梁桂嘉Liang, Kuei-Chia張天樸Chang, Ashley Tempo2022-06-082023-08-162022-06-082021https://etds.lib.ntnu.edu.tw/thesis/detail/0287ae1b27b0582ac97166961c5db2a0/http://rportal.lib.ntnu.edu.tw/handle/20.500.12235/118431科技演進讓人類與機器間除了能透過觸控，還可以肢體、語音等方式與機器互動。語音使用者介面（voice user interface）是讓人類能夠透過對話和機器互動的橋樑。隨著語音技術進步，未來語音的操作將不再侷限執行單一任務，運用層面應能更多元，因此如何設計良好體驗的語音使用者介面及對話內容將是目前的挑戰。然而目前缺乏從使用者經驗設計師角度的語音使用者介面產品設計流程與相關研究。因此本研究將以使用者經驗為中心探討設計語音使用者介面的過程，最後整合案例分析與研究為設計流程模板，並實際運用該流程發展產品設計進行檢視。而本研究因語音能藉聲音傳達同理及其虛擬角色可以提供安慰與陪伴感、隨叫隨到的特性，選擇以應用於新冠肺炎防疫為例，透過語音互動陪伴居家隔離檢疫者度過隔離14天，讓防疫旅程變得更有溫度、正向。本研究拆分為研究與產品設計階段。研究階段旨在釐清研究背景與動機、文獻回顧與案例分析等。而設計階段將運用研究的語音使用者介面設計流程，完成產品設計並檢核流程。使用者經驗研究階段透過訪談了解居家隔離檢疫者的需求，抓住語音痛點並發想概念。使用者經驗設計階段時定義目標使用者，並以使用者旅程圖洞見與延伸為關鍵用例。透過用例情境及推測角色矩陣圖構想語音系統角色，並以人物誌呈現與以其角度撰寫示例對話，最後描繪為對話流程圖。使用者經驗測試階段進行兩次產品設計的評估，並針對回饋進行迭代，以此貼合需求。第一次測試藉綠野仙蹤法對語音對話流暢、用詞、認知負荷，以及語音對話代理人進行評估。其中語音對話代理人的設計表現最佳，然對話脈絡有待提升，並以此為依據調整，完成第二次測試用原型。第二次測試以評估語音使用者介面的易用性、滿意度為旨，最終在易用性得到 85.83 分，而易用性表現較佳的題向，多與學習使用介面有關，顯示目前的設計無學習瓶頸。在滿意度調查時了解到，與現有機制相比，虛擬語音助手「居米」也讓受試者更願意且放心的告知真實狀況。本研究提供語音使用者介面的設計流程建議，對於未來使用者經驗設計師能夠透過本研究流程進行開發帶來幫助與參考價值。With the advance of science and technology, humans and machines can not only interact by touch but also by gesture and voice. Voice user interface is a connector that allows humans to interact with machines by talking. Along with the progressing of voice technology, voice interaction will no longer be limited to performing a single task in the future, and it should grow at a broader application level. Therefore, how to design a voice user interface with a good user experience will be the challenge. However, it was a lack of the design process of voice user interface products that form in the perspective of user experience designers, and the related study is also lacking. Because of this, the study will base on the user experience to explore the design process of the voice user interface, and integrate case analysis into a set of design process templates, furthermore using the process to develop product design for inspection. While voice can convey empathy, its virtual character can provide comfort and companionship, and always being on call. In that case, this study chose to take the COVID-19 pandemic prevention as an example, by accompanying the home isolation quarantine person with the voice assistant app through the quarantine for 14 days, and make the quarantine journey more heartwarming and positive.This study is divided into the research and product design phases. The research phase aims to clarify the research background and motivation, problems, goals, literature review, and case analysis. The design phase will use the results of the research phase - the process of designing a voice user interface to complete the product design and verify the process at the same time. In the user experience research phase, this research uses interviews to understand the needs of the quarantine at home and grasp the voice pain points to develop concepts. In order to define the target user, this research uses the characters and the user journey map to gain insights and extends them as key use cases during the user experience design stage. The Voice Conversational Agent is conceived through use case scenarios and the Voice Conversational Agent Speculative Role Matrix. The sample dialogue will be written from the perspective of the voice conversational agent, and finally depicted as the voice user interface dialogue diagram. In the user experience testing phase, two product design evaluations are performed, and the feedback from the testing is used to iterate to meet users' requirements. The first test implements the Wizard of Oz test to evaluate the fluency, use of vocabulary, the cognitive load of voice sample dialogs, and the design of the voice conversational agent. Among them, "the design of voice conversational agent" performed best, but the dialogue context needs to be improved. Subsequently, based on the feedback from the first test, the prototype was adjusted to complete the prototype for the second test. The second test was designed to evaluate the effectiveness of the voice user interface and the satisfaction of the target users and finally scored 85.83 points on the system usability scale. The topics with better usability were mostly related to learning the voice interface, showing that the current design had no learning bottleneck. In the satisfaction survey, the study learned that compared with the existing service and product, the virtual voice assistant"Gemi" also allows the subjects to be more willing and more cautious to tell the real situation. The research provides suggestions on the design process of the voice user interface so that user experience designers can develop through the process of this research.語音使用者介面使用者經驗設計語音助手語音助手應用程式新型冠狀病毒肺炎語音設計語音產品設計voice user interfaceuser experience designvoice assistantvoice assistant applicationvoice app designcovid-19 prevention學術論文