CN104702543A - Precoding method and device - Google Patents

Abstract

The embodiment of the invention provides a precoding method comprising the following steps: a pilot signal transmitted by a transmitting end is received; channel state information estimation is performed according to the pilot signal to acquire a channel mean and a channel covariance; a precoding matrix is calculated according to the channel mean and the channel covariance; and the transmitting end precodes a data signal needing to be transmitted according to the precoding matrix, and the receiving end de-precodes a received precoded signal according to the precoding matrix. Better MIMO precoding reliability performance and higher MIMO system capacity can be obtained for MIMO-OFDM precoding based on the average signal to interference and noise ratio criterion under the condition of channel state information statistics, and reliability performance and MIMO system capacity better than those of codebook-based precoding can be obtained in a low-correlation environment.

Description

Method for precoding and device

Technical field

The present invention relates to moving communicating field, be specifically related to a kind of method for precoding and device.

Background technology

Multiple-input and multiple-output (MIMO) technology can improve system space throughput under the condition not increasing system bandwidth, also significantly can improve signal transmission quality by means of to effective utilization of space diversity.In mimo systems, precoding technique utilizes channel condition information (channel statementinformation at transmitting terminal, CSI) preliminary treatment is carried out to transmission signal and eliminate intersymbol interference (ISI) or inter-user interference to reach, improve the signal processing technology for the purpose of power system capacity.Common precoding type mainly comprises linear predictive coding and nonlinear precoding.Linear predictive coding mainly contains based on code book and non-code book two kinds of linear predictive codings.Precoding based on code book is exactly share a known codebook set in sending and receiving end, code book is concentrated and is comprised multiple pre-coding matrix, receiving terminal concentrates the pre-coding matrix selecting to make systematic function optimum at code book with a certain performance index according to the channel matrix of channel estimating, again its code book sequence number is fed back to transmitting terminal, transmitting terminal selects pre-coding matrix to carry out precoding processing according to associated sequence numbers.Adopt and comprise based on the problem existing for the precoding of code book, the design of code book should be match channels characteristic, and in actual applications, system may be operated in different fading environment, it wire spacing is also not quite similar, antenna pattern and polarization are also not quite similar, and this makes the design of code book may become very complicated; In addition, demand fulfillment constant modulus property, limited character and Nested property etc. are gone back in the design of code book usually.These factors affect to a great extent and the precoding technique scheme that govern based on code book meets the different requirements of real system more neatly.Compared with the MIMO precoding based on code book, non-code book MIMO precoding can according to the dynamic change of channel condition, selects the pre-coding matrix matched with the characteristic of channel neatly, MIMO precoding requirement under changing environment when can meet complexity better.

Existing MIMO precoding is difficult to take into account under middle Low SNR these two large performance index of mimo system capacity that precoding reliability and system realize simultaneously.

Summary of the invention

The object of this invention is to provide a kind of method for precoding, to realize the method for precoding based on average Signal to Interference plus Noise Ratio criterion, thus under low relevant environment, obtain the unfailing performance and mimo system capacity that are better than based on codebook precoding.

For achieving the above object, the embodiment of the present invention provides a kind of method for precoding on the one hand, and described method comprises:

The uplink/downlink pilot signal that receiving end/sending end sends;

Channel condition information estimation is carried out, to obtain channel average and channel covariancc according to described uplink/downlink pilot signal;

According to described channel average and channel covariancc, calculate pre-coding matrix;

When row channel does not meet symmetry over/under, give described transmitting terminal by described precoding matrix feedback, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix;

According to described pre-coding matrix, solution precoding is carried out to the data-signal received.

Based on first aspect, in the execution mode that the first is possible, described according to uplink/downlink pilot signal carry out channel condition information estimation, to obtain channel average and channel covariancc, comprise further:

The channel status pilot signal received being carried out to pilot frequency symbol position is estimated, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel;

To described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with obtain data to be estimated channel estimation results on sub-carriers;

According to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers;

Obtain the statistical channel state information in correlation time according to described Noise Variance Estimation value, described statistical channel state information comprises described channel average and described channel covariancc.

Based on the execution mode that the first is possible, in the execution mode that the second is possible, the channel status that the described pilot signal to receiving carries out pilot frequency symbol position is estimated, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel, specifically comprises:

By the N in a subframe _pindividual frequency pilot sign Y _k,p(i _m)=X _k,p(i _m) H _k,p(i _m)+W _k,p(i _m), 0≤m≤N _p-1

Be integrated into: Y _k,p=X _k,ph _k,p+ W _k,p

Wherein, k represents a kth subcarrier; i _mthe location index existed in the OFDM symbol of pilot signal, N _pit is the frequency pilot sign number on a subcarrier; for the frequency pilot sign received, for the pilot tone that transmitting terminal sends; N _rfor reception antenna number, N _tfor number of transmit antennas; it is corresponding Frequency domain noise; H _k,p(i _m) represent the mimo channel frequency domain response on a kth subcarrier, and have

$H_{k,p}\left(i_m\right)={\left[\begin{array}{cccc}{h}_{k,p}^{\mathrm{1,1}}\left(i_m\right)& h_k^{\mathrm{1,2}}\left(i_m\right)& ...& h_k^{N_t,1}\left(i_m\right)\\ h_{k,p}^{\mathrm{2,1}}\left(i_m\right)& h_k^{\mathrm{2,2}}\left(i_m\right)& ...& h_k^{N_t,2}\left(i_m\right)\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ h_{k,p}^{N_t,1}\left(i_m\right)& h_k^{N_t,2}\left(i_m\right)& ...& h_k^{N_t,N_r}\left(i_m\right)\end{array}\right]}_{N_t\×N_r}$

At described Y _k,p=X _k,ph _k,p+ W _k,pin,

$Y_{k,p}={\left[\begin{array}{c}{Y}_{k,p}\left(i_0\right)\\ Y_{k,p}\left(i_1\right)\\ ...\\ Y_{k,p}\left(i_{N_p-1}\right)\end{array}\right]}_{N_p\×N_r},W_{k,p}={\left[\begin{array}{c}{W}_{k,p}\left(i_0\right)\\ W_{k,p}\left(i_1\right)\\ ...\\ W_{k,p}\left(i_{N_p-1}\right)\end{array}\right]}_{N_p\×N_r}$

According to described Y _k,p=X _k,ph _k,p+ W _k,p, obtain the frequency domain response that a mimo channel kth subcarrier estimates wherein, for X _k,ppseudoinverse;

According to described obtain the reception pilot matrix after the cascade in coherence time:

${\tilde{Y}}_{k,p}={\tilde{X}}_{k,p}{H}_{k,p}\left(i_0\right)+{\tilde{W}}_{k,p}$

Wherein, H _k,p(i _m) ≈ H _k,p(i ₀), m=1 ..., N _c, N _c>N _pfor the OFDM symbol number of equal value that the correlated time of channel is corresponding, wherein:

${\tilde{Y}}_{k,p}={\left[\begin{array}{c}{Y}_{k,p}\left(i_0\right)\\ Y_{k,p}\left(i_1\right)\\ ...\\ Y_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\×N_r},{\tilde{W}}_{k,p}={\left[\begin{array}{c}{W}_{k,p}\left(i_0\right)\\ W_{k,p}\left(i_1\right)\\ ...\\ W_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\×N_r},{\tilde{X}}_{k,p}={\left[\begin{array}{c}{X}_{k,p}\left(i_0\right)\\ X_{k,p}\left(i_1\right)\\ ...\\ X_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\×N_t}$

Reception pilot matrix according to after the cascade in described coherence time:

The LS channel estimation results obtained based on frequency pilot sign within the correlated time of channel is:

wherein, for pseudoinverse.

Based on the execution mode that the second is possible, in the execution mode that the third is possible, described to described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with the channel estimation results of data estimator symbol position, comprise further:

In time domain to the concrete steps of the channel estimating of described data symbol position to be estimated be:

H _k,data(l′)＝α ₁×H _k,p,LS(l)+β ₁×H _k,p,LS(l+Δl)

Wherein l ' represents the time Domain Index of Data Position to be estimated, and have l < l ' < l+ Δ l, l represents pilot signal position in the time domain, Δ l represents the time domain interval of two pilot signals that time domain is adjacent, k represents pilot signal place subcarrier sequence number, interpolation coefficient α ₁, β ₁calculate by the following method

${\mathrm{\&alpha;}}_1=\frac{l+\mathrm{\&Delta;l}-l^{\&prime;}}{\mathrm{\&Delta;l}},{\mathrm{\&beta;}}_1=\frac{l^{\&prime;}-l}{\mathrm{\&Delta;l}};$ And

To the step of the channel condition information on described Data Position frequency domain index to be estimated in frequency domain, be specially:

H _k′(l)＝α ₂×H _k,S(l)+β ₂×H _k+Δk,S(l),S∈[p,data]；

Wherein k ' represents the frequency domain index of Data Position to be estimated, and have k < k ' < k+ Δ k, k represents pilot signal position in a frequency domain, and Δ k represents the frequency interval of two pilot signals that frequency domain is adjacent, interpolation coefficient α ₂, β ₂calculate by the following method

${\mathrm{\&alpha;}}_2=\frac{k+\mathrm{\&Delta;k}-k^{\&prime;}}{\mathrm{\&Delta;k}},{\mathrm{\&beta;}}_2=\frac{k^{\&prime;}-k}{\mathrm{\&Delta;k}}.$

Based on the execution mode that the third is possible, in the 4th kind of possible execution mode, described according to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers, be specially:

${\mathrm{\&sigma;}}_{w,k}^2=\frac{1}{N_c-1}\underset{m=0}{\overset{N_c-2}{\mathrm{\&Sigma;}}}{\left|\right|H_{k,p,\mathrm{LS}}\left(i_{m+1}\right)-H_{k,p,\mathrm{LS}}\left(i_m\right)\left|\right|}_F^2;$

Wherein || H _{k, p, LS}(i _m+1)-H _{k, p, LS}(i _m) || _frepresent Frobenius norm, i _m+1, i _mrepresent adjacent pilot frequencies subcarrier sequence number on same Pilot OFDM symbols, final Noise Variance Estimation value can be expressed from the next:

${\mathrm{\&sigma;}}_w^2=\frac{1}{M}\underset{k=1}{\overset{M}{\mathrm{\&Sigma;}}}{\mathrm{\&sigma;}}_{w,k}^2$

The wherein total number of sub-carriers of M shared by frequency pilot sign.

