REFERENCES 143
0.1
.-.. 0
E
-0.1
IBC:
beta=0.02 rad; noise:2 pixels
.....
object
0.4 0.6 0.8
(a) tracking traj. in x-y plane x (m)
IBCDC: beta=0.02 rad; noise:2 plxels
0.1 - 9 9 9
0.4 0.6 0.8 1
(b)
tracking traj. in x-y plane x (m)
.04
.02
._:
o
IBC: beta=0.02 rad; noise:2 plxels
~
.
4 6 8
(c) tracking error t (sec)
10
~ .04
.02
.-
o
IBCDC: beta=O.02 rad; noise:2 pixels
4 6 8
(d)
tracking error
t (sec)
IBC:
beta=O.02
tad;
noise:2
pixels
_
,
0.4
0.2
-0.2
0 2 4 6 8
(e) estimated 2D motion para. t (sec)
IBCDC: beta=0.02 tad; noise:2 pixels
0.6
0.4
0.2
0
-0.2
0 2 4 6 8
(f) estimated 2D motion para. t (sec)
FIGURE 4.7
Performance comparison of IBC and IBCDC.
IBCDC: beta--0.02 tad; noise:2 pixels
o~ 0.2
0
.0.2
-0.4
o- 0 2 4 6 8
(g) part of estimated cali.
data
other hand, using the OAT orientation representation of the object, the control design for
grasping is solved as an orientation tracking problem.
In our analysis of tracking and grasping, regular polyhedral objects are assumed. The
approach needs to be extended to deal with objects of arbitrary shape in order to improve
its applicability. Furthermore, although it has been demonstrated by simulations that the
tracking and grasping algorithms have good performance in terms of speed and tracking
accuracy, we believe that further effort should be made to utilize video-rate pipelined
hardware or parallel processor arrays if we are ever to achieve real-time capabilities.
REFERENCES
[1] L. E. Weiss, A. C. Sanderson, and C. P. Neuman, Dynamic sensor-based control of robots with visual feedback.
IEEE J. Robot. Automat.
RA-3(5):404-417, 1987.

Get Control in Robotics and Automation now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.