Stochastic Adaptive Control and Estimation Enhancement.
Final rept. Feb 80-Apr 84,
CONNECTICUT UNIV STORRS
Pagination or Media Count:
This report presents the results of the investigations conducted over a period of four years on control algorithms designed for stochastic systems. The main feature of these algorithms is that these account for 1 the current uncertainty in the system and 2 the anticipated future uncertainty in the system, which is in general control-dependent. The first feature leads to the control to have the cautious property in order to minimize the effect of the current uncertainties of the systems performance. The second feature allows the control to affect in addition to the systems state also the systems uncertainty. Such a controller is called dual controller because, by taking advantage of its dual effect has the capability of reducing the future uncertainties. These uncertainties can pertain to the systems state or its unknown parameters. Both continuous-valued and discrete-valued uncertainties have been considered.
- Statistics and Probability