MSMingjun Sun

Mingjun Sun

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Mingjun Sun has a total of 12 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • GNN-Aided Over-the-Air Reciprocity Calibration for Cell-Free Massive MIMO Systems

    by Mingjun Sun, Xidong Mu, Shaochuan Wu, Chongjun Ouyang, Yuanwei Liu

    Over-the-air (OTA) reciprocity calibration (RC) is essential in cell-free massive multiple-input multiple-output (CF-mMIMO) systems to enable the downlink coherent beamforming from uplink channel estimates. Conventional least-squares (LS) calibration typically treats all pilot links equally and relies on a singular value decomposition-based solution, which can be sensitive to low signal-to-noise ratio (SNR) observations and incurs high computational cost. This letter proposes a graph neural network (GNN)-aided OTA RC framework, where the distributed access-point network is modeled as a directed graph and calibration coefficients are learned from bidirectional pilot measurements. In particular, the proposed dual-head attention mechanism effectively captures link importance, thereby mitigating the adverse impact of low-SNR links. Moreover, the forward inference only involves simple matrix multiplications, yielding low computational complexity. Simulation results demonstrate that the proposed GNN consistently outperforms LS across all considered system configurations. Moreover, it generalizes well across different data distributions, network scales, and topologies, highlighting its robustness and practical applicability in dynamic CF-mMIMO systems.

  • Robust Beamforming for Pinching-Antenna Systems-Based Multi-User Communications

    by Mingjun Sun, Xidong Mu, Chongjun Ouyang, Guanyu Shang, Yuanwei Liu

    Pinching-antenna systems (PASS)-enabled communications are a promising paradigm. However, obtaining accurate channel state information (CSI) is challenging due to the limited pilots, model mismatch, and other hardware impairments. In this paper, we study the robust beamforming design for PASS-based multi-user communication under the CSI uncertainty. In particular, we model the CSI error using a complex-ball uncertainty set and study the worst-case beamforming design to minimize the transmit power for satisfying users’ minimum achievable rate requirements. To address this problem, the S-procedure is first applied to convert the resulting semi-infinite formulation into a tractable finite set of constraints. Then, an alternating optimization framework is developed to design the baseband beamforming and pinching beamforming. The base-band beamforming subproblem is handled via the semidefinite programming, while a penalty dual decomposition-based method is developed to address the non-convex pinching beamforming subproblem. Finally, the effectiveness of the proposed algorithm is verified through simulations.


  • Robust Beamforming for Pinching-Antenna Systems

    by Mingjun Sun, Chongjun Ouyang, Shaochuan Wu, Yuanwei Liu

    Pinching-antenna system (PASS) mitigates large-scale path loss by enabling flexible placement of pinching antennas (PAs) along the dielectric waveguide. However, most existing studies assume perfect channel state information (CSI), overlooking the impact of channel uncertainty. This paper addresses this gap by proposing a robust beamforming framework for both lossy and lossless waveguides. For baseband beamforming, the lossy case yields an second-order cone programming-based solution, while the lossless case admits a closed-form solution via maximum ratio transmission. The PAs' positions in both cases are optimized through the Gauss-Seidel-based method. Numerical results validate the effectiveness of the proposed algorithm and demonstrate that PASS exhibits superior robustness against channel uncertainty compared with conventional fixed-antenna systems. Notably, its worst-case achievable rate can even exceed the fixed-antenna baseline under perfect CSI.


  • Secure Beamforming for Continuous Aperture Array (CAPA) Systems

    by Mingjun Sun, Chongjun Ouyang, Zhaolin Wang, Shaochuan Wu, Yuanwei Liu

    Continuous aperture array (CAPA) is considered a promising technology for 6G networks, offering the potential to fully exploit spatial degrees of freedom (DoFs) and achieve the theoretical limits of channel capacity. This paper investigates the performance gain of a CAPA-based downlink secure transmission system, where multiple legitimate user terminals (LUTs) coexist with multiple eavesdroppers (Eves). The system’s secrecy performance is evaluated using a weighted secrecy sum-rate (WSSR) under a power constraint. We then propose two solutions for the secure current pattern design. The first solution is a block coordinate descent (BCD) optimization method based on fractional programming (FP), which introduces a continuous-function inversion theory corresponding to matrix inversion in the discrete domain. This approach derives a closed-form expression for the optimal source current pattern. Based on this, it can be found that the optimal current pattern is essentially a linear combination of the channel spatial responses, thus eliminating the need for complex integration operations during the algorithm’s optimization process. The second solution is a heuristic algorithm based on zero-forcing (ZF), which constructs a zero-leakage current pattern using the channel correlation matrix. It further employs a water-filling approach to design an optimal power allocation scheme that maximizes the WSSR. In high signal-to-noise ratio regions, this solution gradually approaches the first solution, ensuring zero leakage while offering lower computational complexity. Simulation results demonstrate that: 1) CAPA-based systems achieve better WSSR compared to discrete multiple-input multiple-output (MIMO) systems. 2) The proposed methods, whether optimization-based or heuristic, provide significant performance improvements over existing state-of-the-art Fourier-based discretization methods, while considerably reducing computational complexity.


