Is this the way to use lidar & IMU efficiently?

Posted by Mikor on
URL: http://official-rtab-map-forum.206.s1.nabble.com/Is-this-the-way-to-use-lidar-IMU-efficiently-tp7496.html

Hello,

I am using the data pulled from kitti sequence 0117 with kitti2bag and here is the information that the bag contains.

path:        kitti_2011_09_26_drive_0117_synced.bag
version:     2.0
duration:    1:08s (68s)
start:       Sep 26 2011 15:40:43.44 (1317040843.44)
end:         Sep 26 2011 15:41:51.59 (1317040911.59)
size:        3.5 GB
messages:    9240
compression: none [2641/2641 chunks]
types:       geometry_msgs/TwistStamped [98d34b0043a2093cf9d9345ab6eef12e]
             sensor_msgs/CameraInfo     [c9a58c1b0b154e0e6da7578cb991d214]
             sensor_msgs/Image          [060021388200f6f0f447d0fcd9c64743]
             sensor_msgs/Imu            [6a62c6daae103f4ff57a132d6f95cec2]
             sensor_msgs/NavSatFix      [2d3a8cd499b9b4a0249fb98fd05cfa48]
             sensor_msgs/PointCloud2    [1158d486dd51d683ce2f1be655c3c181]
             tf2_msgs/TFMessage         [94810edda583a504dfda3829e70d7eec]
topics:      /kitti/camera_color_left/camera_info    660 msgs    : sensor_msgs/CameraInfo    
             /kitti/camera_color_left/image_raw      660 msgs    : sensor_msgs/Image        
             /kitti/camera_color_right/camera_info   660 msgs    : sensor_msgs/CameraInfo    
             /kitti/camera_color_right/image_raw     660 msgs    : sensor_msgs/Image        
             /kitti/camera_gray_left/camera_info     660 msgs    : sensor_msgs/CameraInfo    
             /kitti/camera_gray_left/image_raw       660 msgs    : sensor_msgs/Image        
             /kitti/camera_gray_right/camera_info    660 msgs    : sensor_msgs/CameraInfo    
             /kitti/camera_gray_right/image_raw      660 msgs    : sensor_msgs/Image        
             /kitti/oxts/gps/fix                     660 msgs    : sensor_msgs/NavSatFix    
             /kitti/oxts/gps/vel                     660 msgs    : geometry_msgs/TwistStamped
             /kitti/oxts/imu                         660 msgs    : sensor_msgs/Imu          
             /kitti/velo/pointcloud                  660 msgs    : sensor_msgs/PointCloud2  
             /tf                                     660 msgs    : tf2_msgs/TFMessage        
             /tf_static                              660 msgs    : tf2_msgs/TFMessage


I am interested in using only the lidar + imu data and I want to learn how to do that efficiently as the title says. At this time all I do is run rtabmap.launch with the following parameters.

roslaunch rtabmap_ros rtabmap.launch    \
   use_sim_time:=true \
   depth:=false \
   subscribe_scan_cloud:=true \
   frame_id:=base_link \
   scan_cloud_topic:=/kitti/velo/pointcloud \
   imu_topic:=/kitti/oxts/imu \
   scan_cloud_max_points:=131072  \
   icp_odometry:=true \
   approx_sync:=false \
   args:="-d  \
      --RGBD/CreateOccupancyGrid false \
      --Rtabmap/DetectionRate 2 \
      --Odom/ScanKeyFrameThr 0.8 \
      --OdomF2M/ScanMaxSize 10000  \
      --OdomF2M/ScanSubtractRadius 0.5   \
      --Icp/PM true \
      --Icp/VoxelSize 0.5   \
      --Icp/MaxTranslation 2   \
      --Icp/MaxCorrespondenceDistance 1.5 \
      --Icp/PMOutlierRatio 0.7 \
      --Icp/Iterations 10 \
      --Icp/PointToPlane false \
      --Icp/PMMatcherKnn 3 \
      --Icp/PMMatcherEpsilon 1 \
      --Icp/Epsilon 0.0001 \
      --Icp/PointToPlaneK 0 \
      --Icp/PointToPlaneRadius 0 \
      --Icp/CorrespondenceRatio 0.01"

This parameters are taken from this post . In order to add the IMU data I added the imu_topic and it seems to give slightly better results so I guess that the IMU data are included.

My questions are the following.
I think that this is the way to subscribe to the imu topics that the bag contain. Did I understand this correctly ?
Also, I need some more info about the ICP parameters, can you point me somewhere in order to do so ?

I am new to rtabmap and ros in general, every information that will help me improve my understanding will be much appreciated. Thank you in advance.