Effect of Channel Fading and Time-to-Trigger Duration on Handover Performance in UAV Networks

Abstract

Owning to the agility and the transform between line-of-sight (LoS) and non-LoS (NLoS) links, the time-varying channel in unmanned aerial vehicle (UAV) networks leads to the frequent handover (HO). Hence, existing HO models only considering path-loss are unable to analyze the unnecessary HO caused by fading. In this letter, a path-loss-plus-fading model with HO parameters is considered for HO in UAV networks and the impact of channel fading on HO performance is theoretically characterized, where HO states during time-to-trigger (TTT) duration is discretized to analyze the HO trigger probability. By modeling the fading as Nakagami-$m$ distribution, a discrete time Markov chain is utilized where the user's states along the trajectory is discretized into multiple HO states by coherence time. Based on the discretization, HO failure (HOF) and ping-pong (PP) probabilities are derived by analyzing the HO state probabilities during the TTT duration. The results show that the transform between LoS and NLoS links increases HOF probability, and there is a tradeoff between HOF and PP probabilities when configuring TTT duration.