Movement of the NICO Robot. Trustworthiness and Predictability of NICO Robot Movement
Abstract
The study comprises both theoretical and experimental components, with the primary objective of investigating the capabilities of the NICO robot platform within the context of human-robot interaction. The theoretical phase of the research elucidates fundamental concepts crucial to this investigation, notably the predictability and trustworthiness of humanoid robots. Predictability refers to the extent to which the robot's actions can be anticipated or foreseen by human users, influencing their ability to interact effectively with the robot [1].Trustworthiness, on the other hand, can be defined as the degree to which users perceive the robot as reliable, dependable, and worthy of confidence in its actions and behaviour [2]. Furthermore, accuracy within this context refers to the precision and correctness of the robot's responses and movements in relation to human input and expectations. Our study consists of two experiments to comprehensively examine these factors. In the upcoming experiments, we'll explore two main ways of moving the robot. First, there's inverse kinematics, where we figure out the angles needed from certain point coordinates. Then, there's retargeting, which imitates human motion.
Experiment 1 aims to assess the predictability of a robot's movements using two distinct methods: inverse kinematics and retargeting. Participants will be positioned opposite the robot, with a rectangular touch grid placed on the table between them. The robot will initiate movement towards the touch grid using its arm and head, halting at a randomly selected point. Participants will then endeavor to predict the specific point on the touch grid where the robot will stop. This process will be repeated under both conditions: one utilizing inverse kinematics and the other employing retargeting. Each participant will engage in both conditions. The predictability of the robot's movements will be evaluated based on the participants' accuracy in guessing the stopping point on the touch grid in each condition.
Experiment 2 aims to measure the accuracy and trustworthiness of the robot's actions. Similar to Experiment 1, participants and the robot will be situated in the same environment with the touch grid. This time, however, the human participants will perform movements and indicate random coordinates on the touch grid. The robot's task will be to mimic the human's movement and point to the same coordinates indicated by the participant. Accuracy will be determined by measuring the distance between the location pointed to by the human and the robot's pointing location. A smaller distance indicates higher accuracy on the part of the robot. As in Experiment 1, inverse kinematics and retargeting methods will be compared. Additionally, participants will complete a questionnaire after each condition to gauge their perception of the robot's trustworthiness.
The objective of this study is to investigate whether there are significant differences in accuracy, trustworthiness, and predictability between retargeting and inverse kinematics (IK). Demonstrating that retargeting outperforms IK would indicate that further investment in the development of retargeting technology is justified. It is important to note that retargeting is a more complex and time-consuming process compared to IK.
References
[1] A. D. Dragan, K. C. T. Lee, and S. S. Srinivasa, "Legibility and predictability of Robot Motion," 8th ACM/IEEE International Conference on Human-Robot Interaction (HRI), 2013. doi:10.1109/hri.2013.6483603
[2] Kok, B.C., Soh, H. Trust in Robots: Challenges and Opportunities. Curr Robot Rep 1, 297–309 (2020). https://doi.org/10.1007/s43154-020-00029-y
