Search

A new method for modeling the thermal response of semiconductor devices such as light-emitting diodes (LEDs) from temperature measurements is presented. The method uses a realization-theoretic approach to identification combined with convex optimization methods. Linear matrix inequalities are constructed to guarantee that the identified discrete-time model has strictly real eigenvalues between 0 and 1, so that, when converted to continuous time, the model will have strictly real time constants. Additional optional time-domain constraints are developed to guarantee a predetermined steady-state value, guarantee no undershoot or overshoot in the transient response, and/or guarantee a positive time-constant spectrum. The method is applied to the thermal response of a high-power LED. Temperature measurements of the semiconductor device are used to model the time constants of the thermal response and characterize the relative contribution of each time constant to temperature increase. Experiments indicate how the proposed method can be used to detect thermal defects. It is shown that models with five time constants can model the thermal effects of the LEDs used with high accuracy. The proposed method is applicable to larger order systems with multiple simultaneous temperature measurements.

An input shaping algorithm based on convex optimization techniques is presented for the design of reference and feedforward signals in a closed-loop discrete-time linear time-invariant system. The proposed algorithm allows closed-loop signals to be subjected to linear constraints on amplitude and rate of change. As an illustrative example the seeking process in a hard disk drive is investigated. The closed-loop system response to both shaped and non-shaped inputs are compared. The computational scheme was experimentally tested on a modified hard disk drive.

In this paper, we study the limitations of scheduling an internal model to reject disturbances with a time-varying frequency. Hence, any adaptive method that uses scheduling and frequency estimation is also limited in the same manner. The limitations of scheduling are to investigate by posing and solving the problem of rejecting periodic disturbances from a multichannel system when the frequencies of the periodic disturbances are changing rapidly in time by designing an scheduled controller that satisfies the internal model principle. The periodic disturbances are modeled by a sum of sinusoids and the frequencies of the disturbances are used for scheduling the controller. It is shown that a controller that regulates input additive disturbances may not regulate the same disturbances added to the output of the system. This is in contrast to the classical case where the frequency of the disturbances is constant. Copyright © 2011 John Wiley & Sons, Ltd.

A dynamic model of the resistance heater element in a thermal flying height control (TFC) slider of a hard disk drive is identified and used for dynamic flying height control. Experimental data obtained on a spin stand and a generalized realization algorithm are used for identification of a discrete-time dynamic model of the thermal actuator. The flying height change is measured in two different ways: using servo burst information written onto the disk surface and using information written in the data sectors only. The resistance change of the thermal actuator based on the input power level is measured. Based on the identified discrete-time model of the heater and using convex optimization techniques, a computational scheme is proposed to obtain optimized feedforward input profiles to the heater element that minimize repeatable flying height variations and enable low flying heights.

Active noise control Feedforward control Parameter estimation Digital filters Control applications a b s t r a c t This paper discusses the design methodology for the active noise control of sound disturbances in a forced-air cooling system. The active sound cancellation algorithm uses the framework of output-error based optimization of a linearly parametrized filter for feedforward sound compensation to select microphone location and demonstrate the effectiveness of active noise cancellation in a small portable data projector. Successful implementation of the feedforward based active noise controller on a NEC LT170 data projector shows a 20–40dB reduction per frequency point in the spectrum of external noise of the forced-air cooling system can be obtained over a broad frequency range from 1 to 5kHz. A total noise reduction (unweighted) of 9.3dB is achieved.

In recent years, track densities of magnetic hard disks have continued to grow. A promising approach to continuing the trend to ultrahigh density is using bit-patterned media (BPM). However, the implementation of BPM in hard disk drives (HDDs) to achieve high recording density is challenging and requires various new techniques, such as new servo pattern designs and position error signal (PES) decoding schemes. In applying BPM in HDDs, it is important to select a servo pattern providing sufficient PES quality for head positioning. In this paper, we discuss evaluation of PES quality and servo pattern performance from a closed-loop (servo) point of view in order to evaluate the quality of several servo patterns. We consider three servo patterns (the amplitude pattern, chevron pattern, and differential frequency pattern) as case studies. We developed a PES simulation tool to provide a realistic HDD track-following simulation. Because of PES nonlinearity in the amplitude servo pattern, we considered time-based servo patterns as alternatives. For time-based servo patterns, we found that readback signal sampling and transition jitter greatly affect PES quality. Therefore, we conclude that the differential frequency servo pattern is superior to other patterns, since it is less sensitive to transition jitter and readback signal sampling.

