A Table-Driven (Feedback) Decoder

We present a table-driven technique for decoding of the convolutionally encoded data. The approach can be used in either the feedback or the direct mode. A data-independent syndrome vector is generated on the transmitted bits. If it is dgferent from zero it is used to address pre-computed tables of corrections for the encoded bits. The decoding is performed as a separate step. When the feedback mode is used the error propagation is minimized by appropriate choice of the codes, the table construction and the decoding algorithms. The approach allows hardware simplicity comparable to majority-logic decoding, but has error correcting capabilities that can be made as close to the optimal as desired through the adjustment of the syndrome vector length and the correction table size. This makes the method very attractive for high speed satellite and network applications. The performance of the method can be further enhanced through soft detection.