Anti-Interference Adaptive Admittance Control for Collaborative Robots Considering Friction Compensation

The widespread integration of collaborative robots, or cobots, into unstructured industrial environments necessitates control strategies that ensure both high precision and compliant physical human-robot interaction. Traditional fixed-parameter admittance control schemes often fail to maintain stability and performance when subjected to varying environmental stiffness, non-linear friction disturbances, and external interference. This paper proposes a novel anti-interference adaptive admittance control framework designed to enhance the transparency and robustness of collaborative manipulation. The proposed method integrates a variable admittance law with a high-order disturbance observer to estimate and reject lumped disturbances, including unmodeled dynamics and external shocks. Furthermore, a dynamic friction compensation strategy based on the LuGre model is incorporated to mitigate the stick-slip phenomena and hysteresis effects that degrade low-velocity tracking accuracy. Experimental results demonstrate that the proposed control architecture significantly reduces position tracking error and interaction force oscillation compared to conventional methods. The system exhibits superior adaptability to unknown environmental changes and effectively suppresses interference, thereby ensuring safe and smooth collaborative operations.