CN101257324B - Linear combined channel estimation method in TD-SCDMA system - Google Patents

Abstract

The invention discloses a linearity combined channel estimating method in TD-SCDMA system, comprising that: signals of K districts are in real time received by a receiver, a data divide unit separates the receiving signal x of P code from the receiving signal r as the real time input of a channel estimater to perform the Fourier transform, the Fourier transform of basic intermediate code in K districts is used as the constant input value to compute the least square solution of channel impulse response of each district, finally the result of channel estimate h and the data symbol are together used in the combined detection to obtain the estimate of the sending data symbol, and reestimate after selecting a taps.

Description

Linear combined channel estimation method in the TD-SCDMA system

Technical field

The present invention relates to the channel estimation methods of receiver in a kind of wireless communication system, particularly the linear combined channel estimation method in a kind of TD-SCDMA system.

Background technology

Next TD-SCDMA system up-link data will carry out channel estimating and joint-detection after the base station receives by antenna.Associated detection technique then refers to and takes full advantage of multiple access interference (MA I), a kind of signal separation techniques of within the step all users' signal all being separated.The very suitable use associated detection technique of TD-SCDMA system, because code channel number is limited and be confined on the most suitable number in this system, and all users in the same time slot are synchronous.Associated detection technique has become the focus in the 3G (Third Generation) Moblie technology.The associated detection technique that adopts in the TD-SCDMA system is on the basis of traditional detection technology, take full advantage of and cause all subscriber signals that multiple access disturbs and the prior information of multipath thereof, the separation of subscriber signal being used as the joint detection process that is mutually related of a unification finishes, thereby has good interference free performance, reduced the requirement of system to the power control precision, therefore can effectively utilize the uplink spectrum resource more, improve power system capacity significantly.

Channel estimating is a vital part in joint detection process, and it will influence the testing process performance in many aspects.The accuracy of the detection information of combined detector output is directly related with the quality of channel estimating, and how obtaining accurately efficiently, each user's channel impulse response is the key that follow-up base band algorithm is realized.As a rule, channel estimating is more accurate, just can obtain more good systematic function.

In the TD-SCDMA system under the identical networking situation, the receiver channel method of estimation generally has two big class, i.e. single cell channel estimation method and multiple cell combined channel estimation methods.

Single cell channel estimation method is exactly to continue to use Steiner estimator in the past, will be used as to disturb from the signal of other residential quarter and handle, and obviously its performance seriously is subject to interference, usually can't operate as normal in the TD-SCDMA of identical networking system.

Summary of the invention

The purpose of this invention is to provide the linear combined channel estimation method in a kind of TD-SCDMA system, to improve performance for estimating channel, make it to be applicable to the TD-SCDMA system of identical networking.

For solving the problems of the technologies described above, this multiple cell combined channel estimation method is applied to channel estimating with multiuser detection, to be used as different users from the signal of different districts, and adopt suitable multiuser detection partly to eliminate the phase mutual interference between the different districts.Technical scheme of the present invention is:

Linear combined channel estimation method in a kind of TD-SCDMA system is characterized in that this method may further comprise the steps:

(1) the real-time signal that receives K residential quarter from receiver;

(2) data splitting block is isolated the reception signal x of P chip of corresponding intermediate code part from the reception signal r of a time slot, as the real-time input of channel estimator;

(3) basic midamble code according to K the residential quarter that participates in the combined channel estimation constructs a P * KP rank complex matrix

The cyclic shift of classifying k residential quarter basic midamble code as;

(4) the Fourier transform FFT{M of the basic midamble code of K residential quarter of calculating participation combined channel estimation _k(:, 1) }, M wherein _k(:, 1) representing matrix M _kFirst row, i.e. basic midamble code itself, being expressed as matrix form is Λ _k=[λ _{K, 1}λ _{K, P}] ^T=FFT{M _k(:, 1), p=1 ..., P, storage result of calculation Λ _kFixedly input value as channel estimator;

(5) make S=MM ^H, characteristic value Λ=[λ of calculating S ₁λ _P] ^T, wherein ^HThe expression complex-conjugate transpose, ^TThe expression transposition, computational methods are ${\mathrm{\λ}}_p=\underset{k}{\mathrm{\Σ}}{\left|{\mathrm{\λ}}_{k,p}\right|}^2,$ λ _{K, p}Be M _kCharacteristic value, storage vector Λ is as the fixedly input value of channel estimator;

