CN101277279B - Method and apparatus for eliminating serial interference of multi-aerial system - Google Patents

Abstract

A method for removing serial interference of a multi-antenna system and a device for realizing the method are disclosed, the system comprises N<t> sending antennas, the method comprises the following steps of: (a) judging quality of signals of each path of data stream in receiving signals; (b) performing minimum mean square error MMSE detection with a path of data flow having the optimal quality of signals in order to detect a path of data stream; (C) soft-demodulating the detection result to acquire soft-bit information, judging whether the N<t> paths of data streams are detected, if so, ending; if not, performing the next step; (d) according to the soft-bit information acquired in the step (c), obtaining possibility of the detection result mapping on each constellation point of modulation constellation map, and multiplying the possibility by the sum of values of the corresponding constellation points to attain the soft-symbol information, making use of the soft-symbol information to intervene for rebirth, canceling, judging the quality of signals of each path of data stream in the cancelled signals, and entering into the step (B).

Description

A kind of method for eliminating serial interference of multiaerial system and device thereof

Technical field

The present invention relates to successive interference cancellation techniques, relate in particular to the method and the device thereof of the realization serial interference elimination in list/many code words multi-input multi-output system.

Background technology

In Long Term Evolution (LTE) downlink system, for multiple-input and multiple-output (MCW-MIMO) system (as shown in Figure 1) of many code words, general recommendations adopts interference cancellation receiver to improve the detection performance.As shown in Figure 2, detect the channel condition signal of data flow preferably earlier, it is carried out demodulation, deciphers the information bit that obtains this stream at receiving terminal.It is encoded, modulates, experiences these streams of operation recovery such as channel again, and it is eliminated from received signal, iteration is carried out serial interference elimination, to improve the link performance of system.

For multiple-input and multiple-output (SCW-MIMO) structure (as shown in Figure 3) of single codeword, the data of transmitting terminal are dispensed on each antenna and transmit through after encoding.Generally adopt linearity test (as shown in Figure 4) at receiving terminal, after promptly the data of a plurality of reception antennas being detected according to least mean-square error (MMSE) criterion, after each flow point do not carried out operation such as demodulation, each soft bit stream is carried out sending into decoder after the parallel/serial conversion obtain transmission information.

In the SCW-MIMO system, because the data message of each antenna of transmitting terminal belongs to an encoding block, like this at receiving terminal, can't also just can not use the serial interference elimination receiver of MCW-MIMO according to live again the separately interference signal in certain tributary of traditional decoding re-encoding method.This just requires, and for MCW and two kinds of launch scenarios of SCW, needs two kinds of different receiver operations.

Summary of the invention

The technical problem to be solved in the present invention is to provide a kind of serial interference elimination algorithm and corresponding device thereof that the MCW-MIMO system also can be used for the SCW-MIMO system that both can be used for.

In order to solve the problems of the technologies described above, the present invention proposes a kind of method for eliminating serial interference of multiaerial system, and this system comprises N _tThe root transmitting antenna comprises following steps:

(a) judge the signal quality that each circuit-switched data flows in the received signal;

(b) the top-quality circuit-switched data stream of the number of winning the confidence carries out least mean-square error MMSE detection, detects circuit-switched data stream;

(c) testing result in the step (b) is carried out soft demodulation, obtain soft bit information, judge whether to detect the N of all transmitting antenna correspondences _tCircuit-switched data stream is if finish; Otherwise, change next step over to;

(d) soft bit information that obtains in the step (c) is carried out soft-decision, obtain the probability that described testing result is mapped in each constellation point of modulation constellation according to getting soft bit information in the step (c), the value summation of multiply by constellation point with this probability obtains soft symbol information, utilizing soft symbol information to disturb lives again, disappear mutually, judge the signal quality of each circuit-switched data stream in the signal after disappearing mutually, change step (b) over to.

Further, said method also can have following characteristics: after detecting all data flow, all soft bit informations of output in the step (c) are deciphered obtaining original transmission bit.

Further, said method also can have following characteristics: in the described step (b), after detecting circuit-switched data stream, this has been detected the channel matrix row zero setting of data flow correspondence.

Further, said method also can have following characteristics: signal quality is judged by the reprocessing signal to noise ratio SINR that calculates each circuit-switched data stream in the described step (a), and the maximum data flow of SINR value is the best data flow of signal quality in the described step (b).

