Showing papers on "Robustness (computer science) published in 1985"

Robust techniques for signal processing: A survey

Abstract: In recent years there has been much interest in robustness issues in general and in robust signal processing schemes in particular. Robust schemes are useful in situations where imprecise a priori knowledge of input characteristics makes the sensitivity of performance to deviations from assumed conditions an important factor in the design of good signal processing schemes. In this survey we discuss the minimax approach for the design of robust methods for signal processing. This has proven to be a very useful approach because it leads to constructive procedures for designing robust schemes. Our emphasis is on the contributions which have been made in robust signal processing, although key results of other robust statistical procedures are also considered. Most of the results we survey have been obtained in the past fifteen years, although some interesting earlier ideas for minimax signal processing are also mentioned. This survey is organized into five main parts, which deal separately with robust linear filters for signal estimation, robust linear filters for signal detection and related applications, nonlinear methods for robust signal detection, nonlinear methods for robust estimation, and robust data quantization. The interrelationships among many of these results are also discussed in the survey.

Regression Diagnostics: An Expository Treatment of Outliers and Influential Cases

Abstract: Gauging the robustness of regression estimates is especially important in small-sample analyses. Here, we examine recent developments in the detection and analysis of outliers and influential cases...

Robust control of linear time-invariant plants using periodic compensation

Abstract: This paper considers the use and design of linear periodic time-varying controllers for the feedback control of linear time-invariant discrete-time plants. We will show that for a large class of robustness problems, periodic compensators are superior to time-invariant ones. We will give explicit design techniques which can be easily implemented. In the context of periodic controllers, we also consider the strong and simultaneous stabilization problems. Finally, we show that for the problem of weighted sensitivity minimization for linear time-invariant plants, time-varying controllers offer no advantage over the time-invariant ones.

Robustness of continuous-time adaptive control algorithms in the presence of unmodeled dynamics

Abstract: This paper examines the robustness properties of existing adaptive control algorithms to unmodeled plant high-frequency dynamics and unmeasurable output disturbances. It is demonstrated thai there exist two infinite-gain operators in the nonlinear dynamic system which determines the time-evolution of output and parameter errors. The pragmatic implication of the existence of such infinite-gain operators is that 1) sinusoidal reference inputs at specific frequencies and/or 2) sinusoidal output disturbances at any frequency (including dc), can cause the loop gain to increase without bound, thereby exciting the unmodeled high-frequency dynamics, and yielding an unstable control system. Hence, it is concluded that existing adaptive control algorithms as they are presented in the literature referenced in this paper, cannot be used with confidence in practical designs where the plant contains unmodeled dynamics because instability is likely to result. Further understanding is required to ascertain how the currently implemented adaptive systems differ from the theoretical systems studied here and how further theoretical development can improve the robustness of adaptive controllers.

Optical implementation of the Hopfield model.

Abstract: Optical implementation of content addressable associative memory based on the Hopfield model for neural networks and on the addition of nonlinear iterative feedback to a vector–matrix multiplier is described. Numerical and experimental results presented show that the approach is capable of introducing accuracy and robustness to optical processing while maintaining the traditional advantages of optics, namely, parallelism and massive interconnection capability. Moreover a potentially useful link between neural processing and optics that can be of interest in pattern recognition and machine vision is established.

Optical information processing based on an associative-memory model of neural nets with thresholding and feedback.

Abstract: The remarkable collective computational properties of the Hopfield model for neural networks [Proc. Nat. Acad. Sci. USA 79, 2554 (1982)] are reviewed. These include recognition from partial input, robustness, and error-correction capability. Features of the model that make its optical implementation attractive are discussed, and specific optical implementation schemes are given.

The gap metric: Robustness of stabilization of feedback systems

Abstract: In this paper we introduce the gap metric to study the robustness of the stability of feedback systems which may employ not necessarily stable open-loop systems. We elaborate on the computational aspects of the gap metric and provide upper bounds to the gap in cases where the exact formulas do not apply, By admissible uncertainties we mean those which preserve closed-loop stability and a specified small tolerance on the I/O behavior of a feedback system. We show that admissible uncertainties are precisely those which are constrained in the gap. Finally, we conclude that any metric which preserves a continuous relationship between open-loop systems and the corresponding stable feedback interconnections must have the topology of the gap metric.

Minimax frequency domain performance and robustness optimization of linear feedback systems

Abstract: It is shown that feedback system design objectives, such as disturbance attenuation and rejection, power and bandwidth limitation, and robustness, may be expressed in terms of required bounds of the sensitivity function and its complement on the imaginary axis. This leads to a minimax frequency domain optimization problem, whose solution is reduced to the solution of a polynomial equation.

