CN100373840C - Method and apparatus for detecting normalized iterative soft interference cancelling signal - Google Patents

Abstract

The present invention relates to a normalized iteration soft interference cancellation signal detection method, particularly to a signal detection method for an iterative receiver of a multi-antenna wireless communication system. The detection method is an interative detection decoding receiving process based on iterative soft interference cancellation detection, and firstly, received multi-antenna signals are matched and merged in a normalized mode on space time domain at the receiving end in a multi-antenna wireless transmission iterative receiver. Secondly, iterative interference cancellation detection is carried out to obtain the estimation value of the signals and estimated interference noise variance, thirdly, demodulation and decoding are carried out, and a decoder carries out decoding to obtain the soft information of bits. The signals carry out mean value and variance reconstruction, is fed back to a detector to newly carry out interference cancellation and are demodulated and decoded, and the method and an interative detection decoding method of a device have higher performance. A normalized method makes the dynamic range of the data in a computing process diminished and is favorable to fixed point realization, and the number of argument is reduced when mapping is known about, which is favorable to look-up realization.

Description

Normalized iterative soft interference cancelling signal detecting method and device

Technical field

The present invention relates to a kind ofly, relate in particular to a kind of signal detecting method that is used for the multi-aerial radio communication system iterative receiver by using a plurality of send/receive antennas to come the wide-band mobile communication system of transmitting high speed data.Related wireless communication system has one or more reception antenna and one or more transmitting antenna.

Background technology

Under the condition of certain serious forgiveness, many many antennas of antenna transmission reception techniques are used to improve the efficiency of transmission of wireless communication system now, in the condition decline low transmission power of given transmission rate, perhaps improve the transmission rate of signal under the condition of given through-put power.At the carrier frequency height, and between the antenna under the distance situation far away, each transmitting antenna to the path loss between each reception antenna can be similar to thinks independent, in this case, transmission/the method for reseptance that multiple many antennas occurred mainly can space division multiplexing and Space Time Coding.SDM system sends independent data stream at each transmitting antenna, such as V-BLAST when dblast (vertical structure), so high Space Time Coding utilizes the orthogonality design, same signal at each transmitting antenna with multi-form transmission, to obtain to send diversity, the complexity of simplified receiver.

In broadband system, symbol time is expanded less than the multidiameter delay of multipath channel at interval, so can there be intersymbol interference in received signal in time.Multicarrier system can address this problem to a certain extent such as OFDM or the like.Owing to have a plurality of transmitting antennas, on each reception antenna, can have the interference between signals of many antennas.We establish number of transmit antennas is N, and the reception antenna number is M, and the maximum multipath delay spread of channel is L, uses s _n(k) expression n root transmitting antenna is used r at k transmission signal constantly _m(k) and z _m(k) represent signal and the additive white Gaussian noise that m root reception antenna k receives constantly, h respectively _{M, n}N root transmitting antenna is to the channel impulse response coefficient of m root reception antenna when (l) being l for time delay.Then have:

$r_m\left(k\right)=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}\underset{N=1}{\overset{N}{\mathrm{\Σ}}}{h}_{m,n}\left(l\right)\·s_n(k-l)+z_m\left(k\right)---\left[1\right]$

Iterative detection decoding method is the efficient ways that obtains to detect near optimal joint decoding in the multiaerial system with low complex degree.In this method, carry out repeatedly between detector and the decoder just adjudicating after the iteration to sending information bit.And in the process of non-last iteration, decoder calculates the soft information (recently representing with log-likelihood usually) of each bit and feeds back to detector and helps detector to detect, provide at decoder under the condition of the soft information of feedback, detector can obtain to export more accurately the result, thereby make decoder obtain more excellent decoding performance, successive iteration is to obtain more excellent performance like this.There are a lot of detectors to be done research based on iterative detection decoding method at present, mainly contain maximum a posteriori probability (MAP) algorithm, based on minimum mean square error criterion (MMSE) algorithm, approximate maximum likelihood method is globular decoding or the like for example, these method performances all compare better, but complexity is still very high.It is to reduce effective detection method of complexity that the present invention provides the iterative soft interference cancelling detection method that merges based on coupling.

Summary of the invention

Technical problem: the purpose of this invention is to provide a kind of normalized iterative soft interference cancelling signal detecting method and device that is used for wireless communication receiver, this method and apparatus tool mates to merge at one's leisure to be compared other multi-antenna receiving methods with iterative soft interference cancelling and has lower complexity; The method of iterative detection decoding has higher performance; The method for normalizing Data Dynamic scope in the computational process that not only makes diminishes, and is beneficial to fixed point and realizes, and reduce the number of understanding the mapping time-varying element, and realization helps tabling look-up.

Technical scheme: in multi-antenna wireless transmission iterative receiver, the detection method of the iterative soft interference cancelling that coupling merged when the normalization that detector can use the present invention to propose was empty, at first the normalization coupling that the multi-antenna signal that receives carries out on the sky time domain is merged at receiving terminal, carry out the iterative interference cancellation detection then and obtain the estimated value of signal and the interference noise variance of estimation, separate again to be in harmonious proportion and decipher, decoder is deciphered the soft information that obtains bit, signal is carried out average and variance reconstruction, feed back to detector and carry out Interference Cancellation again, separate to be in harmonious proportion and decipher.

The step of the Turbo method of reseptance that detects based on iterative soft interference cancelling is as follows:

Step 1: coupling merged when normalization was empty;

Step 2: if detect for the first time the decoding iteration, below the execution step by step, successive iterations execution in step 3;

21) the signal average is initialized as 0, variance is initialized as 1,

22) carry out multistage counteracting serial interference,

23) calculate the bit likelihood ratio,

24) enter step 4;

Step 3:

31) bit likelihood ratio reconstruction signal average and the variance of feeding back according to decoder,

32) carry out the one-level Parallel Interference Cancellation,

33) calculate the bit likelihood ratio,

34) enter step 4;

Step 4:

41) the bit likelihood ratio is knitted in reciprocal cross, decipher,

42) if the last decoding iteration that detects changes step 5 over to, if not, change step 43 over to,

43) carry out the iterative decoding of certain number of times, soft information does initialization for next time and detect the decoding of decoding iteration in the middle of keeping; Output decoding back bit likelihood ratio is also carried out reciprocal cross and is knitted,

44) enter step 3;

Step 5:

The bit likelihood ratio is adjudicated the output bit sequence after obtaining adjudicating.

Described checkout gear comprises: the soft inputting and soft output detector interweaves and anti-interleaver soft input soft output decode device, decision device; Merge cells when the soft inputting and soft output detector comprises sky, the Interference Cancellation unit, soft demodulating unit, the mean variance unit, the input termination input signal of merge cells when empty, the output termination Interference Cancellation unit of merge cells when empty, the soft demodulating unit of output termination of Interference Cancellation unit, the output of soft demodulating unit divides two-way, wherein one the tunnel connect the mean variance unit, the output termination Interference Cancellation unit of mean variance unit, another road of soft demodulating unit output connect interweave and anti-interleaver in anti-interleaver, the output termination soft input soft output decode device of anti-interleaver, the output of soft input soft output decode device divides two-way, wherein one the tunnel connect interweave and anti-interleaver in interleaver, another road outputs to decision device, the output termination mean variance unit of interleaver.

Merge cells when empty: multi-antenna signal is mated merging, under multipath channel, multipath signal is mated merging.Merge cells uses normalized merging method when empty during its hollow.

Detector cell: remove to disturb, the variance of the signal after obtaining disturbing and the interference noise of estimation, and calculate the soft information (recently representing with log-likelihood usually) of each bit.Wherein detector cell uses the detection method of soft inputting and soft output, and promptly detector cell can use the soft information of decoder feedback.Detector cell uses iterative soft interference cancelling to detect, and when detecting first, the mean variance of signal uses the method for tabling look-up to realize fast.

Interleaver and anti-interleaver unit: the bit likelihood ratio that anti-interleaver obtains detection is lined up the order of decoder.And interleaver is arranged the likelihood ratio of decoder output again according to the requirement of detector.

