Most talks focused on V2X communications and the pros and cons of 802.11p vs 4G/5G cellular communications. Most seem to agree on that side-channel (i.e. device-to-device) communications without infrastructure support is needed if it should entirely replace 802.11p for V2X communications. In urban cities, 5G could be more suitable due to guaranteed coverage for example around corners and large buildings where IEEE 802.11p performance is shown to be poor in presence of reflectors.
Here follows the highlights of the conference.
Keynote and Panel discussions
Keynote talk “V2X Communications for Autonomous Driving – Roadmap for WiFi-V2X and Cellular V2X” by Jérôme Härri (Eurecom) and Shailesh Patil (Qualcomm). They argued that the advantages with cellular communications are enhanced range, high speed support, high density support, reuse of DSRC/C-ITS higher (stack) layers. This is also covered by the 3GPP Release-14 document. They argued that the path forward is from current LTE V2X communications to 5G but in order to do that, 5G needs to have some key features to serve autonomy such as low latency, side channel communication, better performance at high speed (500 km/h speed difference between two vehicles), high speed support using MIMO and high reliability with feedback from Phy/Mac layers.
ITS-G5 rel. 2 is currently under discussion in the C2C consortium mainly about enhanced channel usage, enhanced Phy and Mac and enhanced capacity using mm-Wave bands. They strongly argued for that 802.11p should be enhanced (802.11px) since it is based on the 802.11a technology which is 10+ years old and a new version should be based on, for example, 802.11ac.
The panel discussion later had the title “The future of V2X: Where are and where are we going?”
It was led by Jim Lansford from Qualcomm (who manufactures 802.11p hardware). There were long and at times intense discussions about why 802.11p was most suitable for autonomous vehicles. On one side, there was Qualcomm who argued that their QCA6584 chip was superb, commercially available, etc., and the other side who really wanted to see LTE-5G technologies instead. The discussion was at times a little strange since Qualcomm will also soon release the 9150 Cellular V2X chipset. There was also a comparison between LTE-V2X and ITS-G5 where we move from centralized management do decentralized management. In short, the future was believed to be 5G C-V2X + ITS-G5.
The following slides are not directly taken from the conference but they are very similar and show the subject being covered by both the keynote and the panel discussion:
https://www.qualcomm.com/invention/technologies/lte/advanced-pro/cellular-v2x
A Framework for Policy Based Secure Intra Vehicle Communication
Talk given by Technische Univerität Braunschweig: They propose a “Policy developing framework” that supports gradual introduction of a policy by introducing a trust management module in vehicles that controls all connections between components like a distributed firewall. A “policy evaluation module (PEM)” is used together with a “decision repository” and a communication module which take care of all communication and make sure it complies with the policy. All apps that want to communicate need to go through this security module before reaching the network (i.e. it acts like a proxy).
A Comparison of Cellular Vehicle-to-Everything and Dedicated Short Range Communication
Talk given by Shallesh Patil at Qualcomm: Focus was on how to overcome the challenges of V2X communications such as high relative speeds, high node densities, random resource allocation and lack of time synchronization when out of coverage. Again, the need for direct vehicle-to-vehicle communication using C-V2X was stressed. No direct communication = no adoption. Work is ongoing with reusing existing security and transport layers defined by ISO, ETSI and IEEE 1609.
The slides from the Qualcomm presentation above also cover this talk.
Privacy-Enhanced Capabilities for VANETs using Direct Anonymous Attestation
Talk by Jorden Whitefield, Univ. of Surrey, U.K. They propose a method called “Direct Anonymous Attestation”, DAA, to enhance V2X security. It is a way to shift trust from the infrastructure into the vehicles in order to protect privacy. They use certificates to guarantee user anonymity and that multiple signatures cannot be linked by using a hardware-based TPM attestation system. This simplifies the PKI infrastructure which is now mainly used to revoke misbehaving vehicles.
http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8275615
Privacy-Preserving Attribute-Based Credentials in Cooperative Intelligent Transport Systems
Paper by IBM Researchand Joint Research Centre (EU Commission): This talk discusses the problem with storing large numbers of preloaded pseudonym certificates in vehicles vs. letting them download certificates on-line. They also touch the problem with letting vehicles create pseudonym certificates themselves using group signatures, arguing that they open up for Sybil attacks (the use of multiple pseudonym identities at the same time). They present a scheme for how to generate pseudonyms locally and to guarantee that only one can be generated at any given time.
Vehicular Blocktrees
Joshua Joy, UC Los Angeles, USA. This talk presented the idea of using blocktrees to broadcast and verify vehicle communication signatures. The local view of one vehicle should always match the view in its neighbors, if not, a malicious case is detected. Blocktrees can be used for example to verify the authenticity of logs from V2X communications for example after a crash. Using their method, other vehicles can verify the correctness and detect if they logs have been manipulated.
For questions or access to papers, please contact tomas.olovsson@chalmer.se
