Data‐driven tuning of linear parameter‐varying precompensators

Methods for direct data‐driven tuning of the parameters of precompensators for linear parameter‐varying (LPV) systems are developed. Since the commutativity property is not always satisfied for LPV systems, previously proposed methods for LTI systems that use this property cannot be directly adapted. When the ideal precompensator giving perfect mean tracking exists in the proposed precompensator parameterization, the LPV transfer operators do commute and an algorithm using only two experiments on the real system is proposed. It is shown that this algorithm gives consistent estimates of the ideal parameters despite the presence of stochastic disturbances. For the more general case, when the ideal precompensator does not belong to the set of parameterized precompensators, another technique is developed. This technique requires a number of experiments equal to twice the number of precompensator parameters and it is shown that the calculated parameters minimize the mean‐squared tracking error. The theoretical results are demonstrated in simulation. Copyright © 2009 John Wiley & Sons, Ltd.     

Iterative learning control for robotic manipulators: A bounded‐error algorithm

This paper presents a model‐based nonlinear iterative learning control (NILC) for nonlinear multiple‐input and multiple‐output mechanical systems of robotic manipulators. An algorithm of a new strategy for the NILC implementation is proposed. This algorithm ensures that trajectory‐tracking errors of the proposed NILC, when implemented, are bounded by a given error norm bound. Both standard and bounded‐error learning control laws with feedback controllers attached are considered. The NILC synthesis is based on a dynamic model of a six degrees of freedom robotic manipulator. The dynamic model includes viscous and Coulomb friction and input generalized torques are bounded. With respect to the bounded‐error and standard learning processes applied to a virtual PUMA 560 robot (Unimation, Inc. Danburry, CT, USA), simulation results are presented in order to verify maximal tracking errors, convergence and applicability of the proposed learning control. Copyright © 2013 John Wiley & Sons, Ltd.     

Non‐minimum phase switched systems: HOSM‐based fault detection and fault identification via Volterra integral equation

In this paper, the problem of continuous and discrete state estimation for a class of linear switched systems with additive faults is studied. The class of systems under study can contain non‐minimum phase zeroes in some of their ‘operating modes’. The conditions for exact reconstruction of the discrete state are given using structural properties of the switched system. The state space is decomposed into the strongly observable part, the non‐strongly observable part, and the unobservable part, to analyze the effect of the unknown inputs. State observers based on high‐order sliding mode to exactly estimate the strongly observable part and Luenberger‐like observers to estimate the remaining parts are proposed. For the case when the exact estimation of the state cannot be achieved, the ultimate bounds on the estimation errors are provided. The proposed strategy includes a high‐order sliding‐mode‐based fault detection and a fault identification scheme via the solution of a Volterra integral equation. The feasibility of the proposed method is illustrated by simulations. Copyright © 2013 John Wiley & Sons, Ltd.     

Data‐driven model reference control with asymptotically guaranteed stability

This paper presents a data‐driven controller tuning method that includes a set of constraints for ensuring closed‐loop stability. The approach requires a single experiment and can also be applied to nonminimum‐phase and unstable systems. The tuning scheme generates an estimate of the closed‐loop output error that is used to minimize an approximation of the model reference control problem. The correlation approach is used to deal with the influence of measurement noise. For linearly parameterized controllers, this leads to a convex optimization problem. A sufficient condition for closed‐loop stability is introduced, which can be included in the optimization problem for control design. As the data length tends to infinity, closed‐loop stability is guaranteed. The quality of the estimated controller is analyzed for finite data length. The effectiveness of the proposed method is demonstrated in simulation as well as experimentally on a laboratory‐scale mechanical setup. Copyright © 2010 John Wiley & Sons, Ltd.     

