June 1995, Volume 4, Number 2
When designing SC filters, the question of charge transfer functions and chain parameters becomes very interesting. A complete analysis of this question based on standard charge conservation equations and their solution by algebraic cofactors is presented here.
First, the analytic method used is reviewed (both a standard and a modified nodal voltage method are used), second, the voltage and charge transfer functions of a standard SC four-port model are discussed (including some special cases, such as SC circuits using full bilinear simulation of resistors); and, finally, a chain parameters analogy for SC circuit is derived (including special cases with non existing input/output nodes in one phase).
Further possible ways to establish chain parameters are presented, and the relation between trace parameters of a chain matrix and standard transfer functions is discussed. The entire solution is demonstrated with some simple examples, and possible methods of computer-aided evaluation of the transfer functions are briefly discussed.
The most significant problems of acoustic echo canceller (AEC) realizations are high computational complexity and insufficient convergence rate of the applied adaptive algorithms.
From the analysis of the frequency domain block adaptive filter [2,3] realization and the modified subband acoustic echo canceller  the generalized frequency domain adaptive filter [8,9] has been derived. The result of simulations is demonstrated the efficiency of this algorithm for a stationary noise and real speech signal excitation.
Automatic detection of epileptiform patterns is highly desirable during continuous monitoring of patients with epilepsy. This paper describes an unconvential system for automatic off-line recognition of epileptic sharp transients in the human electroencephalogram (EEG), based on a standard neural network architecture - multi-layer perceptron (MLP), and implemented on a Silicon Graphics Indigo workstation. The system makes comprehensive use of wide spatial contextual information available on 12 channels of EEG and takes advantage of discrete dyadic wavelet transform (DDWT) for efficient parameterisation of EEG data. The EEG database consists of 12 patients, 7 of which are used in the process of training of MLP. The resulting MLP is presented with the testing data set consisting of all data vectors from all 12 patients, and is shown to be capable to recognise a wide variety of epileptic signals.
A method of the calculation of active two-port equivalent noise parameters from n measured noise figures for different values of signal source output admittance is given.
This paper deals with the possibility of learning the neural networks by the use of training patterns having the form of both the crisp numerical data as well as fuzzy numbers.
A short analysis of spread spectrum communication system is performed with respect to the determination of its fundamental system parameters for transmitted signal hidden in the noise. The paper shows dependence of processing gains on the correlation circuits at the receiver side. The distance, from which the transmitter signal is hidden in noise, depends on the spread spectrum factor and the transmitted power. At a given interference power density and at a given requirement of transmitted signal hiding we can calculate the spreading factor of system. A processing gain of spread spectrum system determines a depth of signal hiding in the noise. A transmission range of spread spectrum communication system is given by the transmitted power and the processing gain of the system.