Showing papers on "Robustness (computer science) published in 1987"
TL;DR: A novel method—generalized predictive control or GPC—is developed which is shown by simulation studies to be superior to accepted techniques such as generalized minimum-variance and pole-placement and to be a contender for general self-tuning applications.
3,576 citations
TL;DR: It is shown that a simple scaling of the projection of tentative weights, in the subspace orthogonal to the linear constraints, can be used to satisfy the quadratic inequality constraint.
Abstract: Adaptive beamforming algorithms can be extremely sensitive to slight errors in array characteristics. Errors which are uncorrelated from sensor to sensor pass through the beamformer like uncorrelated or spatially white noise. Hence, gain against white noise is a measure of robustness. A new algorithm is presented which includes a quadratic inequality constraint on the array gain against uncorrelated noise, while minimizing output power subject to multiple linear equality constraints. It is shown that a simple scaling of the projection of tentative weights, in the subspace orthogonal to the linear constraints, can be used to satisfy the quadratic inequality constraint. Moreover, this scaling is equivalent to a projection onto the quadratic constraint boundary so that the usual favorable properties of projection algorithms apply. This leads to a simple, effective, robust adaptive beamforming algorithm in which all constraints are satisfied exactly at each step and roundoff errors do not accumulate. The algorithm is then extended to the case of a more general quadratic constraint.
1,851 citations
TL;DR: In this article, simulated vegetation data were used to assess the relative robustness of ordination techniques to variations in the model of community variation in relation to environment, and the results clearly demonstrated the ineffectiveness of linear techniques (PCA, PCoA), due to curvilinear distortion.
Abstract: Simulated vegetation data were used to assess the relative robustness of ordination techniques to variations in the model of community variation in relation to environment. The methods compared were local non-metric multidimensional scaling (LNMDS), detrended correspondence analysis (DCA), Gaussian ordination (GO), principal components analysis (PCA) and principal co-ordinates analysis (PCoA). Both LNMDS and PCoA were applied to a matrix of Bray-Curtis coefficients. The results clearly demonstrated the ineffectiveness of the linear techniques (PCA, PCoA), due to curvilinear distortion. Gaussian ordination proved very sensitive to noise and was not robust to marked departures from a symmetric, unimodal response model. The currently popular method of DCA displayed a lack of robustness to variations in the response model and the sampling pattern. Furthermore, DCA ordinations of two-dimensional models often exhibited marked distortions, even when response surfaces were unimodal and symmetric. LNMDS is recommended as a robust technique for indirect gradient analysis, which deserves more widespread use by community ecologists.
1,501 citations
10 Jun 1987
TL;DR: Among the problems solved are: simultaneous arbitrary pole assignment for a finite number of systems by a single GSHF controller, exact model matching, and decoupling, and optimal noise rejection.
Abstract: This paper investigates the use of generalized sampled-data hold functions (GSHF) in the control of linear time-invariant systems. The idea of GSHF is to periodically sample the output of the system, and generate the control by means of a hold function applied to the resulting sequence. The hold function is chosen based on the dynamics of the system to be controlled. This method appears to have several advantages over dynamic controllers: it has the efficacy of state feedback without the requirement of state estimation; it provides the control system designer with substantially more freedom; and it requires few on-line computations. This paper focuses on four questions: pole assignment, specific behavior, noise sensitivity, and robustness. Among the problems solved are: simultaneous arbitrary pole assignment for a finite number of systems by a single GSHF controller, exact model matching, decoupling, and optimal noise rejection. Examples are given.
444 citations
TL;DR: This paper rigorously derive several basic properties of a simple discrete-time single integrator loop sigma-delta modulator with an accumulate-and-dump demodulator and shows that when the input is constant, the state sequence of the integrator in the encoder loop can be modeled exactly as a linear system in an appropriate space.
Abstract: Oversampled sigma-delta modulation has been proposed as a practical implementation for high rate analog-to-digital conversion because of its simplicity and its robustness against circuit imperfections. To date, mathematical developments of the basic properties of such systems have been based either on simplified continuous-time approximate models or on linearized discrete-time models where the quantizer is replaced by an additive white uniform noise source. In this paper, we rigorously derive several basic properties of a simple discrete-time single integrator loop sigma-delta modulator with an accumulate-and-dump demodulator. The derivation does not require any assumptions on the correlation or distribution of the quantizer error, and hence involves no linearization of the nonlinear system, but it does show that when the input is constant, the state sequence of the integrator in the encoder loop can be modeled exactly as a linear system in an appropriate space. Two basic properties are developed: 1) the behavior of the sigma-delta quantizer when driven by a constant input and its relation to uniform quantization, and 2) the rate-distortion tradeoffs between the oversampling ratio and the average mean-squared quantization error.
373 citations
01 Mar 1987
TL;DR: These experimental results demonstrate that the adaptive controller enjoys essentially the same level of robustness to unmodelled dynamics as a PD controller, yet achieves much better tracking accuracy than either PD or computed-torque schemes.
Abstract: Earlier work (Slotine and Li, 1986) exploits the particular structure of manipulator dynamics to develop a simple, globally convergent adaptive algorithm for trajectory control problems. The algorithm does not require measurements or estimates of the manipulator's joint accelerations, nor inversion of the estimated inertia matrix. This paper demonstrates the approach on a high-speed 2 d.o.f. semi-direct-drive robot. It shows that the manipulator mass properties, assumed to be initially unknown, can be precisely estimated within the first half second of a typical run. Similarly, the algorithm allows large loads of unknown mass properties to be precisely manipulated. Further, these experimental results demonstrate that the adaptive controller enjoys essentially the same level of robustness to unmodelled dynamics as a PD controller, yet achieves much better tracking accuracy than either PD or computed-torque schemes. Its superior performance for high speed operations, in the presence of parameter uncertainties, and its relative computational simplicity, make it a attractive option both to address complex industrial tasks, and to simplify high-level programming of more standard operations.
372 citations
TL;DR: In this paper, the robust stability analysis problem in linear state-space models with structured uncertainty is considered and lower bounds on allowable perturbations which maintain the stability of a nominally stable system are derived.
Abstract: In this note, we consider the robust stability analysis problem in linear state-space models. We consider systems with structured uncertainty. Some lower bounds on allowable perturbations which maintain the stability of a nominally stable system are derived. These bounds are shown to be less conservative than the existing ones.
367 citations
10 Jun 1987
TL;DR: It is shown that the required minimax optimization can be recast as a linear program for uncertainty descriptions which provide impulse response models as affine functions of uncertain parameters.
Abstract: Concepts of model predictive control are extended to uncertain linear systems. An on-line optimizing control scheme is developed which has as its objective the minimization of the worst-case tracking error for a family of linear plants. For uncertainty descriptions which provide impulse response models as affine functions of uncertain parameters, it is shown that the required minimax optimization can be recast as a linear program. Situations which lead to such an uncertainty description are discussed. An example is presented to demonstrate the properties of the proposed control scheme.
340 citations
TL;DR: In this paper, a simple adaptive controller can be implemented in a large number of complex control systems, without requiring the order of the plant or the pole excess as prior knowledge, in order to satisfy the almost positivity condition.
Abstract: Simple adaptive control systems were recently shown to be globally stable and to maintain robustness in the presence of disturbances if the controlled plant is ‘almost strictly positive real’, namely, if there exists a positive definite static output feedback (unknown and not needed for implementation) such that the resulting closed-loop transfer function is strictly positive real.
This paper is an attempt to show in an intuitive way how to use parallel feedforward and the stabilizability properties of systems in order to satisfy the ‘almost positivity’ condition. The feedforward configuration may be stationary, if some prior knowledge is given, or adaptive, in general.
