搜尋功能  
   

    • 主題分類
    • 篇名
    • 作者
    • 下載
    • 人物專訪
    • 人工智慧國防運用先驅－－專訪交通大學前校長張懋中博士
    • 張懋中／口述‧張素美、楊淯丰／整理‧林大秋／拍攝

    卷期 ： 47 / 4

    出版年 ： 2019/10

    近年來人工智慧(Artificial Intelligence, AI)技術運用是人人耳熟能詳，透過此技術訓練機器具有如同人一般的邏輯思考與行為反應能力。在機器內部置入有如人類的自我感知、學習、校正與推理等功能模組，使機器具備可依指令下達執行交付任務。讓機器具備擬人化功能是一件相當困難的事情，但科技日新月異，目前人工智慧在軍事科技運用上均有明顯進展。伴隨著仿腦晶元、人機交互技術、腦機介面技術等領域的突破，可能帶來真正的革命性進展。如今透過人工智慧技術結合大數據信息處理等，已能處理包含視覺（人臉辨識）、聽覺（語音識別）與觸覺等對外界的感官知覺，目前應用範圍也擴及機器人、智慧型機器、手機、具休閒、運動保健等穿戴裝置、智慧車、無人機、無人載具等等。而世界科技大國、國際大廠與新創業者們，亦均不斷投注預算與精力在防範人們因執行危險任務或取代人類因先天受限而無法執行的「智慧機器」研究領域。

    關鍵詞 ： 5G,人工智慧

    • 中心論題
    • 「人工智慧科技與智慧國防運用」序言
    • 林高洲(Gau-Joe Lin)‧

    卷期 ： 47 / 4

    出版年 ： 2019/10

    眾所皆知，5G結合人工智慧（Artificial Intelligence, AI）科技，將以強大的高速連網、寬頻傳輸及智慧化處理能力，在金融服務、工業應用、醫療照護、教育娛樂、公共安全及能源與交通運輸等方面，利用AI的管理、監控、學習、分析等能力，改善系統效能並提高運作效率，加速智慧應用的普及塑造出智慧型、專屬客製化的互動模式，滿足每個人、不同應用層面的需求和要求；因此AI對國家、社會的影響將遠大於過往各類型的新興科技。於此同時，AI科技也改變國防建軍備戰的思維，世界科技大國不約而同將AI列為國防重點發展方向，期能在下一波軍事革新上取得科技領先與軍事優勢；AI納入國防運用，主要是著重於「智慧國防」的建置，特別在不對稱通資電戰力方面，全面聚焦於「自我感知、自主學習、智慧決策及快速反應」的運用和處理能力，期能即時掌握敵我態勢、高效發揮聯合戰力、自動確認攻擊目標和精準戰果評估分析。

    關鍵詞 ： 人工智慧

    • 中心論題
    • 運用人工智慧建置智慧國防不對稱戰力發展芻議A Preliminary Study of Building an Artificial-Intelligence-based Smart Defense for Asymmetric Warfare
    • 林高洲(Gau-Joe Lin)‧楊淑君(Shu-Chun Yang)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    不對稱作戰之勝算與智慧科技自主程度息息相關，而智慧科技的核心就是人工智慧(AI)；近年來，AI已帶動戰場的用兵觀念轉變為自動化、智慧化和無人化的戰爭模式，也改變國與國之間的產業與經濟的競爭模式，各國皆針對人工智慧研發佈局投入龐大資源，期能在競爭的浪潮中，取得先機、掌握智慧科技的優勢。基於國軍推動創新不對稱戰力提升研發已頗具成效，提出運用AI建置智慧國防不對稱戰力發展芻議，以資通電智慧科技運用構想出發，建置開放式「智慧國防研發平台」架構，著重於掌握應用核心技術，結合軍民科技能量，持續創新思維與前瞻技術發展，戮力強化並精進智慧科技的研發能量和跨領域的整合運用，完成智慧國防不對稱戰力的提升。Prevailing asymmetric warfare is intimately connected to the levels of autonomous smart technology, which is in turn the core of artificial intelligence. In recent years, AI is driving the military operation concepts that are becoming automated, intelligent, being unmanned battlefield, and changing the types of industrial and economic competition worldwide. Thus, most countries have been significantly invested resources in the AI R&D strategic planning intending to direct the development of smart technology and achieve the global competitive advantages. Having been obtaining considerable effectiveness of building the capabilities of innovative and asymmetric warfare of our armed forces, for which we propose to establish an open framework of smart defense development employing AI by integrating the smart technologies of spectrum management, intelligent communications, cognitive radio, cloud computing, and information security. Moreover, combining the civilian capabilities, we will continuously develop innovative and forward-looking technologies including the research of smart defense, the cross-field integration, to enhance the defense forces for asymmetric warfare.

