(21) * << * >> * Russian * English * Content * All Issues
Pages: 164-175
Full text of article: English language.
Abstract:
Digital signal processing (DSP) puts forward a number of specific demands to processors for realization of DSP. First of all the DSP-processors are arithmetical processors intended for arithmetical processing of big number information (signal samplings), which enter to the DSP-processor in a real scale of a time. At that arithmetical processing is reduced as a rule to realization of arithmetical operations of addition, subtraction and multiplication. Therefore, the main demand to the DSPprocessor is a high processing speed. Since the DSP-processors are realized technically as a rule inform of micro-electronic chips the next demand following from micro-electronic technology is a high homogeneity of the DSP-processor structure. Since the DSP-processors are used as a rule as elements of control systems an additional demand is a high reliability and highability to control errors arising in the DSP-processors. The presents article is devoted to statement of new approach to design of the DSP-processors based on the original number system called "Ternary Mirror Symmetrical Number System". The latter is a synthesis of the ternary symmetrical number system [1] used by Nikolay Brousentsov for design of the ternary computer "Setun" and the number system with an irrational base("Tau System") suggested in 1957 by the American mathematician George Bergman [2].
Citation:
Stakhov A.Ternary mirror-symmetrical arithmetic and its applicationto digital signal processing. Computer Optics 2001; 21: 164-175.
References:
© 2009, IPSI RAS
151, Molodogvardeiskaya str., Samara, 443001, Russia; E-mail: journal@computeroptics.ru ; Tel: +7 (846) 242-41-24 (Executive secretary), +7 (846) 332-56-22 (Issuing editor), Fax: +7 (846) 332-56-20