CN102412931A - Serial interference cancellation and detection method and system for multi-antenna bit interleaved coding system - Google Patents

Abstract

The invention relates to a serial interference cancellation and detection method and system for a multi-antenna bit interleaved coding system. The system disclosed by the invention comprises an MMSE (Minimum Mean Square Error) equalization filter module, a post-detection signal-to-noise ratio computing module, a sequencing module, a data channel section module, a soft demodulation module, a data revaluation module, an interference cancellation module, a residual interference estimation module, a similar MMSE equalization filter module, a similar post-detection signal-to-noise ratio computing module and a judgment module. According to the method and the system, the technical problem that the residual interference of 'eliminated' interference is considered to be zero existing in the conventional serial interference cancellation method is solved. In the invention, a correct flow device is provided for an MMSE serial interference cancellation method, a residual interference estimation method is further provided, and a correct form for estimating the MMSE of data in residual layers and a corresponding computation form of a post-detection signal-to-noise ratio used in the presence of 'residual interference' are provided.

Description

The counteracting serial interference detection method and the system of many antennas bit interweaving encoding system

Technical field

The present invention relates to wireless communication field, relate in particular in the MIMO antenna system (MIMO) Interference Cancellation detection technique multi-layer data.

Background technology

MIMO radio communication system can utilize multiplexing to the space, is sending a multi-layer data with a frequency domain resource simultaneously, thereby is increasing the power system capacity of wireless communication system.Yet the detection of every layer data has great influence to the inhibition ability from the interference of other transmission antennas transmit data to systematic function in its testing process.Least mean-square error (MMSE) method is a kind of receiving terminal detection method commonly used, and the present invention has added soft-decision counteracting serial interference detection method and system on the basis of least mean-square error method, progressively eliminate the interference of each layer data to other layer data.

What the present invention was different with the traditional serial interference cancellation method is, the traditional serial interference cancellation method to the interference of " elimination " often think that its residual interference is zero.Yet " elimination " residual interference of disturbing is very important to playing a part in the equalization filtering of follow-up other layer data and the calculating of detection signal-to-noise ratio thereafter.We know that soft bit LLR is used in decoding in the Bit Interleave coded system, and back detection signal-to-noise ratio provides the weighting to the LLR confidence level in the calculating of LLR, if the calculating of back detection signal-to-noise ratio is incorrect, and the very big loss that can bring systematic function.The present invention not only provides the correct flow process and the system of least mean-square error counteracting serial interference method; The method of estimation of residual interference more is provided; Also provide under the situation that has ' residual interference '; The correct form that rest layers data least mean-square error is estimated, and the form of calculation of corresponding back detection signal-to-noise ratio.

Summary of the invention

For the interference of " eliminations " to that solves that existing counteracting serial interference method exists often thinks that its residual interference is zero technical problem; The present invention provides counteracting serial interference detection method and the detection system based on many antennas bit interweaving encoding system; Be applicable to that transmitting terminal and receiving terminal dispose respectively under the situation of many antennas in the radio communication; The correct flow path device of least mean-square error counteracting serial interference method not only is provided, the method for estimation of residual interference more is provided, also provide under the situation that has " residual interference "; The correct form that rest layers data least mean-square error is estimated, and the form of calculation of corresponding back detection signal-to-noise ratio.

Technical solution of the present invention:

The present invention is based on many antennas bit interweaving encoding system the counteracting serial interference detection method based on system configuration comprise: MMSE equalization filter module, back detection signal-to-noise ratio computing module, order module; Choose the data channel module, soft demodulation module, data revaluation module; Interference cancellation module; The residual interference estimation module, type MMSE equalization filter module, type back detection signal-to-noise ratio computing module.

