SpeedSense helps telecom operators understand and improve mobile internet speed using the power of artificial intelligence.
This project was developed as part of an Integrated Engineering Project at ESPRIT School of Engineering.
In an era where 5G adoption is surging, optimizing network quality is no longer a luxury, it's a necessity. SpeedSense empowers telecom operators by predicting throughput performance offering clear, data-driven insights through AI-powered explainability.
Our best model achieves an R² score of 0.949, offering accurate and robust throughput predictions to assist in proactive network optimization.
Trained on a real-world dataset containing over 18,000 samples, our model learns from a wide range of network and environmental conditions.
We leverage a rich, multidimensional dataset to ensure accurate, real-time throughput predictions. Each feature plays a unique role in understanding and optimizing 5G network performance.
PCell & SCell RSRP / RSRQ / RSSI / SNR : Indicators of signal strength, quality, and noise for primary and secondary cells. RSRP_SNR_ratio / RSRQ_SNR_ratio: Combined indicators enhancing prediction reliability
Num_RBs / Average_MCS / Bandwidth (MHz): Provide insight into resource blocks, modulation, and bandwidth usage for data transmission.
Latitude / Longitude / Area / location_cluster: Help model performance by physical region. Timestamp / Hour / Day / Weekday / is_weekend: Temporal features to understand variations across time.
Temperature / ApparentTemperature / DewPoint / Humidity / Pressure / CloudCover / PrecipProbability: External factors that can impact wireless signal quality.
Speed_kmh: Device mobility, crucial for handover and throughput changes. Traffic Jam Factor: Assesses road congestion, affecting network behavior in mobile environments.
Operator / SCell_Active: Network-specific features used for classification and behavior understanding. Target: The predicted throughput class — what the model learns to predict. .
Our project is aligned with the United Nations’ 2030 Agenda for Sustainable Development. SpeedSens contributes to smarter infrastructure and urban resilience by addressing 5G optimization needs.
SpeedSense supports digital transformation in the telecom sector by introducing intelligent tools to monitor and forecast throughput performance, improving the reliability and efficiency of 5G networks.
With better network prediction comes improved connectivity in urban and rural zones. SpeedSense helps enable smarter mobility, better digital services, and more resilient city infrastructure