CN102148782A - Method and device for estimating multi-cell channels - Google Patents

Abstract

The invention provides a method for estimating high-accuracy multi-cell channels in a code division multiple access (CDMA) wireless communication system. When a channel estimation algorithm in the CDMA wireless communication system such as a time division-synchronization code division multiple access (TD-SCDMA) system is extended to multi-cell channel estimation, because the channel estimation accuracy for neighboring cells is low so as to seriously influence the performances of a multi-cell joint detection algorithm, in the invention, a channel estimation value using matched filtering and minimum mean squared error (MMSE) is put forward by changing the structure of a channel estimation matrix, and a high-accuracy channel estimation algorithm for multi-cell joint detection is put forward in accordance with interference reconfiguration and a parallel elimination principle. On the basis of the method, by combining with strong path selection, a more accurate high-accuracy channel estimation algorithm based on strong path interference configuration is put forward. Therefore, the algorithm estimation accuracy is improved, and the performances of the CDMA wireless communication system under the multi-cell joint detection are enhanced effectively.

Description

A kind of multi-cell channel estimation method and device

Technical field

The present invention relates to a kind of wireless communication technology, relate in particular to a kind of detection method of wireless signal, relate to the link multi-user test method in a kind of code division multi-address radio communication system in particular.

Background technology

Along with the appearance and the popularization and application of communication in the personal communication field of commercial mobile radio system, the capacity that how to improve mobile radio system under the finite wireless resources condition is one of important topic in the wireless communication technology evolution always.

In commercial mobile radio system the earliest, promptly generally be used as in the simulation standard system of first generation mobile radio system, adopted the access mode of frequency division multiple access (FDMA).Under the frequency division multiple access access way, different user in the system in sub-district is distinguished by different frequency ranges, and the different districts in the system is configured to different frequency ranges or multiplexing identical frequency range according to the frequency planning of system with channeling (Frequency reuse) mode.In the second generation mobile radio system that is called as global system for mobile communications (GSM), adopted the wireless multiple access access way of time division multiple access (TDMA).Under the time division multiple access access way, different user in the system in sub-district is distinguished by different time slot (Time slot), and the different districts in the system is configured to different frequency ranges or multiplexing identical frequency range according to the frequency planning of system with frequency multiplexing mode.Thus, in the first generation and second generation mobile radio system of described employing frequency division multiple access and time division multiple access access way, power system capacity will be limited by the product of system bandwidth and time.In being called as the second generation mobile radio system of IS-95 code division multi-address radio communication system, adopted the wireless multiple access access way of code division multiple access (CDMA).Under the code division multiple access access way, different user in the system in sub-district is distinguished by different address code (Signature sequence), different districts in the system is distinguished by different scrambler (Scrambling code), in the ideal case, the frequency resource in the system bandwidth can be carried out multiplexing in intrasystem each sub-district.Thus, be with the first generation of described employing frequency division multiple access and time division multiple access access way and the remarkable difference part of second generation mobile radio system, in the second generation mobile radio system of described employing code division multiple access access way, there is not the upper limit that product limited by system bandwidth and time in power system capacity, but, have the characteristics of soft capacity only by wanted to interfering signal ratio level in the system and desired systemic the subject of knowledge and the object of knowledge decision.Because the described characteristics that on power system capacity, have, the code division multiple access access way is in present third generation mobile radio system, as the access mode that extensively adopts in CDMA 2000 (CDMA2000) wireless communication system, Wideband Code Division Multiple Access (WCDMA) (WCDMA) wireless communication system and Time Division-Synchronous Code Division Multiple Access (TD-SCDMA) wireless communication system.

In a code division multi-address radio communication system, interference can be divided into four major parts: internal system noise, intersymbol interference (ISI), multiple access disturb (MAI) and adjacent area interference (ACI).The internal system noise refers to the electronic device thermal noise of internal system, except adopting low noise amplifier, does not have other method for it; Intersymbol interference refers to a phase mutual interference between the distinct symbols in the subscriber signal, and its unfavorable autocorrelation performance by described address code and described scrambler is caused.Multiple access disturbs and adjacent area interference refers to phase mutual interference between the different user signal in the sub-district and the phase mutual interference between the different user signal in the different districts respectively, and its unfavorable their cross correlation by described address code and described scrambler is caused.As mentioned above, the power system capacity of code division multi-address radio communication system is only by wanted to interfering signal ratio level in the system and desired systemic the subject of knowledge and the object of knowledge decision.Therefore, can effectively suppress intrasystem intersymbol interference, multiple access interference and adjacent area interference, and the detection performance of raising wireless signal receiver will become the key factor that determines the code division multi-address radio communication system capacity under certain intersymbol interference, multiple access interference and adjacent area interference level.

