3792 IEEE INTERNET OF THINGS JOURNAL, VOL. 5, NO. 5, OCTOBER 2018
(a)
(b)
Fig. 3. OQAM-OFDM versus OFDM for discontinuous spectrum.
C. Superiority of OQAM-OFDM-Based Wireless
Communications for IoT
The major spectral drawback of OFDM originates from the
use of the time domain rectangular pulse shaping. Therefore,
as the evolution of OFDM, OQAM-OFDM uses a well
designed prototype filter on each subcarrier to obtain ultralow
frequency domain sidelobes. With the improved spectral char-
acteristics, OQAM-OFDM has much better compatibility with
the following IoT scenarios than OFDM.
1) Discontinuous Spectrum: Some applications in IoT
transmit small data packets such as sensors and data acquisi-
tion, or transmit even several bits of control information such
as the smart home. For the above applications, allocating fixed
spectrum resources to IoT terminals is not the best way since
the resources are idle in most time. Therefore, flexibly utiliz-
ing the spectrum holes and the unlicensed spectrum is a good
choice. Note that, the spectrum bands in the spectrum holes
and unlicensed spectrum are typically discontinuous and the
leaked interferences should be strictly controlled in order not
to interfere other users.
In Fig. 3, OQAM-OFDM is compared with OFDM for
the discontinuous spectrum. We can see that there are max-
imum three continuous sub-bands available for the current
user. The neighboring sub-bands are unavailable due to the
usage by other users. Since OFDM has very high frequency
domain sidelobes, large guard bands are required at both the
two sides of the available spectrum bands to avoid the leaked
out-of-band interferences to other users. As a result, two sub-
bands are wasted and only one sub-band can be utilized in
the OFDM-based wireless communication system. On the con-
trary, OQAM-OFDM has ultralow frequency domain sidelobes
and is able to further control the out-of-band interferences
through the prototype filter design. Therefore, almost no guard
band is needed in the OQAM-OFDM-based wireless com-
munication system, which greatly improves the utilization
efficiency of spectrum resources.
2) Asynchronous Transmission: Since the objective of IoT
is to connect everything in the word, it is expected that the
number of IoT terminals will be dramatically increasing year
by year. Different from traditional cellular users, it is diffi-
cult to manage the IoT terminals in a synchronous way due
to the large number of terminals and their uncertain activities.
(a)
(b)
Fig. 4. OQAM-OFDM versus OFDM for asynchronous transmission.
Moreover, many IoT terminals only transmit a few of bits as
the control information. The signaling used in the synchronous
transmission is a significant overhead for those terminals with
small data packets. Therefore, the asynchronous transmission
can significantly improve the system capacities for IoT termi-
nals. In addition, some applications in IoT require real-time
and accurate information, and their tolerances on the latency
are very low. Note that, the asynchronous transmission dis-
cards the synchronization signaling and waiting, which greatly
reduces the signal overhead and decreases the transmission
latency.
In Fig. 4, OQAM-OFDM is compared with OFDM for
the asynchronous transmission. As an orthogonal technique,
OFDM strongly relies on the strict time synchronization
between all users to ensure their signals arrive at the BS simul-
taneously. Otherwise, the orthogonality among the received
signals from different users will be damaged by the high
frequency domain sidelobes, resulting in strong interferences
among asynchronous users. Even with long CP inserted in
OFDM signals, the time offsets originated from large num-
ber of IoT terminals cannot be fully absorbed. Moreover, the
CP insertion will decrease the effective data rate. Therefore,
OFDM cannot support the asynchronous transmission, and
the overhead of signaling and waiting greatly increases
the transmission latency. On the contrary, OQAM-OFDM
is a nonorthogonal technique and depends on the ultralow
frequency domain sidelobes to reconstruct signals. Then, a
certain range of time offsets among received signals from
different users at the BS is affordable for the OQAM-
OFDM-based wireless communication system, and the strict
time synchronization among users is unnecessary. Therefore,
OQAM-OFDM can well support the asynchronous transmis-
sion and greatly reduce the transmission latency caused by the
signaling and waiting.
3) Various Types of Services: In the future, more and more
IoT terminals will be connected to the IoT. Various types of
services have different requirements on the communication
performances such as the data rate and transmission latency.
For example, the rate requirement of wearable devices is on the
level of 10 Mb/s, while the rate requirement of the automation
and monitoring is about 50–500 kb/s. Regarding the tolerance
of the communication latency, most IoT applications require