Based on the 4th kind of possible execution mode, in the 5th kind of possible execution mode, the described statistical channel state information obtained according to described Noise Variance Estimation value in correlation time, described statistical channel state information comprises described channel average and described channel covariancc, specifically comprises:

According to the correlation time of communication channel, select the sliding window length L of time domain, wherein said sliding window length corresponds to the number of pilot symbols that in sliding window, subframe comprises, according to the correlation bandwidth of channel, at the M that frequency domain comprises from each coherence bandwidth ₀the channel estimation results of individual pilot sub-carrier calculates moment i's the average of individual channel condition information:

And covariance:

Wherein vec () represents matrix vector, () ^hrepresent conjugate transpose.

Based on the 5th kind of possible execution mode, in the 6th kind of possible execution mode, described according to described channel average and channel covariancc, calculate pre-coding matrix, comprise further:

According to the described system channel average estimating to obtain channel covariancc and noise variance the impact of noise is added channel average with channel covariancc Φ _h, obtain new channel average and channel covariancc wherein:

${\stackrel{\&OverBar;}{H}}_{m,n}^{\left(s\right)}=R_n^{-1/2}{\stackrel{\&OverBar;}{H}}_m^{\left(s\right)};$

${\mathrm{\&Phi;}}_{H,n}^{\left(s\right)}=(I_{N_r}\&CircleTimes;R_n^{-1/2}){\mathrm{\&Phi;}}_H^{\left(s\right)}(I_{N_r}\&CircleTimes;R_n^{-1/2});$

Wherein noise covariance matrix

According to and Φ _h,n, compute matrix wherein

$Z_{[i,j]}^{\left(s\right)}=\mathrm{Tr}\left(X_{[i,j]}^{\left(s\right)}\right),x^{\left(s\right)}=({\mathrm{\&Phi;}}_H^{\left(s\right)}+(1/N_r){\left(\right({\stackrel{\&OverBar;}{H}}_m^{\left(s\right)})}^H{\stackrel{\&OverBar;}{H}}_m^{\left(s\right)}\&CircleTimes;I_{N_r})),i=1,...,N_t,j=1,...,N_t;$

Wherein X _{[i, j]}an i-th j dimension of expression X is N _r× N _rsub-square;

By matrix Z ^(s)carry out Cholesky decomposition, obtain matrix L ^(s):

Z ^(s)＝L ^(s)·(L ^(s)) ^H；

To L ^(s)matrix carries out singular value decomposition:

$L^{\left(s\right)}=U_L^{\left(s\right)}{\mathrm{\&Lambda;}}_L^{\left(s\right)}{\left(V_L^{\left(s\right)}\right)}^H;$

Obtain Z ^(s)eigenvalues Decomposition expression formula

$Z^{\left(s\right)}=U_Z^{\left(s\right)}{\mathrm{\&Lambda;}}_Z^{\left(s\right)}{\left(V_Z^{\left(s\right)}\right)}^H=U_L^{\left(s\right)}{\left({\mathrm{\&Lambda;}}_L^{\left(s\right)}\right)}^2{\left(U_L^{\left(s\right)}\right)}^H;$

Pre-coding matrix based on average Signal to Interference plus Noise Ratio criterion calculates by the following method

$B^{\left(s\right)}=U_Z^{\left(s\right)}{\mathrm{\&Lambda;}}_Q^{\left(s\right)};$

Wherein $U_Z^{\left(s\right)}=U_L^{\left(s\right)},{\mathrm{\&Lambda;}}_Q^{\left(s\right)}={\mathrm{\&Lambda;}}_L^{\left(s\right)};$

Wherein $P_i=\max (\mathrm{\&mu;}-\frac{{\mathrm{\&sigma;}}_w^2}{{\mathrm{\&lambda;}}_i},0),i=\mathrm{1,2},...,r,\underset{i=1}{\overset{r}{\mathrm{\&Sigma;}}}{P}_i=\frac{P}{N},({\mathrm{\&lambda;}}_1,{\mathrm{\&lambda;}}_2,...{\mathrm{\&lambda;}}_r)$ It is matrix r nonzero eigenvalue, μ is the water filling factor.P is that transmitting terminal carries all subcarriers of user data and the summation of transmitting antenna transmitted power.

Based on first aspect, in the 7th kind of possible execution mode, the signal that after described employing precoding, receiving terminal receives is:

Y _k(l)＝H _k(l)·B·X _k(l)+W _k(l)

Wherein $Y_k\left(l\right)={\tilde{Y}}_k^T\left(l\right),X_k\left(l\right)={\tilde{X}}_k^T\left(l\right),H_k\left(l\right)=H_k^T\left(l\right),W_k\left(l\right)={\tilde{W}}_k^T\left(l\right);$

The channel condition information H obtaining non-pilot character position is estimated by time domain and frequency domain two-dimensional linear interpolation _kl (), in conjunction with described pre-coding matrix B, determines the channel matrix H of equivalence _e,k(l), namely

H _E,k(l)＝Β·H _k(l)

According to the combined channel matrix H of equivalence _e,kl () take least mean-square error as criterion, receiving terminal can carry out separating the joint-detection that precoding and MIMO detect and receive

${\hat{X}}_k\left(l\right)=G_{\mathrm{MMSE}}\&CenterDot;Y_k\left(l\right)={[H_{E,k}^H\left(l\right)H_{E,k}\left(l\right)+{\mathrm{\&sigma;}}_n^2{I}_{N_t}]}^{-1}{H}_{E,k}^H\left(l\right)Y_k\left(l\right)$

Second aspect, embodiments provides a kind of pre-coding apparatus, and described device comprises:

Receiving element, for the pilot signal that receiving end/sending end sends;

Estimation unit, for carrying out channel condition information estimation according to uplink and downlink pilot signal, to obtain channel average and channel covariancc;

Computing unit, for according to described channel average and channel covariancc, calculates pre-coding matrix;

Feedback unit, for giving described transmitting terminal when uplink and downlink channel does not meet symmetry by described precoding matrix feedback, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix.

Decoding unit, for according to described pre-coding matrix, carries out solution precoding to the data-signal received.

Based on second aspect, in the execution mode that the first is possible, described estimation unit specifically for:

The channel status pilot signal received being carried out to pilot frequency symbol position is estimated, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel;

To described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with obtain data to be estimated channel estimation results on sub-carriers;

According to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers;

Obtain the statistical channel state information in correlation time according to described Noise Variance Estimation value, described statistical channel state information comprises described channel average and described channel covariancc.

By the method for precoding that the embodiment of the present invention provides, the pilot signal sent by receiving end/sending end; Channel condition information estimation is carried out, to obtain channel average and channel covariancc according to described pilot signal; According to described channel average and channel covariancc, calculate pre-coding matrix; Described transmitting terminal is given, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix by described precoding matrix feedback.The MIMO-OFDM precoding that can be implemented in based on average Signal to Interference plus Noise Ratio criterion under statistical channel state information condition can process MIMO precoding unfailing performance and mimo system capacity better, under low relevant environment, obtain the unfailing performance be better than based on codebook precoding and mimo system capacity.

Accompanying drawing explanation

In order to be illustrated more clearly in the technical scheme in the embodiment of the present invention, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Fig. 1 implements schematic diagram based on the MIMO precoding technique of average Signal to Interference plus Noise Ratio criterion under TD-LTE system condition.

Fig. 2 be under TD-LTE system condition receiving terminal based on average Signal to Interference plus Noise Ratio criterion MIMO precoding receive process schematic block diagram.

