CN103117787A - Self-adaptive bit distributing method and self-adaptive bit distributing device in cooperative multi-aerial system - Google Patents

Abstract

The present invention provides an adaptive bit allocation method and device in a cooperative multi-antenna system, the method includes extracting the received power of data signals and interference signals and the time delay information of each channel according to channel estimation; according to the signal received power and each The delay information of the channel calculates the optimal number of feedback bits allocated for the data channel and each interference channel under the premise that the total number of feedback bits is fixed, and minimizes the rate loss caused by the non-ideal state of the channel information; according to the optimal The number of feedback bits determines the size of the codebook, and uses the ideal channel information to select the corresponding codeword quantized channel state information from the codebook according to the minimum chord distance criterion and send it; receive and generate zero-forcing based on the quantized interference channel information precoding matrix. In the present invention, the base station uses the obtained non-ideal channel information to perform zero-forcing precoding operation, thereby suppressing multi-user interference in the local cell and inter-cell interference to neighboring cells.

Description

Self-adaptive bit allocation method and device in a kind of collaborative multiaerial system

Technical field

The present invention relates to self-adaptive bit allocation method and device in a kind of collaborative multiaerial system, belong to wireless communication technology field.

Background technology

In wireless communication system, presence of intercell interference is the key factor that causes the Cell Edge User communication quality to reduce.For the user of center of housing estate, itself is closer from the distance of base station, and the interference signal of outer residential quarter is distant, and its Signal to Interference plus Noise Ratio is relatively large.User for cell edge, the user who takies same bandwidth resources due to neighbor cell is larger to its interference ratio, itself is far away apart from the base station in addition, its Signal to Interference plus Noise Ratio is relatively just less, although cause the throughput of residential quarter integral body higher, but the QoS of customer of cell edge is relatively poor, and throughput is lower.Therefore, in wireless communication system, it is extremely important that presence of intercell interference suppresses technology.

The method of traditional minimizing presence of intercell interference is the channeling technology.When cell cluster is more greatly frequency duplex factor as one when larger, presence of intercell interference can be reduced effectively, but spectrum efficiency also can correspondingly reduce simultaneously.Current, OFDM (Orthogonal Frequency Division Multiplexing, OFDM) has become an important technology of Next-Generation Wireless Communication Systems.In ofdm system, in order to improve power system capacity, frequency duplex factor as one should be close to 1, and this moment, spectrum efficiency was very high, but presence of intercell interference is but very serious.Therefore the problem that solves presence of intercell interference is most important for the raising of wireless communication system overall performance.

In the multi-cell coordination communication system based on frequency division multiplexing, the user can estimate complete interference channel information by channel estimation technique, and this information exchange is crossed based on feedback link feeds back to this cell base station.This cell base station passes to interference channel information by X2 interface the base station of this community user being caused interference.The base station of interfered cell utilizes the information of the interference channel that obtains to carry out the ZF pre-encode operation can eliminate interference to interfered cell fully.Yet in the Limited Feedback system based on code book, user feedback is to CSI (the Channel Statement Information of base station, channel condition information) be after quantizing, and there is channel delay to exist, not desirable CSI, therefore the ZF precoding still can't be eliminated interference fully, causes the existence of residual interference, and residual interference can cause the decline of user's transmission rate.

Summary of the invention

The embodiment of the present invention provides self-adaptive bit allocation method and the device in a kind of collaborative multiaerial system, obtains maximum transmission rate by utilizing fixing feedback overhead, to solve the problem of presence of intercell interference.For this reason, the present invention proposes following technical scheme:

Self-adaptive bit allocation method in a kind of collaborative multiaerial system comprises the following steps:

Extract the received power of this cell data channel information, adjacent cell interference channel information, data-signal and interference signal and the Delay of each channel according to channel estimating;

Calculate respectively according to the Delay of described signal received power and each channel the Optimal Feedback bit number that distributes for data channel and each interference channel under the prerequisite that total number of bits of feedback is fixed, and the rate loss that the channel information nonideality causes is carried out minimization, to obtain ideal communication channel information;

Determine the size of code book according to described Optimal Feedback bit number, and utilize described ideal communication channel information to choose channel condition information after corresponding code word quantizes according to the minimum chordal distance criterion from described code book, and described channel condition information is fed back to serving BS by the based on feedback link that time delay is arranged;

Receive the channel condition information of described quantification by the X2 interface that time delay is arranged, and generate the ZF pre-coding matrix according to the interference channel information of described quantification.

