Safe Reinforcement Learning on Autonomous Vehicles

There have been numerous advances in reinforcement learning, but the typically unconstrained exploration of the learning process prevents the adoption of these methods in many safety critical applications. Recent work in safe reinforcement learning uses idealized models to achieve their guarantees, but these models do not easily accommodate the stochasticity or high-dimensionality of real world systems. We investigate how prediction provides a general and intuitive framework to constraint exploration, and show how it can be used to safely learn intersection handling behaviors on an autonomous vehicle.

Authors

• David Isele
• Alireza Nakhaei
• Kikuo Fujimura

Keywords

• Safety
• Autonomous vehicles
• Trajectory
• Games
• Pipelines
• Noise measurement
• Standards
• Safe Reinforcement Learning
• Learning Process
• General Framework
• Safety-critical
• Safety-critical Applications
• Collision
• Time Step
• Minimum Distance
• Physical System
• Single Activity
• Traffic Flow
• Number Of Agents
• Multiple Agents
• Reward Function
• Markov Decision Process
• Safety Activity
• Policy Learning
• Safety Policies
• Safety Constraints
• Fixed Time Window
• Safety Guarantees
• Prediction Module
• Chebyshev’s Inequality
• Real Vehicle