Malysheva J. Aircraft integrated navigation system with optical sensors

An algorithm for aircraft roll and pitch values estimation based on optical horizon line recognition is developed. The algorithm makes use only of machine vision techniques. It has required accuracy of horizon line detection, in particular due to circular mask usage for image distorted areas elimination and processing rate. Such performances allow the algorithm to be used in onboard computers as well as obtaining of processed information in real time. The optical horizon sensor is designed based on onboard camera and computer which is able to calculate roll and pitch angles with accuracy ?0,5°. In contrast with available models such a sensor does not require additional expenses. An algorithm defined the framework of optical system for distance estimation to the obstacles is developed, which is based on stereovision principals. Such a system allows to increase the safety of an UAV during the flight. It is shown that the usage of optical horizon sensor as a part of integrated navigation system built with usage of microelectromechanical sensors proposes an opportunity of accuracy enhancement in tree times for pitch and roll estimation. This can be crucial factor in UAV navigation system design in existing conditions of cost and weight limitation