Re: regarding Constraints of my arm configuration

First, read the announcement of Peter regarding the new release of Orocos 0.24 
and the release notes of that version:
	http://www.orocos.org/releases/orocos-changes-0.24.0.html
It tells you about the new taskcontext component model that is/will be used 
instead of the controlkernel components.

I do not now the exact state of orocos 0.20.  Perhaps some other people know.

What kinematic functions you need, depend on the controller you want to build.
We haven't used ourself a dynamic model in Orocos (yet ?).  We commonly use 
FORWARD POSITION KINEMATICS to calculated the actual cartesian position of 
the robots. A cartesian controller that takes this actual position and the 
desired position and calculates a cartesian velocity.
The INVERSE VELOCITY KINEMATICS are used to transform this cartesian 
velocities into joint space velocities that are sent to the robot hardware.

I'll point you to the relevant controlkernel components ;
packages/control_framework/control_kernel_components/motion_control/cartesian/current



On Tuesday 27 June 2006 12:21, H.S. Randhawa wrote:
> I have calculated the inverse kinematics for my 5 DOF robotic arm. I want
> to know that is it necessary to calculate the differential kinematics and
> dynamics for the same. I want to make it run through Orocos 0.20.
>
> On 6/15/06, Erwin Aertbelien <[email protected]> wrote:
> > This type of robot is not used for completely general robot tasks, but
> > for task that only need 5 d.o.f. e.g. tasks with a rotation symmetric
> > tool (like
> > a grinding disk).
> >
> > Kinematics for this are not provided in Orocos.
> >
> > There are two problems to solve :
> > (1) solving the inverse kinematics.
> >         -For your type of robot there still exists a closed form
> > solution.:
> >                 - out of the specified orientation you can determine the
> > POSITION of the
> > intersection between axes 4 and 5.
> >                 - This position determines the first three axes.
> >                 - Once the first three axes are determined, you can
> > determine the two last.
> > (2) Specification :
> >         you have to specify your task in 5 dof task space.
> >         I know of two ways to do this :
> >                 - using a 6 dof frame specification, and ignore a degree
> > of freedom (e.g.
> > rotation around X-axis).
> >                 - using a 5-d.o.f. specification, like in Numerical
> > Control (NC) milling
> > machines :
> >                         * position (3)
> >                         * two direction cosinuses A and B,
> >                    i.e. your 2 dof orientation can be specified with a
> > normalized vector.
> > The projection of this vector in some plane are A and B.
> >         This choice of specification is important because it will
> > determine your
> > ability to cooperate with other components.
> >
> > Best regards,
> > Erwin
> >
> > On Thursday 15 June 2006 07:28, H.S. Randhawa wrote:
> > > I mailed yesterday regarding the configuration which is littlebit
> >
> > different
> >
> > > from Serial 321 configuration.
> > > In this configuration the following spec. along with motor rotation are
> > > there:
> > >
> > > LINK                 LENGTH           MOTOR (rotation angle)
> > > Rotation axis(according to OROCOS)
> > >                                                   [ Counter-Clockwise
> >
> > +ve ]
> >
> > > BASE              3"                       0-180 Degree
> > >       Z
> > >
> > > SHOULDER     4.75"                  0-180 Degree
> > > X
> > >
> > > ELBOW           4.75"                  -150  to  +30 Degree
> > > X
> > >
> > > WRIST            5.75"                  -90   to
> > > 90                             parallel to Shoulder and elbow
> > >
> > > WRIST ROTATION -----             -90  to 90
> > > Perpendicular to Wrist.
> > >
> > > GRIPPER                               opening 2" max.
> > > Perpendicular to both wrist and wristRotation axes.
> > >
> > >
> > >
> > > The serial family has six *revolute* joints: a "wrist" with *three*
> > > revolute joints with intersecting axes; an "arm" with *two* parallel
> >
> > axes;
> >
> > > and a "shoulder" with *one* revolute joint whose axis is orthogonal to
> >
> > the
> >
> > > axes of the "arm" set. but my arm is having three axes parallel (2nd,
> >
> > 3rd
> >
> > > and 4th) and more over  last three are  not intersecting  at a point.
> >
> > Only
> >
> > > 4th and 5th are intersecting. Sixth is at some offset of 2.2" from that
> > > intersecting point.
> > >
> > > As this is not exactly a 6 DOF arm Its a 5 + gripper but still if we
> >
> > ignore
> >
> > > the last joint and want to move it through desire points, it should
> > > move accordingly.
> > > For some given point which is already in its workspace, Orocos is
> > > calculating joint positions (radians) which are contradicting with
> > > motor constraints so robot can't move to that point. eg  -ve value for
> > > base,
> >
> > or
> >
> > > some other value which is out of range for that particular joint motor.
> > > How it can be resolved?
> > >
> > > Thanks
> > > H.S.Randhawa
> >
> > Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

_______________________________________________
Orocos mailing list
[email protected]
http://lists.mech.kuleuven.be/mailman/listinfo/orocos

Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm