Search

The main goal of this research was to develop and present a general, efficient, mathematical, and theoretical based methodology to model nonlinear forced-vibrating mechanical systems from time series measurements. A system identification modeling methodology for forced dynamical systems is presented based on a dynamic system theory and a nonlinear time series analysis that employ phase space reconstruction (delay vector embedding) in modeling dynamical systems from time series data using time-delay neural networks. The first part of this work details the modeling methodology, including background on dynamic systems, phase space reconstruction, and neural networks. In the second part of this work, the methodology is evaluated based on its ability to model selected analytical lumped-parameter forced-vibrating dynamic systems, including an example of a linear system predicting lumped mass displacement subjected to a displacement forcing function. The work discusses the application to nonlinear systems, multiple degree of freedom systems and multiple input systems. The methodology is further evaluated on its ability to model an analytical passenger rail car predicting vertical wheel/rail force using a measured vertical rail profile as the input function. Studying the neural modeling methodology using analytical systems shows the clearest observations from results, providing prospective users of this tool an understanding of the expectations and limitations of the modeling methodology. [DOI: 10.1115/1.2981096]

This paper presents a wave-based modeling and control approach for suppressing torsional vibrations of an elastic shaft driven by a motor at one end and an inertial load at the other end. A two-port network model representing the dynamics of torsional displacements is obtained starting from the partial differential equations governing a shaft in torsion. By incorporating appropriate boundary conditions, the infinite dimensional transfer function of the system is obtained. Furthermore, the system is represented by delay elements that can be used for simulation purposes. An inversion-based controller that can be used to suppress vibrations while rotating the shaft according to a specific trajectory is then developed. The performance of the controller is further studied using numerical simulations. [DOI: 10.1115/1.2776338] Keywords: torsional vibrations, infinite dimensional systems, vibration control two-port networks, time-delay systems

Sufficient conditions for passivity (positive realness) are obtained for continuous-time, linear, retarded, and neutral-type systems. A delay-dependent solution is given in terms of linear matrix inequalities (LMIs) by using a descriptor model transformation of the system and by applying Park's inequality for bounding cross terms. A memoryless state-feedback solution is derived. Numerical examples are given which illustrate the effectiveness of the new theory. Index Terms--Delay-dependent criteria, linear matrix inequalities (LMIs), positive-real lemma, time-delay systems.

New delay-dependent/delay-independent bounded real criteria are derived for linear continuous-time systems with delay in the dynamics and in the objective function. A sufficient condition for the system to possess [H.sub.[infinity]]-norm that is less than a prescribed level, is given in terms of a linear matrix inequality (LMI). The proposed criteria are less conservative than other existing criteria since they are based on an augmented model that is equivalent to the original system and since they require bounds for fewer terms. We apply the new bounded real criteria to state-feedback [H.sub.[infinity]] control. The advantage of the new criteria is demonstrated by four examples. The first two compare our results with those obtained in the literature for the bounded real lemma and the other two examples treat the state-feedback control problem and compare our result with recently published designs. Index Terms--Bounded real lemma, delay-dependent criteria, [H.sub.[infinity]]-state-feedback control, linear matrix inequalities (LMIs), time-delay systems.

This note considers the class of discrete-time Markovian jump linear system with norm-bounded uncertainties and time-delay, which is dependent on the system mode. Linear matrix inequality (LMI)-based sufficient conditions for the stability, stabilization and [H.sub.[infinity]] control are developed. A numerical example is worked out to show the usefulness of the theoretical results. Index Terms--Discrete-time Markovian jump linear system, [H.sub.[infinity]] control, linear matrix inequality (LMI), time-delay system.

The use of a frequency-selective method for time delay estimation between two channel data was described. The proposed algorithm, based on a stochastic gradient technique, exhibits a low numerical complexity and provides an alternative to data prefiltering and time delay estimation when the source and noise signals have different spectral properties. Time delay estimation was achieved through parameterization of the impulse response in the delay. Experimental results showed that the time delay estimate can be easily performed with an appropriate choice of prefilter.

A novel procedure based on a deconvolution approach with an additive regularization term was proposed for multipath time-delay detection and estimation. The objective is to estimate the number of paths and the associated attenuation and delay for each path. The approach allows the transformation of the underlying optimization problem into a quadratic program that makes it easily and quickly solved by standard programs. The procedure proved capable of handling more severe cases than other methods.

In this paper, a new comb filter design method using fractional sample delay is presented. First, the specification of the comb filter design is transformed into that of fractional delay filter design. Then, conventional finite impulse response (FIR) and allpass filter design techniques are directly applied to design fractional delay filter with transformed specification. Next, we develop a constrained fractional delay filter design approach to improve the performance of the direct design method. Finally, several design examples and an experiment of the power line interference removal in electrocadiogram (ECG) signal are demonstrated to illustrate the effectiveness of this new design approach. Index Terms - Comb filter, fractional delay filter, harmonic interference removal.