Fig. 3 is the flow chart that the embodiment of the present invention provides method for precoding;

Fig. 4 be under TD-LTE system condition transmitting terminal based on the MIMO precoding transmission processing schematic block diagram of average Signal to Interference plus Noise Ratio criterion;

Fig. 5 is the unfailing performance schematic diagram based on the MIMO precoding of average Signal to Interference plus Noise Ratio criterion under high relevant ETU1 channel condition;

Fig. 6 is the MIMO pre-coding system capacity schematic diagram based on average Signal to Interference plus Noise Ratio criterion under high relevant ETU1 channel condition;

Fig. 7 is the unfailing performance schematic diagram based on the MIMO precoding of average Signal to Interference plus Noise Ratio criterion under low relevant ETU1 channel condition;

Fig. 8 is the MIMO pre-coding system capacity schematic diagram based on average Signal to Interference plus Noise Ratio criterion under low relevant ETU1 channel condition;

Fig. 9 is the MIMO precoding unfailing performance schematic diagram based on average Signal to Interference plus Noise Ratio criterion under high relevant EVA1 channel condition;

Figure 10 is the MIMO pre-coding system capacity schematic diagram based on average Signal to Interference plus Noise Ratio criterion under high relevant EVA1 channel condition;

Figure 11 is the MIMO precoding unfailing performance schematic diagram based on average Signal to Interference plus Noise Ratio criterion under low relevant EVA1 channel condition;

Figure 12 is the MIMO pre-coding system capacity schematic diagram based on average Signal to Interference plus Noise Ratio criterion under low relevant EVA1 channel condition;

Figure 13 is the MIMO precoding unfailing performance schematic diagram based on average Signal to Interference plus Noise Ratio criterion under high relevant EPA channel condition;

Figure 14 is the MIMO pre-coding system capacity schematic diagram based on average Signal to Interference plus Noise Ratio criterion under high relevant EPA channel condition;

Figure 15 is the MIMO precoding unfailing performance schematic diagram based on average Signal to Interference plus Noise Ratio criterion under low relevant EPA channel condition;

Figure 16 is the MIMO pre-coding system capacity schematic diagram based on average Signal to Interference plus Noise Ratio criterion under low relevant EPA channel condition;

The structural representation of the pre-coding apparatus that Figure 17 provides for the embodiment of the present invention;

The another kind of structural representation of the pre-coding apparatus that Figure 18 provides for the embodiment of the present invention.

Embodiment

Below by drawings and Examples, technical scheme of the present invention is described in further detail.

The technical scheme that the embodiment of the present invention provides can be applied to various cordless communication network, such as: global mobile communication (global system for mobile communication, referred to as GSM) system, code division multiple access (code division multiple access, referred to as CDMA) system, Wideband Code Division Multiple Access (WCDMA) (wideband code division multiple access, referred to as WCDMA) system, universal mobile communications (universal mobile telecommunication system, referred to as UMTS) system, GPRS (general packet radio service, referred to as GPRS) system, Long Term Evolution (long term evolution, referred to as LTE) system, advanced Long Term Evolution (long term evolutionadvanced, referred to as LTE-A) system, global interconnection inserting of microwave (worldwide interoperabilityfor microwave access, referred to as WiMAX) system etc.Term " network " and " system " can be replaced mutually.

In embodiments of the present invention, base station (base station, referred to as BS) can be as relay with subscriber equipment (user equipment, referred to as UE) or other communication site, carry out the equipment communicated, base station can provide the communication overlay of specific physical region.Such as, base station can be specifically base transceiver station (base transceiver station, referred to as BTS) in GSM or CDMA or base station controller (base station controller, referred to as BSC); Also can be the Node B (node B, referred to as NB) in UMTS or the radio network controller (radio network controller, referred to as RNC) in UMTS; It can also be the evolved base station (evolutional node B, referred to as ENB or eNodeB) in LTE; Or also can be other access network equipments providing access service in cordless communication network, the present invention limit.

In embodiments of the present invention, UE can be distributed in whole wireless network, and each UE can be static or movement.UE can be called terminal (terminal), travelling carriage (mobile station), subscriber unit (subscriber unit), platform (station) etc.UE can be cell phone (cellular phone), personal digital assistant (personal digital assistant, referred to as PDA), radio modem (modem), Wireless Telecom Equipment, handheld device (handheld), kneetop computer (laptopcomputer), cordless telephone (cordless phone), wireless local loop (wireless local loop, referred to as WLL) platform etc.

Fig. 1 shows the embodiment schematic diagram implementing the MIMO precoding based on average Signal to Interference plus Noise Ratio criterion proposed by the invention in TD-LTE system.In order to implement the MIMO precoding based on average Signal to Interference plus Noise Ratio criterion, transmitting terminal source signal is after scrambling and modulation treatment, send transmitting terminal precoding module, the pre-coding matrix determined is calculated according to pre-coding matrix computing module, transfer pre-coding process is performed to transmission data, then send resource particle to map, then after ofdm signal generation module, produce OFDM symbol send transmitting antenna port through MIMO-OFDM transmission.In order to implement the MIMO precoding based on average Signal to Interference plus Noise Ratio criterion, TD-LTE system receiving terminal is via reception antenna port accepts signal, through past CP, after OFDM demodulation, frequency-region signal is switched back to from time-domain signal, then solution resource particle maps, and then performs the joint-detection of MIMO and precoding by separating precoding module, eventually passes the source signal restoring base station end transmission after demodulation is disturbed.In embodiments of the present invention, base station and subscriber equipment all can be used as the main body performing method provided by the invention, such as, when base station is as receiving terminal, base station receives the pilot signal of subscriber equipment transmission by up channel, and obtains channel average and channel covariancc, and calculates pre-coding matrix.On the contrary, if base station is as transmitting terminal, then subscriber equipment is then receiving terminal, and subscriber equipment receives the pilot signal of base station transmission by down channel, and calculates pre-coding matrix.

Fig. 2 shows under TD-LTE system condition receiving terminal and receives process schematic block diagram based on the MIMO precoding of average Signal to Interference plus Noise Ratio criterion.Receiving terminal, via reception antenna port accepts signal, is removed cyclic prefix CP, after OFDM demodulation, is switched back to frequency-region signal from time-domain signal; Then map through separating resource particle; Channel estimation module estimates the channel gain of pilot frequency symbol position according to the pilot signal received, and count channel average and channel covariancc, then precoding computing module calculates according to channel average and channel covariancc and determines pre-coding matrix, transmitting terminal can not utilize channel reciprocity estimate channel condition information time, receiving terminal will by feedback channel by precoding matrix feedback to transmitting terminal to implement precoding processing.Pre-coding matrix and will send solution precoding and MIMO joint-detection module according to the channel estimating that pilot frequency symbol position channel estimating implements the data symbol positions that two-dimensional linear interpolation calculates.Separate precoding and MIMO joint detection results layer maps through separating, demodulation, after descrambling, restore transmitting terminal transmission message.

Based on said system, embodiments provide a kind of method for precoding, in the present embodiment, transmitting terminal can be mobile phone or base station, corresponding, receiving terminal can be base station or mobile phone, and when transmitting terminal is mobile phone, receiving terminal is base station, when transmitting terminal is base station, receiving terminal is mobile phone, and as shown in Figure 3, described method comprises:

301, the uplink/downlink pilot signal that receiving terminal receiving end/sending end sends;

302, receiving terminal carries out channel condition information estimation according to described uplink/downlink pilot signal, to obtain channel average and channel covariancc;

More specifically, this step comprises further: receiving terminal is estimated the channel status that the pilot signal received carries out pilot frequency symbol position, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel;

To described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with obtain data to be estimated described pilot signal channel estimation results on sub-carriers;

According to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers;

Obtain the statistical channel state information in correlation time according to described Noise Variance Estimation value, described statistical channel state information comprises described channel average and described channel covariancc.

303, receiving terminal, according to described channel average and channel covariancc, calculates pre-coding matrix;

304, when row channel does not meet symmetry over/under, described precoding matrix feedback is given described transmitting terminal by receiving terminal, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix;

305, receiving terminal, according to described pre-coding matrix, carries out solution precoding to the data-signal received.

Specifically, the described channel status estimation pilot signal received being carried out to pilot frequency symbol position, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel, can be realized by following mode:

TD-LTE system receiving terminal is via reception antenna port accepts, and frequency domain pilot signal received after removing cyclic prefix CP, OFDM demodulation is

Y _k,p(i _m)＝X _k,p(i _m)H _k,p(i _m)+W _k,p(i _m),0≤m≤N _p-1 (1)

Wherein i _mthe location index existed in the OFDM symbol of pilot signal, N _pit is the frequency pilot sign number on a subcarrier; for the frequency pilot sign received, N _rfor reception antenna number, k represents a kth subcarrier; for the pilot matrix that transmitting terminal sends, N _tfor number of transmit antennas; it is corresponding Frequency domain noise; H _k,p(i _m) represent the mimo channel frequency response on DFT frequency k, and have

$H_{k,p}\left(i_m\right)={\left[\begin{array}{cccc}{h}_{k,p}^{\mathrm{1,1}}\left(i_m\right)& h_k^{\mathrm{1,2}}\left(i_m\right)& ...& h_k^{N_t,1}\left(i_m\right)\\ h_{k,p}^{\mathrm{2,1}}\left(i_m\right)& h_k^{\mathrm{2,2}}\left(i_m\right)& ...& h_k^{N_t,2}\left(i_m\right)\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ h_{k,p}^{N_t,1}\left(i_m\right)& h_k^{N_t,2}\left(i_m\right)& ...& h_k^{N_t,N_r}\left(i_m\right)\end{array}\right]}_{N_t\&times;N_r}---\left(2\right)$

By the N in a subframe _pindividual frequency pilot sign combines

Y _k,p＝X _k,pH _k,p+W _k,p(3)

