A Kalman filter approach to LVDT-based heat deformation test

In this paper, Kalman filter is designed and implemented on a heat deformation tester (HDT) to estimate the LVDT sensor information from a noisy measured data sequence. In simulation and test, the output response of Kalman filter can separate noise from reference signal and the outcome is closely matched with reference signal. The response of low-pass IIR filter is smooth and can also separate noise but some time delay is presented. The mean squared error (MSE) of low-pass IIR filter is 0.6354 while the MSE of Kalman filter reduces to 0.1467, a significant improvement by 4 times. Also, the computational efficiency of Kalman filter is better than that of low-pass IIR filter in this application. Both filtering algorithms are implemented and programmed into a heat deformation tester (HDT). The performance is impressive.