Based on the counteracting serial interference detection method of many antennas bit interweaving encoding system, its special character is: may further comprise the steps:

1] in the MIMO model, calculates MMSE and detect data

and back detection signal-to-noise ratio w according to MMSE equalization filter

Wherein: MIMO model: y=Hx+n, transmitting terminal configuration Nt transmit antennas, receiving terminal configuration Nr root reception antenna;

G _MMSEBe the MMSE equalization filter, its expression formula is: G _MMSE=H ^H(HH ^H+ σ _n ²I _Nr) ^-1

H is a channel matrix, the element h of H _IjRepresent the channel coefficients of j transmit antennas, x=[x to i root reception antenna _k] ^T, k=1,2 ... Nt, expression transmission data, y=[y _k] ^T, k=1,2 ... Nr, expression receives data, n=[n _k] ^T, k=1,2 ... Nr, represent the noise of every reception antenna,

Calculate back detection signal-to-noise ratio w:w=1/ (1-diag (G _MMSEH))-1;

2] sort from big to small according to back detection signal-to-noise ratio w, obtain ordered series of numbers m=[m ₁, m ₂... m _Nt], satisfy $w_{m_1}>w_{m_2}>...>w_{m_{\mathrm{Nt}}};$

3] choose maximum back detection signal-to-noise ratio data channel m _iAs the current data channel of choosing, make i=1, export corresponding back detection signal-to-noise ratio of the current data channel of choosing and corresponding detection data;

4] to step 3] in the current data channel of choosing carry out soft demodulation process, according to

Calculate the LLR of this data channel,

Represent the current m that chooses _iThe LLR value that data channel k bit is corresponding;

5] according to step 4] in the LLR that obtains the detection data of the current data channel of choosing are carried out revaluation, obtain revaluation and detect data

$P\left(\mathrm{\θ}\right|{\hat{x}}_{m_i})=\underset{m=1}{\overset{M}{\mathrm{\Π}}}\frac{1}{2}(1+(2c_{k,m}-1)\tanh \left(\frac{L_k^{m_i}}{2}\right))$

Wherein θ representes the constellation point value of modulator approach that this data channel adopts, c _{K, m}The corresponding m bit value (0/1) of expression constellation point value θ;

6] from receive data, eliminate the current interference of choosing data channel, obtain the reception data model

after the interference eliminated

7] the current residual interference energy value of choosing data channel after basis

the calculating interference eliminated;

8] utilize step 6] in reception data model after the interference eliminated that obtains, use a type MMSE equalization filter, calculate m _I+1Detection data on the circuit-switched data passage:

And output

Type MMSE equalization filter wherein $G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}})}^{-1},$

H _M=[h _k], k=1,2 ... Nt, k ≠ m ₁, m ₂... m _i, eliminate the data channel that had detected, the channel matrix of remaining data passage among the expression H;

9] basis

Calculate m _I+1The back detection signal-to-noise ratio of circuit-switched data passage;

10] judge, if i+1 is less than Nt, i++, execution in step 4]; Otherwise if i+1 equals Nt, each data channel all detects, then finishes.

Step 8] middle G ' _MMSEAdopt following method to calculate:

Order $A_i=(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}}),$ Class MMSE equalization filter is written as following form:

$G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{A_i}^{-1}$

A wherein _i ^-1Obtain through following alternative manner:

$A_i^{-1}=A_{i-1}^{-1}-\frac{A_{i-1}^{-1}{h}_{m_i}{\left(A_{i-1}^{-1}{h}_{m_i}\right)}^H}{\frac{1}{q_{m_i}-1}+h_{m_i}^H\left(A_{i-1}^{-1}{h}_{m_i}\right)}$

$A_0^{-1}={({\mathrm{HH}}^H+{{\mathrm{\σ}}_n}^{2}I)}^{-1}.$

Comprise the MMSE equalization filter module that connects successively, back detection signal-to-noise ratio computing module, order module; Choose the data channel module, soft demodulation module, data revaluation module; Interference cancellation module, residual interference estimation module, type MMSE equalization filter module; Detection signal-to-noise ratio computing module and judge module after type

MMSE equalization filter module according to the model used in MIMO MMSE equalization filter

calculate MMSE detector data

where: MIMO models:

Y=Hx+n, transmitting terminal configuration Nt transmit antennas, receiving terminal configuration Nr root reception antenna;

G _MMSEBe the MMSE equalization filter, its expression formula is: G _MMSE=H ^H(HH ^H+ σ _n ²I _Nr) ^-1

H is a channel matrix, the element h of H _IjRepresent the channel coefficients of j transmit antennas, x=[x to i root reception antenna _k] ^T, k=1,2 ... Nt, expression transmission data, y=[y _k] ^T, k=1,2 ... Nr, expression receives data, n=[n _k] ^T, k=1,2 ... Nr, represent the noise of every reception antenna,

