Model uncertainty is a serious problem facing control engineers.Often it is not possible to adequately represent system characteristics such as non-linearity, time delay, saturation, time-varying parameters, and overall complexity. This is particularly true in automotive control and modelling. To tackle this problem this paper discusses an artificial intelligence method called artificial neural networks. After discussing the basics of artificial neural network, a controller that incorporates a neural network with a self-tuning controller is formulated. Then the neural network is applied to a particular engine problem. The network was used to track the system's dynamic using only the input and output data. (A).
Samenvatting