96boards: Autoware everywhere | Autoware.Auto, bridge with .AI and Dragonboard-845c

Servando German Serrano


In our previous entry we showed how to get Autoware.AI running on the Qualcomm® Robotics (RB3) Dragonboard-845c Development Platform. In this post we will look at Autoware.Auto and how to bridge Autoware.AI and Autoware.Auto in the same way as we did for the Hikey970.

The post is organized as follows:


The steps outlined in this blog posts build on our previous posts and as such you need to:

In addition, if you plan on developing real-time applications in the future your board should be running a RT-enabled kernel as we outlined here.

For visualization purposes we will also use a separate laptop, where we need to have ROS2 available.

Getting the Docker images

We need to get the following Docker images:

$ docker pull 96boards/autoware:auto_20200121
$ docker pull 96boards/ros:ros1_bridge

Getting the demo data

In addition to the Docker images we need to get the demo data for the Autoware.Auto 3D Perception Stack demo. To do so:

  • Move into the shared_dir folder that we created as part of the previous post and download the demo data and parameters file:
$ cd ~/shared_dir
$ wget http://people.linaro.org/~servando.german.serrano/autoware/autoware.auto_get_demo_data
$ chmod +x autoware.auto_get_demo_data
$ ./autoware.auto_get_demo_data
$ wget http://people.linaro.org/~servando.german.serrano/autoware/vlp16_test.param_ai.yaml

Running the demos

We are now set to start with the different demos. Let’s find out how far we can push the RB3 board.

.Auto 3D Perception demo

To run the Autoware.Auto 3D Perception Stack demo we will roughly follow the steps here but adapted to our setup.

In the laptop we get the config files and open 2 terminals to run 2 instances of rviz2 as:

$ wget https://gitlab.com/autowarefoundation/autoware.auto/AutowareAuto/raw/master/src/tools/autoware_auto_examples/rviz2/autoware.rviz
$ wget https://gitlab.com/autowarefoundation/autoware.auto/AutowareAuto/raw/master/src/tools/autoware_auto_examples/rviz2/autoware_voxel.rviz
  • Laptop: Terminal 1:
$ rviz2 -d autoware.rviz
  • Laptop: Terminal 2:
$ rviz2 -d autoware_voxel.rviz

Now we need to ssh into the RB3 in 4 terminals and do the following:

  • In Terminal 1:
$ docker run -it --rm --privileged --net=host -u linaro -v ~/shared_dir:/home/linaro/shared_dir:rw 96boards/autoware:auto_20200121 /bin/bash
$ cd ~
$ source AutowareAuto/install/setup.bash
  • In terminals 2 to 4 we need to access the running container as we did here but using the linaro user:
$ docker exec -it -u linaro CONTAINER_NAME /bin/bash
$ cd ~
$ source AutowareAuto/install/setup.bash

And then:

  • RB3: Terminal 1:
$ udpreplay ~/shared_dir/route_small_loop_rw-
  • RB3: Terminal 2:
$ ros2 run velodyne_node velodyne_cloud_node_exe __params:=/home/"${USER}"/AutowareAuto/src/drivers/velodyne_node/param/vlp16_test.param.yaml
  • RB3: Terminal 3:
$ ros2 run ray_ground_classifier_nodes ray_ground_classifier_cloud_node_exe __params:=/home/"${USER}"/AutowareAuto/src/perception/filters/ray_ground_classifier_nodes/param/vlp16_lexus.param.yaml
  • RB3: Terminal 4:
$ ros2 run voxel_grid_nodes voxel_grid_cloud_node_exe __params:=/home/"${USER}"/AutowareAuto/src/perception/filters/voxel_grid_nodes/param/vlp16_lexus_centroid.param.yaml

If everything went fine we will be able to visualize the demo point cloud and downsampled one in the running rviz2 GUIs as can be seen in the image below.

.AI and .Auto bridge

We will now replicate the steps that were outlined here for the Hikey970. Hence, for a full description of the steps please follow that blog post.

As we did for the Hikey970, we will use 5 terminals in the RB3, terminals 1 and 2 for the .AI container, terminal 3 for the ros1_bridge container and 4 and 5 for the .Auto container.

.AI terminals

  • Terminal 1
$ cd docker/generic
$ ./run.sh -c off -i autoware/arm64v8 -t 1.13.0
$ mkdir ~/.autoware
$ cp -r ~/shared_dir/data ~/.autoware/data
$ roscore &
$ rosparam load ~/shared_dir/headless_setup.yaml &
$ roslaunch lidar_localizer ndt_mapping.launch
  • Terminal 2
$ docker exec -it -u autoware CONTAINER_NAME /bin/bash
$ cd ~
$ source Autoware/install/setup.bash

ros1_bridge terminal

  • Terminal 3
$ docker run -it --rm --privileged --net=host -u linaro 96boards/ros:ros1_bridge /bin/bash -c "source /opt/ros/melodic/setup.bash && source /opt/ros/dashing/local_setup.bash && ros2 run ros1_bridge dynamic_bridge --bridge-all-topics"

.Auto terminals

  • Terminal 4
$ docker run -it --rm --privileged --net=host -u linaro -v ~/shared_dir:/home/linaro/shared_dir:rw 96boards/autoware:auto_20200121 /bin/bash
$ source AutowareAuto/install/setup.bash
$ ros2 run velodyne_node velodyne_cloud_node_exe __params:=/home/"${USER}"/shared_dir/vlp16_test.param_ai.yaml
  • Terminal 5
$ docker exec -it -u linaro CONTAINER_NAME /bin/bash
$ source AutowareAuto/install/setup.bash
$ udpreplay ~/shared_dir/route_small_loop_rw-

A typical look of the terminals running the different nodes is shown in the image below.

Stopping the mapping process

The computational needs for the mapping task are quite heavy and so the processing will take longer as the number of points in the map grows. We can stop the pcap replaying by doing Ctrl+C in terminal 5. To save the map, in terminal 2 we do:

$ rostopic pub /config/ndt_mapping_output autoware_config_msgs/ConfigNDTMappingOutput "header:
  seq: 0
    secs: 0
    nsecs: 0
  frame_id: 'map'
filename: 'auto_map.pcd'
filter_res: 0.2"

Once the mapping process is complete the pcd map will be generated in the .ros folder. We need to move it to shared_dir if we want to keep it after we stop the container.

$ mv ~/.ros/auto_map.pcd ~/shared_dir/

Visualizing the pointcloud map

To visualize the pointcloud map we need to:

  • scp the pcd from the RB3 to the laptop and store it in shared_dir.
  • Start an Autoware.AI container as we did here.
$ cd ~/docker/generic
$ ./run.sh -c off -t 1.13.0
$ cd shared_dir
$ roscore &
$ rosrun map_file points_map_loader noupdate `pwd`/auto_map.pcd &
$ rviz

Within rviz we can select the /points_map topic and, as we did previously, increase the Size to 0.1. The result is displayed below.


With this post we now have 2 different boards (Hikey970 and Qualcomm® Robotics (RB3) Dragonboard-845c Development Platform) that we can use for development of Autoware.

This article is Part 5 in a 10-Part Series.

comments powered by Disqus