Dissertations / Theses on the topic 'Autonomous Train'

Le déploiement des trains autonomes soulève de nombreuses questions et défis, notamment ceux liés au niveau de sécurité visé, qui doit être globalement au moins équivalent à celui du système existant, ainsi que les moyens à mettre en œuvre pour l'atteindre. Conventionnellement, la mise en sécurité d'un système ferroviaire global ou d'un sous-système défini comprend une phase d'analyse des risques et une phase de maîtrise des situations dangereuses. Ainsi, pour tout système technique ferroviaire, qu'il soit classique, automatique ou autonome, un niveau de sécurité acceptable doit être assuré. Dans le contexte des trains autonomes, les défis liés à leur sécurité incluent les aspects émergents de l'intelligence artificielle, le transfert de tâches et de responsabilités du conducteur vers des systèmes décisionnels automatiques, ainsi que les aspects liés à l'autonomisation, tels que la transition entre les modes et la gestion des modes dégradés. La méthodologie de démonstration de sécurité des trains autonomes, doit ainsi prendre en compte les risques engendrés par l'ensemble de ces aspects. Autrement dit, elle doit définir l'ensemble les activités de sécurité (liées à l'introduction de l'autonomie et des Systèmes d'Intelligence Artificielle), complémentaires à la démonstration de sécurité conventionnelle. Dans ce cadre, l'objectif de cette thèse est de contribuer à l'élaboration d'une démarche d'assurance de sécurité pour les trains autonomes. Concrètement, cette thèse propose trois contributions principales. Premièrement, nous proposons une méthodologie globale de haut niveau pour la structuration et la présentation de l'argumentation de sécurité pour les trains autonomes. La méthodologie est basée sur une approche orientée objectifs de sécurité (goal-based safety) en utilisant le formalisme graphique GSN (Goal Structuring Notation). Ensuite, nous proposons une modélisation de la conscience de situation (situational awareness) d'un système de conduite autonome d'un train, intégrant le processus de l'analyse dynamique des risques ferroviaires. Ce modèle permettra au système de conduite autonome de percevoir, de comprendre, d'anticiper et de s'adapter à des situations inconnues dans son environnement tout en prenant des décisions sûres. Le modèle est illustré à travers un cas d'étude concernant la détection et l'évitement d'obstacles sur la voie ferroviaire. Dernièrement, nous élaborons une approche de prise de décision basée sur l'évaluation dynamique des risques. L'approche utilise le Processus Décisionnel de Markov Partiellement Observable (POMDP) et vise à assurer une surveillance continue de l'environnement pour garantir la sécurité opérationnelle, en particulier la prévention des collisions. L'approche repose sur le maintien d'un niveau de risque acceptable grâce à une estimation et une actualisation continues de l'état opérationnel du train et des données de perception de l'environnement
The deployment of autonomous trains raises many questions and challenges, particularly concerning the required safety level, which must be globally at least equivalent to that of the existing systems, along with how to achieve it. Conventionally, ensuring the safety of a global railway system or a defined subsystem includes analyzing risks and effectively handling dangerous situations. Therefore, for any technical railway system, whether it is conventional, automatic, or autonomous, an acceptable level of safety must be ensured. In the context of autonomous trains, safety challenges include aspects related to the use of artificial intelligence models, the transfer of tasks and responsibilities from the driver to automatic decision-making systems, and issues related to autonomy, such as mode transitions and management of degraded modes. Thus, the safety demonstration methodology for autonomous trains must take into account the risks generated by all these aspects. In other words, it must define all the safety activities (related to the introduction of autonomy and artificial intelligence systems), complementary to conventional safety demonstration. In this context, this dissertation proposes three main contributions towards the development of a safety assurance methodology for autonomous trains. Firstly, we establish a high-level framework for structuring and presenting safety arguments for autonomous trains. This framework is based on a goal-based approach represented by the graphical modeling Goal Structuring Notation (GSN). Then, we propose a model for the situational awareness of the automated driving system of an autonomous train, that integrating the process of dynamic risk assessment. This model enables the automated driving system to perceive, understand, anticipate and adapt its behavior to unknown situations while making safe decisions. This model is illustrated through a case study related to the obstacle detection and avoidance. Finally, we develop a decision-making approach based on dynamic risk assessment. The approach is based on Partially Observable Markov Decision Processes (POMDP) and aims to ensure continuous environmental monitoring to guarantee operational safety, particularly collision prevention. The approach is based on maintaining an acceptable level of risk through continuous estimation and updating of the train's operational state and environmental perception data

Now and then train accidents occur. Collisions between trains and objects such as animals, humans, cars, and fallen trees can result in casualties, severe damage on the train, and delays in the train traffic. Thus, train collisions are a considerable problem with consequences affecting society substantially. The company Termisk Systemteknik AB has on commission by Rindi Solutions AB investigated the possibility to detect anomalies on the railway using a trainmounted thermal imaging camera. Rails are also detected in order to determine if an anomaly is on the rail or not. However, the rail detection method does not work satisfactory at long range. The purpose of this master’s thesis is to improve the previous rail detector at long range by using machine learning, and in particular deep learning and a convolutional neural network. Of interest is also to investigate if there are any advantages using cross-modal transfer learning. A labelled dataset for training and testing was produced manually. Also, a loss function tailored to the particular problem at hand was constructed. The loss function was used both for improving the system during training and evaluate the system’s performance during testing. Finally, eight different approaches were evaluated, each one resulting in a different rail detector. Several of the rail detectors, and in particular all the rail detectors using crossmodal transfer learning, perform better than the previous rail detector. Thus, the new rail detectors show great potential to the rail detection problem.

La thèse aborde les défis du monitoring de l’environnement et de détection des anomalies, notamment des obstacles, pour un train de fret autonome. Bien que traditionnellement, les transports ferroviaires étaient sous la supervision humaine, les trains autonomes offrent des perspectives d’avantages en termes de coûts, de temps et de sécurité. Néanmoins, leur exploitation dans des environnements complexes pose d’importants enjeux de sûreté. Au lieu d’une approche supervisée nécessitant des données annotées onéreuses et limitées, cette recherche adopte une technique non supervisée, utilisant des données non étiquetées pour détecter les anomalies en s’appuyant sur des techniques capables d’identifier les comportements atypiques.Deux modèles de surveillance environnementale sont présentés : le premier, basé sur un autoencodeur convolutionnel (CAE), est dédié à l’identification d’obstacles sur la voie principale; le second, une version avancée incorporant le transformeur de vision (ViT), se concentre sur la surveillance générale de l’environnement. Tous deux exploitent des techniques d’apprentissage non supervisé pour la détection d’anomalies.Les résultats montrent que la méthode mise en avant apporte des éléments pertinents pour le monitoring de l’environnement du train de fret autonome, ayant un potentiel pour renforcer sa fiabilité et sécurité. L’utilisation de techniques non supervisées démontre ainsi l’utilité et la pertinence de leur adoption dans un contexte d’application pour le train autonome
The thesis addresses the challenges of monitoring the environment and detecting anomalies, especially obstacles, for an autonomous freight train. Although traditionally, rail transport was under human supervision, autonomous trains offer potential advantages in terms of costs, time, and safety. However, their operation in complex environments poses significant safety concerns. Instead of a supervised approach that requires costly and limited annotated data, this research adopts an unsupervised technique, using unlabeled data to detect anomalies based on methods capable of identifying atypical behaviors.Two environmental surveillance models are presented : the first, based on a convolutional autoencoder (CAE), is dedicated to identifying obstacles on the main track; the second, an advanced version incorporating the vision transformer (ViT), focuses on overall environmental surveillance. Both employ unsupervised learning techniques for anomaly detection.The results show that the highlighted method offers relevant insights for monitoring the environment of the autonomous freight train, holding potential to enhance its reliability and safety. The use of unsupervised techniques thus showcases the utility and relevance of their adoption in an application context for the autonomous train