The MMSE that back detection signal-to-noise ratio computing module is used for calculating according to MMSE equalization filter module detects data

With formula w=1/ (1-diag (G _MMSEH))-1 calculate back detection signal-to-noise ratio w,

Order module is used for according to back detection signal-to-noise ratio w each data passages being sorted from big to small, obtains ordered series of numbers data m=[m ₁, m ₂... m _Nt], wherein

Choose the data channel module and be used to choose maximum back detection signal-to-noise ratio data channel m _iAs the current data channel of choosing, make i=1, export corresponding back detection signal-to-noise ratio of the current data channel of choosing and corresponding detection data;

Soft demodulation module is used for basis

Calculate the LLR of this data channel,

Represent the current m that chooses _iThe LLR value that data channel k bit is corresponding;

The LLR that data revaluation module is used for obtaining according to soft demodulation module carries out revaluation to the detection data of the current data channel of choosing, and obtains revaluation detection data

revaluation principle to do

$P\left(\mathrm{\θ}\right|{\hat{x}}_{m_i})=\underset{m=1}{\overset{M}{\mathrm{\Π}}}\frac{1}{2}(1+({2c}_{k,m}-1)\mathrm{Tanh}\left(\frac{L_k^{m_i}}{2}\right)),$ Wherein θ representes the constellation point value of modulator approach that this data channel adopts, c _{K, m}The corresponding m bit value (0/1) of expression constellation point value θ;

Interference cancellation module is used for eliminating the current interference of choosing data channel from receiving data, obtains the reception data model

after the interference eliminated

The residual interference estimation module is used for according to

the current residual interference energy value of choosing data channel after the calculating interference eliminated;

Class MMSE equalization filter module is utilized the reception data model after the interference eliminated that obtains in the interference cancellation module, calculates m _I+1Detection data on the circuit-switched data passage:

And output

Type MMSE equalization filter wherein $G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}})}^{-1},$

H _M=[h _k], k=1,2 ... Nt, k ≠ m ₁, m ₂... m _i, eliminate the data channel that had detected, the channel matrix of remaining data passage among the expression H;

The detection signal-to-noise ratio computing module is used for basis after type

Calculate m _I+1The back detection signal-to-noise ratio of circuit-switched data passage;

Judge module is used to judge whether current selected data passage is last data channel, if the time, detection of end then, if not, then need get back to soft demodulation module.

The advantage that the present invention had:

1, existing serial interference detection mode is often thought detected data perfect rejecting from receive data; Do not consider the influence of residual interference; The present invention has overcome prior art prejudice, and the influence and the corresponding processing method of residual interference have been carried out comprehensive consideration.

2, to the ordering of detection order, the present invention has adopted the back detection signal-to-noise ratio sort method of MMSE, and this method no longer need be calculated extra ordering parameter, has reduced operand, and obtains than other sort method more performance.

3, the present invention uses alternative manner calculated for subsequent class MMSE coefficient of equalizing wave filter to reduce operand, and has considered the influence that residual interference is brought in the alternative manner, and result of calculation is more accurate, thereby brings performance boost.

4, provide with the present invention in equalization methods corresponding accurately the back detection signal-to-noise ratio the complete computation method.

Description of drawings

Fig. 1 is system model figure of the present invention;

Fig. 2 is the inventive method flow chart;

Fig. 3 is a system construction drawing of the present invention.

Embodiment

A is balanced about least mean-square error

The model of multiple input, multiple output wireless communication system can be used model description as shown in Figure 1.At transmitter terminal configuration Nt transmit antennas, receiver end configuration Nr root reception antenna.Be formulated as:

y＝Hx+n (1)

Wherein H is a channel matrix, the element h of H _IjRepresent the channel coefficients of j transmit antennas to i root reception antenna.This paper uses h at the back _iThe i row of expression channel matrix H are promptly represented the channel coefficients vector of i transmit antennas to all reception antennas.

