Method and Apparatus for Efficient Multi-Symbol Detection

A receiver apparatus for decoding a group of symbols from a set of received signal samples that depend on each symbol of the group to an extent depending on an associated dependence coefficient, said receiver apparatus comprising one or more processing circuits configured to:

• compute first modified signal samples and modified dependence coefficients using said dependence coefficients and said received signal samples;
• for each possible combination of all except at least one symbol in said group of symbols, subtract from said first modified signal samples the dependence on said combination of symbols as described by said modified dependence coefficients in order to obtain second modified signal samples dependent principally on the remaining at least one symbol;
• determining a value for the remaining at least one symbol that is the nearest symbol value in the symbol alphabet to obtain quantized values for said at least one remaining symbol, and associate the quantized symbol values with the combination of other symbols to obtain a candidate symbol group;
• and
• for each candidate symbol group, compute a metric that describes how well the candidate symbol group explains said set of received signal samples, and select the candidate group giving the best metric.

A method for decoding a group of symbols from a set of received signal samples that depend on each symbol of the group to an extent depending on an associated dependence coefficient comprising, for each selected subset of symbols to be decoded from the group:

• forming hypothesized subsets representing different combinations of possible symbol values for those remaining symbols in the group not included in the selected subset;
• for each hypothesized subset, generating an estimate of the symbols in the selected subset;
• forming candidate symbol sets from the hypothesized subsets and the corresponding estimates, each candidate symbol set thus representing a different combination of possible symbol values for the group of symbols;
• computing a metric for each candidate symbol set, the value of which indicates how well the candidate symbol set corresponds to the received signal samples;
• and
• outputting as decoded symbols the estimate of the symbols in the selected subset that corresponds to the candidate symbol set having a best one of the metrics.