程序代写代做代考 data mining case study algorithm Excel flex prolog chain untitled

untitled

INV ITED
P A P E R

Real-Time Wireless
Sensor-Actuator Networks
for Industrial Cyber-Physical
Systems
Despite their success in industrial monitoring applications, wireless sensor actuator

network (WSAN) technologies face significant challenges in supporting control systems.

This paper surveys recent advances in real-time WSANs for industrial control systems.

By Chenyang Lu, Senior Member IEEE, Abusayeed Saifullah, Member IEEE, Bo Li,

Mo Sha, Member IEEE, Humberto Gonzalez, Member IEEE, Dolvara Gunatilaka,

Chengjie Wu, Lanshun Nie, and Yixin Chen, Senior Member IEEE

ABSTRACT | With recent adoption of wireless sensor-actuator
networks (WSANs) in industrial automation, industrial wireless

control systems have emerged as a frontier of cyber-physical

systems. Despite their success in industrial monitoring

applications, existing WSAN technologies face significant

challenges in supporting control systems due to their lack of

real-time performance and dynamic wireless conditions in

industrial plants. This article reviews a series of recent

advances in real-time WSANs for industrial control systems:

1) real-time scheduling algorithms and analyses for WSANs;

2) implementation and experimentation of industrial WSAN

protocols; 3) cyber-physical codesign of wireless control

systems that integrate wireless and control designs; and

4) a wireless cyber-physical simulator for codesign and

evaluation of wireless control systems. This article concludes

by highlighting research directions in industrial cyber-physical

systems.

KEYWORDS | Cyber-physical systems; industrial wireless net-
works; real-time systems; wireless control systems; wireless

sensor-actuator networks

I . INTRODUCTION

Wireless sensor-actuator networks (WSANs) are gaining
rapid adoption in process industries due to their advantage

in lowering deployment effort in harsh industrial environ-

ments. Industrial standard organizations such as ISA [1],

HART [2], WINA [3], and ZigBee [4] have been actively

pushing the application of wireless technologies in

industrial automation and manufacturing. While early

success of industrial WSANs focused on monitoring

applications, there is significant value in exploring WSANs
for process control applications to take full advantage of

wireless technology in industrial plants.

However, wireless control systems face unique chal-

lenges that distinguish them from traditional control

systems. First, it is challenging to meet the stringent

latency requirements of feedback control in WSANs. IEEE

802.15.4 radios commonly used in WSANs have a

maximum bandwidth of only 250 kbps. Multihop commu-
nication over mesh networks further increases communi-

cation delays. Furthermore, channel conditions in

industrial environments can change dynamically due to

Manuscript received June 6, 2015; revised September 17, 2015; accepted October 23,

2015. Date of publication December 7, 2015; date of current version April 19, 2016. This

work was supported in part by the U.S. National Science Foundation under Grants

1320921 (NeTS), 1035773 (CPS), 1017701 (NeTS), and 1144552 (NeTS), and by the

National Natural Science Foundation of China (NSFC) under Grant 61273038.

C. Lu, B. Li, D. Gunatilaka, and Y. Chen are with the Department of Computer Science
and Engineering, Washington University in St. Louis, Saint Louis, MO 63130 USA

(e-mail: lu@wustl.edu; b.li@wustl.edu; dgunatilaka@wustl.edu; ychen25@wustl.edu).

A. Saifullah is with the Department of Computer Science, Missouri University of
Science and Technology, Rolla, MO 65409 USA (e-mail: saifullaha@mst.edu).

M. Sha is with the Department of Computer Science, State University of New York at
Binghamton, Binghamton, NY 13902 USA (e-mail: msha@binghamton.edu).

H. Gonzalez is with the Department of Electrical and System Engineering, Washington
University in St. Louis, Saint Louis, MO 63130 USA (e-mail: hgonzale@wustl.edu).

C. Wu is with Yahoo!, Sunnyvale, CA 94089 USA (e-mail: wuchjie@gmail.com).
L. Nie is with the Department of Computer Science and Technology, Harbin Institute of
Technology, Harbin 150001, China (e-mail: nls@hit.edu.cn).

Digital Object Identifier: 10.1109/JPROC.2015.2497161

0018-9219 � 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

Vol. 104, No. 5, May 2016 | Proceedings of the IEEE 1013

external interference, moving obstacles and weather
conditions. As a result, WSANs demand real-time sched-

uling algorithms and analysis techniques that are both

effective and efficient. In contrast to wired industrial

networks (e.g., control area networks [5]) with well

established real-time scheduling techniques, there have

been limited results on real-time scheduling for WSANs.

The lack of analytical methods for achieving real-time

performance in WSANs hinders their adoption in control
systems. Wireless control demands a new real-time

scheduling theory for WSANs [6].

Second, design of wireless control systems must deal

with interdependencies between control and communica-

tion. For example, while it is well known in digital control

that a low sampling rate usually degrades control

performance [7], a high sampling rate may increase

resource contention in bandwidth-constrained WSANs
leading to long communication delays, which again may

lead to degraded control performance [8]. The coupling

between wireless communication and control therefore

motivates a cyber-physical codesign approach that integrates
wireless networks and control designs.

This article provides a review of recent advances in real-

time WSANs for industrial control systems, an increasingly

important class of cyber-physical systems (CPS) in the dawn
of Industrial Internet [9] and Industry 4.0 [10].

Section II describes real-time scheduling for industrial

WSANs. We first introduce WirelessHART, an industrial

standard for wireless sensing and actuation. We then

present a suite of real-time scheduling algorithms, delay

analyses and protocol implementation to achieve real-time

performance in WirelessHART networks. Section III

introduces our recent efforts to design industrial wireless
control systems following a cyber-physical codesign

approach. We first study the problem of selecting sampling

rates to optimize control performance. We then investigate

how to incorporate emergency alarms in wireless control

systems. To support the design and evaluation of wireless

control systems, we introduce the wireless cyber-physical

simulator (WCPS) for holistic studies of wireless control

systems. We conclude this article by highlighting prom-
ising research directions.

II . REAL-TIME INDUSTRIAL WSANs

We first provide an overview of the WirelessHART

standard and then describe our implementation and

experimentation of a WirelessHART protocol stack.

Finally, we summarize our real-time scheduling algorithms
and analyses for WirelessHART.

A. WirelessHART
A wireless control system comprises feedback control

loops connecting sensors, controllers and actuators

through a wireless mesh network. Sensors measure

variables of the plant and send the measurements to a

controller over the wireless mesh network. The controller

then sends control commands to the actuators in order to

control the physical processes. Industry plants pose harsh

environments for wireless communication due to signif-
icant channel noise, physical obstacles, multipath fading,

and interference from coexisting wireless devices [11].

These harsh and dynamic environments make it difficult

for WSANs to meet the stringent reliability and real-time

requirements of industrial control applications.

There exist multiple industrial WSAN standards, e.g.,

WirelessHART [2], ISA100 [1], WINA [3], and ZigBee [4].

In this article we will focus on the WirelessHART
standard, although the design principles may be extended

for the other standard technologies. WirelessHART

devices are now adopted worldwide for industrial process

management and control [13], [14]. Unique features that

make WirelessHART particularly suitable for industrial

process control are described below.

A WirelessHART network is managed by a centralized

network manager. The network manager is responsible for
the management, scheduling, creating the routes, and

optimizing the network. Network devices include a

Gateway, a set of field devices, and several access points

as shown in Fig. 1. The network manager and the

controllers are installed or connected to the Gateway.
The field devices are wireless sensors and actuators. Each
field device is equipped with a half-duplex omnidirectional

radio transceiver compliant with the IEEE 802.15.4
standard. Multiple access points are wired to the Gateway
to provide redundant paths between the wireless network

and the Gateway.

A WirelessHART network adopts various mechanisms

to ensure reliable communication in unreliable industrial

environments. Time is globally synchronized and slotted

Fig. 1. Architecture of a WirelessHART network (Credit: HART
Communication Foundation [12]).

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

1014 Proceedings of the IEEE | Vol. 104, No. 5, May 2016

using a built-in time synchronization protocol. Each time
slot is 10 ms, which is sufficient to send or receive one

packet and an corresponding acknowledgement. Transmis-

sions are scheduled based on a multichannel time division

multiple access (TDMA) protocol. A time slot can be either

dedicated or shared. In a dedicated slot, only one sender is
allowed to transmit to the receiver. In a shared slot, more
than one sender can attempt to transmit to the same

receiver. Since shared slots are assigned to multiple
senders, collisions may occur within a shared slot. To

reduce collisions, senders adopt carrier sense multiple
access with collision avoidance (CSMA/CA) to contend in a
shared slot. CSMA/CA allows multiple nodes to share a

common channel through carrier sensing, where nodes

attempt to avoid collisions by transmitting only when the

channel is idle.