Acting as of linear equalization filter obtains sending the estimated value of data to receiving data through the linear equalization filter, can be formulated as:

$\hat{x}=\mathrm{Gy}$

The citation form of least mean-square error equalization filter is:

G _MMSE＝E{xy ^H}[E{yy ^H}] ^-1(2)

In the system model of describing like formula (1) of the present invention, the least mean-square error equalization filtering can be written as:

G _MMSE＝H ^H(HH ^H+σ _n ²I _Nr) ^-1(3)

After supposing that a circuit-switched data passage (i circuit-switched data passage) is eliminated in estimation from the primary reception signal, residue receives data and is expressed as:

Its effect is:

According to formula 2, can get rest layers least mean-square error equalization filter and be:

$G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}})}^{-1}$

H _M=[h _k], k=1,2 ... Nt, k ≠ m ₁, m ₂... m _iEliminate the data channel that had detected, the channel matrix of remaining data passage among the expression H.

B is about soft bit demodulation

The output LLR of soft demodulation module is defined as:

$L_k^{m_i}=\ln \mathrm{\Λ}=\ln \underset{\mathrm{\θ}\∈\{\mathrm{\Θ}:{\mathrm{\θ}}_k=1\}}{\mathrm{\Σ}}\exp (-\frac{{(x_{m_i}-\mathrm{\θ})}^2}{{2\mathrm{\σ}}^2})-\ln \underset{\mathrm{\θ}\∈\{\mathrm{\Θ}:{\mathrm{\θ}}_k=0\}}{\mathrm{\Σ}}\exp (-\frac{{(x_{m_i}-\mathrm{\θ})}^2}{{2\mathrm{\σ}}^2})$

Be that tentation data is made up of useful data and white Gaussian noise, after receiving data x, judge that x is that 1 probability and x are the natural logrithm value of the ratio of 0 probability.Adopt approximate linear approximation methods can be written as:

$L_k^{m_i}=\frac{1}{{\mathrm{\σ}}^2}[\underset{\mathrm{\θ}\∈\{\mathrm{\Θ}:{\mathrm{\θ}}_k=1\}}{\min }\left({|x_{m_i}-\mathrm{\θ}|}^2\right)-\underset{\mathrm{\θ}\∈\{\mathrm{\Θ}:{\mathrm{\θ}}_k=0\}}{\min }\left({|x_{m_i}-\mathrm{\θ}|}^2\right)]---\left(5\right)$

Wherein Θ representes the set of each constellation point on the planisphere.

expression back detection signal-to-noise ratio, the absolute value of the LLR of linear approximation is directly proportional with detection signal-to-noise ratio thereafter.

Method flow diagram embodiment below in conjunction with as shown in Figure 2 among the present invention is made detailed description to the present invention:

A) MMSE equalization filter module

This module is used for the MMSE that receives data is detected, and MMSE testing process described in MMSE filter such as the formula (3) is divided into following three steps:

Step 1: calculate a Nr * Nr inverse of a matrix, definition:

This inverse of a matrix also will use for the iteration of matrix inversion in the subsequent calculations class MMSE equalization filtering module as an output of MMSE equalization filtering module.

Step 2: calculate the MMSE equalization filter through matrix multiplication according to formula (3).

Step 3: the MMSE that obtains each circuit-switched data passage according to

receives the estimated value of data.

B) back detection signal-to-noise ratio computing module

This module functions is to calculate the back detection signal-to-noise ratio that detects each layer data of back through MMSE, and this signal to noise ratio can be used as the reliability of each layer data, is used for the reliability ordering to the transmission data, thus the order of specified data Interference Cancellation.

The back detection signal-to-noise ratio that detects each layer data of back through MMSE can calculate through formula (6):

w＝1/(1-diag(G _MMSEH))-1(6)

Wherein the diagonal entry of matrix is got in diag () expression.W will use at the soft timing of separating of constellation; W has represented the confidence level of each layer data simultaneously; Utilize vector w that data to be detected are sorted according to confidence level from big to small; The data high according to confidence level detect earlier, eliminate it can improve the counteracting serial interference method to the method for the interference of other layer data performance then.

C) order module

This module according to from big to small sequence arrangement, obtains the maximum circuit-switched data of w to the output w of back detection signal-to-noise ratio computing module, supposes it is i circuit-switched data passage.

