Closed ibergonzani closed 2 years ago
Hey, I added a comparison of the CoM Jacobian computed in Hyrodyn vs. your KDL based solution. It is a bit weird. In the fixed-based test case (compare_kdl_vs_hyrodyn), the KDL CoM Jacobian is different from the one computed in Hyrodyn. In the floating base case (compare_kdl_vs_hyrodyn_floating_base), they are identical
add comJacobian method to RobotModelKDL the implementation is naive atm: weighted average of spaceJacobians for each segment with mass in the KDL tree
the method made a copy of the original KDL tree, add a frame(segment) to the tree for each body with mass, placed at the COG . the COM jacobian is computed using the added new frames