Chair: Richard C. Rose, AT&T Bell Laboratories (USA)
Alle-Jan van der Veen, Delft University of Technology (THE NETHERLANDS)
Shilpa Talwar, Stanford University (USA)
Arogyaswami Paulraj, Stanford University (USA)
The finite alphabet property of digital communication signals, along with oversampling techniques, enables the blind identification and equalization of an unknown FIR channel carrying a superposition of such signals, provided they have the same (known) period. Applied to multi-user wireless communications, the same framework allows the blind separation of multiple finite alphabet signals received at an arbitrary antenna array through an unknown multipath propagation environment with finite delay spread. An algorithm is proposed and tested on simulated data.
N. Magotra, University of New Mexico
W. McCoy, University of New Mexico
F. Livingston, University of New Mexico
S. Stearns, Sandia National Laboratory (USA)
This paper describes the application of adaptive filters in a two stage lossless data compression algorithm. The term lossless implies that the original data can be recovered exactly. The first stage of the scheme consists of a lossless adaptive predictor while the second stage performs arithmetic coding. The unique aspects of this paper are (a) defining the concept of a reversible filter as opposed to an invertible filter (b) performing lossless data compression using floating-point arithmetic (c) designing lossless adaptive predictors (d) using a modified arithmetic coding algorithm that can handle input data word sizes exceeding 14 bits.
Nihal I. Wijeyesekera, Schlumberger Houston Product Center
Ram G. Shenoy, Schlumberger-Doll Research (USA)
A new procedure for the minimax estimation of quadratic functionals of signals is described. The estimates are optimum when the signals satisfy a quadratic constraint, a common assumption made for estimation of linear functionals. The method will, for example, provide best minimax estimates of signal energy in a time-window and of pointwise evaluations of Fourier transform magnitude, in contrast to earlier methods, which first obtain optimum minimax estimates of linear functionals, and subsequently form a suboptimum quadratic estimate by evaluating a weighted sum of the squared linear estimates.
F. Marvasti, King's College London (UK)
In this presentation, we shall describe interpolation of low pass signals from a class of stable sampling sets at half the Nyquist rate. Practical reconstruction algorithms are also suggested.
L. Vandendorpe, UCL Telecommunications & Remote Sensing Laboratory (BELGIUM)
B. Maison, UCL Telecommunications & Remote Sensing Laboratory (BELGIUM)
L. Cuvelier, UCL Telecommunications & Remote Sensing Laboratory (BELGIUM)
The generalized sampling theorem states that any analogue signal whose spectrum is limited to 1/T can be exactly recovered from N sequences of samples taken at a rate 2/NT and all having a different sampling phase. When N=2, the exact interpolation formula can be derived quite easily. The ideal interpolation filters have infinite impulse responses. This paper addresses the theoretical question of recovering from the 2 initial sequences, any other sequence taken at the same rate 1/T and with a different sampling phase. FIR filters optimized for a mean squared error criterion have been derived in ICASSP 94. In the present paper, FIR filters are derived for a least square interpolation error. Moreover, an adaptive implementation is proposed and formulated as a Kalman algorithm. Simulation results obtained for AR processes show the effectiveness of the solution compared to static solutions.
Ramamurthy Mani, Boston University (USA)
S. Hamid Nawab, Boston University (USA)
The Exponential Rate Operator (ERO) is presented for determining the instantaneous exponential rate of the amplitude modulation during musical transients. Its extension to multiband signal representations such as STFT and wavelet transforms is also described. Sensitivity of the ERO to white noise is examined and computational efficiency of the STFT-based ERO is discussed. Examples involving synthetic and real musical transients illustrate the usefulness of ERO analysis.
R.A. Gopinath, IBM T.J. Watson Research Center (USA)
This paper describes the connection between a certain signal recovery problem and the decoding of Reed-Solomon codes. It is shown that any algorithm for decoding Reed-Solomon codes (over finite fields) can be used to recover wide-band signals (over the real/complex field) from narrow- band information.
Stephane Chretien, CNRS-Supelec (FRANCE)
Ioannis Dologlou, CNRS-Supelec (FRANCE)
In this paper, we show how successive projection-like algorithms may be used for approximation or exact modelling of a signal. For that purpose, we propose a new efficient algorithm providing adequate linear difference equations satisfied by the original signal. The projection operators at each step of the approximation algorithm and the new procedure are shown to be orthogonal. Finally, a pyramidal structure summarises the possibilities offered by the combinations of both procedures.
Li-Chien Lin, Feng Chia University (TAIWAN)
C.-C. Jay Kuo, University of Southern California (USA)
A new signal extrapolation technique based on the wavelet representation, known as scale/time-limited extrapolation, was recently studied by Xia, Kuo and Zhang. However, the extrapolated result may be unstable for noisy data due to the ill-posedness of the extrapolation problem. We extend the previous wavelet extrapolation framework and examine a regularization technique for robust extrapolation. We first formulate the regularization problem and characterize the properties of its solution. Then, a practical iterative algorithm is proposed to achieve robust extrapolation. Compared with the regularized band-limited extrapolation, the major advantage of this new extrapolation approach is that it provides a large class of wavelet bases for signal modeling.
Wooshik Kim, Korea Telecom Systems Development Center (SOUTH KOREA)
In this paper, we consider the problem of making a minimum phase signal from an arbitrary one-dimensional signal by adding a point signal and its application to a two-dimensional phase retrieval problem. In particular, we show that a two-dimensional phase retrieval problem can be decomposed into several one-dimensional phase retrieval problems so that a M x N two-dimensional signal can be reconstructed from its Fourier transform magnitude by solving min {M,N} + 2 one dimensional phase retrieval problems.