06.02.2013 Views

Underwater Robots - Gianluca Antonelli.pdf

Underwater Robots - Gianluca Antonelli.pdf

Underwater Robots - Gianluca Antonelli.pdf

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

6.5 Singularity-Robust Task Priority 113<br />

the vehicle position, but, for seek of simplicity, let us consider only the joints<br />

positions by defining<br />

n� 1 q i,max − q i,min<br />

H ( q )=<br />

c i ( q i,max − q i )(q i − q i,min)<br />

i =1<br />

with c i > 0and the subscript max and min that obviously denotes the two<br />

joint limits. This function ([252]) inherits the concepts developed in [44].<br />

Its partial derivative with respect to the joint positions isgiven by<br />

∂H( q )<br />

∂qi<br />

= 1 ( q i,max − q i,min)(2q i − q i,max − q i,min)<br />

c i ( q i,max − q i ) 2 ( q i − q i,min) 2 .<br />

The elements of the weight matrix are then defined as<br />

� �<br />

�<br />

W i,i =1+ �<br />

∂H( q ) �<br />

�<br />

� ∂qi<br />

� ,<br />

in fact, it can be easily observed that the element goes to1when the joint<br />

is in the center of its allowed range and goes toinfinity when the joint is<br />

approaching its limit.<br />

As afurther improvement itispossible to relate the weight also to the<br />

direction of the joint bydefining<br />

⎧ � � � �<br />

�<br />

⎪⎨<br />

1+ � ∂H( q ) � �<br />

�<br />

� ∂qi � if ∆ � ∂H( q ) �<br />

�<br />

� ∂qi � > 0<br />

W i,i =<br />

� �<br />

⎪⎩<br />

�<br />

1 if∆�∂H( q ) �<br />

�<br />

� ∂qi � ≤ 0<br />

In Section 6.6, the joint limits are part of anumber oftasks handled with<br />

afuzzy approach. In [163], B.H. Jun, P.M. Lee and J. Lee propose afirst<br />

order task priority approach with the optimization ofspecific cost functions<br />

developed for the ROV named KORDI. The joints constraints are taken into<br />

account also.<br />

6.5 Singularity-Robust Task Priority<br />

To achieve an effective coordinated motion ofthe vehicle and manipulator<br />

while exploiting the redundant degrees of freedom available, G.<strong>Antonelli</strong><br />

and S.Chiaverini, in[20], resort tothe singularity-robust task priority redundancy<br />

resolution technique. The velocity vector ζ r is then computed as<br />

shown in (6.7).<br />

In the case of aUVMS, the primary task vector will usually include the<br />

end-effector task vector, while the secondary task vector might include the<br />

vehicle position coordinates. This choice isaimed at achieving station keeping<br />

of the vehicle as long as the end-effector task can be fulfilled with the sole<br />

manipulator arm. It is worth noticing that this approach isconceptually

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!