1,720,994 research outputs found
Classification of Drivers' Workload Using Physiological Signals in Conditional Automation
Full text linkThe use of automation in cars is increasing. In future vehicles, drivers will no longer be in charge of the main driving task and may be allowed to perform a secondary task. However, they might be requested to regain control of the car if a hazardous situation occurs (i.e., conditionally automated driving). Performing a secondary task might increase drivers' mental workload and consequently decrease the takeover performance if the workload level exceeds a certain threshold. Knowledge about the driver's mental state might hence be useful for increasing safety in conditionally automated vehicles. Measuring drivers' workload continuously is essential to support the driver and hence limit the number of accidents in takeover situations. This goal can be achieved using machine learning techniques to evaluate and classify the drivers' workload in real-time. To evaluate the usefulness of physiological data as an indicator for workload in conditionally automated driving, three physiological signals from 90 subjects were collected during 25 min of automated driving in a fixed-base simulator. Half of the participants performed a verbal cognitive task to induce mental workload while the other half only had to monitor the environment of the car. Three classifiers, sensor fusion and levels of data segmentation were compared. Results show that the best model was able to successfully classify the condition of the driver with an accuracy of 95%. In some cases, the model benefited from sensors' fusion. Increasing the segmentation level (e.g., size of the time window to compute physiological indicators) increased the performance of the model for windows smaller than 4 min, but decreased for windows larger than 4 min. In conclusion, the study showed that a high level of drivers' mental workload can be accurately detected while driving in conditional automation based on 4-min recordings of respiration and skin conductance
Relevant Physiological Indicators for Assessing Workload in Conditionally Automated Driving, Through Three-Class Classification and Regression
Full text linkIn future conditionally automated driving, drivers may be asked to take over control of the car while it is driving autonomously. Performing a non-driving-related task could degrade their takeover performance, which could be detected by continuous assessment of drivers' mental load. In this regard, three physiological signals from 80 subjects were collected during 1 h of conditionally automated driving in a simulator. Participants were asked to perform a non-driving cognitive task (N-back) for 90 s, 15 times during driving. The modality and difficulty of the task were experimentally manipulated. The experiment yielded a dataset of drivers' physiological indicators during the task sequences, which was used to predict drivers' workload. This was done by classifying task difficulty (three classes) and regressing participants' reported level of subjective workload after each task (on a 0–20 scale). Classification of task modality was also studied. For each task, the effect of sensor fusion and task performance were studied. The implemented pipeline consisted of a repeated cross validation approach with grid search applied to three machine learning algorithms. The results showed that three different levels of mental load could be classified with a f1-score of 0.713 using the skin conductance and respiration signals as inputs of a random forest classifier. The best regression model predicted the subjective level of workload with a mean absolute error of 3.195 using the three signals. The accuracy of the model increased with participants' task performance. However, classification of task modality (visual or auditory) was not successful. Some physiological indicators such as estimates of respiratory sinus arrhythmia, respiratory amplitude, and temporal indices of heart rate variability were found to be relevant measures of mental workload. Their use should be preferred for ongoing assessment of driver workload in automated driving
Model of the driver's physiological state in conditionally automated driving
No full textRoad accidents are still one of the main causes of death in the world, despite all the technological advances made on cars since its invention. In particular, the driver's state is often the cause of these road accidents. To assist the driver, the level of automation of cars has increased in recent years, including advanced driver assistance systems. The next level of automation should be conditionally automated driving, where the driver is no longer responsible for the main driving task. In theory, this should reduce the number of accidents. But the fact that the driver can engage in non-driving related tasks could be very dangerous if the car suddenly requires him or her to take control. In addition, long periods of automated driving can also reduce drivers' alertness and ability to take over control in critical situations, up to causing an accident. In this regard, this thesis aims at proposing an approach to assess the driver's state continuously in the specific context of conditionally automated driving. This would allow to know if he or she is able to take over control when the car asks him or her. To achieve that goal, machine learning techniques and physiological signals were used to assess the driver's state. In particular, the prediction of four predictive risk factors was done as they are critical at this level of automation: fatigue, mental workload, affective state and situation awareness. The main contribution of this thesis is the design, implementation and validation of a model that assesses continuously the driver's state using physiological signals in conditionally automated driving (L3-SAE). This main contribution encompasses the realisation of sub-contributions to address several formulated research questions: the collection of a physiological dataset in the specific context of conditionally automated driving, the creation of a pipeline to train machine learning models able to predict the selected risk factors from the collected data, and a system to measure breathing non-intrusively. The results showed that the different risk factors could be predicted with an accuracy ranging from 73 to 99%. The fusion of physiological signals generally increased the accuracy, as did the segmentation of the signals. The physiological signals (or features) and the optimal time window to use to predict each risk factor are proposed from the results obtained with machine learning. The continuous predictions of the model in the final experiment were globally consistent and are encouraging