Back to school… time to learn: VEDECOM launches MobTips, the first digital mapping of new mobility in France

With the arrival of new services and uses of electric, hybrid, autonomous and shared mobility, the transport and automotive ecosystem is undergoing unprecedented changes, both social, economic and technological. A “Mobility” sector is emerging, more complex and multifaceted. There is a growing need to share and understand a common language between all mobility actors, from designer to user. To clarify the issues related to new mobility, VEDECOM, a French Institute for Energy Transition (ITE), is launching Mobtips, the first reference system dedicated to “the words of mobility”. Mobtips’s mission is to share a French common vocabulary and make the changes and new mobility concepts understandable to a large audience in a fun way.To visit VEDECOM’s Mobtips site: https://mobtips.fr This initiative received financial support from the SGPI (French State investments for future programs) and the Automotive Industry Platform (PFA).

Read the press release

From pedestrian to autonomous vehicle… A public survey is on line!

L’évolution du véhicule autonome passera nécessairement par la communication, l’échange d’informations et l’amélioration de ses capacités de perception à l’environnement. Il parait ainsi essentiel que les véhicules autonomes et les piétons cohabitent et interagissent entre eux en toute sécurité.

Cependant, aujourd’hui de nombreuses questions subsistent notamment une : quel sera le comportement d’un piéton face au véhicule autonome ?

C’est dans ce cadre que les chercheurs de l’institut VEDECOM lancent un questionnaire en ligne portant sur le véhicule autonome et la communication avec les autres usagers de la route.

Ce questionnaire dure une trentaine de minutes environ, nous comptons sur vous !

Toutes vos réponses sont essentielles pour poursuivre notre travail de recherche.


Pour répondre à notre questionnaire :

Participer à l’enquête

Merci à tous pour votre participation et la diffusion de cette enquête autour de vous !

The best delivery robots selected by VEDECOM’S new Mobility Camp site!

In these times of lockdown, VEDECOM’s launch of its new Mobility Camp site is well timed, as it offers information about the best teaching resources for new mobility methods. The goal: to imagine a post-coronavirus world which places a strong emphasis on innovative, sustainable mobility methods. And the icing on the cake: if you’re a player in this field, you can contribute your own content to it!

This week, we focus on Europe’s Top 5 delivery robots. TwinswHeel, Starship, Kar Go, AMR and Eliport: in the current health crisis, these delivery solutions could provide a solution to the difficult balancing act of maintaining economic activity while also protecting delivery company employees. These are technologies which, in more “normal” times, will also help to address issues such as pollution reduction, shared use of space and efficiency gains.

You can find our selection of delivery robots at https://mobilitycamp.fr/experimentations/

To contribute, click the Contribute button on the left side of the page.

Final conference for the European CoExist project held in video conference format

The final conference for the European CoExist project was held on 25 and 26 March 2020. Originally intended to take place in Milton Keynes, one of the project’s partner towns, the Covid-19 crisis ultimately forced it to be conducted in video conference form instead. The purpose of the CoExist project, launched in 2017 with a budget of 3.5 million euros, is to prepare for the transition phase during which roads will be shared by self-driving and conventional vehicles. The goal: to assist road authorities and local governments to plan for a road network incorporating various levels of automation, traditional vehicles, and other road users. By simulating the incorporation of self-driving vehicles into traffic flows in 4 European towns and cities, the project analysed the consequences of the presence of these new vehicles on urban road infrastructure. We take a close look at a project in which Vedecom’s expertise with self-driving vehicles proved enormously beneficial.

Building a bridge between self-driving vehicles and infrastructure planning

If the introduction of self-driving vehicles is to live up to its promises in terms of reducing road space and improving traffic efficiency and safety, vehicle design and urban infrastructure planning are key issues to be considered. This is the challenge addressed by the European CoExist project, bringing together 17 partners from 7 European countries (Germany, Belgium, France, Italy, Netherlands, the UK and Sweden), representing industry, academic institutions and local government. Their task was to study the feasibility and consequences of introducing automated smart vehicles in terms of impacts on traffic flow, the environment (greenhouse gases and CO2) and noise pollution in an urban environment, in a few highly specific use cases.

Three key stages in modelling traffic for 4 European cities

The project followed three key stages in the development of transport and infrastructure. The first step was to validate extensions of existing transport models at micro and macro level, including different types of self-driving vehicles with varying levels of automation. The second step was to develop an assessment tool that would simulate the impacts of these autonomous vehicles on safety, traffic flow and infrastructure changes – using a variety of different ramp-up scenarios for self-driving vehicles in such environments. Lastly, these tools were applied to eight use cases in four European towns and cities: Gothenburg (SWE), Stuttgart (DE), Helmond (NL) and Milton Keynes (GB). The end goal was to produce directives for designing hybrid infrastructure that is equally capable of supporting conventional and automated vehicles, thus providing the smoothest possible transition phase.

 

VEDECOM : the CoExist project’s self-driving vehicle expert

VEDECOM’s partnership and expertise was invaluable in implementing the project. The Institute agreed to the use of control models for autonomous vehicles that were exactly the same as for VEDECOM’s self-driving vehicle prototypes, and their software adaptations. These were incorporated into the VISSIM application produced by PTV Group, which was then able to simulate fleets of self-driving vehicles and shuttles on the roads of partner towns, which were themselves modelled using the solution.

