Traffic Load Balancing Based on User Data Rate Estimation in Heterogeneous Cellular Networks

The traffic load in heterogeneous cellular networks (HCNs) is often unevenly distributed, which results in unbalanced resource allocation among users and underutilized access points. In this paper, we focus on the user association problem in HCNs, which refers to the problem of assigning mobile users to diverse cells. We extend our previous research on the handover scheme for traffic load balance based on user data rate estimation in realistic HCNs. We formulate an independent interference analytical model and derive the closed-form distribution of the user SINR upper bound. The analytical results are verified very close to the numerical ones. Based on the obtained solution of the distribution, we further estimate user data rates under proportional fairness scheduling for our traffic load balancing scheme. The performance is evaluated by simulations which adopt real site locations in Los Angeles. The numerical results of our balancing scheme indicate considerable improvement of data rate in comparison to the conventional maximum biased-received-power association scheme, especially the data rates of the low-rate users. The impact of user distribution in hotspots on the balancing results is also discussed.