D) choose the data channel module

This module is chosen

and is sent into soft demodulation module with w (i) according to the result that order module provides.And output

E) soft demodulation module

This module is done soft bit demodulation to the data based detection signal-to-noise ratio thereafter of sending into this module.The module effect can be represented by formula (7):

$L_k^{m_i}=w_{m_i}[\underset{\mathrm{\θ}\∈\{\mathrm{\Θ}:{\mathrm{\θ}}_k=1\}}{\min }\left({|x_{m_i}-\mathrm{\θ}|}^2\right)-\underset{\mathrm{\θ}\∈\{\mathrm{\Θ}:{\mathrm{\θ}}_k=0\}}{\min }\left({|x_{m_i}-\mathrm{\θ}|}^2\right)]---\left(7\right)$

Wherein

sends into the back detection signal-to-noise ratio value corresponding with data of this module.

F) data revaluation module

This module reappraises current circuit-switched data passage based on the principle (Soft Decision) of minimized average residual interference power:

Wherein

can do following derivation:

Assume that the first k? Bit sentenced to a probability

sentenced to 0 with probability and the first k? bit of LLR is defined as

$L_k^i=\ln \frac{{P_{1|x_r}}^k}{{P_{0|x_r}}^k}$

Have again, this bit judges that non-0 is 1, therefore has

$P_{1|x_r}^m+P_{0|x_r}^m=1$

Derivation can get:

$P_{1|x_r}^m=\frac{e^{L_k^{m_i}}}{e^{L_k^{m_i}}+1}=\frac{1}{2}(1+\frac{e^{L_k^{m_i}-1}}{e^{L_k^{m_i}}+1})=\frac{1}{2}(1+\tanh \left(\frac{L_k^{m_i}}{2}\right))$

$P_{0|x_r}^m=\frac{1}{e^{L_k^{m_i}}+1}=\frac{1}{2}(1-\frac{e^{L_k^{m_i}-1}}{e^{L_k^{m_i}}+1})=\frac{1}{2}(1-\tanh \left(\frac{L_k^{m_i}}{2}\right))$

Therefore, receive x _r, be judged to constellation point θ probability P (θ | x _r) be:

$P\left(\mathrm{\θ}\right|{\hat{x}}_{m_i})=\underset{m=1}{\overset{M}{\mathrm{\Π}}}\frac{1}{2}(1+(2c_{k,m}-1)\tanh \left(\frac{L_k^{m_i}}{2}\right))---\left(8\right)$

G) interference cancellation module

This module functions is from receive data, to eliminate the interference that i circuit-switched data passage is introduced other layer data, is formulated as:

Its effect is:

H) residual interference estimation module

The function of this module is after estimating that ' eliminations ' i road is disturbed, and estimates that the i circuit-switched data estimates the residual interference energy of all the other reception data.

Make

and be evaluated error, then

is the average energy of residual interference.Estimate the residual interference energy through formula (9).

Wherein

obtained by formula (8).

I) type MMSE equalization filter module

The function of this module is to after the interference of having eliminated part (a road or several roads layer emission data), utilizes the linear minimum mean-squared error equalization filter to carry out equalization filtering to the data of rest layers.The output of this module is the detection order that obtains according to order module, the estimated value of following one deck data to be checked.

Suppose and detected l circuit-switched data passage.Last time detected data i road, earlier need be according to the sequence detection j layer data of order module.Then a type MMSE detection filter device is:

${G^{\′}}_{\mathrm{MMSE}}=h_{m_{i+1}}^H{(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}})}^{-1}---\left(10\right)$

H wherein _M=[h _k], k=1,2 ... Nt, k ≠ m ₁, m ₂... m _i

Order

$A_0={\mathrm{HH}}^H+{{\mathrm{\σ}}_n}^2{I}_{\mathrm{Nr}}=\underset{k}{\mathrm{\Σ}}{h}_k{h}_k^H+{{\mathrm{\σ}}_n}^2{I}_{\mathrm{Nr}}$

$A_1=\underset{k\≠m_1}{\underset{k}{\mathrm{\Σ}}}{h}_k{h}_k^H+{{\mathrm{\σ}}_n}^2{I}_{\mathrm{Nr}}+(q_{m_1}-1)h_{m_1}{h}_{m_1}^H$

$=A_0+(q_{m_1}-1)h_{m_1}{h}_{m_1}^H$

$A_2=A_1+(q_{m_2}-1)h_{m_2}{h}_{m_2}^H$

...

