Wavelet Thresholding Method For Accelerometers Denoising Under Vibration.

Abstract: Aimed at MEMS accelerometers are sensitive to vehicle maneuver vibration, a modified wavelet thresholding denoising method is proposed as an efficient pre-filter for low cost inertial accelerometers under vibration. Wavelet denoising can separate signal from noises according to their different characteristics and wipe off noises spreading entire frequency domain, which is widely used in denoising in the field of signal processing. This paper reveals how wavelet thresholding method to eliminate noises of the accelerometers. The choosing of thresholding operator and thresholding value is a key in wavelet thresholding denoising. Thus, an adaptive thresholding value and a modular square thresholding operator are proposed, and the specific steps of this method are given. The experiment shows that the proposed method exhibits a significant performance improvement, and can filter out noises. [Copyright &y& Elsevier]