For the TD-SCDMA system of identical networking, can be near the adjacent community user at Target cell edge to the stronger interference of this sub-district generation.In the community associated detection algorithm of traditional list, all adjacent areas are disturbed and are handled as white noise.In order to improve power system capacity, the multi-plot joint detection algorithm that the adjacent community user of further elimination disturbs has been proposed, this algorithm can reduce the adjacent area interference, has promoted systematic function.

Yet also there is following shortcoming in associating multi-cell channel algorithm for estimating: (1) algorithm computational complexity is bigger.Owing to changed the structure of channel estimate matrix, lost the Teoplitz characteristic of original matrix cyclic shift, so can not adopt Steiner shortcut calculation based on FFT/IFFT.Associating multi-cell channel algorithm for estimating is after decomposition waits processing through Cholesky, the complexity of algorithm has improved about 2.7 times than the Steiner algorithm that expands to a plurality of adjacent areas, even if select an interference user every 5 subframes, the complexity of algorithm has still improved 1.8 times.(2) user in this sub-district operates at full capacity, and during 8 maximum activation numbers of users promptly reaching system and allowed, the total length of all user CIR has taken all skews of midamble sign indicating number.At this moment, associating multi-cell channel algorithm for estimating can't be introduced the interference user signal in the neighbor cell again, and associating multi-cell channel degradation estimation is estimated for this cell channel.In addition, no matter be for user in this sub-district or neighbor cell user, its multipath signal all has the branch of strong footpath signal and weak footpath signal.In neighbor cell, during not selected many cell null of participation among user's (interference user that total signal power is relatively low) of Combined Treatment, also might there be strong footpath signal.These strong footpath signals have higher power than the weak footpath signal of choosing the user.And this part interference is considered as noise in existing associating multi-cell channel algorithm for estimating.Therefore the algorithm that this strong footpath is selected is wanted to combine with associating multi-cell channel algorithm for estimating, can improve the performance of original associating multi-cell channel algorithm for estimating greatly.

Summary of the invention

The objective of the invention is at the above-mentioned problems in the prior art, a kind of high accuracy channel estimation method that multi-plot joint detects that is used for is proposed, and in conjunction with strong footpath system of selection, this kind algorithm is optimized, form a kind of high accuracy channel estimation method of selecting based on strong footpath.Thereby can be in a code division multi-address radio communication system, improve the estimated accuracy of system simultaneously and reduce the algorithm complex of system greatly, and can reduce the error rate of system to a certain extent, improve greatly and detect performance.

Above-mentioned goal of the invention is realized by following technical scheme: be used for the high accuracy channel estimation method that multi-plot joint detects, comprise that step is as follows:

Chosen at least one user as interference user in the active user at least one neighbor cell of investigation user other active user in its persistent district and its persistent district for (1) one;

(2) described matched filter (MF) channel estimation value that obtains with the described sytem matrix of being investigated user and described interference user composition With described least mean-square error (MMSE) channel estimation value

Utilize

Disturb reconstruct, and carry out interference eliminated, obtain final channel estimation value:

${{\hat{h}}_2}^{\left(m\right)}={{\hat{h}}_0}^{\left(m\right)}-{\left(\mathrm{diag}\left(G^{H}G\right)\right)}^{-1}\left(\stackrel{\‾}{\mathrm{diag}}\left(G^{H}G\right)\right){{\hat{h}}_{1p}}^{\left(m\right)}$ Expression formula 6

Can carry out noise reduction process and obtain more accurate estimated value.

(3) in MMSE channel estimating part, select in conjunction with described strong footpath, obtain new estimated value:

${\hat{h}}^{\left(m\right)}={{\hat{h}}_0}^{\left(m\right)}-{\left(\mathrm{diag}\left({G_r}^H{G}_r\right)\right)}^{-1}\left(\stackrel{\‾}{\mathrm{diag}}\left({G_r}^H{G}_r\right)\right){{\hat{h}}_{r1p}}^{\left(A_m\right)}$ Expression formula 7

Carry out noise reduction process then, obtain more accurate estimated value.

