Combining neural and conventional paradigms for modeling, prediction, and control
Monika Agarwal · 2002
Promising research using neural networks for modeling, prediction, and control, exploits the complementarity of the two paradigms to address realistic problem situations. This paper develops a general framework for identifying the possible ways of combining neural networks with physical models, model-based estimators, and conventional controllers. The framework presented not only naturally leads to the previously proposed schemes in the literature, but also reveals several new possibilities.