Safe Reinforcement Learning with Model Uncertainty Estimates

TL;DR

This paper introduces a safe reinforcement learning approach that uses computationally efficient model uncertainty estimates to improve robustness and safety in autonomous navigation around pedestrians, especially in unseen scenarios.

Contribution

It integrates MC-Dropout and Bootstrapping for uncertainty estimation into reinforcement learning, enhancing safety and robustness in autonomous systems.

Findings

More robust to novel observations

Takes safer actions around unseen pedestrians

Outperforms uncertainty-unaware baseline

Abstract

Many current autonomous systems are being designed with a strong reliance on black box predictions from deep neural networks (DNNs). However, DNNs tend to be overconfident in predictions on unseen data and can give unpredictable results for far-from-distribution test data. The importance of predictions that are robust to this distributional shift is evident for safety-critical applications, such as collision avoidance around pedestrians. Measures of model uncertainty can be used to identify unseen data, but the state-of-the-art extraction methods such as Bayesian neural networks are mostly intractable to compute. This paper uses MC-Dropout and Bootstrapping to give computationally tractable and parallelizable uncertainty estimates. The methods are embedded in a Safe Reinforcement Learning framework to form uncertainty-aware navigation around pedestrians. The result is a collision avoidance policy that knows what it does not know and cautiously avoids pedestrians that exhibit unseen behavior. The policy is demonstrated in simulation to be more robust to novel observations and take safer actions than an uncertainty-unaware baseline.