A Survey on Imitation Learning Techniques for End-to-End Autonomous Vehicles

Luc  Le Mero; Dewei Yi; Mehrdad Dianati; Alexandros  Mouzakitis

doi:10.1109/TITS.2022.3144867

A Survey on Imitation Learning Techniques for End-to-End Autonomous Vehicles

Luc Le Mero, Dewei Yi, Mehrdad Dianati, Alexandros Mouzakitis

Research output: Contribution to journal › Article › peer-review

61 Citations (Scopus)

2 Downloads (Pure)

Abstract

The state-of-the-art decision and planning approaches for autonomous vehicles have moved away from manually designed systems, instead focusing on the utilisation of largescale datasets of expert demonstration via Imitation Learning
(IL). In this paper, we present a comprehensive review of IL approaches, primarily for the paradigm of end-to-end based systems in autonomous vehicles. We classify the literature into three distinct categories: 1) Behavioural Cloning (BC), 2) Direct Policy Learning (DPL) and 3) Inverse Reinforcement Learning (IRL). For each of these categories, the current state-of-theart literature is comprehensively reviewed and summarised, with future directions of research identified to facilitate the development of imitation learning based systems for end-to-end autonomous vehicles. Due to the data-intensive nature of deep
learning techniques, currently available datasets and simulators for end-to-end autonomous driving are also reviewed.

Original language	English
Pages (from-to)	14128-14147
Number of pages	20
Journal	IEEE Transactions on Intelligent Transportation Systems
Volume	23
Issue number	9
Early online date	1 Feb 2022
DOIs	https://doi.org/10.1109/TITS.2022.3144867
Publication status	Published - Sept 2022

Bibliographical note

Funding Agency:
10.13039/100016335-Jaguar Land Rover
10.13039/501100000266-U.K. Engineering and Physical Sciences Research Council (EPSRC) (Grant Number: EP/N01300X/1)
jointly funded Towards Autonomy: Smart and Connected Control (TASCC) Program

Keywords

Intelligent vehicles
Autonomous vehicles
machine learning
autonomous systems
Learning
neural networks

Access to Document

10.1109/TITS.2022.3144867Licence: Unspecified

Mero_etal_IEEETOITS_A_Survey_On_AAM
© 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Accepted author manuscript, 4.46 MBLicence: Unspecified

Cite this

@article{c9a3a9b702744118a4daa6b5a84705ba,

title = "A Survey on Imitation Learning Techniques for End-to-End Autonomous Vehicles",

abstract = "The state-of-the-art decision and planning approaches for autonomous vehicles have moved away from manually designed systems, instead focusing on the utilisation of largescale datasets of expert demonstration via Imitation Learning(IL). In this paper, we present a comprehensive review of IL approaches, primarily for the paradigm of end-to-end based systems in autonomous vehicles. We classify the literature into three distinct categories: 1) Behavioural Cloning (BC), 2) Direct Policy Learning (DPL) and 3) Inverse Reinforcement Learning (IRL). For each of these categories, the current state-of-theart literature is comprehensively reviewed and summarised, with future directions of research identified to facilitate the development of imitation learning based systems for end-to-end autonomous vehicles. Due to the data-intensive nature of deeplearning techniques, currently available datasets and simulators for end-to-end autonomous driving are also reviewed. ",

keywords = "Intelligent vehicles, Autonomous vehicles, machine learning, autonomous systems, Learning, neural networks",

author = "{Le Mero}, Luc and Dewei Yi and Mehrdad Dianati and Alexandros Mouzakitis",

note = "Funding Agency: 10.13039/100016335-Jaguar Land Rover 10.13039/501100000266-U.K. Engineering and Physical Sciences Research Council (EPSRC) (Grant Number: EP/N01300X/1) jointly funded Towards Autonomy: Smart and Connected Control (TASCC) Program",

year = "2022",

month = sep,

doi = "10.1109/TITS.2022.3144867",

language = "English",

volume = "23",

pages = "14128--14147",

journal = "IEEE Transactions on Intelligent Transportation Systems",

issn = "1524-9050",

publisher = "IEEE Press",

number = "9",

}

TY - JOUR

T1 - A Survey on Imitation Learning Techniques for End-to-End Autonomous Vehicles

AU - Le Mero, Luc

AU - Yi, Dewei

AU - Dianati, Mehrdad

AU - Mouzakitis, Alexandros

N1 - Funding Agency: 10.13039/100016335-Jaguar Land Rover 10.13039/501100000266-U.K. Engineering and Physical Sciences Research Council (EPSRC) (Grant Number: EP/N01300X/1) jointly funded Towards Autonomy: Smart and Connected Control (TASCC) Program

PY - 2022/9

Y1 - 2022/9

N2 - The state-of-the-art decision and planning approaches for autonomous vehicles have moved away from manually designed systems, instead focusing on the utilisation of largescale datasets of expert demonstration via Imitation Learning(IL). In this paper, we present a comprehensive review of IL approaches, primarily for the paradigm of end-to-end based systems in autonomous vehicles. We classify the literature into three distinct categories: 1) Behavioural Cloning (BC), 2) Direct Policy Learning (DPL) and 3) Inverse Reinforcement Learning (IRL). For each of these categories, the current state-of-theart literature is comprehensively reviewed and summarised, with future directions of research identified to facilitate the development of imitation learning based systems for end-to-end autonomous vehicles. Due to the data-intensive nature of deeplearning techniques, currently available datasets and simulators for end-to-end autonomous driving are also reviewed.

AB - The state-of-the-art decision and planning approaches for autonomous vehicles have moved away from manually designed systems, instead focusing on the utilisation of largescale datasets of expert demonstration via Imitation Learning(IL). In this paper, we present a comprehensive review of IL approaches, primarily for the paradigm of end-to-end based systems in autonomous vehicles. We classify the literature into three distinct categories: 1) Behavioural Cloning (BC), 2) Direct Policy Learning (DPL) and 3) Inverse Reinforcement Learning (IRL). For each of these categories, the current state-of-theart literature is comprehensively reviewed and summarised, with future directions of research identified to facilitate the development of imitation learning based systems for end-to-end autonomous vehicles. Due to the data-intensive nature of deeplearning techniques, currently available datasets and simulators for end-to-end autonomous driving are also reviewed.

KW - Intelligent vehicles

KW - Autonomous vehicles

KW - machine learning

KW - autonomous systems

KW - Learning

KW - neural networks

U2 - 10.1109/TITS.2022.3144867

DO - 10.1109/TITS.2022.3144867

M3 - Article

SN - 1524-9050

VL - 23

SP - 14128

EP - 14147

JO - IEEE Transactions on Intelligent Transportation Systems

JF - IEEE Transactions on Intelligent Transportation Systems

IS - 9

ER -

A Survey on Imitation Learning Techniques for End-to-End Autonomous Vehicles

Abstract

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this