FPGA-Based 3D Terrain Modeling and Disaster Simulation for Real-Time Response in Complex Environments

This paper presents a fully integrated, on-board FPGA-enabled perception-to-navigation framework for real-time disaster response using UAVs. Unlike conventional UAV systems that rely on off-board or cloud-based processing, the proposed architecture enables edge intelligence directly on the drone, significantly reducing latency and enabling operation in communication-constrained environments. The system combines FPGA-accelerated image processing, real-time obstacle detection, and dynamically adaptive path planning into a unified closed-loop pipeline. Experimental results demonstrate up to 85% reduction in processing latency compared to GPU-based approaches, enabling near-instantaneous decision-making. This work shifts UAVs from passive sensing platforms to autonomous decision-making agents, providing a scalable solution for time-critical disaster management.