Wherein

$Y_{k,p}={\left[\begin{array}{c}{Y}_{k,p}\left(i_0\right)\\ Y_{k,p}\left(i_1\right)\\ ...\\ Y_{k,p}\left(i_{N_p-1}\right)\end{array}\right]}_{N_p\&times;N_r},W_{k,p}={\left[\begin{array}{c}{W}_{k,p}\left(i_0\right)\\ W_{k,p}\left(i_1\right)\\ ...\\ W_{k,p}\left(i_{N_p-1}\right)\end{array}\right]}_{N_p\&times;N_r}$

On the basis of (3) formula measuring-signal, the LS channel estimation results that can obtain frequency pilot sign is:

Wherein for X _k,ppseudoinverse.Consider that in real system, channel gain does not have significant change, i.e. H in channel coherency time _k,p(i _m) ≈ H _k,p(i ₀), m=1 ..., N _c, wherein N _c>N _pfor the OFDM symbol number of equal value that the correlated time of channel is corresponding, then (3) formula can be rewritten as following cascade observation signal sequence:

${\tilde{Y}}_{k,p}={\tilde{X}}_{k,p}{H}_{k,p}\left(i_0\right)+{\tilde{W}}_{k,p}---\left(5\right)$

Wherein

${\tilde{Y}}_{k,p}={\left[\begin{array}{c}{Y}_{k,p}\left(i_0\right)\\ Y_{k,p}\left(i_1\right)\\ ...\\ Y_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\&times;N_r},{\tilde{W}}_{k,p}={\left[\begin{array}{c}{W}_{k,p}\left(i_0\right)\\ W_{k,p}\left(i_1\right)\\ ...\\ W_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\&times;N_r},{\tilde{X}}_{k,p}={\left[\begin{array}{c}{X}_{k,p}\left(i_0\right)\\ X_{k,p}\left(i_1\right)\\ ...\\ X_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\&times;N_t}$

On the basis of (5) formula measuring-signal, the LS channel estimation results that can obtain based on frequency pilot sign within the correlated time of channel is:

Wherein for pseudoinverse.

On the basis estimating pilot frequency symbol position channel condition information, can be derived by time domain and frequency domain two-dimensional linear interpolation and draw the channel condition information of non-pilot character position; Namely to described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with obtain data to be estimated described pilot signal channel estimation results on sub-carriers;

First, in time domain to described until former residence Data Position time Domain Index on the step of channel condition information, be specially:

H _k,data(l′)＝α ₁×H _k,p,LS(l)+β ₁×H _k,p,LS(l+Δl) (7)

Wherein l ' represents the time Domain Index and have l < l ' < l+ Δ l of Data Position to be estimated, l represents pilot signal position in the time domain, Δ l represents the time domain interval of two reference signals that time domain is adjacent, k represents pilot signal place subcarrier sequence number, interpolation coefficient α ₁, β ₁calculate by the following method

${\mathrm{\&alpha;}}_1=\frac{l+\mathrm{\&Delta;l}-l^{\&prime;}}{\mathrm{\&Delta;l}},{\mathrm{\&beta;}}_1=\frac{l^{\&prime;}-l}{\mathrm{\&Delta;l}}---\left(8\right)$

The channel estimation results of frequency pilot sign place subcarrier at time domain non-pilot character position can be obtained after temporal interpolation.

Afterwards, in frequency domain to described until former residence Data Position time Domain Index on the step of channel condition information, be specially:

H _k′(l)＝α ₂×H _k,S(l)+β ₂×H _k+Δk,S(l),S∈[p,data] (9)

Wherein k ' represents the frequency domain index of Data Position to be estimated and has k < k ' < k+ Δ k, k represents pilot signal position in a frequency domain, Δ k represents the frequency interval of two reference signals that frequency domain is adjacent, interpolation coefficient α ₂, β ₂calculate by the following method

${\mathrm{\&alpha;}}_2=\frac{k+\mathrm{\&Delta;k}-k^{\&prime;}}{\mathrm{\&Delta;k}},{\mathrm{\&beta;}}_2=\frac{k^{\&prime;}-k}{\mathrm{\&Delta;k}}---\left(10\right)$

The channel estimation results obtaining all data symbol positions can be estimated after Frequency domain interpolation.

From (6) formula

Afterwards, according to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers;

Realize especially by such as under type, due to adjacent pilot frequencies subcarrier upper signal channel in same OFDM symbol gain closely, i.e. H _k,p(i _m) ≈ H _k,p(i _m+ 1), can utilize this feature adopt following methods direct derivation go out frequency pilot sign Noise Variance Estimation on sub-carrierk

${\mathrm{\&sigma;}}_{w,k}^2=\frac{1}{N_c-1}\underset{m=0}{\overset{N_c-2}{\mathrm{\&Sigma;}}}{\left|\right|H_{k,p,\mathrm{LS}}\left(i_{m+1}\right)-H_{k,p,\mathrm{LS}}\left(i_m\right)\left|\right|}_F^2---\left(12\right)$

Wherein || H _{k, p, LS}(i _m+1)-H _{k, p, LS}(i _m) || _frepresent Frobenius norm.Finally total Noise Variance Estimation value can be expressed from the next:

${\mathrm{\&sigma;}}_w^2=\frac{1}{M}\underset{k=1}{\overset{M}{\mathrm{\&Sigma;}}}{\mathrm{\&sigma;}}_{w,k}^2---\left(13\right)$

Wherein M represents the total number of sub-carriers shared by frequency pilot sign.

Finally, according to the correlation time of communication channel, the sliding window length L(of choose reasonable time domain corresponds to the number of pilot symbols that in sliding window, subframe comprises), according to the correlation bandwidth of channel, at the M that frequency domain comprises from each coherence bandwidth ₀the channel estimation results of individual pilot sub-carrier is estimated to calculate moment i's by the following method the average of individual channel condition information and covariance wherein N is the total number of sub-carriers of carrying user data.

Wherein, vec () represents the vector quantization of matrix, () ^hrepresent conjugate transpose.

Above in the average calculating channel condition information and covariance time, according to M adjacent in coherence in frequency domain bandwidth ₀the average of channel condition information in individual frequency pilot sign place sub-carrier channels estimated result statistical computation coherence in frequency domain bandwidth and covariance because the different sub carrier characteristic of channel is similar to identical in coherence bandwidth, can think that the average of the channel condition information that each subcarrier is corresponding is in correlation bandwidth covariance is

Examine in MIMO pre-coding system, B represents pre-coding matrix, and H represents N _r× N _tmimo channel, wherein N _rfor reception antenna number, N _tfor number of transmit antennas, A representative receives monitoring matrix, then corresponding error vector covariance matrix be

$\mathrm{MSE}=E\left\{(\hat{x}-x){(\hat{x}-x)}^H\right\}{=A}^H{R}_{y}A+I_{N_r}-A^H\mathrm{HB}-B^H{H}^{H}A---\left(17\right)$

Wherein

R _y＝E{yy ^H}＝HBB ^HH ^H+R _n(18)

If the covariance matrix of white complex gaussian noise receiving terminal adopts linear minimum mean-squared error LMMSE monitoring matrix, namely

$A={({\mathrm{HBB}}^H{H}^H+I_{N_r})}^{-1}\mathrm{HB}---\left(19\right)$

Covariance matrix can be written as further:

$\mathrm{MSE}=E\{(\hat{x}-x){(\hat{x}-x)}^H\}={(I_{N_r}+B^H{H}^H\mathrm{HB})}^{-1}---\left(20\right)$

The Signal to Interference plus Noise Ratio SINR of i-th then corresponding transmission symbol _iwith the MSE of its correspondence _ithere is following relation:

${\mathrm{SINR}}_i=\frac{1}{{\mathrm{MSE}}_i}-1=b_i^H{H}^H{\mathrm{Hb}}_i---\left(21\right)$

Wherein

${\mathrm{MSE}}_i={\left[{(I+B^H{H}^H\mathrm{HB})}^{-1}\right]}_{i,i}=\frac{1}{1+b_i^H{H}^H{\mathrm{Hb}}_i}---\left(22\right)$

Wherein b _ii-th column vector of representing matrix B, [X] _i,jthe capable j column element of i of representing matrix X.Then the Signal to Interference plus Noise Ratio sum of all transmission symbols can be expressed as follows:

$\mathrm{SINR}=\underset{i}{\mathrm{\&Sigma;}}{\mathrm{SINR}}_i=\underset{i}{\mathrm{\&Sigma;}}{b}_i^H{H}^{H}H{b}_i=\underset{i}{\mathrm{\&Sigma;}}{\left[B^H{H}^H\mathrm{HB}\right]}_{i,i}=\mathrm{Tr}\left(B^H{H}^H\mathrm{HB}\right)---\left(23\right)$

Under total transmitted power constraints, optimization aim can be set to maximize the Signal to Interference plus Noise Ratio sum of all transmission symbols to determine pre-coding matrix B, namely corresponding precoding optimization design problem can be described as following optimization problem

$\begin{array}{c}\max \underset{B}{\arg }{E}_H\left\{\mathrm{Tr}\left(B^H{H}^H\mathrm{HB}\right)\right\}\\ s.t.\mathrm{tr}\left({\mathrm{BB}}^H\right)=\frac{P}{N}\end{array}---\left(24\right)$

Wherein P/N represents that transmitting terminal distributes to total transmitted power of each subcarrier, and Tr () represents mark.According to commutative properties Tr (the AB)=Tr (BA) of mark, target function can be equivalent as follows:

E _H{tr(B ^HH ^HHB)}＝E _H{tr(HBB ^HH ^H)}＝E _H{tr(HQH ^H)} (25)

Wherein Q=BB ^hrepresent the autocorrelation matrix sending precoding.If the covariance matrix of white complex gaussian noise then the form of above-mentioned target function remains unchanged, and only need do following replacement:

$\tilde{H}=R_n^{-1/2}H,\stackrel{\widetilde{\&OverBar;}}{H}=R_n^{-1/2}\stackrel{\&OverBar;}{H},{\widetilde{\mathrm{\&Phi;}}}_H=(I_{N_r}\&CircleTimes;{R_n}^{-1/2}){\mathrm{\&Phi;}}_H(I_{N_r}\&CircleTimes;{R_n}^{-1/2})---\left(26\right)$

Wherein represent that Kronecker amasss.

