VRAIN-UPV researchers advise a project that will create a digital twin of terrestrial satellite data

VALENCIA, 26 Sep. (EUROPA PRESS) -

A team of researchers from the Valencian University Institute for Research in Artificial Intelligence (VRAIN) of the Polytechnic University of Valencia has advised on the most effective Artificial Intelligence (AI) techniques to develop a software platform that will create a digital twin of the Earth, DTE (Digital Twin Earth).

The central element of this platform is made up of a set of intelligent computing modules, organized in processing layers, which will facilitate the intelligent capture of data from satellites, mainly from the European Space Agency (ESA), and ground sensors.

In addition, it includes a second layer of data modeling in which, using AI techniques, a digital model is defined that emulates the real and physical behavior of the observed terrestrial phenomenon. And finally, a third layer of decision support in which AI techniques compute a probable future scenario about the observed terrestrial phenomenon to predict and recommend future actions or events.

The objective of the AI4DTE (Artificial Intelligence for Digital Twin Earths) project is to develop a set of software tools that, through AI, facilitate the development of digital twin Earth applications.

The project "uses artificial intelligence to generate a digital copy of all the parameters that satellites and sensors have data on."

By processing this information, a computational model is built that is a copy of what is really happening on Earth. Finally, from this model estimates are made and the future of some of these variables is predicted, such as, for example, air quality or the state of certain critical points on the coast or anything that can be predicted from this. information" explains the main researcher of this project at VRAIN of the UPV, Adriana Giret

The potential of this technology "also lies in the fact that through AI models combined with physical models we can observe the interdependencies between the observed variables and deduce patterns of behavior that emerge from them. Thus, the predictions of a DTE can be more precise," adds Giret.

Three application cases

In this sense, and within AI4DTE, three case studies are being carried out that will serve to validate all the technology and set of software tools that the project is developing.

The first of them will help to recommend what type of crops are most appropriate for the future, according to the quality of the land and in certain areas of the earth, specifically with data from Africa.

The second will analyze satellite images obtained from busy roads to know what is the most optimal time, both economically and environmentally, to carry out their maintenance.

And the third, also with satellite images, to see a very common phenomenon in coastal areas such as landslides and to be able to predict it very accurately and prevent it from affecting transport infrastructure or urban areas.

In addition to the AI ​​techniques that have to be used in this project and the most effective mechanisms to develop the tools, the VRAIN team of researchers from the UPV has also collaborated in carrying out some of the case study modules. And it has created a satellite data quality control module since within all of them "noise must be discerned from correct and significant data to ensure that the AI ​​models of the simulation layers and help in decision-making decision to work with quality data," emphasizes researcher Adriana Giret.

In its most ambitious vision, a DTE "could have, on a global and real scale, all the data about what is happening on Earth, thus being able, in an interconnected way, to observe absolutely all terrestrial phenomena, accurately predict the future of these phenomena and recommend beneficial actions to improve this probable future.

Today we are working to achieve this ambitious vision of a DTE and AI4DTE is one of those steps. We still need to process much more data, generate models of many more terrestrial phenomena, with a lot of computing power and develop many DTE applications," highlights Adriana Giret.

Six agents from companies and research centers participate in the AI4DTE project, which will culminate at the end of the year, and is coordinated by the company Telespazio.