Further, said method also can have following characteristics: the i time detected k in the described step (d) _iThe soft symbol information of root transmitting antenna obtains by following formula,

Wherein Be soft symbol information,

The value of expression constellation point,

Be the value of each constellation point, G is the number of constellation point, Expression MMSE testing result Z ^lBe mapped in the probability of each constellation point, at first obtain by soft demodulation in the step (c)

Calculate again

Promptly

λ bit equal the probability of m, wherein Obtain

Represent k _iλ bit of bar data flow.

Disturb the signal Synchronization of living again, the delay submodule of second branch road links to each other with received signal;

The submodule that disappears mutually and postpones the submodule and the interference of the last branch road submodule of living again and links to each other, eliminates the signal that interference is lived again the signal after postponing.

Further, said apparatus also can have following characteristics: described device further comprises parallel serial conversion submodule and decoding submodule, wherein, the parallel serial conversion submodule links to each other with all soft demodulation submodules, is used for the soft bit information of soft demodulation submodule output is carried out parallel serial conversion; The decoding submodule links to each other with the parallel serial conversion submodule, is used for the parallel serial conversion submodule [0016]The ordering submodule is used to judge signal quality, by signal quality data is flow to line ordering; The signal that is input as the receiver reception of the ordering submodule of first branch road, the ordering submodule links to each other with the submodule that disappears mutually in other branch roads;

Least mean-square error MMSE submodule links to each other with the ordering submodule, is used for the top-quality circuit-switched data stream of detection signal, to N _tBranch road, MMSE submodule and this branch road submodule that disappears mutually links to each other;

Soft demodulation submodule links to each other with decoding module with the MMSE submodule, and the signal that is used for the MMSE submodule is detected carries out soft demodulation, the output soft bit information;

The soft-decision submodule, link to each other with soft demodulation submodule, be used for the soft bit information of soft demodulation submodule output is carried out soft-decision, handle the probability that testing result that this soft bit information obtains the MMSE submodule is mapped in each constellation point of modulation constellation, the value summation of multiply by constellation point with this probability obtains soft symbol information;

The interference submodule of living again links to each other with the soft-decision submodule, is used for the soft symbol information of soft-decision submodule output disturbed living again;

Postpone submodule, link to each other with the submodule that disappears mutually of a last branch road, receive the signal after a last branch road disappears mutually, be used for this signal is postponed, a signal and a last branch road after guaranteeing to postpone disturb the signal Synchronization of living again, and the delay submodule of second branch road links to each other with received signal;

The submodule that disappears mutually and postpones the submodule and the interference of the last branch road submodule of living again and links to each other, eliminates the signal that interference is lived again the signal after postponing.

Further, said apparatus also can have following characteristics: described device further comprises parallel serial conversion submodule and decoding submodule, wherein, the parallel serial conversion submodule links to each other with all soft demodulation submodules, is used for the soft bit information of soft demodulation submodule output is carried out parallel serial conversion; The decoding submodule links to each other with the parallel serial conversion submodule, is used for the output signal of parallel serial conversion submodule is deciphered.

Said apparatus comprises N _tThe road decoding module is with N _tIndividual soft demodulation submodule is corresponding one by one to link to each other, and is used for this N _tThe soft bit information of individual soft demodulation submodule output is deciphered.

Further, said apparatus also can have following characteristics: described soft-decision submodule obtains soft symbol information by said method, at first obtains soft bit information from soft demodulation submodule Calculate

Wherein

Obtain MMSE testing result Z ^lBe mapped in the probability of each constellation point

Calculate soft symbol information at last

Be the value of each constellation point, G is the number of constellation point, i=1～N _t, N _tBe system's number of transmit antennas, k _iBe illustrated in the index of data flow detected in the i time iteration, λ is the integer more than or equal to 1, Represent k _iλ bit of bar data flow,

Be λ bit equal the probability of m.

Further, said apparatus also can have following characteristics: described ordering submodule calculates the reprocessing signal to noise ratio SINR of each data flow of input, by SINR each data flow of input is sorted, and described MMSE submodule detects the circuit-switched data stream of SINR maximum.

