IEEE Wireless Communications - April 2017 - page 21

IEEE Wireless Communications • April 2017
19
this technique is preferred by spectrum regula-
tors to provide non-interfering spectrum access to
competing CR operators in unlicensed TVWS
bands [8, 10].
Many novel smart grid applications, including
demand response management, energy monitor-
ing for shared facilities (e.g. lift, corridor light),
environmental sensing, and so on, would require
higher data rates with more stringent quality of
service (QoS) requirements. To cater for these
situations, already there are proposals to assign
certain channels in the TVWS spectrum as high
priority channels (HPCs) [8]. The HPCs could be
leased temporarily by interested CR operators
for exclusive short-term usage by paying a small
license fee to the geo-location database provider.
Maintaining economic viability would then require
appropriate HPC purchase decisions by a CR
operator, and in a highly dynamic environment,
with varying spectrum availability, channel condi-
tions, and incoming data traffic, balancing HPC
leasing cost and QoS guarantees could become
quite challenging.
In this article, we study the performance of a
communication network architecture for M2M
applications in smart grids using TVWS. Following
the approach in [11], a multi-tier communication
network is established. In the proposed architec-
ture, a building area network (BAN) is based on
power line communications (PLC), a neighbor-
hood area network (NAN) is based on TVWS,
and a wide area network (WAN) relies on cable
or fiber optics. We implement the communication
network in large residential complexes in Singa-
pore that could be used to transmit the aggregat-
ed data of several hundred smart meters using
TVWS spectrum to the remote servers of the
network operator. We launched a measurement
campaign in the dense urban environment of Sin-
gapore to determine the achieved data rates. The
lowest data rates that we obtained using TVWS
spectrum exceeded 1 Mb/s, which are sufficient
for several smart grid applications. It is important
to note that currently in Singapore, TVWS is in
the trial phase, and we did not face any spectrum
competition from other CR operators.
However, considering future growth and inter-
est in smart grid applications, we anticipate over-
crowding of TVWS amid stiff competition among
several CR operators. CR operators would be
required to lease HPCs to guarantee the QoS
demands of numerous smart grid applications.
1
In this context, we develop a simple thresh-
old-based real-time decision support mechanism
based on the Lyapunov optimization framework
to help the CR operator in making the HPC pur-
chase decisions. The proposed mechanism cap-
tures the trade-off between the HPC leasing cost
and transmission delay and is called the Dynam-
ic-CD-LYAPUNOV scheme. We then consider
data quality as another QoS dimension. This con-
sideration is based on the fact that in some smart
grid applications, data quality, measured in terms
of the size of data unit (packets, frames), could be
reduced without any significant impact on overall
performance. For example, depending on spec-
trum availability, the least significant bits of smart
meter readings or environmental sensor readings
could be dropped (exploiting redundancy in the
readings). Considering two QoS dimensions, we
study the trade-off between spectrum cost, trans-
mission delay, and data quality, and develop a
Dynamic-CD-QUALITY scheme. Simulation results
indicate that the schemes developed in this work
provide significant opportunities for CR operators
to maintain a balance between spectrum cost and
QoS for several emerging smart grid applications.
To summarize, there are three main contribu-
tions of this work. First, we propose a communi-
cation network architecture for CR operators that
exploit TVWS spectrum and present the results of
a measurement campaign in a dense urban envi-
ronment to determine the achievable data rates
for smart grid applications. Second, we develop
a threshold-based real-time decision rule for the
CR operator to make HPC leasing decisions and
adjustable parameters to trade off spectrum leas-
ing cost and queuing delay. Finally, we propose
an algorithm that exploits data quality along with
the transmission delay to reduce the HPC leasing
cost. These novel ideas and the resulting optimi-
zation problems are not considered in any prior
work on smart grids.
The rest of the article is organized as fol-
lows. In the following section, we describe the
proposed communication network architecture
and present the achievable data rates in a dense
urban environment using TVWS spectrum. The
Dynamic-CD-LYAPUNOV scheme is then devel-
oped, and the Dynamic-CD-QUALITY scheme is
presented following that. Next, we present the
simulation results and conclude the article in the
final section.
P
roposed
S
mart
G
rid
C
ommunication
N
etwork
A
rchitecture
and
the
R
esults
of
a
M
easurement
C
ampaign
In this section, we discuss a generic communica-
tion network architecture, exploiting TVWS, for
dense M2M communication networks supporting
smart grid applications, such as, demand response
management, home automation, and data anal-
ysis. Based on the proposed architecture, we
also present the achievable data rates in a dense
urban environment using TVWS spectrum.
P
roposed
S
mart
G
rid
C
ommunication
N
etwork
A
rchitecture
The proposed architecture can readily be
deployed in any residential, commercial, or indus-
trial complex, having numerous buildings and
a large number of machine type devices. The
choice of an appropriate communication tech-
nology at the building, neighborhood, and wide
area network levels that could leverage existing
infrastructure or freely available resources, while
satisfying the data rate requirements of the tar-
geted application, is essential for a cost-effective
solution.
Usually, each machine (sensor, smart meter,
etc.) inside a building generates or collects some
data, which is transmitted by establishing a BAN.
To create a BAN, we propose wired access (i.e.,
PLC) which reuses existing power lines for com-
munication. Unlike WiFi and other wireless per-
sonal area technologies (e.g., Zigbee, Bluetooth),
PLC is more secure, does not create interference,
The choice of an appro-
priate communication
technology at the build-
ing, neighborhood and
wide area network lev-
els that could leverage
existing infrastructure or
freely available resourc-
es, while satisfying the
data rate requirements
of the targeted applica-
tion, is essential for a
cost effective solution.
1
It is important to note
that in this scenario our
work focuses on the TVWS
spectrum purchase and QoS
issues faced by CR operators,
regardless of the nature of
the underlying protocol (IEEE
802.11af, etc.).
1...,11,12,13,14,15,16,17,18,19,20 22,23,24,25,26,27,28,29,30,31,...132
Powered by FlippingBook