Working Title of Thesis:
Risk Assessment and Data-driven Insurance Pricing Models for Automated and Connected Vehicles (Project acronym: RADICAL)
The proliferation of connected and automated vehicles introduces several challenges in the vehicular ecosystem. One of its main impacts is on the automobile insurance industry since the risk of being involved in an accident is expected to change drastically. On the one hand, the number of accidents is expected to shrink with the removal of the human factor. On the other hand, given the likelihood of more expensive vehicles and potential reputational damage, the severity incurred in each accident is expected to increase.
To cover traditional driving risks, vehicle insurers have been proposing statistical models for premium schemes based on the vehicle model and demographic variables (e.g., place of residence, age of the driver). There has also been a plurality of data-driven risk assessment studies on conventional vehicles able to segment drivers according to their risk. Nevertheless, the predictive power of current models is put into question because of the emerging risks variables caused by a highly connected and automated driving scenario. These risk variables can be divided into three groups: (i) those related to vehicle dynamics and usage of the automated features, (ii) those related to potential failures of the automation components (e.g., sensors, radars, algorithms, connectivity), (iii) those related to cybersecurity.
The proposed industrial fellowship between the University of Limerick and Motion-S S.A., funded by the Luxembourg National Research Fund, aims to fill the gap in the transition in the automobile insurance market by providing a car risk assessment model for a fully digitalized ecosystem along with an interface to a data-driven pricing model for car insurance premiums.