Wenbin-Xu
commented
7 years ago `--- transforms:
-
header: seq: 0 stamp: secs: 1503993101 nsecs: 130777577 frame_id: LeftHandPinky2 child_frame_id: LeftHandPinky3 transform: translation: x: -4.22016831861e+35 y: 0.0 z: 1.03139144699e-16 rotation: x: 3.28584871706e-40 y: 3.92363570011e-43 z: 3.3791611818e-40 w: 1.0 ---` One part of the data I've printed out I think is a little inaccurate because of the 'e+35'. But for other part it is acceptable.
According to the file which serves as a support on the official websites, I think there must only be 57 pairs of child and parent frames. However, there is three more in your code, which are one fixed world frame and two frames, "WorldPerceptionNeuron“，”HipsPosition". One fixed world frame to fix the relative location and angle of one root node is understandable, but why you use these two more frames? (According to the file ) I think the addition of two extra frames could cause all the results to be wrong. Also using your code, I've found something wrong.