Design of a Low Complex Adaptable SOVA Decoder for Convolutional Codes

This paper describes the design of a practical, low complex adaptable soft-output Viterbi algorithm (SOVA) decoder suitable for convolutional encoders. The decoder has 3 look-up tables, which are generated by the encoder. The first table contains the state-transition information while the second table contains information needed by the decoder to back track through the trellis to determine the most likely signal transmitted. The third table contains a list of all possible signals that can be generated by the encoder to facilitate the calculation of the minimum squared Euclidean distance. The adaptability of this design stamps from the advantage of using the three decoding tables i.e. for the same value of the encoder constraint length, the SOVA decoder need not be redesigned for each new scheme to accommodate more/less states and possible waveforms or different symbol alphabet size