An Affordance-Based Approach to Visually Guided Overtaking

Abstract

International audienceWhen an automobile driver overtakes a lead vehicle while avoiding oncoming traffic, does he or she do so with reference to the limits of his or her car? We investigated overtaking from the perspective of the theory of affordances. We define the overtake-ability affordance as a ratio of the minimum satisfying velocity required for safe overtaking (MSV) to the maximum velocity of the driver's car (V-max). Two groups of participants performed overtaking maneuvers, if deemed possible, by driving either a slow (V-max=25m/s) or a fast (V-max=32.5m/s) virtual car in overtaking situations constrained by 14 values of MSV. For any given MSV condition, participants in the fast car group were more likely to attempt an overtaking maneuver. However, when MSV was expressed in intrinsic units as a ratio of V-max for both groups, the frequency of overtaking was not significantly different across groups. Furthermore, overtaking frequency dropped to near 0% for both groups when MSV exceeded V-max. In accordance with the affordance-based framework (Fajen, 2007), our results suggest that participants select their overtaking maneuvers by perceiving an overtake-ability affordance

Information–Movement Coupling in The Control of Driver Approach to an Intersection

“…Not only is this mechanism based on the consideration of an informational support that allows the agent to access, at any time, the state of their relationship with the environment, but task-related variables such as TTA and AC also modulate the effect of the coupling and consequently the adjustments that are produced. We believe our results also provide a useful complement to recent studies that put action capabilities at the heart of the control process in goal-directed action (e.g., Fajen, 2005Fajen, , 2007Morice, Diaz, Fajen, Basilio & Montagne, 2015;Marti, Morice & Montagne, 2015). A goal of studies that aim to formalize control laws could therefore be to consider the capacity for action in the operationalization of the information-movement relation.…”

Hopf Bifurcations in Complex Multiagent Activity: The Signature of Discrete to Rhythmic Behavioral Transitions

“…As referenced throughout this paper so far, this is perhaps best detailed in Saltzman and Kelso’s [ 9 ] Task Dynamics approach, and more recently in Warren’s [ 8 ] Behavioral Dynamics approach, with the latter approach also emphasizing how a modelling task directed perceptual-motor behavior involves identifying control laws that minimize some task quantity of the agent–environment system. For example, reaching for an object involves the minimization of one’s hand to the target location [ 9 ], or the deceleration of a vehicle to avoid collision involves maintaining the control variable tau under a specific value [ 44 , 45 ]. Accordingly, modelling the behavioral dynamics of individuals has been particularly successful with regard to easily quantifiable task objectives, such as reaching and catching [ 46 ], walking through a cluttered environment [ 47 ], and juggling [ 48 ].…”

Modelling vehicles acceleration during overtaking manoeuvres

“…However, the estimation of the speed of opposing vehicles is extremely difficult, because of the very low rate of expansion of objects located so far from the observer. In addition, Basilio et al [3] and Morice et al [4] proposed an overtaking decision model based on the overtaking ability affordance, defined as the quotient between the minimum speed required to overtake and the maximum speed of the vehicle at that time, depending on the vehicle performance. After a driving simulator experiment with only 16 drivers, they evidenced that drivers accurately perceived whether a lead vehicle can be safely overtaken, since overtaking attempt decreased with the real possibility to overtake.…”

“…The use of driving simulator might limit the validity of findings, as risk taken by drivers depends on their immersion in the virtual world during the experiment, and the detection of opposing vehicles which at long distances is complicated, due to the limited resolution of screens. Besides, driving simulator usually only accounts for a very limited (or null) variability of acceleration capabilities of vehicles, because only one type [3] or two types [4] of vehicle are implemented. Consequently, a field study is solely able to study drivers’ behaviour across a wide range of vehicles in real conditions.…”

“…As expressed by Morice et al [4], ADAS for overtaking should be calibrated to be effective. It means that they should be coherent with drivers’ behaviour.…”