industrial environments. However, due to the space constraints of the robot arm, the path planning in complex large scenes is hard to achieve high efficiency, and it is easy for falling into local minimum and stagnation. In this paper, a novel AGV path planning algorithm is proposed on the basis of map preprocessing to improve the planning efficiency and guarantee operating safety. First, by integrating the multiple constraints including the obstacles and robotic position/posture, the preprocessing method of the environmental map is proposed utilizing obstacle expansion and Delaunay triangulation. Then, to achieve better convergence and global optimization capacity, the global path searching is modified by (1) constructing the OpenList with priority queues; (2) exploring the adaptive rules for step size and dynamic weighted heuristic function. Finally, combining linearization and cubic Hermite interpolation method, the planned path is smoothed to enhance the movement stability and energy consumption rate. Simulation analysis and experimental results verify the feasibility, efficiency and superiority of the proposed path planning method.