DETERMINING THE PATH SEARCH GRAPH AND FINDING A COLLISION-FREE PATH BY THE MODIFIED A* ALGORITHM FOR A 5-LINK CLOSED CHAIN
Qing Xue, Yung-Ping Chien · Applied Artificial Intelligence · 1995
Two tightly coordinated 2-link planar manipulators and the straight line between their two bases can be considered as a 5-link closed chain. Since the coordination of robot manipulators has broad applications in manufacturing, hazardous material handling, undersea operation, and space exploration, automatic collision-free path planning for a 5-link closed chain is an important unsolved engineering problem. This paper describes a collision-free path planning algorithm for a 5-link closed chain with revolute joints. In the algorithm the 5-link closed chain is first represented by a path search graph that is built on the basis of the concept of the newly developed C subspace model. Subsequently, a collision-free path is searched upon the graph by the modified A* algorithm. The significance of this path planning algorithm is its convergence and efficiency. The convergence is guaranteed by the C subspace model, which constructs unique mapping between the planned path in C subspaces and that in the world space. The way we build the path search graph and evaluation functions of the A* algorithm is designed to increase the search speed and to preserve the maneuverability of the 5-link closed chain.