Non-Euclidian metrics and the robust stabilization of systems with parameter uncertainty

Abstract: This paper considers, from a complex function theoretic point of view, certain kinds of robust synthesis problems. In particular, we use a certain kind of metric on the disk (the "hyperbolic" metric) which allows us to reduce the problem of robust stabilization of systems with many types of real and complex parameter variations to an easily solvable problem in non-Euclidian geometry. It is shown that several apparently different problems can be treated in a unified general framework. A new result on the gain margin problem for multivariable plants is also given. Finally, we apply our methods to systems with real zero or pole variations.

Improved measures of stability robustness for linear state space models

Abstract: In this paper, the aspect of "stability robustness" of linear systems is analyzed in the time domain. A bound on the structured perturbation of an asymptotically stable linear system is obtained to maintain stability using a Lyapunov matrix equation solution. The resulting bound is shown to be an improved bound over the ones recently reported in the literature. Also, special cases of the nominal system matrix are considered, for which the bound is given in terms of the nominal matrix, thereby, avoiding the solution of the Lyapunov matrix equation. Examples given include comparison of the proposed approach with the recently reported results.

Robust experiment design via stochastic approximation

Abstract: A new methodology is proposed for robust experiment design. It allows uncertainly in the nominal parameters of the model under study to be taken into account by assuming that these parameters belong to some population with known statistics. The mathematical expectation of the determinant of the Fisher information matrix over this population is here taken as a measure of optimality, but the expectation of other nonrobust criteria could have been considered as well. Stochastic approximation techniques are advocated as the simplest tools for optimizing these robust criteria. The efficiency of the proposed algorithms is demonstrated on simple examples—for which an analytical solution exists—as well as on more complex ones. A comparison is made with Landaw's suboptimal approach, which supports an interesting conjecture about the robustness of replicate samples.

Robust stability of systems with integral control

Abstract: A number of necessary and sufficient conditions are derived, which must be satisfied by the plant d.c. gain matrix of a linear time invariant system in order for an integral controller to exist for which the closed loop system is stable. Based on these results, the robustness of integral control systems is analyzed, i.e., the family of plants is defined which are stable when controlled with the same integral controller. Conditions for actuator/sensor failure tolerance of systems with integral control are also given. Finally, parallels are drawn between the results of this paper and the bifurcation theory of nonlinear systems.

Robustness of discrete-time direct adaptive controllers

Abstract: The problem of preserving stability of discrete-time adaptive controllers in spite of reduced-order modeling and output disturbances is addressed in this paper. Conditions for global stability (convergence of the tracking error with bounded signals) are derived for a discrete-time pole-zero placement adaptive controller where the parameter estimator is modified in terms of normalized signals. Following an input-output perpective, the overall system is decomposed into two subsystems reflecting the parameter estimation and modeling errors, respectively, and its stability is studied using the sector stability and passivity theorems. First the analysis is carried for the class of disturbances and reference inputs that are either decaying or can be exactly hulled by a linear controller of the chosen structure. In this L 2 -framework, it is shown that the only substantive assumption to assure stability is the existence of a linear controller such that the closed-loop transfer function verifies certain conicity conditions. The convergence speed and alertness properties of various parameter adaptation algorithms regarding this condition are discussed. The results are further extended to a broader class of L_{\infty} disturbances and reference inputs.

The Use of Maximin Efficiency Robust Tests in Combining Contingency Tables and Survival Analysis

Abstract: A test is the maximin efficiency robust test for a family of possible models underlying the data if no other test has higher minimum efficiency relative to the asymptotically optimum test for each model. Methods used to examine the efficiency robustness of rank tests for the two-sample problem are adapted to obtain maximin efficiency robust procedures for testing the equality of proportions across several 2×2 tables, for combining the results of tests for trend in several 2 x J tables in which the dose-response function is one of a set of possible monotone functions, and to analyze censored survival data when either the Wilcoxon or log-rank may be appropriate. In the survival setting the robust test has maximin efficiency 93.3% relative to the Wilcoxon or long-rank when each is optimum, in contrast to the 75% relative efficiency each statistic has when the other is optimum.

Robust adaptive control: Conditions for global stability

Abstract: An input-output approach is presented for analyzing the global stability and robustness properties of adaptive controllers to unmodeled dynamics. The concept of a tuned system is introduced, i.e., the control system that could be obtained if the plant were known. Comparing the adaptive system to the tuned system results in the development of a generic adaptive error system. Passivity theory is used to derive conditions which guarantee global stability of the error system associated with the adaptive controller, and ensure boundedness of the adaptive gains. Specific bounds are presented for certain significant signals in the control systems. Limitations of these global results are discussed, particularly the requirement that a certain operator be strictly positive real (SPR)-a condition that is unlikely to hold due to unmodeled dynamics.

Theory and application of an extended horizon self‐tuning controller

Abstract: A robust version of the self-tuning regulator is developed. The regulator, which requires relatively little knowledge of system characteristics (estimated order of transfer function polynomials and an upper bound for transportation delays), has been shown to yield stable control and convergence for linear, time-invariant systems. Simulations and practical tests on a large pilot-scale process have shown that the inclusion of a variable forgetting factor and an “extended horizon” control criterion provides the regulator with a sufficient degree of robustness and flexibility to perform well in a nonlinear time-varying environment. The regulator makes use of intuitively easy-to-understand concepts and leaves few degrees of freedom for the potential user. Furthermore, extensive experiments and simulation studies have shown it to be insensitive to choice of initial conditions and dynamic characteristics set by the user.

