Uncertainty model unfalsification with simulation

There has been much recent interest in model unfalsification for robust control of dynamic systems. In particular, it has been shown how to establish a tradeoff between sensor/process disturbances and linear time-invariant dynamic uncertainty using prediction error modeling. Unfalsification has also been integrated into adaptive control algorithms, in which a key component is the ability to trade between disturbances and dynamic uncertainty. The main contribution presented here is the calculation of model unfalsification tradeoff curves for a simple simulated system. The unfalsified dynamic uncertainties are compared to the difference between the identified models and the true model