Spatial-Semantic Reasoning using Large Language Models for Efficient UAV Search Operations
RoboticsArtificial Intelligence
Summary
The authors created a system that helps drones find objects faster by understanding instructions given in plain language. They use a large language model to make smart guesses about where the object might be based on what the drone sees and where it is. The drone plans smooth flying paths in real time, updating its search as it explores. Tests show their method quickly finds objects without losing accuracy, working both in simulations and real life.
Unmanned Aerial Vehicles (UAVs)Object Goal Navigation (ObjectNav)Large Language Model (LLM)semantic reasoningobject detection3D spatial mappingtrajectory planningpolynomial spline interpolationreal-time navigation
Authors
Marin Maletic, Marijana Peti, Tamara Petrovic, Stjepan Bogdan
Abstract
We present a real-time semantic navigation framework for Unmanned Aerial Vehicles (UAVs) focused on improving time efficiency in the Object Goal Navigation (ObjectNav) task. Central to our approach is a Large Language Model (LLM) that interprets user-provided natural language instructions and performs semantic reasoning over detected objects and spatial context to prioritize high-probability search regions. The system combines real-time object detection, 3D spatial mapping, and polynomial spline interpolation for smooth and feasible UAV trajectory planning. Unlike prior methods that rely on offline reasoning or simulator-constrained action spaces, our framework can operate in real time, continuously updating semantic relevance based on new observations. Experiments in both simulated and real-world settings demonstrate reductions in mission duration while maintaining high search accuracy, underscoring the effectiveness of LLM-guided reasoning for time- efficient UAV-based ObjectNav.