Network‐based fault detection for discrete‐time state‐delay systems: A new measurement model

In this paper, the fault detection problem is studied for a class of discrete‐time networked systems with multiple state delays and unknown input. A new measurement model is proposed to account for both the random measurement delays and the stochastic data missing (package dropout) phenomenon, which are typically resulted from the limited capacity of the communication networks. At any time point, one of the following cases (random events) occurs: measurement missing case, no time‐delay case, one‐step delay case, two‐step delay case, …, q‐step delay case. The probabilistic switching between different cases is assumed to obey a homogeneous Markovian chain. We aim to design a fault detection filter such that, for all unknown input and incomplete measurements, the error between the residual and weighted faults is made as small as possible. The addressed fault detection problem is first converted into an auxiliary H_∞ filtering problem for a certain Markovian jumping system (MJS). Then, with the help of the bounded real lemma of MJSs, a sufficient condition for the existence of the desired fault detection filter is established in terms of a set of linear matrix inequalities (LMIs). A simulation example is provided to illustrate the effectiveness and applicability of the proposed techniques. Copyright © 2007 John Wiley & Sons, Ltd.     

Robust fault detection and adaptive parameter identification for DC-DC converters via switched systems

In this paper, the problem of fault detection and identification for DC-DC converters is presented. First, switched systems model and fault model are analyzed based the switched characteristics of the DC-DC converters, taking the DC-DC buck converter as an example. According to the switched Lyapunov function technique, a fault detection observer and a bank of linear switched fault identification observers are designed for the switched systems. Next, the fault detection observer detects the fault based on the residual produced by the observer output and actual output. After the fault is detected, fault identification observers are activated. The location of fault is identified by comparing the residual evaluation functions. Meanwhile, the adaptive parameter identification is achieved by choosing an appropriate adaptive law. Finally, in order to show the feasibility of the fault detection and identification, the simulation results are given in this article.     

On identification methods for direct data‐driven controller tuning

In non‐iterative data‐driven controller tuning, a set of measured input/output data of the plant is used directly to identify the optimal controller that minimizes some control criterion. This approach allows the design of fixed‐order controllers, but leads to an identification problem where the input is affected by noise, and not the output as in standard identification problems. Several solutions that deal with the effect of measurement noise in this specific identification problem have been proposed in the literature. The consistency and statistical efficiency of these methods are discussed in this paper and the performance of the different methods is compared. The conclusions offer a guideline on how to solve the data‐driven controller tuning problem efficiently. Copyright © 2010 John Wiley & Sons, Ltd.     

Process fault prognosis using a fuzzy‐adaptive unscented Kalman predictor

By monitoring the future process status via information prediction, process fault prognosis is able to give an early alarm and therefore prevent faults, when the faults are still in their early stages. A fuzzy‐adaptive unscented Kalman filter (FAUKF)‐based predictor is proposed to improve the tracking and forecasting capability for process fault prognosis. The predictor combines the strong tracking concept and fuzzy logic idea. Similar to the standard adaptive unscented Kalman filter (AUKF) that employs an adaptive parameter to correct the estimation error covariance, a Takagi–Sugeno fuzzy logic system is designed to provide a better adaptive parameter for smoothing this regulation. Compared with the standard AUKF, the proposed FAUKF has the same strong tracking ability but does not suffer from the drawback of serious tracking fluctuation. Two simulation examples demonstrate the effectiveness of the proposed predictor. Copyright © 2011 John Wiley & Sons, Ltd.     

Integration of multiple model based fault detection and nonlinear model predictive fault‐tolerant control

In this paper, we present a new fault‐tolerant control system for a class of nonlinear systems with input constraints. Because of many important factors that stabilize a nonlinear model predictive controller, it can be used as a powerful controller in the event of fault occurrence. So, the reconfigurable controller is designed based on the quasi‐infinite model predictive control (QIMPC) approach as a fault‐tolerant approach. On the other hand, a fault detection and diagnosis (FDD) system is designed based on the multiple model method. The bank of extended Kalman filters (EKFs) is used to detect the predefined actuator fault and estimate the unknown parameters of a fault. When a fault is detected, the proposed FDD information is used to correct the model of the faulty system recursively and reconfigure the controller. Delay on FDD decision may lead to performance degradation or even instability for some systems. The timely proposed FDD approach will preserve stability. Moreover, a framework is presented to ensure stability when a fault occurs. The effectiveness of this method is demonstrated, in comparison with conventional nonlinear model predictive control, by two practical examples. © 2015 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

