Multi-Agent_SelfDriving

This is a ongoing project.

Currently, This code support DDPG not MADDPG.

You can run ddpg merge scenario simulator using ddvg_v3.py.

and This repository requires many refinement and code refactoring.

I will try to update this repository as soon as possible.

Thanks.

Overview

A project that trains a multi-agent to simulate double merge scenario.

Python 2.7 To install Python, I just install Anaconda here
Pygame pip install pygame
Pymunk. Use Pymunk 4. Using version 5 will not work as it was a major refactor. V4 is here

**Unpack it:

***tar zxvf pymunk-4.0.0.tar.gz then ***cd pymunk-pymunk-4.0.0 install pymunk ***python setup.py install

Install tensorflow ** pip install tensorflow **if you want to use GPU-version, check this page
Install Keras **pip install keras

*To be Continue

This is a multi agent simulation of self driving in double merge scenario.

Language:Python 100.0%