We describe modeling, system identification and control design for a magnet manipulator. Two of these manipulators, placed on either side of a patient, will articulate their large permanent magnets to generate a magnetic field in an anatomical region (e.g. the heart) as part of a medical navigation system under the remote control of a medical professional. We derive the equations of motion using the Lagrangian formulation; identify motor, amplifier and gear friction characteristics using data from a single manipulator prototype; design the servo control algorithm based on feedback linearization and decoupling followed by H∞ control and test it on the prototype. Developing a mathematical model of the mechanism and using advanced controls algorithms initially take more time than tuning a PID controller. But the resulting design yields much smoother motor torques compared to highgain controllers, does not excite high frequency dynamics and potentially improves reliability.

No abstract