IEEE Consumer Electronics Magazine - January/February 2024 - 81

the critical roles of SAGINs in providing varieties
of vertical applications by connecting
enormous heterogeneous devices, machines,
and industrial processes. Through MFA by artificial
intelligence (AI) techniques, including
supervised learning, unsupervised learning,
and reinforcement learning, trusted communications
and services can be accomplished to
adapt to dynamic environments.
Driven by the limitations of conventional
authentication mechanisms and the induced
demands in an integrated network, this article
presents a novel authentication framework for
SAGINs, where an AI-oriented two-phase multifactor
authentication scheme (ATMAS) is proposed.
In ATMAS, a machine-learning-based continuous
authentication, which refers to " Phase II " is performed
to intelligently grant or deny access reliably
by capturing user-profiles and traits,
following a conventional cryptographic authentication
named " Phase I. " The contributions of this
article are summarized as follows:
1) We propose the ATMAS based on the characteristics
of SAGINs. For the ATMAS, four communication
entities, i.e., mobile user (MU),
base station (BS), satellite, and network control
center (NCC) are considered, in which
satellite and NCC cooperate with each other
to implement authentication.
2) Through the design of ATMAS, we leverage
the unique spatial-temporal features of
SAGIN in Phase II, including service and geographic
features. This allows us to ensure
the security of the SAGIN from login to
logout.
3) To evaluate the robustness and validate the
security of the proposed scheme, we conduct
security analysis by analyzing the security
characteristics. Moreover, from the perspective
of performance evaluation, we shed light
on lightweight and efficient authentication
mechanism design by a proper combination
of spatial-temporal factors.
The remainder of this article is organized as
follows. Challenges of authentication for the
SAGINs are first introduced. Then, we present
the advantages of AI-enhanced multifactor continuous
authentication. The design of the
January/February 2024
proposed ATMAS and its security analysis is
described in the following sections. In addition,
simulation results are presented to evaluate the
performance of authentication accuracy. Finally,
concluding remarks and future research directions
are provided.
CHALLENGES OF AUTHENTICATION
FOR THE SAGIN
Comprised of intersatellite and intersatellite
links, the SAGIN is a highly open wireless system.
As satellites are exposed to the open environment,
it is easy to cause attacks such as link
hijacking and entity counterfeiting, which seriously
affects the secure communication of the
SAGIN. Reliable access control plays a vital role
in distinguishing the source nodes and addressing
identity-based attacks such as spoofing and
Sybil attacks. In general, access control can be
decomposed into user authentication and authorization.
Authentication is a process that verifies
that someone or something is who they say they
are and authorization is the security process
that determines a user or service's level of
access. The latter can be accomplished by methods
such as enforcing allow or deny rules based
on the user's authorization level, e.g., general
user, super user, and administrator, which is relatively
less complicated in the implementation
stage. However, authentication brings about a
number of SAGIN-specific research questions
that should be carefully considered in this
article.
Long Propagation and Processing Latencies.
In SAGIN, the introduction of high-altitude satellites
leads to unidirectional propagation delays
of around 15 ms, even for the links between LEO
satellites and IoT devices. This results in roundtrip
delays of approximately 50 ms for IoT devices,
which significantly reduces quality of service
(QoS) in latency-sensitive scenarios.
Additionally, due to high mobility, the propagation
delay is not permanent and varies with the
satellite's position. Apart from propagation
delay, conventional cryptographic authentication
requires increased communication and
computation overhead to cope with the increasing
requirements of security, leading to long
processing latencies.7 Such prolonged latencies
81
IEEE Consumer Electronics Magazine - January/February 2024

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - January/February 2024
