Abstract

International audience; City systems, consisting of the various social and technical infrastructures required for urban areas to function, are facing major challenges in the coming decades as global urbanization and the pressure on resource efficiency and sustainability continue. Agent-based simulation and modelling techniques enable us to study these systems from various levels of abstraction looking at actors as well as system level perspectives, and are particularly suited to understand the dynamics of underlying structures and experiment with various scenarios and the effect of local decisions on emerging system properties. This special issue aims to showcase innovative use of agent-based modelling (ABM) and multi-agent system (MAS) approaches for effective decision support regarding urban innovations in fields linked with urban development and to explore and identify how researchers in different disciplines integrate their knowledge through agent models. We are inviting researchers from across a wide range of disciplines (including computer science, physics, engineering, economics, sociology and others) to submit papers to this special issue that exhibit state-of the art innovation in ABM/MAS applied to cities. This innovation may come in the form of computational, conceptual, analytical, participatory or epistemological advances for a range of aims and motivations associated with the challenge of combining competitiveness and sustainable urban development in city systems. Possible topics include, but are not limited to, the following:[...]

Abstract

International audience; The production of PhotoVoltaic (PV) energy depends on the solar irradiance level. The PV power plant fluctuations may have a significant impact on the frequency regulation in sufficiently small power systems, such as islands. The objective of this paper is to present a method using cooperative multi-agent systems to reduce the frequency fluctuations due to the unpredicted fluctuations of the PV production using electric vehicles as electricity storage units in an isolated power system.

Abstract

International audience; The modes of reasoning which are used in the context of safety analysis and the very nature of knowledge about safety mean that a conventional computing solution is unsuitable and the utilization of artificial intelligence techniques would seem to be more appropriate. Our research has involved three specific aspects of artificial intelligence: knowledge acquisition, machine learning and knowledge based systems (KBS). Development of the knowledge base in a KBS requires the use of knowledge acquisition techniques in order to collect, structure and formalizes knowledge. It has not been possible with knowledge acquisition to extract effectively some types of expert knowledge. Therefore, the use of knowledge acquisition in combination with machine learning appears to be a very promising solution. This paper presents the result of these two research activities which are involved in the methodology of safety analysis of guided rail transport systems.

Abstract

International audience; To sustain the rapidly increasing air traffic demand, the future air traffic management system will rely on a concept, called Trajectory-Based Operations (TBO), that will require aircraft to follow an assigned 4D trajectory (time-constrained trajectory) with high precision. TBO involves separating aircraft via strategic (long-term) trajectory deconfliction rather than the currently-practicing tactical (short-term) conflict resolution. In this context, this chapter presents a strategic trajectory planning approach aiming at minimizing the number of conflicts between aircraft trajectories for a given day. The proposed methodology allocates an alternative departure time, a horizontal flight path, and a flight level to each aircraft at a nation-wide scale.In real-life situations, aircraft may arrive at a given position with some uncertainties on its curvilinear abscissa due to external events. To ensure robustness of the strategic trajectory plan, the aircraft arrival time to any given position will be represented here by a probabilistic distribution over its nominal assigned arrival time.The proposed approach optimizes the 4D trajectory of each aircraft so as to minimize the probability of potential conflicts between trajectories. A hybrid-metaheuristic optimization algorithm has been developed to solve this large-scale mixed-variable optimization problem. The algorithm is implemented and tested with real air traffic data taking into account uncertainty over the French airspace for which a conflict-free and robust 4D trajectory plan is produced.

