Showing papers by "South China University of Technology published in 1991"

An adaptive variable structure model following control design for robot manipulators

Abstract: An adaptive variable structure model-following control (AVSMFC) design is proposed for trajectory tracking in a nonlinear robot system which ensures the stability of the intersection of the surfaces without necessarily stabilizing each individual one. The approach avoids the difficulties linked to the strict positive realness requirement in traditional AMFC by taking advantage of the inherent positive definiteness of the manipulator's inertia matrix, and is easily extendible to a higher number of links. The controller does not require any knowledge of a nonlinear robotic system and does not necessarily need the occurrence of a sliding mode at each individually stable discontinuity surface. It thus greatly reduces the complexity of design. In the closed loop, the joint angles asymptotically converge to the reference trajectory with a prescribed transient response. Chattering is eliminated by restricting the state of the system to slide within a boundary layer rather than along the intersection of the surfaces. >

Modification of AlSi alloys with sodium

Abstract: Effect of sodium on the solidification process of high purity Al-12 wt% Si alloys and the epitaxial growth of Si phase from (111) Si or (100) Si substrate has been studied by means of thermal analysis, microscopy, X-ray photoelectron spectroscopy and Auger electron spectroscopy. Results show that normal modification with Na has little influence on the eutectic arrest temperature, melting point, latent heat of freeze and heat of fusion of the Al-12 wt% Si alloy, and that Na adsorbs on (111) Si and affects both nucleation and growth of the Si. Kinetic features and mechanism of the modification are discussed and it is suggested that modification of AlSi eutectic with Na is a kinetic problem rather than a thermodynamic one and the transition of growth morphology of eutectic Si from platelike to fibrous results from the effects of Na on the nucleation and growth of both Al and Si phase.

2-D DCT systolic array implementation

Abstract: The algorithm and architecture of a 2-D systolic array processor for the DCT (discrete cosine transform) are proposed. It is based on the relationship between DCT and cosine DFT and sine DFT. Two systolic architectures of 1-D DCT data and control flow computation are discussed. By use of the main feature of the two systolic 1-D arrays for DCT, a full 2-D systolic DCT array is presented.

Spectral structure for a class of one-dimensional three-tile quasilattices

Abstract: Using a decomposition-decimation method based on the renormalization-group technique, we have studied the spectral structure of a class of one-dimensional three-tile quasiperiodic lattice models, for which the (concurrent) substitution rules are S --> M, M --> L, and L --> LS, where S, M, and L represent, respectively, the short, medium, and long tiles. Branching rules for the electronic energy spectrum have been analytically obtained and confirmed by numerical simulations. It is found that three kinds of branching patterns alternately appear in the spectrum, which displays a kind of self-similarity very different from the trifurcating self-similarity of one-dimensional Fibonacci quasilattices.

Electronic properties of two-dimensional quasicrystals with near-neighbor interactions

Abstract: We study the electronic properties of the vertex model of two-dimensional Penrose lattices in the framework of the tight-binding Hamiltonian with near-neighbor interactions. For Penrose tiling, the first three shortest interatomic distances are the short diagonal of a thin rhombus, the edge of a rhombus, and the short diagonal of a fat rhombus. In this paper the effects of different ranges of interactions on the energy spectrum, degenerate confined state, and localization of electronic states are studied in detail. A similarity transformation is introduced to reduce the Hamiltonian, and the degeneracies of eigenstates are analytically determined. It is found that if first- and second-neighbor interactions are considered, three kinds of wave functions (extended, localized, and intermediate states) coexist. If third-neighbor interactions are included, the degeneracy of zero-energy eigenstates and localized electronic states completely disappear, and only extended and intermediate states are observed.

Variable structure control of linearizable systems with applications to robot manipulators

Abstract: Results concerning the variable structure control of nonlinear systems which are exactly input-output linearizable by state feedback are presented. Stability conditions are given for the bounded tracking. The scheme is applied to control a two joint robot manipulator. The simulation results demonstrate that this scheme can achieve fast precise robot motion control under substantial modeling errors. >

Method and equipment for electromagnetic dynamic plasticating extrusion of polymer materials

Abstract: Method and apparatus for the electromagnetic dynamic plasticating of polymer materials includes a magnetic moving body which is supported both axially and radially by a magnetic levitation gap. The magnetic body is rotated in non-steady state vibration within an electromagnetic winding and cooperates with a stationary body to form a processing space for the polymer material.

