Problem Definition:

Magnetic robotics can enable ultra-minimally invasive surgeries by moving small magnetic devices inside the body—without any incisions or physical connections. However, two major challenges slow down adoption:

1. Tracking the robot – Since the surgeon can’t directly see inside the body, we need an imaging system to track the robot’s position in real time.

2. Hardware complexity – Most magnetic actuation systems require large, custom-built electromagnetic coil setups, which take up space and make them less practical for use in operating rooms.

Hypothesis:

MRI machines are already widely available in hospitals and can:

1. Capture 3D images inside the body in real time.

2. Produce strong, controllable magnetic fields.

If we could use a single MRI machine for both imaging and actuation, we could track and move magnetic robots inside the body without introducing bulky new equipment.

Solution:

I converted a commercial MRI scanner into an autonomous robotic manipulation platform that:

• Provides real-time image feedback for closed-loop control.

• Receives intuitive user commands from a Custom-GUI.

• Generates actuation forces to move magnetic objects inside the body.

• Autonomously controls the motion of a magnetic rigid body object at high precision.

This system allows real-time, precise control of position and orientation for magnetic devices—enabling navigation through the body in an ultra-minimally invasive way.

I also built a teleoperation mode so surgeons can monitor live MRI scans and guide the robot using a standard handheld controller.