Search

We present WASDI, an open-source, federated multi-cloud platform for Earth Observation (EO) data that has been used successfully in several impactful international projects in the field of Natural Hazards (NH), in particular with floods and wildfires. WASDI helps EO developers turn their algorithms into operational services in the cloud. These can also be published in a marketplace where end users can exploit them. The platform automatically interoperates with several data providers. The accessible data includes observations (such as those from the Copernicus and Landsat programmes), derived products (such as the Copernicus DEM and the ESA World Cover), and the output of simulations (such as ECMWF ERA5 reanalysis). A single interface simplifies the access and thus the development of algorithms that operate data fusion using different sources.Thanks to these capabilities, WASDI is used to daily monitor floods in several South East Asian countries, in the context of a World Bank initiative to promote parametric reinsurance of the financial risks associated with floods. The extent of the floods on bare soil are mapped mainly using an algorithm developed by the Luxembourg Institute of Science and Technology (LIST), using Sentinel-1 data, and integrated by an algorithm that uses Sentinel-2 data developed by the CIMA Foundation. Urban floods are mapped with great accuracy using Sentinel-1 thanks to an innovative algorithm developed by LIST. These algorithms have been deployed to the platform and are now available in its marketplace.Recently, these technologies have been used to perform an extensive assessment of the effects caused by the very large floods in Pakistan in 2022 to support the Asian Development Bank (ADB) recovery efforts in the framework of ESA’s Global Development Assistance (GDA) Disaster Resilience program.The platform features also other algorithms, such as those for the detection of active fires based on Sentinel-3 data, which have been used for the assessment of wildfires in the Greek island of Evia in 2022, and for mapping burned areas using couples of Sentinel-2 images, developed by CIMA Foundation and used by the Civil Protection Department of Italy and other countries. This algorithm was also used to help assess the impact of the wildfires of 2020 in Ukraine, close to the Chernobyl power plant.These cases demonstrate the effectiveness of Earth Observation data, algorithms and cloud technology in case of natural hazards for prevention, response and assessment.

The growing number of deep space exploration missions operating simultaneously in the solar system triggers an increasing demand for large ground antennas capable of tracking distant spacecraft. Several space agencies have their own deep space tracking networks, where each antenna belonging to a ground station complex is meant to be shared among different deep space missions in flight, significantly constraining the tracking schedule. A typical ranging and Doppler radio tracking session requires long tracking passes and a single ground antenna, while angular (Delta-DOR) observations require at least two antennas but usually for much shorter tracking sessions. However, during Delta-DOR observations, the baseline between the two receiving antennas should be kept as large as possible, thus reducing the time windows in which Delta-DOR observations are actually feasible. This leaves little room for adaptation of the tracking schedules of these antennas and calls for the need for possible alternatives for the receiving stations. The antennas belonging to the very long baseline interferometry (VLBI) network worldwide meet the requirements to carry out Delta-DOR tracking sessions. Here, we present an experimental activity carried out tracking ESA’s GAIA spacecraft using a mixed deep space antenna configuration involving an ESA ESTRACK antenna at New Norcia (Australia) and a VLBI antenna at Medicina (Italy). This baseline was used to form Delta-DOR observables with the aim of demonstrating that VLBI antennas offer the capability to track deep space missions, thus increasing the number of possible baselines and observation time windows.

Automating the design of elevator systems poses interesting challenges when it comes to deal with configuration constraints and to express "quality-of-design" targets that must also comply to specific norms and regulations. Our automated elevator design tool LIFTCREATE is based on an AI engine featuring special-purpose heuristic techniques which are successful in handling complex real-world designs, but are lacking flexibility and thus are difficult to extend to, e.g., different classes of systems or different norms and regulations. In this paper we describe the development of a replacement engine for L IFTCREATE which seeks to overcome the limitations of the current one. The new approach is based on satisfiability modulo theories (SMT) with theories of cost. We describe encodings to solve choice and placement of critical components in the elevator, and we experiment with our encodings using a state-of-the-art SMT solver. Our experiments show that the SMT-based approach is competitive with respect to the heuristic-based one, both in terms of efficiency and in terms of design quality.