Inria's ambitions for Europe
Many Inria project teams are involved in the FP7. They are involved in around a hundred projects, including some sixty on information and communication technology. These teams are working with a large number of academic and industrial partners. Inria is the scientific coordinator for eight European projects.
The following presents these programmes, with major scientific and political challenges for European research.
CONTRAIL: computing resources at your service
With cloud computing, which emerged in 2007, companies are able, with the ebb and flow of their data storage or computing requirements, to rent out their own IT resources or, conversely, augment them from other companies. The purpose of the three-year Contrail project, initiated in October 2010, is to make such virtual resource management on demand transparent.
Researchers are using the results from a previous European project (XtreemOS). They are designing open source software, easily programmable, enabling resources to be managed very flexibly, but also enabling several clouds to be combined and viewed as a single platform. Several areas for application are being looked into, such as general public services and high-performance computing.
HUMAVIPS: social robots
Locating and identifying one or more people from among a noisy crowd, with background music. Something most of us can do, but beyond a robot's abilities. This is one of the objectives of the three-year Humavips project, started in February 2010.
Using two cameras and two microphones, researchers want to build some intelligent behaviour technology based on visual and auditory recognition into their humanoid robots. Using image and sound processing techniques, combined with statistical learning methods, they want to demonstrate that a humanoid robot can interact naturally with a group of people, capturing information from its surroundings, interpreting it and acting upon it.
Radu Horaud believes thatnbsp;"Setting up European collaboration firstly ensures a certain reputation. Several countries are involved, with four university teams from France, the Czech Republic, Switzerland and Germany in our case, plus one French industrial partner. Institutions, governments and business are keeping a close eye on project progress. Projects are appraised by scientists from outside Europe, which gives this type of project an international dimension. A European project is also an opportunity to work with diverse teams, the best in each field, and to exchange students (PhD students and post-doctoral researchers). Ties between teams are strengthened. This synergy typically continues well beyond the project's lifespan."
FEEDNETBACK: controlling with wireless sensor networks
Automatically running or controlling the operation of a building, road traffic or high-risk areas such as nuclear power plants. This is now potentially achievable using the latest wireless sensors which can be networked and communicate, compute and control on a large scale.
These sensors are however causing new problems connected to their small size and wireless communication. The control loop needs to be redesigned to make better use of their capacities and develop robust automatic management systems. This is the objective of the three-year Feednetback project started in September 2008. It combines research teams, organisations supplying case studies and partners to disseminate the results.
HITCH: the ability to exchange electronic healthcare data
The complexity of the electronic data used in healthcare (test results, X-rays, MRI scans, etc.) and the systems to process it is growing continuously. For patient safety, it must be possible to exchange this data between various healthcare specialists, across Europe.
The Hitch project is developing a road map to harmonise electronic healthcare data systems and ensure interoperability for patient data across Europe. Out of concern for reliability, these recommendations are tested on genuine medical data and actual electronic healthcare systems. The project, started in February 2010 to last a year and a half, is run by Inria and brings together the leading European names in healthcare (IHE Europe, EuroRec, MedCom).
GEONET: smart cars
An instant warning to cars behind that the car in front has just braked sharply, traffic information, or space availability in a car park. In the hope of reducing road accidents, vehicles could soon communicate such information with each other and the highways infrastructure.
The purpose of the two-year Geonet project, initiated in February 2008, is to combine current geopositioning network solutions with the most recent Internet communication protocol (IPv6) to create this unique, transparent communication architecture. It is expected that a standard will be proposed and tested on two prototypes.
CONNECT: adapting to non-uniform equipment
A variety of equipment such as computers, mobile telephones and wireless sensors may be networked up for personal or business use. Hitherto, the disparate nature of these systems was handled by the middleware software layer where data is interchanged between various applications. A solution that is not suited to the massive deployment of such networks.
The Connect project, started in February 2009 to run for three and a half years, is developing a revolutionary approach, a software solution which should make it possible to generate the appropriate interchange protocol on demand for systems that need to communicate.
FOX: a sound theory for XML
Internet users do not restrict themselves to viewing pages. They contribute to discussion forums, social networks, put videos on line, connect their digital equipment, etc. The Internet's flexibility and dynamics depend on communication between different websites. XML is the best-known communication language. Developed in 1998, it needs to be adjusted to current requirements.
Such is the aim of the three-year Fox project, initiated in May 2009. It will strengthen the theoretical foundations of XML and of the World Wide Web Consortium's (W3C) communication standards more generally.
SMALL: optimised signal processing
What do multimedia, audio, astronomy and medical imaging have in common? All are applications using signals (images, videos, audio) that need to be compressed in response to the growing quantity of data to be handled.
The objective of the three-year Small project, started in February 2009, is to develop signal representation models (new processing algorithms) using very few parameters. What mathematicians call a sparse representation. A very noisy signal can consequently be retrieved while eliminating interference, and acquisition can also be done in a both minimal and meaningful way.
The FP7 (Framework programme for research and development) is a funding scheme set up by the European Union to support and encourage European research for the purposes of keeping industry competitive.