HomeChair: V. John Mathews, University of Utah, USA
Home
Xiao-Ping Zhang, University of Texas, San Antonio (U.S.A.)
Mita D Desai, University of Texas, San Antonio (U.S.A.)
The conventional linear adaptive filters are not effective for discriminating the transient wideband signal components from noise. A recently developed wavelet shrinkage approach is able to maintain the function local regularity while suppressing noise however, it has only been used in function estimation problems. In this paper, a new type of nonlinear filtering method for adaptive noise suppression is presented, based on shrinkage method. A new class of shrinkage functions is also presented. The filtering structure and the learning algorithm are developed. The theoretical analysis proves convergence in certain statistical sense. The numerical results of our system are presented for both the standard and the new shrinkage function and compared with the conventional linear adaptive filter based techniques. Results indicate that both the optimal solution and the learning performance are superior to the conventional methods. It is shown that our new shrinkage function performs better than the standard shrinkage function.
ic981107.pdf (From Postscript)
TOP
Yoo-Sok Saw, University of Bristol (U.K.)
Peter M. Grant, University of Edinburgh, Scotland (U.K.)
John M. Hannah, University of Edinburgh, Scotland (U.K.)
Data rate management of compressed digital video has been treated mainly from the teletraffic control point of view, i.e. by modelling congestion control via network protocols. Relatively less attention has been focused on video rate management in the source coding side. In this paper we consider that it is more efficient and less costly to control video rate at the video source than handling network congestion (or overloading) due to an extremely large quantity of incoming variable bit rate (VBR) video traffic. Thus this paper investigates effective rate control algorithms for video encoders. Considering the non-stationary nature of video rate derived from scene variations (i.e. the wide band nature of digital video), we adopted and compare the performance of two nonlinear approaches; radial basis function (RBF) estimation using a neural network-based approach and fuzzy logic control as a nonlinear feedback control.
ic981252.pdf (Scanned) TOP
Takis Zourntos, University of Toronto (Canada)
David A Johns, University of Toronto (Canada)
We present a globally stable arbitrary-order single-bit delta-sigma modulator architecture with continuous-time loop filtering. Using Lyapunov arguments and the method of equivalent control, it is shown that stability is guaranteed for any input signal with peak magnitude less than L > 0, where -L and +L denote the quantization levels. The design augments the conventional delta-sigma modulator with switching feedback and the use of distinct operating modes; the additionalcircuitry required for the implementation of these stabilizing measures is nominal. For a given noise transfer function and fixed oversampling ratio, the new architecture achieves the same peak signal-to-noise-plus-distortion ratio as a traditional delta-sigma modulator. The proposed design can also yield near-peak performance for inputs which destabilize the conventional delta-sigma data converter. Simulation results are provided for the proposed modulator and a comparable standard interpolative design.
ic981299.pdf (From Postscript) TOP
Minyue Fu, University of Newcastle (Australia)
Lihua Xie, Nanyang Technological University (Singapore)
Huaizhong Li, Laboratoire d'Automatique de Grenoble (France)
The problem of passivity analysis finds important applications in many signal processing systems such as digital quantizers, decision feedback equalizers and digital and analog filters. This paper considers the passivity analysis problem for a large class of systems which involve uncertain parameters, time delays, quantization errors, and unmodeled high order dynamics. By characterizing these and many other types of uncertainty using a general tool called integralquadratic constraints (IQCs), we present a solution to the problem of robust passivity analysis. More specifically, we determine if a given uncertain system is robustly passive. The solution is given in terms of the feasibility of a linear matrix inequality (LMI) which can be solved efficiently.
ic981626.pdf (From Postscript) TOP
Luigi Agarossi, Philips Research Monza (Italy)
Sandro Bellini, Politecnico di Milano (Italy)
Alberto Canella, CEFRIEL (Italy)
Pierangelo Migliorati, University of Brescia (Italy)
This paper presents a study aiming to define a nonlinear model, based on the Volterra series, of the high density optical discread out process. Under high density condition, because of the high linear density and reduced track pitch, the signal read out is not a linear process and suffers from cross talk. To cope with such a problem the identification of a suitable nonlinear model is required. According to the Hopkins analysis, a physical model based on the optical scalar theory was mplemented. The results of this analysis have then been used to identify the kernels of a nonlinear model based on the Volterra series. The obtained results show that a second order bidimensional model is sufficient to accurately describe the read out process. The nonlinear Volterra model is a convenient starting point to devise and analyze nonlinear equalization and cross talk cancellation techniques.
ic981713.pdf (Scanned) TOP
Souad Meddeb, ENSEEIHT/GAPSE (France)
Jean Yves Tourneret, ENSEEIHT/GAPSE (France)
Francis Castanie, ENSEEIHT/GAPSE (France)
A large class of non-linear phenomena can be described using bilinear systems. Such systems are very attractive since they usually require few parameters, to approximate most non-linearities (compared to other systems). This paper addresses the problems of bilinear system identification using Bayesian inference. The Gibbs sampler is used to estimate the bilinear system parameters, from measuremements of the system input and output signals.
ic982009.pdf (From Postscript) TOP
Alexander C Wang, MIT RLE (U.S.A.)
Gregory W. Wornell, MIT RLE (U.S.A.)
The 1/f family of fractal processes provides useful models for the extraordinary variety of natural and man-made phenomena that exhibit long-term dependence. Using algorithms based on a multiscale state-space representation, we address the problems of parameter estimation of discrete 1/f signals in white noise, estimation of deterministic signals in 1/f noise, and prediction of discrete 1/f processes. Among other results, distant past data are shown to have a dramatically greater effect on these estimators than when ARMA processes are involved.
ic982111.pdf (From Postscript) TOP
Alberto Carini, University of Trieste (Italy)
Giovanni L Sicuranza, University of Trieste (Italy)
V. John Mathews, University of Utah (U.S.A.)
This paper presents a theory for the exact and the p-th order equalization or linearization of nonlinear systems with known recursive or nonrecursive polynomial input-output relationships. The equalizing and linearizing filters have simple and computationally efficient structures., An experimental result that illustrates the good properties of the technique we propose is also included in this paper.
ic982215.pdf (From Postscript) TOP