Abstract

Technological advancements made it possible for Electric vehicles (EVs) to have onboard computation, communication, storage, and sensing capabilities. Nevertheless, most of the time these EVs spend their time in parking lots, which makes onboard devices cruelly underutilized. Thus, a better management and pooling these underutilized resources together would be strongly recommended. The new aggregated resources would be useful for traffic safety applications, comfort related applications or can be used as a distributed data center. Moreover, parked vehicles might also be used as a service delivery platform to serve users. Therefore, the use of aggregated abundant resources for the deployment of different local mobile applications leads to the development of a new architecture called Vehicular Fog Computing (VFC). Through VFC, abundant resources of vehicles in the parking area, on the mall or in the airport, can act as fog nodes. In another context, mobile applications have become more popular, complex and resource intensive. Some sophisticated embedded applications require intensive computation capabilities and high-energy consumption that transcend the limited capabilities of mobile devices. Throughout this work, we tackle the problem of achieving an effective deployment of a VFC system by aggregating unused resources of parked EVs, which would be eventually used as fog nodes to serve nearby mobile users’ computation demands. At first, we present a state of the art on EVs and resource allocation in VFC. In addition, we assess the potential of aggregated resources in EVs for serving local mobile users’ applications demands by considering the battery State of Health (SOH) and State of Charge (SOC). Here, the objective is to choose EVs with a good condition of SOH and SOC so that owners secure tolerable amount of energy for mobility. Then, we address the problem of resource allocation scheme with a new solution based on Markov Decision Process (MDP) that aims to optimize the use of EVs energy for both computing users’ demands and mobility. Hence, the novelty of this contribution is to take into consideration the amount of aggregated EVs resource for serving users’ demands. Finally, we propose a stochastic theoretical game approach to show the dynamics of both mobile users’ computation demands and the availability of EVs resources.; Résumé : Les progrès technologiques ont permis aux véhicules électriques d’avoir des capacités à la fois de calcul, de communication, de stockage et de perception. Néanmoins, la plupart du temps, ces véhicules électriques sont en stationnement, ce qui engendre une sous utilisation de leurs capacités embarquées. Ainsi, une meilleure gestion et une mise en commun de ces ressources sous-utilisées deviennent fortement recommandées. Les ressources agrégées seraient utiles pour des applications de sécurité routière, des applications liées au confort ou pourraient même être utilisées en tant que centre de calcul distribué. En outre, les véhicules en stationnement pourraient également être utilisés comme plate-forme de fourniture de services. Par conséquent, l’utilisation de ressources abondantes agrégées pour le déploiement de différentes applications mobiles locales a conduit au développement du concept d’informatique en brouillard véhiculaire (an anglais, Vehicular Fog Computing - VFC). Grâce à ce dernier, les véhicules dans les aires de stationnement, les centres commerciaux ou les aéroports vont agir en tant que nœuds fog. Dans un autre contexte, les applications mobiles sont devenues de plus en plus populaires, complexes et gourmandes en ressources. Certaines applications mobiles nécessitent des capacités de calcul intensives et une consommation d'énergie élevée qui transcendent les capacités limitées des appareils mobiles. Tout au long de ce travail, nous abordons les verrous liés au déploiement efficace d’un système VFC agrégeant les ressources inutilisées des véhicules électriques en stationnement pour être utilisées comme nœuds fogs répondants aux demandes de calcul des utilisateurs mobiles à proximité. Notre travail commence par un état de l'art sur les véhicules électriques et l'allocation de ressources dans le système VFC. En outre, nous évaluons le potentiel des ressources agrégées dans les véhicules électriques pour répondre aux demandes d'applications d'utilisateurs mobiles locaux en prenant en compte l'état de santé de la batterie (en anglais, State of Health - SOH) et son état de charge (en anglais, State of Charge - SOC). Notre objectif est de choisir des VEs ayant un état de santé et de charge satisfaisants pour faire partie du VFC tout en permettant aux propriétaires de ces véhicules de disposer d’une quantité d’énergie suffisante pour leur mobilité. Nous abordons, par la suite, le problème d’allocation de ressources avec une nouvelle solution basée sur le processus de décision Markovien (en anglais, Markov Decision Process - MDP) qui vise à optimiser l’utilisation de l’énergie des véhicules électriques pour répondre à la fois à aux demandes de calcul et de mobilité des utilisateurs. Enfin, nous proposons une approche basée sur un jeu stochastique pour montrer la dynamique de la demande de calcul des utilisateurs mobiles et la disponibilité des ressources des véhicules électriques.

