Traffic Management for Heterogeneous Networks With Opportunistic Unlicensed Spectrum Sharing

Abstract

This paper studies how to maximize the per-user-based throughput in an $M$ -tier heterogeneous wireless network (HetNet) by optimally managing traffic flows among the access points (APs) in the HetNet. The APs in the first $M-1$ tiers can use the licensed spectrum at the same time whereas they share the unlicensed spectrum with the APs in the $M$ th tier by the proposed opportunistic CSMA/CA protocol. We characterize the statistical property of the cell load and channel access probability of each AP using a general AP association scheme. For an AP in each tier, the tight bounds on its mean spectrum efficiencies in the licensed and unlicensed spectra are derived in a low-complexity form for general random channel gain and AP association weight models and they can give some insights on how channel gains, AP association weights and void AP probabilities affect the mean spectrum efficiencies. We define the per-user link throughput and per-user network throughput based on the derived the mean spectrum efficiencies and maximize them by proposing the decentralized and centralized traffic management schemes for the APs in the first $M-1$ tiers under the constraint that the per-user link throughput of the tier- $M$ APs must be above some minimum required value. Finally, a numerical example of coexisting LTE and Wi-Fi networks is provided to validate our derived results and findings.