Projects

Printer-friendly versionPDF version

To give you a flavour of the type of work I do here is a collection of some current and previous projects.

HIANIC - Human Inspired Autonomous Navigation In Crowds. Financed by ANR. (2017-2020)

Autonomous vehicles typically function well on large roads where there are few pedestrians. However and increasing trend in urban design is to promote shared spaces where segregation between pedestrains and vehicles is minimised by removing features such as kerbs, road markings, traffic signs and traffic lights. Autonomous car navigation in such shared spaces raises grreat challenges. The HIANIC cybercar will analyze its environment by detecting people, evaluating crowd flows, recognizing typical scenarios. It will infer the reaction of the passengers in order to navigate in a way that makes them feel comfortable. Models of human and crowd behavior will be built in order to give the autonomous car knowledge of its social environment in the present and
the near future. These models will integrate knowledge from social sciences (proxemics, behaviorism) and will use agent based simulation.

SWIFT - Simulations with Intelligence for Fire Training. Financed by AGIR Pole MSTIC. (2016-2019)

The SWIFT project addresses the problem of modelling human behaviour during a bushfire in the state of Victoria, Australia. The model takes into account psychological behaviours, such as risk averseness and attachment to property, etc. and is based on field studies and interviews conducted after the tragic Black Saturday busfires on the 7th February 2009. An agent based approach is used following a BDI (Belief, Desire, Intention) architecture. The simulator is implemented using the GAMA simulation platform.

SNICS - Social Norms and their Impact in Crisis Situations. Financed by LIG Emergence scheme. (2016-2017)

The project is concerned with studying the impact of norms on the behaviour of populations during a crisis. Faced with a crisis, individuals react not only to institutional norms (e.g. evacuation procedures), but their own perception of the situation and their own beliefs, which may be in conflict with the norms (e.g. find your child rather than evacuating yourself to a safe place). The goal is to produce a develop a model and simulation of the normative behaviours in order to provide decision makers with a support tool. This will allow them to understand how people really behave in crisis situations and to develop more targeted procedures and communications. We will use a agent based approach and model mental attitudes that allows us to create different individual profiles taking into account the individual's perception of the situation, beliefs, personality, values, and norm adherence etc. The project is a collaboration between MAGMA-LIG and AMA-LIG.

MUDAMO - Integration of multidisciplinary data for crisis modelling. Financed by IXXI Complex Systems Institute, Rhône-Alpes. (2015-2017)

The strongly multidisciplinary MUDAMO project aims to promote interoperability between disciplines (human and social sciences, computer science) that are involved in simulation for crisis management. It will develop a methodology to integrate the results of questionnaires and social surveys into multi-agent models. It will also produce a survey protocol that is specifically aimed at multi-agent modelling. The MUDAMO project uses the specific example of modelling post-seismic pedestrian mobility. A novel tool will be developed to promote dialogue between researchers from diffrent disciplines. These methods and tools will be disseminated as part of a research workshop.  Partners include: PACTE Lab, MAGMA-LIG, Images Solidaires.

MaGIL - Modélisation générique et intégrrée de la gestion de la zone littorale. "Generic and integrated modelling coastal zone management". Financed by CNRS DEFI Littoral  (2016-2017)

Plus d'un milliard de personnes vivent à moins de 100 km de la côte, et il est estimé que 800 millions d’entre elles résident à moins de 10 m du niveau actuel de la mer. De ce tropisme pour le littoral, découle une vulnérabilité accrue (populations, ports, raffineries, infrastructures routières, milieux naturels...) qui, combinée à des aléas dont les intensités et les régimes pourraient évoluer avec les changements climatiques, vont renforcer le risque littoral. Face à les phénomènes d’érosion et de submersion marine, qui pourraient s’amplifier du fait du changement climatique et de la montée du niveau marin associé, il convient d’élaborer des mesures d’adaptation, entre autres au travers d’une gouvernance particulière, qui impliquent une évolution des doctrines relatives à la fixation du trait de côte, au profit de logiques axées sur la réduction de la vulnérabilité, l’amélioration de la qualité de vie et la soutenabilité du développement territorial et la reconquête d’écosystèmes.  S’il est aujourd’hui clair que la relocalisation, la recomposition des territoires littoraux passe par un recul stratégique (Séminaire MEDDE 2013), les politiques locales et étatiques pour l’accompagner restent globalement à définir tant l’adhésion à celles-ci est complexe et « la recherche [...] insuffisamment développée et l’intégration déficiente ». Le système littoral sera donc non seulement impacté sur un plan physique par des pertes de territoires mais ces zones fortement anthropisées vont aussi être soumises à des changements d’usages et de gouvernance (locale ou étatique).

