Decomposed Physical Workload Estimation for Human-Robot Teams

摘要

Human-robot teams operate in uncertain environments to accomplish a wide range of tasks. A dynamic under-standing of the human’s workload can enable fluid interactions between team members. Workload can be decomposed into workload components (e.g., cognitive, visual, speech, auditory, gross motor, fine motor, and tactile). A system that seeks to adapt interactions for a human-robot team needs to understand the distribution of workload across the different components. Prior work treated the gross motor, fine motor, and tactile components as a joint physical workload. The presented algorithm estimates gross motor, fine motor, and tactile workload for a human-robot team operating in a non-sedentary supervisory environment; however, noise and task uncertainty lead to mixed results. The metrics for this algorithm were collected using a diverse set of wearable sensors, including heart rate monitors, motion trackers, and surface electromyography sensors.