Turbo equalisation

1. A method of operating a receiver to receive encoded data transmitted simultaneously as a plurality of M different sequences of transmitted symbols from different transmit antennas using a plurality of m modulation levels, where M and m comprise integers and each of the transmitted symbols represents a plurality of bits of the encoded data, the method comprising:

providing N received symbol combinations by receiving, at a plurality of N receive antennas, the plurality of M different sequences of transmitted symbols, wherein each received symbol combination comprises M simultaneously received ones of the transmitted symbols, and wherein N comprises an integer;

generating, from the N received symbol combinations, M pre-processed signals by performing interference cancellation, in which interference cancellation a different symbol of a respective received symbol combination comprises a wanted signal and the other symbols of the respective received symbol combination comprise interfering signals;

generating, from each of the M pre-processed signals, a list of initial candidate symbol combinations, each of the M generated lists including up to m initial candidate symbol combinations, for each of the up to m initial candidate symbol combinations in each of the M generated lists, selecting a first initial symbol indicative of a different one of the m modulation levels and selecting M−1 further initial symbols;

generating, from the lists of initial candidate symbol combinations, an initial estimate of the transmitted symbols;

decoding bits represented by the initial estimate of the transmitted symbols; and

generating M lists of secondary candidate symbol combinations by updating the corresponding M lists of initial candidate symbol combinations, wherein for each list of secondary candidate symbol combinations, the secondary candidate symbol combinations have smaller Euclidean distances than the initial candidate symbol combinations in the corresponding list of initial candidate symbol combinations, relative to the respective pre-processed signal.

2. The method as claimed in claim 1 , wherein:

generating the M pre-processed signals from the N received symbol combination comprises performing interference cancellation by, for each of the pre-processed signals:

generating a channel matrix of channel transfer function elements indicative of a channel transfer function between each of the transmit antennas and each of the receive antennas, wherein the channel transfer function elements are positioned in the channel matrix dependent on which of the symbols of the respective received symbol combination is the wanted signal;

performing QR decomposition of the channel matrix into a matrix Q and a matrix R, the matrix Q comprising N rows and M columns and for which Q −1 =Q H , where Q H denotes a Hermitian transpose of Q, and the matrix R comprising an upper triangular matrix having M rows and M columns; and

generating the respective pre-processed signal by multiplying a vector representative of the respective received symbol combination by Q H ; and

generating the list of up to m initial candidate symbol combinations comprises generating the list of up to m initial candidate symbol combinations, from each of the M pre-processed signals, dependent on the respective upper triangular matrix R.

3. The method as claimed in claim 1 , further comprising selecting the M−1 further initial symbols dependent on Euclidean distances of bits of each of the selected first and further initial symbols relative to the respective pre-processed signal from which the list of initial candidate symbol combinations is generated.

4. The method as claimed in claim 3 , wherein selecting the M−1 further initial symbols comprises minimizing a sum of squares of the Euclidean distances of the bits of each of the selected first and further initial symbols relative to the respective pre-processed signal from which the list of initial candidate symbol combinations is generated.

5. The method as claimed in claim 4 , wherein generating each of the M lists of up to m initial candidate symbol combinations comprises including only initial candidate symbol combinations for which the minimized sum of squares of the Euclidean distances of the bits of each of the selected first and further initial symbols relative to the respective pre-processed signal from which the list of initial candidate symbol combinations is generated is less than a first threshold.

6. The method as claimed in claim 1 , wherein the initial candidate symbol combinations comprise a soft value for each bit represented by each symbol of each initial candidate symbol combination, and wherein generating the initial estimate of the transmitted symbols comprises merging the initial candidate symbol combinations dependent on the soft values of the initial candidate symbol combinations.

7. The method as claimed in claim 6 , wherein the soft value for each bit represented by each symbol of each initial candidate symbol combination is indicative of a minimum Euclidean distance of the respective bit relative to the respective pre-processed signal from which the list of initial candidate symbol combinations is generated, and wherein the merging comprises selecting, for each bit represented by each symbol of the initial estimate of the transmitted symbols, the bits of the M lists of initial candidate symbol combinations for which the soft value is indicative of the smallest Euclidean distance.

8. The method as claimed in claim 1 , wherein decoding the bits represented by the initial estimate of the transmitted symbols comprises generating soft information for each of the bits represented by the initial estimate of the transmitted symbols.

