The possibility is considered of constructing arithmetic error-correcting codes (intended for digital automata) that use number systems with arbitrary base m. Unlike the binary system, in an m-nary system it is not enough to know the position of error; its order and sign should also be determined. An (m-1)-order error is the maximum error in the m-nary system; it may turn 0 into m-1 or vice versa. The number of types of error are tabulated. For detecting and correcting errors of a given order, a code with a lower redundancy than for errors of the order given can be used. The modulo check is used for correcting arithmetic errors in the AN plus B codes, where N = the number to be closed, A and B = constants. Construction of the AN plus B codes that detect and correct single errors is considered.
展开▼