In sum, the soft decision information that the present invention obtains by the utilization of SD-SIC module disturbs lives again, replace traditional information flow being carried out the regeneration of recompile modulation carrying out symbol, so not only can reduce the complexity of interference cancellation receiver but also reduce processing delay.This receiver can be used under SCW and the MCW emission mode simultaneously.Under the MCW emission mode, performance loss is arranged slightly with respect to traditional interference cancellation receiver.But in the system of reality, because the diversity gain that system obtains by alternate manner can be very big, in the prerequisite of the high slope of curve, it is very little that the distance between SD-SIC and the traditional interference cancellation algorithm curve can become.

Description of drawings

Fig. 1 is existing MCW-MIMO transmitter (N _t=2);

Fig. 2 is existing MCW-MIMO tradition interference cancellation receiver (N _t=2);

Fig. 3 is existing SCW-MIMO transmitter (N _t=2);

Fig. 4 is existing SCW-MIMO linear receiver (N _t=2)

Fig. 5 is the MMSE-SE SIC receiver that the present invention is used for the MCW/SCW-MIMO system;

Fig. 6 is an embodiment of the invention QPSK planisphere;

Fig. 7 is that the performance of existing SCW-MIMO system linear receiver and SD-SIC receiver of the present invention compares;

Fig. 8 is that the performance of existing MCW-MIMO system linear receiver, traditional interference cancellation receiver and SD-SIC receiver of the present invention compares.。

Embodiment

Below in conjunction with the drawings and specific embodiments technical scheme of the present invention is described in detail.

Introduce the concrete operations of the linear receiver of SCW-MIMO at first in detail:

Suppose that has a N _tRoot is sent out antenna and N _rRoot is received the SCW-MIMO system of antenna.The source bit stream passes through N through coding _tRoot is sent out antenna multiplexed, flow through ovennodulation, emission of coded data then.System model can be expressed as:

$Y=H\&CenterDot;X+n=\underset{i}{\mathrm{\&Sigma;}}{H}_i\&CenterDot;X_i+n---\left(1\right)$

$X={[X_1,\&CenterDot;\&CenterDot;\&CenterDot;X_{N_t}]}^T\&Element;C^{N_t},$ $Y={[Y_1,\&CenterDot;\&CenterDot;\&CenterDot;Y_{N_r}]}^T\&Element;C^{N_r}$ Be respectively that recurrence is penetrated and the receiving symbol vector. $n={[n_1,\&CenterDot;\&CenterDot;\&CenterDot;n_{N_r}]}^T\&Element;C^{N_r}$ Be that a covariance is $R_{\mathrm{nn}}={\mathrm{\&sigma;}}_n^{2}I$ Being added with property of independent same distribution (i.i.d) Gaussian noise vector. $H\&Element;C^{N_r\&times;N_t}$ Be the complex channel transfer matrix of a Rayleigh fading,

Be to receive antenna is sent out antenna to the j root channel coefficients from the i root.Every channel path supposing to send out antenna and receive between the antenna is independently, E (| h _{J, i}| ²)=1.

The MMSE detection module can be represented as

$W^T={(H^{H}H+{\mathrm{\&sigma;}}_n^{2}I)}^{-1}{H}^H---\left(2\right)$

ρ _kFor MMSE detects the SINR (post-SINR, reprocessing signal to noise ratio) of k bar data flow afterwards:

${\mathrm{\&rho;}}_k=\frac{1}{{\mathrm{\&sigma;}}_n^2{\left[{(H^{H}H+{\mathrm{\&sigma;}}_n^{2}I)}^{-1}\right]}_{k,k}}-1---\left(3\right)$

K=1 wherein ..., N _t

We obtain by the MMSE detection module

$Z_k=B_k{W}_k^T\&CenterDot;Y=B_k{W}_k^T{H}_k{X}_k+B_k{W}_k^T\&CenterDot;\underset{i\&NotEqual;k}{\mathrm{\&Sigma;}}{H}_i\&CenterDot;X_i+B_k{W}_k^T\&CenterDot;n---\left(4\right)$

Wherein $B_k=\frac{1}{\sqrt{{\left|\right|W_k^T{H}_k\left|\right|}^2}}$ It is normalization coefficient.

(4) formula first is the symbol X that hope obtains _i, second portion is the interference that comes from other circuit-switched data streams after detecting, last part is the noise after detecting.

