MeX FX algorithm in temperature sensor data reconstruction

The problem that is still often encountered in temperature sensors and other sensors implementation is data processing. The data processing aims so that the data obtained from the sensor is valid. In this study, we have proposed the MeX FX algorithm to process error measurements on the LM35 (Linear Monolithic 35) temperature sensor when using a long data cable. MeX FX is a combination of an Exponential Weighted Moving Average (EWMA) filter specially modified for this study (EWMA custom build), an averaging algorithm, and a 1-dimensional Gaussian filter. The purpose of the MeX FX algorithm is to process data from the LM35 temperature sensor with the lowest possible implementation costs and memory requirements while maintaining the quality of data processing results. From the simulation test, it has been found that the MeX FX algorithm was superior to other tested algorithms when the noise level was from -150 mV to +150 mV. When we tested the algorithm on the microcontroller chip, there were differences in results caused by improper noise modeling in simulation testing. However, the MeX FX algorithm has the potential to be applied as a temperature sensor data processing algorithm when the noise characteristics are the same as in the simulation test.