An Evaluation of Operator Workload, During Partially-Autonomous Vehicle Operations

A series of three field experiments were conducted to evaluate the level of autonomous mobility for the Army’s Experimental Unmanned ground Vehicle (XUV), during which an assessment of operator mental workload was performed. Workload data collection methods employed were the “NASA Task Load Index”, the “Overall Workload” scale, and experimenter observation during partially autonomous operations, conducted within relevant operating environments including open arid, vegetated, and urban terrains. Although the vehicle was able to successfully traverse terrain unaided at a rate of almost 95%, the level of mental workload increased significantly during periods when human intervention became necessary. Terrain difficulty revealed the most significant effect, followed by an effect for distances traveled. Topography changes, resulting from inclement weather, caused unexpected increases in workload during voyages over what was thought to be the less difficult terrain, articulating the benefit for conducting tests within actual environments in exposing critical operational issues. Additionally, perceived mental workload was highly influenced during conduct of the more deliberate type missions (cautious approches to and from points of interest). The NASA-TLX subscale categories “Temporal” (the amount of time pressure felt), followed closely by “Mental” (the degree of recalling or calculating require), revealed highest workload demand, and comparison of NASA-TLX “Global” ratings with the “Overall Workload” collection method show acutely high correlation, demonstrating the latter an advantageous less obtrusive collection technique. Though the information exposed may be considered most beneficial as baseline performance criterion, it is reasonable to anticipate that as future operators become expected to perform ancillary assignments, reserve human mental capacities should