arXiv subjects

Nghia Thinh Nguyen

Publications and source records attributed to Nghia Thinh Nguyen.

4 recordsLinked to original sources

arXiv2026

Physics-Informed Unsupervised Learning for Reliable Beamforming in Cluttered MIMO-ISAC

In this paper, we consider joint beamforming and power allocation for a mobile multiuser MIMO integrated sensing and communication (ISAC) network with clutter, residual self-interference, and outdated channel state information, under per-user rate and sensing signal-to-interference-plus-noise ratio (SINR) constraints that must hold on every realization. The design must be recomputed every block, and distant or weak targets can make the sensing constraint unmeetable. We propose a physics-informed unsupervised learning (PI-UnSL) network that corrects the analytical design, meets the power and norm constraints exactly by construction, and is conditioned on closed-form interference statistics, spanning the whole tradeoff in one pass. We derive a closed-form sensing SINR at the interference-aware combiner, yielding a necessary range-radar-cross-section-power condition and a priority-independent analytic violation floor, so the problem is intermittently infeasible. We then show that a dual targeting below the attainable floor has no fixed point, and calibrate the label-free objective against the measured floor. Numerical results confirm the floor and the bounded calibrated dual, and show that PI-UnSL substantially raises per-realization joint quality of service over the closed-form baseline, delivers more rate than time-division ISAC at matched detection over most of the tradeoff, and exceeds the reliability of an uncalibrated alternating-optimization solver at a fraction of its latency. More broadly, the physical knowledge given to the network, not its architecture, makes such problems learnable without labels: five architectures fail alike from scratch and perform alike when started from the closed-form design and conditioned on its interference statistics.

eess.SP↗

arXiv2026

Probabilistic Denoising-Enhanced ISAC for Stochastic Cluttered Mobile Environments

In this paper, we propose Probabilistic Denoising ISAC (PDISAC), a framework built on a multi-bit slot-partitioned ISAC waveform: by partitioning each maximal-length sequence into alternating pilot and data slots, we embed multiple bits per sequence through symbol-level spreading, multiplying the data rate while every chip retains the deterministic radar code. The added throughput, however, injects data-dependent, non-white sidelobes into the range-Doppler (RD) heatmap that degrade matched-filter (MF) sensing. Rather than modifying the MF receiver, we develop RDPDNet, a lightweight probabilistic denoising network inserted between RD-map formation and constant-false-alarm-rate detection; training it with an adversarial frequency-mixup mechanism, we suppress the data-induced sidelobes and thermal noise without knowledge of the embedded symbols. We further characterize the statistics of the geometry-determined channel. The fundamental performance limits of the design are then analyzed through an analytical lower bound, a semi-analytical bit error rate (BER), and an average capacity that tie the slot allocation and sequence length to the sensing-communication trade-off. Through analytical and numerical results over a realistic urban geometry, we show that RDPDNet absorbs most of the data-embedding sensing penalty and markedly lowers the RMSE at low SNR, while the conventional data-free chain attains the bias-adjusted benchmark at high SNR. Moreover, increasing the slot allocation raises the data rate at the expense of a higher BER, exposing a tunable sensing--communication trade-off.

eess.SP↗

arXiv2025

Temporal Micro-Doppler Spectrogram-based ViT Multiclass Target Classification

In this paper, we propose a new Temporal MDS-Vision Transformer (T-MDS-ViT) for multiclass target classification using millimeter-wave FMCW radar micro-Doppler spectrograms. Specifically, we design a transformer-based architecture that processes stacked range-velocity-angle (RVA) spatiotemporal tensors via patch embeddings and cross-axis attention mechanisms to explicitly model the sequential nature of MDS data across multiple frames. The T-MDS-ViT exploits mobility-aware constraints in its attention layer correspondences to maintain separability under target overlaps and partial occlusions. Next, we apply an explainable mechanism to examine how the attention layers focus on characteristic high-energy regions of the MDS representations and their effect on class-specific kinematic features. We also demonstrate that our proposed framework is superior to existing CNN-based methods in terms of classification accuracy while achieving better data efficiency and real-time deployability.

eess.SP↗

arXiv2025

Generative and Explainable AI for High-Dimensional Channel Estimation

In this paper, we propose a new adversarial training framework to address high-dimensional instantaneous channel estimation in wireless communications. Specifically, we train a generative adversarial network to predict a channel realization in the time-frequency-space domain, in which the generator exploits the third-order moment of the input in its loss function and applies a new reparameterization method for latent distribution learning to minimize the Wasserstein distance between the true and estimated channel distributions. Next, we propose an explainable artificial intelligence mechanism to examine how the critic discriminates the generated channel. We demonstrate that our proposed framework is superior to existing methods in terms of minimizing estimation errors. Additionally, we find that the critic's attention focuses on the high-power portion of the channel's time-frequency representation.

eess.SP↗