Ce travail de thèse contribue au projet collaboratif "Train de Fret Autonome'' porté par la SNCF et a pour but d’automatiser la conduite du train de Fret afin de garantir une conduite autonome fiable et sécurisée dans toutes conditions environnementales. Dans ce contexte, nos contributions concernent le développement d’estimateurs permettant de reconstruire avec précision la position et la vitesse longitudinale du train dans des conditions variables d’adhérence roue-rail. La difficulté majeure provient du fait que les capteurs odométriques embarqués fournissent une mesure de la rotation des roues au niveau des essieux mais ne détectent pas directement le phénomène de glissement des roues, ce qui entraîne une grande imprécision au niveau de l’estimation de la vitesse longitudinale du train. Pour faire face à ce problème, il est nécessaire de faire un recalage précis en utilisant la position du train mesurée par des radio-balises installées sur le rail. Néanmoins, cette mesure est discrète avec une période d’échantillonnage variable. L’un des challenges est de considérer les mesures du train qui combinent à la fois des mesures considérées continues et des mesures apériodiquement échantillonnées. Ainsi, la principale contribution théorique de cette thèse est la conception d’un observateur continu-discret pour une classe de systèmes multi-entrées/multi-sorties avec des sorties continues entachées de bruits de mesures et des sorties apériodiquement échantillonnées. Cet observateur est conçu pour répondre au cahier de charge du projet ``Train de Fret Autonome'' afin d’estimer la vitesse du train dans des conditions variables d’adhérence. Les performances de cet observateur sont montrées en simulation et comparées avec d’autres approches d’estimation de la vitesse du train puis ensuite validées expérimentalement via une campagne d’essais réalisée au Centre d’Essai Ferroviaire de Tronville-en-Barrois
This thesis contributes to the collaborative project "Train de Fret Autonome'' led by the SNCF and aims to control autonomous freight trains in order to guarantee reliable and safe autonomous driving in all environmental conditions. In this context, our contributions concern the development of estimators for accurately reconstructing the train position and longitudinal speed under variable wheel-rail adhesion conditions. The major difficulty arises from the fact that on-board odometric sensors provide a measurement of wheel rotation at axle level, but do not directly detect the phenomenon of wheel slippage, resulting in inaccurate estimation of the longitudinal speed of the train. To overcome this problem, it is necessary to make a precise recalibration using the position of the train measured by radio beacons installed on the rail. Nevertheless, this measurement is discrete with a variable sampling period. One of the challenges is to consider train measurements that combine both continuous and aperiodically sampled measurements. Thus, the main theoretical contribution of this thesis is the design of a continuous-discrete observer for a class of multi-input/multi-output systems with continuous noisy outputs and aperiodically sampled outputs. This observer is designed to meet the specifications of the "Autonomous Freight Train" project, in order to estimate train speed under variable adhesion conditions. The performance of this observer is shown in simulation and compared with other approaches to train speed estimation, then validated experimentally via an experimental test program carried out at the Centre d'Essai Ferroviaire of Tronville-en-Barrois

Optimizing routes for multiple autonomous vehiclesin complex traffic situations can lead to improved efficiency intraffic. Attempting to solve these optimization problems centrally,i.e. for all vehicles involved, often lead to algorithms that exhibitthe curse of dimensionality: that is, the computation time andmemory needed scale exponentially with the number of vehiclesresulting in infeasible calculations for moderate number ofvehicles. However, using a numerical framework called tensordecomposition one can calculate and store solutions for theseproblems in a more manageable way. In this project, we investi-gate different tensor decomposition methods and correspondingalgorithms for solving optimal control problems, by evaluatingtheir accuracy for a known solution. We also formulate complextraffic situations as optimal control problems and solve them.We do this by using the best tensor decomposition and carefullyadjusting different cost parameters. From these results it canbe concluded that the Sequential Alternating Least Squaresalgorithm used with canonical tensor decomposition performedthe best. By asserting a smooth cost function one can solve certainscenarios and acquire satisfactory solutions, but it requiresextensive testing to achieve such results, since numerical errorsoften can occur as a result of an ill-formed problem.
Att optimera färdvägen för flertalet au-tonoma fordon i komplexa trafiksituationer kan leda till effekti-vare trafik. Om man försöker lösa dessa optimeringsproblemcentralt, för alla fordon samtidigt, leder det ofta till algorit-mer som uppvisar The curse of dimensionality, vilket är då beräkningstiden och minnes-användandet växer exponentielltmed antalet fordon. Detta gör många problem olösbara för endasten måttlig mängd fordon. Däremot kan sådana problem hanterasgenom numeriska verktyg så som tensornedbrytning. I det här projektet undersöker vi olika metoder för tensornedbrytningoch motsvarandes algoritmer för att lösa optimala styrproblem,genom att jämföra dessa för ett problem med en känd lösning.Dessutom formulerar vi komplexa trafiksituationer som optimalastyrproblem för att sedan lösa dem. Detta gör vi genom attanvända den bästa tensornedbrytningen och genom att noggrantanpassa kostnadsparametrar. Från dessa resultat framgår det att Sequential Alternating Least Squaresalgoritmen, tillsammans medkanonisk tensornedbrytning, överträffade de andra algoritmersom testades. De komplexa trafiksituationerna kan lösas genomatt ansätta släta kostnadsfunktioner, men det kräver omfattandetestning för att uppnå sådana resultat då numeriska fel lätt kan uppstå som ett resultat av dålig problemformulering.
Kandidatexjobb i elektroteknik 2020, KTH, Stockholm

