Accurately measuring precipitation remains a major challenge, as reliable rainfall data is only available for selected locations equipped with weather stations. This data gap results in inaccurate weather forecasts and makes it difficult to respond swiftly and effectively to extreme weather events such as heavy rainfall. To mitigate the impact of flash floods caused by intense precipitation, municipalities, urban planners, and disaster management teams need precise and comprehensive precipitation data. This enables targeted planning and monitoring of flood-prone areas.

Developing an Acoustic Real-Time Rainfall Measurement System

The research partners in the "lokalRAIN" project aim to tackle these challenges by developing an acoustic rain measurement system for surface applications. The core idea is to capture the acoustic vibrations generated when raindrops impact surfaces using vibration sensors and analyze them in real-time with machine learning and deep learning techniques.

Building on this concept, researchers will develop a novel sensor system for detecting and analyzing spatially distributed rainfall using photovoltaic modules as a test platform. This involves not only designing efficient machine learning technologies but also developing cost- and energy-optimized computational hardware for sensor applications.

Another critical aspect is the reliable transmission of precipitation data from individual sensor nodes via connected infrastructures to a central server. On the server, the data is aggregated and analyzed in real-time based on precipitation parameters such as quantity, drop size, and spatial distribution to improve forecasting capabilities.

Application Areas of lokalRAIN

Jakob Bergner, project manager at Fraunhofer IDMT, highlights some key applications of lokalRAIN: "By collecting precipitation data across large areas, weather services can develop more accurate forecast models. In cases of extreme rainfall, real-time precipitation data can help disaster management teams plan emergency response measures effectively. Insurance companies can access detailed risk assessment information, while agricultural businesses can optimize their irrigation strategies, improving efficiency and conserving resources."

Contributions of the Project Partners

Fraunhofer IDMT is responsible for detecting precipitation events acoustically. By applying signal processing and machine learning techniques, researchers analyze the interaction between the vibrations caused by raindrops and the amount, intensity, and composition of precipitation.

Additionally, appropriate acoustic sensor concepts are being developed to enable integrated processing and analysis of recorded acoustic signals. The networking of individual sensor nodes will allow precise real-time assessments of local precipitation data.

Pan Acoustics GmbH is responsible for developing the sensor node hardware and wireless data transmission technology.