A Multimodal Attention Tracking in Human-Robot Interaction in Industrial Robots for Manufacturing Tasks

The field of human-robot interaction has seen tremendous growth in recent years, and the use of robots in manufacturing tasks has become increasingly common. However, the success of human-robot interaction is highly dependent on the ability of the robot to understand and adapt to the human operator's actions and attention. In this paper, we propose a novel approach that uses a context-aware natural language interface and position tracker to track the operator's attention and improve interaction with the robot. The system integrates multimodal inputs such as head pose estimation and intent recognition to accurately predict the operator's attention and adjust the robot's behaviors. The proposed approach is evaluated in a manufacturing logistic scenario, and the results show a significant improvement in collaboration and a reduction in errors in task completion. The approach is expected to have broad applicability in industrial manufacturing settings, where it can enhance productivity and efficiency by improving human-robot interaction.