    關鍵詞 ： 人工智慧(artificial intelligence, AI)、智慧國防(smart defense)、不對稱作戰(asymmetric warfare)、資通訊科技(communication and information technology)、感知無線電(cognitive radio)

    • 中心論題
    • 人工智慧理論與趨勢Theories and Trends of Artificial Intelligence
    • 郭立言(Li-Yen Kuo)‧湯士堅(Shi-Chian Tan) ‧沈辰峯(Cheng-Feng Shen)‧林品均(Pin-Jun Lin)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    人工智慧已然成為新潮流，在生活的各層面均改變了人們的習慣，並揭櫫了開放、共享的社群生態。本文中，將簡述人工智慧的歷史與演進，其後就臺灣與本院面臨的資料缺乏之現況與端運算之需求，介紹生成對抗網路、主動學習與模型壓縮的理論與趨勢。最後，在結語中陳述了本院發展人工智慧應用的機會、缺口與方向。The artificial intelligence has been a buzzword in our daily lives changing people's behavior and presenting an open and shared ecosystem. In this paper, the history and progress of AI will be described, and followed by the introduction to the GANs, active learning and model compression under the conditions of lacking data and demand of edge computing. At last, it is recommended to review the opportunity, research gaps and direction of AI development for NCSIST.

    關鍵詞 ： 關鍵詞︰人工智慧(artificial intelligence)、機器學習(machine learning)

    • 中心論題
    • 人工智慧應用 Applications of Artificial Intelligence
    • 劉于碩(Yu-Shuo Liu)‧陳韻如(Yun-Ru Chen)‧戴君翰(Chun-Han Tai).林余峻(Yu-Chun Lin)‧葉佳龍(Chia-Lung Yeh)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    「智慧」可以突破自然界的侷限，並能以本身的想法改變世界。人工智慧的目的是為了讓電腦能夠模仿人類，根據環境感知理解周圍，並採取行動以達成目標，而人工智慧從過去發展至今，在數種工作中替代了人類，但仍有許多問題須依靠人類智慧才能解決。隨著科技日新月異及網際網路的興起，所帶來的巨量資訊深刻影響著我們的生活，以及計算機效能演進，從過去傳統序向中央處理單元到平行高效運算單元，使得許多新穎的機器學習技術得以發展與應用。其中最具代表性為深度學習的崛起，在許多領域中皆產生顯著大幅的躍進。本文以「人工智慧」領域之深度學習技術為主軸，探討影像辨識、自然語言處理與腦機介面等關鍵技術。Intelligence transcends the limitations of human imagination that will changes the world. The purpose of artificial intelligence is making machines that imitate human being, can sense surroundings and act to reach goals. Nowadays, even artificial intelligence has been developed for some skills that substituted human employees, a lot of problems still need to rely on human intelligence itself. Moreover, with the ever-increasing speed of technology changes and the emergence of websites worldwide, the massive information has deeply affected our lives. The progress of big data analytics and high-performance computing, from central processing unit to parallel computing, have made development and application of new machine learning techniques, of which the most significant one is the rise of deep learning that has led to a huge leap in many related areas as well. This article emphasizes on the deep learning in expert systems development including image recognition, natural language processing, and brain-computer interfaces.

