CN101826946A - Method for preprocessing information in multi-input multi-output system - Google Patents

Abstract

The invention relates to a method for preprocessing information in a multi-input multi-output system, which belongs to the technical field of mobile communication. The method comprises the following steps of: obtaining channel information at each subcarrier at the current time; identifying the number of layers of sending data; selecting a codebook for a preprocessing matrix, to obtain an equivalent channel matrix consisting of a channel information matrix and the preprocessing matrix; obtaining the sum of equivalent signal-to-noise radio and symbol error rates on each layer of data, and using the sum of the symbol error rates as the weight of the selected preprocessing matrix; obtaining the weight of other preprocessing matrixes in the selected codebook respectively; and selecting the preprocessing matrix with the minimum weight, and feeding the index of the matrix with the minimum weight in the codebook back to a sending end which uses the selected preprocessing matrix to preprocess the sending data in the next sending process so as to complete the sending process. The invention has a low system symbol error rate, is applicable to any conditions, and overcomes the disadvantage that the prior art can not work when the preprocessing matrix is a unitary matrix.

Description

The preprocess method of information in the multi-input multi-output system

Technical field

What the present invention relates to is the method in a kind of mobile communication technology field, specifically is the preprocess method of information in a kind of multi-input multi-output system.

Background technology

Extensively adopted MIMO (Multi-Input Multi-Output, multiple-input and multiple-output) technology in the mobile communication system of a new generation.Preconditioning technique then is to realize the closed loop multiple-input and multiple-output, improves a necessary link of systematic function.Supposing has N in the multi-input multi-output system _TTransmit antennas and N _RRoot reception antenna, transmitting terminal are launched K flow data (K≤N at every turn _T), launch then that data x needs earlier and a preconditioning matrix W multiplies each other is mapped to N _TOn the transmit antennas, and then launch.Thereby preconditioning matrix W need design according to the channel information of system and improve systematic function.

In frequency division duplex system,,, and feed back to transmitting terminal so preconditioning matrix W is generally definite by receiving terminal because channel does not have reciprocity.But in real system, feed back accurate preconditioning matrix W and need take a large amount of uplink feedback channel resources.In order to reduce feedback overhead, adopted feedback scheme in LTE (Long Time Evolution, the Long Term Evolution) agreement under the 3GPP (3rd Generation Partnership Project, third generation partner program) based on code book.In the pretreatment mode based on code book, the channel information that the receiving terminal utilization estimates carries out the selection of preconditioning matrix in predetermined preliminary treatment code book, and the sequence number of selected matrix is fed back to transmitting terminal.Transmitting terminal is determined preconditioning matrix again according to the sequence number of receiving terminal feedback from code book.How selecting an appropriate pretreatment matrix from code book is a key issue that promotes the multi-input multi-output system performance.

Find through retrieval existing document, being published in " IEEE TRANSACTION ON INFORMATION THEROTY " goes up name and is called the paper of " Limited Feedback Unitary Precoding for Spatial Multiplexing Systems (Limited Feedback in the SDM system merges precoding) " and Chinese patent application and number is: 200810116917.2, and name is called: the method for choosing pre-coding matrix in the multi-input multi-output system, equipment and system " document in provided several method: it is the method for the minimum singular value of maximization equivalent channel that first method is selected the criterion of preconditioning matrix.Wherein, equivalent channel Heff is the product of actual channel H and preconditioning matrix W.The criterion of second method is the mean square error that minimizes decoded signal and send signal.The criterion of the third method is the capacity of maximization equivalent channel.The criterion of the 4th kind of method is to select from code book and the preconditioning matrix of desirable preconditioning matrix Euclidean distance minimum.Because these four kinds of methods are not all not directly optimized at the decoding characteristic of system, so their decoding performance all is not unusual ideal.In addition, first three methods is inoperative at preconditioning matrix during for unitary matrice, therefore, certain limitation is arranged when in use.

Summary of the invention

The objective of the invention is to overcome the prior art above shortcomings, the preprocess method of information in a kind of multi-input multi-output system is provided.The present invention directly adopts the error sign ratio of system as the criterion of selecting preconditioning matrix, makes the error sign ratio of system be able to further decline, and under any circumstance can use do not have limitation.

The present invention is achieved by the following technical solutions, may further comprise the steps:

Step 1, receiving terminal carries out channel estimation process to the pilot data that receives, and obtains the channel information at each subcarrier place of current time, and each channel information is a N _R* N _TComplex matrix H, wherein: N _TBe the number of transmitting antenna, N _RIt is the number of reception antenna.

