A soft decision syndrome decoding algorithm for convolutional codes

Bit decoding algorithms of convolutional codes are used for real-time systems. A syndrome decoder is shown to be an efficient way for very high data rate implementation. With soft decision signals, the performance of the decoder can be improved. Utilizing a signal-plus-noise channel model to specify a Q-level quantized AWGN channel or a binary input Q-ary output channel, one is able to define both errors and syndromes through an r-dimensional binary vector representation. A Q-level soft decision syndrome decoder based on this model is built. The binary vector representation is found to be unique under this construction. Performance is also shown via simulation.<<ETX>>