Dynamics calibration and real-time state estimation of a redundant flexible joint robot based on encoders and gyroscopes

Abstract

We show the identification of dynamic parameters for a redundant flexible joint robot with a flexible bearing in the first joint. We compare a standard model (without flexible bearing) and an extended model (with flexible bearing). The bearing leads to distortions of the link velocities measured by gyroscopes. Joint positions measured by encoders. We show how to pack the calibration problem into the sparse least-squares on manifolds toolkit (SLOM) and how easy it is to adapt to a new description – our extended model. In the second part we implement a Kalman filter on a microcontroller to estimate the 16 states in real-time using the extended model.