390
W. Chang et al.
Fig. 7.15 Security-aware system design by the PBD paradigm [36]
abstract attacks to outsider attacks and insider attacks. To protect against them, we
introduced the key management system and the intrusion detection systems and
pointed out the need of system integration due to the limited resource and strict
system constraints. We admit that they are just part of the big picture of automotive
security, but we believe that they can provide important insights to secure connected
autonomous vehicles and even other IoT systems.
7.4 Concluding Remarks
CPS are fast developing, demanding a new design methodology that unifies all the
layers, aiming for safety, security, robustness, and resource efficiency. This chapter
gives the technical background on the cyber components, physical components,
and their interactions. A perspective from CAVs is taken with two case studies on
assuring the safety of machine-learning-based perception and assuring the security
and robustness, respectively.
Research and development along the direction of CPS require multidisciplinary
expertise. There are still many challenges that need to be addressed. For instance,
in the perception function of autonomous systems, machine learning algorithms are
dominating in the performance. However, they have obvious drawbacks. First, they
are vulnerable to adversarial attacks. Second, they are hard to analyze and provide
guarantee.
W. Chang et al.
Fig. 7.15 Security-aware system design by the PBD paradigm [36]
abstract attacks to outsider attacks and insider attacks. To protect against them, we
introduced the key management system and the intrusion detection systems and
pointed out the need of system integration due to the limited resource and strict
system constraints. We admit that they are just part of the big picture of automotive
security, but we believe that they can provide important insights to secure connected
autonomous vehicles and even other IoT systems.
7.4 Concluding Remarks
CPS are fast developing, demanding a new design methodology that unifies all the
layers, aiming for safety, security, robustness, and resource efficiency. This chapter
gives the technical background on the cyber components, physical components,
and their interactions. A perspective from CAVs is taken with two case studies on
assuring the safety of machine-learning-based perception and assuring the security
and robustness, respectively.
Research and development along the direction of CPS require multidisciplinary
expertise. There are still many challenges that need to be addressed. For instance,
in the perception function of autonomous systems, machine learning algorithms are
dominating in the performance. However, they have obvious drawbacks. First, they
are vulnerable to adversarial attacks. Second, they are hard to analyze and provide
guarantee.