Bit soft value normalization

1. An LTE uplink bit soft value normalization method for Log-Max based de-mapping, including the steps of:

de-mapping modulated data symbols of each corresponding SC-FDMA symbol in a slot to un-normalized bit soft values that are independent of noise power;

individually determining a noise power measure for each said SC-FDMA symbol in the slot by exploiting noise information carried within modulated data symbols of the corresponding SC-FDMA symbol, wherein said exploiting is based on using a statistical model that describes the statistical distribution of received modulated data symbols obtained from received SC-FDMA symbols; and

individually normalizing each of the un-normalized bit soft values, said normalized bit soft values representing reliability of received bits obtained from the SC-FDMA symbols in the slot, based on the determined noise power measures for each corresponding SC-FDMA symbol.

2. The method of claim 1 , wherein said normalizing comprises normalizing the bit soft values obtained from each SC-FDMA symbol in the slot with a normalization factor based on the determined noise power measure for that SC-FDMA symbol.

3. The method of claim 1 , wherein said normalizing comprises normalizing the bit soft values obtained from the SC-FDMA symbols in the slot with a common normalization factor based on the average of determined noise power measures.

4. The method of claim 1 , further comprising controlling said normalization by:

determining a ratio between a smallest determined noise power measure and a largest determined noise power measure;

if the ratio exceeds a predetermined value, normalizing the bit soft values obtained from each SC-FDMA symbol in the slot with a normalization factor based on the corresponding determined noise power measure; and

otherwise, normalizing the bit soft values obtained from the SC-FDMA symbols in the slot with a common normalization factor based on the average of the determined noise power measures.

5. The method of claim 1 , further including:

selecting a closest modulation constellation point for each modulated data symbol; and

estimating the noise power measure for each SC-FDMA symbol based on errors between modulated data symbols and corresponding closest modulation constellation points.

6. The method of claim 5 , wherein selecting a closest modulation constellation point for each modulated data symbol comprises

finding closest modulation constellation points by hard decoding the un-normalized bit soft values.

7. The method of claim 1 , further comprising equalizing the modulated data symbols before determining the noise power measures.

8. A Log-Max based LTE uplink soft demapper, including:

a preliminary demapper configured to demap modulated data symbols of each corresponding SC-FDMA symbol in a slot to un-normalized bit soft values that are independent of noise power;

an SC-FDMA symbol noise power meter configured to individually determine a noise power measure for each said SC-FDMA symbol in the slot by exploiting noise information carried within modulated data symbols of the corresponding SC-FDMA symbol, wherein said exploiting is based on using a statistical model that describes the statistical distribution of received modulated data symbols obtained from received SC-FDMA symbols; and

a bit soft value normalizer connected to the SC-FDMA symbol noise power meter and configured to individually normalize each of the un-normalized bit soft values, representing reliability of received bits obtained from the SC-FDMA symbols in the slot, based on the determined noise power measures for each corresponding SC-FDMA symbol.

9. The demapper of claim 8 , wherein the bit soft value normalizer includes a multiplying unit configured to normalize the bit soft values obtained from each SC-FDMA symbol in the slot with a normalization factor based on the corresponding determined noise power measure for that SC-FDMA symbol.

10. The demapper of claim 8 , wherein the bit soft value normalizer includes a multiplying unit configured to normalize the bit soft values obtained from the SC-FDMA symbols in the slot with a common normalization factor based on the average of determined noise power measures.

11. The demapper of claim 8 , wherein the bit soft value normalizer includes a normalization controller configured to determine a ratio between a smallest determined noise power measure and a largest determined noise power measure, and to control a multiplying unit to:

normalize the bit soft values obtained from each SC-FDMA symbol in the slot with a normalization factor based on a corresponding determined noise power measure for that SC-FDCMA symbol if the determined ratio exceeds a predetermined value; and

normalize the bit soft values obtained from the SC-FDMA symbols in the slot with a common normalization factor based on the average of the determined noise power measures otherwise.

12. The demapper of claim 8 , wherein the SC-FDMA symbol noise power meter includes:

a closest modulation constellation point selector configured to find a closest constellation point for each modulated data symbol; and

an SC-FDMA symbol noise power estimator configured to estimate the noise power measure for each SC-FDMA symbol based on errors between modulated data symbols and corresponding closest modulation constellation points.

13. The demapper of claim 12 , wherein the SC-FDMA symbol noise power meter includes:

a closest constellation point selector connected to the preliminary demapper and configured to find closest modulation constellation points for each modulated data symbol by hard decoding the un-normalized bit soft values.

14. An LTE receiver including the soft demapper of claim 8 .

15. The LTE receiver of claim 14 , including an equalizer connected to the soft demapper and configured to equalize the modulated data symbols before determining the noise power measures.