Approximation property of multi-layer neural network (MLNN) and its application in nonlinear simulation

Abstract: Presents the approximation properties of multilayer neural networks (MLNN) and their application in nonlinear simulation and analysis. The authors give a direct proof of the approximation ability for a single input-output MLNN with hidden units with sigmoid activation functions, and also give the relationship between the best polynomial approximation and the number of MLNN hidden units. Based on the analysis, the authors propose an MLNN model with hybrid sigmoid-Gaussian activation functions. To verify the idea, they present experiments and results of nonlinear simulation and analysis by MLNN for a solid-state power amplifier. These results prove that the proposed method has general application in nonlinear engineering simulations. >

On the bound of the approximation capacity of multi-layer neural network

Abstract: The authors deal with the approximation capacity of a multilayer neural network (MLNN) and present a strict mathematical expression for the bound of the approximation ability of a single input-output MLNN with hidden units of sigmoid activation functions. The influence of the number of hidden units, connections, and the threshold on the approximation accuracy is also discussed. The relationship between the order of the polynomial and the number of MLNN hidden units used to approximate the polynomial is given. >

The exact integral equation of Hertz's contact problem

Abstract: This paper presents the exact integral equation of Hertz's contact problem, which is obtained by taking into account the horizontal displacement of points in the contacted surfaces due to pressure.

Thermo‐oxidative aging in surface layer of cis‐1,4‐polybutadiene vulcanizates

Abstract: The structural changes occurring in the surface layer and in the center bulk of butadiene (BR) vulcanizates has been studied in hot air at 120-180°C. The thermo-oxidative aging takes place only in the surface layer. The thickness of the oxidized surface layer is almost independent on the aging time after a period of aging time, but decreases with an increase of aging temperature. The crosslink on oxidation in the surface layer of BR vulcanizates increases with aging time, while the thermo-cyclization is formed in the center bulk

Self-condensation of oxidized lignosulfonate. II. Condensation behaviors

Abstract: Studies have been made on the principal parameters that affect the self-condensation of oxidized lignosulfonate, such as lignin content, pH, temperature, time of hot pressing, and oxidation. During condensation, the hydroxyl groups in both the lignosulfonate and carbohydrate molecules could react with carboxyl, sulphonic acid, and carbonyl groups to form an ester bond and result in (hemi) ketal and (hemi) acetal structures. These structures made condensation products having excellent resistance in water and in alkaline solutions. The results also demonstrated that the oxidized lignosulfonate could be condensed at a higher rate than could the untreated one.

Stability and attractivity analysis of bidirectional associative memory from the matched-filtering viewpoint

Abstract: The authors study the bidirectional associative memory (BAM) model from the matched-filtering viewpoint, getting an intuitive understanding of its information processing mechanism. They analyze the problem of stability and attractivity, in BAM and propose some sufficient conditions. The shortcomings of BAM, that is, low memory capacity and weak attractivity, are pointed out. A revised BAM model is proposed by taking an exponential function operating on the related correlations between a probing vector and its neighbor library pattern vectors. From the analysis, it was found that stability and attractivity in the modified model are much better than in the original BAM if all the conditions are the same. >

Mean-square practical stability of stochastic large-scale dynamical systems

Abstract: Mean-square practical stability for stochastic large-scale dynamical systems is studied. The vector Lyapunov functions and the basic comparison principle of stochastic systems as well as the decomposition-aggregation method for large-scale systems were used to obtain results for various types of mean-square practical stability of nonlinear stochastic large-scale dynamical systems. The objective is to analyze stochastic large-scale systems in terms of their lower-order subsystems and in terms of their interconnecting structure. These results make it possible to determine the mean-square practical stability of stochastic large-scale dynamical systems by testing the corresponding practical stability of the auxiliary deterministic systems. >

Comparison of Kohonen feature map against K-mean clustering algorithm with application to reversible image compression

Abstract: A proposed criteria based on the concept of normalized entropy is used to evaluate the performances of the Kohonen feature map and the K-mean clustering algorithm and the experimental results are discussed. Then a newly proposed efficient reversible image compression method is described briefly. It turned out that a problem which severely handicaps the practical realization of the method can be resolved by the Kohonen's algorithm at a satisfactory level. >