Class MMSE equalization filter formula (10) is rewritten as formula (11)

$G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{A_i}^{-1}---\left(11\right)$

Can get according to matrix inversion lemma

$A_i^{-1}=A_{i-1}^{-1}-\frac{A_{i-1}^{-1}{h}_{m_i}{\left(A_{i-1}^{-1}{h}_{m_i}\right)}^H}{\frac{1}{q_{m_i}-1}+h_{m_i}^H\left(A_{i-1}^{-1}{h}_{m_i}\right)}---\left(\text{12}\right)$

The inverse matrix iteration that only needs to calculate the space of matrices opened through the vector

of matrix inversion last time and previous iteration just can obtain this layer MMSE equalization filtering filter need inverse matrix.

J) type back detection signal-to-noise ratio computing module

The function of this module is to calculate m _I+1The back detection signal-to-noise ratio of circuit-switched data passage,

Calculate through formula (13).

$w_{m_{i+1}}=1/G_{\mathrm{MMSE}}^{,}{h}_{m_i}-1---\left(13\right)$

K) judge module

Judge whether current selected data passage is last data channel, if the time, detection of end then, if not, then need get back to soft demodulation module.

The invention provides a system based on method, as shown in Figure 3,

Based on the counteracting serial interference detection system of many antennas bit interweaving encoding system, comprise the MMSE equalization filter module that connects successively, back detection signal-to-noise ratio computing module; Order module is chosen the data channel module, soft demodulation module; Data revaluation module, interference cancellation module, residual interference estimation module; Class MMSE equalization filter module, type back detection signal-to-noise ratio computing module and judge module, MMSE equalization filter module is used at the MIMO model according to the MMSE equalization filter

Calculate MMSE and detect data

Wherein: MIMO model: y=Hx+n, transmitting terminal configuration Nt transmit antennas, receiving terminal configuration Nr root reception antenna; G _MMSEBe the MMSE equalization filter, its expression formula is: $G_{\mathrm{MMSE}}=H^H{({\mathrm{HH}}^H+{{\mathrm{\σ}}_n}^2{I}_{\mathrm{Nr}})}^{-1}$

H is a channel matrix, the element h of H _IjRepresent the channel coefficients of j transmit antennas, x=[x to i root reception antenna _k] ^T, k=1,2 ... Nt, expression transmission data, y=[y _k] ^T, k=1,2 ... Nr, expression receives data, n=[n _k] ^T, k=1,2 ... Nr, represent the noise of every reception antenna,

The MMSE that back detection signal-to-noise ratio computing module is used for calculating according to MMSE equalization filter module detects data

With formula w=1/ (1-diag (G _MMSEH))-1 calculate back detection signal-to-noise ratio w,

Said order module is used for according to back detection signal-to-noise ratio w each data passages being sorted from big to small, obtains ordered series of numbers data m=[m ₁, m ₂... m _Nt], wherein

Choose the data channel module and be used to choose maximum back detection signal-to-noise ratio data channel m _iAs the current data channel of choosing, make i=1, export corresponding back detection signal-to-noise ratio of the current data channel of choosing and corresponding detection data;

Soft demodulation module is used for basis

Calculate the LLR of this data channel,

Represent the current m that chooses _iThe LLR value that data channel k bit is corresponding;

The LLR that data revaluation module is used for obtaining according to soft demodulation module carries out revaluation to the detection data of the current data channel of choosing, and obtains revaluation detection data

revaluation principle to do

$P\left(\mathrm{\θ}\right|{\hat{x}}_{m_i})=\underset{m=1}{\overset{M}{\mathrm{\Π}}}\frac{1}{2}(1+(2c_{k,m}-1)\mathrm{Tanh}\left(\frac{L_k^{m_i}}{2}\right)),$ Wherein θ representes the constellation point value of modulator approach that this data channel adopts, c _{K, m}The corresponding m bit value (0/1) of expression constellation point value θ;

Interference cancellation module is used for eliminating the current interference of choosing data channel from receiving data, obtains the reception data model

after the interference eliminated

The residual interference estimation module is used for according to

the current residual interference energy value of choosing data channel after the calculating interference eliminated;

