Variance-Constrained Filtering for Stochastic Complex Systems

During the past two decades, the variance-constrained filtering and estimation issues have been one of the dominant research topics on account of their great application potentials in a variety of research areas, such as imagine processing, target tracking and environmental monitoring. Along with the rapid advancements of network communication techniques, compared with the traditional control systems, the networked control systems possess prominent merits including flexible architecture, decreased maintenance cost, high reliability and so on. While the network communication brings some convenience, a series of network-induced phenomena generally emerge in the process of signal transmission due to constrained network bandwidth, unreliable communication link and complicated network environment, such as missing measurement, fading measurement, sensor saturation, sensor delay, network attacks, and so on. In addition, it is worth emphasizing that because of the modeling error or implementation error, some imperfect internal features regarding the dynamical systems might appear including stochastic uncertainty, stochastic nonlinearity and so on. On the other hand, it is noted that the systems are always time-varying in practical engineering applications on account of the complicated network media. As a result, it is crucial to tackle the filtering and estimation issues for time-varying dynamical systems affected by the incomplete information (networked-induced phenomena, modeling uncertainty and so on) under variance-constrained sense, which serves as a meaningful yet challenging topic.

In this chapter, the optimized recursive filtering issue is discussed for networked time-varying nonlinear systems with ROQ. A sequence of binary random variables is used to describe the ROQ, in which the logarithmic quantizer is employed to describe the phenomenon of quantized measurements. The purpose of this chapter is to construct a desirable filter of recursive form such that, for all ROUs, ROQ and stochastic nonlinearity, an upper bound regarding the FEC is solved and the expected filter gain with expression form is derived and given. Furthermore, a sufficient condition is provided to guarantee the boundedness of the filtering error dynamics. Lastly, some comparative simulation examples are provided to validate the usefulness of established recursive filtering algorithm.

In this chapter, the resilient filtering problem is addressed for networked time-varying nonlinear systems with stochastic uncertainties and incomplete measurements.

In this chapter, we aim to discuss the effects from the stochastic uncertainties and limited communication capability onto the whole filtering algorithm accuracy.

In this chapter, the objective is to tackle the filtering problems for two classes of nonlinear systems in the presence of ETM and missing measurements. Firstly, the RONs and missing measurements are examined. Here, the binary random variables are employed to characterize the RONs and the missing measurements, in which the occurrence probabilities of them are assumed to be uncertain. With the aim to reduce resource consumptions, the ETM is utilized during the data transmissions through the network.

In this chapter, the fault estimation problem is handled for time-varying nonlinear stochastic systems subject to RONs and RODAs. Two mutually independent binary random variables are employed to characterize the RONs and RODAs, respectively.

In this chapter, the fault estimation issue is addressed for time-varying networked systems influenced by ROUs and packet dropouts. Two sets of mutually independent random variables are employed to depict the ROUs and packet dropouts. The chief purpose is to develop a novel fault estimation scheme such that, for ROUs and packet dropouts, an optimized upper bound in relation to the EEC is obtained and the estimator gain with expression form is derived.

In this chapter, the fault estimation issue is handled for time-varying nonlinear stochastic systems with ROFs and sensor saturations.

In this chapter, the state estimation issue is handled for time-varying coupled stochastic complex networks involving missing measurements in the variance-constrained framework.

In this chapter, we aim to provide a novel recursive state estimation algorithm for stochastic complex networks in the presence of uncertain inner coupling and measurement quantization mechanism. The coupling strengths are varying over certain intervals and the measurements are quantized before being delivered to the remote estimator side.

In this chapter, we address the state estimation problem for a class of time-varying complex networks subject to ETM and stochastic coupling with UOPs. Firstly, two sets of Bernoulli random variables are used to model the stochastic coupling and missing measurements, where the UOPs are clearly described here.

In this chapter, the state estimation issue is addressed for time-varying uncertain complex networks involving switching topology, ETM and DPDs. To save network resources and reduce communication cost, the ETM is employed to arrange signal transmission.

In this chapter, the state estimation issues are solved for time-varying nonlinear complex networks with missing measurements and stochastic inner coupling under variance-constrained framework.

In this chapter, the main conclusions of this book are given and the potential research directions are pointed out accordingly.