Dottorato in INGEGNERIA INFORMATICA

Titolo della tesi: Developing Scenarios and Research Methodologies for Evaluation Human-Robot Interaction in Social Contexts

As a newly emerging field, Human-Robot Interaction (HRI) offers the unique opportunity to study the multi-discipline regarding robots, humans and interactions in social contexts. Understanding how people of different ages, gender and abilities best perceive from robot’s behaviors remains an important challenge in the field. In this thesis, we aim to develop an HRI research approach that investigates human behaviors toward robots and robot’s behaviors perceived by humans in social contexts. In this direction, the first part of this thesis introduces an overview of HRI, from its origins, and discusses the current state-of-the-art related to our research targets. The second part of the thesis presents research studies on understanding human factors, developing evaluation techniques, designing task scenarios and conducting experimental studies. Specifically, we first introduce a confirmatory research study in a social context to investigate if user perception in HRI may be affected by human factors. Leveraging the standardized ques- tionnaire Robot Social Attribute Scale (RoSAS), we analyze the quantitative data collected by the questionnaire and report the relevant findings (see Chapter 3). Secondly, since questionnaires are among the most used evaluation technique in the field of social HRI, we develop an approach to design a new type of questionnaire as a task-driven evaluation technique for measuring user perception in social contexts. The approach consists of two steps. First, it relies on interviewing experts on HRI to understand which robot’s behaviors can potentially affect the user perceptions during an HRI task. Then, it leverages a user survey to filter out those robot’s behaviors that are not significantly relevant from the end user perspective. The results of the survey have allowed us to derive a final list of 17 behaviors to be captured in the questionnaire, which has been finally developed relying on a 5-point Likert-scale (see Chapter 4). Thirdly, we employ the SciRoc challenge (that is a repeatable and general-purpose test method developed for HRI performance evaluation in a realistic social context) to introduce an exploratory research study to analyze the outcomes of the most sociable task scenario of SciRoc, namely "E4 - Take the elevator". The main novelty concerns the implementation of the research study in a realistic social robotic competition environment. We investigate robot’s behaviors perceived by human by adopting the questionnaire developed and validate the reliability of the questionnaire as well (see Chapter 5). Fourthly, we introduce a novel approach for designing HRI task scenarios in the context of the SciRoc challenge. The new approach consists of two steps, analyzing the elements of a task scenario and sketching out its layout (see Chapter 6). We conclude the second part of the thesis by introducing qualitative research approaches for exploratory research study, quantitative research approach for confirmatory research study, and qualitative and quantitative (mixed-method) research approach in the field of social HRI (see Chapter 7). Finally, the third part summarizes the strengths and limitations of the research presented in this thesis, discussing potential future works in the field.