Autonomous Sensory Meridian Response (ASMR) is a rapidly emerging but understudied sensory concept which according to earlier studies can help relieve symptoms of for exampledepression, anxiety, pain and stress while creating feelings of relaxation. The experience begins with a triggering audio or audio-visual stimulus which causes a physical reaction described as a tingling sensation beginning at the back of the head and further towards the periphery of the body. A previous study suggested that ASMR and neuroticism are related. Anxiety is a subfacet of neuroticism, which is why this study explored the relationship between State- and Trait-anxiety (S- and T-Anxiety) and ASMR to investigate if and in what way they are related. The research questions were “Is the level of State and Trait Anxiety associated with the level of ASMR experience?” and “Are there certain ASMR stimuli that are stronger related to State and Trait Anxiety than other ASMR stimuli?”. A total of 35 participants completed the questionnaires, of which the majority were between the ages 25-34, female, employed, coliving and answered that their highest level of education were studies at college/university. The results from the current self-report correlational study showed that ASMR and T-Anxiety are significantly negatively correlated, however, it failed to show any significant correlation between S-Anxiety and ASMR. Results also showed that one ASMR stimulus, finger flutters, significantly correlated negatively with T-Anxiety. The rest of the individual stimuli failed to show significant correlations with either S- or T-Anxiety. This suggests that anxiety may not be the driving sub-facet when searching for explanations for what makes ASMR and neuroticism associated. It also raises the question whether or not ASMR could be suitable as a therapeutic method for managing high T-Anxiety. Finally, it seems that individual ASMRstimulus vary in their connection to anxiety. However, it is only possible to draw conclusions on this sample, not the population, due to the small sample size. Further studies are needed to at least verify these results.
Autonomous Sensory Meridian Response (ASMR) är ett snabbt växande men understuderatsensoriskt koncept som enligt tidigare studier kan underlätta symptom av exempelvis depression, ångest, smärta och stress samtidigt som det skapar en avslappnande känsla. Upplevelsen börjar med ett triggande audio eller audiovisuellt stimuli som orsakar en fysisk reaktion beskrivet som en pirrig känsla med början vid bakhuvudet och vidare mot kroppens periferi. En studie om ämnet föreslog att ASMR och neuroticism är relaterade. Ångest är en del av neuroticism, varför den här studien undersöker relationen mellan ASMR och State- och Trait-Anxiety (S- och T-Anxiety) för att undersöka hur och på vilket sätt de är relaterade. Forskningsfrågorna var ”Är nivån av State och Trait Anxiety associerad med nivån av ASMRupplevelse?” och ”Finns det specifika ASMR stimuli som är starkare relaterade till State och Trait Anxiety än andra ASMR stimuli?”. Totalt slutförde 35 deltagare enkäterna, av vilka majoriteten var mellan åldrarna 25–34, kvinnor, anställda, sammanlevande och rapporterade att studier på universitet/högskola var deras högsta nivå av utbildning. Resultaten från den här självrapporterings-korrelationsstudien visade att ASMR och T-Anxiety är signifikant negativt korrelerade, men det kunde inte visas att S-Anxiety och ASMR är signifikant korrelerade. De visade också att ett ASMR stimuli, finger flutters, är signifikant negativt korrelerad med TAnxiety. Resterande individuella ASMR stimuli hade ingen signifikant korrelation med vare sig S- eller T-Anxiety. Det här tyder på att ångest kanske inte är den drivande underkategorin i jakt på förklaring till vad som gör ASMR och neuroticism associerade. Det väcker också frågan om ASMR verkligen hade varit en lämplig terapeutisk metod för hantering av hög T-Anxiety. Till sist verkar det som att individuella ASMR stimuli varierar i sin relation till ångest. Viktigt att tillägga är dock att det bara är möjligt att dra slutsatser om urvalet, inte populationen, på grund av storleken på urvalet. Ytterligare studier behövs för att åtminstone verifiera dessa resultat.

Ce travail de thèse concerne la conception d'un système de supervision appliqué aux véhicules tout électriques autonomes à quatre roues motrices. Deux méthodes de surveillance à base de modèle sont développées en utilisant le modèle non linéaire du véhicule électrique: La première est basée sur La technique de multi observateurs pour détecter et isoler les défauts. La deuxième approche est la génération des relations de redondance analytique non linéaires (espace de parité non linéaire). En plus de la génération des résidus par la méthode de l'espace de parité, des observateurs différentiels à modes glissants sont intégrés pour l'estimation des dérivées successives des entrées et des sorties. Pour réaliser un train de trois véhicules, une commande longitudinale (commande de l'inter-distance) basée sur la mesure de l'écart inter véhiculaire est développée. La commande latérale est basée sur un suivi de trajectoire de référence qu'on impose pour tous les véhicules. Une commande de braquage est alors calculée en fonction de la différence entre la trajectoire mesurée et celle désirée. Enfin, quelques stratégies de reconfiguration sont développées pour le train de véhicules. Le principe de la reconfiguration est le choix de la stratégie à adopter pour le train de véhicules en fonction de type de défaut détecté
This work concerns the design of a supervision system applied to autonomous electric vehicle. Two based model monitoring methods are developed using the nonlinear model of electric vehicle: The first is based on multiple observers to detect and to isolate faults. The second approach is the generation of nonlinear analytical redundancy relations (nonlinear parity space). In addition to the generation of residuals by the method of parity space, differential sliding mode observers are integrated to estimate the successive derivatives of inputs and outputs. To make a train of three vehicles, a longitudinal control (control of the safety space) based on measuring the distance between vehicles is developed. The lateral control is based on a reference path tracking imposed for all vehicles. A steering command is then calculated based on the difference between the measured and the desired trajectory. Finally, some strategies are developed for the reconfiguration of the train of vehicles. The principle of the reconfiguration is the choice of strategy for the train of vehicles according to kind of detected fault

L'automobile est un moyen de transport de prédilection puisqu'elle confère beaucoup de liberté aux utilisateurs. Cependant, les coûts sociaux et économiques reliés aux accidents de la route, ainsi que les émissions polluantes produites par la combustion des hydrocarbures, représentent une problématique à laquelle le Réseau de centres d'excellence AUTO21 s'intéresse. Notamment, AUTO21 finance ce projet qui porte sur les architectures décisionnelles appliquées aux trains de véhicules. Un train de véhicule est un système de transport intelligent regroupant un ensemble de véhicules qui se suivent les uns derrière les autres, conservant une faible distance entre eux.L'architecture DCD est une architecture décisionnelle complète avec communication bidirectionnelle responsable de la prise de décision distribuée d'un groupe de véhicules opérant au sein d'un train de véhicules.L'architecture réalise la plupart des manoeuvres réalisables dans les trains de véhicules tels l'insertion ou le retrait d'un véhicule dans le train, ainsi que la gestion des situations d'urgences telles les collisions et la défaillance de certains capteurs. La validation de l'architecture est effectuée avec un groupe de quatre robots mobiles équipés d'un système de positionnement relatif qui permet de localiser les véhicules les uns par rapport aux autres. Un ensemble de 18 scénarios de tests, couvrant un large éventail des situations susceptibles de survenir dans les trains de véhicules, est réalisé. Les résultats obtenus indiquent que la prise de décision des véhicules, ainsi que la coordination des véhicules via la communication sont robustes et permettent l'exécution des manoeuvres attendues.

Northwestern Arctic Canada is one of the most rapidly warming regions in the Arctic (Serreze et al., 2009). Retrogressive thaw slumps (RTS) are one of the most dramatic thermokarst features in permafrost terrain (Kokelj et al., 2013). Many studies have focused on describing the distribution of thermokarst landscapes (i.e., Olefeldt et al., 2016), as well as change in thermokarst terrain over the historical record (i.e., Kokelj and Jorgenson, 2013). However, improved high temporal and spatial resolution monitoring of thaw slump activity is required to enhance our understanding of factors governing their growth. Recent advances in aerial and ground-based Structure from Motion (SfM), a photogrammetry application, allow for temporal and spatial high-resolution characterization of landscape changes. This thesis explores two methods in SfM photogrammetry: 1) aerial imaging using an unmanned aerial vehicle (UAV) and 2) ground-based imaging using stationary multi-camera time-lapse installations, to derive high-resolution temporal and spatial data for change detection. A trend in mean elevation change was produced, and agrees with the RTS behaviour over the study period, which supports the viability of the proposed capture method. The lack of congruency in data range suggests need for further development in terms of analyses and differencing algorithms employed. The proposed method may be feasible for employment in other fields of science in which high temporal resolution change detection is desired. This proof of concept study was conducted at a small slump on the Peel Plateau, NWT, Canada, and aims to enhance understanding of the development and perpetuation of thaw slumps, to better anticipate landscape and ecosystem responses to future climate change.

