Machine-Learning-Aided Massive Hybrid Analog and Digital MIMO DOA Estimation for Future Wireless Networks

Abstract

Due to a high spatial angle resolution and low circuit cost of massive hybrid analog and digital (HAD) multiple-input multiple-output (MIMO), it is viewed as a valuable green communication technology for future wireless networks. Integrating the massive HAD-MIMO with direction of arrival (DOA) will provide an even ultra-high performance of DOA measurement, which can the fully-digital (FD) MIMO. However, phase ambiguity is a challenge issue for a massive HAD-MIMO DOA estimation. In this paper, we consider three parts: detection, estimation, and Cramer-Rao lower bound (CRLB). First, a multi-layer-neural-network (MLNN) detector is proposed to infer the existence of emitters. Then, a two-layer HAD (TLHAD) MIMO structure is proposed to estimate the DOA and eliminate phase ambiguity using only one time block. Simulation results show that the proposed MLNN detector is much better than both the existing generalized likelihood ratio test (GRLT) and the ratio of maximum eigen-value (Max-EV) to minimum eigen-value (R-MaxEV-MinEV) in terms of detection probability. Additionally, the proposed TLHAD structure can achieve the corresponding CRLB.