Described channel estimation process is: at first adopt least square method (Least Square at the pilot sub-carrier place, LS) obtain channel estimation value, adopt linear minimum mean-squared error (Linear Minimum Mean Square Error at frequency domain then, LMMSE) the algorithm interpolation goes out the channel estimation value at data subcarrier place, adopts linear interpolation to obtain the channel estimation value at all data subcarrier places in time domain at last.

Step 2 according to the size of the order l of channel information matrix H, determines to send the number of plies K of data, K≤l.

Step 3 is selected any one preconditioning matrix code book from existing communication standard, selecting a line number from this code book is N _T, columns is the preconditioning matrix F of K, channel information matrix H and this preconditioning matrix F are multiplied each other obtains equivalent channel matrix H _Eff

Step 4 according to the balanced matrix and the equivalent channels information of receiving terminal equalizer, obtains the equivalent signal-to-noise ratio of every layer data.

Described equivalent signal-to-noise ratio is meant:

${\mathrm{SNR}}_k=\frac{{\mathrm{\&delta;}}_{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HDA0000021445430000011.png,HDA0000021445430000012.png"\; state="\{\{state\}\}">s</figure-callout>}}^2{\left|g^k{h}_k\right|}^2}{{{\mathrm{\&delta;}}_n}^2{\left({\mathrm{GG}}^H\right)}_{\mathrm{kk}}+{\mathrm{\&delta;}}_{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HDA0000021445430000011.png,HDA0000021445430000012.png"\; state="\{\{state\}\}">s</figure-callout>}}^2\underset{j=1,j\&NotEqual;k}{\overset{k}{\mathrm{\&Sigma;}}}{\left|g^k{h}_j\right|}^2},$

Wherein:

SNR _kBe the equivalent signal-to-noise ratio of k layer data, G represents balanced matrix, δ _n ²Expression receives the power of white noise in the data, δ _s ²Be signal power, g ^kFor the k of balanced matrix G is capable, h _kBe the equivalent channels information matrix H _EffK row, () _KkThe k of representing matrix is capable, the element of k row, and I is a unit matrix.

Step 5 according to the equivalent signal-to-noise ratio of modulation system and every layer data of transmitting terminal, obtains the error sign ratio of every layer data, and the error sign ratio sum of all layer data is exactly the weight of this preconditioning matrix.

Described error sign ratio is meant:

Wherein: SER _KBe the error sign ratio of k layer data, SNR _kBe the equivalent signal-to-noise ratio of k layer data, M is first number of quadrature amplitude modulation.

Step 6, selecting remaining each line number from the preconditioning matrix code book is N _T, columns is the preconditioning matrix of K, repeating step three obtains the weight of each preconditioning matrix respectively to step 5.

Step 7 is selected the preconditioning matrix of weight minimum, and its index in code book is fed back to transmitting terminal, and transmitting terminal uses in emission process next time chooses preconditioning matrix to carry out preliminary treatment to sending data, and finishes emission process.

Compare with prior art, the invention has the beneficial effects as follows: directly adopt the error sign ratio criterion to select preconditioning matrix, have the error sign ratio of lower system, improved the performance of system; And the present invention under any circumstance can be suitable for, and overcome the shortcoming that prior art can not play a role during for unitary matrice at preconditioning matrix.

Description of drawings

Fig. 1 is the error sign ratio performance schematic diagram relatively of preconditioning matrix embodiment method and art methods when being unitary matrice;

Fig. 2 is the performance of BER schematic diagram relatively of preconditioning matrix embodiment method and art methods when being unitary matrice;

Fig. 3 is the error sign ratio performance schematic diagram relatively of preconditioning matrix embodiment method and art methods when being non-unitary matrice;

Fig. 4 is the performance of BER schematic diagram relatively of preconditioning matrix embodiment method and art methods when being non-unitary matrice.

Embodiment

Below in conjunction with accompanying drawing method of the present invention is further described: present embodiment is being to implement under the prerequisite with the technical solution of the present invention, provided detailed execution mode and concrete operating process, but protection scope of the present invention is not limited to following embodiment.

