СИСТЕМА ВИЗНАЧЕННЯ КООРДИНАТ АВТОНОМНОГО ПОЛЬОТУ БПЛА ЗА ДОПОМОГОЮ ШТУЧНОГО ІНТЕЛЕКТУ

Abstract

The master’s thesis is dedicated to the research and development of a system for determining the spatial coordinates and trajectory of UAV motion based on video stream data and motion sensors using computer vision and artificial intelligence methods. The goal of the work is to study the subject area of autonomous UAV navigation, formalize the problem of determining coordinates and trajectory in conditions of limited or absent satellite navigation, and develop the algorithmic and software foundations of the corresponding system. The research focuses on the analysis of visual odometry methods, marker-based localization, and sensor fusion approaches, as well as on building a mathematical model of UAV motion and an algorithm for estimating spatial position over time. During the course of the work, system requirements were analyzed, input/output data formats were defined, and a structural and functional model of the software product was developed, ensuring visual data processing, trajectory formation, and result visualization. The outcome of the work is a UAV coordinate and trajectory determination system that generates structured processing results in the form of coordinate time series. The developed system has potential for further development, including integration of inertial sensors, improvement of trajectory filtering methods, and application of more advanced artificial intelligence models.