CN101237434A - A soft judgement method for Graham M-PSK modulation - Google Patents

Abstract

The invention provides a soft demodulation algorithm with low complexity for high price gray code mapping M-PSK demodulation, consisting of a balancing procedure, a channel estimation procedure, a balancing procedure on a received signal r, a hard decision procedure, a bit likelihood ratio calculation procedure and a coding procedure. The soft demodulation algorithm is characterized in that the method for calculating the bit likelihood ratio is easier to realize compared with the prior constellation search method and greatly decreases the calculated mount.

Description

A kind of soft decision method of gray mappings M-PSK modulation

Technical field

This invention belongs to the modulation field of radio communication.

Background technology

Along with the quick increase of wireless communication user, how to improve the main target that the availability of frequency spectrum has become the design wireless communication system.

Gray mappings M-PSK (multi-system phase shift keying) modulation is widely used in the radio digital communication system owing to the availability of frequency spectrum is high.Wherein, what the modulation constellation that gray mappings representative is adopted was carried out is Gray code, thereby can guarantee to have only a bit symbols difference between any two adjacent constellation point, thus can with communication system that efficient coding and decoding combines in effectively improve systematic function.

In wireless system with FEC (forward error correction), the relative hard decision demodulation method of soft decision demodulation method can provide bigger performance improvement, see document: F.Tosato and P.Bisaglia; " Simplifed soft-output demapper for binary interleavedCOFDM with application to HIPERLAN/2; " in proc.IEEE ICC ' 02,2002, pp, 664-668.

In sum, in the transmitting terminal use gray mappings M-PSK of wireless communication system modulation, use the wireless modulation-demodulation technology of soft decision demodulation to be used widely at receiving terminal simultaneously.

Traditional soft demodulation decision method is as follows: received signal r can represent as the formula (1) usually.

r＝ax+n (1)

Wherein x transmits, and it belongs to the constellation point S set={ s of gray mappings M-PSK modulation ₁, s ₂... s _M, wherein M represents the exponent number of modulation, and a is a channel gain, and n is that average is 0, and variance is σ ²Additive white Gaussian noise.Each constellation point s _iRepresent k=log ₂The information of M bit is designated as: b ₀, b ₁..., b _K-1

According to document M.M.Wang, Weimin Xiao, and T.Brown, " Soft decision metric for QAM withchannel estimation error. " IEEE trans, Commun, vol.50.pp.1058-1061.July 2002.I bit b among the received signal r _iThe bit likelihood ratio can be expressed as:

$\mathrm{LLR}\left(b_i\right)=\frac{1}{{\mathrm{\σ}}^2}({|r-{\mathrm{as}}_i^{-}|}^2-{|r-{\mathrm{as}}_i^{+}|}^2),i=\mathrm{0,1}...,k-1---\left(2\right)$

LLR (b wherein _i) expression b _iThe bit likelihood ratio, s _i ^-Represent that i bit is 0 and from the nearest constellation point of r/a, s _i ⁺Represent that the i bit is 1 and from the nearest constellation point of r/a, σ ²It is the noise power of received signal.

According to formula (2), calculate bit likelihood ratio LLR (b _i) that will seek is s _i ^-And s _i ⁺Suppose to use s ^*The constellation point that expression is determined by hard decision, s so ^*Be the nearest constellation point of M-PSK planisphere middle distance r/a, obviously, if s ^*I bit be 0, then $s_i^{-}=s^{*};$ If s ^*I bit be 1, then $s_i^{*}=s^{*}.$ Like this, in traditional soft demodulating method, s _i ^-And s _i ⁺One of them can obtain by hard decision, but another need among the searching for constellation figure have a few and could obtain, for high order modulation, its amount of calculation is very big.

Summary of the invention

At in the traditional soft demodulation method, calculating bit likelihood ratio LLR (b _i) time, because s _i ^-Or s _i ⁺Need among the searching for constellation figure have a few and could obtain, thus for high order modulation, the problem that its amount of calculation is very big the invention provides the low complexity soft demodulating algorithm that a kind of high-order Gray code mapping M-PSK modulates, and comprises transmitting terminal processing procedure and receiving terminal processing procedure.

Transmitting terminal processing procedure such as following step are finished:

Step 1: the binary data to input is encoded, to reduce the random and burst error that it may occur in transmission course.

Step 2: the data behind the coding are carried out gray mappings M-PSK modulation obtain sending signal x ∈ S.

Receiving terminal processing procedure following steps are to received signal finished:

Step 3: channel estimating.According to certain estimation criterion, from received signal In carry out channel estimating, for next step equilibrium provides condition.

Step 4: channel equalization is eliminated and is disturbed.The estimation channel that obtains according to step 3 r to received signal carries out equilibrium, in the hope of eliminating channel to sending the interference that signal x is caused, obtains the signal after the equilibrium

Step 5: hard decision.Right

Carry out hard decision, obtain to declare firmly constellation point $s^{*}=(b_0^{*},b_1^{*},...,b_{i-1}^{*},b_i^{*},b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}).$ B wherein _i ^*(i=0,1,2...k-1) expression s ^*I bit.

