Intelligent drone navigation using a hybrid neighborhood dragonfly algorithm and probability fuzzy logic (NDA−PFL) in uncertain environments

Purpose This paper aims to develop efficient path-planning technique for a single and multiple drones navigating in complex indoor environments filled with obstacles. With the increasing presence of drones in different domains, there should be a way to ensure that they move safely and efficiently across uncertain environments. To achieve this, a hybrid scheme that integrates the neighborhood dragonfly algorithm and probability fuzzy logic (NDA–PFL) to obtain a smooth, collision-free path using less time and energy is introduced. The integration of both techniques allows drones to adapt their path dynamically, converting complex, uncertain environments into clear navigation decisions. Design/methodology/approach A hybrid NDA–PFL controller is proposed, combining global search (NDA) with local adaptability (PFL). It is tested on single/multiple unmanned aerial vehicles (UAVs) in static and dynamic environments. Simulations and real-time experiments were conducted, with parameter tuning and sensitivity analysis ensuring optimal performance. Findings The NDA–PFL outperforms standalone NDA and PFL. In static environments, it shows only 7.25% path length deviation (versus PFL’s 8.55%). In multi-UAV trials, it minimizes travel time deviation to 7.2% and maintains under 7.4% deviation in dynamic cases, ensuring superior robustness and reproducibility. Originality/value This research introduces a novel hybrid NDA–PFL framework for UAV navigation. Unlike existing metaheuristic–fuzzy hybrids, NDA–PFL uniquely combines neighborhood-guided global optimization with probabilistic fuzzy reactivity, enabling adaptive decision-making under uncertainty. The framework is among the first to be experimentally validated in GPS-denied indoor environments, demonstrating reliable and collision-free navigation across diverse scenarios.