Institute for Systems Research Technical Reports

Permanent URI for this collection

http://hdl.handle.net/1903/4376

This archive contains a collection of reports generated by the faculty and students of the Institute for Systems Research (ISR), a permanent, interdisciplinary research unit in the A. James Clark School of Engineering at the University of Maryland. ISR-based projects are conducted through partnerships with industry and government, bringing together faculty and students from multiple academic departments and colleges across the university.

Browse

Now showing 1 - 20 of 2409

Ambiguous Behavior of Logic Bistable Systems
(1975-10-04) Hurtado, Marco; Elliott, David L.
The standard specification of logic bistable devices do not specify the behavior under conditions in which the input is logically undefined or in which certain kinds of multiple input changes occur. These conditions are unavoidable in logic synchronizers and arbiters. A general deterministic model of bistable devices is proposed, consisting of a non-liner differential system with some adequate properties. Analysis of this model shows that bistable devices can be driven into a logically undefined region by certain admissible inputs and can remain in this region for an unbounded length of time.
DYNAMAN: A Tool for Manipulator Design and Analysis.
(1985) Sreenath, N.; Krishnaprasad, Perinkulam S.; ISR
This report describes a method to formulate the dynamic equations of a robot manipulator with N links and with either revolute or prismatic joints using a Newton-Euler formulation. A general- purpose tool for analysis and design of robot manipulators and related control problems is being developed in the form of a software package DYNAMAN using this method.
Interleaving and Channels with Unknown Memory.
(1985) Hughes, Brian; Narayan, Prakesh; ISR
Communications channels which have unknown, time-varying parameters frequently arise in practice. In certain instances, such as in jammed channels, the duration of memory is not known, and memory is not known, and may also be time-varying. Most models of jammed channels which appear in the literature, however, assume that the jammer is restricted to memoryless jamming techniques; this restriction is usually rationalized by arguing that the use of interleaving by the transmitter is sufficient to destroy channel memory. In this paper, the ability of interleaving to eliminate channel memory is investigated by comparing the capacity of jammed channels with memory and random interleaving with that of similar memoryless, jammed channels.
Real-Time Default Reasoning, Relevance, and Memory Models.
(1985) Drapkin, J.; Miller, Michael; Perlis, D.; ISR
We describe a working example of mechanical default reasoning that is rather robust, in that it can detect a conflict between a default conclusion and another assertion (or conclusion), can (under suitable conditions) decide between them, and can maintain this decision indefinitely until overridden by information to do so. Our model contains five key elements: STM, LTM, ITM, QTM, and RTM. STM, LTM, and ITM are standard parts of cognitively- based models of memory. QTM is a technical device that controls the flow of information into STM, and RTM is the repository of default resolution and relevance.
Gaussian Arbitrarily Varying Channels.
(1985) Hughes, Brian; Narayan, Prakesh; ISR
The Arbitrarily Varying Channel (AVC) can be interpreted as a model of a channel jammed by an intelligent and unpredictable adversary. In this paper, we investigate the asymptotic reliability of optimum random block codes on Gaussian Arbitrarily Varying Channels (GAVCs). A GAVC is a discrete-time, memoryless Gaussian channel with input power P_T and noise power N_e which is further corrupted by an additive "jamming signal. The statistics of this signal are unknown and may be arbitrary, except that it is subject to a power constraint, P_J. We distinguish between two types of power constraints: peak , average. For peak constraints on the input power , the jamming power, we show that the GAVC has a capacity. For the remaining cases, in which the transmitter , /or the jammer are subject to average power constraints, only LAMBDA- capacities are found. The asymptotic error probabilities suffered by optimal random codes in these cases are determined. Our results suggest that if the jammer is subject only to an average power constraint, reliable communication is impossible at any positive code rate.
Optimal Quantizer Design for Noisy Channels: An Approach to Combined Source-Channel Coding.
(1985) Farvardin, Nariman; Vaishampayan, Vinay; ISR
In this paper, we present an analysis of the zero-memory quantization of memoryless sources when the quantizer output is to be encoded and transmitted across a noisy channel. Necessary conditions for the joint optimization of the quantizer and the encoder/decoder pair are presented and a recursive algorithm for obtaining a locally optimum system is developed. The performance of this locally optimal system, obtained for the class of generalized Gaussian distributions and the Binary Symmetric Channel is compared against the Optimum Performance Theoretically Attainable (using Rate-Distortion theoretic arguments), as well as against the performance of Loyd-Max quantizers encoded using Natural Binary Codes. It is shown that this optimal design could result in substantial performance improvements. The performance improvements are more noticeable at high bit rates and for more broad-tailed densities.
A Continuous-Time Distributed Version of Wald's Sequential Hypothesis Testing Problem.
(1985) Baras, John S.; ISR
