三月 13

2017 實驗室簡介/Lab tour for 2017 new students


我們將在 3/13 日下午兩點舉辦研究室簡介並開放參觀。地點為資訊系442研究室(以及 B04 機器人足球場)。有興趣的同學請聯絡 林坤力同學 。

六月 30

Software Tools for Computer Science Writing

Using right research tools can allow us to effectively conduct researches. It is difficult, not impossible, to have ONE tool for all researchers. Below are the tools used by the members of our research group. While selecting the tools, we try to choose the tools that are cross-platforms, open-source, and low-cost with education discount.

Tools we are using:

  • Bibliography: mendeley (free) is the reference tool to collect reference materials. With this, one can easily share his/her reference materials with colleagues. Desktop tools are provided to edit reference materials off-line. Web-Plugin are also available to clip information on web. It is one of few reference tools that can export reference to latex bibliography format.
  • Mind Map: Xmind (free) is the software to manage your thoughts. Xmind is very likely the most used mind mapping software at time of this writing. In addition, it is cross-platform.
  • Taking note: evernote  (free version is available) is the tool to capture notes and reference materials, in particular, information on the web. It can be used with iPad for hand drawing notes.
  • Writing: latex  (free) is the document formatting tool for scientific publication. It allows us to focus on the content, not the format.
  • Editor: Emacs (free) is the cross-platform editor, highly recommended to programmers and researchers.
  • Graphic: gnuplot (free) and OmniGraffle (Commercial). gunplot is a tool to plot research results and is cross-platform. OmniGraffle is a commercial drawing tool to provide nice looking vector graphs.
  • Project Management: newslabx.csie.ntu.edu.tw/redmine (free)

六月 30

Timeline to Prepare Your Thesis

Writing a thesis should not be a three week project and deserves a long term preparation. Here is the suggested timeline for a master thesis in the group. Suppose that you are expected to take the exam in June at year Y.

  • July/August, Y-1: discuss with the professor for the research topic. The discussion should be held intensively, either once or twice a week.
  • September, Y-1: narrow down the topic and conduct the preliminary experiments to show that the targeted topic is real. Prepare 5 page slides to describe the motivation, preliminary experiments, and targeted topic.
  • October, Y-1: Write down the motivation, formulate the targeted topic, survey the literature and start to develop the solution.
  • November, Y-1: The solution should be carefully analyzed and write down the notes.
  • December, Y-1: Write down introduction and related work chapter, and start to implement the solution
  • January, Y: submit the introduction and related work for review.
  • February, Y: Revise the solution and implementation after the first implementation is ready for testing.
  • March, Y: Submit system architecture and approach for review
  • April, Y: Collect performance evaluation results and write the performance evaluation chapter.
  • May, Y: Submit performance evaluation chapter for review.
  • June, Y: defend your thesis and submit final version to NTU.

The above schedule does not give you the flexibility to fail and do it again. So, you should start earlier and be prepared for a delay schedule in your approach does not work.

十月 06

四歲孩童 也能開發物聯網服務

智慧聯網 應用與商機探究(四)


台灣大學資訊工程研究所施吉昇教授表示,悟空(WuKong)團隊的目標在開發一個可以連結物聯網產業鏈與使用者的智慧型虛擬中介軟體,包含物聯網服務(IoT Services)開發環境、虛擬機、物聯網服務(IoT Services)安裝與服務更新。


施吉昇指出,現有物聯網大多以客製化方式設計與開發,其主要原因為新興應用系統缺乏可重複使用的軟體元件與服務,多樣性的硬體平台,以及缺乏物聯網軟體服務標準與通訊標準。此一設計與開發方式與 90 年代後期智慧手機應用軟體的開發遭遇相同的困境,而首波智慧手機應用軟體開發的失敗開發經驗,也見證此一方法的確不可行。

2007 年後,智慧手機透過平台標準化,包含標準虛擬平台(如 Android) 以及標準硬體平台(如 iPhone 與 Blackberry),降低智慧手機應用軟體開發成本,引導建立完整、自給自足的產業生態圈,才開啟了行動應用軟體的蓬勃發展。

「悟空(WuKong)物聯網服務開發環境」專門為沒有經過長期訓練及軟體開發能力的應用程式開發人員而設計,悟空團隊的目標在建立一個開放的物聯網服務開發、安裝、與維護的架構。藉由提供流程導向架構(Flow-Base Programming, FBP)服務開發環境,應用服務的開發者,重在如何將現有服務組合為適切的物聯網應用服務與資料流動,其開發過程,完全與軟硬體平台獨立。

