Skip to main content

Über dieses Buch

This book constitutes the refereed post-conference proceedings of the 10th International Conference on Broadband Communications, Networks, and Systems, Broadnets 2019, which took place in Xi’an, China, in October 2019. The 19 full papers presented were carefully reviewed and selected from 61 submissions. The papers are thematically grouped as follows: Wireless Networks and Applications, Communication and Sensor Networks, Internet of Things, Pervasive Computing, Security and Privacy.



Wireless Networks and Applications


Design and Implementation of Non-intrusive Stationary Occupancy Count in Elevator with WiFi

Wi-Fi Sensing has shown huge progress in last few years. Multiple Input and Multiple Output (MIMO) has opened a gateway of new generation of sensing capabilities. This can also be used as a passive surveillance technology which is non-intrusive meaning it is not a nuisance as it is not need the subjects to carry any dedicated device. In this thesis, we present a way to count crowd in the elevator non-intrusively with 5 GHz Wi-Fi signals. For this purpose, Channel State Information (CSI) is collected from the commercially available off-the-shelf (COTS) Wi-Fi devices setup in an elevator. Our goal is to Analyze the CSI of every subcarrier frequency and then count the occupancy in it with the help of Convolutional Neural Network (CNN). After CSI data collection, we normalize the data with Savitzky Golay method. Each CSI subcarrier data of all the samples is made mean centered and then outliers are removed by applying Hampel Filter. The resultant wave is decimated and divided into 5 equal length segments representing the human presence recorded in 5 s. Continuous wavelet frequency representations are generated for all segments of every CSI sub-carrier frequency waves. These frequency pattern images are then fed to the CNN model to generalize and classify what category of crowd they belong to. After training, the model can achieve the test accuracy of more than 90%.
Wei Shi, Umer Tahir, Hui Zhang, Jizhong Zhao

Human Activity Recognition Using Wi-Fi Imaging with Deep Learning

Robots have been increasingly used in production line and real life, such as warehousing, logistics, security, smart home and so on. In most applications, localization is always one of the most basic tasks of the robot. To acquire the object location, existing work mainly relies on computer vision. Such methods encounter many problems in practice, such as high computational complexity, large influence by light conditions, and heavy crafting of pre-training. These problems have become one of the key factors that constrains the precise automation of robots. This paper proposes an RFID-based robot navigation and target localization scheme, which is easy to deploy, low cost, and can work in non-line-of-sight scenarios. The main contributions of this paper are as follows: 1. We collect the phase variation of the tag by a rotating reader antenna, and calculate the azimuth of the tag relative to the antenna by the channel similarity weighted average method. Then, the location of the tag is determined by the AoA method. 2. Based on the theory of tag equivalent circuit, antenna radiation field, and cylindrical symmetry oscillator mutual impedance, the phenomenon of RSS weakening of adjacent tags is analyzed. Based on this phenomenon, we achieve accurate target localization and multi-target relative localization by utilizing region segmentation and dynamic time warping algorithms. 3. The proposed scheme is lightweight and low-cost. We built a prototype system using commercial UHF RFID readers and passive tags, and conduct extensive experiments. The experimental results show that the model can effectively achieve the precise location of the robot and the object with an average error of 27 cm and 2 cm.
Yubing Li, Yujiao Ma, Nan Yang, Wei Shi, Jizhong Zhao

Convex Optimization Algorithm for Wireless Localization by Using Hybrid RSS and AOA Measurements

With the development of new array technology and smart antenna, it is easier to obtain the angle of arrival (AOA) measurements. The hybrid received signal strength (RSS) and AOA measurement techniques are proposed for the wireless localization in the paper. By converting the measurement equations and relaxing the optimization function, a second order cone programming and semidefinite programming (SOCPSDP) algorithm is put forward to obtain the position estimate by considering the known or unknown transmit power. The proposed SOCPSDP algorithm provides a solution to the source position estimate and avoids the initialization process. The simulations show that the SOCPSDP algorithm performs better than the semidefinite programming (SDP) algorithm. The accuracy performance of the proposed SOCPSDP algorithm degrades as the measurement noises increase.
Lufeng Mo, Xiaoping Wu, Guoying Wang

