Download or read book State Inference and Bayesian Identification of Non-linear State-space Models written by Aditya Tulsyan and published by . This book was released on 2013 with total page 285 pages. Available in PDF, EPUB and Kindle. Book excerpt: The main focus of this thesis is on state inference and identification of non-linear dynamical systems, which can be represented by discrete-time, stochastic state-space models (SSMs). We consider the state inference and identification as related, but two distinct problems. For identification of SSMs, we restrict ourselves only to the class of Bayesian methods. In this thesis, we develop a novel sequential Monte Carlo (SMC) based Bayesian method for simultaneous on-line state inference and identification of non-linear SSMs. Extension of the method to handle missing measurements in real-time is also provided. Using posterior Cramþer-Rao lower bound (PCRLB), a minimum mean square error (MMSE) simultaneous state inferencing and identification strategy is developed for general nonlinear systems. The PCRLB used here is derived for discrete-time, stochastic non-linear SSMs with unknown model parameters. It is shown that under some conditions, performing simultaneous state inferencing and identification according to the developed PCRLB based strategy yields a minimum mean square error state and parameter estimates. To allow assessment of the quality of the parameter estimates, a PCRLB based tool is developed for error analysis. A distinct advantage of the developed tool is that it is general, and can be used to perform error analysis for an entire class of on-line Bayesian identification methods. In addition to the above developments, the problems of input design and prior design are also considered in this thesis. The input design problem helps to design optimal inputs for Bayesian identification of non-linear SSMs; whereas, the prior design problem helps to effectively organize a priori information available on the process and model parameters. In this thesis, the problem of prior design is only considered in the context of designing optimal inputs for Bayesian identification of non-linear SSMs. For state inferencing in non-linear SSMs, we develop a PCRLB based performance assessment and diagnosis tool for different non-linear state filters. The proposed assessment and diagnosis tool makes use of PCRLB, derived for discrete-time, stochastic non-linear SSMs with known model parameters. The utility of the above developments in devising an optimal state inferencing strategy for non-linear systems is also provided. To avoid using the true states in the computation of the PCRLB, an SMC based method is also developed to allow computation of the PCRLB in absence of true state information. Finally, we show how the tools developed in this thesis can be put together into a unified framework to allow for efficient state inferencing and Bayesian identification of non-linear dynamical systems.