Scalability of Robotic Controllers: An Evaluation of Controller Options- Experiment III.

摘要

This experiment is the third in a series to investigate scaling robotic controllers for use by dismounted warfighters. A 2 -2 within-subjects design crossing two controllers (game controller GC) and virtual joystick (VJ) with two levels of robotic autonomy (manual control MC) and click to navigate (CN) was used. Twenty-two Soldiers conducted reconnaissance tasks on a 200-m course. As a secondary task, Soldiers were told to report course times when prompted. Neither the main effects for controller nor autonomy were statistically significant for driving errors or the number of unexploded ordnance rounds detected, nor were the controller and autonomy interactions. There was a significant controller and autonomy interaction regarding secondary task performance. With the CN feature, there was no significant difference in the efficiency of task performance between the GC and VJ. When using MC, the Soldiers' secondary task performance was significantly more efficient with the VJ than with the GC because the GC required two hands for operation. However, the participants generally preferred the GC over the VJ; the VJ required visual attention because it provided no haptic feedback. They also preferred MC to CN. Suggestions are offered for improving the design of both types of controllers.