Adaptive bit distributor in a kind of collaborative multiaerial system comprises:

Receiving element is used for extracting the received power of this cell data channel information, adjacent cell interference channel information, data-signal and interference signal and the Delay of each channel according to channel estimating;

The feedback computing unit, calculate respectively the Optimal Feedback bit number that distributes for data channel and each interference channel under the prerequisite that total number of bits of feedback is fixed for the Delay according to described signal received power and each channel, and the rate loss that the channel information nonideality causes is carried out minimization, to obtain ideal communication channel information;

Quantize computing unit, be used for determining according to described Optimal Feedback bit number the size of code book, and utilize described ideal communication channel information to choose channel condition information after corresponding code word quantizes according to the minimum chordal distance criterion from described code book, and the index information that described channel condition information is corresponding feeds back to serving BS by the based on feedback link that time delay is arranged;

The base station precoding unit is used for receiving by the X2 interface that time delay is arranged the channel condition information of described quantification, and generates the ZF pre-coding matrix according to the interference channel information of described quantification.

The present invention utilizes the non-ideal communication channel information of obtaining to carry out the ZF pre-encode operation by the base station, thus suppress in this residential quarter multi-user interference and to the presence of intercell interference of adjacent cell.

Description of drawings

Accompanying drawing is used to provide a further understanding of the present invention, and consists of the part of specification, is used for together with embodiments of the present invention explaining the present invention, is not construed as limiting the invention.In the accompanying drawings:

Fig. 1 is the schematic flow sheet of the self-adaptive bit allocation method in collaborative multiaerial system provided by the invention;

Fig. 2 is the system block diagram of collaborative MIMO in the embodiment of the present invention;

Fig. 3 is the structural representation of the adaptive bit distributor in collaborative multiaerial system provided by the invention.

Embodiment

Below in conjunction with accompanying drawing, the preferred embodiments of the present invention are described, should be appreciated that preferred embodiment described herein only is used for description and interpretation the present invention, is not intended to limit the present invention.

As shown in Figure 1, the embodiment of the present invention provides the self-adaptive bit allocation method in a kind of collaborative multiaerial system, comprises the following steps:

Step 1 is extracted the received power of this cell data channel information, adjacent cell interference channel information, data-signal and interference signal and the Delay of each channel according to channel estimating;

Step 2, calculate respectively according to the Delay of described signal received power and each channel the Optimal Feedback bit number that distributes for data channel and each interference channel under the prerequisite that total number of bits of feedback is fixed, and the rate loss that the channel information nonideality causes is carried out minimization, to obtain ideal communication channel information;

Step 3, determine the size of code book according to described Optimal Feedback bit number, and utilize described ideal communication channel information to choose channel condition information after corresponding code word quantizes according to the minimum chordal distance criterion from described code book, and described channel condition information is fed back to serving BS by the based on feedback link that time delay is arranged;

Step 4 receives the channel condition information of described quantification, and generates the ZF pre-coding matrix according to the interference channel information of described quantification by the X2 interface that time delay is arranged.

In this embodiment, suppose that in the multi-cell coordination multi-input multi-output system M adjacent cell arranged around a reference cell, the base station in reference cell arranges K user to carry out simultaneously transfer of data.In system, each base station and each user configure respectively n _TAnd n _RWith antenna.When the multiplexing factor of system frequency was 1, each user can be subject to two kinds of interference: 1. from the inter-user interference of this residential quarter, and 2 presence of intercell interference from adjacent cell.In reference cell, k user receives, and (1≤m≤M) signal power of individual interference base station is respectively P from serving BS and m _kWith Wherein

Be power coefficient, namely the power of this cell data signal is better than the power of the interference signal of interfered cell. Value depend on and channel large scale decline comprise path loss and shadow fading.K user of reference cell is at the reception signal of time point n $y_k\left[n\right]\∈G^{n_R\×1}$ Can be written as:

$y_k\left[n\right]=H_k\left[n\right]T_k\left[n\right]x_k\left[n\right]+H_k\left[n\right]\underset{j=1,j\≠k}{\overset{K}{\mathrm{\Σ}}}{T}_j\left[n\right]x_j\left[n\right]+\underset{m=1}{\overset{M}{\mathrm{\Σ}}}{G}_{k,m}\left[n\right]T_m\left[n\right]x_m\left[n\right]+z_k\left[n\right]---\left(1\right)$

Wherein

With

Be respectively reference cell and m interfered cell to the channel matrix of user k.

Be the transmitting data vector, satisfy maximum transmit power limit:

Be piece diagonal angle precoding unitary matrice.z _k[n] expression is obeyed and is distributed as CN (0, σ ²) multiple Gaussian noise vector.