    關鍵詞 ： 人工智慧(artificial intelligence, AI)、大數據(big data)、影像辨識(image recognition)、自然語言處理(natural language processing, NLP)、腦機介面(brain-computer interface, BCI)

    • 中心論題
    • GPU系統開發環境簡介Introduction to GPU System Development Environment
    • 李樂賓(Yueh-Bin Lee)‧簡國芳(Kuo-Fang Chien)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    人工智慧是目前熱門的研究技術，而GPU因具備平行運算能力，非常適合人工智慧的深度學習神經網絡運算，其運算能力比較一般的CPU，可提升數十倍，甚至百倍。NVIDIA是全球GPU領導廠商。它的GPU含蓋了行動裝置、車用、桌上型、工作站、伺服器等解決方案；其發展軟體環境包含深度學習框架(Framework)、SDK(System Development Kit)及CUDA ToolKit等。本文針對Nvidia的硬體產品及軟體發展環境作簡要介紹，期能讓初入GPU領域的讀者有所幫助。Artificial Intelligence is one of the current hot topics in computer science. Featuring the parallel computing, the graphics processing unit (GPU) is a suitable solution for AI deep learning. NVIDIA, a global leader of GPU chip manufacturers, provides GPU hardware solutions for various platforms, such as mobile device, desktop, workstation, servers; as well as those for software including deep learning framework, system development kit (SDK) and CUDA toolkit. This article briefly introduces the NVIDIA's hardware products and software development environments to facilitate the study for readers interested in the GPU field.

    關鍵詞 ： 圖像處理單元(graphic processing unit, GPU)、層積神經網路(convolutional neutral network, CNN)、統一計算架構深度神經網路(compute unified deep neutral network, cuDNN)、統一計算架構(compute unified device architecture, CUDA)、人工智慧(artificial intelligence, AI)

    • 中心論題
    • 認知無線電於動態頻譜管理之研究Research of Cognitive Radio Application for Dynamic Spectrum Management
    • 廖恪應(Ke-Ying Liao)‧林士閔(Shih-Min Lin)‧陳德育(Te-Yu Chen)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    由於頻譜的使用率日益增加，在資源有限的情形下，必須找出有效提升頻譜使用率的方法。隨著無線電新技術的發展，認知無線電的出現，使得動態頻譜共享成為可能，使用者可在不同時間、頻段和地理位置等多維度上共享頻譜，實現頻譜資源的再利用與增進使用效率，解決當前頻譜資源匱乏的問題。It is increasingly using the spectrum. Under resources constraints, it is necessary to obtain a solution to enhance the efficiency of spectrum uses. With the advancing development of radio technology and the emergence of cognitive radio, it is made possible to share dynamic spectrum at different times, frequency bands, and locations to deal with the lack of resources, and implement spectrum reuse and efficiency enhancement.

    關鍵詞 ： 認知無線電(cognitive radio)、動態頻譜管理(dynamic spectrum management)

    • 中心論題
    • 人工智慧於網路攻擊之應用Applying Artificial Intelligence Technologies to Cyber Attack
    • 高永威(Yung-Wei Kao)‧徐丞謙(Cheng-Qian Xu) ‧孫敬智(Ching-Chuh Sun)‧趙亞略(Ya-Lyue Jhao)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    近年來，由於人工智慧技術的突破，引起各界熱烈討論人工智慧技術運用於各種不同領域（包括資訊安全領域）之可行型。傳統資訊安全相關研究多以探討如何使用人工智慧技術進行惡意程式或行為之辨識或分群。本文則著重於探討如何利用人工智慧技術於資安攻擊之應用中。The breakthrough of Artificial Intelligence has been increasingly discussed in terms of the feasibility of application in various fields, including cybersecurity, and becoming people's common interests. Conventionally, it is more frequently investigated on how to identify or group malware and malicious behaviors applying AI. However, in this paper, how AI technologies are applied to cybersecurity attacks in defensive scenarios are discussed.

    關鍵詞 ： 人工智慧(artificial intelligence)、網路攻擊(cybersecurityattack)

    • 中心論題
    • 淺談人工智慧在資安攻防發展之應用趨勢An introduction to the trend of Cyberscurity of Artificial Intelligence
    • 江啟賓(Ci-Bin Jiang)‧張家駿(Jia-Jun Zhang)‧呂臣恩(Chen-En Lu)‧賴均霖(Chun-Lin Lai)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    自AlphaGo橫空出世打敗世界棋王後，有長足發展的人工智慧再度成為資訊領域一項炙手可熱的技術。近年重大資安事件層出不窮，網路犯罪的手法日新月異且規模日益擴大。人工智慧在資安領域上扮演著輔助偵測之角色，主要還是要從專家經驗、惡意程式行為分析、寫偵測規則，到後續人工智慧偵測學習模型，以達到自動化偵測。本研究將帶領讀者一窺究竟人工智慧是怎麼被運用在網路戰的攻擊方與防守方，提供給資安研究人員一份參考，了解最新人工智慧資安攻防發展之應用趨勢。Since AlphaGo defeated the world chess king, the artificial intelligence is getting hotter, even it has been developing for a very long time in the information field. In recent years, cybersecurity incidents have been emerging endlessly, and more severely, the cybercrime technique are increasingly changing and expanding. When it comes to mentioning artificial intelligence to play a role of detection in cybersecurity, the expert experience, malware behavior analysis, detection rules, and detection models will be the major components of AI to achieve automatic detection. This paper instructs readers to see how artificial intelligence is applied to the cyber attackers and defenders.

