Design of a Reconfigurable Sensor Array for Localization of Magnetic Capsule Endoscope

Magnetic capsule endoscope (MCE) has been well used in clinical diagnosis. However, localization for MCE was still lacking in clinical scenarios. In this study, we developed a reconfigurable magnetic localization system (MLS) and a localization algorithm for MCE in clinical scenarios. The magnetic sensor array consists of six sensor boards. Configuration of the array, which is defined as homogeneous transformation matrices between these boards, can be adjusted based on anatomical structures to adapt to different patient sizes. The proposed localization algorithm includes initial value estimation and optimization. The initial value estimation method is based on an automatic sensor selection process, which can provide an effective iterative initial value for the optimization algorithm. In vitro experiments with commercial magnetic capsules were conducted in static and dynamic manners. The average position error in the static localization experiment was 1.2 mm with a standard deviation of 0.6 mm. The average position error in the dynamic tracking experiment was 2.4 mm with a standard deviation of 2.0 mm. Results showed that the proposed method can locate the capsule robot effectively, making it a valuable addition to noninvasive gastrointestinal diagnosis and treatment.