Performance Analysis of an LAA and WiFi Coexistence System Using the LAA Category-4 LBT Procedure With GAP

LTE LAA listen-before-talk (LBT) procedures have been considered a promising solution to reasonable coexistence of LTE with other access technologies in unlicensed spectrum. An issue that arises with the LAA LBT procedures is the so-called GAP period, which may largely degrade the performance of a coexistence system. There are two methods to deal with the GAP period. Method 1 is to fill in the channel during a GAP period by using a reservation signal, while Method 2 takes no action but just listens to the channel during a GAP period. Most existing analytical models for the LAA LBT procedures are inaccurate because they do not take into account the GAP period. This paper studies the modeling of the LAA category-4 LBT procedure in an LAA and WiFi coexistence system and investigate the impact of the GAP period on the system performance. A 2-D Markov chain is built to model the category-4 LBT procedure of an LAA eNodeB, considering the effect of the GAP period. Based the Markov model, analytical models are further derived to investigate the throughput and packet delay performance of the coexistence system with Method 1 and Method 2, respectively. Numerical results show that with Method 1, the performance of an LAA eNodeB is similar to that of a WiFi AP in terms of the transmission probability and packet delay. With Method 2, the performance of an LAA eNodeB is worse than that of a WiFi AP.