The method according to this invention, described step (1) comprising:

(1a) describedly investigated the user and carried out channel estimating at described active user;

(1b) describedly investigated the user and in described active user, chosen at least one user as described interference user with the strongest channel impulse response power.

The method according to this invention, described step (1) comprising:

(1a) describedly investigated the user and carried out channel estimating at described active user;

(1b) describedly investigated the user and in described active user, choose at least one and have user greater than the channel impulse response power of a predetermined threshold value P as described interference user.

The method according to this invention, described step (1) repeats every a default time T.

The method according to this invention, in described step (2), the described up link training sequence code of user based on itself and described interference user of being investigated carries out combined channel at itself and described interference user and estimates.

The method according to this invention, described step (2) comprising:

(2a) describedly investigated the up link training sequence code matrix G that the user constructs itself and described interference user, wherein, the line number of described matrix G is L, and representative is used for the length of the described training sequence code received signal of channel estimating; The columns of described matrix G is WK, represents channel impulse response length and the described product of being investigated the sum of user and described interference user;

(2b) the described received signal e=Gh+n of user that investigated according to described training sequence code, channel impulse response h is united estimation, wherein, described channel impulse response h is made up of the channel impulse response of K wireless channel, and n is the additive white Gaussian noise signal that is additional on the described signal e.

The method according to this invention is carried out the MF channel estimating with the described sytem matrix of being investigated user and described interference user composition in described step (2b), obtain estimated value:

${{\hat{h}}_0}^{\left(m\right)}={\left(\mathrm{diag}\left(G^{H}G\right)\right)}^{-1}{G}^H{e_{\mathrm{mid}}}^{\left(m\right)}$ Expression formula 8

The method according to this invention is carried out the MMSE channel estimating with the described sytem matrix of being investigated user and described interference user composition in described step (2b), obtain estimated value:

${{\hat{h}}_1}^{\left(m\right)}={(G^{H}G+{\mathrm{\σ}}^{2}I)}^{-1}{G}^H{e_{\mathrm{mid}}}^{\left(m\right)}$ Expression formula 9

And carry out noise reduction process, obtain estimated value

The method according to this invention, described step (3) comprising:

(3a) carry out described strong footpath earlier and select, obtain new sytem matrix G _r, G _rDimension be far smaller than the dimension of G;

(3b) described new sytem matrix G _rCarry out the MMSE channel estimating, obtain estimated value:

${{\hat{h}}_{r1}}^{\left(A_m\right)}={({G_r}^H{G}_r+{\mathrm{\σ}}^{2}I)}^{-1}{G_r}^H{e_{\mathrm{mid}}}^{\left(A_m\right)}$ Expression formula 10

And carry out noise reduction process, obtain estimated value

(3c) in described step (2) with described MMSE channel estimation value

Described in step (3b)

Replace, sytem matrix G is with new sytem matrix G _rReplace, disturb reconstruct, obtain new estimated value:

${\hat{h}}^{\left(m\right)}={{\hat{h}}_0}^{\left(m\right)}-{\left(\mathrm{diag}\left({G_r}^H{G}_r\right)\right)}^{-1}\left(\stackrel{\‾}{\mathrm{diag}}\left({G_r}^H{G}_r\right)\right){{\hat{h}}_{r1p}}^{\left(A_m\right)}$ Expression formula 11

Below will be described in detail purpose of the present invention and characteristic in conjunction with the accompanying drawings by specific embodiment, these specific embodiments are illustrative, do not have restricted.

Description of drawings

Fig. 1 is the flow chart according to the embodiment of the inventive method;

Fig. 2 is simulation result figure of the present invention.

Embodiment

Core concept of the present invention is: by redefining the structure of channel estimate matrix G, the sytem matrix that user and adjacent sub-district interference user are included joint-detection simultaneously in is inserted in this sub-district, utilize matched filtering principle and mean square error minimum criteria algorithm to obtain estimated value respectively

With

Utilize this two results, the thought that adopts data reconstruction and parallel interference to eliminate obtains more precise channels estimated result, thereby has reduced the adjacent area interference, the elevator system performance.In addition, no matter be for user in this sub-district or neighbor cell user, its multipath signal all has the branch of strong footpath signal and weak footpath signal.In neighbor cell, during not selected many cell null of participation among user's (interference user that total signal power is relatively low) of Combined Treatment, also might there be strong footpath signal.These strong footpath signals have higher power than the weak footpath signal of choosing the user.And this part interference is considered as noise in existing associating multi-cell channel algorithm for estimating.Therefore the algorithm that this strong footpath is selected is wanted to combine with associating multi-cell channel algorithm for estimating, can improve the performance of original associating multi-cell channel algorithm for estimating greatly.

