Deep–adaptive fuzzy predictive navigation framework for stable and intelligent mobile robot control

Autonomous mobile robot navigation in dynamic and uncertain environments demands control architectures that are simultaneously robust, adaptive, and provably stable. This work introduces a hierarchical predictive navigation framework that combines adaptive fuzzy decision-making with forward-looking motion optimization and explicit stability constraints. Simulation studies conducted in static, mixed, and dynamic environments demonstrate that the proposed framework achieves approximately 30–40 % higher average velocity, a 25–35 % reduction in traversal time, and 5–10 % lower energy consumption per unit distance compared with conventional fuzzy–potential field and optimization-tuned fuzzy navigation baselines. Across all evaluated scenarios, the robot maintained collision-free navigation and bounded control behavior. Selective human supervision was required in fewer than 10 % of operating intervals, reducing operator involvement while preserving safety. These results indicate that the proposed framework provides a quantitatively validated and interpretable alternative to existing fuzzy-based and predictive navigation approaches for autonomous mobile robots.