En France, il existe deux modes de traction pour les trains : la traction diesel ou la traction électrique. Chaque mode fait face à des problématiques qui lui sont propres. Dans le cas du diesel, les émissions de gaz à effet de serre étant de plus en plus contrôlées, il devient nécessaire de faire évoluer ce type de train vers une solution moins polluante. Dans le cas de la traction électrique, la consommation d’énergie entraine une chute de tension qui peut imposer un ralentissement des trains, empêchant ainsi le développement du trafic. La solution étudiée par la SNCF est l’hybridation des trains (ajout de systèmes de stockage en embarqué).Ces travaux de thèse ont pour objectif de mettre en place une méthode permettant de faire le pré-dimensionnement des systèmes de stockage embarqués dans le train. De plus, afin de tenir compte de l’influence réciproque de la gestion sur le dimensionnement, celle-ci est incluse dans le modèle de dimensionnement. La résolution du modèle global se fait à l’aide d’un algorithme d’optimisation.La méthode a été mise en place sur les deux modes de traction ferroviaire (diesel et électrique) et l’optimisation a été faite avec l’algorithme SQP (Sequential Quadratic Programming)
In France, there are two traction modes for railway: the diesel and electric traction. Each mode has its own issues. For diesel, the increasing control of the greenhouse gas emissions imposes to evolve this type of train to a less polluting solution. For electric traction, the energy consumption creates a voltage drop which can cause a traffic slowdown, which will limit the traffic development. The studied solution by SNCF is the hybridization of the train (adding storage system).Thus, these works have the objective to build a method to do the pre-sizing of storage systems embedded in trains. Moreover, to take into account the mutual influence of the sizing and the energy management, this last one is included in the sizing model. An optimization algorithm solves the global model.The method has been developed for the two traction modes (diesel and electric) and the optimization has been made with SQP algorithm (Sequential Quadratic Programming)

Made available in DSpace on 2016-04-27T18:15:49Z (GMT). No. of bitstreams: 1 Jose Carlos Brito Silva.pdf: 10582738 bytes, checksum: cf5493c30207a7fbda5b7e5f62e4d477 (MD5) Previous issue date: 2009-10-01
This research analyzes and seeks to understand the increase in informal activities in general and mobile commerce as a particular case and the conflict between the public and employees of these itinerant trade. This analysis is performed as a result of the dynamics inherent in the capitalist system that is based on socioeconomic contradictions, among them the need to generate a surplus of labor to thus have its disposal an army reserve. In that sense this analysis takes into account the urbanization process as RMSP suffered by one of the promotion of unemployment which had their causes in several variables. The phenomenon of the increase in practice of these activities, named in this study as "independent work", practiced by itinerant workers in trains and other public spaces of RMSP, has widened from the 80's resulting from economic crises that hit the country causing huge recession unemployment and an economy that can be described as the results of politics applied at that moment. This unemployment is particularly accentuated in the 90s resulting from the economic opening process undertaken by the Brazilian Government as a way of integrating the country in the process of globalization. The way this integration, was made, marked by unbridled economic openness, reflected in several aspects of the economy of the country: it has a new territoriality of plants provided by the development of technical-scientific-media informational affectry RMSP mainly because this the most industrialized in the country, it also involves a process of productive restructuring in which the industrial sector loses its economic space and it is substituted by tertiary activities; as a consequence implemented a new model of administrative organization and the production process adopted by enterprises, the model flexible also called post-fordist, requiring a new profile with more and better professional qualifications to perform the tasks, now with a greater degree of complexity and flexibility. These changes reflected directly in the levels of unemployment now reaching ever higher rates. Informal activities in general and labor have become mobile, thus a way to ensure survival for the huge number of unemployed, to consolidate is not as alternatives but rather as a temporary work, by effective work. Just as the growing number of than practitioners in itinerant trade grew the conflicts that came to rely even violently with police actions undertaken. The territory of the CPTM trains as public spaces and great appeal to the trade due to the large flow of passengers, has become one of those places where this picture was reflected socioeconomic and conflict is fierce. This territory, which offers easier to enable practitioners of this activity a condition of legalization of the organization and activities to provide decent conditions for a significant number of workers was the locus chosen in this research as part of the complex and contradictory dynamics of capitalist economy
Esta pesquisa analisa e procura compreender o aumento das atividades informais de maneira geral e do comércio ambulante como um caso específico e os conflitos entre o poder público e esses trabalhadores do comércio ambulante. Esta análise é realizada enquanto resultante de uma dinâmica inerente ao próprio sistema capitalista que se assenta em contradições socioeconômicas, dentre elas a necessidade de geração de um excedente de mão-de-obra para, dessa maneira, ter a sua disposição um exército de reserva. Nesse sentido essa análise leva em consideração o processo de urbanização sofrido pela RMSP como um dos pressupostos de promoção do desemprego que teve suas causas em várias variáveis. O fenômeno do aumento na prática dessas atividades, denominadas nessa pesquisa como trabalho autônomo , praticada pelos trabalhadores ambulantes nos trens e outros espaços públicos da RMSP, se acentuou a partir das décadas de 80 resultantes de crises econômicas que atingiram o Brasil provocando enorme recessão econômica e um desemprego que pode ser caracterizado como conjuntural. Esse desemprego acentua-se de sobremaneira na década de 90 resultante do processo de abertura econômica empreendido pelo Estado brasileiro como forma de integração do País ao processo de globalização. A forma como se deu essa integração, marcada por uma abertura econômica desenfreada, repercutiu em vários aspectos da economia do País: empreendeu-se uma nova territorialidade das plantas industriais proporcionado pelo desenvolvimento dos meios técnico-científico-informacionais atingindo principalmente a RMSP por ser esta a mais industrializada do País; a mesma também passa por um processo de reestruturação produtiva em que perde espaço o caráter industrial e ganham espaço as atividades terciárias; implanta-se um novo modelo de organização administrativa e do processo de produção adotado pelas empresas, o modelo flexível também denominado pós-fordista, que exigem um novo perfil de profissional com maior e melhor qualificação para executar as tarefas, agora com um maior grau de complexidade e flexibilidade. Essas transformações repercutiram diretamente nos níveis de desemprego que passaram a atingir índices cada vez maiores. As atividades informais em geral e o trabalho ambulante tornaram-se, dessa forma, numa forma de garantir a sobrevivência para esse enorme contingente de desempregados, passando a se consolidarem não como alternativas temporárias e sim como trabalho de fato. Da mesma forma que crescia o número de praticantes no comércio ambulante cresciam os conflitos que passaram a contar inclusive com ações policiais empreendidas violentamente. O território dos Trens da CPTM, enquanto espaços públicos e de grande apelo para o comércio devido ao grande fluxo de passageiros, passou a ser um desses lugares onde este quadro socioeconômico se refletiu e os conflitos se acirraram. Esse território que oferece condições mais fáceis de possibilitar aos praticantes dessa atividade uma condição de organização e legalização da atividade visando proporcionar condições dignas a um número significativo de trabalhadores foi o lócus escolhido nesta pesquisa como uma das partes da complexa e contraditória dinâmica da economia capitalista

En France, il existe deux modes de traction pour les trains : la traction diesel ou la traction électrique. Chaque mode fait face à des problématiques qui lui sont propres. Dans le cas du diesel, les émissions de gaz à effet de serre étant de plus en plus contrôlées, il devient nécessaire de faire évoluer ce type de train vers une solution moins polluante. Dans le cas de la traction électrique, la consommation d’énergie entraine une chute de tension qui peut imposer un ralentissement des trains, empêchant ainsi le développement du trafic. La solution étudiée par la SNCF est l’hybridation des trains (ajout de systèmes de stockage en embarqué).Ces travaux de thèse ont pour objectif de mettre en place une méthode permettant de faire le pré-dimensionnement des systèmes de stockage embarqués dans le train. De plus, afin de tenir compte de l’influence réciproque de la gestion sur le dimensionnement, celle-ci est incluse dans le modèle de dimensionnement. La résolution du modèle global se fait à l’aide d’un algorithme d’optimisation.La méthode a été mise en place sur les deux modes de traction ferroviaire (diesel et électrique) et l’optimisation a été faite avec l’algorithme SQP (Sequential Quadratic Programming)
In France, there are two traction modes for railway: the diesel and electric traction. Each mode has its own issues. For diesel, the increasing control of the greenhouse gas emissions imposes to evolve this type of train to a less polluting solution. For electric traction, the energy consumption creates a voltage drop which can cause a traffic slowdown, which will limit the traffic development. The studied solution by SNCF is the hybridization of the train (adding storage system).Thus, these works have the objective to build a method to do the pre-sizing of storage systems embedded in trains. Moreover, to take into account the mutual influence of the sizing and the energy management, this last one is included in the sizing model. An optimization algorithm solves the global model.The method has been developed for the two traction modes (diesel and electric) and the optimization has been made with SQP algorithm (Sequential Quadratic Programming)

