Simulated annealing based algorithm for smooth robot path planning with different kinematic constraints

Abstract

In this paper, we present a simulated annealing (SA) based algorithm for robot path planning. The kernel of our SA engine is based on Voronoi diagram and composite Bezier curve to obtain the shortest smooth path under given kinematic constraints. In our algorithm, a Voronoi diagram is constructed according to the global environment. The piecewise linear path in the Voronoi diagram which keeps away from the obstacles is obtained by performing Dijkstra's shortest path algorithm. The control points on the reference path are used to create the control variables for our SA engine. Our SA engine then updates the control variables to obtain the shortest composite Bezier curve path while satisfying given kinematic constraints. Experimental results on two maps containing sharp turns demonstrate the effectiveness of the proposed SA-based smooth path planning algorithm.