The General Realization Algorithm (GRA) is developed to identify modal parameters of linear multi-degree-of-freedom dynamic systems subjected to measured (known) arbitrary dynamic loading from known initial conditions. The GRA extends the well known Eigensystem Realization Algorithm (ERA) based on Hankel matrix decomposition by allowing an arbitrary input signal in the realization algorithm. This generalization is obtained by performing a weighted Hankel matrix decomposition, where the weighting is determined by the loading. The state-space matrices are identified in a two-step procedure that includes a state reconstruction followed by a least squares optimization to get the minimum prediction error for the response. The statistical properties (i.e., bias, variance, and robustness to added output noise introduced to model measurement noise and modeling errors) of the modal parameter estimators provided by the GRA are investigated through numerical simulation based on a benchmark problem with non-classical damping.

The design of an optimal internal model-based (IMB) controller by extending standard discrete time optimal control theory for IMB controllers is described. The optimal observer and state feedback gains of the IMB controller are given via the solution of discrete time algebraic Riccati equations. The design method is applied to an acoustic system that is subjected to disturbances from a server fan. Periodic disturbances from the server fan appear as harmonics of the fundamental frequency of the fan. Parametric models for the plant and non-periodic part of the disturbance are identified from experimental data. An internal model is designed in discrete time and the internal model principle is used to design a feedback controller that rejects periodic disturbances in the acoustic system. The controller is implemented in real-time and successfully attenuates the first four harmonics of the fan noise.

In this paper, we consider the iterative learning control (ILC) framework to design a reference signal that directly cancels periodic disturbances in a feedback measurement. Cancellation of periodic disturbances is useful in reducing undesirable repeatable tracking errors in applications such as the two-stage servo track writing process for disk drives. A general problem description is given for a linear discrete-time periodic system and convergence results for the learning system are derived. A learning filter is designed with the use of a finite-impulse response model approximation for the inverse of the closed-loop sensitivity such that convergence is achieved in learning a reference signal that provides cancellation with periodic perturbations affecting the system measurement. The ILC algorithm is applied to a disk drive system where experimental results demonstrate the effectiveness of the method in reducing periodic measurement disturbances.

The basic principle of feedforward noise cancellation for broadband active noise reduction is based on the availability of a reliable measurement of the noise source. Reliability is compromised in case the noise measurement is acoustically coupled to the actual active noise cancellation (ANC), as stability and performance of the feedforward compensation becomes ambiguous. This paper presents a framework to recursively estimate a feedforward filter in the presence of acoustic coupling, addressing both stability and performance of the active feedforward noise cancellation algorithm. The framework is based on fractional model representations in which a feedforward filter is parameterized by coprime factorization. Conditions on the parameterization of the coprime factorization formulated by the existence of a stable perturbation enables stability in the presence of acoustic coupling. In addition, the paper shows how the stable perturbation can be estimated on-line via a recursive least-squares estimation of a generalized FIR filter to improve the performance of the feedforward filter for ANC.

In a hard disk drive with a dual-stage actuator the standard rotary actuation of the voice coil motor (VCM) is combined with an additional micro- or milli-actuation to accomplish high-bandwidth and highly accurate track following. For extremely high density recording, the servo control algorithm designed for a specific dual-stage actuator needs to maintain high performance track following in lieu of external disturbances and uncertainties in the dynamic response induced by product manufacturing tolerances. Essential in the robust design of a servo control algorithm is to bound these uncertainties and design a robust track following servos accordingly. In this paper, we present an experiment based methodology to characterize and bound the uncertainties in the dynamic response of a dual-stage actuator in the form of an uncertainty model. Subsequently, the uncertainty model is used for the analysis and synthesis of a robust dual-stage servo controller that is able to optimize position error variance in the presence of uncertainty information.

