IEEE Systems, Man and Cybernetics Magazine - January 2023 - 27

devices) can be fully utilized and, thus, the overall performance
of existing cognitive networks can be improved.
We explore a data-driven cognitive network architecture
to achieve the aforementioned goal. In this cognitive
network architecture, the cognitive network brokers collect
the spectrum status data of primary users and rental
intention information of idle smart devices (e.g., smartphones,
media tablets, laptops, wearable devices, embedded
in telligent devices, smart robots, au -
tonomous vehicles, and unmanned
aerial vehicles) over a payment and
then obtain the desired available
spectrum and smart devices by
analyzing and mining the collected
data. Also, the brokers collect network
performance requirements
from the demand side for network
construction and then provide
cognitive network construction
services through matching the
obtained available resources (i.e.,
radio spectrum and smart devices)
and the collected network performance
requirements. In summary,
we focus on the following two
important aspects.
1) We explore a data-driven cognitive network architecture
to deal with the construction problem of cognitive
networks, where available spectrum and idle
smart devices can form a virtual resource platform
and, thus, it is convenient for users to make full use
of various network resources. The proposed architecture
contains a spectrum and device cognitive
engine as well as a networking cognitive engine. The
spectrum and device cognitive engine achieves the
available spectrum and idle smart devices by collecting
spectrum status data from primary users and
device status data from device owners over a payment
and then mining spectrum and device using
rules from these data. The networking cognitive
engine achieves the required intelligence for building
cognitive networks by deeply analyzing secondary-user-oriented
big data and optimally allocating
the available spectrum as well as idle smart devices
for networking services.
We focus on how to
integrate spectrum
resources and smart
devices to construct
cognitive network
services through
designing a novel
data-driven cognitive
network architecture.
An Overview of Cognitive Network
Architecture Systems
To achieve low latency, high reliability, and a big data
rate, many research works have been continuously
pushed forward, and some typical studies [7]-[14] are
elaborated in Table 1. These studies did not introduce
intelligent cognition services to improve network performance.
However, other studies [15]-[20] have made useful
explorations in this regard, as
shown in Table 1. Although the
studies in [15]-[20] considered
how to build an effective cognitive
network architecture system
to achieve low-latency or highreliability
communications, they
did not jointly consider the use of
cognitive spectrum resources
and idle device resources to build
cognitive network services. Therefore,
we focus on how to integrate
spectrum resources and
smart devices to construct cognitive
network services through
designing a novel data-driven
cognitive network architecture.
Since the connected smart devices
are increasing exponentially, the data-intensive applications
in future wireless networks will need ultrahigh
throughput, ultrareliable transmission, ultralow latency,
and high energy efficiency. To this end, the new spectrum
resources (especially in the terahertz band) should be
allocated for future wireless networks. However, there is
no widely accepted propagation model that can comprehensively
depict the propagation characteristics of the
new spectrum band. Therefore, it is necessary to make
use of the existing spectrum resources to reduce the pressure
on the demand for new spectrum resources.
With the aid of cognitive radio technology, spectrum2)
On the basis of the above architecture, we discuss a
game-based incentive mechanism to deal with the incentive
problem of primary users to providing spectrum
usage information, the owners of idle smart devices
offering rental intention information, and secondary
users competing for desirable networking services.
Furthermore, we also summarize the future challenges
for the design of incentive mechanisms in cognitive
architecture systems that ensure low-latency and highreliability
requirements for data-intensive and contentrich
applications.
sharing techniques enable cognitive networks to integrate
the underutilized spectrum into existing wireless
network layouts. Currently, there is an argument that
secondary users can obtain the availability of dynamic
spectrum resources and decide how to use them without
interfering with primary users [21]. In fact, due to the
limitations of the information acquired by secondary
users and their computing resources, the actual effect is
very limited when compared with the aforementioned
desirable effect. In particular, when a primary user
decides to use its licensed spectrum again, any secondary
user that is using this spectrum should timely detect
the primary user's activity and then decide to either
select another frequency band and switch to it or adjust
its transmission parameters to avoid interfering with the
primary user, which will be a difficult task for a secondary
user. Besides the underutilized spectrum resources,
a large number of idle smart devices also need to be fully
January 2023 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE 27
IEEE Systems, Man and Cybernetics Magazine - January 2023

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