Chung Cheng University Institutional Repository
Not a member yet
889 research outputs found
Sort by
雲端控制與無線家用路由器基於IoTivity框架之管理方法;Management Method of Cloud Access Controller Wireless Home Gateway Based on IoTivity Framework
No full text[[abstract]]隨著Internet的興起，網絡人口不斷增長，勢不可擋。大量的Internet連網服務已經擴展到家庭網絡。基於上述趨勢，對廣泛使用的家庭/住宅 路由器 (IP 分享器) 的效率運行和管理都成為非常重要的角色，尤其以家用路由器的管理。近年來，所有類型的物聯網應用的發展正在迅速發展。本文的主要目的是開發雲端控制器 Cloud Access Controller（Cloud AC）與家庭網關之間的物聯網通信架構。 IoTivity是一個開放源始碼的軟體框架，可以實現無縫設備到設備 (Device to Device) 的連接，並滿足物聯網的新興需求。本文實現了IoTivity和資料庫授權(Authentication) ，增加了Wireless Home Gateway和Cloud AC之間的安全認證機制。本研究結束之後，在Wireless Home Gateway和Cloud AC之間建立了一種高效安全的通信架構，包括自動開通(Auto Provision) ，註冊 (Registration) ，認證 (Authentication)，連接 (Connection) 和管理 (Management) ，以及具有IoTivity的大規模部署機制，顯示不僅提供了比以前更安全和更好的性能優勢，並以Cloud AC 平台提供快速安裝大量部署，安全管控等…管理Wireless Home Gateway的完整系統。
Following the rise of the Internet, the networking population grows continuously and is unstoppable. A lot of Internet accessing services have been extended to home networks. Based on the above trend, not only efficiency operation but excellent management to be widely used Home / Residential Gateway becomes a very important topic in the Internet. In recent years, the development of all types of IoT application is growing rapidly. The main purpose of this study is to develop a comprehensive communication architecture of IoT connection between Cloud Access Controller (Cloud AC) and Home Gateways. That is why the IoTivity is adopted in this study to enable the seamless device-to-device connectivity possible and satisfy the emerging requirement of IoT. The IoTivity and Database authorization are developed in this study to improve the security and authentication between Wireless Home Gateways and Cloud AC. Finally, an efficient and secure communication architecture between Wireless Home Gateway and Cloud AC is realized, including the auto provision, registration, authentication, connection and management. At the same time, the large scale deployment mechanisms based on IoTivity framework which obviously doesn’t only provides advanced secure and performance, but also more mature operation design for large scale wireless networking in the near future
先進FPGA之延遲變異評估;Assessment of Timing Variations in Modern FPGA
No full text[[abstract]]現場可程式化閘陣列（field-programmable gate array；FPGA）除了作為ASIC之快速雛型設計以外，近年更被廣泛應用在各式系統中，因為市場需求量大更成為新一代的製程先驅（technology driver），先進製程提供高效能、低功耗等優點，但製程微縮同時也造成嚴重變異。本論文提出以一般用於自適性電壓調控之時序餘裕偵測器（timing margin detector；TMD）來量測資料路徑的延遲變異。在28nm FPGA上可觀察到1.27%的WID（within-die）及2.04%的D2D（die-to-die）的延遲變異量。另外，我們也提出新式的TMD設計：pulse generation flip-flop （PGFF），與現有的DTFF（double-edge-triggered flip-flop）與PGL（pulse-generation latch）設計比較，可分別省下9.8%、7.3%硬體資源。最後與保守的FPGA設計工具比較，TMD可協助探勘佔整體時脈32.5%的動態時脈餘裕（dynamic timing slack），約等同於25.6%的功率消耗。
