Abstract
An artificial potential method has two problems of local minimum and computational cost. In order to resolve such problems, this paper proposes the utilization of a spline surface and a domain reduction method. A spline surface that interpolates arbitrary shaped boundaries is used as an artificial potential to guide a robot for global motion planning of a mobile robot. The domain reduction method resolves the local minimum problem by reducing a domain that maps on a free space in a configuration space. The spline surface is generated over the free space by assigning appropriate boundary conditions to the boundary of the free space. A moving obstacle is represented as an inner boundary of the domain. Some simulation experiments are applied to some complicated environments. From these experiments, this paper shows that a mobile robot can move to a target position selecting a wait action and an evasive action automatically in order to avoid a moving obstacle in a maze. Additionally, this paper shows that the proposed methods are applicable to a wall obstacle with arbitrary shaped boundaries.
