As Internet of Things systems become widely adopted, cyber-physical systems offer smarter environments and services than earlier networked systems by leveraging and integrating sensed data of various modalities. Despite adopting the benefits of the IoT technologies, critical infrastructures and cyber-physical systems have various weaknesses due to existing vulnerabilities in wireless networks, particularly in the 5G Era. With massively connected nodes that push sensed data, the cyber threat surface leaves more nodes as target points to enter the monitoring network.
By considering connected and autonomous vehicles as a special use case of critical cyber-physical infrastructures, this talk will consist of the following parts:
1) A brief overview of the vulnerabilities and countermeasures in IoT and IIoT cybersecurity with a focus on threat models and attacks surfaces in cyber-physical critical infrastructures,
2)Threat vectors in 5G and their interplay with cyber-physical critical infrastructures
3) State of the art in machine/deep learning (ML/DL)-based intrusion detection in IoT and cyber-physical system networks,
4) The interplay between artificial intelligence and system/application-level security issues in cyber-physical settings,
5) Security-by design solutions for critical infrastructures by bridging ML/DL and Radio Frequency Domain.
The talk will wrap up by presenting open issues, challenges and opportunities in this field are going to be presented for professionals, researchers and developers who are interested in this field. Furthermore, the talk will present remedies for performance degradation of ML/DL models in the case of dynamically changing physical/environmental conditions such as changing channel conditions since the training of an ML model to recognize the signatures of radio signals. At the end of the talk, the audience will have information on how to tackle the following problems from various angles: How can AI be leveraged to meet the requirements of holistic security / safety of cyber-physical critical infrastructures? What are the challenges and opportunities 5G introduces in empowering trained AI models for dependable decisions in a cyber-physical critical infrastructure?