The best pre-coding matrix B that can calculate and determine under average Signal to Interference plus Noise Ratio criterion is solved for (24) formula optimization problem.Definition wherein unvec () is the inverse operator of vec (), and utilizes variable zero-mean and mark commutative properties, and based on the statistics DCSIT channel model in (16) formula, the target function in (24) formula can be written as further:

$\begin{array}{c}{E}_H\left\{\mathrm{Tr}\left({\mathrm{HQH}}^H\right)\right\}\\ =E_{H_w}\left\{\mathrm{Tr}\left(Q^{1/2}{H}^H{\mathrm{HQ}}^{1/2}\right)\right\}\\ =E_{H_w}\left\{\mathrm{Tr}\right(Q^{1/2}{(\stackrel{\&OverBar;}{H}+{\tilde{H}}_w)}^H(\stackrel{\&OverBar;}{H}+{\tilde{H}}_w)Q^{1/2}\left)\right\}\\ =E_{H_w}\left\{\mathrm{Tr}\left(Q^{1/2}{\left({\tilde{H}}_w\right)}^H{\tilde{H}}_w{Q}^{1/2}\right)\right\}+\mathrm{Tr}\left({\stackrel{\&OverBar;}{H}}^H\stackrel{\&OverBar;}{H}Q\right)\end{array}---\left(27\right)$

Wherein

$\begin{array}{c}{E}_{H_w}\left\{\mathrm{Tr}\left(Q^{1/2}{\left({\tilde{H}}_w\right)}^H{\tilde{H}}_w{Q}^{1/2}\right)\right\}\\ =E_{H_w}\{{\left(\mathrm{vec}\left({\tilde{H}}_w\right)\right)}^H({\left[Q^{1/2}\right]}^T\&CircleTimes;I_{N_r})\\ =\mathrm{Tr}\left({\mathrm{\&Phi;}}_H(Q^T\&CircleTimes;I_{N_r})\right)\end{array},({\left[Q^{1/2}\right]}^T\&CircleTimes;I_{N_r})\mathrm{vec}\left({\tilde{H}}_w\right)\}---\left(28\right)$

Wherein () ^tthe transposition of representing matrix, make use of trace function character Tr (A here ^hb)=(vec (A)) ^hvec (B), and in conjunction with channel covariancc Φ _hdefinition.(28) formula is substituted into (27) formula, and the Operation Nature long-pending according to Kronecker, target function can be transformed to:

$\begin{array}{c}\mathrm{Tr}\left({\mathrm{\&Phi;}}_H(Q^T\&CircleTimes;I_{N_r})\right)+\mathrm{Tr}\left({\stackrel{\&OverBar;}{H}}^H\stackrel{\&OverBar;}{H}Q\right)\\ =\mathrm{Tr}\left({\mathrm{\&Phi;}}_H(Q^T\&CircleTimes;I_{N_r})\right)+\frac{1}{N_r}\mathrm{Tr}({\stackrel{\&OverBar;}{H}}^T{\stackrel{\&OverBar;}{H}}^{*}\&CircleTimes;I_{N_r})(Q^T\&CircleTimes;I_{N_r})\\ =\mathrm{Tr}\left(\right[{\mathrm{\&Phi;}}_H+\frac{1}{N_r}({\stackrel{\&OverBar;}{H}}^T{\stackrel{\&OverBar;}{H}}^{*}\&CircleTimes;I_{N_r})\left](Q^T\&CircleTimes;I_{N_r})\right)\\ =\mathrm{Tr}\left(X(Q\&CircleTimes;I_{N_r})\right)\end{array}---\left(29\right)$

Wherein () ^*the conjugation of representing matrix, (29) equation character Tr (A)=Tr (A of mark is applied in formula ^t).N _rn _t× N _rn _tdimension matrix (i, j) individual N _r× N _rsubmatrix is proportional units battle array, namely wherein q _ijrepresent the i-th row j column element of Q, δ _rskronecker delta function (δ _rs=1, r=s; δ _rs=0, r ≠ s), basis in formula the special shape of matrix and matrix trace equal this definition of matrix diagonals line element sum.(29) formula can simplify as follows further:

$\begin{array}{c}\mathrm{Tr}\left(X(Q\&CircleTimes;I_{N_r})\right)=\mathrm{Tr}\left((Q\&CircleTimes;I_{N_r})X\right)=\underset{i=1}{\overset{N_t}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_t}{\mathrm{\&Sigma;}}}\underset{r=1}{\overset{N_r}{\mathrm{\&Sigma;}}}\underset{s=1}{\overset{N_r}{\mathrm{\&Sigma;}}}{q}_{\mathrm{ij}}{\mathrm{\&delta;}}_{\mathrm{rs}}{X}_{(i-1)N_r+r,(j-1)N_r+s}\\ =\underset{i=1}{\overset{N_t}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_t}{\mathrm{\&Sigma;}}}{q}_{\mathrm{ij}}\underset{r=1}{\overset{N_r}{\mathrm{\&Sigma;}}}\underset{s=1}{\overset{N_r}{\mathrm{\&Sigma;}}}{\mathrm{\&delta;}}_{\mathrm{rs}}{X}_{(i-1)N_r+r,(j-1)N_r+s}=\underset{i=1}{\overset{N_t}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_t}{\mathrm{\&Sigma;}}}{q}_{\mathrm{ij}}\mathrm{Tr}\left(X_{[i,j]}\right)\\ \stackrel{z_{i,j}=\mathrm{Tr}\left(X_{[i,j]}\right)}{\&RightArrow;}\underset{i=1}{\overset{N_t}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{N_t}{\mathrm{\&Sigma;}}}{q}_{\mathrm{ij}}{z}_{i,j}=\mathrm{Tr}\left(\mathrm{QZ}\right)=\mathrm{Tr}\left(\mathrm{ZQ}\right)\end{array}---\left(30\right)$

Wherein X _{[i, j]}represent (i, j) individual N of X _r× N _rsubmatrix.Therefore the target function in (24) formula can be rewritten as

$\begin{array}{c}\max \underset{Q}{\arg }\mathrm{Tr}\left(\mathrm{ZQ}\right)\\ s.t.\mathrm{Tr}\left(Q\right)=\frac{P}{N}\end{array}---\left(31\right)$

Because Z entry of a matrix element is the mark of positive semidefinite matrix sub-block, Z matrix is also positive semi-definite.Can prove that the base of matrix Q is consistent with Z, if namely represent the Eigenvalues Decomposition of Z, then the Eigenvalues Decomposition of Q is

${Q=U}_Z{\mathrm{\&Lambda;}}_Q{U}_Z^H---\left(32\right)$

Wherein, Λ _qrepresent the eigenvalue matrix of Q.The Eigenvalues Decomposition of Z and above formula are substituted into target function, Λ _qsolve problems just become following optimization problem

$\begin{array}{c}\max \underset{{\mathrm{\&lambda;}}_{Q_i}}{\arg }\underset{i=1}{\overset{N_t}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{Z_i}{\mathrm{\&lambda;}}_{Q_i}\\ s.t.{\mathrm{\&lambda;}}_{Q_i}\&GreaterEqual;0,{\mathrm{\&Sigma;}}_{i=1}^{N_t}{\mathrm{\&lambda;}}_{Z_i}\&le;\frac{P}{N}\end{array}---\left(33\right)$

Wherein λ _zi, λ _qirepresent Z respectively, the characteristic value of Q.Because Z matrix is represented by channel average and channel covariancc.If consider subchannel mean allocation transmitted power, i.e. Λ _qdiagonal entry equal if consideration transmitted power every root transmitting antenna is distributed to, then Λ by power water filling _qcalculate by the following method

Wherein, $P_i=\max (\mathrm{\&mu;}-\frac{{\mathrm{\&sigma;}}_w^2}{{\mathrm{\&lambda;}}_i},0),i=\mathrm{1,2},...,r,\underset{i=1}{\overset{r}{\mathrm{\&Sigma;}}}{P}_i=\frac{P}{N_t},({\mathrm{\&lambda;}}_1,{\mathrm{\&lambda;}}_2,...{\mathrm{\&lambda;}}_r)$ It is matrix r nonzero eigenvalue, μ is the water filling factor.

