A Human-Robot Sub-dialogues Structure Using XML Document Object Model

Abstract

Improving remote interactivity between patients and their health carers is a key requirement for emerging eHealth systems. This paper presents a new human-robot sub-dialogues structure based on the XML Document Object Model (DOM). The dialogue structure which aims at improving patient-carer interactivity enables both the patients and their carers to control behavior of humanoid robot locally through verbal instructions or tactile sensors as well as remotely through textual Web-based instructions. The textual instructions which are prepared in the form of interactive dialogues are converted to XML DOM and sent to the in-home robot via the Internet. The robot will then verbalize these dialogues and present them to the patient along with appropriate gesture/posture attributes. The XML DOM standard is adopted in this study to facilitate the structure and navigation flexibility between the internal nodes of the dialogue. Numerous human-robot interactions based on the proposed dialogue structure have been designed, built and successfully performed by the robot using both local and remote instructions. These dialogues are tailored to support diabetes self-management in children. The prototype evaluation results and observations showed (i) a seamless and accurate data transfer between the robot and a remote Web-based health portal, and (ii) a high level of navigation flexibility between different nodes of the experimented dialogues.