Skip to content

Aakriti05/Prop-Fail-Detect-Control-RL

 
 

Repository files navigation

Mid-flight Propeller Failure Detection and Control of Propeller-deficient Quadcopter using Reinforcement Learning

Installation

For installation instructions please see https://github.com/leggedrobotics/raisimGym.

Usage

scripts/quadcopter_position tracking.py can be used to train RL agents as explained in the link above. To test a model, scripts/plotting_stuff.py can be used with the following arguments:

--scene:
0 - 4 propeller system (Default)
1 - 3 propeller system
2 - 2 propeller system
3 - 4 to 3 propeller system
4 - 3 to 2 propeller system

--plot:
0 - Suppress plotting (Default)
1 - Enable plotting

The accompanying paper can be found at https://arxiv.org/abs/2002.11564.

For any queries please contact us at juhi05aakritiagrawal@gmail.com.

Contributors

  • Rohitkumar Arasanipalai
  • Aakriti Agrawal
  • Debasish Ghose

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 93.3%
  • Dockerfile 4.1%
  • Shell 1.4%
  • CMake 1.2%