Analysis and detection of incipient faults in post‐fault systems subject to adaptive fault‐tolerant control

This paper presents an effective scheme for detecting incipient faults in post‐fault systems (PFSs) subject to adaptive fault‐tolerant control (AFTC). Through a survey of existing techniques, it is shown that the adaptivity of the AFTC counteracts the effect of an incipient fault in the PFS. This makes some of the conventional fault‐detection strategies, such as Beard–Jones detection filters and adaptive observers, ineffective in this situation. It is shown that the unknown input observer (UIO) is an effective tool; hence, the UIO is designed to decouple the incipient fault from the AFTC such that the fault‐detection residual is sensitive only to the incipient fault. Extensive simulation study is presented using an aircraft example to test three fault‐detection approaches; it is demonstrated that the UIO is the most effective tool in detecting the incipient fault in a PFS subject to AFTC. Copyright © 2007 John Wiley & Sons, Ltd.     

Neural observer-based small fault detection and isolation for uncertain nonlinear systems

Small faults (some weak faults with a tiny magnitude) are difficult to detect and may cause severe problems leading to degrading the system performance. This paper proposes an approach to estimate, detect, and isolate small faults in uncertain nonlinear systems subjected to model uncertainties, disturbances, and measurement noise. A robust observer is developed to alleviate the lack of full state measurement. Using the estimated state, a dynamical radial basis function neural networks observer is designed in form of LMI problem to accurately learn the function of the inseparable mixture between modeling uncertainty and the small fault. By exploiting the knowledge obtained by the learning phase, a bank of observers is constructed for both normal and fault modes. A set of residues is achieved by filtering the differences between the outputs of the bank of observers and the monitored system output. Due to the noise dampening characteristics of the filters and according to the smallest residual principle, the small faults can be detected and isolated successfully. Finally, rigorous analysis is performed to characterize the detection and isolation capabilities of the proposed scheme. Simulation results are used to prove the efficacy and merits of the proposed approach.     

Threshold computation for fault detection in linear discrete‐time Markov jump systems

This paper proposes a threshold computation scheme for an observer‐based fault detection (FD) in linear discrete‐time Markovian jump systems. An observer‐based FD scheme typically consists of two stages known as residual generation and residual evaluation. Even information of faults is contained inside a residual signal, a decision of faults occurrence is consequently made by a residual evaluation stage, which consists of residual evaluation function and threshold setting. For this reason, a successful FD strongly depends on a threshold setting for a given residual evaluation function. In this paper, Kalman filter (KF) is used as a residual generator. Based on an accessibility of Markov chain to KF, two types of residual generations are considered, namely mode‐dependent and mode‐independent residual generation. After that threshold is computed in a residual evaluation stage such that a maximum fault detection rate is achieved, for a given false alarm rate. Without any knowledge of a probability density function of residual signal before and after fault occurrence, a threshold is computed by using an estimation of residual evaluation function variance in a fault‐free case. Finally, a detection performance is demonstrated by a numerical example. Copyright © 2013 John Wiley & Sons, Ltd.     

A robust output error identifier for continuous‐time systems

This paper shows that the adaptive output error identifier for linear time‐invariant continuous‐time systems proposed by Bestser and Zeheb is robust vis‐à‐vis finite energy measurement noise. More precisely, it is proven that the map from the noise to the estimation error is –stable—provided a tuning parameter is chosen sufficiently large. A procedure to determine the required minimal value of this parameter is also given. If the noise is exponentially vanishing, asymptotic convergence to zero of the prediction error is achieved. Instrumental for the establishment of the results is a suitable decomposition of the error system equations that allows us to strengthen—to strict—the well‐known passivity property of the identifier. The estimator neither requires fast adaptation, a dead‐zone, nor the knowledge of an upperbound on the noise magnitude, which is an essential requirement to prove stability of standard output error identifiers. To robustify the estimator with respect to non‐square integrable (but bounded) noises, a prediction error‐dependent leakage term is added in the integral adaptation. –stability of the modified scheme is established under a technical assumption. A simulated example, which is unstable for the equation error identifier and the output error identifier of Bestser and Zeheb, is used to illustrate the noise insensitivity property of the new scheme. Copyright © 2014 John Wiley & Sons, Ltd.     

