Autoware.Auto with LGSVL Simulator
Table of Contents
Overview top#
This guide describes setting up and using Autoware.Auto with the LGSVL simulator. As Autoware.Auto is still under-development, full self-driving is not yet possible. This guide will focus on running individual modules which have been implemented.
Setup top#
Requirements top#
- Linux operating system
- Nvidia graphics card
Installing Docker CE top#
To install Docker CE please refer to theĀ official documentation. We also suggest following through with theĀ post installation steps to run docker as a non-root user.
Installing Nvidia Docker top#
- Before installing nvidia-docker make sure that you have an appropriate Nvidia driver installed. To test if nvidia drivers are properly installed enter
nvidia-smi
in a terminal. If the drivers are installed properly an output similar to the following should appear.
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 390.87 Driver Version: 390.87 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 108... Off | 00000000:65:00.0 On | N/A |
| 0% 59C P5 22W / 250W | 1490MiB / 11175MiB | 4% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1187 G /usr/lib/xorg/Xorg 863MiB |
| 0 3816 G /usr/bin/gnome-shell 305MiB |
| 0 4161 G ...-token=7171B24E50C2F2C595566F55F1E4D257 68MiB |
| 0 4480 G ...quest-channel-token=3330599186510203656 147MiB |
| 0 17936 G ...-token=5299D28BAAD9F3087B25687A764851BB 103MiB |
+-----------------------------------------------------------------------------+
-
Install nvidia docker.
Note: For docker 19.03 and newer nvidia GPUs are natively supported as devices in docker runtime, and nvidia-docker2 is deprecated, however, because Autoware.auto uses the
--runtime nvidia
argument nvidia-docker2 will need to be installed even for newer docker versions.
Installing Autoware.auto top#
- Follow the installation and development setup guide for Autoware.auto.
Simulator installation top#
- Download and extract the latest simulator release under the ~/adehome folder.
- (Optional) Download the latest PythonAPI release (make sure the release version matches the simulator) and install it using pip:
cd PythonAPI
pip3 install --user .
Install ROS2 LGSVL Bridge top#
Downloading
cd adehome
git clone https://github.com/lgsvl/ros2-lgsvl-bridge.git
Building
Refer to README.md in the repo.
cd ros2-lgsvl-bridge
colcon build --cmake-args '-DCMAKE_BUILD_TYPE=Release'
Running
Refer to README.md in the repo.
source ~/ros2-lgsvl-bridge/install/setup.bash
lgsvl_bridge
Install ROS2 LGSVL Messages top#
Downloading
mkdir -p ~/adehome/AutowareAuto/src/external/lgsvl_msgs
cd ~/adehome/AutowareAuto/src/external/lgsvl_msg
git clone https://github.com/lgsvl/lgsvl_msgs.git
Building
# In the ade container
cd ~/AutowareAuto
colcon build --cmake-args '-DCMAKE_BUILD_TYPE=Release'
# You may want to build only lgsvl_msgs package with the following command.
colcon build --packages-select lgsvl_msgs --cmake-args '-DCMAKE_BUILD_TYPE=Release'
Testing
cd ~/AutowareAuto
source install/setup.bash
ros2 msg list |grep lgsvl_msgs
# If you can see the list of lgsvl_msgs, they're ready to be used.
Run Simulator alongside Autoware.Auto top#
The ROS2 web bridge allows the simulator and Autoware.auto to communicate. To test this connection we can visualize sensor data from the simulator in rviz2 (running in the Autoware.auto container).
- Start the Autoware.Auto containers without NVIDIA setup:
cd ~/adehome/AutowareAuto
source .aderc
ade start
- Start the Autoware.Auto containers with NVIDIA setup:
- ~/adehome/AutowareAuto/.aderc-nvidia:
export ADE_DOCKER_RUN_ARGS="--cap-add=SYS_PTRACE --net=host --privileged --add-host ade:127.0.0.1 -e RMW_IMPLEMENTATION=rmw_cyclonedds_cpp --runtime=nvidia -ti --rm -e DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix \
-e NVIDIA_VISIBLE_DEVICES=all \
-e NVIDIA_DRIVER_CAPABILITIES=compute,utility,display"
export ADE_GITLAB=gitlab.com
export ADE_REGISTRY=registry.gitlab.com
export ADE_DISABLE_NVIDIA_DOCKER=false
export ADE_IMAGES="
registry.gitlab.com/autowarefoundation/autoware.auto/autowareauto/ade:master
registry.gitlab.com/autowarefoundation/autoware.auto/autowareauto:master
nvidia/cuda
"
cd ~/adehome/AutowareAuto
source .aderc-nvidia
ade start
- Enter the container and start rviz2:
ade enter
cd ~/AutowareAuto
colcon build
source ~/AutowareAuto/install/setup.bash
rviz2 -d /home/"${USER}"/AutowareAuto/install/autoware_auto_examples/share/autoware_auto_examples/rviz2/autoware.rviz
- Start the LGSVL Simulator in ADE container by launching the executable and click on the button to open the web UI.
$ ~/AutowareAuto/lgsvl_simulator/simulator
-
When you see Open Browser in a window, enter localhost:8080 in browser.
-
In the Vehicles tab look for
AWF Lexus2016 RX Hybrid
. If not available download it from here and follow these instructions to add it.- Click on the wrench icon for the Lexus vehicle:
- Change the bridge type to
ROS2
- Use the following JSON configuration Autoware Auto JSON Example
-
Switch to the Simulations tab and click the
Add new
button:- Enter a name and switch to the
Map & Vehicles
tab - Select a map from the drop down menu. If none are available follow this guide to get a map.
- Select the
Lexus2016RXHybrid
from the drop down menu. In the bridge connection box to the right enter the bridge address (default:localhost:9090
) - Click submit Select the simulation and press the play button in the bottom right corner of the screen
- Enter a name and switch to the
-
Launch ROS2 LGSVL bridge in a new terminal:
NOTE* ROS2 LGSVL Bridge needs to be running.
You should now be able to see the lidar point cloud in rviz (see image below).
If the pointcloud is not visible make sure the Fixed Frame (under Global Options) is set to lidar_front
and that a PointCloud2 message is added which listens on the /lidar_front/points_raw
topic.