Generating robot action plans by means of an heuristic search

Abstract

A planning concept is proposed which meets the requirements of flexibility and simplicity of the resulting plan structures and efficiency of replanning. The concept is based on a heuristic search process (branch-and-bound) by which a plan generation tree containing (as nodes) the most promising solution graphs is partly developed. This is achieved by a heuristic evaluation function which estimates the quality of the intermediate plans during the plan synthesis phase. The concept is shown for the area of automatic assembly using a small assembly task as an example. A short description of a robot planning system which is based on the concept is included.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>