hku-mars / fast_lio Goto Github PK

Hi @XW-HKU
In function h_share_model(), what does 's' mean after calculating plane parameters?

Hi,

When I tested this package with Mid70 Lidar, I got the following error:

Leaf size is too small for the input dataset. Integer indices would overflow.[pcl::VoxelGrid::applyFilter] Leaf size is too small for the input dataset. Integer indices would overflow.

Then, the published odometry would drift to very far distance. Is this normal?

Thanks.

Hello, thank you again for this amazing project, I've been seeing pretty good results using it so far. One small issue I've had is in RViz when visualizing the data, I usually get 3-10 fps only and it works slowly even though the map looks great, I think it's just a bit too many points for my machine to handle. I'm running on an NVidia Jetson Xavier with an OS1-128, any suggestions about maybe somehow reducing the number of points RViz visualizes so I can improve performance a bit?

Hello, I would like to ask you how the topic was configured in the fixed frame and pointclouds in RVIZ when you made the drawing. I did not find the poingcloud data
thank you for your answer

The FAST-LIO described in your paper uses both planar/surf/flat features and edge/corner/sharp features for state estimation.

In your implementation, the feature processing part extracts and publishes both flat and sharp features src/feature_extract.cpp#L127#L128. But in state estimation part, it seems that only flat features are subscribed and used src/laserMapping.cpp#L627.

May I ask why the sharp features are not used in the implementation?

我用这个算法构建了一个非常满意的地方， 可是应该用什么方法保存下来用来定位导航呢？

Hello,

Thanks for sharing such amazing work. I am using fastlio 2 for map generation using Velodyne vlp16 and intel l515 (for imu). For some scenarios, it is generating an amazing map. But for some, there is a large drift during mapping. Can you please help me if there is any fine-tuning needed for the same?

Snapshot (Scene: Car parking basement)

HI:
Does Fast-lio-1.0 support MId-70?
if it supports, do you have turning good parameter for Mid-70? parameters is " edgea edgeb smallp_ratio point_filter_num ..."

Is there any other way to use raw livox lidar data such as /livox/lidar_xxxxxxxxxx | Type: sensor_msgs/msg/PointCloud2 directly?

I get the following error when catkin building. Does anyone know the fix?

error: invalid conversion from ‘const char*’ to ‘int’ [-fpermissive]
omp_set_num_threads(MP_PROC_NUM);

Hi, so I would like to try this package, I've been having decent results with Ouster's built-in IMU but I wanna try it using a Microstrain IMU that I have. The problem is the sensor has different coordinates for acceleration and attitude so I'm not sure how to get all the transformations needed correctly in one matrix. Is there a way to adapt the code to have 2 separate ext matrices? Mine would look something like this:
extrinsicRot: [-1, 0, 0,
0, 1, 0,
0, 0, -1]
extrinsicRPY: [0, 1, 0,
-1, 0, 0,
0, 0, 1]
There should be a way to make this possible but I'm not exactly sure where to start as I am not yet familiar with FAST-LIO

Hi, is there a way currently to save the trajectory from a scan in something like a pcd file like the map is saved?

I would like to buy an AVIA, but I tried to install your FAST_LIO package first, but once the ros_melodic - PCL OPENCV and eigen packages are installed, when I download your ghitub it gives me a privacy error I had to comment the line "### cmake_policy (SET CMP0074 NEW) "otherwise the compilation does not proceed and now I have the problem that gives me an error because you have not inserted in the [include] folder a library #include <fast_lio / States.h>, now you can correct the problem and insert all the right libraries? . Another thing seems that the vs ghitub must reside inside the src of the space where livox_ros_driver also resides otherwise if you put it in a different space at the time of catkin_make it does not find the ros_drive directory and the compilation stops. Can you explain better and be more precise in all the steps otherwise we will not be able to proceed, unfortunately we are not as expert programmers as you are? Thank you
this is the terminal error:
[ 64%] Building CXX object FAST_LIO/CMakeFiles/loam_laserMapping.dir/src/laserMapping.cpp.o
In file included from /home/livox2/ws_livox/src/FAST_LIO/src/laserMapping.cpp:47:0:
/home/livox2/ws_livox/src/FAST_LIO/include/common_lib.h:8:10: fatal error: fast_lio/States.h: File o directory non esistente
#include <fast_lio/States.h>
^~~~~~~~~~~~~~~~~~~
compilation terminated.
FAST_LIO/CMakeFiles/loam_laserMapping.dir/build.make:62: recipe for target 'FAST_LIO/CMakeFiles/loam_laserMapping.dir/src/laserMapping.cpp.o' failed
make[2]: *** [FAST_LIO/CMakeFiles/loam_laserMapping.dir/src/laserMapping.cpp.o] Error 1
CMakeFiles/Makefile2:3616: recipe for target 'FAST_LIO/CMakeFiles/loam_laserMapping.dir/all' failed
make[1]: *** [FAST_LIO/CMakeFiles/loam_laserMapping.dir/all] Error 2
Makefile:140: recipe for target 'all' failed
make: *** [all] Error 2
Invoking "make -j1 -l1" failed
livox2@livox2-VirtualBox:~/ws_livox$