Fig. 5 is one of the present invention based on disappear the mutually MMSE receiver apparatus schematic diagram of (SD-SIC) of soft-decision.This receiver comprises SD-SIC module and decoding module, and supposing the system has N _tRoot is sent out antenna, so,

The SD-SIC module has N _tRoad branch road, first branch road comprise the ordering submodule that links to each other successively, the MMSE submodule, and soft demodulation submodule, the soft-decision submodule disturbs the submodule of living again;

Middle N _t-2 branch roads (are worked as N _t2 o'clock), each branch road comprises the delay submodule that links to each other successively, the submodule that disappears mutually, ordering submodule, MMSE submodule, soft demodulation submodule, soft-decision submodule and the interference submodule of living again;

N _tBranch road comprises the delay submodule that links to each other successively, the submodule that disappears mutually, ordering submodule (optional), MMSE detection sub-module and soft demodulation submodule;

To all branch roads, the connection in the following explanation is except that particularly pointing out with all the other branch roads link to each other, all refer to and being connected of other submodules of same branch road, wherein:

The ordering submodule is used for to the received signal, calculates the reprocessing signal to noise ratio SINR of each circuit-switched data stream, sorts; The signal that is input as the receiver reception of the ordering submodule of first branch road, the ordering submodule links to each other with the submodule that disappears mutually in other branch roads;

Wherein, the ordering submodule also can calculate the parameter that other can represent signal quality, is used for ordering, is not limited to and calculates the reprocessing signal to noise ratio;

The MMSE submodule links to each other with the ordering submodule, is used for selecting the data flow (being the best data flow of signal quality) of SINR maximum to detect to the data flow of ordering submodule output; To N _tBranch road, when this branch road does not comprise ordering during submodule, MMSE submodule and this branch road submodule that disappears mutually links to each other;

Soft demodulation submodule links to each other with decoding module with the MMSE submodule, and the signal that is used for the MMSE submodule is detected carries out soft demodulation, the output soft bit information;

The soft-decision submodule links to each other with soft demodulation submodule, is used for the soft bit information of soft demodulation submodule output is carried out soft-decision, obtains soft symbol information;

The interference submodule of living again links to each other with the soft-decision submodule, is used for the soft symbol information of soft-decision submodule output disturbed living again;

Postpone submodule, be used for signal is postponed, a signal and a last branch road after guaranteeing to postpone disturb the signal Synchronization of living again; Link to each other with the submodule that disappears mutually of a last branch road, receive the signal after a last branch road disappears mutually, to second branch road, the submodule because first branch road does not disappear mutually, the delay submodule of second branch road links to each other with received signal;

The submodule that disappears mutually and postpones the submodule and the interference of the last branch road submodule of living again and links to each other, eliminates the signal that interference is lived again the signal after postponing.

Decoding module if be used for the SCW system, comprises the parallel serial conversion submodule, the decoding submodule, and wherein the parallel serial conversion submodule links to each other with all soft demodulation submodules, is used for the soft bit information of soft demodulation submodule output is carried out parallel serial conversion; The decoding submodule links to each other with the parallel serial conversion submodule, is used for the output signal of parallel serial conversion submodule is deciphered;

If be used for the MCW system, corresponding to the multiple branch circuit output of SD-SIC module, comprise multiple branch circuit decoding submodule, the decoding submodule is corresponding one by one continuous with the soft demodulation submodule of SD-SIC module, is used for the soft bit information of soft demodulation submodule output is deciphered.

The interference elimination method that disappears mutually based on soft-decision that the present invention proposes is the SINR of first calculated data stream and it is sorted, and selects the wherein data flow (k of maximum SINR then _iBar) carry out MMSE and detect, obtain testing result, this result is carried out soft demodulation, soft-decision (obtaining soft symbol information), the soft-decision that obtains (soft symbol information) disturbs and lives again.Begin next iteration then, till all transmission data flow all were detected, concrete steps were as follows:

Step 610: calculate the reprocessing signal to noise ratio SINR of each circuit-switched data stream in the received signal, beginning is iteration for the first time;

Step 620: the data flow of getting the SINR maximum is carried out least mean-square error MMSE detection, detects circuit-switched data stream, and these channel matrix H row that detected the data flow correspondence are put 0;

The present invention is not limited to and calculates SINR, if the parameter of other gauge signal quality is arranged, as based on the MMSE parameter, also can be used for input, as long as the best circuit-switched data stream of signal quality is detected.

