Abstract:
The wide bandwidth available at the millimeter wave (mm Wave) frequencies
is expected to offer high data rates in the fifth generation (5G) of cellular
networks. The technology implements directional transmission to overcome
increased path loss at high frequencies. The dependence on directionality
urges to establish new control layer protocols because the algorithms implemented in omnidirectional long term evolution (LTE) systems are not suitable for these networks. The mmWave base station (BS) and user equipment
(UE) need to be properly aligned for directional communication constituting
long-lasting initial access (IA) phase. Recently, several research works have
been done to devise smart IA procedures for mmWave systems. Some of
these schemes periodically sweep across the cell area while others make use
of the contextual information regarding BS and UE profiles and propagation
environment to establish the link. This paper proposes a smart machine
learning-based context-aware sequential algorithm for IA in 5G mm Wave
systems and analyzes its performance in comparison to the conventional exhaustive and iterative search algorithms. The algorithm is shown to provide
a comparatively lower misdetection probability and a smaller discovery delay.
