TL;DR
This paper introduces a more realistic model for analyzing the coverage and energy efficiency of dense cellular networks by using the Matérn Hard-Core Process (MHCP) instead of the traditional Poisson Point Process (PPP). The MHCP model better captures the spatial exclusion between base stations, leading to improved performance predictions and demonstrating significantly higher energy efficiency compared to PPP-based models.
What Problem Does This Paper Address?
Modern cellular networks are becoming increasingly dense with the deployment of various base station (BS) types (macro, pico, femto) to meet rising data demands. This density, however, leads to significant inter-cell interference, impacting both coverage and energy efficiency. While the Poisson Point Process (PPP) is widely used for its analytical tractability, it assumes independent BS locations, which doesn't accurately reflect real-world deployment strategies. Interference management techniques often enforce spatial separation between BSs, creating a "repulsion" effect that PPP fails to capture. This paper addresses this gap by proposing and analyzing a two-tier heterogeneous cellular network (HCN) model that incorporates this spatial exclusion using the Matérn Hard-Core Process (MHCP).
Methods
The authors extend the Approximate Signal-to-Interference Ratio (SIR) analysis based on PPP (ASAPPP) method to a two-tier network modeled by two independent MHCPs (MHCP-MHCP). This "Each Tier ASAPPP" (ET-ASAPPP) approach approximates each MHCP tier as a PPP tier with the same density but adjusts the SIR threshold using an empirically derived SIR gain. This gain quantifies the improvement in SIR performance due to the MHCP's spatial exclusion compared to a PPP. The paper first derives the asymptotic coverage probability using this ET-ASAPPP method and then formulates the network's energy efficiency (EE) based on the derived coverage and a linear power consumption model. The SIR gain for the MHCP model is obtained through simulation and data fitting.
Datasets
Not discussed in the paper. The study relies on simulated data and parameters defined in Table 1 for analysis.
Key Insights and Ideas
- MHCP is a better fit for dense HCNs: The spatial exclusion inherent in the Matérn Hard-Core Process (MHCP) more accurately models real-world BS deployments, where interference management necessitates a minimum distance between BSs.
- Spatial exclusion boosts performance: The repulsion between BSs in the MHCP model significantly reduces interference compared to the random placement assumed by PPP. This leads to higher coverage probabilities and, consequently, better energy efficiency.
- ET-ASAPPP bridges the gap: The proposed "Each Tier ASAPPP" (ET-ASAPPP) method effectively extends existing PPP-based analysis techniques to MHCP-modeled networks, providing a tractable way to derive performance metrics.
- EE is convex with respect to transmit power: The study reveals that energy efficiency exhibits a convex relationship with the transmit power of pico base stations (PBSs) and macro base stations (MBSs), indicating the existence of optimal power levels for maximizing EE.
Key Results
- The coverage probability of the MHCP-MHCP network is shown to be significantly higher than that of a comparable PPP-PPP network. Theoretical results closely match simulation outcomes, validating the ET-ASAPPP method.
- The energy efficiency of the MHCP-MHCP network is approximately 1.2x10^6 to 1.3x10^6 bits/Joule higher than that of the PPP-PPP network under similar parameters.
- The paper demonstrates that even with fixed BS densities, the MHCP-MHCP network's EE surpasses that of a PPP-PPP network using optimized BS densities.
- The study identifies optimal transmit power levels for both MBSs and PBSs that maximize energy efficiency, suggesting joint optimization as a promising avenue for further EE improvements.
Limitations and Future Work
- Tractability vs. Accuracy: The paper acknowledges that the ET-ASAPPP method is an approximation. While it provides tractable analytical results, there might be a gap between theoretical approximations and simulation values, especially for small SIR thresholds where the equivalent SIR threshold is not significantly greater than 1.
- Simplified assumptions: The analysis assumes universal frequency reuse and equal bandwidth allocation per BS to MUEs. It also doesn't explicitly consider MUE QoS requirements in the coverage analysis.
- Future Work: The authors suggest that joint optimization of MBS and PBS transmit powers, along with BS densities, can further improve energy efficiency. Exploring more complex power consumption models and incorporating QoS constraints could also be valuable future research directions.
Summary updated