  • Multiuser Beamforming for Pinching-Antenna Systems: An Element-Wise Optimization Framework

    by Mingjun Sun, Chongjun Ouyang, Shaochuan Wu, Yuanwei Liu

    The pinching-antenna system (PASS) reconstructs wireless channels through pinching beamforming, i.e., optimizing the activated locations of pinching antennas (PAs) along the waveguide. The aim of this article is to investigate the joint design of baseband beamforming and pinching beamforming. A low-complexity element-wise sequential optimization framework is proposed to address the sum-rate maximization problem in PASS-enabled downlink and uplink channels. i) For the downlink scenario, maximum ratio transmission (MRT), zero-forcing (ZF), and minimum mean square error (MMSE) beamforming schemes are employed as baseband beamformers. For each beamformer, a closed-form expression for the downlink sum-rate is derived as a single-variable function with respect to the pinching beamformer. Based on this, a sequential optimization method is proposed, where the positions of the PAs are updated element-wise using a low-complexity one-dimensional search. ii) For the uplink scenario, signal detection is performed using maximum ratio combining (MRC), ZF, and MMSE combiners. A closed-form sum-rate expression is derived for each linear combiner, and a similar element-wise design is applied to optimize the pinching beamforming. Numerical results are provided to validate the effectiveness of the proposed method and demonstrate that: (i) For all considered linear beamformers, the proposed PASS architecture outperforms conventional fixed-antenna systems in terms of sum-rate performance; (ii) in both downlink and uplink channels, ZF achieves performance close to that of MMSE and significantly outperforms MRT or MRC; and (iii) the proposed element-wise design eliminates the need for alternating updates between the baseband and pinching beamformers, thereby ensuring low computational complexity.


  • A Homogeneous Graph Neural Network for Precoding and Power Allocation in Scalable Wireless Networks

    by Mingjun Sun, Shaochuan Wu, Haojie Wang, Yuanwei Liu, Guoyu Li, Tong Zhang

    Deep learning is widely used in wireless communications but struggles with fixed neural network sizes, which limit their adaptability in environments where the number of users and antennas varies. To overcome this, this paper introduced a generalization strategy for precoding and power allocation in scalable wireless networks. Initially, we employ an innovative approach to abstract the wireless network into a homogeneous graph. This primarily focuses on bypassing the heterogeneous features between transmitter (TX) and user entities to construct a virtual homogeneous graph serving optimization objectives, thereby enabling all nodes in the virtual graph to share the same neural network. This “TX entity” is known as a base station (BS) in cellular networks and an access point (AP) in cell-free networks. Subsequently, we design a universal graph neural network, termed the information-carrying graph neural network (ICGNN), to capture and integrate information from this graph, maintaining permutation invariance. Lastly, using ICGNN as the core algorithm, we tailor the neural network's input and output for specific problem requirements and validate its performance in two scenarios: 1) in cellular networks, we develop a matrix-inverse-free multi-user multi-input multi-output (MU MIMO) precoding scheme using the conjugate gradient (CG) method, adaptable to varying user and antenna numbers; 2) in a cell-free network, facing dynamic variations in the number of users served by APs, the number of APs serving each user, and the number of antennas per AP, we propose a universal power allocation scheme. Simulations demonstrate that the proposed approach not only significantly reduces computational complexity but also achieves, and potentially exceeds, the spectral efficiency (SE) of conventional algorithms.


  • A Matrix-Inverse-Free WMMSE Algorithm to MISO Beamforming Based on Quasi-Newton

    by Mingjun Sun, Zeng Li, Shaochuan Wu, Ruofei Ma, Litong Jiang

    This letter propose a quasi-Newton based weighted minimum mean square error (WMMSE) algorithm without matrix inverse to solve the weighted sum rate (WSR) maximization problem in multi-user multi-input single-output (MU-MISO) beamforming. On one hand, the quasi-Newton method can replace the first-order optimal condition to solve the extremum problem of the convex quadratic function, without involving matrix inverse. One the other hand, compared to projected gradient descent (PGD) approach, it can achieve a faster convergence under the guidance of approximate Hessian matrix and avoid performance loss under the condition of high transmit power. Furthermore, a learning strategy is adopted to replace the linear searching process to obtain the optimal step size that satisfies the Wolfe condition. Simulation results validate that the proposed algorithm can achieve the same performance as WMMSE, but with a reduced computation complexity.