Studies on iterative identification and model based control design have shown the necessity for identifying models on the basis of closed-loop data. Estimating models on the basis of closed-loop data requires special attention due to cross correlation of noise and input signals and the possibility to estimate unstable systems operating under a stabilizing closed-loop controller. This paper provides a method to perform an approximate identification of normalized coprime factorization from closed-loop data. During the identification, a constrained linear regression parametrization is used to estimate the normalized coprime factors. A servomechanism case study illustrates the effectiveness of the proposed algorithm.

This paper details the design of an adaptive internal model-based controller to attenuate periodic disturbances in the presence of random noise. The internal model-based controller is designed in two steps, making use of the separation principle to design a controller with the desired properties that stabilizes the closed-loop system. Periodic disturbances appear as harmonics of a fundamental frequency. The fundamental frequency of the disturbance is estimated with a magnitude/phase-locked loop and is used to adjust the parameters of the controller, specifically the natural frequency of the internal model and the feedback gain. An active noise control study shows the ability of the proposed method to cancel time varying disturbances in an acoustic system.

To achieve frequency separation between actuators of a dual-stage servo actuator system in a hard disk drive, the PQ control design method can be used. In the PQ method, the servo controller is designed by focusing on structural interference of the actuators. This paper provides a specific parametrization for fixed order feedback controllers used in the PQ-method for which the structural interference between actuators can be minimized. The proposed control design method is applied to a dual-stage instrumented suspension, where the additional sensor of the instrumented suspension is used to attenuate high frequency disturbance caused by windage turbulence in a hard disk drive.

In this paper a method is proposed to design an optimal controller that uses an internal model and an additive noise model to reject periodic disturbances. The proposed design method allows for the design of a feedback controller that rejects periodic disturbances in the presence of random noise. An active noise cancellation application for an HP DL380 G3 rack server cooling fan is presented to demonstrate the efiectiveness of the proposed design method. System identiflcation techniques are used to flnd a model of the non-periodic disturbances. The feedback controller was implemented in real-time to show the efiectiveness of periodic blade pass frequency sound reduction. Copyright c ∞ 2005 IFAC.

A feedforward control algorithm for active noise control based on the recursive estimation of a generalized finite impulse response (FIR) filter is presented in this paper. Recursive least square estimation (RLSE) with variable forgetting factors is applied to a commercial air ventilation silencer to provide the online estimation of the generalized finite impulse response (FIR) filter to obtain active noise compensation in an airduct. The advantage of the generalized FIR filters lies in the possibility to include prior knowledge of system dynamics in the tapped delay line of the filter. It is shown that a significant improvement in noise cancellation is obtained with the implementation of generalized FIR filter for feedforward ANC.

Commercially available mechanical systems are available to teach and demonstrate the principles behind dynamics and control. A single system can be used for basic dynamic analysis in an undergraduate class to teaching and applying sophisticated identification techniques in a graduate class. In this paper it shown how a commercial system is used at the undergraduate level to estimate lumped parameter coefficients using multiple step responses and first principle modeling. At the graduate level. the same commercial system is used to teach concepts of system identification for the estimation of models for a multi-degree of freedom mechanical system.

This paper summarizes the organization of a graduate course on control related topics in identification. The course is taught at the department of Mechanical and Aerospace Engineering at the University of California, San Diego and analyzes the problems of estimation on closed-loop data and control relevant approximation of systems. As part of the course, several case studies are reviewed and include the well documented examples of a sugar cane crushing mill and the identification of the marginally stable electromechanical system found in a CD-ROM player.

In this paper, we consider the problem of estimating low order and control relevant models of plant dynamics and additive noise dynamics on the basis of closed-loop experiments. Estimating low order models for both the plant and noise dynamics is important in control design applications that focus on disturbance rejection. Several methods for low order model estimation on the basis of closed-loop data exist in the literature, but fail to address the simultaneous estimation of low order noise models that are relevant in disturbance control problems. In this paper we evaluate and compare some of these methods and propose a new methodology that extends the results to low order noise model estimation. The new methodology is an extended two-stage method where the first stage is used to estimate high order models for filtering purposes. In the second stage, filtered signals are used for low order model approximation. The methodology is illustrated in a realistic simulation study based on the windage disturbance reduction of a flexible hard disk drive suspension.

