Watermarking of time‑series data for weather forecasts

Important results from the project

The project developed and evaluated a method for protecting mmWave measurement data through digital watermarking. The results showed that data manipulation can be detected reliably while preserving data quality and usability. The project also demonstrated that watermarking did not affect the performance of an AI-based rainfall classification model, confirming that data protection mechanisms can be introduced without reducing the value of the data for future analytics and sensing applications.

Expected long term effects

The project is expected to contribute to safer and more reliable use of sensor data in the long term by enabling verification of the authenticity and origin of the data without affecting its usability. The results create a foundation for future solutions in secure data sharing, AI-based analysis and networked sensor services, which can strengthen trust in data in both research and industrial applications.

Approach and implementation

The project was implemented as planned through the development, implementation and evaluation of methods for digital watermarking of mmWave measurement data. The work included signal processing, development of watermarking algorithms and verification of the results using AI-based rain classification. The project mainly followed the planned schedule and the objectives were achieved.