for the use of this kind of model in our future cars. Furthermore, the sensor developed to measure breathing in a non-intrusive way proved that the breathing rate could be measured with an error of more or less one breath per second on average compared to a reference sensor. The work proposed in this thesis suggests that machine learning models can accurately predict various risk factors related to conditionally automated driving. This work can serve as a basis for improving driver condition assessment for automotive manufacturers. It can also be used in academic research, especially to further optimize the prediction of the different risk factors selected in this thesis.Les accidents de la route restent l'une des principales causes de décès dans le monde, malgré tous les progrès technologiques réalisés sur la voiture depuis son invention. En particulier, l'état du conducteur est souvent la cause de ces accidents de la route. Pour aider le conducteur, le niveau d'automatisation des voitures a augmenté ces dernières années, notamment avec les systèmes avancés d'aide à la conduite. Le prochain niveau d'automatisation devrait être la conduite conditionnellement automatisée, où le conducteur n'est plus responsable de la tâche principale de conduite. En théorie, cela devrait réduire le nombre d'accidents. Mais le fait que le conducteur puisse s'engager dans des tâches non liées à la conduite pourrait s'avérer très dangereux si la voiture lui demande soudainement de prendre le contrôle. En outre, de longues périodes de conduite automatisée peuvent également réduire la vigilance du conducteur et sa capacité à reprendre le contrôle dans des situations critiques, jusqu'à provoquer un accident. A cet égard, cette thèse vise à proposer une approche permettant d'évaluer l'état du conducteur en continu dans le contexte spécifique de la conduite automatisée conditionnelle. Pour atteindre cet objectif, des techniques d'apprentissage automatique et des signaux physiologiques ont été utilisés pour évaluer l'état du conducteur. En particulier, la prédiction de quatre facteurs de risque a été effectuée, étant jugés critiques à ce niveau d'automatisation : la fatigue, la charge mentale, l'état affectif et la conscience de situation. La principale contribution de cette thèse est la conception, l'implémentation et la validation d'un modèle qui évalue en continu l'état du conducteur à l'aide de signaux physiologiques en conduite conditionnellement automatisée (L3-SAE). Cette contribution principale comprend la réalisation de sous-contributions pour répondre à plusieurs questions de recherche formulées : la collecte d'un ensemble de données physiologiques dans le contexte spécifique de la conduite conditionnellement automatisée, la création d'un pipeline pour entraîner des modèles d'apprentissage automatique capables de prédire les facteurs de risque sélectionnés à partir des données collectées, et un système pour mesurer la respiration de manière non-intrusive. Les résultats ont montré que les différents facteurs de risque pouvaient être prédits avec une précision allant de 73 à 99%. La fusion des signaux physiologiques a généralement augmenté la précision, tout comme la segmentation des signaux. Les signaux physiologiques (ou les caractéristiques) et la fenêtre temporelle optimale à utiliser pour prédire chaque facteur de risque sont proposés à partir des résultats obtenus avec l'apprentissage automatique. Les prédictions continues du modèle dans l'expérience finale étaient globalement cohérentes et sont encourageantes pour l'utilisation de ce type de modèle dans nos futures voitures. En outre, le capteur développé pour mesurer la respiration de manière non intrusive a prouvé que le rythme respiratoire pouvait être mesuré avec une erreur de plus ou moins une respiration par seconde en moyenne par rapport à un capteur de référence. Le travail proposé dans cette thèse suggère que les modèles d'apprentissage automatique peuvent prédire avec précision divers facteurs de risque liés à la conduite conditionnellement automatisée. Ce travail peut servir de base pour améliorer l'évaluation de l'état du conducteur pour les constructeurs automobiles. Il peut également être utilisé dans la recherche académique, notamment pour optimiser davantage la prédiction des différents facteurs de risque sélectionnés dans cette thèse
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Appropriate Similarity Measures for Author Cocitation Analysis
Full text linkWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Empathic Interactions in Automated Vehicles #EmpathicCHI
No full textAutomation in driving will change the role of the drivers from actor to passive supervisor. Although the vehicle will be responsible for driving manoeuvres, drivers will need to rely on automation and understand its decisions to establish a trusting relationship between them and the vehicle. Progress has been made in conversational agents and affective machines recently. Moreover, it seems to be promising in this establishment of trust between humans and machines. We believe it is essential to investigate the use of emotional conversational agents in the automotive context to build a solid relationship between the driver and the vehicle. In this workshop, we aim at gathering researchers and industry practitioners from different fields of HCI, ML/AI, NLU and psychology to brainstorm about affective machines, empathy and conversational agent with a particular focus on human-vehicle interaction. Questions like "What would be the specificities of a multimodal and empathic agent in a car?", "How the agent could make the driver aware of the situation?"and "How to measure the trust between the user and the autonomous vehicle?"will be addressed in this workshop
Dispelling the Myths Behind First-author Citation Counts
Full text linkWe conducted a full-scale evaluative citation analysis study of scholars in the XML research field to explore just how different from each other author rankings resulting from different citation counting methods actually are, and to demonstrate the capability of emerging data and tools on the Web in supporting more realistic citation counting methods. Our results contest some common arguments for the continued
use of first-author citation counts in the evaluation of scholars, such as high correlations between author rankings by first-author citation counts and other citation
counting methods, and high costs of using more realistic citation counting methods that are not well-supported by the ISI databases. It is argued that increasingly available digital full text research papers make it possible for citation analysis studies to go beyond what the ISI databases have directly supported and to employ more
sophisticated methods
- …
