Applying Means-Ends Analysis to Spatial Planning

Abstract

Currently known methods for robot planning fall far behind human capabilities: they require approximations of shapes, and they cannot generate plans which involve moving obstacles to clear a path for the moving object. In this paper, we explore the hypothesis that means-ends analysis based on a world model involving mental imagery allows more human-like solutions. Our method is based on a novel way of representing planning constraints which makes it possible to incrementally generate the symbolic representations for means-ends planning using only imagery operations. 1 Spatial Planning Problems The problem of planning a sequence of actions to move objects into a desired configuration is fundamental to intelligent robotics. Path planning algorithms have reached