slam with d455

[karimrahimi76]

Hi all ,
i am trying to develop slam with intelRealsense camera d455 for my robot
i converted the 3d point cloud data to 2d lidar data using depth_to_laser package
now i want to develop nav2 but i am confused
for localization , i think amcl is not good choice because realsense han not 360 degree fov and amcl may not work correctly
also for odometry data i have wheel encoder and i want to fuse it with realsense imu data
is it good to fuse realsense imu with odometry for get accurate odometry or use something like rgdb odometry from rtabmap package
also for localization in nav2 should i use robot localization package instead of amcl ?
and what if i use robot localization , what about map->odom tf
is robot localization publish it correctly ?
thank you all

[MartyG-RealSense]

Hi @karimrahimi76 Thanks very much for your questions.

I am not familiar with using nav2 unfortunately, so do not have advice to provide about questions involving nav2. The recent nav2 related cases at IntelRealSense/librealsense#13468 and #3243 may be helpful references for you though.

RealSense 400 Series stereo depth cameras that have an IMU are not usually used with wheel encoders. There was a now-retired camera product called the RealSense T265 Tracking Camera that supported that feature.

However, the link below has a guide to using a RealSense D435i with IMU data and wheel encoding. The information should be applicable to the D455 camera model.

https://github.com/fazildgr8/realsense_explorer_bot

robot_localization should be able to publish odom, yes - please see #3173