Description

Vehicular communications is a key function to enable future autonomous vehicles. Vehicular communication networks are being developed to allow vehicles, e.g. cars, vans, trucks, and other traffic participants, to exchange sensory, traffic, and other information to enhance autonomy and safety in autonomous driving scenarios. While communication functions and protocols are being standardised by standards bodies such as ETSI, the question whether and how a message received from another traffic participant, e.g. a vehicle, can be trusted remains unanswered. The current standard specifies a public-private encryption approach to authenticate a message received from another vehicle. But it does not deal with the situation whether the vehicle can be trusted. Trust is key here, however, as vehicles advise each other on other objects on the road, their direction of travel, on traffic accidents, and other road situations. Transmission of wrong information can lead to accidents.

This project aims at studying trustworthiness in vehicular communications and devising a trust management approach to allow a receiving vehicle to judge whether a received message can be trusted or not. A range of trust estimation approaches exists, such as Dempster-Shafer theory, machine learning models, etc., leveraging, for example, social network-based recommendations. Evaluation and analysis of devised approaches will be carried out using computer simulation models.

This project suits a student who is interested in computer simulation, computer networks, and aspects of cyber security and trust.

Objectives

• review the current trust management approaches in the literature
• devise a simulation environment for vehicular communications using simulation frameworks such as SimPy or others
• implement trust management approaches within the simulation environment
• analyse and evaluate trust management approaches