Wi-Fi Floor Localization in an Unsupervised Manner

In recent decades, with the development of computer, indoor positioning applications have been developed rapidly. GPS has become one of the standards for outdoor positioning. However, there are great conditions for the use of GPS, GPS cannot be used indoors. At the same time, the indoor positioning scene has a great application prospect, through the use of indoor accessible signals (such as Wi-Fi, ZigBee, Bluetooth, UWB, etc.), according to the indoor environment and application, can be created based on the indoor positioning system. In the indoor positioning, there are two challenges, first of all, floor positioning, if the building has more than two layers, the second is planar positioning.
This paper solves the problem of floor positioning, and floor positioning based on Wi-Fi unsupervised recognition has attracted wide attention because it can get positioning results at a lower cost. In this paper, we try unsupervised indoor positioning methods, using only Wi-Fi crowdsourcing data. We get four months of data from seven-story buildings, by scanning the router’s information. The application of neural network model can achieve unsupervised indoor positioning.
This clustering model aggregates all signals from the same floor into one class, and we use convolution neural networks, descending dimension feature extraction functions. The experiments show our solution obtains very high precision clustering results, so it can be summed up in this sense that the Wi-Fi crowdsourcing data can be used to locate in some way as the future direction of indoor positioning development.
Liangliang Lin, Wei Shi, Muhammad Asim, Hui Zhang, Shuting Hu, Jizhong Zhao

Communication and Sensor Networks


Virtual Network Embedding Algorithm Based on Multi-objective Particle Swarm Optimization of Pareto Entropy

Virtual network embedding/mapping refers to the reasonable allocation of substrate network resources for users’ virtual network requests, which is a key issue for virtual resource leasing in Cloud computing. Most of the existing researches only aim to maximize the revenue. As the scale of hardware network expands, the energy consumption of substrate network also needs to be paid more attention. In this paper, a multi-objective virtual network mapping algorithm based on particle swarm optimization with Pareto entropy (VNE-MOPSO) is proposed. It combines energy consumption and revenue. The algorithm controls the energy consumption of the substrate network as much as possible to achieve the goal of energy saving on the premise of ensuring a small resource cost. By introducing the Pareto entropy based multi-objective optimization model, it can calculate the difference of entropy and evaluate the evolutionary state. With this as feedback information, a dynamic adaptive particle velocity updating strategy is designed to achieve the goal of solving the approximate optimal multi-objective optimization mapping scheme. Simulation results show that the proposed algorithm has certain advantages over the typical single target mapping algorithm in cost, energy consumption and average return.
Ying Liu, Cong Wang, Ying Yuan, Guo-jia Jiang, Ke-zhen Liu, Cui-rong Wang

Measuring and Analyzing the Burst Ratio in IP Traffic

The burst ratio is a parameter of the packet loss process, characterizing the tendency of losses to group together, in long series. Such series of losses are especially unwelcome in multimedia transmissions, which constitute a large fraction of contemporary traffic. In this paper, we first present and discuss results of measurements of the burst ratio in IP traffic, at a bottleneck link of our university campus. The measurements were conducted in various network conditions, i.e. various loads, ports/applications used and packet size distributions. Secondly, we present theoretical values of the burst ratio, computed using a queueing model, and compare them with the values obtained in the measurements.
Dominik Samociuk, Marek Barczyk, Andrzej Chydzinski

WiCLR: A Sign Language Recognition System Framework Based on Wireless Sensing

The non-intrusion and device-free sign language recognition (SLR) is of great significance to improve the quality of life, broaden living space and enhance social service for the deaf and mute. In this paper, we propose a SLR system framework, called WiCLR, for identifying isolated words in Chinese sign language exploring the channel state information (CSI). WiCLR is made up entirely of commercial wireless devices, which does not incur significant deployment and maintenance overhead. In the framework we devise a signal denoising method to remove the environment noise and the internal state transitions in commercial devices. Moreover, we propose the multi-stream anomaly detection algorithm in action segmentation and fusion. Finally, the extreme learning machine (ELM) is utilized to meet the accuracy and real-time requirements. The experiment results show that the recognition accuracy of the approach reaches 94.3% and 91.7% respectively in an empty conference room and a laboratory.
Wang Lin, Liu Yu, Jing Nan

