Fig. 6: Static grasp experiments results.
From: High-speed aerial grasping using a soft drone with onboard perception
A Success rate for each experiment, comparing motion-capture-based performance (when available) with onboard-vision-based performance. Ten flights were conducted for each experiment, as indicated by the horizontal dashed line. B Refined target pose estimation errors for each object at the time of trajectory planning, averaged across all runs. C Actual forward velocity as measured by motion capture and VIO drift for varying commanded speeds, averaged across runs. D, E Position and velocity tracking errors over the grasp trajectory for the three objects. F, G Position and velocity errors for four different speeds. The vertical dashed line denotes the time of grasp and the shaded regions represent one standard deviation from the mean. The errors here are computed as the mismatch between the vision-based estimate and desired setpoint.