Step 6: calculate received signal The 0th bit b ₀Bit likelihood ratio LLR (b ₀), concrete steps are as follows:

Substep 1: judge described bit b ₀ ^*,

If ● $b_0^{*}=0,$ Then order $s_0^{-}=s^{*}=(0,b_1^{*},...,b_{i-1}^{*},b_i^{*},b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ And make constellation point s ₀ ⁺The 0th bit get 1, s ₀ ⁺The 1st bit with declare constellation point s firmly ^*The 1st bit identical, s ₀ ⁺All the other bits be 0 entirely, promptly $s_0^{+}=(1,b_1^{*},\mathrm{0,0},...,\mathrm{0,0,0},...,\mathrm{0,0});$

If ● $b_0^{*}=1,$ Then $s_0^{+}=s^{*}=(1,b_1^{*},...,b_{i-1}^{*},b_i^{*},b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ Constellation point s ₀ ^-The 0th bit get 0, s ₀ ^-The 1st bit with declare symbol s firmly ^*The 1st bit is identical, s ₀ ^-All the other bits be 0 entirely, promptly

$s_0^{-}=(0,b_1^{*},\mathrm{0,0},...\mathrm{0,0,0},...,\mathrm{0,0});$

Substep 2: with substep 1 resulting s ₀ ^-With s ₀ ⁺Substitution formula (3) calculates received signal The likelihood ratio of the 0th bit:

$\mathrm{LLR}\left(b_0\right)=\frac{1}{{\mathrm{\&sigma;}}^2}({|\hat{r}-s_0^{-}|}^2-{\left|\hat{r}{-\mathrm{<figure-callout\; id="0"\; label="s"\; filenames="S2008100449138E00011.png"\; state="\{\{state\}\}">s</figure-callout>}}_0^{+}\right|}^2---\left(3\right)$

Step 7: calculate received signal All the other bit b _i(i=1,2..., likelihood ratio LLR (b k-1) _i) method is:

Substep 1: the b that calculates gained in the decision steps 5 _i ^*,

If ● $b_i^{*}=0,$ Then order $s_i^{-}=s^{*}=(b_0^{*},b_1^{*},...,b_{i-1}^{*},0,b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ And make constellation point s _i ⁺The the 0th to i-1 bit with declare symbol s firmly ^*The the 0th to i-1 bit equate that i bit is that 1, an i+1 bit is 1, all the other bits are 0 entirely, promptly $s_i^{+}=(b_0^{*},b_1^{*},...,b_{i-1}^{*},\mathrm{1,1,0,0},...,0);$

If ● $b_i^{*}=1,$ Then order $s_i^{+}=s^{*}=(b_0^{*},b_1^{*},...,b_{i-1}^{*},1,b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ And make constellation point s _i ^-The the 0th to i-1 bit with declare symbol s firmly ^*The the 0th to i-1 bit equate that i bit is that 0, an i+1 bit is 1, all the other bits are 0 entirely, promptly $s_i^{-}=(b_0^{*},b_1^{*},...,b_{i-1}^{*},\mathrm{0,1,0,0},...,0);$

Substep 2: with s _i ^-With s _i ⁺Substitution formula (4) calculates

In i bit b _iThe bit likelihood ratio:

$\mathrm{LLR}\left(b_i\right)=\frac{1}{{\mathrm{\&sigma;}}^2}({|\hat{r}-s_i^{-}|}^2-{\left|\hat{r}{-s}_i^{+}\right|}^2),i=\mathrm{1,2}...k-1---\left(4\right)$

Step 8: utilize the likelihood ratio of each bit that step 6 and step 7 calculate, decipher, then the data after the output decoding.

Need to prove:

1) the gray mappings M-PSK modulator approach in the step 2 must adopt 3GPP predetermined rule in TR 25.848 V4.0.0 standards, as shown in Figure 4; Details are seen TR 25.848 V4.0.0;

2) simplification of received signal correspondence was declared algorithm firmly and is seen document under gray mappings M-PSK modulated in the step 5: " C.K; Yuen; A fast analog to gray code converter; " in Proceedings of the IEEE, vol.65, pp.1510-1511, Oct.1977 utilizes this algorithm can reduce the hard decision complexity of gray mappings M-PSK modulation effectively.

Essence of the present invention and effect:

Essence of the present invention is: determine s with the hard decision method _i ^-With s _i ⁺After in the symbol one of them, just can be according to s _i ^-With s _i ⁺Between internal relation directly determine s _i ^-With s _i ⁺In the another one symbol, calculate the likelihood ratio LLR (b of each bit of received signal then _i), thereby avoided complicated planisphere search procedure.Therefore, this algorithm can reduce the complexity of soft demodulation effectively, is convenient to hardware and realizes, reduces hardware cost simultaneously, makes it to be used widely in following high-speed mobile communication system.

