Modeling of a humidity system using ARX model and feed forward neural network - A comparative study

An experimental study was conducted on a laboratory scale humidity system to obtain the empirical model (process). A pseudo random sequence was introduced for the obtained model to get the input and output data. Different models such as autoregressive exogenous input model (ARX), non linear ARX and a feed forward neural network (FNB) model using back propagation algorithm were tried in MATLAB environment with the help of input and output data. The models obtained were analyzed for root mean squared error (RMSE) and FNB model resulted the best fit with minimum RMSE.