Privacy-aware service provisioning in V2X networks

Intelligent Transport System (ITS) service providers typically use C-V2X data to perform analytics to achieve and maintain service innovations and standards. However, as illustrated in Figure 1, there is little to stop those with access to the data from abusing it, leading to a privacy breach.

Data privacy refers to the protection of raw data from the service provider while inference privacy refers to preventing the service provider from making certain statistical inferences it has not been authorised to perform.

In this framework, by focusing on preserving fine location, trajectory or speed information of the user while allowing certain location-specific metrics to be inferable by the service provider, the project aims to develop privacy mechanisms that restrict utility of C-V2X data to a service provider’s authorised mode of usage.

The main objectives are:

1. To develop a state space model for location and telemetric information in the context of service provisioning in a V2X network.
2. To develop a machine learning method for both inference and data privacy preservation in an ITS network, with application to the specific use-cases of protecting location, trajectory, and speed information of a vehicle.
3. To develop a privacy mechanism architecture that can be efficiently implemented in vehicle on-board units (OBU), road-side units (RSU) and 5G base stations (gNodeB) to achieve monitoring and service provisioning while ensuring inference and data privacy with low computational overhead.
4. To perform experiments on the V2X network testbed and develop a trusted embedded system for V2X privacy with industry stakeholders.

V2X networks and agents (such as vehicles and pedestrians) are highly dynamic and safety is of the highest priority. Given that most privacy methods are tailored for cloud computing and IoT networks, which are largely different, a new privacy framework needs to be developed specifically for V2X networks, to bring inference privacy mechanisms to ITS networks.

The outcomes of the project will include the following:

1. Machine learning model with both inference and data privacy preservation for V2X networks.
2. Privacy architecture and protocol, for monitoring and service provisioning in relation to location information and other telemetric data.
3. Trusted embedded system implemented in hardware, in collaboration with industry partners.

With that, the project will ultimately develop mechanisms and an architectural framework to allow industry partners to provide ITS services while minimising leakage of location and telemetric information of the user.