Session: 16-01-01: Government Agency Student Poster Competition

Paper Number: 150195

150195 - Characterization of Human Trust in Robot Through Multimodal Physical and Physiological Biometrics in Human-Robot Partnerships 

With the rise of the AI boom, we find ourselves circumnavigating the unprecedented, entering a new era
of unforeseen technological advancements and social implications. One domain in particular growing at a
rapid pace is robotics, as the shift in technology has expedited robotic capabilities, sparking their gradual
incorporation into the workforce. Different sectors across a variety of industries are finding an increase in
productivity, by replacing repetitive, labor-intensive tasks with collaborative robots. This can also be
known as smart manufacturing, which in parallel, leads to a higher involvement of human-robot
collaboration. Human-robot collaboration in short is a symbiotic partnership between humans and robots
to accomplish a shared goal. This process may be a new paradigm to most, which may lead to an
unfamiliar, distrustful, and uncomfortable situation for inexperienced people to navigate. With these
sentiments in mind, our goal is to have a comprehensive understanding of the factors that affect people’s
trust in robots. Trust is an attribute that many people use daily, whether consciously or subconsciously.
Although commonly referred to as a firm belief in reliability, trust can be rather complex, and depending
on the individual has different meanings. Instead, trust can be an emotion, feeling, or choice, and can be
influenced in a variety of ways. To evaluate factors that may affect an individual's trust in robots, we
developed a novel trust database by exploring the trust between human collaborators and a robot. Each
collaborator wore four biological sensors and would conduct exercises with a robot performing
collaborative tasks. During these tasks, with the use of these sensors, we collected trust-related physical
and physiological human biometrics from the brain (EEG), heart (ECG), forearm (EMG), and eyes.
Additionally, we collected trust ratings through a questionnaire, enabling the production of a multimodal
human-robot trust database (TrustBase). TrustBase provides insightful guidance to optimize and improve
the environment deployment and robot configuration in human-robot partnerships within smart
manufacturing contexts. With the data from TrustBase, data analysis techniques were used to produce
time series and correlation graphs, to deliver a better understanding of the data. We were able to identify
outliers and make assumptions about psychological elements that may have occurred during the tasks
shared with the robot. Additionally, further computational and analytical approaches, including TabPFN,
XGBoost, and SVM, were used to investigate the correlation between robot performance factors and
humans’ trust levels and to characterize humans’ trust in robots during human-robot collaboration.
Results and their analysis suggest the effectiveness of the developed models, providing new findings to
the human factors and cognitive ergonomics in human-robot interaction.

Presenting Author: Jesse Parron Montclair State University

Presenting Author Biography: Jesse Parron is a research associate in collaborative robotics and smart systems laboratory at Montclair State University. His research interests include Robotics, Human-Robot Collaboration, Artificial Intelligence, Vision Systems.

Authors:

Jesse Parron Montclair State University
Weitian Wang Montclair State University