(Twizy) Collect launch files for sensors

[dtonda8]

• x2 people Goal: Collate/create/find launch files (.xml or .launch.py) for launching sensors

Ensure sensors work and publish correct Message Types:

• [x] LiDAR as sensor_msgs/PointCloud2.msg
• [x] IMU as sensor_msgs/Imu.msg
• [x] Cameras as sensor_msgs/Image.msg
• [x] GNSS-INS as sensor_msgs/msg/NavSatFix.msg

Launch file tasks:

• [x] Create a camera.launch.xml or .py launch file
• [ ] Create a lidar.launch.xml or .py launch file

Helpful resources

• Twizy Sensor's Notion Doc
• Autoware's Sensor launch
• Sensor component design details more about each sensor
• Consult this unresolved issue for more details about what's currently working

[dtonda8]

SSH

ssh mcav@192.168.11.127

password: mcav01

[dtonda8]

For Velodyne, Alienware Laptop (via ethernet or wirelessly) and LIDAR (via ethernet) must be connected to in-built router First terminal:

ros2 launch velodyne_driver velodyne_driver_node-VLP16-launch.py

Second terminal:

ros2 launch velodyne_pointcloud velodyne_transform_node-VLP16-launch.py

Visualising

Use RViz (type rviz2 on terminal):

1. Set Fixed Frame to velodyne
2. Click "Add" on the bottom left and add the PointClound2 topic (if you can't see this then the lidar is not properly working)

[dtonda8]

Docs for starting the Vehicle Interface

[jxcksonli]

http://wiki.ros.org/video_stream_opencv https://github.com/clydemcqueen/h264_image_transport

[dtonda8]

Current GStreamer files

https://github.com/Monash-Connected-Autonomous-Vehicle/stream_sample_camera_gmsl/tree/stream_directly

Testing GStreamer locally

Server

# From file
cat test1.h264 | gst-launch-1.0 -e fdsrc ! h264parse ! \
    avdec_h264 ! videoconvert ! videorate ! video/x-raw,framerate=30/1 ! \
    x264enc tune=zerolatency bitrate=1000 speed-preset=ultrafast ! \
    video/x-h264,profile=baseline ! rtph264pay ! udpsink \
    host=<CHANGE TO YOUR IP> port=5000

# From mac
gst-launch-1.0 avfvideosrc device-index=1 ! videoscale ! videoconvert ! x264enc tune=zerolatency bitrate=3000 speed-preset=ultrafast ! h264parse ! rtph264pay ! udpsink host=<CHANGE TO YOUR IP> port=5000

Client

# Get stream 
gst-launch-1.0 -e udpsrc port=5000 caps="application/x-rtp" ! rtpjitterbuffer \
    ! rtph264depay ! avdec_h264 ! fpsdisplaysink

For cameras, either:

1. Publish to ROS from PX 2
2. Stream and publish to ROS form Alienware Laptop

PX 2 SSH:

1. Connect to StreetDrone wifi
2. SSH

   ssh nvidia@192.168.1.180

3. Password: nvidia

Sample files and stream repo (switch to stream_directly)

[dtonda8]

gscam2 might be useful

[jxcksonli]

Attempt using gscam2, located in mcav@twizy-x86:~/dbat/test1$

Ran using ros2 run gscam2 gscam_main --ros-args --remap /image_raw:=/my_camera/image_raw --params-file src/gscam2/gscam_params.yaml -p camera_info_url:=file://$PWD/src/gscam2/cfg/my_camera.ini

Error: Failed to Pause Stream, Check GScam config

[dtonda8]

ROS answer on h264 streaming i.e. what we want

[dtonda8]

Hi guys gscam2 works! I've tried in my laptop. The tasks for you guys is to:

• [ ] Create a camera.launch.xml or .py launch file that runs the script. If you remember from last Thursday, gscam required a local workspace, so modify the script to source the correct workspace location.
• [ ] Create a lidar.launch.xml or .py launch file to launch the velodyne that we tested a 2 weeks ago

We won't need to create launch files for IMU since that's already done by the vehicle interface

gscam details

• Installation. Then

  source install/local_setup.bash && export GSCAM_CONFIG='udpsrc port=5000 caps="application/x-rtp" ! rtpjitterbuffer ! rtph264depay ! avdec_h264 ! videoconvert' && ros2 run gscam2 gscam_main --ros-args -p sync_sink:=true -p preroll:=true

ROS 2 tutorial on creating launch files