Les robots sont l’avenir. Mais comment pouvons-nous leur apprendre de nouvelles compétences utiles? Ce travail couvre une variété de sujets, ayant tous pour but commun de faciliter l’entraînement des robots. La première composante principale de cette thèse est notre travail sur le transfert de modélisation sim2real. Lorsqu’une stratégie a été entièrement apprise en simulation, ses performances sont généralement considérablement inférieures à celles du vrai robot. Cela peut être dû à du bruit aléatoire, à des imprécisions ou à des effets non modélisés, tels que des réactions en retour. Nous introduisons une nouvelle technique pour apprendre la différence entre le simulateur et le vrai robot et pour l’utiliser afin de corriger le simulateur. Nous avons constaté que pour plusieurs de nos idées, aucune simulation appropriée n’était disponible. Par conséquent, pour la deuxième partie principale de la thèse, nous avons créé un ensemble de nouvelles simulations robotiques et de nouveaux environnements de test. Nous fournissons (1) plusieurs nouvelles simulations pour des robots existants, ainsi que des variantes d’environnements existants, qui permettent un ajustement rapide de la dynamique du robot. Nous avons également co-créé (2) le défi AIDO de Duckietown, qui est un concours de robotique en direct à grande échelle pour les conférences NIPS 2018 et ICRA 2019. Pour ce défi, nous avons créé l’infrastructure de simulation, qui permet aux participants d’entraîner leurs robots en simulation avec ou sans ROS. Il leur permet également d’évaluer automatiquement leurs soumissions sur des robots en direct dans un ”Robotarium”. Afin d’évaluer la compréhension et l’acquisition continue de langage par un robot, nous avons développé le (3) Test d’Interaction Multimodal Homme-Robot (MHRI). Cet ensemble de tests contient plusieurs heures d’enregistrements annotés de différentes personnes montrant et pointant des objets ménagers courants, le tout du point de vue d’un robot. La nouveauté et la difficulté de cette tâche découlent du bruit réaliste inclus dans le jeu de données: la plupart des personnes n’était pas de langue maternelle anglaise, certains objets étaient obstrués et personne n’avait reçu d’instructions détaillées sur la manière de communiquer avec le robot, entraînant des interactions très naturelles. Nous avons constaté un manque flagrant de simulations d’environnements domestiques réalistes, avec annotations sémantiques, qui permettraient à un agent d’acquérir les compétences nécessaires pour maîtriser une telle tâche. C’est pourquoi nous avons créé (4) HoME, une plate-forme de formation de robots domestiques à la compréhension du langage. L’environnement a été créé en encapsulant la base de données existante SUNCG 3D, composée de maisons, dans un moteur de jeu pour permettre aux agents simulés de parcourir ces dernières. Il intègre un moteur acoustique très détaillé et un moteur sémantique pouvant générer des descriptions d’objets en relation avec d’autres objets, meubles et pièces. La troisième et dernière contribution principale de ce travail prend en considération le fait qu’un robot peut se trouver dans un nouvel environnement non couvert par la simulation. Dans un tel cas, nous fournissons une nouvelle approche qui permet à l’agent de reconstruire une scène 3D à partir d’une seule image 2D en apprenant l’intégration d’objets. Le principal inconvénient de ce travail est qu’il ne prend actuellement pas en charge de manière fiable la reconstruction de couleur et de texture. Nous avons testé cette approche sur une tâche de rotation mentale, courante dans les tests de QI, et avons constaté que notre modèle arrivait nettement mieux à reconnaître et à faire pivoter des objets que plusieurs modèles de référence
Robots are the future. But how can we teach them useful new skills? This work covers a variety of topics, all with the common goal of making it easier to train robots. The first main component of this thesis is our work on model-building sim2real transfer. When a policy has been learned entirely in simulation, the performance of this policy is usually drastically lower on the real robot. This can be due to random noise, to imprecisions, or to unmodelled effects like backlash. We introduce a new technique for learning the discrepancy between the simulator and the real robot and using this discrepancy to correct the simulator. We found that for several of our ideas there weren’t any suitable simulations available. Therefore, for the second main part of the thesis, we created a set of new robotic simulation and test environments. We provide (1) several new robot simulations for existing robots and variations on existing environments that allow for rapid adjustment of the robot dynamics. We also co-created (2) the Duckietown AIDO challenge, which is a large scale live robotics competition for the conferences NIPS 2018 and ICRA 2019. For this challenge we created the simulation infrastructure, which allows participants to train their robots in simulation with or without ROS. It also lets them evaluate their submissions automatically on live robots in a ”Robotarium”. In order to evaluate a robot’s understanding and continuous acquisition of language, we developed the (3) Multimodal Human-Robot Interaction benchmark (MHRI). This test set contains several hours of annotated recordings of different humans showing and pointing at common household items, all from a robot’s perspective. The novelty and difficulty in this task stems from the realistic noise that is included in the dataset: Most humans were non-native English speakers, some objects were occluded and none of the humans were given any detailed instructions on how to communicate with the robot, resulting in very natural interactions. After completing this benchmark, we realized the lack of simulation environments that are sufficiently complex to train a robot for this task. This would require an agent in a realistic house settings with semantic annotations. That is why we created (4) HoME, a platform for training household robots to understand language. The environment was created by wrapping the existing SUNCG 3D database of houses in a game engine to allow simulated agents to traverse the houses. It integrates a highly-detailed acoustic engine and a semantic engine that can generate object descriptions in relation to other objects, furniture, and rooms. The third and final main contribution of this work considered that a robot might find itself in a novel environment which wasn’t covered by the simulation. For such a case we provide a new approach that allows the agent to reconstruct a 3D scene from 2D images by learning object embeddings, since especially in low-cost robots a depth sensor is not always available, but 2D cameras a common. The main drawback of this work is that it currently doesn’t reliably support reconstruction of color or texture. We tested the approach on a mental rotation task, which is common in IQ tests, and found that our model performs significantly better in recognizing and rotating objects than several baselines

