Presentation
Do Tired Humans Trust Robots Differently? Effects of Physical Fatigue and Robot Reliability in pHRI
DescriptionWith advancements in artificial intelligence, robots are increasingly common in industrial settings; however, in applications such as custom manufacturing, full automation remains challenging due to unstructured environments. Humans and robots possess complementary skills, and combining the two can address these challenges, making it critical to understand how operators interact with robots under real-world conditions. Workers often experience physical fatigue over the course of a shift, yet little is known about how fatigue shapes trust, performance, and neuromuscular responses during physical human-robot collaboration. We conducted a within-subjects experiment with 40 participants performing a co-lifting task with a UR10e robot under two fatigue conditions (absent, present) and two levels of robot reliability (high, low). We measured subjective trust, perceived exertion, trajectory similarity, and muscle activity from four upper-limb muscles using surface EMG. Results revealed that both fatigue and unreliable robot behavior significantly reduced trust, even though the robot's actual capability remained unchanged across fatigue conditions. Fatigue also degraded trajectory similarity and produced clear neuromuscular signatures of compensatory motor strategies. These findings suggest that robots should sense and adapt to operator fatigue in real time, since ignoring human state can erode trust and coordination even when the robot performs flawlessly.
Event Type
Lecture
TimeThursday, October 22nd4:50pm - 5:10pm PDT
Location
Similar Presentations