METHODS AND DEVICE FOR PULSE-SIGNAL DATA STREAM FRAGMENTATION, STRUCTURE ANALYSIS, AND COMPRESSION Russian patent published in 1998 - IPC

FIELD: computer engineering and other industries using microprocessor units. SUBSTANCE: fragmentation method involves separation of source fragment into desired number of sub-fragments followed by computing ratios of difference between next and preceding blocks; then number of maximum ratios is chosen from set of ratios of difference obtained sot that length of each sub-fragment deviates from mean length of sub-fragment by not more than P. During fragment structure analyses choice of maximum ratios of difference is followed by computation of values characterizing randomness of fragments of bit, byte, and word streams, respectively, which are used to calculate randomness standards according to desired functions, and decision is made concerning structure of source fragment. Each of sub-fragments of word item streams obtained is compressed by entering in summary fragments for words contained in storage library their lengths and positions in storage library, for words absent in storage library, these words as a whole; when stream of byte item is compressed, each of sub-fragments obtained is compressed by comparing each of probable values of byte items with number of bytes having such value in sub-fragment; results of this comparison are used to select code corresponding to each value of byte; each of fragments of byte item stream is compressed by repeated conversion of source sub-fragment into four summary fragments and of main of these four fragments into same four summary fragments until this conversion results in decrease in total number of bits in all four summary fragments. EFFECT: improved conditions for storage, transmission, and processing of pulse-signal information. 6 cl, 5 dwg