Research Methodology Survey
about User Analysis & Evaluation
from the perspective of social robot
Research Keywords
HRI(Human-Robot Interaction);
HMI(Human-Machine Interaction);
Social Robot;
Emotion Interaction;
UX Design;
User Analysis;
Research Statement
This research examines the current UX design of social and service robots and discusses its limitation and focusing on user research among UX design, it aims to investigate and provide user analysis, requirements derivation method, and user evaluation method. Methods to objectify and quantify user analysis and evaluation methods for the application of social robots in life are mainly covered in the research.
This study will analyze the causes of success and failure in the aspects of literature review & survey research, exploratory research, and experimental research, which have been used for user analysis, targeting the currently developed social robots, and will suggest ways to expand the application of social robots. In addition, by analyzing and supplementing the limitations of the existing user analysis method, his research will present an objectified and quantified user analysis method from the perspective of HRI applicable to the actual development process
The expected result aims to create a social robot that can share emotions and communicate with them, which can be applied in real life, by suggesting user-defined and user analysis methods suitable for social robots to be developed.
Yong Seop Kwon
Figure 1
An example experimental setup where a participant is involved in an interaction with the humanoid robot iCub , while neuro-cognitive mechanisms are measured through performance data, as well as eye tracking and EEG., 2020
1
2
Figure 2
The research process of scientific methods
As shown in Figure 2 , HRI research usually uses the scientific method used in social science research.
A scientific method is a method of observing and verifying phenomena to explain phenomena.
Based on the logic of deductive and inductive, the purpose is to observe, theory, experiment and reproduce phenomena.
Social science is an academic discipline with a scientific approach to explain the phenomena of human society, and it is in line with the purpose of HRI research to explain robots in human society, so HRI studies utilize the scientific methods used in social science research.
3
_Measurement%20u.png)
Figure 3
NARS(Negative Attitude towards Robots Scale) Questionnaire
HRI Research Method - I
Measurement using a Questionnaire
Questionnaire is a research tool consisting of questions to measure the response of the participants in the experiment in a standardized manner.
The Negative Attitude Towards Robots Scale (NARS) is a survey developed to measure psychological factors that prevent interaction with robots in everyday life (Nomura et al., 2006). It was first developed in Japanese by Tatsuya Nomura of Ryukoku University in Japan, and is translated into English and widely used.
NARS measures respondents' negative attitudes from three perspectives: first, respondents' negative attitude toward the situation in which they interact with the robot, second, respondents' negative attitude toward the social impact of the robot, and third, their negative attitude toward the robot's feelings in interaction with the robot.
_Measurement.png)
_Measurement.png)
HRI Research Method - II
Measurement using a Biological Signals
Researchers use a method of measuring the response of participants through biometric signals.
At the University of British Columbia in Canada, experiment participants' anxiety about the movement of robot arms with path generation algorithms for safety-considered robot arms and path generation algorithms for non-objected robot arms was measured using a physiological sensor.
At RWTH Aachen University in Germany, as the human-likeness of the interactee increases, The cerebral cortex related to Theory of Mind was expected to be activated in the human brain, and this was measured using MRI.
6
6
_Measurement.png)
7
_Measurement.png)
HRI Research Method - III
Measurement using a Non-Constrained Sensor
Since measurement using bio-signals uses various changes occurring in the body, it is advantageous for directly measuring human reactions, but information is inevitably insufficient for analysis considering the contextual context.
In order to overcome these limitations, video and audio information can be used for reaction analysis that considers contextual contextual information. These information can be collected from non-Constrained sensors (cameras, microphones, etc.), and human behavioral reactions interacting with robots can be analyzed. It allows real-time analysis according to context.