Robot Normal Posture Tracking Method Based on Force Information in Unknown Human Working Environments

In human-robot interaction scenarios such as medical rehabilitation and cosmetic care, a robot must maintain a stable contact posture normal to the skin surface. Uncertainties from inter-individual differences, variations in working areas, posture changes during the process, and real-time deformation severely limit the robot’s ability to track the flexible curved contours of the human body. To solve this problem, multiple auxiliary coordinate frames are established on the force-motion interaction system, the kinematics as well as the force/torque relationships during the contact are described and analyzed, and the corresponding transformation matrices of the coordinate frames are constructed. By combining Hertzian elastic contact theory with a biomechanical adhesion-friction model, a normal-vector relationship model based on six-axis force information is developed for a rigid spherical end-effector interacting with soft tissue, which is then used to obtain the current normal attitude in real time. To ensure the accuracy of the six-axis force data, a dual compensation scheme integrating secondary gravity compensation with periodic torque-error compensation is implemented. The proposed method enables real-time tracking of the unknown surface normal posture of human soft tissue through force-sensor feedback. Finally, experiments on a facial model, tracking the trajectory from the glabella to the nose tip along the nasal dorsum, are carried out. The results demonstrate that the normal-attitude error remains within 1.12°~3.20°under an impedance controller that regulates a compliant normal force, meaning that the proposed control strategy is effective and is helpful to enhance the adaptability of robots in unknown human working environments.