The network uses 16 channels defined in IEEE 802.15.4,
and adopts channel hopping in every time slot. Channel
hopping provides frequency diversity to mitigate interfer-

ences and reduce multipath fading effects. Any excessively

noisy channel is blacklisted.
A WirelessHART network supports two types of routing

approaches: source routing and graph routing. Source
routing provides a single directed path for routing from a
source to a destination device. Graph routing involves a
routing graph consisting of a directed list of paths between

the two devices, and is adopted for enhanced end-to-end

reliability. In graph routing, packets from field devices are

routed to the Gateway through the uplink graph. To every
field device, there is a downlink graph from the Gateway.
The end-to-end communication between a source (sensor)

and destination (actuator) happens in two phases. In the

sensing phase, on one path from the source to the Gateway
in the uplink graph, the scheduler allocates a dedicated slot

for each device starting from the source, followed by

allocating a second dedicated slot on the same path to

handle a retransmission. Then, to offset failure of both

transmissions along a primary link, the scheduler allocates

a third shared slot on a separate path to handle another

retry. Then, in the control phase, using the same method,
the dedicated links and shared links are scheduled in the
downlink graph of the destination.

B. Implementation and Experimentation
To study and evaluate WSAN protocols, we have

developed an experimental WSAN testbed [15]. The system

comprises a network manager on a server and a network

protocol stack implementation on TinyOS 2.1.2 [16] and

TelosB motes [17]. Each mote is equipped with a TI
MSP430 microcontroller and a TI CC2420 radio compat-

ible with the IEEE 802.15.4 standard. Fig. 2 shows the

motes deployment in our campus buildings. In an

experiment, our motes can be designated as access points

and field devices, which form a multihop wireless mesh

network running WSAN protocols. To support experi-

mentation and measurements, all motes in our system are

physically connected to a wired backplane network that

can be used for managing wireless experiments and

measurements without interfering with wireless com-

munication.

Our network manager implements a route generator

and a schedule generator. The route generator is

responsible for generating source routes or graph routes
based on the collected network topology, while the

schedule generator is responsible for generating packet

transmission schedules. Our protocol stack adopts the

CC2420x radio driver [19] as the radio core, which is

responsible for transmitting and receiving packets. A

multichannel TDMA MAC protocol, RT-MAC, is imple-
mented on top of the radio core. As shown in Fig. 3, RT-

MAC divides time into 10 ms slots based on the
WirelessHART standard. The clocks of all the motes in

the network are synchronized using Flooding Time

Synchronization Protocol (FTSP) [20] during a Sync
window. The network then transmits packets based on

recurring superframes (transmission schedules). A 2 ms

guard time is reserved in the beginning of each slot to

accommodate the clock synchronization error and channel

switching delay. Both dedicated and shared slots specified
in the WirelessHART standard are supported by RT-MAC.

Only one sender is allowed to transmit and the packet

transmission occurs immediately after the guard time in a

dedicated slot, while more than one sender can contend

for the channel in a shared slot.

Fig. 2. WSAN testbed in Bryan Hall and Jolley Hall of Washington
University in St. Louis [18].

Fig. 3. Time frame format of RT-MAC.

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

Vol. 104, No. 5, May 2016 | Proceedings of the IEEE 1015

We have performed a series of experiments on our
WSAN testbed [15]. For instance, we have conducted a

comparative study of the two alternative routing ap-

proaches adopted by WirelessHART, namely source

routing and graph routing, and investigated the tradeoff

among reliability, latency, and energy consumption under

the different routing approaches.

We first run our experiments in a clean environment

where we blacklist the four 802.15.4 channels overlapping
with our campus Wi-Fi network and run the experiments

on the remaining 802.15.4 channels. We then repeat our

experiments in a noisy environment where we configure

the network to use channels overlapping with our campus

Wi-Fi network and in a stress-testing environment where

we generate controlled Wi-Fi interference. In our study,

we use reliable links with PRR higher than 80%, where

PRR stands for Packet Reception Ratio, i.e., the fraction of
transmissions that are successful over a wireless link. We

set up 8 data flows, and run our experiments long enough

such that each flow delivers at least 500 packets from its

source to its destination. We measure the packet delivery

rate (PDR) of a flow, defined as the percentage of packets

that are successfully delivered to their destination. Table 1

shows the minimum, median and maximum PDR among

all the flows in the clean, noisy and stress-testing
environments. Graph routing improves the median PDR

by 15.9%, and 21.4% compared to source routing in noisy

and stress testing. Moreover, graph routing drastically

improves the minimum PDR, with 35.5% and 63.5%

increase in noisy and stress testing environments. This

result shows graph routing can effectively enhance the

predictability of real-time performance, especially under

significant interference. However, our study also shows
that route diversity incurs costs for latency and energy

consumption, with graph routing suffering from an

average of 80% increase in end-to-end latency and

consuming an average of 130% more energy over source

routing.

Our study concludes that it is important to employ

graph routing algorithms specifically designed to optimize

latency and energy efficiency. This problem was recently
explored in [21] and [22], which proposed a set of real-time

and energy-efficient routing algorithms for WirelessHART

networks. Both source routing and graph routing have

been implemented in our WSAN testbed and the Wireless

Cyber-physical Simulator (WCPS) (to be discussed in

Section III-B), which enable systematic comparison between

the alternative routing approaches.

C. Real-Time Scheduling
To save cost and enhance flexibility, it is desirable to

support many control loops in a same network. The

stability and control performance of these systems heavily

depends on real-time communication over the shared

wireless network. Feedback control loops in a WSAN

therefore impose strict requirements on reliability and

real-time guarantees in wireless communication in order

to avoid plant shutdowns and/or accidents. For example, in
oil refineries, spilling of oil tanks has to be avoided by

controlling oil level in real-time. However, industrial

plants pose a harsh environment for wireless communica-

tion due to unpredictable channel conditions, limited

bandwidth, physical obstacles, multipath fading, and

interference from coexisting wireless devices [11]. With

the adoption of industrial wireless standards such as

WirelessHART [2], process monitoring and control
applications have seen the feasibility of achieving reliabil-

ity and real-time wireless communication through spatial

and spectrum diversity. Real-time communication in these

wireless networks pose new and important challenges.

Unlike wired networks, there have been limited results on

real-time scheduling theory for wireless networks. Real-

time transmission scheduling and analysis for WSANs

require new methodologies to deal with unique wireless
characteristics.

1) Overview of State of the Art: Real-time communication
in wireless networks has been explored in many earlier

efforts [23]–[34], [34]–[40]. The survey in [41] provides a

comprehensive review of these works. However, these are

not suitable for industrial applications that usually need

multichannel communication, multipath routing and real-
time performance analysis results. Earlier works [42]–[46]

on real-time performance analysis for wireless sensor

networks focused on data collection through a routing tree

[42], [45].

Real-time WSAN for control purposes has been studied

for single hop networks in [47]–[52]. For multihop

wireless networks, a mathematical framework to model

and analyze schedules for WirelessHART networks has
been proposed in [53]. Real-time scheduling for Wireless-

HART networks has received attention in recent works [6],

[54]–[60]. These scheduling policies can be broadly

classified into two categories: fixed priority scheduling

and dynamic priority scheduling [61]. In a fixed priority
scheduling policy, each data flow has a fixed priority, and
every transmission has the priority of its flow. Due to its

simplicity and low scheduling overhead, fixed priority
scheduling is a common class of real-time scheduling

policies in practice [61]. On the other hand, dynamic
priority scheduling policy refers to the policy where there is
no fixed priority, i.e., priorities of transmissions during the

scheduling change dynamically according to some chosen

criterion. While such policies have higher scheduling

overhead, they provide better schedulability since

TABLE 1 Minimum, Median, and Maximum pdr Among All Flows

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

1016 Proceedings of the IEEE | Vol. 104, No. 5, May 2016

priorities are changed dynamically to enhance the
schedulability.

In the following we summarize our recent results on

real-time scheduling for industrial WSANs under fixed

priority and dynamic priority scheduling, respectively.

2) Fixed Priority Scheduling: Delay analysis can be used to
determine whether real-time data flows can meet their

deadlines. When used offline, the analysis helps users
design a WSAN to meet the real-time performance

requirements. For example, we used schedulability analysis

to select the sampling rates of wireless control loops to

optimize control performance under communication dead-

lines [62]. In [57], we used schedulability analysis to assign

priorities to real-time flows in order to meet their

deadlines. When used online, the network manager can

adjust the workloads in response to network dynamics.
Industrial environments can cause frequent changes in

network topologies and channel conditions. If a WSAN can

no longer guarantee the deadlines for all the flows based on

delay analysis, the network manager may stop or adjust the

data rates of noncritical flows in order to maintain real-time

guarantees to the remaining flows. The Wireless Cyber-

Physical Simulator (WCPS) introduced in Section III-B can

be used to assess the pessimism of delay analysis and
wireless control performance under realistic settings.

