Analysis of Modern Approaches to Autonomous UAV Navigation

Abstract

The development of autonomous navigation methods for unmanned aerial vehicles (UAVs) is driven by the need to ensure reliable positioning, orientation, and trajectory planning without dependence on external infrastructure. This task is particularly critical in environments with partial or complete uncertainty—such as indoor, underground, or GPS-denied areas. Over the last decade, numerous approaches to autonomous UAV navigation have been proposed and implemented, ranging from traditional geometric and algorithmic techniques to intelligent systems and sensor fusion architectures. The objective is t analyze and systematize modern approaches to autonomous UAV navigation and to highlight the potential of correlation-extreme navigation methods in uncertain environments. The paper presents a structured overview of state-of-the-art navigation methods, their classification by operational principles, and their strengths and limitations in different use cases. Special attention is paid to correlation-extreme navigation methods, which rely on optimizing similarity functions between sensory data and reference maps. These methods demonstrate high potential for fully autonomous navigation in GPS-denied or signal-jammed environments. A comparative matrix highlights the trade-offs between accuracy, computational cost, robustness, and flexibility for each method. Correlation-extreme navigation is a promising direction for achieving robust autonomy in UAV systems operating in uncertain and complex conditions.