This way, simple adaptive controllers can be implemented in a large number of complex control systems, without requiring the order of the plant or the pole excess as prior knowledge
299 citations
TL;DR: The algorithmic changes made do ensure the robustness of the approach, but introduce additional algorithmic difficulties, the solutions of which are also presented.
Abstract: A technical description of the algorithms employed in the modified quadtree mesh generator is given. Although the basis of the mesh generator is the same as the original version developed by Yerry and Shephard,1,2 the actual algorithms on which it is built have been entirely changed for the purpose of ensuring the robustness of the technique. As demonstrated in the paper the algorithmic changes made do ensure the robustness of the approach, but introduce additional algorithmic difficulties, the solutions of which are also presented. In addition to examples showing the capability of the mesh generator, the linear computational growth rate of the mesh generator is demonstrated.
TL;DR: It is shown that a minimax method of efficiency measurement through chance-constrained programming methods can be suitably applied for the case of chance constraints on the basis of stochastic variations of input and output data.
TL;DR: An adaptive algorithm is developed for finding the intersection curve(s) of pairs of rectangular parametric patches which are continuously differentiable, controlled by a set of tolerances.
TL;DR: A modification of the classical Simulated Annealing algorithm for the macro-cell placement problem is proposed for implementation on multiprocessor systems and experimental results show that the new algorithm obtains results comparable in quality to those of the single processor version.
Abstract: A modification of the classical Simulated Annealing algorithm for the macro-cell placement problem is proposed for implementation on multiprocessor systems. The algorithm has been implemented on the Sequent Balance 8000, a multiprocessor system with a shared-memory architecture. Experimental results show that the new algorithm obtains results comparable in quality to those of the single processor version; processor utilization is greater than 80 percent using up to eight processors.
TL;DR: This paper demonstrates under general conditions the robustness of the t-test in that the maximum actual level of significance is close to the declared level.
Abstract: One may encounter the application of the two independent samples t-test to ordinal scaled data (for example, data that assume only the values 0, 1, 2, 3) from small samples. This situation clearly violates the underlying normality assumption for the t-test and one cannot appeal to large sample theory for validity. In this paper we report the results of an investigation of the t-test's robustness when applied to data of this form for samples of sizes 5 to 20. Our approach consists of complete enumeration of the sampling distributions and comparison of actual levels of significance with the significance level expected if the data followed a normal distribution. We demonstrate under general conditions the robustness of the t-test in that the maximum actual level of significance is close to the declared level.
01 Mar 1987
TL;DR: An approach to designing controllers for dynamic hybrid position/force control of robot manipulators is presented, and preliminary experimental results are given.
Abstract: An approach to designing controllers for dynamic hybrid position/force control of robot manipulators is presented, and preliminary experimental results are given. Dynamic hybrid control is an extension of an approach proposed by M.H. Raibert and J.J. Craig (1981) to the case where the full manipulator dynamics is taken into consideration and the end-effector constraint is explicitly given by the constraint hypersurfaces. This design method consists of two steps. The first step is the linearization of the manipulator dynamics by nonlinear state feedback. Formulation of the constraint by the constraint hypersurfaces plays an essential role in establishing the linearizing law. The second step is the design of position and force controllers for the linearized model using the concept of two-degrees-of-freedom servocontroller. The merit of this servocontroller is that it can take account of both the command response and the robustness of the controllers to modeling errors and disturbances. Preliminary experiments using a SCARA robot show the validity of the approach. >
01 Jan 1987
TL;DR: This paper gives a survey on methods for the detection and localization of sensor and component faults of uncertain dynamic systems that make use of analytical redundancy and allow to detect and localize faults with the aid of a digital computer.
Abstract: This paper gives a survey on methods for the detection and localization of sensor and component faults of uncertain dynamic systems. In contrast to the commonly used techniques of hardware redundancy these methods make use of analytical redundancy and, thereby, allow to detect and localize faults with the aid of a digital computer. They comprise single, multiple or hierarchical state estimation using Luenberger observers or Kalman filters. An issue of particular relevance is the consideration of parameter uncertainties or parameter variations of the process. Several proposals are discussed to reduce the effects of parameter variations. Moreover, results from computer simulations and a practical technical application to the control of an inverted pendulum are reported.
TL;DR: This Wiener-based displacement estimation algorithm provides a linear least-squares estimate of the update using N observations to obtain a reliable displacement estimate, and has proven to be very successful to compensate motion in some typical video conferencing scenes.
TL;DR: In this paper, a robustness approach for the single period plant layout problem under uncertainty is proposed. But the robustness method is not suitable for the case of single period plants, and the effectiveness of this approach is illustrated by a numerical example.
Abstract: The single period plant layout problem under uncertainty is discussed. Solution procedures for dealing with this problem, mainly the robustness approach, are recommended. The effectiveness of the robustness approach to this problem is illustrated by a numerical example.
TL;DR: In this article, a new method for combining multiple forecasts is proposed, which offers either theoretical richness or empirical robustness, but not both together, and can be used to combine multiple forecasts.
Abstract: Existing approaches to combining multiple forecasts generally offer either theoretical richness or empirical robustness, but not both together. In this paper, we propose a new method for combining ...
TL;DR: A nonlinear feedback multivariable controller is used to implementMultivariable tracking in a nonlinear system and it appears that the resulting control scheme may have advantages over others which have been proposed in the robotics literature.
Abstract: A nonlinear feedback multivariable controller is used to implement multivariable tracking in a nonlinear system. The tracking error is measured by a general function of system state and the input command. The controller is robust in the sense that the tracking error is ultimately bounded in the presence of modeling errors. Free parameters, which affect the form of the controller, allow flexibility in determining such factors as: the size of the ultimate bound, the rate of error decay, excursion of the control, conditions on the class of modeling errors, and the level of system gain. Restrictive assumptions on the structure of the model and the modeling errors are required but they are treated in a transformation framework which allows the generalization of similar conditions which appear in the prior literature. These assumptions hold for robotic manipulators. This application is investigated at some length and it appears that the resulting control scheme may have advantages over others which have been proposed in the robotics literature.
10 Jun 1987
TL;DR: In this article, a method for designing Smith predictor controllers that provide robust performance despite real parameter uncertainties in the process model is presented, and an exact bound for the magnitude of multiplicative uncertainty used to approximate simultaneous uncertainties in process gain, time-constant, and time-delay is derived.
Abstract: A method is outlined for designing Smith predictor controllers that provide robust performance despite real parameter uncertainties in the process model Insight into the design process is gained by viewing the Smith predictor from the perspective of Internal Model Control Performance requirements are written in terms of a frequency-domain weight restricting the magnitude of the closed-loop sensitivity function A general method for approximating multiple parameter uncertainties by a single multiplicative uncertainty is developed - an exact bound is derived for the magnitude of multiplicative uncertainty used to approximate simultaneous uncertainties in process gain, time-constant, and time-delay Three controller design methods are demonstrated The first method locates loop transfer-function uncertainty regions to test for robust performance - real parameter uncertainties are considered exactly The second tuning method approximates real parameter uncertainties by multiplicative uncertainty and uses structured singular value analysis to guarantee robust performance Finally, the Smith predictor controller is compared with the Structured-Singular-Value-optimal controller
01 Dec 1987-Graphical Models \/graphical Models and Image Processing \/computer Vision, Graphics, and Image Processing
TL;DR: Three examples, namely template matching and absolute and relative orientation of cameras, demonstrate that the measures provide objective quality measures that make intuitive evaluation precise and that they seem to besuitable for automatic quality control of mensuration problems encountered in computer vision.
Abstract: The analysis of a mensuration problem aims at an evaluation of the suitability of the design of the measuring process for a specific task and at an assessment of the actually obtained measurements and of their influence onto the result. The concept of quality control, as it has been developed by the Netherlands geodesist W. Baarda is outlined. This theory provides objective quality measures, which take the geometry of the design and the used estimation and testing procedure into account: The evaluation of the design is based on measures for the precision, the controllability, and the robustness, which themselves can be used for planning purposes. The evaluation of the data is based on a statistical test, the estimated size of possible blunders and on the influence of the observed values onto the result. Three examples, namely template matching and absolute and relative orientation of cameras, demonstrate that the measures make intuitive evaluation precise and that they seem to besuitable for automatic quality control of mensuration problems encountered in computer vision.
01 Oct 1987
TL;DR: In this article, an algorithm for performing regularized set operations on polyhedral solids is described, which is achieved by adding symbolic reasoning as a supplemental step that compensates for possible numerical uncertainty.
Abstract: We describe an algorithm for performing regularized set operations on polyhedral solids. Robustness of this algorithm is achieved by adding symbolic reasoning as a supplemental step that compensates for possible numerical uncertainty. The algorithm has been implemented, and our experience with the implementation is discussed.
TL;DR: A survey of robustness of nonlinear state feedback is given, showing that under mild restrictions an optimal controller can tolerate an infinite increase in gain and loss and some results for systems linear in the control.
TL;DR: A review of recent results in robust control is presented in this paper, where only nonadaptive or non-self-tuning solutions to the problem of controlling uncertain systems are reviewed.
Abstract: A historical review of recent results in robust control is presented. The robust control problem, i.e., the problem of designing accurate control systems in the presence of significant plant uncertainties, is classical. However, over the past 15 years, significant new theory has been developed for the solution of this problem, especially with respect to linear multivariable systems characterized in the frequency domain, and the term robust control for this classical problem is only of recent vintage (1972). Some of the major contributions of modern robust control theory include the development of synthesis techniques for robust stabilization, and H2and H^{\infty} sensitivity optimization of multivariable systems. In this review, we confine the term robust control to the design of fixed controllers. Thus, only nonadaptive or nonself-tuning solutions to the problem of controlling uncertain systems are reviewed. Finally, it should be noted that the review is largely limited to the literature published in IEEE journals and conference proceedings, and some English-language journals. It is, of course, recognized that many significant related contributions have appeared elsewhere.
TL;DR: In this paper, the authors examined the robust, multivariable control of large space structures by controllers designed on a reduced-order model using positivity concepts, and compared three different controller methodologies: the familiar multivariability control, individual modal control, and individual sensor control.
Abstract: This paper examines the robust, multivariable control of large space structures by controllers designed on a reduced-order model using positivity concepts. Controllers are designed using the DRAPER I and DRAPER II structures. Three different controller methodologies are compared: the familiar multivariable control, individual modal control, and individual sensor control. Controller robustness is measured qualitatively from the plots of the minimum singular value of the return difference matrix as a function of the frequency. All controllers, when designed to give the same total average control cost, have a very similar line-of-sight response. In addition, closed-loop stability can be maintained in the event of sensor and/or actuator failure.
Proceedings Article•
01 Jan 1987TL;DR: A network for matrix inversion based on the concept of Hopfield's neural network was designed, and implemented with electronic hardware, and is readily applicable to solving a linear simultaneous equation efficiently.
Abstract: Inverse matrix calculation can be considered as an optimization. We have demonstrated that this problem can be rapidly solved by highly interconnected simple neuron-like analog processors. A network for matrix inversion based on the concept of Hopfield's neural network was designed, and implemented with electronic hardware. With slight modifications, the network is readily applicable to solving a linear simultaneous equation efficiently. Notable features of this circuit are potential speed due to parallel processing, and robustness against variations of device parameters.
TL;DR: Using the theory of uncertain dynamical systems, robust non-linear control strategies, with guaranteed tracking properties that can be quantified given bounds on the extent of model uncertainty, sensor noise, input disturbances, etc., are derived.