Data set for Neural-Network-Based Digital Predistortion for Active Antenna Arrays Under Load Modulation

The dataset contains over-the-air measurements on a 64 active antenna array (Anokiwave AWMF-0129) operating at 28 GHz carrier frequency and transmitting a 200 MHz OFDM waveform with FFT size of 4096, 3168 active subcarriers, subcarrier spacing of 60 kHz and 5 times oversampling w.r.t the critical sampling rate. The dataset contains the I/Q samples of the TX waveform as well as the corresponding over-the-air received signals when the electrical beam is steered toward different beamforming directions. This dataset allows to observe and study the so-called beam-dependent load modulation, which is the phenomenon that causes the nonlinear characteristics of the antenna array to change with the beamforming direction. A Matlab script for data visualization is also provided.

For further details please refer to the following papers:

A. Brihuega et al., "Piecewise Digital Predistortion for mmWave Active Antenna Arrays: Algorithms and Measurements," in IEEE Transactions on Microwave Theory and Techniques, doi: 10.1109/TMTT.2020.2994311.

A.  Brihuega et al., “Neural-Network-Based Digital Predistortion for Active Antenna Arrays Under   Load Modulation,” in IEEE Microwave and Wireless Components Letters, doi: 10.1109/LMWC.2020.3004003