WO2006136052A1

WO2006136052A1 - A pre-filtering sdma transmitting and receiving method, system and a filter thereof

Info

Publication number: WO2006136052A1
Application number: PCT/CN2005/000887
Authority: WO
Inventors: Junfeng Zhang
Original assignee: Zte Corporation
Priority date: 2005-06-21
Filing date: 2005-06-21
Publication date: 2006-12-28
Also published as: CN101147349B; CN101147349A

Abstract

A pre-filtering SDMA transmitting and receiving method and system, and a filter thereof and its generating method, the SDMA transmitting system calculates the complete diagonalization pre-filter matrix composed of U sub-blocks satisfying block diagonalization condition and complete diagonalization condition; at the same time, after the data stream which the said transmitting system sends to the users passes the pre-filter, it is transmitted through numbers of antennas synchronously, and reaches each user terminal via air channel; after the terminal receives the transmitting signal, judges the received signal directly, and recovers the data stream sent to itself, at the same time, feeds the measured channel characteristic back to the said transmitting system ,the channel is the sub-channel from the said transmitting system to the terminal. In addition to eliminate the multi-user interference, the present invention can also eliminate the interference between various data streams of the user, then simplifies the design of the receiver.

Description

Pre-filtered spatial division multiple access transmission and reception method, system and 顸 filter used

The present invention relates to a Space Division Multiple Access (SDMA) system in a mobile communication system, and more particularly to a space division for downlink pre-filtering of a MIMO (Multiple Input Multiple Output) system. The prefilter method, system, prefilter, and its generation method. Background technique

Compared with single-antenna systems, ΜΒίΟ technology has received considerable attention in recent years due to the use of multiple parallel spatial channels, resulting in significant capacity or diversity gain. To date, research in this area has focused primarily on single-user, point-to-point multi-antenna communication systems, without regard to co-channel interference between multiple users. Until recently, the focus of research has shifted to the multi-user MIM0 system. Compared with the single-user MIM0 system, the multi-user MIM0 system has a large difference. Co-channel interference that is not considered in a single-user system will cause serious deterioration of system performance if it is not handled well. Both TDMA technology and CDMA technology have been proposed as methods for combating co-channel interference in multi-user systems, but both have obvious deficiencies. The spectral efficiency of TDMA technology is too low, so the system and capacity are not high, which does not reflect the advantages of the large capacity of the MIM0 system. The problem with CDMA technology is that multi-user MB10 systems need to consume a lot of code resources, and the system can hardly meet this requirement.

Compared with TDMA and CDMA, SDMA (Spatial Division Multiple Access) technology is a promising development to eliminate interference in multi-user MIM0 systems due to its inherent high spectrum utilization and no code resource consumption. direction. In addition, if SDMA is combined with TDMA or CDMA, the flexibility of the system application can be further improved. The key technology of SDMA is how to eliminate the interference between MUI (Multi-user Interference) and each user's internal data stream. Due to the asymmetry of the data service itself, the demand for high-speed and large-capacity downlinks is often more urgent than uplink, so this technology is also mainly used for the downlink. The following three references propose ways to eliminate MUI:

"Joint Tx-Rx optimization for MIMO SDMA based on a null-space constraint^ , A. Bourdoux, N. Khaled, Joint Tx-Rx optimization for MIMO - SDMA based on a null-space constraint, VTC 2002-Fall, Vancouver, Canada, Sept. 2002 : 171 - 174.

"A transmit preprocessing technique for Multi-user MIMO systems using a decomposition approach", L. Choi, RD Murch, A transmit preprocessing technique for Multi-user MIMO systems using a decomposition approach, IEEE Transactions On Wireless Communications, Jan. 2004, 3 (1): 20 - 24. as well as

"Zero-forcing methods for downlink spatial multiplexing in multi-user MIMO channels", QH Spencer, AL Swindlehurst, M. haardt, Zero-forcing methods for downlink spatial multiplexing in multi-user MIMO channels, IEEE Transactions On Signal Processing, Feb. 2004, 52 (2): 461-471.

The proposed method is to design a pre-filter on the transmitting end so that the equivalent channel is block diagonalized, that is, the equivalent channel matrix after multiplication of the channel matrix and the pre-filter is block diagonalization. Hereinafter, this algorithm is simply referred to as a BD algorithm. Under ideal conditions, such processing can completely eliminate the MUI at the transmitting end, leaving only the interference between the various data streams within the user. This interference needs to be processed at the user terminal by the detection algorithm of the single-user MIM0 system.