Specifically, according to the MIMO-OFDM system channel average estimating to obtain channel covariancc and noise variance the impact of noise is added in channel average and channel covariancc, obtains new channel average and channel covariancc wherein

${\stackrel{\&OverBar;}{H}}_{m,n}^{\left(s\right)}=R_n^{-1/2}{\stackrel{\&OverBar;}{H}}_m^{\left(s\right)}---\left(35\right)$

${\mathrm{\&Phi;}}_{H,n}^{\left(s\right)}=(I_{N_r}\&CircleTimes;R_n^{-1/2}){\mathrm{\&Phi;}}_H^{\left(s\right)}(I_{N_r}\&CircleTimes;R_n^{-1/2})---\left(36\right)$

Wherein, noise covariance matrix according to and Φ _h,n, compute matrix wherein:

$Z_{[i,j]}^{\left(s\right)}=\mathrm{Tr}\left(X_{[i,j]}^{\left(s\right)}\right),x^{\left(s\right)}=({\mathrm{\&Phi;}}_H^{\left(s\right)}+(1/N_r){\left(\right({\stackrel{\&OverBar;}{H}}_m^{\left(s\right)})}^H{\stackrel{\&OverBar;}{H}}_m^{\left(s\right)}\&CircleTimes;I_{N_r})),i=1,...,N_t,j=1,...,N_t---\left(37\right)$

Wherein X _{[i, j]}each and every one dimension of i-th j of expression X is N _r× N _rsub-square.By matrix Z ^(s)carry out Cholesky decomposition, obtain matrix L ^(s)

Z ^(s)＝L ^(s)·(L ^(s)) ^H(38)

Then to L ^(s)matrix carries out singular value decomposition

$L^{\left(s\right)}=U_L^{\left(s\right)}{\mathrm{\&Lambda;}}_L^{\left(s\right)}{\left(V_L^{\left(s\right)}\right)}^H---\left(39\right)$

Thus obtain Z ^(s)eigenvalues Decomposition expression formula

$Z^{\left(s\right)}=U_Z^{\left(s\right)}{\mathrm{\&Lambda;}}_Z^{\left(s\right)}{\left(V_Z^{\left(s\right)}\right)}^H=U_L^{\left(s\right)}{\left({\mathrm{\&Lambda;}}_L^{\left(s\right)}\right)}^2{\left(U_L^{\left(s\right)}\right)}^H---\left(40\right)$

Pre-coding matrix then based on average Signal to Interference plus Noise Ratio criterion calculates by the following method

$B^{\left(s\right)}=U_Z^{\left(s\right)}{\mathrm{\&Lambda;}}_Q^{\left(s\right)}---\left(41\right)$

Wherein

$U_Z^{\left(s\right)}=U_L^{\left(s\right)},{\mathrm{\&Lambda;}}_Q^{\left(s\right)}={\mathrm{\&Lambda;}}_L^{\left(s\right)}---\left(42\right)$

If consider, transmitted power P is averagely allocated to every root transmitting antenna, then calculate by the following method

${\mathrm{\&Lambda;}}_{Q,\mathrm{ave}}^{\left(s\right)}=\sqrt{\frac{P}{N_{t}N}}{I}_{N_t}---\left(43\right)$

If consideration transmitted power every root transmitting antenna is distributed to, then by power water filling calculate by the following method

Wherein $P_i=\max (\mathrm{\&mu;}-\frac{{\mathrm{\&sigma;}}_w^2}{{\mathrm{\&lambda;}}_i},0),i=\mathrm{1,2},...,r,\underset{i=1}{\overset{r}{\mathrm{\&Sigma;}}}{P}_i=\frac{P}{N},({\mathrm{\&lambda;}}_1,{\mathrm{\&lambda;}}_2,...{\mathrm{\&lambda;}}_r)$ It is matrix r nonzero eigenvalue, μ is the water filling factor.

Observe the form of pre-coding matrix B, for power division matrix, reflect power allocation case, matrix Z ^(s)unitary matrice, multiplexed signals is decomposed the transmission of several mutually orthogonal direction by the mutually orthogonal characteristic of its column vector, can see multimode wave beam formed matrix as, and input forming matrix can think unit matrix I.And restrictive condition then meet gross power permanent character.Obviously, adopt the pre-coding matrix shown in (41) formula, the average Signal to Interference plus Noise Ratio of receiving terminal is

$\begin{array}{c}E\left\{\mathrm{SINR}\right\}=E\left\{\underset{i}{\mathrm{\&Sigma;}}{\mathrm{SINR}}_i\right\}=E\left\{\underset{i}{\mathrm{\&Sigma;}}{b}_i^H{H}^H{R}_{n,i}^{-1}{\mathrm{Hb}}_i\right\}=E\left\{\underset{i}{\mathrm{\&Sigma;}}{\left[B^H{H}^H{R}_n^{-1}\mathrm{HB}\right]}_{i,i}\right\}\\ =E\left\{\mathrm{Tr}\left(B^H{H}^H{R}_n^{-1}\mathrm{HB}\right)\right\}=\mathrm{Tr}\left(\mathrm{ZQ}\right)=\underset{i}{\overset{N_t}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{Z_i}{\mathrm{\&lambda;}}_{Q_i}=\frac{P}{N_t}\underset{i}{\overset{N_t}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_{Z_i}=\frac{P}{N_t}\mathrm{Tr}\left(Z\right)\end{array}---\left(45\right)$

From Such analysis, when determining each subcarrier pre-coding matrix according to the calculating of statistics DCSIT information, pre-coding scheme mainly depends on the channel average with each sub-carrier channels channel covariancc consider the channel average of different sub carrier channel in coherence in frequency domain bandwidth channel covariancc substantially identical, when the pre-coding matrix of Practical Calculation determination different sub carrier channel, can by coherence in frequency domain bandwidth, channel can be divided into accordingly group calculates determines corresponding pre-coding matrix.

At receiving terminal, after the data that the pre-coding matrix receiving the employing feedback that transmitting terminal sends sends, after adopting precoding, receiving terminal Received signal strength is:

Y _k(l)＝H _k(l)·B·X _k(l)+W _k(l) (46)

Wherein $Y_k\left(l\right)={\tilde{Y}}_k^T\left(l\right),X_k\left(l\right)={\tilde{X}}_k^T\left(l\right),H_k\left(l\right)=H_k^T\left(l\right),W_k\left(l\right)={\tilde{W}}_k^T\left(l\right).$ Receiving terminal, on the basis estimating pilot frequency symbol position channel condition information, estimates by time domain and frequency domain two-dimensional linear interpolation the channel condition information H obtaining non-pilot character position _kl (), in conjunction with calculating the pre-coding matrix B determined, can calculate the channel matrix H determining equivalence _e,k(l), namely

H _E,k(l)＝Β·H _k(l) (47)

According to the combined channel matrix H of equivalence _e,kl () take least mean-square error as criterion, receive can carry out separating the joint-detection that precoding and MIMO detect and receive

${\hat{X}}_k\left(l\right)=G_{\mathrm{MMSE}}\&CenterDot;Y_k\left(l\right)={[H_{E,k}^H\left(l\right)H_{E,k}\left(l\right)+{\mathrm{\&sigma;}}_n^2{I}_{N_t}]}^{-1}{H}_{E,k}^H\left(l\right)Y_k\left(l\right)---\left(48\right)$

Fig. 4 shows under TD-LTE system condition to implement the MIMO precoding transmitting terminal end process schematic block diagram based on average Signal to Interference plus Noise Ratio criterion.When uplink downlink satisfying reciprocity, transmitting terminal is via antenna port Received signal strength, after SC-FDMA demodulation, then map through separating resource particle, channel estimation module estimates the channel gain of pilot frequency symbol position according to the uplink pilot signal received, and count channel average and channel covariancc, then precoding computing module calculates according to channel average and channel covariancc and determines downlink precoding matrix.When transmitting terminal can not utilize channel reciprocity to estimate channel condition information, transmitting terminal calculates the pre-coding matrix determined so that transmitting terminal enforcement precoding processing by being received by feedback channel by receiving terminal.

Fig. 5-Figure 16 compares SVD linear predictive coding (average power allocation SVD-ave and power water filling SVD-wf), (SINR mean) scheme of the MIMO precoding based on average Signal to Interference plus Noise Ratio criterion under DCSIT condition and based on maximizing the codebook precoding (average power allocation SINRave and power water filling SINRwf) of SINR codebook selecting scheme at ETU1, bit error rate performance under EVA1 and EPA channel condition and power system capacity, wherein the feedback delay of channel condition information is the transmitting time of a subframe, feedback granularity is a subframe, in closed loop code book feedback scheme, the time domain feedback granularity of PMI is a subframe, frequency domain feedback granularity is 2 Resource Block, the time domain feedback granularity of linear predictive coding SVD is an OFDM symbol, frequency domain feedback granularity is 1 Resource Block, simulated conditions is as shown in table 1.Can draw to draw a conclusion from Fig. 5-Figure 16:

First aspect, from the system goes capacity that precoding unfailing performance and employing precoding realize, SVD-wf can obtain best compromise between precoding performance and power system capacity, SVD-ave takes second place, and the precoding performance of Signal to Interference plus Noise Ratio average SINRave and SINRwf is basic suitable with the codebook precoding based on SINR criterion.

Second aspect, under low associated channel condition, the pre-coding scheme based on Signal to Interference plus Noise Ratio average SINRave and SINRwf can obtain the unfailing performance close to SVD pre-coding scheme, and power system capacity is suitable with the codebook precoding based on SINR criterion.This shows, the pre-coding scheme based on Signal to Interference plus Noise Ratio average is that the precoding under low associated channel condition is implemented to provide a kind of feasible technical scheme.