Le projet vise à comprendre comment les usages actuels du littoral, relevant d’intérêts individuels différenciés, de principes de justice diversifiés et de doctrines passées (urbanisation, défenses côtières, rechargements de plage …) mises en oeuvre par des acteurs institutionnels variés, confrontés au changement global vont évoluer et susciter des rétroactions (résistance/adhésion) politico-socio-économiques favorisant une nouvelle gouvernance d’adaptation à long terme. Dans ce contexte, le projet MaGIL vise à aller vers l’intégration dans un outil de modélisation « système multi-agents » (SMA) du littoral, des dynamiques physiques incertaines et affectées par le changement global, des caractéristiques socio-économiques d’un territoire à risques et de leur gouvernance. Le projet MaGIL vise donc à modéliser au travers de scénarios génériques sur des territoires virtuels, les comportements à long terme de cette interface que constitue le littoral ce
qui, à notre connaissance, a peu été abordé aussi bien sur le plan national qu’international.

MIMICS - Multi-agent modelling of mobility in seismic crises. ADR Financed by ARC 7 Rhône-Alpes (2015-2018)

The MIMICS is interested in pedestrian mobility in earthquake crisis situations. After an earthquake, citizens usually try to reach safe areas, but in the dense urban environment they become highly vulnerable to falling debris as they try to reach safety. In order to measure the impact of urban policies, evacuation strategies and information campaignes (that can influence human behaviours), we are interested in developing a dynamic multi-agent model. The multi-agent model and simulator will form the basis of a serious game in order to transfer project results to society. Partners: PARN, Images Solidaires, PACTE and MAGMA-LIG.

VUSIM-AMS Urban vunerability to earthquakes of Mendoza, Argentina - from acquisition to modelling for descision-making (Vulnérabilité urbaine aux séismes de la ville de Mendoza, Argentine - de l'acquisition à la modélisation pour la décision. (2014-2015)

This PEPS funded project looks at acquiring data for modelling human behaviours immediately after an earthquake. In the urban context of Mendoza (Argentina), vulnerability to earthquakes may be underestimated due to there being only a slight probability of a major event. However, as history shows from the 1861 earthquake that destroyed the city, there is a real risk of a large earthquake.  To reduce the population's vulnerability, either by following building construction norms or by emergency management procedures, this exploratory international partnership project mobilizes researchers from the social sciences, earth sciences and computer science to study urban vulnerability. VUSIM-AMS aims to: 1) determine the physical vulnerability of buildings for crisis manageent 2) develop a multi-agent model of human mobility in a crisis that could help to development more efficient evacuation plans, assess the effects of a public information campaign, etc.

PROMIS: PROfessional network of Masters degrees in Informatics as a Second Competence (2013-2017)

PROMIS is a project sponsored by Tempus Program and is based on a previous ERAMIS project. Project number: 544319-TEMPUS-1-2013-1-FR-TEMPUS-JPCR. This is a teaching project that aims to: Extend the ERAMIS network in Central Asia; Improve the academic quality of the network; Improve the vocational focus of the master degrees in computer sciences in different ways; Establish strong relationships between companies and universities; Adapt teaching process to the students working in parallel of the studies; Improve the network by sharing courses and pedagogic material. It groups together: 5 unversities from Europe; 9 universities from Central Asia; 3 Europen companies. The partners are: 1. From Europe, Grenoble Alps University, Grenoble, France, Beuth University of Applied Sciences, Berlin, Germany, Kaunas University of Technology, Kaunas, Lithuania, Lublin University of Technology, Lublin, Poland, Savonia University of Applied Sciences, Kuopio, Finland. 2. From Kazakhstan, Al-Farabi Kazakh National University, Almaty, Kazakhstan, Gumilyov Eurasian National University, Astana, Kazakhstan. 3. From Kyrgyzstan, Kyrgyz National Technical University, Bishkek, Kyrgyzstan, Osh Technological University, Osh, Kyrgyzstan. 4. From Tajikistan, Khujand State University, Khudjand, Tadjikistan, Technological University of Tajikistan, Dushanbe, Tajikistan. From Turkmenistan, Turkmen State Institute of Economics and Management, Ashgabat, Turkmenistan, Turkmen State Institute of Transportation and Communication, Ashgabat, Turkmenistan. 5. From Uzbekistan, Bukhara State University, Bukhara, Uzbekistan, National University of Uzbekistan, Tashkent, Uzbekistan. 6. Industrial partners, eLeDia, enterprise, Berlin, Ilmi Solutions Oy, enterprise, Kuopio, Finland, SYMETRIX, enterprise, Grenoble, France.

