SWARM-INTELLIGENT BIOINSPIRED DRONE SYSTEMS FOR AUTONOMOUS NAVIGATION IN GPS-DENIED AND CLIMATE-CHALLENGED TERRAINS: A MULTI-AGENT DEEP REINFORCEMENT LEARNING FRAMEWORK WITH REAL-TIME ENVIRONMENTAL PERCEPTION AND ENERGY-AWARE PATH OPTIMIZATION

This paper presents a novel swarm-intelligent, bioinspired drone system designed to operate in GPS-denied and climate-challenged terrains using a multi-agent deep reinforcement learning (DRL) framework. Drawing inspiration from collective animal behavior, the proposed architecture integrates real-time environmental perception, energy-aware path optimization, and decentralized coordination. It leverages Multi-Agent Proximal Policy Optimization (MAPPO) and neural sensory fusion to enable robust and adaptive navigation strategies. Experimental simulations across varied terrains (forest, desert, mountain) demonstrate enhanced energy efficiency and increased mission success rates compared to conventional algorithms. The system’s resilience, agility, and autonomy provide a transformative capability for disaster response, ecological monitoring, and military reconnaissance missions.