Search

2e Rencontres Francophones Transport Mobilité, Montréal, CANADA, 11-/06/2019 - 13/06/2019; Le projet IVA (Information Voyageurs Augmentée) s'intéresse notamment à l'impact de l'information voyageurs sur le comportement des usagers réguliers des transports en commun franciliens en situation de perturbation inopinée. Dans ce cadre, nous avons réalisé une étude pilote visant à tester la faisabilité (acquisition de données en situation réelle, confort des participants, gestion des contraintes de terrain...) et la pertinence (nature des résultats obtenus) d'une méthodologie innovante couplant l'oculométrie en situation réelle à l'entretien et au questionnaire. L'objectif est d'étudier l'effet de l'information voyageur sur le comportement des usagers et leur prise de décision en situation perturbée. Après un rapide état de l'art, nous présentons la méthodologie, les résultats et les principales conclusions de cette étude pilote.

Planning routes using transportation network maps is a common task that has received little attention in the literature. Here, we present a novel eye-tracking paradigm to investigate psychological processes and mechanisms involved in such a route planning. In the experiment, participants were first presented with an origin and destination pair before we presented them with fictitious public transportation maps. Their task was to find the connecting route that required the minimum number of transfers. Based on participants' gaze behaviour, each trial was split into two phases: (1) the search for origin and destination phase, i.e., the initial phase of the trial until participants gazed at both origin and destination at least once and (2) the route planning and selection phase. Comparisons of other eye-tracking measures between these phases and the time to complete them, which depended on the complexity of the planning task, suggest that these two phases are indeed distinct and supported by different cognitive processes. For example, participants spent more time attending the centre of the map during the initial search phase, before directing their attention to connecting stations, where transitions between lines were possible. Our results provide novel insights into the psychological processes involved in route planning from maps. The findings are discussed in relation to the current theories of route planning.

La nature du lien entre le mouvement des yeux et la tâche est un problème ouvert (Yarbus, 1967). Les méthodes statistiques de classification sont utilisées depuis plus de vingt ans dans de nombreux domaines, mais elles n’ont été introduites pour l’analyse des mouvements oculaires par Greene et al. qu’en 2012. Lors d’une expérience, ils proposaient différentes tâches pour un même stimulus. Ils ont obtenu des scanpaths, puis ils ont utilisé un classifieur pour prédire la tâche à partir des mouvements oculaires (avec un résultat négatif). L’an dernier, Henderson et al. (2013) ont repris la même question avec un protocole et des tâches différentes. Leur classifieur a réussi à distinguer des tâches de recherche visuelle et de mémorisation. Suite à ces travaux, nous nous sommes demandé s’il était possible d’utiliser un classifieur pour prédire une tâche à partir du mouvement des yeux, mais dans une situation dynamique.

An experiment was conducted in a driving simulator to test how eye-movements patterns evolve over time according to the decision-making processes involved in a driving task. Participants had to drive up to crossroads and decide to stop or not. The decision-making task was considered as the succession of two phases associated with cognitive processes: Differentiation (leading to a prior decision) and Consolidation (leading to a final decision). Road signs (Stop, Priority and GiveWay) varied across situations, and the stopping behavior (Go and NoGo) was recorded. Saccade amplitudes and fixation durations were analyzed. Specific patterns were found for each condition in accordance with the associated processes: high visual exploration (larger saccade amplitudes and shorter fixation durations) for the Differentiation phase, and lower visual exploration (smaller saccades and longer fixations) for the Consolidation phase. These results support that eye-movements can provide good indexes of underlying processes occurring during a decision-making task in an everyday context.

Two main factors seem to contribute to the development of a riding configuration, and consequently of a motorcycle simulator: the trajectory control modality and the leaning rendering. The goal of this study was to compare two riding simulator configurations through the assessment of the underlying mental workload adopting a multidimensional approach based on psychophysiological, performance, and subjective measures. In the first configuration (reduced motion), the trajectory control is obtained by means of positive steering, while the leaning is produced just by tilting the visual scene. Like a real motorbike, the second configuration (dynamic) allows a progressive transition between positive and counter steering as the speed increases, whereas the leaning is rendered by splitting the rolling angle between the tilting of the visual scene and the rolling of the platform. Each participant completed six lane-change tasks per configuration, of which the first three and the last one were single tasks, and the remaining two were dual tasks. The occurrence of three single-task runs at the beginning of the experiment allowed us to examine the process of adaptation to each configuration, which is a critical precondition for simulator validity. The dynamic configuration proved to have higher validity, as confirmed by psychophysiological and subjective measures. Findings might have implications for the development of future riding simulators.

