用神经网络学习可观测条件下的安全屏障函数，提升机器人在复杂环境中的安全性。
ORN-CBF: Learning Observation-conditioned Residual Neural Control Barrier Functions via Hypernetworks
- 基于哈密顿-雅可比可达性分析，设计可观测条件下的神经屏障函数
- 在仿真与硬件实验中，安全成功率显著优于基线方法
- 适用于部分可观测场景，适合需高安全性的自主系统研发
控制屏障函数（CBFs）已被证明是保障自主系统安全控制的有效方法。尽管CBFs部署简单，但其设计仍具挑战性，促使了基于学习的方法发展。然而，子最优安全集、部分可观测环境下的适用性以及缺乏严格的安全保证等问题依然存在。本文提出基于哈密顿-雅可比（HJ）可达性分析的观测条件神经CBFs，近似恢复最大安全集。利用HJ值函数的某些数学特性，确保预测安全集永不与已观测到的故障集相交。此外，采用超网络架构，特别适合设计观测条件下的安全过滤器。所提方法在地面机器人和四旋翼无人机的仿真与硬件实验中进行了验证，结果表明相比基线方法，成功率达更高且对域外环境具有更强泛化能力。
原文摘要 · Abstract (English)
Control barrier functions (CBFs) have been demonstrated as an effective method for safety-critical control of autonomous systems. Although CBFs are simple to deploy, their design remains challenging, motivating the development of learning-based approaches. Yet, issues such as suboptimal safe sets, applicability in partially observable environments, and lack of rigorous safety guarantees persist. In this work, we propose observation-conditioned neural CBFs based on Hamilton-Jacobi (HJ) reachability analysis, which approximately recover the maximal safe sets. We exploit certain mathematical properties of the HJ value function, ensuring that the predicted safe set never intersects with the observed failure set. Moreover, we leverage a hypernetwork-based architecture that is particularly suitable for the design of observation-conditioned safety filters. The proposed method is examined both in simulation and hardware experiments for a ground robot and a quadcopter. The results show improved success rates and generalization to out-of-domain environments compared to the baselines.
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。