Class MMSE equalization filter module is utilized the reception data model after the interference eliminated that obtains in the interference cancellation module, calculates m _I+1Detection data on the circuit-switched data passage:

And output

Type MMSE equalization filter wherein $G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}})}^{-1},$

H _M=[h _k], k=1,2 ... Nt, k ≠ m ₁, m ₂... m _i, eliminate the data channel that had detected, the channel matrix of remaining data passage among the expression H;

The detection signal-to-noise ratio computing module is used for basis after type

Calculate m _I+1The back detection signal-to-noise ratio of circuit-switched data passage;

Judge module is used to judge whether current selected data passage is last data channel, if the time, detection of end then, if not, then need get back to soft demodulation module.

Claims (3)

1. based on the counteracting serial interference detection method of many antennas bit interweaving encoding system, it is characterized in that: may further comprise the steps:

1] in the MIMO model, calculates MMSE and detect data

and back detection signal-to-noise ratio w according to MMSE equalization filter

Wherein: MIMO model: y=Hx+n, transmitting terminal configuration Nt transmit antennas, receiving terminal configuration Nr root reception antenna;

G _MMSEBe the MMSE equalization filter, its expression formula is: G _MMSE=H ^H(HH ^H+ σ _n ²I _Nr) ^-1

H is a channel matrix, the element h of H _IjRepresent the channel coefficients of j transmit antennas, x=[x to i root reception antenna _k] ^T, k=1,2 ... Nt, expression transmission data, y=[y _k] ^T, k=1,2 ... Nr, expression receives data, n=[n _k] ^T, k=1,2 ... Nr, represent the noise of every reception antenna,

Calculate back detection signal-to-noise ratio w:w=1/ (1-diag (G _MMSEH))-1;

2] sort from big to small according to back detection signal-to-noise ratio w, obtain ordered series of numbers m=[m ₁, m ₂... m _Nt], satisfy $w_{m_1}>w_{m_2}>...>w_{m_{\mathrm{Nt}}};$

3] choose maximum back detection signal-to-noise ratio data channel m _iAs the current data channel of choosing, make i=1, export corresponding back detection signal-to-noise ratio of the current data channel of choosing and corresponding detection data;

4] to step 3] in the current data channel of choosing carry out soft demodulation process, according to

Calculate the LLR of this data channel,

Represent the current m that chooses _iThe LLR value that data channel k bit is corresponding;

5] according to step 4] in the LLR that obtains the detection data of the current data channel of choosing are carried out revaluation, obtain revaluation and detect data

$P\left(\mathrm{\θ}\right|{\hat{x}}_{m_i})=\underset{m=1}{\overset{M}{\mathrm{\Π}}}\frac{1}{2}(1+(2c_{k,m}-1)\tanh \left(\frac{L_k^{m_i}}{2}\right))$

Wherein θ representes the constellation point value of modulator approach that this data channel adopts, c _{K, m}The corresponding m bit value (0/1) of expression constellation point value θ;

6] from receive data, eliminate the current interference of choosing data channel, obtain the reception data model

after the interference eliminated

7] the current residual interference energy value of choosing data channel after basis

the calculating interference eliminated;

8] utilize step 6] in reception data model after the interference eliminated that obtains, use a type MMSE equalization filter, calculate m _I+1Detection data on the circuit-switched data passage:

And output

Type MMSE equalization filter wherein $G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}})}^{-1},$

H _M=[h _k], k=1,2 ... Nt, k ≠ m ₁, m ₂... m _i, eliminate the data channel that had detected, the channel matrix of remaining data passage among the expression H;

9] basis

Calculate m _I+1The back detection signal-to-noise ratio of circuit-switched data passage;

10] judge, if i+1 is less than Nt, i++, execution in step 4]; Otherwise if i+1 equals Nt, each data channel all detects, then finishes.