    關鍵詞 ： 人工智慧(artificial intelligence, AI)、機器學習(machine learning, ML)、網路安全(cybersecurity)、惡意程式(malware)

    • 中心論題
    • 智慧化決策指管系統發展趨勢探討Investigation and Development Trend of Intelligent Decision Support for C2 Systems
    • 曾國銘(Kuo-Ming Tseng)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    在錯綜複雜的戰場中，敏捷力是指揮管制一項重要能力。「指管敏捷力」是北約組織軍事研究熱門話題，它所探討範圍主要著重於組織結構對戰場的應變能力。但自從 AlphaGo 機器擊敗棋士高手，人工智慧與資訊技術不斷地迅速發展與演進，未來的戰爭將是人類高度善用智慧機器之型態。因此指管敏捷力不但提昇決策品質與調適效率，並將大幅強化指揮管制決策系統能力。本文將從指揮管制敏捷力觀點，探討智慧化決策指管系統發展趨勢。In facing the complex battlespace, agility is a critical capability of the command and control (C2). C2 agility, having been the hot topics of military research in NATO, focuses mainly on the adaptability of organizational structure to changing battlespace. Since AlphaGo defeated the world chess champion, and with the increasingly developing AI and IT, the future war will be most likely featuring great use of smart technology. Therefore, the higher the C2 agility, the much more enhanced the quality of adaptation efficiency and the capability of the C2 decision support system. This paper is to investigate the development trend of intelligent decision support for C2 systems from the C2 agility perspective.

    關鍵詞 ： 指管決策支援系統(command and control decision support system, C2DSS)、指管敏捷力(C2 agility)、人工智慧(artificial intelligence, AI)

    • 中心論題
    • 監督式深度學習於風災識別之應用Applying Supervised Deep Learning Techniques to Typhoon Damage Recognition
    • 劉于碩(Yu-Shuo Liu)‧古蕙媜(Hui-Chen Ku)‧林彥佑(Yen-Yu Lin)‧沈辰峯(Cheng-Feng Shen)‧陳家正(Jia-Jheng Chen)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    近年來，由於深度學習技術的突破，克服了許多傳統影像識別方法上遭遇的挑戰，使得理解影像上的高階語意，例如颱風災損程度，變成可能，本文探討如何利用監督式深度學習技術，建構一個風災威脅評估系統，能自動識別颱風災情圖片並回報災害種類。The breakthrough of deep learning techniques has tackled challenges of image recognition from traditional methods and made it possible to understand high level semantics of images, for example, the damage status from Typhoon. In this paper, it is investigated how to create an evaluation system applying supervised deep learning techniques to automatically recognizing the damage images from Typhoon and reporting the results.

    關鍵詞 ： ︰深度學習(Deep learning)、高階語意(high level semantics)、影像識別(image recognition)

    • 中心論題
    • 認知無線電與人工智慧技術Cognitive Radio with Artificial Intelligent Technology
    • 陳文士(Wen-Shyh Chen)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    認知無線電(CR)是一種智慧型通訊系統，具有認知及智慧學習能力，能夠感受外界環境變化，進而適應改變無線電參數，CR被認為可有效解決頻譜不足的問題。認知無線電之基本核心為認知引擎(CE)內建學習與推理決策功能，並實現了各種人工智慧(AI)法則，逐漸在無線電認知功能及應用上扮演重要角色，如頻譜感測、頻譜共享、動態頻譜接取(DSA)等功能。本文討論各種AI技術及演算法則如ANN、MA、HMM等如何應用於解決認知無線電各種問題，並比較分析其優點及限制因素。Cognitive radio (CR) is an intelligent wireless communication system being able to sense and learn its environments that CR is emerged as a solution to the spectrum scarcity and under utilization. Learning & reasoning engines play essential roles of cognitive radio, such as spectrum sensing, spectrum sharing, dynamic spectrum access (DSA), and implement different artificial intelligent (AI) algorithms. In this paper, we discuss various AI algorithms such as ANN, MA, HMM to solve specific problems of cognitive radio, which will be compared and analyzed on factors that influence the choice of them, in terms of responsiveness, complexity, and limitation.

