부산대학교 도서관

Seamless human-robot collaboration in industrial assembly requires 1) dynamic robot path planning in response to unpredictable human behavior, and 2) communication of the robot's intentions to the human to avoid surprises. This paper presents initial findings from a study which utilizes multiple communication modalities to convey the robot's state to humans. Our approach incorporates a motion planner that takes the presence of both humans and the workspace into account. This integration allows the robot to make real-time decisions - it can plan a new path when an obstacle is encountered, or it can temporarily pause until the obstacle clears. The robot's decisions are communicated to the human either by Light, Voice, or Images+Text. We present preliminary results from a pilot user study with respect to execution time and cognitive workload. The results will be used to inform the design of a more comprehensive study.