FPGA (field-programmable gate array) has been widely used in various systems, beyond its original application in ASIC prototyping, and the high volume has pushed it as the new technology driver. Advanced technologies bring the advantages on performance and power dissipation but also incur significant design variations due to shrink device sizes. This thesis presents the characterization of delay variations on datapaths using timing margin detector (TMD) originally designed for adaptive voltage scaling. 1.27% WID (within die) and 2.04% D2D (die to die) delay variations are observed on 28nm FPGA. Moreover, a new TMD design: PGFF (pulse-generation latch) has been proposed, of which the hardware resources can be reduced by 9.8% and 7.3% respectively, compared with conventional DTFF (double-edge-triggered flip-flop) and PGL (pulse-generation latch) TMD designs. Finally, the proposed method can explores 32.5% dynamic timing slack from the conservative cycle time estimated by the FPGA designs tools, which can be approximated as 25.6% power savings
微處理器動態路徑中因變異引發之時脈餘裕測量;On-Chip Measurement of Variation-Induced Dynamic Path Slacks in Microprocessors
No full text[[abstract]]先進製程中電路速度受製程(process)、電壓(voltage)、溫度(temperature)以及資料(data)的影響極大，尤其是區域變異(local variations；within-die (WID))及供電網路因電流變化造成動態IR drop讓時序模擬幾乎不可能。本論文實作一般用於自適性電壓調控(adaptive voltage scaling；AVS)之時脈餘裕偵測器(timing margin detector；TMD)在微處理器資料路徑中，藉以量測電路在不同工作負擔(workload)下的實際時脈餘裕。量測數據也可與處理器之指令同步，回饋至指令模擬器進行比對分析。最後對於處理器時脈與面積之額外負擔也以類似scan之序列連接有效控制，以實現在Xilinx Atrix7 FPGA之MIPSfpga核心為例，其面積與時脈負擔各自為12.5%與2.1%。
PVTD (process, voltage, temperature and data patterns) variations effect strongly in advanced technology. It’s almost impossible to run timing simulation because the local variations (within-die (WID)) and dynamic IR drop cause by current varying on the supply network. This paper implement timing margin detector (TMD), which is generally used in adaptive voltage scaling (AVS), to detect real timing slack under different workload on the microprocessor’s data path. By synchronizing the measured result and instructions, it’s able to analyze the result in the instruction simulator. The timing and area overhead can be effectively decreased by serial scan chain. On our implement MIPSfpga design using Xilinx Atrix7 FPGA, the timing and area overhead are 12.5% and 2.1%
加速比特幣區塊鏈計算的新穎方法;A novel approach to accelerate the computation of the bitcoin blockchain
No full text[[abstract]]隨著比特幣蓬勃發展，漸漸讓大家知道電子貨幣，由於比特幣是虛擬的，嚴格說起來他連貨幣都不太算，在比特幣的世界裡面，底層有一個公開透明的帳本，紀錄著每筆發生過的交易，也就是其中最重要的技術，名字就叫區塊鏈 (blockchain)，顧名思義它就是一個長長的鏈，由區塊 (block) 一個接著一個所組成的，而每個區塊裡面記錄著很多發生過的交易 (transaction)。在比特幣交易中主要會由兩個部分構成，分別是 inputs 和 outputs，每筆交易都有若干交易輸入，也就是資金來源，也都有若干筆交易輸出，也就是資金去向。一般來說，每一筆交易都要花費（spend）一筆輸入，產生一筆輸出，而其所產生的輸出，就是「未花費過的交易輸出」，也就是 UTXO，而交易的inputs就是由一堆UTXO (Unspent Transaction Output)的組合所組成的，而如何讓交易更快速知道選擇哪些UTXO來支付，也就是本論文所要研究的目標。本論文希望設計出一個更快速選擇出要支付的UTXO組合，因此本論文所提出的方法可解決原來比特幣選擇UTXO的複雜判斷條件的問題，利用一些技術來達成選擇交易支出的組合，之後也與原來比特幣選擇交易UTXO演算法進行比較，實驗結果顯示所有評估標準較佳於原比特幣所提出之方法
設計與實現基於叢集圖像晶片處理器之影片物件辨識系統;Design and Implementation of Video Object Detection System Based on Cluster Graphic Processing Unit
