No description available.

No description available.

Alcohol intoxication represents one situation an individual might increase their amount of risk taking when driving. This dissertation is comprised of three studies that investigate the mechanisms by which alcohol increases driver risk-taking. Study 1 examined the effect of alcohol on driver risk-taking using a proxemics approach. The study also tested whether alcohol-induced increases in risky driving co-occurred with pronounced impairment in the driver’s skill. The study also examined whether the most disinhibited drivers were also the riskiest. Indeed, alcohol increased driver risk-taking and impaired driving skill. The study also revealed risky driving can be dissociable from impairing effects on driver skill and that poor inhibitory control is selectively related to elevated risky driving. Studies 2 and 3 built on this work by addressing whether the apparent dissociation between behavioral measures of driver risk and skill was mediated by perceptions the drivers held. While maintaining the distinction between driver risk and skill, Study 2 tested the relationship between drivers’ BAC estimations and their tendency to take risks on the roadway. Drivers who estimated their BAC to be lower were the riskiest drivers following both alcohol and placebo. Study 3 addressed whether risky driving could be increased by environmental factors that shape perceptions the driver holds. There is evidence post-licensure training programs might inadvertently generate overconfidence in drivers’ perceived ability to operate a motor vehicle and thus fail to perceive dangers normally associated with risky driving behavior. To test this hypothesis, twenty-four drivers received either advanced skill training or no training in a driving simulator. Drivers who received skill training showed increased risky driving under alcohol whereas those who received no training tended to decrease their risk taking. Trained drivers also self-reported more confidence in their driving ability. Taken together, these studies represent a large step towards the betterment of laboratory-based models of driving behavior. The work highlights the importance of distinguishing between driver risk-taking and driving skill. The studies also identified that drivers’ personal beliefs influence alcohol-induced risky driving; this suggests training programs focused on correcting drivers’ misconceptions might be most efficacious in reducing their risk taking on the roadway.

The utilisation of navigation systems in cars has given rise to road safety concerns, and the design and functionality of such systems must therefore be adjusted to the users’ needs, since they have to divide their attention between driving and the operation of the navigation system. The study was aimed at finding the optimum system response time (SRT) which would enable a driver to focus as much as possible on the road while attaining an efficient task completion time using an electronic navigational system. The research project consists of two separate experiments and was completed by 10 subjects. Experiment 1 included a temporal reproduction task and a secondary memory task. The subjects had to memorise two symbols and then reproduce six time spans ranging from 1 to 30 s to provide a baseline measurement of their time estimation abilities. Experiment 2 consisted of a simulated automobile driving task. While driving in the simulator the subjects completed a memorising task displayed on a touch screen. The task was presented with seven different system response times (SRTs) ranging from 0 to 30 s. The effects of different SRTs on the eye movement from road to monitor, regarding the duration of fixation and the frequency of change were evaluated. The distribution of gazes to the secondary task was analysed to provide information about the time estimation performance in the driving simulator. Other dependent variables tested were the accuracy of selected items, memory game performance, drive performance and the subjective state of the test person. The results of this study can be employed to find the optimum duration of inter-task delays for in-vehicle technical devices. / Psychology / M.A. (Psychology)