This paper addresses the problem of checking the consistency of experimental closed-loop frequency-domain data with uncertainty models that are structured using coprime factorizations. The uncertainty models presented in this paper use the knowledge of a stabilizing feedback controller to structure and formulate the uncertainty on a model. Subsequently, the controller dependent coprime factor uncertainty model can be used to formulate model (in)validation tests on the basis of closed-loop data. Closed-loop model validation results are developed for the cases of noise-free and noise perturbed closed-loop data. The model validation tests involve the computation of a structured singular value over a finite frequency grid. It is also shown that the computation of the structured singular value simplifies considerably when the feedback controller used for the closed-loop experiments is the same as the controller used for formulating the controller dependent coprime factor uncertainty model.

To maximize bandwidth for a hard disk servo system in the presence of modeling uncertainty, a tradeoff between performance and controller complexity is addressed by control-relevant modeling and robust control. Control-relevant modeling is accomplished by using the information of the appropriate weighting functions used in the control design to find a model via frequency domain curve fitting. This paper illustrates control relevant modeling using frequency domain data and design of a high bandwidth servo system in the presence of modeling uncertainty. Results using real-time implementation show agreement between the designed and achieved bandwidths

In this paper a framework for an iterative procedure of identification and robust control design is introduced wherein the robust performance is monitored during the subsequent steps of the iterative scheme. By monitoring the performance via a model-based approach, the possibility to guarantee performance improvement in the iterative scheme is being employed. In order to monitor achieved performance (for a present controller) and to guarantee robust performance (for a future controller), an uncertainty set is used where the uncertainty structure is chosen in terms of model perturbations in the dual Youla parametrization. This uncertainty structure is shown to be particularly suitable for the control performance measure that is considered. The model uncertainty set can be identified by an uncertainty estimation procedure on the basis of closed-loop experimental data. To obtain performance robustness, robust control design tools are used to synthesise controllers on the basis of the identified uncertainty set.

A systematic design procedure is presented that can be used to construct servo controllers for dual-stage actuation in a hard disk drive. The systematic approach is based on optimal control theory using an H/sub /spl infin//-norm-based control design, where servo design specifications are formulated in terms of relatively simple weighting functions. Different possible servo control strategies are designed and compared on performance and energy consumption. Results based on a piezoelectric milli-actuator from Hutchinson Technology Inc. illustrate the design of a dual-stage servo controller that is able to perform high-density track recording, with a significant reduction in power consumption.

This paper discusses the approximate and feedback relevant parametric identification of the radial servo system present in a Compact Disc player. In this application the problem of approximate identification based on data from closed loop experiments will be analyzed to find a finite dimensional linear time invariant discrete time model, suitable for model-based control design. The feedback relevant identification in this paper is based on the algebraic theory of fractional representations, which has lead to a framework for equivalent open loop identification of (normalized) coprime plant factors and a manageable approximate transfer function estimation of the feedback controlled plant. A mixed worst-case/probabilistic approach to model uncertainty quantification is used to construct an upper bound on the uncertainty of the normalized coprime factors being estimated. Both the nominal model and the uncertainty description are used to design an enhanced robust controller.

Control in an uncertain environment often involves a trade-off between exploratory actions, whose goal is to gather sensory information, and "regular" actions which exploit the information gathered so far and pursue the task objectives. In principle both types of action can be modeled by minimizing a single cost function within the framework of stochastic optimal control. In practice however this is difficult, because the control law must be sensitive to estimation uncertainty which violates the certainty-equivalence principle. In this paper we formalize the problem in a way which captures the essence of the exploration-exploitation trade-off and yet is amenable to numerical methods for optimal control. The key to our approach is augmenting the dynamics of the partially-observable plant with the Kalman filter dynamics, thus obtaining a higher-dimensional but fully-observable plant. The resulting control laws compare favorably to other more ad-hoc approaches. Our formalism is also suitable for modeling human behavior in tasks which benefit from active exploration.

