Workers’ Trust in Collaborative Construction Robots: EEG-Based Trust Recognition in an Immersive Environment

Abstract

Today, advances in robotics and autonomous systems enable construction workers to collaboratively work with robots, assigning physically demanding tasks to them. However, working alongside an industrial robot is a novel experience that may take a heavy toll on workers’ bodies and minds, particularly in dynamic and uncertain environments, such as construction sites. Given that trust is identified as a critical element for successful cooperation between humans and robots, effective measurement of workers’ trust in collaborative robots can lead to practical evaluation of human-robot partnership. In this context, most studies of trust have relied on self-reports. Nevertheless, questionnaires are unable to determine the trust promptly, impeding early preventive interventions. Furthermore, they are invasive, interfering with workers’ daily operations. To address these issues, this study proposes a procedure to non-invasively and continuously recognize workers’ trust in collaborative construction robots using electroencephalogram (EEG) signals. To that end, an experiment was conducted in which human workers performed a collaborative construction task (i.e., handling materials) with a virtual robot in an immersive environment. Meanwhile, workers’ EEG signals were recorded using a wearable sensor. Subsequently, the level of trust of the workers in collaborative robots was measured using the Trust Perception Scale-HRI. By analyzing acquired signals and applying different machine learning algorithms, it was found that EEG signals can be implemented to differentiate levels of trust of construction workers in their robotic counterparts. These findings suggest the feasibility of using workers’ EEG signals as a reliable, real-time indicator of trust in collaborative construction robots, which can be regarded as a practical approach for evaluating human-robot collaboration.