Step 630: the testing result in the step 620 is carried out soft demodulation, obtain soft bit information, judge whether to detect the N of all transmitting antenna correspondences _tCircuit-switched data stream, if, change step 650 over to, otherwise, change step 640 over to;

Step 640: the soft bit information that obtains in the step 630 is carried out soft-decision, obtain soft symbol information, utilize soft symbol information to disturb and live again, disappear mutually, calculate the SINR of each circuit-switched data stream in the signal after disappearing mutually, carry out next iteration, change step 620 over to;

Being implemented as follows of soft-decision, wherein k _i(i=1～N _t) be illustrated in the index of data flow detected in the i time iteration, ρ _kBe SINR.

At first, for serial interference (SIC) receiver that disappears mutually, the l time iteration can be reduced following process:

$Y^l=Y-\underset{i=k_{l-1}}{\overset{k_1}{\mathrm{\&Sigma;}}}{\tilde{X}}_i\&CenterDot;H_i$

$=X_{k_l}\&CenterDot;H_{k_l}+\underset{i\&NotEqual;k_1,\&CenterDot;\&CenterDot;\&CenterDot;,k_l}{\mathrm{\&Sigma;}}{X}_i\&CenterDot;H_i+\underset{i=k_{l-1}}{\overset{k_1}{\mathrm{\&Sigma;}}}(X_i-{\tilde{X}}_i)\&CenterDot;H_i+n---\left(5\right)$

Represent k _i(1＜i＜l) judgement of transmit antennas symbol, it estimates in the i time iteration.Then with the k of H _iRow zero setting obtains new channel matrix , to k _lIndividual data stream detects:

$Z^l=B_{k_l}{{\left(W^l\right)}_{k_l}}^T\&CenterDot;Y^l$

$=B_{k_l}{{\left(W^l\right)}_{k_l}}^T\&CenterDot;[X_{k_l}\&CenterDot;H_{k_l}+\underset{i\&NotEqual;k_1,\&CenterDot;\&CenterDot;\&CenterDot;{,k}_l}{\mathrm{\&Sigma;}}{X}_i\&CenterDot;H_i+n]+B_{k_l}{{\left(W^l\right)}_{k_l}}^T\&CenterDot;\underset{i=1}{\overset{l-1}{\mathrm{\&Sigma;}}}(X_{k_i}-{\tilde{X}}_{k_i})\&CenterDot;H_{k_i}---\left(6\right)$

Be (W ^l) ^TK _lRow.

Soft-decision is handled: in order to above-mentioned Z ^lConditional expectation for condition Estimate

The present invention obtains Z by soft demodulation result is handled ^lBe mapped in the probability of each constellation point of modulation constellation, multiply by the value of constellation point with this probability after summation obtain soft symbol information.Usually, soft-decision is obtained by following formula:

${\tilde{X}}_{k_i}=E\left(X_{k_i}\right|Z^l)$

$=\underset{g=1}{\overset{G}{\mathrm{\&Sigma;}}}{X}_{k_i}^{g}p\left(X_{k_i}^g\right|Z^l)$

Wherein,

The value of representing each constellation point, G is the number of constellation point in the modulation constellation of received signal,

Expression MMSE testing result Z ^lBeing mapped in the probability of each constellation point, is following formula (9) by soft demodulation result, calculates through following formula (8), passes through again $p\left(X_{k_i}^g\right|Z^l)=\underset{\mathrm{\&lambda;}=1,m\&Element;\left(\mathrm{0,1}\right)}{\overset{{\log }_2\left(G\right)}{\mathrm{\&Pi;}}}P{r}_m\left(b_{k_i}^{\mathrm{\&lambda;}}\right)$ Calculate.

With quarternary phase-shift keying (QPSK) QPSK modulation (Fig. 6) is example:

At this moment, G=4, $X_{k_i}^g\&Element;(\frac{\sqrt{2}}{2}+\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07187587420070406E000032.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{\sqrt{2}}{2},\frac{\sqrt{2}}{2}-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07187587420070406E000032.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{\sqrt{2}}{2},-\frac{\sqrt{2}}{2}+\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07187587420070406E000032.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{\sqrt{2}}{2},-\frac{\sqrt{2}}{2}-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="S07187587420070406E000032.png"\; state="\{\{state\}\}">j</figure-callout>}\frac{\sqrt{2}}{2}),$

Expression

The probability of four values in the difference codomain.

