Strategic Initialization of Genetic Algorithm used in Digital Pre-Distortion of mmWave Power Amplifiers for Hybrid Beamforming

As wireless cellular infrastructure experiences a need for wider bandwidths and higher datarates, power am-plifiers (PAs) are an area of intense interest. PAs when supplied with high power signals can generate spectral regrowth which can interfere with adjacent channels. These unwanted emissions can be eliminated with the help of Digital Pre-Distortion (DPD), however the challenge of using a single DPD function to remove the unwanted emissions across a number of PAs is still being investigated. In this paper we present a novel method to linearize the combined response of a set of array amplifiers using a single DPD block. In the DPD block, a combination of least squares (LS) and genetic algorithm (GA) are used to find a common set of DPD coefficients for multiple paths. Traditionally the initial population for a genetic algorithm takes random values but in this work a more computationally efficient approach to selection is taken for the initial population and a unique coefficient adaptation strategy is successfully employed. This approach is validated experimentally and results from this work demonstrate a 14 dB of adjacent channel power ratio (ACPR) improvement.