A new Software Driven external Sensor System for Industrial Robots

For decades, laser tracker and total stations have been the state of the art to measure externally the position disturbances in robotic systems. High system costs limit their usage for control systems in common production machines. We present details for an alternative software-driven approach. First, we derive a metrological error model for a new self-referencing, high-precision photogrammetry sensor system. Second, we propose a heuristic software approach, which combines an optical simulation pipeline, with motion planning and camera placement to achieve the best possible accuracy. Finally, we outline the hardware implementation and integration in a closed loop control system.