Adaptive Noise Reduction Model for MEMS Based Inertial Sensors

M. El-Diasty, A. El-Rabbany

Abstract:	The output signals of low cost, MEMS-based inertial sensors are characterized by their high noise level. Suppressing the high frequency noise component is essential for optimising the pre-filtering methodology. This paper proposes an adaptive neural network-based model for sequential noise reduction. A modular, threelayer feedforward neural network trained using the back-propagation algorithm is used for this purpose. Simulated data as well as real data collected at various rates from the Crossbow’s AHRS300CA IMU are used to validate the model. A comparison between the developed sequential model and another neural network-based de-noising model, namely the autoassociative approach, is also presented.
Published in:	Proceedings of the 2003 National Technical Meeting of The Institute of Navigation January 22 - 24, 2003 Disneyland Paradise Pier Hotel Anaheim, CA
Pages:	636 - 640
Cite this article:	El-Diasty, M., El-Rabbany, A., "Adaptive Noise Reduction Model for MEMS Based Inertial Sensors," Proceedings of the 2003 National Technical Meeting of The Institute of Navigation, Anaheim, CA, January 2003, pp. 636-640.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In