SmartRescue: Smartphones for coordinated threat assessment and evacuation planning. (2012-2015)

This is a project managed by the University of Agder, Norway and funded by Agder Utviklings og Kompetansefond. The project explores how mobile wireless devices can be used in acute crisis situations where individuals need awareness of immediate threats, as well as plans for evacuation from the affected area in the safest possible way. I am collaborating with the team at the CIEM (Centre for Integrated Emergency Management) at the University of Agder on the task of simulating human behaviours in crisis situations.

SmartEnergy: Federative Research project (4 research labs) concerning developing new approaches and decision-making tools for undersatnding and managing energy from the SmartGrid to the SmartHome (to start end 2012)

Funded by Grenoble-INP, this project will look at the forthcoming new generation of energy systems. The work involves developing multi-scale and multi-form demonstrators and simulation systems that model the breadth and diversity of the various actors involved in the energy supply and consumation chain.

MoCA (My Little Artificial Companions World) (2010-2014)

This ANR funded project looks at developing hyrid societies of human and artificial agents (robots, virtual characters, software agents). The challenges involved include analysing how humans can interact and develop live long relationships with artificial characters. The project involves members from Lab-STICC, LIG (MAGMA and IIHM teams), LIMSI and LTCI. 

SUPERBAT: Simulation tools for Energy Management in Buildings (2010-2014)

In Europe, buildings account for 40-45% of total energy consumption. Simulation is a powerful tool not only for helping us to reduce energy consumption in existing buildings but to help us design new low consumption buildings. Inhabitants' behaviours have a huge impact on energy consumption. However current simulators tend to ignore this aspect and largely focus on simulating physical aspects based for example on thermal models. Funded by ANR, the SUPERBAT project aims to incorporate human behaviour modelling in order to better predict energy consumption. Furthermore, the project aims to improve the physical modelling of buildings by dynamically taking into account the energy load. Ayesha Kashif, my current PhD student, works on this project in the area of modelling dynamic human behaviour using an agent-based approach. The project includes 6 partners and is coordinated by EDF.

ISSUE: Integrated Spatial decision support Systems for Urban Emergencies (2009 - 2011)

Funded by STIC-Asie, the project brings together computer scientists, geophysicists and remote sensing specialists from Vietnam, China, Japan and France in order to investigate building spatial decision support systems for natural disasters in urban areas. Coordinated by Alexis Drogoul, IRD IFI/MSI, I am involved as the PI for MAGMA-LIG looking at human behaviour modelling in disaster situations.

C2EC2: Cognitive Complexity and Error in Critical Care (2007 - 2012)

The project is funded by the James S. McDonnell Foundation and is coordinated by Vimla L. Patel, co-director of the Center for Cognitive Informatics and Decision Making at the University of Texas at Houston. I am involved as one of the European collaborators. This multidisciplinary project is concerned with with modelling complexity and errors in critical care (emergency medicine and intensive care) environments. The project's approach emphasizes the cognitive properties of the system within which the error occurs and the notion of distributed cognition. Further information on the project is available here. The main themes of the project are:

  • to develop models of complexity and distributed cognition and to use them to develop intervention strategies to predict and prevent future error.
  • to develop team training protocols, simulators and on-the-job tools to aid in prevention of medical error.
  • to develop a robust theoretical model of error that incorporates concepts derived from the study of complex systems, distributed cognition, and medical error.

COSI: Complexity in Social Science (2000 - 2004)

Although this project is quite old now I've included it as it greatly influenced my approach and it formed the basis for a lot of my current work. COSI was an EC funded TMR Network (Framework 5), involving 8 European partners and two non-European partners. My previous research team (GRIC-IRIT) was coordinator of this project and in addition to working as a researcher I was also co-project manager. The goal of the COSI Project was to understand and model socio-cognitive processes in the context of real organisational systems. The main themes of the project were:

  • Complexity
  • Socio-organisational systems
  • Modelling and Simulation (predominately using an agent-based approach)