• Other associations
• Contact
• Selling conditions
• Legal notice

Numéro : 2545 - Year : 2008

Technologies for data fusion applied to robust obstacle detection

Gilles MICHEL, INTEMPORA
Nicolas DU LAC, INTEMPORA
Claire DELAUNAY, INTEMPORA

International Autonomous Surface Symposium (ASSS 2008)

Aerospace, ground or naval Intelligent Robotics Systems or Drones, rely on a robust perception of their localization in order to appraise their route and dynamics and to detect in due time any obstacle or adverse mobile which could represent a real danger for their integrity.

Data fusion algorithms allow the use of multiple sensor information to enhance processing efficiency and accuracy and get such robust perception. Hence it is possible to increase the perception quality to its optimum by merging information from high quality sensors or, conversely, to get a standard quality level using low cost sensors.

Each data fusion problem remains unique. But some characteristics are always the same ones: the multisensor and real time processing context, the need to consider simultaneously the data values and their time stamps, i.e. the date at which the values have been measured by the system. A simple and efficient approach is offered by RTMaps, a software developed by Intempora S.A. the associated methodology of which is presented in this article which stresses the pragmatics of data fusion techniques.

(Text in French only)