High-Resolution Image Reconstruction Array of Based on Low-Resolution Infrared Sensor

As the time is progressing the number of wireless devices around us is increasing, making Wi-Fi availability more and more vibrant in our surroundings. Wi-Fi sensing is becoming more and more popular as it does not raise privacy concerns in compare to a camera based approach and also our subject (human) doesn’t have to be in any special environment or wear any special devices (sensors).
Our goal is to use Wi-Fi signal data obtained using commodity Wi-Fi for human activity recognition. Our method for addressing this problem involves capturing Wi-Fi signals data and using different digital signal processing techniques. First we do noise reduction of our sample data by using Hampel filter then we convert our data from frequency domain into time domain for temporal analysis. After this we use the scalogram representation and apply the above mentioned steps to all our data in terms of sub carriers. Finally we use those sub carriers in combined for one activity sample as all the sub carriers combined form up an activity so we shall use the combined signal in the form of power spectrum image as input for the neural network.
We choose Alexnet for classification of our data. Before feeding our data into pre-trained CNN for training we first divided the data into two portions first for training which is 85% secondly for validation which is 15%. It took almost 18 h on single CPU and finally achieved an accuracy of above 90%.
Yubing Li, Hamid Hussain, Chen Yang, Shuting Hu, Jizhong Zhao

Internet of Things


Analysis and Implementation of Multidimensional Data Visualization Methods in Large-Scale Power Internet of Things

In the large-scale power Internet of things, a large amount of data is generated due to its diversity. Data visualization technology is very important for people to capture the mathematical characteristics, rules and knowledge of data. People tend to get limited and less valuable information directly form large data when rely only on human-being’s cognition. Therefore, people need new means and technologies to help display these data more intuitively and effectively. Data visualization mainly aims at conveying and communicating information clearly and effectively in term of graphical display, which can make data more human-readable and intuitive. Multidimensional data visualization refers to the methods to project multidimensional data to two-dimensional plane. It has important applications in exploratory data analysis, and verification of clustering or classification problems. This paper mainly studies the data visualization algorithm and technology in large-scale power Internet of things. Specifically, the traditional Radviz algorithm is selected and improved. The improved radviz-t algorithm is designed and implemented, and the unknown information of data transmission is obtained by analyzing its visualization effect. Finally, the methods used to study fault detection ability of radviz-t algorithm are discussed in detail.
Zhoubin Liu, Zixiang Wang, Boyang Wei, Xiaolu Yuan

Device-Free Gesture Recognition Using Time Series RFID Signals

A wide range of applications can benefit from the human motion recognition techniques that utilize the fluctuation of time series wireless signals to infer human gestures. Among which, device-free gesture recognition becomes more attractive because it does not need human to carry or wear sensing devices. Existing device-free solutions, though yielding good performance, require heavy crafting on data preprocessing and feature extraction. In this paper, we propose RF-Mnet, a deep-learning based device-free gesture recognition framework, which explores the possibility of directly utilizing time series RFID tag signal to recognize static and dynamic gestures. We conduct extensive experiments in three different environments. The results demonstrate the superior effectiveness of the proposed RF-Mnet framework.
Han Ding, Lei Guo, Cui Zhao, Xiao Li, Wei Shi, Jizhong Zhao

Dynamic IFFSM Modeling Using IFHMM-Based Bayesian Non-parametric Learning for Energy Disaggregation in Smart Solar Home System

Recently, the analysis and recognition of each appliance’s energy consumption are fundamental in smart homes and smart buildings systems. Our paper presents a novel Non-Intrusive Load Monitoring (NILM) recognition method based on Bayesian Non-Parametric (BNP) learning approach to solve the problem of energy disaggregation for smart Solar Home System (SHS). Several researches assumed that there is prior information about the household appliances in order to restrict those that do not hold the maximum expectation for inference. Therefore, to deal with the unknown number of electrical appliances in a SHS, we have adapted a dynamic Infinite Factorial Hidden Markov Model (IFHMM) -based Infinite Factorial Finite State Machine (IFFSM) to our NILM times-series modeling as an unsupervised BNP learning method. Our suggested method can grip with few or nappropriate learning data as well as to standardize electrical appliance modeling. Our proposed method outperforms FHMM-based FSM modeling results illustrated in literature.
Kalthoum Zaouali, Mohamed Lassaad Ammari, Amine Chouaieb, Ridha Bouallegue

