Showing papers on "Robustness (computer science) published in 1983"

Order and stepsize control in extrapolation methods

Abstract: The paper presents a new theory for joint order and stepsize control in extrapolation methods. This theory defines a locally optimal order that can be determined along any trajectory to be computed. In addition, Shannon's information theory is applied to derive some ideal convergence model that is expected to describe the behavior of an extrapolation method over a large set of test problems. Extensive numerical comparisons document a drastic acceleration in stiff integration and a mild acceleration in non-stiff integration by the new device. Moreover, a significant increase in reliability, robustness, and portability of the extrapolation codes is achieved.

Robustness and efficiency properties of scatter matrices

Abstract: The robustness and efficiency properties of likelihood ratio tests for functions of the population covariance matrix are studied. An alternative class of tests based upon affineinvariant M-estimates of scatter is proposed whenever the function of the covariance matrix is invariant under a common scale change. For such inferences, it is shown that a single scalar-valued index of efficiency is sufficient.

Application of vector performance optimization to a robust control loop design for a fighter aircraft

Abstract: For a McDonnell-Douglas F-4C aircraft a robust, fixed gain controller is designed, which provides satisfying handling qualities of the longitudinal motion or the aircraft over the complete flight range without gain scheduling. Robustness is achieved in the sense of covering largo paramoter variations and providing good gain and phase margins, Only low control rates and low feedback gains are involved, The results are obtained by application of a performance vector optimization design method which accounts for a great many different design objectives simultaneously in a highly systematic fashion. Two different designs are presented placing emphasis on pitch rate control (pointing) and normal acceleration control (manoeuvring) respectively.

Achieving robustness in distributed database systems

Abstract: The problem of concurrency control in distributed database systems in which site and communication link failures may occur is considered. The possible range of failures is not restricted; in particular, failures may induce an arbitrary network partitioning. It is desirable to attain a high “level of robustness” in such a system; that is, these failures should have only a small impact on system operation.A level of robustness termed maximal partial operability is identified. Under our models of concurrency control and robustness, this robustness level is the highest level attainable without significantly degrading performance.A basis for the implementation of maximal partial operability is presented. To illustrate its use, it is applied to a distributed locking concurrency control method and to a method that utilizes timestamps. When no failures are present, the robustness modifications for these methods induce no significant additional overhead.

Architecture and Design of a Reliable Token-Ring Network

Abstract: Architecture, performance, transmission system, and wiring strategy of a token-ring local area network implemented at the IBM Zurich Research Laboratory are described. In the design of the system, particular emphasis was placed on high reliability, availability, and serviceability. To ensure robustness of the token-access protocol, we employ the concept of a monitor function which is responsible for fast recovery from access-related errors. Our protocol supports asynchronous transmission of data frames concurrently with full-duplex synchronous channels, e.g., for voice services or other applications requiring guaranteed delay. The delay-throughput performance of the token ring is shown to depend very little on data rate and distance. The transmission system of the ring is fully bit synchronous and allows insertion/removal of stations in/from the ring at any time. A mixed ring/star wiring strategy is used which provides the means for both fault detection and isolation, and system reconfiguration, and allows wiring of a building systematically.

Robustness of queuing network formulas

Abstract: Analyuc models of queuing networks have been observed to give good results for systems which do not conform to the assumptions of classical queuing theory This has been explained to some extent by the recently proposed concept of operational analysis However, the use of operational analysis for prediction of system performance revolves a restrictive assumption, that of homogeneous service tnnes (HST), and practical systems often deviate from HST behavior. It is shown that the main system performance measures are surpnsmgly insensitive to violations of the HST assumption. This further explains the robustness of analyacal models for predicting performance of queuing networks Some issues regarding the operational versus stochasuc approaches are resolved, since it is clatmed that the present analysis, which is carried out m the operational framework, cannot be meaningfully carried out m the stochastic framework. Categones and Subject Descnptors: D.4.8 [Operating Systems] Performance--modeling andpredtctwn; operatwnal analysis;, stochastic analysts General TermsExperimentation, Measurement, Performance Additional