Soft input soft output decode device unit: the bit likelihood ratio after reciprocal cross knitted is deciphered according to the constraint of encoder according to encoding block, obtain the likelihood ratio after the new decoding, in non-last iteration, it is fed back to detector, and in the last iteration, will be used for the judgement of information bit.

Iterative receiver based on the iterative soft interference cancelling detection

Iterative detection decoding receiving system in the multiaerial system of the present invention mainly can be divided into the soft inputting and soft output detector, interweaves and anti-interleaver several parts such as soft input soft output decode device.

Coupling merged when 1, the soft inputting and soft output detector can be divided into the normalization sky, Interference Cancellation, and the bit likelihood ratio is calculated and the signal mean variance is rebuild four parts.During its hollow the coupling assembling section received signal according on the room and time dimension signal energy being collected, the coefficient of the signal after obtaining merging and corresponding signal component and interference signal.Interference Cancellation part merges signal and the corresponding coefficient that obtains during according to sky, also has the mean variance of rebuilding, and removes and disturbs, and calculate the variance of residual noise and interference.The signal after bit likelihood ratio calculating section disturbs according to removal and the variance of residual noise and interference are calculated the likelihood ratio of bit, and it is delivered to decoder.The mean variance rebuilding module of signal is with mean variance decoder feedback or that detect the likelihood ratio reconstruction signal that obtains.

2, anti-interleaver bit likelihood ratio that detection is obtained is lined up the order of decoder.And interleaver is arranged the likelihood ratio of decoder output again according to the requirement of detector.

3, the bit likelihood ratio after decoder is knitted reciprocal cross is deciphered according to the constraint of encoder according to encoding block, obtains the likelihood ratio after the new decoding, in non-last iteration it is fed back to detector, and will be used for the judgement of information bit in the last iteration.

Beneficial effect: the iteration receiving method that the present invention proposes has following advantage:

1. compare with the receiver of traditional detection decoding cascade, performance is improved greatly, has improved power efficiency greatly under the constant situation of spectrum efficiency.Use high order modulation under the constant situation of power efficiency, to improve spectrum efficiency greatly.

2. compare with other iterative detection decoding method, the iterative soft interference cancelling detection method that coupling merges during based on normalized sky has reduced the complexity of detector effectively, reduced the dynamic range of data, not only make complexity with number of transmit antennas and the linear growth of footpath number, and avoided the necessary inversion operation of MMSE, make algorithm more stable.

3. the method for normalizing Data Dynamic scope in the computational process that not only makes diminishes, and is beneficial to fixed point and realizes, and reduce the number of understanding the mapping time-varying element, and realization helps tabling look-up.

The method of reseptance that the present invention proposes is suitable for all kinds of wireless transmitting systems, mainly comprises

1. arrowband multi-aerial transmission system, promptly channel is the frequency-flat decline for signal.

2. single carrier multi-aerial transmission system, promptly channel is a frequency selective fading for signal, and the single carrier mode is adopted in transmission.

3. multicarrier and multi-antenna transmission system, such as OFDM, generalized multi-carrier system.

4. not only spread spectrum system can be used for, also non-spread spectrum system can be used for.

5. multi-access mode can adopt CDMA, TDMA, and FDMA etc. are various.

6. can be used for the communication between any two multi-antenna wireless devices.

Description of drawings

The iterative receiver block diagram that Fig. 1 detects for iterative soft interference cancelling.Wherein have: merge cells 11, Interference Cancellation unit 12, soft demodulating unit 13, mean variance unit 14 when soft inputting and soft output detector 1, sky interweave and anti-interleaver 2, anti-interleaver 21, interleaver 22, soft input soft output decode device 3, decision device 4.

Embodiment

The normalized iterative soft interference cancelling signal detecting method is in multi-antenna wireless transmission iterative receiver, at first the normalization coupling that the multi-antenna signal that receives carries out on the sky time domain is merged at receiving terminal, carry out the iterative interference cancellation detection then and obtain the estimated value of signal and the interference noise variance of estimation, separate again to be in harmonious proportion and decipher, decoder is deciphered the soft information that obtains bit, signal is carried out average and variance reconstruction, feed back to detector and carry out Interference Cancellation again, separate to be in harmonious proportion and decipher.

