An Ergonomic Evaluation of Robot-Initiated Safety Guidance in Retail Using Surface Electromyography

While retail mobile cobots typically handle routine service tasks, they may also serve as proactive safety-intervention systems, such as guiding individuals away from hazardous areas. However, the neuromuscular impact of such proactive robotic guidance on targeted users remains poorly understood. We conducted a mixed-reality experiment with 16 participants, simulating a retail shelf-picking task across three robot-initiated guidance modes: no robot, low-speed, and high-speed robotic intervention. Surface electromyography (EMG) recorded from 10 muscle groups showed that robotic guidance significantly increased muscle activity compared to the baseline ( p = .001). However, increasing robot speed did not significantly raise muscle activation ( p = .064). These findings suggest that carefully designed faster proactive guidance may help reduce hazard exposure time while maintaining comparable levels of neuromuscular demand relative to slower interventions. Future research would systematically evaluate the broader impacts of robotic guidance behaviors on human movements, neuromuscular, and safety outcomes.