DICE - Driver interaction with cyclists and e-scooterists at intersections

Research Project, 2021 – 2022

From 2023, Euro NCAP will include new scenarios to test the capability of advanced driving assistance systems (ADAS) to improve cyclist safety (CBTA, CBDA) at intersections. In both these scenarios, new knowledge about how cyclists interact with cars and the environment will lead to design more realistic tests and improve ADAS activation (both in terms of acceptability and safety benefits).

Quantitative models of road-user behaviour can provide this knowledge by predicting the future intentions and actions of surrounding road-users, so that their behaviour can be mathematically described and then included in ADAS and virtual testing. It is worth noticing that understanding cyclist intention is a critical aspect for CBTA because, if the cyclist intends to stop at the intersection, then ADAS are not needed, and once if the cyclist does not intend to (or cannot) stop, ADAS may immediately start acting (providing information, warning, or intervention, depending on the kinematic threat).

Recently, e-scooters became popular in several cities worldwide and Euro NCAP may include scenarios with e-scooters in the future. However, the extent to which ADAS for cyclists will also be effective for e-scooterists is today unknown because we do not know how e-scooterist’s behaviour compares to the one of cyclists. Quantitative models of e-scooters behaviour may guide engineers tuning up ADAS that, today, avoid collisions with cyclists so that these ADAS may also prevent crashes with e-scooterists.

DICE will model cyclist and e-scooterist behaviour to 1) inform CBTA and CBDA test scenarios in Euro NCAP and 2) support the development of ADAS that may prevent cars colliding with cyclists or e-scooterists. While our previous project DIV focused on modelling a driver going straight at an intersection, this project will model the cyclist as well, to explain the interplay between these road users. The project will also investigate the extent to which cyclist behaviour differs from e-scooterist behaviour and the possible implications for ADAS design and testing.