Robustness optimization of linear feedback systems

Abstract: The robustness of a linear, finite-dimensional, time-invariant, single-input single-output feedback system with respect to plant variations is analyzed and expressed in terms of the sensitivity function of the feedback system and its complement. The analysis leads to defining optimally robust control systems as control systems for which the supremum over all frequencies of a weighted combination of the sensitivity function and its complement is minimal. Besides accomodating robustness considerations, the approach allows handling specifications on the closed-loop bandwidth and controller pole excess. A complete solution of the problem is given. Specialization of the general formulation to the minimization of the supremum of the sensitivity function or its complement alone affords determining the achievable performance of a single-input single-output plant. This performance depends on the plant pole-zero configuration and is characterized by the so-called regulability number of the plant. The ideas are illustrated with an example involving a double pendulum system in four different equilibrium positions.

Robust stability of systems with integral control

Abstract: A necessary and sufficient condition is derived which must be satisfied by the plant steady state gain matrix of a linear time invariant system in order for an integral controller to exist for which the closed loop system is unconditionally stable. Based on this theorem 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.

Score Tests for Regression Models

Abstract: Many score statistics used for testing trends in means, either based on the observed values of the response variables or their ranks, can be expressed in terms of quadratic forms. This article explores systematically the robustness of optimality of these statistics, as well as the optimal properties of the associated tests.

Recent Computational Studies in the Interpretation of Structure from Motion

Abstract: The interpretation of structure from motion is the recovery of three-dimensional shape from the transformation induced in the image by objects in motion. This paper reviews, compares, and extends the results obtained to date in the computational study of the interpretation of structure from motion. Comparisons are also made between perspective versus orthographic, velocity-based versus position-based, and discrete versus continuous interpretation schemes. In comparing the different schemes, two main issues are discussed. The first is the uniqueness problem: under what conditions a two-dimensional transformation has a unique three-dimensional interpretation. The second is the robustness problem: whether the interpretation be stable under small changes in the input. With respect to the uniqueness problem, the conditions that are known to guarantee the uniqueness of the solution are summarized. With respect to the robustness problem, it is argued that schemes based on the instantaneous velocity field are unstable when applied locally.

Robust Control of Flexible Spacecraft

Abstract: A well-designed feedback control system exhibits the properties of external disturbance attenuation and performance robustness with respect to plant uncertainty. The plant uncertainties of flexible spacecraft include unmodeled dynamics and parameter uncertainties. Singular value robustness measures are used to compare performance and stability robustness properties of different control design techniques in the presence of residual modal interaction (control and observation spillover) for a design example which is representative of a practical flexible spacecraft system. The control designs evaluated include linear quadratic geometry (LQG) control, integral feedback, bias removal control, innovations feedthrough, and frequency-shaped LQG.

Non-Parametric Regression

Abstract: One transformation that get us around many assumptions about distributions, is the rank distribution. In this, one or more variables are replaced by their ranks. The rank transformation simply assigns the value 1 to the smallest observed value, 2 to the next smallest, etc. Of course in practice there are often ties, and some sort of intelligent tie-breaker needs to be employed. Usually this is to assign both values their average rank.

Minimax Robust Discrete-Time Matched Filters

Abstract: The problem of designing finite-length discrete-time matched filters is considered for situations in which exact knowledge of the input signal and/or noise characteristics is not available. Such situations arise in many applications due to channel distortion, incoherencies, nonlinear effects, and other modeling uncertainties. In such cases it is often of interest to design a minimax robust matched filter, i.e., a nonadaptive filter with an optimum level of worst-case performance for the expected uncertainty class. This problem is investigated here for three types of uncertainty models for the input signal, namely, the mean-absolute, mean-square, and maximum-absolute distortion classes, and for a wide generality of norm-deviation models for the noise covariance matrix. Some numerical examples illustrate the robustness properties of the proposed designs.

