Localization of Camera-free Capsule Robot within the Gastrointestinal Tract

Capsule robots have emerged as a promising tool for minimally invasive sampling of the gastrointestinal (GI) tract for early disease detection. However, effective localization of these capsule robots, especially the battery-free capsule robot without an on-board camera, within the GI tract remain challenging due to its complex, deformable geometry and limited accessibility. Accurate localization is essential for clinical deployment, for targeting disease site, and for improving the safety of the procedure. Compared with existing imaging techniques such as X-ray or CT, magnetic localization provides a radiation-free alternative for precise capsule robot localization. In this project, we designed a localization method for a magnetic battery-free capsule robot using a Hall effect sensor array, magnetic field superposition, and particle filtering. To enable localization, a robotic arm positions a permanent magnet (robotic arm magnet) with pre-calculated position and orientation. A magnetic capsule robot containing a magnetic compartment (capsule magnet) is placed within the setup mimicking the GI tract, and the Hall effect sensor array measures the combined field generated by both sources. By applying a magnetic field superposition model, supported and validated by finite element analysis of the generated magnetic field, the system can separate the capsule’s contribution (i.e., from capsule magnet) from the known actuator field (i.e., from robotic arm magnet). These refined measurements are then processed by a particle filter algorithm to calculate the capsule’s position in real-time. Future work will focus on tuning parameters used in the particle filter algorithm, expanding the sensor array to increase the workspace area, and implementing closed-loop force control to enable capsule navigation relying on the localization information.