Andrzej Miekina

The use of deconvolution and iterative optimization for spectrogram interpretation

The problem of spectrogram interpretation is considered under the assumption that the parameters of spectral peaks-their positions and magnitudes-contain the information essential for spectrometric analysis. The subsequent use of Tikhonov deconvolution and iterative correction of the estimates of those parameters is proposed. Deconvolution is used for transforming the processed spectrogram in such a way as to facilitate finding initial estimates of its parameters. The advantages of the proposed approach, i.e., gains in accuracy of estimating the parameters of peaks, are demonstrated using both synthetic and real-world spectrophotometric data.

Combined use of Tikhonov deconvolution and curve fitting for spectrogram interpretation

Abstract The problem of numerical correction of spectrograms is addressed. A new method of correction is developed which consists of sequential use of the Tikhonov deconvolution algorithm, for estimating the positions of spectral peaks, and a curve-fitting algorithm, for estimating their magnitudes. The metrological and numerical properties of the proposed method for spectrogram interpretation are assessed by means of spectrometry-based criteria, using synthetic and real-world spectrograms. Conclusions are drawn concerning computational complexity and accuracy of the proposed method and its metrological applicability.

Curve-fitting algorithms versus neural networks when applied for estimation of wavelength and power in DWDM systems

This paper is on optical performance monitors for applications in dense wavelength-division multiplexing (DWDM) communication systems. Two algorithms for estimation of central wavelength and signal power of DWDM channels, on the basis on raw measurement data provided by a low-resolution spectrometric transducer, are compared. The first algorithm is based on the use of a curve-fitting and constrained-optimization technique; the second on application of a superposition of simple feedforward neural networks. The comparison is carried out using semisynthetic data. Conclusions are drawn concerning the applicability of compared algorithms in engineering practice.

Improving Absorbance Spectrum Reconstruction via Spectral Data Decomposition and Pseudo-Baseline Optimization

This paper addresses the problem of reconstruction of the optical absorption spectra on the basis of low-resolution data provided by a spectrophotometric transducer. It is shown that the performance of the algorithms of (generalized) deconvolution with a built-in positivity constraint, which is frequently used for this purpose, deteriorates for data containing a considerable baseline (background). A new method for overcoming that limitation is proposed. It consists of the subtraction of a rough estimate of the baseline (called pseudo-baseline) from the data, reconstruction of the spectrum on the basis of the data modified in this way, and restoration of the pseudo-baseline in the final result of spectrum reconstruction. The proposed amelioration is illustrated by means of an algorithm of spectrum reconstruction based on Janssons method of deconvolution.

Regularized differentiation of measurement data using a priori information on signal and noise spectra

An algorithm previously proposed by the authors (see Proc. of Conf. EMISCON, Prague, 153-5, June 13-15, 1989) for real-time differentiation of discrete measurement data is further developed and studied. The effectiveness of this algorithm depends on a regularization parameter whose value should be fitted to the level of disturbances to which the data are subject. A simple method for choosing this value is proposed which requires only scanty a priori information on the data, namely an estimate of the signal bandwidth and an estimate of the signal-to-noise ratio. The effectiveness of this method is demonstrated using synthetic data and computer experimentation methodology. It is shown that the attainable accuracy of differentiation is very close to the optimum which may be reached by empirical optimization of the parameter of regularization. >

Neural-network-based calibration of a mini-spectrophotometer

A new method for calibration of mini-spectrophotometers is proposed. The method is designed to overcome two important drawbacks of existing methods, viz. their inability to deal with the problems implied by insufficiency of the number of output data and the effects of light polarization. It is based on the use of a tunable laser for acquisition of calibration data, and an RBF (radial basis function) neural network for modeling the polarization effects. The results of a preliminary study of this method, based on semi-synthetic data, are given.

Using variational approach and spectrometry-specific criteria for calibration of spectrometric systems

The problem addressed in this paper is related to numerical correction of spectrometric data aimed at improving the resolution of spectrometric analyses. An idea of weighted optimization is put forward and examined with respect to its power of spectrogram interpretation.

Selected algorithms for measurement data processing in impulse-radar-based system for monitoring of human movements

The importance of research on new technologies that could be employed in care services for elderly and disabled persons is highlighted. Advantages of impulse-radar sensors, when applied for non-intrusive monitoring of such persons in their home environment, are indicated. Selected algorithms for the measurement data preprocessing - viz. the algorithms for clutter suppression and echo parameter estimation, as well as for estimation of the twodimensional position of a monitored person - are proposed. The capability of an impulse-radar- based system to provide some application-specific parameters, viz. the parameters characterising the patients health condition, is also demonstrated.

Monitoring of the OSNR in DWDM Systems Using a Low-resolution Spectrometric Transducer

The paper is on the monitoring of optical telecommunication channels by means of specialized spectrophotometers, called optical performance monitors. An algorithm for data processing is proposed that makes possible the estimation of central wavelength, power and optical signal-to-noise ratio for all channels in parallel. It is demonstrated that it is guaranteeing an acceptable level of measurement uncertainty when applied to the data provided by a low-resolution spectrometric transducer being the constitutive element of an optical performance monitor

Polarization-versed calibration of spectrophotometric transducers

A new method for the calibration of spectrophotometric transducers is proposed. The method, using a radial-basis-function neural network, is designed to overcome two important drawbacks of existing methods, viz. their inability to deal with the problems implied by insufficiency of the number of output data, and light polarization. The results of a study of this method, based on semi-synthetic data, have been given, and conclusions, pointing out its practical usefulness, are drawn.