We present a data-based approach for modeling and controller design of a dual-stage tape servo actuator. Our method uses step response measurements and a generalized realization algorithm to identify a multivariable discrete-time model of the actuator. The data acquisition and modeling can be implemented in the servo firmware of a tape drive. We have designed a dual-stage controller, based on the model, using loop shaping techniques adopted for multivariable control problems. We applied the procedure to the prototype of a dual-stage actuator tape head to reduce the effect of lateral tape motion. The prototype consists of a conventional voice coil motor for coarse positioning and a micro-actuator for fine positioning. The micro-actuator, which is mounted on the voice coil motor, uses a piezo crystal to follow high-frequency lateral tape motion (up to the kilohertz regime), while the voice coil motor follows only low-frequency lateral tape motion. Compared to a single-stage design, the dual-stage servo design provides a 25% bandwidth improvement and a voice coil motor control signal that is much smaller in magnitude.

Research in control design on the basis of identified models has lead to many iterative algorithms where performance weights in a control objective function are determined a-priori and remain fixed in order to evaluate performance improvement during iterations. This paper considers the case where performance weights are adjusted in iterative identification and control schemes where performance robustness is maintained during each iteration to yield a cautious control design. In order to adjust performance weights, a nominal model along with uncertainty description obtained from a system identification procedure are used to adjust the performance weighting function conjointly with robust control synthesis. A framework is developed for measuring performance comparison between iterations which allows adjustment of the performance weights during the iterative cautious control design.

This paper presents a method of finding a control relevant, low-order weighting filter that over bounds the uncertainty variations in measured frequency domain data. Frequency response measurements of hard disk drive actuators are used to model manufacturing and mounting variation. A polynomial positivity condition is use to guarantee that a minimal phase representation of the uncertainty that is found and linear constraints in the frequency domain are used to bound the magnitude of uncertainty model. The positivity condition can be formulated as a linear matrix inequality and linear constraints are used to shape the filter. Manufacturing and mounting variations in hard disk drives are modeled as a multiplicative uncertainty and a low-order over bound is found from data on a finite grid. This application compares a linear programming based methods for finding uncertainty models with the purpose of designing a robust controller.

This paper presents two distinct methods of designing scheduled repetitive controllers for the case where the frequency of the disturbance is slowly varying and when it is rapidly varying. In both cases the plant is subjected to random and periodic disturbances. The first method uses set points to design a family of controllers and then interpolates between them online. The conditions for designing sub-optimal gains are given. The second method utilizes results from time-varying Riccati equations to design a time-varying observer. It is shown via simulation that the time-varying Riccati method is superior but at a greater computational cost.

This paper revisits a pair of linear matrix inequalities (LMIs) that are related to checking a frequency domain inequality (FDI) over a finite interval. The first contribution is to show that the proposed pair of LMIs contain the original formulation of the Kalman-Yakubovich-Popov Lemma when the coefficient matrix is constant. The coefficient matrix can be made affine on the frequency variable at no extra computational cost. The second contribution is to show how to transform the frequency variable in order to extend the proposed results to infinite frequency intervals. In applications such as robustness analysis, allowing for frequency dependent coefficient matrices can be significant in reducing conservatism, a feature which is illustrated with a simple numerical example.

The contribution of this work is to incorporate product variability into a low-order control design based on model uncertainty for hard disk drives. A method for including uncertainty into control synthesis is to design the servo controller using a worst-case Hinfin approach and mixed-sensitivty approach.

In this study we consider the iterative learning control (ILC) framework to design a reference signal that reduces the periodic component of disturbances in a feedback measurement containing both repeatable and non-repeatable components. Reduction of periodic disturbances is useful in alleviating undesirable repeatable tracking errors in applications such as the two-stage servo track writing process for disk drives. A general problem description is given for a linear discrete time system and convergence robustness results for the learning system are derived. A learning filter is designed with the use of an FIR model approximation for the inverse of the closed-loop sensitivity for fast nominal convergence while robustness to modeling errors and non-repeatable disturbances is achieved through additional filtering. The ILC algorithm is applied to a disk drive system where experimental results demonstrate the effectiveness of the design method in reducing periodic measurement disturbances.

