You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[FEATURE]: Publish estimates as TwistWithCovarianceStamped to support estimators #7
The current implementation publishes the state information using the marine_msgs DVL message. This should continue to be used. However, this message is not accepted by various estimators supported by ROS.
Feature Description
Create a new second publisher, which publishes the velocity estimates transformed into the appropriate coordinate frame and as a TwistWithCovarianceStamped. This will ensure that estimators such as the robot_localization and fuse can interpret the message.
Alternative Solutions
I need to verify that the model used by fuse will accept this message type. The message type supported by fuse should be preferred over the robot_localization message type as the robot_localization package has been deprecated in favor of fuse.
Oh I completely forgot that I added that. Maybe we should put the existing publisher on the odom_fsd frame and the new one on odom. We should double check how Fuse handles transforms. If it expects everything to be at the base frame, we might also want to think about a way to apply the adjoint transformation to that frame
Activity
rakeshv24 commentedon Nov 15, 2024
@evan-palmer is the expected coordinate frame for the estimators
base_linkorbase_link_fsd?evan-palmer commentedon Nov 15, 2024
This would be
base_linkrakeshv24 commentedon Nov 15, 2024
Also,
TwistWithCovarianceis part ofOdometry. Do we still need a separate publisher?evan-palmer commentedon Nov 15, 2024
Oh I completely forgot that I added that. Maybe we should put the existing publisher on the
odom_fsdframe and the new one onodom. We should double check how Fuse handles transforms. If it expects everything to be at the base frame, we might also want to think about a way to apply the adjoint transformation to that framerakeshv24 commentedon Nov 15, 2024
Makes sense!