For fixed priority scheduling in WirelessHART net-

works, we proposed a suite of delay analyses for real-time

flows [55], [56]. These analyses determine upper bounds

on the end-to-end communication delays and provide

sufficient conditions for schedulability. Two factors con-

tribute to the communication delay of a control loop. A

lower priority flow can be delayed by higher priority flows:
a) due to channel contention (when all channels are assigned
to transmissions of higher priority flows in a time slot); and

b) due to transmission conflicts (when a transmission of the
flow and a transmission of a higher priority flow involve a

common node). A key insight underlying this analysis is to

map the real-time transmission scheduling in Wireless-

HART networks to real-time multiprocessor scheduling. By

incorporating the unique characteristics of WirelessHART
networks into the state-of-the-art worst case response time

analysis for multiprocessor scheduling, the analysis can

efficiently compute a safe and tight upper bound of the end-

to-end delay of every flow.

Priority assignment is an important problem in fixed-

priority scheduling as it has a significant impact on the

schedulability of real-time flows. An ideal priority assign-

ment should not only enable real-time flows to meet their
deadlines, but also work synergistically with real-time

schedulability tests to support effective network capacity

planning and efficient online admission control and

adaptation. In [57] we studied optimal and near optimal

fixed priority assignment using local search. A salient

feature of this local search approach is that it exploits the

delay bounds provided by a schedulability analysis to

efficiently search for feasible priority assignments by
discarding unnecessary branches in the search space.

3) Dynamic Priority Scheduling: Dynamic priority sched-
uling of transmissions was studied in [58] and [60] for

WirelessHART networks of with tree topologies. Our work

in [54] and [59] studied dynamic priority scheduling of

transmissions for general WirelessHART networks. The

work in [54] has shown that the real-time transmission
scheduling for WirelessHART networks is NP-hard. We

observed that transmission conflicts play a major role in the

communication delays and the schedulability of the control

loops, which makes the traditional real-time scheduling

policies such as least laxity first (LLF) less effective. We

propose an optimal local search scheduling algorithm that

exploits the necessary condition to effectively discard

infeasible branches in the search space. We also proposed a
faster heuristic called conflict-aware least laxity first

(C-LLF) for dynamic priority scheduling. The algorithm

identifies the critical time windows in which too many

conflicting transmissions have to be scheduled, thereby

determining the criticality of each transmission. Criticality

of a transmission is quantified by its conflict-aware laxity,
which is its laxity after discarding the (estimated) time slots

that can be wasted through waiting to avoid transmission
conflict. Transmissions exhibiting the lowest conflict-

aware laxity are assessed to be more critical. C-LLF gives

the highest priority to the transmissions exhibiting lower

conflict-aware laxity. Thus C-LLF integrates LLF and the

degree of conflicts associated with a transmission, and

outperforms traditional real-time scheduling policies.

Our work in [59] has provided a schedulability analysis

for earliest deadline first (EDF), a common dynamic
priority scheduling in WirelessHART networks.

We evaluated our schedulability analyses against

experimental results on our WSAN testbed [18] as well

as in simulations. In our evaluation, both fixed-priority

scheduling [57] and dynamic priority scheduling [54]

policies outperformed the traditional real-time scheduling

policies by significant margins. All experimental results

and simulations showed that our analyses provide safe
upper bounds of communication delays in the network and

enable effective schedulability tests [56]. For tighter upper

bounds under graph routing, we established probabilistic

delay bounds in [6] that represent upper bounds with

probability � 0.90. The worst-case and probabilistic
bounds can be used in different application scenarios

depending on the level of predictability required. Our

analysis hence can be used for an effective schedulability
test and admission control of real-time flows in Wireless-

HART networks.

III . WIRELESS CONTROL CODESIGN

Due to limited resources in a WSAN shared by multiple

control loops, it is critical to optimize the overall control

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

Vol. 104, No. 5, May 2016 | Proceedings of the IEEE 1017

performance. However, in a wireless control system, the
control performance not only depends on the control

algorithms, but also relies on real-time communication

over the shared wireless network. The coupling between

real-time communication and control requires a cyber-

physical codesign approach for a holistic optimization of

control performance. This section summarizes our recent

efforts on cyber-physical codesign for industrial wireless

control systems. We first study the problem of selecting
sampling rates to optimize the control performance of

multiple feedback control loops sharing a WSAN. To

support cyber-physical codesign of wireless control

systems, we present wireless cyber-physical simulator

(WCPS), an integrated simulation environment for holistic

studies of wireless control systems. We then investigate

how to incorporate emergency alarms in wireless process

control systems. We wrap up this section with a case study
of wireless process control for coupled water tanks using

both regular and emergency control loops.

A. Rate Selection
In control systems, the impacts of sampling rate have

been studied for robot control [63], disturbance rejection

[64], and various control performance indices [65]–[69].
The choice of sampling rates needs to balance control and

communication. While a high sampling rate is desirable

from a pure digital control perspective, it also has the

undesirable effect of heavier network load and longer

communication delay, which can also degrade the control

performance.

The control performance can be quantified as a

function of sampling rates using the formulation proposed
by Seto et al. [65], which characterizes the performance
difference between the continuous-time control system

and that of its digital implementation.

In [62] and [70] we apply these ideas to optimize the

performance of a set of feedback control loops over a

WSAN, where the control cost of the i-th control loop at the
sampling rate fi is approximated as �i e

��i fi , where
�i; �i > 0 are the magnitude and decay rate coefficients,
respectively. Given the sampling rates of all of the control

loops ff1; f2; . . . ; fng, the total control cost of the system is
defined by

Xn

i¼1
wi �i e

��i fi (1)

where wi is the weight of the i-th control loop. We then use
the total control cost as the performance index of the

system and as the objective of an optimization problem to

determine the best sampling rates of a multiloop control

system sharing a WirelessHART network under stability

and delay constraints. In particular, the objective is to

minimize the cost in (1) subject to two constraints. First,

the delay bound Ri of the i-th loop, which can be computed
based on the sampling rate in [55] and [56], must be

smaller than its predetermined deadline Di. Second, each
control loop must have a sampling rate fi within its
minimum and maximum rates, denoted f mini and f

max
i

respectively, to ensure stability. Thus, we mathematically

formulate this optimization problem as follows:

min
ffigni¼1

Xn

i¼1
wi �i e

��i fi

subject to : Riðf1; . . . ; fnÞ � Di; 8 i 2 f1; . . . ; ng;
f mini � fi � f

max
i ; 8 i 2 f1; . . . ; ng: (2)

The resulting constrained optimization problem is

challenging since the delay Ri is nondifferentiable,
nonconvex, and not in closed-form. We explored four

methods to solve this problem: a) a subgradient-descent

algorithm; b) a greedy heuristic; c) a penalty method using
simulated annealing (SA); and d) a convex relaxation

method based on a new delay bound that is convex and

smooth but more pessimistic than previous analyses [55],

[56] using a different approach. Notably, the convex

relaxation greatly simplifies the optimization problem,

since the overall problem becomes convex and differen-

tiable by simplifying the delay bound analysis.

In [62], we evaluate the methods through simulations
based on the topology of our WSAN testbed [18]. The

results demonstrate that SA achieves the lowest control

cost but requires the longest execution time. Both the

greedy heuristic and the subgradient method lead to high

control costs because the optimization problem is highly

nonlinear with a large number of local extrema. For

example, for 30 control loops, the greedy heuristic incurs a

control cost up to 2.67 times that of SA. In contrast, the
convex relaxation approach using an interior point method

incurs a control cost no more than 35% higher than that of

SA, while incurring a much shorter execution time. The

convex relaxation approach therefore achieves the desir-

able balance between control performance and run-time

efficiency. This result demonstrates the significant

advantage of cyber-physical codesign that integrates

control optimization and schedulability analysis in
wireless control systems.

B. WCPS
Existing wireless control system research often relies

on lab-scale equipment and simulations. However, lab-

scale equipment usually suffers from limited physical size,

which cannot capture delays and data loss in real WSANs.

Simulation tools for control systems often lack realistic

models of WSANs that exhibit complex and stochastic

behavior in real-world environments. The lack of tools that

can capture both the cyber wireless network and physical

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

1018 Proceedings of the IEEE | Vol. 104, No. 5, May 2016

aspects of control systems has been a hurdle to wireless
control research.

Early experimental work on wireless control [71]–[76]

usually relied on a lab testbed where wireless sensors are

within a single hop and experience no data loss due to the

physical proximity of the devices. The challenge in realistic

experimentation with wireless control systems motivated

the development of simulation tools for wireless control

systems. For example, NCSWT [77] and Gisso [78] are two
simulators designed for wireless control systems. Truetime

[7] is a well established control system simulator capable of

holistic studies of CPU scheduling, communication, and

control algorithms. Unfortunately, none of these simula-

tors provides a realistic wireless radio model or a state-of-

the-art WSAN protocol stack.