The third aspect, in low signal-to-noise ratio region, SINRave precoding unfailing performance is better than SINRwf precoding, and power system capacity is slightly inferior to SINRwf, and in high s/n ratio region, the precoding performance of two kinds of power allocation schemes is suitable.This shows, power waterflood project is little to the performance gain based on the precoding of average SINR criterion, and mean allocation power is implemented simple, and SINRave precoding is a kind of simply useful technical scheme.

Consider the unfailing performance achieved by system and power system capacity, system realizes the requirement to channel condition information in addition, is one rational precoding technique scheme under DCSIT condition based on Signal to Interference plus Noise Ratio average criterion pre-coding scheme.

Accordingly, the embodiment of the present invention additionally provides a kind of pre-coding apparatus, and described device can be base station also can be subscriber equipment, when described device is base station, base station is receiving terminal, transmitting terminal is subscriber equipment, on the contrary, if described device is subscriber equipment, then base station is transmitting terminal, receiving terminal is subscriber equipment, and as seen from Figure 17, described device comprises:

Receiving element 701, for the pilot signal that receiving end/sending end sends;

Estimation unit 702, for carrying out channel condition information estimation according to uplink and downlink pilot signal, to obtain channel average and channel covariancc;

Computing unit 703, for according to described channel average and channel covariancc, calculates pre-coding matrix;

Feedback unit 704, for giving described transmitting terminal when uplink and downlink channel does not meet symmetry by described precoding matrix feedback, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix.

Decoding unit 705, for according to described pre-coding matrix, carries out solution precoding to the data-signal received.

Further, described estimation unit specifically for:

The channel status pilot signal received being carried out to pilot frequency symbol position is estimated, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel;

To described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with obtain data to be estimated channel estimation results on sub-carriers;

According to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers;

Obtain the statistical channel state information in correlation time according to described Noise Variance Estimation value, described statistical channel state information comprises described channel average and described channel covariancc.

How to calculate channel average and covariance about pre-coding apparatus in the present embodiment, and obtain the method for pre-coding matrix, with reference to aforesaid embodiment, seldom can repeat.

As shown in figure 18, the embodiment of the present invention additionally provides a kind of pre-coding apparatus, described device can be base station also can be subscriber equipment, when described device is base station, base station is receiving terminal, transmitting terminal is subscriber equipment, on the contrary, if described device is subscriber equipment, then base station is transmitting terminal, receiving terminal is subscriber equipment, as seen from Figure 18, and the processor 184 that described device comprises transmitter 182, receiver 181, memory 183 and is connected with transmitter 182, receiver 181 and memory 183 respectively.Certainly, base station can also comprise the universal components such as antenna, baseband process component, middle radio frequency processing parts, input/output unit, and the embodiment of the present invention is in this no longer any restriction.

Wherein, in memory 183, store batch processing code, and processor 184 is for calling the program code stored in memory, for performing following operation:

The uplink/downlink pilot signal sent by receiver receiving end/sending end;

Channel condition information estimation is carried out, to obtain channel average and channel covariancc according to described uplink/downlink pilot signal;

According to described channel average and channel covariancc, calculate pre-coding matrix;

When row channel does not meet symmetry over/under by transmitter, give described transmitting terminal by described precoding matrix feedback, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix;

According to described pre-coding matrix, solution precoding is carried out to the data-signal received.

It should be noted that, the pre-coding apparatus shown in Figure 17 and Figure 18 may be used for realizing any one method that above embodiment of the method provides, and does not repeat them here.

Professional should recognize further, in conjunction with unit and the algorithm steps of each example of embodiment disclosed herein description, can realize with electronic hardware, computer software or the combination of the two, in order to the interchangeability of hardware and software is clearly described, generally describe composition and the step of each example in the above description according to function.These functions perform with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme.Professional and technical personnel can use distinct methods to realize described function to each specifically should being used for, but this realization should not thought and exceeds scope of the present invention.

The software module that the method described in conjunction with embodiment disclosed herein or the step of algorithm can use hardware, processor to perform, or the combination of the two is implemented.Software module can be placed in the storage medium of other form any known in random asccess memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field.

Above-described embodiment; object of the present invention, technical scheme and beneficial effect are further described; be understood that; the foregoing is only the specific embodiment of the present invention; the protection range be not intended to limit the present invention; within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. a method for precoding, is characterized in that, described method comprises:

The uplink/downlink pilot signal that receiving end/sending end sends;

Channel condition information estimation is carried out, to obtain channel average and channel covariancc according to described uplink/downlink pilot signal;

According to described channel average and channel covariancc, calculate pre-coding matrix;

When row channel does not meet symmetry over/under, give described transmitting terminal by described precoding matrix feedback, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix;

According to described pre-coding matrix, solution precoding is carried out to the data-signal received.

2. method for precoding as claimed in claim 1, is characterized in that, described according to uplink/downlink pilot signal carry out channel condition information estimation, to obtain channel average and channel covariancc, comprise further:

The channel status pilot signal received being carried out to pilot frequency symbol position is estimated, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel;

To described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with obtain data to be estimated channel estimation results on sub-carriers;

According to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers;

Obtain the statistical channel state information in correlation time according to described Noise Variance Estimation value, described statistical channel state information comprises described channel average and described channel covariancc.

3. method for precoding as claimed in claim 2, it is characterized in that, the channel status that the described pilot signal to receiving carries out pilot frequency symbol position is estimated, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel, specifically comprises:

By the N in a subframe _pindividual frequency pilot sign Y _k,p(i _m)=X _k,p(i _m) H _k,p(i _m)+W _k,p(i _m), 0≤m≤N _p-1

Be integrated into: Y _k,p=X _k,ph _k,p+ W _k,p

Wherein, k represents a kth subcarrier; i _mthe location index existed in the OFDM symbol of pilot signal, N _pit is the frequency pilot sign number on a subcarrier; for the frequency pilot sign received, for the pilot tone that transmitting terminal sends; N _rfor reception antenna number, N _tfor number of transmit antennas; it is corresponding Frequency domain noise; H _k,p(i _m) represent the mimo channel frequency domain response on a kth subcarrier, and have

$H_{k,p}\left(i_m\right)={\left[\begin{array}{cccc}{h}_{k,p}^{\mathrm{1,1}}\left(i_m\right)& h_k^{\mathrm{1,2}}\left(i_m\right)& ...& h_k^{N_t,1}\left(i_m\right)\\ h_{k,p}^{\mathrm{2,1}}\left(i_m\right)& h_k^{\mathrm{2,2}}\left(i_m\right)& ...& h_k^{N_t,2}\left(i_m\right)\\ .& .& & .\\ .& .& ...& .\\ .& .& & .\\ h_{k,p}^{N_t,1}\left(i_m\right)& h_k^{N_t,2}\left(i_m\right)& ...& h_k^{N_t,N_r}\left(i_m\right)\end{array}\right]}_{N_t\&times;N_r}$

At described Y _k,p=X _k,ph _k,p+ W _k,pin,

$Y_{k,p}={\left[\begin{array}{c}{Y}_{k,p}\left(i_0\right)\\ Y_{k,p}\left(i_1\right)\\ ...\\ Y_{k,p}\left(i_{N_p-1}\right)\end{array}\right]}_{N_p\&times;N_r},W_{k,p}={\left[\begin{array}{c}{W}_{k,p}\left(i_0\right)\\ W_{k,p}\left(i_1\right)\\ ...\\ W_{k,p}\left(i_{N_p-1}\right)\end{array}\right]}_{N_p\&times;N_r}$

According to described Y _k,p=X _k,ph _k,p+ W _k,p, obtain the frequency domain response that a mimo channel kth subcarrier estimates wherein, for X _k,ppseudoinverse;

According to described obtain the reception pilot matrix after the cascade in coherence time:

${\tilde{Y}}_{k,p}={\tilde{X}}_{k,p}{H}_{k,p}\left(i_0\right)+{\tilde{W}}_{k,p}$

Wherein, H _k,p(i _m) ≈ H _k,p(i ₀), m=1 ..., N _c, N _c>N _pfor the OFDM symbol number of equal value that the correlated time of channel is corresponding, wherein:

${\tilde{Y}}_{k,p}={\left[\begin{array}{c}{Y}_{k,p}\left(i_0\right)\\ Y_{k,p}\left(i_1\right)\\ ...\\ Y_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\&times;N_r},{\tilde{W}}_{k,p}={\left[\begin{array}{c}{W}_{k,p}\left(i_0\right)\\ W_{k,p}\left(i_1\right)\\ ...\\ W_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\&times;N_r},{\tilde{X}}_{k,p}={\left[\begin{array}{c}{X}_{k,p}\left(i_0\right)\\ X_{k,p}\left(i_1\right)\\ ...\\ X_{k,p}\left(i_{N_c-1}\right)\end{array}\right]}_{N_c\&times;N_t}$

Reception pilot matrix according to after the cascade in described coherence time:

The LS channel estimation results obtained based on frequency pilot sign within the correlated time of channel is:

wherein, for pseudoinverse.

