Robotic arm self-taught path planning using transportation model heuristics

Abstract

The objective of this experimental research was to determine whether or not an integrated vision and robotic arm system, using grey level detection and a transportation model heuristic, could plan and execute its own point-to-point navigation. The heuristic is the shortest-route algorithm which deals with finding the smallest possible distance to visit multiple locations only once. The only point taught by a human operator was a starting position and all other points within the workspace, to be visited by the robotic arm to perform tasks such as screw-fastening, adhesive application and detailed inspection of assemblies, were learned by the system.