An Optimal Design for Passive Magnetic Localization System Based on SNR Evaluation

Passive magnetic localization is a feasible method to perform a small-scale localization for robotics and biomedical devices, for instance, wireless capsule endoscopy. Suffering from the magnetic noises such as geomagnetic noise and magnetic disturbance in the environment, the localization accuracy of the passive magnetic localization is drastically influenced without a careful system design. In this article, the signal-noise ratio (SNR) of each sensor measurement is defined and applied in the design of both the localization algorithm and the measurement system. Simulations are conducted to verify the relationship between the measurement SNR and the accuracy of the distance estimation. In addition, strategies on the selection of the target magnet, effective SNR threshold of the localization algorithm, and the optimal sensors' distribution are proposed based on the SNR evaluation results. Extended experimental results show that by setting the SNR threshold at 6 dB and adopting the optimal sensors' distribution, the root-mean-square localization error is improved from 0.0529 to 0.0364 mm and the orientation error is improved from 17.3° to 13.7° compared to the uniform sensors' distribution. The proposed algorithm can be used for accurate localization and tracking for untethered devices or biomedical devices, e.g., wireless capsule endoscopy.