Experimental evidence supporting simpler Action Point paradigms for car-following

•Action Point (AP) models are systematically reviewed.•Actual APs, based on pedals observations, are identified.•Results support the adoption of simple AP paradigms.•Qualitatively similar phenomena are observed in different experimental contexts. The Action Point theory is one of the paradigms that...

Full description

Saved in:
Bibliographic Details
Published in:Transportation research. Part F, Traffic psychology and behaviour Traffic psychology and behaviour, 2015-11, Vol.35, p.1-15
Main Authors: Pariota, Luigi, Bifulco, Gennaro Nicola
Format: Article
Language:English
Subjects:
Action Point Model
Analogies
Car-following behaviour
Driving behaviour
Driving simulator
Instrumented vehicle
Psychology
Simulators
Thresholds
Traffic engineering
Traffic flow
Transportation
Virtual environments
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Action Point (AP) models are systematically reviewed.•Actual APs, based on pedals observations, are identified.•Results support the adoption of simple AP paradigms.•Qualitatively similar phenomena are observed in different experimental contexts. The Action Point theory is one of the paradigms that can be applied to understand and reproduce car-following behaviour. Several different approaches to this theory have been proposed, some more simple and others more complex. In particular, the reference point in this field is still the paradigm from Wiedemann, which requires the identification of four action-point thresholds. In this paper we review Action Point theories in order to highlight similarities and differences and to ascertain whether all the thresholds proposed by Wiedemann actually bind the driving behaviour. Based on a large-scale experiment in which car-following data were collected, we identified all candidate action points assuming that the more complex (four-threshold) theory holds. Then we tested these points with respect to the large data set of available observations, in order to check whether actual actions are performed at the points. The results show that very often simpler approaches better match the observed data and that in order to explain car-following behaviour it is sufficient in most cases to refer to two thresholds. The results obtained by real-world observation were also tested in virtual environments (two different kinds of driving simulators) and were confirmed.
ISSN:1369-8478
1873-5517
DOI:10.1016/j.trf.2015.08.002