TL;DR
This paper introduces a novel non-OFDM based Non-Orthogonal Multiple Access (NOMA) scheme for Visible Light Communication (VLC) that simplifies implementation and improves spectral efficiency compared to existing OFDM-based NOMA approaches. The proposed method offers flexible decoding options for Multiple Access Channels (MAC) and efficient decoding for Broadcast Channels (BC), demonstrating competitive performance with reduced computational complexity.
What Problem Does This Paper Address?
Visible Light Communication (VLC) offers a promising alternative for high-speed wireless data transmission using everyday LED lighting. However, integrating Non-Orthogonal Multiple Access (NOMA) – a technique that allows multiple users to share the same frequency band – into VLC presents unique challenges. Traditional NOMA schemes, often based on Orthogonal Frequency Division Multiplexing (OFDM), require complex signal processing steps like Hermitian symmetry and inverse Fast Fourier Transforms (IFFTs) to convert complex signals into the non-negative real values required for LED intensity modulation. This not only increases computational complexity but also reduces spectral efficiency. Furthermore, the non-negative constraint of LED intensity modulation makes direct application of radio-frequency NOMA techniques difficult. This paper addresses the need for a more efficient and simpler NOMA scheme for VLC that overcomes these limitations.
Methods
The authors propose a generalized non-OFDM based NOMA scheme that avoids the complexities of OFDM. Instead of relying on IFFTs and Hermitian symmetry, their approach directly assigns power allocation coefficients to each transmitter (Tx) or user, tailored to their desired spectral efficiency. These coefficients form the constellation points, eliminating the need for complex signal transformations.
For the Multiple Access Channel (MAC) scenario, where multiple Txs communicate with a single receiver, the paper evaluates three decoding strategies:
- Successive Interference Cancellation (SIC): A standard technique where users are decoded sequentially, removing interference from previously decoded users.
- Joint Maximum Likelihood (JML) Decoding: A more computationally intensive but potentially more accurate method that considers all possible signal combinations.
- M JML + (L-M) SIC Decoding: A hybrid approach that uses JML for a subset of users and SIC for the rest, offering a trade-off between performance and complexity.
For the Broadcast Channel (BC) scenario, where a single Tx communicates with multiple users, the paper focuses on SIC-based decoding.
Datasets
Not discussed in the paper. The evaluation relies on simulated scenarios with parameters defined in Table II.
Key Insights and Ideas
- Non-OFDM NOMA for VLC is Feasible and Beneficial: The core idea is that by avoiding OFDM, the scheme significantly simplifies the system and improves spectral efficiency, directly addressing the limitations of existing VLC NOMA approaches.
- Flexible Decoding for MAC: The paper highlights that for MAC, JML decoding offers superior Bit Error Rate (BER) performance compared to SIC, albeit at a higher computational cost. The hybrid M JML + (L-M) SIC decoding provides a tunable knob to balance this trade-off.
- Computational Efficiency is Key: A major takeaway is that the proposed non-OFDM NOMA scheme is computationally much less demanding than OFDM-based NOMA and even OMA in certain scenarios, making it more practical for real-world VLC deployments.
- Trade-offs in BC: For BC, the SIC-based approach is computationally efficient, though it involves a trade-off with BER performance.
Key Results
- MAC Performance: For the MAC scenario, the proposed scheme with JML decoding demonstrates better BER performance than state-of-the-art Orthogonal Multiple Access (OMA) with comparable or lower computational complexity. The hybrid decoding strategy allows for optimization based on specific requirements.
- BC Efficiency: In the BC scenario, the proposed scheme is computationally efficient, offering a practical solution with a manageable BER trade-off.
- Complexity Reduction: The paper quantifies the computational complexity, showing that the proposed non-OFDM NOMA is significantly less complex than DCO-OFDM NOMA and OMA, especially for higher spectral efficiencies. For instance, Table III shows a dramatic reduction in ML computations for BC compared to DCO-OFDM NOMA.
Limitations and Future Work
- BER vs. Complexity Trade-off: The paper acknowledges that while JML decoding offers better BER, it comes with higher computational complexity. The hybrid M JML + (L-M) SIC decoding is presented as a way to manage this, but finding the optimal M for specific scenarios is an ongoing consideration.
- BC BER Performance: While computationally efficient, the BER performance for the BC scenario using SIC is noted to have a trade-off.
- Future Work: The paper implicitly suggests further exploration into optimizing the M JML + (L-M) SIC decoding strategy for MAC and investigating more advanced decoding techniques for BC to further improve BER performance while maintaining low complexity. The paper also mentions that for identical channel gains in BC, NGDPA is not applicable, suggesting that further research into handling such scenarios might be beneficial.
Summary updated