Le Vagabondage de la Pensée (VP) est un état d’inattention particulier que chacun expérience au quotidien et qui consiste,pour un individu en un changement de l’orientation de son attention vers des pensées personnelles. Dans le cas de la conduiteautomobile, le VP a de nombreux effets négatifs générant un risque accru d’être impliqué et responsable d’un accident.L’objectif de cette thèse était d’abord de mieux décrire l’impact du VP sur les variables physiologiques et comportementales duconducteur afin de dresser une liste des indicateurs sensibles à cet état. Le VP en agissant comme une tâche additionnelle à laconduite, consommerait une partie des ressources cognitives du conducteur. Il convenait donc de rechercher, par une approcheexpérimentale, un indicateur de l’effort cognitif en conduite automobile. Comme la multiplicité des pensées et de leurscaractéristiques pourraient créer différents niveaux de risque pour le conducteur, il apparaissait pertinent d’étudier l’impact de ces différentes formes d’inattention sur la conduite.Les résultats des différents travaux menés montrent que la variation du rythme cardiaque est sensible à l’effort cognitif enconduite simulée. Une solution possible pour améliorer cet indicateur en vue d’atteindre une détection temps réel de l’effortcognitif, a été proposée à un concours international : le Valeo Innovation Challenge de 2016. Un second indicateur a étéidentifié : le regard du conducteur lorsque celui-ci est en VP est plus fixe que lors d’épisodes d’attention vers la conduite. Enfin,l’analyse des mesures électroencéphalographiques a révélé que le VP diminue l’attention allouée aux différentes étapes (sensorielle et cognitive) du traitement de l’information visuelle en conduite simulée. Ainsi, l’effet de différents types de pensée sur le comportement oculaire, le traitement de l’information visuelle et sur l’indicateur d’effort mis en évidence a renseigné le degré d’interférence de différents types de pensée sur l’activité de conduite.Les résultats de cette thèse montrent qu’il semble prometteur d’aller encore plus loin dans la recherche d’indicateurs sensiblesau vagabondage et possiblement aux autres états internes du conducteur pour pallier leurs effets négatifs sur l’activité deconduite. La mise en commun par une technique de fusion de données, des indicateurs sensibles au VP pourrait permettre, dans le futur, de créer des algorithmes à même de superviser le conducteur. L’automatisation progressive des véhicules couplée à une surveillance intelligente des conducteurs pourrait permettre de limiter le risque d’accident et les coûts humains, sociétaux et économiques qui leur sont imputables. / Mind-Wandering (MW) is a particular state of inattention that everyone experience in daily tasks. It involves a change of anindividual attentional focus, from the main activity to personal thoughts. During driving, MW has many negative effectsenhancing the risk of being involved and responsible for a car accident.The objective of this thesis was to better describe the impact of MW on drivers’ physiological and behavioral variables to drawup a list of sensitive indicators to this state. MW, acting as an additional task to driving, would consume some of driver'scognitive resources. It was therefore necessary to search, by an experimental approach, an indicator of cognitive effort in driving.As the multiplicity of thoughts and their characteristics could create different levels of risk for the driver, it seemed relevant tostudy the impact of these different forms of inattention on driving.Results of various work conducted, show that heart rate change is sensitive to cognitive effort during simulated driving. A possible solution to improve this indicator and try to achieve a real-time detection of cognitive effort has been proposed to an international competition: the 2016 Valeo Innovation Challenge. Another indicator has been identified: driver's eyes appeared to be more staring during MW than during episodes of attention dedicated to driving. Lastly, analysis of electroencephalographicmeasurements revealed that MW decreases attention allocated to the different stages (sensory and cognitive) of visual information processing during simulated driving. Thus, the effect of different types of thoughts on ocular behavior, visual information processing, and on the cognitive effort indicator previously highlighted, informed about the degree of interference of different types of thoughts on the driving activity.The results of this thesis show that it seems promising to go even further in the search for indicators sensitive to MW and possibly other internal states to mitigate their negative effects on the driving activity. Gathering MW sensitive indicators, using atechnique of data fusion, could allow, in the future, the possibility to create algorithms able to supervise the driver. The gradualautomation of vehicles coupled with intelligent monitoring of drivers could limit the risk of accidents and human, societal andeconomic costs that are attributable to them.

The utilisation of navigation systems in cars has given rise to road safety concerns, and the design and functionality of such systems must therefore be adjusted to the users’ needs, since they have to divide their attention between driving and the operation of the navigation system. The study was aimed at finding the optimum system response time (SRT) which would enable a driver to focus as much as possible on the road while attaining an efficient task completion time using an electronic navigational system. The research project consists of two separate experiments and was completed by 10 subjects. Experiment 1 included a temporal reproduction task and a secondary memory task. The subjects had to memorise two symbols and then reproduce six time spans ranging from 1 to 30 s to provide a baseline measurement of their time estimation abilities. Experiment 2 consisted of a simulated automobile driving task. While driving in the simulator the subjects completed a memorising task displayed on a touch screen. The task was presented with seven different system response times (SRTs) ranging from 0 to 30 s. The effects of different SRTs on the eye movement from road to monitor, regarding the duration of fixation and the frequency of change were evaluated. The distribution of gazes to the secondary task was analysed to provide information about the time estimation performance in the driving simulator. Other dependent variables tested were the accuracy of selected items, memory game performance, drive performance and the subjective state of the test person. The results of this study can be employed to find the optimum duration of inter-task delays for in-vehicle technical devices. / Psychology / M.A. (Psychology)

In this thesis I deal with the problem of navigation and autonomous driving using convolutional neural networks. I focus on the main approaches utilizing sensory inputs described in literature and the theory of neural networks, imitation and reinforcement learning. I also discuss the tools and methods applicable to driving systems. I created two deep learning models for autonomous driving in simulated environment. These models use the Dataset Aggregation and Deep Deterministic Policy Gradient algorithms. I tested the created models in the TORCS car racing simulator and compared the result with available sources.

This thesis studies navigation and autonomous driving using convolutional neural networks. It presents main approaches to this problem used in literature. It describes theory of neural networks and imitation and reinforcement learning. It also describes tools and methods suitable for a driving system. There are two simulation driving models created using learning algorithms DAGGER and DDPG. The models are then tested in car racing simulator TORCS.