Mobile users can open program in new tab for better viewing.

Open program in new tab

Day 1 06/11/2020
Room #1

Welcome Message by the General Chair 09:00 - 09:05

Conference starts at 9:00am (Changsha, China, Time Zone)

Welcome Message by the EAI Conference Manager 09:05 - 09:10

Welcome Message by the EAI Community Manager 09:10 - 09:15

Keynote Speaker 1: Prof. Xuemin (Sherman) Shen 09:15 - 09:55

Reinforcement Learning for Resource Management in Space-Air-Ground (SAG) Integrated Vehicular Networks

Coffee Break 09:55 - 10:05

10-minute coffee break

Session 1 10:05 - 12:05

After each paper presentation, please join us for a short Q&A session at Slack platform
10:05 - 10:25
SmartDis: Near-Optimal Task scheduling in Multi-Edge Networks

In multi-edge networks, as the bandwidth and computing resources of edge servers are limited, transmission and processing of large amounts of data could bring significant pressure, leading to violations of service agreements. Thus, it is very important to schedule tasks in edge network efficiently for better performance. In this paper, we formulate the problem as minimizing the overall completion time of tasks in edge networks. Since the problem can be proved to be NP-hard, we propose a novelty algorithm, SmartDis, for scheduling tasks accross multiple edges. The main idea of SmartDis is to select offload slots of tasks based on the principle of choosing the smallest sum of added value of the overall completion time. We show theoretically that the system transmission time of SmartDis is within a constant times of the optimal result, as long as the data upload is scheduled according to the transmission order. The evaluation results illustrate that SmartDis is superior to other cross- domain job scheduling algorithms at this stage, achieving a performance improvement of at least 25%.
Authors: Weiwei Miao (State Grid Jiangsu Electric Power Co., Ltd. Information & Telecommunication Branch), Zeng Zeng (State Grid Jiangsu Electric Power Co., Ltd. Information & Telecommunication Branch), Chuanjun Wang (State Grid Jiangsu Electric Power Co., Ltd. Information & Telecommunication Branch), Zhuzhong Qian,
Hide Authors & Abstract

Show Authors & Abstract
10:25 - 10:45
Energy Efficient Service Composition with Delay Guarantee in a Cloud-Edge System

In a cloud-edge system, mobile users submit comprehensive service requests, on-the-fly service composition to orchestrate service components from different edge nodes is a promising way to achieve a quick response to these requests. Since several mobile applications consume large amount of energy during waiting for the responses, it is critical to achieve less service delay for energy saving as well as improve QoE (Quality of Experience). However, the service completion time in an edge is quite unstable, which increases the overall response time of the composite service. This paper argues that we may accelerate services through service clone via different edges, so that we can guarantee the overall response time of the composite service. And since the data fetch is also time consuming, we propose an effective data-aware service composition algorithm via service cloning to minimize the overall response time. We implement the algorithm and evaluate the performance with extensive simulations. The simulation results show that the proposed algorithm has a good performance improvement on service delay and energy consumption reduction, compared to the traditional algorithms.
Authors: Quan Fang (State Grid Jiangsu Electric Power Co.,LTD.), Menghan Xu (State Grid Jiangsu Electric Power Co.,LTD.), Hao Li (Nanjing University of Aeronautics and Astronautics), Jun Yu (State Grid Jiangsu Electric Power Research Institute), Xin Li (NUAA), Zhuzhong Qian (Nanjing University),
Hide Authors & Abstract

Show Authors & Abstract
10:45 - 11:05
Characterization of OFDM Based Free Space Optical Transmission System Under Heavy Rain Weather

We investigate the performance of OFDM-based FSO transmission link under heavy rain weather conditions in Ségou region, Mali. The proposed system is consisted of a single MZM modulator and a PIN photodiode that performs the optical direct detection. By selecting the convenient values for beam divergence angle and lunch power, the simulation results prove that the generated 42 Gbps OFDM data could be sent up to 1.90 km, using the Carbonneau rain attenuation Model under the worst rain conditions for the considered location.
Authors: Drissa Kamissoko (College of Computer Science and Electronic Engineering, Hunan University), Jing He (College of Computer Science and Electronic Engineering, Hunan University), Macki Tall (College of Computer Science and Electronic Engineering, Hunan University), Hassana Ganamé (School of Electronics and Information Engineering, Huazhong University of Science and Technology),
Hide Authors & Abstract

Show Authors & Abstract
11:05 - 11:25
TouchSense: Accurate and Transparent User Re-authentication via Finger Touching