No full text[[abstract]]本論文設計一叢集式影片物件辨識平台，藉由NVidia 公司所推出的TX1嵌入式開發板所提供的圖形處理器(GPU)進行OpenCV影像辨識，並使用Python的分散式套件dispy在多台機器上進行分散運算。因圖形處理器其結構相異於傳統中央處理器(CPU)，對於多媒體資料具有相當高的運算效率。藉由dispy套件的串聯，使得多台機器能夠同時進行平行運算。本論文同時設計針對辨識結果進行權重統計之方法，以符合人類在主觀意識上對於影片內容之想法，並進行不同影格之間之相似比對，用於過濾相似影格，節省運算總量，最後對於進行加速時所造成的效能損失進行探討。本論文實作之平台，在無失真之情況下相較於傳統辨識方法可以獲得12倍的效能提升，在允許部分失真之情況下，忽略 20%影格差異能達到約45倍效能提升，忽略 40%影格差異能達到約85倍效能提升，大幅度降低多重物件辨識所需時間，使得相關領域能夠大幅度節省辨識所花費的時間。
Designing an object detection platform cluster by using Jetson TX1, which developed by NVidia. Jetson TX1 provides an environment with Graphics processing unit(GPU) supporting CUDA. By the GPU we can process OpenCV object detection faster than CPU. We can also connect these development kits by using Python extension – dispy, and make development kits parallel compute the data. GPU is quite different than CPU, its’ multi-core can process media data with high-performance.We also design an algorithm to calculate detection result, according to its’ frequency and multiplied by weights. Filtering similar frame can also reduce process time. And finally make a discussion about performance loss. The platform we make can get 12 time faster than traditional detection. Ignoring 20% of frame difference can get 45 time faster. Ignoring 40% of frame difference can get 85 time faster. By our experiment, this platform can significantly reduce detection time
將C語言程式轉換為具有執行緒層級之備援執行程式模型之研究;A Study on Translating C Code to Software-based Thread-Level Redundancy Programming Model
No full text[[abstract]]隨著硬體技術不斷進步， 電晶體的體積越做越小，使得同一個晶片內的晶體密度得以增加。但這個趨勢會增加我們在運算時出現軟體錯誤的機率。因此，關於計算機容錯機制研究的重要性與日俱增。本篇論文主要探討的目標為軟體錯誤之中的瞬時故障，這是一種無法預測何時會發生的錯誤，而且並不會造成硬體或軟體上永久性的損壞。對一般使用者來說，容錯機制通常會採用冗餘執行的方式，把重要的計算部分多做幾次來保證結果的正確性，但是設計一套具有容錯能力的程式是很複雜的。 在本篇論文中，我們設計了一個能將C語言程式轉換為相對應的執行續層級冗餘計算程式模型轉換器。我們利用編譯器技術來完成這個程式轉換器的實作。而實驗結果顯示，本篇論文提出的轉換器可以正確的將所有的測試程式轉換為具有自我保護能力的程式。
With the new advances in computer hardware technologies, transistor size is shrinking while its density of transistor on a single chip is growing. This trend increases soft error rate (SER) of computer systems. The transient fault is one of soft errors, which occurs at unpredictable times and does not result in permanent hardware or software damage. In order to tolerate the transient faults, users usually need to leverage redundant execution technique to protect critical computations. However, writing such programs is complicated. In this thesis, we develop a translator that translates C codes to the corresponding thread-level redundant execution programming model. The front-end compiler of LLVM compiler infrastructure, Clang is used to implement the translator. For the tested benchmark programs, the experimental results show that the proposed translator can correctly generate protected codes
工會與國際併購決策分析;Labor Union and International Merger
No full text[[abstract]]本論文使用子賽局完美均衡(subgame perfect equilibrium)概念，並用倒解法(backward induction)求得廠商的最適的庫諾數量均衡，分析工會對兩國廠商併購的影響。首先，我們從資方進行分析：A國廠商的工會先決定工資，之後廠商再決定在這個工資水準下併購是否有利可圖，最後在B國市場進行數量競爭，其中，A國工會能不能透過工資的調整影響廠商的決策，是我們研究的重點之一。另一方面，我們從受資方廠商工會的角度進行分析：在B國廠商面臨被併購的情況時，B國工會能不能透過工資的調整，來拒絕或是促使B國廠商進行併購；而在兩種情況下，社會福利又是如何變動，則是我們另一個重點。經過模型的推導，我們發現當廠商的產品替代性低時，工會完全無法阻止廠商併購；如果廠商的產品在中等的替代性程度，工會就應該使用策略性的調高工資，阻止廠商併購；只有在產品替代性高時，工會就不應該使用策略性方法阻止廠商併購，否則會降低自己的福利。而對被併購的工會而言，當產品差異程度較低時，接受併購所帶來的，可以說是利大於弊，工資水準雖然必須調降，但產量的增加帶來的就業量完全可以彌補工會的損失，所以在產品同質性高的情況下，工會一定會接受國外廠商對於國內廠商的併購。除此之外，因為併購會帶來社會福利的上升，所以政府對於併購也需要一些政策來鼓勵