Wireless cyber-physical simulator (WCPS) [79] is

designed to provide a holistic and realistic simulation of
wireless control systems. WCPS employs a federated

architecture that integrates: a) simulink for simulating the

physical system dynamics and controllers; and b) TOSSIM

for simulating WSANs. Simulink is commonly used by

control engineers to design and study control systems, while

TOSSIM [80] has been widely used in the sensor network

community to simulate WSANs based on wireless link

models that have been validated in diverse real-world
environments [81]. WCPS provides an open-source middle-

ware to orchestrate simulations in Simulink and in TOSSIM.

WCPS 2.0 implements the WirelessHART network

protocol stack at the routing and MAC layers [82]. To

support WirelessHART we also extended TOSSIM to

simulate wireless communication over multiple channels.

We have implemented both Source Routing and Graph
Routing as specified in the WirelessHART standard. To our
knowledge, WCPS 2.0 is the first simulator that supports

WirelessHART protocols based on a realistic wireless link

model.

Thanks to WCPS we have been able to develop case

studies that simulate wireless control systems for civil

infrastructure and process plants. In [83] we simulated two

wireless structural control systems. Wireless structural

control systems offer an attractive approach to protect civil
infrastructures from natural hazards such as earthquakes

and other natural disasters. In the first simulation we

studied a benchmark building model, where the wireless

traces were collected in a multistory building. In the

second simulation we studied the structural model of the

Cape Girardeau bridge over the Mississippi River, where

the wireless traces were collected from a similar bridge in

South Korea. These case studies shed light on the
limitations of traditional structural control approaches

under realistic wireless conditions. They further allowed

us to validate the advantages of cyber-physical codesign

algorithms for wireless communication protocols [84],

[85]. Based on our experience with WCPS, we have

recently enhanced the wireless building control study,

transforming it into a benchmark that can be used by the

structural control community to explore and evaluate
different wireless control approaches [86], allowing

practitioners to easily generate and configure realistic

nuisances such as network induced delay, data loss,

measurement noise, and control constraints.

C. Emergency Communication
In [82], we considered the case of WirelessHART

control networks comprised of: a) regular flows, which are
periodically generated and typically stabilize a desired part

of a physical system; and b) emergency flows, which are
infrequent and typically signal that an unsafe situation is

about to occur. Both regular and emergency flows have

predetermined deadlines to transmit a packet. Since

emergency flows carry information related to unsafe

situations, these flows have a higher criticality than

regular flows. Therefore, we proposed the following
approach to transmit packets through the network: a) in

regular mode, i.e., when there is no emergency, all regular
flows should meet their deadlines; and b) in emergency
mode, i.e., when an emergency occurs, all existing
emergency flows should meet their deadlines, while

regular flows are delivered on a best-effort basis.

Periodic scheduling (PS) is a baseline scheduling

approach that reserves periodic slots dedicated to emergen-
cy packets in the transmission schedule. Emergency packets

can be transmitted only in the reserved slots. When there is

no emergency, the reserved slots are left unused. We take a

two-level priority assignment approach for scheduling,

where we first schedule emergency flows to meet their

deadlines, and then we schedule the regular flows using the

remaining capacity. Within each group, flows are scheduled

using a rate monotonic policy. This basic approach is simple
in design, yet it wastes (a potentially large amount of)

network bandwidth, reserving time slots for emergency

flows even when there is no emergency.

To avoid wasting resources when no emergencies

occur, we designed a slot stealing (SS) method that allows
emergency flows to ‘‘steal’’ slots from regular schedules

when emergencies arise, removing the need to allocate

exclusive time slots for emergency flows. The ‘‘stealing’’
process is implemented by adding a random backoff and a

clear channel assessment (CCA) to the transmission of all

regular flow packets, and allowing emergency packets to

be transmitted immediately at the beginning of the time

slot. Hence, if an emergency packet is available, then it

will take the slot, and any regular flow packet trying to use

that slot will fail the CCA, forcing it to wait for the next

available time slot or drop the packet. By allowing the
overlapping emergency and regular schedules, SS is able to

schedule the same number of flows in a shorter time frame

when compared to PS, while ensuring timely delivery of

emergency packets in case of an emergency.

We also explored two alternative approaches to send

emergency signals. For systems that need to periodically

monitor and control the emergency state, an emergency

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

Vol. 104, No. 5, May 2016 | Proceedings of the IEEE 1019

flow is activated whenever emergencies occur, and the
emergency flow then periodically generates data until the

emergency is over. For systems that do not need to

periodically monitor and control the emergency state, the

system can adopt an event-based approach to communicate

the emergency alarms, i.e., an emergency sensor only

sends one alarm-start packet and one alarm-stop packet at

the beginning and end of the emergency, respectively.

While the event-based communication results in the same
transmission schedule, it significantly reduces the number

of regular transmissions that are dropped or delayed by

emergency transmissions, leading to better control

performance. In [82], simulation results showed the

combination of slot stealing and event-based emergency

communication produced the best results in terms of

emergency handling and control performance.

D. Emergency Control Case Study
We tested our emergency communication protocol on a

set of two identical water tank systems sharing a common
wireless network simulated using WCPS 2.0. A diagram of

each of water tank systems is shown in Fig. 4. Our choice is

based on the simple, yet representative, dynamics of water

tanks, their hybrid dynamical nature (since the evolution

of the system changes when the water tanks are either full

or empty), and more importantly, its similarity to systems

commonly used in industrial applications. Also, we

simulated two identical water tank systems sharing a
wireless network to study the effect of one system’s

emergency over the second system’s regular flow.

Following the software architecture in WCPS 2.0, the

sensor data generated by Simulink is fed into the WSAN

simulated using TOSSIM. TOSSIM then returns the

packets delayed or dropped according to the behavior of
the network, which are then fed to the controller

implemented in Simulink. Controller commands are

then fed again into TOSSIM, which delays or drops the

packets and sends the outputs to the actuators in the water

tanks, closing the loop.

For this study we collect wireless traces from 21 nodes

in a WSAN testbed [18]. We then use the wireless traces as

inputs to the TOSSIM simulator to generate realistic
wireless characteristics in a simulated WSAN. The route in

the simulated WSAN is 6 hops. We used further adjusted

received signal strength (RSS) to test the system perfor-

mance with different wireless signal strengths.

On the control side, we used a PID controller to

regulate the water level of Tank 2, denoted L2, by actuating
the Pump taking water from the Basin, as shown in Fig. 4.

If no emergencies occur, then the Valve stays closed. We
also defined a discrete emergency controller setting the

values of the Pump and Valve, with the objective of

avoiding water spillage. Hence, if any of the emergency

level sensors, denoted LH1 , L
H
2 , L

L
b, and L

H
b , detect a

dangerous water level, then an emergency signal is sent

and the controller switches from the PID controller to the

discrete emergency controller, switching back when the

emergency is cleared. We defined a system failure using
two principles: a) if the system cannot stabilize at a regular

configuration (i.e., within nonemergency water levels) in a

fixed 100 second interval; and b) if any of the water levels

exceed the height of its associated water tank

Our simulations show that the combination of the slot-

stealing and event-based emergency communication is

highly effective in avoiding system failures, since it

produces a tighter schedule that results in a faster update
frequency of the regular control flows. It also reduces

communication load and mitigates the impact of emer-

gency communication on regular control flows. Our case

study demonstrates that even for a 6-hop lossy wireless

network (with 5.8% median packet loss), successful

system control can still be achieved through a combination

of regular and emergency control.

IV. RESEARCH DIRECTIONS

A. Enhancing Scalability
A major limitation of current industrial WSANs is their

limited scalability due to the centralized network archi-

tecture. In WirelessHART, when a node or link fails, the

centralized network manager must generate a new global
TDMA schedule, which requires the network manager to

collect the current topology of the network, create a new

schedule, and distribute the schedule among all field

devices. As WSANs are subject to frequent changes in

channel quality and link condition, global changes to the

network schedule create excessive communication

overhead in a large network. As a result, while the

Fig. 4. Diagram of the coupled water tank system. The water levels
of Tank 1, Tank 2, and Basin are denoted L1, L2, and Lb, respectively.

The water emergency levels are denoted LH1 , L
H
2 , L

L
b, and L

H
b . The water

flows between tanks are denoted u, v12, v1b, and v2b. The state of the

valve is denoted d 2 f0; 1g, where d ¼ 1 if the valve is open.

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

1020 Proceedings of the IEEE | Vol. 104, No. 5, May 2016

centralized architecture has proven to be sufficient for
small-scale installations, it can become a significant

limitation as WSANs start to be deployed over large

geographic areas (e.g., thousands of devices over an oil

field). A key research direction is to make WSANs scalable

while achieving end-to-end real-time communication.