An Investigation of the Robustness of the t Test to Unit of Analysis Violations

Abstract: In the first part of this study, sampling experiments were used to assess the Type I error rates of the t test in situations where classes or schools were randomly assigned to groups but analyses were carried out on individual student scores. It was found that Type I error rates were grossly inflated under all conditions investigated. In the second part of the study, mathematical analysis was used to demonstrate the fact that even small amounts of between class (or school) variation can cause large inflations in the Type I error rate of the t test when analyses are incorrectly carried out on individual student scores.

Design Of The Measurement System For State Estimation In The Norwegian High-Voltage Transmission Network

Abstract: A method proposed by Koglin has been used to find an optimal measurement system for static state estimation in the Norwegian high-voltage transmission network. The method is extended in order to obtain a more robust solution in situations with loss of measurements and bad data, and to reduce the computational effort required. The significance of choice of estimation method, load flow situation and network topology is discussed. In addition the optimal measurement system has been compared with manually designed and arbitrarily generated systems.

Bayesian Approaches to Outliers and Robustness

Abstract: A general, Bayesian approach to robustification via model elaboration is introduced and discussed. The approach is illustrated by considering the elaboration of standard models to incorporate the possibility of non-standard distributional shapes or of individual aberrant observations (outliers). Influence functions are then considered from a Bayesian point of view and an approach to robust time series analysis is outlined.

Application of the one-dimensional Fourier transform for tracking moving objects in noisy environments

Abstract: In Riddle and Rajala (1981), an algorithm was presented which operates on an image sequence to identify all sets of pixels having the same velocity. The algorithm operates by performing a transformation in which all pixels with the same two-dimensional velocity map to a peak in a transform space. The transform can be decomposed into applications of the one-dimensional Fourier transform and therefore can gain from the computational advantages of the FFT. The aim of this paper is the concern with the fundamental limitations of that algorithm, particularly as relates to its sensitivity to image-disturbing parameters as noise, jitter, and clutter. A modification to the algorithm is then proposed which increases its robustness in the presence of these disturbances.

Fractional representation theory: Robustness results with applications to finite dimensional control of a class of linear distributed systems

Abstract: This paper reviews and extends fractional representation theory. In particular, new and powerful robustness results are presented. This new theory is utilized to develop a preliminary design methodology for finite dimensional control of a class of linear evolution equations on a Banach space. We design for stability in an input-output sense but pay particular attention to internal stability as well.

The Robustness of a Hierarchical Model for Multinomials and Contingency Tables

Abstract: Publisher Summary This chapter focuses on the robustness of a hierarchical model for multinomial and contingency tables. Hierarchical Bayesian models can be used in a completely Bayesian manner or in a manner that has been called semi-, or pseudo-, or quasi Bayesian. In the latter case, the hyperparameters, or possibly the hyperhyperparameters, etc., are estimated by non-Bayesian methods, or at least by some not purely Bayesian method such as by maximum likelihood. In the so-called non-Bayesian statistics, the use of the Ockham-Duns razor is sometimes called the principle of parsimony, and it encourages one to avoid having more parameters than are necessary. In hierarchical Bayesian methods, one similarly uses a principle of parsimony or hyper-razor. There are at least two different ways to test a model. One is by means of significance tests after observations are made. Another is by examining the robustness of a model, that is, by seeing if small changes in the model lead to small changes in the implications. Tests for robustness can sometimes be carried out by the device of imaginary results before making observations. When calculating a Bayes factor F1, reasonable robustness with respect to the choice of hyperhyperparameters has also been found.

Advanced methods of recursive time-series analysis

Abstract: Two of the most advanced procedures for recursively estimating the parameters in linear, observation space models of stochastic dynamic systems are the prediction error (PE) and optimal generalized equation error (OGEE) methods. This paper discusses the relationship between these methods in the case of the transfer function (TF) or Box-Jenkins model ; and compares their performance in terms of optimality, computational complexity and practical robustness. While the methods are quite similar for the TF model, it is shown that the instrumental-variable inspired OGEE approach yields algorithms that are computationally simpler and more robust when applied in practical situations.

