Elastic-Tracker

0. Overview

Elastic-Tracker is a flexible trajectory planning framework that can deal with challenging tracking tasks with guaranteed safety and visibility.

Authors: Jialin Ji, Neng Pan and Fei Gao from the ZJU Fast Lab.

Paper: Elastic Tracker: A Spatio-temporal Trajectory Planner Flexible Aerial Tracking, Jialin Ji, Neng Pan, Chao Xu, Fei Gao, Accepted in IEEE International Conference on Robotics and Automation (ICRA 2022).

Video Links: youtube or bilibili

<img src="https://github.com/ZJU-FAST-Lab/Elastic-Tracker/raw/main/figs/cover.png" width="500"/>

1. Simulation of Aerial Tracking

[NOTE] remember to change the CUDA option of src/uav_simulator/local_sensing/CMakeLists.txt

Preparation and visualization:

git clone https://github.com/ZJU-FAST-Lab/Elastic-Tracker.git
cd Elastic-Tracker
catkin_make
source devel/setup.zsh
chmod +x sh_utils/pub_triger.sh
roslaunch mapping rviz_sim.launch

A small drone with the global map as the chasing target:

roslaunch planning fake_target.launch

Start the elastic tracker:

roslaunch planning simulation1.launch

Triger the drone to track the target:

./sh_utils/pub_triger.sh

<img src="https://github.com/ZJU-FAST-Lab/Elastic-Tracker/raw/main/figs/sim1.gif" width="500"/>

Comparision of the planners with and without visibility guarantee:

roslaunch planning simulation2.launch

<img src="https://github.com/ZJU-FAST-Lab/Elastic-Tracker/raw/main/figs/sim2.gif" width="500"/>

2. Simulation of Aerial Landing

First start the stage of tracking:

roslaunch planning fake_car_target.launch
roslaunch planning simulation_landing.launch
./sh_utils/pub_triger.sh

Triger the drone to land on the moving vehicle:

./sh_utils/land_triger.sh

<img src="https://github.com/ZJU-FAST-Lab/Elastic-Tracker/raw/main/figs/sim_landing.gif" width="500"/>

3. Acknowledgement

We use MINCO as our trajectory representation.

We use DecompROS for safe flight corridor generation and visualization.