• Proposes a lightweight, interpretable mapless navigation method using only 2D LiDAR, mimicking human strategies to escape dead ends without requiring prior maps.
• Effectively suppresses oscillations and enables autonomous recovery from local minimum traps, addressing key challenges in mapless navigation.
• Achieves map-like performance without a map, reducing average path length by 50.51% compared to the classical Bug2 algorithm.
• Increases path length by only 17.57% compared to map-based navigation, demonstrating robustness across diverse environments.