The step of the iteration receiving method that detects based on iterative soft interference cancelling is as follows:

Step 1: coupling merged when normalization was empty;

Step 2: if detect for the first time the decoding iteration, below the execution step by step, successive iterations execution in step 3,

21) the signal average is initialized as 0, variance is initialized as 1,

22) carry out multistage counteracting serial interference,

23) calculate the bit likelihood ratio,

24) enter step 4;

Step 3:

31) bit likelihood ratio reconstruction signal average and the variance of feeding back according to decoder,

32) carry out the one-level Parallel Interference Cancellation,

33) calculate the bit likelihood ratio,

34) enter step 4;

Step 4:

41) the bit likelihood ratio is knitted in reciprocal cross, decipher,

42) if the last decoding iteration that detects changes step 5 over to, if not, change step 43 over to,

43) carry out the iterative decoding of certain number of times, soft information does initialization for next time and detect the decoding of decoding iteration in the middle of keeping; Output decoding back bit likelihood ratio is also carried out reciprocal cross and is knitted,

44) enter step 3;

Step 5:

The bit likelihood ratio is adjudicated the output bit sequence after obtaining adjudicating.

Supposing the system has N root transmitting antenna, M root reception antenna.Information bit is through error correction coding (comprise the cascaded code that uses iterative decoding or based on the sign indicating number of figure), and then passes through Bit Interleave and modulation, is distributed to each transmitting antenna.On each transmitting antenna, can directly send symbol sebolic addressing, also can insert Cyclic Prefix and process IFFT (anti-fast fourier transform), become OFDM (OFDM) symbol and send.

Order:

$\underset{\_}{r}\left(k\right)=\left[\begin{array}{c}r\left(k\right)\\ r(k+1)\\ .\\ .\\ .\\ r(k+L-1)\end{array}\right],u\left(k\right)=\left[\begin{array}{c}s(k-L+1)\\ .\\ .\\ .\\ s\left(k\right)\\ .\\ .\\ .\\ s(k+L-1)\end{array}\right]\text{,}\underset{\_}{z}\text{(k)=}\left[\begin{array}{c}z\left(k\right)\\ z(k+1)\\ .\\ .\\ .\\ z(k+L-1)\end{array}\right]$

Then formula [1] can be expressed as

r(k)＝H·u(k)+z(k) [2]

Wherein, r(k) be the complete observation signal vector of the signal s (k) of ML * 1, H is the channel convolution matrix of N of ML * (2L-1), and u (k) is the transmission signal vector of (2L-1) N * 1, z(k) be the noise vector of ML * 1.

At receiving terminal, need estimate channel impulse response earlier with pilot tone or known array, after obtaining channel impulse response, the beginning iterative detection decoding.Coupling merges when at first the signal that sends being carried out normalized sky, promptly

$x_n\left(k\right)={\mathrm{\ρ}}_n^{-1}\·h_{N(L-1)+n}^H\·\underset{-}{r}\left(k\right)---\left[3\right]$

H in the formula [3] _{N (L-1)+n}Be N (the L-1)+n row of matrix H, ${\mathrm{\ρ}}_n=h_{N(L-1)+n}^H\·h_{N(L-1)+n}.$ And order ${\tilde{H}}_{N(L-1)+n}$ For matrix H is removed h _{N (L-1)+n}The part that the back is remaining,

For u (k) removes s _n(k) the remaining part in back, that is:

${\tilde{H}}_{N(L-1)+n}=[h_1,\·\·\·,h_{N(L-1)+n-1},h_{N(L-1)+n+1},\·\·\·,h_{N(2L-1)}]$

${\underset{\‾}{\overset{\‾}{u}}}_n\left(k\right)={[s_1(k-L+1),\·\·\·,s_{n-1}\left(k\right),s_{n+1}\left(k\right),\·\·\·,s_N(k+L-1)]}^T$

1, detects first

With the average of initializing signal is 0, and variance is 1.Utilize the interference cancellation signal after the mean variance of rebuilding is combined, remove and disturb, and calculate the variance of residual interference noise.

