Set-membership filtering for complex networks with constraint communication channels☆
Introduction
With the expansion of system scale and the increasing of its complexity, complex networks that can be used to describe many complex systems are penetrating into many different fields such as mathematics, life, engineering, and so on (Basaras et al., 2019, Chu and Iu, 2017, Liu and Ye, 2021, Yang, Hu et al., 2021, Zhang et al., 2021). In most literature concerning complex networks, the sampling rates for different components (such as plant, sensor, and so on) are assumed to be the same for analysis convenience (Li and Xiao, 2021, Vega et al., 2020). In practical, however, for different components, or even for the same component, such an assumption is often unrealistic since it is difficult to guarantee that the sampling rate is the same under different physical constraints. Furthermore, it is clear that a higher sampling rate gives more available information to improve system performance by sacrificing more resources, which is often undesired for resource-limited case (Shen, Wang, Shen and Han, 2020). Consequently, it is of practical importance to adopt the multi-rate (MR) sampling scheme, and the research of MR sampling systems has received considerable attention (Lin and Sun, 2019, Shen, Wang, Shen, Alsaadi, 2020, Tian and Sun, 2021). Nevertheless, to our best knowledge, the research in MR sampling complex networks has been rarely addressed, which motivates our further study.
Wireless sensor networks, which consist of plenty of sensor nodes and are widely distribute over a variety of environments, are usually used to collect and process information as well as communicate with filter (Liu, Zhao et al., 2020, Rao et al., 2020, Wang et al., 2021, Xu et al., 2021). For the distributed sensor networks, the energy of sensors is always limited and difficult to charge. Furthermore, the capacity of communication channel is constrained, which implies that the transmitted information would inevitably suffer from data collisions or packets dropouts when all system components try to access a shared communication channel simultaneously. In order to overcome this constraint and achieve the goal of improving the transmission efficiency and saving energy, a weighted try-once-discard (WTOD) protocol (Shen, Wang, Shen, & Dong, 2021) is introduced in this paper, which is a dynamic scheduling scheme in accordance with different weights. Compared with stochastic communication protocol (Song, Wang, Liu, & Wei, 2020) and round-robin protocol (Xu, Lv, Lin, Lu, & Quevedo, 2022), the WTOD protocol assigns the transmission right to the most needed sensor according to the given quadratic selection principle (Ju et al., 2020, Zou et al., 2016). However, once the WTOD protocol is employed to select one for transmission, the rest of data is discarded, which is undesirable for filter (Liu, Wang, Chen, & Wei, 2020b). Therefore, how to compensate for the other measurements except the assigned one becomes a challenging problem. Some compensation methods, including the hold on compensation (Zou et al., 2016) and the zero compensation (Ju et al., 2020), are frequently used. It should be pointed out that the hold on compensation method and zero compensation method cannot ensure which one is closer to the actual value. Therefore, how to design a mixed compensation method by combining these methods is an interesting issue.
Due to the large scale of complex networks, it is extremely difficult to directly acquire the full state information of complex networks (Liu et al., 2020a, Liu, Wang, Lu et al., 2021, Wang and Luo, 2021), and the measurement is also subject to the noise. Thus the filter is necessary to be designed by the available measurement of complex networks (Liu, Wang, Zhou, 2021, Liu et al., 2022). Up to now, various filtering techniques have been presented, which include Kalman filtering, robust filtering, set-membership filtering, and other approaches (Ierardi, Orihuela, & Jurado, 2021). Kalman filtering is good at solving systems with Gaussian noise (Rocha and Terra, 2021, Xia and Wang, 2015), and robust filtering is mainly used to handle energy bounded noise (Dong et al., 2014, Dong et al., 2016, Yang, Tu et al., 2021). The set-membership filtering method, dated back to the 1960s, has an advantage in dealing with the unknown but bounded noise, and such a method has been recognized in research communities in the past decades (Liu, Wang, Wang et al., 2021, Orihuela et al., 2017). It needs to be emphasized that the unknown but bounded noise only requires to know the upper and lower bounds rather than its distribution, which is more conformable to practical situation in MR sampling complex networks. Nevertheless, little attention has been paid to set-membership issue for MR sampling complex networks, which motivates our current investigation.
Motivated by the above considerations, this work investigates the set-membership filtering for MR sampling complex networks with WTOD protocol. Different from the existing ones (Ju et al., 2020, Zou et al., 2016), a new mixed compensation method and corresponding filter design method are proposed for better filtering performance. The main contributions are as follows.
- 1.
For complex networks, MR sampling and communication capacity constraint are considered simultaneously, the WTOD protocol is introduced to overcome the constrained communication channel, and the Pseudo measurement (PM) approach is utilized for modeling synchronization. Compared with the existing ones (Lin and Sun, 2019, Shen, Wang, Shen, Alsaadi, 2020, Tian and Sun, 2021), the set-membership filtering for the model that concentrates on MR sampling time-varying complex networks and constrained communication channel is meaningful.
- 2.
A new mixed compensation method is proposed to obtain the measurement that is closer to the actual value. An indication function is introduced to describe the used compensation method, which can be adjusted by a chosen compensation weight coefficient . Based on the measurement under the mixed compensation, a mixed compensation dependent filter is designed.
- 3.
