In-vehicle data recording, storage and access management in autonomous vehicles

Viktoras Kabir Veitas,Simon Delaere
DOI: https://doi.org/10.48550/arXiv.1806.03243
2018-05-28
Computers and Society
Abstract:Transport sector is in the process of being rapidly and fundamentally reshaped by autonomous and collaborative driving technologies. This reshaping promises huge economic and social benefits as well as challenges in terms of developing and deploying secure and safe transportations systems, their smooth integration to social fabric. We have employed Policy Scan and Technology Strategy Design methodology in order to identify concrete societal expectations and problems and map them with mitigating technological availabilities in the domain of autonomous driving and smart mobility. Event Data Recorder for Autonomous Driving (EDR/AD) is an envisioned subsystem of a vehicular Controller Area Network which ensures the confidentiality, integrity and availability of data related to operation of a vehicle in order to permit recovery of exact situation following the occurrence of an event or on demand. The exact technical and regulatory requirements for the device are still in the development internationally, but it is clear that it will be included into vehicle type-approval requirements at UNECE level. We present an analysis of the context of the usage of the EDR/AD in collaborative intelligent transport systems, related security, data provenance and privacy, other regulatory and technical issues considering many interest groups and stakeholders involved. We present a concrete proposal for developing a EDR/AD proof of the concept prototype with clear market deployment potential and urge security researchers, vehicle manufacturers, and component suppliers to form a collaboration towards implementing important technology for making future autonomous vehicles more socially acceptable and legally compliant. Furthermore, EDR/AD technology, apart from its immediate use in autonomous driving and smart mobility domain has a potential to be extended to general autonomous robot and AI applications.
What problem does this paper attempt to address?