Magnetic Medical Capsule Robots

Abstract

Over the last decade, researchers have started exploring the design space of Medical Capsule Robots (MCRs): devices that can operate within the human body performing functionalities such as diagnosing, monitoring, and treating diseases. Clinical applications for MCRs span from abdominal surgery to endoscopy. MCRs are severely resource constrained devices in size and in available mechanical/electrical power. Magnetic manipulation is becoming one of the main strategies for MCRs navigation and remote actuation, as it allows the transmission of forces across a physical barrier without requiring an internal source of power. The current state of the art lacks proprioceptive systems able to control magnetic actuation to enable MCRs to perform medical procedures. This dissertation presents controllable strategies for remote and local magnetic manipulation of MCRs combined with real-time proprioceptive sensing. The study focuses on two main applications: magnetic pose detection for navigation of Wireless Capsule Endoscope (WCE) along the gastrointestinal (GI) tract, and Local Magnetic Actuation (LMA) for powering Degrees of Freedoms (DOFs) of abdominal surgical robots. The proposed magnetic pose detection algorithms were successfully applied to remote navigation on several WCE prototypes during in-vivo trials. Two different approaches for magnetic pose detection compatible with magnetic actuation based on sensor fusion are presented and evaluated. The first approach takes advantage of the magnetic field mathematical derivation based on cylindrical symmetry, performing absolute pose detection for 50 Hz real-time systems. The second algorithm achieves a refresh rate of 1 kHz applying a least square interpolation to the finite element solution of the magnetic field, to obtain Jacobians closed-form expression function of capsule poses changes. A third algorithm estimates the force generated by magnetic coupling along the magnetization direction to study the force required to drag WCEs along the GI tract. The second topic covers the design and control for the LMA. This concept allows the transmission of rotary motion by applying the magnetic spur gear concept to actuate robotic DOFs of MCRs. The proposed dynamic model is used to design closed-loop control strategies. Two surgical tools based on the LMA are presented: a single-DOF retractor and a 4-DOF manipulator.