Signal Processing - March 2016 - 22

0.7
Five APs
Ten APs
15 APs
20 APs
One AP

0.6
0.5
0.4
0.3
0.2
0.1
0

6
8 10 12 14 16 18
Average Number of Users per AP
(a)

Normalized Achievable Data Rate

Normalized Achievable Rate

better performance than the TR
intended location and reduces
1.8
system with basic TR waveleakage to other locations,
form. Nevertheless, with optileading to a reduction in both
TR System
1.6
mal waveform, the TR system
the required transmit power
802.11 System
can still outperform LTE and
consumption and cochannel
1.4
LTE-A in most SNR regions,
interference to other locations.
which demonstrates that the TR
Defining energy efficiency
1.2
system can achieve higher ca(in bits/Joule) of a system as the
pacity than OFDM systems
spectral efficiency (sum-rate in
1
when the bandwidth is wide
bits/channel use) divided by the
enough. Note that there is a
transmit power expended (in
0.8
large amount of spectrum at
Joules/channel use), and using
1
2
3
4
5
millimeter-wave frequencies [3]
real-world channel measureNumber of Access Points
that can be utilized by TR.
ments in a typical indoor environment, we compare the figure 8. The normalized achievable rate comparison between the TR
energy efficiency of a TR sys- system and 802.11 system.
Scalability for extreme
tem with that of a direct transnetwork densification
mission system without TR. The results
trum is fully reused by different users.
With a high capacity, a single TR AP has
are shown in Figure 6. It can be seen that
Such a full spectrum reuse feature,
the potential to serve many users while
with TR, the energy efficiency can be
together with wide bandwidth, have the
creating little interference to other wireimproved by up to 7 dB. Note that a wide
potential to provide high capacity [15].
less users. However, in some scenarios,
bandwidth is generally required for a TR
This is validated in Figure 7, where we
the density of users may be so high that
system to resolve the rich multipaths and
show the performance comparison in
one single AP is insufficient to support all
fully harvest energy from the environment.
terms of achievable rate between the
of them. We will show that the TR system
As 5G technology is expected to be able to
TR system and two OFDM systems.
is highly scalable and extra APs can be
support larger bandwidth, the benefits and
It can be seen that for the one-user
added with simple reconfiguration.
unique advantages of TR due to the tempocase, even with basic TR waveform, the
In conventional wireless communicaral and spatial focusing effects in a richTR scheme can achieve much better pertion systems, a mechanism is needed to
scattering environment promise a great
formance than long-term evolution (LTE)
prevent or alleviate the interference intropotential for achieving high energy-effiin all SNR regions and better perforduced by adding more APs due to the nearciency in next-generation networks.
mance than LTE-advanced (LTE-A) in
far effect. This near-far effect is solely the
most SNR regions. With optimal waveresult of the distance between the AP and
form,
the
performance
of
the
TR
system
the users. In the TR system, however, difHigh capacity when
can
be
further
improved.
When
there
are
ferent users have different resonances,
bandwidth is available
ten
users,
due
to
the
selectivity
among
difwhich are the result of location-specific
By utilizing spatial focusing, a TR
ferent users, the achievable rate of LTE
channel impulse responses instead of the
access point (AP) can communicate
and LTE-A can be enhanced, and LTE-A
distance only. With such a strong-weak fowith multiple users simultaneously
can achieve comparable and even slightly
cusing effect, there is no clear definition of
within the same spectrum, i.e., the spec-

0.7
Five APs
Ten APs
15 APs
20 APs
One AP

0.6
0.5
0.4
0.3
0.2
0.1
0

6
8 10 12 14 16 18
Average Number of Users per AP
(b)

figure 9. A comparison between the TR system and 802.11 system: (a) graceful performance degradation of TR and (b) performance degradation of 802.11.
22

IEEE Signal Processing Magazine

March 2016

Table of Contents for the Digital Edition of Signal Processing - March 2016