A Network Optimization Approach to Three-Dimensional Collision-Free Robot Path Planning
Abstract This paper presents a network optimization approach to the generation of tool tip paths in three-dimensional space for robot manipulators working in the presence of obstacles. The developed algorithm relies on a graph structure enumeration of possible path segments between discrete points inside the workspace. An intelligent heuristic scheme is used to select a small searchspace over which the search for the optimal path is carried out. Paths which are optimal with respect to pre-determined objective functions based on robot kinematic and dynamic solutions are synthesized by applying Dijkstra’s minimum cost algorithm. Collision avoidance is checked for the tool tip as well as for the robot body. The methodology is robot independent and can be applied to any robot whose kinematics and dynamics can be solved. A computer program has been developed to implement this approach for three-axis manipulators. Results of the application of this scheme to some robots in this class are also presented.