An Uncalibrated and Accurate Robotic Puncture Method Under Respiratory Motion

Abstract

Robotic puncture system (RPS) has been widely concerned in thoracoabdominal puncture surgery, but it usually requires cumbersome hand-eye calibration (HEC) to ensure accuracy. An uncalibrated and accurate robotic puncture method under respiratory motion is proposed to determine the relationship between the manipulator and optical tracking system (OTS) in real-time. This method explored RPS to achieve the accurate location of angle and position without calibration respectively, and constructed the angle Jacobian matrix and position Jacobian matrix. An efficient approach that used Square-Root Unscented Kalman Filter (SR-UKF) to online estimate the Jacobian matrices of the RPS was presented. Moreover, this method is less time-consuming and can achieve a dynamic following, making it more efficient. Experimental studies on our designed RPS show that the proposed method has good accuracy and robustness in angle and position location; the average error of puncture results on the respiratory simulation model is 2.2382mm, which can meet the accuracy requirements of the actual thoracoabdominal puncture surgery.