Information processing using a soft output Viterbi algorithm

A method of decoding encoded binary information symbols comprising the steps of:

• updating, without any backwards computation, for successive symbol times a vector of logarithmic probability ratios for each state of a decoding trellis at a respective symbol time, each vector of logarithmic probability ratios corresponding to a survivor path through the decoding trellis, each logarithmic probability ratio representing in a logarithmic domain a ratio of the relative probabilities of a symbol representing a binary one or a binary zero, the updating comprising:
• (a) determining probabilities for reaching each state via respective transition paths corresponding to binary one and zero values of the information symbols from respective states at a previous symbol time;
• (b) combining said determined probabilities to determine a total probability of reaching the state;
• and
• (c) for binary one and zero information symbol values at each state, merging respective probability vectors for respective states at said previous information symbol time in accordance with at least one respective probability of the transition paths from such states of a decoding trellis;
• and
• (d) determining a the probability ratio for the information symbol at a start of the survivor path from elements of at least one vector for all of possible states of the decoding trellis at the respective information symbol time;
• wherein steps (a), (b), (c) and (d) are performed for each and every transition path and steps (b), (c) and (d) are performed for each and every information symbol.

A decoder arranged to carry out a method comprising the steps of:

• updating, without any backwards computation, for successive symbol times a vector of logarithmic probability ratios for each state of a decoding trellis at a respective symbol time, each vector of logarithmic probability ratios corresponding to a survivor path through the decoding trellis, each logarithmic probability ratio representing in a logarithmic domain a ratio of the relative probabilities of a symbol representing a binary one or a binary zero, the updating comprising:
• (a) determining probabilities for reaching each state via respective transition paths corresponding to binary one and zero values of the information symbols from respective states at a previous symbol time;
• (b) combining said determining probabilities to determine a total probability of reaching the state;
• and
• (c) for binary one and zero information symbol values at each state, merging respective probability vectors for respective states at said previous information symbol time in accordance with at least one respective probability of the transition paths from such states of a decoding trellis;
• and
• (d) determining the probability ratio for the information symbol at a start of the survivor path from elements of at least one vector for all of possible states of the decoding trellis at the respective information symbol time;
• wherein steps (a), (b), (c) and (d) are performed for each and every transition path and steps (b), (c) and (d) are performed for each and every information symbol.

A method of decoding encoded binary information symbols comprising the steps of:

• updating, without any backwards computation, for successive symbol times two vectors of logarithmic probabilities for each state of a decoding trellis at a respective symbol time, each vector corresponding to a survivor path through the decoding trellis, each logarithmic probability of the two vectors representing in a logarithmic domain a probability of an information symbol representing a binary one or a binary zero respectively, the updating comprising:
• determining probabilities for reaching each state via respective transition paths corresponding to binary one and zero values of the information symbols from respective states at a previous symbol time;
• combining said determined probabilities to determine a total probability of reaching the state;
• and
• for binary one and zero information symbol values at each state, merging respective probability vectors for respective states at said previous information symbol time in accordance with the respective probabilities of the transition paths from such states;
• and
• determining a probability ratio for the information symbol at a start of the survivor path from elements of the vectors for all of possible states of the decoding trellis at the respective information symbol time.

A decoder arranged to carry out a method comprising the steps of:

• updating, without any backwards computation, for successive symbol times two vectors of logarithmic probabilities for each step of a decoding trellis at a respective symbol time, each vector corresponding to a survivor path through the decoding trellis, each logarithmic probability of the two vectors representing in a logarithmic domain a probability of an information symbol representing a binary one or a binary zero respectively, the updating comprising:
• determining probabilities for reaching each state via respective transition paths corresponding to binary one and zero values of the information symbols from respective states at a previous symbol time;
• combining said determined probabilities to determine a total probability of reaching the state;
• and
• for binary one and zero information symbol values a each state, merging respective probability vectors for respective states at said previous information symbol time in accordance with the respective probabilities of the transition paths from such states;
• and
• determining a probability ratio for the information symbol at a start of the survivor path from elements of the vectors for all of possible states of the decoding trellis at the respective information symbol time.