Self-driving vehicles are the future of transportation, offering convenience, safety, and efficiency. However, as these vehicles become more connected and collaborative, they also become vulnerable to cyber attacks. A recent study led by the University of Michigan highlights the risks associated with data fabrication attacks on self-driving vehicle networks and proposes preventive measures to protect fleet operators and passengers.
The study focuses on collaborative perception, a network of connected and autonomous vehicles that share information to enhance their ability to „see“ the environment. This collective sensing power allows vehicles to detect objects and obstacles beyond their individual sensors‘ range. However, this collaboration opens up opportunities for hackers to introduce fake objects or manipulate perception data, potentially causing accidents or disruptions in traffic flow.
According to Z. Morley Mao, a professor of computer science and engineering at U-M, understanding and countering these attacks is crucial for advancing the security of connected and autonomous vehicles. The study, led by doctoral student Qingzhao Zhang, introduces sophisticated, real-time attacks that were tested in both virtual simulations and real-world scenarios at U-M’s Mcity Test Facility.
The researchers administered falsified LiDAR-based 3D sensor data with malicious modifications to the system, simulating cyber attacks that could be carried out by hackers. These attacks, known as zero-delay attack scheduling, were designed to introduce malicious data without any lag or delay, making them difficult to detect.
In virtual simulated scenarios, the attacks were highly effective, with success rates reaching 86%. On-road attacks conducted at the Mcity Test Facility triggered collisions and hard brakes, highlighting the potential dangers of data fabrication attacks on self-driving vehicle networks.
To counter these threats, the researchers developed a system called Collaborative Anomaly Detection, which leverages shared occupancy maps to cross-check data and detect abnormal or inconsistent information. This system achieved a detection rate of 91.5% with a false positive rate of 3% in virtual simulated environments, effectively reducing safety hazards in real-world scenarios.
The findings of the study provide a robust framework for improving the safety and security of connected and autonomous vehicles. By open-sourcing their methodology and providing comprehensive benchmark datasets, the researchers aim to set a new standard for research in this domain, fostering further development and innovation in autonomous vehicle safety and security.
As self-driving vehicles continue to evolve and become more prevalent on our roads, it is essential to address the cybersecurity challenges associated with these advanced technologies. By implementing preventive measures and developing sophisticated detection systems, we can ensure that connected and autonomous vehicles remain safe, reliable, and resilient in the face of potential cyber threats.