Combined detection-estimation (DE) schemes for estimating the state of a dynamical stochastic system with uncertainties are considered. Several different performance criteria such as the weighted minimax mean-squared error and the incremental mean-squared error are used. The uncertainties in the system model too are modeled in several different ways such as single or multiple bounds on the unknown parameter values, or Markov chain nature of the unknown jump parameters. The DE schemes presented exhibit adaptivity to a certain degree for large observation record, while, on the other hand, possess small-sample robustness too.
|Original language||English (US)|
|Number of pages||6|
|Journal||National Conference Publication - Institution of Engineers, Australia|
|State||Published - Jan 1 1979|
|Event||Proc, Jt Autom Control Conf - Denver, CO, USA|
Duration: Jun 17 1979 → Jun 21 1979
ASJC Scopus subject areas