Cultural Influences on Human-Robot Interaction: Effects of Robot Appearance and Control Modes

Abstract

This thesis investigates how cultural differences affect human-robot interaction (HRI), with a focus on physical HRI (pHRI) in collaborative tasks. It presents two empirical studies exploring the influence of cultural background on user perceptions, comfort, and acceptance when interacting with robots. Specifically, three cultural groups were considered: Western, Middle Eastern, and Chinese.

The first study examines how the robot’s physical form—humanoid (TALOS) versus non-humanoid (MOVO)—shapes user experiences among participants from the three cultural backgrounds. Results indicate that the TALOS robot tended to be perceived as more comfortable and approachable, especially during independent tasks, possibly due to its human-like features and movements. However, cultural background played a significant role in moderating these perceptions: Western and Chinese participants responded more positively to TALOS, while Middle Eastern participants showed more caution and discomfort. Physiological data, including heart rate and galvanic skin response, indicated higher stress levels during collaborative tasks across all groups, with lower stress levels typically observed when interacting with TALOS.

The second study explores how cultural background affects user preferences for robot control modes—autonomous versus human-controlled—and their responses to robot errors. Using the Sawyer collaborative robotic arm, participants performed tasks in both modes (users were deceived in believing the modes were different, while in reality, they were the same), with an intentional error introduced during one of the interactions. Most participants preferred the autonomous mode and reported higher comfort when the robot operated independently. However, Western and Chinese participants generally demonstrated higher trust in autonomous systems, whereas Middle Eastern participants tended to exhibit greater caution, particularly following errors in the human-controlled condition. Interestingly, users were generally more forgiving of mistakes in the human-controlled condition, often attributing errors to the human operator. Physiological responses supported these observations, showing increased stress during error conditions, with Western and Chinese participants recovering more quickly than Middle Eastern participants.

Together, these studies highlight the importance of considering cultural background in the design and deployment of robots in several applications, like healthcare, rehabilitation, and industrial automation. This work aims to support incorporating cultural awareness into robotic systems, leading to the future development of more inclusive, trustworthy, and user-friendly technologies.