Stochastic system identification is of great interest in the areas of control and communication. In stochastic system identification, a model of a dynamic system is determined based on given inputs and received outputs from the system, where stochastic uncertainties are also involved. The scope of the report is to consider continuous-time models used within control and communication and to estimate the model parameters from sampled data with high accuracy in a computational efficient way. Continuous-time models of systems controlled in a networked environment, stochastic closed-loop systems, and wireless channels are considered. The parameters of a transfer function based model for the process in a networked control system are first estimated by a covariance function based approach, relying upon the second order statistical properties of the output signal. Some other approaches for estimating the parameters of continuous-time models for processes in networked environments are also considered. Further, the parameters of continuous-time autoregressive exogenous models are estimated from closed-loop filtered data, where the controllers in the closed-loop are of proportional and proportional integral type, and where the closed-loop also contains a time-delay. Moreover, a stochastic differential equation is derived for Jakes's wireless channel model, describing the dynamics of a scattered electric field with the moving receiver incorporating a Doppler shift. / <p>Article I was still in manuscript form at the time of the defense.</p>

Many important factors affect the spread of childhood infectious disease. To understand better the fundamental drivers of infectious disease spread, several researchers have estimated seasonal transmission coefficients using discrete-time models. This research addresses several shortcomings of the discrete-time approaches, including removing the need for the reporting interval to match the serial interval of the disease using infectious disease data from three major cities: New York City, London, and Bangkok. Using a simultaneous approach for optimization of differential equation systems with a Radau collocation discretization scheme and total variation regularization for the transmission parameter profile, this research demonstrates that seasonal transmission parameters can be effectively estimated using continuous-time models. This research further correlates school holiday schedules with the transmission parameter for New York City and London where previous work has already been done, and demonstrates similar results for a relatively unstudied city in childhood infectious disease research, Bangkok, Thailand.

abstract: Cigarette smoking remains a major global public health issue. This is partially due to the chronic and relapsing nature of tobacco use, which contributes to the approximately 90% quit attempt failure rate. The recent rise in mobile technologies has led to an increased ability to frequently measure smoking behaviors and related constructs over time, i.e., obtain intensive longitudinal data (ILD). Dynamical systems modeling and system identification methods from engineering offer a means to leverage ILD in order to better model dynamic smoking behaviors. In this dissertation, two sets of dynamical systems models are estimated using ILD from a smoking cessation clinical trial: one set describes cessation as a craving-mediated process; a second set was reverse-engineered and describes a psychological self-regulation process in which smoking activity regulates craving levels. The estimated expressions suggest that self-regulation more accurately describes cessation behavior change, and that the psychological self-regulator resembles a proportional-with-filter controller. In contrast to current clinical practice, adaptive smoking cessation interventions seek to personalize cessation treatment over time. An intervention of this nature generally reflects a control system with feedback and feedforward components, suggesting its design could benefit from a control systems engineering perspective. An adaptive intervention is designed in this dissertation in the form of a Hybrid Model Predictive Control (HMPC) decision algorithm. This algorithm assigns counseling, bupropion, and nicotine lozenges each day to promote tracking of target smoking and craving levels. Demonstrated through a diverse series of simulations, this HMPC-based intervention can aid a successful cessation attempt. Objective function weights and three-degree-of-freedom tuning parameters can be sensibly selected to achieve intervention performance goals despite strict clinical and operational constraints. Such tuning largely affects the rate at which peak bupropion and lozenge dosages are assigned; total post-quit smoking levels, craving offset, and other performance metrics are consequently affected. Overall, the interconnected nature of the smoking and craving controlled variables facilitate the controller's robust decision-making capabilities, even despite the presence of noise or plant-model mismatch. Altogether, this dissertation lays the conceptual and computational groundwork for future efforts to utilize engineering concepts to further study smoking behaviors and to optimize smoking cessation interventions. / Dissertation/Thesis / Doctoral Dissertation Bioengineering 2014