BURA Community:

EEG-GANet: Simulating Electroencephalogram Data to Address Sample Imbalance in P300 Speller

2026-04-22T00:00:00Z

Title: EEG-GANet: Simulating Electroencephalogram Data to Address Sample Imbalance in P300 Speller Authors: Lou, Xicheng; Li, Xinwei; Meng, Hongying; Li, Peiyang; Zhong, Lisha; Li, Zhangyong Abstract: The P300 speller is a brain−computer interface (BCI) system that enables character selection using event-related potentials (ERPs). However, the inherent imbalance between target and non-target stimuli biases deep learning models. Reducing non-target samples to mitigate this issue risks underutilizing valuable data. We propose EEG-GANet, a generative adversarial network leveraging domain transformation to augment target samples and mitigate class imbalance in P300 speller datasets. Additionally, we develop EEG-DBNet-V2, a compact dual-branch network that efficiently extracts temporal and spectral features from electroencephalography (EEG) signals, serving both as a classification model and as the discriminator within EEG-GANet. Extensive experiments with ten-fold cross-validation on three public datasets demonstrated EEG-DBNet-V2’s superior classification accuracy compared to state-of-the-art models, achieving this with significantly fewer parameters. Integrating EEG-GANet’s augmented data via fine-tuning further improved classification performance. EEG-GANet effectively addresses sample imbalance by generating physiologically accurate augmented data, substantially enhancing EEG-DBNet-V2’s classification capability. Our study introduces a targeted generative adversarial network (GAN)-based EEG data augmentation framework, enabling balanced model training without discarding valuable non-target samples. The integration of EEG-GANet and EEG-DBNet-V2 offers a lightweight yet robust solution, advancing the practical deployment of EEG-based BCIs under data-constrained conditions. The source code is publicly available at https://github.com/xicheng105/EEG-GANet.

Scale-Aware Gated Routing Distillation for Lightweight Remote Sensing Object Counting

2026-08-31T00:00:00Z

Title: Scale-Aware Gated Routing Distillation for Lightweight Remote Sensing Object Counting Authors: Zhou, Zhiyi; Zheng, Junjing; Gao, Yihang; Meng, Hongying Abstract: Remote sensing object counting remains challenging for lightweight models due to large scale variation, dense object distribution, and complex background interference. Existing lightweight counting methods usually rely on static multi-scale fusion, which limits adaptive receptive-field selection under heterogeneous scenes. To address this issue, this paper proposes Scale-Aware Gated Routing Distillation (SAGRD), a lightweight framework that reformulates multi-scale feature fusion as an adaptive routing problem. At its core, a scale-aware gated routing pyramid generates dynamic branch weights from three complementary cues—global scale descriptors, branch-response feedback, and spatial routing signals—to adaptively reweight dilated receptive-field responses before weighted aggregation, enabling more flexible scale-aware representation under limited model capacity. In addition, a routing-oriented distillation strategy is introduced to transfer not only density estimation knowledge but also the teacher’s scale-selection distributions to the lightweight student, allowing the compact model to inherit structured receptive-field adaptation strategies. Extensive experiments on multiple remote sensing object counting datasets demonstrate that the proposed framework achieves a favorable accuracy-efficiency trade-off. On the RSOC Building benchmark, the student model achieves an MAE/RMSE of 4.712/6.367 using only 1.395 M parameters and 1.29 G FLOPs, closely approaching the teacher performance. Description: Data Availability Statement: The datasets used in this study are publicly available benchmark datasets, as described and cited in the experimental section of the manuscript. No new private dataset was generated in this study.

An Ultrawideband and Low-Profile Phased Array Design for Satellite-assisted Internet of Things Vehicular Communication

2026-05-25T00:00:00Z

Title: An Ultrawideband and Low-Profile Phased Array Design for Satellite-assisted Internet of Things Vehicular Communication Authors: Zhao, Liangliang; Zhu, Hailiang; Li, Aidong; Wen, Lehu; Wang, Shi-Tong; Wang, Kai-Cheng; Zhao, Huiling; Hu, Chufeng Abstract: This article presents a low-profile dual-polarized phased array antenna featuring an improved feed structure and a multiresonant wide-angle impedance matching (WAIM) layer for satellite-assisted Internet-of-Things (SATA-IoT) vehicular communications across the L-, S-, C-, and X-bands. First, an improved feed structure based on a double-sided multi-notch balun (DMB) is developed to provide stable impedance transformation and tailored field distribution, thereby enhancing impedance matching in the low- and mid-frequency bands. Second, a planar nonuniform metasurface (PNM)-based multiresonant WAIM layer is introduced. By properly engineering its effective permittivity and permeability, two desired resonances together with inductive coupling are achieved, enabling wide-angle scanning performance in the high-frequency band. These two key techniques jointly allow the proposed array to realize ultrawide bandwidth, wide-angle scanning capability, and a low-profile configuration. To validate the design, a 5×5 prototype was fabricated and measured. The measured results show a 9:1 impedance bandwidth from 1.4 to 12.6 GHz. In particular, both array configurations exhibit an active VSWR below 3 at broadside, and maintain active VSWR below 3.3 for scan angles up to ±70° across the entire operating band. The overall profile is only 0.075 λlow. These results demonstrate that the proposed array is a promising candidate for SATA-IoT and related vehicular IoT communication systems.

An Ultrawideband Magnetoelectric Dipole Antenna With Gain Improvement and Its Application in Array Design for Sub-6G Intelligent IoT Communication

2026-08-19T00:00:00Z

Title: An Ultrawideband Magnetoelectric Dipole Antenna With Gain Improvement and Its Application in Array Design for Sub-6G Intelligent IoT Communication Authors: Zhao, Liangliang; Zhu, Hailiang; Li, Aidong; Wen, Lehu; Wang, Shi-Tong; Wang, Kai-Cheng; Zhao, Huiling; Hu, Chufeng Abstract: An ultrawideband magnetoelectric (ME) dipole antenna with stable gain is proposed for sub-6G Internet of Things (IoT) communication systems. While several structures can effectively achieve a wide impedance bandwidth, a critical limitation is that an excessively wide beamwidth often results in gain deterioration over the corresponding frequency range. To address this issue, vertically placed metal plate (VPMP) are introduced to generate out of phase currents and narrow the beamwidth, while folded metal branch (FMB) provide additional vertical radiation, thereby improving gain stability. Consequently, the proposed ME dipole antenna exhibits a 108% impedance bandwidth (1.2–4 GHz) with a measured gain of 7.2 ± 1.9 dBi across the band. Furthermore, based on the proposed antenna element, vertical shorting post (VSP) are employed to introduce out of phase currents for mutual coupling suppression, enabling a 100% operating bandwidth (1.26–3.85 GHz) for the array. In addition, a one-dimensional array prototype is fabricated to validate the proposed design strategy. It demonstrates a scanning range of up to ±45° across 1.26–2 GHz and ±30° across 2–3.85 GHz. Covering most of the L- and S-bands while maintaining stable radiation characteristics, the proposed antenna is a promising candidate for multi-service IoT communication systems including smart home devices, agricultural monitoring, and industrial manufacturing.