Beam Alignment for Non-Stationary Environments Using a Novel Time-Varying Structured Bandit

In this paper, we study beam alignment which is effective in non-stationary environments. Specifically, we first present a novel multi-armed bandit method, time-varying structured bandits, and then, propose an adaptive beam alignment algorithm based on it. The algorithm identifies the best beam by not only narrowing down but also widening beam search regions adaptively, exploiting the correlation structure over beams which is time-varying due to the non-stationarity. We show that the proposed algorithm is asymptotically optimal in theory. Furthermore, we demonstrate via simulations that our proposed algorithm outperforms other state-of-the-art algorithms.