A promising approach to enhance scalability is through

a hierarchical network architecture, where a large WSAN

is divided into multiple subnetworks. Each subnetwork
employs its own manager to manage local operations, and a

global manager coordinates with the subnetwork managers

to manage the entire network in a hierarchical fashion. An

advantage of this architecture is that it can meet industrial

needs for both network visibility and scalability. As a

submanager may deal with the wireless dynamics within

its subnetwork, the hierarchical architecture can scale

effectively. A challenge in designing a hierarchical
architecture is to deal with the interdependencies among

the subnetworks that share the wireless spectrum and need

to support flows traversing multiple subnetworks subject

to end-to-end deadlines.

B. Exploring White Spaces
Another limiting factor of today’s WSANs stems from

the short communication range of the IEEE 802.15.4
radios adopted by industrial standards such as Wireless-

HART. To overcome the short communication range,

many WSANs form multihop mesh networks resulting in

long communication delays and limited scalability. An

opportunity to enhance the scalability of real-time WSANs

arises from the opening of a new spectrum resulting from

the transition to digital TV broadcasting globally and

freeing up the VHF/UHF spectrum. White spaces refer to
the allocated but locally unused TV spectrum. Since TV

transmissions are in lower frequencies, white space

transmissions have excellent propagation characteristics

over long distance. They can easily penetrate walls and

other obstacles, and hence hold enormous potential for

industrial applications that need real-time communication

over large geographic areas. Thanks to its long communi-

cation range and wall-penetration capability, a white space
network will have small hop counts and will drastically

reduce communication delays and protocol complexities.

While white spaces are mostly being tapped into for

wireless broadband access to date, they also open up the

opportunity to support highly scalable real-time industrial

applications that have been challenging under existing

WSAN technologies [87]. The characteristics and the

application demands in industrial sensing and actuation
pose unique challenges in adopting white spaces for

industrial WSANs. Instead of high-throughput traffic, a

WSAN should exploit the wide spectrum in white spaces to

support low-data-rate communication from numerous field

devices. Furthermore, a WSAN over white spaces must

achieve high degrees of energy efficiency that is compara-

ble to today’s WSAN technologies. It also needs to handle

changes and variations in spectrum availability in white
spaces while maintaining desired real-time performance.

C. Cyber-Physical Codesign
While earlier work has shown the promise of cyber-

physical codesign in wireless control systems through

point designs and case studies, there exist opportunities to

develop a theory and practice of cyber-physical codesign

through a broader exploration of the interaction between
the network protocol stack and control design. For

example, if a controller employs a state observer to

estimate system states based on intermittent observations,

the controller can tolerate a certain degree of data loss

from the sensors. A WSAN can exploit the controller’s

resiliency against sensor data loss by reducing the route

redundancy from sensors to controllers, while dedicating

more network resources to enhance the reliability of
communication to actuators. That is, the allocation of

network resources should be dependent on the control

design. Conversely, an actuator may buffer the control

inputs from a model predictive controller and use

previously received control inputs when the network fails

to deliver the new control inputs. When the wireless

condition deteriorates causing more data loss, the

controller may increase its sampling rates or control
horizon to increase its tolerance to data loss to the

actuators. That is, the controller should adapt to network

conditions. Note that the adaptation of network resources

and controller sampling rates can be designed using

approximated mathematical models, or using distributed

extensions of the data-driven methods in [88]. Therefore, a

wireless control system will need to codesign wireless

networks and control in an interdependent fashion. To
establish a unified cyber-physical codesign approach to

industrial wireless control, it will be crucial to develop

interfaces between the WSAN and the control components

to maintain optimal control performance by adapting to

changing network and physical dynamics in the wireless

control systems. We need to further develop a suite of

algorithms and analytical techniques to assure the safety of

the control system under dynamic conditions. Finally, the
research community will need long-term collaboration

among computer science, control theory and domain

experts for this inherently interdisciplinary research.

D. Cyber-Physical Testbed
The advancement and adoption of new WSAN

technologies have been hindered by the lack of testbeds

under realistic industrial settings. While cyber-physical
simulators such as WCPS are useful for studying wireless

control systems, they require realistic industrial plant

models and wireless traces collected from real-world

industrial settings. Despite best efforts on modeling

techniques, simulations cannot replace physical compo-

nents and networks with complex and sometimes unex-

pected dynamics. Unfortunately, full-scale industrial

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

Vol. 104, No. 5, May 2016 | Proceedings of the IEEE 1021

plants are often difficult to replicate for research due to

space and safety constraints. A step beyond simulations

might be hybrid testing technology that combines physical

components and simulations for hardware-in-the-loop

experiments. For example, a physical WSAN testbed may

be carefully integrated with simulations of an industrial

plant to study the impact of real wireless dynamics on

control systems. Conversely, an industrial plant may be

integrated with a WSAN simulator to evaluate potential

wireless control designs without a physical WSAN

deployment. This would require extending tools such as

WCPS to support hardware-in-the-loop testing where both

the WSAN and the plant can be replaced by physical

implementations. Close partnership between industry and

researchers will be critical to develop realistic cyber-

physical testbeds for industrial wireless control systems.

V. CONCLUSION

Real-time WSANs are poised to play a key role in industrial

automation in the era of Industry 4.0 [10] and Industrial
Internet [9]. Recent research and industrial developments

have demonstrated the promise of supporting real-time

communication and control over WSANs. This article

reviews recent advances in two related fronts: a) real-time

scheduling and analytical techniques for achieving real-

time performance in industrial WSANs and b) cyber-

physical codesign of wireless control systems. We highlight

the significant challenges and opportunities in cyber-
physical systems research that crosscut wireless and

control domains. Interdisciplinary collaboration and part-

nership among wireless and control researchers and

industry communities will be crucial in realizing the

vision of wireless industrial control. h

REF ERENCE S

[1] ISA100. [Online]. Available: https://www.isa.
org/isa100/.

[2] WirelessHART. [Online]. Available: http://
en.hartcomm.org/.

[3] Wireless Industrial Networking Alliance.
[Online]. Available: http://www.wina.org.

[4] ZigBee Alliance. [Online]. Available: http://
www.zigbee.org.

[5] R. I. Davis, A. Burns, R. J. Bril, and
J. J. Lukkien, ‘‘Controller area network (can)
schedulability analysis: Refuted, revisited and
revised,’’ Real-Time Syst., vol. 35, no. 3,
pp. 239–272, 2007.

[6] A. Saifullah, D. Gunatilaka, P. Tiwari,
M. Sha, C. Lu, B. Li, C. Wu, and Y. Chen,
‘‘Schedulability analysis under graph routing
for wirelesshart networks,’’ presented at the
RTSS, , 2015.

[7] A. Cervin, D. Henriksson, B. Lincoln, J. Eker,
and K.-E. Årzén, ‘‘How does control timing
affect performance? Analysis and simulation
of timing using Jitterbug and TrueTime,’’
IEEE Control Syst., vol. 23, no. 3, pp. 16–30,
Jun. 2003.

[8] F.-L. Lian, J. Moyne, and D. Tilbury,
‘‘Network design consideration for distributed
control systems,’’ IEEE Control Syst. Technol.,
vol. 10, no. 2, pp. 297–307, Mar. 2002.

[9] P. Evans and M. Annunziata, ‘‘Industrial
Internet: Pushing the boundaries of minds
and machines, General Electric (GE),
Nov. 26, 2012. [Online]. Available:
http://www.industrialinternetconsortium.
org/white-papers.htm.

[10] H. Kagermann, W. Wahlster, and J. Helbig,
‘‘Recommendations for implementing the
strategic initiative industrie 4.0: Final report
of the industrie 4.0 working group,’’
Apr. 2013, Tech. Rep.

[11] J. Song, A. K. Mok, D. Chen, and M. Nixon,
‘‘Challenges of wireless control in process
industry,’’ presented at the CRTES, 2006.

[12] WirelessHART Architecture. [Online].
Available: http://en.hartcomm.org/hcp/tech/
wihart/images/Expanded_WirelessHART_
Mesh_Network.jpg.

[13] Siemens, ‘‘Process instrumentation and
weighing technology.’’ [Online]. Available:
http://www.industry.usa.siemens.com/
automation/us/en/process-instrumentation-
and-analytics/process-instrumentation/
brochures/Documents/PIBR-0A940-1014-
one-stop-shop.pdf.

[14] HART Communication Foundation,
‘‘Mitsubishi chemical connects HART
technology to control room savings up to
/20–30,000 a day.’’ [Online]. Available:
http://en.hartcomm.org/hcp/tech/
applications/applications_success_
mitsubishi_chemical.html.

[15] M. Sha, D. Gunatilaka, C. Wu, and C. Lu,
‘‘Implementation and experimentation of
industrial wireless sensor-actuator network
protocols,’’ presented at the EWSN, 2015.

[16] TinyOS. [Online]. Available: http://www.
tinyos.net/.

[17] Memsic, TelosB Mote Platform. [Online].
Available: http://www.memsic.com/userfiles/
files/Datasheets/WSN/telosb_datasheet.pdf.

