Unified Admittance Control for Accurate Puncture and Respiration Following Based on Disturbance Observation and Model Predictive Control

Abstract

Percutaneous puncture is the clinical standard for diagnosis and therapy of lung tumors. Needle placement accuracy and safety are of great significance but severely affected by respiration. In this letter, a unified admittance control method for accurate puncture and respiration following is proposed. For lung puncture robots, an admittance puncture model considering remote center of motion (RCM) constraint, respiration and needle bending is first proposed, in which angular velocity feedforward is used to decrease the tearing force. Then, under real-time images, extended state observer (ESO) is introduced to observe needle bending and respiration motion velocity relative to the robot, which are further compensated. The resulting integral system is used to designed dynamic response for accurate puncture. Finally, robot output force is optimized with model predictive control (MPC) and designed in a unified form to be compliant with the needle force for respiration following and overcome it for accurate puncture. Experiment results show accurate puncture under static condition with small steady-state error, designed dynamic response and small tearing force, and respiration following with small interaction force and nearly constant needle angle and position relative to lung, which indicate high clinical value for robotic lung puncture.