Thesis (Ph.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1979. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. / Vita. / Includes bibliographical references. / by Richard Tse-min Ku. / Ph.D.

by Michael Athans and Sanjoy K. Mitter. / Research supported by Air Force Office of Scientific Research (AFSC) Research Grant AF-AFOSR 72-2273. Report for 1975/76 distributed through Industrial Liaison Program.

The observability measure based on the mutual information between the last state and the measurement sequence originally proposed by Mohler and Hwang (1988) is analyzed in detail and improved further for linear time invariant discrete-time Gaussian stochastic systems by extending the definition to the observability measure of a state sequence. By using the new observability measure it is shown that the unobservable states of the deterministic system have no effect on this measure and any observable part with no measurement uncertainty makes it infinite. Other distance measures i.e., Bhattacharyya and Hellinger distances are also investigated to be used as observability measures. The relationships between the observability measures and the covariance matrices of Kalman filter and the state sequence conditioned on the measurement sequence are derived. Steady state characteristics of the observability measure based on the last state is examined. The observability measures of a subspace of the state space, an individual state, the modes of the system are investigated. One of the results obtained in this part is that the deterministically unobservable states may have nonzero observability measures. The observability measures based on the mutual information are represented recursively and calculated for nonlinear stochastic systems. Then the measures are applied to a nonlinear stochastic system by using the particle filter methods. The arguments given for the LTI case are also observed for nonlinear stochastic systems. The second moment approximation deviates from the actual values when the nonlinearity in the system increases.

Use of model-based path planning and navigation is a common strategy in mobile robotics. However, navigation performance may degrade in complex, time-varying environments under model uncertainty because of loss of prediction ability for the robot state over time. Exploration and monitoring of ocean regions using autonomous marine robots is a prime example of an application where use of environmental models can have great benefits in navigation capability. Yet, in spite of recent improvements in ocean modeling, errors in model-based flow forecasts can still significantly affect the accuracy of predictions of robot positions over time, leading to impaired path-following performance. In developing new autonomous navigation strategies, it is important to have a quantitative understanding of error in predicted robot position under different flow conditions and control strategies. The main contributions of this thesis include development of an analytical model for the growth of error in predicted robot position over time and theoretical derivation of bounds on the error growth, where error can be attributed to drift caused by unmodeled components of ocean flow. Unlike most previous works, this work explicitly includes spatial structure of unmodeled flow components in the proposed error growth model. It is shown that, for a robot operating under flow-canceling control in a static flow field with stochastic errors in flow values returned at ocean model gridpoints, the error growth is initially rapid, but slows when it reaches a value of approximately twice the ocean model gridsize. Theoretical values for mean and variance of error over time under a station-keeping feedback control strategy and time-varying flow fields are computed. Growth of error in predicted vehicle position is modeled for ocean models whose flow forecasts include errors with large spatial scales. Results are verified using data from several extended field deployments of Slocum autonomous underwater gliders, in Monterey Bay, CA in 2006, and in Long Bay, SC in 2012 and 2013.

Diss. Uppsala : Univ., 2004.

Thesis (M.Sc. (Hons.)) -- University of Western Sydney, 2001. / "A thesis presented to the University of Western Sydney in partial fulfilment of the requirements for the degree of Master of Science" Bibliography : leaves 65-69.

Thesis (Ph. D.)--Rijksuniversiteit te Groningen. / At head of title: Rijksuniversiteit te Groninge. Summary in Dutch. Parts of this thesis are based on material originally appearing in Statistica Neerlandica, 1978, Automatica, 1981 and the proceedings of Dynamic Modelling and Control of National Economics 1981 ... Includes index. Bibliography: p. 153-159.

The stochasticity of cellular processes and the small number of molecules in a cell make deterministic models inappropriate for modelling chemical reactions at the single cell level. The Chemical Master Equation (CME) is widely used to describe the evolution of biochemical reactions inside cells stochastically but is computationally expensive. The Linear Noise Approximation (LNA) is a popular method for approximating the CME in order to carry out inference and parameter estimation in stochastic models. Data from stochastic systems is often aggregated over time. One such example is in luminescence bioimaging, where a luciferase reporter gene allows us to quantify the activity of proteins inside a cell. The luminescence intensity emitted from the luciferase experiments is collected from single cells and is integrated over a time period (usually 15 to 30 minutes), which is then collected as a single data point. In this work we consider stochastic systems that we approximate using the Linear Noise Approximation (LNA). We demonstrate our method by learning the parameters of three different models from which aggregated data was simulated, an Ornstein-Uhlenbeck model, a Lotka-Voltera model and a gene transcription model. We have additionally compared our approach to the existing approach and find that our method is outperforming the existing one. Finally, we apply our method in microscopy data from a translation inhibition experiment.

abstract: Resource allocation in communication networks aims to assign various resources such as power, bandwidth and load in a fair and economic fashion so that the networks can be better utilized and shared by the communicating entities. The design of efficient resource-allocation algorithms is, however, becoming more and more challenging due to the precipitously increasing scale of the networks. This thesis strives to understand how to design such low-complexity algorithms with performance guarantees. In the first part, the link scheduling problem in wireless ad hoc networks is considered. The scheduler is charge of finding a set of wireless data links to activate at each time slot with the considerations of wireless interference, traffic dynamics, network topology and quality-of-service (QoS) requirements. Two different yet essential scenarios are investigated: the first one is when each packet has a specific deadline after which it will be discarded; the second is when each packet traverses the network in multiple hops instead of leaving the network after a one-hop transmission. In both scenarios the links need to be carefully scheduled to avoid starvation of users and congestion on links. One greedy algorithm is analyzed in each of the two scenarios and performance guarantees in terms of throughput of the networks are derived. In the second part, the load-balancing problem in parallel computing is studied. Tasks arrive in batches and the duty of the load balancer is to place the tasks on the machines such that minimum queueing delay is incurred. Due to the huge size of modern data centers, sampling the status of all machines may result in significant overhead. Consequently, an algorithm based on limited queue information at the machines is examined and its asymptotic delay performance is characterized and it is shown that the proposed algorithm achieves the same delay with remarkably less sampling overhead compared to the well-known power-of-two-choices algorithm. Two messages of the thesis are the following: greedy algorithms can work well in a stochastic setting; the fluid model can be useful in "derandomizing" the system and reveal the nature of the algorithm. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2015

Orientador: Jose Claudio Geromel / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-06T08:10:58Z (GMT). No. of bitstreams: 1 Goncalves_AlimPedrodeCastro_M.pdf: 293253 bytes, checksum: 4365f5ae19ae33e8d97053c2c57ad15d (MD5) Previous issue date: 2006 / Resumo: Este trabalho aborda alguns dos aspectos mais relevantes relacionados à estabilidade e norma H2 de sistemas lineares discretos sujeitos a saltos markovianos, bem como as estratégias para a síntese de controle por realimentação de estado. A maior contribuição apresentada é um método para calcular os ganhos de realimentação de estado sem a necessidade de observar, em cada instante, todos os estados da cadeia de Markov / Abstract: This work discusses some of the most relevant aspects of stability and H2 norm of discrete-time markov jump linear systems, as well as a method for state feedback contraI design. Our major contribution is on the definition of a procedure to determine the state feedback gains without the complete knowledge, at each instant of time, of the Markov chain state / Mestrado / Automação / Mestre em Engenharia Elétrica