Adversarial VR: An Open-Source Testbed for Evaluating Adversarial Robustness of VR Cybersickness Detection and Mitigation

TL;DR

This paper presents Adversarial-VR, an open-source real-time VR testbed that evaluates the robustness of DL-based cybersickness detection and mitigation methods against adversarial attacks, highlighting vulnerabilities and providing a tool for improvement.

Contribution

It introduces Adversarial-VR, a novel open-source testbed integrating state-of-the-art models and adversarial attacks to assess and improve the robustness of cybersickness detection systems in VR.

Findings

Adversarial attacks significantly reduce model accuracy.

C&W attack causes a 5.94x decrease in accuracy.

The testbed is open-source for community use.

Abstract

Deep learning (DL)-based automated cybersickness detection methods, along with adaptive mitigation techniques, can enhance user comfort and interaction. However, recent studies show that these DL-based systems are susceptible to adversarial attacks; small perturbations to sensor inputs can degrade model performance, trigger incorrect mitigation, and disrupt the user's immersive experience (UIX). Additionally, there is a lack of dedicated open-source testbeds that evaluate the robustness of these systems under adversarial conditions, limiting the ability to assess their real-world effectiveness. To address this gap, this paper introduces Adversarial-VR, a novel real-time VR testbed for evaluating DL-based cybersickness detection and mitigation strategies under adversarial conditions. Developed in Unity, the testbed integrates two state-of-the-art (SOTA) DL models: DeepTCN and Transformer, which are trained on the open-source MazeSick dataset, for real-time cybersickness severity detection and applies a dynamic visual tunneling mechanism that adjusts the field-of-view based on model outputs. To assess robustness, we incorporate three SOTA adversarial attacks: MI-FGSM, PGD, and C&W, which successfully prevent cybersickness mitigation by fooling DL-based cybersickness models' outcomes. We implement these attacks using a testbed with a custom-built VR Maze simulation and an HTC Vive Pro Eye headset, and we open-source our implementation for widespread adoption by VR developers and researchers. Results show that these adversarial attacks are capable of successfully fooling the system. For instance, the C&W attack results in a $5.94x decrease in accuracy for the Transformer-based cybersickness model compared to the accuracy without the attack.