Extended Kalman Filter Calibration Method Based on Lightweight Robotic Arm

Abstract

Most researches on position-based visual servo system pay more attention to control method rather than calibration. However, the impact of calibration on performance cannot be ignored, and its accuracy needs to be improved as much as possible. At the same time, some calibration methods rely heavily on high-precision sensors. In practical applications, some small devices such as lightweight robotic arm does not require high accuracy during the task, and it is not cost-effective to use high-precision sensors. Aiming at the above situation, a handeye calibration algorithm based on extended Kalman filter is designed to achieve the desired accuracy without relying on high-precision measuring devices. First, state equation and observation equation are constructed according to camera model and translation-rotation. Then use extended Kalman filter to eliminate accumulated error in state parameters. The result proves feasibility of the algorithm.