PROBLEMS OF INFORMATION TRANSMISSION
A translation of Problemy Peredachi Informatsii


Volume 3, Number 2, April–June, 1967
Back to contents page

CONTENTS                   Powered by MathJax

 

Amount of Information and Entropy of Segments of Stationary Gaussian Processes
R. L. Stratonovich
pp. 1–17

Abstract—An asymptotic method is proposed for calculating the quantities of information and entropy of large segments of stationary processes. The quantities arriving per unit time, and also the quantities relating to the limits of the interval are calculated.

 

Asymptotic Transmission Capacity of Channels with a Countable Alphabet and Application to Asynchronous Channels
V. V. Prelov
pp. 18–29

Abstract—Discrete memoryless communications channels are considered with a countable number of symbols at input and output. It is assumed that the signal distortion at the input is subject to the restriction $$ \sum^\infty_{k=1}\varphi(k)p_k\le\mu, $$ where $\varphi(\cdot)$ is some positive function satisfying fairly general conditions. For such channels with low noise, for fairly general conditions, formulas are obtained asymptotic (with respect to a small parameter characterizing the noise power) for the transmission capacity. The method described is used to calculate the asymptotic transmission capacity of asynchronous channels with synchrosymbol and low noise.

 

Two-Way Discrete Memoryless Communications Channels
L. M. Libkind
pp. 30–36

Abstract—In the present note a study is made of the range of the transmission capacity of a two-way memoryless channel. It is shown that for the coding of two statistically independent sources the utilization of data obtained on each direction of the channel does not make it possible to transmit along the channel more information than when these data are not used.

 

Sequential Signal Transmission in a Gaussian Channel with Feedback
R. Z. Khas'minskii
pp. 37–42

Abstract—The effect of feedback signal delay and channel discreteness on message transmission rate in a Gaussian channel with noise-free feedback and symbol-by-symbol coding is examined.

 

Limits for the Decoding Error Probability when Linear Codes Are Used in Memoryless Channels
E. M. Gabidulin
pp. 43–48

Abstract—Upper and lower limits are found for the error probability when linear codes are used in memoryless channels. It is shown that no serious loss in transmission rate or error probability occurs when a linear code is used in a binary completely asymmetric channel.

 

Algorithms for Converting a Binary Code into a Uniform Binary Code Giving an Increased Data Transmission Rate
V. M. Mikhelev and V. Yu. Vershubskii
pp. 49–55

Abstract—Algorithms are described for the simple and effective conversion of any uniform binary code into another such code, in which any code combination has not less than a previously assigned number of zeros between any two ones. By using this latter code, the data transmission rate can be increased in channels with certain specific properties.

 

Memory Contraction for an Automaton Stable to Injuries and Competitions in Its Internal Elements
Yu. L. Sagalovich
pp. 56–64

Abstract—The states of an asynchronous finite automaton are coded in accordance with each input signal (Liu’s method), thereby eliminating critical competitions. An exhaustive method is developed for allocating the places in the individual codes, thereby minimizing (for the given coding method) the number of elements in the automaton storage. The symbols of the resulting code words are regarded as informational symbols, and the automaton is stabilized to injuries as well as competitions among its memory elements by means of a conventional linear coding.

 

Optimum Filtering Taking Quantum Effects into Account
V. V. Mityugov
pp. 65–68

Abstract—The quantum problems of the detection of electromagnetic signals on a background of noise when the receiver measures the intensity of the radiation field or its square are considered. Using the maximum disinformation principle, the matrices of the density of states of the radiation with known spectral components of the signal and noise and given form of the total output signal are obtained. Calculation of the parameters of the useful signal using these matrices leads to a quantum formula of optimum filtering, which reduces to the classical formula at low frequencies.

 

Synaptic Transmission of Information in the Nervous System. Transmission of Signals through the Synaptic Plate
A. I. Podkovyrov, I. T. Kruglikov, and L. V. Idel's
pp. 69–73

Abstract—The problems involved in the passage of a nerve impulse through the synaptic plate are examined. A mathematical model is constructed. The processes of transmitting a signal through the synaptic plate and their dependence on some initial parameters of the system and on the signal characteristics are analyzed. The possibility that the system will reach a stable operating condition, and the dependence of the stabilization level on the signal parameters are investigated. The parameters and operating conditions of the system are considered.

 

Data Transmission over a Gaussian Channel with Feedback
K. Sh. Zigangirov
pp. 74–76

Abstract—Methods for transmitting Gaussian random variables and binary sequences over a memoryless discrete Gaussian channel with 100% feedback are outlined. The transmission rate is equal to the channel capacity.

 

An Algorithm for the Correction of Independent Errors by Cyclic Codes
O. F. Dmitriev
pp. 76–78

Abstract—An algorithm is presented for the correction of independent errors which combines permutation decoding and decoding with compensating polynomials.

 

The Problem of Detecting Coherent Optical Radiation in Thermal Noise
A. G. Sheremet'ev and R. M. Kochetkov
pp. 79–80

Abstract—An expression is derived for the detection threshold of coherent radiation in a thermal background using a Neyman–Pearson detector for a single reading. The probability of detecting coherent radiation in a thermal background is calculated for a particular case.

 

Pattern Recognition by Linear Programing
B. A. Golovkin
pp. 80–84

Abstract—The problem of pattern recognition through preliminary learning is examined. The method of linear programing is used to construct closed finite decision regions.