【CDC 2023 Cooperative Aerial Robots Inspection Challenge】

CDC 2023 Cooperative Aerial Robots Inspection Challenge 网址

Install the CARIC packages

Once the dependencis have been installed, please create a new workspace for CARIC, clone the necessary packages into it, and compile:

# Create the workspace
mkdir -p ~/ws_caric/src
cd ~/ws_caric/src

# Download the packages:

# Manager node for the mission
git clone https://github.com/ntu-aris/caric_mission

# Simulate UAV dynamics and other physical proccesses
git clone https://github.com/ntu-aris/rotors_simulator

# GPU-enabled lidar simulator, modified from: https://github.com/lmark1/velodyne_simulator
git clone https://github.com/ntu-aris/velodyne_simulator

# Converting the trajectory setpoint to rotor speeds
git clone https://github.com/ntu-aris/unicon

# To generate an trajectory based on fixed setpoints. Only used for demo, to be replaced by user's inspection algorithms
git clone https://github.com/ntu-aris/traj_gennav

# Build the workspace
cd ~/ws_caric/
catkin build

The compilation may report errors due to missing depencies or some packages in CARIC are not yet registered to the ros package list. This can be resolved by installing the missing dependencies (via sudo apt install < package > or sudo apt install ros-$ROS_DISTRO-)). Please try catkin build again a few times to let all the compiled packages be added to dependency list.

Run the flight test

To make sure the code compiles and runs smoothly, please launch the example flight test with some pre-defined fixed trajectories as follows:

source ~/ws_caric/devel/setup.bash
roscd caric_mission/scripts
bash launch_demo_paths.sh

You should see 5 UAVs take off, follow fixed trajectories, and fall down when time is out.

The benchmark design

The UAV fleet

A 5-UAV team is designed for the challenge, two of the explorer class (nicknamed jurong and raffles), and three of the photographer class (changi, sentosa, and nanyang), plus one GCS (Ground Control Station). Each unit has an intended role in the mission.

An illustration of one GCS, one explorer, and two photographers in Gazbo environment：

• The explorer drone carries a rotating lidar apparatus and a gimballed camera.
• The photographer is smaller and only carries gimballed camera.
• The GCS is where the images will be sent back and used to update the final score.
  Note that the explorer is twice the size and weight of the photographer. Thanks to the bigger size, it can carry the lidar and quickly map the environment, at the cost of slower speed. In contrast, the photographers have higher speed, thus they can quickly cover the surfaces that have been mapped by the explorer to obtain images of higher score. The GCS’s role is to compare the images taken by the drones. For each interest point, the GCS can select the image with the best quality to assign the score to it.