An admittance adaptive force feedback device and its interaction stability involving coupling with humans and uncertain environments

Abstract

Admittance control is a widely used approach for ensuring compliant robot behavior in physical human–robot interaction (pHRI) tasks. The selection of admittance parameters is crucial, as it directly affects interaction stability. However, this process becomes challenging when the robot’s interaction with objects involves not only the human hand but also the task environment. This is because the task environment is often unknown and hard to predict while the models of the human hand have been extensively studied in previous work. To address this issue, this paper proposes an admittance adaptive algorithm that ensures stability in human–robot-environment interaction tasks. This algorithm can adjust damping online without prior information about environments. Specifically, we consider the coupling between the robot, human hand, and task environment, treating them as a whole to analyze interaction stability and construct an energy function. Then, based on the energy function, a passive observer is designed to monitor unstable behaviors during the interaction process. Finally, the algorithm adjusts the damping online based on the observed values. The algorithm was experimentally validated using a custom admittance force feedback device. Experimental results indicate that the algorithm can ensure interaction stability without prior information about environments. In the experiment of writing letters, compared to a constant-parameter admittance controller, the algorithm reduces operator effort while maintaining stability.