https://github.com/rsgoksel/mechopter
PyGame-based quadcopter simulator & Reinforcement Learning Project
https://github.com/rsgoksel/mechopter
a2c dqn ipynb ppo pygame python quadcopter reinforcement-learning
Last synced: 7 months ago
JSON representation
PyGame-based quadcopter simulator & Reinforcement Learning Project
- Host: GitHub
- URL: https://github.com/rsgoksel/mechopter
- Owner: RsGoksel
- License: apache-2.0
- Created: 2023-10-22T07:34:49.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-11-08T14:56:27.000Z (over 2 years ago)
- Last Synced: 2024-10-11T15:40:12.689Z (over 1 year ago)
- Topics: a2c, dqn, ipynb, ppo, pygame, python, quadcopter, reinforcement-learning
- Language: Python
- Homepage:
- Size: 481 KB
- Stars: 5
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Mech-opter 🎮
This repository provides code exercise and solution for Reinforcement Learning. All code is written in Python and uses custom RL environment from OpenAI Gym.
Goal was train the Reinforcement Learning Drone Agent mastering the flying and accurately hitting a target balloon. A2C was used primaly, PPO and DQN also can be used. Explanations and details are down below ↓
## Game Screen (Before training, trained one is down below!)
________________________________________________________________________________________________________________
# Environment & Game Mechanism
### _**Reward:**_
* -score | Proportional to the distance to the target
* Hitting the target is +score, & restart
* Flying out of the game screen is -score & restart
### _**Observation Space:**_
* Agent Y Coordinate
* Angle to Up
* Velocity
* Angular Velocity
* Angle to Target
* Angle Between Target and Velocity
* Distance to Target
### _**Action Space:**_
Action space is discrete(5). It means there is certain 5 moves the Agent has to do. *__Do Nothing, Up, Down, Right, Left__*
________________________________________________________________________________________________________________
## Files
* [test.py](https://github.com/RsGoksel/Train-PPO-Agent_PongGame/blob/main/test.py): Testing for the environment. You can display game screen.
* [train.py](https://github.com/RsGoksel/Train-PPO-Agent_PongGame/blob/main/train.py): Trains the Agent. You can change total_steps from Constants.py.
* [Agent.py](https://github.com/RsGoksel/Train-PPO-Agent_PongGame/blob/main/Agent.py): Drone class.
* [evaluate.py](https://github.com/RsGoksel/Train-PPO-Agent_PongGame/blob/main/evaluate.py): Evaluate the trained model. Elaborated usage is at down below.
* [env.py](https://github.com/RsGoksel/Train-PPO-Agent_PongGame/blob/main/Env.py) : Environment class. You can alter the game rules, Reward mechanism etc.
* [Constants.py](https://github.com/RsGoksel/Train-PPO-Agent_PongGame/blob/main/Constants.py) : Constant variables of the Game. Screen width, hyperparameters etc.
________________________________________________________________________________________________________________
## Install required libraries:
```
$ pip install -r requirements.txt
```
## Test the environment and check everything is OK:
```
$ python test.py
```
## Let's Train!
Default step is 300k. You can alter it from Constants.py
(Loading Libraries may take a while)
```
$ python train.py
```
## Evaluting the Agent
! After the training, your model will be saved in 'models' file.
Evaluate your trained model with adding --model parameter to terminal,
Or use pretrained models Which in __*models*__ folder.
There is a already pretrained model which named "Crayz_bill" and it is default model.
```
$ python evaluate.py
```
__*If you would like to evaluate your custom model:*__
```
$ python evaluate.py --model models/yourmodel
$ python evaluate.py --model models/200k
```
# After the Training 🦾
________________________________________________________________________________________________________________