Abstract

Connected objects of everyday living have made their way into our lives. Known as Internet of Things, the various technologies inspire a vast variety of applications. One of the pioneer applications is the concept and development of a smart home. This is now spreading outdoors; making vehicles, buildings, and even large cities smart. Moreover, the technology is getting more personal as well – as wearing smart clothes and other self-tracking devices become increasingly common and popular. This is often referred to as the quantified self.One particular case of a smart environment is ambient assisted living, which is designed to enhance elderly people’s day-to-day life. Such a ubiquitous and unobtrusive computer system can also be ported to other domains and age groups. For instance, the tracking of daily activities can also help younger adults to improve their lifestyle. Everyone can be encouraged to maintain a healthy lifestyle, perform sufficient physical activity, and make more informed decisions about their mobility. These are direct factors in preventing health risks, such as metabolic diseases like the type 2 diabetes, and allow a better control over respiratory diseases like the asthma.Driven by these ideas, this thesis explores the possibilities of a web-based platform with a semantic rule-based reasoning. The thesis details the work on technical improvements, enhancements in activity recognition, extensions for data analysis, and a mobility-oriented application.Following a user-centric approach, a real life deployment of the described technologies is necessary. Two use cases are examined. First, I enhanced and built upon a pre-existing system, which consists of sensors and a gateway placed into elderly participants' homes. The second use case is the deployment of a mobile phone application for active mobility assistance. Collecting relevant and timely data, the application then outputs a level of recommendation for every type of mobility. The recommendations are based on each user’s exercise tracking device, which incorporates their goals, their profiles, and other publicly available data sources such as weather and air quality.This thesis describes the outcomes and lessons learnt from these deployments. In addition, this thesis provides an in-depth discussion as well as analytical insights on the results of the deployments.; Les objets connectés de la vie courante ont trouvé leur chemin dans notre quotidien. Connu sous le terme d'Internet des choses, une pluralité de technologies inspire une vaste variété des cas d'utilisation. Une des applications innovatrices est le concept et le développement de la maison intelligente. Actuellement, ce concept est en train de s'étendre vers l'extérieur ; il rend futés des véhicules, bâtiments et même des villes entières. De plus, la technologie devient aussi plus personnelle - comme le port des vêtements futés et d'autres appareils pour l'autosurveillance personnelle devient de plus en plus courant et populaire. Ce phénomène est souvent appelée le soi quantifié.Un cas particulier de l'environnement futé est l'assistance à l'autonomie à domicile (ambient assisted living) conçue pour améliorer la qualité de la vie quotidienne des personnes âgées. Un tel système informatique, omniprésent et discret, peut être porté dans d'autres domaines et tranches d'âge. Par exemple, le suivi des activités du quotidien peut aider jeune adultes à améliorer leur mode de vie. Tout le monde peut être encouragé à maintenir un mode de vie sain, à exercer une activité physique suffisante et prendre les décisions fondées quant à la mobilité. Ce sont les facteurs direct contribuant à la prévention des risques de santé, tel que les maladies métaboliques comme le diabète de type 2 ; et permettent un meilleur maîtrise des maladies respiratoires comme l'asthme.Dirigée par ces idées, cette thèse explore les possibilités d'une plate-forme web avec un raisonnement sémantique basé sur les règles. La thèse détaille le travail sur les améliorations techniques, avancements dans la reconnaissance d'activités, les extensions pour l'analyse des données et une application orientée mobilité.Suivant l'approche centrée utilisateur, un déploiement dans les conditions réelles est necessary. Deux cas d'utilisation sont abordés. Le premier est l'amélioration du système existant, consistant des capteurs et d'une passerelle tous placés dans le domicile de la personne âgée. Le deuxième cas est le déploiement d'une application pour téléphones portables pour l'assistance à la mobilité active. En collectant les donnés relevantes et actuelles, l'application affiche un niveau de recommandation personnalisée pour chaque type de mobilité. Ces recommandations sont basées sur l'appareil de soi quantifié qui incorpore les objectifs personnels, le profil et autres données librement accessible, tel que la météo ou la qualité d'air.Cette thèse décrit les sorties et les leçons tirées des déploiements de ces technologies. Cette thèse apporte une discussion et des analyses des résultats de ces déploiements.

Abstract

Day by day the number of vehicles on roads is growing, increasing also the number of accidents, traffic jams and carbon dioxide emissions. For this reason, the European Union has adopted an action plan for improving urban mobility where the public transport is the main key. To present day, solutions as special bus lanes, electrical and hydrogen buses and rail-guided buses have been tested and implemented in several cities as London and Madrid. In this article an automated stopping system for electrical buses is presented. Stopping system have been implemented by the AUTOPIA program on an electric minibus and tested on a private circuit with satisfactory results. Peer Reviewed

Abstract

International audience; A prerequisite to the design of future Advanced Driver Assistance Systems for cars is a sensing system providing all the information required for high-level driving assistance tasks. Carsense is a European project whose purpose is to develop such a new sensing system. It will combine different sensors (laser, radar and video) and will rely on the fusion of the information coming from these sensors in order to achieve better accuracy, robustness and an increase of the information content. This paper demonstrates the interest of using probabilistic reasoning techniques to address this challenging multi-sensor data fusion problem. The approach used is called Bayesian Programming. It is a general approach based on an implementation of the Bayesian theory. It was introduced first to design robot control programs but its scope of application is much broader and it can be used whenever one has to deal with problems involving uncertain or incomplete knowledge.

Abstract

acknowledged by the SESAR program, current ATC systems must be drastically improved to accommodate the predicted traffic growth in Europe. This study aims at assessing the performance of 4D-trajectory planning and strategic deconfliction, two of the key concepts identified to meet SESAR capacity objectives. Among the possible degrees of freedom, we focus here on the flight level (FL) optimization to avoid en-route conflicts between intersecting flights. The resulting problem can be reduced to Graph Coloring with a specific cost function minimizing the discrepancies to requested FLs. This FL allocation leads to very large combinatorial optimization problems when applied at the continental scale, especially when considering temporal uncertainties. The instances were solved with a Tabu Search algorithm in a few seconds to a few minutes, depending on size and conflict density. Our results shows that the global conflict resolution workload is alleviated by at least 20 %, while bounding the individual FL discrepancies to three levels for a small proportion of the traffic.

Abstract

International audience; As mobility grow in urban cities, traffic congestion become more frequent and troublesome. traffic signal is one way to decrease traffic congestion in urban areas but needs to be adjusted in order to take into account the stochasticity of traffic. Reinforcement learning (RL) has been the object of investigation of many recent papers as a promising approach to control such a stochastic environment. The goal of this paper is to analyze the feasibility of RL, particularly the use of Q-learning algorithm for adaptive traffic signal control in different traffic dynamics. A RL control was developed for an isolated multi-phase intersection using a microscopic traffic simulator known as Paramics. The novelty of this work consists of its methodology which uses a new generalized state space with different known reward definitions. The results of this study demonstrate the advantage of using RL over fixed signal plan, and yet exhibit different outcomes depending on the reward definitions and different traffic dynamics being considered.