Planning continuous-curvature paths for car-like robots

摘要

This paper presents a continuous-curvature path planner (CCPP) for a car-like robot. Previous collision-free path planners for car-like robots compute paths made up of straight segments connected with tangential circular arcs. The curvature of this type of path is discontinuous so much so that if a car-like robot were to actually follow such a path, it would have to stop at each curvature discontinuity so as to reorient its front wheels. CCPP is one of the first planner to compute collision-free paths with continuous curvature profiles. These paths are made up of clothoid arcs, i.e. curves whose curvature is a linear function of their arc length. CCPP uses a general planning technique called the Ariadne's Clew algorithm. It is based upon two complementary functions: SEARCH and EXPLORE. EXPLORE builds an approximation of the region of the configuration space reachable from a start configuration by incrementally placing a set of reachable landmarks in the configuration space. SEARCH checks the existence of a solution path between a landmark newly placed and the goal configuration.