Performance of Predictive Display Teleoperation under Different Delays with Different Degree of Freedoms

In teleoperation systems, operator performance can be significantly degraded even from small time-delays in visual feedback. This paper evaluated quantitatively tele-manipulation performance under different time delays from 0 to 4 seconds while controlling from 2 to 6 robot arm DOF's in an object alignment task. In our user study 15 subjects teleoperate 1650 alignment motions. Performance is quantitatively characterized by time delays, and qualitatively using NASA TLX analysis. We found that when time delay is small, the number of degree of freedoms has little effect on operating performance. However delay becomes larger the DOF influence is significant, with Higher DOF tasks taking longer. Furthermore, we found that graphical predictive display, that generates synthesized timely visual feedback, corrected almost all the degrading effects of time delay.