Explicit Modelling and Performance Analysis of Cell Group Selection With Backhaul-Aware Biasing

Abstract

In existing works on performance analysis, cell selection is distance-based, and the relation of backhaul capacity variance and cell selection has not been explicitly modelled. Using stochastic geometry, this letter analyzes the performance of backhaul-aware cell selection by deriving the analytical expression for the outage probability, where high-backhaul-capacity small base stations (SBSs) have larger association biases to offload users from low-backhaul-capacity SBSs. Specifically, the general case of cell group (CG) selection is considered instead of single cell selection, and a CG consists of multiple SBSs with the largest biased reference signal received power, where the bias value is the SBS's backhaul capacity. In addition, the serving SBS changes adaptively which is the one with the largest transmission rate (minimum of access and backhaul rate) inside CG, thus exploiting the diversity of SBSs within CG. The results show that the outage probability is decreased by 30.6% when the CG size equals 5, compared with the distance-based CG formation strategy.