Massimo D'Apuzzo

Wavelet network-based detection and classification of transients

A methodology is presented for developing a digital signal-processing architecture capable of simultaneous and automated detection and classification of transient signals. The basic unit of the aforementioned architecture is the wavelet network, which combines the ability of the wavelet transform of analyzing nonstationary signals with the classification capability of artificial neural networks. By exploiting the modularity as well as original strategies concerning wavelet network implementation and training, the method succeeds in enhancing the classification performance with respect to other available solutions.

A neural network approach for identification and fault diagnosis on dynamic systems

The possibilities offered by neural networks for overcoming both system identification and fault diagnosis problems in dynamic systems are investigated. In particular, an original neural fault diagnosis procedure is illustrated. Its sensitivity and response time enables it to be used to great advantage in online applications. Some applications are also reported which, although pertaining to a simple linear dynamic system, highlight the general applicability and advantages of a neural approach. >

A method for the automatic detection and measurement of transients. Part I: the measurement method

A digital signal-processing method for the automatic detection and measurement of transients is presented. The method executes a time measurement first in order to estimate the duration of the transient. Successively, thanks both to the result of the time measurement and a suitable subband filtering approach, the transient is extracted from the waveform on which it is superimposed in order to carry out accurate amplitude measurements. Both the phases are based on the use of the Wavelet Transform, which is able to assure accurate results for different kinds of transients also in the presence of a low signal-to-noise ratio. In the paper, a detailed description of the operating steps of the proposed measurement method is given along with useful suggestions aimed at its proper implementation. Furthermore, a complete illustration of the theoretical background underlying the measurement method is provided.

A digital signal-processing approach for phase noise measurement

A digital signal-processing method for phase noise measurement is presented. By properly over-sampling the input signal and adopting an optimized coherent demodulation scheme, the method grants acceptable performance in analyzing sinusoidal carriers the frequencies of which range from fractions of Hertz up to hundreds of megahertz. Moreover, the method shows itself a valid alternative both to analog measurement systems, especially for the evaluation of close-to-the-carrier phase noise, and time interval analyzers, particularly for carrier frequencies greater than few units of megahertz. At first, the fundamental stages of the proposed method are described in detail. Its theoretical performance is then derived and compared to that granted by other measurement solutions already available on the market. The results of experiments carried out on actual signal sources are finally presented.

The unscented transform: a powerful tool for measurement uncertainty evaluation

An original approach for uncertainty evaluation in indirect measurements is presented hereinafter. The approach applies the unscented transform to the measurement model (i.e., the functional relationship between output and input quantities) in order to gain a reliable estimate of output expectation and standard deviation (measurement uncertainty). Thanks to some useful properties of the transform, notable limits of the current GUM recommendations can be overcome. In particular, reliable estimates are also granted in the presence of nonlinear and/or nonanalytical measurement models or complex digital signal processing algorithms. A number of numerical tests are conducted on simulated and actual measurement data. Remarkable concurrence between obtained estimates and those granted by Monte Carlo simulations confirms the efficacy of the proposed approach

Poer Measurement in Digital Wireless Communications Systems through Parametric Spectral Estimation

Power measurement in digital wireless communication systems often suffers from poor repeatability, usually accompanied by a low accuracy. To face the problem, the use of parametric spectral estimators is investigated in this paper. In particular, a new method is proposed, which first estimates the power spectral density (PSD) of the analyzed signal through Burgs solution, and then evaluates the power by applying straightforward measurement algorithms to the estimated PSD. The results of a number of experiments, carried out on both laboratory and actual signals peculiar to digital wireless communication systems, assess the efficacy and reliability of the method. Moreover, a comparison of the achieved performance to that offered by an alternative measurement solution, already proposed by the authors and based on nonparametric PSD estimation, shows that the method allows for a significant reduction of measurement time, while exhibiting the same repeatability.

A measurement method based on time-frequency representations for testing GSM equipment

A digital method for the measurement and characterization of Global System for Mobile Communications (GSM) equipment is presented here. The method, based on the use of time-frequency representations, shows itself suitable to carry out a whole range of measurements needed for compliance tests both for base stations and mobiles. It is so possible to avoid the employment of different instruments, complex procedures, and qualified users. Some aspects regarding GSM measurements are first presented to clearly describe the operative stages of the method. Then, its performance is assessed by analyzing reference signals. Experimental results obtained from the application of the method to actual GSM signals are finally given.

The detection of echoes from multilayer structures using the wavelet transform

A method for measuring unknown thicknesses of multilayer structures, based on echo detection by means of the wavelet transform (WT), is presented. A brief discussion of the theoretical considerations underlying the method is first given. This highlights the excellent performance shown by the WT as a powerful tool for the analysis of echoes in a noisy environment. A suitable operating procedure for validation of the method is then set up. To this end, tests on 1) simulated signals and 2) actual signals received from known thicknesses are carried out: the obtained results are finally given and discussed.

A method based on wavelet networks for the detection and classification of transients

The paper deals with a new method for automatically detecting and classifying transient signals. The method is based on wavelet networks which combine the aptitude of the wavelet transform in analyzing nonstationary signals with the classification capabilities of artificial neural networks. A detailed description of the wavelet network structure is first given. Then, the fundamental stages of the proposed method are presented. The first results obtained from the application of the method to transients typical of a significant field are finally shown.

New digital signal-Processing approach for transmitter measurements in third generation Telecommunications systems

The rapid growth of third generation (3G) telecommunications systems has created the need for getting new test equipment as well as getting measurement techniques up and running in a very short time. Trying to satisfy this exigency, a new measurement method for testing 3G transmitters is proposed here. Thanks to the use of time-frequency distributions, the method provides a unified approach for carrying out, automatically and in a very straightforward manner, most measurements needed to fulfill the aforementioned task. After a brief outline of 3G technology basics, some details concerning transmitter tests are given. The fundamental steps of the proposed method are then described with references to a clarifying example, also highlighting advantages with respect to other measurement solutions. At the end, the performance of the method is assessed by means of several experiments on both simulated and emulated signals.