Stochastic Geometry based Handover Probability Analysis in Dense Cellular Networks

Abstract

In the ultra-dense network (UDN), terminals may be exposed to more frequent handover than ever due to smaller cell radius. Employing the handover protocol of the received signal strength with handover hysteresis (RSSH), the ping-pong effect can be significantly mitigated. In this work, we propose a stochastic geometry framework for handover analysis in UDNs and derive the theoretical expression for handover probability under such handover protocol. However, the handover probability becomes tricky to handle because the hysteresis margin makes the user association state strongly correlated, and UE does not any longer associate with the nearest BS consistently. Using the law of total probability, we derive the theoretical expression for handover probability by addressing its conditional probabilistic events of handover (HO) or non-handover $(\overline{\mathrm{H}\mathrm{O}})$ in the former stage and obtain the simplified expression in the low mobility case. Both analytical and simulation results demonstrate the correctness and effectiveness of our analysis and show that higher hysteresis is tolerable for a denser network. Furthermore, the simplified expression for the special case of low mobility is shown to be quite accurate, and thus can be used to capture first-order insights for general cases.