Abstract

This paper proposes a study of existing environments used to enact data science pipelines applied to graphs. Data science pipelines are a new form of queries combining classic graph operations with artificial intelligence graph analytics operations. A pipeline defines a data flow consisting of tasks for querying, exploring and analysing graphs. Different environments and systems can be used for enacting pipelines. They range from graph NoSQL stores, programming languages extended with libraries providing graph processing and analytics functions, to full machine learning and artificial intelligence studios. The paper describes these environments and the design principles that they promote for enacting data science pipelines intended to query, process and explore data collections and particularly graphs.

Abstract

dvanced Driver Assistance Systems (ADAS) are embedded systems in the automotive domain which play an important role for safety enhancement. They are subject to several complex hardware and software constraints, and executed in a highly variable environment. Thus, the design of these systems is a very important task but becomes increasingly difficult, especially with the lack of generic reusable templates for ADAS. So, in this paper, we propose three design patterns covering design problems relative to the following key tasks: sensing, processing and control of sensor data, and taking actions (warning and actuation). These patterns are specified using the Unified Modeling Language (UML) to represent structural and behavioral information. This paper also describes two examples of ADAS that illustrate how we have applied these patterns.

Abstract

TRA 2018, 7th Transport Research Arena, Vienne, AUTRICHE, 16-/04/2018 - 19/04/2018; The decrease of road accidents these recent years has induced the interest to work on road driving hazardous situations called incidents. Otherwise, in France, users of Power Two-Wheeler (PTW) are less than 2% of traffic but, in the accidents represent 43% of serious injuries and 23% of killed persons. This paper provides a descriptive analysis of the incident database acquired during one year with a fleet of 30 motorcycles and a few cars equipped with an Event Data Recorder. The objective is to acquire knowledge on their accident mechanisms and improve their safety. For the 395 motorbikes' and 32 cars' incidents collected, the analysis of the triggering criteria, the dynamics parameters and the road configurations illustrates the great variety of situations and shows how the behaviours of riders are different from those of car drivers. The investigations on the causes of the incidents indicate that the more frequent cause is the driver (94%), but the road environment and the other road users could have been involved in 40% of the cases each.; La baisse du nombre d'accidents de la route de ces dernières années a amené les chercheurs à s'intéresser aux situations de conduite critiques que l'on appelle des incidents. Par ailleurs, en France, les usagers de deux roues motorisés représentent moins de 2% du trafic, et pourtant 43% des blessés graves et 23% des décès sur la route. Ce papier propose une analyse descriptive de la base de données des incidents recueillis pendant un an sur une flotte de 30 motos et quelques voitures équipées d'enregistreur de données de la route. L'objectif était d'acquérir des connaissances sur leurs mécanismes d'accident et d'améliorer leur sécurité. Pour les 395 incidents de 2-roues et les 32 incidents de voitures, l'analyse des critères de déclenchement, des paramètres de la dynamique du véhicule et des configurations de la route illustre la grande variété des situations et montre les différences de comportement des motards par rapport aux automobilistes. Les investigations sur les causes des incidents indiquent que la cause la plus fréquente est liée au conducteur, mais l'environnement et les autres usagers pourraient être impliqués dans 40% des cas.

Abstract

International audience; The increase in air traffic and the limited capacity of air traffic control services force us to think of a new way in which to control aircraft. An innovative ATM concept, called Sector-Less Air Traffic Management, has been defined by the Eurocontrol Experimental Centre. In this concept the role of the controller is radically different from the current one : instead of having controllers controlling a sector, controllers would be responsible for a certain number of aircraft from departure to arrival in terminal areas. The aim of this paper is to redesign the European en-route network in order to make this new way of controlling aircraft possible and to provide an aeroplane with the shortest route possible by providing a new route network. The approach proposed here is, starting from scratch, to provide a very simple route network and to improve it by using optimisation algorithm. A generated route network will be evaluated in terms of length of trajectories weighted by the amount of aeroplanes using them and then compare it with the direct route network and the current one. Using this tool and a fast time air traffic simulator, the Sector-Less concept will be able to be evaluated in terms of capacity and delay due to conflict. This also provides the route density of the network for this concept. The results of various simulations performed on the best route network generated so far will be presented.

