A Radar-Based Real-Time Cyberattack Detection, Classification, And Notification System Based on Learning Driving-Simulated Vehicle Trajectory Data Under

Investigator(s):

• Zhixia (Richard) Li, University of Cincinnati, ORCID # 0000-0002-7942-4660 (PI)
  
  

Project Description:

Figure 1 Cyberattack for red-light countdown application enabled by connected vehicles

Connected vehicles communicate wirelessly with other vehicles and transportation infrastructure. Attacking wireless communication is the most likely occurrence during an cyberattack. While previous researchers have focused on delay and congestion attacks, from a hacker's perspective, red light running behavior is the most dangerous and could lead to more serious consequences. Therefore, a wireless attack on the red-light countdown application is the most likely scenario that hackers would consider. If a cyberattack occurs, the driver will get a falsified message to let them know there is a short red light. The driver may still maintain the speed when the vehicle enters the intersection and results into a severe accident. To detect such safety issue. The objectives of the research are: (1) create a radar-based real-time cyberattack detection, classification, and notification algorithm that can detect and classify vehicle trajectory data under cyberattacks; (2) use the driving-simulated cyberattack scenario experiment data including the collected vehicle trajectories and driver behavior data to train and test the cyberattack detection model.

Figure 2 Hidden Markov Models used in detection of cyberattacks from the vehicle trajectory data