Electrical computers: arithmetic processing and calculating – Electrical digital calculating computer – Particular function performed
Reexamination Certificate
1997-11-06
2001-07-03
Ngo, Ohuong Dinh (Department: 2121)
Electrical computers: arithmetic processing and calculating
Electrical digital calculating computer
Particular function performed
Reexamination Certificate
active
06256652
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to computer technology and can be particularly used to compress the binary code in the devices of processing, storage, transmitting, receiving and recording as well as in information systems, for example, communication and multimedia systems.
2. Description of the Related Art
Known are the methods of compressing the binary code for its transmission through the communication channels. These methods include the translation of the initial binary code to a compressed data flow by orthogonal functions which is effected in synchronous multiplication of each initial code bit by corresponding readings of the orthogonal function used (see F. Hartmouth. Transmitting Inforrnation by Orthogonal Functions.—Moscow, Svjaz, 1975; Abmed, Nasir, Rao, Kamiseti-Ramamahan. Orthogonal Transformations in Processing of Digital Signals.—Transl. fm English by T. E. Krenel.—Moscow: Svjaz, 1980; S. A. Kuritsyn, E. P. Perfiliev, V. I. Ponomarev. Formation of Signal Spectrum at Data Transmittance. Electrosvjaz, 1975, 12, 41-46).
Also known are the methods of compressing the binary code for its recording to data storage devices, such as floppy or hard disks, which include the transformation of a data flow to a compressed code by using one of the known codes of compression, in particular, a method using “prediction” coding (see U.S. Pat. No. 5,146,221 MPK
5
H03M 13/00 publ.8.09.1992; U.S. Pat. No. 5,126,736 MPK
5
H03M 13/00 publ.17.09.1992).
The deficiency of these known methods is a low degree of compression, which varies from about 20 to 80 percent of the initial binary code length, as well as a contradiction which appears between the degree of compression and the pass band of the channel used (track), loss of information (data).
There are methods of compressing data with loss of information, for example, compressing videoimage (laser disks, multimedia). Considering that the information to be transmitted is specific, they allow to suppress the relative (subjective) excessiveness of information. Such algorithms allow to reach the degree of compression of 400 to 4,000 percent. However, such methods are not considered here, as the task set does not allow any loss of information.
As a prototype for the method of compressing the binary code, a method was selected, which includes the transformation of the initial binary data flow by an orthogonal function, which is effected by multiplying the initial binary code by the corresponding readings of the basis function with subsequent summation, and the sums received are transformed to an analogue signal with a predefined frequency spectrum (see B. V. Fesenko, A. D. Chemavin. Modem in KAMAK Standard with a Digital Mode of Forming a Signal.—Avtometria, 1980,4, pp.24-28).
The deficiency of this method is a low degree of signal compression due to the dependence of compression degree on the frequency of the carrier signal. According to Kotelnikov's Theorem, on the increase of the compression degree, the frequency of the carrier is proportionally increased, hence, the public communication lines can be used with limitations for computer communications.
As a prototype of the method of decompression a method has been chosen which is effected by inverse transformation of signal by an orthogonal function with subsequent averaging of signal within predefined time interval. (V. A. Tamm, P. G. Fritzler, Method of Suppressing Discontinuous Jamming and Short-time Intermissions at Data Transmittance.—Electrosvjaz, 1984, 10, pp.52-55).
As a prototype of a compression and decompression processor, the processor has been chosen which comprises an arithmetic and logic unit (ALU), random access memory (RAM), control unit (CU), data registers (DR) and exchange registers (ER) (Lestes, Sanders. Monocrystal Coder-Decoder for Code Translation without Return to Manchester Code Zero.—Electronica, 1982, 15; U.S. Pat. No. 5,126,739 MPK
5
H03M 13/00 publ.17.09.1992).
The deficiency of this processor prototype is a stiff connection of the spectral band to the speed of data transmission (compression degree), resulting in the necessity of expansion of spectrum of signal passing through the track during the transmission, whereas this in turn needs special communication lines or leads to information losses in usual communication lines.
SUMMARY OF THE INVENTION
The task being fulfilled by this invention is to considerably increase (by more than 100-500 times) the compression degree of information transmitted by parallel or serial binary code with high speed of processing (compression) and data information storage under conditions of a noisy track, speeding up the signal coding and decoding with no considerable increase of the carrier frequency.
The task thus set is fulfilled by using, in the known method of binary code compression including transformation of initial data flow by serial multiplication of signal bits to be coded by readings of orthogonal coding function and summation of all products for some time period, according to this invention, as a coding function, of discrete values of function which actually is a piecewise continuous function in form of a Gaussian pulse with frequency filling, and as a coding transformation, a computation of Duamel sequence is used, whereas the data input flow and discrete values of the coding function are used as arguments of this sequence.
The problem is also solved by changing the frequency of Gaussian pulse fill according to exponential law, and the phase in each impulse point is computed as an integral of the frequency filling, whereas of all Gaussian pulse duration at coding only the section approximated by a polynomial function with the polynomial degree not lower than four is used.
The problem is also solved by computing the Duamel sequence on principle of superposition of data initial flow and corresponding values of Gaussian pulse.
The problem is also solved by putting correspondence of the serial input data flow to serial values of the piecewise continuous Gaussian pulse, and the process of computing of Duamel sequence is done parallel-serially, whereas deparalleling of the process of data initial flow compression is effected by impulse processing overlap.
The problem is also solved by using, according to this invention, in a known method of the binary code decompression including the inverse transformation of the compressed values by an orthogonal coding function, as decoding, of the process of bit by bit subtraction of decoded values from undecoded compressed ones, and the remaining values are used to superposition of the data initial flow and the corresponding values of Gaussian pulse within the section of a polynomial function, whereas the remaining values are transformed to the data initial flow by conversion tables, setting several new values of the decompressed (initial) function with a correspondence for several serial values of the function remained after subtraction.
The problem is also solved in parallel compression and decompression processor, comprising an arithmetic and logic unit, random access memory, control unit, data registers and exchange registers by using, in accordance with this invention, as an random access memory, of a multiple-order shift register with input and output logic, besides, whereto a unit of forming codes for values of piecewise continuous Gaussian pulse is introduced.
The problem is also solved by using, in accordance with this invention, as an random access memory, of a multiple-order shift register with input and output logic, additionally are introduced an address counter and a unit to form values of piecewise continuous Gaussian pulse.
The problem is also solved by using a read only memory to store codes of values of piecewise continuous Gaussian pulse in form of tables and an address counter controlling a serial selection of values from the read only memory as a unit of forming codes of values of piecewise continuous Gaussian pulse.
The problem is also solved by using, as a unit of forming codes of values of piecewise continuou
Krot Nikolay Felixovich
Saperov Anatoly Grigorjevich
Carella Byrne Bain Gilfillan Cecchi Stewart & Olstein
Hand, Esq. Francis C.
Milto Jury P.
Ngo Ohuong Dinh
LandOfFree
Binary code compression and decompression and parallel... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Binary code compression and decompression and parallel..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Binary code compression and decompression and parallel... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2505999