[18] The WUSTL Wireless Sensor Network
Testbed. [Online]. Available: http://wsn.cse.
wustl.edu/index.php?title=Testbed.

[19] CC2420x Radio Driver. [Online]. Available:
https://github.com/tinyos/tinyos-main/tree/
master/tos/chips/cc2420x.

[20] M. Maróti, B. Kusy, G. Simon, and Á. Lédeczi,
‘‘The flooding time synchronization protocol,’’
in Proc. 2nd Int. Conf. Embed. Netw. Sensor
Syst., 2004, pp. 39–49.

[21] C. Wu, D. Gunatilaka, A. Saifullah, M. Sha,
P. Tiwari, C. Lu, and Y. Chen, ‘‘Maximizing
network lifetime of wireless sensor-actuator
networks under graph routing,’’ All Comput.
Sci. Eng. Res., Tech. Rep. WUCSE-2015-004,
(2015). [Online]. Available: http://openscho-
larship.wustl.edu/cse_research/508/

[22] C. Wu, D. Gunatilaka, M. Sha, C. Lu, and
Y. Chen, ‘‘Conflict-aware real-time routing for
industrial wireless sensor-actuator
networks,’’ All Comput. Sci. Eng. Res., Tech.
Rep. WUCSE-2015-005, (2015). [Online].
Available: http://openscholarship.wustl.edu/
cse_research/507/

[23] H. Li, P. Shenoy, and K. Ramamritham,
‘‘Scheduling messages with deadlines
in multi-hop real-time sensor networks,’’
presented at the RTAS, 2005.

[24] X. Wang, X. Wang, X. Fu, G. Xing, and
N. Jha, ‘‘Flow-based real-time communication
in multichannel wireless sensor networks,’’
presented at the EWSN, 2009.

[25] V. Kanodia, C. Li, A. Sabharwal, B. Sadeghi,
and E. Knightly, ‘‘Distributed multi-hop
scheduling and medium access with delay and
throughput constraints,’’ presented at the
MobiCom, 2001.

[26] C. Lu, B. M. Blum, T. F. Abdelzaher,
J. A. Stankovic, and T. He, ‘‘RAP: A real-time
communication architecture for large-scale
wireless sensor networks,’’ presented at the
RTAS, 2002.

[27] K. Karenos, V. Kalogeraki, and
S. V. Krishnamurthy, ‘‘A rate control
framework for supporting multiple classes of
traffic in sensor networks,’’ presented at the
RTSS, 2005.

[28] K. Karenos and V. Kalogeraki, ‘‘Real-time
traffic management in sensor networks,’’ in
Proc. RTSS, 2006, pp. 422–434.

[29] T. He, B. M. Blum, Q. Cao, J. A. Stankovic,
S. H. Son, and T. F. Abdelzaher, ‘‘Robust and
timely communication over highly dynamic
sensor networks,’’ Real-Time Syst., vol. 37,
no. 3, pp. 261–289, 2007.

[30] G.-S. Ahn, A. T. Campbell, A. Veres, and
L.-H. Sun, ‘‘SWAN: Service differentiation in
stateless wireless ad hoc networks,’’ in Proc.
INFOCOM, 2002, vol. 2, pp. 457–466.

[31] T. He, J. A. Stankovic, C. Lu, and
T. Abdelzaher, ‘‘SPEED: A stateless protocol
for real-time communication in sensor
networks,’’ presented at the ICDCS, 2003.

[32] E. Felemban, C.-G. Lee, and E. Ekici,
‘‘MMSPEED: Multipath multi-SPEED
protocol for QoS guarantee of reliability
and timeliness in wireless sensor networks,’’
IEEE Trans. Mobile Comput., vol. 5, no. 6,
pp. 738–754, Jun. 2006.

[33] B. D. Bui, R. Pellizzoni, M. Caccamo,
C. F. Cheah, and A. Tzakis, ‘‘Soft real-time
chains for multi-hop wireless ad-hoc
networks,’’ presented at the RTAS, 2007.

[34] Y. Gu, T. He, M. Lin, and J. Xu,
‘‘Spatiotemporal delay control for
low-duty-cycle sensor networks,’’ presented at
the RTSS, 2009.

[35] A. Koubaa, M. Alves, and E. Tovar, ‘‘i-GAME:
An implicit GTS allocation mechanism in
IEEE 802.15.4 for time-sensitive wireless
sensor networks,’’ presented at the ECRTS,
2006.

[36] M. Caccamo, L. Y. Zhang, L. Sha, and
G. Buttazzo, ‘‘An implicit prioritized
access protocol for wireless sensor networks,’’
presented at the RTSS, 2002.

[37] K. Liu, N. Abu-Ghazaleh, and K.-D. Kang,
‘‘JiTS: Just-in-time scheduling for real-time
sensor data dissemination,’’ presented at the
PerCom, 2006.

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

1022 Proceedings of the IEEE | Vol. 104, No. 5, May 2016

[38] B. Li, D. Wang, F. Wang, and Y. Q. Ni,
‘‘High quality sensor placement for shm
systems: Refocusing on application demands,’’
presented at the INFOCOM, 2010.

[39] R. Mangharam, A. Rowe, R. Rajkumar, and
R. Suzuki, ‘‘Voice over sensor networks,’’
presented at the RTSS, 2006.

[40] T. Facchinetti, L. Almeida, G. C. Buttazzo,
and C. Marchini, ‘‘Real-time resource
reservation protocol for wireless mobile
ad hoc networks,’’ presented at the RTSS,
2004.

[41] J. A. Stankovic, T. F. Abdelzaher, C. Lu, L. Sha,
and J. C. Hou, ‘‘Real-time communication
and coordination in embedded sensor
networks,’’ Proc. IEEE, vol. 91, no. 7,
pp. 1002–1022, Jul. 2003.

[42] O. Chipara, C. Lu, and G.-C. Roman,
‘‘Real-time query scheduling for wireless
sensor networks,’’ presented at the RTSS,
2007.

[43] T. F. Abdelzaher, S. Prabh, and R. Kiran,
‘‘On real-time capacity limits of multihop
wireless sensor networks,’’ presented at the
RTSS, 2004.

[44] P. Jurcı́k, R. Severino, A. Koubâa, M. Alves,
and E. Tovar, ‘‘Real-time communications
over cluster-tree sensor networks with mobile
sink behaviour,’’ presented at the RTCSA,
2008.

[45] J. B. Schmitt and U. Roedig, ‘‘Sensor network
calculusVA framework for worst case
analysis,’’ presented at the DCOSS, 2005.

[46] N. Pereira, B. Andersson, E. Tovar, and
A. Rowe, ‘‘Static-priority scheduling over
wireless networks with multiple broadcast
domains,’’ presented at the RTSS, 2007.

[47] H.-J. Korber, H. Wattar, and G. Scholl,
‘‘Modular wireless real-time sensor/actuator
network for factory automation applications,’’
IEEE Trans. Ind. Inform., vol. 3, no. 2,
pp. 111–119, May 2007.

[48] X. Liu and A. Goldsmith, ‘‘Cross-layer design
of distributed control over wireless network,’’
in Advances in Control, Communication
Networks, and Transportation Systems,
E. H. Abed, Ed. Berlin, Germany: Springer-
Verlag, 2005, pp. 111–136.

[49] X. Liu and A. Goldsmith, ‘‘Wireless network
design for distributed control,’’ presented at
the CDC, 2004.

[50] L. Xiao, M. Johansson, H. Hindi, S. Boyd, and
A. Goldsmith, ‘‘Joint optimization of wireless
communication and networked control
systems,’’ in Switching and Learning in
Feedback Systems, R. Murray-Smith and
R. Shorten, Eds. Berlin, Germany:
Springer-Verlag, 2005, pp. 248–272.

[51] T. W. Carley, M. A. Ba, R. Barua, and
D. B. Stewart, ‘‘Contention-free periodic
message scheduler medium access control in
wireless sensor/actuator networks,’’ presented
at the RTSS, 2003.

[52] A. Saifullah, S. Sankar, J. Liu, C. Lu,
R. Chandra, and B. Priyantha, ‘‘Capnet:
A real-time wireless management network
for data center power capping,’’ presented at
the RTSS, 2014.

[53] R. Alur, A. D’Innocenzo, K. H. Johansson,
G. J. Pappas, and G. Weiss, ‘‘Modeling and
analysis of multi-hop control networks,’’
presented at the RTAS, 2009.

[54] A. Saifullah, Y. Xu, C. Lu, and Y. Chen,
‘‘Real-time scheduling for WirelessHART
networks,’’ presented at the RTSS, 2010.

[55] A. Saifullah, Y. Xu, C. Lu, and Y. Chen,
‘‘End-to-end delay analysis for fixed
priority scheduling in wirelesshart networks,’’
presented at the RTAS, 2011.