    關鍵詞 ： 認知無線電(cognitive radio, CR)、認知核心引擎(cognitive engine, CE)、軟體無線電(SDR)、頻譜感測、動態頻譜接取(DSA)、頻譜共享、人工智慧(AI)、類神經網路(ANN)、隱馬可夫模型(HMM)

    • 中心論題
    • 運用類神經網路於魚雷輻射噪音辨識研究Research of Torpedo Radiation Noise Identification using Neural Network
    • 吳柏賢(Bo-Hsien Wu)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    傳統聲納系統，目標辨識是由操作手憑藉噪音的節拍、起伏等等節奏，搭配方位、頻譜紀錄圖跡進行判讀。一般來說，在複雜的海洋環境中識別各種目標噪音是件非常困難的事，尤其當目標噪音疑似為魚雷時，操作手需在短時間內做出攸關生存的判斷，心理更是承受極大的壓力。本研究依據艦艏環形聲納架構，提出被動式魚雷輻射噪音辨識技術，研究包括：波束合成、目標偵測、方位追蹤、特徵析出、特徵辨識等五項技術，期望能以電腦輔助操作手進行半人工或是全自動化的魚雷預警任務。結果顯示，運用類神經網路建構之魚雷辨識模型經過945筆數據訓練後，於420筆數據驗測中，其辨識性能對於水面艦目標有高達98%以上的成功率，水下目標則有88%以上的成功率。In a traditional sonar system, the operator identifies targets by listening to the noise tempo and watching the bearing frequency records. Generally, it is often difficult to identify targets in a complex underwater environments. The operator needs to make judgments within a short time, especially as the noise is suspected to be a torpedo. In this research, based on hull-mounted sonar, a technology of passive identification to the torpedo radiation noise is proposed, including beamforming, target detection, bearing tracking, feature extraction, and feature identification. It is expected that with the aided computer, the operator can perform semi-manual or automated tasks dealing with torpedo warning. Finally, a torpedo identification model is built using neural network, for which 945-item training data and 420-item testing data were conducted respectively. The results that show a high identification probability more than 98% for the surface ships, and 88% for the underwater targets were obtained from the model developed.

    關鍵詞 ： 倒傳遞類神經網路(neural network by reverse delivery)、魚雷預警(torpedo warning)、水下目標噪音辨識(noise identification of underwater targets)

    • 一般論述
    • 後量子密碼發展現況The Development of Post Quantum Cryptography
    • 李明倫(Ming-Lun Lee)‧陳陌凱(Mo-Kai Chen)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    由於量子計算／量子電腦之快速發展，未來傳統公鑰密碼系統皆有可能在短時間遭到破解，引起各國政府高度重視，並積極討論可能的因應作為。目前分為二種技術，一為以量子力學物理特性為基礎的量子密鑰分發技術(Quantum Key Distribution, QKD)，此方式安全性高，但須搭配光纖網路甚至是衛星的輔助才能有遠距離的傳輸；另一方式為以數學難題為基礎的新式加密演算法，又稱為後量子密碼(post-quantum cryptography, PQC)，除了具量子電腦無法有效破解的高安全性外，也能直接取代傳統的加密演算法，不須改變現有的傳輸架構。因此，美國NIST(技術標準局)即在2016年公開徵選PQC。根據其最新公布之時程，可能於2021年進入第三輪決選或完成決選。追蹤掌握徵選現況，有助於提升本院保密裝備研製能量與系統升級。With the increasingly rapid development of quantum computing, the public key cryptosystems might be destructed in a quick way that has caused extensive attention and corresponding countermeasures worldwide. They can be categorized into two types of techniques: one is quantum key distribution (QKD) based on the physical property of quantum mechanics featuring high security, however, needing fiber network and satellites for distant transmission; the other is post-quantum cryptography (PQC) with a new encryption algorithm embedded complex mathematics providing security against quantum computer attacks, and can be incorporated into existing protocols. It will facilitate production capability of NCSIST's security equipment to keep track of the progress in the development of the NIST competition to determine the future PQC schemes.

    關鍵詞 ： 後量子密碼(post-quantum cryptography)、量子抵抗(quantum resistant)、量子安全(quantum safe)、量子電腦(quantum computer)

    • 一般論述
    • 脈衝電場下射頻天線感應電流評估及防護測試Current Threat Assessment and Protection Test for the High-Altitude Electromagnetic Pulse (HEMP) Coupling to Antenna Systems
    • 黃仲強(Chung-Chiang Huang)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    指管通資情監偵系統C4ISR，是一套戰場自動化支援，命令下達及作戰回報的連貫程序系統，使指揮得以運籌帷幄，達到的作戰效果。高空電磁脈衝(HEMP)通常做為同時、全區域癱瘓對方C4ISR無線通裝之首選利器，因此本文介紹使用頻率掃頻照射(illumination test)方式，計算天線曝露高空核爆電磁脈衝(HEMP)威脅量，進一步使用脈衝電流注入(PCI)耐受性測試。C4ISR systems are essential to automated battlefield management integrating command, control, communications, computers, intelligence, surveillance, and reconnaissance to achieve a preemptive and preventive military action. The High-Altitude Electromagnetic Pulse (HEMP) is an instantaneous electromagnetic energy field that is produced in the atmosphere by the power and radiation of a nuclear explosion, and that is damaging to electronic equipment over a very wide area. The HEMP will be a preferable and crucial element to neutralize the enemy's C4ISR wireless communication system and effect systemic paralysis. This study provides insights on the test method of frequency sweep illumination assessing the threat level of the HEMP coupling to antenna systems by means of fast Fourier transform and extrapolation for the Application of PCI acceptance test.

    關鍵詞 ： 掃頻照射(illumination test)、電流威脅量(Ithreat(ω))、高能暫態測試(high-level transient testing)、脈衝電流注入(pulse current injection, PCI)

    • 一般論述
    • 多合金灰口鑄鐵材料與其元件製造技術開發The Manufacture and Fabrication of Multi-alloy Gray Cast Iron
    • 林孟泓(Meng-Hung Lin)‧李乘清(Cheng-Ching Lee)‧李基銓(Chih-Chuan Lee)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    在中小型無人飛行載具的動力選擇上，因轉子引擎具有體積小與大馬力等優點，故相當適合做為動力選擇，然而因馬力大且氣封件形狀特殊等特性，對於氣封件材料在耐磨耗與高溫機性的要求相對更高，傳統使用不銹鋼或球墨鑄鐵已經無法滿足引擎性能需求。因此本院為滿足轉子引擎性能需求，以灰口鑄鐵為基底透過合金設計添加多種元素使材料成為多合金灰口鑄鐵，多合金灰口鑄鐵為高合金化鑄鐵材料並具有高硬化能，加上所添加的元素可在顯微組織中產生片狀石墨組織、高強度麻田散鐵組織、史帝田鐵耐磨相以及碳化鈦／釩的高溫機性穩定析出相，使本材料相較其他氣封材料具有更優異的磨潤性能與高溫強度，極適合使用在轉子引擎中。此外為將本材料應用在轉子引擎中，藉由精密線切割技術製作出精密氣封零件並安裝於轉子引擎上執行性能測試，其測試結果引擎性能明顯優於由球墨鑄鐵所製作之氣封零件，滿足研發需求。Being small volume and highpower, the rotary engine is commonly used to power small to medium UAV. However, to meet high power needs, it is most requested that the rotary engine seals materials have high wear resistance and good performance in high temperature than conventional ones such as stainless steel and nodular graphite cast iron which are no longer effective as needed. Thus, to meet the performance requirements of rotary engine, we add some other elements into gray cast iron to become multi-alloy gray cast iron (MAG cast iron), and yield the microstructure that is characterized by high hardenability, flake graphite, martensite, steadite, and carbide. That consequently makes the MAG cast iron have higher machining strength, better wear resistance and lubricity. We've made rotary engine seals with MAG cast iron by using wire electrical discharge machining (WEDM). Finally, the rotary engines installed with MAG cast iron seals were tested and, from the test results, they revealed more excellent performance than those with nodular graphite cast iron, and met the requirements of the rotary engine development.

    關鍵詞 ： 轉子引擎(rotary engine)、密封圈(seals)、合金配方(alloy formula)、高淬透性(high hardenability)

    • 一般論述
    • 新一代頭盔顯示器之設計研究Research on the Design of Next Generation Helmet Mounted Display System
    • 詹明蕙(Ming-Hui Chan)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    本篇文章將介紹頭盔顯示器(helmet mounted display system, HMDs)的起源與發展，由情境認知(situation awareness)切入，探討頭盔顯示器在空軍戰鬥行動中的重要性；並以人因工程角度，依力學及光學設計分析，探討頭盔顯示器如何兼具技術及人因工程設計，提供飛行員最佳的戰鬥裝備。This paper introduces the background and development of Helmet Mounted Display System (HMDS). With a cognitive approach to situation awareness, the importance of HMDS is described for the air force operations. Furthermore, from the perspectives of human factors engineering, mechanics, and optics, the paper investigates how to integrate them to design next generation HMDS providing pilots with more suitable flight equipment.

    關鍵詞 ： 頭盔顯示器(helmet mounted display system, HMDS)、情境認知(situation awareness)

    • 一般論述
    • 超穎結構於雷達系統之應用Metamaterials Application in Radar System
    • 歐陽良昱(Liang-Yu Ou Yang)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    超穎結構(Metamaterial)是一呈現出不尋常電磁特性的子波長週期性結構。超穎結構的終極精神是透過設計電磁週期（子波長等級）結構，合成出任意的等效材料特性，進而達到任意的電磁目的。近來，超穎結構的領域不斷延伸擴張，遍及電磁（光）、熱、聲學。本報告將簡介超穎結構在電磁領域，尤其是雷達系統方面的應用。A Metamaterial is a subwavelength periodic structure with unusual electromagnetic (EM) properties. The ultimate spirit of Metamaterials is to synthesize arbitrary characters of effective materials by designing certain EM periodic structure, and then to achieve any goals. Recently, the field of Metamaterials is extended to electromagnetics, optics, heat, and acoustics. The application of Metamaterials in EM region, especially for radar systems, will be introduced in this report.

    關鍵詞 ： 超穎結構(Metamaterial)、雷達(radar)、微波(microwave)、週期性結構(periodic structure)

    • 一般論述
    • 應用AHP法及DEA法於本院專案績效評估模式研究Applying AHP and DEA to Project Performance Evaluation
    • 李衍榮(Yen-Jung Lee)‧盧健勝(Lu-ChienSheng)‧張家翊(Chia-I Chang)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    績效評估與組織競爭力有密切的關係，其評估結果對企業而言能提供有效的激勵，同時協助員工成長並達到組織之目標，提昇組織效能並促使組織永續經營發展，透過建全的績效評估機制，企業可瞭解自身之競爭優勢及需加以改進之處。本文提供兩種業界常用之績效評估方法：資料包絡分析法(data envelopment analysis, DEA)及分析層級法(analytic hierarchy process, AHP)，並考量本院各所中心之特性將績效評估因子分為定量及定性因子，分別評估各單位之暫時績效，再綜合此二種績效作為最後績效評估之結果，此兩種績效評估方式皆以學術理論為基礎，使決策者在評選單位績效時有一參考依據，進而客觀判斷出績效良好之單位。Performance evaluation has a strong positive relationship with organizational competitiveness. Evaluation results provide effective incentives for companies, help employees grow, facilitate achievement of organizational goals and effectiveness, and gain sustainable development. By conducting the comprehensive performance evaluation mechanism, a company can realize its competitive advantage and what needs improving. This paper proposes two commonly used performance evaluation methods, data envelopment analysis (DEA) and analytic hierarchy process (AHP). On considering characteristics of NCSIST, the performance evaluation factors can be classified into quantitative and qualitative for intermediate performance evaluation, which are then combined to final evaluation results. Both of performance evaluation methods are based on academic theories allowing decision makers to objectively evaluate the performance of each division.

    關鍵詞 ： 資料包絡分析法(data envelopment analysis)、分析層級法(analytic hierarchy process)、績效評估(performance evaluation

    • 一般論述
    • 全視域影像產生系統建置Establishment of Panorama Generation System
    • 張順惠(Shun-Hui Chang)‧劉建良(Jian-Liang Liu)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    龍園科技展示館為展示中科院研發成果與能量，航空所模擬系統組負責「模擬體驗區」系統建置，本研究係在模擬體驗區展示自製的影像產生器(image generator, IG)之應用成果，透過360度環場銀幕及聲光效果，播放模擬能量、視效系統場景及全臺視效資料庫能量。本文將介紹360度環場顯示的硬體系統設計、IG展示內容、功能開發及成果展示，以實現IG高效能全視域的場景模擬。The Science and Technology Exhibition Center in Lung Yuan Research Park presents the accomplishments NCSIST has made over the past decades. This study elaborates on the self-developed image generator and the applications in the Simulation Experience Area that ASRD (Aeronautical Systems Research Division) established, with the 360-degree panoramic display system, demonstrating simulation capability, audiovisual scene, and the terrain database rendering capabilities covering the whole Taiwan regions. Additionally, this paper introduces the hardware design, functionality development, content editing and results presentation of 360-degree panoramic display system for achieving high performance IG in full field of view.

    關鍵詞 ： 影像產生器(image generator)、360度環場顯示系統(360-degree panoramic display system)

    • 一般論述
    • 解繳物料標籤之精進作為Improvement of Goods Labeling in Product Delivery
    • 張漢湘(Han-Hsiang Chang)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    解繳軍種的物料，除了需包裝防護之外，另一個重點就是須標示清楚，因此產生標籤的製作需求。而為方便客戶直接辨識及掃描建檔，除文字的標示之外，還需加上條碼。本研究先簡單介紹幾種常用的標籤紙及碳帶的種類與用途，接著再介紹常見的一維與二維條碼。最後說明自製二維條碼標籤的開發過程中，所遭遇的困難與解決方法。In addition to package protection, clear labeling is critical to ensure good delivery to customers. Consequently, customer and regulatory label demands are on the rise. To meet customers specific label requirements, the barcode labels are required other than text for identification of goods and data establishment. This study briefly introudces commonly used labels, carbon ribbons, one- and two- dimensional barcodes, and how to tackle problems during self-development of 2D barcode.

    關鍵詞 ： 標籤(label)、條碼(barcode)

    • 一般論述
    • 在WMC未來演進中的AIThe role of AI in the Future Revolution of WMC
    • 呂錫民(Shyi-Min Lu)

    卷期 ： 47 / 4

    出版年 ： 2019/10

    隨著3GPP1第五代(fifth generation, 5G)新空中介面(New Radio, NR2)規範[1-2]第一版(Rel-15)的完成，研究下一步重點應該轉向無線行動通訊(Wireless Mobile Communication, WMC)發展。與前幾代產品類似，NR預計在未來十年內逐步展開，從引入創新技術組件到進一步增強功能，以及擴大5G無線接取範圍等等。從更長遠角度來看，我們預期將會有全新的“超越5G”無線接取技術(Radio Access Technology, RAT)出現。重新塑造或重新定義未來無線通訊潛力的新興技術領域，將非人工智慧與機器學習(machine learning, ML)莫屬。在本文中，我們簡要概述了WMC未來發展的一些重要研究面向，並探討WMC系統中不同部門使用AI/ML的潛在影響。Upon completion of the first edition of the 3GPP 5G/NR specification (Rel-15), the following steps of research will emphasize future development of wireless mobile communications (WMC). Like previous generation products, new radio (NR) is expected to be progressively launching, introducing its components of technological innovation and more powerful functions, and extending the range of 5G wireless access. From a long-term perspective, the innovative "Beyond 5G" radio access technology (RAT) is emerging and will be radically redefined and reformed by using artificial intelligence (AI) and machine learning (ML). This paper briefly highlights the critical dimension of future WMC development and investigate the potential impacts of AI and ML on different sectors.

    關鍵詞 ： 人工智慧(artificial intelligence, AI)、無線行動通訊(wireless mobile communications, WMC)、5G(fifth generation)、機器學習(machine learning, ML)