| 摘要: | 本論文提出一個全面的機器人輔助學習框架，結合先進技術、認知策略與心理方法，以增強學生的學習參與度與學習成效。本研究基於人機互動（Human-Computer Interaction, HCI）理論，探討互動式運算的認知、情感與行為層面，重點關注學生如何與智慧學習系統進行互動、反思與建立聯結。本研究回應當前教育系統面臨的幾個關鍵挑戰，包括機器人資源有限、僅限課堂內的互動、缺乏反思工具、無法建立情感聯結的非個性化機器人，以及難以維持長期參與等問題。本框架由四個相互關聯的研究構成，並在每一階段基於前一階段的成果進行發展。研究1將機器人與數位實境整合，創造沉浸式情境學習環境。研究2引入數位分身寵物機器人，讓學生能在課堂外持續互動。研究3實施自我對話機制，促進學生的反思與互動式自我評估。研究4探討情感聯結，設計個性化機器人，以增強情感互動並保持長期的學習參與。這項研究融合計算機科學、人機互動與認知心理學，設計並評估針對學生學習的認知、情感與行為層面進行互動系統的設計，並在機器人如何運用於教育領域方面帶來了轉型性突破。數位雙胞胎技術擴展了實體機器人的功能，使其可廣泛應用並提高可及性；自我對話機制與情感聯結則有助於促進學生的反思並維持學習動機。準實驗研究顯示，本框架在認知層面（提升學習成效）、情感層面（增強學習動機、自我效能、情感依附與學習興趣）以及行為層面（提高互動頻率與延長參與時間）都有顯著改善。該框架具備模組化和可擴展性，適用於AI輔助教學、聊天機器人及虛擬助理等多種應用，並能靈活適應不同學科與學習環境。此研究為未來HCI領域的發展奠定了基礎，並突破了計算系統在教學與學習中的應用邊界，帶來教育機器人應用領域的重大變革。;This dissertation presents a comprehensive framework for robot-assisted learning that integrates advanced technologies, cognitive strategies, and psychological methods to enhance student engagement and learning outcomes. Grounded in Human-Computer Interaction (HCI), the research explores cognitive, affective, and behavioral dimensions of interactive computing, focusing on how students engage, reflect, and form connections with intelligent learning systems. The study addresses key challenges in educational systems, including limited robot availability, classroom-restricted interactions, lack of reflective tools, non-personalized robots that hinder emotional connections, and difficulties in sustaining long-term engagement. The proposed framework was developed through four interconnected studies, each building upon the previous. Study 1 integrated robots with digital reality for immersive, situated learning. Study 2 introduced digital twin pet robots to enable continuous engagement beyond classroom walls. Study 3 implemented self-talk mechanisms for reflective, interactive self-assessment. Study 4 explored affectional bonds, designing personalized robots that foster emotional connections and sustain long-term engagement. Bridging computer science, HCI, and cognitive psychology, this research designs interactive systems that address cognitive, affective, and behavioral aspects of student learning, marking a transformational shift in how robots are used in education. The use of digital twins extends physical robot capabilities, enabling scalable, ubiquitous access, while self-talk mechanisms and affectional bonding promote reflection and sustained engagement. Quasi-experimental studies demonstrated improvements in cognitive outcomes (enhanced learning performance), affective outcomes (increased motivation, self-efficacy, emotional attachment, and learning interest), and behavioral outcomes (higher interaction frequency and longer engagement durations). The framework’s modular and scalable design supports diverse applications, extending to AI tutors, chatbots, and virtual assistants, making it adaptable to various subjects and learning environments. This research lays a foundation for future advancements in HCI, pushing the boundaries of how computational systems can transform teaching and learning paradigms. |