Figure 1 shows a spatial division multiple access transmission and reception system using a partial block diagonalization pre-filter. The space between the signal source ( ¹ ~ ^ and the channel matrix H is part of the block diagonalization). Matrix filter. If the partial block diagonalization matrix filter K is a horizontal splicing of C7 pre-filtering matrices, and the column of T is selected to be in the null space, then the partial block diagonalization condition can be In this case, it is necessary to know the channel conditions of all users at the same time in order to receive their own signals. Therefore, although the above block diagonalization method can completely eliminate the MUI, it is very effective for overcoming the near-far effect, but it has the following drawbacks: In order to eliminate interference between data flows inside the user (each data stream corresponds to one each) In the case of an antenna, it is equivalent to eliminating interference between antennas within the user. Each mobile station needs to know the channel conditions of all other mobile stations at any time. Due to the time-varying nature of the wireless channel, this is not realistic at all in terms of implementation of the user terminal. More often, the fading between mobile stations is independent at any one time, and each mobile station can only know its own channel condition. , and cannot know the channel status of other mobile stations. The block diagonalization method is actually difficult to eliminate interference between various data streams within the user.

In addition, when the block diagonalization method is adopted, the user terminal needs an additional receiving filter, which also makes the structure relatively complicated. Summary of the invention

The technical problem to be solved by the present invention is to provide a pre-filtered spatial division multiple access transmission and reception method for a multi-user system downlink, which can eliminate multi-user interference and internal data streams between users. Interference, easy to implement, and simplifies the design of user terminals. The present invention also provides a space division multiple access transmitting and receiving system that can implement the method.

In order to achieve the above technical problem, the present invention provides a pre-filtered spatial division multiple access transmission and reception method for signal transmission and reception of a downlink of a multi-user multiple input multiple output system, including the following steps:

(Α) The space division multiple access transmitting system calculates in real time a fully diagonalized pre-filtering matrix F consisting of U sub-blocks satisfying H. F" = 0 and "= J, /^ [FL^] , where / Unit

Array, 0 is a zero matrix, is the channel characteristic of the user terminal M feedback, H is the matrix obtained after deleting all the rows corresponding to the user M from the channel matrix H, " = l, 2, ~, U;

At the same time, the transmitting system sends the source data stream C: ¹ to C ^y to the user, and the data stream sent to the user is recorded as C, and the data stream is pre-filtered, that is, completely diagonalized. The pre-filtering matrix is multiplied to obtain a transmitted signal consisting of the U-group data stream C"E", "= 1,2, ·■■,[/ ; and the transmitted signal is simultaneously transmitted from multiple antennas, passing through the air. The channel reaches every user; (B) after receiving the transmission signal, each user terminal directly determines the received signal, and recovers the data stream sent to itself, and at the same time, the user terminal "measures the measured sub-channel from the transmitting system to the terminal. The channel characteristic H" is fed back into the transmitting system, _U = 1, 2, -, U.

Further, the above-described spatial division multiple access transmission and reception method may have the following feature: when the spatial division multiple access transmission system determines the transmission power of the user signal, the water injection method is used to perform power unequal distribution between users. Further, the above-described spatial division multiple access transmission and reception method may have the following feature: the sub-block F" = N"E", where N" is an orthogonal basis corresponding to zero space, and E" is passed The following formula is calculated -

In the formula, it means to find the Frobenius norm, "is the user", and u = \,2,-,U, " ( · ) ⁺ " represents the pseudo-inverse operation on the matrix.

The pre-filtered spatial division multiple access transmitting and receiving system provided by the invention comprises a space division multiple access transmitting system and U user terminals including U users, and the transmitting system comprises U user source generators and multiple antennas a signal transmission power distribution module, a pre-filter and a generating module thereof, each user terminal comprising a receiver, a wireless channel matrix measuring device and a plurality of receiving antennas, wherein:

The pre-filter is a fully diagonalized pre-filter, the matrix of which is composed of U sub-blocks, ie ^ = ^Ρ^··^], F" satisfies the block diagonalization condition H:.F"=0 and completely The diagonalization condition H ^u F" =1, where / is a unit matrix, 0 is a zero matrix, is a matrix composed of rows corresponding to user u in the entire channel matrix ^, H is deleted from the channel matrix H all corresponding to the user The matrix obtained after u's line, u = \, 2, -, U;

The pre-filter generating module is configured to calculate, in real time, the fully diagonalized pre-filtering matrix that satisfies a block diagonalization condition and a full diagonalization condition according to the received channel characteristics measured by each user terminal. The receiver of the user terminal is configured to directly judge the received signal and resume the data stream sent to itself.

