DriveGuard: Robustification of Automated Driving Systems with Deep Spatio-Temporal Convolutional Autoencoder

TL;DR

DriveGuard is a lightweight spatio-temporal autoencoder designed to enhance the robustness of autonomous vehicle perception systems against noisy and adverse visual conditions, reducing performance degradation.

Contribution

We introduce DriveGuard, a novel autoencoder architecture that improves perception robustness without retraining existing models, leveraging spatio-temporal data and multi-component loss functions.

Findings

Significantly increased robustness against adverse image effects.

Performance within 5-6% of original model on clean images.

Effective across diverse real and synthetic datasets.

Abstract

Autonomous vehicles increasingly rely on cameras to provide the input for perception and scene understanding and the ability of these models to classify their environment and objects, under adverse conditions and image noise is crucial. When the input is, either unintentionally or through targeted attacks, deteriorated, the reliability of autonomous vehicle is compromised. In order to mitigate such phenomena, we propose DriveGuard, a lightweight spatio-temporal autoencoder, as a solution to robustify the image segmentation process for autonomous vehicles. By first processing camera images with DriveGuard, we offer a more universal solution than having to re-train each perception model with noisy input. We explore the space of different autoencoder architectures and evaluate them on a diverse dataset created with real and synthetic images demonstrating that by exploiting spatio-temporal information combined with multi-component loss we significantly increase robustness against adverse image effects reaching within 5-6% of that of the original model on clean images.