Associate Professor of Computer and Information Science, College of Engineering and Computer Science, The University of Michigan-Dearborn
The objective of the research was to gain a deep understanding of the security risks/needs of automated-vehicle (AV) platooning, to understand the ability
of sensor fusion to counter cyberattacks, and to develop platooning-control algorithms that take into account both safety and cybersecurity. The proposed work was visualized, demonstrated and evaluated through large-scale vehicle network simulator. It is widely believed that platooning requires dedicated short range communications (DSRC) to work properly. We believed that the DSRC channel is most vulnerable to forging attacks, and hence cannot be leveraged for any safety-critical operations by itself. It is also possible to spoof radar and Light Detection and Ranging (LiDAR) sensors and blind cameras. We analyzed security risks of automated platooning and evaluated the potential of sensor fusion in the context of cybersecurity by a structured risk assessment. We then analyzed the potential of sensor fusion by assigning input sensors confidence levels (e.g., camera is harder to spoof than DSRC), and establishing a model to utilize such confidence levels for a platooning-control algorithm to account for safety and cybersecurity. We tested the effectiveness of our algorithms in a simulation.
The main project outcomes are a platooning-risk analysis, an evaluation of sensor fusion in the context of cybersecurity, control algorithms to incorporate safety and cybersecurity aspects, and a simulation environment for further development and evaluation.