Adaptive Passivation of Admittance Controllers by Bypassing Power to Null Space on Redundant Manipulators

The significance of physical human-robot interaction (pHRI) with collaborative robots in the industry is growing steadily. Within this domain, an admittance controller is crucial for enabling robots to follow or assist human intentions. However, a persistent challenge with admittance controllers is ensuring their passivity. Various strategies have been developed to address this issue by adjusting the control signals derived from the admittance model. While these strategies achieve passivity, they often inadvertently impact collaborative performance, preventing the system from accurately aligning with the intended dynamics. Accordingly, this paper introduces an adaptive hierarchical control approach for redundant robots to handle this problem. This approach diverts non-passive power into the null space without diminishing the robot's responsiveness to human input. Implementing this null-space controller involves the dynamic adjustment of compliance control error, ensuring joint limit avoidance while facilitating integration with energy tanks for enhanced reliability. Moreover, the method enables the calculation of adaptive error gain in a closed form, simplifying its real-time application. Experimental validation with a 7-DOF manipulator showed a reduction of non-passive energy from 1.61 J to 0.12 J without compromising task performance.