In large chips, the propagation delay of the data and clock signals can limit performance due to long resistive interconnect. The insertion of repeaters alleviates the quadratic increase in propagation delay with interconnect length while decreasing power dissipation by reducing short-circuit current. In order to develop a repeater design methodology, a timing model characterizing a complementary metal-oxide-semiconductor (CMOS) inverter driving a resistance-capacitance (RC) load is presented. The model is based on the Sakurai short-channel [Alpha]-power law model of transistor operation. The inverter model is applied to the problem of repeaters to produce design expressions for determining the optimum number of uniformly sized repeaters to be inserted along a resistive interconnect line for reduced delay. For a wide variety of typical RC loads, this analytical repeater model exhibits a maximum error of 16% as compared to a dynamic circuit simulator (SPICE). The advantage of uniformly sized repeaters versus tapered-buffer repeaters is also investigated using the repeater model presented in this paper. It is shown that uniform repeaters remain advantageous over tapered buffers and tapered-buffer repeaters even with relatively small resistive RC loads. An expression for the short-circuit power dissipation of a repeater driving an RC load is presented. A comparison of the short-circuit power dissipation to the dynamic power dissipation in repeater chains and related power/delay tradeoffs are made. Index Terms - Buffer insertion, delay optimization, RC interconnect, repeaters.

A multipath signal is considered as a multiple regression model, with each time sample a possible arrival time of a pulse. The regression stepwise procedure is used to select true pulses and eliminate false pulses. A Monte Carlo simulation study using sinusoidal and trapezoidal pulses was undertaken, to show the accurate estimation of arrival times achieved by the regression stepwise procedure.

Logic-level simulation of a digital circuit is commonly effected using a discrete event-driven strategy. We have, alternatively, employed a demand-driven simulation strategy in which simulation speed is improved by avoiding unnecessary component activations and signal evaluations. In this paper we describe the application of this method to the simulation of both logic and timing behaviour. Preliminary results are presented, showing significant performance gains in comparison with event-driven simulation.

High-Performance Super Self-Aligned Process Technology (HSST), a new deep submicrometer super self-aligned silicon bipolar technology, in combination with electron-beam direct writing technology, helps achieve a 0.3-micrometer minimum patterning. Inclusion of parasitic capacitances of polyelectrode and polyresistors simulates ECL delay time accurately for low-power operations. High-speed communication systems will find the HSST technology promising.

Numerous physical systems can be described by higher order models including a small time delay. Simpler reduced-order models are needed for control system design and implementation. In this paper, the problem of optimal model reduction for multi-input - multi-output (MIMO) systems in the presence of a small time delay is addressed. The necessary conditions for solving the optimization problem are developed. An example demonstrating the feasibility of the approach is presented.

The definition of the Teager energy operator over complex-valued signals was determined with respect to the ambiguity function. The relationship between the Teager energy operator and the ambiguity function allows the use of the Teager energy in estimating the second moment angular bandwidth and the moments of a signal duration as well as that of its spectrum. Furthermore, the relation of the second moment angular bandwidth to the time delay estimation was derived.

As the speed and number of Local Area Networks continue to increase, higher and higher demands are placed on Wide Area Networks. To support the tremendous resulted traffic and prevent their links from congestion, such networks should utilize a robust routing strategy. This paper proposes an efficient routing algorithm and a new model for evaluating the average packet delay in high-speed networks. The routing strategy takes into account the level of congestion over each link in order to choose the best possible routes for each source/destination pair, whereas the proposed model is directly derived from the characteristics of high-speed links. Numerical results show that the combination of such a strategy with the proposed model gives a good behavior to high-speed networks and induces the generalization of the average packet delay. [C] 2000 Elsevier Science Ltd. All rights reserved. Keywords: Delay analysis; Distributed networks; End-to-end delay; Flow assignment; High-speed networks; Packet-switching technology; Routing strategy

An analytical delay model of a uniform wire and driver, based on a combination of a switch-level model and Elmore delay is developed for the on-chip interconnections. The model is utilized for the optimization of delay, power, area, or combinations thereof, subject to different constraints. A linear relation between wire width and transistor widths is proposed, which leads to closed-form solutions to the optimization problems. These solutions are valid also for a multistage uniform wire and driver and are easily used in practical design. The solutions are compared to SPICE simulated results, using practical process parameters. Results show that the root of mean square (rms) errors of delay and power between SPICE simulation and the proposed method are around 3%. Index Terms - Area, delay, high-speed interconnect, power, switch level, VLSI.

Three networks are compared for low false alarm stock trend predictions. Short-term trends, particularly attractive for neural network analysis, can be used profitably in scenarios such as option trading, but only with significant risk. Therefore, we focus on limiting false alarms, which improves the risk/reward ratio by preventing losses. To predict stock trends, we exploit time delay, recurrent, and probabilistic neural networks (TDNN, RNN, and PNN, respectively), utilizing conjugate gradient and multistream extended Kalman filter training for TDNN and RNN. We also discuss different predictability analysis techniques and perform an analysis of predictability based on a history of daily closing price. Our results indicate that all the networks are feasible, the primary preference being one of convenience. Index Terms - Conjugate gradient, extended Kalman filter, financial engineering, financial forecasting, predictability analysis, probablistic neural network, recurrent neural network, stock market forecasting, time delay neural network, time series analysis, time series prediction, trend prediction.