Stochastic Modeling and Performance Analysis of Intelligent Reflecting Surface assisted Wireless Communication Systems

Abstract

In a wireless terrain, presence of various obstacles leads to compliance of Non-line- of-sight (NLOS) norms between a transmitter (base station, BS) and a receiver (user), which renders severe degradation in the Quality of service (QoS) of the wireless links and thereby limits the anticipated performance. Presence of IRS panels stimulates a supplementary wireless channel that enhances reliable end-to-end connectivity between a transmitter and a receiver, especially when a direct path (line-of-sight, LOS) experi- ences either severe signal degradation or completely lost in shorter epochs. Using Intel- ligent Reflecting Surface (IRS), performance of a wireless network can be improved by smartly re-configuring the passive reflecting elements which are embedded on a planar surface. In this thesis, as geographical terrain experience ubiquitous presence of obstacles, using estimate pf performance measures such as Achievable rate, Outage probability, Coverage probability and Ergodic capacity, it is validated that in such scenarios IRS- assisted wireless communication system performance is significantly enhanced with re- spect to a conventional wireless system. A mathematical framework for Time-invariant (TI) and Time-variant (TV) channel models is proposed. A comprehensive analysis is performed in terms of Achievable rate, Outage probability, Ergodic capacity and Cov- erage probability estimate. During these performance measures, distance between a BS and a receiver is considered as a primary variable while regulating parameters such as; transmit power, number of Embedded Reflecting Elements (EREs), separation distance between EREs and receiver speed for an Urban Micro-street canyon (UMi-SC) and an Urban Micro-open square (UMi-OS) wireless terrain. In addition, terrain equipped with multiple IRS panels are studied to achieve a better network coverage enhancement. The thesis investigates the network outage performance of multiple IRS-assisted wireless terrain. Different case studies are presented in which a receiver experiences a v direct (LOS) link and/or IRS reflected links, and all these links are characterized using kappa-mu (κ−μ) shadowed fading. Here, an IRS panel resulting in maximum instanta- neous signal-to-noise ratio (SNR) is selected to participate in establishing an end-to-end wireless link. An exact analysis comprising κ − μ shadowed fading impact on outage probability and ergodic capacity of a wireless link and subsequently, approximating the κ − μ shadowed fading channel by Nakagami-m fading model is presented. An ana- lytical framework is presented to estimate the outage probability and ergodic capacity of an end-to-end wireless link for different wireless terrain equipped with multiple IRS panels and experiencing different degree of fading along with varying data/vehicular- traffic statistics. While incorporating geographical attributes driven spatial aspects and the distinct vehicular traffic originated in distinct time-intervals exact κ − μ shadowed fading model is considered. For all these scenarios, numerical results demonstrate the precision achieved in approximation accuracy. Urban terrain that characterizes with high rise buildings and huge size sign boards along the road can be equipped with IRS panels at appropriate locations to enable Vir- tual Line-Of-Sight (VLOS) links between BSs and users through these IRS panels. Due to surrounding spatial attributes, temporal characteristics of wireless terrain and users’ behavioural aspects, a single stochastic model doesn’t suffice to approximate the ter- rain’s characteristics and the associated underlying propagation mechanism. In this thesis, to accommodate diversified spatial and temporal characteristics of terrain in a reasonable manner, usage of Poisson Point Process (PPP) and Poisson Cluster Process (PCP) is proposed to model the users’ distribution and IRS panels deployment. Cover- age probability is estimated to validate the proposed stochastic models, while regulating the cluster size. Furthermore, the efficacy of IRSs equipped wireless terrain is demon- strated in terms of enhanced coverage probability for a broad ranging SINR threshold. An analytical framework is presented to estimate the coverage probability of a wireless link for varying fading conditions in multiuser environment with multiple IRS panels. Mapping of real world vehicular traffic scenarios with PPP and PCP is established and accordingly proposed models of PPP and PCP are considered to address location aspect of IRS panels and BSs.