Further, the above-described space division multiple access transmitting and receiving system may further have the following feature: the signal transmission power distribution module uses the water injection method to perform power unequal distribution between users.

Further, the above-described spatial division multiple access transmitting and receiving system may further have the following feature: a sub-block F" = iV" E" in the pre-filtering matrix, where N" is an orthogonal basis corresponding to zero space, E" is The pseudo inverse of H"XN".

Further, the above-described space division multiple access transmitting and receiving system may further have the following feature: the pre-filter generating module calculates E" as follows:

In the formula, it means to find the Frobenius norm, P" is the transmission power for the user _M , u = \X-, U, "( · ) ⁺ " represents the pseudo inverse of the matrix.

Another technical problem to be solved by the present invention is to provide a pre-filtered spatial division multiple access transmission system for a downlink of a multi-user MIM0 system, which eliminates multi-user interference and completely diagonalized pre-filter on the network side. Interference between data flows within the user.

In order to solve the above technical problem, the present invention further provides a pre-filtered spatial division multiple access transmission system, comprising a space division multiple access transmission system including U users, the system comprising U user source generators, multiple roots An antenna, a signal transmission power distribution module, a pre-filter and a generating module thereof, wherein:

The pre-filter is a fully diagonalized pre-filter, and the matrix F is composed of U sub-blocks, that is, /^[/^••^], which satisfies the block diagonalization condition ^" = 0 and the complete diagonalization condition H" F ^U = I , where J is a unit matrix, 0 is a zero matrix, and ^ is a matrix of rows corresponding to user u in the entire channel matrix, H. After "is to delete all rows _¾ corresponding to the user from the channel matrix H obtained in the matrix, Μ = 1,2, ·, ί /..; The pre-filter generation module is configured to calculate, in real time, the fully diagonalized pre-filtering matrix satisfying a block diagonalization condition and a full diagonalization condition according to the received channel characteristics measured by each user terminal.

The technical problem to be solved by the present invention is to provide a method for generating a fully diagonalized pre-filtering matrix for a multi-user multiple input multiple output system, and a pre-filter using the generation matrix thereof to transmit by transmitting at the transmitting end The pre-filtering of the signal eliminates interference between multi-user interference and the various data streams within the user.

In order to achieve the above technical problem, the present invention provides a method for generating a fully diagonalized pre-filtering matrix in a multi-user multiple input multiple output system, including the following steps:

(a) The space division multiple access transmitting system including U users receives the channel characteristics H" of the feedback measured by the U user terminals within a set time interval, and then vertically splicing them to obtain a channel matrix, where _Μ = 1 ,2,.··, ί/;

(b) For any value in the set [1.17], remove all rows corresponding to the user _W from the matrix, and obtain a matrix.

(c) U "respectively obtain the orthogonal basis N of its corresponding zero space", and then splicing the U N" horizontally to form a matrix V = [- N ] _;

(d) According to the obtained H" and N", obtain a pseudo inverse E", u = l, 2, -, U that satisfies ¹ ^^;

(e) Let F"H splicing the t/pre-filtering matrices horizontally to obtain complete diagonalization

The matrix of the pre-filter F = [ H ^U ], when the next time interval comes, returns to step (a).

The invention also provides a fully diagonalized pre-filter in a multi-user MIMO system, comprising an N-matrix filter for eliminating inter-user interference, characterized in that it further comprises a method for eliminating internal users. An E-matrix filter for interference between multiple data streams, where:

The matrix N of the N matrix filter is horizontally spliced by U sub-blocks, N = [N'...N ^U , N" is the orthogonal basis of ff' corresponding to zero space, and II: is the slave channel matrix# Remove all corresponding to The matrix obtained after the line of the user,

The matrix of the E matrix filter, E" is the inverse of H" X iV"

The matrix of the fully diagonalized pre-filter = NxE.

It can be seen from the above that the present invention has the following advantages over the existing transmission pre-filtering system based on the block diagonalization joint algorithm -

1) In addition to eliminating multi-user interference, it also eliminates interference between various data streams within the user.

2) It is more realistic and feasible. It is enough for each receiving user to know its own channel conditions. It is not necessary to know the channel conditions of all users at the same time. This is crucial for the application of SDMA MIM0 technology.

3) Since the interference between the MUI and the internal antennas of the user is completely eliminated at the transmitting end, the user terminal does not need a more complicated receiving filter for processing, which greatly simplifies the receiver design.

