In a virtual reality simulation, 16 participants manipulated an arm mounted on rover with joystick or look. The second mode is much faster and less tiring, but less precise for removal.
A robotic arm attached to a rover can extend the astronauts' hand to collect samples from the Moon, but continuously piloting it with a joystick requires constant hand-eye coordination, while the operator must also monitor the status of the systems. Researchers have developed a supervised gaze-assisted control: the operator stares at a rock or a deposit point, explicitly confirms his choice, and the robot performs the input and placement on its own, within predefined limits.
Faster, less mental load
Tested with 16 participants in a simulated virtual reality mission, with several levels of difficulty and simultaneous sound monitoring, this mode reduces the average time per rock from 47.65 to 18.43 seconds, with most of the gain occurring during input. The back and forth of the gaze between areas of interest drops from nearly 50 to less than 28 per minute, and the perceived workload (NASA-TLX index) drops from 32.5 to 18.3, without degrading the surveillance performance.
However, the joystick retains the advantage for precision dispensing, with fewer errors and more successes. The authors therefore recommend a division of roles: the gaze to quickly indicate the intention and let the machine act, the manual control for steps that require finesse.
Read the full article on the original source
Go to the source →