Stability of Multiparameter Singular Perturbation Problems with Parameter Bounds, 11. Time-Invariant Systems with Arbitrary Perturbations.
(1985) Abed, Eyad H.; ISR
The asymptotic stability of a general linear time-invariant multiparameter singular perturbation problem is studied with no a priori assumptions on the relative magnitudes of the small parameters. Thus the results apply, for example, to multiple time scale problems as well as to multiparameter singular perturbation problems possessing only two time scales. An explicit upper bound is obtained for a weighted norm of the vector of singular perturbation parameters such that asymptotic stability of the system is ensured if this bound is respected. In a companion paper [1], the time-varying case is studied using Liapunov stability theory, assuming that the small parameters have bounded mutual ratios (two time scales). The present paper makes use of fixed-point methods, in addition to Liapunov techniques, to arrive at the desired upper bound.
Design of a Flight Controller for an F14 Aircraft Using the DELIGHT.Mary.Lin Optimization-Based CACSD System.
(1985) Fan, Michael K-H.; Walrath, C.D.; Lee, C.; Tits, A.L.; Rimer, M.; Grant, R.; Levine, William S.; ISR
A Superlinearly Convergent Feasible Method for the Solution of Inequality Constrained Optimization Problems.
(1985) Panier, E.R.; Tits, A.; ISR
When iteratively solving optimization problems arising from engineering design applications, it is sometimes crucial that all iterates satisfy a given set of 'hard' inequality constraints, and generally desirable that the objective function value improve at each iteration. In this paper, we propose an algorithm of the successive quadratic programming (SQP) type which, unlike other algorithm of this type, does enjoy such properties. Under mild assumptions, the new algorithm is shown to converge from any initial point, locally superlinearly. Numerically tested, it has proven to be competitive with the most successful currently available nonlinear programming algorithms, while the latter do not exhibit the desired properties.
An Optimal Scheme for Two Competing Queues with Constraints.
(1985) Shwartz, A.; Makowski, Armand M.; ISR
Two types of traffic, e.g., voice and data, share a single synchronous and noisy communication channel. This situation is modeled as a system of two discrete-time queues with geometric service requirements which compete for the attention of a single server. The problem is cast as one in Markov decision theory with long-run average cost and constraint. An optimal strategy is identified that possesses a simple structure, and its implementation is discussed in terms of an adaptive algorithm that is extremely simple, recursive and easily implementable, with no prior knowledge of the actual values of the statistical parameters. The derivation of the results combines martingale arguments, results on Markov chains, O.D.E. characterization of the limit of stochastic approximations , methods from weak convergence. The ideas developed here are of independent interest , should prove useful in studying broad classes of constrained Markov decision problems.
Estimation of the Rate of a Doubly-Stochastic Time-Space Poisson Process.
(1985) Gubner, John A.; Narayan, P.; ISR
We consider the problem of estimating the rate of a doubly- stochastic, time-space Poisson process when the observations are restricted to a region D subset of R^2. In the general case, we obtain a representation of the minimum mean-square-error (MMSE) estimate in terms of the conditional characteristic function of an underlying state process. In the case D=R^2, we extend a known result to compute the MMSE estimate explicitly. For a special form of the rate process, a well-defined integral equation is presented which defines the linear MMSE estimate of the rate.
Two Case Studies in Optimization-Based Computer-Aided Design of Control Systems.
(1985) Fan, Michael K-H.; Walrath, C.D.; Tits, A.L.; Nye, W.T.; Rimer, M.; Grant, R.; Levine, William S.; ISR
There have been many approaches proposed for the computer-aided design of control systems. The co-authors of this paper include several strong proponents of a design methodology emphasizing designer's intuition, man-machine interaction and sophisticated optimization techniques. This methodology has been implemented by our group as part of the DELIGHT.Mary.Lin system, an offshoot of Berkeley's DELIGHT systems. A logical way to test this approach to computer-aided design is to apply DELIGHT.Mary.Lin to a number of real control system design problems and see how well it performs. This paper describes just such applications to two control problems supplied by industry. In both cases, the problems had been previously solved by other techniques. Thus, we are able to compare the solution obtained with DELIGHT.Mary.Lin to solutions that are acceptable in actual practice.
Decomposition and Stability of Multiparameter Singular Perturbation Problems.
(1985) Abed, Eyad H.; ISR
Time-scale separation and stability of linear time-varying and time-invariant multiparameter singular perturbation problems are analyzed. The first problem considered in the paper is that of deriving upper bounds on the small parasitic parameters ensuring the existence of an invertible, bounded transtormation exactly separating test and slow dynamics. This problem is most interesting for the time-varying case. The analysis of this problem in the time-varying case requires the two time-scale setting introduced by H.IC. Khalil and P.V. Kokotovic (SLIAM J. Control Optim., 17, 56-65, l979). This entails that the mutual ratios of the small parameters are bounded by known positive constants. The second problem considered is to derive parameter bounds ensuring that the system in question is uniformly asymptotically stable. The results on decomposition are used to facilitate the derivation of these latter bounds. Fortunately, the analysis of decomposition and stability questions for time- invariant multiparameter singular perturbation problems requires no restriction on the relative magnitudes of the small parameters. A concept of 'strong D-stability' is introduced and shown to greatly simplify the stability analysis of time- invariant multiparameter problems.
