IMU intergration fail
cshang412 opened this issue · 6 comments
Everything going smooth with lidar alone mapping , but when set use_imu: true,and publish IMU topic with 200HZ, Lidar 10HZ, i got error like below:
ros2 launch lidarslam lidarslam_tukuba.launch.py
[INFO] [launch]: All log files can be found below /home/david/.ros/log/2022-04-27-22-21-37-612097-david-pc-1003649
[INFO] [launch]: Default logging verbosity is set to INFO
[INFO] [scanmatcher_node-1]: process started with pid [1003651]
[INFO] [static_transform_publisher-2]: process started with pid [1003653]
[INFO] [graph_based_slam_node-3]: process started with pid [1003655]
[static_transform_publisher-2] [INFO] [1651069297.746841314] [static_transform_publisher_Bp2zNgpNGCEbeiRh]: Spinning until killed publishing transform from 'base_link' to 'laser_link'
[scanmatcher_node-1] registration_method:NDT
[scanmatcher_node-1] ndt_resolution[m]:2
[scanmatcher_node-1] ndt_num_threads:1
[scanmatcher_node-1] gicp_corr_dist_threshold[m]:5
[scanmatcher_node-1] trans_for_mapupdate[m]:2
[scanmatcher_node-1] vg_size_for_input[m]:1
[scanmatcher_node-1] vg_size_for_map[m]:0.1
[scanmatcher_node-1] use_min_max_filter:true
[scanmatcher_node-1] scan_min_range[m]:1
[scanmatcher_node-1] scan_max_range[m]:100
[scanmatcher_node-1] set_initial_pose:true
[scanmatcher_node-1] use_odom:false
[scanmatcher_node-1] use_imu:true
[scanmatcher_node-1] scan_period[sec]:0.1
[scanmatcher_node-1] debug_flag:false
[scanmatcher_node-1] map_publish_period[sec]:100
[scanmatcher_node-1] num_targeted_cloud:30
[scanmatcher_node-1] ------------------
[scanmatcher_node-1] [INFO] [1651069297.773546138] [scan_matcher]: initialize Publishers and Subscribers
[scanmatcher_node-1] [INFO] [1651069297.781154965] [scan_matcher]: initialization end
[graph_based_slam_node-3] [INFO] [1651069297.788144665] [graph_based_slam]: initialization start
[graph_based_slam_node-3] registration_method:NDT
[graph_based_slam_node-3] voxel_leaf_size[m]:0.2
[graph_based_slam_node-3] ndt_resolution[m]:5
[graph_based_slam_node-3] ndt_num_threads:3
[graph_based_slam_node-3] loop_detection_period[Hz]:5000
[graph_based_slam_node-3] threshold_loop_closure_score:2.5
[graph_based_slam_node-3] distance_loop_closure[m]:100
[graph_based_slam_node-3] range_of_searching_loop_closure[m]:20
[graph_based_slam_node-3] search_submap_num:4
[graph_based_slam_node-3] num_adjacent_pose_cnstraints:5
[graph_based_slam_node-3] use_save_map_in_loop:true
[graph_based_slam_node-3] ------------------
[graph_based_slam_node-3] [INFO] [1651069297.788516988] [graph_based_slam]: initialize Publishers and Subscribers
[graph_based_slam_node-3] [INFO] [1651069297.791048893] [graph_based_slam]: initialization end
[scanmatcher_node-1] [INFO] [1651069301.453226030] [scan_matcher]: create a first map
[scanmatcher_node-1] [pcl::NormalDistributionsTransform::setInputTarget] Invalid or empty point cloud dataset given!
[scanmatcher_node-1] [pcl::VoxelGridCovariance::applyFilter] No input dataset given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
[scanmatcher_node-1] [pcl::registration::NormalDistributionsTransform::compute] No input target dataset was given!
Is there any tricks about intergrate IMU to this project? seems IMU is highly relavant with time, should i sync my IMU's time and Lidar's time strictly, Thanks! looking forward for your replying.
Is the IMU 6-axis?
This package currently supports only 9-axis IMUs
If not, the terminal output alone does not tell us what the problem is.
Can you provide rosbag?
Do you know the tf between imu and lidar? Also, for the imu compound is referring to lego_loam, that could be a reference.
https://github.com/RobustFieldAutonomyLab/LeGO-LOAM
Do you know the tf between imu and lidar? Also, for the imu compound is referring to lego_loam, that could be a reference. https://github.com/RobustFieldAutonomyLab/LeGO-LOAM
Sorry, i forgot about that,
Here is TF(imu is 0.94m lower than lidar)
extrinsicTrans: [ 0.0204919, 0.00745582, -0.94 ]
extrinsicRot: [1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
0.0, 0.0, 1.0]
imuAccNoise: 6.1411237845862649e-03
imuGyrNoise: 4.1727751978347269e-04
imuAccBiasN: 2.4105859025256234e-04
imuGyrBiasN: 1.4751283004207574e-05
BTW, i didn't test imu compound of Lego-LOAM, i tested imu preintergration module of LIO-SAM same as li_slam_ros2,
the difference is i used this IMU preintergration part for NDT Localization init guess in localization module, works pretty well.
Fixing the bug about imu in lidarslam_ros2 is going to take a bit of time.
I have confirmed that li_slam_ros2 works with your rosbag, so please use that one.
https://github.com/rsasaki0109/li_slam_ros2
I am closing this issue because I did not feel the need to support imu in this package. if you want to use imu, use li_slam_ros2.