Robust and efficient beam training scheme for millimetre wave indoor communications
Antenna beamforming is a key enabler for the deployment of reliable millimetre wave communication systems. Indoor millimetre wave network standards have adopted a multi-stage codebook-based beam training protocol with an objective to reduce the number of preamble transmissions required to identify the optimum transmit–receive beam pair. Multi-stage beam training schemes entail moderate search complexity; however certain implementation level challenges have an adverse impact on the search success efficiency. Moreover, it is desirable to reduce the complexity further in a dynamic environment with limited mobility where beam training needs to be repeated often. In this tudy, he authors develop a low-complexity algorithm for millimetre wave beam training within a heuristic numerical optimisation framework. Based on the classical Rosenbrock direct search method, he authors propose a threshold acceptance feature augmented with a divide-and-conquer strategy and direction of arrival aided initialisation, specifically to provide resilience to link blockage and enhance the search success performance, respectively. The method facilitates ease of implementation, and is applicable to generic phased array architectures irrespective of array geometry. he authors further evaluate the system bit error rate to benchmark the performance of the proposed scheme vis-a-vis optimal beamforming schemes.
