A SUPERVISED LEARNING NEURAL-NETWORK COPROCESSOR FOR SOFT-DECISION MAXIMUM-LIKELIHOOD DECODING

A supervised learning neural network (SLNN) coprocessor which enhances the performance of a digital soft-decision Viterbi decoder used for f orward error correction in a digital communication channel with either fading plus additive white Gaussian noise (AWGN) or pure AWGN has bee n investigated and designed. The SLNN is designed to cooperate with a phase shift keying (PSK) demodulator, an automatic gain control (AGC) circuit, and a three-bit quantizer which is an analog to digital conve rtor (ADC). It is trained to learn the best uniform quantization step- size Delta(BEST) as a function of the mean and the standard deviation of various sets of Gaussian distributed random variables, The channel cutoff rate (R(O)) of the channel is employed to determine the best qu antization threshold step-size (Delta(BEST)) that results in the minim ization of the Viterbi decoder output bit error rate (BER). The quanti zation thresholds in the quantizer are adaptively adjusted by the neur al network according to the statistical characteristics of the analog outputs of the PSR demodulator (for a fading channel) or the AGC (for an AWGN channel), For a digital communication system with a SLNN copro cessor, consistent and substantial BER performance improvements are ob served. The performance improvement ranges from. a minimum of 9% to a maximum of 25% for a pure AWGN channel and from a minimum of 25% to a maximum of 70% for a fading channel. This neural network coprocessor a pproach can be generalized and applied to any digital signal processin g system to decrease the performance losses associated with quantizati on and/or signal instability.