Curvature continuous path smoothing based on cubic Bezier curves for car-like vehicles

This paper focuses on the generation of a feasible path for a car-like vehicle working in a large scientific facility environment. Considering the curvature continuity and maximum curvature constraints, a novel path-smoothing algorithm is proposed in terms of cubic Bezier curves. In the algorithm, Bezier turn and Bezier path are respectively developed. Bezier turn is firstly designed to connect two arbitrary configurations. Then Bezier path could be obtained by using Bezier turns to fit a sequence of objective points offered by a collision-avoidance planner. With guidance of the algorithm, the vehicle could reach the target point with predefined direction. Simulation experiments are well conducted to demonstrate that the planned path is practical, and is accordant with human expert experience.