Embodiment

4 transmit antennas (N are arranged in the multiple-input and multiple-output frequency division duplex system in the present embodiment _T=4), 4 reception antenna (N _R=4), the modulation system of transmitting terminal is the quaternary quadrature amplitude modulation, and receiving terminal is the least mean-square error equalizer, and concrete pre-treatment step is as follows:

Step 1, receiving terminal is according to the channel information at each subcarrier place of existing channel estimation methods acquisition current time, and each channel information is one 4 * 4 a complex matrix.

To be transmitting terminal send pilot data on the frequency to described channel estimation methods when specific, and receiving terminal at first obtains the channel information at pilot tone point place according to the signal that receives, estimate the channel information of other all subcarriers then according to the channel estimating criterion.

Step 2 according to the size of the order l of channel information matrix H, determines to send the number of plies K of data, K≤l.

The order l of channel information matrix H is 4 in the present embodiment, is the characteristic that also can be suitable under the unitary matrice in order to embody the present invention at preconditioning matrix, the number of plies K=4 of the transmission data x that present embodiment is selected.

Step 3 is selected any one preconditioning matrix code book from existing communication standard, select a line number and be 4 from this code book, columns is 4 preconditioning matrix F, and channel information matrix H and this preconditioning matrix F are multiplied each other obtains equivalent channel matrix H _Eff

Present embodiment is selected the linear predictive coding matrix code book that adopts in LTE (Long Time Evolution, the Long Term Evolution) standard for use, and line number is 4 in this code book, and columns is that 4 preconditioning matrix has 16, and is unitary matrice.

Step 4 according to the balanced matrix and the equivalent channels information of receiving terminal equalizer, obtains the equivalent signal-to-noise ratio of every layer data.

Described equivalent signal-to-noise ratio is meant:

${\mathrm{SNR}}_k=\frac{{\mathrm{\&delta;}}_{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HDA0000021445430000011.png,HDA0000021445430000012.png"\; state="\{\{state\}\}">s</figure-callout>}}^2{\left|g^k{h}_k\right|}^2}{{{\mathrm{\&delta;}}_n}^2{\left({\mathrm{GG}}^H\right)}_{\mathrm{kk}}+{\mathrm{\&delta;}}_{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="HDA0000021445430000011.png,HDA0000021445430000012.png"\; state="\{\{state\}\}">s</figure-callout>}}^2\underset{j=1,j\&NotEqual;k}{\overset{k}{\mathrm{\&Sigma;}}}{\left|g^k{h}_j\right|}^2},$

Wherein: SNR _kBe the equivalent signal-to-noise ratio of k layer data, G represents balanced matrix, δ _n ²Expression receives the power of white noise in the data, δ _s ²Be signal power, g ^kFor the k of balanced matrix G is capable, h _kBe the equivalent channels information matrix H _EffK row, () _KkThe k of representing matrix is capable, the element of k row.

In the present embodiment least mean-square error equalizer, so G=(H _Eff ^HH _Eff+ δ _n ²I) ^-1H _Eff ^H, wherein: I is a unit matrix.

Step 5 according to the equivalent signal-to-noise ratio of modulation system and every layer data of transmitting terminal, obtains the error sign ratio of every layer data, and the error sign ratio sum of all layer data is exactly the weight of this preconditioning matrix.

The modulation system quaternary quadrature amplitude modulation of transmitting terminal in the present embodiment, so:

${\mathrm{SER}}_K=1-{[1-\frac{1}{\sqrt{2\mathrm{\&pi;}}}\underset{\sqrt{{\mathrm{SNR}}_K}}{\overset{+\&infin;}{\&Integral;}}{e}^{(-\frac{{\mathrm{<figure-callout\; id="2"\; label="t"\; filenames="HDA0000021445430000011.png,HDA0000021445430000012.png"\; state="\{\{state\}\}">t</figure-callout>}}^2}{2})}\mathrm{dt}]}^2,$

Wherein: SER _KBe the error sign ratio of k layer data, SNR _kIt is the equivalent signal-to-noise ratio of k layer data.

Step 6 selects from the preconditioning matrix code book that remaining 15 line number is 4, columns is 4 preconditioning matrix, and repeating step three obtains the weight of each preconditioning matrix respectively to step 5.

Step 7 is selected the preconditioning matrix of weight minimum, and its index in code book is fed back to transmitting terminal, and transmitting terminal uses in emission process next time chooses preconditioning matrix to carry out preliminary treatment to sending data, and finishes emission process.

