Drone Vision Landing Project on PX4 SITL Simulation

Our project is to make the drone autonomously land on a launch pad by ROS.

The task could be mostly split into two parts: Vision and Control.

The first part is vision system which would detect the pad on the ground and compute the relative distance. This is based on opencv aruco marker algorithm. (/catkin_ws/src/opencvtest_py/src/test.py) The input of the system is the raw image of the camera and the output is the relative distance between drone and the launch pad in camera reference frame.

The second part is the PID control system(catkin_ws/control/src/control.py). When the vision system publishes the relative distance , the control system would subscribe it and use it as the input. A 3 dimension velocity loop PID control is built as we would control the x, y, z linear velocity to approach the mark.

Have fun about our project!!!

Quick Start

Open the PX4 simulation environment

cd ~/src/Firmware

source Tools/setup_gazebo.bash $(pwd) $(pwd)/build/px4_sitl_default export ROS_PACKAGE_PATH=$ROS_PACKAGE_PATH:$(pwd) export ROS_PACKAGE_PATH=$ROS_PACKAGE_PATH:$(pwd)/Tools/sitl_gazebo

roslaunch px4 drone.launch

Run the offboard program so that the drone could enter offboard mode and be armed

rosrun offboard offboard_node

Run the keyboard program so that we could use keyboard to make the drone fly

rosrun keyboard keyboard_ndoe

Run the vision system

rosrun opencvtest_py test.py

Run the control system(Only when the mark is within the view of drone camera)

rosrun control control.py

Then you could use keyboad to move the drone and let it automously landing on the mark by control program!