The boundary belt that comprises the data field of 2 352chip, the middle trained sequence of 1 144chip (midamble sign indicating number) and 1 16chip at the pulse structure of TD-SCDMA physical layer master time slot.1 radio channel resource unit (Ru) is defined as 1 spreading code in 1 main time slot.Contained midamble sign indicating number is the known sign indicating number sequence of receiving terminal among each RU, can be used for channel estimating.This paper studies up link, and base station (Node B) adopts the linear antenna battle array structure of 8 antennas.

The channel impulse response of k user on reception antenna m is designated as

H (k, m)=(h ₁(k, m), h ₂(k, m) ..., h _w(k, m)) ^T Expression formula 12

Wherein, () ^TThe computing of expression transposition, the length of channel impulse response is W.

Because multipath effect, preceding (W-1) position that receives in the midamble sign indicating number is subjected to the influence of preceding half segment data among the RU, so only does channel estimating with the back L bit that receives in the midamble sign indicating number, and the maximum length of L is (144-W).For reception antenna m, the midamble sign indicating number that is used for the received signal of channel estimating partly is defined as

e _Mid(m)=(e ₁, e ₂..., e _L) ^TExpression formula 13

Wherein, L is the length that is used for the midamble sign indicating number of channel estimating.

The signal of the defined midamble sign indicating number part that is used for channel estimating can further segment the signal that the cost community user produced, signal and other interference and the noise sum that adjacent sub-district interference user is produced in the expression formula 13:

${e_{\mathrm{mid}}}^{\left(m\right)}=\underset{k=1}{\overset{K_0}{\mathrm{\Σ}}}{M_0}^{\left(k\right)}{h}^{(k,m)}+\underset{k=K_0+1}{\overset{K_0+K_1}{\mathrm{\Σ}}}{M_0}^{\left(k\right)}{h}^{(k,m)}+n^{\left(m\right)}$ Expression formula 14

Wherein, other interference and noise item are n ^(m)=(n ₁, n ₂..., n _L) ^T

Matrix M ₀ ^(k)Be defined as

Expression formula 15

K ₀Be this sub-district access customer number, K ₁Be adjacent sub-district interference user number.

In the expression formula 14, first is this sub-district access signal that the user produced, second signal that is produced for adjacent sub-district interference user, and the 3rd is other interference and noise.

Redefine sytem matrix G, unite estimation with the channel impulse of this sub-district active user and adjacent sub-district interference user is corresponding.With this sub-district K in the expression formula 14 ₀Individual access user and adjacent sub-district K ₁The pairing matrix M of individual interference user ₀ ^(k)Be arranged in matrix G

$G=({M_0}^{\left(1\right)},\·\·\·,{M_0}^{\left(K_0\right)},\·\·\·,{M_0}^{(K_0+K_1)})$ Expression formula 16

Corresponding channel impulse respective sequence is arranged as:

$h^{\left(m\right)}={(h^{(1,m)T},\&CenterDot;\&CenterDot;\&CenterDot;,h^{({\mathrm{<figure-callout\; id="0"\; label="K"\; filenames="HSA00000502689000021.png"\; state="\{\{state\}\}">K</figure-callout>}}_0,m)T},\&CenterDot;\&CenterDot;\&CenterDot;,h^{({\mathrm{<figure-callout\; id="0"\; label="K"\; filenames="HSA00000502689000021.png"\; state="\{\{state\}\}">K</figure-callout>}}_0+{\mathrm{<figure-callout\; id="1"\; label="K"\; filenames="HSA00000502689000021.png"\; state="\{\{state\}\}">K</figure-callout>}}_1,m)T})}^T$ Expression formula 17

Expression formula 14 can be rewritten into:

e _Mid ^(m)=Gh ^(m)+ n ^(m)Expression formula 18

Expression formula 18 is carried out respectively obtaining expression formula 8 and expression formula 9 based on the estimation of MF with based on the estimation of MMSE, utilize the result who disturbs reconstruct and expression formula 8,9, the principle according to parallel interference is eliminated obtains more accurate estimated value, as expression formula 6.Step 2 among this step such as Fig. 1.

