Search guided by skill in motion planning using dynamic programming

摘要

Basically, two ways of motion planning are considered. One is a mathematical programming method and the other is the use of human skill. This paper presents a new efficiently planning method which incorporates human skill into dynamic programming (DP). DP is a useful mathematical method for motion planning but its computation time and memory are enormous. For its efficient computation, a motion planning method using DP is formulated as a search in a configuration space of a robot, where joint velocities and values of the objective function are attributes of a configuration. Then, a heuristic search method called A* algorithm developed in artificial intelligence is introduced to the search. Human skill can be represented as a heuristic function. Thus DP and human skill are combined. This method is applied to example tasks to explain the proposed method.