Local Delay Analysis and Improvement for User-SBS Correlated Deployments With Random DTX

Abstract

Stochastic geometry has emerged as a well-established analysis tool for dense wireless networks. However, the commonly assumed homogeneous Poisson point process (PPP) model, where users are distributed in the space uniformly, may not be accurate enough, since the users are more likely to be clustered at densely populated sites. Driven by capacity needs, small base stations (SBSs) are deployed at the user hotspots, which results in a user-SBS correlated deployment. In order to capture this correlation, user locations are modeled as Poisson cluster process (PCP) with SBSs serving as the cluster centers. In this paper, by using the discontinuous transmission (DTX) mode, the local delay has been analyzed. We carried out the analysis under two classic PCP models—Thomas cluster process and Matérn cluster process. Moreover, it is found that under PCP the local delay is more dependent on base station (BS) density and transmit power than under the conventional PPP model. Finally, we found that the local delay increases as the cluster size increases, and it converges to the PPP result when the cluster size approaches the infinity.