其次,悟空的虛擬中介軟體負責物聯網服務的安裝與維護。悟空虛擬機(NanoKong Virtual Machine)在多樣性的硬體平台上提供標準化的軟體執行環境,目前已完成 Arduino 與 Intel Galileo 平台適用的虛擬機。




(Original Report: http://www.wa-people.com/2014/08/blog-post_12.html)

六月 20

IEEE Signal Processing Society 2014 Summer School on Internet of Things and Machine-to-Machine Systems

2014 Summer school on Internet of Things and Machine-to-Machine will be hosted in National Taiwan University from August 26 to 29. The students are encouraged to attend the summer school and learn the latest development on IoT and M2M. Below are the summary of the summer school and the details can be found at this link.



There here has been a great deal of interest in the machine-to-machine (M2M) systems and Internet-of-Things (IoT) recently. Billions of M2M devices are expected to be connected and a major portion of them are through wireless media. M2M communication is expected to open doors for new type of applications and services that include smart metering, telemetry, surveillance, healthcare, transportation, utilities, and remote maintenance and control.

Smart M2M systems and IoT are the key technologies to enable applications that involved in interacting with a large number of devices. An increasing number of researches are working to create better technologies for machine-to-machine and IoT systems. Signal processing will play a critical role in the advance of the smart IoT and M2M technologies. Signal processing is essential for ubiquitous information sensing and collection. Advanced signal processing techniques improve M2M communications capability. Distributed signaling processing techniques provide opportunities for scalable IoT solutions. Video signal processing enables multimedia IoT and surveillance. Statistical signaling processing enables information forensic and data analytics to handle the large amount of available information in IoT world. In this summer school, the lecturers will covers various aspects of signal processing M2M systems, and IoT applications.


Technical Program:

  • Wireless Network Economics and Games, Prof. Jianwei Huang, The Chinese University of Hong Kong, Hong Kong
  • Convergence of IoT and Social Networks Issues, Prof. Antonio lera, University of Reggio Calabria Cagliari, Italy
  • Security, Privacy and Trust in IoT and M2M, Dr. Joseph Teo, Institute for Infocomm Research, Singapore
  • A Cyber-Physical System for Oil Applications, Prof. Magdy Bayoumi, University of Louisiana at Lafayette, USA
  • Optimization and Inference for Cyber Security in Complex Engineered Networks, Prof. Chee Wei Tan, City University of Hong Kong, Hong Kong
  • Sensor Networking and Decentralized Signal Processing for Energy Management, Prof. Anna Scaglione, UC Davis, USA
  • The IoT Big Data Analytics, Prof. Mahmoud Daneshmand, Stevens Institute of Technology, USA


  • Student Travel Grant Application: July 1, 2014
  • Early Registration: August 1, 2014






六月 02

102學年度 中國電機工程學會 青年論文獎

102學年度 中國電機工程學會青年論文獎 開始接受申請。2013 年本實驗室畢業學生陳欣毅同學,以其碩士論文獲得『青年論文獎』



六月 02




‧為激勵國內研究生對於資訊相關領域研究工作之投入,特設置 碩、博士論文獎。
‧碩、博士論文獎每年頒發乙次,博士類 優等獎 壹名、佳作 若干名;碩士類 優等獎 壹名、佳作 若干名。
‧得獎者獲頒 獎章乙枚及獎牌乙面。
‧得獎人指導教授及共同指導教授獲贈 獎牌乙面及紀念品。
1.台灣境內的資訊類 碩、博士論文。

‧申請者每年八月十五日前備齊申請表件及相關資料向社團法人中華民國資訊學會提出申請。為遴選年度得獎者,由社團法人中華民國資訊學會邀請 學者專家組成評審小組,就當年度申請者中從嚴擇最優者,向社團法人中華民國資訊學會常務理事會推薦,惟得從缺。

‧本獎項之申請資料電子檔(ZIP檔格式,請以 “申請者英文姓名.zip” 為檔名):內含 1.申請人簡歷表(線上報名))(請線上報名印出報名表簽名後隨申請資料寄出)
2.畢業論文電子檔(PDF 格式),須包含論文封面。
3.學位論文審定書掃描檔( PDF 或 JPG 格式)。
4.簡報檔(PPT 格式或 PDF 格式)
5.學歷證件(學生證影本、畢業證書影本或在學證明等)掃描檔( PDF 或 JPG 格式)
6.其他對評審審查有幫助的內容,如可安裝執行之軟體 程式、系統展示影音檔、相關著作、比賽得獎之獎狀掃描檔。
7.指導教授 (需為本會會員)推薦書

