Application of an optimal look-up table to sensor data processing

This paper presents a new method for calculating a look-up table to evaluate the measurand when a sensor pair is considered. A least mean squares technique has been used to calculate look-up table entries directly from sparse calibration data. The algorithm has been implemented by using the eight-bit MC68HC11 microcontroller; it reduces to a half the maximum value of the approximation error referred to the full scale if compared with the classical methods. A resistive network with end point linearity referred to the full scale of 15% has been used to emulate a sensor pair and the two input signals have been sampled with a 8-bit A/D converter.