BRIEF abstract

Figure 1 is a schematic diagram of a spatial division multiple access transmission and reception system using a spatial division multiple access partial block diagonalization pre-filter.

2 is a schematic diagram of a space division multiple access transmission and reception system employing a fully diagonalized pre-filter in accordance with an embodiment of the present invention.

Fig. 3 is a flow chart showing the method for solving the fully diagonalized pre-filter F matrix of the embodiment. 4 is a flow chart of a pre-filtered spatial division multiple access transmission and reception method of the present embodiment.

Figure 5 is a comparison of BER performance of the BD algorithm and the CD algorithm in an independent MIM0 channel. Figure 6 shows the BER performance of the BD algorithm and CD algorithm in the Nokia spatial correlation IOM0 channel. Compare the graphs.

Figure 7 is a comparison of the Shannon capacity of the BD algorithm and the CD algorithm under the independent MIM0 channel. Figure 8 is a comparison of the interrupt capacity (interrupt rate: 0.1) of the BD algorithm and the CD algorithm in the independent MIM0 channel.

Figure 9 shows the interrupt capacity of the BD algorithm and the CD algorithm in the Nokia space-dependent MIM0 channel.

(Interrupt rate: 0.1) Comparison chart.

BEST MODE FOR CARRYING OUT THE INVENTION

The present invention proposes a new pre-filtered spatial division multiple access transmission and reception method and system, and the fully diagonalized pre-filter and its implementation method in the system, which is simply referred to as CD algorithm. The core idea is to further strengthen the partial block diagonalization conditions mentioned in the above three references to Complete diagonalization, that is, the interference between the data streams inside the user is completely eliminated at the transmitting end. .

The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

2 is a spatial division multiple access transmission and reception system using a fully diagonalized pre-filter according to an embodiment of the present invention, including a space division multiple access transmission system including U users and U user terminals. The space division multiple access transmission system includes U user source generators, an antenna group including A antennas, a signal transmission power distribution module (not shown), a fully diagonalized pre-filter, and a generation module thereof.

The user source generator is configured to generate a source data stream to be sent to each user, and the data stream corresponding to each user includes a multiplexed symbol stream obtained after demultiplexing.

The antenna group is used to simultaneously transmit the pre-filtered multiple data stream signals by the A antennas. A fully diagonalized pre-filter is used to filter the source data stream, the matrix F of which is U

Sub-block construction, with =| .^], F" satisfies block diagonalization condition " "=0 and fully diagonalized

Condition Γ. "=/. where / is the unit matrix, 6» is a zero matrix, ^ is the matrix of the row corresponding to the user U in the entire channel matrix, and all the corresponding to the user a are deleted from the channel matrix # The matrix obtained after the line.

The pre-filter generation module is configured to calculate a fully diagonalized pre-filtering matrix F satisfying the block diagonalization condition and the full diagonalization condition in real time according to the received channel characteristics Τ of each user terminal.

The signal transmission power distribution module is configured to allocate the power of the signals transmitted by the respective users. In this embodiment, the water injection method is used to perform power unequal distribution between users.

Each user terminal includes a plurality of receiving antennas, a 匪 SE (minimum mean square error) receiver, and a radio channel matrix measuring device. among them:

A plurality of receiving antennas are used to receive signals transmitted by the network side, the content of which is a superposition of data sent to all users.

The MMSE receiver is configured to determine the received signal and recover the data stream sent to the user terminal;

The radio channel matrix measuring device is configured to measure the MO subchannel from the space division multiple access transmitting system to the user terminal, obtain channel characteristics and feed back to the space division multiple access transmitting system.

Compared with Fig. 1, this embodiment replaces the partial block diagonalization matrix filter of the prior art with a fully diagonalized pre-filter F. The fully diagonalized pre-filter F further includes a unitary matrix filter and a unitary matrix filter. The unitary matrix filter may also be referred to as an inter-user interference canceller, and the process performed is to cancel the interference between users; the matrix may also be called The inter-data inter-stream interference canceller in the user performs the process of canceling the interference between the single-user majority data streams.

Both filters are followed by real-time changes in the channel characteristics change spatial channel, which varies depending on the user _W from the first measurements obtained from spatial division multiple access transmitting system to the second user's MIM0 〃 sub channel characteristic " , _M = l, 2, ··.,[/.

