Adaptive Admittance Control for Internal-Force Tracking in Dual-Arm Palm Grasping

This work presents an adaptive admittance control aimed at regulating internal forces in dual-arm robotic manipulation under palm-grasping contact. Traditional admittance controllers, which utilize fixed parameters, often exhibit difficulties in ensuring reliable internal-force tracking, particularly during dynamic tasks or when faced with time-varying force references. To address these challenges, this paper adopts an adaptive law that adjusts the damping gain in real time based on the internal-force tracking error. The approach has been experimentally validated using a dual-arm setup that performs high-dynamics 3D and 6D cooperative motions. Results indicate that the proposed method improves force-tracking performance and responsiveness compared to a constant-gain admittance controller even when the choice of parameters is suboptimal.