Description of drawings

Fig. 1 is LLR (b of the present invention _i) realization flow that calculates.

Fig. 2 is the 16-PSK gray mappings planisphere of 3GPP regulation.

Embodiment:

Provide the implementation method of a concrete this patent below, need to prove: the parameter in the following example does not influence the generality of this patent.

Make a start

It is 300 that input signal carries out code length, and code check carries out gray mappings after being 1/2 LDPC coding, carries out the 16-PSK modulation then, after the signal power normalization after will modulating again through wireless channel.

As shown in Figure 2, M=16 here, k=4.

Receiving end

Carry out channel estimating, equilibrium to received signal successively, obtain the signal after equilibrium

Right

Declare acquisition firmly and declare symbol s firmly ^*Method according to step 7 and step 8 in the summary of the invention is calculated reception

The bit likelihood ratio of each bit.After at last the bit likelihood ratio of received signal being delivered to decoding module decoding, obtain judgement output.

Claims (1)

1, the soft decision method of a kind of gray mappings M-PSK modulation by equilibrium, channel estimating, r carries out equilibrium, hard decision, calculating bit likelihood ratio, decode procedure and forms to received signal, is characterized in that: obtain constellation point at hard decision $s^{*}=(b_0^{*},b_1^{*},...,b_{i-1}^{*},b_i^{*},b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*})$ Afterwards, the method for calculating bit likelihood ratio is:

Step 1: calculate received signal

The 0th bit b ₀Bit likelihood ratio LLR (b ₀), concrete steps are as follows:

Substep 1: judge described bit b ₀ ^*,

If ● $b_0^{*}=0,$ Then order $s_0^{-}=s^{*}=(0,b_1^{*},...,b_{i-1}^{*},b_i^{*},b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ And make constellation point s ₀ ⁺The 0th bit get 1, s ₀ ⁺The 1st bit with declare constellation point s firmly ^*The 1st bit identical, s ₀ ⁺All the other bits be 0 entirely, promptly $s_0^{+}=(1,b_1^{*},\mathrm{0,0},...,\mathrm{0,0,0},...,\mathrm{0,0});$

If ● $b_0^{*}=1,$ Then $s_0^{+}=s^{*}=(1,b_1^{*},...,b_{i-1}^{*},b_i^{*},b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ Constellation point s ₀ ^-The 0th bit get 0, s ₀ ^-The 1st bit with declare symbol s firmly ^*The 1st bit is identical, s ₀ ^-All the other bits be 0 entirely, promptly $s_0^{-}=(0,b_1^{*},\mathrm{0,0},...,\mathrm{0,0,0},...,\mathrm{0,0});$

Substep 2: with substep 1 resulting s ₀ ^-With s ₀ ⁺Substitution formula (1) calculates received signal The likelihood ratio of the 0th bit:

$\mathrm{LLR}\left(b_0\right)=\frac{1}{{\mathrm{\&sigma;}}^2}({|\hat{r}-s_0^{-}|}^2-{|\hat{r}-s_0^{+}|}^2)---\left(3\right)$

Step 2: calculate received signal

All the other bit b _i(i=1,2..., likelihood ratio LLR (b k-1) _i) method is:

Substep 1: the b that calculates gained in the decision steps 5 _i ^*,

If ● $b_i^{*}=0,$ Then order $s_i^{-}=s^{*}={(b}_0^{*},b_1^{*},...,b_{i-1}^{*},0,b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ And make constellation point s _i ⁺The the 0th to i-1 bit with declare symbol s firmly ^*The the 0th to i-1 bit equate that i bit is that 1, an i+1 bit is 1, all the other bits are 0 entirely, promptly

$s_i^{+}={(b}_0^{*},b_1^{*},...,b_{i-1}^{*},\mathrm{1,1,0,0},...,0);$

If ● $b_i^{*}=1,$ Then order $s_i^{+}=s^{*}={(b}_0^{*},b_1^{*},...,b_{i-1}^{*},1,b_{i+1}^{*},...,b_{k-2}^{*},b_{k-1}^{*}),$ And make constellation point s _i ^-The the 0th to i-1 bit with declare symbol s firmly ^*The the 0th to i-1 bit equate that i bit is that 0, an i+1 bit is 1, all the other bits are 0 entirely, promptly $s_i^{-}=(b_0^{*},b_1^{*},...,b_{i-1}^{*},\mathrm{0,1,0,0},...,0);$

Substep 2: with s _i ^-With s _i ⁺Substitution formula (2) calculates

In i bit b _iThe bit likelihood ratio:

$\mathrm{LLR}\left(b_i\right)=\frac{1}{{\mathrm{\&sigma;}}^2}({|\hat{r}-s_i^{-}|}^2-{|\hat{r}-s_i^{+}|}^2),i=\mathrm{1,2}...k-1---\left(2\right)$

Step 3: utilize the likelihood ratio of each bit that step 1 and step 2 calculate, decipher, then the data after the output decoding.

Images