Closed Thazz closed 2 years ago
Hi @Thazz
Are you talking about a specific node, for example loc2d_ros
, or should this feature be available in all nodes?
I think it only makes sense to implement it in loc2d_ros
as it is used for localization.
For slam2d_ros
it could be useful in case od lifelong mapping as discussed in https://github.com/iris-ua/iris_lama_ros/issues/38, where this node would also be used for localization, but as I understand from discussion, LaMa is currently not suitable for this.
Another question, is it possible to estimate covariance matrix for localization, so that pose could be published as geometry_msgs::PoseWithCovarianceStamped
message?
I think I can add the option to disable tf broadcasting for loc2d_ros
.
Another question, is it possible to estimate covariance matrix for localization, so that pose could be published as geometry_msgs::PoseWithCovarianceStamped message?
This already exist for loc2d_ros
I would propose to add a parameter to optionally disable publishing of
map->odom
tf transfrom, such astf_broadcast
parameter in amcl package.Use case for this is when one would be using robot_localization package for fusing multiple localization sources with Kalman filter (e.g localization with lama + bluetooth beacons or GPS). In this case
map->odom
is (or can be) published byrobot_localization
nodes and should be disabled forrobot_localization
input sources, i.e lama ros.Edit: I see that using Kalman filter is also discussed in https://github.com/iris-ua/iris_lama_ros/issues/5.