Range assisted inertial navigation system for multi-rotor micro aerial vehicles

Fernando, Hondanaidelage C. T. Eranga (2022) Range assisted inertial navigation system for multi-rotor micro aerial vehicles. Doctoral (PhD) thesis, Memorial University of Newfoundland.

Full text not available from this repository.


Pose estimation of multi-rotor micro-aerial vehicles (MAVs) in indoor environments is a key challenge in the development of autonomous MAV applications. The main focus of this research is to develop robust, and accurate localization systems for MAVs utilizing the least number of sensors. This research develops three range assisted inertial navigation system (RINS) designs for quadrotor MAVs. Range measurement is selected as the key measurement due to the robustness, accuracy of the emerging ranging technologies, and ease of deployment in various types of environments even under challenging conditions. The first part of this thesis presents the development of a RINS that utilizes three or two range measurements along with the accelerometer and gyroscope measurements. The proposed RINS incorporates the effects of aerodynamic drag forces on MAV, which allows the RINS to operate without using a velocity sensor. A nonlinear observability study is carried out to evaluate the feasibility, and identify the limitations of the proposed RINS. The observability analysis is conducted based on two cases. Case 1 : RINS with three range measurements, and Case 2 : RINS with two range measurements. For each case, the analysis shows that the RINS is locally weakly observable for a generic trajectory. Additionally, several specific trajectories are identified that render the RINS unobservable. The unobservable directions for each unobservable trajectory are determined analytically and validated through numerical simulations. The performance of the proposed RINS during an observable trajectory is validated through experiments conducted on a quadrotor MAV. The second part of this thesis analyzes the consistency of the error-state extended Kalman filter (EKF) implementation of the proposed two and three range assisted INS. The analysis shows that the EKF-RINS suffers from inconsistencies when the MAV is flying on the unobservable trajectories identified through the observability study. The consistency of the estimator under unobservable trajectories is improved by applying observability constraints during the estimation process. The novelty of the proposed approach is that the observability constraints are dependent on the unobservable scenarios and applied only during any unobservable trajectory. A Monte Carlo analysis is performed to validate the improvement of the localization performance of EKF-RINS achieved through selectively applying the observability constraints. Finally, this thesis presents two RINS designs that use a single range measurement. The first design uses just the range measurement. However, the observability analysis shows that this design is unobservable under any trajectory. In order to develop a locally weakly observable trajectory, the second design incorporates heading information in addition to the single range measurement. The observability study identified several unique trajectories under which the magnetometer and single range assisted INS (M-RINS) becomes unobservable. Performance evaluation of the M-RINS and the validation of unobservable directions are carried out using numerical simulations.

Item Type: Thesis (Doctoral (PhD))
URI: http://research.library.mun.ca/id/eprint/15867
Item ID: 15867
Additional Information: Includes bibliographical references (pages 142-151) -- Restricted until August 15, 2023
Keywords: micro aerial vehicles, inertial navigation system, range assisted localization, UWB
Department(s): Engineering and Applied Science, Faculty of
Date: August 2022
Date Type: Submission
Digital Object Identifier (DOI): https://doi.org/10.48336/7HHQ-PT74
Library of Congress Subject Heading: Micro air vehicles; Inertial navigation systems; Indoor positioning systems (Wireless localization)

Actions (login required)

View Item View Item