(6) calculate the Fourier transform FFT{x} that receives signal x;

(7) calculate the least square solution of estimation of the channel impulse response of each residential quarter ${\hat{h}}_k=K*\mathrm{IFFT}\{\mathrm{conj}\left\{{\mathrm{\Λ}}_k\right\}.*\mathrm{FFT}\left\{x\right\}./\mathrm{\Λ}\},$ Wherein * represents multiplication, and IFFT{} represents contrary fast fourier transform, and conj{} represents complex conjugate, and .* represents that corresponding element multiplies each other, and FFT{} represents fast fourier transform ./expression corresponding element is divided by;

(8) based on the result of above-mentioned channel estimating, selected tap is used for estimating again;

(9) selected tap forms a new channel impulse response vector h ', new matrix M of row formation of extraction correspondence from matrix M ' and, the least square solution of channel impulse response estimation is

Wherein

Expression Moore-Penrose pseudoinverse;

(10) h as a result with channel estimating is applied to joint-detection with the data symbol part that receives signal r, to obtain sending the estimation of data symbol.

The criterion of the selected tap when estimating again in the described step 9 is based on the hypothesis to the multidiameter delay expansion.

The condition that described M ' for estimation again is a row non-singular matrix is that the columns of M ' is less than or equal to P.

Described P=128, set basis is referring to 3GPP TS25.221 physical channel and transmission channel mapping (TDD) the 5A.2.3 joint to physical channel.

Described reception signal x can be expressed as

x＝Mh+n

Wherein

Be the channel impulse response of k residential quarter,

Be noise;

Minimize the Euclidean norm of evaluated error The least square solution that can obtain the estimation of channel impulse response is

$\hat{h}={\mathrm{KM}}^H{\left({\mathrm{MM}}^H\right)}^{-1}x$

Wherein ^-1Expression is inverted.

Described new matrix M for estimation again ' be a row non-singular matrix, the least square solution of channel impulse response estimation is $\hat{h}\′=K{(M^{\′H}M\′)}^{-1}{M}^{\′H}x.$

Adopt shortcut calculation in the described estimation again: calculate M ' ^HIn M ' time, calculate M by the method for fast fourier transform ^HM therefrom extracts desired element then; Calculate (M ' ^HM ') ^-1The time, earlier with the partitioning of matrix, carry out Cholesky then and decompose; Calculate M ' ^HDuring x, calculate M by the method for fast fourier transform equally ^HX therefrom extracts desired element then.

Compared with prior art, beneficial effect of the present invention is as follows:

In the channel estimation results of technical solution of the present invention, arbitrary residential quarter accesses supression preferably to the Recombinant Interferon of other residential quarter, and the channel estimation window of the user's correspondence in the residential quarter can reflect the multi-path environment that this user experiences more exactly.Therefore, compare with single cell channel estimated result, performance for estimating channel of the present invention has greatly improved.

Description of drawings

Fig. 1 is a typical TD-SCDMA structure of time slot schematic diagram.

Fig. 2 is channel estimating and joint-detection schematic diagram.

Fig. 3 is the simulation result figure of single cell channel estimation method and multiple cell combined channel estimation method.

Embodiment

The present invention will be further described below in conjunction with drawings and Examples, but should not limit protection scope of the present invention with this.

As shown in Figure 1, provide a typical TD-SCDMA structure of time slot schematic diagram.A time slot is made of at interval the intermediate code of the data symbol of two sections 352 chips, 144 chips and the protection of 16 chips, altogether 864 chips.

In general different basic midamble codes is used in different residential quarters, and basic midamble code generates intermediate code and saves referring to 3GPP TS25.221 physical channel and transmission channel mapping (TDD) 5A.2.3 to physical channel.

Consider one and can receive the situation that K residential quarter transmits simultaneously.From receiver, receive the signal of K residential quarter in real time.

Fig. 2 provides a channel estimating and joint-detection schematic diagram, as shown in Figure 2, the reception signal r of a time slot comprises 864 chips, pass through data splitting block, intercepting is the reception signal x of corresponding intermediate code part wherein, namely 496 chips of the 369th chip to the have P chip altogether as the real-time input of channel estimator, get P=128 here.

The reception signal x of corresponding intermediate code part can be expressed as

x＝Mh+n

Wherein

Be that M is a P * KP rank complex matrix,

The cyclic shift of classifying k residential quarter basic midamble code (Basic Midamble Code) as.

