The field of autonomous navigation is moving towards more efficient and safe exploration of cluttered environments. Recent developments have focused on integrating reinforcement learning with safety mechanisms and adaptive confidence updates to improve navigation reliability. Notable advancements include the use of graph neural networks, potential fields, and change-aware sensing representations to enable autonomous systems to navigate complex dynamic environments. Noteworthy papers include: ClutterNav, which presents a novel decision-making framework for efficient 3D clutter removal. CUTE-Planner, which proposes a confidence-aware uneven terrain exploration planner for planetary exploration robots. Platform-Agnostic Reinforcement Learning Framework, which integrates a graph neural network-based policy with a safety filter for safe exploration of cluttered environments. Flow-Aided Flight, which empowers reinforcement learning with single LiDAR sensing to realize an autonomous flight system directly from point to motion.