IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications 21-23 September 2009, Rende (Cosenza), Italy Data Coding Method on the Basis of M - Sequences Vasyl Yatskiv 1 , Nataliya Yatskiv 2 1 Ternopil National Economic University, 3 Peremoga Square, Ternopil, Ukraine, 46004, jazkiv@ukr.net 2 Ternopil National Economic University, 3 Peremoga Square, Ternopil, Ukraine, 46004, jatskiv@ukr.net Abstract – This paper presents data coding method in computer networks, which provides opportunity of errors detection in data bursts and partial errors correcting due to use of pseudorandom sequence and four signal elements for signal introduction in physical level of communication channel. Keywords – coding, networks, communications channel, M – sequence, M – sequence. I. INTRODUCTION A considerable increase of data transmission rate in the telecommunication systems and computer networks raises requirements to coding methods at physical level of communication channel. These requirements concern effective use of channel capacity and increase of data transmission reliability. When we use rectangular impulses for discrete information transmission it is necessary to choose such coding method, which would answer the followings criteria simultaneously: 1) provide synchronization between a transmitter and receiver; 2) occupy the least width of spectrum at identical bit rate; 3) absence of permanent constituent; 4) find errors; 5) have a low cost of realization. The resulted criteria conflict with each other, that is why them simultaneous implementation is a problem. The conducted analysis of the known coding methods at physical level of computer networks makes it clear that they did not provide detecting and correction of errors in the trasmission process [1]. All existent coding methods can be classified to potential or impulsive. The shortcomings of potential methods is absence of internal synchronization and presence of permanent constituent. The elimination of the potential methods shortcomings occurs in the following ways. The additional bits of reverse polarity insert in the starting code. In that way long zero sequences cut short and we get self-synchronizing code. But this method reduces transmission line capacity, because the additional bits don’t carry any information. Another method is based on premixing of source information in such a way as to occurrence probability of zeros and units are approximate. The basic shortcoming of impulsive methods is the large spectrum width in relation to potential coding methods. The actual problem is elaboration of coding method for error detecting and error correcting on physical layer of communication channels of the specialized computer networks and systems with a high electromagnetic noise level. One of the known coding methods , which provide of errors detection (only in single bits) is the code of Bipolar Alternate Mark Inversion (AMI) (Fig. 1) [2]. The violation of strict alternating of signals in Bipolar Alternate Mark Inversion code is error impulse or disappearing of correct impulse line. The signal with incorrect polarity is forbidden. It allows to detect an error. The shortcoming of this method is a presence of permanent constituent and impossibility of error detecting in a few digits, which do not result in violation of polarity. In the quasi-ternary code (QT-1) information unit is transmitted by potential “+1”, information zero by potential “-1”, the iteration of unit or zero by potential “0” (fig.2). The quasi-ternary code provides self-synchronization during data transmission and permit partial errors correcting, only if we have the same potential signals at the receiver input. II. PROPOSED APPROACH This work is a development of coding method on the basis of M-sequence. In the offered method informative bits is replaced by one or two bits of M- sequence. Thus, when data bit value is zero we replace it by one bit of M - sequence, and when data bit value is one we replace it by two bits of M - sequences (table 1). For data transmission on the basis of the offered method we’ll use four signal element ) ( ), ( 2 1 t S t S , ) ( 4 ), ( 3 t S t S (Fig. 3). As a pseudorandom sequence M – sequences is chosen. M – sequences – has good correlational characteristics and simple generation algorithm on the base of feedback shift register [3, 4].