Hybrid decoding for digital-communication-feedback systems

Abstract

Forward error correction (f.e.c.) and decision feedback (a.r.q.) are combined using a hybrid decoding scheme of partial error correction and partial error detection. This is shown to have advantages when a low error probability is required and efficient use of transmitter power is important. The performance of the scheme is analysed, assuming transmission over a channel disturbed primarily by white Gaussian noise. Results are presented for systems employing various BCH codes showing the effects on system performance of varying code length, code ratio, and of varying the maximumerror weight the decoder is allowed to correct. The hybrid scheme is shown to need less transmitter power than systems using either pure forward error correction or decision feedback with error detection alone, although the ability to reject error bursts is retained. A design procedure for choosing the code parameters and maximum error-correcting weight is given, and one method of implementing the decoding is suggested.