Dear XW-HKU:

I use to have a low cpu consuming yaml file for data capture (recording a bag file) and a highly optimized for accuracy yaml file for postprocessing the bag recorded in field.
My question, regarding this, is:

• point_filter_num: I understand that this downsamples the pointcloud, grouping the number of points indicated in the parameter. Is it right? If I want the best result should I use value 1?
• max_iteration: this referes to the maximum iterations on ICP adjustment per scan. Is it right? So I should use around 10?
• filter_size_surf: This I think that is the voxelization size. So 0.5 means that every plane of 0.5x0.5 should be used in the adjustment. Is it right?
• filter_size_map: This value I have no idea what means.

If I want the best fitting, even playing the bag at rate 0.5 or 0.1:

• What parameters do you recomend for the best accuracy?
• Should I modify any of the ones hardcoded in the sources?

Thanks in advance

the slam system has the above error when running and the map built has wrong transform when the robot itself rotates, nothing wrong when running in straight line

尊敬的港大团队好：

先恭喜恭喜，新作品的进展让人激动。

这次尝试编译Fast Lio2遇到困难， 请多多指教 😀：

0. 安装

在说明文稿 “2.build” 里的git clone https://github.com/XW-HKU/fast_lio.git,
可能要改为：git clone https://github.com/hku-mars/FAST_LIO.git

1. 第一次编译,不论是用终端clone，还是直接下载都没有Ikd-treed文件夹内的文件，导致编译失败。可能是文件夹名存在@符号造成的？

简化的报错信息如下：

make[2]: *** 没有规则可制作目标“
~  /ikd-Tree/ikd_Tree.cpp”，

fatal error: ikd-Tree/ikd_Tree.h: 
没有那个文件或目录 #include <ikd-Tree/ikd_Tree.h>

2. 再次尝试编译, 把Ikd-treed的档案手动copy到文件夹，CMD输入catkin_make有大量的报错。怀疑是编码问题 或 cmake版本问题，多多指教啊~

报错档案如下， 共三类。

• 第一类是编码报错

简化的报错信息如下：

[ 29%] Building CXX object FAST_LIO/ ... /ikd_Tree.cpp.o

~ ikd-Tree/ikd_Tree.cpp:70:76: 

error: stray ‘\302’ in program
   <title>ikd-Tree/ikd_Tree.cpp at 3d115a41377243420a74fc15dd7cf7ef337730df 
�� hku-mars/ikd-Tree · GitHub</title>

• 第二类是语法报错

简化的报错信息如下：

~ ikd_Tree.cpp:320:12: 

warning: missing terminating ' character
       <!-- '"` --><!-- </textarea></xmp> --></option></form><form class="js-site-search-form" role="search" aria-label="Site" data-scope-type="Repository" data-scope-id="341097199" data-scoped-search-url="/hku-mars/ikd-Tree/search" data-owner-scoped-search-url="/orgs/hku-mars/search" data-unscoped-search-url="/search" action="/hku-mars/ikd-Tree/search" accept-charset="UTF-8" method="get">

• 第三类似乎是建议改写

简化的报错信息如下：

~ /laserMapping.cpp:859:82: 
error: ‘kdtree_incremental_time’ was not declared in this scope
                 
aver_time_incre = 
aver_time_incre * (frame_num - 1)/frame_num + (kdtree_incremental_time)/frame_num;

------------------------------------------
~/laserMapping.cpp:859:82: 

note: suggested alternative: ‘map_incremental’
                
 aver_time_incre = 
aver_time_incre * (frame_num - 1)/frame_num + (kdtree_incremental_time)/frame_num;
                                                                                 

We recently started using AVIA devices.
We tried to Play use Livox Viewer and set the frame time to 3000ms, the output is below.