[56] A. Saifullah, Y. Xu, C. Lu, and Y. Chen,
‘‘End-to-end communication delay analysis in
industrial wireless networks,’’ IEEE Trans.
Comput., vol. 64, no. 5, pp. 1361–1374,
May 2013.

[57] A. Saifullah, Y. Xu, C. Lu, and Y. Chen,
‘‘Priority assignment for real-time flows in
WirelessHART networks,’’ presented at the
ECRTS, 2011.

[58] P. Soldati, H. Zhang, and M. Johansson,
‘‘Deadline-constrained transmission
scheduling and data evacuation in
wirelesshart networks,’’ presented at the
ECC, 2009.

[59] C. Wu, M. Sha, D. Gunatilaka, A. Saifullah,
C. Lu, and Y. Chen, ‘‘Analysis of EDF
scheduling for wireless sensor-actuator
networks,’’ presented at the IWQoS, 2014.

[60] H. Zhang, F. Osterlind, P. Soldati, T. Voigt,
and M. Johansson, ‘‘Rapid convergecast on
commodity hardware: Performance limits and
optimal policies,’’ presented at the SECON,
2010.

[61] G. C. Butazzo, Hard Real-Time Computing
Systems, 2nd ed. Berlin, Germany:
Springer-Verlag, 2005.

[62] A. Saifullah, C. Wu, P. B. Tiwari, Y. Xu,
Y. Fu, C. Lu, and Y. Chen, ‘‘Near optimal rate
selection for wireless control systems,’’
ACM TECS, vol. 13, no. 4s, 2014.

[63] P. Khosla, ‘‘Choosing sampling rates for
robot control,’’ in Proc. ICRA, 1987, vol. 4,
pp. 169–174.

[64] H. Hirata and J. Powell, ‘‘Sample rate effects
on disturbance rejection for digital control
systems,’’ in Proc. ACC, 1990, pp. 1137–1145.

[65] D. Seto, J. P. Lehoczky, L. Sha, and K. G. Shin,
‘‘Trade-off analysis of real-time control
performance and schedulability,’’ Real-Time
Syst., vol. 21, pp. 199–217, 2001.

[66] A. Horch and A. J. Isaksson, ‘‘Assessment of
the sampling rate in control systems,’’ Contr.
Eng. Practice, vol. 9, no. 5, pp. 533–544.

[67] E. Bini and A. Cervin, ‘‘Delay-aware period
assignment in control systems,’’ presented at
the RTSS, 2008.

[68] D. Henriksson and A. Cervin, ‘‘Optimal
on-line sampling period assignment for
real-time control tasks based on plant state
information,’’ in Proc. CDC-ECC, 2005,
pp. 4469–4474.

[69] J. Bai, E. P. Eyisi, F. Qiu, Y. Xue, and
X. D. Koutsoukos, ‘‘Optimal cross-layer design
of sampling rate adaptation and network
scheduling for wireless networked control
systems,’’ presented at the ICCPS, 2012.

[70] A. Saifullah, C. Wu, P. B. Tiwari, Y. Xu,
Y. Fu, C. Lu, and Y. Chen, ‘‘Near optimal
rate selection for wireless control systems,’’
presented at the RTAS, 2012.

[71] Y. Wang, A. R. Swartz, J. P. Lynch, K. H. Law,
and C.-H. Loh, ‘‘Performance evaluation of
decentralized wireless sensing and control in
civil structures,’’ in Proc. SPIE, Nondestructive
Characterization Composite Mater., Aerosp.
Eng., Civil Infrastructure, Homeland Security,
vol. 6531, 2007, pp. 653113–653113.

[72] Y. Wang and K. H. Law, ‘‘Structural control
with multi-subnet wireless sensing feedback:
Experimental validation of time-delayed
decentralized H1 control design,’’ Adv.
Structural Eng., vol. 14, no. 1, pp. 25–39, 2011.

[73] J. P. Lynch, Y. Wang, A. R. Swartz, K. Lu, and
C. Loh, ‘‘Implementation of a closed-loop
structural control system using wireless
sensor networks,’’ Structural Contr. Health
Monitoring, vol. 15, no. 4, pp. 518–539, 2008.

[74] P. Tabuada, ‘‘Event-triggered real-time
scheduling of stabilizing control tasks,’’ IEEE
Trans. Autom. Control, vol. 52, no. 9,
pp. 1680–1685, Sep. 2007.

[75] M. Pajic, S. Sundaram, J. Le Ny, G. J. Pappas,
and R. Mangharam, ‘‘Closing the loop: A
simple distributed method for control over
wireless networks,’’ presented at the IPSN,
2012.

[76] A. Bemporad, S. Di Cairano, E. Henriksson,
and K. H. Johansson, ‘‘Hybrid model
predictive control based on wireless sensor
feedback: An experimental study,’’ Int. J.
Robust Nonlinear Contr., vol. 20, no. 2,
pp. 209–225, 2010.

[77] E. Eyisi, J. Bai, D. Riley, J. Weng, W. Yan,
Y. Xue, X. Koutsoukos, and J. Sztipanovits,
‘‘Ncswt: An integrated modeling and
simulation tool for networked control
systems,’’ Simulation Modelling Practice
Theory, vol. 27, pp. 90–111, 2012.

[78] B. Aminian, J. Araújo, M. Johansson, and
K. H. Johansson, ‘‘GISOO: A virtual
testbed for wireless cyber-physical systems,’’
presented at the IECON, 2013.

[79] Wireless Cyber Physical Simulator. [Online].
Available: http://wcps.cse.wustl.edu.

[80] P. Levis, N. Lee, M. Welsh, and D. Culler,
‘‘TOSSIM: Accurate and scalable simulation
of entire TinyOS applications,’’ presented at
the Sensys, 2003.

[81] H. Lee, A. Cerpa, and P. Levis, ‘‘Improving
wireless simulation through noise modeling,’’
presented at the IPSN, 2007.

[82] B. Li, L. Nie, C. Wu, H. Gonzalez, and C. Lu,
‘‘Incorporating emergency alarms in reliable
wireless process control,’’ in ICCPS, 2015.

[83] B. Li, Z. Sun, K. Mechitov, G. Hackmann,
C. Lu, S. J. Dyke, G. Agha, and
B. F. Spencer, Jr., ‘‘Realistic case studies of
wireless structural control,’’ in Proc. ICCPS,
2013, pp. 179–188.

[84] Z. Sun, B. Li, S. Dyke, and C. Lu, ‘‘Evaluation
of performances of structural control
benchmark problem with time delays from
wireless sensor network,’’ presented at the
EMI/PMC, 2012.

[85] Z. Sun, B. Li, S. Dyke, and C. Lu, ‘‘A novel
data utilization and control strategy for
wireless structural control systems with tdma
network,’’ presented at the ASCE IWCCE,
2013.

[86] Z. Sun, B. Li, S. J. Dyke, C. Lu, and
L. Linderman, ‘‘Benchmark problem in active
structural control with wireless sensor
network,’’ Structural Contr. Health Monitoring,
2015. [Online]. Available: http://dx.doi.org/
10.1002/stc.1761.

[87] A. Saifullah, C. Lu, J. Liu, R. Chandra, and
S. Sankar, ‘‘Challenge: Wireless Sensor
Networking Over White Spaces,’’ Microsoft
Research, Tech. Rep. MSR-TR-2014-6, 2014.
[Online]. Available: http://research.
microsoft.com/pubs/208183/
whitespace_vision.pdf

[88] S. Yin, H. Luo, and S. X. Ding, ‘‘Real-time
implementation of fault-tolerant control
systems with performance optimization,’’
IEEE Trans. Ind. Inform., vol. 61, no. 5,
pp. 2402–2411, May 2014.

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

Vol. 104, No. 5, May 2016 | Proceedings of the IEEE 1023

ABOUT THE AUT HORS

Chenyang Lu (Senior Member, IEEE) received the

B.S. degree from the University of Science and

Technology of China, Hefei, China, the M.S. degree

from the Chinese Academy of Sciences, Beijing,

China, and the Ph.D. degree from University of

Virginia, Charlottesville, VA, USA, in 1995, 1997,

and 2001, respectively.

He is currently a Fullgraf Professor in the

Department of Computer Science and Engineering

at Washington University, St. Louis, MO, USA. His

research interests include real-time systems, wireless sensor networks,

cyber-physical systems, and Internet of Things. He is the author and

coauthor of over 150 research papers with over 14 000 citations and an

h-index of 52.

Dr. Lu is Editor-in-Chief of the ACM Transactions on Sensor Networks,

an Area Editor of the IEEE INTERNET OF THINGS JOURNAL, and an Associate

Editor of the new ACM Transactions on Cyber-Physical Systems and Real-

Time Systems. He also chaired premier conferences such as the IEEE

Real-Time Systems Symposium (RTSS), the ACM/IEEE International

Conference on Cyber-Physical Systems (ICCPS), and the ACM Conference

on Embedded Networked Sensor Systems (SenSys).

