How to configure odom frame for move_base with only LiDAR and IMU odometry (no wheel odometry)?

[emirhancibir]

I am using the hdl_localization package with move_base, but I don't have wheel odometry. My robot relies solely on LiDAR and IMU odometry for navigation.

Here is my current setup:

The hdl_localization node publishes the transform between the map frame and my sensor frame (laser) via the /odom topic.
I am using a static transform tree:
laser -> base_link -> ...
However, move_base requires an odom frame. I am unsure how to configure this in my current setup.

Should I use the odom frame directly from hdl_localization, or is there a better approach for integrating the odometry output with move_base?
How do I ensure the odom frame is consistent with move_base requirements in a system with no wheel odometry?
Any guidance or best practices would be greatly appreciated.

<!-- globalmap_server_nodelet -->
<node pkg="nodelet" type="nodelet" name="globalmap_server_nodelet" args="load hdl_localization/GlobalmapServerNodelet $(arg nodelet_manager)"  output="screen">
  <param name="globalmap_pcd" value="$(find servant_t)/pcd/liorf2/GlobalMap.pcd" />
  <param name="convert_utm_to_local" value="true" />
  <param name="downsample_resolution" value="0.1" />
</node>

<!-- hdl_localization_nodelet -->
<node pkg="nodelet" type="nodelet" name="hdl_localization_nodelet" args="load hdl_localization/HdlLocalizationNodelet $(arg nodelet_manager)">
  <remap from="/velodyne_points" to="$(arg points_topic)" />
  <remap from="/gpsimu_driver/imu_data" to="$(arg imu_topic)" />
  <!-- <remap from="/odom" to="/lidar_odom"/> -->
  <!-- odometry frame_id -->
  <param name="odom_child_frame_id" value="$(arg odom_child_frame_id)" />
  <!-- imu settings -->
  <!-- during "cool_time", imu inputs are ignored -->
  <param name="use_imu" value="$(arg use_imu)" />
  <param name="invert_acc" value="$(arg invert_imu_acc)" />
  <param name="invert_gyro" value="$(arg invert_imu_gyro)" />
  <param name="cool_time_duration" value="2.0" />
  <!-- robot odometry-based prediction -->
  <param name="enable_robot_odometry_prediction" value="$(arg enable_robot_odometry_prediction)" />
  <param name="robot_odom_frame_id" value="$(arg robot_odom_frame_id)" />
  <!-- ndt settings -->
  <!-- available reg_methods: NDT_OMP, NDT_CUDA_P2D, NDT_CUDA_D2D-->
  <param name="reg_method" value="NDT_OMP" />
  <!-- if NDT is slow for your PC, try DIRECT1 serach method, which is a bit unstable but extremely fast -->
  <param name="ndt_neighbor_search_method" value="DIRECT1" />
  <param name="ndt_neighbor_search_radius" value="2.0" />
  <param name="ndt_resolution" value="1.0" />
  <param name="downsample_resolution" value="0.1" />
  <!-- if "specify_init_pose" is true, pose estimator will be initialized with the following params -->
  <!-- otherwise, you need to input an initial pose with "2D Pose Estimate" on rviz" -->
  <param name="specify_init_pose" value="true" />
  <param name="init_pos_x" value="0.0" />
  <param name="init_pos_y" value="0.0" />
  <param name="init_pos_z" value="0.0" />
  <param name="init_ori_w" value="1.0" />
  <param name="init_ori_x" value="0.0" />
  <param name="init_ori_y" value="0.0" />
  <param name="init_ori_z" value="0.0" />

  <param name="use_global_localization" value="$(arg use_global_localization)" />
</node>

<node pkg="hdl_localization" type="plot_status.py" name="plot_estimation_errors" if="$(arg plot_estimation_errors)" />