ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

Unfortunately, the only way that you can approximate this is when generating TP programs in the traditional, off-line manner that CAM tools do this. Fanuc robots don't have the motion interfaces that would allow an on-line, closed loop implementation. The Insititute Maupertuis has made a ..

Unfortunately, the only way that you can approximate this is when generating TP programs in the traditional, off-line manner that CAM tools do this. Fanuc robots don't have the motion interfaces that would allow an on-line, closed loop implementation. implementation.

The Insititute Maupertuis has made a ..package available that implements a basic 'post processor' for Fanuc TP (ie: .ls programs): InstitutMaupertuis/fanuc_post_processor.

You could use that to generate programs and then upload them to the controller.

Note: I wrote 'approximate this' as even with this the robot will still not always do exactly as you'd like it to do, but at least the limitation is now on the robot controller side, not on the ROS side.

Unfortunately, the only way that you can approximate this is when generating TP programs in the traditional, off-line manner that CAM tools do this. Fanuc robots don't have the motion interfaces that would allow an on-line, closed loop implementation.implementation. This is not really a ROS limitation, it's just the way Fanuc controllers work.

The Insititute Maupertuis has made a package available that implements a basic 'post processor' for Fanuc TP (ie: .ls programs): InstitutMaupertuis/fanuc_post_processor.

You could use that to generate programs and then upload them to the controller.

Note: I wrote 'approximate this' as even with this the robot will still not always do exactly as you'd like it to do, but at least the limitation is now on the robot controller side, not on the ROS side.

Unfortunately, the only way that you can approximate this is when generating TP programs in the traditional, off-line manner that CAM tools do this. Fanuc robots don't have the motion interfaces that would allow an on-line, closed loop implementation. This is not really a ROS limitation, it's just the way Fanuc controllers work.

The Insititute Maupertuis has made a package available that implements a basic 'post processor' for Fanuc TP (ie: .ls programs): InstitutMaupertuis/fanuc_post_processor.

You could use that to generate programs and then upload them to the controller.

Note: I wrote 'approximate this' as even with this the robot will still not always do exactly as you'd like it to do, do (due to interpolation/rounding/blending/etc), but at least the limitation is now on the robot controller side, not on the ROS side.

Unfortunately, the only way that you can approximate this is when generating TP programs in the traditional, off-line manner that CAM tools do this. Fanuc robots don't have the motion interfaces that would allow an on-line, closed loop implementation. This is not really a ROS limitation, it's just the way Fanuc controllers work.

The Insititute Maupertuis has made a package available that implements a basic 'post processor' for Fanuc TP (ie: .ls programs): InstitutMaupertuis/fanuc_post_processor.

You could use that to generate programs and then upload them to the controller.

Note: I wrote 'approximate this' as even with this the robot will still not always do exactly as you'd like it to do (due to interpolation/rounding/blending/etc), but at least the limitation is now on the robot controller side, not on the ROS side.

Note2: one of the major drawbacks of off-line program generation and upload is the inherent limitation of the length of the programs that can be used this way and the overhead experienced by the controller when it's parsing the new program. It's quite easy to quickly run into the limitations wrt maximum trajectory length and program size. An on-line, externally controlled, approach avoids these limitations (but of course is also not perfect).