Parameter Estimation and Hybrid Precoding Design for Millimeter Wave Mobile Networks

University of Technology Sydney. Faculty of Engineering and Information Technology.With the exponential rise of mobile data rates, millimeter wave (mmWave) mobile networks (mmWMNs) have become the trend in the 5th generation mobile cellular networks and beyond. In mmWMNs, the mmWave band can provide the ultra-high data rates due to its extremely wide frequency band resources, and the densely deployed base stations (BS) can significantly improve the network throughput per cell. However, the severe path loss and fading issues of the mmWave band dramatically impair the received signal-to-interference-plus-noise ratio and limit the network throughput. A revolution in the hardware architecture and the signal processing has been occurring for years. Numerous novel channel estimation and precoding techniques were proposed. In particular, angular sparsity is an intensified property for conducting mmWave channel estimations and hybrid precoding is a promising technique to realize mmWave communications. Existing hybrid precoding schemes either require full channel state information (CSI) or use codebook-based design. The former one requires highly accurate estimated channels while the latter has a degraded system performance. On the other hand, mmWave radar sensing has been successfully and commercially adopted for decades. With the number of electric devices increasing rapidly, there exist more and more demands to fuse the radar functions into the mmWave communication mobile networks. The primary issue is to realize a robust mmWave communication system. Issues following this include how to jointly estimate the communication channel and the radar channel, and how to address the interferences between radar waveforms and communication waveforms. Under this background, this doctoral thesis mainly focuses on signal processing techniques that can realize mmWave channel estimation for both radar and communication purposes, and hybrid beamforming/precoding algorithms that can increase the communication data rates. This thesis will include: 1) Subarray-based angle-of-arrivals (AoAs) estimation, where the AoAs can refer to both the line-of-sight (LOS) angles coming from users and the non-line-of-sight (NLOS) angles coming from targets; 2) Energy-efficient hybrid precoding and sparse precoding (virtual array), where both fully-connected and partially-connected hybrid precoders are optimized based on the metric of energy efficiency; 3) Adaptive hybrid precoding and the quantization of radio-frequency (RF) precoder using minimum subspace distortion (MSD), where the adaptive precoding aims to adjust the precoding matrix based on the transmit power, and the MSD quantization aims to improve the system performance loss caused by the quantization; 4) Uplink radar sensing fused in mmWMNs, where a radar sensing scheme is proposed without requiring synchronization between BS and user equipment