Manipulation Facing Threats: Evaluating Physical Vulnerabilities in End-to-End Vision Language Action Models
This addresses safety concerns for robotic manipulation tasks in open-vocabulary scenarios, but it is incremental as it builds on existing safety research for Multimodal Large Language Models.
The paper tackles the problem of evaluating physical vulnerabilities in Vision Language Action Models (VLAMs) for robotic manipulation, proposing a pipeline to assess robustness against threats like adversarial patches and showing performance fluctuations under attack.
Recently, driven by advancements in Multimodal Large Language Models (MLLMs), Vision Language Action Models (VLAMs) are being proposed to achieve better performance in open-vocabulary scenarios for robotic manipulation tasks. Since manipulation tasks involve direct interaction with the physical world, ensuring robustness and safety during the execution of this task is always a very critical issue. In this paper, by synthesizing current safety research on MLLMs and the specific application scenarios of the manipulation task in the physical world, we comprehensively evaluate VLAMs in the face of potential physical threats. Specifically, we propose the Physical Vulnerability Evaluating Pipeline (PVEP) that can incorporate as many visual modal physical threats as possible for evaluating the physical robustness of VLAMs. The physical threats in PVEP specifically include Out-of-Distribution, Typography-based Visual Prompt, and Adversarial Patch Attacks. By comparing the performance fluctuations of VLAMs before and after being attacked, we provide generalizable \textbf{\textit{Analyses}} of how VLAMs respond to different physical threats.