IR-Observer

Real-time process forces are required for force control, for which force/torque sensors are used in the state of the art. Since this additional sensor technology leads to both higher costs and a higher probability of failure, it is desirable to replace it with an estimation of the contact force. So far, no state-of-the-art method has been able to achieve a sufficient quality of contact force estimation in industrial robots so that it can be used for mechanical production work. The aim of the research project is to investigate an optimal contact force estimation based on the moving horizon approach and the recently available joint-side position measurement. In contrast to classical disturbance observers (Kalman filter), the method is characterized by the fact that not only the instantaneous value, but a longer time window of the signals is used for state estimation, whereby a better and more robust estimate can be achieved. Moving horizon observers (MHE) are based on the idea of minimizing a cost functional that is defined on the basis of a moving window with a finite number of time steps.

The goal of the project is to replace force/torque sensors with a contact force estimation system with high accuracy, dynamics and robustness. The optimization problem of the disturbance observer should be solvable on typical industrial robot controllers without compromising the real-time requirements.