Regressor generator additions #246
Conversation
| parent.getInertia().getMass() == 0.0) | ||
| { | ||
| framesNames.push_back(parentLinkName); | ||
| } |
There was a problem hiding this comment.
Basically if you have two consecutive massless links, connected by two fixed joints, consider both of them "frames"? This make sense, I noticed that you have some of those in the Walkman model.
There was a problem hiding this comment.
Just for safety, I would add a check that the parent has only one child, otherwise we risk to have the situation in which a "fake link (frame)" is a parent to a "real link". Adding such check should move us closer to the behavior that we get in non-KDL classes by two calls to removeFakeLinks, as described in #247 .
There was a problem hiding this comment.
As per f2f discussion, adding the single children check would not be sufficient to replicate the behavior we are interested in. Merging like this for now.
* Initial commit of HumanDynamicsEstimation Library * Fix CMake errors * Move Utilities into HumanDynamicsEstimation Library * Fix compilation of HumanStateProvider * Add HumanDynamicsEstimationLibrary namespace to Utils * minor fix in HumanDynamicsEstimation CMakeLists * update namespaces for HDE library * minor refactoring * Remove unused interface * change interfaces includes * Add algorithms namespace * Remove unused file * change library name to hde * minor refactoring * Reorganize cmake project and include scoping * Fix project cmake * fix visualizer includes * Clean include libraries * cleanup
This should include fake links that have fake children and ignore them (this sounds a bit weird. these links have a fake life!). There also is some output from the getDescriptionOfLinks() method now