This paper presents a collection of questions and a brief example directed towards the development of a methodology for the incorporation of a priori information (‘prejudice’ to some) into System Identification for ...
详细信息
This paper presents a collection of questions and a brief example directed towards the development of a methodology for the incorporation of a priori information (‘prejudice’ to some) into System Identification for the purpose of feedback control design. The route that we take is to follow through a sequence of questions which attempt to extract from the physical problem information which guides the selection of design parameters or choices. By concentrating on the external imposition of performance or model veracity measures, we hope to expose the wealth of unresolved issues associated with modelling for control, which are by-passed in more statistical modelling approaches based on finding the ‘true’ linear model from amongst the noise. While our title, tone and content might be deliberately provocative, our resort to procedures such as these in formulating practical model-based controllers, helps to vindicate our approach.
We propose a method for the identification of a nonlinear plant under possibly nonlinear feedback. This procedure is a nonlinear extension of a method known as the Hansen scheme in the literature. It is shown that usi...
详细信息
We propose a method for the identification of a nonlinear plant under possibly nonlinear feedback. This procedure is a nonlinear extension of a method known as the Hansen scheme in the literature. It is shown that using nonlinear left fractional descriptions one can convert a general nonlinear closed-loop identification problem to one of open-loop identification by parametrizing the model using a Youla-Kucera parameter. The open-loop problem can be implemented by parametrizing the nonlinear Youla parameter in terms of a model of the plant. We provide gradient expressions for implementation in a steepest descent algorithm.
We present gradient expressions for a closed-loop parametric identification scheme. The method is based on the minimization of a standard identification criterion and a parametrization that is tailored to the closed-l...
详细信息
We present gradient expressions for a closed-loop parametric identification scheme. The method is based on the minimization of a standard identification criterion and a parametrization that is tailored to the closed-loop configuration. It is shown that for both linear and nonlinear plants and controllers, the gradient signals can be computed exactly.
Presents an identification-based mechanism for introducing guaranteed stability when using a data-driven model-free iterative control design method known as iterative feedback tuning. Also, the use of unbiased estimat...
详细信息
Presents an identification-based mechanism for introducing guaranteed stability when using a data-driven model-free iterative control design method known as iterative feedback tuning. Also, the use of unbiased estimates of the Hessian is shown to significantly improve the user control over the tuning procedure.
A computationally inexpensive suboptimal blind equalization algorithm is presented for noisy IIR channels. It is based on combining a recursive hidden Markov model (HMM) estimator with a relaxed SPR (strictly positive...
详细信息
A computationally inexpensive suboptimal blind equalization algorithm is presented for noisy IIR channels. It is based on combining a recursive hidden Markov model (HMM) estimator with a relaxed SPR (strictly positive real) extended least squares (ELS) scheme. Simulation studies show that the algorithm yields satisfactory results.< >
robust exponential stability of full state and observer feedback controlled systems is explored in the context of predictive control. The total stability theorem is used to ensure exponential stability via bounds on m...
详细信息
robust exponential stability of full state and observer feedback controlled systems is explored in the context of predictive control. The total stability theorem is used to ensure exponential stability via bounds on model, modelling error and control gain characteristics in two scenarios of differing complexities. Sufficient conditions are obtained to ensure an observed, nonlinear plant's state trajectory stays within a fixed distance of an exponentially stable, controlled, model state trajectory. These results have applications in predictive control, where model-based receding-horizon designs are used to yield robust solutions on the infinite-horizon.
In this paper we investigate some important issues in system identification for the windsurfer approach to robustadaptive control. In particular, correlation function estimates an Power spectrum estimates are compare...
详细信息
In this paper we investigate some important issues in system identification for the windsurfer approach to robustadaptive control. In particular, correlation function estimates an Power spectrum estimates are compared as methods for model validation. This comparison leads us to suggest a reliable procedure for deciding (1) when should we identify a better model, and (2) whether we have identified a good model for our purposes.
The use of the extended Kalman filter (EKF) is heavily entrenched in non-linear signal processing applications. However linearisation errors inherent in the specification of an EKF can severely degrade its performance...
详细信息
暂无评论