This paper reviews a cluster of recent and current researches in the field of visual attention conducted at the LEPSIS Lab (IFSTTAR). Taking advantage of the multidisciplinary structure of the research team, contributions have been proposed in the fields of image processing, artificial intelligence, virtual reality, cognitive psychology and ethology. In image processing, we have contributed to predictive models of the gaze fixations in a driving task, namely, looking for road signs, proposing a computational model of top-down visual attention. In artificial intelligence, a simple computational model of visual attention was proposed and implemented in an agent-based microscopic traffic simulation model. In virtual reality, the blink frequency was proposed in order to assess the relevance of driving simulators in terms of the driver's mental workload. In cognitive psychology, the visual conspicuity of powered two wheels was assessed on a driving simulator, leading to recommendations about their lighting system design; another contribution gave evidence that oculomotor patterns can reveal crossing decisions at crossroads. In cognitive ethology, a joint work is in progress with the CEREMA in order to study oculomotor patterns at crossroads in a field test, thanks to a dedicated vehicle., Cet article présente un ensemble de recherches récentes et en cours dans le domaine de l’attention visuelle, menées au laboratoire LEPSIS de l’IFSTTAR. Grâce à la structure pluridisciplinaire du laboratoire, des contributions ont été proposées dans les domaines du traitement d’images, de l’intelligence artificielle, de la réalité virtuelle, de la psychologie cognitive et de l’éthologie. En traitement d’images, nous avons proposé un modèle de calcul de la saillance visuelle des panneaux routiers, en référence à une tâche de conduite. En intelligence artificielle, nous avons implémenté un modèle d’attention visuelle pour améliorer la simulation de trafic microscopique. En réalité virtuelle, nous avons proposé d’utiliser les clignements pour mesurer la charge mentale sur simulateur de conduite. En psychologie cognitive, la saillance visuelle des deux roues motorisés a été étudiée, ce qui a permis d’aboutir à des recommandations sur le design des phares; nous avons également montré que les patterns oculomoteurs peuvent nous informer sur la décision de traverser en carrefour. En ethologie, des travaux conjoints avec le CEREMA sont en cours pour étudier le comportement oculomoteur en carrefour, grâce à un véhicule instrumenté.

Holmqvist, K., Mulvey, F., & Johannson, R. (2013). Abstracts of the 17th European Conference on Eye Movements 2013. Journal of Eye Movement Research, 6(3). https://doi.org/10.16910/jemr.6.3.1; Eye measures have been shown to be relevant indicators of mental workload. Empirical results revealed a relationship between mental workload and eye blinks, depending on the task constraints. In a single-task, the blink rate seems to drop as a function of greater task complexity, while the opposite result has been found in a dual-task. The goal of this study was to examine the effect of driving-task complexity on blink rate, using a single versus a dual task. For the experiment, 20 participants drove in 3 driving environment complexities: low (dual-carriageway), medium (rural-road) and high (urban-road). Each condition was presented alone, or associated with a vocal reaction-time-task to an auditory stimulus. Results showed a strong effect of driving-task complexity, as blink rate decreased from low to high environment complexity. By contrast and differing from previous research, the dual-task condition didn’t affect blink rate. This suggests that the reaction time and the driving task were performed simultaneously, and that the reaction -time -task didn’t affect the driving -task. The effects of a dual task on blink rate might probably occur in case of attentional switching. It could be valuable to examine in a future experiment the effect of environment complexities on blink rate in a new dual-task paradigm: driving and mental-arithmetic.

With the aim of studying attentional leftward bias in a driving context, we combined recording of gaze behaviour with a simulated driving task (Lane Change Test – LCT) in spatial symmetry conditions. The LCT requires driving along a straight traffic-free three-lane road, changing lanes according to the information provided by two identical road signs displayed concurrently on both left and right sides of the road. Participants directed most of their attention to the left-hand signs. The prevalent visuospatial nature of the driving task – which is supposed to mostly activate the right cerebral hemisphere – could have caused a contralateral attention shift. The introduction of a secondary task – performed with the right hand, concurrently with the LCT – attenuated the leftward bias.

Visual attention is interrelated with the research of visuals information, especially in complex tasks such as driving. In this task, especially when approaching a crossroads, top-down processes (goal-directed) are dominant. To reach his goal, the driver needs to take decisions, such as Go/NoGo at the crossroads. Therefore, driving is a relevant task to study decisional processes using eyes movement. In our study, we describe the decision-making with three cognitive processes, and postulate three associated oculomotor patterns. These processes are: information-taking (collect information to end in a decision); control (soundness of the decision); anticipation (information-taking attached to the next decision and control, depending on the risk level and the associated attentional resource). Road signs and traffic densities were manipulated in order to modulate the decision. We tested two hypotheses, one before and one after the Go/NoGo decision in the give-way condition: 1/ before the decision, oculomotor patterns are distinct between give-way (information-taking), stop (anticipation) and priority (control); 2/ after the decision, they are similar between priority and give-way with Go decision (control), and between stop and give-way with NoGo decision (anticipation). Thirty-four participants, equipped with an eye-tracker, drove in a simulator. Saccade amplitudes and fixation durations were in good agreement with our hypotheses. Finally, we focus on the dynamics of the processes.

