Wideband signal detection and recognition (WSDR) is considered an effective technical means for monitoring and analyzing spectra. The mainstream technical route involves constructing time–frequency representations for wideband sampled signals and then achieving signal detection and recognition through deep learning-based object detection models. However, existing methods exhibit insufficient attention on the prior information contained in the time–frequency domain and the structural features of signals, leaving ample room for further exploration and optimization. In this paper, we propose a novel model called TFDP-SANet for the WSDR task, which is based on time–frequency distribution priors and structure-aware adaptivity. Initially, considering the horizontal directionality and banded structure characteristics of the signal in the time–frequency representation, we introduce both the Strip Pooling Module (SPM) and Coordinate Attention (CA) mechanism during the feature extraction and fusion stages. These components enable the model to aggregate long-distance dependencies along horizontal and vertical directions, mitigate noise interference outside local windows, and enhance focus on the spatial distributions and shape characteristics of signals. Furthermore, we adopt an adaptive elliptical Gaussian encoding strategy to generate heatmaps, which enhances the adaptability of the effective guidance region for center-point localization to the target shape. During inference, we design a Time–Frequency Clustering Optimizer (TFCO) that leverages prior information to adjust the class of predicted bounding boxes, further improving accuracy. We conduct a series of ablation experiments and comparative experiments on the WidebandSig53 (WBSig53) dataset, and the results demonstrate that our proposed method outperforms existing approaches on most metrics.

In recent years, object detection methods have achieved promising results in the field of wideband signal detection and recognition (WSDR). However, for resource-constrained airborne platforms and degraded electromagnetic environments, existing methods still face challenges such as high computational overhead and insufficient noise robustness. In this article, we propose a lightweight cascade network based on YOLO (LC-YOLO) and apply it to tackle the WSDR challenge under complex electromagnetic environments or resource-constrained scenarios. Specifically, we design a lightweight hybrid attention enhancement module (HAEM), which is constructed by channel attention (CA), pixel attention (PA), and mixed-shape separable convolution groups. It is cascaded at the front of the backbone network to enhance the representation ability of targets in images. To further improve the network's detection performance for targets with extreme aspect ratios, we propose a feature extraction module for optimizing the backbone network, named cross stage partial with strip module (C2SM). This module inherits the gradient shunting advantage of the efficient layer aggregation network (ELAN) and integrates the receptive field characteristics of large strip convolution, significantly improving the detection accuracy of slender signal targets while reducing the number of parameters. Furthermore, we reveal that the coordinate discretization in distribution focused loss (DFL) limits complete boundary regression for slender targets, and propose an enhanced approach integrating coordinate continuous regression with Smooth L1 Loss. Experimental results show that the N, S, and M configurations of LC-YOLO outperform the current mainstream lightweight state-of-the-art (SOTA) object detection methods on the WidebandSig53 dataset (WBSig53). The smallest configuration, LC-YOLO-N, outperforms the baseline model YOLOv8-N by 3.9% in mAP, while maintaining only 2.5 M parameters and an inference speed of 113 FPS.

A photonic-aided wide-band high-precision spectrum analysis system is proposed and demonstrated. The polarization multiplexing is employed to perform concurrent operation of fast frequency-to-time mapping (FTTM)-based frequency measurement (FM) and neural network-assisted signal analysis. Dual-band FTTM is implemented via chirp multiplexing and matched filtering, which enables simultaneous FM covering two 6-GHz bands. A two-step filtering architecture is designed to mitigate the impact of laser frequency drift and achieve a measurement accuracy of ±2.6 MHz with 30 kHz spectral resolution. By utilizing a YOLOv8-based network, high-precision multiparameter estimation to the signal under test (SUT) is realized with a frequency error under 0.9 MHz, a bandwidth error under 2.8 MHz, a time-center error under $1~\mu $ s, a duration error under $2.1~\mu $ s, and a modulation recognition error rate under 0.025. The proposed system combines the advantages of photonic and electrical technologies, which holds great promise for applications in intelligent RF systems for cognitive radio and spectrum management.

In open, noncooperative electromagnetic environments, prior information about target signals, such as their modulation types and symbol rates, is often difficult to obtain. Consequently, both conventional methods and explicitly label-dependent jamming waveform generation methods face substantial limitations in practical applications. To address this issue, we propose a Multiscale Aggregation Generative Adversarial Network with Masked Pre-training (MA-GAN), a communication jamming waveform generation method based on masked self-supervised learning and multiscale feature fusion. We construct a masked waveform modeling pre-training framework for In-phase/Quadrature (I/Q) time-series signals, through which the Masked Waveform Modeling encoder (MWM-encoder) learns hierarchical and multiscale representations from unlabeled data. The generator is guided by multiscale statistical features and combined with a multi-head multiscale discrimination mechanism and physical constraints to improve the structural consistency between the generated jamming waveforms and the target signals. Experimental results show that, under Additive White Gaussian Noise (AWGN), Rayleigh fading and Rician fading channels and across different Signal-to-Noise Ratio (SNR) and Jamming-to-Signal power Ratio (JSR) conditions, MA-GAN achieves effective jamming performance for five modulation types observed during training and five unseen modulation types, with performance approaching that of the Signal-Like method, which assumes complete prior information. These results indicate that MA-GAN can generate communication jamming waveforms with a certain degree of cross-modulation and robustness to the tested channel conditions without relying on explicit prior information.

As the electromagnetic environment grows increasingly crowded with coexisting wireless systems, overlapping signals have become commonplace. Signal overlap in a limited space leads to feature mixing, boundary blurring, and annotation difficulty, which existing deep learning methods cannot effectively handle. To tackle these challenges, we propose a two–stage detection framework that combines SimMIM self-supervised pretraining with a repulsion loss. In the pretraining phase, a Swin-Tiny backbone is trained on unlabeled time–frequency images through masked reconstruction. In the fine-tuning phase, the backbone is transferred to a YOLOv8 detector, where the repulsion loss is optimized jointly with the base detection loss. In the component ablation, the complete framework improves mAP50 by 9.0 points over the random-initialized baseline on the validation set. On the held-out main and overlap-stress sets, it obtains 0.875 mAP50, 0.798 mAP50-95, and 0.873 mAR, and 0.838, 0.756, and 0.829, respectively. Compared with the YOLOv8 baseline, the observed mAP50 gains are 1.95 and 2.18 points, respectively. The main high-overlap subset is small; its observed 1.64-point margin is therefore reported as descriptive evidence. These results indicate that in-domain self-supervised pretraining combined with spatially constrained box regression provides an effective representation and detection strategy for overlapping signals.