We propose a trajectory-based flight assistive system specifically designed for assisting drone racing that ensures the timeliness of executing user intents in high-speed scenarios, while maintaining accurate intent capture.

[ICRA 2024] Open source code for paper A Trajectory-based Flight Assistive System for Novice Pilots in Drone Racing Scenario

Video Links: Bilibili.

The project has been tested on Ubuntu 20.04(ROS Noetic). run the following commands to setup:

  git clone https://github.com/ZJU-FAST-Lab/Assistive-Racing.git
  catkin_make

jsk-rviz-plugins is used to simulate the FOV in rviz, and joy package is used to receive user input. you can install them by running the following commands:

  sudo apt-get install ros-noetic-jsk-rviz-plugins
  sudo apt-get install ros-noetic-joy

before running code, ensure joystick is connect to your pc successfully!

Firstly, generate the optimal reference trajectory and flight corridors offline:

  roslaunch map_generator offline_generate.launch

check if generated files exist in plan_manage/offlines, then run the following command to start flying:

  roslaunch plan_manage test.launch

trigger the 5 channel on your joystick to start fly!

Attention: We provide the param files for Blacksheep joystick and XBOX joystick, you can check the joystick type in test.launch, for other joystick, you need to modify the realted param files.

our code support loading your own map, you can place your map(.pcd) in plan_manage/pcd/, and your gate information(gate.yaml) in plan_manage/config/. Then running the code agian to fly in your own map!

We use MINCO as our trajectory representation. and We use DecompROS for safe flight corridor generation and visualization.