Distributed Integrated Modular Avionics Resource Allocation and Scheduling Algorithm Supporting Task Migration

At present, the avionics system tends to be modularized and integrated, and the distributed integrated modular avionics system (DIMA) is proposed as the development direction of the next generation avionics system. In order to support the operation of complex tasks, DIMA needs to have an effective resource allocation and scheduling algorithm for task migration and reorganization to achieve reconstruction. However, many current resource allocation and scheduling algorithms, used in traditional avionics systems, are not available for DIMA. In view of the above problems, the paper analyzes the characteristics of the DIMA avionics system architecture model and builds abstract models of the computing resources, computing platforms and tasks. Based on the established model, an efficient task scheduling algorithm, resource allocation algorithm and task migration algorithm for DIMA avionics architecture are designed. And we do simulation experiments to establish the model, and compare the designed EWSA algorithm with the mainstream algorithm JIT-C. The results show better performance in terms of workflow average completion time, successful scheduling completion rate and optimization rate. In addition, considering the failure in the process of executing the mission, we proposed a mission migration and reorganization algorithm WMA and set different time and number of fault resources of the aircraft in the simulation experiments to evaluate the performance of WMA algorithm.
Qing Zhou, Kui Li, Guoquan Zhang, Liang Liu

Pervasive Computing


An Intelligent Question and Answering System for Dental Healthcare

The intelligent question and answering system is an artificial intelligence product that combines natural language processing technology and information retrieval technology. This paper designs and implements a retrieval-based intelligent question and answering system for closed domain, and focuses on researching and improving related algorithms. The intelligent question and answering system mainly includes three modules: classifier, Q&A system and Chatbots API. This paper focuses on the classifier module, and designs and implements a classifier based on neural network technology, mainly involving word vector, bidirectional long short-term memory (Bi-LSTM), and attention mechanism. The word vector technology is derived from the word2vec tool proposed by Google in 2013. This paper uses the skip-gram model in word2vec.The Q&A system mainly consists of two modules: semantic analysis and retrieval. The semantic analysis mainly includes techniques such as part-of-speech tagging and dependency parsing. The retrieval mainly relates to technologies such as indexing and search. The Chatbots API calls the API provided by Turing Robotics. The intelligent question and answering system designed and implemented in this paper has been put into use, and the user experience is very good.
Yan Jiang, Yueshen Xu, Jin Guo, Yaning Liu, Rui Li

Classifier Fusion Method Based Emotion Recognition for Mobile Phone Users

With the development of modern society, people are paying more and more attention to their mental situation. An emotion is an external reaction of people’s psychological state. Therefore, emotion recognition has attached widespread attention and become a hot research topic. Currently, researchers identify people’s emotion mainly based on their facial expression, human behavior, physiological signals, etc. These traditional methods usually require some additional ancillary equipment to obtain information. This always inevitably makes trouble for users. At the same time, ordinary smart-phones are equipped with a lot of sensor devices nowadays. This enables researchers to collect emotion-related information of mobile users just using their mobile phones. In this paper, we track daily behavior data of 50 student volunteers using sensors on their smart-phones. Then a machine learning based classifier pool is constructed with considering diversity and complementary. Base classifiers with high inconsistent are combined using a dynamic adaptive fusion strategy. The weights of base classifiers are learned based on their prior probabilities and class-conditional probabilities. Finally, the emotion status of mobile phone users are predicted.
Luobing Dong, Yueshen Xu, Ping Wang, Shijun He

