Integral resonant filter-based chattering reduction in switched human-robot interaction with hybrid admittance-adaptive impedance control

Physical human-robot interaction (pHRI) is a type of interaction that implies that robots should demonstrate dynamic and compliant behavior that is dependent on the interaction environment. A prominent solution is compliant control, which includes either Admittance Control in the case of soft contact or Impedance Control for stiff contact. However, if the nature of the environment is unknown or dynamic, then a hybrid control scheme that switches between the admittance and impedance regimes is adopted. However, switching from admittance to impedance and vice versa results in discontinuity in the control signal. The discontinuity inherent in the abrupt switching mechanism causes high-frequency oscillations (chattering) in the control signal. The effect of this phenomenon is poor performance of the system, mechanical stress, and high energy consumption. To effectively attenuate this chattering and its negative effects, this work proposes an Integral Resonant Control (IRC) command filter, which is incorporated into the switching loop of the hybrid admittance and adaptive impedance control. The IRC filter, normally employed to damp structural vibrations, is designed here to act as a command filter for robust signal smoothing by damping the high-frequency components of the control signal introduced by the switching action. The proposed hybrid controller is compared to a standard hysteresis (dead-zone) chattering mitigation strategy. Comparative simulation results indicate that the proposed filtering technique can reduce the Chattering Index on average by over 76% for both joints of a planar rehabilitation robot while maintaining trajectory tracking accuracy. The results obtained indicate that the proposed switched hybrid control is Lyapunov stable and can be employed to achieve effective, energy-efficient, and compliant robotic rehabilitation in uncertain environments.