flowdrone build guide (in-progress)

Follow the instructions from the px4.io website. To focus the lens:

1. Plug in the PX4Flow by itself while QGroundControl is open. Navigate to the firmware tab and update the firmware by following the instructions.
2. In Vehicle Setup, you should see 'PX4Flow' appear. Clicking on that, you can see the black and white camera output for focusing.
3. To focus the lens, loosen the outer ring (around the lens) which will allow you to screw - unscrew the lens itself, changing the focal distance and focusing the text in the camera view.
4. In Analyze Tools -> MavLink Inspector -> OPTICAL_FLOW, you can see the range reading from the PX4Flow in ground_distance.

Pluggin the Pixhawk 4 back in, you will have to toggle the SENS_EN_PX4FLOW parameter value from 0->1. There are multiple ways to do this:

• Use QGroundControl -> Vehicle Setup -> Parameters -> SENS_EN_PX4FLOW
• QGroundControl -> Analyze Tools -> MAVLink console: param set SENS_EN_PX4FLOW 1

Once you have the PX4Flow sensor connected to the Pixhawk 4 via I2C1 -> I2CA, you can access the PX4Flow via QGroundControl -> Analyze Tools -> MAVLink console. The following commands are useful:

1. > px4flow info for general info
2. > px4flow start -X works to get the PX4Flow started (on I2C bus 4 (external). In the future, this may or may not start automatically upon booting up the Pixhawk 4.
3. MAVLink Insector should show new topics: OPTICAL_FLOW_RAD and DISTANCE. Keep in mind that DISTANCE from the PX4Flow is quite imprecise, hence the need for the LIDAR-Lite-v3.

Finally, in order to use the PX4Flow for state estimation, you must change the EKF2_AID_MASK parameter to use Optical Flow instead of GPS. This can be done in QGroundControl -> Vehicle Setup -> Parameters -> EKF2_AID_MASK.

Follow the instructions from the px4.io website. Similarly to with the PX4Flow, the LIDAR-Lite-v3 status can be checked using MAVLink console with ll40ls status. With the LIDAR installed, you should see the quality of the optical flow estimates be consistently high (~255). You can check the quality either through MAVLink Inspector -> OPICAL_FLOW_RAD or listener optical_flow in MAVLink console.

First, boot Ubuntu 20.04 onto the Raspberry Pi (youtube) using an SD card. I installed ubuntu-20.04.5-preinstalled-server-arm64+raspi.img.xz. If you would like the desktop version, you can upgrade via sudo apt-get install ubuntu-desktop. The youtube tutorial recommends modifying the netplan to enable connection to wifi upon boot. I prefer to install the default .iso and connect the Raspberry Pi to a monitor, mouse, and keyboard, and to edit the netplan from there.

Here is an example of how you could configure the netplan via sudo nano /etc/netplan/50-cloud-init.yaml : (be very careful about indentation in .yaml files!)

            
    network:
        ethernets:
            eth0:
                dhcp4: true
                optional: true
        wifis:
            wlan0:
                addresses:
                - 192.168.0.175/24
                gateway4: 192.168.0.1
                nameservers:
                    addresses: [192.168.0.1, 8.8.8.8]
                optional: true
                access-points:
                    "WIFI NAME":
                        password: "WIFI PASSWORD"
        version: 2
            
        

Follow the main tutorial: Fast DDS Installation. Here is the exact order in which I installed Fast-DDS and FastRTPSGen on a fresh Ubuntu 20.04 install:

        
sudo apt install openjdk-11-jre-headless
sudo apt install curl
sudo apt install zip
curl -s "https://get.sdkman.io" | bash
source "/home/ubuntu/.sdkman/bin/sdkman-init.sh"
sdk install gradle 6.3
sudo apt install cmake
sudo apt-get install build-essential

git clone https://github.com/eProsima/foonathan_memory_vendor.git
cd foonathan_memory_vendor
mkdir build && cd build
cmake ..
sudo cmake --build . --target install

sudo apt install libssl-dev
git clone --recursive https://github.com/eProsima/Fast-DDS.git -b v2.0.2 ~/FastDDS-2.0.2
cd ~/FastDDS-2.0.2
mkdir build && cd build
cmake -DTHIRDPARTY=ON -DSECURITY=ON ..
make -j$(nproc --all)
sudo make install

git clone --recursive https://github.com/eProsima/Fast-DDS-Gen.git -b v1.0.4 ~/Fast-RTPS-Gen \
    && cd ~/Fast-RTPS-Gen \
    && ./gradlew assemble \
    && sudo ./gradlew install
        
    

You can check your installation with which fastrtpsgen.

Follow the Ubuntu (Debian) installation instructions for ROS2 Foxy. I opted for the ROS-Base Install (Bare Bones) as opposed to the Desktop install. After installing and sourcing your installation, you should be able to run ros2 without error. Make sure you install the additional dependencies in the PX4 tutorial.

The px4_ros_com and px4_msgs messaging libraries provide the infrastructure to communicate with the Pixhawk 4 over ROS2. The official repositories do not (at the time of writing) support publisher/subscriber nodes written in Python, which we found very helpful in our development. If you wish to have this Python capability, I recommend installing our forks (and specifically, using the correct branch of our px4-ros-com fork). As a warning, our forks may not be maintained.

        
    mkdir -p ~/px4_ros_com_ros2/src
    git clone https://github.com/irom-lab/px4_ros_com.git ~/px4_ros_com_ros2/src/px4_ros_com
    git clone https://github.com/irom-lab/px4_msgs.git ~/px4_ros_com_ros2/src/px4_msgs
    git fetch origin other-python
    git checkout other-python
    cd ~/px4_ros_com_ros2/src/px4_ros_com/scripts
    ./build_ros2_workspace.bash