Quasiperiodic predictive filtering for robot-assisted beating heart surgery

Beating heart procedures promise significant health benefits to patients but the fast motion of the heart poses a serious challenge to the surgeon. Robotic motion synchronization to heart movements could facilitate these surgeries, although for intracardiac procedures this requires the development of a predictive filter to compensate for the measurement noise and time delay present in 3D ultrasound imaging. In this paper, we present a quasiperiodic cardiac motion model and apply the extended Kalman filter to estimation of its parameters in real-time. We experimentally demonstrate high accuracy robot tracking to heart motion using this filter.