Centralized Multi-Sensor Robust Recursive Least-Squares Wiener Estimators in Linear Discrete-Time Stochastic Systems with Uncertain Parameters

Abstract

Seiichi Nakamori

Among the multi-sensor information fusion estimation problems, this paper proposes the centralized robust recursive least-squares (RLS) Wiener filter and fixed-point smoother for estimating the signal and the state in linear wide-sense stationary stochastic systems with the uncertain parameters in the system and observation matrices. Previously, the robust RLS Wiener filter and fixed-point smoother are proposed by the author in the case of the single-sensor observation for linear discrete-time stochastic systems with uncertain parameters. This paper extends the robust RLS Wiener estimators in the case of the single-sensor observation to the centralized multi-sensor robust RLS Wiener estimators. The signal is observed at each station as degraded by the uncertain parameters in the observation matrix. The centralized multi-sensor robust RLS Wiener filter and fixed-point smoother, proposed in this paper, have the advantage of not using information such as probabilities about the uncertain parameters in the system and observation matrices. Related to the centralized multi-sensor robust RLS Wiener filter, the recursive algorithm for the filtering error variance function of the state is proposed. The estimation accuracies of the centralized multi-sensor robust RLS Wiener filter and fixed-point smoother are superior to the centralized multi-sensor RLS Wiener filter and fixed-point smoother, respectively.

PDF

Share this article