Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.650675
Title: Widely linear state space filtering of improper complex signals
Author: Dini, Dahir Hersi
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2013
Availability of Full Text:
Access through EThOS:
Full text unavailable from EThOS. Please try the link below.
Access through Institution:
Abstract:
Complex signals are the backbone of many modern applications, such as power systems, communication systems, biomedical sciences and military technologies. However, standard complex valued signal processing approaches are suited to only a subset of complex signals known as proper, and are inadequate of the generality of complex signals, as they do not fully exploit the available information. This is mainly due to the inherent blindness of the algorithms to the complete second order statistics of the signals, or due to under-modelling of the underlying system. The aim of this thesis is to provide enhanced complex valued, state space based, signal processing solutions for the generality of complex signals and systems. This is achieved based on the recent advances in the so called augmented complex statistics and widely linear modelling, which have brought to light the limitations of conventional statistical complex signal processing approaches. Exploiting these developments, we propose a class of widely linear adaptive state space estimation techniques, which provide a unified framework and enhanced performance for the generality of complex signals, compared with conventional approaches. These include the linear and nonlinear Kalman and particle filters, whereby it is shown that catering for the complete second order information and system models leads to significant performance gains. The proposed techniques are also extended to the case of cooperative distributed estimation, where nodes in a network collaborate locally to estimate signals, under a framework that caters for general complex signals, as well as the cross-correlations between observation noises, unlike earlier solutions. The analysis of the algorithms are supported by numerous case studies, including frequency estimation in three phase power systems, DIFAR sonobuoy underwater target tracking, and real-world wind modeling and prediction.
Supervisor: Mandic, Danilo Sponsor: Defence Science and Technology Laboratory ; Ministry of Defence
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.650675  DOI: Not available
Share: