TY - JOUR

T1 - Using deep learning for visual navigation of drone with respect to 3D ground objects

AU - Kupervasser, Oleg

AU - Kutomanov, Hennadii

AU - Levi, Ori

AU - Pukshansky, Vladislav

AU - Yavich, Roman

N1 - Publisher Copyright:
© 2020 by the authors. Licensee MDPI, Basel, Switzerland.

PY - 2020/12

Y1 - 2020/12

N2 - In the paper, visual navigation of a drone is considered. The drone navigation problem consists of two parts. The first part is finding the real position and orientation of the drone. The second part is finding the difference between desirable and real position and orientation of the drone and creation of the correspondent control signal for decreasing the difference. For the first part of the drone navigation problem, the paper presents a method for determining the coordinates of the drone camera with respect to known three-dimensional (3D) ground objects using deep learning. The algorithm has two stages. It causes the algorithm to be easy for interpretation by artificial neural network (ANN) and consequently increases its accuracy. At the first stage, we use the first ANN to find coordinates of the object origin projection. At the second stage, we use the second ANN to find the drone camera position and orientation. The algorithm has high accuracy (these errors were found for the validation set of images as differences between positions and orientations, obtained from a pretrained artificial neural network, and known positions and orientations), it is not sensitive to interference associated with changes in lighting, the appearance of external moving objects and the other phenomena where other methods of visual navigation are not effective. For the second part of the drone navigation problem, the paper presents a method for stabilization of drone flight controlled by autopilot with time delay. Indeed, image processing for navigation demands a lot of time and results in a time delay. However, the proposed method allows to get stable control in the presence of this time delay.

AB - In the paper, visual navigation of a drone is considered. The drone navigation problem consists of two parts. The first part is finding the real position and orientation of the drone. The second part is finding the difference between desirable and real position and orientation of the drone and creation of the correspondent control signal for decreasing the difference. For the first part of the drone navigation problem, the paper presents a method for determining the coordinates of the drone camera with respect to known three-dimensional (3D) ground objects using deep learning. The algorithm has two stages. It causes the algorithm to be easy for interpretation by artificial neural network (ANN) and consequently increases its accuracy. At the first stage, we use the first ANN to find coordinates of the object origin projection. At the second stage, we use the second ANN to find the drone camera position and orientation. The algorithm has high accuracy (these errors were found for the validation set of images as differences between positions and orientations, obtained from a pretrained artificial neural network, and known positions and orientations), it is not sensitive to interference associated with changes in lighting, the appearance of external moving objects and the other phenomena where other methods of visual navigation are not effective. For the second part of the drone navigation problem, the paper presents a method for stabilization of drone flight controlled by autopilot with time delay. Indeed, image processing for navigation demands a lot of time and results in a time delay. However, the proposed method allows to get stable control in the presence of this time delay.

KW - Alex-Net

KW - Artificial neural network

KW - Autopilot

KW - Deep learning convolution network

KW - Drone

KW - Machine learning

KW - Quaternions

KW - Stability of differential equations

KW - Time delay

KW - Vision-based navigation

KW - Visual navigation

UR - http://www.scopus.com/inward/record.url?scp=85097303373&partnerID=8YFLogxK

U2 - 10.3390/math8122140

DO - 10.3390/math8122140

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

AN - SCOPUS:85097303373

SN - 2227-7390

VL - 8

SP - 1

EP - 13

JO - Mathematics

JF - Mathematics

IS - 12

M1 - 2140

ER -