Adaptive sliding mode control of constrained robot manipulators

Abstract: The design of sliding mode control for trajectory tracking of an end-effector on a constrained surface with specified constraint forces is considered. Using sliding mode control, the only information on uncertain parameters which is needed for the design is its possible bounds. An adaptive scheme which is capable of performing the estimation of the bounds is introduced. The feedback control will then be dependent on the estimation of the bounds. This greatly reduces the complexity of the design. A numerical example is presented to illustrate the method developed. A two-link robotic manipulator with a circular path constraint is used to verify the validity of the control approach outlined. >

Decentralized exponential stabilization of deterministic and stochastic large scale systems: a unified approach

Abstract: A unified approach to decentralized exponential stabilization for both deterministic and stochastic large-scale interconnected systems is proposed. An algorithm is given for designing local state feedback controllers to stabilize the large scale-systems. Two examples are given to show the applicability of the approach. >

Multiplicative complexity of discrete cosine and sine transform

Abstract: By use of the closure operation for an odd integer set, the minimal number of real multiplications needed to compute length-2m four-type discrete cosine and sine transforms over the field Q, of rational numbers has been demonstrated.

Nonlinear Dynamic Analysis of Flexible Marine-Risers

Abstract: A nonlinear dynamic analysis model is estabilished on the basis of "lumped mass" approach, which takes the influence of the fluid flow within the pipe into consideration. Numerical results are compared with the published works, and the effects of internal fluid flow, internal pressure, dyanmics as well as the nonlinear characteristics on the behavior of flexible risers are discussed. From this work, some useful conclusions are drawn.

Image reconstruction by an extended autocorrelation method

Abstract: The authors propose a method using a new Extended Autocorrelation (EAC) technique to improve the traditional 2D AR-model algorithms. They use the EAC technique to achieve better parameter estimation that provides exact matching between the first P+1 autocorrelation coefficients of the 20 image signal and those of the 20 model while the high-order coefficients mismatch is minimized. >

A learning algorithm for MLN with dynamic neurons

Abstract: A multilayer network architecture with dynamic neurons which have multilocal feedbacks is built. The proposed network can be trained to memorize sequential patterns. A learning algorithm which is more effective and easier to implement is derived. Some experiments on speech recognition of Chinese digits designed to explore the capabilities of the proposed networks to learn dynamic properties of time-varying data are discussed. The performance of dynamic neurons with different time delay periods is also shown. >

MICROSTRUCTURE OF SiOrN_y THIN FILMS AND A NEW MODEL FOR THE CURRENT TRANSPORT MECHANISM

Abstract: The microstructure and current transport behavior in the thermally nitrided silicon oxide (SiO_xN_y) thin films have been studied and a new model is proposed to explain the current transport behavior of the films. In particular, the model includes the case in which the external applied high field rises up to the value just before the occurance of intrinsic breakdown. The theoretical result of the new model agrees well with the experiment. Finally, the significant current enhancement and trapping ledge phenomenon in the I-V characteristics of the films have been explained satisfactorily using this new model, and the influences of the external applied field and the electron trap on the current transport behavior in the films are also discussed.

The Bound on the Critical Point in the Nearest-Neighbor Majority Vote Model

Abstract: In this paper, the relation of the nearest-neighbor majority vote model to the contact process by coupling is studied. Using this method, we obtain that if the parameter of the nearest-neighbor d-dimensional majority vote model is δ ∈ ((2d − 1)/4d, (2d + 1)/4d), then there is no phase transition for this model. It is an improvement of the result δ ∈ (d/(2d + 1), (2d + 1)/4d) in Ref. 1.

Visual Evoked Potential Measurement by Maximum Length Sequence Technique

Abstract: The maximum length sequence (MLS) technique is applied to improve data collection of a clinically important long latency visual potential. Given a limited time for measuring the response and with low number of stimulus presentation, the MLS technique exceeded the conventional method of ensemble averaging in attaining a higher signal-to-noise ratio (SNR). The theoretical consideration of such improvement is developed along with the system implementation. Typical examples of the results obtained in a human subject are presented.