Rapid GNSS-Based Calibration and Target Localization Strategy for Deployable Optical UAV Detection Systems

Abstract

This article presents the design and experimental evaluation of the feasibility of a rapid Global Navigation Satellite System (GNSS)-based calibration and target localization strategy in a process structure for deployable optical uncrewed aerial vehicle (UAV) detection systems in combination with complementary sensors, such as radar. The calibration strategy utilizes a cooperative UAV with standard GNSS point positioning to determine unknown orientation parameters of a deployable telescope-based system in a global reference system. During a calibration flight, the cooperative UAV with a GNSS receiver is automatically tracked using a deep learning-based detector and a pan/tilt mount. By combining flight log data of the UAV, detector, and mount data, orientation parameters of the system relative to the World Geodetic System 1984 are determined. Uncertainties of internal and external sources are investigated, and the proposed process structure is enhanced by a target localization strategy based on Lissajous search trajectories, which is also validated during field tests in combination with a radar system after applying the proposed calibration procedure. During field tests, a remaining angular calibration root-mean-square (rms) error of 0.255° in elevation and 0.277° in azimuth is achieved using a DJI Mini 2 and a calibration flight duration of 100s. By locating and tracking a validation UAV, an rms error of 0.376° in elevation and 0.336° in azimuth is achieved.