Abstract
Robot teleoperation systems are critical for ensuring operator safety in hazardous environments. However, precise manipulation faces dual challenges in pose and contact force control, which may induce operator cognitive overload and workpiece damage. To address this, we propose a hierarchical shared control system comprising master, middle, and slave layers, integrating adaptive robust force control (ARFC), intelligent posture control, and variable-impedance human-robot interaction. The ARFC is established by comprehensively modeling environmental position, velocity, and disturbances. Using the end-effector's normal velocity as input and combining adaptive laws with Lyapunov stability analysis, it ensures accurate and robust force control during grinding and polishing. An implicit surface modeling-based posture control strategy is then proposed to predict optimal postures on complex curved surfaces in real time, ensuring the grinding head consistently conforms to surface geometry. On the master side, variable-impedance control and Kalman filtering are integrated to filter hand tremor, forming a hierarchical shared architecture for flexible manufacturing. Under this shared-control paradigm, the operator retains high-level strategic authority while the robot executes low-level force/pose adjustments, synergistically fusing human expertise with machine precision. Extensive experiments validate the proposed method in constant-force maintenance, surface adaptability, and surface quality improvement, achieving quality gains of 94.84% and 76.6% in blade polishing and painted metal sheet grinding, respectively.

This publication has 26 references indexed in Scilit: