Article
Transactions of the Institute of
Measurement and Control
1–8
Ó The Author(s) 2017
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0142331217741959
journals.sagepub.com/home/tim
Event-triggered based distributed
H
N
consensus filtering for
discrete-time delayed systems over
lossy sensor network
Shenquan Wang, Yuenan Wang, Yulian Jiang and Yuanchun Li
Abstract
This paper investigates the issue of event-triggered distributed H
N
consensus filtering for discrete time-varying delay systems over lossy sensor net-
works with stochastic switching topologies. For each sensor node, the event-triggering mechanism is given by an event detector, which determines
whether to transmit the output measurement or not. The communication links between the event detector and the distributed filter are assumed to
be over a lossy network, and the missing probability is governed by a set of random variables. Through available output measurements from not only
the individual sensor but also its neighbouring sensors, according to the interconnection topology to estimate the system states, a sufficient condition
is established for the desired distributed filter to ensure that the overall filtering dynamics are stochastically stable and achieve a prescribed distributed
H
N
average performance. Meanwhile, the corresponding solvability conditions for the desired distributed filter gains are characterized in terms of feasi-
bility linear matrix inequalities. Finally, a simulation example is provided to demonstrate the effectiveness of the proposed approaches.
Keywords
Event-triggered control, distributed filtering, stochastic switching topologies, lossy sensor networks, packet dropout, time-varying delays
Introduction
Over the past decades, sensor networks have received increas-
ing attention, owing to their huge potential applications in
various areas, including environment monitoring, industrial
automation, information collection and wireless networks
(Dong et al., 2013, 2014; Liang et al., 2017; Liu et al., 2015;
Shen et al., 2010, 2011; Su et al., 2013; Yu et al., 2009; Zhu
et al., 2016). A typical sensor network is composed of a large
number of sensor nodes and also a few control nodes, where
each sensor is capable of performing simple tasks of sensing,
computation and wireless communication. As an important
issue in the research of sensor networks, distributed filtering
or estimation has received increasing interest recently (Dong
et al., 2013, 2014; Liu et al., 2015; Olfati, 2007; Shen et al.,
2010, 2011; Su et al., 2013; Ugrinovskii, 2011; Yu et al., 2009;
Zhu et al., 2016, and references therein). Compared with tra-
ditional filtering for a single sensor, each sensor for a
distributed-filters design can receive not only the information
from its own measurement but also that from its neighbour-
ing sensors’ measurements, according to the topology of the
given sensor network. However, the distributed filters in
the classical H
N
sense have been designed by allowing for the
fixed topology of sensor networks (Dong et al., 2013, 2014;
Shen et al., 2010, 2011; Su et al., 2013), although a rare result
published in the current literature uses a switching topology,
(Zhu et al., 2016). Hence, it is significant to study the problem
of distributed filtering with stochastic switched topologies.
Owing to the inherently limited bandwidth of the commu-
nication channel and unnecessary communications, distribu-
ted filtering in a sensor network inevitably leads to some
degree of lossy network environment, including packet drop-
outs (Dong et al., 2013, 2014; Shen et al., 2010; Su et al.,
2013) and time delays (Gu, 2000; Li and Gao, 2011; Seuret
et al., 2015; Shao and Han, 2011). These packet dropout and
time delays are regarded as the main sources of performance
degradation or even divergence of the implemented distribu-
ted filtering algorithms. Conversely, note that the packet
dropout phenomenon is considered an extended delays, which
can be convenient in research of distributed estimation (Jiang
et al., 2014; Milla
´
n et al., 2012). Thus, it is noteworthy that
the problem of how to deal with packet dropouts and time-
varying delay among interacting sensors in a proper way has
become a main challenge in distributed H
N
filtering designs.
Unfortunately, at the time of writing, when it comes to sensor
networks, the corresponding distributed filtering problem
with data packet dropouts and time-varying delay has not
College of Electrical and Electronic Engineering, Changchun University of
Technology, China
Corresponding author:
Yulian Jiang, College of Electrical and Electronic Engineering, Changchun
University of Technology, Changchun, 130012, China.
Email: jiangneu@gmail.com