Abstract: Autonomous drone navigation in dynamic and uncertain environments remains a major research challenge due to limitations in robustness, adaptability, and scalability of existing solutions. This paper introduces a Cognitive AI-driven Navigation Framework that integrates bio-inspired learning models with adaptive cognitive reasoning for efficient trajectory planning and real-time decision-making. The proposed system leverages swarm-inspired optimization and biologically motivated learning principles to enhance path efficiency, collision avoidance, and mission success rates. A systematic literature review (SLR) guided the framework design, highlighting gaps in computational efficiency and scalability across state-of-the-art approaches. The model was evaluated through simulation in complex obstacle-rich environments and benchmarked against deep reinforcement learning (DRL), PSO-only optimization, and hybrid CNN-RL methods. Results demonstrate a 95.3% mission success rate, 28.9 ms decision latency, and 92.6% robustness, outperforming baseline methods in both single-drone and multi-drone scenarios. Discussion emphasizes the significance of combining cognitive adaptability with bio-inspired mechanisms, offering a pathway to real-time, scalable, and resilient UAV operations. This work contributes to advancing UAV navigation towards next-generation autonomous systems for applications in disaster response, urban mobility, and cooperative surveillance.

Keywords: Autonomous Drones, AI Navigation, Bio-Inspired Learning, Cognitive Systems, Swarm Intelligence, Adaptive Path Planning, UAV Scalability, Real-Time Decision Making

Pages: 258-267

WSEAS Transactions on