Generating Transferable Adversarial Simulation Scenarios for Self-Driving via Neural Rendering

TL;DR

This paper introduces a neural rendering-based method to generate adversarial scenarios for self-driving cars, enabling the evaluation of safety and robustness by transferring simulated failures to real-world environments.

Contribution

It presents a novel neural rendering approach to create transferable adversarial scenarios for autonomous driving policies, improving safety testing.

Findings

Adversarial scenarios can be generated via neural rendering.

Transferability of adversarial scenarios from surrogate to real scenes.

Effective in both simulated and real environments.

Abstract

Self-driving software pipelines include components that are learned from a significant number of training examples, yet it remains challenging to evaluate the overall system's safety and generalization performance. Together with scaling up the real-world deployment of autonomous vehicles, it is of critical importance to automatically find simulation scenarios where the driving policies will fail. We propose a method that efficiently generates adversarial simulation scenarios for autonomous driving by solving an optimal control problem that aims to maximally perturb the policy from its nominal trajectory. Given an image-based driving policy, we show that we can inject new objects in a neural rendering representation of the deployment scene, and optimize their texture in order to generate adversarial sensor inputs to the policy. We demonstrate that adversarial scenarios discovered purely in the neural renderer (surrogate scene) can often be successfully transferred to the deployment scene, without further optimization. We demonstrate this transfer occurs both in simulated and real environments, provided the learned surrogate scene is sufficiently close to the deployment scene.