Improved state estimation for jump Markov linear systems

Orguner, Umut
This thesis presents a comprehensive example framework on how current multiple model state estimation algorithms for jump Markov linear systems can be improved. The possible improvements are categorized as: -Design of multiple model state estimation algorithms using new criteria. -Improvements obtained using existing multiple model state estimation algorithms. In the first category, risk-sensitive estimation is proposed for jump Markov linear systems. Two types of cost functions namely, the instantaneous and cumulative cost functions related with risk-sensitive estimation are examined and for each one, the corresponding multiple model estate estimation algorithm is derived. For the cumulative cost function, the derivation involves the reference probability method where one defines and uses a new probability measure under which the involved processes has independence properties. The performance of the proposed risk-sensitive filters are illustrated and compared with conventional algorithms using simulations. The thesis addresses the second category of improvements by proposing -Two new online transition probability estimation schemes for jump Markov linear systems. -A mixed multiple model state estimation scheme which combines desirable properties of two different multiple model state estimation methods. The two online transition probability estimators proposed use the recursive Kullback-Leibler (RKL) procedure and the maximum likelihood (ML) criteria to derive the corresponding identification schemes. When used in state estimation, these methods result in an average error decrease in the root mean square (RMS) state estimation errors, which is proved using simulation studies. The mixed multiple model estimation procedure which utilizes the analysis of the single Gaussian approximation of Gaussian mixtures in Bayesian filtering, combines IMM (Interacting Multiple Model) filter and GPB2 (2nd Order Generalized Pseudo Bayesian) filter efficiently. The resulting algorithm reaches the performance of GPB2 with less Kalman filters.


A Bayesian Approach to Learning Scoring Systems
Ertekin Bolelli, Şeyda (2015-12-01)
We present a Bayesian method for building scoring systems, which are linear models with coefficients that have very few significant digits. Usually the construction of scoring systems involve manual efforthumans invent the full scoring system without using data, or they choose how logistic regression coefficients should be scaled and rounded to produce a scoring system. These kinds of heuristics lead to suboptimal solutions. Our approach is different in that humans need only specify the prior over what the ...
Analytical modeling of a time-threshold based multi-guard bandwidth allocation scheme for cellular networks
Candan, Idil; Salamah, Muhammed (2009-05-28)
In this paper, the analytical modeling of a time-threshold based multi-guard bandwidth allocation scheme is presented using a two-dimensional markov chain. The main idea of the scheme is based on monitoring the elapsed real time of handoff calls and according to a time threshold (t(e)), a handoff call is either slightly-prioritized or fully-prioritized. A slightly-prioritized handoff call has higher priority than a new call and lower priority than a fully-prioritized handoff call. Also in this paper, the nu...
An approach for decentralized process modeling
Turetken, Oktay; Demirörs, Onur (2007-05-20)
This paper describes a method for organizations to perform process modeling in a decentralized and concurrent manner. The approach is based on the idea that modeling organizations' processes can be performed by individuals actually performing the processes. Instead of having a central and devoted group of people to analyze, understand, model and improve processes, real performers are held responsible to model and improve their own processes concurrently. The paper also summarizes the lessons learned by appl...
On investigating the performance of various turbo decoding algorithms based on the finite-EXIT chart method
Yilmaz, AO (2004-04-30)
Finite length analysis of iterative decoders can be done by using probabilistic models based on EXIT charts. The validity of these models will be investigated by checking the performance of iterative decoding under various scenarios.
Improving Computational Efficiency of Particle Swarm Optimization for Optimal Structural Design
Hasançebi, Oğuzhan (2013-06-01)
This paper attempts to improve the computational efficiency of the well known particle swarm optimization (PSO) algorithm for tackling discrete sizing optimization problems of steel frame structures. It is generally known that, in structural design optimization applications, PSO entails enormously time-consuming structural analyses to locate an optimum solution. Hence, in the present study it is attempted to lessen the computational effort of the algorithm, using the so called upper bound strategy (UBS), wh...
Citation Formats
U. Orguner, “Improved state estimation for jump Markov linear systems,” Ph.D. - Doctoral Program, Middle East Technical University, 2005.