Bit-loading method and system for a DMT transceiver
View Patent ↗A system and method for a bit-loading algorithm based on an equal error rate principle for each sub-channel (bin) used in a discrete multi-tone modulation (DMT) system. The system and method of the present invention provides a bit-loading algorithm that provides long reach, high data rate and good error performance at a given condition. The bit-loading method and system depends on the Shannon channel capacity formula to allocate bits to different sub-channels. For DMT, the system and method tabulates different SNR GAP for different sub-channels (bins), and constructs a look-up table for the SNR GAP with different bits to achieve a better performance for the DMT system. In addition, Trellis coding across all the sub-channels are optionally supported in DMT, and likewise, the system and method creates another look-up table for variable coding gain provided for the Trellis coding gain in the bit-loading algorithm to further improve the system performance, knowing that the Trellis coding gain is different for different sub-channels with different bit allocation.
1. A method performed in a DMT transceiver for loading bits into discrete multi-tone (DMT) sub-channels, the method comprising:
computing signal-to-noise ratio (SNR i ) for all available sub-channels;
computing a variable sub-channel coding gain according to an actual constellation size;
computing a variable sub-channel signal-to-noise ratio gap (SNR-GAP) according to the actual constellation size;
computing the number of bits to be allocated into each sub-channel;
allocating bits to sub-channels based on an equal error rate.
2. The method of claim 1 , wherein the total bits calculated for loading into all the sub-channels is computed using:
b
=
∑
i
log
(
1
+
SNR
i
*
C
i
/
M
*
Γ
i
)
.
where b is the total bits, SNR i is the signal-to-noise ratio (SNR) in an i th sub-channel, C is a coding gain, M is a margin, and Γ is an SNR gap.
3. The method of claim 1 , wherein the probability of the symbol error is computed by:
P s =K ( i )* Q ( d ( i )/2σ)
where P is the probability, K(i) is a nearest symbol neighbor for an i th sub-channel, d(i) is a corresponding minimum distance, and σ is a 1-dimensional Additive White Gaussian Noise (AWGN) standard deviation.
4. The method of claim 1 , wherein the channel capacity is computed using:
b =log 2 (1+SNR c ),
where SNR c is a received signal-to-noise ratio (SNR) per 2-dimensions.
5. The method of claim 1 , wherein the real signal-to-noise ratio is computed by:
SNR r =A[Q −1 *( P b /K )] 2
where A is a number related to a size and shape of a constellation, P b is a required bit error rate, Q is an error function, and K is a nearest symbol neighbor.
6. The method of claim 1 , wherein the signal-to-noise ratio gap (SNR GAP) for each sub-channel is computed using:
Γ=SNR r −SNR c .
where SNR r is a received signal-to-noise ratio (SNR) and SNR c is an actual SNR.
7. The method of claim 6 , wherein the signal-to-noise ratio gap (SNR GAP) for each sub-channel is used to estimate the difference between the channel capacity and actual signal-to-noise ratio.
8. A system for loading bits into discrete multi-tone (DMT) sub-channels, the system comprising a DMT transceiver adapted to
compute a signal-to-noise ratio (SNR i ) for all available sub-channels;
compute a variable sub-channel coding gain according to an actual constellation size;
compute a variable sub-channel signal-to-noise ratio gap (SNR-GAP) according to the actual constellation size;
compute the number of bits to be allocated into each sub-channel; and
allocate bits to sub-channels based on an equal error rate.
9. The system of claim 8 , wherein the total bits calculated for loading into all the sub-channels is computed using:
b
=
∑
i
log
(
1
+
SNR
i
*
C
i
/
M
*
Γ
i
)
.
where b is the total bits, SNR i is the signal-to-noise ratio (SNR) in an i th sub-channel, C is a coding gain, M is a margin, and Γ is an SNR gap.
10. The system of claim 8 , wherein the probability of the symbol error is computed by:
P s =K ( i )* Q ( d ( i )/2σ).
where P is the probability, K(i) is a nearest symbol neighbor for an i th sub-channel, d(i) is a corresponding minimum distance, and σ is a 1-dimensional Additive White Gaussian Noise (AWGN) standard deviation.
11. The system of claim 8 , wherein the channel capacity is computed using:
b =log 2 (1+SNR c ).
where SNR c is a received signal-to-noise ratio (SNR) per 2-dimensions.
12. The system of claim 8 , wherein the real signal-to-noise ratio is computed by:
SNR r =A[Q −1 *( P b /K )] 2 .
where A is a number related to a size and shape of a constellation, P b is a required bit error rate, Q is an error function, and K is a nearest symbol neighbor.
13. The system of claim 8 , wherein the signal-to-noise ratio gap (SNR GAP) for each sub-channel is computed using:
Γ=SNR r −SNR c ,
where SNR r is a received signal-to-noise ratio (SNR) and SNR c is an actual SNR.
14. The system of claim 13 , wherein the signal-to-noise ratio gap (SNR GAP) for each sub-channel is used to estimate the difference between the channel capacity and actual signal-to-noise ratio.
15. A method for bit loading performed in a DMT transceiver based on an equal error rate principle for each sub-channel comprising:
computing a signal-to-noise ratio (SNR i ) for all available sub-channels;
computing a variable sub-channel coding gain according to an actual constellation size;
computing a variable sub-channel signal-to-noise ratio gap (SNR-GAP) according to the actual constellation size;
computing the number of bits to be allocated into each sub-channel; and allocating bits to sub-channels based on an equal error rate.
16. The method of claim 15 , wherein the actual constellation size varies from 1 to 15 bits per sub-channel.
17. The method of claim 15 , wherein the signal-to-noise ratio gap (SNR GAP) for each sub-channel is used to estimate a difference between channel capacity and an actual capacity usage.
18. The method of claim 17 , wherein computing the variable SNR-GAP depends on a shape and number of bits allocated to each sub-channel.
19. The method of claim 18 , further comprising tabulating different SNR-GAP values for different sub-channels.
20. The method of claim 19 , wherein tabulating different SNR-GAP values further comprises constructing a look-up table for the SNR-GAP with different bits.