| 摘要 |
A sequential encoder/decoder pair for error correction using block and convolutionally encoded data. The encoder includes a block coder with up to t bit error correction capability and a convolutional coder, so that the block encoded bits are also convolutionally coded. The convolutional decoder includes a block decoder, which renders the convolutional decoding process more efficient by reducing the amount of computation required to correct t or fewer errors in a received block of data. Since the most likely error patterns involve the least number of bits, most of the correction is accomplished by block decoding, with only minimal use of the convolutional-encoding algorithm, thus saving valuable computation time for more complex but less likely error patterns. The decoder employs an error pattern estimator and branch metric generator to correct all error patterns using the Fano algorithm, using stored values which are derived from known probability of random errors in the data channel. The decoder feeds an encoded block of data into a replica of the convolutional encoder. The replica encoder calculates a syndrome bit pattern from the combination of the received parity together with parity data calculated using the received block and a given number of previous data blocks, i.e., calculated convolutionally. The syndrome bit can then be used to correct the received data.
|