We found some noise points, is there any way to reduce the noise points?

Any anther question is, is there any way to increase the points' density?
Thank you.

Hey, do you guys have any pointers on how to integrate L515 with FAST_LIO

I've seen

And also seen the specific handers for each lidar, but i am not able to figure out how to write a l515_handler, in specific i can't figure out the processing function to convert the PCL data from L515 into PointCloudXYZI format.

尊敬的港大团队好：

因为avia型号买不到，只能尝试用mid-40和70绑定外接imu，遇到些困难，请指点下~

• 尝试把loam_livox里，mid系列的launch_file 拿到fast_lio尝试。可以看到点云，但优化可能不明显。
  您会考虑 mid系列外接imu 的优化么？

• 还试了按过去的相关回复修改，但mid系列的点云没有出现在rviz界面里。请问还适用么？

相关回复：`Does Fast-lio-1.0 support MId-70?` 
FAST LIO可以外接IMU（修改laser_mapping.cpp里面的ros 订阅topic name），
不过需要保证旋转上没有偏差，也就是rotation的外参是单位阵。
平移外参可以直接在common_lib.h里面修改。

Hi, @XW-HKU

I compiled Fast-LIO in ubuntu16.04 and I tried to test it with Avia. But the result may be totally wrong, I'm curious about what I may have done to make this kind of strange phenomenon?

The pose may drift even thought it is still.

Looking forward to your help ;>

Hey, I noticed in FAST-LIO sometimes when I play a ROS bag with normal speed, the bag finished playing and everything is okay, but it takes FAST-LIO more time and it keeps processing the data even after the bag finishes playing, is this normal behaviour?

Hey, thank you again for the amazing work that you guys have put in! I have a couple of questions about performance and optimization of the algorithm during longer periods of scanning. I noticed that my RAM usage goes up higher and higher each second the map is being populated, and after a while I suppose it would overflow and crash the program ( I have 32GB on my machine so it would take a few minutes but still), is there any way to stop this or at least clear some of the scans after a certain time so we can free some RAM? I'm guessing this is mainly due to RViz as well using a lot to visualize the whole map with that many points (my lidar has 128 channels so I get around 2mil points per second), which is my second question as well. I'm trying to think of a way to visualize some sparser version of the map in rviz on the screen so I can save on PC resources but I would still like the end result in the pcd to be with the same quality as if I had publish_dense set to 1 with all the points, possibly one solutions that comes to my mind is having a separate topic that is more filtered and less dense just for visualization in rviz? Let me know if you have suggestions, thanks!

您好：
我在读您的论文的时候，关于论文中的公式18有点不太懂，这个公式似乎和一般的IEKF更新不太一样？请问在哪里能找到相关的资料或文献说明呢？非常感谢您的解答！

Hi, thank you for sharing this great job!
I use a velodyne VLP-32C lidar with external imu. For a outdoor data, if I set extract feature to true, the program runs well, but has small drift in z axis. Then I set extract feature to false, the trajectory is still right, but the pointcloud in rviz is abnormal. The points far from ground are lost, I mean, only the base of a building is remain, points about 2 meters high are gone.


The first image is with extract feature on, and second is off.
Is this a normal phenomenon？

the mid-70 1000 rosbag can be used in livox_mapping and livox_loam.

but in Fast_LIO, after rosbag play XXX.bag, a md5sum error will be show, (the bag will play anyway), and no pointclouds in rviz.

目前的1.0程序版本可以驱动livox-horizon，但是由于垂直视场太小，（horizon 是25°），所以在闭环检查时，问题就表现得比较严重，平面误差近20米，高程误差近2米。（大概介绍一下，我采集得情况，是个地下车库，采集不足10分钟，rosbag 11.7G），附上采用rosbag回放

Cannot run catkin_make for below reason, anybody can help?

It says
Could NOT find livox_ros_driver (missing: livox_ros_driver_DIR)

I've already installed livox_ros_direver.

