Search

With the development of 5G and future 6G communication technology, millimeter-wave communication (MMWC) has attracted widespread attention because it can provide a wide frequency band and high data transmission rate. However, the effective application of MMWC is challenged by the severe path loss and environmental sensitivity caused by its high-frequency characteristics. In addition, most of the existing beam prediction methods are limited by excessive communication resource overhead and low prediction accuracy. To tackle these problems, this article proposes for the first time to use Wi-Fi channel state information (CSI) to assist beam prediction. By combining ubiquitous and low-communication-resource-cost Wi-Fi sensing technology with deep learning technology, an efficient deep learning model called dynamic channel attention convolutional neural network (DCACNN) is designed. This model aims to reduce the computational complexity and high resource overhead while improving prediction accuracy. Then, a real-world dataset based on millimeter-wave (mmWave) power and CSI is made. Finally, extensive experiments are conducted to demonstrate the proposed method’s potential for efficient and reliable beam prediction in dynamic environments. Specifically, the prediction accuracy can be achieved up to 94.8%, 86.81%, and 70.18% under quiet, normal, and noisy environments, respectively. The results showcase the exceptional predictive accuracy and superior environmental adaptability of the proposed deep learning model compared to other deep learning models.

Localization has become a prominent use case for 6G mobile networks, and the integrated localization and communication (ILAC) system represents an inevitable trend. To design and manage such ILAC systems effectively and efficiently, quantification of its performance bound, i.e., an analytical model that reveals the trade-off between communication and localization performance, is a crucial yet unresolved task. To address this, we proposed an analytical framework for an ILAC system that achieves communication and localization. Specifically, we derived a closed-form expression of the capacity loss versus localization Cramer-Rao lower bound (CRB) loss through time-domain and frequency-domain resource allocation. Through simulations, we validated our analytical model and observed that frequency-domain resource allocation is preferable in scenarios with fewer antennas at the next generation nodeB (gNB) and a larger distance between user equipment (UE) and gNB. Conversely, time-domain resource allocation is preferable in scenarios with more antennas and a smaller distance between UE and the gNB.

Machine analysis of underwater images is essential to most underwater applications. However, both the limitation of communication bandwidth and underwater degradation bring much difficulty to accurate machine recognition at the end system. Few existing underwater compression methods consider unique underwater prior knowledge to better serve for machine vision under low bit-rates. To address this problem, we propose a novel underwater image compression framework for machines, which utilizes underwater priors to contrastively enhance degraded features by contrastive learning and efficiently compress machine-friendly features under low bit-rates. A dataset is built to provide positive and negative samples for contrastive learning based on machine analysis performance. At the encoder side, a feature extractor and a feature encoder are employed to extract machine-related features and compress them into compact representations. To alleviate the effect of underwater degradation on machine vision, a prior-guided contrastive feature enhancement module is proposed to learn more machine-friendly features based on positive and negative samples from our dataset. Then a feature refinement block is designed to remove channel-wise redundancy and focus spatial-wise importance based on high similarity of machine-related features and characteristics of underwater images. More compact representations are obtained without degrading analysis performance under low bit-rates. At the decoder side, both machine-friendly features and image are reconstructed to support different types of analysis tasks. Experimental results demonstrate the superiority of our framework in machine vision tasks compared with traditional compression methods and learned-based methods. Besides, our method still preserves basic capability of human perception.