Dynamic estimates of confidence and informational sets in statistically indeterminate systems (Q1921521)

Consider a discrete-time stochastic linear control system driven by additive white noise terms, and having also a linear state-output (observation) equation, where it is only known that the controls in the dynamic equation and in the state-output equation are contained in certain ranges. Using a minimax approach, from the a posteriori state distribution, confidence sets are constructed for the state of the system with respect to a given probability level \(\alpha\). The computation of the confidence sets in real time is discussed.