Variable Admittance Control in Feature Space Based on Human Intention

This work presents a variable admittance control method within the image-based visual servoing feature space to improve the robot’s adaptability to complex trajectories and enhance compliance and impact resistance in physical human–robot interaction (pHRI), integrating multiple sensors and accounting for various human intentions. A robust adaptive damper based on shear thickening fluid is designed to enhance compliance and impact resistance in pHRI, leveraging the virtual velocity of feature points and human direct intentions. To ensure smooth and stable motion during complex trajectory tasks, the desired force is computed using human indirect intentions and curvature information, enabling the robot to accurately follow and guide the forces applied by the human. An adaptive virtual stiffness coefficient is introduced, dynamically adjusting in response to the contact force, ensuring rapid convergence of the servoing process even under external disturbances. Experimental results demonstrate that the proposed method achieves rapid convergence and high impact resistance in visual servo positioning. The method significantly enhances compliance and impact resistance in pHRI across various trajectories, while the integration of admittance control in feature space further improves alignment success rates and stability.