AI RESEARCH
ORN-CBF: Learning Observation-conditioned Residual Neural Control Barrier Functions via Hypernetworks
arXiv CS.LG
•
ArXi:2509.16614v3 Announce Type: replace-cross Control barrier functions (CBFs) have been nstrated 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.