Uncertainty propagation for efficient model-based control solutions

The aim of this work is to represent and propagate model parameter uncertainty so that a model-based control solution is more accurate and robust. To meet this aim, the random fuzzy variable approach is used to provide an accurate range of the uncertainty in the input parameters. In these ranges, the Latin hypercube Hammersley sampling technique is applied for efficient uncertainty propagation. The resulting distribution of the outputs can be used to identify suitable models from which to develop model-based controller solutions. An example is provided that demonstrates the potential of this type of approach for more robust regulation when compared to models that exclude the effect of parameter uncertainty.