A novel positioning method for magnetic spiral-type capsule endoscope using an adaptive LMS algorithm

Abstract

In this paper, a novel positioning method for the wireless capsule endoscope (WCE) is proposed. An up-down symmetric array of magnetic sensors is used to detect magnetic signals from an external permanent magnet (EPM) for active control and mixed magnetic signals (the EPM and the WCE), and the adaptive least mean squares (LMS) algorithm is applied. Firstly, the number of iterations is determined by comparing the cancellation effect of input signals of different lengths. Subsequently, to separate the WCE’s magnetic signals from the mixed magnetic signals, the data obtained from the magnetic sensor arrays are processed in weighted iterations. The method has been applied to the actual experimental platform. From the experimental results achieved in this work, the average relative errors of the WCE’s triaxial signals were found to be 2.04 %, 2.20 %, and 1.47 % respectively. Achieved results demonstrate the feasibility and rationality of the positioning method discussed in this work. Moreover, the method can solve the problem of strong magnetic interference when the EPM provides active control to the WCE. It plays an essential role in driving the realization of closed-loop control of the WCE.