Simulation-based reinforcement learning for real-world autonomous driving

11/29/2019

∙

We use synthetic data and a reinforcement learning algorithm to train a system controlling a full-size real-world vehicle in a number of restricted driving scenarios. The driving policy uses RGB images as input. We analyze how design decisions about perception, control and training impact the real-world performance.

READ FULL TEXT

Simulation-based reinforcement learning for real-world autonomous driving

Virtual to Real Reinforcement Learning for Autonomous Driving

Perception as prediction using general value functions in autonomous driving applications

Scaling Planning for Automated Driving using Simplistic Synthetic Data

Sim-to-Real via Sim-to-Seg: End-to-end Off-road Autonomous Driving Without Real Data

Data generation using simulation technology to improve perception mechanism of autonomous vehicles

Data and Knowledge for Overtaking Scenarios in Autonomous Driving

PCG-based Static Underground Garage Scenario Generation

Simulation-based reinforcement learning for real-world autonomous driving

Related Research

Virtual to Real Reinforcement Learning for Autonomous Driving

Perception as prediction using general value functions in autonomous driving applications

Scaling Planning for Automated Driving using Simplistic Synthetic Data

Sim-to-Real via Sim-to-Seg: End-to-end Off-road Autonomous Driving Without Real Data

Data generation using simulation technology to improve perception mechanism of autonomous vehicles

Data and Knowledge for Overtaking Scenarios in Autonomous Driving

PCG-based Static Underground Garage Scenario Generation