${\tilde{X}}_{k_i}=E\left(X_{k_i}\right|Z^l)$

$=[\frac{1}{\sqrt{2}}\&CenterDot;P{r}_1\left(b_{k_i}^1\right)-\frac{1}{\sqrt{2}}\&CenterDot;P{r}_0\left(b_{k_i}^1\right)]+[\frac{1}{\sqrt{2}}\&CenterDot;P{r}_1\left(b_{k_i}^2\right)-\frac{1}{\sqrt{2}}\&CenterDot;P{r}_0\left(b_{k_i}^2\right)]\&CenterDot;j---\left(7\right)$

$=\frac{1}{\sqrt{2}}[P{r}_1\left(b_{k_i}^1\right)-P{r}_0\left(b_{k_i}^1\right)]+\frac{1}{\sqrt{2}}[P{r}_1\left(b_{k_i}^2\right)-P{r}_0\left(b_{k_i}^2\right)]\&CenterDot;j$

Be

λ bit equal the probability of m.This probability tables is shown:

$\left\{\begin{array}{c}P{r}_1\left(b_{k_i}^{\mathrm{\&lambda;}}\right)=\frac{\exp \left(L\left(b_{k_i}^{\mathrm{\&lambda;}}\right|Z^l)\right)}{\exp \left(L\left(b_{k_i}^{\mathrm{\&lambda;}}\right|Z^l)\right)+1}\\ P{r}_0\left(b_{k_i}^{\mathrm{\&lambda;}}\right)=\frac{1}{\exp \left(L\left(b_{k_i}^{\mathrm{\&lambda;}}\right|Z^l)\right)+1}\end{array}\right.---\left(8\right)$

Wherein, max log posterior probability (Max-log-MAP) method is calculated the log-likelihood (LLR) of each bit, is expressed as:

$L\left(b_{k_i}^{\mathrm{\&lambda;}}\right|Z^l)={\mathrm{\&rho;}}_i\&times;(\underset{\tilde{a}\&Element;A_{\mathrm{\&lambda;}}^0}{\min }{\left|\right|Z^l-\tilde{a}\left|\right|}^2-\underset{\tilde{a}\&Element;A_{\mathrm{\&lambda;}}^1}{\min }{\left|\right|Z^l-\tilde{a}\left|\right|}^2)---\left(9\right)$

$A_{\mathrm{\&lambda;}}^m=\{\mathrm{\&alpha;}:b_{k_l}^{\mathrm{\&lambda;}}=m\}$ Be the subclass of modulation symbol set { α }, its k _iλ bit of bar data flow equals m.

(8), (9) substitution (7) are got

${\tilde{X}}_{k_i}=\frac{1}{\sqrt{2}}\&CenterDot;\frac{\exp \left(L\left(b_{k_i}^1\right|Z^l)\right)-1}{\exp \left(L\left(b_{k_i}^1\right|Z^l)\right)+1}+\frac{1}{\sqrt{2}}\&CenterDot;\frac{\exp \left(L\left({\mathrm{<figure-callout\; id="2"\; label="b\; k\; i"\; filenames="S07187587420070406E000032.png"\; state="\{\{state\}\}">b</figure-callout>}}_{k_i}^{{}_{}}\right|Z^l)\right)-1}{\exp \left(L\left({\mathrm{<figure-callout\; id="2"\; label="b\; k\; i"\; filenames="S07187587420070406E000032.png"\; state="\{\{state\}\}">b</figure-callout>}}_{k_i}^{{}_{}}\right|Z^l)\right)+1}\&CenterDot;j$

(10)

$=\frac{1}{\sqrt{2}}\&CenterDot;\tanh \left[\frac{L\left(b_{k_i}^1\right|Z^l)}{2}\right]+\frac{1}{\sqrt{2}}\&CenterDot;\tanh \left[\frac{L\left({\mathrm{<figure-callout\; id="2"\; label="b\; k\; i"\; filenames="S07187587420070406E000032.png"\; state="\{\{state\}\}">b</figure-callout>}}_{k_i}^{{}_{}}\right|Z^l)}{2}\right]\&CenterDot;j$

The value that soft-decision SIC promptly adopts formula (10) to estimate according to signal expectation is carried out the interference eliminated between antenna, carries out MMSE then and detects.

