3.6 Simulations and experiments
Figure 3.11: Trajectories of the fingers generated by the local grasp planner (green:
approach, red: grasp) and the corresponding sequence of floating grasp points
achieved during the reconstruction process (yellow) for the two objects, both eval-
uated with k
c
m
= 1 (left) and k
c
m
= 0 (right).
surface. The green lines represent the trajectories that the planner generates for
the fingertips after spatial filtering and considering the safety distance. Finally,
the red lines show the last part of the grasp trajectory, when the safety distance is
progressively reduced achieving a perpendicular approach to the object's surface.
It is worth noticing that in the case k
c
m
= 0 the fingers aim to reach a larger
polygon area to grasp since they are not anymore constrained to have the grasp
plane containing the center of mass.
For the case of the little bottle, four fingers of the hand have been considered.
The final grasp configuration corresponds to the equilateral best grasp (90
o
apart)
for the object. Moreover, the achieved trajectories are very regular due to the
good choice of the initial grasp configuration evaluated by the preshaping module.
This result is common when the object is symmetric with respect to one or more
axes, and so it is well represented by an ellipsoidal surface. Of course, for the
particular shape of the bottle, the results do not change significantly when k
c
m
is
set to 0.
Furthermore, to validate the proposed method, a comparison between the re-
39
Summary :
3.6 Simulations and experiments Figure 3.11: Trajectories of the fingers generated by the local grasp planner (green: approach, red: grasp) and the corresponding sequence of floating grasp points achieved during the reconstruction process (yellow) for the two objects, both eval- uated with k c m = 1 (left) and k c m = 0 (right).
Tags :
grasp,fingers,surface,trajectories,safety,distance,planner,object,case,configuration,achieed,hae,green