In this embodiment, H _k[n] and G _{K, m}[n] has respectively the code book W that quantizes (RVQ) based on random vector _kAnd W _k,mQuantize.Generate in advance under each user code book used is all online and transmitting-receiving two-end all obtains the full content of code book in advance.Code book W _kAnd W _k,mSize be respectively

With

Namely

Wherein in code book, each element is random half unitary matrice that generates.Suppose that the total feedback bandwidth of each user is B _totBit, wherein B _kAnd B _k,mBit quantizes respectively H _k[n] and G _{K, m}[n].The user goes out corresponding code word as the channel matrix after quantizing from codebook selecting

With

System of selection is according to the minimum chordal distance criterion:

${\hat{H}}_k\left[n\right]=\underset{W_{k\∈}{W}_k}{{\arg \min d}^2}\left(H_k\right[n],W_k)$ （2）

${\hat{G}}_{k,m}\left[n\right]=\underset{W_{k,m\∈}{W}_{k,m}}{{\arg \min d}^2}\left(G_{k,m}\right[n],W_{k,m})$

The channel delay that causes due to feedback and the information interaction of X2 interface also can cause systemic loss of energy.Become the piece fading channel model when considering in this concrete enforcement, namely channel in a mark space fix and adjacent-symbol between channel response be correlated with.Channel relevancy can be expressed as:

H _k[n]＝η _kH _k[n-D _k]+E _k[n] (3)

G _k，m[n]＝η _k,mG _k，m[n-D _k,m]+E _k，m[n]

E wherein _k[n] and E _{K, m}[n] is the channel errors matrix, and and H _k[n-D _k] and G _{K, m}[n-D _k,m] relatively independent.D _kAnd D _k,mBe respectively the sampled point number of channel delay.E _k[n] and E _{K, m}Element in [n] is obeyed and is distributed

With

And satisfy respectively

Coefficient correlation η _kAnd η _k,mDetermined by carat card autocorrelation model:

η _k＝J ₀(2πf _dD _kT _s) (4)

η _k,m＝J ₀(2πf _dD _k,mT _s)

Wherein, J ₀() is 0 rank Bessel function of the first kind, f _dBe Doppler frequency shift, T _sBe mark space.

When transmitting terminal can obtain desirable channel condition information without time delay, adopt based on the collaborative piece diagonal angle method for precoding of ZF and can eliminate presence of intercell interference and inter-user interference fully, its pre-coding matrix satisfies following restriction:

H _j[n]T _k[n]＝0,forj＝1,2…Kand j≠k (5)

G _k，m[n]T _k[n]＝0,for1≤m≤M

Wherein in (5) formula, first equation represents that the ZF of disturbing in the residential quarter limits, and second equation represents the ZF restriction of presence of intercell interference.Thereby in the residential quarter, k user's pre-coding matrix should be included in

Kernel in.

May be defined as

Wherein

With Be respectively the polymer matrix of multi-user interference channel and presence of intercell interference channel:

${\tilde{H}}_k\left[n\right]={\left(H_1^T\right[n]\·\·\·H_{k-1}^T[n\left]H_{k+1}^T\right[n]\·\·\·H_K^T[n\left]\right)}^T$ (6)

${\tilde{G}}_k\left[n\right]={\left(G_{k,1}^T\right[n]\·\·\·G_{k,m}^T[n]\·\·\·G_{k,M}^T[n\left]\right)}^T$

Right

Carrying out the standard singular value decomposition can get

${\widetilde{\mathrm{\Π}}}_k\left[n\right]={\tilde{U}}_k\left[n\right]{\widetilde{\mathrm{\Σ}}}_k\left[n\right]{\left({\tilde{V}}_k^{\left(1\right)}\right[n\left]{\tilde{V}}_k^{\left(0\right)}\right[n\left]\right)}^H---\left(7\right)$

Formed

One group of orthonormal basis of right null space.Order

Can satisfy the restriction in (5) formula.In this embodiment, concentrate on n _T〉=(K-1) n _R+ Mn _RSituation to satisfy the restriction of antenna dimension.Can adopt dynamic clustering cooperation and multi-subscriber dispatching algorithm to reduce the value of K and M.When transmitting terminal had the ideal communication channel state information, piece diagonal angle method for precoding can be deleted multi-user interference and presence of intercell interference fully, and user's average transmission rate expression formula is:

$C_k=E\left[{\log }_2\det (I_{n_R}+\frac{P_k}{{\mathrm{\σ}}^2}{H}_k[n\left]T_k\right[n\left]T_k^H\right[n\left]H_k^H\right[n\left]\right)\right]---\left(8\right)$

Consideration is in real system, and transmitting terminal can't obtain desirable channel condition information without time delay, the base station n constantly only can obtain quantize and the time data channel information delayed

And interference channel information

Employing piece diagonal angle pre-coding matrix

The phase mutual interference can't be eliminated fully, only following restriction can be satisfied:

${\hat{H}}_j[n-D_k]{\hat{T}}_k\left[n\right]=0,\mathrm{forj}=\mathrm{1,2}\·\·\·\mathrm{Kandj}\≠k\·$ (9)

Under non-ideal communication channel, user's average transmission rate expression formula is:

${\hat{G}}_k=E\left[\mathrm{lo}{\mathrm{<figure-callout\; id="2"\; label="g"\; filenames="HDA00002790498900011.png"\; state="\{\{state\}\}">g</figure-callout>}}_2\det (I_{n_R}+\frac{P_k}{{\mathrm{\&sigma;}}^2}{H}_k[n\left]{\hat{T}}_k\right[n\left]{\hat{H}}_k^H\right[n\left]R_k^{-1}\right[n\left]\right)\right]---\left(10\right)$

R wherein _k[n] is the covariance matrix of k user's interference plus noise, is expressed as:

$R_k\left[n\right]=P_k{H}_k\left[n\right]\underset{j=1,j\&NotEqual;k}{\overset{K}{\mathrm{\&Sigma;}}}{\hat{T}}_j\left[n\right]{\hat{T}}_j^H\left[n\right]$ (11)

$+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{\&PartialD;}_{k,m}{P}_k{G}_{k,m}\left[n\right]{\hat{T}}_m\left[n\right]{\hat{T}}_m^H\left[n\right]G_{k,m}^H\left[n\right]+{\mathrm{\&sigma;}}^2{I}_{n_R}$

Because non-ideal communication channel information exists, the interference in (10) formula can't be removed fully, namely

With

Thereby residual interference can worsen systematic function.

At first derive when existing based on Limited Feedback and channel delay the closed expression of the difference of user's transmission rate under user's transmission rate and ideal communication channel situation in this embodiment, and proposed a kind of adaptive bit allocation algorithm based on this closed expression.Wherein, the closed expression of the undesirable rate loss that causes of channel condition information is derived as follows:

${\mathrm{\&Delta;C}}_k=C_k-{\hat{C}}_k$

$=E\left[{\log }_2\det (I_{n_R}+\frac{P_k}{{\mathrm{\&sigma;}}^2}{H}_k[n\left]T_k\right[n\left]T_k^H\right[n\left]H_k^H\right[n\left]\right)\right]$

$-E\left[{\log }_2\det \left(R_k\right[n]+\frac{P_k}{{\mathrm{\&sigma;}}^2}{H}_k[n\left]{\hat{T}}_k\right[n\left]{\hat{T}}_k^H\right[n\left]\right)\right]+E\left[{\log }_2\det \left(R_k\right[n\left]\right)\right]$

$\stackrel{\left(a\right)}{\&le;}E\left[{\log }_2\det \left(R_k\right[n\left]\right)\right]$ (12)

$\stackrel{\left(b\right)}{\&le;}{\mathrm{<figure-callout\; id="2"\; label="log"\; filenames="HDA00002790498900011.png"\; state="\{\{state\}\}">log</figure-callout>}}_2\det \left[E\left(R_k\right[n\left]\right)\right]$

$\stackrel{\left(c\right)}{=}{\mathrm{<figure-callout\; id="2"\; label="log"\; filenames="HDA00002790498900011.png"\; state="\{\{state\}\}">log</figure-callout>}}_2\det [\frac{P_k{\mathrm{\&Phi;}}_k{\mathrm{\&eta;}}_k^2(K-1)}{n_T-n_R}{n}_T{I}_{n_R}+(K-1){\mathrm{\&epsiv;}}_k^2{n}_R{I}_{n_R}+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}(\frac{{\&PartialD;}_{k,m}{P}_k{\mathrm{\&Phi;}}_{k,m}{\mathrm{\&eta;}}_{k,m}^2}{n_T-n_R}{n}_T{I}_{n_R}+{\mathrm{\&epsiv;}}_{k,m}^2{n}_R{I}_{n_R})+{\mathrm{\&sigma;}}^2{I}_{n_R}]$

Wherein ignored R in (a) formula _kThe part positive semidefinite item of [n].T _k[n] and

Be isotropic distribution, and and H _k[n] is separate, therefore $H_k\left[n\right]T_k\left[n\right]T_k^H\left[n\right]H_k^H\left[n\right]$ With $H_k\left[n\right]{\hat{H}}_k\left[n\right]{\hat{T}}_k^H\left[n\right]H_k^H\left[n\right]$ Expectation equate.(b) formula is inequality gloomy according to letter.The derivation of equation (c) is as follows:

Have for the S1 item:

$E\left(H_k\right[n\left]\underset{j=1,j\&NotEqual;k}{\overset{K}{\mathrm{\&Sigma;}}}{T}_j\right[n\left]T_j^H\right[n\left]H_k^H\right[n\left]\right)=$

$\stackrel{\left(d\right)}{=}E\left({\mathrm{\&eta;}}_k^2{H}_k\right[n-D_k\left]\underset{j=1,j\&NotEqual;k}{\overset{K}{\mathrm{\&Sigma;}}}{T}_j\right[n\left]T_j^H\right[n\left]H_k^H\right[{n-D}_k\left]\right)+E\left(E_k\right[n\left]\underset{j=1,j\&NotEqual;k}{\overset{K}{\mathrm{\&Sigma;}}}{T}_j\right[n\left]T_j^H\right[n\left]E_k^H\right[n\left]\right)$ (13)

$\stackrel{\left(e\right)}{=}E\left({\mathrm{\&eta;}}_k^2{\tilde{H}}_k\right[{n-D}_k\left]\underset{j=1,j\&NotEqual;k}{\overset{K}{\mathrm{\&Sigma;}}}{T}_j\right[n\left]T_j^H\right[n\left]{\tilde{H}}_k^H\right[{n-D}_k\left]{\mathrm{\&Lambda;}}_K\right)+(K-1){\mathrm{\&epsiv;}}_k^2{n}_R{I}_{n_R}$

$\stackrel{\left(f\right)}{=}\frac{{\mathrm{\&Phi;}}_k{\mathrm{\&eta;}}_k^2(K-1)}{n_T-n_R}{n}_T{I}_{n_R}+(K-1){\mathrm{\&epsiv;}}_k^2{n}_R{I}_{n_R}$

Wherein (d) formula is according to equation (3) and E _k[n] and H _k[nD _k] independent distribution.(e) formula is according to H _k[n-D _k] the Eigenvalues Decomposition expression formula: $H_k^H[n-D_k]H_k\left[{n-D}_k\right]={\tilde{H}}_k^H\left[{n-D}_k\right]{\mathrm{\&Lambda;}}_k{\tilde{H}}_k\left[{n-D}_k\right],$ Wherein ${\tilde{H}}_k\left[{n-D}_k\right]$ Formed H _k[nD _k] the column vector subspace of opening in one group of orthonormal basis.Λ _kBe diagonal matrix, its diagonal entry is

N _RIndividual non-zero, the characteristic value of descending.(e) formula is according to E _k[n] is multiple Gauss's matrix and T _j[n] is unitary matrice, thereby E _k[n] T _j[n] also obeys for each element

The Gauss's matrix that distributes.Matrix

Can decompose as follows:

${\tilde{H}}_k\left[{n-D}_k\right]=A_k{B}_k{\hat{H}}_k\left[{n-D}_k\right]+S_k{C}_k---\left(14\right)$

Wherein

Be the orthonormal basis of one group of isotropic distribution, open into

Left kernel in n _RDimensional plane.C _kN _R* n _RUpper triangular matrix, its diagonal entry are on the occasion of also satisfying: $\mathrm{tr}\left(C_k{C}_k^H\right)=d^2\left({\tilde{H}}_k\right[n],{\hat{H}}_k[n\left]\right).$ And, S _kAnd C _kSeparate.According to (14) Shi Kede:

$E\left({\mathrm{\&eta;}}_k^2{\tilde{H}}_k\right[{n-D}_k\left]\underset{j=1,j\&NotEqual;k}{\overset{K}{\mathrm{\&Sigma;}}}{T}_j\right[n\left]T_j^H\right[n\left]{\tilde{H}}_k^H\right[{n-D}_k\left]{\mathrm{\&Lambda;}}_K\right)$

$={\mathrm{\&eta;}}_k^2(K-1)E\left[C_k\left(S_k{T}_j\right[n\left]T_j^H\right[n\left]S_k^H\right)C_k^H{\mathrm{\&Lambda;}}_K\right]---\left(15\right)$

$={\mathrm{\&eta;}}_k^2(K-1)\frac{n_T}{n_T-n_R}E\left(C_k{C}_k^H{\mathrm{\&Lambda;}}_K\right)$

(15) in formula last equation according to S _kAnd T _j[n] isotropic distribution exists

Left kernel in and both are separate, thereby For matrix immediately variable to obey parameter be n _RAnd n _T-2n _RBeta distribute, namely $E\left[S_k{T}_j\right[n\left]T_j^H\right[n\left]S_k^H\right]=\frac{n_R}{n_T-n_R}.$ Character according to (14) formula matrix decomposition has $E\left({\mathrm{\&Lambda;}}_k\right)=n_T{I}_{n_R},$

Φ wherein _kFor quantization error can be approximately equal to:

${\mathrm{\&Phi;}}_k\&ap;\frac{\mathrm{\&Gamma;}\left(\frac{1}{T}\right)}{T}{\left(C_{n_T}{n}_R\right)}^{-\frac{1}{T}}{2}^{-\frac{B_k}{T}}---\left(16\right)$

T=n wherein _R(n _T-n _R), $C_{n_T{n}_R}=\frac{1}{T!}\underset{i=1}{\overset{n_R}{\mathrm{\&Pi;}}}\frac{(n_T-i)!}{(n_R-i)!},$ Thereby obtain (f) formula.

Derivation and the S1 of S2 are similar, as follows:

$E\left(G_{k,m}\right[n\left]\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{\hat{T}}_m\right[n\left]{\hat{T}}_m^H\right[n\left]G_{k,m}^H\right[n\left]\right)=$

$E\left(\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{\mathrm{\&eta;}}_{k,m}^2{G}_k\right[{n-D}_{k,m}\left]{\hat{T}}_m\right[n\left]{\hat{T}}_m^H\right[n\left]G_{k,m}^H\right[{n-D}_{k,m}]+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{E}_{k,m}[n\left]{\hat{T}}_m\right[n\left]{\hat{T}}_m^H\right[n\left]E_{k,m}^H\right[n\left]\right)$ (17)

$=E\left(\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{\mathrm{\&eta;}}_{k,m}^2{\tilde{G}}_k\right[{n-D}_{k,m}\left]{\hat{T}}_m\right[n\left]{\hat{T}}_m^H\right[n\left]{\tilde{G}}_{k,m}^H\right[{n-D}_{k,m}\left]{\mathrm{\&Lambda;}}_{k,m}\right)+M{\mathrm{\&epsiv;}}_{k,m}^2{n}_R{I}_{n_R}$

$=\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&Phi;}}_{k,m}{\mathrm{\&eta;}}_{k,m}^2}{n_T-n_R}{n}_T{I}_{n_R}+{\mathrm{\&epsiv;}}_{k,m}^2{n}_R{I}_{n_R}$

Wherein

Form G _k[n-D _k,m] column vector open into one group of orthonormal basis in space, Λ _{K, m}Be diagonal matrix, its diagonal entry is N _RIndividual non-zero, the characteristic value of descending.Φ _{K, m}Be quantization error:

${\mathrm{\&Phi;}}_{k,m}\&ap;\frac{\mathrm{\&Gamma;}\left(\frac{1}{T}\right)}{T}{\left(C_{n_T{n}_R}\right)}^{-\frac{1}{T}}{2}^{-\frac{B_{k,m}}{T}}---\left(18\right)$

The closed expression that finally obtains the rate loss that Limited Feedback and channel delay cause is:

${\mathrm{\&Delta;C}}_k\&le;n_R\mathrm{lo}{g}_2(\frac{P_k{\mathrm{\&Phi;}}_k{\mathrm{\&eta;}}_k^2(K-1)}{n_T-n_R}{n}_T+(K-1){\mathrm{\&epsiv;}}_k^2{n}_R+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}\frac{{\&PartialD;}_{k,m}{P}_k{\mathrm{\&Phi;}}_{k,m}{\mathrm{\&eta;}}_{k,m}^2}{n_T-n_R}{n}_T+{\mathrm{\&epsiv;}}_{k,m}^2{n}_R+{\mathrm{\&sigma;}}^2)---\left(19\right)$

By (19) formula as can be known, rate loss causes mainly due to two factors: residual multi-user interference and presence of intercell interference.A key issue is the total number of bits of feedback B as each user _totIn the time of fixedly, in order to minimize rate loss, should be how many number of bits of feedback B of data channel assignment _kAnd should be interference channel and distribute how many bit number B _{K, m}Next this embodiment will propose a kind of novel optimal bit allocative decision, to minimize the rate loss of (19) formula.

According to (19) formula, notice that logarithmic function is monotonic increasing function, ignore constant term, can get target function as follows:

$\min {\stackrel{\&OverBar;}{P}}_k{2}^{-\frac{B_k}{T}}+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{P}}_{k,m}{2}^{-\frac{B_{k,m}}{T}}$ (20)

$s\&CenterDot;t\&CenterDot;B_k+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{B}_{k,m}=B_{\mathrm{tot}}$

Wherein, ${\stackrel{\&OverBar;}{P}}_k\frac{P_k{\mathrm{\&eta;}}_k^2{n}_T\mathrm{\&Gamma;}\left(\frac{1}{T}\right){\left(C_{n_T{n}_R}\right)}^{-\frac{1}{T}}(K-1)}{T(n_T-n_R)},$ ${\stackrel{\&OverBar;}{P}}_{k,m}=\frac{{\&PartialD;}_{k,m}{P}_k{\mathrm{\&eta;}}_{k,m}^2{n}_T\mathrm{\&Gamma;}\left(\frac{1}{T}\right){\left(C_{n_T{n}_R}\right)}^{-\frac{1}{T}}}{T(n_T-n_R)}.$

(20) formula is convex function, available protruding Optimization Method.Lagrangian can be written as:

