Improving transparency in physical human–robot interaction via unknown dynamics compensation

Physical human–robot interaction (pHRi) often involves tasks with unknown load dynamics, such as transport and assembly, which can reduce transparency, efficiency, and operator comfort. This study presents a compensator for real-time adjustment of unknown load dynamics, enhancing transparency through admittance control. Transparency was quantified as energy per distance generated by interaction forces, with the compensator’s design addressing unaccounted physical dynamics affecting both operator and robot, modeled as an equivalent physical system. Utilizing interaction data from the end-effector coordinate system and a time delay control approach, the compensator was mathematically formulated to mitigate dynamic impacts, with stability verified via the Lyapunov criterion. Simulations and empirical tests demonstrated improved transparency over existing controllers across varying motion speeds and load dynamics. This study highlights the role of dynamic compensation in advancing pHRi transparency and proposes future work to refine low-level dynamic adjustments.