Diffusion PS: DRL for gyroscope control (master)

Edit
PS: DRL for gyroscope control
ActivePublic

Semester project

Title: Deep reinforcement learning for gyroscope control
Supervisor: Prof. Alireza Karimi
Direct supervisor: Paul-Arthur Dreyfus
Section: Robotics
Author: Matthieu Le Cauchois

PS: DRL for gyroscope control (master)

.ipynb_checkpoints/
README.ipynb
README.md
README.pdf
code/
report/

Recent Commits

Commit	Author	Details	Committed
6d81585206bf	MatthieuLC	added everything	Jun 28 2020
e072711e8645	MatthieuLC	delete	Jun 28 2020
a28b1ddcdcc2	MatthieuLC	removed previous env folder	Mar 25 2020
79122de832f8	MatthieuLC	param search up and running with GyroscopeEnv added to Gym lib folder	Mar 25 2020
b33e00fde1b8	MatthieuLC	grid search works	Mar 25 2020
4e69f4222150	MatthieuLC	added support for gyro env added to gym api	Mar 25 2020
e8362b851456	MatthieuLC	added support for gym env downloaded to gym api	Mar 25 2020
7daad2e32ff0	MatthieuLC	code for param search and baseline iterations, resources not included and must…	Mar 25 2020
dfb40e8f4cd2	MatthieuLC	param search notebook	Mar 25 2020
87458acae87e	MatthieuLC	code files,not resources	Mar 24 2020

README.md

DRLGyro

DRLGyro is the project folder for my semester project *Deep Reinforcement Learning to Control a Gyroscope* conducted during the 2020 spring semester at EPFL.

For more information, please contact me at: <lecauchoismatthieu@gmail.com>

Installation

The project highly depends on two frameworks, namely OpenAI's *Gym* and *Spinning Up* frameworks, which can be installed as such:

For Gym:

pip install gym

For Spinning Up:

git clone https://github.com/openai/spinningup.git 
cd spinningup 
pip install -e .

Files

The project folder consists of two main directories:

bash
├── code 'All code files and model files of the project'
└── report 'Report and pictures of the report'

The *code* directory contains the utilities, experiments and results of the project. Not all directories and sub-directories are shown here, such as the many model folders that were generated during the experimentation process, but the most relevant ones are. To understand what the many model experimentation folders correspond to, please refer to the notebooks where the experiments are performed.

bash
├── custom_functions 'Custom python library implementing useful functions for the project'
├── environment
│   ├── Gyroscope 'Environment folder ready for import as a library'
│   │   └── gym_GyroscopeEnv
│   │       └── envs
│   │           ├── gyroscopediscontinuous_env.py 'Environment with discontinuity'
│   │           ├── gyroscope_env.py 'Simple environment'
│   │           ├── gyroscopeintegral_env.py 'Environment with additionam integral term'
│   │           └── gyroscoperobust_env.py 'Robust environment'
│   └── gyroscope_environment_testing.ipynb 'Simulation code (python class) transferred to the Gym API'
├── simulation 'Initial simulation experiments'
│   ├── AgramFrame.png
│   └── gyroscope_simulation.ipynb 'Simulation coding and experiments'
├── training_spinuplib 'Experiments using the spinning up library'
│   ├── activations_visualization.ipynb 'Visualization notebook to cluster activations using t-SNE'
│   ├── baseline_norm_ddpg_td3_sac.ipynb 'Testing notebook between DRL algorithms for norm. reward'
│   ├── baseline_quadreward_ddpg_td3_sac.ipynb 'Testing notebook between DRL algorithms for quad. reward'
│   ├── discontinuity_ddpg.ipynb 'Testing notebook for discontinuous environment'
│   ├── envvariants_quadreward.ipynb 'Testing notebook with additional integral term'
│   ├── paramsearch_iter0_norm_td3.ipynb 'Testing notebook for hyperparamater search iteration 0'
│   ├── paramsearch_iter1_norm_td3.ipynb 'Testing notebook for hyperparamater search iteration 1'
│   ├── paramsearch_iter2_norm_td3.ipynb 'Testing notebook for hyperparamater search iteration 2 and 3'
│   ├── reward_testing_ddpg.ipynb 'Testing notebook for reward function analysis'
│   ├── robust_agent_testing.ipynb 'Testing notebook for the robust agent'
│   ├── simple_agent_testing.ipynb 'Testing notebook for the simple agent'
│   ├── td3_n_it2_an_var1_extralay_extralay_s0 'Model folder (load w/ pytorch) of the final simple agent'
│   └── best_agent_robustified 'Model folder (load w/ pytorch) of the final robust agent'
└── training_udalib 'Preliminary experiments using the udalib library before moving to spinning up'

python

EditPS: DRL for gyroscope controlActivePublic