2. the counteracting serial interference detection method based on many antennas bit interweaving encoding system according to claim 1 is characterized in that: step 8] middle G ' _MMSEAdopt following method to calculate:

Order $A_i=(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}}),$ Class MMSE equalization filter is written as following form:

$G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{A_i}^{-1}$

Wherein

obtains through following alternative manner:

$A_i^{-1}=A_{i-1}^{-1}-\frac{A_{i-1}^{-1}{h}_{m_i}{\left(A_{i-1}^{-1}{h}_{m_i}\right)}^H}{\frac{1}{q_{m_i}-1}+h_{m_i}^H\left(A_{i-1}^{-1}{h}_{m_i}\right)}$

$A_0^{-1}={({\mathrm{HH}}^H+{{\mathrm{\σ}}_n}^{2}I)}^{-1}.$

3. according to the counteracting serial interference detection system based on many antennas bit interweaving encoding system of the said method of claim 1, it is characterized in that: comprise the MMSE equalization filter module that connects successively, back detection signal-to-noise ratio computing module; Order module is chosen the data channel module, soft demodulation module; Data revaluation module, interference cancellation module, residual interference estimation module; Class MMSE equalization filter module, type back detection signal-to-noise ratio computing module and judge module

The MMSE equalization filter module according to the model used in a MIMO MMSE equalizer filter

calculate MMSE detector data

where: MIMO models:

Y=Hx+n, transmitting terminal configuration Nt transmit antennas, receiving terminal configuration Nr root reception antenna;

G _MMSEBe the MMSE equalization filter, its expression formula is: G _MMSE=H ^H(HH ^H+ σ _n ²I _Nr) ^-1

H is a channel matrix, the element h of H _IjRepresent the channel coefficients of j transmit antennas, x=[x to i root reception antenna _k] ^T, k=1,2 ... Nt, expression transmission data, y=[y _k] ^T, k=1,2 ... Nr, expression receives data, n=[n _k] ^T, k=1,2 ... Nr, represent the noise of every reception antenna,

The MMSE that said back detection signal-to-noise ratio computing module is used for calculating according to MMSE equalization filter module detects data

With formula w=1/ (1-diag (G _MMSEH))-1 calculate back detection signal-to-noise ratio w,

Said order module is used for according to back detection signal-to-noise ratio w each data passages being sorted from big to small, obtains ordered series of numbers data m=[m ₁, m ₂... m _Nt], wherein

The said data channel module of choosing is used to choose maximum back detection signal-to-noise ratio data channel m _iAs the current data channel of choosing, make i=1, export corresponding back detection signal-to-noise ratio of the current data channel of choosing and corresponding detection data;

Said soft demodulation module is used for basis

Calculate the LLR of this data channel,

Represent the current m that chooses _iThe LLR value that data channel k bit is corresponding;

The LLR that said data revaluation module is used for obtaining according to soft demodulation module carries out revaluation to the detection data of the current data channel of choosing, and obtains revaluation detection data

revaluation principle to do

$P\left(\mathrm{\θ}\right|{\hat{x}}_{m_i})=\underset{m=1}{\overset{M}{\mathrm{\Π}}}\frac{1}{2}(1+({2c}_{k,m}-1)\mathrm{Tanh}\left(\frac{L_k^{m_i}}{2}\right)),$ Wherein θ representes the constellation point value of modulator approach that this data channel adopts, c _{K, m}The corresponding m bit value (0/1) of expression constellation point value θ;

Said interference cancellation module is used for eliminating the current interference of choosing data channel from receiving data, obtains the reception data model

after the interference eliminated

Said residual interference estimation module is used for according to

the current residual interference energy value of choosing data channel after the calculating interference eliminated;

Said type of MMSE equalization filter module utilized the reception data model after the interference eliminated that obtains in the interference cancellation module, calculates m _I+1Detection data on the circuit-switched data passage:

And output

Type MMSE equalization filter wherein $G_{\mathrm{MMSE}}^{\′}=h_{m_{i+1}}^H{(H_M{H}_M^H+q_{m_i}{h}_{m_i}{h}_{m_i}^H+{\mathrm{\σ}}_n^2{I}_{\mathrm{Nr}})}^{-1},$

H _M=[h _k], k=1,2 ... Nt, k ≠ m ₁, m ₂... m _i, eliminate the data channel that had detected, the channel matrix of remaining data passage among the expression H;

Said type of back detection signal-to-noise ratio computing module is used for basis

Calculate m _I+1The back detection signal-to-noise ratio of circuit-switched data passage;

Said judge module is used to judge whether current selected data passage is last data channel, if the time, detection of end then, if not, then need get back to soft demodulation module.

Images