Fixed Budget Beam Identification in Millimeter Wave Systems via Unimodal Bandits

Debamita Ghosh, Manjesh Kumar Hanawal

Abstract: The millimeter-wave communication systems suffer from significant attenuation loss due to their small wavelengths, which significantly reduces the data rate from gigabits per second to a few megabits per second in 5G networks. To enable their use in the 5G network, it requires that the transmission energy be focused on sharp pencil beams. As any misalignment between the transmitter and receiver beam pair can reduce the data rate significantly, it is important that they are aligned as much as possible. Recent works of Beam Alignment (BA) propose to adaptively select the beams such that the cumulative reward measured in terms of received signal strength or throughput is maximized that continuously performs exploration and exploitation, and hence can incur significant BA latency, reduction in throughput, or outages. In this paper, we develop an algorithm that exploits the unimodal structure of the received signal strengths of the beams to identify the best beam in a finite time using pure exploration strategies, which are more suitable for wireless network protocol design. Our algorithm, named Unimodal Bandit for Best Beam (UB3), identifies the best beam with a high probability in a few rounds. We demonstrate that it outperforms the state-of-the-art algorithms through extensive simulations. Moreover, our algorithm is simple to implement and has low computational complexity.