multisense joint state runaway
Issue #46
open
In monitoring the /multisense/joint_states topic, it seems like the position data is drifting dramatically over time even though the head is not moving in the sim.
Here's an rxplot screenshot over about a 5 minutes time period:
Unless i'm confused about something I would expect 1) this data to be the neck tilt data, and 2) the value to stay within reasonable bounds in radians (whatever the RoM of the neck is).
Comments (4)
-
-
- changed status to open
-
-
assigned issue to
-
assigned issue to
-
reporter @btshrewsbury ah, sorry i thought it was neck q, but obviously this is spindle q. Regardless, the accumulation should probably be turned into a wrapped number. Thanks!
- Log in to comment
/multisense/joint_states should only contain the spindle q. I believe it is always spinning as we have an internal controller that we haven't exposed set to a constant velocity. We can wrap the q which should fix the runaway. I'll look into both the constant velocity and the wrap bug