AI-Powered Rainfall Detection System Using TV Signal Processing

Use Case:

Integration of artificial intelligence into SRS devices to improve the accuracy and responsiveness of rainfall detection based on opportunistic signal data from TV antennas.

Outcome:

Feasibility of embedded AI-augmented SRS validated in real-world pilot; Demonstration of AI-enhanced rainfall detection using neural networks; Deployment plan for TRL7 prototype targeting municipalities and civil protection; Integrated hybrid model combining adaptive algorithms and machine learning; Embedded processing near the sensor, reducing bandwidth and central server load.

Ecosystem Support:

Mentorship via StairwAI program enabled delivery of a tailored business model for AI adoption, strategic alignment of AI feasibility with operational infrastructure, and exploration of new use cases such as flood forecasting and snow detection.

AI Relevance:

This project advances AI democratization by: I) applying AI to real-time environmental sensing on edge devices; II) leveraging AI to enhance low-cost, non-invasive weather monitoring systems; III) reducing infrastructure dependency via decentralized data processing; IV) enabling cost-effective deployment to smart cities and local authorities.

Summary:

Artys, a technology innovator in environmental sensing, developed the AI4RAIN system by embedding AI models into its Smart Rainfall System (SRS) – a non-intrusive, low-cost sensor leveraging TV signal attenuation for rain detection. The objective was to enhance accuracy and enable edge-based AI processing directly on embedded systems. The feasibility study validated the integration of neural networks for real-time precipitation detection, replacing or complementing existing adaptive algorithms. The pilot, planned for the Polcevera river basin in Genoa, involves a network of 11 SRS stations and targets TRL7 through full operational demonstration.
From a business standpoint, a dedicated model for AI adoption was developed using the StairwAI canvas. This framework guided the definition of a clear product vision, risk testing strategy, and long-term market fit. The project aims to serve civil protection agencies and local authorities with improved rainfall monitoring capabilities. Post-program plans include broader applications such as surface soil moisture estimation via satellite fusion, and solid precipitation detection through a collaboration with the University of Montreal. AI4RAIN shows how AI, paired with opportunistic sensing, can provide scalable and sustainable solutions for climate adaptation and urban resilience.