A Three-Dimensional EM-Based Implant Device Localization Method Improved by Genetic Algorithm

Abstract

These years, implant devices has been attracting great attention. Wireless capsule endoscope is one good example of the implant devices. Nowadays, the implant devices are usually used to just examine one’s diseases. However, in the near future, they may be used not only for the examination but also to treatment of a disease. To realize the next generation capsule endoscope, it is strongly required to add location information to the implant device. Such localization techniques enable us to control the movement of the implant device or to improve the treatment efficiency. Localization techniques has been studied widely, which often use radio frequency (RF) signals from the device. When we use this technique, there is a problem that we need a burdensome pre-measurements before the localization. To avoid such a burdensome pre-measurement, we turn our eyes on electromagnetic (EM) imaging which estimates the internal structure of an subject. The imaging can realize an examination without any burdensome pre-measurements, however, the EM imaging with high resolution will need huge computational complexity. To solve the problem, we applied genetic algorithm (GA) into the EM imaging-based localization method. Furthermore, we evaluated the localization performance of the proposed GA-based method compared with the conventional RF-based methods through computer simulations using a simple human body model. In addition, we show a possibility of using the EM imaging as an promising implant device localization method without any burdensome pre-measurement.