The method for generating the fully diagonalized pre-filter F in this embodiment is shown in Fig. 3. The pre-filter generation module in the space division multiple access transmission system performs the following steps:

Step 110: Receive feedback from the U user terminals during the set time interval. The channel characteristic T, the channel characteristic T is only for the MIM0 subchannel from the space division multiple access transmission system to the "user", u = l, 2 -; U;

^ is the row corresponding to the user u in the entire channel matrix, that is, H" is the MIM0 subchannel from the base station to the uth user.

Step 120: Vertically splicing the channel matrices of the U users to obtain a channel matrix jy. For any _M value in the set [1.1/], delete all the rows corresponding to the user from the matrix to obtain a matrix H _£ "; ί言道 Matrix:

. "Having ∑B

Step 130: Obtain the orthogonal basis V«, _w = l, 2, ..., f/ of the corresponding zero space by U H, respectively, and then splicing the u sub-blocks horizontally to form a matrix N,

, the dimension of its "sub-block N" is

Step 140, according to the obtained 'and N', find the condition that satisfies the condition .^ =0 and the condition "·Τ "· "=/", where / is a unit matrix, (? is a zero matrix, the E is H The pseudo inverse of «XN"; this embodiment is calculated according to the Frobenius norm normalization algorithm, as shown in the following equation, but the present invention is not limited to this:

In the formula, |b|| _F denotes the Frobenius norm, which is the transmit power for the user M, u = \X-, U, "( · ) ⁺ " represents the pseudo inverse of the matrix.

Step 150, let F" = N ^U E ^u , u = l, 2, -, U, and then a pre-filtering matrix F "horizontally stitched, that is, obtain the matrix F of the fully diagonalized pre-filter, and then at the next time When the interval comes, the process returns to step 110. The matrix of fully diagonalized pre-filters, F = j H ], can also be expressed as:

The visible matrix E is composed of U block diagonalization matrices, and the dimension of the u-th diagonal diagonal sub-block on the diagonal is D" x C" ₀

After the matrix F is obtained, the total number of rows is C ^U (C" indicates the data stream sent to the user W), and the data source matrix multiplication process with the number of columns infinity is the pre-filtering process. The above solution method is based on two Constraint derivation

, O is a zero matrix, which can also be expressed as condition one: ^ . ^ . E^ O and condition two: HH = I. Condition one is the block diagonalization condition, which can completely eliminate the MUI. Condition 2 is a fully diagonalized condition, which satisfies the interference between the various data streams within the user. This will be further clarified below.

As shown in FIG. 4, the pre-filtered spatial division multiple access transmission and reception method for the downlink of the multi-user MB© system in this embodiment includes the following steps:

Step 210: The spatial division multiple access transmitting system sends multiple data streams to the user/user, and the data stream sent to the user is C (including the multi-path data stream obtained by multiplexing), _U = l, 2, - , U, at the same time, the transmitting system calculates a fully diagonalized pre-filtering matrix F satisfying the block diagonalization condition and the complete diagonalization condition in real time according to the channel characteristics fed back by the user terminal, the method is described above;

Step 220, the source data stream through each C ¹ ~ ^ corresponding block diagonal matrix subblock E "filtering, i.e. the filtering process stream C" block and sub-block diagonalization by multiplying E "moment, array Process; Step 230, each C''and the multiplied result C"DE" continues to be filtered by the N-matrix filter, that is, multiplied by N" in the matrix N, and the output A path is pre-filtered by the transmitted signal C"D£ OV", u = \, 2, -, U, respectively transmitted to the A antenna;

Step 240, the pre-filtered A-channel transmitting signal is simultaneously sent by the A antenna, and The air channel reaches every user;

Step 250: Each user terminal simultaneously receives the superposition of data sent to all users, and directly determines the received signal by using the 匪SE receiver, and restores the data stream C" sent to itself; meanwhile, the "user terminal measurement" From the space division multiple access transmission system to the MIM0 subchannel of the wth user, the channel characteristics are obtained and fed back into the space division multiple access transmission system.

The following uses symbolic vector operations to illustrate why the above method eliminates interference between the MUI and the user's internal data streams:

Suppose that the user's symbol vector is xW at any one time, where: the symbol vector and the other representation of the data stream c"

c is the number of symbol streams obtained by demultiplexing one input bit stream.

The signal received by the Wth client of this embodiment can be expressed as: x ^u (k) = H ^u -Fx(k) + n ^u (k) , for all users satisfying J3⁄4^F"=O, it means H"-F _c "=0 , O is a zero matrix, the so-called orthogonal guarantee, so x ^u (k) = H ^u -F"-x"(k) + n ^u (k), at this point, MUI has Was completely eliminated.

