The team is focusing on challenges linked to transport services, systems and infrastructures for safer mobility. It is tackling issues to do with safety, operations, sustainable mobility and energy linked to technological developments: the road of the future, the digital revolution, driving aids, autonomous vehicles, etc.
Sites: Clermont-Ferrand, Toulouse ; 18 people involved, including 6 permanent researchers.
The aim of the team is to respond to the challenges of public policies implemented in the fields of mobility, digital, transport, infrastructure, road safety and regional cohesion. Mobility of goods and people represents a crucial aspect in the lives of people and our society which is confronted by challenges in terms of mobility and technological innovations.
To prepare for the mobility of tomorrow, we must:
- anticipate issues linked to development of ITS, digital mobility and vehicle automation;
- contribute to enhancing road safety by grasping the opportunities afforded by megadata to prevent accidents and anticipate the challenges of tomorrow: •being at the heart of innovation and supporting adoption of technological innovations;
- contribute to development and cohesion of sustainable and resilient regions.
The activities of the team are focused on Intelligent Transport Systems backed by transport infrastructure, and more broadly travel management systems on the regional scale. The team will adopt a systemic approach to mobility of goods and people. It will take account of the upheavals expected due to the introduction of autonomous vehicles and vehicles driven by automatic means as part of transport systems. It will explore in great depth the question of the impact these new technologies will have on human behaviour, and will evaluate their performance in order to support their gradual deployment.
Its research activities are organised around two scientific areas:
- Area 1 – towards safer efficient infrastructures for the mobility of the future;
- Area 2 – towards automated driving.
The research relating to area 1 aim to help improve intelligent transport systems by adopting an approach on three levels:
- the first level focuses on observation of infrastructure via the associated network of sensors, with the aim of proposing innovative solutions to deduce its use, its condition and its safety;
- the second level takes advantage of the knowledge acquired about the infrastructure to propose decision support solutions in order to better organise, plan and prepare for its use;
- the third level adopts a more forward-looking approach, seeking to imagine the infrastructure of tomorrow, whether physical or virtual.
The research carried out in area 2 will examine the upheavals caused by introduction of vehicles driven by automatic means and autonomous vehicles within transport systems. To tackle this issue, the approach chosen in this area of work will be based on the levels of driving automation defined by the SAE (Society of Automotive Engineers), and will be divided into three sub-areas of research:
- the first sub-area will seek to improve knowledge about human behaviour in the absence of any automation;
- the second will explore the new interactions brought about by driving automation and its connections;
- the third will conduct an assessment of autonomous vehicles, regarding the technical aspects, and in the context of projects to deploy autonomous vehicles.
The complementarity of these two areas of research, which each have a human and social science component, will enable the team to fully explore a wide variety of subjects linked to ITS. The main challenges are:
- collection and appraisal of the data generated by digital mobility;
- production of information and decision support algorithms to organise, plan and prepare to use an intelligent transport system;
- multi-dimensional design and assessment of innovative technological solutions for tomorrow’s mobility and its infrastructure;
- consideration of human factors in order to become familiar with and predict the behaviour of users;
- consideration of degraded operational conditions in order to improve robustness of transport systems
- the impact of autonomous vehicles on mobility in the future.
The main identified scientific obstacles:
- development of techniques to process the «big data» generated by digital mobility;
- modelling of indicators associated with the performance of an intelligent transport system;
- development of deep learning methods to identify specific situations or events in large amounts of data;
- design of robust detection and analysis methods in degraded situations;
- development of tools for experimentation and modelling of new services provided by the transport infrastructure, in terms of both energy and digital technology;
- changes to statistical tools to measure the individual road risk;
- behaviour of drivers, and more generally of users, when using autonomous vehicles;
- design of methodologies to assess an autonomous vehicle;
- change management for vulnerable people, operators and users of the transport system.
