Spectrum efficiency is one of the key performance metrics in 5G communication networks. To enhance spectrum efficiency, advanced spectrum sharing techniques are normally used. In this article, we provide a survey of the recent development of advanced techniques for spectrum sharing. In particular, we elaborate cognitive radio, device-todevice communication, in-band full-duplex communication, non-orthogonal multiple access, and Long Term Evolution on unlicensed spectrum. For each technique, we present the basic principle and research methodology of the state of the art. By considering various promising evolutions in 5G networks, we emphasize challenges to deploy each technique. Finally, we discuss the integration issue of multiple spectrum sharing techniques and identify potential challenges.

The world of connected devices has led to the rise of the Internet of Things paradigm, where applications rely on multiple devices, gathering and sharing data across highly heterogeneous networks. The variety of possible mechanisms, protocols, and hardware has become a hindrance in the development of architectures capable of addressing the most common IoT use cases, while abstracting services from the underlying communication subsystem. Moreover, the world is moving toward new strict requirements in terms of timeliness and low latency in combination with ultra-high availability and reliability. Thus, future IoT architectures will also have to support the requirements of these cyber-physical applications. In this regard, edge computing has been presented as one of the most promising solutions, relying on the cooperation of nodes by moving services directly to end devices and caching information locally. Therefore, in this article, we propose a modular and scalable architecture based on lightweight virtualization. The provided modularity, combined with the orchestration supplied by Docker, simplifies management and enables distributed deployments, creating a highly dynamic system. Moreover, characteristics such as fault tolerance and system availability are achieved by distributing the application logic across different layers, where failures of devices and micro-services can be masked by this natively redundant architecture, with minimal impact on the overall system performance. Experimental results have validated the implementation of the proposed architecture for on-demand services deployment across different architecture layers.

MEC is able to be used to execute the compute- intensive applications on the edge of transportation networks directly. As a result, the communications traffic is substantially increased among the connected edge devices. Therewith, communications security is emerging as a serious problem, and as an important research issue of the communications security, active feature learning is studied in this article for actively detecting unknown attacks. A model based on deep learning is designed to learn attack features. This model uses unsupervised learning to accomplish the active learning process. In the evaluation, 10 datasets are used to conduct experiments. We compare our model with four other machinelearning- based algorithms, and the comparative results illustrate that our model has a 6 percent gain in accuracy.

In recent years, the success of deep learning in natural scene image processing boosted its application in the analysis of remote sensing images. In this paper, we applied Convolutional Neural Networks (CNN) on the semantic segmentation of remote sensing images. We improve the Encoder- Decoder CNN structure SegNet with index pooling and U-net to make them suitable for multi-targets semantic segmentation of remote sensing images. The results show that these two models have their own advantages and disadvantages on the segmentation of different objects. In addition, we propose an integrated algorithm that integrates these two models. Experimental results show that the presented integrated algorithm can exploite the advantages of both the models for multi-target segmentation and achieve a better segmentation compared to these two models.