Improving the response of an accelerometer by using optimal filtering

In this paper a recursive least squares adaptive noise canceling method is applied to estimate electrical signals coming from an accelerometer embedde d in a bus under a performance test. Noise and/or interference corrupt this electrical signal, and it is necessary to diminish the noise in measuring the acceleration of the bus. The recursive least squares algorithm used in this paper has the advantage of computational simplicity and is well suited to represent time varying features in real-time measurements of the accele ration. In this experiment the signal of interest and the noise have very c lose bands of frequency to each other and it is very difficult to diminish the noise using fixed filters. The experimental results have shown that the adaptive filter acts as an adaptive notch filter whose null points are det ermined by the frequency of the noise signal, and it achieved an improvemen t of 20 dB of the signal-to-noise ratio. The experimental results demonstra te the importance of using both analog signal conditioning and digital sign al processing when we have to deal with signals corrupted by noise. (C) 200 1 Elsevier Science B.V. All rights reserved.