BER Analysis and Optimization for Continuous RIS-Enabled NOMA

This paper proposes and validates a joint optimization framework for uplink PD-NOMA systems assisted by continuous reconfigurable intelligent surfaces (CRIS) that derives an accurate bit error rate (BER) expression under spatially correlated fading to eliminate BER floors and outperform conventional OMA and non-optimized NOMA schemes.

The Gist

Imagine you are trying to host a massive, chaotic dinner party where everyone wants to talk to the host (the Base Station) at the exact same time. In the old days (Orthogonal Multiple Access, or OMA), you'd have to give each guest a different time slot to speak. Guest A talks for 5 minutes, then Guest B, then Guest C. It's orderly, but slow.

NOMA (Non-Orthogonal Multiple Access) is like letting everyone talk at once. The host has to listen to a jumbled mess of voices and figure out who said what. This is much faster and uses less energy, but it's risky: if the voices are too similar, the host gets confused, and the conversation breaks down (errors happen).

Now, enter the RIS (Reconfigurable Intelligent Surface). Think of this as a giant, magical wall of mirrors placed between the guests and the host. In the past, these mirrors were made of distinct, separate tiles (Discrete RIS). You could tilt a few tiles here and there to help, but it was a bit clunky.

This paper introduces a CRIS (Continuous RIS). Imagine the wall isn't made of tiles at all, but is a single, seamless sheet of liquid metal or a holographic surface. You can bend and shape the entire surface perfectly to guide every single voice clearly to the host.

The Problem: The "Noise Floor"

Even with this magical wall, when everyone talks at once, there's a problem. The host can usually hear the loudest guest clearly, but the quieter guests get drowned out by the "echo" of the louder ones. In technical terms, this is called Inter-User Interference (IUI).

No matter how much you turn up the volume (transmit power), the conversation quality hits a "ceiling" or a floor. You can't get better than a certain point because the interference is just too messy. This is the "BER Floor" (Bit Error Rate Floor) mentioned in the paper. It's like trying to hear a whisper in a room where someone is shouting; turning up the whisper doesn't help if the shouting gets louder too.

The Solution: A Smart Dance of Power and Space

The authors of this paper came up with a two-part strategy to fix this, which they call Joint Optimization. Think of it as a choreographer directing the dinner party:

1. Power Allocation (Who shouts louder?): The system decides exactly how loud each guest should speak. The guests who are naturally closer to the host or have a better path don't need to shout as hard. The distant ones get a little boost, but not so much that they drown out the others.
2. RIS Partitioning (Who gets which part of the mirror?): This is the clever part. Since the CRIS is a continuous wall, the system slices it up. It gives the "stronger" guests (those who are decoded first) a huge section of the mirror to focus their voice. It gives the "weaker" guests a smaller, but perfectly tuned, section.

By combining these two moves, the system ensures that the "shouting" guests don't accidentally drown out the "whispering" ones. The interference is managed so well that the "noise floor" disappears. The conversation becomes crystal clear, even at very high volumes.

The Analogy: The Orchestra Conductor

Imagine an orchestra where every musician is playing the same instrument at the same time.

• Old Way (OMA): They take turns playing solos.
• NOMA without help: They all play at once, and it sounds like a cacophony.
• NOMA with Discrete RIS: They have a few sound-dampening panels, but it's not perfect.
• This Paper's Solution (CRIS + Optimization): The conductor (the algorithm) has a magical, seamless sound-shaping wall. The conductor tells the loud violins to play slightly softer and assigns them a specific, large section of the wall to focus their sound. Simultaneously, the quiet flutes get a smaller, highly focused section of the wall to amplify their tone just enough to be heard without being drowned out.

Why This Matters

The paper proves mathematically and through simulations that this new method:

1. Eliminates the "Error Floor": You can keep improving the quality of the connection indefinitely by adding more power or a bigger wall, without hitting a dead end.
2. Beats the Competition: It works better than the old "take turns" method (OMA) and better than using the clunky, tiled mirrors (Discrete RIS).
3. Handles Real-World Messiness: It accounts for the fact that in the real world, signals bounce off things in correlated ways (spatial correlation), making the solution robust for actual future 6G networks.

In a nutshell: This paper teaches us how to use a "smart, seamless mirror wall" to let many people talk to a computer at once without them interrupting each other, ensuring that even the quietest voice can be heard perfectly, no matter how loud the room gets.

Technical Summary

Here is a detailed technical summary of the paper "BER Analysis and Optimization for Continuous RIS-Enabled NOMA".

1. Problem Statement

The paper addresses the limitations of current mobile networks in handling data-intensive applications (e.g., AR, AI) by exploring the integration of Non-Orthogonal Multiple Access (NOMA) with Continuous Reconfigurable Intelligent Surfaces (CRIS) in the Uplink (UL) direction.

• The Gap: Existing research on continuous aperture surfaces (like CRIS or Holographic RIS) and NOMA has primarily focused on Downlink (DL) scenarios or discrete RIS (DRIS) systems. There is a lack of analytical work on UL NOMA systems aided by CRIS, particularly under realistic spatially correlated fading conditions.
• The Challenge: In UL NOMA, users share the same time-frequency resource, leading to Inter-User Interference (IUI). Successive Interference Cancellation (SIC) is used to decode signals, but it often results in Bit Error Rate (BER) floors at high Signal-to-Noise Ratios (SNR) due to residual interference. Furthermore, the infinite element density of a CRIS introduces complex spatial correlations that make traditional discrete channel models inaccurate.