9. The method as claimed in claim 8 , wherein the soft information representative of each of the bits represented by the initial estimate of the transmitted symbols is dependent on a square of a Euclidean distance of the respective bit, after decoding, from the respective pre-processed signal from which the respective initial symbol that represents the respective bit was selected.

10. The method as claimed in claim 8 , comprising:

generating the M lists of secondary candidate symbol combinations, from each of the M pre-processed signals, each of the M generated lists including up to m secondary candidate symbol combinations, for each of the up to m secondary candidate symbol combinations in each of the M generated lists, selecting a first secondary symbol indicative of a different one of the m modulation levels and selecting M−1 further secondary symbols dependent on the generated soft information;

generating, from the M lists of secondary candidate symbol combinations, a secondary estimate of the transmitted symbols; and

decoding bits represented by the secondary estimate of the transmitted symbols.

11. The method as claimed in claim 8 , wherein updating the corresponding M lists of initial candidate symbol combinations further comprises calculating the secondary candidate symbol combinations by employing the decoded bits and the corresponding generated soft information.

12. The method as claimed in claim 10 , further comprising selecting the M−1 further secondary symbols dependent on Euclidean distances of bits of each of the selected first and further secondary symbols relative to the respective pre-processed signal from which the list of secondary candidate symbol combinations is generated.

13. The method as claimed in claim 10 , wherein selecting the M−1 further secondary symbols dependent on the soft information comprises minimizing the sum of:

squares of Euclidean distances of bits of each of the selected first and further secondary symbols relative to the pre-processed signal from which the list of secondary candidate symbol combinations is generated; and

the soft information representative of each of the bits represented by the initial estimate of the transmitted symbols.

14. The method as claimed in claim 13 , wherein generating each of the M lists of up to m secondary candidate symbol combinations comprises including only secondary candidate symbol combinations for which the minimized sum of squares of Euclidean distances of bits of each of the selected first and further secondary symbols relative to the pre-processed signal from which the list of secondary candidate symbol combinations is generated and the soft information representative of each of the bits represented by the initial estimate of the transmitted symbols is less than a second threshold.

15. The method as claimed in claim 10 , wherein the secondary candidate symbol combinations comprise a soft value for each bit represented by each symbol of each secondary candidate symbol combination, and wherein generating the secondary estimate of the transmitted symbols comprises merging the secondary candidate symbol combinations dependent on the soft values of the secondary candidate symbol combinations.

16. The method as claimed in claim 15 , wherein the soft value for each bit represented by each symbol of each secondary candidate symbol combination is indicative of a minimum Euclidean distance of the respective bit relative to the respective pre-processed signal from which the list of secondary candidate symbol combinations is generated, and wherein the merging comprises selecting, for each bit represented by each symbol of the secondary estimate of the transmitted symbols, the bits of the M lists of secondary candidate symbol combinations for which the soft value is indicative of the smallest Euclidean distance.

17. A receiver for receiving encoded data transmitted simultaneously as a plurality of M different sequences of transmitted symbols from different transmit antennas using a plurality of m modulation levels, where M and m comprise integers and each of the transmitted symbols represents a plurality of bits of the encoded data, the receiver comprising:

a demodulator configured to provide N received symbol combinations by receiving, at a plurality of N receive antennas, the plurality of M different sequences of transmitted symbols, wherein each received symbol combination comprises M simultaneously received ones of the transmitted symbols, and wherein N comprises an integer;

an equalizer configured to:

generate, from each of the received symbol combinations, M pre-processed signals by performing interference cancellation, in which interference cancellation a different symbol of the respective received symbol combination comprises a wanted signal and the other symbols of the respective received symbol combination comprise interfering signals;

generate, from each of the M pre-processed signals, a list of initial candidate symbol combinations, each of the M generated lists including up to m initial candidate symbol combinations, for each of the up to m initial candidate symbol combinations in each of the M generated lists, selecting a first initial symbol indicative of a different one of the m modulation levels and selecting M−1 further initial symbols;

generate, from the lists of initial candidate symbol combinations, an initial estimate of the transmitted symbols;

a decoder configured to decode bits represented by the initial estimate of the transmitted symbols;

wherein the equalizer is further configured to generate M lists of secondary candidate symbol combinations by updating the corresponding M lists of initial candidate symbol combinations, wherein for each list of secondary candidate symbol combinations, the secondary candidate symbol combinations have smaller Euclidean distances than the initial candidate symbol combinations in the corresponding list of initial candidate symbol combinations, relative to the respective pre-processed signal.