Estimation of the Rate of a Discrete-Time Multivariate Point Process.
(1985) Gubner, John A.; Narayan, P.; ISR
We introduce the notion of a discrete-time multivariate point process which can arise in the modeling of an optical communication system. We wish to estimate the rate of this process at time t given the past of the process up to time t-l. This requires the computation of a certain conditional expectation: we perform this computation by introducing an absolutely continuous change of measure and then applying the generalized Bayes' rule.
Convergence of Implicit Discretization Schemes for Linear Differential Equations with Application to Filtering.
(1985) Piccioni, M.; ISR
This paper presents a generalization of results on convergence and robustness of discretization schemes for nonlinear filtering proposed by Kushner. This is made possible by a general theorem on the convergence of semigroups of operators on a Banach space, which gives sufficient conditions for a semidiscretization scheme to remain convergent, once the time is implicitly discretized. As a consequence, sufficient conditions can be given for selecting space discretizations of the state process generator to construct computable nonlinear filters converging to the optimal one.
Local Feedback Stabilization and Bifurcation Control, I.Hopf Bifurcation.
(1985) Abed, Eyad H.; Fu, Jyun-Horng; ISR
Local bifurcation control problems are defined and employed in the study of the local feedback stabilization problem for nonlinear systems in critical cases. Sufficient conditions are obtained for the local stabilizability of general nonlinear systems whose linearizations have a pair of simple, nonzero imaginary eigenvalues. The conditions show, in particular, that generically these nonlinear critical systems can be stabilized locally. The analysis also yields a direct method for computing stabilizing feedback controls. Use is made of bifurcation formulae which require only a series expansion of the vector field. The results are easily applied since they do not involve preliminary state transformations, center manifold reduction, or Liapunov functions.
Stochastic Unit Commitment Scheduling and Dispatch of Electric Power Systems.
(1985) Yan, I.; Blankenehip, Gilmer L.; ISR
Unit commitment, including economic dispatch, is a key component of short term operation scheduling of an electric energy system. Common industry practice is based on the use of a 'priority list' for generation scheduling and a deterministic model for power/energy demand. The priority list specifies the next unit to be started or shutdown in response to an increase or decrease in load. A common problem in the use of priority lists is that the next unit is improperly sized to meet the actual change in load. The algorithm proposed here is more accurate than the priority list method , much faster than dynamic programming which can hardly be applied to systems of more 5 machines. For a system of 41 machines, the algorithm can determine schedules in 0.1 second which is test enough for on-line control. Furthermore the total generating cost is superior to that obtained by dynamic programming successive approximations.
Integration of Computer Aided Design and Computer Aided Process Planning.
(1985) Ssemakula, M.E.; ISR
In recent years, CAD/CAM systems have gained wide acceptance in industry. As experience in the use of these systems builds up, it is becoming increasingly evident that for most effective operation, they must be integrated. Process Planning, being responsible for the conversion of design speciflcation to manufacturing instructions, is essential to this integration process. Research is being conducted to enable geometric data trom a CAD database to be accessed by process planning programs for use in generating appropriate process plans for a given part. The resulting process planning data is then translated into an NC program which can be used to machine the part.
A Simple Proof of Stability on the Center Manifold for Hopf Bifurcation.
(1985) Abed, Eyad H.; ISR
A simple proof is presented for a well known fact about Hopf bifurcation: if the loss of an equilibrium point results in periodic solutions via Hopf bifurcation, then the stability ot these periodic solutions is determined by their stability on an associated center manifold. More precisely, it is shown that the characteristic exponent determining the stability of the periodic solutions is the same whether computed for the original systems or the system restricted to the center manifold. Attention is focused on the finite dimensional case of a one parameter family of ordinary differential equations. The proof consists of exhibiting a similarity transformation which uncovers the relationship between the linearized flow of the original system and that of its restriction to the center manifold. (This paper combines and revises papers TR-85-5 and TR-85-6; to appear in IEEE Trans. Automat. Cont.).

Browse

Browsing Institute for Systems Research Technical Reports by Issue Date

Results Per Page

Sort Options