This repository contains the flight controller, and the vision-based state estimator for a quadrotor subjected to complete failure of a single rotor. The program runs in a ROS environment.
If you use this work in an academic context, please cite the following RA-L publication:
S. Sun, G. Cioffi, C. de Visser and D. Scaramuzza, "Autonomous Quadrotor Flight despite Rotor Failure with Onboard Vision Sensors: Frames vs. Events," IEEE Robot. Autom. Lett. (RA-L). 2021. PDF, VIDEO
@ARTICLE{SunAutonomous2021,
author={S. {Sun} and G. {Cioffi} and C. {de Visser} and D. {Scaramuzza}},
journal={IEEE Robotics and Automation Letters},
title={Autonomous Quadrotor Flight despite Rotor Failure with Onboard Vision Sensors: Frames vs. Events},
year={2021},
volume={6},
number={2},
pages={580-587},
doi={10.1109/LRA.2020.3048875}}
Vision-based pose estimator using a standard camera.
Rotation-corrected Complementary Filter and Extended Kalman Filter providing full pose and velocity estimates for the fault-tolerant flight controller.
Fault-tolerant flight controller that generates thrust command of individual rotors, using a Nonlinear Dynamic Inversion approach.
This installation guide was tested with Ubuntu 18.04
Requirements:
- CMake >= 3.0
- ROS (>= Kinetic) (see installation guide).
You also need to install the following packages for Ceres Solver:
sudo apt install liblapack-dev libblas-dev
Create a ROS workspace
mkdir -p ~/catkin_ws/src && cd ~/catkin_ws
catkin init
Configure the ROS environment
catkin config --extend /opt/ros/melodic --cmake-args -DCMAKE_BUILD_TYPE=Release
Clone the repository and dependencies
cd src/
git clone https://github.com/uzh-rpg/fault_tolerant_control.git
vcs-import < fault_tolerant_control/dependencies.yaml
Build the workspace
cd ~/catkin_ws && catkin build
If you run the program in arm64 architecture, please add
set(ENV{ARM_ARCHITECTURE} aarch64)
in /vio_estimator/common/ze_common/CMakeLists.txt and /vio_estimator/common/ze_cmake/cmake/modules/ze_setup.cmake
in /fast_neon package, use branch /test/aarch64-compilation by
git checkout /test/aarch64-compilation
You can test the vision-based state estimator by running this rosbag recorded from real-flights using a standard camera (mvBLuefox).
roslaunch ze_vio_ceres live_Bluefox.launch
rosbag play <directory_of_the_bag>/data_bluefox.bag
Due to the copyright reasons, the event-based estimator will not be released at the moment.
You can test the flight controller in RotorS, a MAV gazebo simulator. First of all, please install Gazebo.
Then run the following launch file
roslaunch ftc_ctrl testSim.launch
The drone is ready to fly in the simulator. We provide some simple commands in the /scripts folder
roscd ftc_ctrl && cd scripts
To take off, please type
./start_rotors.sh hummingbird
Then you can switch one motor off by running the script
./stop_rotor.sh hummingbird
You may decide which rotor to be turned off in simulation.yaml.
To control the quadrotor to a way point, e.g., xyz = [1.0 2.0 3.0], use the script
./reference.sh 1.0 2.0 3.0 hummingbird
You may reproduce our experiments in real flights, if you have
- A mvBluefox camera and it's ROS driver.
- A Terarranger-One range sensor and it's ROS_driver.
- A quadrotor that you can control the thrust of each individual rotor.
- A Joystick.
- Safety nets ;)
To start the flight controller, please run
roslaunch ftc_ctrl test.launch
The controller will publish rotor thrust command in a rostopic
/control_command/rotor_thrust
You can feed these command to your low-level motor controller.
Once the drone starts hovering before one motor is switched off, you can start the state estimator by running the following launch files:
roslaunch ze_vio_ceres live_Bluefox.launch
The estimated state will be published in the ROS topic
/state_est
which is used by the flight controller.