8.若有共同指導教授,其姓名須表列於學位論文封面上,方認定受獎。 * 申請人應負責電子檔之可讀性,並自行說明內容(請內含一 README 文字檔)使審查程序得以順利進行。
* 每位教授至多推薦博士兩名、碩士兩名。

‧上述資料請燒錄製成光碟 (三片) 寄至:
「社團法人中華民國資訊學會碩、博士最佳論文獎 評審委員會」收
地址:106 台北市羅斯福路四段1號
臺灣大學 資訊工程學系 賴飛羆教授 收轉
黃國軒教授 (大同大學資訊工程系)
Email : khhuang@ttu.edu.tw
TEL : (02)2182-2928 ext 6551

二月 17

SIGBED Review newsletter, volume 10 issue 4


This issue of SIGBED review is a special issue for 5th Workshop on Adaptive and Reconfigurable Embedded Systems (APRES 2013). Below are lists of papers in the issue.


  • Towards Runtime Adaptation in AUTOSAR
  • Fair Bandwidth Sharing among Virtual Machines in a Multi-criticality Scope
  • Fault-Tolerant Hierarchical Real-Time Scheduling with Backup Partitions on Single Processor
  • Towards a Resilient Deployment and Configuration Infrastructure for Fractionated Spacecraft
  • Realizing a Fault-tolerant Embedded Controller on Distributed Real-Time Systems
  • Transition-aware Task Scheduling and Configuration Selection in Reconfigurable Embedded Systems
  • Adaptive EDF: Using Predictive Execution Time
  • On the Conceptual Design of a Dynamic Component Model for Reconfigurable AUTOSAR Systems
  • Formal Verification of Service-Oriented Adaptive Driver Assistance Systems
  • Toward Predictable, Efficient, System-Level Tolerance of Transient Faults
  • A New Concept for System-Level Design of Runtime Reconfigurable Real-Time Systems
  • Reconfigurable Industrial Process Monitoring using the CHROMOSOME Middleware

二月 17

CFP: ACM SenSys 2014

ACM SenSys 2014

Memphis, TN, USA, November 3-6, 2014



C A L L  F O R   P A P E R S


Sensors have become an essential part of computing systems and applications. Computing today is increasingly characterized by ubiquitous, information-rich sensors that produce massive quantities of data about the physical world. This new era of computing is driving important new systems issues, and requires new system-level approaches and design principles.

The ACM Conference on Embedded Networked Sensor Systems (SenSys 2014) is a computer systems conference focused on the architecture, design, implementation, and performance of sensor systems. ACM SenSys brings together academic, industry, and government professionals to a single-track, highly selective forum on sensor network design, implementation, and application. It is the premier forum to discuss systems issues that arise specifically due to sensing. SenSys takes a broad view on the areas of computing that are relevant to the future of sensor systems, and topics of interest include but are not limited to the following:

Topics of Interest

. Compelling challenge papers grounded in technology trends
. Applications and deployment experiences
. Knowledge discovery from sensor data
. Emerging sensor systems, such as Kinect, LIDAR, and cameras
. RFID computation, communications, storage, and networking
. Mobile and wearable sensing
. Ubiquitous and pervasive sensing
. Internet of Things, Cyber-Physical Systems, and Sensor Swarms
. Software for sensor systems
. Communication and networking for sensor systems
. Sensor context such as time and location estimation
. Energy harvesting and management for long-term sensor operation
. Storage, retrieval, processing, and management of sensor data
. New sensor technology and hardware designs
. Fault-tolerance and reliability of sensor systems
. Sensor data quality, integrity, and trustworthiness
. Security and privacy of sensor systems

We invite technical papers describing original ideas, ground-breaking results, and/or real-world experiences involving innovative sensor systems. Successful submissions will explain why the topic is relevant to a vision of the future of sensing systems. Submissions will be judged on originality, significance, clarity, relevance, and correctness. In addition to citing relevant, published work, authors must cite and relate their submissions to relevant prior publications of their own.