Strong footpath is selected to be attached in the step 2 among Fig. 1: the channel estimate matrix of introducing after strong footpath is selected only keeps the pairing row of strong path position selected among the matrix G.For example, G ^(k)The channel estimate matrix of expression user k is established the position of the the 3rd, the 5 and the 6th footpath for signal CIR place, the strong footpath of this selected user, then G ^(k)In only keep the the 3rd, the 5 and the 6th column vector.For user k, the matrix that remains is designated as G ^(k)With this sub-district K ₀Individual access user and adjacent sub-district K ₁The pairing matrix G of individual interference user ^(k)Be arranged in matrix G _r:

$G_r=({G_r}^{\left(1\right)},\&CenterDot;\&CenterDot;\&CenterDot;,{G_r}^{\left(K_0\right)},\&CenterDot;\&CenterDot;\&CenterDot;,{G_r}^{({\mathrm{<figure-callout\; id="0"\; label="K"\; filenames="HSA00000502689000021.png"\; state="\{\{state\}\}">K</figure-callout>}}_0+K_1)})$ Expression formula 19

The sytem matrix G that the combined channel of selecting according to the new strong footpath of determining is estimated _r, utilize the MMSE criterion to estimate, then

${{\hat{h}}_{r1}}^{\left(A_m\right)}={({G_r}^H{G}_r+{\mathrm{\&sigma;}}^{2}I)}^{-1}{G_r}^H{e_{\mathrm{mid}}}^{\left(A_m\right)}$ Expression formula 20

Because matrix G _rReduced than matrix G columns, then matrix G _r ^HG _rDimension just reduced.For example, choose the total K of this sub-district and adjacent area interference user ₀+ K ₁=8, the length W=16 of CIR, the received signal length that is used for channel estimating is L=128, then the size of matrix G is 128 * 128.According to the representative channel model of ITU definition, the number in useful strong footpath generally is no more than 3.More conservatively consider, if on average each user selects 4 useful multipaths, then matrix G _rSize be 128 * 32, G _r ^HG _rSize reduces to 32 * 32.In final CIR result, the pairing component of all not selected weak footpath signals all is changed to zero, and other components also will be changed to zero noise reduction process process by the component that will be lower than noise gate.

Described according to step 3 among Fig. 1, the MMSE that carries out again selecting based on strong footpath estimates, disturbs reconstruct again, according to the parallel principle of eliminating, obtains comparatively accurate estimated value, and expression formula 7 as described.

According to explanation of the present invention, those skilled in the art should be not difficult to find out, to be used for the high accuracy channel estimation method that multi-plot joint detects suitable equally at the code division multi-address radio communication system that uses multi-antenna technology or multi-transceiver technology according to of the present invention.

The content that is not described in detail in the present patent application book belongs to this area professional and technical personnel's known prior art.

Claims (8)

1. the multi-cell channel estimation method in the code division multi-address radio communication system is characterized in that comprising that step is as follows:

(1) at least one neighbor cell of a quilt investigation sub-district, chooses at least one active user as interference user;

(2) this sub-district is inserted the sytem matrix that user and adjacent sub-district interference user are included the multi-cell channel estimation simultaneously in, carried out combined channel and estimate;

(3) characteristics of the branch in strong, weak footpath are arranged at channel estimation results, become and select interference user, combine with described step (2) for selecting by force directly.

2. multi-cell channel estimation method according to claim 1 is characterized in that described step (1) comprising:

(1a) carry out multiuser channel and estimate, perhaps carry out channel estimating respectively at each active user in the described neighbor cell at the active user in the described neighbor cell;

(1b) in the active user of described adjacent sub-district, choose at least one Target cell is had more strongly disturbing user as described interference user.

3. multi-cell channel estimation method according to claim 1 is characterized in that described step (1) comprising:

(1a) the described active user of sub-district at described neighbor cell of being investigated carries out channel estimating;

(1b) the described sub-district of being investigated is chosen at least one and is had user greater than the channel impulse response power of a predetermined threshold value P as described interference user in the active user of described neighbor cell.