${\tilde{x}}_n\left(k\right)=x_n\left(k\right)-F_n^H\·E\left[{\underset{\‾}{\overset{\‾}{u}}}_n\left(k\right)\right]---\left[4\right]$

${\mathrm{\σ}}_{I,n}^2\left(k\right)=F_n^H\·\mathrm{cov}\left({\underset{\‾}{\overset{~}{u}}}_n\left(k\right)\right)\·F_n+{\mathrm{\ρ}}_n^{-1}\·{\mathrm{\σ}}_n^2---\left[5\right]$

Wherein $F_n^H={\mathrm{\ρ}}_n^{-1}\·h_{N(L-1)+n}^H{\tilde{H}}_{N(L-1)+n}.$ The interference that formula [4] has been removed other signals, formula [5] has calculated the variance of residual interference noise.We suppose the residual interference Gaussian distributed, then signal transmission equivalence are become Gaussian channel, and are as follows:

${\tilde{x}}_n\left(k\right)=s_n\left(k\right)+n^{\′}---\left[6\right]$

According to the signal modulation system it is carried out soft demodulation then, obtain the likelihood ratio of each coded-bit.Recently the mean variance of reconstruction signal is as follows with the bit likelihood:

$E\left[s_n\left(k\right)\right]=\underset{\mathrm{\α}}{\mathrm{\Σ}}\mathrm{\α}\·P(s_n\left(k\right)=\mathrm{\α})---\left[7\right]$

$\mathrm{cov}\left[s_n\left(k\right)\right]=\underset{\mathrm{\α}}{\mathrm{\Σ}}|\mathrm{\α}{|}^2\·P(s_n\left(k\right)=\mathrm{\α})-E{\left[s_n\left(k\right)\right]}^2---[8]$

If symbol is by bit d ₀, d ₁... d _Mc-1Mapping forms, then

$P(s_n\left(k\right)=\mathrm{\α})=\underset{k=0}{\overset{M_c-1}{\mathrm{\Π}}}\frac{1}{2}[1+{\stackrel{\‾}{d}}_k\·\tanh (L\left(d_k\right)/2)]---\left[9\right]$

Utilize formula [7], after the average and variance of formula [8] and formula [9] reconstruction signal, use formula [4] again, formula [5] carries out Interference Cancellation, carries out soft demodulation again.So enter decoding after the loop iteration 2～4.It should be noted that here that when detecting first what use is counteracting serial interference, promptly after current sign is by demodulation, just upgrade its average and variance, in the Interference Cancellation of successive character, promptly use the mean variance after upgrading.

Utilize formula [7], formula [8], formula [9] carry out mean variance to be rebuild and need carry out soft demodulation to signal earlier, and the computation complexity height, and along with order of modulation is exponential increase.We can further simplify the mean variance process of reconstruction according to normalized characteristic, and average and variance only are in the formula [6]

Function, promptly

$E\left[s_n\left(k\right)\right]=f_1({\tilde{x}}_n\left(k\right),{\mathrm{\σ}}_{I,n}^2\left(k\right))---\left[10\right]$

$\mathrm{cov}\left[s_n\left(k\right)\right]=f_2{(\tilde{x}}_n\left(k\right),{\mathrm{\σ}}_{I,n}^2\left(k\right))---[11]$

Wherein

Be two-dimensional function, in the realization of real system, can realize with the method for tabling look-up or tabling look-up in conjunction with interpolation.This method not only has lower complexity, and system only need change mapping table when adopting the different modulating mode, has the advantage that complexity is low and complexity is constant.

2, subsequent detection

The bit likelihood ratio that decoder is fed back is according to formula [7], the average and the variance of formula [8] and formula [9] reconstruction signal.Carry out all signals are carried out Interference Cancellation according to formula [4] and formula [5] again.Obtain carrying out soft demodulation again after the variance of signal estimated value and corresponding noise jamming, obtain the bit likelihood ratio.In subsequent detection, we have used 1 grade of Parallel Interference Cancellation.Be not according to the average and the variance of new testing result update signal before the Interference Cancellation.