General Chair
Akos Ledeczi (Vanderbilt University)

Program Chairs
Prabal Dutta (University of Michigan)
Chenyang Lu (Washington University in St. Louis)

Submission Guidelines

Submissions must be full papers, at most 14 single-spaced 8.5″ x 11″ pages, including figures, tables, and references, two-column format, using 10-point type on 11-point (tight single-spaced) leading, with a maximum text block of 7″ wide x 9″ deep with an intercolumn spacing of .25″. Authors must make a good faith effort to anonymize their submissions. Papers that do not meet the size, formatting, and anonymization requirements will not be reviewed. Accepted submissions will be available on the ACM digital library at least one week before the conference.


二月 17

Prepare your research thesis

Guidance of preparing your research thesisThesis

This article tries to illustrate my guidance for preparing research thesis for master students. It serves the guidance and minimal requirements for the students in my research group. Although I try my best to assure that my minimal requirements also meet NTU’s requirements on research thesis, I may miss some.  In case of submitting thesis to NTU for graduation, one should refer to NTU’s requirements on research thesis. (In particular, NTU has clear guidance on thesis format and one should follow NTU’s guidance. This article provides guidance on content, not format.)

The guidance are written for each component of a research thesis: overall requirements, introduction chapter, literature review and background, system architecture, designed algorithm, and evaluation.

Overall Requirements:

All the materials should be prepared in portable and editable formats. To be specific,

  • All the writing must be formatted in LaTeX or pure text. The information on tex can be found on tex-live site.
    • (NEW) You can also use ShareLaTeX(www.sharelatex.com) to use web browser as an editor for collaborative latex edit.
  • All the figures must be editable in SVG format, excepting screenshot. A list of SVG editors on different platforms can be founded on wiki site. If you use Mac, I recommend OmniGraffle and InkScape.
  • All the experiments should come with raw data and source code.
  • The experiment results should be reproducible from raw data. The recommended way is using GNUPLOT scripts to visualize the experiment results.

Not to do:

  • Use powerpoint to draw and convert to eps/pdf. The size of these files are unnecessary huge and misses certain fonts for publication.
  • Use excel to draw the experiment results. The format of these figures are terrible: extreme thin lines, wasted white space on the sides, and poor quality when scaling it up.

Introduction Chapter

The purpose of this chapter is to motivate the readers to finish reading your thesis. Hence, Introduction chapter (Chapter 1) usually illustrates the motivation, contribution, and overview of the research thesis. You should assume that the readers only have a bachelor degree in computer science and avoid using any math equations in this chapter. Here are few guidance to prepare this chapter.

  • Motivation: one should use certain applications to motivate the researches. Although the applications can be real application or applications in future, one should clearly describe the use scenario and/or challenges of these applications. At least two applications should be provides.
  • Challenges: one should describe the challenges or difficulties to provide or realize the application scenario described in motivation. One should describe all the challenges but only targets few of them in the thesis.
  • Contributions: contribution refers to how the described works change the world or the development of the technology. It does not refer to what algorithm or application you developed in your thesis. As an example, what’s the contribution of google search engine. Its contribution is NOT a cluster of computers and ranking algorithm. Its contribution is allowing one to find the information scattered around the world in front of the computer.
  • Thesis organization: in a thesis, each chapter deserves one paragraph to illustrate the goal of other chapters in the thesis.

Not to do in Introduction Chapter:

  • Your research target is chose because of the recent/rapid development of technology.
  • Collecting whatever you do, including algorithm, software, and application, to be the contribution of the thesis.
  • Listing the title of each chapter to describe thesis organization.

Literature review and Background

The purpose of the second chapter is to bridge the readers and you so that they have the capability of understanding your works if they decide to read your thesis. Hence, it usually provides the literature survey and background. If the readers are familiar with the targeted challenges described in the introduction chapter, they may skip this chapter. This chapter also demonstrate how well you understand the targeted challenges and state-of-the-art.

To prepare the background, you should assume that the readers have at least the bachelor degree in computer science. Hence, there is no need to repeat to content in textbooks. You should describe the following:

  • technology or concept specific to the targeted applications.
  • technology developed in last few years or not presented in textbook.

Literature review is a major task in doing research. It not only demonstrates how well you understand the challenges but also demonstrates why your work is in needed. Without a thorough literature review, one SHOULD NOT start his/her researches.

(To be finished)

Not to do in literature review:

  • Copy/paste either abstract, introduction, or summary of related works to your thesis. (This violates the copyright and research ethics.)
  • Only summarizes the related works.
  • Criticizes what are missing in the related works in bad wording.

(To be continued.)

舊文章 «