autowarefoundation / autoware_ai_common Goto Github PK

View Code? Open in Web Editor NEW

39.0 39.0 46.0 1.1 MB

License: Apache License 2.0

CMake 2.86% C++ 96.62% Shell 0.01% Makefile 0.13% Python 0.39%

autoware_ai_common's People

Contributors

Stargazers

Watchers

autoware_ai_common's Issues

Lanelet2 Map Serialization not portable

Bug report

Required information:

Operating system and version: Ubuntu 20 and Ubuntu 18
Autoware installation type: Built lanelet2_extension from source
Autoware version or commit hash: Latest develop
ROS distribution and version: Noetic (on Ubuntu 20) and Melodic (on Ubuntu 18)
ROS installation type: Docker
Package or library, if applicable: lanelet2_extension

Description of the bug

A lanelet2 map that is serialized (see https://github.com/Autoware-AI/common/blob/master/lanelet2_extension/lib/message_conversion.cpp#L52) on the Ubuntu 20 machine cannot be deserialized (see https://github.com/Autoware-AI/common/blob/master/lanelet2_extension/lib/message_conversion.cpp#L76) on the Ubuntu 18 machine because boost::archive::binary_iarchive is not portable. Specifically:

terminate called after throwing an instance of 'boost::archive::archive_exception'
  what():  unsupported version

Would it be possible to switch the serialization to text_iarchive/text_oarchive or would that be too much on a hit on performance?

Loading vector map failed

Bug report

Required information:

Operating system and version:
- Ubuntu 18.04
Autoware installation type:
- Docker with mounted source code directory
Autoware version or commit hash
- Autoware common package commit hash is 1fa3767
ROS distribution and version:
- melodic
ROS installation type:
- Docker
Package or library, if applicable:

Description of the bug

Loading point maps succeeded.
Loading vector map failed with the following error.

['roslaunch', '/home/autoware/shared_dir/autoware-data/BorregasAve/my_launch/my_map.launch']
pid=21670
sched policy=OTHER prio=0
load /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/pointcloud_map/BorregasAve.pcd
/home/autoware/Autoware/install/map_file/lib/map_file/vector_map_loader: symbol lookup error: /home/autoware/Autoware/install/map_file/lib/map_file/vector_map_loader: undefined symbol: _ZN10vector_map9VectorMap9subscribeERN3ros10NodeHandleEy
[vector_map_loader-4] process has died [pid 21702, exit code 127, cmd /home/autoware/Autoware/install/map_file/lib/map_file/vector_map_loader /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/lane.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/dtlane.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/point.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/line.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/pole.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/signaldata.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/stopline.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/vector.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/whiteline.csv /home/autoware//shared_dir/autoware-data/BorregasAve/data/map/vector_map/node.csv __name:=vector_map_loader __log:=/home/autoware/.ros/log/3fd75d90-d746-11ea-b6c2-54bf647d3a59/vector_map_loader-4.log].
log file: /home/autoware/.ros/log/3fd75d90-d746-11ea-b6c2-54bf647d3a59/vector_map_loader-4*.log

(runtime_manager_dialog.py:21408): Gtk-WARNING **: 18:06:32.942: Negative content width -2 (allocation 32, extents 17x17) while allocating gadget (node button, owner GtkToggleButton)

Steps to reproduce the bug

1.Download vector, point map.

$ cd ~/shared_dir
$ git clone https://github.com/lgsvl/autoware-data.git

2.Run docker container

# Change directory to docker/generic
$ ./run.sh -t local # in my case, I used local tag. Instead you may use 1.14.0 tag.

3.Apply commit to make bug.

# Change directory to autoware common one.
$ git reset --hard 1fa376783ff584528510587bf80257dafd6d6fb3
# In df8f9fa commit, loading vector map succeeded, In 1fa3767 commit, it failed.

4.Start Runtime Manager

$ roslaunch runtime_manager runtime_manager.launch

5.Select map launch file.

/home/autoware/shared_dir/autoware-data/BorregasAve/my_launch/my_map.launch

6.Click Map in Quick Start tab and start Rviz

Expected behavior

Vector map should be loaded like point map is loaded.

Actual behavior

Vector map is not loaded with the error.

