Adaptive Tracking Control of One-Dimensional Respiration Induced Moving Targets by Real-Time Magnetic Resonance Imaging Feedback

One of the main challenges in magnetic resonance imaging (MRI) guided interventions is for an instrument to access and track a respiration-induced moving target inside the patient. In this article, we present a mechatronic system that demonstrates such capability for one-dimensional motion in phantom experiments. Our system consists of real-time sensing by MRI acquisition, detection, and localization, MRI-conditional instrument actuation by remotely and hydrostatically actuated fluid actuators, and adaptive control algorithms for tracking real-time measured respiratory motions subject to the sensing and actuation dynamics and uncertainties. Model-based simulations are performed to analyze the tracking performance pertaining to system parameters such as sensing latency and image tracking error. Experimental results obtained from moving phantoms are presented to demonstrate the control system's tracking performance.