Be the channel impulse response of k residential quarter, wherein ^TThe expression transposition.

Be noise.

The least square of the estimation of channel impulse response (LS, Least Square) solution is

$\hat{h}={\mathrm{KM}}^H{\left({\mathrm{MM}}^H\right)}^{-1}x$

Wherein ^HThe expression complex-conjugate transpose, ^-1Expression is inverted.

Make S=MM ^H, M so _kWith S all be circular matrix (Circulant Matrices).According to the character of circular matrix, M _kCharacteristic value be λ _{K, p}, p=1 ..., P is expressed as matrix form, i.e. Λ _k=[λ _{K, 1}λ _{K, P}] ^T=FFT{M _k(:, 1) }, Mk (:, 1) representing matrix M wherein _kFirst row, i.e. basic midamble code itself.Similarly, the characteristic value of S can be expressed as Λ=[λ ₁λ _P] ^T, wherein ${\mathrm{\λ}}_p=\underset{k}{\mathrm{\Σ}}{\left|{\mathrm{\λ}}_{k,p}\right|}^2,$ p＝1，…，P。

Store described result of calculation Λ _k, Λ is as the fixedly input value of channel estimator.Calculate the Fourier transform FFT{x} that receives signal x.

Therefore the least square solution of the estimation of channel impulse response also can basis

${\hat{h}}_k=K*\mathrm{IFFT}\{\mathrm{conj}\left\{{\mathrm{\Λ}}_k\right\}.*\mathrm{FFT}\left\{x\right\}./\mathrm{\Λ}\}$ Calculate channel estimation results.Wherein * represents multiplication, and IFFT{} represents contrary fast fourier transform, and conj{} represents complex conjugate, and .* represents that corresponding element multiplies each other, and FFT{} represents fast fourier transform ./expression corresponding element is divided by.

Based on above-mentioned channel estimation results, according to certain criterion, for example based on the hypothesis to the multidiameter delay expansion, select some may exist the tap of multipath signal to estimate again.Selected tap can form a new channel impulse response vector h ', new matrix M of row formation of extraction correspondence from matrix M ' and, the least square solution of channel impulse response estimation can be write as so

Wherein Expression Moore-Penrose pseudoinverse.

When above-mentioned new matrix M ' columns when being less than or equal to P, M ' is a row non-singular matrix, following formula can be write as so

$\hat{h}\′=K{(M^{\′H}M\′)}^{-1}{M}^{\′H}x$

In the above-mentioned process of estimating again, can use some shortcut calculations.For example, calculating M ' ^HIn M ' time, can calculate M by the method for fast fourier transform ^HM therefrom extracts desired element then; Calculate (M ' ^HM ') ^-1The time, can carry out Cholesky then and decompose earlier with the partitioning of matrix; Calculating M ' ^HIn the time of x, can calculate M by the method for fast fourier transform equally ^HX therefrom extracts desired element then.

As shown in Figure 2, the result of channel estimating will be applied to joint-detection, obtain sending the estimation of data symbol with the data symbol part that receives signal.

By an emulation good effect of the technical program is described, simulated conditions as shown in Table 1 and Table 2, set basis is referring to 3GPP TS25.102 subscriber equipment wireless transmission and reception (TDD).

Table 1 simulated conditions

Table 2 multipath fading environments propagation conditions

Based on above-mentioned simulated conditions, Fig. 3 has provided the Matlab simulation result figure of single cell channel estimation method and multiple cell combined channel estimation method, and abscissa is tap among the figure, and ordinate is intensity.Two figure in the left side are single cell channel results estimated, and last figure is single cell channel estimated result of residential quarter 1, and figure below is single cell channel estimated result of residential quarter 2; Two results that figure is technical solution of the present invention in the right, last figure is the linear combined channel estimation results in many residential quarters of residential quarter 1, figure below is the multi-plot joint linear channel estimated result of residential quarter 2.

Obviously, in the channel estimation results of single residential quarter, residential quarter 2 has been subjected to the interference from residential quarter the last 1, and in the channel estimation results of technical solution of the present invention, interference from residential quarter 1 is restrained preferably, can see the channel estimation window of user's correspondence in the residential quarter 2 more clearlyly, reflect the multi-path environment of this user's experience more exactly.Therefore, compare with single cell channel estimated result, the present invention has bigger performance improvement.

The above is preferred embodiment of the present invention only, is not to limit practical range of the present invention.Be that all equivalences of doing according to the content of the present patent application claim change and modification, all should be technology category of the present invention.

