[ITU AI/ML Challenge 2021] Dataset IEEE 802.11ax Spatial Reuse

This dataset has been created for the problem statement ITU-ML5G-PS-004 of the ITU AI/ML Challenge (2021 edition). More information can be found here: https://challenge.aiforgood.itu.int/ and https://www.upf.edu/web/wnrg/2021-edition. 

The dataset contains the information of 2.000 IEEE 802.11ax deployments (divided into two different scenarios) at which the Basic Service Set (BSS) of interest applies different possible OBSS/PD thresholds in the context of the Spatial Reuse (SR) operation. In total, 21 OBSS/PD values are considered for each deployment, so the dataset includes 42.000 simulations, which are expected to be used for training Machine Learning (ML) and Federated Learning (FL) algorithms.

More specifically, the dataset is divided as follows:

• Scenario 1: 1,000 different deployments with 2-6 APs and 1 STA per AP. A minimum distance limitation is applied, so that each AP different from AP_A is located at a minimum distance of 10 meters from that one. Each deployment is simulated for each of the 21 possible OBSS/PD thresholds in 11ax. Two files are considered:
  1. output_11ax_sr_simulations_sce1.txt: contains the output generated by the simulator for the deployments in Scenario 1.
  2. simulator_input_files_sce1.zip: contains the input files used by the simulator to simulate the deployments in Scenario 1.
• Scenario 2: 1,000 different deployments with 2-6 APs and 1-4 STAs per AP. No distance limitation is applied. Each deployment is simulated for each of the 21 possible OBSS/PD thresholds in 11ax. Two files are considered:
  1. output_11ax_sr_simulations_sce2.txt: contains the output generated by the simulator for the deployments in Scenario 2.
  2. simulator_input_files_sce2.zip: contains the input files used by the simulator to simulate the deployments in Scenario 2.