An adaptive Kalman filter for physiological tremor suppression in robot-assisted minimally invasive surgery

Abstract

Physiological tremors of surgeons’ hands bring vibration to the tip of surgical instruments, which may severely damage the accuracy of surgical procedures and even put patients’ lives at risk. Thus robot-assisted surgery using a teleoperation system was introduced to solve this problem. Many filter algorithms were implemented into the robot-assisted surgery system to suppress the inevitable tremors, but always overlooking the traits of hand motion. An innovation/residual-based floating forgetting-factor adaptive Kalman filter (IFFAKF) is proposed to address this problem. A floating forgetting-factor is brought into the innovation/residual-based adaptive estimation algorithm to adjust the process noise covariance (Q) and measurement noise covariance (R) of the system, thus generating estimation of voluntary hand movement and suppressing the vibration at the tip of the surgical instrument caused by the physiological tremor during surgeries. The proposed filter has been contrasted with various filters in simulations and experiments. The results confirm that the proposed filter has obvious advantages in terms of time delay, estimation error, and tremor suppression performance.