Claims (6)

1. the linear combined channel estimation method in the TD-SCDMA system is characterized in that this method may further comprise the steps:

(1) the real-time signal that receives K residential quarter from receiver;

(2) data splitting block is isolated the reception signal x of P chip of corresponding intermediate code part from the reception signal r of a time slot, as the real-time input of channel estimator;

(3) basic midamble code according to K the residential quarter that participates in the combined channel estimation constructs a P * KP rank complex matrix M=[M ₁M ₂M _K] ∈ C ^{P * KP}, M _k∈ C ^{P * P}The cyclic shift of classifying k residential quarter basic midamble code as; Wherein: k=1 ..., K;

Receiving signal x is expressed as

x＝Mh+n

Wherein $h={\left[\begin{array}{cccc}{h}_1^T& h_2^T& ...& h_K^T\end{array}\right]}^T\∈C^{\mathrm{KP}\×1},$ h _k∈ C ^{P * 1}Be the channel impulse response of k residential quarter, n ∈ C ^{P * 1}Be noise;

The least square solution of the estimation of channel impulse response is

$\hat{h}={\mathrm{KM}}^H{\left({\mathrm{MM}}^H\right)}^{-1}x$

Wherein ^-1Expression is inverted, ^HThe expression complex-conjugate transpose;

(4) the Fourier transform FFT{M of the basic midamble code of K residential quarter of calculating participation combined channel estimation _k(:, 1) }, M wherein _k(:, 1) representing matrix M _kFirst row, i.e. basic midamble code itself, being expressed as matrix form is Λ _k=[λ _{K, 1}λ _{K, P}] ^T=FFT{M _k(:, 1), p=1 ..., P, storage result of calculation Λ _kFixedly input value as channel estimator;

(5) make S=MM ^H, characteristic value Λ=[λ of calculating S ₁λ _P] ^T, wherein, ^TThe expression transposition, computational methods are λ _{K, p}Be M _kCharacteristic value, storage vector Λ is as the fixedly input value of channel estimator;

(6) calculate the Fourier transform FFT{x} that receives signal x;

(7) calculate the least square solution of estimation of the channel impulse response of each residential quarter ${\hat{h}}_k=K*\mathrm{IFFT}\left\{\mathrm{conj}\right\{{\mathrm{\Λ}}_k\}.*\mathrm{FFT}\{x\}./\mathrm{\Λ}\},$ Wherein * represents multiplication, and IFFT{} represents contrary fast fourier transform, and conj{} represents complex conjugate, and .* represents that corresponding element multiplies each other, and FFT{} represents fast fourier transform ./expression corresponding element is divided by;

(8) based on the result of above-mentioned channel estimating, selected tap is used for estimating again;

(9) selected tap forms a new channel impulse response vector h ', new matrix M of row formation of extraction correspondence from matrix M ' and, the least square solution of channel impulse response estimation is

Wherein

Expression Moore-Penrose pseudoinverse;

(10) h as a result with channel estimating is applied to joint-detection with the data symbol part that receives signal r, to obtain sending the estimation of data symbol.

2. the linear combined channel estimation method in the TD-SCDMA according to claim 1 system, the criterion that it is characterized in that described selected tap when estimating again is based on the hypothesis to the multidiameter delay expansion.

3. the linear combined channel estimation method in the TD-SCDMA according to claim 1 system is characterized in that described is that the condition of a row non-singular matrix is that the columns of M ' is less than or equal to P for the M ' that estimates again.

4. according to the linear combined channel estimation method in claim 1 or the 3 described TD-SCDMA systems, it is characterized in that described P=128.

5. according to the linear combined channel estimation method in claim 1 or the 3 described TD-SCDMA systems, it is characterized in that describedly for the new matrix M of estimating again ' be a row non-singular matrix, the least square solution of channel impulse response estimation is

6. the linear combined channel estimation method in the TD-SCDMA according to claim 5 system is characterized in that describedly adopting shortcut calculation in estimating again: calculate M ' ^HIn M ' time, calculate M by the method for fast fourier transform ^HM therefrom extracts desired element then; Calculate (M ' ^HM ') ^-1The time, earlier with the partitioning of matrix, carry out Cholesky then and decompose; Calculate M ' ^HDuring x, calculate M by the method for fast fourier transform equally ^HX therefrom extracts desired element then.

Images