The checkout gear of normalized iterative soft interference cancelling signal detecting method of the present invention comprises: soft inputting and soft output detector 1 interweaves and anti-interleaver 2 soft input soft output decode device 3, decision device 4; Merge cells 11, Interference Cancellation unit 12, soft demodulating unit 13, mean variance unit 14 when soft inputting and soft output detector 1 comprises sky, the input termination input signal " r of merge cells 11 when empty _m(K) "; the output termination Interference Cancellation unit 12 of merge cells 11 when empty; the soft demodulating unit 13 of the output termination of Interference Cancellation unit 12; output of soft demodulating unit 13 divides two-way; wherein a tunnel connect mean variance unit 14; the output termination Interference Cancellation unit 12 of mean variance unit 14, another road of soft demodulating unit 13 outputs connect interweave and anti-interleaver 2 in anti-interleaver 21, the output termination soft input soft output decode device of anti-interleaver 21, the output of soft input soft output decode device divides two-way, wherein one the tunnel connect interweave and anti-interleaver 2 in interleaver 22, another road outputs to decision device 4.

The iteration that soft dried counteracting detects based on iteration receives and handles is in multi-antenna wireless transmission iterative receiver, at first the normalization coupling that the multi-antenna signal that receives carries out on the sky time domain is merged at receiving terminal, carry out the iterative interference cancellation detection then and obtain the estimated value of signal and the interference noise variance of estimation, separate again to be in harmonious proportion and decipher, decoder is deciphered the soft information that obtains bit, signal is carried out average and variance reconstruction, feed back to detector and carry out Interference Cancellation again, separate to be in harmonious proportion and decipher.

As follows based on the iteration reception treatment step that iterative soft interference cancelling detects:

Step 1: coupling merged when normalization was empty;

Promptly according to formula [3], first design factor ${\mathrm{\ρ}}_n=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}\underset{m=1}{\overset{M}{\mathrm{\Σ}}}\·\left|h_{m,n}\left(l\right)\right|$ And coefficient $f_{n,n^{\′}}\left(l\right)={\mathrm{\ρ}}_n^{-1}\·\underset{l^{\′}=\max \{0,l\}}{\overset{\min \{L-1,l+L-1\}}{\mathrm{\Σ}}}\underset{m=1}{\overset{M}{\mathrm{\Σ}}}{h}_{m,n}^{*}\left(l^{\′}\right)\·h_{m,n^{\′}}(l^{\′}-l),$ L=-(L-1) wherein ..., L-1 calculates then $x_n\left(k\right)={\mathrm{\ρ}}_n^{-1}\·\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}\underset{m=1}{\overset{M}{\mathrm{\Σ}}}{h}_{m,n}^{*}\left(l\right)\·r_m(k+l).$

Step 2: if detect for the first time the decoding iteration, below the execution step by step, successive iterations execution in step 3,

2.1) the signal average is initialized as 0, variance is initialized as 1, promptly ${\stackrel{\‾}{s}}_n\left(k\right)=0,$ ${\mathrm{\σ}}_{s,n}^2\left(k\right)=1.$

2.2) carry out multistage counteracting serial interference, establishing B is the signal block length, T is an Interference Cancellation progression, makes n=0, k=0, t=0.

2.2.1) calculate ${\tilde{x}}_n\left(k\right)=x_n\left(k\right)-\underset{l=1}{\overset{L-1}{\mathrm{\Σ}}}{f}_{n,n^{\′}}(-l)\·s_n(k-l)-\underset{l=1}{\overset{L-1}{\mathrm{\Σ}}}{f}_{n,n^{\′}}\left(l\right)\·s_n(k+l)$ With

${\mathrm{\σ}}_{l,n}^2\left(k\right)={\mathrm{\ρ}}_n^{-1}\·{\mathrm{\σ}}_z^2+\underset{l=1}{\overset{L-1}{\mathrm{\Σ}}}\left[\right|f_{n,n^{\′}}\left(l\right){|}^2\·{\mathrm{\σ}}_{s,n^{\′}}^2(k+l)+{\left|f_{n,n^{\′}}(-l)\right|}^2\·{\mathrm{\σ}}_{s,n^{\′}}^2(k-l)]$