網路創業成功關鍵因素之研究─以教育部U-start計畫獲補助新創公司之Facebook粉絲專頁為例;A Study of Key Factors for Success of Internet Venturing : A Case of Subsidized Start-ups by U-start Plan via Facebook Fan-pages
No full text[[abstract]]本研究旨在刻畫出良好評斷出網路創業成功關鍵要素之方法，在網路社群平台蓬勃發展下，冀望能在未得知該公司營利數據資料下，仍可藉由網路社群的經營與發展勾勒出其創業之成功程度以及影響成功的關鍵因素。 此次研究以「大專畢業生創業服務計畫」中獲補助新創公司之Facebook粉絲專頁作為探究網路創業成功的樣本，共75家新創公司之有效樣本，並定義「粉絲專頁主頁按讚人數」為衡量「成功」的標準。計量分析方法以普通最小平方法為主，並使用主成分分析法作為第二種評估資料的作法。 研究蒐集各新創公司粉絲專頁的按讚數、留言數、分享數、貼文數四個主要衡量成功的指標。本研究將蒐集的四個指標作為量化屬質文獻中的創新度、忠誠度之憑據，發展出本研究的創新度與忠誠度，並加入資本額、經營存續時間後提出研究假說，再採線性結構方程式驗證之。研究結果顯示：(1)創新度正向顯著影響成功；(2)忠誠度正向顯著影響成功；(3)資本額負向不顯著影響成功；(4)經營存續時間正向不顯著影響成功。 研究主要貢獻在於除提供問卷調查之外的一種新方法─蒐集網路行為之數據，透過蒐集使用者在Facebook上按讚數、留言數、分享數，讓研究能夠更直接地觀察使用者的行為，再透過不同的行為去找出使用者的偏好，最後透過偏好進行分析。
This research studies the key factors for success of internet venturing,hoping to find the effect of key factors on the degree of success without observing income data of companies. The data used in this research are Facebook Fan-pages of Subsidized Start-ups by U-start plan. There are seventy-five samples of companies in the data set. The research defines Home-page Like Count as Success, and uses ordinary least squares and principal components analysis as research methods. The research collects four main indexes, which are Like Count, Share Count, Comment Count, and Post Count. The purpose of the research is first to quantify Loyalty and Innovation that both have been discussed by former scholars, and then to develop new explanatory variables belonging to our own. Moreover, the research adds Capital and Duration Time to the linear regression to analyze their effects and comes up with six hypotheses. The results are: (1)Innovation has significantly positive impact on Success. (2)Loyalty has significantly positive impact on Success. (3)Capital has negative impact on Success with no significance. (4) Duration Time has positive impact on Success with no significance. The main contribution of the research is to provide a new method, which is collecting data on network site besides filling questionnaires. By collecting Like Count, Comment Count and Share Count that revealed on Facebook Fan-pages, the research directly observes actions made by Facebook users. Also we can figure out the different preferences of Facebook users through their actions. Finally, the research can generate the conclusion by Facebook users' preferences
個人股東可扣抵稅額減半與現金減資：公司特性差異;The Relationship between Halved Imputation Tax Credit and Capital Reduction By Cash：Company Characteristics difference
