This is the first entry of our “96boards: Autoware everywhere” blog series. With this blog series we will show how to use 96boards hardware to run different components of the Autoware software stack for autonomous vehicles.
Autoware.AI is the world’s first “All-in-One” open-source software for autonomous driving technology. It is based on ROS 1 and available under Apache 2.0 license. It contains the following modules:
- Localization is achieved by 3D maps and SLAM algorithms in combination with GNSS and IMU sensors.
- Detection uses cameras and LiDARs with sensor fusion algorithms and deep neural networks.
- Prediction and Planning are based on probabilistic robotics and rule-based systems, partly using deep neural networks as well.
- The output of Autoware to the vehicle is a twist of velocity and angular velocity. This is a part of Control, though the major part of Control stack commonly reside in the by-wire controller of the vehicle.
Autoware.Auto is a project supported by the Autoware Foundation. It is a clean slate rewrite of Autoware. Autoware.Auto applies best-in-class software engineering practices which include pull request reviews, pull request builds, comprehensive documentation, 100% code coverage, a coding style guide, and a defined development and release process, all managed by an open source community manager.
It also has crisply defined interfaces for different modules including messages and APIs and a software architecture designed for determinism such that will be possible to reproduce behaviors live and on development machines.
To kick things off we will first start outlining how to get Autoware.AI’s rosbag demo running on the headless Hikey970 and use a different machine for visualization.
The post is organized as follows:
For the following steps you will need your Hikey970 to be accesible via SSH and, for an easier setup, have Docker installed on it. So, to get the board ready you can follow the steps here if you plan to also develop real time applications using ROS/ROS2 or just use the Bionic Builder Tool to get Ubuntu 18.04 working.
In addition, to install Docker you just need to follow the official steps to install docker on arm64 target for Ubuntu. Though it is also possible to install ROS Melodic natively on the board running Ubuntu 18.04.
If you have opted for installing ROS natively on the board you can follow the Autoware.AI source build instructions to get it installed natively.
If you have decided to go down the Docker route there are pre-compiled Autoware images in the autoware/arm64v8 Dockerhub repository. To pull and run one we just need to SSH to the board and follow the next steps:
- Create a
shared_dirfolder that will be mounted as a volume inside the docker container:
$ mkdir ~/shared_dir
we will use this folder to store the rosbag demo data.
- Clone the Autoware.AI’s
dockerrepository to make use of the convenience scripts:
$ git clone https://gitlab.com/autowarefoundation/autoware.ai/docker.git $ cd docker/generic $ ./run.sh -c off -i autoware/arm64v8 -t 1.13.0
Which will pull and run
1.13.0-melodicAutoware image. Once inside the container if we do
lswe will see:
$ ls Autoware shared_dir
Running the rosbag demo
NOTE: Due to memory constraints on the Hikey970 we will not be able to run all the rosbag demo components in the board. So we will run
my_localization.launch in the Hikey970 and will use a separate laptop for visualization and to run
my_motion_planning.launch. This means you will also need to get Autoware.AI’s docker repo in the laptop as explained above.
If you have chosen to just install Ubuntu 18.04 using the Bionic Buil Tool and use a monitor with the board you can follow the default rosbag demo steps since you have access to the GUI. For the headless approach please keep reading.
To be able to run the demo we need to download the demo data and extract it inside the
shared_dir folder. We will then have the following:
$ ls shared_dir data sample_moriyama_150324.bag
To run headless we will also need to setup some Autoware parameters. We can
wget the setup file into the
$ cd ~/shared_dir $ wget http://people.linaro.org/~servando.german.serrano/autoware/headless_setup.yaml
To run the each component of the demo independently we will need 4 terminals where we have ssh’ed into the Hikey970 and 4 more terminals in the separate laptop.
In Terminal 1 we run our container and, within it, make an
.autoware folder and copy the
data folder over:
$ ./run.sh -c off -i autoware/arm64v8 -t 1.13.0 $ mkdir ~/.autoware $ cp -r ~/shared_dir/data ~/.autoware/data
We also need to set some environment variables following the ROS in Multiple Machines tutorial to be able to run the visualization in a laptop. To avoid repeating the same setup steps in each of the terminals we can do:
$ echo "export ROS_MASTER_URI=http://BOARD_IP_ADDRESS:11311" >> ~/.bashrc $ source ~/.bashrc
inside the container in Terminal 1, where
BOARD_IP_ADDRESS is the current IP address of your Hikey970.
