Sites Inria

Version française


Fundamental research and applications with societal impact

Plateforme Vgate : immersion et interaction 3D grandeur nature © INRIA / Photo Kaksonen

Through its bases in Grenoble and Lyon, Inria Grenoble - Rhône-Alpes is a major player in research and innovation in computational sciences in the Rhône-Alpes region. The research centre contributes especially to the fields of embedded software and imagery in Grenoble and to information technologies applied to life sciences in Lyon.

The research centre contributes especially to the fields of embedded software and imagery in Grenoble and to information technologies applied to life sciences in Lyon.

The main research topics of Inria Grenoble - Rhône-Alpes, defined within its strategic plan for 2013-2017 "Towards Inria 2020" are:

  • Distributed systems and mobile networks.
  • Reliable software and embedded systems for ambient computing.
  • Modeling and simulation of multi-scale and multi-component phenomena.
  • Perception and interaction with the real and virtual worlds.

On this research topics, have been identified six scientific priorities:

Robots sharing our workspace and living space

Robots are moving into our complex, every-changing daily environments with us. They need to be robust, safe and effective in their perception and mutual understanding. This also requires anticipation, decision-making, action and social interaction.There is a range of different robots including humanoid robots, smart buildings, electric cars, drones and robotic arms. What they all have in common is that they share our living and working environments with us, which leads to fundamental scientific problems needing generic, robust solutions for perception, understanding, decision-making, action and interaction.

The Internet of things and the Internet of data: digital society

The Internet of Things can be defined as an extension of the Internet where everyday items, housing, cities and even the environment are equipped with communication and interaction skills.The first challenge is to model, design and develop new applications and to be able to test them in situ through large-scale tools (infrastructure broadband networks, clouds, sensors, wireless networks, cognitive radio components) in order to address any issues related to the Internet of Things.The second challenge concerns the massive quantity of varied data, generated by the components of the Internet of Things. It is particularly important to know how to associate semantics to such data, and ensure its privacy and safety with the capacity to calculate and analyse.

Modelling interactions in biology

The functional mechanisms of living systems are closely linked to existing interactions between atoms, molecules, cells, organisms and species on different scales of time and space. To understand how biological systems work, we must look into how these interactions are organized.There are a number of experimental and methodological challenges in integrating different types of data, variability and the temporal dynamics of interaction networks and approximations in comparison to the actual system. Various kinds of expertise in models and methods are required to face these challenges and better understand the fundamental role of intra-and inter-organisational interactions on evolutionary health and biodiversity processes.

Forms, appearances and movements for virtual worlds

The enhanced performance of video games, virtual worlds, 3D animation and interactive applications has opened up new areas of investigation in the design and analysis of 3D models. The main challenges concern establishing a model of living forms to be able to generate and capture animated models as close as possible to reality and the establishment of models for natural phenomena.

We are exploring two areas of research:

  • Creating static and animated models working from observed data or procedurally.
  • Analysing natural forms and phenomena from their observations.

The data exploitation produced with the Kinovis platform allows us to experiment with innovative applications for movement analysis, medical diagnostic assistance and even animation reports for cinema.

Hardware-software interface

The main challenge of this scientific priority is making the programming of new multicore architectures as easy and transparent as sequential programming. This involves addressing any issues with multicore programming and hardware accelerators, as well as any issues with memory architectures’ design and operation and assembling hardware components in complex infrastructures.The first research area involves providing new adapted programming languages (e.g. data-flow), developing compilers, code generators and specialised circuits, and developing code analysis and formal verification. The second research area involves developing of implementing “light” cores and memory coherence mechanisms, addressing any issues with portability and virtualisation, observability and predictability, and scheduling and adaptive control.

Learning and distributed optimization for large scale systems

The data deluge generated by the Internet and the digitisation of society (Big Data) is transforming a range of domains such as imaging, biology, medicine and power generation systems through the influence of massive, heterogeneous and distributed data.Efficient data use requires learning and optimisation tools. This is so we can make use of the convergence between game theory, learning and distributed optimisation. In order to do so, we will use non-regular and stochastic optimisation to deal with problems concerning high dimensionality. We expect to see the possibility of processing terabytes of data, with, for example, benefits of several orders of magnitude in image classification, the effective establishment of a model for the visual complexity of the real world, and application to network routing protocols and “smart grids”.  

Keywords: Grenoble Lyon Areas of research