Imaging vehicle-to-vehicle communication using visible light

Abstract

With advances in automated and connected driving, secure communication is increasingly becoming a safety-critical function. Injection of manipulated radio messages into traffic can cause severe accidents in the foreseeable future, and can currently be achieved without having to manipulate on-board vehicle systems directly, for example by hijacking cellphones instead and using these as senders. Thereby, large-scale attacks on vehicles can be executed remotely, and target relatively vulnerable devices. To mitigate remaining vulnerabilities in current automotive security architectures, this paper proposes a secondary communication channel using vehicle head and taillights. In contrast to existing approaches, this method allows both to achieve a sufficient data rate and to extract the angular position of the sender, by means of an imaging process which only requires close-to-market, cost-efficient technology. Through this, injecting false messages by masquerading as a different sender is considerably more challenging: The receiver can verify a message’s source position with the supposed position of the sender, e.g. by using onboard sensors or communicated information. Thereby, reliably faking both the communicated messages and the position of the sender will require direct manipulation of on-board vehicle systems, raising the security level of the function accordingly, and precluding low-threshold, wide-range attacks.