CE56 - Interfaces : mathématiques, sciences du numérique - sciences du système Terre et de l’environnement

Rockfall risk: the use of Artificial Intelligence for the operational management of the risk – C2R-IA

Submission summary

Currently, the management of rockfall risk is mainly dealt with the construction of protective structures, which often represents a disproportionate cost compared to the financial resources of municipalities and private operators. Morever, it is often impossible to develop such solutions over an entire basin exposed to rockfall risk. A more sustainable and effective strategy for rockfall risk management would be to take into account the influence of weather conditions on hazard level, thus allowing infrastructure managers to anticipate an increase of hazard level in order to implement risk mitigation systems (restriction of access, monitoring, mobilization of emergency kits, predictive maintenance). Such dynamic risk management is potentially associated with high socio-economic costs. Thus, its implementation requires a well-justified decision-making procedure.
The aim of this project is to overcome the necessity to construct decision-making procedure based on expertise by improving our understanding of the cliff behaviour under climate forcing in order to produce efficient predictive models.
This issue is addressed through the 3 main axes of the project:
• Make the detection of rockfalls more reliable, in particular by taking advantage of technological progress to acquire massive and varied data, by using them with improved processing methods, as well as by comparing different data sources (data fusion).
• Develop effective predictive models which results can be interpreted by the expert using recent innovations in Artificial Intelligence (AI), making it possible to translate the results into potential socio-economic costs and operational risk management rules.
• Transfer predictive models from a site to a new site at a lower cost by applying innovative AI transfer learning methods to these predictive models, as well as by testing the relevance of data from low-cost devices for training of these models.

Project coordination


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.


GéoCoD Centre d'études et d'expertise sur les risques, l'environnement, la mobilité et l'aménagement
LISTIC Université Savoie Chambéry
ISTERRE Institut des Sciences de la Terre
LIRIS Laboratoire de Recherche en Image et Systemes d'Information

Help of the ANR 703,274 euros
Beginning and duration of the scientific project: February 2023 - 48 Months

Useful links

Explorez notre base de projets financés



ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter