Loading...
|
Please use this identifier to cite or link to this item:
https://nccur.lib.nccu.edu.tw/handle/140.119/134037
|
Title: | 人工智慧在影像監控的應用探討 – 商務車載產業個案研究 The Application of Artificial Intelligence in Video Surveillance – Case Study of Commercial Vehicle Industry |
Authors: | 林家聰 Lin, Chia-Tsung |
Contributors: | 鄭至甫 Jeng, Jr-Fu 林家聰 Lin, Chia-Tsung |
Keywords: | 人工智慧 商用車載 商用車載監控 影像監控 自駕車 行車安全 AI application in commercial vehicle Artificial Intelligence Video surveillance Commercial Vehicle |
Date: | 2021 |
Issue Date: | 2021-03-02 14:22:34 (UTC+8) |
Abstract: | 所謂的人工智慧(Artificial Intelligence, AI)就是「擁有人類智慧的程式」(三津村直貴,2018),要能執行人類所能執行的工作,能擁有人類一樣的意識。人工智慧在近幾年已躍升為產業應用的顯學,舉凡金融保險、製造、交通、城市安全、旅館餐飲、運輸倉儲、零售、醫療、娛樂、不動產、農業、礦業、教育、能源、行政等…產業都有AI應用的實例。然而真正落實到產業產品及服務還是需要切合市場的需求或是創造新的商業價值。目前AI的主要應用還是模擬、執行特定的功能,偏重在取代可重複的現場工作、資料的蒐集整理分析、一般的溝通(涉及的情感程度低)。透過深度學習以及越來越豐富的大數據資料蒐集分析,技術正朝向下一階段能「理解事物現象」、「捕捉語言涵義」、「解決事情的能力」的方向進展。
過去影像監控受到終端設備運算能力的限制,大部分資料的分析仰賴雲端的運算,而大量的影像資料傳輸受網路頻寬限制,以致即時的影像分析能力也受到限制更遑論即時反饋的行動建議。隨著硬體運算能力提升、網路傳輸頻寬進步以及整合人工智慧演算法,監控產業過去被詬病的誤警報、誤動作已獲得大幅改善。過去商務車載在影像監控及GPS路徑監控是不同的設備系統,這幾年因應市場需求逐漸朝整合成一個監控系統發展。商務車載著重在提升營運的成本效率、降低行車風險、提升車輛安全性及可靠度,針對行車安全性及降低風險更需要即時的回饋。藉由人工智慧結合物聯網(IoT, Internet of Thing. ADAS、監控相機等也被歸類為IoT)、5G傳輸,以及晶片運算能力的大幅提升,有機會達成商務車載對即時監控及即時行動回饋的要求,甚至開發出新的應用產生新的商務模式。
本研究以創新擴散理論及客戶價值模型、技術市場矩陣等相關文獻來探討,以形成本研究的研究理論及研究工具的基礎。研究方法為針對商務車載價值鏈上下游廠商進行訪談及問卷調查,以了解供應端及需求端對市場需求短、中、長期的認知。進而解析人工智慧在商務車載應用市場面臨的挑戰及未來發展的契機。
研究發現人工智慧導入商務車載市場尚屬創新採用階段。供應端及需求端對人工智慧導入市場長期的期待是一致的,但是雙方的短期需求認知差異頗大(不論是期待的成本或者功能),各垂直市場破碎化的產業需求,人工智慧研發能量不足,各國法規以及市場對安全考量是導致市場滲透率低的主要因素。本研究以技術市場矩陣及客戶價值模型來解析創新擴散理論實務上在市場採用新產品/技術所面臨的挑戰及困難,以及探討從創新擴散初期的創新採用階段進入早期採用階段需要克服的市場差距。不論是商用自駕車或者搭載人工智慧裝置的商務車,能滿足市場短期對提升安全及增進營運效率的需求,人工智慧才能跨入創新擴散理論的早期採用階段,進而發展中長期創造新市場、新商務模式。 The Artificial Intelligence (AI) is a "program with human intelligence" (Naoki Mitsumura, 2018). It must be able to perform tasks as humans do and have the same consciousness as humans. In recent years, Artificial Intelligence has become a prominent industry application in the filed such as finance, insurance, manufacturing, transportation, urban security, hotels and restaurants, transportation and storage, retail, medical, entertainment, real estate, agriculture, mining, education, energy , administration, etc. However, the actual implementation of products and services still needs to meet market demand or create new commercial value. Currently, the main application of AI is to simulate and perform specific functions, focusing on replacing repeatable field tasks, data collection and analysis, and general communication (with low emotional level involved). Through deep learning and increasingly big data collection and analysis, technology is moving towards the next stage of "understanding things and phenomena", "capturing language meaning", and "ability to solve things".
In the past, video surveillance was limited by the computing power of terminal equipment. Most data analysis is relied on cloud computing. The transmission of large amount image data was restricted by the network bandwidth, so that real-time image analysis capabilities were also limited. The recommended real-time feedback action is even unfeasible . With the improvement of hardware computing power, network transmission bandwidth and the large progress on artificial intelligence algorithms, the false alarms and malfunctions criticized by the surveillance industry in the past have been greatly improved. In the past, commercial vehicles used different operation systems for video surveillance and GPS route tracking. Recently, in response to market demand, they have gradually integrated into a single operation system. Commercial vehicles focus on improving the cost efficiency of operations, reducing driving risks, and improving vehicle safety and reliability. It is required for immediate feedback for driving safety to reduce risks. With integration of Artificial Intelligence, Internet of Things(IoT) (ADAS, surveillance cameras, etc. are also classified as IoT), 5G transmission, and substantially increasing in hardware computing power, it is feasible to achieve real-time monitoring and action feedback to meet commercial vehicles requirements, and even develop new applications to generate new business models.
The research uses the innovation diffusion theory, customer value model, technology versus market matrix and other related literature to explore, to form the basis of the research theory and research tools of the research. The research method is to conduct interviews and questionnaire surveys with upstream and downstream in the commercial vehicle value chain to understand the short, medium and long-term perceptions of market demand and expectation on the supply side and demand side. Then analyze the challenges faced by artificial intelligence in the commercial vehicle industry and the opportunities for future development.
The study found that the introduction of artificial intelligence into the commercial vehicle industry is at the stage of innovative adoption. The supply side and the demand side have the same long-term expectations for the introducing artificial intelligence into the market. However, the perceptions of short-term demand on both sides are quite different. Especially, on the expected cost and function, variant demands of fragmented vertical markets, and the lack of artificial intelligence research and development capability, national regulations and market safety considerations are the main factors caused to low market penetration of AI. This research uses the technology market matrix and customer value model to analyze the challenges and difficulties faced by the adoption of new products/technologies in the market in terms of innovation diffusion theory and practice, and explores the market gap that needs to be overcome from the initial stage of innovative adoption to the early adoption stage. No matter commercial self-driving car or a commercial vehicle equipped with artificial intelligence devices, it must meet the short-term needs of the market to improve driving safety and operational efficiency. Then it is possible for artificial intelligence to enter the early adoption stage, develop new markets and create new business model in the long-term. |
Reference: | Rogers E. M. (2006)。創新的擴散(唐錦超譯)。台北市: 遠流。(原 著出版年1995) 三津村直貴(2018)。AI人工智慧大未來(陳子安譯)。台北市: 旗標。 (原著出版年2017) 古明地政俊、長谷佳明作(2018)。AI人工智慧的現在.未來進行式(林 仁惠譯)。 台北 市: 遠流。(原著出版年2017) 李開復(2019)。AI新世界。台北市: 遠見天下文化。 李開復、王詠剛(2017) 。人工智慧來了。台北市: 遠見天下文化。 呂惠聰、強南囡、王微微(2018)。客戶關係管理。台北市: 財經錢 線。 陳昇瑋、溫怡玲(2019)。人工智慧在台灣。台北市: 天下雜誌。 裴有恆、陳玟錡(2020)。AIoT人工智慧在物聯網的應用與商機。台北 市: 碁峰。 裴有恆(2019)。白話AIoT數位轉型。台北市: 博碩。 台灣車載資通訊產業協會(TTIA) (2015)。智慧巴士 4G 專用車載機產 業標準。台北市: 台灣車載資通訊產業協會。 資策會產業情報研究所楊蕣仿(2018)。中華電信AI技術發展 與應用。台北市:中華電信研究院。 交通部公路總局(2019)。交通科技產業政策白皮書草案,智慧電動巴 士。台北市: 交通部公路總局。 何心宇、周士雄、林信亨、翁嘉德、張旭佑、張家維、鄭兆倫、富士 Chimera總研(2017) 。車聯網大廠布局與新創案例剖析。台北市: 資策會產業情報研究所。 李綱(2019)。自駕車商業化的挑戰。新北市: 國家運輸安全調查委員 會。 呂坤龍、張國樑、黃立仁、紀坤明、張雍昌、楊智仁。車用電子的功 能安全需求–ISO26262國際安全規範簡介及其應用。電腦與通訊, 163期,Page 10-16。 林信亨、徐育群、陳賜賢、童啟晟、熊嘉彬、Nikkei Electronics、 富士Chimera總研(2017) 。車聯網發展趨勢暨技術分析。台北市: 資策會產業情報研究所。 林明彥(2019)。從新創看自駕車發展趨勢與機會。新北市: iAUTO 徐志偉(2019)。台中花博自駕車情境體驗與應用服務。新竹縣: 工研 院資通所 徐志偉(2019)。自駕車結合車聯網之創新應用。新竹縣:工研院資通所 許銘德(2019)。以點對點連線實現具有即時影像監控的車隊管理系 統。 輔仁大學電機工程學研究所,新北市。 施聰評、林信賢(2015)。先進駕駛輔助系統 (ADAS) 法規趨勢。 車輛研測資訊 107期,Page 8-13。 億歐智庫(2018)。2018年中國AI+安防行業發展研究報告。中國北京: 億歐智庫。 顧馨文(2018)。2018 ICT產業白皮書(下)人工智慧/5G/資訊安全/電子 商務。 余至浩 (2018/9/18)。全球自駕車發展歷程。 檢自https://www.ithome.com.tw/news/125897都保杰 (2018/8 /26)。 英視特安防 (2019/02/15)。2019年AI安防監控的視覺應用。 檢自https://kknews.cc/tech/gazxavm.html 郜敏 (2019/12/03) 。專家預測2023年自駕車增加逾74萬台,但仍受3 大因素影響。檢自https://newtalk.tw/news/view/2019-12-03 /335181 。 邱莉燕 (2020/7/23)。台北首輛無人駕駛夜間巴士來了!最快9月開放 搭乘。檢自https://www.seinsights.asia/article/3290/3270 /7225 陳佑怡 (2019/1/15)。迎接人機介面新商機,掌握全球AI語音助理下 一步。檢自 https://www.2cm.com.tw/2cm/zh-tw/tech /B73A5BE5A4954D5988102C7D8A5B8A68?type=26 商湯科技:AI 公司從技術到商業化的 5 個關鍵。 檢自https://www.stockfeel.com.tw/author/lieyunwang/ 。 愛范兒 (2019/08/20) 。世上最聰明的自駕車,低估了人類這個「障 礙物」。檢自https://technews.tw/2019/08/20/waymo-human- problem/ 楊安琪 (2020/09/28)。北市自駕巴士開放民眾試乘,信義路測試路段 由遠傳 5G 全線覆蓋。檢自 https://technews.tw/2020/09/28/taipei-city-to-start- autonomous-bus-trial-service/ 廖專崇 (2019/04/02)。跨越Level 2邁向自駕,ADAS汽車感測系統再 進化。檢自 https://www.2cm.com.tw/2cm/zh-tw/market/ 2FECE0BCF90A43B98557BF56CF979F77?type=14 盧佳柔 (2020/03/30) 。封閉場域/固定路線更易落地,商用無人車發 展超前部署。檢自 https://www.2cm.com.tw/2cm/zh-tw/market /29372115E3FF4BEA9BC6F580E2BC6E06?type=14 簡鈺璇 (2020/07/24)。強化決策力,微星打造無人車隊的勝利方程 式。檢自https://www.ankecare.com/2020/20980 蕭佑和 (2018/4/11) 。大和有話說:完整解析AI人工智慧:3大浪 潮+3大技術+3大應用。檢自 https://dahetalk.com/2018/04/08/%e5%ae%8c%e6%95%b4%e8%a7%a3 %e6%9e%90ai%e4%ba%ba%e5%b7%a5%e6%99%ba%e6%85%a7%ef%bc%9a3%e5 %a4%a7%e6%b5%aa%e6%bd%ae%ef%bc%8b3%e5%a4%a7%e6%8a%80%e8%a1 %93%ef%bc%8b3%e5%a4%a7%e6%87%89%e7%94%a8%ef%bd%9c/。 Koopman P. & Wagner M. (2016). Challenges in Autonomous Vehicle Testing and Validation. Carnegie Mellon University Edge Case Research LLC, 16AE-0265 Paddock N. M. (2016). Driving to safety: How many miles of driving would it take to demonstrate autonomous vehicle reliability?. Transportation Research Part A: Policy and Practice, Volume 94, Pages 182-193 Roberts J. M., Duff E. S., Corke P. I. (2002). Reactive navigation and opportunistic localization for autonomous underground mining vehicles. Information Sciences,Volume 145, Page 127–14 Waytz A., Heafner J., Epley N. (2014). The mind in the machine: Anthropomorphism increases trust in an autonomous vehicle. Experimental Social Psychology Volume 52,Pages 113-117 Wang B, Wang H., Liu, B., Meng X., Yang G. H. (2017). Pedestrian recognition and tracking using 3D LiDAR for autonomous vehicle. Robotics and Autonomous Systems, Volume 88, Pages 71-78 a&s (2019/7/29). 8 AI-based video analytics platforms advance security implementation.Asmag. Retrieved from https://www.asmag.com/showpost/28633.aspx?name=news a&s (2019/7/26). AI development in the US might take time. Asmag.Retrieved from https://www.asmag.com/showpost /28628.aspx?name=news Dahua Technology (2019/10/2). Dahua Smart Transportation Solutions to address security and efficiency. Asmag. Retrieved from https://www.asmag.com/showpost /27169.aspx?name=news Jonas A. A. (2018/3/20). Uber Accident Impact On Self-Driving Tech. Youtube. Retrieved from https://www.youtube.com /watch?v=AH5YeEYhijs Liu S. (2019/11/25). Cumulative revenue of top 10 use cases/segments of artificial intelligence (AI) market worldwide, between 2016 and 2025. Statista. Retrieved from https://www.statista.com/statistics/607835/worldwide- artificial-intelligence-market-leading-use-cases/ MarketsandMarkets (2019/6). fleet management market report. MarketsandMarkets.Retrieved from https://www.marketsandmarkets.com/Market-Reports/fleet- management-systems-market-1020.html McHale A. (2018/11/12). AI video analytics has crossed the Rubicon. Asmag. Retrieved from https://www.asmag.com/showpost/27374.aspx Montaqim A. (2019/5/24). Top 25 AI chip companies: A macro step change inferred from the micro scale. Robotics & Automation News. Retrieved from https://roboticsandautomationnews.com/2019/05/24/top-25-ai- chip-companies-a-macro-step-change-on-the-micro- scale/22704/ Nokia (2016/11/29). IoT Video Analytics for Smart Cities. Youtube. Retrieved from https://www.youtube.com/watch?time_continue=149& v=__MdWBpFDGg&feature=emb_logo OSP Labs (2019/3/20). Artificial Intelligence (AI) Based Video Analytics & Motion Analytics. Youtube. Retrieved from https://www.youtube.com/watch?v=73wc0eTrQqA Pao W. (2019/11/20). What benefits does video surveillance on buses provide?. Asmag.Retrieved from https://www.asmag.com /showpost/30837.aspx?name=news Pao W. (2019/11/21). How public transit can benefit from data analytics. Asmag. Retrieved from https://www.asmag.com /showpost/30842.aspx?name=news Pettersson E. (2020/4/30). Tesla’s Autopilot Blamed in Lawsuit for Fatal Accident in Japan. Blooberg. Retrieved from https://www.bloomberg.com /news/articles/2020-04-29/tesla-s-autopilot-blamed-in- lawsuit-for-fatal-accident-in-japan Rice D. (2018/11/19). A New Wave of Video Analytics. SDM. Retrieved from https://www.sdmmag.com/articles/95908-a-new- wave-of-video-analytics Rice D. (2020/4/20). Is AI Video Ready for Its Close-Up?. SDM. Retrieved from https://www.sdmmag.com/articles/97855- is-ai-video-ready-for-its-close-up Scheinman J. & Cummings M. (2018/3/20). Are We Ready For Self-Driving Cars? Experts Debate. Youtube. Retrieved from https://www.youtube.com/watch?v=TQRXzVGmjlI Shein M. M. (2019/2/18). Artificial Intelligence in Video Surveillance. SDM. Retrieved from https://www.sdmmag.com /articles/96235-artificial-intelligence-in-video- surveillance Strom E. (2019/8/29). How machine, deep learning are making traffic management smarter. Asmag. Retrieved from https://www.asmag.com/showpost/30450.aspx?name=news Strom E. (2019/10/25). Regulations posing obstacles to railways adoption of AI and edge computing. Asmag. Retrieved from https://www.asmag.com/showpost /30704.aspx?name=news Thomas P. A. (2018/12/19). Challenges and solutions in public transport fleet management. Asmag, Retrieved from https://www.asmag.com/showpost/27567.aspx?name=news Thomas P. A. (2019/7/9). 8 trends in fleet management solutions you just cannot miss. Asmag. Retrieved from https://www.asmag.com/showpost/28554.aspx?name=news Thomas P. A. (2019/12/25). How AI-based fleet management arms police to the teeth. Asmag. Retrieved from https://www.asmag.com/showpost/30990.aspx?name=news William Pao (2019/5/3). How ‘Logistics 4.0’ contributes to smart logistics. Asmag, Retrieved from https://www.asmag.com/showpost/28276.aspx?name=news |
Description: | 碩士 國立政治大學 經營管理碩士學程(EMBA) 106932082 |
Source URI: | http://thesis.lib.nccu.edu.tw/record/#G0106932082 |
Data Type: | thesis |
DOI: | 10.6814/NCCU202100198 |
Appears in Collections: | [經營管理碩士學程EMBA] 學位論文
|
Files in This Item:
File |
Description |
Size | Format | |
208201.pdf | | 4454Kb | Adobe PDF2 | 0 | View/Open |
|
All items in 政大典藏 are protected by copyright, with all rights reserved.
|