A number of computational models of visual attention have been proposed based on the concept of saliency map. Some of them have been validated as predictors of the visual scan-path of observers looking at images and videos, using oculometric data. They are widely used for Computer Graphics applications, mainly for image rendering, in order to avoid spending too much computing time on non salient areas, and in video coding, in order to keep a better image quality in salient areas. However, these algorithms were not used so far with High Dynamic Range (HDR) inputs. In this paper, we show that in the case of HDR images, the predictions using algorithms based on Itti, Koch and Niebur [1] are less accurate than with 8-bit images. To improve the saliency computation for HDR inputs, we propose a new algorithm derived from Itti and Koch [3]. From an eye tracking experiment with a HDR scene, we show that this algorithm leads to good results for the saliency map computation, with a better fit between the saliency map and the ocular fixation map than Itti, Koch and Niebur's algorithm. These results may impact image retargeting issues, for the display of HDR images on both LDR and HDR display devices.

Traffic signs are designed to be clearly seen by drivers. However a little is known about the visual influence of the traffic sign environment on how it will be perceived. Computer estimation of the conspicuity from images using a camera mounted on a vehicle is thus of importance in order to be able to quickly make a diagnosis regarding conspicuity of traffic signs. Unfortunately, our knowledge about the human visual processing system is rather incomplete and thus conspicuity visual mechanisms remain poorly understood. A complete model for conspicuity is not known, only specific features are known to be of importance. It makes sense to assume that an important task for drivers is to search for traffic signs. We therefore propose a new paradigm for conspicuity estimation in search tasks based on statistical learning of the visual features of the object of interest.

De nombreuses recherches menées antérieurement dans le milieu professionnel et en psychologie expérimentale, ont montré que le vieillissement, de part son action sur les processus perceptifs (tels que la vision) et cognitifs (comme la mémoire et l'attention), pouvait interférer sur la prise et le traitement de l'information. L'introduction d'écrans embarqués à bord des automobiles, nécessitant de la part du conducteur la gestion d'informations nouvelles en plus de la tâche de conduite, amène à s'interroger sur les modalités de présentation de cette information (taille, couleur, contraste,localisation dans le champ visuel), compte tenu notamment des capacités fonctionnelles spécifiques aux personnes âgées. Une étude en laboratoire de paramètres physiques de présentation de l'information en situation d'attention partagée sur des sujets 'jeunes' et 'âgés' a été faite. Les sujets ont effectué une tâche de poursuite d'une cible se déplacant sur un écran d'ordinateur en utilisant une souris. La seconde tâche consistait à répondre à des signaux directionnels (flèches) en appuyant sur les boutons de la souris. Chaque tâche est effectuée seule puis en compétition avec l'autre. Les performances sont appréhendées en terme de temps de réponse, de pourcentage de bonnes réponses, de pourcentage de non-réponses pour la tâche de réponse aux flèches. Dans le cas de la tâche de poursuite, la distance entre la cible et le curseur souris est retenue comme critère de performance. Thèse de doctorat : Spécialité Neurosciences.

Negotiating curves in driving situations creates considerable difficulties. We do not fully understand the way that drivers perceive curve characteristics while approaching curves, and how they adjust their speed. Various studies have shown that the radius and angle of curves are two important parameters. Unfortunately we do not know all the cues, their use, their weight in the driving scene and their connections. Existing analyses were obtained on the road are in the laboratory; our studies used simulated driving situations (moving base simulator) for a new approach. Using comparisons of the behaviors observed in actual and simulated driving situations, we studied the anticipation in perception of four curves, by analyzing the distribution of visual fixation and scan path in connection with characteristic conditions of curves (real and simulated situations) and driving experience. Overall, there was an increase in the function of some zones: some indices were missing in the environment, road lines and roadsides were used more for controlling trajectory. This was not totally sufficient, however, and caused a reduction in anticipated driving. In simulated situations, the speedometer was used more than in real situations. Mean fixation time was shortest in real situations then in simulated situations. This can serve to demonstrate that indices are more difficult to extract from synthesized images. Experienced drivers had shorter mean fixation times than beginners and the location of fixations seems to be more adapted to curve characteristics. Finally, in methodological terms, we investigated the use of simulators and driver behavior and driving task analysis.

Graphic information displays have the potential to communicate public information in situations where normal announcement types are ineffective. This study used eye tracking techniques to analyze comprehension mechanism of event-related information on railway traffic disruptions presented via different graphic formats. One hundred thirteen participants were asked to understand and compare series of traffic disruption messages delivered via three purely visual formats. Animated displays were the most effective presentation type. Eye tracking data showed why an animation facilitates comprehension: it enhances processing strategies which provide the best condition for segmenting the causal chain of the events provided in the message.