chengwei0427 / lidar_imu_localization

Lidar-IMU Localization System with Prior Map Constraint and Lio Constraint

Hi, 👋 I'm Chengwei Zhao(赵成伟)

About me:

• 🏫 An autonomous driving practitioners in HangZhou.
• ❤️ I’m currently working on 3d LiDAR SLAM.
• 🌱 I’m looking to collaborate on laser mapping and localization.
• 💬 Ask me about robot state estimation involved in slam.
• 📫 How to reach me: [email protected].

Robotics | Sensor Fusion | 2d/3d SLAM

lidar_imu_localization's Issues

void Eigen::JacobiSVD<MatrixType, QRPreconditioner>::allocate(Eigen::Index, Eigen::Index, unsigned int) [with _MatrixType = Eigen::Matrix<double, 3, 3>; int QRPreconditioner = 2; Eigen::Index = long int]: Assertion `(!(m_computeThinU || m_computeThinV) || (MatrixType::ColsAtCompileTime==Dynamic)) && "JacobiSVD: thin U and V are only available when your matrix has a dynamic number of columns."' failed.

使用SVD求解的时候会有问题：

手动给初始位姿后报错，程序没有反应

load map successful...
ROOT_DIR: /home/qt/catkin_ws/src/Lidar_IMU_Localization/
publish corner map,size: 146129
Get initial pose: 0.675896 -0.407225 0 0 -0 0
-----extract surround keyframes ------
extractSurroundKeyFrames,takes: 74.9178ms
free(): double free detected in tcache 2
[LIO_Localization_maplocalization-5] process has died [pid 44200, exit code -6, cmd /home/qt/catkin_ws/devel/lib/LIO_Localization/LIO_Localization_maplocalization __name:=LIO_Localization_maplocalization __log:=/home/qt/.ros/log/44814162-ff61-11ed-89a2-4b6b0e2f477c/LIO_Localization_maplocalization-5.log].
log file: /home/qt/.ros/log/44814162-ff61-11ed-89a2-4b6b0e2f477c/LIO_Localization_maplocalization-5*.log
^C[LIO_Localization_rviz-6] killing on exit

编译失败

作者你好，我没有在Lidar_IMU_Localization中找到cloud_info.h

WINDOWSIZE，SLIDEWINDOWSIZE意义

如果我不用local map,初始化WINDOWSIZE = 1；
//if IMU success initialized
WINDOWSIZE = SLIDEWINDOWSIZE; // SLIDEWINDOWSIZE =2
问题：
1）这里注释表示如果IMu初始化成功，全局变量又变成了2，怎里理解这个SLIDEWINDOWSIZE，为什么要存两帧激光雷达，是前后帧吗？和WINDOWSIZE有什么关系和联系？

2）松耦合（IMU_Mode<2）的时候这个LidarIMUInited没有地方赋值true;

你好，请问定位精度有多少呀

作者你好，请问定位精度有几cm呢，还有就是doc中的lio_xyz中的图例TUM是什么意思呢，三次测试的数据吗

定位方法以及建图端pcd保存

你好，有两个疑问：
1.定位部分是分成基于地图的定位和基于LIO的两种模式吗，这两种模式是分开的吗？
2.定位加载的那几个pcd文件具体格式是啥，比如将姿态的三个角度信息放在normalxyz里，关键帧的索引放在intensity？

run failed

ROOT_DIR: /home/zzy/fast_lio2/src/Lidar_IMU_Localization-main/
file dir: /home/zzy/LOAM_58
Failed to find match for field 'normal_x'.
Failed to find match for field 'normal_y'.
Failed to find match for field 'normal_z'.
Failed to find match for field 'curvature'.
Failed to find match for field 'normal_x'.
Failed to find match for field 'normal_y'.
Failed to find match for field 'normal_z'.
Failed to find match for field 'curvature'.
Failed to find match for field 'normal_x'.
Failed to find match for field 'normal_y'.
Failed to find match for field 'normal_z'.
Failed to find match for field 'curvature'.
Failed to find match for field 'normal_x'.
Failed to find match for field 'normal_y'.
Failed to find match for field 'normal_z'.
Failed to find match for field 'curvature'.
99646, 35754

程序的bug

程序 bool extractSurroundKeyFrames(const PointType &p) 少了return true；运行报错；

Device support

Does this code support Mid360?And can I use the map output of Fast-LIO for localizate?

src/lio文件夹的作用是？

${PROJECT_NAME}_poseEstimate 这个节点有用吗？

关于特征点是否有效

if (std::fabs(vLineFeatures[f][cntFtu].error) > 1e-5) { problem.AddResidualBlock(e, loss_function, para_PR[f]); vLineFeatures[f][cntFtu].valid = true; } else { vLineFeatures[f][cntFtu].valid = false; }
作者你好，这段代码为什么误差大于1e-5的特征才有效，而不是小于

nonfeature error on neotic

when i start the mapping launch on noetic, i got an error message on poseEstimate:
Estimator::FeatureNon::ComputeError(Eigen::Matrix<double, 4, 4, 0, 4, 4> const&)()

I don't know why it would happened when I debug the code, and couldn't find the resean.

please help me, thanks!