Loading...

Proposing an Empirical Motion-Time Ppattern for Hhuman Gaze Behavior in Bifferent Social Situations and Implementing the Pattern on RASA Social

Mashaghi, Mohammad Hossein | 2022

78 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 56158 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Taheri, Alireza; Behzadipour, Saeed
  7. Abstract:
  8. Social robots that are designed to interact with people in order to fulfill purposes like education, healthcare, etc. have to behave interactively like human. One of the human’s interactive behaviors is eye gaze. Studying the literature, we found out that in previous researches conducted to control the social robots’ gaze behavior, human gaze behavior was investigated in some limited situations such as two- or three-way conversation in order to extract the pattern of this behavior. Therefore, increasing the variety of studied social situations is a way to fill this gap. In this research we intend to propose an empirical motion-time pattern for human gaze behavior in some different social situations. These situations include scenes with 2 to 4 people performing “speaking”, “waving”, “pointing”, “entering” and “leaving” social behaviors in an organized way. The data we used for pattern extraction were all obtained from eye tracking experiments. In the first phase of this research, we tried to extract the pattern from pre-collected eye tracking data. Despite of being promising, the results were not practical and unable to implement. In the second phase, by performing a targeted eye tracking experiment and using the genetic optimization algorithm, we were able to extract the significance coefficient of each of the mentioned social behaviors, the main advantage of it compared to the first phase was the ability of implementation. Finally, by implementing these coefficients on the real robot and validating its performance by means of a survey, the effectiveness of this pattern extraction was confirmed. The positive effect of implementing the model in this survey consisting of 10 questions is generally evident, and in particular, it made a significant difference on 3 out of 10 questions
  9. Keywords:
  10. Social Robotics ; Genetic Algorithm ; Eye Tracking ; Motion-Time Pattern ; Social Eye Gaze ; RASA Social Robot

 Digital Object List

 Bookmark

No TOC