Under the condition identical (preconditioning matrix is a unitary matrice) with present embodiment, adopt respectively the preconditioning matrix method of existing four kinds of methods and picked at random obtains in present embodiment method, the background technology error sign ratio relatively schematic diagram as shown in Figure 1, relatively schematic diagram is as shown in Figure 2 for the bit error rate that obtains.By Fig. 1 and Fig. 2 as seen, error sign ratio and performance of BER that the four kinds of methods of prior art and the method for picked at random preconditioning matrix obtain are suitable, and the performance of the error sign ratio of present embodiment method and bit error rate obviously improves, so the decoding performance of present embodiment method is best.

When preconditioning matrix is not under the situation of unitary matrice, under identical condition, adopt respectively the preconditioning matrix method of existing four kinds of methods and picked at random obtains in present embodiment method (this moment number of data layers be 2), the background technology error sign ratio relatively schematic diagram as shown in Figure 3, relatively schematic diagram is as shown in Figure 4 for the bit error rate that obtains.By Fig. 3 and Fig. 4 as seen, at preconditioning matrix is under the situation of non-unitary matrice, the error sign ratio of present embodiment method and the performance of the bit error rate still first three methods than prior art are superior, and the superiority of comparing present embodiment method then with the method for picked at random preconditioning matrix is more obvious.

Claims (4)

1. the preprocess method of information in the multi-input multi-output system is characterized in that, may further comprise the steps:

Step 1, receiving terminal carries out channel estimation process to the pilot data that receives, and obtains the channel information at each subcarrier place of current time, and each channel information is a N _R* N _TComplex matrix H, wherein: N _TBe the number of transmitting antenna, N _RIt is the number of reception antenna;

Step 2 according to the size of the order l of channel information matrix H, determines to send the number of plies K of data, K≤l;

Step 3 is selected any one preconditioning matrix code book from existing communication standard, selecting a line number from this code book is N _T, columns is the preconditioning matrix F of K, channel information matrix H and this preconditioning matrix F are multiplied each other obtains equivalent channel matrix H _Eff

Step 4 according to the balanced matrix and the equivalent channels information of receiving terminal equalizer, obtains the equivalent signal-to-noise ratio of every layer data;

Step 5 according to the equivalent signal-to-noise ratio of modulation system and every layer data of transmitting terminal, obtains the error sign ratio of every layer data, and the error sign ratio sum of all layer data is exactly the weight of this preconditioning matrix;

Step 6, selecting remaining each line number from the preconditioning matrix code book is N _T, columns is the preconditioning matrix of K, repeating step three obtains the weight of each preconditioning matrix respectively to step 5;

Step 7 is selected the preconditioning matrix of weight minimum, and its index in code book is fed back to transmitting terminal, and transmitting terminal uses in emission process next time chooses preconditioning matrix to carry out preliminary treatment to sending data, and finishes emission process.

2. the preprocess method of information in the multi-input multi-output system according to claim 1, it is characterized in that, channel estimation process described in the step 1 is: at first adopt least square method to obtain channel estimation value at the pilot sub-carrier place, adopt linear minimum mean-squared error algorithm interpolation to go out the channel estimation value at data subcarrier place at frequency domain then, adopt linear interpolation to obtain the channel estimation value at all data subcarrier places in time domain at last.

3. the preprocess method of information is characterized in that in the multi-input multi-output system according to claim 1, and the equivalent signal-to-noise ratio described in the step 4 is meant:

${\mathrm{SNR}}_k=\frac{{\mathrm{\&delta;}}_s^2{\left|g^k{h}_k\right|}^2}{{{\mathrm{\&delta;}}_n}^2{\left({\mathrm{GG}}^H\right)}_{\mathrm{kk}}+{\mathrm{\&delta;}}_s^2\underset{j=1,j\&NotEqual;k}{\overset{K}{\mathrm{\&Sigma;}}}{\left|g^k{h}_j\right|}^2},$

Wherein:

SNR _kBe the equivalent signal-to-noise ratio of k layer data, G represents balanced matrix, δ _n ²Expression receives the power of white noise in the data, δ _s ²Be signal power, g ^kFor the k of balanced matrix G is capable, h _kBe the equivalent channels information matrix H _EffK row, () _KkThe k of representing matrix is capable, the element of k row, and I is a unit matrix.

4. the preprocess method of information is characterized in that in the multi-input multi-output system according to claim 1, and the error sign ratio described in the step 5 is meant:

Wherein: SER _KBe the error sign ratio of k layer data, SNR _kBe the equivalent signal-to-noise ratio of k layer data, M is first number of quadrature amplitude modulation.

Images