Vous êtes sur la page 1sur 1

CODE ERROR DETECTION AND CORRECTION Codes and raw digital data are being transmitted with increasing

volume every year. Unless some means of error detection is used, it is not possible to know when errors have occurred. These errors are caused by noise and transmission system impairments. In contrast, it is obvious when a voice transmission has been impaired by noise or equipment problems. Redundancy is used as the means of error detection when codes and digital data are transmitted. A basic redundancy system is to transmit everything twice and to make sure that exact correlation exists. Fortunately, schemes have been developed that do not require such a high degree of redundancy. Parity The most common mode of method of error detection is the use of parity. A single bit called the parity bit is added to each code representation. If it makes the total number of 1s even, it is termed even parity, and an odd number of 1s is odd parity. For example, if the ASCII code for A is to be generated, the code is P1000001 and P is the parity bit. Odd parity would be 11000001 since the number of 1s is now 3. The receiver checks the parity. If an even number of 1s occur in a character grouping (word), an error is indicated and the receiver usually requests a retransmission. Unfortunately, if two errors (an even number) occur, parity systems will not indicate an error. In many systems a burst of noise causes two or more errors, so that more elaborate error detection schemes may be required. Block Check Character A more sophisticated method of error detection than simple parity is needed in high-data-rate systems (typically, 2400 b/s and above). At these speeds, telephone data transmission is usually synchronous and blocked. A block is a group of characters transmitted with no time gap between them. It is followed by an end of message indicator (EOM) and then a block check character (BCC). A block size is typically 256 characters. The transmitter uses a predefined algorithm to compute the BCC. The same algorithm is used at the receiver based on the block of data received. The two BCCs are compared, and if identical, the next block of data is transmitted. Cyclic Redundancy Check A more sophisticated set of algorithms than parity schemes for BCC calculation is the cyclic redundancy check (CRC). It is easily the most powerful error-detection technique in common use. The entire message block is treated as a long binary number. Binary division is performed by some binary constant and the remainders are subsequently compared at the receiver. A difference between the remainder transmitted by the transmitter and generated at the receiver causes that message block to be retransmitted. The division process usually involves special 12 or 16 bit shift registers at the transmitter and receiver. The CRC technique does not burden the transmitter with the continuous sending of parity bits and thus offers more efficient data transmission. The CRC pattern is only 16 bits or less, and it need only be sent at the end of blocks containing several thousand bits.

Vous aimerez peut-être aussi