透過您的圖書館登入
IP:34.230.68.214
  • 學位論文

陪伴機器人以提出情境感知之寒暄話題來開啟與人互動

Context-aware Topic Proposal in Phatic Communion for Interaction Initiation of a Robot Companion

指導教授 : 傅立成

摘要


隨著家用機器人逐漸出現在市場上,人們對於機器人的期望不再僅限於單純的提供服務,而是希望機器人能夠更人性化的陪伴孩童遊戲,關心家中成員,甚至是看護長者。此類陪伴型機器人不僅要能以自然的方式與使用者互動,還需要更進一步的與使用者建立社交關係,才能夠藉由陪伴達到提供使用者社交支持的目的。為使機器人更快的與使用者建立社交關係,機器人必須具備主動與使用者互動的能力。本論文參考人際互動中寒暄式對話之特性,使機器人仿效人們在開啟寒暄互動時話題選擇的策略,根據當下情境、雙方共同的認知以及社交距離來選擇適當的話題來開始與使用者的互動。首先, 機器人必須能夠藉由有限的感測器資訊判斷當下的情境。其次,機器人必須以符合社交規範與常識的方式找出與情境相關的話題。最後,機器人必須根據與使用者之間的社交距離選擇適當的話題。因此,我們所提出之話題選擇系統包含三大部分。第一,我們建立了一個常識知識庫作為情境判斷與話題選擇的依據,此知識庫包含情境中「人、事、時、地、物」之間各種可能的關係。第二,使用機率軟性邏輯(Probabilistic Soft Logic)將知識庫中的邏輯規則轉換為機率模型以考慮感測器與環境的不確定性。第三,根據社會語言學之研究,考慮當下情境與使用者的親近程度,選擇適當的話題以避免過於唐突的話題冒犯到使用者。最後,我們邀請八位志願者參與我們設計的實驗,實驗結果顯示所提出之話題選擇系統確實能夠依據不同情境與社交關係選擇符合當下情境與社交關係的話題。

並列摘要


The researches on robot companions that provide personal service at home have become more and more popular. Robot companions are expected to provide not only physical assistance in daily tasks but also social supports for human. In order to provide social connections and enhance the positive user experience, the capability for such a robot companion to actively initiate an interaction in a natural way is important. When people try to engage with others, they often start from topics that related to the context involved: themselves, their partner or the environment. In this work, we aim to develop a context-aware robot that initiate a non-task-oriented interaction with the similar strategy. We focus on how a robot could interpret its environment as semantic concepts and generate proper topic proposal to invite the user to start an interaction. To achieve context awareness, we use the help of object recognition components that developed in computer vision field to detect semantic entities in the environment. Then, we combine the detection results with each other using the commonsense knowledge that consist of concepts best representing the environment to establish a context model based on hinge-loss Markov random field. Finally, the robot scores each concept in the context according to the social relationship with the user and generates a sentence to initiate the interaction with the user. In the experiments, we first examine the performance of the context model. Then, four topic selection strategies are compared to examine the effectiveness of the context-aware topics in the field experiments.

參考文獻


[1] Cynthia Breazeal and B. Scassellati. How to build robots that make friends and influence people. Intelligent Robots and Systems, pages 858–863, 1999.
[2] David Feil-Seifer and Maja J. Mataric. Socially assistive robotics: Ethical issues related to technology. IEEE Robotics and Automation Magazine, 18(1):24–31, 2011.
[3] Muhammad Ali, Samir Alili, Matthieu Warnier, and Rachid Alami. An Architecture Supporting Proactive Robot Companion Behavior. In New Frontiers in HumanRobot Interaction at AISB, 2009.
[4] Khaoula Youssef, P . Ravindra De Silva, and Michio Okada. Exploring the Four Social Bonds Evolvement for an Accompanying Minimally Designed Robot. 6th
International Conference, ICSR 2014, Proceedings, (C):412, 2014.

延伸閱讀