This research demonstrates that robots can achieve socially acceptable interactions, using loosely synchronized head gaze-speech, without understanding the semantics of the dialog. Prior approaches used tightly synchronized head gaze-speech, which requires significant human effort and time to manually annotate synchronization events in advance, restricting interactive dialog, and requiring the operator to act as a puppeteer. This approach has two novel aspects. First, it uses affordances in the sentence structure, time delays, and typing to achieve autonomous synchronization of head gaze-speech. Second, it is implemented within a behavioral robotics framework derived from 32 previous implementations. The efficacy of the loosely synchronized approach was validated through a 93-participant 1 x 3 (loosely synchronized head gaze-speech, tightly synchronized head gaze-speech, no-head gazespeech) between-subjects experiment using the "Survivor Buddy" rescue robot in a victim management scenario. The results indicated that the social acceptance of loosely synchronized head gaze-speech is similar to tightly synchronized head gazespeech (manual annotation), and preferred to the no head gaze-speech case. These findings contribute to the study of social robotics in three ways. First, the research overall contributes to a fundamental understanding of the role of social head gaze in social acceptance, and the production of social head gaze. Second, it shows that autonomously generated head gaze-speech coordination is both possible and acceptable. Third, the behavioral robotics framework simplifies creation, analysis, and comparison of implementations.
