[Feature Proposal] Add map to odom TF publishing capability

[abudori3939]

Overview

This pull request adds the capability to publish the map->odom transform, complementing the existing map->base_link transform.

A new parameter, enable_map_odom_tf, has been added to param/localization.yaml. Setting this to true enables the new behavior. When false (the default), the node maintains its original behavior of publishing the map->base_link transform.

This feature allows for easier integration with standard ROS robot configurations that publish an odom->base_link transform (e.g., from wheel odometry), enabling a complete map->odom->base_link TF chain.

Implementation Details

The map->odom transform is calculated based on the following formula:
Transform(map→odom) = Transform(map→base_link) × Transform(odom→base_link)⁻¹

The process is as follows:

1. The map->base_link pose is estimated via NDT/GICP.
2. The odom->base_link transform is looked up from the TF tree.
3. The final map->odom transform is calculated from these two transforms and published.

Verification

I verified this feature using a robot that publishes wheel odometry (odom->base_footprint).

1. Original Behavior (enable_map_odom_tf: false)

With an existing odom->base_footprint transform, the map->base_link transform published by this node is not used, resulting in a disconnected TF tree.

2. New Behavior (enable_map_odom_tf: true)

With this feature enabled, the node publishes the map->odom transform, correctly forming a single, continuous TF chain: map->odom->base_footprint->base_link. This allows the localization result to coexist properly with the robot's odometry.

Notes

If you believe this feature would be beneficial to the project, I would appreciate your review and merge.

[rsasaki0109]

Thanks for the PR. I’ll check it later, hopefully one evening this week.

[rsasaki0109]

ref

オドメトリフレーム（/odom）とは何か
https://qiita.com/NaokiAkai/items/08deb338956c8da83aa9

[rsasaki0109]

Thank you. Everything looks good, so I’ve merged it. Great PR — thanks a lot!