650: pass through soft bit information that soft demodulation operation obtains and decipher and obtain original transmission bit to all.

At last, the sign N that will obtain by the SD-SIC module _tThe LLR value of the binary bits data flow of individual antenna is sent into the channel decoding module, and as shown in Figure 5, for the SCW launch scenario, these data flow through parallel/serial conversion, are sent into decoder earlier then; For MCW, each data flow is sent into own corresponding decoder respectively.

SD-SIC of the present invention disturbs the soft symbol information that obtains according to soft-decision of living again to carry out, and is with the different of traditional SIC, and traditional SIC adopts and detects the signal demodulation, the decoding that obtain and encode, modulate and obtain symbolic information, disturbs then and lives again.

Below by emulation the advantage of SD-SIC receiver with respect to linear receiver is described.Suppose that flat Rayleigh fading channel channel in a frame is constant.Antenna configurations is 4 * 4, and each stream all adopts the QPSK modulation, adopts the 1/2turbo coding of MAX-log-MAP decoding.

Fig. 7 is that the performance of SCW-MIMO system linear receiver and SD-SIC receiver compares, and as can be seen from Figure 7, for the SCW-MIMO system, the interference eliminated of soft-decision (SD-SIC) is at 1/2 code check BER=10 ^-3The place compares the gain that linear MMSE detects to be had above 1dB.When code check was 3/4, this gain can surpass 3dB.Owing to carried out interference eliminated, for back detected data flow, improved diversity gain, so the slope of curve of SD-SIC will be obviously greater than linearity test.

As can be seen from Figure 8, in the MCW-MIMO system, the performance of SD-SIC is better than the performance of linearity test equally greatly, but for traditional interference cancellation algorithm, performance can be poor slightly.This gap is probably about 0.5dB.Mainly be because in traditional interference cancellation algorithm interference signal want decoded after more again decoding live again, decipher the accuracy that this module can improve the signal of living again.But this process has increased the time delay of interference eliminated, and has increased the complexity of receiver.

In sum, SD-SIC utilizes the soft decision information obtain to disturb and lives again, and replaces traditional information flow is carried out recompile modulation carrying out symbol regenerate, and so not only can reduce the complexity of interference cancellation receiver but also reduce processing delay.This receiver also can be used under SCW and the MCW emission mode simultaneously.Under the MCW emission mode, performance loss is arranged slightly with respect to traditional interference cancellation receiver.But in the system of reality, because the diversity gain that system obtains by alternate manner can be very big, in the prerequisite of the high slope of curve, it is very little that the distance between SD-SIC and the traditional interference cancellation algorithm curve can become.

Claims (9)

1. the method for eliminating serial interference of a multiaerial system, this system comprises N _tThe root transmitting antenna comprises following steps:

(a) judge the signal quality that each circuit-switched data flows in the received signal;

(b) the top-quality circuit-switched data stream of the number of winning the confidence carries out least mean-square error MMSE detection, detects circuit-switched data stream;

(c) testing result in the step (b) is carried out soft demodulation, obtain soft bit information, judge whether to detect the N of all transmitting antenna correspondences _tCircuit-switched data stream is if finish; Otherwise, change next step over to;

(d) soft bit information that obtains in the step (c) is carried out soft-decision, obtain the probability that described testing result is mapped in each constellation point of modulation constellation according to the soft bit information in the step (c), the value summation of multiply by constellation point with this probability obtains soft symbol information, utilizing soft symbol information to disturb lives again, disappear mutually, judge the signal quality of each circuit-switched data stream in the signal after disappearing mutually, change step (b) over to.

2. the method for claim 1 is characterized in that: after detecting all data flow, all soft bit informations of output in the step (c) are deciphered obtaining original transmission bit.

3. the method for claim 1 is characterized in that: in the described step (b), after detecting circuit-switched data stream, this has been detected the channel matrix row zero setting of data flow correspondence.

4. method as claimed in claim 1 or 2, it is characterized in that: signal quality is judged by the reprocessing signal to noise ratio SINR that calculates each circuit-switched data stream in the described step (a), and the maximum data flow of SINR value is the best data flow of signal quality in the described step (b).

