Automaton Presentations of Cyclic Codes
Keywords
Abstract
Known methods for representing cyclic codes (polynomial, matrix, and algebraic) are suitable for all classes of linear block error-correcting codes, but they do not take into account the particularities of specific classes of the codes. For example, the cycle property of the cyclic codes contains large potential features that are not nearly used in the specified methods of code representation.
The automaton representations of cyclic codes using finite automaton in Galois fields – linear finite-state machine (LFSM) – are proposed. This type of finite automaton belongs to the systems the processes of which are developing in time cyclically, i.e. to dynamic systems. The automaton-analytic and automaton-graphical models are considered. Accordingly, the definition of cyclic codes based on these automaton models is given. The relationship between the automaton representation and the known representations of cyclic codes is shown.
The classification of LFSM from the positions of the automaton representation of cyclic codes is done. For the first time, two characteristic LFSM matrices are taken into account for classification, which makes it possible to distinguish four basic LFSM types: recursive LFSM and non-recursive LFSM of Galois and Fibonacci. When taking into account the direction of data movement, it is possible to distinguish between left-hand and right-hand LFSM, i.e. eight types of LFSM.
A study of the procedures of systematic encoding and decoding of cyclic codes based on their automaton-analytical models is carried out. It is shown that all types of LFSM give an identical result at encoding and decoding, but with different labor intensiveness. The hardware implementation for each type of LFSM is theoretically substantiated. Criteria over type selection LFSM of relatively physical time and hardware-software expenses are brought.
Basic advantage of methods of encoding and decoding of cyclic codes based on offered mathematical models is the linear complexity of computations and simple hardware-software implementation.
