This letter studies multiuser (MU) downlink polarforming with equal-amplitude, phase-only polarization-reconfigurable antennas (PRAs), each driven by a single radio-frequency (RF) chain. At each base station (BS) antenna, one polarization phase is shared by all transmitted streams, while digital precoding remains user-specific. Exploiting this structure, we show that each transmit phase reduces to an exact scalar unit-modulus problem and each receive phase to an exact one-dimensional problem, for both continuous and finite-resolution control. These updates are integrated into a single weighted minimum mean-square error (WMMSE) loop, avoiding the additional penalty-dual-decomposition procedure required by more general polarforming formulations. We prove monotonic convergence of the weighted sum rate (WSR) value. Extensive simulations under cross-polarization discrimination (XPD), imperfect polarized channel state information (CSI), phase quantization, and solver comparisons demonstrate the resulting rate–complexity tradeoff against fixed-polarization, as well as amplitude-and-phase designs.
Automatic modulation recognition (AMR) is a key technology in cognitive radio and non-cooperative communication systems. Existing AMR methods suffer from limited recognition performance under label-scarce scenarios. To address this issue, this letter proposes a Multi-view Twin Contrastive Learning Framework (MTCLF). MTCLF maps the signals into four representation views: IQ, frequency, wavelet, and constellation. A shared Vision Transformer encoder is designed for cross-view feature extraction. The framework constructs cross-view positive pairs from the four signal representation views and extends SimSiam to multi-view contrastive learning. Experiments on benchmark datasets demonstrate that MTCLF outperforms existing methods under both linear evaluation and few-shot fine-tuning scenarios.
This paper investigates the joint antenna position and beamforming design for movable antenna (MA)-enabled cooperative secure integrated sensing and communication (ISAC) systems. In the considered setup, MAs enable flexible antenna position adaptation at the base station (BS), which cooperates with a relay to serve legitimate users in the presence of a potential eavesdropper. To fully exploit the spatial degrees of freedom (DoFs) offered by MAs, we formulate a secrecy rate maximization problem by jointly optimizing antenna positions, transmit beamforming, and artificial noise (AN), subject to constraints on transmit power, antenna deployment, and sensing accuracy. The resulting problem is highly non-convex due to the intricate coupling among discrete antenna position variables and continuous transmission variables. To tackle this challenge, we develop a low-complexity algorithm based on an alternating optimization (AO) framework, where the continuous variables are optimized via convex programming, while a greedy search strategy is employed to efficiently update antenna positions. Numerical results demonstrate that the proposed design significantly enhances secrecy performance compared to conventional fixed-position array systems, highlighting the benefits of adaptive spatial configuration enabled by MAs.
This letter proposes a partial-band constructive interference (PB-CI) scheme for multi-user visible light communications. The total bandwidth is divided into a shared sub-band for CI and exclusive sub-bands for orthogonal transmission. Then, a joint optimization problem is formulated to maximize the spectral efficiency (SE) subject to users’ quality-of-service requirements and limited linear dynamic-range constraints of light-emitting diodes. To solve this mixed-integer non-convex problem, a two-stage method is developed in which waveform design is solved first, followed by signal scaling to fully exploit the dynamic range. Simulation results demonstrate that the PB-CI scheme effectively balances bandwidth utilization and dynamic-range consumption, achieving up to 37.7% higher SE over several benchmarks, with an energy efficiency improvement of up to 50.3%.
To meet the stringent information freshness requirements of beyond fifth-generation (B5G) and sixth-generation (6G) Internet of Things (IoT) scenarios, we propose a proactive hybrid automatic repeat request (HARQ)-based cell-free massive multiple-input multiple-output (CF-mMIMO) system to minimize the average peak age of information (AoI). By leveraging the derived signal-to-interference-plus-noise ratio (SINR) and incorporating the effective transmission success probability, a closed-form expression for the average peak AoI is derived. To enhance information freshness, we jointly optimize the transmit power and the maximum number of retransmissions, formulate a mixed-integer nonconvex optimization problem, and employ a genetic algorithm (GA) to efficiently solve the problem. Finally, simulation results verify the accuracy of the theoretical analysis and demonstrate the effectiveness of the proposed algorithm.