Toward Detection of Driver Drowsiness with Commercial Smartwatch and Smartphone

In the life, there are always many objects that are unable to actively contact with us, such as keychains, glasses and mobile phones. In general, they are referred to non-cooperative targets. Non-cooperative targets are often overlooked by users while being hard to find. It will be convenient if we can localize those non-cooperative targets. We propose a non-cooperative target localization system which based on MEMS. We detect the arm posture changes of the user by using the MEMS sensors which embedded in the smart watch. First distinguish the arm motions, identify the final motion, and then perform the localization. There are two essential models in our system. The first step is arm gesture estimation model which based on MESE sensor in smart watch. we first collect the MEMS sensor data from the watch. And then the arm kinematic model and formulate the mathematical relationship between arm degrees of freedom with and the gestures of watch. We compare the results of the four actions which are important in the later model with the Kinect observations. The errors in the space are less than 0.14 m. The second step is non-cooperative target localization model that based on the first step. We use the 5-degrees data of the arm to train the classification model and identify the key actions in the scene. In this step, we estimate the location of non-cooperative targets through the type of interactive actions. To demonstrate the effectiveness of our system, we implement it on tracking keys and mobile phones in practice. The experiments show that the localization accuracy is >83%.
Liangliang Lin, Hongyu Yang, Yang Liu, Haoyuan Zheng, Jizhong Zhao

Security and Privacy


Fast Algorithm for the Minimum Chebyshev Distance in RNA Secondary Structure

Minimum Chebyshev distance computation between base-pair and structures cost most time while comparing RNA secondary structures. We present a fast algorithm for speeding up the minimum Chebyshev distance computation. Based on the properties of RNA dot plots and Chebyshev distance, this algorithm uses binary search to reduce the size of base pairs and compute Chebyshev distances rapidly. Compared with O(n) time complexity of the original algorithm, the new one takes nearly [O(log2n), O(1)] time.
Tiejun Ke, Changwu Wang, Wenyuan Liu, Jiaomin Liu

Fine-Grained Access Control in mHealth with Hidden Policy and Traceability

Ciphertext-Policy Attribute-Based Encryption (CP-ABE) is a well-received cryptographic primitive to securely share personal health records (PHRs) in mobile healthcare (mHealth). Nevertheless, traditional CP-ABE can not be directly deployed in mHealth. First, the attribute universe scale is bounded to the system security parameter and lack of scalability. Second, the sensitive data is encrypted, but the access policy is in the plaintext form. Last but not least, it is difficult to catch the malicious user who intentionally leaks his access privilege since that the same attributes mean the same access privilege. In this paper, we propose HTAC, a fine-grained access control scheme with partially hidden policy and white-box traceability. In HTAC, the system attribute universe is larger universe without any redundant restriction. Each attribute is described by an attribute name and an attribute value. The attribute value is embedded in the PHR ciphertext and the plaintext attribute name is clear in the access policy. Moreover, the malicious user who illegally leaks his (partial or modified) private key could be precisely traced. The security analysis and performance comparison demonstrate that HTAC is secure and practical for mHealth applications.
Qi Li, Yinghui Zhang, Tao Zhang

Construction of Laboratory Refined Management in Local Applied University

The refined management of laboratory in local applied university is conducive to the formation of mode to innovative talent training and the improvement the ability to innovate. To ensure the efficient operations of laboratory in local applied university, in this paper we put forward to s new construction method of laboratory refined management in local applied university on the basis of analyzing the current actual situation of laboratory managements. Finally, taking the Computer laboratory of School of Electrical and Electronic Engineering in Shanghai University of Engineering and Technology as an example, the refined management case of laboratory in local applied university is constructed in terms of developing and designing of the refined laboratory management information system (MIS), and future works on the following refined managements of laboratory in local applied university was presented. It has a certain impact on promoting the management efficiency of local applied university and the service ability of the society.
Lan Liu, Xiankun Sun, Chengfan Li, Yongmei Lei

Design of VNF-Mapping with Node Protection in WDM Metro Networks

Network Function Virtualization (NFV) is considered to be one of the enabling technologies for 5G. NFV poses several challenges, like deciding the virtual network function (VNF) placement and chaining, and adding backup resources to guarantee the survivability of service chains. In this paper, we propose a genetic algorithm that jointly solves the VNF-placement, chaining and virtual topology design problem in WDM metro ring network, with the additional capacity of providing node protection. The simulation results show how important is to solve all of these subproblems jointly, as well as the benefits of using shared VNF and network resources between backup instances in order to reduce both the service blocking ratio and the number of active CPUs.
Lidia Ruiz, Ramón J. Durán, Ignacio de Miguel, Noemí Merayo, Juan Carlos Aguado, Patricia Fernández, Rubén M. Lorenzo, Evaristo J. Abril


Weitere Informationen

Premium Partner