Bionyms: Driver−centric Message Authentication using Biometric Measurements

Abstract

The technology of self-driving cars and driver-assistance systems has reached a point where vehicles start to make decisions on behalf of drivers and operate autonomously. The introduction of Vehicular Ad-hoc Networks (VANETs) will increase this autonomy to greatly improve efficiency and safety on the road. However, when relying on vehicle-to-vehicle communication to make life-critical decisions, such as emergency braking, information authenticity and integrity is of paramount importance. Current schemes that satisfy these properties tie the identity of a vehicle's owner to its messages and discourage malicious behavior under the penalty of prosecution. But if driver and owner are not the same, it is difficult to identify the person causing an accident or committing a traffic offence. This is particularly relevant for increasingly popular car sharing schemes and transportation as a service where vehicles are owned by mobility providers. In this paper, we propose a novel message authentication scheme based on biometric information that provides traceability of each message to the driver. This enables accountability and exclusion from the network on a per-individual basis, while at the same time preserving drivers' privacy. To evaluate functional protocol properties, such as computation overhead, we simulate traffic in a realistic road network. We implement our scheme and demonstrate its feasibility using a driver's body impedance, an unobtrusive biometric modality that can be acquired while holding the steering wheel. Our evaluation is supported by data gathered in a user study with 33 subjects conducted under simulated driving conditions.