Different results after every excecution.

[Mikor]

Hello,

I noticed that after running rtabmap.launch with the same parameters multiple times I get different results but I didn't notice that when I was using it previously. For example, with these parameters:

roslaunch rtabmap_ros rtabmap.launch       use_sim_time:=true    depth:=false    subscribe_scan_cloud:=true    frame_id:=velodyne    scan_cloud_topic:=/pointcloud2    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 0.2   \
      --Icp/MaxTranslation 2   \
      --Icp/MaxCorrespondenceDistance 0.5 \
      --Icp/PMOutlierRatio 0.7 \
      --Icp/Iterations 10 \
      --Icp/PointToPlane true \
      --Icp/PMMatcherKnn 3 \
      --Icp/PMMatcherEpsilon 1 \
      --Icp/Epsilon 0.0001 \
      --Icp/PointToPlaneK 20 \
      --Icp/PointToPlaneRadius 0 \
      --Icp/CorrespondenceRatio 0.01"

I got this result at the first run:



and this at the second:



It is not a live capture, the topics are published from a bag.

Do I need to use a clean copy of the rtamap.launch file every time ?

[matlabbe]

With "--Icp/PM true", make sure rtabmap has libpointmatcher support. Otherwise results will be very poor with PCL's default ICP approach with 3D LiDARs.

How is the lidar oriented?

A Icp/VoxelSize of 20 cm is quite small for large-scale environments. OdomF2M/ScanSubtractRadius should be generarly at same value than Icp/VoxelSize.

If you can share the bag, that could easier to reproduce what you are seeing.

cheers,
Mathieu