No abstract

In this paper a 3D ray tracing tool is developed for mmWave outdoor environments. Coverage analysis and system performance is performed for a small cell system at 60GHz. Peak throughputs beyond 3Gbps and cell edge throughputs in excess of 500Mbps are demonstrated at 98% of the mobile test locations. Analogue codebook based beamforming is applied. The analysis of the statistical angular spread at the mobiles is shown to be clustered; leading to many users selecting the similar analogue beamforming code which results is high levels of co-channel interference. It is often assumed that the use of analogue beamforming implies a single beam pattern pointing in a specific direction to support a single user. In this case users that cannot be spatially separated must be scheduled in different time slots. However, due to the clustered aspect of the users' angle spreads, those with the same beam code can still be scheduled in the same time slot using different frequency channels via a single RF chain. The paper analyses a scheduling scheme based on this concept. The same scheduling scheme is applied to systems operating at 28, 38 and 60GHz where different numbers of frequency channels are available. Different user densities and street scenarios are studied. The 60GHz scheme make use of 4 frequency channels and is shown to double the system throughput compared to operation at 28GHz band using a single frequency channel. Finally, the optimal number of RF chains required at the base station is analysed. This is shown to be a critical design issue for mmWave cellular systems.

Spectrum shortage in the conventional microwave bands coupled with the need for multi-gigabit per-cell capacities has motivated the use of the mmWave bands in future 5G networks. In this paper 3D ray tracing is used to derive key statistical parameters for the mmWave channel. These include path loss, LoS probability, K-factor, RMS delay spread and angular spreads. The mmWave parameters are directly compared to the microwave channel present in LTE-A. Cellular deployment in both bands is then simulated for a 500 m by 500 m regqasion of central Bristol. The impact of different antenna geometries is also studied. Due to the trade-offs between the number of required BSs and the cost of network deployment, we propose a converged mmWave/microwave network for improved user experience.Index Terms-MmWave, LTE-A, ray tracing, 5G, beamforming.

No abstract

This paper presents a practical demonstration of the benefits of limited feedback beamforming for mmWave applications. Different beamforming strategies are presented, analysed and compared. Compressive sensing (CS) is investigated to accurately estimate the sparse Power Angular Profile (PAP) of the mmWave channel. This information is then used to estimate the antenna array weights required to steer the beam to maximise received SNR. This scheme is particularly attractive for outdoor mmWave networks where large antenna arrays are necessary to mitigate high observed pathloss. Codebook based beamforming strategies are also analysed. Using a practical mmWave antenna away, measurements are reported for a controlled multipath channel that compare the performance of the proposed beamforming strategies. In terms of received SNR CS based beamforming was shown to outperform more traditional codebook based solutions by to 5dB.