Mixed‐data multi‐modelling for fault detection and isolation

Early recognition/isolation of a faulty behaviour of a dynamic system is the main task of a fault detection and isolation (FDI). FDI methods based on adaptive probabilistic models with multiple modes represent a theoretically well justified way of solution. Their use is severely restricted by an inherent computational complexity. The complexity problem is addressed here by employing an efficient quasi‐Bayes estimation algorithm. It is directly applicable to the mixture of components created as products of factors belonging to the exponential family. It opens a novel way to deal adaptively with mixed continuous–discrete, dynamically related data. The presented theory and algorithmization are illustrated by a simple simulation example. Copyright © 2001 John Wiley & Sons, Ltd.     

Fault accommodation for multi‐input linear sampled‐data systems

A novel observer‐based fault accommodation technique for linear multi‐input multi‐output sampled‐data systems affected by a general class of actuator faults in the presence of quantization errors is addressed in the paper. Only the output signal has been assumed to be available for direct measurement. A simulation study on a three‐tank system supports theoretical developments. Copyright © 2014 John Wiley & Sons, Ltd.     

Bias compensation‐based parameter estimation for output error moving average systems

Identification problems of output error models with moving average noises are considered in this paper. The least‐squares‐based parameter estimation is biased under the colored noises in outputs. Firstly, a bias compensation term is formulated to achieve the bias‐eliminated estimates of the system parameters. Secondly, the bias compensation term is determined by the unknown variance of the noise and the unknown noise model, thus based on the hierarchical identification principle, an unbiased parameter estimation is obtained by interactively estimating noise variance and noise parameters. Finally, the estimated bias compensation term is added to the biased parameter estimates. The simulation examples confirm the effectiveness of the proposed algorithm. Copyright © 2011 John Wiley & Sons, Ltd.     

Two‐stage Kalman filter‐based actuator/surface fault identification and reconfigurable control applied to F‐16 fighter dynamics

In this study, an active fault‐tolerant control technique with reconfiguration against actuator/surface failures is presented. A two‐stage Kalman filter is designed in order to identify the control distribution matrix elements that correspond to the faulty actuator/surface; thus, the control reconfiguration is carried out using this identified control distribution matrix. The actuator/surface fault identification problem is solved through two jointly operating Kalman filters: the first one is for the estimation of the control distribution matrix elements that correspond to the faulty actuator/surface, and the second one is for the estimation of the state variables of the aircraft model. A structure for the active fault‐tolerant aircraft flight control system with reconfiguration against actuator/surface failures is presented. A control reconfiguration action is taken in order to keep the performance of the impaired aircraft the same as that of the unimpaired aircraft. In simulations, the nonlinear flight dynamics of an AFTI/F‐16 fighter model is considered, and the performance of the proposed actuator/surface failure identification and reconfigurable control schemes are examined for this model. Copyright © 2012 John Wiley & Sons, Ltd.     

Robust multiple‐fault detection and isolation: A gradient flow approach

This paper presents a novel solution to the fault detection and isolation observer design problem for linear time‐invariant systems. A gradient flow approach is proposed for synthesizing a residual generator under optimal eigenstructure assignment. This is achieved by minimizing the spectral condition number of the observer eigenvector matrix. The properties of convergence of the gradient flow solution are proved and its efficiency demonstrated via a numerical example. Copyright © 2007 John Wiley & Sons, Ltd.