Sensor Adversarial Traits: Analyzing Robustness of 3D Object Detection Sensor Fusion Models

TL;DR

This paper investigates the robustness of sensor fusion models in autonomous vehicles against adversarial attacks, revealing vulnerabilities despite multimodal inputs and proposing potential defenses.

Contribution

First analysis of adversarial robustness in sensor fusion 3D object detection models, challenging assumptions about sensor redundancy mitigating attack risks.

Findings

Sensor fusion models are vulnerable to image-based adversarial attacks.

Adversarial attacks include disappearance, universal patch, and spoofing.

Potential defenses and recommendations for robust sensor fusion models.

Abstract

A critical aspect of autonomous vehicles (AVs) is the object detection stage, which is increasingly being performed with sensor fusion models: multimodal 3D object detection models which utilize both 2D RGB image data and 3D data from a LIDAR sensor as inputs. In this work, we perform the first study to analyze the robustness of a high-performance, open source sensor fusion model architecture towards adversarial attacks and challenge the popular belief that the use of additional sensors automatically mitigate the risk of adversarial attacks. We find that despite the use of a LIDAR sensor, the model is vulnerable to our purposefully crafted image-based adversarial attacks including disappearance, universal patch, and spoofing. After identifying the underlying reason, we explore some potential defenses and provide some recommendations for improved sensor fusion models.