AI RESEARCH
Decision-Driven Semantic Object Exploration for Legged Robots via Confidence-Calibrated Perception and Topological Subgoal Selection
arXiv CS.CV
•
ArXi:2509.20739v2 Announce Type: replace-cross Conventional navigation pipelines for legged robots remain largely geometry-centric, relying on dense SLAM representations that are fragile under rapid motion and offer limited for semantic decision making in open-world exploration. In this work, we focus on decision-driven semantic object exploration, where the primary challenge is not map consistency but how noisy and heterogeneous semantic observations can be transformed into stable and executable exploration decisions.