In this paper, a scheduled controller is presented by the merging together of internal model-based control and linear parameter-varying control theory. The internal model is used to attenuate periodic disturbances and concepts from linear parameter-varying control theory are used to design the feedback gain that quadratically stabilizes the closed loop system during scheduling. An estimate of the frequency of the disturbance is used as the scheduling variable that updates the internal model. The order of the controller is kept low by making use of the separation principle and designing a reduced order observer for state feedback. A design example of a simple mass-spring-damper system demonstrates the effectiveness of the controller to cancel periodic disturbances with a time-varying frequency.

This paper presents the results of a parametric identification method for the modeling of the dynamics of a servo actuator and the windage disturbance present in a hard disk drive. The results are obtained with an identification method that is based on a recently introduced extended two-stage method, which enables the estimation of low order complexity models of system and disturbance dynamics on the basis of closed-loop data. As a result, experimental data obtained under realistic servo conditions is used to find low order discrete time models that describe the actuator and windage disturbance dynamics at different locations on the hard disk.

A feedforward control algorithm for active noise control based on the recursive estimation of a generalized finite impulse response (FIR) filter is presented in this paper. The feedforward control algorithm is applied to a commercial air ventilation silencer to provide active noise compensation in an airduct. A generalized FIR filter has the same linear parameter structure as a taped delay FIR filter that is favorable for (recursive) estimation purposes. However, the advantage of the generalized FIR filters lies in the possibility to include prior knowledge of system dynamics in the tapped delay line of the filter. By comparison with a conventional FIR filter implementation it is shown that a significant improvement in noise cancellation is obtained.

This paper addresses the problem of checking the consistency of experimental closed-loop frequency-domain data with uncertainty models that are structured using coprime factorizations. The uncertainty models presented in this paper use the knowledge of a stabilizing feedback controller to structure and formulate the uncertainty on a model. Subsequently, the controller dependent coprime factor uncertainty model can be used to formulate model (in)validation tests on the basis of closed-loop data. The model (in)validation is performed on sample data from a flexible structure to illustrate the presented model validation results. An open-loop based uncertainty model is also used to demonstrate the benefits of closed-loop uncertainty modeling over open-loop uncertainty modeling.

This paper presents a new methodology for feedforward filter design in active noise cancellation in the presence of acoustic coupling. This is done via a feedforward filter estimation that uses a dual-Youla parametrization to incorporate the acoustic coupling. The experimental results show that the feedforward filter parametrized and estimated via dual-Youla parametrization provides an effective feedforward compensation for the active noise control in an airduct.

This paper addresses the time domain model validation problem for uncertainty models that are structured using coprime factorizations. A model validation technique is proposed in which measurement data is used to validate a derived uncertainty model. Newly proposed model validation techniques are based on a fractional representation approach and addresses the problem of correlation between the input and output signals inherent to closed-loop systems. The model validation problem for coprime factorizations is considered for closed-loop time domain data in the cases of noise-free and noisy measurements where the results rely on a linear relationship between input and output data.

Bundle methods have been proposed in the literature for solving semi-definite programming (SDP) problems which consist of minimizing a linear objective subject to linear matrix inequality (LMI) constraints. These methods are appealing especially for large scale problems. Many control problems can be cast as SDPs. In contrast fixed order control problems in the LMI framework require an additional rank constraint on the LMIs which makes the problem insolvable with traditional SDP methods. One way of imposing the rank constraint is to formulate the fixed order control problem in terms of the eigenvalues of the LMIs. This makes the problem nondifferentiable, but amenable to bundle methods.

Recently introduced methods of iterative identification and control design are directed towards the design of high performing and robust control systems. These methods show the necessity of identifying approximate models from closed loop plant experiments. In this paper a method is proposed to approximately identify normalized coprime plant factors from closed loop data. The fact that normalized plant factors are estimated gives specific advantages both from an identification and from a robust control design point of view. It will be shown that the proposed method leads to identified models that are specifically accurate around the bandwidth of the closed loop system. The identification procedure fits very naturally into the iterative identification/control design scheme as presented by Schrama (1992). >

Classical indirect methods of closed-loop identification can be applied on the basis of different closed-loop transfer functions. Here the multivariable situation is considered and conditions are formulated under which identified approximative plant models are guaranteed to be stabilized by the present controller. Additionally it is shown in which sense the classical indirect methods are generalized by a previously introduced identification method based on the dual-Youla parametrization. For stable controllers the two methods are shown to be basically equivalent to each other.