Le prochain secteur à être fortement automatisé est sans doute celui du transport routier. D’abord en environnement peu complexe comme c’est déjà le cas sur autoroute, puis à terme dans les scénarios les plus difficiles, comme ceux rencontrés au coeur du trafic urbain. Une étape intermédiaire consiste à introduire des véhicules autonomes en milieu urbain, mais hors du trafic, comme sur des sites fermés. Ces environnements présentent des caractéristiques particulières, comme un encombrement important. Alors, pour être en mesure de naviguer en toute autonomie, les véhicules doivent être à la fois agiles, précis et sûrs, mais aussi efficaces, donc rapides. Ce travail de thèse adresse deux problématiques clés de la navigation autonome en milieu encombré : la planification et le suivi de chemin pour les véhicules à deux trains directeurs (4WS). En effet, la cinématique 4WS est la réponse des constructeurs au besoin d’agilité et nécessite le développement de techniques de commande adaptées. La planification de chemin en temps-réel permet l’évitement d’obstacles, de façon à rendre le service robuste aux aléas de l’environnement. Le contrôle automatique de la direction permet ensuite de suivre ces chemins avec précision. La principale difficulté concernant la planification de chemin est le contraste entre la faible puissance de calcul embarquée sur le véhicule et la forte contrainte temporelle qu’impose le besoin de réactivité. Il s’agit donc de choisir l’approche adéquate et de mettre en place les simplifications algorithmiques les plus efficaces, c’est-à-dire qui réduisent la quantité de calculs nécessaires sans réduire la qualité des chemins générés. De plus, la planification de chemin en milieu encombré pose le problème des impasses dans lesquelles les méthodes les plus réactives restent coincées. De fait, l’approche retenue consiste à construire une fonction de navigation en deux dimensions dont est extrait le chemin. Au sens des critères choisis, il est optimal pour un véhicule circulaire. Des stratégies sont alors mises en place pour adapter cet algorithme aux véhicules 4WS de forme rectangulaire, comme la notion centrale de double-chemin qui permet de représenter la trajectoire d’un mobile à trois degrés de liberté dans un espace en deux dimensions. Enfin, d’autres stratégies et heuristiques sont introduites pour optimiser les performances de l’algorithme de planification et lui permettre de générer des manoeuvres efficaces. Concernant le développement du contrôleur de direction, il s’agit de choisir le modèle d’évolution qui retranscrit au mieux le comportement du véhicule vis-à-vis du double chemin qu’il doit suivre, puis d’en déduire les lois de commande des trains avant et arrière. Ayant été éprouvé par de nombreux travaux, le modèle bicyclette 4WS a été choisi. Ce modèle cinématique permet d’introduire facilement des angles de dérive utiles à la compensation des glissements des pneus sur le sol, qu’ils soient dus à un manque d’adhérence ou à une géométrie imparfaite des trains. Des lois de commande par retour d’état sont ensuite synthétisées et une stratégie de gestion de la saturation des actionneurs est proposée. Enfin, ce premier contrôleur est décliné en une version prédictive qui apporte un suivi de chemin d’une grande stabilité sur les véhicules réels. Les contributions ont toutes été évaluées en simulation et lors d’expérimentations en vraie grandeur sur l’EZ10, une navette électrique 4WS industrielle. Enfin, il se trouve que les performances obtenues dépassent les attentes initiales
The next sector to be highly automated is probably road transport. First in an uncomplicated environment as is already the case on highway, then eventually in the most difficult scenarios, such as those encountered at the heart of the urban traffic. An intermediate step is to introduce autonomous vehicles in urban areas, but outside traffic, as on closed sites. These environments have particular characteristics, they can be very cluttered. So to be able to navigate autonomously, the vehicles have to be agile, precise and safe, but also efficient, i.e. fast. This thesis addresses two key issues of autonomous navigation in cluttered environments: path planning and tracking for bi-steerable (4WS) vehicles. Indeed, the 4WS kinematics is the response of the manufacturers to the need for agility and requires the development of adapted control techniques. Real-time path planning allows the avoidance of obstacles in order to make the service robust to the vagaries of the environment. The automatic control of the direction then makes it possible to follow these paths with precision. The main difficulty about path planning is the contrast between the low onboard computing power and the high temporal constraints imposed by the need for reactivity. It is thus necessary to choose the appropriate approach and to implement the most efficient algorithmic simplifications, i.e. reducing the amount of calculations without reducing the quality of the generated paths. In addition, path planning in cluttered environment raises the problem of dead-ends in which the most reactive methods remain stuck. Then, the adopted approach consists in constructing a navigation function in two dimensions from which the path is extracted. In the sense of the chosen criteria, it is optimal for a circular vehicle. Strategies are then put in place to adapt this algorithm to 4WS vehicles of rectangular shape, such as the central dual-path concept which allows to represent the trajectory of a mobile with three degrees of freedom in a two-dimensional space. Finally, other strategies and heuristics are introduced to optimize the performance of the planning algorithm and allow it to generate efficient maneuvers. Concerning the development of the controller, it is about choosing the evolution model which best reflects the behavior of the vehicle with respect to the dual-path that it must follow and then deducing the front and rear control laws. The 4WS bicycle model has shown itself to be very effective. This kinematic model makes it possible to easily introduce useful side-slip angles allowing to compensate the slippage of the tires on the ground, whether due to a lack of grip or an imperfect geometry of the axles. State feedback control laws are then synthesized and a strategy for managing the saturation of the actuators is proposed. Finally, this first controller is declined in a predictive version that provides a very stable tracking on real vehicles.The contributions were all evaluated in simulation and in full-scale experiments on the EZ10, an industrial 4WS electric shuttle. Finally, it turns out that the performances exceed initial expectations

Background: Work-related stress (WRS) is associated with musculoskeletal pain (MSP), changes in the autonomic nervous system (ANS) and anxiety. Objective: To determine the feasibility of a follow-up study and treatment efficacy of the Reaset Approach on MSP, ANS and State-Trait anxiety. Methods: 15 subjects with WRS and MSP were assigned into 3 groups (Body, Head-Neck, Head-Neck-Body). Each group received a single 25 minute ‘Reaset Approach’ intervention. Heart rate variability (HRV), electro-dermal activity (EDA), State Trait Anxiety (STAI) and MSP were measured. Results: HRV parameters: SDNN increased in 13 of 15 subjects while SD1 and SD2 increased in 12 of 15 subjects. EDA reduced in 10 of 14 subjects. State Anxiety reduced in all subjects and Trait Anxiety reduced in 14 of 15 subjects. MSP reduced in all subjects after the intervention and were still lower three days afterwards. Conclusions: This pilot study determined that a follow-up study can ensue provided minor modifications are implemented and that the ‘Reaset Approach’ has an influence on the ANS, anxiety and MSP. Results do differ between groups. The intervention groups including the head and neck modalities demonstrated better results
Hintergrund: Arbeitsbedingter Stress (ABS) ist verbunden mit muskelschmerzen, Veränderungen im autonomen Nervensystem (ANS) und Angst. Ziel: Machbarkeit einer Follow-up-Studie und Wirksamkeit der Behandlung des Reaset Ansatzes auf ANS, Muskelschmerzen und State und Trait- Angst bestimmen. Methoden: 15 Patienten mit ABS und Muskelschmerzen wurden in 3 Gruppen eingeteilt (Körper, Kopf-Hals, Kopf-Hals-Körper). Jede Gruppe erhielt eine einzige 25 Minuten dauernde 'Reaset Approach’-Behandlung. Herzfrequenzvariabilität (HRV), elektro-dermale Aktivität (EDA), State-Trait-Angstsinventar (STAI) und Muskelschmerzen (SF-MPQ) wurden gemessen. Ergebnisse: Die HRV-wert: SDNN ist bei 13 von 15 Probanden erhöht, während SD1 und SD2 bei 12 von 15 Probanden zugenommen hat. EDA war bei 10 von 14 Probanden reduziert. Die State-Angst hat bei allen Probanden und die Trait-Angst bei 14 der 15 Probanden abgenommen. Muskelschmerzen waren bei alle Probanden anschließend an und drei Tage nach der Intervention reduziert. Schlussfolgerung: Diese Pilotstudie hat gezeigt, dass eine Follow-up-Studie fortgesetzt werden kann, sofern kleinere Änderungen durchgeführt werden. Die 'Reaset Approach’ hat einen günstigen Einfluss auf die ANS, State-Trait-Angst und Muskelschmerzen. Ergebnisse zwischen den Gruppen sind unterschiedlich. Die Interventionsgruppen mit einschließlich der Kopf-Hals-Modalitäten zeigten bessere Ergebnisse

