With the proliferation of wireless technologies in vehicular networks, robust automatic modulation classification (AMC) has become crucial for optimizing spectrum utilization and maintaining reliability. However, AMC in dynamic vehicular channels poses significant challenges for traditional machine learning techniques. This paper proposes a novel CNN-based approach named Reparameterization Causal Convolutional Network (RepCCNet) to achieve highly accurate and noise-robust AMC performance. RepCCNet incorporates causal convolutions and structural reparameterization techniques to extract long-term time-domain features. A bottleneck structure with channel attention dynamically calibrates feature channels, retaining only helpful information. Multi-sample dropout is integrated during training to improve generalization capability. We demonstrate RepCCNet's state-of-the-art classification accuracy on the two widely used datasets, RadioML 2016.10a and RadioML 2018.01a, across varying signal-to-noise ratios. Compared to existing methods, RepCCNet achieves highly competitive results compared to state-of-the-art approaches, utilizing fewer than 40k parameters. Ablation studies validate the contributions of the proposed architectural innovations. This work represents a significant advancement toward developing deep learning solutions for robust wireless signal classification tasks.

The automatic modulation classification for surface and underwater sensors in the perception layer is crucial in the Internet of Underwater Things (IoUT), where Deep Learning (DL) is becoming an important tool to improve classification accuracy. This work focuses on the radio environment in the perception layer. The biggest challenge in popular DL-based methods is deploying the algorithm in edge devices with limited computing power. Network pruning has been found to be a critical and effective method for network lightweight and the improvement of resources, thus mitigating potential interference. While not studied in previous work, this paper fills the hole in algorithm deployment’s criterion selection and accuracy loss. Specifically, we develop a Convolutional Neural Network (CNN) based lightweight framework on distinguishing modulated signals from generated datasets (which is named DLocean) in different Signal-to-Noise Ratios (SNR). The performance of the lightweight framework is tested on the edge device. The experiments demonstrate that the proposed model compensates for accuracy and can successfully classify the modulation schemes with 93.4% accuracy at the SNR=5 dB. Our results also show that the proposed framework can improve performance without exceeding the original network complexity on edge device deployment.

Modulation recognition plays a crucial role in the acoustic communication systems of autonomous underwater vehicles (AUVs). However, deploying accurate modulation recognition models on resource-constrained edge devices remains a significant challenge. To address this issue, we propose GIQNet, an endto-end deep-learning framework for underwater acoustic modulation classification. GIQNet incorporates two key components: a Temporal Large Kernel Convolution (TLKC) module and a Gated I/Q Mixer (GIQM) module. TLKC applies large kernel convolutions to capture long-range dependencies in the temporal domain. At the same time, GIQM facilitates the adaptive fusion of information from the in-phase and quadrature channels using gating mechanisms. Experiments on real-world underwater acoustic datasets demonstrate GIQNet's effectiveness. Moreover, when evaluated on publicly available radio modulation datasets, GIQNet outperforms existing approaches while using 55% fewer parameters. Further analyses validate the effectiveness of TLKC and GIQM in learning robust features for modulation recognition. The results show that the proposed architecture provides an efficient and effective deep-learning solution tailored for underwater acoustic modulation classification, making it suitable for deployment on resource-constrained AUV edge devices.

Underwater Internet-of-Things (UIoT) is an extension of Internet-of-Things technology in underwater. The underwater acoustic network with WiFi architecture (UW-WiFi), as a specific deployment of UIoT, has been proved to be a promising technique for wide-ranging marine applications. However, due to the unique features of underwater acoustic channel, such as long and variable propagation delay, low available bandwidth and high bit error rate, conventional medium access control (MAC) protocols designed for terrestrial WiFi networks need an overhaul to work efficiently for UW-WiFi networks. In consideration of the aforementioned channel features, different demands of nodes to occupy channel resources and the reliability of data transmission, a time sequence-based dynamic on-demand assignment (SDDA) MAC protocol for UW-WiFi networks is proposed in this paper. In SDDA, the collision-free scheduling is integrated with the reservation mechanism, aiming to address the issue of various access requirements of diverse task nodes on channel resources in the UW-WiFi network system. The designed protocol employs propagation delays and the amount of data to be sent by terminal nodes to achieve non-conflict transmissions of control packets and on-demand scheduling of data packets. Additionally, the scheme does not require extra overhead for time synchronization. Comparison simulations demonstrate that SDDA provides considerable benefits in terms of channel utilization, end-to-end delay and packet delivery ratio. At last, the SDDA protocol is implemented and a UW-WiFi network system is set up in the marine environment. The ocean field experiment results agree well with the simulated ones and also verify that the proposed protocol achieve conflict-free on-demand scheduling, fairness among terminal nodes at different ranges and the practicability in actual environments.