IEEE Systems, Man and Cybernetics Magazine - January 2023 - 26

high-frequency band. Therefore, this article introduces
cognitive network brokers based on a data-driven
cognitive network architecture to integrate and make
full use of various resources to provide good network
services for users, including an engine for spectrum
and device cognition and an engine for cognitive network
service construction.
By learning from spectrum and device usage contexts,
the spectrum and device cognitive engine aims at
mining primary users' spectrum
usage rules in primary networks
and device owners' device-using
rules. On the basis of secondary
users' data analytics as requirements
for providing high-quality
network services, it is vital for
the networking cognitive engine
to offer customized network services
toward specific domains.
Moreover, we discuss a gamebased
incentive mechanism for
mot ivat ing pr imary users to
offer spectrum usage information
and device owners to provide
resources (e.g., communication devices, computing
facility, and storage space) for building cognitive network
services and encouraging secondary users to compete
for desirable networking services. Experimental
results show the effectiveness of the presented incentive
mechanism.
The proposed
architecture contains
a spectrum and
device cognitive
engine as well as a
networking cognitive
engine.
due to the immaturity of both propagation models and
radio interface technologies, the expectation for the
terahertz frequency band to accommodate the rapid
growth of mobile data traffic is still questionable. At
present, most wireless network systems still use the traditional
sub-6-GHz frequency band ranging from
300 MHz to 3 GHz, which can reap the benefits from its
reliable propagation characteristics over a long transmission
distance in a variety of radio environments [5]. Future
wireless networks are still expected
to densely deploy diverse small
cells (e.g., micro, pico, and femto),
which can efficiently reuse the
sub-6-GHz frequency band. However,
dense networks will lead to
large cochannel interference, which
will, in turn, reduce spectral efficiency.
Therefore, in an ultradensification
area formed by the overlap
of diverse small cells, it is very
difficult to intelligently manage
the shared spectrum resources to
increase network capacity in future
wireless networks.
Cognitive radio technology is capable of mitigating
Supporting Data-Intensive Applications
The global Internet of Things (IoT) medical devices [1]
market is projected to have a compound annual growth
rate of 25.2% from 2018 to 2023, which is predicted to
reach US$63.43 billion by 2023 [25]. However, rich, realtime,
reliable, cost-effective, and affordable multimedia
services that can support customized IoT-based applications
[2], [3], are scarce and, thus, hinder the further
growth of this type of market. Clearly, network capacity is
a major bottleneck. Therefore, it is very urgent for future
data-intensive applications to adopt next-generation
mobile communication technologies, which are expected
to provide seamless wide-area coverage, ultracapacity
access service, low latency, and high-reliability connections
in future wireless networks.
Based on the aforementioned requirements, both radio
industries and research organizations have been collaborating
on future wireless networks, where spectrum supply
and spectral efficiency are important for capacity
improvement and low delay support. To achieve this goal,
it is necessary to efficiently explore the higher-frequency
band (e.g., the terahertz frequency band) or further optimize
the usage of the sub-6-GHz frequency band [4].
The terahertz frequency band offers a new source for
spectrum supply for future wireless networks. However,
26 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE January 2023
interference and enhancing energy efficiency. Also, it can
ease spectrum scarcity by exploiting the underused
licensed spectrum resources over different spaces and
times or reusing the occupied spectrum resources in an
opportunistic manner [6]. The concept of primary and secondary
users is usually mentioned in cognitive radio systems,
where primary users are the holders of licensed
spectrum resources, while secondary users try to use
them opportunistically. However, the premise of such a
capability is that the nodes with cognitive radio interfaces
can automatically sense and learn from radio environments,
timely adapt the transmission parameters, and fully
share spectrum resources in space, time, frequency, and
modulation mode [6].
In the increasingly congested sub-6-GHz frequency
band, due to the limited spectrum-sensing information
and insufficient computing resources, it is very difficult
for individual nodes serving as secondary users (i.e., cognitive
users or unlicensed users) to obtain the desired
available spectrum resources by actively sensing the
spectrum usage status, gathering information (e.g., radio
frequency type, transmission power, modulation type)
from its surroundings, and learning from its context-aware
information. Furthermore, due to the dynamic nature with
respect to available spectrum resources, the topology stability
of cognitive networks using these spectrum resources
also fluctuates, which increases the maintenance
overhead for network owners or users. Therefore, it is
urgent to make full use of the existing abundant network
resources to build a cognitive network infrastructure,
where network resources (e.g., radio spectrum and smart
IEEE Systems, Man and Cybernetics Magazine - January 2023

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - January 2023
