High-Precision Tracking Control of Modular Dielectric Elastomer Actuators Through Kalman Filter Base Self-Adaptive Direct Inverse Hysteresis Compensation

This work proposes a Kalman filter based direct inverse hysteresis compensation approach for modularized dielectric elastomer actuators (DEAs) with rate-dependent viscoelasticity, mechanical vibration, and individual variation. To this end, we firstly investigate the dynamics of three modularized DEAs, including viscoelasticity, mechanical vibration, and their individual variation. Then, based on a rate-independent phenomenological hysteresis model, a direct inverse hysteresis compensation (DIHC) is established to eliminate the hysteresis effect. Further, a feedback controller is adopted to remove the creep and model uncertainty. Lastly, a Kalman filter is employed to automatically adjust the parameter of the DIHC, capable of achieving high-precision tracking control of modular DEAs within a wide frequency range. To verify the effectiveness of our control strategy, three modularized DEAs are employed to track periodical trajectories with different frequencies (in a range of 0.1Hz to 10Hz). The experimental results demonstrate that the Kalman filter based controller can fully eliminate the rate-dependent viscoelasticity, mechanical vibration, and individual difference, achieving high-precision tracking control of modular DEAs in a wide frequency range. The maximum error and root mean square error are reduced from 20.16% and 11.44% to 9.87% and 5.22%, respectively.