Martin Günther

This is the script you're looking for: https://github.com/android-password-store/Android-Password-Store/issues/173#issuecomment-473557194

What values do you get (you can run `rostopic echo -n1 `) and what were you expecting? Also, did you look at RViz and check that the tracking works correctly?

Small comment: I believe geometry_msgs/Twist gives the speed in m/s. If you expect 40 km/h, that should be about 11.11 m/s. Still a far way off from the 0.85 or...

Perhaps @TontonTremblay has a more specific answer, but I guess the short answer is that something went wrong during the mesh export/training phase. Every tool has its own conventions. For...

> Is this relative transform (the one in Martin's config_pose.yaml file) object specific or constant for all objects? It's object specific. > We are also a bit confused because while...

The easiest way is to use ROS and [TF](http://wiki.ros.org/tf/Tutorials). It's a bit of a learning curve, but absolutely worth it for solving problems such as yours. Since this is more...

> Is the following model_info.json correct? No, it should look something like this: ```json { "symmetries_discrete": [[ 1, 0, 0, 0, 0, -1, 0, 0, 0, 0, -1, 0, 0,...

Did you specify the symmetries? See here: https://github.com/NVlabs/Deep_Object_Pose/tree/master/scripts/nvisii_data_gen#handling-objects-with-symmetries Also please show a visualization of the belief maps.

Damn, when I developed the symmetry handling, I *knew* somebody would try to model a sphere at some point, but I hoped it wouldn't be so soon. Spheres are not...

Ah damn, I was afraid that would happen. The problem is that what my proposal is doing is very inefficient: * we have symmetries on all 3 axes, discretized into...