Adaptive Admittance Control for Robotic Ultrasound Examination Based on a Breast Biomechanical Model.

BACKGROUND: Autonomous Robotic Ultrasound Systems (ARUS) enable standardized breast cancer screening, but accurate breast localization and the nonlinear deformation of soft tissue hinder stable contact-force control. METHODS: We propose a framework integrating visual perception with compliant control. An adaptive gradient-based edge detector extracts the nipple position, and a Variable Damping Admittance Control based on a Breast Biomechanical Model (VDAC-BBM) is developed, which compensates for tissue deformation via a biomechanical prior and dynamically tunes the damping parameter in real time to suppress force error. RESULTS: Phantom and human experiments showed force tracking error ≤ 0.4 N, with disturbance recovery speed improved by 16.5%-21.3% over conventional admittance control across the tested target forces. CONCLUSION: Embedding a biomechanical prior into the control strategy enhances force compliance and stability during robotic scanning of deformable soft tissue, providing a technical basis for clinical translation.