A Method for Gear Impact Suppression In Torsion Torque Control of Two-Inertia Systems with Backlash

Abstract

Torsion torque control has been successfully applied in powertrains for automotive systems, and in robotics for improving the performance of load-side control and increasing backdrivability of robots. However its performance is severely compromised when backlash is present in the transmission, as the reengaging of the mechanical components causes a excessive impact torque in the system. Conventional gear impact suppression methods require the use of load-side encoders and backlash models, which could be costly and difficult to implement, particularly in robots. This paper proposes a gear impact suppression method based on a impact torque suppressor (ITS), which consists of scheduled torque-derivative-feedback directly to the joint torsional dynamics. The effectiveness of the proposed method is verified by simulation and experimental results.