Spectral-Domain Spreading via Hadamard Transform for Robust Downlink Non-Orthogonal Multiple Access

This paper proposes Hadamard-NOMA, a robust downlink non-orthogonal multiple access scheme that leverages the Hadamard Transform at the source level to mitigate fading and channel state information imperfections, achieving significant bit error rate improvements over existing T-NOMA and Usman-NOMA schemes.

The Gist

Here is an explanation of the paper, translated from technical jargon into a story you can visualize.

The Big Problem: The "Noisy Dinner Party"

Imagine a wireless network (like your Wi-Fi or 5G) is a dinner party where one host (the Base Station) is trying to talk to multiple guests (the Users) at the same time.

In the old way of doing things (called T-NOMA), the host tries to talk to everyone simultaneously by shouting different messages at different volumes:

• The Guest in the Corner (Far User): They are far away and can't hear well. The host shouts their message very loudly so they can hear it.
• The Guest at the Table (Near User): They are close and have great hearing. The host whispers their message quietly.

The Catch: Because everyone is listening at the same time, the loud message drowns out the quiet one. To fix this, the "Near Guest" has to use a special trick called SIC (Successive Interference Cancellation). They have to:

1. Listen to the loud message first.
2. Decode it perfectly.
3. Mentally "subtract" or cancel it out of the noise.
4. Then finally hear their own quiet message.

The Flaw: If the "Near Guest" makes even a tiny mistake decoding the loud message (maybe the room is echoey, or the host's voice wavers), that mistake ruins everything. The subtraction goes wrong, and the quiet message becomes garbled. This is what happens when the network has "imperfect Channel State Information" (CSI)—basically, the host doesn't know exactly how clear the air is.

The New Solution: The "Hadamard Magic Trick"

The authors of this paper propose a new system called H-NOMA. Instead of just shouting louder or quieter, they use a mathematical magic trick called the Hadamard Transform (HT).

Think of the Hadamard Transform as a secret decoder ring or a shuffling machine.

1. Before the Shout: Before the host even opens their mouth to speak, they take the messages for all the guests and run them through this "shuffling machine."
2. Spreading the Signal: Instead of sending "Message A" and "Message B" as distinct, separate blocks, the machine mixes them together. It spreads the information of every guest across every part of the signal.
   • Analogy: Imagine instead of handing Guest A a red ball and Guest B a blue ball, you mix red and blue paint into a purple liquid and pour it into cups for everyone. If you spill a little bit of the cup, you don't lose the whole color; you just lose a tiny bit of the mix, and you can still figure out the original colors later.

Why This Changes Everything

By using this "shuffling" technique before the signal is sent, the system becomes incredibly tough against errors.

• The "Safety Net" Effect: In the old system, if the "Near Guest" messed up the subtraction, the "Far Guest" was doomed. In the new H-NOMA system, because the data is spread out and mixed, the receiver doesn't need to be perfect to get the message. Even if part of the signal is distorted by fading (echoes) or bad weather, the receiver can still reconstruct the original message because the information is redundant and spread out.
• The Results:
  • For the "Far Guest": They get a massive boost. The paper says they get a 15 dB improvement. In plain English, this is like going from straining to hear a whisper to hearing a conversation clearly across a noisy room.
  • For the "Near Guest": They also get a huge boost (10 dB). They no longer have to worry as much about making a perfect subtraction; the system is forgiving.
  • Image Quality: The authors tested this by sending pictures (like photos of a baboon or a girl). With the old system, the pictures were blurry and pixelated (low quality). With the new H-NOMA system, the pictures came through crisp and clear, looking almost like the original.

The Bottom Line

This paper introduces a way to make wireless networks more robust.

• Old Way (T-NOMA): Like a fragile house of cards. If one card (one piece of data) slips, the whole tower falls.
• New Way (H-NOMA): Like a woven basket. If you pull on one strand, the whole thing holds together because the strands are interwoven.

By applying this "woven" math (Hadamard Transform) before sending the data, the network can handle bad connections, imperfect equipment, and crowded environments much better. This means faster internet, fewer dropped calls, and clearer video calls for everyone, even in the most challenging conditions.

Technical Summary

Here is a detailed technical summary of the paper "Spectral-Domain Spreading via Hadamard Transform for Robust Downlink Non-Orthogonal Multiple Access."

1. Problem Statement

Non-Orthogonal Multiple Access (NOMA) is a key technology for future wireless networks (5G/6G) due to its ability to allow multiple users to share the same time-frequency resources, thereby enhancing spectral efficiency. However, traditional NOMA (T-NOMA) faces significant challenges:

• Sensitivity to Channel State Information (CSI): Performance degrades significantly under imperfect CSI, which is common in practical dynamic environments.
• Successive Interference Cancellation (SIC) Errors: The reliance on SIC makes the system vulnerable to error propagation. If a user with a stronger channel fails to decode and cancel the signal of a weaker user, it causes cascading errors for subsequent users.
• Hardware and Fading Impairments: T-NOMA struggles with multi-path fading and hardware imperfections, particularly for users far from the base station (Far users) who rely on high power allocation but suffer from poor channel conditions.
• Limitations of Existing Solutions: Previous attempts to improve robustness, such as applying Hadamard Transforms (HT) after modulation or using Multiple Description Coding (MDC) at the symbol level, have introduced unnecessary complexity or failed to fully exploit correlation properties.

