Detailed Description
This dataset is generated from controlled indoor laboratory experiments with a multi-cell 5G setup and a single stationary UE (Google Pixel 5 or Quectel RM5xxG). The line-of-sight distance between the UE and gNB ranges from 1 to 3 meters, with no UE mobility during data collection. Downlink traffic is generated using iPerf, employing TCP with the CUBIC congestion control algorithm. Measurements are recorded at a 1-second granularity and include both downlink and uplink MAC-layer KPIs, such as ACK/NACK statistics, bitrate, MCS, SNR, and the number of LDPC decoding iterations per transmission attempt. In parallel, the dataset captures detailed RAN configuration parameters, including nominal P0, downlink and uplink PRB allocation, transmit and receive gains, power ramping steps, LDPC decoder iteration limits, center frequency, maximum UE transmit power, and TDD patterns. This joint view of configuration and performance enables systematic analysis of how RAN parameter settings affect observed KPIs.
Key Features
- 1-second–granularity time-series RAN measurements
- Joint visibility of RAN configuration parameters and MAC-layer KPIs
- Controlled indoor laboratory setup for reproducibility
- Suitable for modeling, optimization, and control research
Use Cases
- Configuration-to-KPI modeling
- RAN parameter tuning and optimization
- Evaluation of data-driven and learning-based control algorithms
- Benchmarking RAN performance under controlled conditions
