Dynamic systems and optimal control

Navigation of unmanned autonomous vehicles based on registration and transmission of observations of the surrounding landscape

Alexander B. Miller, Karen V. Stepanyan, Boris M. Miller, Alexey K. Popov

A. A. Kharkevich Institute for Information Transmission Problems RAS

Abstract: Navigation by observations of current video images is one of the most promising means for control of unmanned vehicles in conditions of limited access to signals from global positioning satellite systems. In the autonomous motion mode, simply adding a video camera in the absence of recognition and interpretation tools to the inertial navigation system does not provide a significant effect. Therefore, extracting navigation information from a sequence of images plays a key role. An important characteristic of the observed images is the evolution of the generated information flow, the processing of which allows us to estimate the absolute linear and angular velocities of the device, which is a significant expansion of the capabilities of the control system.

Keywords: navigation, unmanned vehicle, image sequence, velocity estimation.

UDC: 519.7

MSC: 93-10, 93C41

Received: 20.09.2024
Revised: 29.11.2024
Accepted: 06.12.2024

DOI: 10.26516/1997-7670.2025.53.18