碩士
元智大學
工業工程與管理學系
107
In recent years, because of the outstanding performance of reinforcement learning in certain cases, and in response to the issue of Industry 4.0. The industry has considered using reinforcement learning to solve some problems. In this study, process control and autonomous car were selected as a case study. And study the difference between reinforcement learning and traditional methods. Nowadays industries uses PID to solve various control problems. Part of the process control, such as the cement process which has a large time delay, the PID does not have the ability to control it accurately . We often need experienced staff to monitor the production status of the process. And this study hopes to use reinforcement learning to improve these problems, using a first-order time delay model to simulate the environment of process control. In the real environment, due to factors such as raw materials and temperature the environment will change, so this situation will also be considered when simulating the environment. Using this environment to train the DQN model, and compare the average and standard deviation of the trained model and the traditional PID model for a period of control behavior. Then, in order to immitate more like the real-world process control, the second-order time delay model is used to simulate the training and to control and compare. In the research of autonomous car, traditional method need to be equipped with guiding equipment such as wires or ribbons, but there will be problems in some workshops. Therefore, the DDPG model is used to solve these problems, establish an environment for simulation training, and increase obstacles and environmental conditions, finally compare differences between traditional method and reinforcement learning. After training, the reinforcement of the learning model in the simple problem is better than the traditional method, and because of the characteristics of the reinforcement learning, the trained model can be monitored without personnel.

Connected and Autonomous Vehicles (CAVs) benefit from both connectivity between vehicles and city infrastructures and automation of vehicles. In this respect, CAVs can improve safety and reduce traffic congestion and environmental impacts of daily commutes through making collaborative decisions. This dissertation studies how to reduce the energy consumption of vehicles and traffic congestion by making high-level routing decisions of CAVs. The first half of this dissertation considers the problem of eco-routing (finding the energy-optimal route) for Plug-In Hybrid Electric Vehicles (PHEVs) to minimize the overall energy consumption cost. Several algorithms are proposed that can simultaneously calculate an energy-optimal route (eco-route) for a PHEV and an optimal power-train control strategy over this route. The results show significant energy savings for PHEVs with a near real-time execution time for the algorithms. The second half of this dissertation tackles the problem of routing for fleets of CAVs in the presence of mixed traffic (coexistence of regular vehicles and CAVs). In this setting, all CAVs belong to the same fleet and can be routed using a centralized controller. The routing objective is to minimize a given overall fleet traveling cost (travel time or energy consumption). It is assumed that regular vehicles (non-CAVs) choose their routing decisions selfishly to minimize their traveling time. A framework is proposed that deals with the routing interaction between CAVs and regular uncontrolled vehicles under different penetration rates (fractions) of CAVs. The results suggest collaborative routing decisions of CAVs improve not only the cost of CAVs but also that of the non-CAVs. This framework is further extended to consider congestion-aware route-planning policies for Autonomous Mobility-on-Demand (AMoD) systems, whereby a fleet of autonomous vehicles provides on-demand mobility under mixed traffic conditions. A network flow model is devised to optimize the AMoD routing and rebalancing strategies in a congestion-aware fashion by accounting for the endogenous impact of AMoD flows on travel time. The results suggest that for high levels of demand, pure AMoD travel can be detrimental due to the additional traffic stemming from its rebalancing flows, while the combination of AMoD with walking or micromobility options can significantly improve the overall system performance.

碩士
國立成功大學
企業管理學系碩博士班
96
Due to the changes in marketing environment, business and academic field have been focusing on the issue of relationship marketing, since 1980s. Unlike traditional marketing theory which puts emphasis on attracting new customers, businesses nowadays try to keep their old customers. Therefore, today, businesses come up with various royalty programs. Usually, scholars focus their research on the classification of royalty programs or on the positive effects like increasing royalty; however, this research will mainly focus on consumer reactance which has only been taken into consideration by some researchers recently. Based on the theory of psychological reactance presented by Brehm (1966) with the classification of relational marketing level by Wendlandt and Schrader (2007), the research elaborates how royalty program bonds influence consumers’ perceived utility from financial, social and contractual variables. The research analyzes the influences of consumers’ personal traits, the trait reactance tendency construct and the importance of autonomous buying behavior construct on consumers’ mental reactance and consumers’ following behavioral intention. The research uses chain Movie Theater as its experiment subject, develops three royalty program bonds, designs 16 situations, and gives out questionnaires randomly. 363 questionnaires are collected and then analyzed by regression analysis and linear structure relation in order to analyze the data and to support the hypothesis. Following are the conclusions: 1.Of all the royalty program bonds, only contractual bond will arouse consumers’ reactance. 2.Both financial and social bonds will enhance perceived utility, but the former is more evident than the later. On the other hand, contractual bond will decrease perceived utility. 3.Consumers who put emphasis on the importance of autonomous buying behavior or have the trait reactance tendency will increase the possibility of reactance. 4.The influence of perceived utility strongly conducts the situational reactance. 5.Situational reactance will cause businesses negative effects on consumers’ behavioral intention, such as engaging in negative word-of-mouth, or refusing to participate in the program or to repurchase. Finally, according to the conclusions, the research suggests that when adopting relationship marketing, businesses might avoid arousing consumers’ reactance which might lead to negative effects. Hence, businesses can try to learn to apply the three bonds of royalty program to strengthen perceived utility, to prevent reactance and to achieve the goal of customer retention.

This thesis addresses the problem of modelling visual attention in the context of autonomous off-road robots. The goal of using visual attention mechanisms is to allow robots to focus perception on the aspects of the environment that are more relevant to the task at hand. As this work will show, this capability is a promoter of robustness and computational parsimony in both obstacle and trail detection. These features are key enablers of fast and energetically efficient field robots. One of the major challenges in modelling visual attention emerges from the need to ensure that the model is able to manage the speed-accuracy trade-off in the face of context and task changes. This thesis shows that this trade-off is handled if the cognitive process of visual attention is modelled as a self-organising process, whose operation is modulated by the robot’s action selection process. By closing the loop from the action selection process to the perceptual one, the latter is able to perform on a by-need basis, anticipating actual robot moves. To endow visual attention with self-organising properties, this work gets inspiration from Nature. Concretely, the mechanisms underlying the ability that army ants have of foraging in a self-organising way are used as metaphor to solve the task of searching, also in a self-organising way, for obstacles and trails in the robot’s visual field. The solution proposed in this thesis is to have multiple covert foci of attention operating as a swarm via pheromone-based interactions. This work represents the first embodied realisation of swarm cognition, which is a new born field that aims to uncover the basic principles of cognition by inspecting the self-organising properties of the collective intelligence exhibited by social insects. Hence, this thesis contributes to robotics as an engineering discipline, and to robotics as a modelling discipline supporting the study of adaptive behaviour.