2. Methodology: Hadamard-NOMA (H-NOMA)

The authors propose a novel framework called H-NOMA, which integrates the Hadamard Transform (HT) at the source level (prior to modulation) rather than after.

• System Model:

  • Downlink Transmission: A Base Station (BS) serves $K$ single-antenna users.
  • Source Processing: Instead of directly modulating user data $d$, the input data vector is first processed through an HT of order $N$ using a Sylvester-Hadamard matrix ($H_N$).
  • Transformation: The data vector $d$ is transformed into $w = H_N d$. The resulting vector is then affine transformed and mapped to a higher-order constellation (e.g., QAM) to generate the transmitted symbols.
  • Superposition Coding: The BS combines the modulated signals of all users with specific power allocation coefficients ($\alpha_k$), where users with weaker channels receive higher power.
  • Receiver Processing: Users employ SIC. Crucially, in H-NOMA, the receiver utilizes the correlation properties of the transformed domain. The estimated transformed vector $\hat{w}$ is recovered, and the original data is retrieved via the Inverse Hadamard Transform ($\hat{d} = H_N^{-1} \hat{w}$).

• Key Mechanism:

  • By applying HT before modulation, the system spreads the user data across multiple "descriptions." This aligns with Multiple Description Coding (MDC) theory.
  • Even if interference or noise corrupts specific parts of the signal, the correlation introduced by the HT allows the receiver to reconstruct the original data more accurately, effectively mitigating the impact of imperfect SIC and CSI errors.

3. Key Contributions

1. Novel Framework: Proposes the first H-NOMA scheme that applies the Hadamard Transform at the source level (pre-modulation), ensuring physical meaning preservation and theoretical soundness compared to post-modulation approaches.
2. Theoretical Analysis: Provides rigorous mathematical derivations for the Bit Error Rate (BER) under both perfect and imperfect SIC conditions. The analysis demonstrates that H-NOMA effectively cancels noise interference (reducing noise impact by a factor of two for near users) and mitigates near-user interference for far users.
3. Robustness Validation: Extensive Monte Carlo simulations validate the system under Rayleigh and Nakagami-$m$ fading channels, proving resilience against channel impairments and hardware errors.
4. Practical Application: Demonstrates the feasibility of H-NOMA for real-world applications like image transmission, showing significant improvements in reconstruction quality.

4. Key Results

The paper compares H-NOMA against Traditional NOMA (T-NOMA) and a variant where HT is applied post-modulation (Usman-NOMA).

• BER Performance (Two-User Scenario):

  • Near User: Achieves a 15 dB gain at a BER of $10^{-2}$ compared to T-NOMA.
  • Far User: Achieves a 10 dB gain at a BER of $10^{-1}$ compared to T-NOMA.
  • Comparison with Usman-NOMA: The proposed method provides a 15 dB improvement for the Far user at BER $10^{-1}$, proving that pre-modulation application is superior to post-modulation.
  • Imperfect SIC: In a two-user scenario with imperfect SIC, User 1 (Far) requires an SNR at least 14 dB lower than User 2 to achieve a BER of $10^{-3}$, highlighting the system's ability to handle residual interference.

• Four-User Scenario:

  • H-NOMA shows performance improvements for near-area users (User 2, 3, 4) and one far-area user (User 2) compared to T-NOMA.
  • Note: For the furthest user (User 1) in the 4-user scenario, T-NOMA showed slightly better BER performance in specific configurations, but the overall system robustness of H-NOMA against fading remains superior.

• Image Transmission (Real-Time Application):

  • Simulations using 512x512 grayscale images showed H-NOMA significantly outperforms T-NOMA in Peak Signal-to-Noise Ratio (PSNR).
  • Far User (User 1): Image quality improved by >6 dB.
  • Near User (User 2): Image quality improved by >17 dB.

5. Significance and Conclusion

The paper establishes that applying the Hadamard Transform at the source level is a highly effective strategy for enhancing the reliability of NOMA systems.

• Resilience: H-NOMA significantly reduces the sensitivity to CSI errors and SIC failures, which are major bottlenecks in current NOMA implementations.
• Efficiency: It improves spectral efficiency and throughput without requiring complex iterative decoding or excessive computational overhead compared to other robust schemes.
• Future Outlook: The results suggest that H-NOMA is a promising candidate for next-generation (6G) wireless networks, particularly for scenarios requiring high reliability, massive connectivity, and robust performance in dynamic, fading environments. The ability to maintain high-quality data transmission (e.g., images) under adverse conditions makes it suitable for diverse real-world applications beyond standard telecommunications.