Deep Reinforcement Learning for Autonomous Driving: A Survey (2020-02-02T00:00:00.000000Z)

TL;DR

This review summarises deep reinforcement learning (DRL) algorithms and provides a taxonomy of automated driving tasks where (D)RL methods have been employed, while addressing key computational challenges in real world deployment of autonomous driving agents.

Abstract

With the development of deep representation learning, the domain of reinforcement learning (RL) has become a powerful learning framework now capable of learning complex policies in high dimensional environments. This review summarises deep reinforcement learning (DRL) algorithms and provides a taxonomy of automated driving tasks where (D)RL methods have been employed, while addressing key computational challenges in real world deployment of autonomous driving agents. It also delineates adjacent domains such as behavior cloning, imitation learning, inverse reinforcement learning that are related but are not classical RL algorithms. The role of simulators in training agents, methods to validate, test and robustify existing solutions in RL are discussed.

Authors

B. R. Kiran

1 papers

Ibrahim Sobh

2 papers

V. Talpaert

1 papers

TL;DR

Abstract

Authors

References159 items

Deep Multi Agent Reinforcement Learning for Autonomous Driving

Distributed Multiagent Coordinated Learning for Autonomous Driving in Highways Based on Dynamic Coordination Graphs

A Survey of Deep Learning Applications to Autonomous Vehicle Control

Multi-Agent Connected Autonomous Driving using Deep Reinforcement Learning

INTERACTION Dataset: An INTERnational, Adversarial and Cooperative moTION Dataset in Interactive Driving Scenarios with Semantic Maps

Multi-objective multi-agent decision making: a utility-based analysis and survey

rlpyt: A Research Code Base for Deep Reinforcement Learning in PyTorch

Conditional Vehicle Trajectories Prediction in CARLA Urban Environment

Behaviour Suite for Reinforcement Learning

RGB and LiDAR fusion based 3D Semantic Segmentation for Autonomous Driving

SoilingNet: Soiling Detection on Automotive Surround-View Cameras

WoodScape: A Multi-Task, Multi-Camera Fisheye Dataset for Autonomous Driving

Generating Adversarial Driving Scenarios in High-Fidelity Simulators

Model-free Deep Reinforcement Learning for Urban Autonomous Driving

Failure-Scenario Maker for Rule-Based Agent using Multi-agent Adversarial Reinforcement Learning and its Application to Autonomous Driving

NeurAll: Towards a Unified Visual Perception Model for Automated Driving

A Systematic Review of Perception System and Simulators for Autonomous Vehicles Research

Yes, we GAN: Applying Adversarial Techniques for Autonomous Driving

Exploring applications of deep reinforcement learning for real-world autonomous driving systems

Soft Actor-Critic Algorithms and Applications

Learning to Drive from Simulation without Real World Labels

ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst

Deep Reinforcement Learning for Autonomous Driving

Urban Driving with Multi-Objective Deep Reinforcement Learning

Reinforcement Learning with A* and a Deep Heuristic

MODNet: Motion and Appearance based Moving Object Detection Network for Autonomous Driving

Microscopic Traffic Simulation using SUMO

A Survey of State-Action Representations for Autonomous Driving

Dynamic Programming

Decoupling feature extraction from policy learning: assessing benefits of state representation learning in goal based robotics

Unity: A General Platform for Intelligent Agents

Recurrent World Models Facilitate Policy Evolution

Large-Scale Study of Curiosity-Driven Learning

Policy Optimization with Demonstrations

Learning to Drive in a Day

A Tour of Reinforcement Learning: The View from Continuous Control

Visual SLAM for Automated Driving: Exploring the Applications of Deep Learning

Planning and Decision-Making for Autonomous Vehicles

Navigating Occluded Intersections with Autonomous Vehicles Using Deep Reinforcement Learning

A Tempt to Unify Heterogeneous Driving Databases using Traffic Primitives

A Reinforcement Learning Based Approach for Automated Lane Change Maneuvers

Simple random search provides a competitive approach to reinforcement learning

Monocular Fisheye Camera Depth Estimation Using Sparse LiDAR Supervision

On First-Order Meta-Learning Algorithms

State Representation Learning for Control: An Overview

Diversity-Driven Exploration Strategy for Deep Reinforcement Learning

VR-Goggles for Robots: Real-to-Sim Domain Adaptation for Visual Control

Transferring Autonomous Driving Knowledge on Simulated and Real Intersections

Policy invariance under reward transformations for multi-objective reinforcement learning

Composing Meta-Policies for Autonomous Driving Using Hierarchical Deep Reinforcement Learning

Mastering the game of Go without human knowledge

Sim-to-Real Transfer of Robotic Control with Dynamics Randomization

CARLA: An Open Urban Driving Simulator

Flow: Architecture and Benchmarking for Reinforcement Learning in Traffic Control

Continuous Adaptation via Meta-Learning in Nonstationary and Competitive Environments

Deep Reinforcement Learning that Matters

Survival-Oriented Reinforcement Learning Model: An Effcient and Robust Deep Reinforcement Learning Algorithm for Autonomous Driving Problem

Formulation of deep reinforcement learning architecture toward autonomous driving for on-ramp merge

Proximal Policy Optimization Algorithms

Deep semantic segmentation for automated driving: Taxonomy, roadmap and challenges

Curiosity-Driven Exploration by Self-Supervised Prediction

AirSim: High-Fidelity Visual and Physical Simulation for Autonomous Vehicles

A Theoretical and Empirical Analysis of Reward Transformations in Multi-Objective Stochastic Games

Explaining How a Deep Neural Network Trained with End-to-End Learning Steers a Car

Virtual to Real Reinforcement Learning for Autonomous Driving

Deep Q-learning From Demonstrations

Deep Reinforcement Learning framework for Autonomous Driving

Recurrent Environment Simulators

Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

Reinforcement Learning with Deep Energy-Based Policies

Query-Efficient Imitation Learning for End-to-End Simulated Driving

Autonomous braking system via deep reinforcement learning

Unsupervised Pixel-Level Domain Adaptation with Generative Adversarial Networks

Learning to Drive using Inverse Reinforcement Learning and Deep Q-Networks

End-to-End Learning of Driving Models from Large-Scale Video Datasets

Combining Deep Reinforcement Learning and Safety Based Control for Autonomous Driving