Statistical Model of Downlink Power Consumption in Cellular CDMA Networks

Technical Summary: Statistical Model of Downlink Power Consumption in Cellular CDMA Networks

Problem Statement
In Code Division Multiple Access (CDMA) cellular networks, transmission power is the primary radio resource, making power consumption a critical metric for network capacity and resource allocation. While the downlink is often the bottleneck due to asymmetric data flows, developing accurate analytical models for downlink power consumption has proven difficult. Existing literature typically relies on Monte Carlo simulations, particularly when Soft Handoff (SHO) is considered, or resorts to simplifying assumptions that neglect the complexity of SHO connection modes. A major analytical obstacle is the presence of sums of log-normally distributed random variables (representing shadowing losses) in interference calculations, which prevents the derivation of closed-form expressions. Consequently, there is a need for a theoretical statistical model that can provide closed-form solutions for power consumption moments while realistically accounting for cell selection and handoff algorithms.

Methodology
The authors propose a theoretical statistical model based on a standard large-scale path loss model, decomposing losses into a distance-dependent term and a log-normal shadowing component. The model analyzes a hexagonal cellular network with a mobile station (MS) in a central cell surrounded by two tiers of neighbors.

The core methodology involves the following steps:

1. Formulation of Power Consumption: The downlink power required for a single connection is derived for three scenarios: Hard Handoff (HHO), 2-way Soft Handoff (SHO), and 3-way SHO. These derivations incorporate intra-cell and inter-cell interference, orthogonality factors, and balanced power allocation among participating Base Stations (BSs).
2. Integration of Handoff Logic: The model explicitly integrates conditions imposed by cell selection and handoff schemes. It defines specific inequalities involving shadowing variables ($\xi$) and thresholds (hysteresis $c_{st}$ and SHO threshold $s_{ht}$) to determine the Active Set (AS) size (1, 2, or 3).
3. Statistical Approximation: To overcome the complexity of summing log-normal random variables, the authors employ a Taylor series expansion of the power consumption expressions around the expected values of the interference terms. By truncating the series after the second order, they derive closed-form expressions for the first (mean) and second (variance) moments of the power consumption fraction ($\beta$).
4. Moment Calculation: The first and second moments are calculated by integrating the joint probability density function of the independent shadowing variables over the specific regions defined by the handoff conditions. The integration utilizes a specific function $A(x,y)$ derived from the error function to handle the log-normal limits analytically.

Key Contributions

• Closed-Form Expressions: The primary contribution is the derivation of closed-form analytical expressions for the first and second moments of downlink power consumption. This avoids the computational load of Monte Carlo simulations while maintaining accuracy.
• Comprehensive Handoff Modeling: Unlike previous works that often ignore soft handoff or simplify its impact, this model rigorously accounts for 2-way and 3-way SHO scenarios, including the specific constraints of the Active Set size and handoff thresholds.
• Taylor Series Approximation: The paper introduces a specific approximation technique using Taylor series expansion to linearize the complex sums of log-normal variables, enabling the calculation of statistical moments without resorting to numerical integration for every scenario.
• Cross-Layer Framework: The model bridges physical layer parameters (propagation, shadowing) with Radio Resource Management (RRM) settings (handoff thresholds, cell selection), facilitating a cross-layer approach to network optimization.

Numerical Results and Discussion
The proposed model was validated against independent numerical simulations involving 100,000 random shadowing samples. The results demonstrated that the theoretical estimations for the expected value ($E[\beta]$) and standard deviation ($\sigma_\beta$) converge closely with simulation data, confirming the validity of the Taylor series approximation.

Key findings from the numerical analysis include:

• Propagation Impact: In Hard Handoff scenarios, power consumption generally increases as the MS approaches the cell border. However, in hostile propagation environments (high path loss factor $\alpha$ and high shadowing $\sigma$), the model predicts a reduction in power consumption near the border because the MS is more likely to hand off to a neighbor cell rather than struggle with a degraded link.
• Benefits of Soft Handoff: The inclusion of Soft Handoff (AS=2 or AS=3) significantly reduces both the mean power consumption and its variance compared to Hard Handoff (AS=1). This benefit is most pronounced when the MS is near the cell border or in areas with good propagation conditions where a neighbor BS can effectively share the load.
• Threshold Sensitivity: Tighter cell selection criteria and appropriate SHO thresholds encourage beneficial handoffs, reducing the system's power burden and variance.
• Variance Reduction: The model shows that increasing the Active Set size (e.g., moving from AS=1 to AS=3) leads to a more location-insensitive behavior and a significant reduction in the standard deviation of power consumption.

Significance and Claims
The paper claims that the proposed theoretical statistical model offers a balanced approach between oversimplified assumptions and the inaccuracies of potential approximations used in prior studies. By providing closed-form expressions, the model facilitates efficient performance evaluation and optimization of CDMA networks without the heavy computational cost of simulations.

The authors position this work as a foundational step toward a cross-layer design approach, where physical layer propagation characteristics can be directly integrated with RRM algorithms. They modestly note that while the current work focuses on link-level statistics and balanced power allocation, the framework can be extended to network-level evaluations, unbalanced power schemes, and the development of optimized soft handoff algorithms that utilize these estimations. The paper does not claim to solve all network optimization problems but provides a necessary analytical tool for understanding the statistical behavior of downlink power under realistic handoff conditions.