V2XP-ASG: Generating Adversarial Scenes for Vehicle-to-Everything Perception
This addresses safety-critical evaluation for autonomous vehicles and infrastructure before deployment, though it is incremental as it builds on existing adversarial generation methods for a specific domain.
The paper tackles the problem of evaluating and improving Vehicle-to-Everything (V2X) perception systems under challenging traffic scenarios by proposing V2XP-ASG, an adversarial scene generator that produces realistic scenes, resulting in a 12.3% accuracy improvement on challenging scenes and 4% on normal scenes when used for training.
Recent advancements in Vehicle-to-Everything communication technology have enabled autonomous vehicles to share sensory information to obtain better perception performance. With the rapid growth of autonomous vehicles and intelligent infrastructure, the V2X perception systems will soon be deployed at scale, which raises a safety-critical question: \textit{how can we evaluate and improve its performance under challenging traffic scenarios before the real-world deployment?} Collecting diverse large-scale real-world test scenes seems to be the most straightforward solution, but it is expensive and time-consuming, and the collections can only cover limited scenarios. To this end, we propose the first open adversarial scene generator V2XP-ASG that can produce realistic, challenging scenes for modern LiDAR-based multi-agent perception systems. V2XP-ASG learns to construct an adversarial collaboration graph and simultaneously perturb multiple agents' poses in an adversarial and plausible manner. The experiments demonstrate that V2XP-ASG can effectively identify challenging scenes for a large range of V2X perception systems. Meanwhile, by training on the limited number of generated challenging scenes, the accuracy of V2X perception systems can be further improved by 12.3\% on challenging and 4\% on normal scenes. Our code will be released at https://github.com/XHwind/V2XP-ASG.