Denoising MEMS accelerometer sensors based on L2-norm total variation algorithm

作者：Abbasi Ke**i R; Nikfarjam A^*

来源：Electronics Letters, 2017, 53(5): 322-323.

DOI：10.1049/el.2016.3811

• microsensors
• accelerometers
• signal denoising
• health care
• wireless sensor networks
• Butterworth low pass filter
• equiripple-pass-band low pass filter
• Savitzky-Golay low pass filter
• noise reduction
• power spectral density function
• signal-to-noise ratio
• accelerometer signal denoising
• LTV algorithm
• wireless accelerometer sensor
• healthcare application
• human movement assessment
• L2-norm total variation algorithm
• denoising MEMS accelerometer sensor

A method for denoising accelerometer data based on the L2-norm total variation (LTV) algorithm is presented. In order to collect accelerometer data, a wireless accelerometer sensor was developed that is directly connected to a central node. By benefiting from the LTV algorithm, the obtained signals from the accelerometer are denoised. The proposed method is tested by denoising in different accelerometer signals and the results are evaluated by signal-to-noise ratio and power spectral density functions of the signals. The obtained results reveal that noise reduction has been implemented satisfactorily. Hence, the measurement accuracy of accelerometer signals for the proposed method have improved approximate to 4-10% than other the three low-pass filters including Savitzky-Golay, equiripple-pass-band and Butterworth.

• 出版日期2017-3-2