5. the method for claim 1, it is characterized in that: the soft symbol information of the i time detected ki root transmitting antenna obtains by following formula in the described step (d),

Wherein

Be soft symbol information,

The value of expression constellation point,

Be the value of each constellation point, G is the number of constellation point, Expression MMSE testing result Z ^lBe mapped in the probability of each constellation point, at first obtain by soft demodulation in the step (c)

Calculate again

Promptly

λ bit equal the probability of m, wherein

Obtain

Represent k _iλ bit of bar data flow.

6. the serial interference elimination device of a multiaerial system, this system comprises N _tThe root transmitting antenna, described device comprises N _tThe road branch road, wherein first branch road comprises the ordering submodule that links to each other successively, least mean-square error MMSE submodule, soft demodulation submodule, the soft-decision submodule disturbs the submodule of living again; N _tBranch road comprises the delay submodule that links to each other successively, the submodule that disappears mutually, MMSE submodule, soft demodulation submodule; All the other branch roads comprise successively the delay submodule that links to each other, the submodule that disappears mutually, ordering submodule, MMSE submodule, soft demodulation submodule, soft-decision submodule and the interference submodule of living again; Described device comprises N _tThe road decoding module is with N _tIndividual soft demodulation submodule is corresponding one by one to link to each other, and is used for this N _tThe soft bit information of individual soft demodulation submodule output is deciphered, wherein:

The ordering submodule is used to judge signal quality, by signal quality data is flow to line ordering; The signal that is input as the receiver reception of the ordering submodule of first branch road, the ordering submodule links to each other with the submodule that disappears mutually in other branch roads;

Least mean-square error MMSE submodule links to each other with the ordering submodule, is used for the top-quality circuit-switched data of detection signal stream, and to the Nt branch road, MMSE submodule and this branch road submodule that disappears mutually is continuous;

Soft demodulation submodule links to each other with decoding module with the MMSE submodule, and the signal that is used for the MMSE submodule is detected carries out soft demodulation, the output soft bit information;

The soft-decision submodule, link to each other with soft demodulation submodule, be used for the soft bit information of soft demodulation submodule output is carried out soft-decision, handle the probability that testing result that this soft bit information obtains the MMSE submodule is mapped in each constellation point of modulation constellation, the value summation of multiply by constellation point with this probability obtains soft symbol information;

The interference submodule of living again links to each other with the soft-decision submodule, is used for the soft symbol information of soft-decision submodule output disturbed living again;

Postpone submodule, link to each other with the submodule that disappears mutually of a last branch road, receive the signal after a last branch road disappears mutually, be used for this signal is postponed, a signal and a last branch road after guaranteeing to postpone disturb the signal Synchronization of living again, and the delay submodule of second branch road links to each other with received signal;

The submodule that disappears mutually and postpones the submodule and the interference of the last branch road submodule of living again and links to each other, eliminates the signal that interference is lived again the signal after postponing.

7. device as claimed in claim 6, it is characterized in that: described device further comprises parallel serial conversion submodule and decoding submodule, wherein, the parallel serial conversion submodule links to each other with all soft demodulation submodules, is used for the soft bit information of soft demodulation submodule output is carried out parallel serial conversion; The decoding submodule links to each other with the parallel serial conversion submodule, is used for the output signal of parallel serial conversion submodule is deciphered.

8. device as claimed in claim 6 is characterized in that: described soft-decision submodule obtains soft symbol information by said method, at first obtains soft bit information from soft demodulation submodule

Calculate Wherein

Obtain MMSE testing result Z ^lBe mapped in the probability of each constellation point Calculate soft symbol information at last

Be the value of each constellation point, G is the number of constellation point, i=1～N _t, N _tBe system's number of transmit antennas, k _iBe illustrated in the index of data flow detected in the i time iteration, λ is the integer more than or equal to 1,

Represent k _iλ bit of bar data flow,

Be λ bit equal the probability of m.

9. device as claimed in claim 6 is characterized in that: described ordering submodule calculates the reprocessing signal to noise ratio SINR of each data flow of input, by SINR each data flow of input is sorted, and described MMSE submodule detects the circuit-switched data stream of SINR maximum.

Images