Page 182 - Human Inspired Dexterity in Robotic Manipulation

180 Human Inspired Dexterity in Robotic Manipulation

Table 9.4 Physical parameters
Three-fingered robotic hand

0.064 m
1st link length, l i1
0.064 m
2nd link length, l i2
0.030 m
3rd link length, l i3
0.023 m
1st center of mass, l gi1
0.035 m
2nd center of mass, l gi2
0.010 m
3rd center of mass, l gi3
0.038 kg
1st mass, m i1
0.024 kg
2nd mass, m i2
0.054 kg
3rd mass, m i3
(Fingertip)
0.015 m
Radius, r i
2
6
2.390 10 (N/m )
Physical properties, s i
Table 9.5 Details of the grasped object
Triangular prism
Mass (m) 0.0015 kg
Material Styrene foam
(Figure)
Length of side of triangle 0.060 m
Height 0.039 m

weight would be sufficiently light that the effect of gravity would be neg-
ligible. Each parameter of the grasped object and each gain are listed in
Tables 9.5 and 9.6. A stereo optical tracking system (Micron Tracker
H3-60, Claron, Inc.) was utilized to measure the position and orientation
of the grasped object. This system could measure the position and orienta-
tion of the grasped object every 50 ms by tracking a specific marker on the
object. The overall system is shown in Fig. 9.9. Two types of time delays
were present in the visual-sensing system, t sample and t image , as mentioned
in Section 9.4.1. The sampling rate of the system, h ¼ 20 Hz, was not suf-
ficient to allow the direct use of a real-time feedback approach. The average
time delay due to the image processing burden and data communication
latency was t image 50 ms. The time delay due to image processing was
observed in real time by measuring the time from when an image was cap-
tured to the beginning of data transmission. The time delay due to the
latency of data transmission from the visual-sensing system to the controller
was measured in advance.