In mountainous and hilly regions, landslides represent a real risk of ground destabilisation that can threaten the lives of people and animals, as well as cause damage to infrastructure.
In most cases, these events are difficult to predict using traditional methods as they are characterised by rapid rock movements or loose deposits along the slopes.
In fact, at the present state of the art, the evaluation of the susceptibility of a terrain to landslides is mainly based on indices linked to rainfall: this allows only macroscopic information to be obtained, inherent to very large geographical areas and without considering the degree of saturation already reached by the soil.
To cope with these limitations, the RASPLAN project led to the development of a compact radiofrequency probe that, inserted in a borehole, allows continuous recording of soil moisture data at various depth levels.
Thanks to specific algorithms developed, based on geological soil models supported by artificial intelligence techniques, the system then allows a landslide risk index to be determined in real time.
The project, financed by the Swiss Agency for the Promotion of Innovation (Innosuisse), involved a consortium consisting of research institutes, institutional and industrial partners: the Institute of Systems and Applied Electronics (ISEA) and the Institute of Earth Sciences (IST) of SUPSI, the Laboratory for Web Science (LWS) of the FFHS, the Office for Forests and Natural Hazards of the Canton of Graubünden and the companies GeoAlps Engineering SA and Solexperts AG.
"It was a very interesting and interdisciplinary project: the fruitful collaboration of electronic engineers, computer scientists and geologists allowed everyone to broaden their horizons and acquire new knowledge by successfully tackling the various challenges", said Prof. Ricardo Monleone, ISEA Scientific Contact Person. "In particular, the project's test site is located in a remote location in the Val Müstair (GR), which is practically unreachable during several months of the year for any technical interventions. It was therefore a mission with no margin for error."
Although the project is now complete, the system will remain operational in Valpaschun for several months, continuing to send data useful for refining the models developed and training the neural networks used.
Further information: