CN102324960A - Interference suppression merging method and receiver - Google Patents

Abstract

The invention discloses an interference suppression merging method and a receiver. The method is applied in a multi-user system, and comprises the steps of: obtaining the signal channel state information of local users; acquiring the quantized precoding vectors of the local users and interference users; and forming a linear weight vector gi according to the signal channel state information, the quantized precoding vectors of the local users and the interference users. The receiver comprises an estimation module, a coding module and a weighting module, wherein the estimation module is used for estimating the signal channel state information of the local users; the coding module is used for obtaining the quantized precoding vectors of the local users and the interference users; and the weighting module is used for forming the linear weight vector gi according to the signal channel state information, the quantized precoding vectors of the local users and the interference users. The signaling information which is sent to the users by a base station does not only contain the purchase management index (PMI) of the local users, a user side can obtain the accurate interference information, direct calculation is carried out, and the interference and noise correlation matrix are obtained, so the multi-user interference is effectively eliminated.

Description

A kind of interference rejection combining method and receiver

Technical field

The present invention relates to wireless communication field, be specifically related to a kind of interference rejection combining method and suppress to merge receiver with disturbing.

Background technology

In follow-on broadband wireless communication network, how to solve in the cordless communication network and to support bigger bandwidth to become a key factor that improves inter-cell user terminal throughput and user terminal average throughput under the bandwidth in current wireless communication system and be faced with formidable challenges.

In multi-user system, for ZF-BF (Zero-forceing beam forming, close-to zero beam forming) pre-coding scheme, along with the increase of signal to noise ratio, user throughput can reach a performance limit.This is the finiteness because of feedback bits (bit) number, causes the inexactness of channel information, and it can't be eliminated between the multi-user fully and disturb.Under the high s/n ratio condition, disturbing between the multi-user becomes the principal element that influences user performance.Therefore how to eliminate inter-user interference, lifting is very helpful to systematic function.

(Interference Reject Combine, IRC) algorithm can utilize receive diversity effectively to eliminate the interference between the user at receiving terminal to disturb the inhibition merging.Traditional diversity combining method, for example high specific merges (MRC), owing to only consider the channel self character, disturbs the principal element that becomes the system of influence between the multi-user.(Minimum Mean Square Error MMSE) receives merge algorithm to least mean-square error, has also only considered the influence of noise item.The difference of IRC algorithm and above-mentioned algorithm is, utilizes receive diversity that interference and noise item are all handled at receiving terminal, has realized the inhibitory action to coloured noise (interference plus noise).The key of using the IRC algorithm to suppress to disturb is to obtain the correlation matrix of interference and noise to obtain weighing vector.The correlation properties of accurate acquisition interference and noise have very big influence to the performance of IRC algorithm.

At present; Long Term Evolution (Long Term Evolution in third generation partner program; LTE) propose two kinds in and calculate the methods of disturbing with noise correlation matrix: a kind of is to adopt the autocorrelation matrix approximate evaluation interference of received signal and the covariance matrix of noise, and carry out on the time-domain and frequency-domain on average; Another method is directly to adopt to disturb with noise to calculate its covariance matrix, also carries out average (can adopt the transmission signal here, also can adopt pilot signal) on the time-domain and frequency-domain.Wherein the advantage of method one is calculate comparatively simply, and complexity is relatively low, and shortcoming is directly to utilize to receive signal and cause that certain error is arranged; The advantage of method two is to utilize real interference and its covariance matrix of Noise Estimation, and performance can be better in theory.But it is higher that shortcoming is a complexity, at first will do one according to a preliminary estimate to sending signal, and the complexity of inverting is also higher relatively.

Summary of the invention

The present invention provides a kind of interference rejection combining method and device, solves existing interference rejection combining method and can not effectively eliminate the interference between the user, is applied to the problem of heterogeneous network underaction.

In order to address the above problem, the present invention proposes a kind of interference rejection combining method, and said interference rejection combining method comprises:

Obtain local user's channel condition information;

Obtain the precoding vector after local user and interference user quantize;

According to the precoding vector after channel condition information, local user and the interference user quantification, construct linear weighing vector g _i

Further, said method also comprises:

Obtain local user's noise vector ni; When the linear weighing vector gi of structure, local user's noise ni is counted.

Further, when the local user is linear the reception, linear weighted function vector g _iFor

${g_i}^H=h_{\mathrm{Ei}}^H{\left\{E\left(f_i{f_i}^H\right)\right\}}^{-1}$ Perhaps ${g_i}^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Ff}}^{-1},$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

f _iExpression interference user channel information vector, $f_i=H_i\·\underset{1\≤j\≤N}{\overset{j\≠i}{\mathrm{\Σ}}}{w}_j;$

R _Ff=E{ff ^HBeing the correlation matrix of interference, f is the interference channel information matrix, H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

Further, when the local user is linear the reception, linear weighted function vector g _iFor:

$g_i^H=h_{\mathrm{Ei}}^H{\left\{E\left(u_i{u_i}^H\right)\right\}}^{-1}$ Perhaps $g_i^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1};$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

u _i=f _iS _i+ n _i, f _iExpression interference user channel information vector,

s _iThe expression base station sends to local user's data matrix;

R _Uu=E{uu ^HBeing the covariance matrix of interference and noise, u is the interference plus noise information matrix, H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

Further, when the local user is linear the reception, linear weighted function vector g _iFor:

$g_i^H=h_{\mathrm{Ei}}^H{(h_{\mathrm{Ei}}{h}_{\mathrm{Ei}}^H+R_{\mathrm{Uu}})}^{-1}$ Perhaps $g_i^H=\frac{1}{1+h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}{h}_{\mathrm{Eu}}}{h}_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

R _Uu=E{uu ^HBeing the covariance matrix of interference and noise, u is the interference plus noise information matrix, u _i=f _iS _i+ n _i, f _iExpression interference user channel information vector,

s _iThe expression base station sends to local user's data matrix;

H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

Further, said method also comprises: adaptation coefficient α is set,

According to the precoding vector after channel condition information, local user and the interference user quantification, construct linear weighing vector g _iBefore also comprise:

Judge interference value Z≤α σ ²The time, then construct linear weighing vector g according to following mode _i: $g_i^H=h_{\mathrm{Ei}}^H;$

Judge interference value Z＞α σ ²The time, the precoding vector after just quantizing according to channel condition information, local user and interference user is constructed linear weighing vector g _i

Z=||f|| wherein ², h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users, σ ²Be noise variance.

Further, said according to the precoding vector after channel condition information, local user and the interference user quantification, construct linear weighing vector g _iThe time:

Calculate acquisition linear weighted function vector g by the local user _i

Or calculate to obtain said user's linear weighted function vector g by the base station _i, and with said linear weighted function vector g _iSend to said user.

In order to address the above problem, the present invention also proposes a kind of interference and suppresses to merge receiver, and said receiver comprises: estimation module, coding module and weighting block, wherein:

Said estimation module is used to estimate to obtain local user's channel condition information;

Precoding vector after said coding module is used to obtain local user and interference user and quantizes;

Said weighting block is used for constructing linear weighing vector g according to the precoding vector after channel condition information, local user and the interference user quantification _i

Further, said receiver also comprises the noise module, and said noise module is used to obtain local user's noise vector n _i

Said weighting block is at the linear weighing vector g of structure _iThe time, the local user's that the noise module is obtained noise n _iCount.

Said further weighting block is constructed linear weighing vector g _iThe time:

When the local user is linear the reception, linear weighted function vector g _iFor

${g_i}^H=h_{\mathrm{Ei}}^H{\left\{E\left(f_i{f_i}^H\right)\right\}}^{-1}$ Perhaps ${g_i}^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Ff}}^{-1},$

Or

$g_i^H=h_{\mathrm{Ei}}^H{\left\{E\left(u_i{u_i}^H\right)\right\}}^{-1}$ Perhaps $g_i^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1};$

Or

$g_i^H=h_{\mathrm{Ei}}^H{(h_{\mathrm{Ei}}{h}_{\mathrm{Ei}}^H+R_{\mathrm{Uu}})}^{-1}$ Perhaps $g_i^H=\frac{1}{1+h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}{h}_{\mathrm{Eu}}}{h}_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

R _Ff=E{ff ^HBeing the correlation matrix of interference, f is the interference channel information matrix, R _Uu=E{uu ^HBeing the covariance matrix of interference and noise, u is the interference plus noise information matrix, u _i=f _iS _i+ n _i, f _iExpression interference user channel information vector,

s _iThe expression base station sends to local user's data matrix; H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

Further, said receiver also comprises adaptation module, and said adaptation module is used to be provided with adaptation coefficient α,

Said weighting block is constructed linear weighing vector g _iBefore also comprise:

Judge interference value Z≤α σ ²The time, then construct linear weighing vector g according to following mode _i: $g_i^H=h_{\mathrm{Ei}}^H;$

Judge interference value Z＞α σ ²The time, the precoding vector after just quantizing according to channel condition information, local user and interference user is constructed linear weighing vector g _i

Z=||f|| wherein ², h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users, σ ²Be noise variance.

(Procoding Matrix Index, enhancing PMI) is notified to disturb and is suppressed to merge to reach the purpose of interference eliminated to interference user precoding indication through the base station in the present invention.Among the present invention, the signaling information that send the user base station does not include only local user's PMI, comprises the PMI of interference user simultaneously, and local user's the PMI and the PMI of interference user form pre-coding matrix W.User side just can obtain accurate interference information like this, and directly calculates interference and noise correlation matrix, is used to eliminate between the multi-user and disturbs.The present invention proposes a kind of self adaptation IRC receiver based on disturbance regime on this basis: transmitting terminal and receiving terminal be estimating noise item and interference item size respectively, when disturbing relative noise big, adopts interference rejection combining method of the present invention; When disturbing relative noise hour, can directly adopt traditional M RC method, complexity reduces greatly under the situation that performance does not obviously descend.

Description of drawings

Fig. 1 is a scene graph in the sub-district of the present invention;

Fig. 2 is the rate capability comparison diagram of different implementation methods of the present invention;

Fig. 3 is the error rate (BER, Bit Error Rate) the performance comparison diagram of different implementation methods of the present invention;

Fig. 4 is the throughput performance curve chart of the present invention's IRC-adaptive method and IRC-SINR method under different alpha parameter values;

Fig. 5 is the bit error rate performance curve chart of the present invention's IRC-adaptive method and IRC-SINR method under different alpha parameter values;

Fig. 6 is a method flow diagram of the present invention.

Embodiment

For making the object of the invention, technical scheme and advantage clearer, hereinafter will combine accompanying drawing that embodiments of the invention are elaborated.Need to prove that under the situation of not conflicting, embodiment among the application and the characteristic among the embodiment be combination in any each other.

For the multi-user system that contains N user, interference rejection combining method of the present invention is following:

User i obtains the channel condition information H of user i through estimating the base station _i, 1≤i≤N wherein, N is a total number of users;

Adopt the precoding vector w after pre-coding scheme obtains user i and interference user j quantification respectively _iAnd w _j

Precoding vector w after said user i and interference user j quantize _iAnd w _jCan obtain for adopting the close-to zero beam forming pre-coding scheme.

Channel condition information H according to said user i _iWith the precoding vector w after the user i quantification _i, calculate the equivalent channel state information matrix h that acquisition comprises each user's channel condition information _e

Equivalent channel state information matrix h _eBe expressed as:

H wherein _EiThe equivalent channel state information vector of expression user i, h _Ei=H _iW _i

Channel condition information H according to said user i _iWith the precoding vector w after the interference user j quantification _j, calculate to obtain interference channel information matrix f, j ≠ i wherein, 1≤j≤N;

Interference channel information matrix f is expressed as:

F wherein _iThe interference channel information vector of expression user i,

Obtain the noise matrix n of receiving terminal;

The noise matrix n of receiving terminal is expressed as:

N wherein _iThe noise of expression user i receiving terminal, n _i∈ CN (0, σ ²).

The signal matrix Y, the equivalent channel state information matrix h that utilize each user to receive _e, interference channel information matrix f and receiving terminal noise matrix n, construct linear weighting matrix g, go back original sender and give user's data matrix s;

The signal matrix Y that each user receives is expressed as: Y=h _eS+fs+n, promptly

Formula one $\left\{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="y"\; filenames="HDA0000061034250000021.png,HDA0000061034250000031.png"\; state="\{\{state\}\}">y</figure-callout>}}_1=h_{e1}{\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="HDA0000061034250000021.png,HDA0000061034250000031.png"\; state="\{\{state\}\}">s</figure-callout>}}_1+{\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA0000061034250000021.png,HDA0000061034250000031.png"\; state="\{\{state\}\}">f</figure-callout>}}_1\&CenterDot;\underset{2\&le;j\&le;N}{\mathrm{\&Sigma;}}{s}_j+{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="HDA0000061034250000021.png,HDA0000061034250000031.png"\; state="\{\{state\}\}">n</figure-callout>}}_1\\ ...\\ y_i=h_{\mathrm{Ei}}{s}_i+f_i\&CenterDot;\underset{1\&le;j\&le;N}{\overset{j\&NotEqual;i}{\mathrm{\&Sigma;}}}{s}_j+n_i\\ ...\\ y_N=h_{\mathrm{EN}}{s}_N+f_N\&CenterDot;\underset{1\&le;j\&le;N-1}{\mathrm{\&Sigma;}}{s}_j+n_N\end{array}\right.,$

Receiving terminal is linear the reception, and the linear weighted function vector is g, and detected signal matrix r is expressed as:

R=g ^H(h _eS+fs+n), promptly

Formula two $\left\{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HDA0000061034250000021.png,HDA0000061034250000031.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=g_1^H(h_{e1}{\mathrm{<figure-callout\; id="1"\; label="s"\; filenames="HDA0000061034250000021.png,HDA0000061034250000031.png"\; state="\{\{state\}\}">s</figure-callout>}}_1+{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="HDA0000061034250000021.png,HDA0000061034250000031.png"\; state="\{\{state\}\}">H</figure-callout>}}_1\&CenterDot;\underset{2\&le;j\&le;N}{\mathrm{\&Sigma;}}{w}_j{s}_j+n_1)\\ ...\\ r_i=g_i^H(h_{\mathrm{Ei}}{s}_i+H_i\&CenterDot;\underset{1\&le;j\&le;N}{\overset{j\&NotEqual;i}{\mathrm{\&Sigma;}}}{w}_j{s}_j+n_i)\\ ...\\ r_N=g_N^H(h_{\mathrm{EN}}{s}_N+H_N\&CenterDot;\underset{1\&le;j\&le;N-1}{\mathrm{\&Sigma;}}{w}_j{s}_j+n_N)\end{array}\right.$

Construct linear weighting matrix

Eliminate to disturb and noise, make detected signal matrix r also original sender give user's data matrix s, wherein g _iExpression user i weighing vector, subscript H representes conjugate transpose.

Construct the method one of linear weighting matrix/vectorial g: the IRC-ZF method:

Multiply by the kernel null (f of corresponding distracter at the various two ends of formula one _i), the signal matrix Y ' behind the distracter that is eliminated;

Multiply by the conjugation null (f of the equivalent channel state information vector of respective user i again at the various two ends of signal matrix Y ' _i) h _Ei, the signal matrix Y that obtains having the local signal to noise ratio of maximization ",

The base station is calculated and is obtained linear weighted function vector g _i, be expressed as

${g_i}^H={\left\{\mathrm{null}\left(f_i\right)h_{\mathrm{ei}}\right\}}^H\mathrm{null}\left(f_i\right)$

$=h_{\mathrm{ei}}^H\left\{\mathrm{null}{\left(f_i\right)}^H\mathrm{null}\left(f_i\right)\right\}$

$=h_{\mathrm{ei}}^H{\left\{f_i{f_i}^H\right)\}}^{-1},$

$=h_{\mathrm{ei}}^H{R}_{\mathrm{ff}}^{-1}$

Wherein, R _Ff=E{ff ^HIt is the correlation matrix that disturbs.

Construct the method two of linear weighting matrix/vectorial g: the IRC-SINR method:

The signal matrix Y that each user is received is expressed as: Y=h _eS+u, wherein u=fs+n represent distracter and noise and matrix;

Make receiving terminal SINR maximum, target function is expressed as:

$\underset{g}{\max }\mathrm{SINR}=\frac{E\left\{{\left|g^H{h}_{e}s\right|}^2\right\}}{E\left\{{\left|g^{H}u\right|}^2\right\}}=\frac{g^H{h}_e{h}_e^{H}g}{g^H{R}_{\mathrm{uu}}{g}^H},$

R wherein _Uu=E{uu ^HBe the covariance matrix of interference and noise,

Obtain making by broad sense rayleigh quotient theory and be expressed as linear weighted function matrix g:

$g^H=h_e^H{R}_{\mathrm{uu}}^{-1}\frac{-b\&PlusMinus;\sqrt{b^2-4\mathrm{ac}}}{2a}.$

Construct the method three of linear weighting matrix/vectorial g: the IRC-MMSE method:

The signal matrix Y that each user is received is expressed as: Y=h _eS+u, wherein u=fs+n represent distracter and noise and matrix;

Based on the objective criteria of least mean-square error, target function is expressed as:

$\underset{g}{\min }E\left\{{|g^{H}y-s|}^2\right\}$

Obtaining linear weighted function matrix g is:

$g^H=h_e^H{(h_e{h}_e^H+R_{\mathrm{uu}})}^{-1}=\frac{1}{1+h_e^H{R}_u^{-1}{h}_e}{h}_e^H{R}_{\mathrm{uu}}^{-1},$

R wherein _Uu=E{uu ^HIt is the covariance matrix of interference and noise.

Construct the method four of linear weighting matrix/vectorial g: the IRC-adaptive method:

The base station is provided with adaptation coefficient α,

$g_{\mathrm{opt}}^H=\left\{\begin{array}{cc}{h}_e^H& Z\&le;\mathrm{\&alpha;}{\mathrm{\&sigma;}}^2\\ h_e^H{R}_{\mathrm{uu}}^{-1}& Z>\mathrm{\&alpha;}{\mathrm{\&sigma;}}^2\end{array}\right.,$

Wherein Z is an interference value, Z=||f|| ², n _i∈ CN (0, σ ²), σ ²Be noise variance.

Calculate linear weighted function vector g _iIn time, can be calculated by the local user and obtained linear weighted function vector g _iOr calculate to obtain said user's linear weighted function vector g by the base station _i, and with said linear weighted function vector g _iSend to said user.

The present invention is the research background with (Intral-cell) in the sub-district among the CoMP (Coordinated Multi Point Transmission/Reception, coordinate multipoint transmission/reception).The Intral-cell scene is as shown in Figure 1, and present embodiment is an example with two base stations and two users, and two base stations are connected by a master controller.Each user can estimate that the base station obtains local channel condition information, representes with H1 and H2 respectively.

The signal indication that the user receives is:

(three)

S wherein _kFor transmitting terminal sends to the data symbol of user k, H _kThe expression transmitting terminal is to the channel condition information of user k, n _k∈ CN (0, σ ²) be the noise of user k receiving terminal.Here adopt the ZF-BF pre-coding scheme, w _kBe the precoding vector after the user k quantification.H _kw _ks _kBe the valid data that receive, because there is certain quantization error, so distracter H ₁w ₂s ₂And H ₂w ₁s ₁Be not 0.

Make h _Ek=H _kW _k, the product of channel condition information and local precoding vector as the equivalent channel state information, is made f _k=H _kW _n, with the product of channel condition information and interference user precoding vector as the interference channel state information, k=1 here, 2, n=1,2.Formula (three) can be reduced to:

y ₁＝h _e1s ₁+f ₁s ₂+n ₁

y ₂=h _E2s ₂+ f ₂s ₁+ n ₂(4)

Receiving terminal is a linearity test, and detected signal definition is r _k:

r ₁＝g ₁ ^H(h _e1s ₁+f ₁s ₂+n ₁)

r ₂=g ₂ ^H(h _E2s ₂+ f ₂s ₁+ n ₂) (five)

Construct linear weighting matrix/vector

Eliminate to disturb and noise, make detected signal matrix r _kAlso original sender is given user's data matrix s _k, g wherein _kExpression user k weighing vector, subscript H representes conjugate transpose.

Construct the method one of linear weighting matrix/vectorial g: the IRC-ZF method:

Multiply by the kernel null (f of corresponding distracter at the two ends of formula (four) _k), to eliminate distracter;

Multiply by the conjugation null (f of corresponding equivalent channel state information again _k) h _Ek, obtaining having the local signal to noise ratio of maximization, purpose is to guarantee to receive the signal energy maximum;

The base station is calculated and is obtained linear weighted function vector g _k, be expressed as:

${g_k}^H={\left\{\mathrm{null}\left(f_k\right)h_{\mathrm{ek}}\right\}}^H\mathrm{null}\left(f_k\right)$

$=h_{\mathrm{ek}}^H\left\{\mathrm{null}{\left(f_k\right)}^H\mathrm{null}\left(f_k\right)\right\}$

$=h_{\mathrm{ek}}^H{\left\{E\left(f_k{f_k}^H\right)\right\}}^{-1},$

$=h_{\mathrm{ek}}^H{R}_{\mathrm{ff}}^{-1}$

Wherein, R _Ff=E{ff ^HIt is the correlation matrix that disturbs.

Does not f in last two formulas need indexing k? How does the correlation matrix that disturbs specifically calculate? (R _Ff=E (ff ^H), when h multiply by in the front _Ek ^HAfter, h _Ek ^HE (ff ^H)=h _Ek ^HE (f _kf _k ^H))

When adopting following formula as weighing vector, multi-user interference can be eliminated fully, so (interference-limited) performance has very big raising under the high s/n ratio scene.

Construct the method two of linear weighting matrix/vectorial g: the IRC-SINR method:

Formula (four) is rewritten the signal indication that receives is:

y ₁＝h _e1s ₁+u ₁

y ₂=h _E2s ₂+ u ₂(6)

U wherein _kExpression distracter and noise and;

In order to make receiving terminal SINR maximum, target function is expressed as:

$\underset{g}{\max }\mathrm{SINR}=\frac{E\left\{{\left|g^H{h}_{e}s\right|}^2\right\}}{E\left\{{\left|g^{H}u\right|}^2\right\}}=\frac{g^H{h}_e{h}_e^{H}g}{g^H{R}_{\mathrm{uu}}{g}^H}$

R wherein _Uu=E{uu ^HIt is the covariance matrix of interference and noise.

Obtain linear weighted function matrix/vector g by broad sense rayleigh quotient theory, be expressed as:

$g^H=h_e^H{R}_{\mathrm{uu}}^{-1}.$

R wherein _Uu=E{uu ^HIt is the covariance matrix of interference and noise.

Construct the method three of linear weighting matrix/vectorial g: the IRC-MMSE method:

The signal that receives is suc as formula (six) expression, and no with method two is the objective criteria that this method is based on least mean-square error (MMSE), and target function is expressed as:

$\underset{g}{\min }E\left\{{|g^{H}y-s|}^2\right\}$

Obtaining linear weighted function matrix/vector g is:

$g^H=h_e^H{(h_e{h}_e^H+R_{\mathrm{uu}})}^{-1}=\frac{1}{1+h_e^H{R}_u^{-1}{h}_e}{h}_e^H{R}_{\mathrm{uu}}^{-1}$

R wherein _Uu=E{uu ^HIt is the covariance matrix of interference and noise.

The difference of this method and traditional M MSE method of reseptance is to have considered distracter, and it is joined in the middle of the noise.It should be noted that the linear weighted function matrix/vector g that method two and method three obtain only differs from a coefficient, IRC-MMSE method and IRC-SINR method are equivalent.

Construct the method four of linear weighting matrix/vectorial g: the IRC-adaptive method:

The base station is provided with adaptation coefficient α,

$g^H=\left\{\begin{array}{cc}{h}_e^H& Z\&le;\mathrm{\&alpha;}{\mathrm{\&sigma;}}^2\\ h_e^H{R}_{\mathrm{uu}}^{-1}& Z>\mathrm{\&alpha;}{\mathrm{\&sigma;}}^2\end{array}\right.,$

Wherein Z is an interference value, Z=||f|| ², n _i∈ CN (0, σ ²), σ ²Be noise variance.

This method is provided with a constant alpha, as Z≤α σ ²The time, directly adopt traditional MRC method of reseptance; As Z＞α σ ²The time, adopt IRC-SINR method of reseptance or IRC-ZF method of reseptance or IRC-MMSE method of reseptance.

Here IRC-adaptive method of reseptance key parameter is α, and its parameter setting hopes under the situation that descending does not appear in performance, to reduce as much as possible the reception complexity.When α select too small, can cause disturbing and still select the IRC method of reseptance under the very low situation, the complexity of the relative IRC-SINR method of reseptance of IRC-adaptive method of reseptance reduces not obvious like this.And excessive when the α selection, can cause disturbing under the very big situation, still select the MRC method of reseptance, systematic function descends to some extent.

Calculate the covariance matrix R of interference and noise _UuThe time, the user can obtain the PMI of interference user, obtains accurate interference channel information f easily, and the correlation matrix that calculating can obtain disturbing is R _Ff=E{ff ^H, the covariance matrix of interference and noise is R _Uu=E{uu ^H.Noise is a white Gaussian noise, with interference be incoherent.Obtain following formula easily:

R _uu＝E{ff ^H}+σ ²I＝R _ff+σ ²I

So just can calculate R _Uu, corresponding linear weighting matrix/vectorial g also just obtains easily.

Is the linear weighted function matrix/vector to be calculated or user's calculating by the base station?

Traditional method of reseptance and each method of reseptance of the present invention more as shown in table 1.

Table 1

The present invention is used for multi-user's heterogeneous network, the present invention is based on the base station to the enhancing of distracter notice, carries out the interference eliminated between the multi-user in conjunction with the IRC method of reseptance.The user not only obtains local user's PMI, obtains the PMI of interference user simultaneously, and the adjacent area interference of base station measurement can obtain the PMI of interference user.In addition, the PMI of the Signalling exchange acquisition interference user between cooperative base station (sub-district) can be passed through in the base station.Theory and combining emulation has proved IRC method (noise limited) under the low signal-to-noise ratio scene, and is approaching with MRC method performance; (interference-limited) has greatly improved to systematic function under the high s/n ratio situation.The IRC-ZF method that the present invention proposes, it can eliminate distracter in theory fully.The IRC method has certain higher with respect to its complexity of traditional method, therefore proposed to carry out MRC and the adaptively selected a kind of method of reseptance of IRC-SINR, i.e. IRC-adaptive method among the present invention according to disturbance regime.Adaptive reception can balance system performance and the complexity of receiving algorithm.

Analyze the throughput performance and the bit error rate performance of several method of the present invention in conjunction with Fig. 2 to Fig. 5: at first the performance of its various method of reseptances is carried out simulating, verifying; Provide the systematic function curve of different alpha parameters under being provided with then, therefrom chosen suitable alpha parameter value.

For simplicity, suppose that the Intral-cell system comprises 2 base stations and 2 users, and base station end and user side all have 2 antennas, the transmitting terminal equivalence is 4 transmitting antennas.The ZF-BF pre-coding scheme is adopted in precoding.Transmitting terminal contains identical random code book with receiving terminal, to the vector quantization that 6bit is carried out in zero precoding of compeling that obtains.Receiving terminal can obtain local channel information here, and knows local precoding vectors and disturb precoding vectors.Receiving terminal can carry out relevant detection so.The system's average throughput and the error rate are carried out emulation respectively.

Can find out that from the simulation result of Fig. 2 throughput of system IRC method (interference-limited) under high s/n ratio has greatly improved to performance.Along with the raising of signal to noise ratio, IRC-ZF method and IRC-SINR method performance are approaching, and wherein IRC-SINR method (consistent with the IRC-MMSE method) performance is best relatively.Under the low signal-to-noise ratio situation (noise limited), IRC-SINR method (IRC-MMSE method) MRC method does not relatively have tangible performance boost, and IRC-ZF method performance has obvious decline.The IRC-ZF method has been eliminated interference in theory fully though this is, disturb this moment is not the main cause that influences systematic function, and it has reduced the energy that receives signal accordingly with respect to the MRC method, so performance has loss.

The result can obtain similar conclusion from Fig. 3 error rate of system performance simulation, and promptly IRC-SINR method (IRC-MMSE method) (interference-limited) under high s/n ratio has tangible lifting with respect to other receiving algorithm performances.

Fig. 4 and Fig. 5 have provided the throughput and the BER performance curve of IRC-adaptive method of reseptance under the different alpha parameter values respectively, and compare with the performance of IRC-SINR method of reseptance.Find easily that from figure the IRC-adaptive performance curve is provided with down at the different alpha parameter that has provided, significantly descending does not all appear in performance, can under the situation that guarantees systematic function, reduce and send and receive complexity.When signal to noise ratio is very high, the adaptively selected IRC method of IRC-adaptive method is therefore consistent with the performance of IRC-SINR; When signal to noise ratio was very low, the adaptively selected MRC method of IRC-adaptive method was because be noise limited this moment, so performance and IRC-SINR method are also very approaching.When interference and noise are comparable (the same order of magnitude), the relative IRC-SINR method of IRC-adaptive method performance descends to some extent, and alpha parameter is set big more, and performance loss is big more.When α was set at 1, performance was approaching, basic not loss; When α is set at 5, the nearly 1dB of performance loss.In actual application, the alpha parameter value can be set as required.Herein, we can be set to 1 by α, when having reduced complexity, have guaranteed systematic function like this.

One of ordinary skill in the art will appreciate that all or part of step in the said method can instruct related hardware to accomplish through program, said procedure can be stored in the computer-readable recording medium, like read-only memory, disk or CD etc.Alternatively, all or part of step of the foregoing description also can use one or more integrated circuits to realize.Correspondingly, each the module/unit in the foregoing description can adopt the form of hardware to realize, also can adopt the form of software function module to realize.The present invention is not restricted to the combination of the hardware and software of any particular form.

Above embodiment is only unrestricted in order to technical scheme of the present invention to be described, only with reference to preferred embodiment the present invention is specified.Those of ordinary skill in the art should be appreciated that and can make amendment or be equal to replacement technical scheme of the present invention, and do not break away from the spirit and the scope of technical scheme of the present invention, all should be encompassed in the middle of the claim scope of the present invention.

Claims (11)

1. interference rejection combining method is applied to comprise in the multi-user system:

Obtain local user's channel condition information;

Obtain the precoding vector after local user and interference user quantize;

According to the precoding vector after channel condition information, local user and the interference user quantification, construct linear weighing vector g _i

2. interference rejection combining method as claimed in claim 1 is characterized in that: said method also comprises:

Obtain local user's noise vector n _iAt the linear weighing vector g of structure _iThe time, with local user's noise n _iCount.

3. interference rejection combining method as claimed in claim 1 is characterized in that:

When the local user is linear the reception, linear weighted function vector g _iFor

${g_i}^H=h_{\mathrm{Ei}}^H{\left\{E\left(f_i{f_i}^H\right)\right\}}^{-1}$ Perhaps ${g_i}^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Ff}}^{-1},$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

f _iExpression interference user channel information vector, $f_i=H_i\&CenterDot;\underset{1\&le;j\&le;N}{\overset{j\&NotEqual;i}{\mathrm{\&Sigma;}}}{w}_j;$

R _Ff=E{ff ^HBeing the correlation matrix of interference, f is the interference channel information matrix, H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

4. interference rejection combining method as claimed in claim 2 is characterized in that:

When the local user is linear the reception, linear weighted function vector g _iFor:

$g_i^H=h_{\mathrm{Ei}}^H{\left\{E\left(u_i{u_i}^H\right)\right\}}^{-1}$ Perhaps $g_i^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1};$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

u _i=f _iS _i+ n _i, f _iExpression interference user channel information vector,

s _iThe expression base station sends to local user's data matrix;

