Open AccessUGV‐UAV robust cooperative positioning algorithm with object detection
Author(s) -
Wang Dongjia,
Lian Baowang,
Tang Chengkai
Publication year - 2021
Publication title -
iet intelligent transport systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.579
H-Index - 45
eISSN - 1751-9578
pISSN - 1751-956X
DOI - 10.1049/itr2.12063
Subject(s) - computer vision , inertial measurement unit , artificial intelligence , gnss applications , computer science , kalman filter , simultaneous localization and mapping , object detection , global positioning system , extended kalman filter , robot , mobile robot , pattern recognition (psychology) , telecommunications
Traditional Global Navigation Satellite Systems (GNSS) experience their limitations in urban canyons. However, it is significant to improve the accuracy of positioning with the rapid development of smart cities. To solve this problem, a UGV‐UAV robust cooperative positioning algorithm with object detection is proposed, which utilises an unmanned aerial vehicle (UAV) to assist an unmanned ground vehicle (UGV) to achieve accurate positioning. When the UAV is in the sky with a good reception of satellite signals, the UGV uses the YOLOv3 object detection method to detect the UAV in images captured by camera, and acquires visual measurements including angles and ranges of the ground camera relative to the UAV through the proposed monocular vision measuring with object detection (ODMVM) model. Then, in order to solve the problem that visual measurement is disturbed by the real world, a robust Kalman filter is introduced that integrates measurements from available GNSS, inertial measurement unit (IMU), monocular camera, and the position broadcast of cooperative UAV to obtain more robust and accurate position estimation. Experimental and simulation results show that the proposed cooperation positioning algorithm can improve the positioning accuracy by 73.63% compared with the traditional cooperation positioning algorithm in urban canyons.