4. method as claimed in claim 3, it is characterized in that, described to described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with the channel estimation results of data estimator symbol position, comprise further:

In time domain to the concrete steps of the channel estimating of described data symbol position to be estimated be:

H _k,data(l′)＝α ₁×H _k,p,LS(l)+β ₁×H _k,p,LS(l+Δl)

Wherein l ' represents the time Domain Index of Data Position to be estimated, and have l < l ' < l+ Δ l, l represents pilot signal position in the time domain, Δ l represents the time domain interval of two pilot signals that time domain is adjacent, k represents pilot signal place subcarrier sequence number, interpolation coefficient α ₁, β ₁calculate by the following method

${\mathrm{\&alpha;}}_1=\frac{l+\mathrm{\&Delta;l}-l^{\&prime;}}{\mathrm{\&Delta;l}},{\mathrm{\&beta;}}_1=\frac{l^{\&prime;}-l}{\mathrm{\&Delta;l}};$ And

To the step of the channel condition information on described Data Position frequency domain index to be estimated in frequency domain, be specially:

H _k′(l)＝α ₂×H _k,S(l)+β ₂×H _k+Δk,S(l),S∈[p,data]；

Wherein k ' represents the frequency domain index of Data Position to be estimated, and have k < k ' < k+ Δ k, k represents pilot signal position in a frequency domain, and Δ k represents the frequency interval of two pilot signals that frequency domain is adjacent, interpolation coefficient α ₂, β ₂calculate by the following method

${\mathrm{\&alpha;}}_2=\frac{k+\mathrm{\&Delta;k}-k^{\&prime;}}{\mathrm{\&Delta;k}},{\mathrm{\&beta;}}_2=\frac{k^{\&prime;}-k}{\mathrm{\&Delta;k}}.$

5. method as claimed in claim 4, is characterized in that, described according to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers, be specially:

${\mathrm{\&sigma;}}_{w,k}^2=\frac{1}{N_c-1}\underset{m=0}{\overset{N_c-2}{\mathrm{\&Sigma;}}}{\left|\right|H_{k,p,\mathrm{LS}}\left(i_{m+1}\right)-H_{k,p,\mathrm{LS}}\left(i_m\right)\left|\right|}_F^2;$

Wherein || H _{k, p, LS}(i _m+1)-H _{k, p, LS}(i _m) || _frepresent Frobenius norm, i _m+1, i _mrepresent adjacent pilot frequencies subcarrier sequence number on same Pilot OFDM symbols, final Noise Variance Estimation value can be expressed from the next:

${\mathrm{\&sigma;}}_w^2=\frac{1}{M}\underset{k=1}{\overset{M}{\mathrm{\&Sigma;}}}{\mathrm{\&sigma;}}_{w,k}^2$

The wherein total number of sub-carriers of M shared by frequency pilot sign.

6. method as claimed in claim 5, is characterized in that, the described statistical channel state information obtained according to described Noise Variance Estimation value in correlation time, and described statistical channel state information comprises described channel average and described channel covariancc, specifically comprises:

According to the correlation time of communication channel, select the sliding window length L of time domain, wherein said sliding window length corresponds to the number of pilot symbols that in sliding window, subframe comprises, according to the correlation bandwidth of channel, at the M that frequency domain comprises from each coherence bandwidth ₀the channel estimation results of individual pilot sub-carrier calculates moment i's the average of individual channel condition information:

And covariance:

Wherein vec () represents matrix vector, () ^hrepresent conjugate transpose.

7. method as claimed in claim 6, is characterized in that, described according to described channel average and channel covariancc, calculates pre-coding matrix, comprises further:

According to the described system channel average estimating to obtain channel covariancc and noise variance the impact of noise is added channel average with channel covariancc Φ _h, obtain new channel average and channel covariancc wherein:

${\stackrel{\&OverBar;}{H}}_{m,n}^{\left(s\right)}=R_n^{-1/2}{\stackrel{\&OverBar;}{H}}_m^{\left(s\right)};$

${\mathrm{\&Phi;}}_{H,n}^{\left(s\right)}=(I_{N_r}\&CircleTimes;R_n^{-1/2}){\mathrm{\&Phi;}}_H^{\left(s\right)}(I_{N_r}\&CircleTimes;R_n^{-1/2});$

Wherein noise covariance matrix

According to and Φ _h,n, compute matrix wherein

$Z_{[i,j]}^{\left(s\right)}=\mathrm{Tr}\left(X_{[i,j]}^{\left(s\right)}\right),x^{\left(s\right)}=({\mathrm{\&Phi;}}_H^{\left(s\right)}+(1/N_r){\left(\right({\stackrel{\&OverBar;}{H}}_m^{\left(s\right)})}^H{\stackrel{\&OverBar;}{H}}_m^{\left(s\right)}\&CircleTimes;I_{N_r})),i=1,...,N_t,j=1,...,N_t;$

Wherein X _{[i, j]}an i-th j dimension of expression X is N _r× N _rsub-square;

By matrix Z ^(s)carry out Cholesky decomposition, obtain matrix L ^(s):

Z ^(s)＝L ^(s)·(L ^(s)) ^H；

To L ^(s)matrix carries out singular value decomposition:

$L^{\left(s\right)}=U_L^{\left(s\right)}{\mathrm{\&Lambda;}}_L^{\left(s\right)}{\left(V_L^{\left(s\right)}\right)}^H;$

Obtain Z ^(s)eigenvalues Decomposition expression formula

$Z^{\left(s\right)}=U_Z^{\left(s\right)}{\mathrm{\&Lambda;}}_Z^{\left(s\right)}{\left(V_Z^{\left(s\right)}\right)}^H=U_L^{\left(s\right)}{\left({\mathrm{\&Lambda;}}_L^{\left(s\right)}\right)}^2{\left(U_L^{\left(s\right)}\right)}^H;$

Pre-coding matrix based on average Signal to Interference plus Noise Ratio criterion calculates by the following method

$B^{\left(s\right)}=U_Z^{\left(s\right)}{\mathrm{\&Lambda;}}_Q^{\left(s\right)};$

Wherein $U_Z^{\left(s\right)}=U_L^{\left(s\right)},{\mathrm{\&Lambda;}}_Q^{\left(s\right)}={\mathrm{\&Lambda;}}_L^{\left(s\right)};$

Wherein $P_i=\max (\mathrm{\&mu;}-\frac{{\mathrm{\&sigma;}}_w^2}{{\mathrm{\&lambda;}}_i},0),i=\mathrm{1,2},...,r,\underset{i=1}{\overset{r}{\mathrm{\&Sigma;}}}{P}_i=\frac{P}{N},({\mathrm{\&lambda;}}_1,{\mathrm{\&lambda;}}_2,...{\mathrm{\&lambda;}}_r)$ It is matrix r nonzero eigenvalue, μ is the water filling factor.P is that transmitting terminal carries all subcarriers of user data and the summation of transmitting antenna transmitted power.

8. the method for claim 1, is characterized in that, the signal that after described employing precoding, receiving terminal receives is:

Y _k(l)＝H _k(l)·B·X _k(l)+W _k(l)

Wherein $Y_k\left(l\right)={\tilde{Y}}_k^T\left(l\right),X_k\left(l\right)={\tilde{X}}_k^T\left(l\right),H_k\left(l\right)=H_k^T\left(l\right),W_k\left(l\right)={\tilde{W}}_k^T\left(l\right);$

The channel condition information H obtaining non-pilot character position is estimated by time domain and frequency domain two-dimensional linear interpolation _kl (), in conjunction with described pre-coding matrix B, determines the channel matrix H of equivalence _e,k(l), namely

H _E,k(l)＝Β·H _k(l)

According to the combined channel matrix H of equivalence _e,kl () take least mean-square error as criterion, receiving terminal can carry out separating the joint-detection that precoding and MIMO detect and receive

${\hat{X}}_k\left(l\right)=G_{\mathrm{MMSE}}\&CenterDot;Y_k\left(l\right)={[H_{E,k}^H\left(l\right)H_{E,k}\left(l\right)+{\mathrm{\&sigma;}}_n^2{I}_{N_t}]}^{-1}{H}_{E,k}^H\left(l\right)Y_k\left(l\right)$

9. a pre-coding apparatus, is characterized in that, described device comprises:

Receiving element, for the pilot signal that receiving end/sending end sends;

Estimation unit, for carrying out channel condition information estimation according to uplink and downlink pilot signal, to obtain channel average and channel covariancc;

Computing unit, for according to described channel average and channel covariancc, calculates pre-coding matrix;

Feedback unit, for giving described transmitting terminal when uplink and downlink channel does not meet symmetry by described precoding matrix feedback, so that described transmitting terminal is need the data-signal sent to carry out precoding according to described pre-coding matrix.

Decoding unit, for according to described pre-coding matrix, carries out solution precoding to the data-signal received.

10. pre-coding apparatus as claimed in claim 9, is characterized in that, described estimation unit specifically for:

The channel status pilot signal received being carried out to pilot frequency symbol position is estimated, to obtain the LS channel estimation result based on frequency pilot sign within the correlated time of channel;

To described acquisition LS channel estimation result based on frequency pilot sign within the correlated time of channel, perform linear interpolation, with obtain data to be estimated channel estimation results on sub-carriers;

According to the LS channel estimation result based on frequency pilot sign in described correlation time, calculate described frequency pilot sign Noise Variance Estimation value on sub-carriers;

Obtain the statistical channel state information in correlation time according to described Noise Variance Estimation value, described statistical channel state information comprises described channel average and described channel covariancc.