Drone's Attitude Estimation in Corridor-Like Environments

Abstract

In this study, we suggest an attitude estimation algorithm for drones flying indoors. In particular, we consider a corridor-like environment and adapt ideas from the aerospace field, where algorithms were developed for satellite's attitude estimation. Many algorithms can be found that estimate satellite's attitude, based on rate gyroscopes and a sensor called, star-tracker. The star-tracker identifies celestial objects, and by that, determines their directions compared to the satellite. Using star maps, the same celestial objects directions, compared to the earth, is known. By comparing the celestial objects directions in the satellite frame and in the earth frame, the attitude of the satellite can be estimated. Complementing the star-tracker with rate gyroscopes provides smooth attitude estimation, while also compensating for the rate gyroscope's drift. The novelty in this paper comes from the implementation of the star-tracker method on a drone in a corridor-like environment, and by finding features, which replace the celestial objects used by a star-tracker.