An Equivalent Model for Handover Probability Analysis of IRS-Aided Networks

Thanks to the reflecting beamforming gain, Intelligent Reflecting Surface (IRS) is expected to increase coverage range and reduce unnecessary handovers (HOs). However, $O({{N^{2}}})$ gains from $N$ elements and double path-loss change HO decision thus IRS cascaded channels need to be considered instead of the direct base station (BS)-user equipment (UE) channel, which makes existing Euclidean distance-based HO models not applicable. This paper proposes a uniform standard of equivalent UE-BS distance for deciding HOs with IRS cascaded channels, where the reference signal from the serving BS is reflecting beamformed while the reference signals from the neighboring BSs are randomly scattered due to the IRS being occupied. Specifically, the geometric point set of possible BS locations satisfying the HO condition is revised via the equivalent UE-BS distance. In addition, the IRS reselection affects the cascaded channel gain and HO decision, so the distribution of UE-IRS distance is modified with the correlation of UE location in unit time. The compact expression of HO probability in IRS-aided networks is obtained and the effect of the degree of IRS distributed deployment on HO is studied. The results show that there exists an optimal distributed deployment factor that minimizes the HO probability, which decreases by 59.6% compared to networks without IRSs when the density of BS is $10^{3}/\rm {km^{2}}$, the IRS serving distance is $20\rm {m}$ and the number of IRS elements per cell is 100.

I. Introduction

With the aid of reflecting beamforming, Intelligent Reflective Surface (IRS) is regarded as a promising technology to provide seamless service [1]. Since handovers (HOs) result in signaling overhead, it is desirable to minimize HOs to avoid service degradation in this new paradigm. HO is formulated as an optimization problem with IRS in [2]. However, existing works have not proposed analytical models to study the impact of IRS reflecting beamforming on HO, so cannot analyze the optimal IRS configuration and mine gains in HO performance.