Abstract
This paper studies a downlink system that combines orthogonal-time-frequency-space (OTFS) modulation and sparse code multiple access (SCMA) to support massive connectivity in high-mobility environments. We propose a cross-domain receiver for the considered OTFS-SCMA system which efficiently carries out OTFS symbol estimation and SCMA decoding in a joint manner. This is done by iteratively passing the extrinsic information between the time domain and the delay-Doppler (DD) domain via the corresponding unitary transformation to ensure the principal orthogonality of errors from each domain. We show that the proposed OTFS-SCMA detection algorithm exists at a fixed point in the state evolution when it converges. To further enhance the error performance of the proposed OTFS-SCMA system, we investigate the cooperation between downlink users to exploit the diversity gains and develop a distributed cooperative detection (DCD) algorithm with the aid of belief consensus. Our numerical results demonstrate the effectiveness and convergence of the proposed algorithm and show an increased spectral efficiency compared to the conventional OTFS transmission.
Introduction
As wireless communications evolve toward 6G, the industry faces the dual challenge of supporting massive connectivity for billions of IoT devices while maintaining reliable links in high-mobility scenarios, such as high-speed rail or unmanned aerial vehicles (UAVs). Standard 5G modulation, Orthogonal Frequency Division Multiplexing (OFDM), often fails in these environments (at speeds up to 1000 km/h) because the Doppler effect destroys subcarrier orthogonality, leading to significant inter-carrier interference.
To address this, Orthogonal Time Frequency Space (OTFS) modulation has emerged as a promising candidate, multiplexing symbols in the Delay-Doppler (DD) domain where the channel remains relatively stable. However, OTFS is inherently orthogonal. To achieve the massive connectivity required for 6G, it must be paired with a Non-Orthogonal Multiple Access (NOMA) scheme. This paper proposes the integration of OTFS with Sparse Code Multiple Access (SCMA) to create a robust downlink system.
System Architecture
The proposed OTFS-SCMA scheme utilizes the strengths of both technologies. OTFS provides resilience against high Doppler shifts, while SCMA allows multiple users to share the same time-frequency resources through sparse codebooks, increasing spectral efficiency.
The challenge in such a system lies in the receiver complexity. Fractional Doppler shifts can make the channel response dense in the DD domain, complicating equalization. The authors address this by proposing a cross-domain approach.
Proposed Methods
1. Cross-Domain Iterative Receiver
The receiver avoids the high complexity of DD-domain-only processing by "hopping" between domains. It performs Linear Minimum Mean Square Error (L-MMSE) equalization in the time domain, where the channel remains sparse and manageable. The processed data is then transformed into the DD domain for decoding via the Message Passing Algorithm (MPA). By iteratively passing extrinsic information (means and variances) through a unitary transformation, the receiver ensures that errors from each domain remain uncorrelated, following the principles of Orthogonal Approximate Message Passing (OAMP).
2. Distributed Cooperative Detection (DCD)
To further improve reliability, the authors propose a cooperative layer. In a downlink scenario, users receive signals that contain information for the entire group. By broadcasting partial local information to neighbors and using a belief consensus algorithm, users can iteratively update their estimates. This exploits spatial diversity, significantly boosting signal quality for all participating nodes.
Performance and Insights
The system was evaluated using Monte Carlo simulations with a 4-path wireless channel model featuring fractional Doppler shifts. Key findings include:
- Reliability Gains: The joint cross-domain and DCD algorithm (Scheme 1) achieved a 5dB gain at a Bit Error Rate (BER) of $10^{-4}$ compared to separate detection schemes.
- Convergence: Theoretical State Evolution (SE) predictions matched simulation results perfectly, with the receiver typically converging in fewer than five iterations.
- Spectral Efficiency: The scheme successfully supported an overloading factor of 150% (e.g., 6 users on 4 resources), demonstrating its capability for massive connectivity.
- Codebook Optimization: A notable insight is that OTFS transforms fading channels into quasi-static AWGN-like channels in the DD domain. Consequently, SCMA codebooks optimized for high Minimum Euclidean Distance (MED) perform better than those designed for Rayleigh fading (e.g., Huawei codebooks).
Limitations and Future Work
While the cross-domain approach is efficient, the L-MMSE step still requires a matrix inversion of $O((MN)^3)$. Future work could explore log-linear complexity approximations. Additionally, this study assumes perfect Channel State Information (CSI), which is difficult to obtain in high-mobility channels. Finally, while this paper focuses on the downlink, the uplink presents unique challenges in multi-user channel estimation.
Summary updated