Screenshots

The following shows that loading vector map failed.

Additional information

Use Build Farm Packages for mrt_cmake_modules and lanelet2

The packages mrt_cmake_modules and lanelet2 have been released to the ROS build farm so there is no need to build them from source.

Fail to build due to pthread linking issue on Autoware 1.14.0

Bug report

The version 1.14.0 of autoware fails to build due a linking failure to the pthread library.

Required information:

Operating system and version: Ubuntu 18.04 x64
Autoware installation type: Source
Autoware version or commit hash: 1.14.0
ROS distribution and version: melodic
ROS installation type: deb
Package or library, if applicable: ros_observer

Description of the bug

Starting  >>> ros_observer                                                                                                                                                                                 
___________________________________________________________________________________________________________________________________________________________________________________________________________
Errors     << ros_observer:make /home/alexis/auto-ws/logs/ros_observer/build.make.005.log                                                                                                                  
/usr/bin/ld: CMakeFiles/ros_observer.dir/src/ros_observer.cpp.o: undefined reference to symbol 'pthread_mutexattr_settype@@GLIBC_2.2.5'
//lib/x86_64-linux-gnu/libpthread.so.0: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status
make[2]: *** [/home/alexis/auto-ws/devel/.private/ros_observer/lib/ros_observer/ros_observer] Error 1
make[1]: *** [CMakeFiles/ros_observer.dir/all] Error 2
make: *** [all] Error 2
cd /home/alexis/auto-ws/build/ros_observer; catkin build --get-env ros_observer | catkin env -si  /usr/bin/make --jobserver-fds=6,7 -j; cd -
...........................................................................................................................................................................................................
Failed     << ros_observer:make           [ Exited with code 2 ]                                                                                                                                           
Failed    <<< ros_observer                [ 0.6 seconds ]

Steps to reproduce the bug

git clone https://github.com/Autoware-AI/autoware.ai -b 1.14.0 --depth 1
cd autoware.ai
vcs import < autoware.ai.repos
cd ../..
rosdep update
rosdep install -y --from-paths src/autoware.ai/ --ignore-src --rosdistro melodic
catkin build ros_observer

Expected behavior

It builds successfully

Actual behavior

It doesn't build.

laserscan2costmap provides a map with wrong orientation

Ubuntu 18.04
Autoware 1.13 installed from source following the wiki page
git rev-parse HEAD : e91b3e0
ROS Melodic
package related: object_map

Description of the bug

When I try to create a costmap from my laserscan, the orientation of the /ring_ogm is not correct. Is it rotated with regards to the sensor frame. This has also been mentioned in this issue.

Steps to reproduce the bug

Launch runtime manager and enable RViZ and laserscan2costmap node from the Computing tab.
Load this bag in the Simulation tab.
In the laserscan2costmap node, set as scan topic /front_laser/scan' and as sensor frame: front_laser_base_link`
In RViZ visualize the /front_laser/scan and the ogm_ring topics.
Run the ROSbag

Expected behavior

The costmap should be rotated anticlockwise 90 degrees to match the sensor frame.

Screenshots

The result is depicted in the image below:

Implement a TVM inference pipeline

Feature request

Description

Proceeding with https://discourse.ros.org/t/unified-ml-inference-in-autoware-a-proposal/14058 plan, we need an implementation for a generic inference pipeline using TVM.

Implementation considerations

The pipeline will be generic, the preprocessing and postprocessing stages will need to be implemented by the user.

Additional information

A validation test will also be provided, running the pipeline on yolo_v2_tiny.

autowarefoundation / autoware_ai_common Goto Github PK

autoware_ai_common's People

Contributors

Stargazers

Watchers

Forkers

autoware_ai_common's Issues

Bug report

Required information:

Description of the bug

Bug report

Required information:

Description of the bug

Steps to reproduce the bug

Expected behavior

Actual behavior

Screenshots

Additional information

Bug report

Required information:

Description of the bug

Steps to reproduce the bug

Expected behavior

Actual behavior

Description of the bug

Steps to reproduce the bug

Expected behavior

Screenshots

Feature request

Description

Implementation considerations

Additional information

Recommend Projects

Recommend Topics

Recommend Org

Jobs