My absolutely first repository on github!
This repository contains my bachelor's degree thesis project: Designing a DDPG algorithm for approaching the lane keeping problem in the autonomous ground vehicle driving.
The work is structured as follows.
βββ src #The main folder of all the code
β βββ DDPG #Folder containing all the classes and functions for implementing the algo.
β | βββ noise #Foler containing the noise classes used fro the DDPG
| | | βββ OU.py #Implementation of the OrnsteinβUhlenbeck noise with the stochastic brake mod
| | | βββ GN.py #Implementation of the Gaussian Noise
| | | βββ TVNoise.py #Implementation of the Time Variant Noise
β β βββ conditioning.py #Implementation of a utility function for interacting with TORCS env.
β β βββ networks.py #Implementation of Actor and Critic networks
β β βββ replayBuffer.py #Implementation of the experience replay required by DDPG
β β βββ reward.py #Implementation of the three env. reward functions
β β βββ config.py #Config file you need to look before running any experiment with DDPG
β β βββ agent.py #Implementation of the agent class and all of its functions
β βββ launch.py #Main python script to launch the experiment
β βββ snakeoil.py #Implementation of the client interface for the interaction with TORCS game
βββ experiment #A copy of the src folder but with the final driver model in it for testing purpose
β βββ DDPG #Folder containing all the classes and functions for implementing the algo.
β | βββ noise #Foler containing the noise classes used fro the DDPG
| | | βββ OU.py #Implementation of the OrnsteinβUhlenbeck noise with the stochastic brake mod
| | | βββ GN.py #Implementation of the Gaussian Noise
| | | βββ TVNoise.py #Implementation of the Time Variant Noise
| βββ models #Folder containing different version of the driver.
β β βββ conditioning.py #Implementation of a utility function for interacting with TORCS env.
β β βββ networks.py #Implementation of Actor and Critic networks
β β βββ replayBuffer.py #Implementation of the experience replay required by DDPG
β β βββ reward.py #Implementation of the three env. reward functions
β β βββ config.py #Config file you need to look before running any experiment with DDPG
β β βββ agent.py #Implementation of the agent class and all of its functions
β βββ launch.py #Main python script to launch the experiment
β βββ snakeoil.py #Implementation of the client interface for the interaction with TORCS game
βββ docs #Folder containing documents
βββ Tesi.pdf #Thesis on DDPG design for lane keeping in TORCS Environment. Italian language
βββ latex_template #Latex template folder of my thesis project