Perturbation-Based Adaptive Beamforming for MU-mMIMO

A perturbation-based adaptive beamforming scheme employing particle swarm optimization is proposed to mitigate multi-user interference by creating deep nulls in the potential interference directions. The paper uses an 8x8 circularly-polarized antenna array prototype as an example and examines the variation in antenna element radiation patterns at different positions on the large array, as well as their impact on the adaptive beamforming performance. The proposed scheme considers the measured element patterns and uses weights generated by the Linearly Constrained Minimum Variance method as a starting point to search for new optimal steering weights, which create deep nulls with a slight tradeoff in the directivity. The illustrative results show the proposed algorithm can achieve normalized nulls lower than -48.1dB with an accurate nulling location control. Compared to the linear constraint adaptive beamforming scheme, the proposed approach demonstrates improved results, achieving on average 22.9dB deeper arbitrary nulls with a tradeoff of 0.7 dB drop in the beam directivity.