Linearizing a thermistor characteristic in the range of zero to 100degree C with two layer artificial neural networks

摘要

Artificial neural networks appear as an efficient tool to correctinput-output nonlinearities of sensors. In this paper, an artificialneural network (ANN) with two hidden layers used to linearize a staticcharacteristic of a thermistor is discussed. The data used were takenfrom the input-output calibrating thermistor bridge. Bothbackpropagation (BP) and random optimization method (ROM) have beencombined to adjust the weights of the neural network. Simulation resultsshow effectiveness and ability of the method suggested to linearize athermistor characteristic in the range of zero to 100 degree C