2.2.2) basis $P\left(x_n\left(k\right)\right|s_n\left(k\right)=\mathrm{\α})=\frac{1}{\sqrt{2\mathrm{\π}}{\mathrm{\σ}}_{l,n}\left(k\right)}\·\exp (-\frac{{({\tilde{x}}_n\left(k\right)-\mathrm{\α})}^2}{{\mathrm{\σ}}_{l,n}^2\left(k\right)})$ Signal calculated s _n(k) send the probability of symbol for each.According to ${\stackrel{\‾}{s}}_n\left(k\right)=\underset{\mathrm{\α}}{\mathrm{\Σ}}\mathrm{\α}\·P\left(x_n\left(k\right)\right|s_n\left(k\right)=\mathrm{\α})$ With

${\mathrm{\σ}}_{s,n}^2\left(k\right)=\underset{\mathrm{\α}}{\mathrm{\Σ}}|\mathrm{\α}{|}^2\·P\left(x_n\left(k\right)\right|s_n\left(k\right)=\mathrm{\α})-{\stackrel{\‾}{s}}_n^2\left(k\right)$

K=k+1, if k=B, n=n+1 then, k=0; If n〉N, n=1 then, t=t+1 is if t=T then enters step 2.3, otherwise enters step 2.2.1.

2.3) calculating bit likelihood ratio $L\left(b\right)=\log \frac{\underset{a:b=+1}{\mathrm{\Σ}}P\left(x_n\left(k\right)\right|s_n\left(k\right)=\mathrm{\α})}{\underset{a:b=-1}{\mathrm{\Σ}}P\left(x_n\left(k\right)\right|s_n\left(k\right)=\mathrm{\α})}$

2.4) enter step 4;

Step 3:

3.1) according to bit likelihood ratio reconstruction signal average and variance, the at first basis of decoder feedback

$P(s_n\left(k\right)=\mathrm{\α})=\underset{k=0}{\overset{M_C-1}{\mathrm{\Π}}}\frac{1}{2}[1+{\tilde{d}}_k\·\tanh (L\left(d_k\right)/2)]$ By the bit likelihood ratio compute sign probability of feedback, basis then ${\stackrel{\‾}{s}}_n\left(k\right)=\underset{\mathrm{\α}}{\mathrm{\Σ}}\mathrm{\α}\·P(s_n\left(k\right)=\mathrm{\α})$ With ${\mathrm{\σ}}_{s,n}^2\left(k\right)=\underset{\mathrm{\α}}{\mathrm{\Σ}}|\mathrm{\α}{|}^2\·P(s_n\left(k\right)=\mathrm{\α})-{\stackrel{\‾}{s}}_n^2\left(k\right)$ Signal calculated average and variance.

3.2) carry out the one-level Parallel Interference Cancellation

Calculate ${\tilde{x}}_n\left(k\right)=x_n\left(k\right)-\underset{l=1}{\overset{L-1}{\mathrm{\Σ}}}{f}_{n,n^{\′}}(-l)\·s_n(k-l)-\underset{l=1}{\overset{L-1}{\mathrm{\Σ}}}{f}_{n,n^{\′}}\left(l\right)\·s_n(k+l)$ With

${\mathrm{\σ}}_{l,n}^2\left(k\right)={\mathrm{\ρ}}_n^{-1}\·{\mathrm{\σ}}_z^2+\underset{l=1}{\overset{L-1}{\mathrm{\Σ}}}\left[\right|f_{n,n^{\′}}\left(l\right){|}^2\·{\mathrm{\σ}}_{s,n^{\′}}^2(k+l)+{|f_{n,n^{\′}}(-l)}^2\·{\mathrm{\σ}}_{s,n^{\′}}^2(k-l)]$

3.3) calculating bit likelihood ratio, (with step 2.3)

3.4) enter step 4;

Step 4:

4.1) reciprocal cross knits the bit likelihood ratio, decipher,

4.2) if the last decoding iteration that detects changes step 5 over to, if not, change step 43 over to,

4.3) carry out the iterative decoding of certain number of times, soft information does initialization for next time and detect the decoding of decoding iteration in the middle of keeping; Output decoding back bit likelihood ratio is also carried out reciprocal cross and is knitted,

4.4) enter step 3;

Step 5:

The bit likelihood ratio is adjudicated the output bit sequence after obtaining adjudicating.

