# ÄLLITÄ: How Weather Influences 5G mmW Signal Quality - Savonia AMK

> Use of 5G signal strength attenuation to detect very local changes in weather.

## Project Overview
In our ÄLLITÄ-project (Älyä lämmitysjärjestelmiin Ilmastoystävällisesti tekoälyä hyödyntämällä, Regional Council of Pohjois-Savo, J10756, co-funded by European Union), Work package 2 is focusing on AI and data in smart energy systems. One item was to study the use of 5G signal strength attenuation to detect very local changes in weather and use that information indirectly to predict for example solar panel energy production. 

## Methodology
The measurements are based on Savonia UAS private 5G NSA network measurements and FMI weather station data. The FMI weather station is located at Savilahti, Kuopio, very close to the Savonia campus area, providing local weather information. The dataset covers data from July 1st, 2025, until September 30th, 2025.

## Key Findings
Recent analysis reveals how environmental factors influence wireless signal performance:

*   **Humidity and SNR (-0.42):** The strongest correlation observed, indicating that higher humidity significantly reduces signal-to-noise ratio, likely due to increased RF absorption and scattering. This suggests operators should consider adaptive modulation and error correction during humid conditions.
*   **Temperature vs RSRP (-0.31):** Elevated temperatures slightly weaken signal strength, possibly affecting equipment efficiency and propagation. Proactive cooling strategies and antenna optimization during heatwaves can mitigate this impact.
*   **Humidity vs RSSI (-0.28):** Moisture degrades received signal strength, reinforcing the need for additional small cells or repeaters in high-humidity zones.

## Conclusion and Future Work
These insights highlight the importance of weather-aware network planning. By leveraging predictive analytics and dynamic resource allocation, operators can maintain robust connectivity even under challenging atmospheric conditions. We aim to further analyze and validate these results; if the findings are promising and support our hypothesis, we plan to submit them to a suitable scientific conference.

## Authors
*   **Aki Happonen**, Digital Development Manager, DigiCenter, Savonia-ammattikorkeakoulu, aki.happonen@savonia.fi
*   **Shahbaz Baig**, RDI Specialist, DigiCenter, Savonia ammattikorkeakoulu, shahbaz.baig@savonia.fi
*   **Mika Leskinen**, RDI Specialist, DigiCenter, Savonia-ammattikorkeakoulu, mika.leskinen@savonia.fi
*   **Laura Leppänen**, RDI Specialist, Savonia-ammattikorkeakoulu Oy, laura.leppanen@savonia.fi