In this paper an approach is presented to estimate a linear multivariable model on the basis of (noisy) frequency domain data via a curve fitting procedure. The multivariable model is parametrized in either a left or a right polynomial matrix fraction description and the parameters are computed by using a two-norm minimization of a multivariable output error. Additionally, input-output or element-wise based multivariable frequency weightings can be specified to tune the curve fitting error in a flexible way. The procedure is demonstrated on experimental data obtained from a 3 input 3 output wafer stepper system.

For consecutive model-based control design, approximate identification of linear models should be performed on the basis of a feedback-relevant criterion, compatible with the control design. For an H/sub /spl infin//-norm based control design, a procedure is presented to estimate a possibly unstable and feedback controlled plant by using an H/sub /spl infin//-norm based feedback-relevant identification criterion. It is shown that the formulated identification problem can be handled by taking (noisy) closed loop frequency domain measurements of the plant and fitting a model of a prespecified McMillan degree, which is parametrized in a stable factorization, using a certain nonlinear constrained minimization. The procedure is illustrated by an example.

The powerful standard representation for uncertainty descriptions in a basic perturbation model based on a standard plant representation can be used to attain necessary and sufficient conditions for stability robustness within various uncertainty descriptions. In this paper these results are employed to formulate necessary and sufficient conditions for stability robustness of several uncertainty sets based on unstructured additive coprime factor uncertainty, gap-metric uncertainty as well as /spl Lambda/-gap uncertainty.

Generally, fault detection is the process of monitoring a physical dynamic system accompanied by conformation and assessment of any degradation of system performance. These systems are modelled and terms that are representative of a specific fault are identified and monitored for detection. In this paper, a fault detection algorithm is developed to isolate and detect friction changes in a high precision positioning mechanism. The designed fault detection algorithm addresses dynamic model estimation, dynamic filtering and recursive parameter estimation techniques to monitor online friction changes. The procedure is illustrated on a high precision servo pneumatic cylinder that drives a translational air bearing apparatus, designed to permit the addition of friction. Side loading of the cylinder rod induced by a friction fault causes significant loss of performance in these applications. It therefore serves to design a simple and effective online fault detection and isolation scheme for the designed experimental set-up.

Crucial to the increase in aerial densities of future hard disk drives (HDD) is the increase of the number of tracks per inch. In order to achieve very high track densities new high bandwidth and very accurate servo systems will be necessary. One possible solution is the use of dual-stage actuators, in which a high bandwidth and a highly accurate micro-actuator (MA) is used in combination with a traditional voice coil motor (VCM). In this case the nature of the servo controller becomes multivariable since both the VCM and the MA have to be controlled simultaneously. The design of servo controllers for track following with dual-stage actuators has been an active area of research in the last few years. Modern control methods were specifically developed for multivariable systems and offer significant advantages over classical control methods. On the down side, they usually lead to controllers which are of order equal to or greater than the plant they were designed for. The aim of the paper is to present a systematic approach for the design of a fixed order servo controller and to illustrate this design on a piezoelectric MA designed by Hutchinson Technologies Inc. The complexity of the servo controller is limited to address costs and implementation issues in a commercial HDD. The systematic approach for such a low order servo controller design is developed in the paper by utilizing H/sub /spl infin//-optimization with a choice of specific closed loop weighting functions and additional constraints on the order of the controller.

In order to provide a formal definition of the model validation problem, it is necessary to provide more details on an uncertainty model that consists of a nominal model and an additional model uncertainty or allowable model perturbation. The following problems are described: model validation via optimisation; uncertainty models based on coprime factor perturbation; and closed-loop model validation using coprime factorisation.

In this paper we introduce a Linear Matrix Inequality (LMI) condition for robust stability analysis. The condition is expressed as a pair of convex inequalities that provides an upper bound for the structured singular that can be used to verify stability and performance robustness. This robust analysis test incorporates a particular class of frequency-dependent multipliers and can be limited to finite frequency intervals, features which can significantly reduce conservatism as compared to existing conditions with similar complexity. The results are illustrated with a simple numerical example that illustrates the improvement of the proposed LMI condition.