Adaptive based Sequential Implementation Resampling Particle Filter (ASIRPF) for Real Time Geophysical Applications

Abstract

The services of geophysical application in the obstructed area can be extremely challenging, especially if the Global Navigation Satellite System (GNSS) signal is unavailable or weak. Usually, the integration with the inertial sensor in the smartphone was used for assisting better navigation. Nonetheless, the usage of particle filter module in optimizing data from positioning sensor will contribute to the low particle sample size or known as sample impoverishment phenomenon, and finally increasing the location positioning error in a certain period. Adaptations towards to particle sample size and noise, must be made, to make the particle filter more intelligent, reliable and robust in a long time it is running. In this paper, we propose a new algorithm of sequential implementation resampling particle filter by adapting the particle sample size and sensor noise measurement. This adaptation will be used to counteract in a different situation. As the results, the paper shows the proposed solution can achieve an average improvement of 24.78% by reducing RMSE of state estimation compared to previous algorithm. In the future, it is expected to contribute to the modernization of geophysical applications.