Hysteresis compensation of a porous silicon relative humidity sensor using ANN technique

This paper presents a simple technique based on well-known multilayer perceptron (MLP) neural network with back propagation training algorithm for compensating the significant error due to hysteresis in a porous silicon relative humidity sensor. The porous silicon humidity sensor has been fabricated, and its hysteresis with increasing and decreasing relative humidity has been determined experimentally by a novel phase detection circuit. Simulation studies show that the artificial neural network (ANN) technique can be effectively used to compensate the hysteresis of the porous silicon sensor for relative humidity (%RH) measurement. A hardware implementation scheme of the hysteresis compensating ANN model using a micro-controller is also proposed. Simulation studies show that the maximum error is within ±1% of its full-scale value.