Fairness and Rate Coverage of Symmetric Transmission over Heterogeneous Cellular Networks under Diverse Coupling and Association Criteria

Abstract

Human communication, received through physical senses, is often a symmetric service. On other occasions, only certain feedback is needed for multisensory communication but even in these cases a minimum throughput is needed in the uplink for signaling purposes. In this paper, the joint uplink (UL) and downlink (DL) binary rate that can be achieved by realistic spectral efficiencies as signaled by channel quality indicators in LTE have been studied. Specifically, symmetric binary rate gathers both uplink and downlink capacities while forcing them to be reached simultaneously in order to carry the bidirectional information flow. We assume a two-tier heterogeneous cellular network with macro and pico Base Stations (BS) and User Equipments (UE) uniformly distributed. Downlink transmission to a UE could be associated to its nearest base station or to that BS from which the average received power is maximum. In the later case, certain bias could be added to the power received from the small cells in order to offload users from macro to pico BSs. Realistic limits to users simultaneously active in a cell are included in the evaluation. Regarding uplink, it has been considered two options. As current networks, uplink association can be coupled to that in downlink. However, uplink association might also be decoupled from that of DL and UL transmission carried out over that path with minimum losses. This possibility would reduce interference after fractional power control (FPC). Results for rate coverage, spectral efficiency, and binary rate for marginal uplink, marginal downlink, and joint uplink/downlink has been obtained from simulations as a function of diverse system parameters. Besides averaging, fairness measurements have also been studied. Results show that taking into account limits on the maximum spectral efficiency and on the number of simultaneous active users within a cell strongly modifies results on joint binary rate. Although spectral efficiency of DL is maximum for average criteria, its DL binary rate is not necessarily the highest as many macro users get inactive. Moreover, for realistic 3GPP path loss models, decoupled DL/UL association does not improve results enough to be worthwhile.