Robot Path Planning by Means of Binary Algebra
Nicholas J. Yannoulakis, Richard A. Wysk · 1991
Abstract The issues of robot path planning and collision avoidance have been addressed extensively in literature. This paper examines these problems for a specific domain: that of a rectangular gripper moving in a two-dimensional space of iso-oriented rectangular obstacles. The free space is represented by a set of binary strings. These strings are manipulated through Boolean algebra to yield the largest convex areas in which the gripper can move without colliding with any obstacles. The largest convex areas then become the nodes of two networks (one for each of the orthogonal orientations of the gripper), the arcs of which indicate the possible gripper paths and roll locations. Based on these networks, a search procedure can produce the shortest path from a start to a target position. This algorithm is easy to implement on a PC and will go from a workspace representation to robot control commands.