Due to the influence of compensation weight coefficient on the compensated measurements, an algorithm is proposed by minimizing the constraint matrix to find an optimized and the corresponding optimized ellipsoid. The mixed compensation method and the optimized algorithm can be directly used to the other communication protocols which need compensation, such as stochastic communication and the round-robin protocol, and so on (Ju et al., 2020, Shen et al., 2021, Song et al., 2020, Xu et al., 2022, Zou et al., 2016).
The organization of this article is as follows. Section 2 formulates the discrete time-varying complex networks with nodes, proposes the WTOD protocol and mixed compensation method, and designs a mixed compensation dependent filter. In Section 3, sufficient conditions of the -dependent constraint are derived. Then a set of optimized ellipsoids and the filter gains are obtained. In Section 4, a numerical simulation is given, and conclusions are given in the end.
Notations: , and denote, respectively, the set of natural numbers, n-dimensional Euclidean space and real matrices. The diagonal matrix is denoted as . The identity matrix and the zero matrix with appropriate dimensions are represented by and . The symbol in a matrix represents the symmetric term. For a matrix , the matrix and symbol denote its transpose and the trace, respectively. The Kronecker delta function is a binary function that equals if and equals otherwise. For a vector , denotes its largest element. The symbol “” stands for the argument of the maximum , where is the set of values of . The notation “” returns the remainder after division of by , where is the dividend and is the divisor.
Section snippets
Complex networks description
As shown in Fig. 1, discrete time-varying complex networks with nodes are given as where is the sampling instant for state and () is the sampling instant for measurement. and are the state vector and the measurement of the th node, respectively. The nonlinear function satisfies Assumption 1. , , ,
Main results
In this section, by employing the recursive matrix inequality technique and the mathematical induction, a sufficient condition for the filtering error system (16) satisfying -dependent constraint is first derived. Then, a set of optimized ellipsoids and the filter gains are obtained by solving the optimization problems.
Numerical example
In this section, the effectiveness of the designed set- membership filtering for time-varying MR sampling complex networks with WTOD protocol and mixed compensation is illustrated by the following numerical example, where , and . The inter-coupling matrix is and the outer-coupling matrix is given by
The system parameters for each node are considered as
Conclusions
In this paper, the set-membership filtering problem has been studied for the MR sampling complex networks with WTOD scheduling protocol. A new mixed compensation method has been proposed to obtain better compensation and a corresponding mixed compensation dependent filter has been designed. Then sufficient conditions have been derived to guarantee the existence of the set-membership filters. By solving the optimization problems with inequality constraints, the set of filtering gains and the
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (37)
- et al.
Finite-horizon estimation of randomly occurring faults for a class of nonlinear time-varying systems
Automatica
(2014) - et al.
A distributed set-membership estimator for linear systems with reduced computational requirements
Automatica
(2021) - et al.
The identification of crucial spreaders in complex networks by effective gravity model
Information Sciences
(2021) - et al.
An overview of multirate multisensor systems: Modelling and estimation
Information Fusion
(2019) - et al.
Dynamic event-based state estimation for delayed artificial neural networks with multiplicative noises: A gain-scheduled approach
Neural Networks
(2020) - et al.
Observer-based synchronization control for complex networks against asynchronous attacks
Information Sciences
(2021) - et al.
Distributed set-membership observers for interconnected multi-rate systems
Automatica
(2017) - et al.
Robust Kalman filter for systems subject to parametric uncertainties
Systems & Control Letters
(2021) - et al.
State estimation for multi-rate artificial neural networks with integral measurements: A switched system approach
Information Sciences
(2020) - et al.
Low-dimensional recurrent neural network-based Kalman filter for speech enhancement
Neural Networks
(2015)
On extended state estimation for nonlinear uncertain systems with round-robin protocol
Automatica
Finite-time cluster synchronization in complex-variable networks with fractional-order and nonlinear coupling
Neural Networks
Synchronization of delayed neural networks via event-triggered dynamic output control
Neural Networks
Intermittent control for finite-time synchronization of fractional-order complex networks
Neural Networks
Set-membership filtering for time-varying systems with mixed time-delays under round-robin and weighted try-once-discard protocols
Automatica
Identifying influential spreaders in complex multilayer networks: A centrality perspective
IEEE Transactions on Network Science and Engineering
Complex networks theory for modern smart grid applications: A survey
IEEE Journal on Emerging and Selected Topics in Circuits and Systems
On H-infinity estimation of randomly occurring faults for a class of nonlinear time-varying systems with fading channels
IEEE Transactions on Automatic Control
Cited by (8)
Event-Triggered Ultimately Bounded Filtering for Two-Dimensional Discrete-Time Systems under Hybrid Cyber Attacks
2024, Journal of the Franklin InstituteSet-membership filtering for two-dimensional shift-varying systems with stochastic communication protocol and uniform quantization
2024, Transactions of the Institute of Measurement and ControlPartial-node-based resilient set-membership filtering for output-coupled complex networks with event-triggered mechanisms
2024, International Journal of Robust and Nonlinear ControlState Estimation for Nonuniformly Sampled Neural Networks with Hidden Information
2023, IEEE Transactions on Systems, Man, and Cybernetics: Systems
- ☆
This work was supported in part by the Key Area Research and Development Program of Guangdong Province (2021B0101410005), the National Natural Science Foundation of China under Grants (62121004, 62006043), the Natural Science Foundation of Guangdong Province, China (2021B1515420008, 2021A1515011634, 2019A1515011061), and the Local Innovative and Research Teams Project of Guangdong Special Support Program (2019BT02X353).