A Mu-Test for Robustness Analysis of a Real-Parameter Variation Problem

Abstract: A new interconnection structure is constructed for applying ?-tests to the robust stability problem associated with real parameter variations. Some remarks are made concerning the computability of upper bounds for real-?, and an application of the theory to analysis of the lateral axis flight control system of the space shuttle is given.

An alternative formulation for an optimum beamformer with robustness capability

Abstract: In the paper, an alternative formulation for an optimum beamformer with a robustness capability against directional errors is presented. With this approach, the width of the main beam can be specified and a compromise can be reached between a reasonable signal acceptance angle and the ability of the beamformer to reject directional interferences. Furthermore, based on a partitioned processor interpretation, the new beam-former gives a clue to a way of reducing the complexity of a full processor.

Linear prediction of speech with a least absolute error criterion

Abstract: Most methods for linear prediction of speech are based on a least square error criterion (L 2 criterion). In this paper, it is shown that the robustness of a least absolute error criterion (L 1 criterion) is particularly well adapted to the analysis of voiced sounds. The well-known linear programming L 1 algorithms have two important drawbacks: the stability of the synthesis filter is not guaranteed, and the computation load is too heavy for real time analysis. All these problems are alleviated by the new L 1 algorithm proposed in this paper. Experiments show comparable efficiency with classical L 2 methods.

Image gathering and processing - Information and fidelity

Abstract: In this paper we formulate and use information and fidelity criteria to assess image gathering and processing, combining optical design with image-forming and edge-detection algorithms. The optical design of the image-gathering system revolves around the relationship among sampling passband, spatial response, and signal-to-noise ratio (SNR). Our formulations of information, fidelity, and optimal (Wiener) restoration account for the insufficient sampling (i.e., aliasing) common in image gathering as well as for the blurring and noise that conventional formulations account for. Performance analyses and simulations for ordinary optical-design constraints and random scences indicate that (1) different image-forming algorithms prefer different optical designs; (2) informationally optimized designs maximize the robustness of optimal image restorations and lead to the highest-spatial-frequency channel (relative to the sampling passband) for which edge detection is reliable (if the SNR is sufficiently high); and (3) combining the informationally optimized design with a 3 by 3 lateral-inhibitory image-plane-processing algorithm leads to a spatial-response shape that approximates the optimal edge-detection response of (Marr's model of) human vision and thus reduces the data preprocessing and transmission required for machine vision.

Robust jump linear quadratic control: A mode stabilizing solution

Abstract: This correspondence is concerned with optimal feedback control of a class of linear stochastic systems. For a quadratic performance index the solution to this problem is known but the resulting optimal regulator does not ensure stability in every mode of operation. To improve robustness, a new mode stabilizing solution is defined. An example illustrates the advantages of the proposed controller.

On the stability robustness of the continuous-time LQG optimal control

Abstract: The robustness properties of the LQG optimal controller that is obtained by the "asymptotic recovery" method are discussed. It is shown that in spite of the good recovered stability margins the resulting closed-loop design is inherently unrobust in the sense that small parameter variations in the plant and/or small realization errors in the controller may lead to an instantaneous instability of the design.

Robust nonlinear feedback control for robotic manipulators

Abstract: Nonlinear feedback control algorithms for manipulators utilise the complete (coupled and nonlinear) dynamic model to decouple the robot joints. In this framework, we present a comparative evaluation of the computed-torque and direct-design control algorithms, and outline the practical problems introduced by modelling inaccuracies, unmodelled dynamics and parameter errors. We then formulate the α-computed-torque nonlinear feedback control algorithm which is robust in the presence of the aforementioned error sources. Numerical experiments with cylindrical robots confirm the robustness and applicability of our α-computed-torque algorithm.

Augmented Models for Statistical Fault Isolation in Complex Dynamic Systems

Abstract: The equation error approach to fault isolation implies the statistical testing of balance equation errors. In this paper, some substantial extensions to the existing methodology are proposed, including - generalized linear dynamic models - the concept of statistical isolability - the idea of and an algorithm for model augmenting - fault sensitivity analysis and filtering

A multiloop robust controller design study using singular value gradients

Abstract: A method for designing robust feedback controllers for multiloop systems is presented. Robustness is characterized in terms of the minimum singular value of the system return difference matrix at the plant input. Analytical gradients of the singular values with respect to design variables in the controller are derived. A cumulative measure of the singular values and their gradients with respect to the design variables are used with a numerical optimization technique to increase the system's robustness. Both unconstrained and constrained optimization techniques are evaluated. Numerical results are presented for a two-input/two-output drone flight control system.