Facial reconstruction

Search LJMU Research Online

Browse Repository | Browse E-Theses

Generalised state estimators for robotic platforms through the use of improved sensor characterisation and variance modelling.

Pointon, H (2020) Generalised state estimators for robotic platforms through the use of improved sensor characterisation and variance modelling. Doctoral thesis, Liverpool John Moores University.

[img]
Preview
Text
2019pointonphd.pdf - Published Version

Download (109MB) | Preview

Abstract

The aim of this research is to develop an improved representation of the sensor variance in a state estimator and assess its viability in conjunction with generalised system models. This would enable the use of a single state estimation system across many different platforms. A key challenge in the safe deployment of \gls{uav} systems is localisation. In built up environments traditional \gls{GNSS} systems become unreliable, and other sensing systems are often limited in application. Deploying \gls{uav} platforms in complex or safety critical operations often requires a legal exemption, with a demonstration of robust, practical operation of the equipment proposed. To this end, a generalised state estimator would allow repeated use of the same, experimentally validated systems. This research presents a methodology to characterise the principle input sensor, in this case, an UWB system through the use of the RTS. The project continues, by demonstrating the implementation of a sensor variance model in the commonly used EKF framework, in both ground and aerial platforms. The work concludes, with a demonstration of a generalised state estimator in use for both a ground and aerial platform, and shows a more stable, noise tolerant output, assessed using the RTS system.

Item Type: Thesis (Doctoral)
Uncontrolled Keywords: State Estimation; EKF; UAV; Probabilistic Filtering; UWB; RTS; MB-EKF
Subjects: T Technology > T Technology (General)
T Technology > TL Motor vehicles. Aeronautics. Astronautics
Divisions: Engineering
Date Deposited: 17 Jun 2020 22:17
Last Modified: 07 Sep 2022 16:04
DOI or ID number: 10.24377/LJMU.t.00013086
Supervisors: Bezombes, F, Matthews, C and Abdullah, B
URI: https://researchonline.ljmu.ac.uk/id/eprint/13086
View Item View Item