Abstract

In an effort to curb air pollution, the city of Delhi (India), known to be one of the most populated, polluted, and congested cities in the world has run a trial experiment in two phases of 15 days intervals. During the experiment, most of four-wheeled vehicles were constrained to move on alternate days based on whether their plate numbers ended with odd or even digits. While the local government of Delhi represented by A. Kejriwal (leader of AAP party) advocated for the benefits of the experiment, the prime minister of India, N. Modi (former leader of BJP) defended the inefficiency of the initiative. This later has led to a strong polarization of public opinion towards OddEven experiment. This real-world urban experiment provided the scientific community with a unique opportunity to study the impact of political leaning on humans perception at a large-scale. We collect data about pollution and traffic congestion to measure the real effectiveness of the experiment. We use Twitter to capture the public discourse about the experiment in order to study people’s opinion within different dimensions: time, location, and topics. Our results reveal a strong influence of political affiliation on how people perceived the outcomes of the experiment. For instance, AAP supporters were significantly more enthusiastic about the success of OddEven compared to BJP supporters. However, taking into account location of people revealed that personal experience is able to overcome political bias.

Abstract

International audience; Traffic congestion causes driver frustration and costs billions of dollars annually in lost time and fuel consumption. This paper presents three traffic re-routing strategies designed to be incorporated in a cost-effective and easily deployable vehicular traffic guidance system that reduces the effect of traffic congestions. This system collects real-time traffic data from vehicles and road-side sensors and computes proactive, individually-tailored re-routing guidance which is pushed to vehicles when signs of congestion are observed on their route. Extensive simulation results over two urban road networks show that all three strategies, namely multipath load balancing considering future vehicle positions (EBkSP), random multipath load balancing (RkSP), and dynamic shortest path (DSP), significantly decrease the average travel time. EBkSP is the best, with as much as 104% improvement compared to the "no re-routing" baseline. Additionally, it lowers with 34% the re-routing frequency compared to the other strategies. Finally, all strategies offer good improvements even when many drivers ignore the guidance or when the system adoption rate is relatively low.

Abstract

International audience; In addition to identifying safety deficiencies evidenced by accidents, the independent investigative body makes recommendations to eliminate or reduce these deficiencies. Its main purpose is to advance transportation safety by conducting investigations of accidents in rail and other modes of transportation. He must answer several questions: what happened, why did it happen, and what can be done to reduce the risk of it happening again? In this process, one of the difficulties involved is finding abnormal accident scenarios which are capable of generating a specific hazard. This paper proposes an original method based on machine learning to assist investigators experts in their crucial task of analysis and assessment of the safety for railway transport systems in France. This contribution is based on the use of artificial intelligence techniques and has involved the development of several approaches and tools which assist in the modeling, storage and assessment of knowledge about safety. The proposed approach has two objectives, firstly to record and store experience concerning safety analyses, and secondly to assist those involved in the development and assessment of the systems in the demanding task of evaluating safety studies.

Abstract

International audience; The air transport sector entered the European Trading Scheme in 2012 (for the intra-European flights). The regulation of CO2 emissions is costly for airlines and modifies the organization of their market. Our paper proposes an economic analysis in which the regulation but also CO2 emissions of airlines are modeled. We show that, in a perfect competition setting, the difference between passengers carried without regulation and when the regulation is considered is negative for the best-performing planes. However, for the less efficient aircraft, the implementation of the regulation entails a reduction of airlines activity, and therefore a low level of carbon emissions.

Abstract

International audience; The main goal of Data Mining is the research of relevant information from a huge volume of data. It is generally achieved either by automatic algorithms or by the visual exploration of data. Thanks to algorithms, an exhaustive set of patterns matching specific measures can be found. But the volume of extracted information can be greater than the volume of initial data. Visual Data Mining allows the specialist to focus on a specific area of data that may describe interesting patterns. However, it is often limited by the difficulty to deal with a great number of multi dimensional data. In this paper, we propose to mix an automatic and a manual method, by driving the automatic extraction using a data scatter plot visualization. This visualization affects the number of rules found and their construction. We illustrate our method on two databases. The first describes one month French air traffic and the second stems from 2012 KDD Cup database.