Edge Computing: Enabling Technologies, Applications, and Services
Honghao Gao,Yuyu Yin,Guangjie Han,Wenbing Zhao
DOI: https://doi.org/10.1002/ett.4197
IF: 3.6
2020-01-01
Transactions on Emerging Telecommunications Technologies
Abstract:The development of Internet of Things (IoT) has enabled the creation of many smart devices and these devices in turn are generating unprecedented amount of data. It is desirable to process these data close to the devices that generated them. Edge computing aims to distribute services away from centralized data centers by exploiting the concept of cloudlets for the data generation side, which could significantly improve the response time and user experiences. This new computing paradigm brings new challenges, such as task scheduling, resource allocation, security privacy, collaborative computing. This special issue accepted ten high quality papers. In the paper entitled “An Edge Computing-based Predictive Evaluation Scheme Towards Geological Drilling Data Using Long Short-term Memory Network” by Yin, Junluo et al, they presented a predictive evaluation scheme to effectively achieve mineral exploration. Specifically, they used the long short-term memory to model certain mineral components. Moreover, they incorporated an evolutionary computing method into the scheme for the purpose of optimizing key parameters. They validated their proposed scheme experimentally using geological drilling data. In the paper entitled “A High-accurate Content Popularity Prediction Computational Modelling for Mobile Edge Computing by Using Matrix Completion Technology” by Tan, Jiawei et al, they designed an effective task prediction and caching scheme to tackle the limited storage capacity issue in the edge nodes for edge-computing. The paper introduced two novel mechanisms to achieve the goal. First, they used matrix completion-based caching for content estimation. Second, they decided on cache placement by considering the benefit of unit storage. They showed the effectiveness of their proposed scheme with comprehensive numerical studies. In the paper entitled “HCAR: Human Continuous Activity Recognition using Latent Structure Features” by Zhao, Kun et al, they proposed a structure-feature-based mechanism for human-activity recognition to reduce latency and transmission cost for edge computing. They demonstrated the advantages of the proposed mechanism experimentally. In the paper entitled “Research on Wireless Robot Path Planning under Edge Computing Considering Multi-step Searching and Inflection Points” by Bi, Zhongqin et al, they proposed an improved ant colony optimization algorithm for robot path planning. Specifically, they introduced a grid method to simulate the environment that includes barriers and paths. Then, they used a multi-step path planning method to optimize the path length that a robot travels. Finally, they integrated a inflection point parameter into the probability formula for robot routing. They showcased the effectiveness of the proposed path planning scheme experimentally. In the paper entitled “Traffic Data Reconstruction based on Compressive Sensing with Neighbor Regularization” by Qin, Zhenquan et al, they presented a compressive sensing technique to reduce the amount of data transmission between roadside units (RSUs) and mobile edge computing (MEC) servers. Specifically, they leveraged the low-rank nature and the neighbor relationship between detectors to improve the reconstruction accuracy of different missing scenes. They experimentally demonstrated that they could reconstruct the complete even when the traffic data from the RSUs and MEC servers are reduced by 50%. In the paper entitled “Data Poisoning Attacks on Neighborhood-based Recommender Systems” by Chen, Liang et al, they introduced a new data poisoning attack toward the neighborhood-based recommender systems. The purpose is to improve the robustness of this type of systems by understanding how an adversary would attack them. This attack involves with the injection of fake users into the system for the purpose of recommending specific items to users. The authors studied the effectiveness and transferability of the attack via extensive experiments. In the paper entitled “Maintaining SmartX Multi-View Visibility for Distributed Cloud-native Edge Boxes” by Rathore, M Ahmad et al, they presented a testbed to help understand how to maintain high quality of service in complex distributed systems. A key contribution is to maintain multi-view visibility of the monitoring data regardless of different environments and scenarios. They evaluated the proposed method experimentally at 14 institutions for a number of months. In the paper entitled “Predicting the Next Location: a Self-attention and Recurrent Neural Network Model with Temporal Context” by Zeng, Jun et al, they investigated how to use machine learning to predict a future location for