Dear friend:
我直接说中文啦！我是个slam初学者，看了您们的论文“FAST-LIO: A Fast, Robust LiDAR-inertial Odometry Package by Tightly-Coupled Iterated Kalman Filter”，有个地方很迷惑,，不知道论文公式14中雅可比H怎么计算的？具体什么形式呢？

/*** calculate the Measuremnt Jacobian matrix H ***/ Eigen::Vector3d A(point_crossmat * state.rot_end.transpose() * norm_vec); Hsub.row(i) << VEC_FROM_ARRAY(A), norm_p.x, norm_p.y, norm_p.z;

Hi,

I have used this package to create a very good 3D map with Livox Mid lidar. With the saved PCD file, is it possible to use this package to do relocalization? Something like the following package:

https://github.com/Livox-SDK/livox_relocalization

Thanks for your great package.

David.

启动了mapping_avia.launch会发布里程计信息，我现在正在给机器狗做定位导航，机器狗没有编码器，所以需要利用雷达和imu得出里程计信息，请问有什么方法可以只发布里程计信息吗

Hi!

This is not an issue with the library, more of a general question. Could you suggest how I would integrate GPS for online georeferencing?

Ideally there would be a transform for the point cloud that is in UTM coordinates.

Any help would be appreciated.

Hi @XW-HKU
"Please make sure the IMU and LiDAR are Synchronized, that's important."

Does it mean to ensure that the time source between LiDAR and IMU is the same?

