Digital Predistortion with Compressive Observations for Cloud-based Learning

In this paper, we propose a novel system architecture for digital predistortion (DPD) of power amplifiers (PA), where the training of the DPD model is done in the cloud or distributed unit (DU). Following the current trend to partition the radio access hardware into a low-complexity frontend, and perform computationally expensive processing separately, we propose to split the DPD system and perform the compute-intensive DPD training in the DU. To enable the distant training the observed PA output must be available, however, passing the data intensive observation signal to the DU adds cost to the system. A low-complexity compression method is proposed to reduce the bit-resolution of the observation signal by removing the known linear part in the observation to use fewer bits to represent the remaining information. Numerical simulations show a reduction from 8 to 4 bits/sample for the accurate training the DPD model to compensate for the distortions of a strongly driven PA operated at 28 GHz with a 200 MHz wide OFDM signal.