R _Uu=E{uu ^HBeing the covariance matrix of interference and noise, u is the interference plus noise information matrix, H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

5. interference rejection combining method as claimed in claim 2 is characterized in that:

When the local user is linear the reception, linear weighted function vector g _iFor:

$g_i^H=h_{\mathrm{Ei}}^H{(h_{\mathrm{Ei}}{h}_{\mathrm{Ei}}^H+R_{\mathrm{Uu}})}^{-1}$ Perhaps $g_i^H=\frac{1}{1+h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}{h}_{\mathrm{Eu}}}{h}_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

R _Uu=E{uu ^HBeing the covariance matrix of interference and noise, u is the interference plus noise information matrix, u _i=f _iS _i+ n _i, f _iExpression interference user channel information vector,

s _iThe expression base station sends to local user's data matrix;

H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

6. like the arbitrary described interference rejection combining method of claim 1 to 5, it is characterized in that: said method also comprises: adaptation coefficient α is set,

According to the precoding vector after channel condition information, local user and the interference user quantification, construct linear weighing vector g _iBefore also comprise:

Judge interference value Z≤α σ ²The time, then construct linear weighing vector g according to following mode _i: $g_i^H=h_{\mathrm{Ei}}^H;$

Judge interference value Z＞α σ ²The time, the precoding vector after just quantizing according to channel condition information, local user and interference user is constructed linear weighing vector g _i,

Z=||f|| wherein ², h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users, σ ²Be noise variance.

7. according to claim 1 or claim 2 interference rejection combining method is characterized in that:

Said according to the precoding vector after channel condition information, local user and the interference user quantification, construct linear weighing vector g _iThe time:

Calculate acquisition linear weighted function vector g by the local user _i

Or calculate to obtain said user's linear weighted function vector g by the base station _i, and with said linear weighted function vector g _iSend to said user.

8. one kind is disturbed inhibition to merge receiver, and it is characterized in that: said receiver comprises estimation module, coding module and weighting block, wherein:

Said estimation module is used to estimate to obtain local user's channel condition information;

Precoding vector after said coding module is used to obtain local user and interference user and quantizes;

Said weighting block is used for constructing linear weighing vector g according to the precoding vector after channel condition information, local user and the interference user quantification _i

9. receiver as claimed in claim 8 is characterized in that: said receiver also comprises the noise module, and said noise module is used to obtain local user's noise vector n _i

Said weighting block is at the linear weighing vector g of structure _iThe time, the local user's that the noise module is obtained noise n _iCount.

10. receiver as claimed in claim 9 is characterized in that:

Said weighting block is constructed linear weighing vector g _iThe time:

When the local user is linear the reception, linear weighted function vector g _iFor

${g_i}^H=h_{\mathrm{Ei}}^H{\left\{E\left(f_i{f_i}^H\right)\right\}}^{-1}$ Perhaps ${g_i}^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Ff}}^{-1};$

Or

$g_i^H=h_{\mathrm{Ei}}^H{\left\{E\left(u_i{u_i}^H\right)\right\}}^{-1}$ Perhaps $g_i^H=h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1};$

Or

$g_i^H=h_{\mathrm{Ei}}^H{(h_{\mathrm{Ei}}{h}_{\mathrm{Ei}}^H+R_{\mathrm{Uu}})}^{-1}$ Perhaps $g_i^H=\frac{1}{1+h_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}{h}_{\mathrm{Eu}}}{h}_{\mathrm{Ei}}^H{R}_{\mathrm{Uu}}^{-1}$

Wherein, h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

R _Ff=E{ff ^HBeing the correlation matrix of interference, f is the interference channel information matrix, R _Uu=E{uu ^HBeing the covariance matrix of interference and noise, u is the interference plus noise information matrix, u _i=f _iS _i+ n _i, f _iExpression interference user channel information vector,

s _iThe expression base station sends to local user's data matrix; H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users.

11. like the arbitrary described receiver of claim 8 to 10, it is characterized in that: said receiver also comprises adaptation module, said adaptation module is used to be provided with adaptation coefficient α,

Said weighting block is constructed linear weighing vector g _iBefore also comprise:

Judge interference value Z≤α σ ²The time, then construct linear weighing vector g according to following mode _i: $g_i^H=h_{\mathrm{Ei}}^H;$

Judge interference value Z＞α σ ²The time, the precoding vector after just quantizing according to channel condition information, local user and interference user is constructed linear weighing vector g _i,

Z=||f|| wherein ², h _EiExpression local user's equivalent channel state information vector, h _Ei=H _iW _i

H _iBe channel condition information, w _iBe the precoding vector after local user's quantification, w _jBe the precoding vector after the interference user quantification; 1≤i wherein, j≤N, N are total number of users, σ ²Be noise variance.

Images