为自动驾驶汽车设计安全可靠的协同感知机制
CONClave -- Secure and Robust Cooperative Perception for CAVs Using Authenticated Consensus and Trust Scoring
- 通过认证、共识与信任评分三者联动保障数据可信
- 可快速检测微小感知错误，显著提升系统鲁棒性
- 适合高安全需求的车联网场景，部署开销极小
联网自动驾驶车辆在协同感知应用中共享感知数据，具有提升交通安全与交通效率的巨大潜力。然而，恶意行为和非故意错误可能引发事故。以往工作多针对特定场景下的单一安全或可靠性问题，未能综合应对各类错误。本文提出CONClave，一种紧密耦合的认证、共识与信任评分机制，全面保障自动驾驶车辆协同感知的安全性与可靠性。该机制利用步骤间的流水线特性，使故障检测更快且计算开销更低。实验表明，CONClave能有效防止安全漏洞，检测出相对微小的感知误差，并显著提升协同感知的鲁棒性与准确性，同时仅引入极小额外开销。
原文摘要 · Abstract (English)
Connected Autonomous Vehicles have great potential to improve automobile safety and traffic flow, especially in cooperative applications where perception data is shared between vehicles. However, this cooperation must be secured from malicious intent and unintentional errors that could cause accidents. Previous works typically address singular security or reliability issues for cooperative driving in specific scenarios rather than the set of errors together. In this paper, we propose CONClave, a tightly coupled authentication, consensus, and trust scoring mechanism that provides comprehensive security and reliability for cooperative perception in autonomous vehicles. CONClave benefits from the pipelined nature of the steps such that faults can be detected significantly faster and with less compute. Overall, CONClave shows huge promise in preventing security flaws, detecting even relatively minor sensing faults, and increasing the robustness and accuracy of cooperative perception in CAVs while adding minimal overhead.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。