A Unified Model for the Design and Analysis of Spatially-Correlated Load-Aware HetNets

• Published in: IEEE transactions on communications Vol. 62; no. 11; pp. 1 - 16
• Main Authors: Shojaeifard, Arman, Hamdi, Khairi Ashour, Alsusa, Emad, So, Daniel K. C., Jie Tang
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2014; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: aggregate network interference; Aggregates; Analytical models; cell selection strategies; Channels; Computer simulation; Design engineering; Downlink heterogeneous networks; Ergodic processes; Fading; Interference; Load modeling; load-awareness; Mathematical models; Monte Carlo methods; Monte- Carlo simulations; Networks; point processes; Shadow mapping; spatial-correlations; stochastic geometry; Stochastic processes; spatial-correlations; point processes; Monte-Carlo simulations; cell selection strategies; stochastic geometry; aggregate network interference; Downlink heterogeneous networks; load-awareness
• ISSN: 0090-6778; 1558-0857
• DOI: 10.1109/TCOMM.2014.2361758

We develop a unified framework for the performance analysis of arbitrary-loaded downlink heterogeneous networks (HetNets) in which interfering sources are inherently spatially-correlated. Considering a randomly-deployed multi-tier cellular network comprised of a diverse set of large-and small-cells, we incorporate the notion of load-awareness and spatial-correlations in characterizing the activities of base stations (BSs) using binary decision variables. A stochastic geometry-based approach is accordingly employed to systematically develop a bounded expression of ergodic rate with different cellular association and load-balancing strategies. Employing the proposed unified framework hence allows for relaxation of several major limitations in the existing state-of-the-art models, in particular the always-transmitting-BSs, uncorrelated interferers, and Rayleigh fading assumptions. We elaborate on the usefulness of adopting this methodology by providing detailed analysis of the aggregate network interference generated by interdependent load-proportional sources over Nakagami-m fading interfering channels. The analytical formulations are validated through Monte-Carlo (MC) simulations for various scenarios and system settings of interest. We observe that the heavily-adopted fully-loaded model as well as the more recent interference-thinning-based approximations are significantly limited in capturing the actual performance curve. The proposed bounded load-aware model and MC trials reveal several important trends and design guidelines for the practical deployment of HetNets.