Nonlinear Multifunctional Sensor Signal Reconstruction Based on Total Least Squares

The least squares method is often used to estimate the parameters in multi-functional sensor signal reconstruction. If the data has been contaminated, the computational result of the method turns out to be insignificant. Two methods presented in this paper are suitable for different nonlinear conditions, which are based on the combination of the total least squares algorithm with the local linearization strategy and Stone-Weierstrass theorem. The two methods evaluate both the sensor output bias and its input error. The results of emulation and theory analysis indicate that the proposed algorithms are more accurate and reliable for signal reconstruction.