Real-time path planning of autonomous vehicles for unstructured road navigation

Abstract

This paper presents a path planning algorithm for autonomous vehicles to generate feasible and smooth maneuvers in unstructured environments. Our approach uses a graph-based search algorithm and a discrete kinematic vehicle model to find a primitive path that is non-holonomic, collision-free, and safe. The algorithm then improves the quality and smoothness of the identified path by making local shortcuts. This shortcut for smoothing the path is determined by connecting the internal state of the primitive path to the initial state based on the Pythagorean Hodograph (PH) cubic curve. The formulation of the PH cubic curves in the smoothing path provides arc-length parameterization of the curve in a closed form. We present examples and experimental results for the real time path planning in unstructured road from an implementation on an autonomous vehicle.