Investigating and modelling the factors affecting cooperative driving behaviours at narrow passage road interactions

Abstract

Cooperative road interactions, road interactions in which drivers work together to complete manoeuvres (Kraft et al., 2019), have received increased attention from researchers looking to model the behaviours exhibited by drivers during the interactions, and those investigating how autonomous vehicles should interact with human-driven vehicles in a composite road transport system. This attention is due to the relative lack of traffic regulations that govern these interactions, with drivers instead relying on informal traffic rules and communications to safely navigate the situations. Narrow passage interactions - a “deadlock” cooperative road interaction in which two (or more) drivers driving in opposite directions encounter one another at a road narrowing – have, however, remained under-researched and thus inadequately represented in microscopic traffic models. For example, it is widely accepted that the vehicle type being interacted with alters driving behaviours, and yet there is limited investigation of this factor on the behaviours exhibited by drivers during narrow passage interactions, with little subsequent incorporation of the factor in driver behaviour models.
This research, therefore, contributes to knowledge by improving the understanding of the factors that affect a car driver’s decision-making during narrow passage interactions and the modelling of this understanding in mathematical driver behaviour models. A questionnaire study was initially carried out to widen the range of factors found to influence narrow passage decision-making. Notably, the study found that the vehicle type and the number of vehicles being interacted with influenced the decision of drivers to give way at a road narrowing. The factors were then further explored in the qualitative analysis of the concurrent and retrospective verbalisations produced in an on-road study, which sought to gain an insight into the chronological relationships between what drivers perceive, how they interpret this information and the actions they subsequently undertake. The findings from the questionnaire study were validated by the on-road study and a theoretical framework of the decision-making processes of drivers during narrow passage interactions was also proposed. Using this theoretical framework, a rule-based driver behaviour model for narrow passage interactions was developed as part of a co-simulator driving simulation package. The bespoke narrow passage driving simulator, which sought to remedy limitations of previous simulator studies, was used in a simulator study consisting of both open-world and controlled drives. The results of the study further validated the findings of the previous studies and led to the development of a novel binary mixed logit model to represent narrow passage decision-making. This model was found to outperform model structures representative of the previous state-of-the-art in the narrow passage human factors and modelling literature, that were trained and evaluated using the same dataset.

More information

Identifiers

ORCID for Peter Refaat Salib Youssef: orcid.org/0000-0001-6755-329X

ORCID for Ben Waterson: orcid.org/0000-0001-9817-7119

ORCID for Katie Plant: orcid.org/0000-0002-4532-2818

Catalogue record

Export record