Improving Robotic Force Control Performance in Devices With Force Measurement and Modeling Error

This paper investigates methods for improving robot force control performance in situations where force measurement error is unavoidable, such as exoskeletons, low-cost sensors, and flexible robotics. When using force control strategies like Impedance control (IC) and Admittance control (AC) in these situations, we propose that it is important to consider their inherent robustness to either force measurement error or modelling error. Improper choice of controller can lead to increased error, undesired interaction behaviour, and even instability. Furthermore, we propose that interpolation with Unified Interaction control (UIC) can trade-off these two robustness properties and, in some cases, can improve performance. A case study using a low-cost force sensitive resistor with less accuracy than traditional force/torque sensors is conducted. The results illustrate that, even when force measurements do not reach the accuracy of traditional sensors, an accurate rendering of the desired impedance can be achieved by careful selection of the controller.