Abusayeed Saifullah (Member, IEEE) received

the Ph.D. degree from Washington University,

St. Louis, MO, USA, in 2014.

He is an Assistant Professor in Computer

Science at the Missouri University of Science and

Technology, Rolla, MO, USA. His research interests

include cyber-physical systems with contributions

spanning real-time systems, embedded systems,

wireless sensor networks, and parallel and

distributed computing.

Dr. Saifullah received the Turner Dissertation Award 2014 of

Washington University, the Best Paper Award at RTSS’14, and the Best

Student Paper Award at RTSS’11.

Bo Li received the bachelor’s and master’s

degrees in electrical engineering from the Harbin

Institute of Technology, Harbin, China, in 2006

and 2008, respectively. He is currently working

toward the Ph.D. degree in the Department of

Computer Science and Engineering at Washington

University, St. Louis, MO, USA.

Between 2008 and 2010, he was a Research

Assistant at the Hong Kong Polytechnic University.

His research interests include wireless system

research, e.g., wireless process control, wireless structural health

monitoring and control, and optimization.

Mo Sha (Member, IEEE) received the B.Eng.

degree from Beihang University, Beijing, China,

the M.Phil. degree from the City University of Hong

Kong, Kowloon Tong, Hong Kong, and the Ph.D.

degree in computer science from Washington

University, St. Louis, MO, USA, in 2007, 2009,

and 2014, respectively .

He is an Assistant Professor in the Department

of Computer Science at Binghamton University-

The State University of New York, Binghamton,

NY, SA. His research interests include cyber-physical systems, wireless

sensor networks, embedded systems, as well as their applications to

smart energy, smart home, wireless health, and industrial process

automation.

Humberto Gonzalez (Member, IEEE) received the

B.S. and M.S. degrees in electrical engineering

from Universidad de Chile, Santiago, Chile, in

2005, and the Ph.D. degree in electrical engineer-

ing and computer sciences from the University of

California at Berkeley, Berkeley, CA, USA, in 2012.

He is currently an Assistant Professor in the

Department of Electrical and Systems Engineering

at Washington University, St. Louis, MO, USA. His

research interests include the theory and imple-

mentation of control algorithms for cyber-physical systems, with an

emphasis on optimal control and numerical algorithms.

Dolvara Gunatilaka received the B.S. degree in

information and communication technology, and

the M.S. degree in computer science from Mahidol

University, Thailand, in 2010 and 2013, respectively.

She is currently working toward the Ph.D. degree in

the Department of Computer Science and Engineer-

ing at Washington University, St. Louis, MO, USA.

Her research interests include industrial wireless

sensor-actuator networks.

Chengjie Wu received the B.S. and M.S. degrees

from Tsinghua University, Beijing, China, in 2006

and 2008, respectively, and the Ph.D. degree in

computer science from Washington University,

St. Louis, MO, USA, in 2014.

He is a Software Engineer at Yahoo!. His

research interests include real-time systems,

cyber-physical systems, wireless sensor networks,

and Internet of Things.

Dr. Wu was a recipient of the McDonnell

International Scholarship and the Emerson Corporate Fellowship

awarded by Washington University in St. Louis. He received the Best

Paper Nomination at the RTAS’12.

Lanshun Nie received the B.S., M.S., and Ph.D.

degrees from the Harbin Institute of Technology,

Harbin, China.

He was a Visiting Scholar of the CPS Laboratory

at Washington University, St. Louis, MO, USA. He is

currently an Associate Professor of Computer

Science at the School of Computer Science and

Technology, Harbin Institute of Technology, China.

His research interests include cyber-physical

systems and wireless sensor networks.

Yixin Chen (Senior Member, IEEE) received B.S.

degree from University of Science and Technology of

China, Hefei, China, in 1999, and the M.S. and Ph.D.

degrees fromUniversity of IllinoisUrbana-Champaign,

Urbana, IL, USA, in 2001 and 2005, respectively.

He is an Associate Professor of Computer

Science at the Washington University in St. Louis.

His research interests include data mining, ma-

chine learning, artificial intelligence, and optimi-

zation. He received the Best Student Paper

Runner-up Award at the KDD’14, and Best Paper Awards at AAAI’10

and ICTAI’05. He has won First Prizes in the International Planning

Competitions (2004 & 2006). He received an Early Career Principal

Investigator Award from the Department of Energy (2006) and a

Microsoft Research New Faculty Fellowship (2007).

Dr. Chen is an Associate Editor for ACM Transactions of Intelligent

Systems and Technology and serves on the Editorial Board of Journal of

Artificial Intelligence Research.

Lu et al. : Real-Time Wireless Sensor-Actuator Networks for Industrial Cyber-Physical Systems

1024 Proceedings of the IEEE | Vol. 104, No. 5, May 2016

<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles false /AutoRotatePages /None /Binding /Left /CalGrayProfile (Gray Gamma 2.2) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.4 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails true /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams true /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments false /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues false /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo true /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 300 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >>
/ColorImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >>
/JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/AntiAliasGrayImages false
/CropGrayImages true
/GrayImageMinResolution 300
/GrayImageMinResolutionPolicy /OK
/DownsampleGrayImages true
/GrayImageDownsampleType /Bicubic
/GrayImageResolution 300
/GrayImageDepth -1
/GrayImageMinDownsampleDepth 2
/GrayImageDownsampleThreshold 1.50000
/EncodeGrayImages true
/GrayImageFilter /DCTEncode
/AutoFilterGrayImages false
/GrayImageAutoFilterStrategy /JPEG
/GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >>
/GrayImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >>
/JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >>
/AntiAliasMonoImages false
/CropMonoImages true
/MonoImageMinResolution 1200
/MonoImageMinResolutionPolicy /OK
/DownsampleMonoImages true
/MonoImageDownsampleType /Bicubic
/MonoImageResolution 600
/MonoImageDepth -1
/MonoImageDownsampleThreshold 1.50000
/EncodeMonoImages true
/MonoImageFilter /CCITTFaxEncode
/MonoImageDict << /K -1 >>
/AllowPSXObjects false
/CheckCompliance [
/None
]
/PDFX1aCheck false
/PDFX3Check false
/PDFXCompliantPDFOnly false
/PDFXNoTrimBoxError true
/PDFXTrimBoxToMediaBoxOffset [
0.00000
0.00000
0.00000
0.00000
]
/PDFXSetBleedBoxToMediaBox true
/PDFXBleedBoxToTrimBoxOffset [
0.00000
0.00000
0.00000
0.00000
]
/PDFXOutputIntentProfile (None)
/PDFXOutputConditionIdentifier ()
/PDFXOutputCondition ()
/PDFXRegistryName ()
/PDFXTrapped /False

/Description << /CHS
/CHT
/DAN
/DEU
/ESP
/FRA
/ITA
/JPN
/KOR
/NLD (Gebruik deze instellingen om Adobe PDF-documenten te maken die zijn geoptimaliseerd voor prepress-afdrukken van hoge kwaliteit. De gemaakte PDF-documenten kunnen worden geopend met Acrobat en Adobe Reader 5.0 en hoger.)
/NOR
/PTB
/SUO
/SVE
/ENU (Use these settings to create Adobe PDF documents best suited for high-quality prepress printing. Created PDF documents can be opened with Acrobat and Adobe Reader 5.0 and later.)
>>
/Namespace [
(Adobe)
(Common)
(1.0)
]
/OtherNamespaces [
<< /AsReaderSpreads false /CropImagesToFrames true /ErrorControl /WarnAndContinue /FlattenerIgnoreSpreadOverrides false /IncludeGuidesGrids false /IncludeNonPrinting false /IncludeSlug false /Namespace [ (Adobe) (InDesign) (4.0) ] /OmitPlacedBitmaps false /OmitPlacedEPS false /OmitPlacedPDF false /SimulateOverprint /Legacy >>
<< /AddBleedMarks false /AddColorBars false /AddCropMarks false /AddPageInfo false /AddRegMarks false /ConvertColors /ConvertToCMYK /DestinationProfileName () /DestinationProfileSelector /DocumentCMYK /Downsample16BitImages true /FlattenerPreset << /PresetSelector /MediumResolution >>
/FormElements false
/GenerateStructure false
/IncludeBookmarks false
/IncludeHyperlinks false
/IncludeInteractive false
/IncludeLayers false
/IncludeProfiles false
/MultimediaHandling /UseObjectSettings
/Namespace [
(Adobe)
(CreativeSuite)
(2.0)
]
/PDFXOutputIntentProfileSelector /DocumentCMYK
/PreserveEditing true
/UntaggedCMYKHandling /LeaveUntagged
/UntaggedRGBHandling /UseDocumentProfile
/UseDocumentBleed false
>>
]
>> setdistillerparams
<< /HWResolution [600 600] /PageSize [612.000 792.000] >> setpagedevice