ASTRID - Accompagnement Spécifique de Travaux de Recherches et d'Innovation Défense

Viability and autonomy of systems in unreliable and constrained environment – VIATIC

Submission summary

The VIATIC project aims at prototyping a robotic coordination system. The mission of the robots is to reach several targets having a slow motion and their position is not precisely known.
This is a multi-robot / multi-target problem which has a strong dual character. The problem is to relocate targets previously indicated and to allocate the most suitable robot for a mission of interception, emergency or other mission depending on the scenario. The imperfect knowledge (incomplete map of the environment) and inaccurate (localization errors) are the essential elements of these scenarios.

In the military field, the algorithms developed can be applied to the problem of interception of mobile targets with re-allocation. These multi-missile / multi-targets scenarios require the development of new strategies of guidance and allocation of cooperative objectives that can be applied to ground attacks in urban areas.

In the civilian field, the application of this problem is the redistribution of electric vehicles with autonomous control functions in order to reach a charging station and / or a parking space. It is also a multi-vehicle / multi-target scenario for which viable trajectories have to be computed in constrained environments with conflicts avoidance and allocation of cooperative objectives.

The coordination of the robots is done using optimization algorithms, embedded if possible, taking into account constraints and the uncertainty on the environment and on the location of the targets. The realization of the system is based on recent developments in viability and dynamic games theory.

This application is a natural continuation of many works that have already been made, both theoretically and practically, regarding the capture of a static target by a robot. The introduction of constraints and obstacles, and the extension to a multi-target / multi-robot context generate more complexity. The existing approaches, to our knowledge, have great difficulties to treat such a problem. This project aims to explore a new solution, which can lead to good results despite the high complexity of the problem.

The analysis and the solution of such problems are the main aims this project. It has the ambition to go from mathematical theory to experimentation on a real multi-robots system.


The four partners involved in the project are among the leaders in their respective fields:
- IRSEEM, as the designer of this project. IRSEEM will be in charge of the realization of the demonstrator and will be the coordinator.
- CREA, as the research lab in viability and complexity theory.
- LASTRE, as an expert in mathematics, software design and consulting.
- MBDA France, worldwide leader in the development and manufacturing of missile systems.

Project coordinator

Monsieur Xavier SAVATIER (Institut de Recherche en Systèmes Electroniques Embarqués) – xavier.savatier@esigelec.fr

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

IMM Institut Marcel Mauss
IRSEEM Institut de Recherche en Systèmes Electroniques Embarqués
MBDA MBDA

Help of the ANR 294,400 euros
Beginning and duration of the scientific project: October 2011 - 36 Months

Useful links

Sign up for the latest news:
Subscribe to our newsletter