IEEE Consumer Electronics Magazine - November/December 2022 - 18

Securing the Future of Transporation
Table 1. Summary of AI-based IDS solutions discussed in this article.
Level
Work
GIDS14
INDRA18
In-vehicle
network
LATTE16
WSG-InV15
TENET17
RACCON21
DD-IDS19
VANET
CS-IDS20
AED-ITS22
GAN-based IDS using CAN images
GRU-based recurrent autoencoder; static-threshold-based attack detection
LSTM-based encoder-decoder with self-attention; OCSVM-based attack detection
WSG-based IDS
Temporal CNN with neural attention; DT-based classifier for attack detection
Explores five different ML techniques; static threshold to detect V2V attacks
CNNs to detect attacks aimed at RSUs
DBN for data reduction; DT classifier to detect attacks in received cloud service
requests
LSTM-based autoencoder; static threshold to detect V2V and V2I attacks
Technique
the near future. Note that the order in which the
key elements in the roadmap is presented does
not represent a recommendation for solving
them sequentially in that order. These key elements
can be addressed concurrently.
Cybersecurity-Aware Design Practices
Traditional automotive systems are designed
by having safety and reliability requirements as
the main focus, and security is added as an additional
layer on the top. This severely limits the
scope of security goals that can be realized, which
exposes the automotive systems to a variety of
attacks. On the other hand, in a cybersecurityaware
design process, the security requirements
are considered from the beginning of the development
cycle. This integration of security as one of
the fundamental requirements in the design process
helps in achieving a diverse set of security
goals, making automotive systems robust and better
protected against cyber-attacks. Additionally,
advanced security strategies, such as the defensein-depth
or castle strategy,27 where a series of layered
security mechanisms are built to provide
confidentiality and integrity goals, need to be
adapted to further enhance security. These
approaches introduce redundancy in the defense
mechanisms that can be highly effective against
complex attacks. However, since the cost is a significant
concern for automakers, it is crucial to
keep the cost of security solutions low. The cybersecurity
economics and analysis models, such as
Rathod and H€am€al€ainen's work,39
18
customized for automotive needs to estimate the
cost of the security solutions.
The modern automotive system design process
also needs to employ proactive security
strategies, such as a zero-trust security model.
This model relies on the " never trust, always
verify " principle, where no participant or
Table 2. Key elements in cybersecurity roadmap for
autonomous vehicles.
Roadmap Elements
Cybersecurity-aware design
practices
Components
Security
requirements
Multilayered
security
Zero-trust security
Secure hardware and
software stack
New security and AI
standards and regulations
Advanced threat
intelligence
HSMs and TPMs
SDL for automotive
software
SOTA and FOTA
updates
ISO/SAE 21434
UNECE WP. 29
AI regulations
Vulnerability
assessment
Penetration testing
Auto ISAC
Open challenges
need to be
Data protection and
privacy
Tamper-proof AI
Securing automotive
IC supply chain
Adopting emerging
technologies
IEEE Consumer Electronics Magazine
IEEE Consumer Electronics Magazine - November/December 2022

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - November/December 2022