$L(B_k,B_{k,m},\mathrm{\&lambda;})={\stackrel{\&OverBar;}{P}}_k{2}^{-\frac{B_k}{T}}+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{P}}_{k,m}{2}^{-\frac{B_{k,m}}{T}}+\mathrm{\&lambda;}(B_k+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{B}_{k,m}-B_{\mathrm{tot}})---\left(21\right)$

According to (21) formula, can get the optimum KKT condition of single order as follows:

$\frac{\&PartialD;L(B_k,B_{k,m},\mathrm{\&lambda;})}{{\&PartialD;B}_k}=\frac{-\ln \left(2\right){\stackrel{\&OverBar;}{P}}_k}{T}{2}^{-\frac{B_k}{T}}+\mathrm{\&lambda;}=0$

$\frac{\&PartialD;L(B_k,B_{k,m},\mathrm{\&lambda;})}{{\&PartialD;B}_{k,m}}=\frac{-\ln \left(2\right){\stackrel{\&OverBar;}{P}}_{k,m}}{T}{2}^{-\frac{B_{k,m}}{T}}+\mathrm{\&lambda;}=0---\left(22\right)$

$\frac{\&PartialD;L(B_k,B_{k,m},\mathrm{\&lambda;})}{\&PartialD;\mathrm{\&lambda;}}=B_k+\underset{m=1}{\overset{M}{\mathrm{\&Sigma;}}}{B}_{k,m}-B_{\mathrm{tot}}=0$

Find the solution the equation in (22) formula, can get optimal solution as follows:

$B_k^{\mathrm{opt}}={[\frac{B_{\mathrm{tot}}}{M+1}+\mathrm{Tlo}{g}_2\left(\frac{{\stackrel{\&OverBar;}{P}}_k}{{\left({\stackrel{\&OverBar;}{P}}_k\underset{m=1}{\overset{M}{\mathrm{\&Pi;}}}{\stackrel{\&OverBar;}{P}}_{k,m}\right)}^{\frac{1}{M+1}}}\right)]}^{+},B_{k,m}^{\mathrm{opt}}{[\frac{B_{\mathrm{tot}}}{M+1}+{T\log }_2\left(\frac{{\stackrel{\&OverBar;}{P}}_{k,m}}{{\left({\stackrel{\&OverBar;}{P}}_k\underset{m=1}{\overset{M}{\mathrm{\&Pi;}}}{\stackrel{\&OverBar;}{P}}_{k,m}\right)}^{\frac{1}{M+1}}}\right)]}^{+}---\left(23\right)$

Wherein

By (23) formula as can be known Along with

Increase and increase.Also namely or data channel time delay strong when the power of data-signal hour, system can distribute more number of bits of feedback to come the quantized data channel.On the contrary, when interference channel stronger, time delay hour, more number of bits of feedback can be distributed to the quantification interference channel.In other words, this embodiment is put forward the algorithm self adaptation must be stronger for power, the feedback bits that the channel allocation that time delay is less is more; For power a little less than, the feedback bits that the channel allocation that time delay is larger is less is to minimize rate loss.

Self-adaptive bit allocation method in the collaborative multiaerial system that this embodiment is provided below by specific embodiment is elaborated.In multiaerial system block diagram shown in Figure 2, each user not only receives the data channel of this residential quarter, can be subject to the interference of adjacent cell signal simultaneously, the user utilizes channel estimating to obtain the channel condition information of data channel and interference channel, and will feed back to serving BS after these information quantizations.Serving BS passes to the interfered cell base station by X2 interface with interference channel information again.

The base station is the broadcast channel periodically.The user utilizes channel estimating to estimate the channel information of this cell data channel and adjacent cell interference channel and the relevant information of signal power and time delay.The user utilizes the information of signal power and time delay to calculate respectively the Optimal Feedback bit number of data channel and the distribution of each interference channel according to (22) formula.Determine the Optimal Feedback bit number, also namely determined the size of code book, as W _kAnd W _k,mSize be respectively

With

The user selects corresponding code word as the channel matrix after quantizing from code book respectively according to the minimum chordal distance criterion

With

And the code book indication information is fed back to serving BS by the feedback channel that time delay is arranged.The base station by X2 interface with interference channel information Pass to corresponding adjacent cell base station.The base station utilize the non-ideal communication channel information obtain carry out the ZF pre-encode operation with suppress in this residential quarter multi-user interference and to the presence of intercell interference of adjacent cell.

