Reducing the Complexity of Robot's Scene for Faster Collision Detection

• Published in: Journal of intelligent & robotic systems Vol. 26; no. 1; pp. 79 - 89
• Main Authors: Nearchou, A C, Aspragathos, N A, Sofotassios, D P
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Nature B.V 01.09.1999
• Subjects: Algorithms; Complexity; Density; Manipulators; Obstacles; Robot arms; Robotics; Robots; Studies
• ISSN: 0921-0296; 1573-0409
• DOI: 10.1023/A:1008190406456

In many real-life industrial applications such as welding and painting, the hand tip of a robot manipulator must follow a desired Cartesian curve while its body avoids collisions with obstacles in its environment. Collision detection is an absolutely essential task for any robotic manipulators in order to operate safely and effectively in cluttered environments. A significant factor that influences the complexity of the collision detection problem is the obstacles' density, i.e., the total number of obstacles in the robot's environment. In this paper, a heuristic algorithm for approximating the collision detection problem into a simpler one is presented. The algorithm reduces the number of obstacles that must be examined during the robot's motion by applying efficient techniques from computational geometry. The algorithm runs in time O(max(n^sup 2^ log m, n m)) , and uses O(n^sup 2^ + m n) space; with n being the number of obstacles in the robot's workspace, and m the total number of obstacles' vertices. Both costs are worst-case bounds.[PUBLICATION ABSTRACT]