Search

Teleoperated robots have proven useful across various domains, as they can more readily search for survivors, survey collapsed and structurally unsound buildings, map out safe routes for rescue workers, and monitor rescue environments. A significant drawback of these robots is that they require the operator to perceive the environment indirectly. As such, camera angles, uneven terrain, lighting, and other environmental conditions can result in robots colliding with obstacles, getting stuck in rubble, and falling over (Casper & Murphy, 2003). To better understand how operators remotely perceive and navigate unmanned ground vehicles, the present work investigated operators’ abilities to negotiate corners of varying widths. In Experiment 1, we evaluated how instruction method impacts cornering time and collisions, looking specifically at the speed-accuracy tradeoff for negotiating corners. Participants navigated a virtual vehicle around corners under the instruction to focus on accuracy (i.e., avoiding collisions) or speed (i.e., negotiating the corners as quickly as possible). We found that as the task became more difficult, subjects’ cornering times increased, and their probability of successful cornering decreased. We also demonstrated that the Fitts’ law speed-accuracy tradeoff could be extended to a cornering task. In Experiment 2, we challenged two of the assumptions of Pastel et al.’s (2007) cornering law and assessed how corner angle and differences in path widths impacted cornering time. Participants navigated a virtual vehicle around corners of varying angles (45°, 90°, and 135°) and varying path widths. We found that increases in corner angle resulted in increased cornering times and a decreased probability of successful cornering. The findings from these experiments are applicable to contexts where an individual is tasked with remotely navigating around corners (e.g., video gaming, USAR, surveillance, military operations, training).