Re-authentication identifies the user during the whole usage to enhance the security of smartphones. To avoid frequent interrupts to users, user features should be imperceptibly collected for identification without user assistance. Conventionally, behavior habits (e.g. movement, trail) during the user operation are commonly considered as the most appropriate features for re-authentication. The behavior features, however, are often fluctuating and inevitably sacrifice the accuracy of re-authentication, which puts the phones at risk increasingly. In this paper, we propose TouchSense, an accurate and transparent scheme for user re-authentication. The basic idea is to leverage the combined information of human biometric capacitance and touching behavior for user identification. When the user touches capacitive-based sensors, both information can be automatically collected and applied in the authentication, which is transparent to the user. Based on the authentication results, we build up user-legitimate models to comprehensively evaluate the user's legitimacy, which reduces misjudgment and further improves accuracy. Moreover, we implement TouchSense on an SX9310 EVKA board and conduct comprehensive experiments to evaluate it. The results illustrate that TouchSense can identify 98% intruders within 10 seconds, but for legitimate users, the misjudgment is less than 0.9% in 2.6-hours-usage.
Authors: Chong Zhang (University of Electronic Science and Technology of China), Songfan Li (University of Electronic Science and Technology of China), Yihang Song (University of Electronic Science and Technology of China), Li Lu (University of Electronic Science and Technology of China), Mengshu Hou (University of Electronic Science and Technology of China),
Hide Authors & Abstract

Show Authors & Abstract
11:25 - 11:45
Low-carbon Emission Driven Traffic Speed Optimization for Internet of Vehicles

Climate change has become a worldwide concern. Reducing CO2 emission is a major challenge for road transportation sector and is of critical importance. This paper, after studying and analyzing the influence of speed on vehicle CO2 emission, proposes a recommended speed calculation scheme based on IoV to obtain vehicle speed and traffic signal phase information. In the recommended speed scenario, the vehicle is informed of the traffic phase information before arriving at the intersection and can set and optimize the current speed. This paper analyzes the three different status of traffic lights and studies the speed that should be adopted in each status. Under the proposed scheme, the recommended speed helps the driver to reach the destination with higher driving efficiency. The average wait time at red traffic lights is shorter than at speeds that are not recommended, resulting in reduced total travel time, higher uninterrupted pass rates, and decreased vehicle fuel consumption and CO2 emissions.
Authors: Wenjie chen (Central South University of Forestry and Technology), Zhende Xiao (Hunan University), Siming Zou (Central South University of Forestry and Technology),
Hide Authors & Abstract

Show Authors & Abstract
11:45 - 12:05
Android Malware Detection using Ensemble Learning on Sensitive APIs

In recent years, with the quiet popularity of mobile payment methods, mobile terminal equipment has gradually occupied an increasingly important position in people's lives, and its potential security risks have also been paid more and more attention by people. Behavior-based Android malware detection is mostly based on permission analysis and API calls. In this paper, we propose a static Android malicious detection scheme based on sensitive API calls. We extracted all APIs called in the experimental samples through decompilation, and then calculated and ranked the threats related to these APIs according to the mutual information model, selected the top 20 sensitive API calls, and generated a 20-dimensional feature vector for each application. In the classification process, an integrated learning model based on DT classifier, kNN classifier and SVM classifier is used to effectively detect unknown APK samples. We collected 516 benign applications and 528 malicious applications. Through a large number of experiments, the results show that the accuracy of our scheme can be up to 94%, and the precision is up to 95%
Authors: Chunlei Zhao (Tianjin University of Technology), Junhui Yu (Tianjin University of Technology), Wenbai Zheng (Tianjin University of Technology), Yunlong Li (Tianjin University of Technology), Chao Chen (Tianjin University of Technology), Chunxiang Zhang (Tianjin University of Technology),
Hide Authors & Abstract

Show Authors & Abstract

Lunch Break 12:05 - 12:35

30-minute lunch break

Keynote Speaker 2: Prof. Schahram Dustdar 12:35 - 13:25

Edge Intelligence – Edge Computing and Artificial Intelligence

Session 2 13:25 - 14:05

After each paper presentation, please join us for a short Q&A session at Slack platform
13:25 - 13:45
Few Shot Learning Based On The Street View House Numbers (SVHN) Dataset

In recent years, deep learning model has made remarkable achievements in image, voice, text recognition and other fields. However, deep learning model relies heavily on large number of labeled data, which limits its application in the special field of data shortage. For the practical situation such as lack of data, many scholars carry out research on the few shot learning methods, and there are many typical research directions, among which model-agnostic meta-learning (MAML) is one of them. Aiming at the few shot learning method, this paper systematically expounds the current main research methods on few shot learning, the algorithm of MAML and implements the MAML on the SVHN dataset.
Authors: Rundong Yang (Hebei University of Technology), Yancong Deng (University of California, San Diego), Anqi Zhu (University of New South Wales), Xin Tong (University of Illinois, Urbana Champaign), Zhihao Chen (Shanghai University),
Hide Authors & Abstract

