H_∞ filtering for uncertain time‐varying systems with multiple randomly occurred nonlinearities and successive packet dropouts

Abstract

SUMMARYThis paper is concerned with the robust H ∞ finite-horizon filtering problem for discrete time-varying stochastic systems with multiple randomly occurred sector-nonlinearities (MROSNs) and successive packet dropouts. MROSNs are proposed to model a class of sector-like nonlinearities that occur according to the multiple Bernoulli distributed white sequences with a known conditional probability. Different from traditional approaches, in this paper, a time-varying filter is designed directly for the addressed system without resorting to the augmentation of system states and measurement, which helps reduce the filter order. A new H ∞ filtering technique is developed by means of a set of recursive linear matrix inequalities that depend on not only the current available state estimate but also the previous measurement, therefore ensuring a better accuracy. Finally, two illustrative examples are used to demonstrate the effectiveness and applicability of the proposed filter design scheme.

Moving Horizon Estimation for Networked Systems With Quantized Measurements and Packet Dropouts

“…By the results in [38], the quantization effects can be transformed into sector bounded uncertainties (5) By using (1), (2) and (5), one obtains the following system model (6) where . The following assumptions are needed in the derivation of the main results: [11], [13]- [18], it is assumed that the statistics of the noises are known. However, it is difficult to know the exact statistics of the noises in practice.…”

“…In [10], the robust filtering problem was studied for stochastic uncertain systems with missing measurements, and the filtering error system was exponentially mean-square stable. The finite-horizon robust filtering problem was considered in [11] for discrete-time stochastic systems with packet dropouts, and a time-varying filter was obtained. Recently, the moving horizon state estimation problem was investigated in [12] for networked systems with packet dropouts, and a sufficient condition for the convergence of the estimation error sequence was presented.…”

“…The quantized filtering problem was investigated in [18] for Markovian jump LPV systems with packet dropouts. It should be noted that the works mentioned above (e.g., [4]- [11], [13]- [18]) have not been concerned with the bounds or constraints on states or noises. However, it is well known that these bounds allow us to add more information about the physical characteristics of the states, such as concentrations or molecular weights that must be positive [19], [20].…”

Finite-horizon estimation of randomly occurring faults for a class of nonlinear time-varying systems

“…If not properly coped with, the ROP would unavoidably deteriorate the operating efficiency or even cause the instability of the plant. So far, a series of results have been reported in the literature concerning the filtering and control problems in the presence of ROP, for example, randomly occurring nonlinearities [4,18], randomly occurring uncertainties [10], randomly occurring sensor delays [4] and randomly occurring sensor saturations [23]. Nevertheless, an important class of networkinduced phenomena, namely, randomly occurring faults (ROFs), have been largely overlooked despite their practical significance in networked control systems.…”

Distributed Synchronization in Networks of Agent Systems With Nonlinearities and Random Switchings

“…In particular, in [32], the concept of randomly occurred nonlinearities (RONs) is proposed to explain nonlinear disturbances in networked control systems and biological networks. The filtering problem of randomly occurred phenomena in networked systems was investigated [35], [36]. The problem of stochastic synchronization analysis for complex networks 2168-2267/$31.00 © 2012 IEEE composed of coupled discrete-time neural networks with RONs and time delays was studied.…”