For realizing said method, the embodiment of the present invention has also proposed the adaptive bit distributor in a kind of collaborative multiaerial system, as shown in Figure 3, comprising:

Receiving element 31 is used for extracting the received power of this cell data channel information, adjacent cell interference channel information, data-signal and interference signal and the Delay of each channel according to channel estimating;

Feedback computing unit 32, calculate respectively the Optimal Feedback bit number that distributes for data channel and each interference channel under the prerequisite that total number of bits of feedback is fixed for the Delay according to described signal received power and each channel, and the rate loss that the channel information nonideality causes is carried out minimization, to obtain ideal communication channel information;

Quantize computing unit 33, be used for determining according to described Optimal Feedback bit number the size of code book, and utilize described ideal communication channel information to choose channel condition information after corresponding code word quantizes according to the minimum chordal distance criterion from described code book, and the index information that described channel condition information is corresponding feeds back to serving BS by the based on feedback link that time delay is arranged;

Base station precoding unit 34 is used for receiving by the X2 interface that time delay is arranged the channel condition information of described quantification, and generates the ZF pre-coding matrix according to the interference channel information of described quantification.

Preferably, comprise at feedback computing unit 32;

The minimization subelement is used to the more feedback bits of channel allocation that signal power is strong, time delay is less; Be the less feedback bits of channel allocation that signal power is weak, time delay is larger.

Preferably, comprise in base station precoding unit 4;

Disturb to suppress subelement, be used for utilizing singular value decomposition to obtain the kernel matrix of interference channel, and utilize this kernel matrix as pre-coding matrix to suppress this two kinds of interference.

The embodiment of the processing capacity of each unit that comprises in the adaptive bit distributor in above-mentioned collaborative multiaerial system is described in method execution mode before, no longer is repeated in this description at this.

Adopt the scheme of the present embodiment, utilize the non-ideal communication channel information of obtaining to carry out the ZF pre-encode operation by the base station, thus suppress in this residential quarter multi-user interference and to the presence of intercell interference of adjacent cell.

The above; only for the better embodiment of the present invention, but protection scope of the present invention is not limited to this, anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; the variation that can expect easily or replacement are within all should being encompassed in protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range of claim.

Claims (6)

1. an adaptive bit allocation method in a cooperative multi-antenna system, is characterized in that, comprises the following steps: Extract the data channel information of the cell, the interference channel information of the adjacent cell, the received power of the data signal and the interference signal, and the delay information of each channel according to the channel estimation; According to the received signal power and the delay information of each channel, calculate the optimal number of feedback bits allocated for the data channel and each interference channel under the premise that the total number of feedback bits is fixed, and calculate the rate loss caused by the non-ideal state of channel information Minimize to obtain ideal channel information; Determine the size of the codebook according to the optimal number of feedback bits, and use the ideal channel information to select the corresponding codeword quantized channel state information from the codebook according to the minimum chordal distance criterion, and store the channel state The information is fed back to the serving base station through a time-delayed feedback link; The quantized channel state information is received through a time-delayed X2 interface, and a zero-forcing precoding matrix is generated according to the quantized interference channel information. 2. The method according to claim 1, wherein the minimizing the rate loss caused by the non-ideal state of channel information comprises: Allocate more feedback bits to channels with stronger signal power and smaller time delay; allocate less feedback bits to channels with weaker signal power and longer time delay. 3. The method according to claim 1, wherein said generating a zero-forcing precoding matrix according to said quantized interference channel information comprises: The null space matrix of the interfering channel is obtained by singular value decomposition, and the null space matrix is used as a precoding matrix to suppress the two kinds of interference. 4. An adaptive bit allocation device in a cooperative multi-antenna system, characterized in that it comprises: The receiving unit is used to extract the data channel information of the current cell, the interference channel information of the adjacent cell, the received power of the data signal and the interference signal, and the delay information of each channel according to the channel estimation; The feedback calculation unit is used to calculate the optimal number of feedback bits allocated for the data channel and each interference channel under the premise that the total number of feedback bits is fixed according to the received signal power and the time delay information of each channel, and the channel information Minimize the rate loss caused by the ideal state to obtain ideal channel information; a quantization calculation unit, configured to determine the size of the codebook according to the optimal number of feedback bits, and use the ideal channel information to select the corresponding codeword quantized channel state information from the codebook according to the minimum chordal distance criterion, and feeding back index information corresponding to the channel state information to the serving base station through a time-delayed feedback link; The base station precoding unit is configured to receive the quantized channel state information through a time-delayed X2 interface, and generate a zero-forcing precoding matrix according to the quantized interference channel information. 5. The device according to claim 4, characterized in that, the feedback calculation unit includes; The minimization processing subunit is configured to allocate more feedback bits to channels with strong signal power and small time delay; allocate less feedback bits to channels with weak signal power and large time delay. 6. The device according to claim 4, wherein the base station precoding unit includes; The interference suppression subunit is used to obtain the null space matrix of the interference channel by using singular value decomposition, and use the null space matrix as a precoding matrix to suppress the two types of interference.

Images