Adaptive controller design of MRI compatible pneumatic teleoperation system -

Abstract

This work presents an adaptive control design of a pneumatic teleoperation system that could be useful for applications like MRI-guided surgery. The system under study is unique because of its reduced number of components compared to other bilateral teleoperation systems, so cost and complexity are reduced. The direct fluidic connection and force feedback that is transferred to the human operator allow the operator to feel as if s-he were having physical contact with the environment situated in another room without the need for a force sensor on the slave actuator. The system under study is implemented in simulation first, transparency and stability were assessed for different operating scenarios, and sensitivity study is conducted to investigate what parameters affect the system performance. A linear controller is optimized for various operating remote environments via frequency response analysis, and yielded satisfactory results for certain operating physical environments, but its tuning is dependent on the impedance characteristic of the environments both on the master and slave sides. Since the system must perform under parametric uncertainties on both sides of the teleoperator, an adaptive control scheme is developed. A self-tuning regulator is designed to allow the teleoperator to cope with variable operating conditions. Using recursive least-square estimation, system parameters are estimated continuously and the controller drives the system output to the commanded input using a specified reference model using pole placement, so it can adapt to any change in environment impedances on both sides of the teleoperator. The controller is validated both in simulation and experiment, and yielded satisfactory performance under multiple operating conditions.