A Localization Method for the Magnetic Levitation Capsule Robot

Abstract

Magnetically controlled capsule endoscope robot has developed rapidly. A key problem of robot is how to precisely control its movement in human body. Thus, the position and orientation parameters of the robot should be detected in real time. This paper proposes a novel localization method based on simulated annealing-particle swarm optimization (SA-PSO) algorithm. A localization system with 4×4 Hall sensor array is designed and used to test the accuracy of the localization method. Static experimental results show that the average total position error is 1.64 mm, the average total orientation error is 1.08°. The total position error and the total orientation error at the center of the working space can reach 1.31 mm and 1.03° respectively. The localization method in real-time localization experiments also has great accuracy. When the localization system is applied in the magnetic levitation capsule robot control system (MLCRCS), the position error and orientation error can reach 1.61 mm and 1.5°.