implement successive interference cancellation

In low-rank MIMO systems where the number of users exceeds the number of AP antennas and channel observations are partially missing, joint channel estimation and data detection becomes highly challenging due to severe rank deficiency and information scarcity. Method: This paper introduces, for the first time, continuous-time diffusion models to this task, proposing a successive interference cancellation (SIC)-assisted score-based joint estimation framework. It designs an SIC-coupled iterative score-gradient update scheme that unifies generative score matching with Bayesian iterative inference, overcoming fundamental performance bottlenecks of conventional methods under low-rank conditions. Contribution/Results: Experiments demonstrate significant improvements over state-of-the-art baselines in normalized mean square error (NMSE) and symbol error rate (SER), especially at medium-to-low SNRs, while maintaining robustness in full-rank regimes. The work pioneers the synergistic integration of diffusion modeling and SIC for low-rank MIMO joint estimation, achieving both theoretical feasibility and substantial empirical gains.

To address inaccurate channel state information (CSI) estimation and degraded detection performance caused by time-varying channels in fast-fading MIMO-OFDM systems for 5G and beyond, this paper proposes a data-aided joint channel estimation and signal detection method. We first formulate a generic data-aided linear minimum mean square error (LMMSE) framework tailored for iterative joint optimization, and then design a low-complexity surrogate algorithm leveraging time-frequency domain modeling and low-dimensional parametric approximation to achieve a superior trade-off between estimation accuracy and computational cost. Experimental results demonstrate that, across diverse MIMO configurations, pilot lengths, and time-varying channel conditions, the proposed method improves detection accuracy by 15–22% over state-of-the-art basis expansion model (BEM)-based receivers, while reducing computational complexity by a factor of 3.8.

This work addresses the challenge of achieving low-complexity, high-order modulation-compatible MIMO detection with reliable soft-output generation in resource-constrained edge scenarios such as 5G RedCap and IoT. The paper proposes recurSIC, a lightweight learning-based MIMO detection framework inspired by successive interference cancellation (SIC). By integrating a tunable-complexity, multi-path hypothesis tracking mechanism within a single forward pass, recurSIC efficiently produces high-quality soft information. The approach embeds a compact neural network into the SIC architecture, requiring only a minimal number of parameters to simultaneously deliver accurate hard decisions and reliable soft outputs. Experimental results demonstrate that recurSIC achieves near-optimal detection performance under real-world wireless channels while significantly reducing computational overhead, making it well-suited for deployment in edge MIMO receivers.

To address the low channel estimation accuracy and high pilot overhead in multi-RIS-aided MIMO systems, this paper proposes a joint iterative detection, decoding, and channel estimation framework. The method innovatively integrates LDPC-coded pilots with data packet parity bits for channel estimation, enabling mutual enhancement between channel information and code constraints across iterations—thereby significantly improving estimation accuracy while reducing reliance on pilot resources. Unlike conventional approaches, it imposes no sparsity assumption on the channel and is applicable to realistic Sub-6 GHz LOS/NLOS propagation environments. Simulation results demonstrate that, under identical pilot overhead, the proposed scheme achieves substantial reductions in both bit error rate and channel estimation mean-square error compared to state-of-the-art methods, exhibiting superior efficiency and robustness.

In low-Earth-orbit (LEO) satellite-to-cellular direct links, geographically proximate users exhibit highly correlated channels, severely degrading the convergence speed of conventional iterative MIMO detectors. To address this, we propose a cluster-aware two-stage detection framework: first, user clusters are explicitly modeled based on channel spatial correlation, and intra-cluster strong interference is mitigated via small-scale matrix inversion; second, inter-cluster interference is suppressed using an enhanced Gauss–Seidel method combined with symmetric successive over-relaxation (SSOR). This work is the first to explicitly incorporate geographic correlation into the MIMO detection architecture, enabling effective interference decoupling and computational dimensionality reduction. Simulation results demonstrate that the proposed method achieves over 12× faster convergence under ideal channel conditions, and maintains a 9× speedup even in the presence of channel estimation errors—significantly enhancing system real-time performance and robustness.

This work proposes a novel SIC-free rate-splitting multiple access (RSMA) receiver that circumvents the error propagation inherent in conventional RSMA schemes relying on successive interference cancellation (SIC). Instead of performing SIC, the proposed receiver employs joint demapping (JD) to directly evaluate bit vectors over a composite constellation, thereby eliminating the performance bottleneck imposed by SIC decoding. The feasibility of this SIC-free approach is experimentally validated for the first time over-the-air in a two-user MISO prototype system. Results demonstrate that the proposed method significantly enhances system robustness and practicality across a wider operational range, while achieving close alignment with theoretical performance predictions.

This work addresses the recovery of $k$-sparse binary signals of length $n$ under extreme undersampling conditions where the number of measurements $m$ is less than the sparsity level $k$, a regime in which conventional compressive sensing methods fail. The authors propose ISDP-MVSIC, a novel approach that integrates randomized semidefinite programming (SDP) sampling, majority voting (MV), and successive interference cancellation (SIC), enhanced by a residual-driven retry mechanism for staged signal reconstruction. This method achieves, for the first time, exact recovery with high probability even when $m < k$, while offering a tunable trade-off between computational complexity and reconstruction performance. Experimental results demonstrate empirically perfect recovery for $n = 100$ and $144$ across a wide range of sparsity ratios, specifically for $m/k \in [0.4, 5.0]$, at the cost of modestly increased computational overhead.

This study addresses interference suppression and bit error rate (BER) optimization in full-duplex reconfigurable intelligent surface (RIS)-assisted multi-user communications. A QPSK-based multi-user system is established, and a successive interference cancellation (SIC) optimization framework tailored for variable-amplitude RISs is proposed. By jointly designing three SIC ordering strategies with phase-shift optimization algorithms for amplitude-variable reflecting elements, the system performance is validated through Monte Carlo simulations. The results demonstrate that mean square error (MSE)-based ordering achieves optimal performance in small-scale RIS configurations, whereas the proposed optimization method excels in large-scale scenarios. This work reveals the optimal matching relationship between RIS scale and SIC ordering mechanisms, significantly reducing the system BER.

This study addresses the challenges of power allocation and receiver design in superimposed pilot transmission for OFDM-MIMO systems. We propose an iterative error analysis framework to optimize the power ratio between pilots and data, alongside a novel Transformer-based AI-ICED receiver architecture. By integrating iterative channel estimation and detection with deep learning mechanisms, this approach effectively mitigates superimposed interference. Simulation results demonstrate that, compared to conventional non-superimposed pilot schemes, the proposed system significantly enhances spectral efficiency while maintaining robust estimation accuracy. Consequently, this work offers a promising solution for achieving high spectral efficiency in next-generation communication systems by synergizing model-driven signal processing with data-driven intelligence.

This work investigates whether practical quantum receivers can surpass the channel capacity limits imposed by conventional successive interference cancellation (SIC) in multi-access networks. To this end, the authors propose a full-stack quantum receiver architecture that integrates front-end quantum sensing with back-end quantum signal processing. By leveraging ensembles of unentangled qubits and employing quantum-correlation measurements alongside parallel processing of superposition states, the approach efficiently extracts multi-user interference signals. Notably, this method achieves performance beyond the SIC limit without requiring complex entanglement resources, significantly enhancing spectral and detection efficiency—particularly at low signal-to-noise ratios—and thereby transcending the theoretical boundaries of classical multi-access channel capacity.