[ 76%] Built target livox_ros_driver_generate_messages
[ 80%] Building CXX object livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/timesync/timesync.cpp.o
[ 84%] Building CXX object livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/timesync/user_uart/user_uart.cpp.o
[ 88%] Building CXX object livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/common/comm/comm_protocol.cpp.o
[ 92%] Building CXX object livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/common/comm/sdk_protocol.cpp.o
[ 96%] Building CXX object livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/common/comm/gps_protocol.cpp.o
[100%] Linking CXX executable /home/dji/ws_livox/devel/lib/livox_ros_driver/livox_ros_driver_node
CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function livox_ros::Lddc::CreateBagFile(std::string const&)': lddc.cpp:(.text+0x41b): undefined reference to rosbag::Bag::open(std::string const&, unsigned int)'
lddc.cpp:(.text+0x4ec): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function livox_ros::Lddc::PrepareExit()':
lddc.cpp:(.text+0x677): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' lddc.cpp:(.text+0x6f0): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)'
CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function livox_ros::Lddc::GetCurrentPublisher(unsigned char)': lddc.cpp:(.text+0x13ca): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)'
lddc.cpp:(.text+0x1459): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o:lddc.cpp:(.text+0x1539): more undefined references to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' follow
CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function livox_ros::Lddc::PublishPointcloud2(livox_ros::LidarDataQueue*, unsigned int, unsigned char)': lddc.cpp:(.text+0x2016): undefined reference to ros::Publisher::getTopic() const'
lddc.cpp:(.text+0x239c): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function livox_ros::Lddc::PublishPointcloudData(livox_ros::LidarDataQueue*, unsigned int, unsigned char)':
lddc.cpp:(.text+0x270e): undefined reference to ros::Publisher::getTopic() const' lddc.cpp:(.text+0x2ad7): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)'
CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function livox_ros::Lddc::PublishCustomPointcloud(livox_ros::LidarDataQueue*, unsigned int, unsigned char)': lddc.cpp:(.text+0x2fcb): undefined reference to ros::Publisher::getTopic() const'
lddc.cpp:(.text+0x312a): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function livox_ros::Lddc::PublishImuData(livox_ros::LidarDataQueue*, unsigned int, unsigned char)':
lddc.cpp:(.text+0x3574): undefined reference to ros::Publisher::getTopic() const' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function void rosbag::Bag::writeMessageDataRecord<sensor_msgs::PointCloud2_<std::allocator > >(unsigned int, ros::Time const&, sensor_msgs::PointCloud2_<std::allocator > const&)':
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs12PointCloud2_ISaIvEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs12PointCloud2_ISaIvEEEEEvjRKN3ros4TimeERKT]+0x104): undefined reference to rosbag::Bag::toHeaderString(ros::Time const*) const' lddc.cpp:(.text._ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs12PointCloud2_ISaIvEEEEEvjRKN3ros4TimeERKT_[_ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs12PointCloud2_ISaIvEEEEEvjRKN3ros4TimeERKT_]+0x246): undefined reference to rosbag::Bag::writeHeader(std::map<std::string, std::string, std::lessstd::string, std::allocator<std::pair<std::string const, std::string> > > const&)'
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs12PointCloud2_ISaIvEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs12PointCloud2_ISaIvEEEEEvjRKN3ros4TimeERKT]+0x280): undefined reference to rosbag::Bag::appendHeaderToBuffer(rosbag::Buffer&, std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > const&)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function void rosbag::Bag::writeMessageDataRecord<pcl::PointCloudpcl::PointXYZI >(unsigned int, ros::Time const&, pcl::PointCloudpcl::PointXYZI const&)':
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN3pcl10PointCloudINS2_9PointXYZIEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN3pcl10PointCloudINS2_9PointXYZIEEEEEvjRKN3ros4TimeERKT]+0x104): undefined reference to rosbag::Bag::toHeaderString(ros::Time const*) const' lddc.cpp:(.text._ZN6rosbag3Bag22writeMessageDataRecordIN3pcl10PointCloudINS2_9PointXYZIEEEEEvjRKN3ros4TimeERKT_[_ZN6rosbag3Bag22writeMessageDataRecordIN3pcl10PointCloudINS2_9PointXYZIEEEEEvjRKN3ros4TimeERKT_]+0x2a1): undefined reference to rosbag::Bag::writeHeader(std::map<std::string, std::string, std::lessstd::string, std::allocator<std::pair<std::string const, std::string> > > const&)'
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN3pcl10PointCloudINS2_9PointXYZIEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN3pcl10PointCloudINS2_9PointXYZIEEEEEvjRKN3ros4TimeERKT]+0x2d9): undefined reference to rosbag::Bag::appendHeaderToBuffer(rosbag::Buffer&, std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > const&)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function void rosbag::Bag::writeMessageDataRecord<livox_ros_driver::CustomMsg_<std::allocator > >(unsigned int, ros::Time const&, livox_ros_driver::CustomMsg_<std::allocator > const&)':
