End-to-end autonomous driving based on reinforcement learning

In this project, an RGB camera will be used as data input to explore an end-to-end method based on visual based reinforcement learning. The project will be carried out with the Unity game engine as the training environment, along with Unity’s ML-Agents package that provides out of the box deep Reinf...

Full description

Saved in:
Bibliographic Details
Main Author: Ong, Chee Wei
Other Authors: Lyu Chen
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2022
Subjects:
Online Access:https://hdl.handle.net/10356/158264
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-158264
record_format dspace
spelling sg-ntu-dr.10356-1582642022-06-02T08:08:32Z End-to-end autonomous driving based on reinforcement learning Ong, Chee Wei Lyu Chen School of Mechanical and Aerospace Engineering lyuchen@ntu.edu.sg Engineering::Mechanical engineering::Mechatronics Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision In this project, an RGB camera will be used as data input to explore an end-to-end method based on visual based reinforcement learning. The project will be carried out with the Unity game engine as the training environment, along with Unity’s ML-Agents package that provides out of the box deep Reinforcement Learning (RL) algorithms to interface with their environment. The results of training a simulated donkey car to drive in its own lane with an on-policy method, Proximal Policy Optimization (PPO), and an off-policy method, Soft Actor-Critic (SAC) will be compared. An ablation study, consisting of adding Generative Adversarial Imitation Learning (GAIL), semantic segmentation and stacked visual inputs, will be performed. Additionally, RL based obstacle avoidance will be explored. The results, based on stability of control and ability to stay in lane, indicate that the best performing method is PPO. Code is available at: https://github.com/MrOCW/Autonomous-Driving-RL-Unity Bachelor of Engineering (Mechanical Engineering) 2022-06-02T08:08:32Z 2022-06-02T08:08:32Z 2022 Final Year Project (FYP) Ong, C. W. (2022). End-to-end autonomous driving based on reinforcement learning. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/158264 https://hdl.handle.net/10356/158264 en C040 application/pdf Nanyang Technological University
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Engineering::Mechanical engineering::Mechatronics
Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
spellingShingle Engineering::Mechanical engineering::Mechatronics
Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
Ong, Chee Wei
End-to-end autonomous driving based on reinforcement learning
description In this project, an RGB camera will be used as data input to explore an end-to-end method based on visual based reinforcement learning. The project will be carried out with the Unity game engine as the training environment, along with Unity’s ML-Agents package that provides out of the box deep Reinforcement Learning (RL) algorithms to interface with their environment. The results of training a simulated donkey car to drive in its own lane with an on-policy method, Proximal Policy Optimization (PPO), and an off-policy method, Soft Actor-Critic (SAC) will be compared. An ablation study, consisting of adding Generative Adversarial Imitation Learning (GAIL), semantic segmentation and stacked visual inputs, will be performed. Additionally, RL based obstacle avoidance will be explored. The results, based on stability of control and ability to stay in lane, indicate that the best performing method is PPO. Code is available at: https://github.com/MrOCW/Autonomous-Driving-RL-Unity
author2 Lyu Chen
author_facet Lyu Chen
Ong, Chee Wei
format Final Year Project
author Ong, Chee Wei
author_sort Ong, Chee Wei
title End-to-end autonomous driving based on reinforcement learning
title_short End-to-end autonomous driving based on reinforcement learning
title_full End-to-end autonomous driving based on reinforcement learning
title_fullStr End-to-end autonomous driving based on reinforcement learning
title_full_unstemmed End-to-end autonomous driving based on reinforcement learning
title_sort end-to-end autonomous driving based on reinforcement learning
publisher Nanyang Technological University
publishDate 2022
url https://hdl.handle.net/10356/158264
_version_ 1735491139083960320