I. F. Akyildiz et al.: 6G and Beyond: The Future of Wireless Communications Systems
Besides these classical approaches for THz band device
development, new materials, including graphene, carbon nan-
otubes, and graphene nanoribbons, are gaining more attention
due to their extremely high electron mobility in the order of
8, 000 − 10, 000 cm
2
/(V · s) at room temperature, as com-
pared to 1, 400 cm
2
/(V · s) of silicon, and 8, 500 cm
2
/(V · s)
of GaAs, which means that the link throughput can be poten-
tially up to ten times higher than is currently achievable
with most semiconductors [28]. The graphene-based devices,
offering outstanding mechanical, electrical, and optical prop-
erties, have been utilized in the development of power detec-
tors at 600 GHz [29] and 200 GHz [30], as well as plasmonic
antenna arrays and transceivers [31], [32]. Graphene-based
devices have the potential to break new ground in reaching
the desired level of performance at much higher frequencies
above 1 THz.
C. PHYSICAL LAYER MODELING AT TH E THz BAND
The realization of wireless communications at THz frequen-
cies requires the development of accurate channel models to
capture the impact of both channel peculiarities including the
high atmospheric attenuation and molecular absorption rates
at various transmission windows, as well as the propagation
effects including reflection, scattering, and diffraction, with
respect to different materials. Current research has reported
several efforts to provide fundamental understanding of such
channels. For example, an early work in [33] demonstrates
the remarkable capacity the THz band channel can support
for short transmission distances. The model provides a detail
analysis on the effect of attenuation caused by the molecular
absorption and spreading loss, on which the performance
of channel throughput is heavily dependent. One step fur-
ther, in order to extend the transmission distance at THz
bands, an idea dubbed as the ultra-massive multiple-input
multiple-output (UM MIMO) communications, enabled by
an element array size of 1024 × 1024 with plasmonic nano-
antennas, can drastically boost the signal strength by steering
and focusing the transmitted beams in both space and fre-
quency [34]. Correspondingly, a UM MIMO channel model
has been developed in [35] which takes into account the role
of such arrays.
Additionally, a stochastic channel model for indoor THz
band communications at 300 GHz has been reported in [36]
which characterizes both spatial and temporal domain chan-
nel information. More recently, based on the aforemen-
tioned applicable scenarios in the THz band, a stochastic
channel model for kiosk applications has been reported
in [37] which ranges from 200 − 340 GHz. The main take-
away from current models validated using either mea-
surements or the ray-tracing technique is that the direct
path between the transmitter and the receiver and the
single-bounce reflected paths dominate the received power,
while other channel effects, including diffraction and scatter-
ing, attenuate power significantly along propagation.
On the basis of the ultra-wideband channel character-
ization, THz band communications faces a critical and
challenging task of synchronization in the receiver design.
Existing pulse-based modulation schemes permit the use
of low-complexity non-coherent analogue detectors, e.g.,
energy detector and auto-correlation receiver, which involve
the multiplication of the received signal with itself, followed
by an integrator. However, a more advanced non-coherent
receiver architecture in [38] based on a continuous time
moving average symbol detection scheme demonstrates bet-
ter performance compared to previous detection schemes
for pulse-based modulations in terms of the symbol error
rate. In addition, robust frequency and timing synchroniza-
tion for multi-carrier communication in the THz band is
desirable to decode multiple incoming signal streams. The
work presented in [39] realizes both a low-rate sampling
scheme for channels with high SNR values and a maximum
likelihood-based algorithm for low SNR channels with sat-
isfying bit-error-rate performance. More recently, a synchro-
nization scheme based on medium access control protocols
is reported in [40], which shows good performance in both a
macro-scale scenario to overcome the distance limitation at
the THz band and a nano-scale scenario for nano-devices for
energy harvesting.
Similarly, given the peculiar ultra-wideband nature and
frequency selectivity of THz band communications, equal-
ization solutions pose another relevant challenge [15], [41].
In [42], three equalization solutions, including the Tomlinson-
Harashima precoding, a waveform with interference manage-
ment for time-reversal systems, and an iterative algorithm
with adaptive soft feedback, are reviewed and compared for
an indoor channel. Results show that the iterative algorithm
with adaptive soft feedback yields the most promising BER
performance [42].
D. MEDIUM ACCESS CONTROL I N THz BAND
COMMUNICATIONS
On top of physical layer channel models, the medium access
control (MAC) schemes in THz band communications should
adopt certain spatial and spectral features in order to pro-
vide solutions in resolving issues such as the deafness prob-
lem and LoS blockage, among others [43], [44]. Different
from commonly used MAC solutions in RF systems that
utilize omnidirectional antennas, such as carrier sense mul-
tiple access with collision avoidance (CSMA/CA), the MAC
protocols designed for THz band rely on handshakes between
transceivers with highly directional beams. These razor-sharp
beams can provide higher power radiation gain and prolong
the transmission distance, but when misalignment happens,
the deafness problem arises. As such, the deafness avoidance
approach is required in MAC scheme design. Existing solu-
tions utilized in IEEE 802.15.3c [45] and others employ a
beam-training phase to estimate and steer beams towards des-
tined devices. Recent works also propose methods based on
angular division multiplexing [46] and a priori aided channel
tracking schemes [47]. The results in such proposed solutions
suggest that with good beam alignment strategies the channel
throughput can be improved significantly.
VOLUME 8, 2020 134001
剩余35页未读，继续阅读
