Optimal Robust Navigation Control of an Endovascular Microrobot

Abstract

Microrobots are considered to be a promising solution for complex sequence of operations in biomedical applications. The paper aims to present an algorithm for path planning and control of an endovascular microrobot navigated by a conventional magnetic resonance imaging (MRI) device. The microrobot is a ferromagnetic spherical bead moving through human vascular system; accordingly, the dynamic model is well developed considering the most dominant resistive force, i.e., hydrodynamic drag. An optimal approach considering a power index is used for path planning of the robot from an initial position to the desired destination avoiding collision with boundaries within a generated robust domain based on the unmodeled dynamic of surface forces. Due to the presence of parameter uncertainty and unmodeled dynamics, an adaptive backstepping controller is designed to navigate the bead through the optimal robust path such that the control signal can be implemented via a conventional MRI device. The control method is proved to have stability and convergence as it manages to maintain the estimated parameters and control inputs bounded in the presence of disturbance and uncertainty. The method of path planning explicitly deals with unstructured uncertainties of the surface forces based on developed comparison and reasoning as it maintains minimum effort optimality. The adaptive backstepping method is designed to handle structural uncertainties and disturbances due to pulsatile flow and non-measured velocity. Several simulations were performed to examine navigation of the ferromagnetic microrobot in the circulatory system of human body, and the tracking errors, estimated parameters, and control signal inputs are represented, accordingly. Also, it is shown how wall effect correlation term in the model could increase the control accuracy.