Scope and Motivation:
The Cyber-Physical Systems (CPSs) have become very complex, more sophisticated, intelligent and autonomous. We cite as example of CPS smart grid in energy sector, smart factory and industry 4.0, intelligent transportation systems, healthcare and medical systems, and robotic systems. The CPSs offer very complex interaction between heterogeneous cyber and physical components; additionally to this complexity they are exposed to important disturbances due to unintentional and intentional events which lead the prediction of their behaviors (categorized as “Normal” or “Faulty”) a very difficult task. Meanwhile, cyber security for CPS is attracting the attention of research scientists in both industry and academia since the number of cyber-attacks have increased and their behaviors have become more sophisticated commonly known as zero-day threats.
Conventional cyber security mechanisms, such Intrusion Detection and Prevention Systems (IDS/IPS), and access control have not the capability to detect, prevent and block this category of cyber-attacks since the zero-day threats exhibit an unknown misbehavior that are not defined in signatures’ database of the security systems. Recently, a new era of cyber security mechanisms based on Artificial Intelligent (AI) are under development to protect the CPSs from these zero-day attacks. In the context of cyber security, the machine learning technologies are used to manage a huge amount of heterogeneous data that come from different sources of information with a goal of generating automatically different attacks patents and hence predict accurately the future attackers’ misbehavior. Meanwhile, game-theoretic approaches have been used in the context of cyber defense to solve the decision-making issues (i.e., the suspect device is an attacker or not) and attacks prediction. In decision-making issue, the cyber security game is used to study the interaction between the security agents (e.g., IDS and IPS) and their opponents (e.g. attackers) with a goal to determine the optimal decision making of security agent to classify the suspected opponent as attacker or not.
Preventing the occurrence of zero-day attacks requires the collaboration between different AI systems including machine learning and game theory, as well as security expert intervention. In fact, the involving of human intervention in the decision-making leads an improvement of attacks detection since the purpose of human-machine interaction is to reduce the number of false positives.
Another example to illustrate the migration of security solutions to use more intelligent principles and technologies, the Identity Management & Access control (IAM) which switch from a simple login/password checking to voice and facial recognition.
This Special Issue (SI) aims to bring together researchers from academic and industrial to share their visions of the AI application in cyber security context, present challenges and recent works and advances related to AI-based cyber security applied to CPSs. Potential topics include, but not limited to the following:
- Design and verification of AI-based security solutions,
- Impact of AI-based security solutions on CPS performances,
- Safety of AI-based security solutions.
- IDS/IPS based on machine learning,
- IDS/IPS based on deep and reinforcement learning,
- Cyber security game to protect the CPS,
- Authentication and Access Control,
- AI modeling for attack behavior,
- Attacks prediction based on machine learning and game theory,
- Human-machine interaction in the context of cyber security,
- Application of AI-based security in internet of things and transportation segments
- AI-based Solution for Physical layer security
- Manuscript Submission Deadline: May 1st 2019
- Initial Decision: August 1st, 2019
- Revised Manuscript Due: September 1st, 2019
- Decision Notification: November 1st, 2019
- Final Manuscript Due: December 1st, 2019
- Publication Date: May 2020
Manuscripts should conform to the standard format as indicated in the Information for Authors section of the Paper Submission Guidelines.
All manuscripts to be considered for publication must be submitted by the deadline through Manuscript Central. Select the “May 2020: Cyber Security based on Artificial Intelligence for Cyber-Physical Systems” topic from the drop-down menu of Topic/Series titles.
Guest Editors of the Special Issue: