Research on Local Path Planning Algorithm for Unmanned Vehicles

The local path planning algorithm of unmanned vehicle has certain requirements for the safety and real-time performance of obstacle avoidance, and the smoothness of obstacle avoidance path. In this paper, a local path planning algorithm based on discrete optimization is proposed, which uses cost function to evaluate the safety and smoothness of discretely generated candidate paths, and then obtains the local optimal path through the weighted calculation of each cost function. Aiming at the randomness of obstacles movement, a moving obstacles safety cost function is designed based on motion estimation combined with Gaussian convolution. Considering the curvature and its continuity of path, a path smoothness cost function is designed. A new coordinate transformation calculation method is adopted to convert the path from the s-ρ coordinate system to the earth Cartesian coordinate system, enhancing real-time performance. Finally, a PreScan / Matlab co-simulation and a real vehicle experiment on "Yuan Fei" unmanned vehicle experimental platform are both carried out. The results show that the path planning algorithm proposed not only enables the unmanned vehicle to safely and reasonably avoid the static and moving obstacles, but also fully meets the real-time requirements of local path planning algorithm. © 2020, Editorial Office of Journal of Building Structures. All right reserved.