Evaluating Variable Autonomy for a Teleoperated Navigation Task in a Habituated Environment

We investigated variable autonomy in a teleoperated robot navigation task within a simulated office environment, focusing on who controls the shift between levels of automation (LoA). Two switch control modes—human-initiated (HI) and robot-initiated (RI)—were examined. The study assessed whether these modes influence operators’ cognitive workload, situational awareness (SA), and task performance. The task involved two LoA conditions: teleoperation (low LoA) and remote monitoring (high LoA), while participants simultaneously performed a secondary task. Task performance was measured through completion time, collision rates, time spent in low LoA, and secondary task performance. Operator-related measures included workload [National Aeronautics and Space Administration Task Load Index (NASA-TLX)], trust, ease of use, and perceived SA. The results revealed a clear tradeoff between operational efficiency and situational awareness. The RI mode improved task efficiency and reduced workload but compromised SA, whereas HI mode enhanced SA at the cost of efficiency. These findings highlight the inherent tension between automation-driven performance gains and operator’s cognitive engagement. The study contributes to the understanding of variable autonomy in human–robot collaboration (HRC) and underscores the need to evaluate both operator experience and task performance when designing teleoperation systems for habituated environments.