No full text[[abstract]]我國自2015年1月1日起適用本國個人股東及國外股東獲配股利淨額之可扣抵稅額減半，及綜合所得稅課稅級距由原本五級修正為六級，最高稅率由原先之40%提高調整至45%，新增綜合所得稅淨額超過1,000萬部分，適用稅率為45%。本研究主要探討因應新稅制下，公司是否為了大股東抵稅權益而執行現金減資，及執行現金減資之公司特性是否有所不同，本研究係以新稅制後2015年至2016年間經金融監督管理委員會證?期貨局受理申報(請)案件情形之辦理現金減資生效結案類型者之上市上櫃公司為研究對象，以羅吉斯迴歸分析現金減資之其公司特性。實證結果發現，當公司的董事會成員中個人大股東的比率較高時，較傾向採取現金減資決策，表示大股東之稅負因素會影響公司是否執行現金減資；另淨值成長率與減資決策為顯著負相關，則代表當公司規模達一定程度，其所屬之產業已進入成熟階段其市場趨近飽和狀態，以致於淨值成長率下降之情形下，公司將傾向以執行現金減資予以提高淨值成長率。關鍵詞：兩稅合一、股東可扣抵稅額減半、現金減資、公司特性
Application for the imputed tax credit of resident individual and foreign shareholders has been halved since January 1, 2015. The tax bracket of individual income tax has been amended from five level to six and the tax rate has raised to 45% from 40%. The new regulation has set up for the net consolidated income exceeding 10 million, the exceeded part shall be applicable to the tax rate of 45%. This study focuses on whether the company implement cash distribution for the major shareholders’ imputation tax credit in the new tax system. Furthermore, analyze firm characteristics of listed companies have made cash distribution declaration to Securities and Futures Bureau, Financial Supervisory Commission from 2015 to 2016 by Logistic regression.The empirical results show that the companies with higher proportion of major shareholders in the board, higher scale and higher developing opportunity apt to adopt the strategies of cash distribution. This indicates that the factor of the major shareholders’ individual income tax will affect the strategies. In addition, the growth rate of net value has significant negative correlation with the strategies of cash distribution. When the scale of company reaches a certain degree, the industry it belonging to has become mature and the market has reached saturation point. The company tends to implement cash distribution to improve when the growth rate of net value declines.Key words: Integrated Income Tax System、Imputation Tax Credit、Cash Distribution、Company Characteristic
法人說明會影音資訊與應計項目盈餘管理之關聯性;
No full text[[abstract]]本研究旨在探討法人說明會影音資訊之時間長度及用詞傾向與公司應計項目盈餘管理行為之關聯。實證結果顯示，企業過去的應計項目盈餘管理行為與法說會的總時間長度及問答階段時間長度呈顯著正相關，說明公司過去盈餘管理上調越多，公司在法說會中所耗費的時間越多，特別是在問答階段的時候。此外，本研究亦探討法說會用詞傾向對公司後續應計項目盈餘管理行為的影響，結果顯示兩者呈顯著正相關，特別是在法說會中的陳述階段的用詞，說明了公司在法說會中的用詞越樂觀越正面時，其越可能增加其盈餘管理程度來達成期望。
The purpose of this thesis is to explore the relation between conference call video information and the accrual-based earnings management, using the time of conference call and the tone of conference call as research variables. The empirical results show that there is a significant positive correlation between the time of conference call and the previous accrual-based earnings management, especially in QA phase. It shows that the more accrual-based earnings management the company did, the longer time company spent in conference call. In addition, this study explores whether the tone of conference call impacts on the accrual-based earnings management. The results illustrate that there is a significant positive correlation, especially in the statement phase. It means that when the company tends to use more optimistic and positive tone, the company is more likely to achieve expectations through upward earnings management