VEDECOM was also involved in research into the interpretation and analysis of data regarding the economic and environmental impacts of the gradual introduction of such autonomous vehicles into urban areas. Lastly, it provided analysis into the system’s feasibility and the infrastructure modifications required.

Find out more at: https://www.h2020-coexist.eu/what-is-coexist/

“The role of acceptability in the interaction between a conventional and an automated vehicle”: thesis defence by Géraldine Van Der Beken

VEDECOM is pleased to announce that Géraldine Van Der Beken has successfully defended her thesis entitled “The role of acceptability in the interaction between a conventional and an automated vehicle”.

The viva took place on Friday 14 February 2020, at the University of Rennes 2.

Supervisors were Alain SOMAT, Sami KRAIEM and Pascal PANSU.

ABSTRACT

The central theme of this thesis deals with the role of acceptability in the interaction between a conventional vehicle driven by a human and an automated vehicle. The initial study consists of a meta-analysis summarising the determinants of acceptability of a new technology. The results showed that acceptability was predicated on six factors: behavioural intent, expected performance, expected ease of use, attitude, social influence and sense of control. A second study was carried out to assess the effect of the decision of acceptability on the difference in behaviour exhibited by the driver of a conventional vehicle when interacting with an automated vehicle. The results showed that a low level of acceptability is associated with cautious behaviour towards the automated vehicle. A third study, using a driving simulator, showed that drivers of conventional vehicles with a high level of acceptability behave identically towards an automated vehicle and a conventional vehicle. In conclusion, this thesis discusses how important acceptability of a technological device is when interacting with it.

Articles récents :

No Results Found

The page you requested could not be found. Try refining your search, or use the navigation above to locate the post.

Weakly supervised learning in the context of autonomous vehicle perception : Florent Chiaroni defended his thesis

 VEDECOM is pleased to announce that Florent Chiaroni successfully defended his thesis entitled “Weakly supervised learning in the context of autonomous vehicle perception”.

The thesis defence took place on Monday, the 3rd of February 2020, at Centrale Supelec. 

JURY

Yap-Peng TAN, Professor, Nanyang Technological University (NTU), Singapore – Rapporteur

Hichem SAHBI, Research Fellow CNRS (HDR), UPMC Sorbonne Université (LIP6) – Rapporteur

Jean-Luc DUGELAY, Professor, EURECOM – Examinateur

Samia BOUCHAFA, Professor, Université d’Evry Val-d’Essonne (IBISC) – Examiner

Camille COUPRIE, Researcher, FACEBOOK (FAIR) – Examiner

Frédéric DUFAUX, Research Director CNRS, CentraleSupelec (L2S) – Thesis Supervisor

Mohamed-Cherif RAHAL, Researcher, Institut VEDECOM – Thesis co-supervisor (référent VEDECOM)

Nicolas HUEBER, Researcher, Institut Saint-Louis Franco-Allemand ISL (ELSI) – Thesis co-supervisor

Féthi BEN OUEZDOU, Professor, Université de Versailles Saint-Quentin en Yvelines (LISV) – Guest

ABSTRACT

 In the context of autonomous vehicle perception applications, the interest of the research community for deep learning approaches has continuously grown since the last decade. This can be explained by the fact that deep learning techniques provide nowadays while requiring only low-cost vision sensors. More specifically, deep learning techniques can provide, from a monocular camera sensor, rich semantic information concerning the complex visual patterns encountered in autonomous driving scenarios. However, such approaches require, as their name implies, to learn on data. In particular, state-of-the-art prediction performances on discriminative tasks often demand hand labeled data of the target application domain. Hand labeling has a significant cost, while, conversely, unlabeled data can be easily obtained in the autonomous driving context. It turns out that a category of learning strategies, referred to as weakly supervised learning, enables to exploit partially labeled data. Therefore, we aim in this thesis at reducing as much as possible the hand labeling requirement by proposing weakly supervised learning techniques.

We start by presenting a type of learning methods which are self-supervised. They consist in substituting hand-labels by upstream techniques able to automatically generate exploitable training labels. Self-supervised learning (SSL) techniques have proven their usefulness in the past for off road obstacles avoidance and path planning through changing environments, by learning at the application time. More recently, they have also been applied for depth map estimation, asphalt road segmentation, and moving obstacles instance segmentation and tracking. However, SSL techniques still leave the door open for detection, segmentation, and classification of static potentially moving obstacles. For instance, the latter can be motionless cars at a road intersection, or pedestrians waiting to cross the street. Consequently, we propose in this thesis three novel weakly supervised learning methods using generative adversarial networks, with the final goal to deal with such road users through an SSL framework.

The first two proposed contributions of this work aim at datasets, such that the labeling effort can be focused only on our class of interest, the positive class. Then, we propose an approach which can deal with training data containing a high fraction of wrong labels, referred to as noisy labels. Next, we propose to demonstrate the potential of such weakly supervised image classification strategies for the two following real application tasks: detection and segmentation of potentially moving obstacles. 

Finally, we draw a conclusion on this thesis research work, followed by future research perspectives in order to motivate further investigations towards the proposed directions.

 

Articles récents :

No Results Found

The page you requested could not be found. Try refining your search, or use the navigation above to locate the post.