Simultaneous state and parameter estimation is essential for control system design and dynamic modeling of physical systems. This capability provides critical real-time insight into system behavior, supports the discovery of underlying mechanisms, and facilitates adaptive control strategies. Surveyed in this review paper are two classes of state and parameter estimation methods: Kalman Filters and Luenberger Observers. The Kalman Filter framework, including its major variants such as the Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), Cubature Kalman Filter (CKF), and Ensemble Kalman Filter (EnKF), has been widely applied for joint and dual estimation in linear and nonlinear systems under uncertainty. In parallel, Luenberger observers, typically used in deterministic settings, offer alternative approaches through high-gain, sliding mode, and adaptive observer structures. This review focuses on the theoretical foundations, algorithmic developments, and application domains of these methods and provides a comparative analysis of their advantages, limitations, and practical relevance across diverse engineering scenarios.
29 May 2006 in Journal of Dynamic Systems, Measurement, and Control
Presented is the detection, isolation, and estimation of faults that occur in the intake air path of internal combustion engines during steady state operation. The proposed diagnostic approach is based on a static air path model, which is adapted online such that the model output matches the measured output during steady state conditions. The resulting changes in the model coefficients create a vector whose magnitude and direction are used for fault detection and isolation. Fault estimation is realized by analyzing the residual between the actual sensor measurement and the output of the original (i.e., healthy) model. To identify the structure of the steady state air path model a process called system probing is developed. The proposed diagnostics algorithm is experimentally validated on the intake air path of a Ford V-8 engine. The specific faults to be identified include two of the most problematic faults that degrade the performance of transient fueling controllers: bias in the mass air flow sensor and a leak in the intake manifold. The selected model inputs include throttle position and engine speed, and the output is the mass air flow sensor measurement.
1 December 2004 in Journal of Dynamic Systems, Measurement, and Control
A method for designing repetitive feedback controllers using nonparametric frequency response plant models is developed. In comparison to the zero-phase-error (ZPE) controller (ASME J. Dyn. Syst. Meas. Control, 111, pp. 353–358), this method has the added benefit of providing improved transient performance when the plant inverse is unstable. In this controller design process, a connection is made between model uncertainty and the desired frequency response of the so-called q filter. Also, it will be shown that an optimal equiripple filter is useful when designing high-order q filters. The entire process was experimentally verified on an engine control application. A repetitive controller was used to determine the dynamic fueling requirements of a fuel injected, spark-ignition engine subjected to periodic changes in the throttle position. This fueling information is necessary when designing feedforward fueling algorithms.
1 March 1996 in Journal of Dynamic Systems, Measurement, and Control
Presented in this paper is a method in which the performance weights W1(s) and W2(s) of the robust performance problem and the H∞ optimization problem can be chosen to directly enforce hard time domain constraints. The class of systems addressed are single-input-single-output (SISO) regulating systems required to maintain the output within a prespecified time domain tolerance despite (i) plant uncertainty, (ii) an external step disturbance, and (iii) actuator output saturation. The performance weights are extracted from the feedback configuration and facilitate performance maximization.
17 August 2005 in Journal of Dynamic Systems, Measurement, and Control
Presented in this paper is a feedforward fueling controller identification methodology for the transient fueling control of spark ignition (SI) engines. The hypothesis of this work is that the feedforward fueling control of SI engines can be separated into steady state and transient phenomena and that the majority of the nonlinear behavior associated with engine fueling can be captured with nonlinear steady state models. The proposed transient controller identification process is built from standard nonparametric identification techniques followed by parametric model recovery. Crank angle serves as the independent variable for these models. Two separate system identification problems are solved to identify the air path dynamics and the fueling path dynamics. The transient feedforward controller is then calculated as the ratio of the air path-over-the fueling path dynamics thereby coordinating the engine fueling with the air path dynamics. It will be shown that a linear transient feedforward-fueling controller operating in tandem with a nonlinear steady state fueling controller can achieve air-fuel ratio regulation comparable to the production fueling controller without the extensive controller calibration process. The engine used in this investigation is a 1999 Ford 4.6L V-8 fuel injected engine.
14 July 2016 in Journal of Dynamic Systems, Measurement, and Control
Presented are reduced-order models of one-dimensional transient two-phase gas–liquid flow in pipelines. The proposed model is comprised of a steady-state multiphase flow mechanistic model in series with a transient single-phase flow model in transmission lines. The steady-state model used in our formulation is a multiphase flow mechanistic model. This model captures the steady-state pressure drop and liquid holdup estimation for all pipe inclinations. Our implementation of this model will be validated against the Stanford University multiphase flow database. The transient portion of our model is based on a transmission line modal model. The model parameters are realized by developing equivalent fluid properties that are a function of the steady-state pressure gradient and liquid holdup identified through the mechanistic model. The model ability to reproduce the dynamics of multiphase flow in pipes is evaluated upon comparison to olga, a commercial multiphase flow dynamic code, using different gas volume fractions (GVF). The two models show a good agreement of the steady-state response and the frequency of oscillation indicating a similar estimation of the transmission line natural frequency. However, they present a discrepancy in the overshoot values and the settling time due to a difference in the calculated damping ratio. The utility of the developed low-dimensional model is the reduced computational burden of estimating transient multiphase flow in transmission lines, thereby enabling real-time estimation of pressure and flow rate.