The team hopes to achieve different types of results: they may be tools, methodologies, algorithms, new knowledge. For instance:
- Algorithms to monitor and detect human behaviour in very confined environments; • Algorithms to detect potentially hazardous situations at level crossings;
- Algorithms enabling meteorological measurement by cameras and assessment of roadside devices in degraded weather conditions;
- Multi-spectrum modelling of light transmission in fog, and experimental validation in ranges of ADAS sensors;
- Definition of use cases and performance indicators to develop a method to assess the supply chain of the Bordeaux Living Lab;
- Methodology for multi-criteria assessment of connected embedded systems in real driving situations and to reduce emissions on a route;
- Tools for constructions/validation of digital mapping of transport infrastructure based on data from tracking vehicles;
- Change management methodology;
- Tools to dimension new energy recovery infrastructures according to local needs;
- Methodology to measure the reflective characteristics of road surfaces;
- Recommendations for managers in order to modify driver behaviour through their perception of the cross-section;
- Results of driving simulator tests, improvements aiming to ensure the driver naturally adopts appropriate behaviour for calmer driving;
- Knowledge of the amount of information required to perform a driving task, in complex situations (noise, overall visual configuration or presence of distractions) or degraded situations including loss of visual information (night, fog, rain);
- Method to evaluate overall and individual risk;
- Methodology to take account of support at an individual or group level (road management departments and ministries);
- Statistical analysis of accident and incident data according to potentially critical use cases for future autonomous vehicles interacting with other road users;
- Assessment methodology and test protocols to objectively assess isolated sensors and sensors built into the vehicle:
- Methodology to assess connected systems, definition of indicators;
- On-site assessment methodology for autonomous shuttles.
- Institutional partners, Public bodies: IFSTTAR, IGN, IRSTEA, Météo France, Vedecom, VTT;
- French and international academic partners: Université de Clermont-Ferrand (Institut Pascal, LAPSCO, LABEX ImobS3), Rennes 2, Paris-Est, Toulouse, Belfort Montbéliard, Carlos III de Madrid (Spain), Institut National d'Optique (INO) (Canada), Newcastle (UK), - Université de Sousse (Tunisia), CERTH research institute, Greece, TNO Netherlands, LAB (Accident, biomechanics and human behaviour research laboratory) PSA/Renault, Ircam Paris...
- Clusters: ViaMéca, CARA, Movéo, Tenerrdis, Indura, ID4CAR, Aerospace Valley;
- Private sector partners: Renault, PSA, Orange, SNCF, Michelin, Continental, MAPTM, GeoLoc Systems, DYNNIQ, Daimler, Autoliv, Hitachi, IBEO , Innoluce, Moduligh, Oplatek, Vaisala, Xenics, ENSIS-Optis, Zehnter, etc.
Mehdi Chahir, St Brieuc (PhD student at Université de Rennes 2 (UR2), LP3C - “Elaboration d’un modèle d’accompagnement multidimensionnel du changement technologique à travers la prise en compte du facteur humain” (Creation of a model for multi-dimensional technological change management taking into account the human factor.) Supervised by Professor Alain Somat and Stéphanie Bordel (Cerema STI). Presentation planned in 2019
Minh Mai Nguyen, 2019-2022 - Toulouse, (PhD student at IRIT Toulouse), « Segmentation sémantique d’image par approche ontologique. Application à la perception de l’environnement d’un véhicule autonome pour la détection d’obstacles par météo tout temps” (Semantic image segmentation by ontological approach. Application to the perception of the environment of an autonomous vehicle for obstacle detection in all-weather conditions), Supervised by : Alain Crouzil, Louahdi Khoudour, Pierre Duthon (Cerema STI)
Mathieu Labussière, 2018-2021 –Clermont-Fd (PhD student at Université de Clermont Auvergne, LABEX ImobS3) « Contributions à la perception multi-sensorielle en milieu perturbé par apprentissage profond » (Contributions to multi-sensory perception in perturbed environments by deep learning ) . Supervised by : Omar Ait Aider, Frédéric Bernardin (Cerema STI), co-supervisor: Céline Teulière
Prince Sévi, 2020-2023 –Clermont-Fd (PhD student at Université de Savoie Mont Blanc, financement Région Dromotherm) « Caractérisation mécanique et énergétique de l’échangeur Dromotherm et valorisation en lien avec les usages du bâtiment » (Mechanical and energetic characterisation of the Dromotherm heat exchanger and valorisation in relation to building uses). Supervised by: Benoit Stutz (USM), Frédéric Bernardin (Cerema STI), Co-supervisor: Alexandre Cuer, Evelyne Toussaint>
Ali Krayem, 2021-2024 –Clermont-Fd (PhD student at Université de Clermont Auvergne, LABEX ImobS3) « Estimation par méthode inverse des propriétés optiques d’un milieu diffusant sous lumière polarisée - Application à l’impact du brouillard sur la perception artificielle » (Inverse estimation of the optical properties of a scattering medium under polarised light - Application to the impact of fog on artificial perception ). Direction : Frédéric Bernardin (Cerema STI), Arnaud Münch (UCA, Laboratoire de Mathématiques), Co-supervisor: Santiago Royo (Université de Catalogne)
Post-doctoral
Chafik Bakey, Toulouse, 2017-2019, Post doctoral: “Projet H2020 Safer-LC : traitement d’images appliqué à la surveillance des passages à niveau” (“H2020 Safer-LC project: image processing applied to monitoring of level crossings”) (Louahdi Khoudour Cerema STI).
Amine Ben-Daoued, Clermont-Fd, 2020-2021, Post-Doctoral : Evaluation probabiliste d’un système de détection en conditions météorologiques dégradées (Probabilistic evaluation of a detection system in degraded weather conditions), (Frédéric Bernardin, Pierre Duthon)
Khouloud Dahmane, Clermont-Fd, (PhD student at Université de Clermont Auvergne, LABEX ImobS3, Institut Pascal): “Analyse d'images par méthode de deep learning appliquée au contexte routier en conditions météorologiques dégradées” (“Image analysis by Deep Learning method applied to the road context in degraded meteorological conditions”) Supervisors: Frédéric Chausse, Frédéric Bernardin(Cerema STI), Co-supervisors: Christophe Blanc, Michèle Colomb (Cerema STI)
Yann Meneroux, (PhD student at Université de Paris-Est & IGN): “Utilisation des véhicules traceurs et des données géographique pour la construction d’une infrastructure routière numérique” (“Use of tracking vehicles and geographical data to construct a digital road infrastructure”). Co-supervisors: Guillaume Saint Pierre (Cerema STI) and Sébastien Mustière. Co-supervisors: Arnaud Le-Guilcher, Olivier Orfila
Jérémy Matias, (PhD student LABEX ImobS3, Université de Clermont-Auvergne): “Perception visuelle et anticipation motrice dans les systèmes de traitement de l’information biologiques et artificiels” (“Visual perception and motor anticipation in biological and artificial information processing systems”) Thesis presented on 16 July 2019 – Jury, Michel Dhome, Laeticia Sylvert, Jean-Charles Quinton, Marie Izaute, Maria Jesus Funes, François Maquestiaux, Michèle Colomb (Cerema STI)
Huy-Hieu Pham, (PhD student Université Paul Sabatier Toulouse III): “Architectures d’apprentissage profond pour la reconnaissance d’actions humaines dans des séquences vidéo RGB-D monoculaires.” (“Deep learning architectures for human action recognition in monocular RGB-D video sequences.”) Application to surveillance in public transport». Thesis presented on 19 September 2019. Co-supervisors: Louahdi Khoudour (Cerema STI) and Alain Crouzil
Boris Quétard, (Doctorant LABEX ImobS3, Université de Clermont-Fd) : “Perception visuelle et anticipation motrice dans les systèmes de traitement de l’information biologiques et artificiels” (“Visual perception and motor anticipation in biological and artificial information processing systems”) Thesis presented on 10 April 2018 – Jury, Michel Dhome, Marial Mermillot, Jean-Charles Quinton, Marie Izaute, Giovanni Pezzullo, Michèle Colomb (Cerema STI)
You have a question ?