Coverage Performance Analysis of Reconfigurable Intelligent Surface-aided Millimeter Wave Network with Blockage Effect

In order to solve spectrum resource shortage and satisfy immense wireless data traffic demands, millimeter wave (mmWave) frequency with large available bandwidth has been proposed for wireless communication in 5G and beyond 5G. However, mmWave communications are susceptible to blockages. This characteristic limits the network performance. Meanwhile, reconfigurable intelligent surface (RIS) has been proposed to improve the propagation environment and extend the network coverage. Unlike traditional wireless technologies that improve transmission quality from transceivers, RISs enhance network performance by adjusting the propagation environment. One of the promising applications of RISs is to provide indirect line-of-sight (LoS) paths when the direct LoS path between transceivers does not exist. This application makes RIS particularly useful in mmWave communications. With effective RIS deployment, the mmWave RIS-aided network performance can be enhanced significantly. However, most existing works have analyzed RIS-aided network performance without exploiting the flexibility of RIS deployment and/or considering blockage effect, which leaves huge research gaps in RIS-aided networks. To fill the gaps, this thesis develops RIS-aided mmWave network models considering blockage effect under the stochastic geometry framework. Three scenarios, i.e., indoor, outdoor and outdoor-to-indoor (O2I) RIS-aided networks, are investigated.

Firstly, LoS propagation is hard to be guaranteed in indoor environments since blockages are densely distributed. Deploying RISs to assist mmWave transmission is a promising way to overcome this challenge. In the first paper, we propose an indoor mmWave RIS-aided network model capturing the characteristics of indoor environments. With a given base station (BS) density, whether deploying RISs or increasing BS density to further enhance the network coverage is more cost-effective is investigated. We present a coverage calculation algorithm which can be adapted for different indoor layouts. Then, we jointly analyze the network cost and coverage probability. Our results indicate that deploying RISs with an appropriate number of BSs is more cost-effective for achieving an adequate coverage probability than increasing BSs only.

Secondly, for a given total number of passive elements, whether fewer large-scale RISs or more small-scale RISs should be deployed has yet to be investigated in the presence of the blockage effect. In the second paper, we model and analyze a 3D outdoor mmWave RIS-aided network considering both building blockages and human-body blockages. Based on the proposed model, the analytical upper and lower bounds of the coverage probability are derived. Meanwhile, the closed-form coverage probability when RISs are much closer to the UE than the BS is derived. In terms of coverage enhancement, we reveal that sparsely deployed large-scale RISs outperform densely deployed small-scale RISs in scenarios of sparse blockages and/or long transmission distances, while densely deployed small-scale RISs win in scenarios of dense blockages and/or short transmission distances.

Finally, building envelope (the exterior wall of a building) makes outdoor mmWave BS difficult to communicate with indoor UE. Transmissive RISs with passive elements have been proposed to refract the signal when the transmitter and receiver are on the different side of the RIS. Similar to reflective RISs, the passive elements of a transmissive RIS can implement phase shifts and adjust the amplitude of the incident signals. By deploying transmissive RISs on the building envelope, it is feasible to implement RIS-aided O2I mmWave networks. In the third paper, we develop a 3D RIS-aided O2I mmWave network model with random indoor blockages. Based on the model, a closed-form coverage probability approximation considering blockage spatial correlation is derived, and multiple-RIS deployment strategies are discussed. For a given total number of RIS passive elements, the impact of blockage density, the number and locations of RISs on the coverage probability is analyzed.

All the analytical results have been validated by Monte Carlo simulation. The observations from the result analysis provide guidelines for the future deployment of RIS-aided mmWave networks.