Show Authors & Abstract
13:45 - 14:05
Efficient Missing Tag Identification in Large High-Dynamic RFID Systems

Missing tag detection is an important for many radio frequency identification (RFID) systems. Most existing methods can only work in relatively static systems, where there are no unknown tags entering the system. For highly dynamic RFID systems where tags move into and out from the system frequently, it is challenging to identify missing tags because of the interference from unknown tags. In this paper, we propose a new time efficient protocol called HDMI to identify missing tags in highly dynamic RFID systems. Our idea is to combine the index of the replying slot and the bits replied by the tag to efficiently filter out unknown tags and identify missing tags simultaneously. We theoretically analyze how to set optimal parameters (e.g., frame length and bit number replied by tags) to minimize the execution time while ensuring the recognition accuracy of missing tags. Extensive experimental results show that HDMI identify missing tags with a high accuracy rate, and achieving higher efficiency than state-of-art solutions.
Authors: Xinning Chen (Hunan University), Xuan Liu (Hunan University), Ying Xu (Hunan University), Jiangjin Yin (Hunan University), Shigeng Zhang (Central South University),
Hide Authors & Abstract

Show Authors & Abstract

Coffee Break 14:05 - 14:15

10-minute coffee break

Session 3 14:15 - 15:30

After each paper presentation, please join us for a short Q&A session at Slack platform.
14:15 - 14:40
Self-Secure Communication for Internet of Things

Cryptographic key management is a challenge for the large scale deployment of Internet of Things (IoT) devices. It is difficult to properly setup and constantly update keys for numerous IoT devices, especially when these devices are restricted by size and lack of the key input interface. This paper proposes a lightweight key management scheme which embeds the key distribution and update process into the communication process. The keys are constantly changing as the communication data flowing back and forth between IoT devices. Therefore even if a key is stolen by the attacker, it will quickly become invalid as the communication goes on. The proposed scheme also contains a key initialization protocol which generates independent keys for multiple IoT devices simultaneously. This paper describes the protocols in detail and analyzes its security properties. The practicality of the protocol is verified by experiments.
Authors: Bin Hao (Hunan University), Sheng Xiao (Hunan University),
Hide Authors & Abstract

Show Authors & Abstract
14:40 - 15:05
Data gathering system based on multi-layer edge computing nodes

The development of the Internet of Things technology has brought new opportu-nities for the development of edge computing. As an emerging computing model, edge computing makes full use of the device resources at the edge of the network, creating a new type of network computing system at the edge of the network. At the same time, the emergence of edge computing has solved the problem of high latency in the wide area network that has not been solved for a long time in the field of cloud computing, bringing users low latency, fast response, and good service experience. This article will use the edge computing architecture to build a multi-layer edge data gathering system and try to solve some problems that may arise in the system. This system has the characteristics of reducing the amount of uploaded data, ensuring the quality of the data, and adapting to the data types with sudden changes, and can perform separate data processing for local areas. Exper-imental data shows that, compared with the original data collection method, on the premise of ensuring data quality, the amount of uploaded data is further com-pressed and communication overhead is reduced.
Authors: Shuzhen Xiang (Hunan University), Huigui Rong (Hunan University), Zhangchi Xu (Hunan University),
Hide Authors & Abstract

Show Authors & Abstract
15:05 - 15:30
Resource allocation method of Edge-side server based on two types of virtual machines in cloud and edge col-laborative computing architecture

The process of large-scale manufacturing workshops is complex, and the tradi-tional fixed resource allocation method will cause unbalanced load. Aiming at this problem, an edge-side server resource allocation algorithm based on cloud col-laborative architecture has been designed and implemented. By defining the three-dimensional information of each IO-intensive virtual machine in the compute node, the priority of the IO-intensive virtual machine is calculated. Through ana-lyzing the relationship between the CPU-intensive virtual machine and the host physical machine, the number of CPU cores for different tasks of the CPU-intensive virtual machine is obtained, and the hardware resources are uniformly allocated in real time according to the maximum priority list. The experimental re-sults show that the proposed algorithm can significantly satisfy the requirements of high throughput and low latency in large manufacturing workshops, and opti-mize the resource allocation for actual production.
Authors: Junfeng Man (Hunan University of Technology), Longqian Zhao (Hunan University of Technology), Cheng Peng (Hunan University of Technology), Qianqian Li (Hunan University of Technology),
Hide Authors & Abstract

Show Authors & Abstract

Closing Message by the General Chair and Best Paper Announcement 15:05 - 15:10

Day 2 07/11/2020
Day 3 08/11/2020