First steps toward natural human-like HRI

Abstract

Natural human-like human-robot interaction (NHL-HRI) requires the robot to be skilled both at recognizing and producing many subtle human behaviors, often taken for granted by humans. We suggest a rough division of these requirements for NHL-HRI into three classes of properties: (1) social behaviors, (2) goal-oriented cognition, and (3) robust intelligence, and present the novel DIARC architecture for complex affective robots for human-robot interaction, which aims to meet some of those requirements. We briefly describe the functional properties of DIARC and its implementation in our ADE system. Then we report results from human subject evaluations in the laboratory as well as our experiences with the robot running ADE at the 2005 AAAI Robot Competition in the Open Interaction Event and Robot Exhibition.

Covert Robot-Robot Communication: Human Perceptions and Implications for HRI

“…While much research in HRI toward this goal has focused on the effects of robot appearance and observable behavior, a significant aspect of natural HRI is communication in natural language (e.g., Scheutz, Schermerhorn, Kramer, & Anderson, 2007), which has only recently received significant attention. Recent research has investigated various social aspects of natural language interactions with robots, such as politeness (e.g., , turn taking (e.g., Nadel, Revel, Andry, & Gaussier, 2004), affective speech (e.g., Scheutz, Schermerhorn, & Kramer, 2006), dialogueappropriate facial movements (e.g., Liu, Ishi, Ishiguro, & Hagita, 2012), pragmatic analysis (e.g., Williams, Briggs, Oosterveld, & Scheutz, 2015), and collaborative control (e.g., Fong, Thorpe, & Baur, 2003).…”

Robust spoken instruction understanding for HRI

“…Our proposed architecture integrates speech recognition, incremental parsing, incremental semantic analysis, disfluency analysis, and situated reference resolution components into our robotic DIARC architecture [5], which provides mechanisms for integrating incremental natural language processing components with action execution [16], [20]. The present work expands the capabilities of the previous work with the inclusion of a more sophisticated parsing mechanism, allowing greater flexibility and more robust natural language understanding.…”

“…In contrast, converging evidence from psycholinguistics suggests that human language understanding is incremental and parallel, depends on the speaker's and listener's contexts, utilizes task and goal knowledge, and involves the perceptions and perspectives of situated, embodied agents. While this difference in processing style may not matter for many NLU applications, it is critical in situations where humans and robots interact naturally as embodied agents co-located in the same environment [5].…”

Resolution of Referential Ambiguity in Human-Robot Dialogue Using Dempster-Shafer Theoretic Pragmatics

“…In this section, we describe how each stage of our HRIoriented clarification request generation framework [26] is handled by components of the DIARC architecture [18].…”