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN16livox_ros_driver10CustomMsg_ISaIvEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN16livox_ros_driver10CustomMsg_ISaIvEEEEEvjRKN3ros4TimeERKT]+0x104): undefined reference to rosbag::Bag::toHeaderString(ros::Time const*) const' lddc.cpp:(.text._ZN6rosbag3Bag22writeMessageDataRecordIN16livox_ros_driver10CustomMsg_ISaIvEEEEEvjRKN3ros4TimeERKT_[_ZN6rosbag3Bag22writeMessageDataRecordIN16livox_ros_driver10CustomMsg_ISaIvEEEEEvjRKN3ros4TimeERKT_]+0x243): undefined reference to rosbag::Bag::writeHeader(std::map<std::string, std::string, std::lessstd::string, std::allocator<std::pair<std::string const, std::string> > > const&)'
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN16livox_ros_driver10CustomMsg_ISaIvEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN16livox_ros_driver10CustomMsg_ISaIvEEEEEvjRKN3ros4TimeERKT]+0x27d): undefined reference to rosbag::Bag::appendHeaderToBuffer(rosbag::Buffer&, std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > const&)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/lddc.cpp.o: In function void rosbag::Bag::writeMessageDataRecord<sensor_msgs::Imu_<std::allocator > >(unsigned int, ros::Time const&, sensor_msgs::Imu_<std::allocator > const&)':
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs4Imu_ISaIvEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs4Imu_ISaIvEEEEEvjRKN3ros4TimeERKT]+0x104): undefined reference to rosbag::Bag::toHeaderString(ros::Time const*) const' lddc.cpp:(.text._ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs4Imu_ISaIvEEEEEvjRKN3ros4TimeERKT_[_ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs4Imu_ISaIvEEEEEvjRKN3ros4TimeERKT_]+0x21b): undefined reference to rosbag::Bag::writeHeader(std::map<std::string, std::string, std::lessstd::string, std::allocator<std::pair<std::string const, std::string> > > const&)'
lddc.cpp:(.text.ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs4Imu_ISaIvEEEEEvjRKN3ros4TimeERKT[ZN6rosbag3Bag22writeMessageDataRecordIN11sensor_msgs4Imu_ISaIvEEEEEvjRKN3ros4TimeERKT]+0x253): undefined reference to rosbag::Bag::appendHeaderToBuffer(rosbag::Buffer&, std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > const&)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/livox_ros_driver.cpp.o: In function main':
livox_ros_driver.cpp:(.text.startup+0x40): undefined reference to ros::console::set_logger_level(std::string const&, ros::console::levels::Level)' livox_ros_driver.cpp:(.text.startup+0x8c): undefined reference to ros::init(int&, char**, std::string const&, unsigned int)'
livox_ros_driver.cpp:(.text.startup+0xf8): undefined reference to ros::NodeHandle::NodeHandle(std::string const&, std::map<std::string, std::string, std::less<std::string>, std::allocator<std::pair<std::string const, std::string> > > const&)' livox_ros_driver.cpp:(.text.startup+0x208): undefined reference to ros::NodeHandle::getParam(std::string const&, int&) const'
livox_ros_driver.cpp:(.text.startup+0x246): undefined reference to ros::NodeHandle::getParam(std::string const&, int&) const' livox_ros_driver.cpp:(.text.startup+0x287): undefined reference to ros::NodeHandle::getParam(std::string const&, int&) const'
livox_ros_driver.cpp:(.text.startup+0x2c8): undefined reference to ros::NodeHandle::getParam(std::string const&, double&) const' livox_ros_driver.cpp:(.text.startup+0x309): undefined reference to ros::NodeHandle::getParam(std::string const&, int&) const'
livox_ros_driver.cpp:(.text.startup+0x345): undefined reference to ros::NodeHandle::getParam(std::string const&, std::string&) const' livox_ros_driver.cpp:(.text.startup+0x383): undefined reference to ros::NodeHandle::getParam(std::string const&, bool&) const'
livox_ros_driver.cpp:(.text.startup+0x3c1): undefined reference to ros::NodeHandle::getParam(std::string const&, bool&) const' livox_ros_driver.cpp:(.text.startup+0x4c9): undefined reference to ros::NodeHandle::getParam(std::string const&, std::string&) const'
livox_ros_driver.cpp:(.text.startup+0x558): undefined reference to ros::NodeHandle::getParam(std::string const&, std::string&) const' livox_ros_driver.cpp:(.text.startup+0x73e): undefined reference to ros::NodeHandle::getParam(std::string const&, std::string&) const'
livox_ros_driver.cpp:(.text.startup+0x929): undefined reference to ros::NodeHandle::getParam(std::string const&, std::string&) const' livox_ros_driver.cpp:(.text.startup+0x9b8): undefined reference to ros::NodeHandle::getParam(std::string const&, std::string&) const'
livox_ros_driver.cpp:(.text.startup+0xdb5): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' livox_ros_driver.cpp:(.text.startup+0xe9c): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)'
livox_ros_driver.cpp:(.text.startup+0xf9c): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' livox_ros_driver.cpp:(.text.startup+0x10d5): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)'
livox_ros_driver.cpp:(.text.startup+0x1178): undefined reference to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' CMakeFiles/livox_ros_driver_node.dir/livox_ros_driver/livox_ros_driver.cpp.o:livox_ros_driver.cpp:(.text.startup+0x126b): more undefined references to ros::console::initializeLogLocation(ros::console::LogLocation*, std::string const&, ros::console::levels::Level)' follow
collect2: error: ld returned 1 exit status
livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/build.make:1074: recipe for target '/home/dji/ws_livox/devel/lib/livox_ros_driver/livox_ros_driver_node' failed
make[2]: *** [/home/dji/ws_livox/devel/lib/livox_ros_driver/livox_ros_driver_node] Error 1
CMakeFiles/Makefile2:1012: recipe for target 'livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/all' failed
make[1]: *** [livox_ros_driver/CMakeFiles/livox_ros_driver_node.dir/all] Error 2
Makefile:140: recipe for target 'all' failed
make: *** [all] Error 2
Invoking "make -j8 -l8" failed

