Low-Complexity Adaptive Beam and Channel Tracking for Mobile mmWave Communications

In this paper, we study low-complexity algorithms for beam and channel tracking for millimeter-wave (mmWave) communications. In particular, the least mean squares (LMS) and bidirectional LMS (BiLMS) algorithms are derived for a mobile mmWave transmission scenario, where channel measurement is a nonlinear function of the unknown angle-of-arrival (AoA) and angle-of-departure (AoD). Numerical results confirm that LMS is superior to widely used Extended Kalman Filter (EKF) algorithm in tracking the mmWave beam, when the initialization of AoA/AoD and channel gains is imperfect (i.e., performed using noisy channel estimates). Moreover, BiLMS exhibits a very good mean square error (MSE) performance as compared to both LMS and EKF, which makes it a promising channel tracking algorithm for a mobile mmWave transmission scenario. We also show that LMS and BiLMS algorithms are more robust against the impairments due to the non-optimal antenna array size as compared to EKF, and show relatively faster convergence characteristic along with increasing signal-to-noise ratio (SNR).