Background: Work-related stress (WRS) is associated with musculoskeletal pain (MSP), changes in the autonomic nervous system (ANS) and anxiety. Objective: To determine the feasibility of a follow-up study and treatment efficacy of the Reaset Approach on MSP, ANS and State-Trait anxiety. Methods: 15 subjects with WRS and MSP were assigned into 3 groups (Body, Head-Neck, Head-Neck-Body). Each group received a single 25 minute ‘Reaset Approach’ intervention. Heart rate variability (HRV), electro-dermal activity (EDA), State Trait Anxiety (STAI) and MSP were measured. Results: HRV parameters: SDNN increased in 13 of 15 subjects while SD1 and SD2 increased in 12 of 15 subjects. EDA reduced in 10 of 14 subjects. State Anxiety reduced in all subjects and Trait Anxiety reduced in 14 of 15 subjects. MSP reduced in all subjects after the intervention and were still lower three days afterwards. Conclusions: This pilot study determined that a follow-up study can ensue provided minor modifications are implemented and that the ‘Reaset Approach’ has an influence on the ANS, anxiety and MSP. Results do differ between groups. The intervention groups including the head and neck modalities demonstrated better results.:I. Abstract (En) III II. Abstract (De) IV III. Table of Contents V IV. Index of figures VIII V. Index of tables IX VI. Index of abbreviations X 1 Introduction 1 2 Background 2 2.1 Work-related musculoskeletal pain 2 2.2 Work-related stress 3 2.3 Osteopathy and the autonomic nervous system 3 2.4 Stress, pain and osteopathy 4 3 Questions 6 3.1 Feasibility 6 3.2 Treatment effect 6 4 Methods 7 4.1 Study design 7 4.2 Participants 8 4.2.1 Inclusion criteria 8 4.2.2 Exclusion criteria 8 4.2.3 Recruitment 8 4.2.4 Randomization 10 4.3 Parameters 11 4.3.1 Heart rate variability 11 4.3.2 Electro-dermal activity 11 4.3.3 State anxiety 11 4.3.4 Trait anxiety 12 4.3.5 Perceived pain 12 4.4 Measuring Instruments 13 4.4.1 Heart rate variability 13 4.4.2 Electro-dermal Activity 13 4.4.3 State-Trait Anxiety Inventory 13 4.4.4 Short-Form McGill Pain Questionnaire 13 4.5 Interventions 14 4.5.1 Intervention ‘B’: Body 14 4.5.2 Intervention ‘HN’: Head and Neck 16 4.5.3 Intervention ‘HNB’: Head, Neck and Body 16 4.6 Study flow 18 4.7 Statistics 20 5 Results 21 5.1 Autonomic nervous system: Heart rate variability 21 5.1.1 SDNN 22 5.1.2 SD1 25 5.1.3 SD2 28 5.2 Autonomic Nervous System: Electro-dermal activity 31 5.3 Anxiety 34 5.3.1 State anxiety 34 5.3.2 Trait anxiety 37 5.4 Musculoskeletal pain 39 5.4.1 Visual analogue scale 40 5.4.2 Total Short-Form McGill Pain Questionnaire 43 6 Discussion 46 6.1 Discussion of the method 46 6.2 Discussion of the results 50 6.2.1 Autonomic nervous system 50 6.2.1.1 Heart rate variability 50 6.2.1.2 Electro-dermal activity 51 6.2.2 Anxiety 51 6.2.2.1 State anxiety 51 6.2.2.2 Trait Anxiety 52 6.2.3 Musculoskeletal pain 52 6.3 Suggestions for future research 53 7 Conclusion 54 8 Literature 55 9 Addendum 63 9.1 Table: SF-MPQ with Sensory, Affective and Evaluative dimension 63 9.2 Patient Information Sheet 64 9.3 Structured telephone interview 70 9.4 Structured pre-treatment interview 72 9.5 SF-MPQ permission 73 9.6 SF-MPQ 74 9.7 STAI License 76 9.8 STAI forms Y-1 and Y-2 77
Hintergrund: Arbeitsbedingter Stress (ABS) ist verbunden mit muskelschmerzen, Veränderungen im autonomen Nervensystem (ANS) und Angst. Ziel: Machbarkeit einer Follow-up-Studie und Wirksamkeit der Behandlung des Reaset Ansatzes auf ANS, Muskelschmerzen und State und Trait- Angst bestimmen. Methoden: 15 Patienten mit ABS und Muskelschmerzen wurden in 3 Gruppen eingeteilt (Körper, Kopf-Hals, Kopf-Hals-Körper). Jede Gruppe erhielt eine einzige 25 Minuten dauernde 'Reaset Approach’-Behandlung. Herzfrequenzvariabilität (HRV), elektro-dermale Aktivität (EDA), State-Trait-Angstsinventar (STAI) und Muskelschmerzen (SF-MPQ) wurden gemessen. Ergebnisse: Die HRV-wert: SDNN ist bei 13 von 15 Probanden erhöht, während SD1 und SD2 bei 12 von 15 Probanden zugenommen hat. EDA war bei 10 von 14 Probanden reduziert. Die State-Angst hat bei allen Probanden und die Trait-Angst bei 14 der 15 Probanden abgenommen. Muskelschmerzen waren bei alle Probanden anschließend an und drei Tage nach der Intervention reduziert. Schlussfolgerung: Diese Pilotstudie hat gezeigt, dass eine Follow-up-Studie fortgesetzt werden kann, sofern kleinere Änderungen durchgeführt werden. Die 'Reaset Approach’ hat einen günstigen Einfluss auf die ANS, State-Trait-Angst und Muskelschmerzen. Ergebnisse zwischen den Gruppen sind unterschiedlich. Die Interventionsgruppen mit einschließlich der Kopf-Hals-Modalitäten zeigten bessere Ergebnisse..:I. Abstract (En) III II. Abstract (De) IV III. Table of Contents V IV. Index of figures VIII V. Index of tables IX VI. Index of abbreviations X 1 Introduction 1 2 Background 2 2.1 Work-related musculoskeletal pain 2 2.2 Work-related stress 3 2.3 Osteopathy and the autonomic nervous system 3 2.4 Stress, pain and osteopathy 4 3 Questions 6 3.1 Feasibility 6 3.2 Treatment effect 6 4 Methods 7 4.1 Study design 7 4.2 Participants 8 4.2.1 Inclusion criteria 8 4.2.2 Exclusion criteria 8 4.2.3 Recruitment 8 4.2.4 Randomization 10 4.3 Parameters 11 4.3.1 Heart rate variability 11 4.3.2 Electro-dermal activity 11 4.3.3 State anxiety 11 4.3.4 Trait anxiety 12 4.3.5 Perceived pain 12 4.4 Measuring Instruments 13 4.4.1 Heart rate variability 13 4.4.2 Electro-dermal Activity 13 4.4.3 State-Trait Anxiety Inventory 13 4.4.4 Short-Form McGill Pain Questionnaire 13 4.5 Interventions 14 4.5.1 Intervention ‘B’: Body 14 4.5.2 Intervention ‘HN’: Head and Neck 16 4.5.3 Intervention ‘HNB’: Head, Neck and Body 16 4.6 Study flow 18 4.7 Statistics 20 5 Results 21 5.1 Autonomic nervous system: Heart rate variability 21 5.1.1 SDNN 22 5.1.2 SD1 25 5.1.3 SD2 28 5.2 Autonomic Nervous System: Electro-dermal activity 31 5.3 Anxiety 34 5.3.1 State anxiety 34 5.3.2 Trait anxiety 37 5.4 Musculoskeletal pain 39 5.4.1 Visual analogue scale 40 5.4.2 Total Short-Form McGill Pain Questionnaire 43 6 Discussion 46 6.1 Discussion of the method 46 6.2 Discussion of the results 50 6.2.1 Autonomic nervous system 50 6.2.1.1 Heart rate variability 50 6.2.1.2 Electro-dermal activity 51 6.2.2 Anxiety 51 6.2.2.1 State anxiety 51 6.2.2.2 Trait Anxiety 52 6.2.3 Musculoskeletal pain 52 6.3 Suggestions for future research 53 7 Conclusion 54 8 Literature 55 9 Addendum 63 9.1 Table: SF-MPQ with Sensory, Affective and Evaluative dimension 63 9.2 Patient Information Sheet 64 9.3 Structured telephone interview 70 9.4 Structured pre-treatment interview 72 9.5 SF-MPQ permission 73 9.6 SF-MPQ 74 9.7 STAI License 76 9.8 STAI forms Y-1 and Y-2 77