https://gitee.com/Rormen/FAST_LIO
下载已成功

First of all, thanks for the great work & research you've done on SLAM, this is definitely the best implementation so far what comes to mapping precision and performance!

I have a GNSS receiver which outputs centimeter accurate position information at 8 Hz synced with lidar & imu clock, and I'd like to try to improve the long-run odometry drift by fusing GNSS location.

What would be the best approach to do that? Is it possible to enter external position estimate from GNSS to Kalman filter?

Optimally, I'd want to fuse GNSS location when it's available (outdoors, with proper GNSS fix) up to 8 Hz. But when it's indoors or with poor GNSS reception, I'd use the existing prediction system.

Also, what would be the best way to alter the initial rotation (heading/yaw) so that it would match the north-referenced XYZ coordinates from GNSS?

Any help would be highly appreciated!

作者你好，能否增加demo数据的百度网盘下载链接，google的资源内地打开比较困难，谢谢。还有r2live的demo数据。

Hi,

Can you please suggest an ideal IMU frequency for this package? Plus, can you please comment on achieving better time synchronization between the IMU and Livox (Mid-70) Custom Messages ?

Thanks.

Hi:

I'm using FAST-LIO with a Livox Horizon and everything is working well, but when I suscribe to /cloud_registered and save pcd files the intensity is always 0. Even in rviz I can't see the intensity.
How could I enable it?

Thanks in advance

Hi,Dr Xu
Thank you for the amazing work，I have a problem about the program cpu.
my computer is Ubuntu16.04 : intel [email protected]*6,GeForce GTX 1660Ti/PCle/SSE2
I have a vlp16, and test the data I recorded use the mapping_velodyne.launch with the default parameters,
I find that, the data processing frequency is lower than 10Hz[Not real time] in the first few minutes, and the CPU usage is also high.
Is this normal?

thanks.

Thank you for that awesome project!

I am experiencing quite long build times for building the project myself within my ros working directory. It takes about 2 minutes at one step. This, of course, would be totally fine if it would just happen ones. However, this happens on any changes that I do to the code with a following build. Like this it is really cumbersome to further code on your project.

Do you also experience this long build times when doing any changes? How do you go about that? Is there any possibility to speed things up? I could not really find out which part is causing this long delay since catkin_make VERBOSE=1 does not give much more info on that build step.

The actual build step that needs to happen on any code manipulation and that takes rougly 2 minutes:

[ 97%] Built target loam_feat_extract
/* at here it takes about 2 minutes until the [100%] apperas */
[100%] Linking CXX executable /home/me/catkin_ws/devel/lib/fast_lio/loam_laserMapping

dji@dji-B85MG:/catkin_ws/src/FAST_LIO/Log$ python plot.py
plot.py:37: UserWarning: loadtxt: Empty input file: "imu.txt"
imu=np.loadtxt('imu.txt')
Traceback (most recent call last):
File "plot.py", line 38, in
time=imu[:,0]
IndexError: too many indices for array
dji@dji-B85MG:/catkin_ws/src/FAST_LIO/Log$ python3 plot.py
Traceback (most recent call last):
File "plot.py", line 25, in
axs[j%4, j/4].plot(time, a_pre[:,i+j*3],'.-', label=lab_pre[i])
IndexError: only integers, slices (:), ellipsis (...), numpy.newaxis (None) and integer or boolean arrays are valid indices
dji@dji-B85MG:~/catkin_ws/src/FAST_LIO/Log$ code ..

Thanks for your great job! I test fast lio2 on KITTI dataset and get nice results.
But I have a question about discrete model when I read fast lio paper.

According to formula (2) (3), we can get P_{t+1} = P_{t} + V_{t} * δt , why not P_{t+1} = P_{t} + V_{t} * δt + 1/2 * a_{t} * δt^{2}?

Hello everyone, thanks for taking the time to answer.

Is it possible to run without IMU because my IMU is unreliable? Also, can somebody recommend any budget IMU ? Thanks in advance. My lidar is a VLP-16.

Would be great to have the covariance approximated in the odometry.

您的工作真的太棒了
我在您的代码的基础上做了一点儿微小的工作，增加了GNSS校准功能，在大场景之下提高了鲁棒性，在进行评估之后会找您进行代码合并。
请问您有这个小型无人机的设计图纸和零件的购买链接么，我也想在进行UAV SLAM验证（ps.horizon特别适合放在无人机上，所以请问第二版什么时候上线哈哈哈

这个是我的一些效果：

Thanks for your excellent work! I notice that in the feature_extract.cpp, there is a lidar type selection among vlp16, horizon... So for the current version, can we run the vlp-16 + imu dataset with adjusting some parameters?

Dear Cris:

When I use point_filter_num=1 and I store the PCD I get a point in the position of the sensor on all of the scans.
Is there a reason for that?
I mean, if I record the raw lidar from the sensor there is not a point in the coords 0,0,0, of the scan.

Thanks

我看了一眼imu数据，静止条件下加速度数据差不多是1，这个数据是有缩放吗？而且也没有看到imu相对于lidar的外参