Wherein step 2.2.2 computation of mean values variance can realize with tabling look-up, promptly

${\stackrel{\‾}{s}}_n\left(k\right)=f_1({\tilde{x}}_n\left(k\right),{\mathrm{\σ}}_{l,n}^2\left(k\right))$ With ${\mathrm{\σ}}_{s,n}^2\left(k\right)=f_2({\tilde{x}}_n\left(k\right),{\mathrm{\σ}}_{l,n}^2\left(k\right)),$ f ₁() and f ₂() all made table in advance and stored.

Claims (3)

1. normalized iterative soft interference cancelling signal detecting method, it is characterized in that this detection method is based on the iterative detection decoding reception processing that iterative soft interference cancelling detects, in multi-antenna wireless transmission iterative receiver, at first the normalization coupling that the multi-antenna signal that receives carries out on the sky time domain is merged at receiving terminal, carry out the iterative interference cancellation detection then and obtain the estimated value of signal and the interference noise variance of estimation, separate again to be in harmonious proportion and decipher, decoder is deciphered the soft information that obtains bit, signal is carried out average and variance reconstruction, feed back to detector and carry out Interference Cancellation again, separate to be in harmonious proportion and decipher.

2. normalized iterative soft interference cancelling signal detecting method according to claim 1, it is as follows to it is characterized in that the iterative detection decoding that detects based on iterative soft interference cancelling receives treatment step:

Step 1: coupling merged when normalization was empty;

Step 2: if detect for the first time the decoding iteration, below the execution step by step, successive iterations execution in step 3,

21) the signal average is initialized as 0, variance is initialized as 1,

22) carry out multistage counteracting serial interference,

23) calculate the bit likelihood ratio,

24) enter step 4;

Step 3:

31) bit likelihood ratio reconstruction signal average and the variance of feeding back according to decoder,

32) carry out the one-level Parallel Interference Cancellation,

33) calculate the bit likelihood ratio,

34) enter step 4;

Step 4:

41) the bit likelihood ratio is knitted in reciprocal cross, decipher,

42) if the last decoding iteration that detects changes step 5 over to, if not, change step 43 over to,

43) carry out the iterative decoding of certain number of times, soft information does initialization for next time and detect the decoding of decoding iteration in the middle of keeping; Output decoding back bit likelihood ratio is also carried out reciprocal cross and is knitted,

44) enter step 3;

Step 5:

The bit likelihood ratio is adjudicated the output bit sequence after obtaining adjudicating.

3. checkout gear that is applicable to the described normalized iterative soft interference cancelling signal detecting method of claim 1, it is characterized in that described checkout gear comprises: soft inputting and soft output detector (1), interweave and anti-interleaver (2) soft input soft output decode device (3), decision device (4); Soft inputting and soft output detector (1) is merge cells (11) when comprising sky, Interference Cancellation unit (12), soft demodulating unit (13), mean variance unit (14), the input termination input signal of merge cells (11) when empty, the output termination Interference Cancellation unit (12) of merge cells (11) when empty, the soft demodulating unit of output termination (13) of Interference Cancellation unit (12), the output of soft demodulating unit (13) divides two-way, wherein one the tunnel connect mean variance unit (14), the output termination Interference Cancellation unit (12) of mean variance unit (14), another road of soft demodulating unit (13) output connect interweave and anti-interleaver (2) in anti-interleaver (21), the output termination soft input soft output decode device of anti-interleaver (21), the output of soft input soft output decode device divides two-way, wherein one the tunnel connect interweave and anti-interleaver (2) in interleaver (22), another road outputs to decision device (4), the output termination mean variance unit (14) of interleaver (22).