Robustness studies in adaptive control

Abstract: This paper examines robustness issues in Model Reference Adaptive Control systems in the presence of unmodeled dynamics and output disturbances. We present an approximate technique, trend analysis, by which we can study the evolution of the parameter error trajectory under periodic excitation. This analysis provides new insights upon the size and spectral content of the excitation sufficient to guarantee local stability.

Algorithms for signal reconstruction from short-time Fourier transform magnitude

Abstract: We have previously established a number of conditions under which a real signal can be uniquely reconstructed from its STFT magnitude. For the STFT magnitude to be a practical signal representation, we need robust reconstruction algorithms. In this paper, we discuss such a class of algorithms within the framework of sequential extrapolation techniques. Such techniques reconstruct the short-time sections of a signal in an order determined by their positions on the time axis. We find that compared to direct reconstruction, the robust algorithms are less sensitive to roundoff errors. To further test the robustness of these algorithms, we applied them to STFT magnitudes which were purposely modified for accomplishing signal processing tasks such as noise reduction and time-scale modification of speech.

Pulse Shape Design for Teletext Data Transmission

Abstract: This paper studies the problem of designing a suitable pulse shape for teletext data transmission. The following four criteria are used: 1) Nyquist I criterion, 2) Nyquist II criterion, 3) degree of overshoots in the channel signal, and 4) robustness to sampling phase jitter. For system bandwidths less than the inverse-baud rate, it is not possible to satisfy all these criteria simultaneously; tradeoffs that have to be made are illustrated. Several candidate pulse shapes are given and a composite criterion developed. A pulse shape, which satisfies the Nyquist I criterion and is closest to satisfying the Nyquist II criterion, in a sum-of-squares-of-deviations sense, is recommended.

Systolic array for recursive least-squares minimisation

Abstract: A systolic array for performing recursive least-squares minimisation is described. It performs an orthogonal triangularisation of the data matrix using a sequence of Givens rotations, and generates the required residual without having to solve the associated triangular linear system by back-substitution. Since the back-substitution process may be ill conditioned and numerically unstable, the reliability and robustness of the method is greatly improved as a result, whilst the amount of circuitry and computation is reduced.

Estimating time and time-lag in time-of-flight velocimetry.

Abstract: Estimating time and time-lag in time-of-flight velocimeters is investigated. Statistics of a filtered Poisson point process is given. A Maximum Likelihood Estimator is compared with suboptimum estimators in terms of robustness. For a dominating background combined spatial and temporal processing can improve the robustness compared with purely temporal processing. Schemes for the spatial filters are given.

Robust stability of interconnected systems

Abstract: The concept of generalized block diagonal dominance is used to extend a number of recently developed robust stability results for multivariable systems to large-scale interconnected systems.

A foundation of information geometry

Abstract: This paper proposes a new geometrical basic theory for the field of information science. the traditional theory for the information system is occupied fully by detailed discussions of the properties of a system, and the importance has not been recognized for the mutual relationship within a set of information systems. There are many cases in engineering where a set of information systems must be considered by a single model (for example, in the cases of statistical and system models). to discuss the properties of a model, the mutual relations among the elements must be described. In such a case, the geometrical structure among the information systems, which are elements of the model (e.g., distance, linearity and curvature), is important. This paper recognizes first the family of probability distributions as a manifold. Then it is shown that the structures such as Riemannian metric, α-pseudo-distance including a real parameter α, and α-affine connection, can be introduced naturally. the differential geometrical structure of the space is described. Using the proposed method, the differential geometrical structure inherent in the set of information sources or set of systems can be recognized. the structure is related closely to the system identification problem (parameter estimation), approximation and robustness problems.

Robust Multivariable Control System Design using Optimal Reduced Order Plant Models

Abstract: A procedure for the design of preccapensators for linear multivariable continuous-time systems based on reduced-order plant models is described. The reducedorder model is obtained via an optimal Hankel-norn approximation method, whose approximation error can be accurately predicted. The closed-loop stability of the full-order plant and compensator can then be deduced from this approximation error and a robust stability result.