Online Path Planning for Aircraft De-Icing Trucks with Dynamic Nonholonomic Constraints

Abstract

This paper presents a method for online path planning for large and heavy de-icing trucks in an airport environment using a continuous hybrid search tree with cost optimization and search heuristics. The vehicle’s dynamic nonholonomic kinematic constraints pose a challenge for traditional planners. The path is significantly constrained by both the minimum turning radius and the limitation on the rate of change of the steering angle. Since the truck requires a lot of space to change its orientation, the entire path needs to be planned in advance to reach the desired goal position and orientation. By building a gridless tree of kinematically feasible path segments and expanding it from the initial position to the goal, the algorithm ensures that a valid trajectory is found which is physically executable. Obstacle avoidance is done using cost maps supporting a dynamic robot footprint. To improve its performance, the tree expansion is guided by a heuristic function and explores multiple paths in parallel. The solution has been implemented within the Robot Operating System (ROS) framework and can run in near real-time on a low-power computing device on the vehicle. Validation of the numerous enhancements are demonstrated in simulation showing good performance.