Computer simulation of sensor-based robot collision avoidance in an unknown environment

Abstract

SUMMARY: Computer simulation is a major tool in validation of robot motion planning systems, since, on the one hand, underlying theory of algorithms typically requires questionable assumptions and simplifications, and, on the other hand, experiments with hardware are necessarily limited by available resources and time. This is especially true when the motion planning system in question is based on sensor feedback and the generated trajectory is, therefore, unpredictable. This paper describes a simulation system ROPAS (for RObot PAth Simulation) for testing one approach — called Dynmic Path Planning (DPP) — to sensor-based robot collision avoidance in an environment with unknown obstacles. Using real time graphics animation of the motion planning system, the user can simulate the behavior of an autonomous vehicle or a robot arm manipulator with a fixed base. The overall structure of the system is described, and examples are presented.