Fast-Drone-250-v2

This document is an extension of Fast-Drone-250, which is closer to the real experiment through Gazebo simulation.

The main updates are:

1. Expanded the single drone to multiple drones and expressed them in the form of MINCO curves
2. Added the one-key takeoff function of the controller
3. Optimized the visual interface

Achieve results：

And here is the video(Bilibili)

How to use:

1. Follow this link to install the simulation platform. For details, please refer to this project

2. Copy the startup files and environment files to the correct directories

   mv multi_vehicle.launch ~/Px4_firmware/launch

   mv forest.world ~/Px4_firmware/Tools/sitl_gazebo/world

3. Start Simulation：

   roslaunch px4 multi_vehicle.launch

4. Follow this link to set the location

5. Start planer

   roslaunch ego_planer swarm_gazebo.launch

   roslaunch ego_planer rviz.launch

6. Start Controller

   roslaunch px4ctrl multi_run.launch

Monitoring UAV State

You can use this tools to monitor the UAV's state and the mode of the planer and controller.

Existing Problems

1. The drone position detection module was not added, which resulted in other drones being modeled as obstacles when building the map, resulting in mediocre planning results.

2. The performance of the linear controller is average, and the tracking effect of the drone is average. The drone trajectory does not overlap with the planned trajectory, and the drone is prone to collision when the speed is high.

Reference Projects

1. XTDroen
   The simulation platform of the experiment is modified based on this project. Interested readers can refer to the document for reproduction
2. Fast-Drone-250
   The controller module comes from this project, but this project is only for real objects and a single drone
3. EGO-Planner-v2
   The cluster planning part comes from this project, but the simulation environment of this project is not consistent with reality