- The new system measures rainfall using acoustic sensors and machine learning technology without moving parts.
- The system provides a low-maintenance alternative that expands meteorological monitoring networks.
- Researchers aim to collect more data to optimize the system for handling heavy rainfall events.
- A denser network of rainfall sensors can improve real-time monitoring of localized precipitation.
Scientists have developed a system that utilizes structural acoustic sensors and machine learning techniques to measure rainfall without any moving parts. This technology can serve as a complement to traditional rain gauges. The system has an average error of 1.88 millimeters per hour. Developed by researchers from the Fraunhofer Institute for Digital Media Technology (Fraunhofer IDMT) and Pan Acoustics GmbH, this new artificial intelligence (AI) rainfall monitoring system can estimate light rainfall with an average error of just 1.88 millimeters per hour.
This new approach offers a low-maintenance alternative to expand meteorological monitoring networks by analyzing the tiny vibrations generated when raindrops hit solar panels.
How the New System Works
The system uses structural acoustic sensors and machine learning to measure rainfall without any moving parts. This technology is capable of capturing highly localized rainfall by creating dense, low-cost measurement points, thus complementing traditional rain gauges. Researchers demonstrated this method through the recently completed lokalRAIN project, testing the system in laboratory experiments and outdoor installations. They are currently seeking industry partners for additional pilot deployments to evaluate the technology's performance under broader operational conditions.
Jakob Bergner, the scientific director of the Fraunhofer IDMT project, stated, "We have demonstrated the feasibility of our approach for light rainfall. Now we hope to collect more data to address less frequent heavy rainfall events and optimize the procedure under different operational conditions." The system installs structural acoustic sensors on photovoltaic (PV) modules. Each raindrop hitting the panel generates microscopic vibrations, creating unique acoustic signatures that vary with rainfall intensity.
The system does not send raw sensor data to a remote server; instead, it processes information directly at the measurement site using an AI model running on energy-efficient microcontrollers. Only the calculated rainfall values are transmitted, significantly reducing power consumption and wireless bandwidth requirements. Pan Acoustics developed the hardware and software for the wireless communication network. In field tests, the system achieved a transmission range of about 2.5 miles (4 kilometers) in urban environments and up to 22 miles (35 kilometers) in unobstructed rural areas. Data collected from multiple sensor nodes is then integrated into a central data infrastructure to generate localized precipitation maps.
Challenges of Traditional Monitoring
Traditional precipitation monitoring relies on meteorological stations equipped with dedicated rain gauges, which are expensive to install and maintain. As a result, monitoring networks are relatively sparse in many areas, making it difficult to capture localized downpours, as rainfall can vary dramatically over short distances. To validate this new method, researchers first conducted experiments in a rainwater simulator, testing different droplet sizes, rainfall intensities, photovoltaic modules, and sensor installation locations. They subsequently installed three outdoor measurement systems at Fraunhofer IDMT in Ilmenau, Pan Acoustics in Wolfenbüttel, and the meteorological garden at Ilmenau University of Technology.
The collected data was compared with reference rainfall measurements to train and validate the machine learning models.
For rainfall rates of 2 to 5 millimeters per hour, the system's average error is 1.88 millimeters per hour, meeting the World Meteorological Organization (WMO) Class C accuracy requirements for low-intensity precipitation measurements. Researchers emphasize that this technology is intended to complement rather than replace high-precision rain gauges. A denser network of rainfall sensors can improve real-time monitoring of localized rainfall and provide valuable data for urban digital twins, enabling cities and infrastructure operators to respond more quickly to flood risks. This information can also support drainage planning, insurance risk assessments, and meteorological forecasting models.
Fraunhofer IDMT is also exploring agricultural applications through another project that integrates sensor networks into agrivoltaic systems, allowing farmers to make more targeted irrigation decisions based on rainfall data specific to their locations. Before commercial deployment, researchers plan to conduct additional pilot installations to collect more data during heavy rainfall events and further refine the sensor enclosures, electronics, and quality assurance processes required for production-ready systems.
How the New Technology Improves Rainfall Monitoring
This new technology provides a low-cost and low-maintenance method for rainfall monitoring through acoustic sensors and machine learning, capable of capturing variations in localized rainfall. Traditional precipitation monitoring systems have sparse networks due to high installation and maintenance costs, making it difficult to effectively respond to localized downpours. The emergence of the new system will help improve the accuracy of meteorological data and provide real-time rainfall information for urban infrastructure, further enhancing flood risk management capabilities.