4. according to the arbitrary described multi-cell channel estimation method of claim 1 to 3, it is characterized in that described step (1) repeats every a default time T.

5. multi-cell channel estimation method according to claim 1, it is characterized in that in described step (2), described this sub-district active user is based on the training sequence code of itself and described adjacent sub-district interference user, carries out combined channel at itself and described adjacent sub-district interference user and estimates.

6. multi-cell channel estimation method according to claim 5 is characterized in that described step (2) comprising:

(2a) describedly investigated the sub-district and construct the training sequence code matrix G of its excited users and described adjacent sub-district interference user, wherein, the line number of described matrix G is L, and representative is used for the length of the described training sequence code received signal of channel estimating; The columns of described matrix G is WK, represents channel impulse response length and the described product of being investigated the sum of sub-district excited users and described adjacent sub-district interference user;

(2b) the described received signal e of sub-district that investigated according to described training sequence code _Mid ^(m)=Gh ^(m)+ n ^(m), to channel impulse response h ^(m)Carry out matched filtering (MF) channel estimating, wherein, described channel impulse response h ^(m)Channel impulse response by K wireless channel is formed, n ^(m)For being additional to described signal e _Mid ^(m)On the additive white Gaussian noise signal, obtain channel estimation value:

${{\hat{h}}_0}^{\left(m\right)}={\left(\mathrm{diag}\left(G^{H}G\right)\right)}^{-1}{G}^H{e_{\mathrm{mid}}}^{\left(m\right)}$ Expression formula 1

(2c) the described received signal e of sub-district that investigated according to described training sequence code _Mid ^(m)=Gh ^(m)+ n ^(m), to channel impulse response h ^(m)Carry out least mean-square error (MMSE) channel estimating, obtain channel estimation value:

${{\hat{h}}_1}^{\left(m\right)}={(G^{H}G+{\mathrm{\&sigma;}}^{2}I)}^{-1}{G}^H{e_{\mathrm{mid}}}^{\left(m\right)}$ Expression formula 2

And carry out noise reduction process, obtain estimated value

(2d) describedly investigated the sub-district and disturbed reconstruct and interference eliminated, obtained channel estimation value according to MF channel estimation value and MMSE channel estimation value:

${\hat{h}}^{\left(m\right)}={{\hat{h}}_0}^{\left(m\right)}-{\left(\mathrm{diag}\left(G^{H}G\right)\right)}^{-1}\left(\stackrel{\&OverBar;}{\mathrm{diag}}\left(G^{H}G\right)\right){{\hat{h}}_{1p}}^{\left(m\right)}$ Expression formula 3

7. multi-cell channel estimation method according to claim 1 is characterized in that comprising in described step (3):

(3a) in described neighbor cell, during not selected many cell null of participation among user's (interference user that total signal power is relatively low) of Combined Treatment, also might there be described strong footpath signal.These described strong footpath signals have higher power than the weak footpath signal of the described interference user of choosing;

(3b) in the interference user of described this sub-district excited users and described neighbor cell, select a most powerful path, again according to choosing thresholding and system's surplus resources is selected the stronger described strong footpath signal of signal from remaining multipath signal, and then utilize described this sub-district and described neighbor cell user's strong footpath signal to carry out the MMSE channel estimating.The estimated value of MMSE part becomes:

${{\hat{h}}_{r1}}^{\left(A_m\right)}={({G_r}^H{G}_r+{\mathrm{\&sigma;}}^{2}I)}^{-1}{G_r}^H{e_{\mathrm{mid}}}^{\left(A_m\right)}$ Expression formula 4

In conjunction with described step (1) and described step (2), obtain the final estimated value of channel:

${\hat{h}}^{\left(m\right)}={{\hat{h}}_0}^{\left(m\right)}-{\left(\mathrm{diag}\left({G_r}^H{G}_r\right)\right)}^{-1}\left(\stackrel{\&OverBar;}{\mathrm{diag}}\left({G_r}^H{G}_r\right)\right){{\hat{h}}_{r1p}}^{\left(A_m\right)}$ Expression formula 5

8. multi-cell channel estimation method according to claim 7 is characterized in that comprising in described step (3b):

(3b1) reduce the dimension of channel estimate matrix, greatly reduced computation complexity;

(3b2) comprised how described strong footpath signal, the interfering energy of consideration is more concentrated, thereby effectively reduces interference, has improved channel estimated accuracy.

Images