2. Methodology

The authors propose a comprehensive framework involving system modeling, statistical channel analysis, and joint optimization.

A. System Model

• Scenario: A UL system with $K$ single-antenna User Equipments (UEs) communicating with a single-antenna Base Station (BS) via a passive rectangular CRIS.
• CRIS Partitioning: The CRIS is divided into $K$ vertical partitions ($P_1$ to $P_K$), where each partition $P_k$ is dedicated to user $U_k$.
• Channel Assumptions:
  • Direct UE-to-BS links are blocked (standard RIS assumption).
  • Channels follow spatially correlated Rayleigh fading.
  • Phase shifts are optimized to align the cascaded channel phases for coherent combining.
• Signal Model: The received signal is a superposition of the desired signal (from the optimized partition) and residual interference (from other partitions).

B. Statistical Analysis & BER Derivation

To derive the BER, the authors analyze the effective channel components:

1. Effective Channel Decomposition: The effective channel for user $k$ is split into a deterministic dominant component ($\gamma_{kk}$) and a random interference component ($\gamma_{ki}$).
2. Moment Calculation: The authors derive the first and second moments (mean and variance) of these components under spatial correlation.
   • $\gamma_{kk}$ is approximated using a Gamma distribution.
   • The real part of the interference $\Re(\gamma_{ki})$ is approximated using a Gaussian distribution.
3. Characteristic Function (CF): Using these approximations, the authors derive the CF of the effective channel. This is crucial because the exact distribution is intractable.
4. BER Expression: An analytical expression for the average BER is derived using the CF and the Q-function, allowing for performance evaluation without exhaustive simulations.

C. Joint Optimization Framework

The core contribution is a joint optimization scheme to minimize the average BER of all users by tuning two variables:

1. Uplink Power Allocation (PA): Adjusting transmit power ($P_k$) for each user.
2. CRIS Partitioning: Dynamically adjusting the width ($W_k$) of the CRIS section assigned to each user.

• Objective: Minimize $\sum BER_{U_k}$ subject to power constraints and total CRIS width constraints.
• Algorithm: The problem is transformed into the log-log domain for stability and solved using a Lagrangian multiplier method combined with gradient descent. This handles the non-linear constraints and ensures convergence.

3. Key Contributions

1. First UL CRIS-NOMA Analysis: This is the first study to analyze the symbol-level performance (BER) of UL NOMA systems aided by CRIS under spatially correlated fading.
2. Analytical BER Derivation: The paper provides a novel, tractable analytical expression for BER by approximating the cascaded channel statistics using Gamma and Gaussian distributions, validated by the derived Characteristic Function.
3. Joint Optimization Strategy: A novel framework that jointly optimizes Power Allocation and CRIS Area Partitioning. Unlike previous works that optimized only one parameter, this approach simultaneously manages interference via power control and resource allocation via surface partitioning.
4. BER Floor Elimination: The proposed scheme successfully eliminates the BER floors typically associated with SIC-based UL NOMA, a significant improvement over non-optimized schemes.

4. Simulation Results

The analytical results were validated against Monte Carlo simulations for 2-user and 3-user systems.

• Accuracy: The analytical BER curves matched simulation results perfectly, confirming the validity of the statistical approximations.
• Performance Gains:
  • Joint Optimization (JO): The JO scheme (optimizing both Power and Partitioning) completely eliminated BER floors down to $10^{-7}$.
  • Comparison with Baselines:
    • vs. Non-Optimized (NO): Non-optimized NOMA suffered from high error floors (>0.1 for strong users).
    • vs. OMA: The optimized CRIS-NOMA outperformed Orthogonal Multiple Access (OMA) in all scenarios, allowing simultaneous access to the full resource block without excessive interference.
    • vs. DRIS: The CRIS system significantly outperformed a Discrete RIS (DRIS) with the same surface area, demonstrating the benefits of continuous aperture control.
• Insight on Partitioning: The optimization naturally assigned larger CRIS partitions to users with stronger channels (who are decoded first in SIC) to suppress interference, while managing power to prevent weaker users from suffering due to reduced channel hardening.

5. Significance

• Theoretical Advancement: It bridges the gap between continuous aperture theory and practical NOMA implementation, providing the first statistical tools to analyze UL performance in this domain.
• Practical Impact: The results demonstrate that combining CRIS with NOMA is a viable path for 6G networks to achieve massive connectivity and high spectral efficiency.
• Solving the Error Floor: By proving that joint optimization can remove the inherent error floors of UL NOMA, the paper offers a concrete solution to a long-standing reliability issue in power-domain NOMA systems.
• Resource Efficiency: The study shows that dynamic partitioning of the RIS surface is as critical as power control, offering a new dimension for network resource management.

In conclusion, the paper establishes that Continuous RIS-enabled UL NOMA, when coupled with joint power and area optimization, offers superior reliability and spectral efficiency compared to traditional OMA and discrete RIS-NOMA systems, effectively overcoming the limitations of interference and error floors.