Error burst detection with high-rate convolutional codes

One crucial requirement for the use of convolutional codes in (only) error detection is low decoding complexity. We show that for this specific application the decoding complexity of convolutional codes is practically equal to the coding complexity, which is very small. Thus, the encoder/decoder can be implemented directly in hardware, or use efficient software decoding techniques like those used for cyclic redundancy check (CRC) error detection codes. Different encoder/decoder implementations are considered. By studying the properties of high-rate convolutional codes for the purpose of error detection, we show their potential advantages over block codes. In addition, this study gives a deeper view of CRC codes-which happen to be a special case in a class of codes that we call unit-rate convolutional codes. Thus, for the extension of CRCs we can employ techniques used for convolutional codes, like the use of unit-memory or cyclic time-varying codes. Certain general error detection capabilities of the convolutional codes are derived.