Skip to Main Content
This paper focuses on wireless capsule endoscope magnetic field based localization by using a linear algorithm, an unconstrained optimization method and a constrained optimization method. Eight sensor populations are employed for performance evaluation. For each of five sensor populations, four different sensor configurations are investigated, which represent potential sensor placements in practice. Accuracy is evaluated over a range of noise standard deviations and the position area is set on a solid cylinder which well represents the realistic scenario of the human body. It is observed that the optimization method greatly outperforms the linear algorithm that should not be used alone in general. The constrained optimization approach outperforms the unconstrained optimization method in presence of large noise. Simulation results show that best position accuracy is achieved when the sensors are uniformly deployed on a 2D plane with some sensors on the boundary of the position area. For the sensor populations considered, when increasing sensor population by one, the accuracy improves by about 0.45 divided by the sensor population. The results provide useful information for the design of wireless endoscope localization systems.