autopilot/README.md

Azaion DroneController and AiController

An autonomous drone controller with AI inference support. Programs uses MAVSDK framework and ArduPilot flight controller. Running the applications requires MAVSDK installation and compiling ArduPilot. Example has been tested in Ubuntu 22.04 and 24.04 environments.

Install necessary dependencies

sudo apt update
sudo apt install ccache git build-essential qmake6 qt6-base-dev

Speed up the compilations

echo "export MAKEFLAGS=\"-j$(($(nproc)))\"" >> ~/.bashrc
echo "export PATH=/usr/lib/ccache:\$PATH"  >> ~/.bashrc

Clone source codes. SSH key must be added before the cloning!

git clone git@github.com:azaion/autopilot.git
git clone --recursive git@github.com:azaion/autopilot.git
git submodule update --init --recursive

Install MAVSDK for Ubuntu 20.04

wget https://github.com/mavlink/MAVSDK/releases/download/v2.12.10/libmavsdk-dev_2.12.10_ubuntu20.04_amd64.deb
sudo dpkg -i libmavsdk-dev_2.12.10_ubuntu20.04_amd64.deb

Install MAVSDK for Ubuntu 22.04

wget https://github.com/mavlink/MAVSDK/releases/download/v2.12.10/libmavsdk-dev_2.12.10_ubuntu22.04_amd64.deb
sudo dpkg -i libmavsdk-dev_2.12.10_ubuntu22.04_amd64.deb

Install MAVSDK for Ubuntu 24.04 (libmavsdk-dev3.0.0)

wget https://github.com/mavlink/MAVSDK/releases/download/v3.0.0/libmavsdk-dev_3.0.0_ubuntu24.04_amd64.deb
sudo dpkg -i libmavsdk-dev_3.0.0_ubuntu24.04_amd64.deb

Install ONNX Runtime for Ubuntu (not required for embedded platforms)

With GPU inference

wget https://github.com/microsoft/onnxruntime/releases/download/v1.18.0/onnxruntime-linux-x64-gpu-cuda12-1.18.0.tgz
sudo tar xf onnxruntime-linux-x64-gpu-cuda12-1.18.0.tgz -C /opt/
sudo ln -s /opt/onnxruntime-linux-x64-gpu-cuda12 /opt/onnxruntime-linux-x64-1.18.0

With CPU inference

wget https://github.com/microsoft/onnxruntime/releases/download/v1.18.0/onnxruntime-linux-x64-1.18.0.tgz
sudo tar xf onnxruntime-linux-x64-1.18.0.tgz -C /opt/

Install MAVSDK for Ubuntu 22.04 or 24.04 PCs

wget https://github.com/mavlink/MAVSDK/releases/download/v2.12.10/libmavsdk-dev_2.12.10_ubuntu22.04_amd64.deb
sudo dpkg -i libmavsdk-dev_2.12.10_ubuntu22.04_amd64.deb

Build ArduPilot

git clone --recursive https://github.com/ArduPilot/ardupilot.git
cd ardupilot
./Tools/environment_install/install-prereqs-ubuntu.sh -y
. ~/.profile
./waf configure --board=sitl
./waf build

Build and install OpenCV 4.10.0

sudo apt update
sudo apt install libgtk-3-dev libpng-dev cmake ffmpeg libavcodec-dev libavformat-dev libavfilter-dev
wget https://github.com/opencv/opencv/archive/refs/tags/4.10.0.zip
unzip 4.10.0.zip
cd opencv-4.10.0
mkdir build && cd build
cmake -DCMAKE_INSTALL_PREFIX=/opt/opencv-4.10.0 -DBUILD_opencv_world=ON -DOPENCV_GENERATE_PKGCONFIG=ON -DBUILD_PERF_TESTS=OFF -DBUILD_TESTS=OFF ..
make -j8 && sudo make install

Install MAVSDK for embedded platforms

Build and install MAVSDK from the sources if needed

sudo apt-get update
sudo apt-get install build-essential cmake git
git clone https://github.com/mavlink/MAVSDK.git
cd MAVSDK
git checkout tags/v2.12.10
git submodule update --init --recursive
cmake -Bbuild/default -DCMAKE_BUILD_TYPE=Release -H.
cmake --build build/default -j8
sudo cmake --build build/default --target install

Build Azaion applications

PC builds (uses ONNX Runtime for the AI inference)

qmake6 && make

OPI5 builds (uses gimbal camera and RKNN libraries for the AI inference)

qmake6 CONFIG+=opi5 CONFIG+=gimbal && make

Run Azaion drone controller with ArduPilot simulator in Ubuntu PC

Launch similator in the ArduPilot directory

./Tools/autotest/sim_vehicle.py --map --console -v ArduCopter

Launch example application in the new terminal window after waiting simulator (around 1 min) to be ready

./drone_controller/drone_controller ./drone_controller/mission.json quadcopter udp

Run Azaion AI controller in Ubuntu PC

Install ffmpeg and mediamtx

sudo apt update
sudo apt install ffmpeg
wget https://github.com/bluenviron/mediamtx/releases/download/v1.8.4/mediamtx_v1.8.4_linux_amd64.tar.gz
mkdir mediamtx
tar xf mediamtx_v1.8.4_linux_amd64.tar.gz -C mediamtx

Launch RTSP server if real camera is not used

cd mediamtx
./mediamtx

Playback of RTSP video stream when no real camera is used

ffmpeg -re -stream_loop -1 -i SOME_MP4_VIDEO_FILE -c copy -f rtsp rtsp://localhost:8554/live.stream

Test RTSP stream with ffplay

ffplay -rtsp_transport tcp rtsp://localhost:8554/live.stream

Compile and run AI controller without camera support

Modify ./ai_controller/aiengineconfig.h and change IP address of RTSP source

qmake6 CONFIG+=opi5 && make && ./ai_controller/ai_controller [ONNX_MODEL_FILE]

Compile and run AI controller with camera support

Modify ./ai_controller/aiengineconfig.h and change IP address of RTSP source

qmake6 CONFIG+=opi5 CONFIG+=gimbal && make && ./ai_controller/ai_controller [ONNX_MODEL_FILE]

Run Azaion AI controller in OPI5

Compile and run AI controller without gimbal camera support

Modify ./ai_controller/aiengineconfig.h and change IP address of RTSP source

qmake6 CONFIG+=opi5 && make && ./ai_controller/ai_controller [RKNN_MODEL_FILE]

Compile and run AI controller without gimbal camera support

Modify ./ai_controller/aiengineconfig.h and change IP address of RTSP source

qmake6 CONFIG+=opi5 CONFIG+=gimbal && make && ./ai_controller/ai_controller [RKNN_MODEL_FILE]