1,721,365 research outputs found
A new approach to teaching signal processing at undergraduate level
No full textThis paper describes the design and implementation of a unique undergraduate program in signal processing at the Queensland University of Technology (QUT). The criteria that influenced the choice of the subjects and the laboratories developed to support them are presented. A recently established Signal Processing Research Centre (SPRC) has played an important role in the development of the signal processing teaching program. The SPRC also provides training opportunities for postgraduate studies and research
Polynomial time-frequency distributions and time-varying higher order spectra: Application to non-stationary signal analysis
No full textThe paper deals with analysis and time-frequency representation of multicomponent signals characterised by non-linear frequency variation in time. A certain type of higher-order time-frequency distribution, referred to as Polynomial Wigner-Ville distribution [2], has been designed to achieve delta function concentration in the time-frequency plane for this class of signals. In the presence of multiplicative noise, the signals are treated as being random and we introduce time-varying higher-order spectra (TV-HOS) as ensemble averaged Polynomial Wigner-Ville distributions. TV-HOS are shown to be natural tools for analysis of non-stationary random signals, and we demonstrate this in the context of FM signals affected by multiplicative noise
Polynomial order reducing property of the lattice filter in the presence of quadratic FM signals
No full textThis paper presents the behaviour of reflection coefficients of FIR lattice filters for quadratic FM signals. We show that the optimal reflection coefficients form linear FM signals with a reduced order compared to that of the input polynomial phase. Similarly, by considering the linear FM signal produced by the coefficients as the input to another lattice filter, sinusoidal signals are generated. This new characteristic, which we call the Polynomial Order Reducing (FOR) property of the lattice filter, is correspondingly held for the adaptive coefficients. This property is used for estimating the Instantaneous Frequency (IF) of a linear FM signal
Time-frequency discriminant analysis for non-stationary Gaussian signals
No full textWe present a new non-stationary signal classification algorithm based on a time-frequency distribution and multiple hypothesis testing. The time-frequency distribution is used to construct a time-dependent quadratic discriminant function. At selected points in time we evaluate the discriminant function and form a set of statistics which are used to test multiple hypotheses. We show that the statistics are a linear combinations of chi square random variables with constant coefficients and hence are not normally distributed. The multiple hypotheses are treated simultaneously using the generalised sequentially rejective Bonferroni test to control the probability of incorrect classification of one class
Performance analysis of fingerprint compression using an efficient wavelet transform algorithm
No full textThis paper presents a new compression algorithm for fingerprint images. A modified wavelet packet scheme based on a fixed decomposition structure, matched to the statistics of fingerprint images, is used to decorrelate the image pixels
Formant detection through Instantaneous-Frequency estimation using Recursive Least Square algorithm
No full textFormant frequencies, represented by major peaks in the spectrum, convey important information about speech. ''Instantaneous Frequency'' (IF) estimation is a methods to track formants. This paper proposes a method to detect the formants of voiced speech using a Recursive Least Square (RLS) algorithm. The method accuracy is assessed by comparing it with conventional formant detection techniques. The method is also analysed from the viewpoint of phonetic conformity using ''Temporal Decomposition''
Design of signal dependent kernel functions for digital modulated signals
No full textOne of the features of a radio monitoring system is the estimation of modulation parameters for digital modulated signals. For low signal-to-noise ratio conditions, time-frequency signal analysis is an attractive method to use. However, analysis of digital modulated signals using existing time-frequency distributions have proven that none of these distributions are suitable for this task. To improve the time-frequency representation of a class of digital modulated signals, a technique for calculating the signal dependent kernel function is derived. The kernel function is derived by minimizing the least-squares error between the calculated and desired time-frequency distributions. Simulations using the bilinear formulation of time-frequency distribution verify the results and the desirable time-frequency distribution is obtained for a given signal class
- …
