Concrete Problems for Autonomous Vehicle Safety: Advantages of Bayesian Deep Learning

TLDR

Autonomous vehicle software is built as a pipeline of components whose erroneous outputs can propagate downstream, so safety must consider the ultimate effects and passengers must also feel secure to trust the system. The study aims to investigate safety, interpretability, and compliance in autonomous vehicles, discuss open research challenges, and propose concrete evaluation metrics, example problems, and potential solutions. The authors propose quantifying component output uncertainties and propagating them through the pipeline to improve safety, explaining observations and decisions to build passenger reassurance, and maintaining passenger control to ensure compliance.

Abstract

Autonomous vehicle (AV) software is typically composed of a pipeline of individual components, linking sensor inputs to motor outputs. Erroneous component outputs propagate downstream, hence safe AV software must consider the ultimate effect of each component’s errors. Further, improving safety alone is not sufficient. Passengers must also feel safe to trust and use AV systems. To address such concerns, we investigate three under-explored themes for AV research: safety, interpretability, and compliance. Safety can be improved by quantifying the uncertainties of component outputs and propagating them forward through the pipeline. Interpretability is concerned with explaining what the AV observes and why it makes the decisions it does, building reassurance with the passenger. Compliance refers to maintaining some control for the passenger. We discuss open challenges for research within these themes. We highlight the need for concrete evaluation metrics, propose example problems, and highlight possible solutions.