In one of the other terminals we check the name of the created container, which will be something like:
$ docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES 3dc4cbcd8519 autoware/arm64v8:1.13.0-melodic "/tmp/entrypoint.sh" 39 seconds ago Up 39 seconds goofy_lamarr
We can now access the same container on each of the terminals so that we do not need set the same environment variables on each individual container. On terminals 2 to 4 we do:
$ docker exec -it -u autoware goofy_lamarr /bin/bash $ cd ~ $ source Autoware/install/setup.bash
- Hikey970 -> Terminal 1
$ roscore & $ rosparam load ~/shared_dir/headless_setup.yaml & $ rosbag play --pause ~/shared_dir/sample_moriyama_150324.bag --clock
- Hikey970 -> Terminal 2
$ roslaunch autoware_quickstart_examples my_map.launch
- Hikey970 -> Terminal 3
$ roslaunch autoware_quickstart_examples my_sensing.launch
- Hikey970 -> Terminal 4
$ roslaunch autoware_quickstart_examples my_localization.launch
We launch the Autoware.AI’s container in terminal 1
$ cd ~/docker/generic $ ./run.sh -c off -t 1.13.0 $ echo "export ROS_MASTER_URI=http://BOARD_IP_ADDRESS:11311" >> ~/.bashrc
After the container is launched, we need to check its name with
docker ps and
docker exec on it followed by sourcing the
Autoware/install/setup.bash file as we did with the Hikey970 for terminals 2 to 4.
- Laptop -> Terminal 1
$ rosrun rviz rviz
and load the default config as per step 5 in here.
- Laptop -> Terminal 2
$ roslaunch autoware_quickstart_examples my_detection.launch
- Laptop -> Terminal 3
$ roslaunch autoware_quickstart_examples my_mission_planning.launch
- Laptop -> Terminal 4
$ roslaunch autoware_quickstart_examples my_motion_planning.launch
Now, it is certainly possible that if you have established the laptop-Hikey970 connection via Wifi attempting to use the default
config file as per step 5 in the rosbag demo steps will show errors and the full demo visualization will not be achieved.
To check that the localization is running appropriately in the board instead of running
rviz in the laptop we can do:
$ roslaunch lidar_localizer ndt_matching_monitor.launch & $ rostopic echo /ndt_monitor/ndt_status
Once the ndt matching algorithm is able to localize based on the lidar readings and the pointcloud map you will see
data: "NDT_OK" in the terminal.
In this blog post we have explored how to run the sensing and localization components of Autoware.AI on the Hikey970. This shows that it is possible to use affordable hardware for development purposes of certain components of Autoware and sets the basis for how to run it headless as opposed to using the GUI. In addition we can also try with another board for the other components that we currently run in the laptop, which is something we will show in other blog post in the near future.
Next time, we will give a go at Autoware.Auto on the Hikey970 and we will follow the Autoware.Auto 3D Perception Stack demo, so keep an eye to this space.
This article is Part 1 in a 15-Part Series.
- Part 1 - 96boards: Autoware everywhere | Autoware.AI and Hikey970
- Part 2 - 96boards: Autoware everywhere | Autoware.Auto and Hikey970
- Part 3 - 96boards: Autoware everywhere | Bridging .AI and .Auto in the Hikey970
- Part 4 - 96boards: Autoware everywhere | Autoware.AI and Dragonboard-845c
- Part 5 - 96boards: Autoware everywhere | Autoware.Auto, bridge with .AI and Dragonboard-845c
- Part 6 - 96boards: Autoware everywhere | Binding Autoware.AI nodes to CPUs
- Part 7 - 96boards: Autoware everywhere | Defaulting to Cyclone DDS
- Part 8 - 96boards: Autoware everywhere | First look at AutoCore's PCU
- Part 9 - 96boards: Autoware everywhere | meta-arm-autonomy in AutoCore's PCU
- Part 10 - 96boards: Autoware everywhere | Running Cyclone DDS on Kubernetes
- Part 11 - 96boards: Autoware everywhere | Xenomai on PCU
- Part 12 - 96boards: Autoware everywhere | K8s-based Autoware deployment on PCU
- Part 13 - 96boards: Autoware everywhere | Autoware.Auto 3D Perception Stack using k8s on PCU
- Part 14 - 96boards: Autoware everywhere | Multi-board Autoware.Auto 3D Perception Stack using k8s
- Part 15 - 96boards: Autoware everywhere | Updating Autoware.Auto 3D Perception Stack modules