However, the interference between the user C's C" way data stream ₍ )...; ^ _W has not been eliminated. By satisfying ^^" = , the interference between the various data streams inside the user u is further eliminated, because The I matrix is an identity matrix with a diagonal element of 1, and other elements of 0. This condition is met, meaning that a certain data stream is only related to itself and completely orthogonal to other data streams.

The following is an application example to further illustrate the implementation process of the downlink of the multi-user MIM0 system using the spatial division multiple access pre-filter.

Assume that the base station has 6 antennas, the system has 3 users, and each user has 2 antennas. Therefore, such a system configuration has 3x2 = 6 parallel data streams simultaneously. The three input bit streams are demultiplexed into two symbol streams after QPSK modulation at the base station. Each symbol stream is subdivided into data frames containing 480 symbols.

Send the user's symbol stream at any one time

2, 3, where; and; c ₂ "(t) is a 2-way symbol stream demultiplexed by one input bit stream. Input user filter vector x of the pre-filter

. The construction of the matrix F of the fully diagonalized pre-filter has been described in detail above, in this example

+ n"(k)

It is possible to recover the original transmitted signal, and the decision is a process from a fuzzy continuous quantity to a quantized digital quantity. Where: is the signal received by the terminal, is the original transmitted signal, is the thermal noise in the air or external interference, "is the natural thermal noise. Restoring the original transmission signal is to recover from (k) ":" (k).

The present invention eliminates the interference between the MUI and the various data streams within the user at the transmitting end, so that no additional processing is required at the user terminal. This is also one of the features of the present invention: Simplified receiver design. After the pre-processing of the transmission filter, the signal after passing through the channel has no interference, and the decision can be directly made.

The Shannon capacity of the fully diagonalized matrix (CD) algorithm of the present invention is as follows:

C 2, 3 )

Where / is the identity matrix, the number of simultaneous transmissions = 3, ^ is the transmission power of user w. σ ² refers to the Gaussian white noise power at the uth user receiving end.

Under equal power allocation conditions, if the total transmit power of the spatial division multiple access transmitter is represented in one symbol period, the Shannon capacity of the CD algorithm becomes:

In fact, for this CD algorithm, we can also distribute the signal transmission power in the system. In the module, the water injection method is used to perform power unequal distribution between users, so that the performance is further improved than the equal power distribution, and the improvement is more obvious at a lower signal to noise ratio.

Note that the rank of the unit matrix / in the above equation is ^".^", and the value is set to L, then the above formula becomes

The water injection method is to solve the maximum value of C under the P„ = power constraint condition (by using the Lagrange multiplier λ, the power of each user can be obtained as:

L.,

σ (Η" · Ν ( u= l , 2 , 3 )

The value of λ\η2 "Τ is obtained by |^, = ^". Figure 5 shows the BER performance comparison between the BD algorithm and the CD algorithm in the independent M-top 0 channel. The simulation results respectively list the BD algorithm and the CD algorithm. The BER performance of the number of base station antennas is 6, 7 and 8, respectively, and the number of UEs is 3, and the number of antennas is 2. It can be seen that the BD algorithm is slightly better than the CD algorithm under the same antenna configuration.

Figure 6 shows the BER performance comparison between the BD algorithm and the CD algorithm under the Nokia space-dependent MIM0 channel. The simulation results show the BER performance of the BD algorithm and the CD algorithm under the condition that the number of base station antennas is 6, 7, 8, the number of UEs is 3, and the number of antennas is 2. It can be seen that the BD algorithm is slightly better than the CD algorithm under the same antenna configuration.

Figure 7 shows the Shannon capacity comparison of the BD algorithm and the CD algorithm on a separate MIM0 channel. The simulation results respectively enumerate the Shannon capacity performance of the BD algorithm and the CD algorithm under the condition that the number of base station antennas is 6, 7, 8, the number of UEs is 3, and the number of antennas is 2. It can be seen that the BD algorithm is slightly better than the CD algorithm under the same antenna configuration.

Figure 8 shows a comparison of the interrupt capacity (interruption rate: 0.1) of the BD algorithm and the CD algorithm under the independent MIM0 channel. The simulation results respectively illustrate the interrupt capacity performance of the BD algorithm and the CD algorithm under the condition that the number of base station antennas is 6, 7, 8, the number of UEs is 3, and the number of antennas is 2. Can It can be seen that the BD algorithm is slightly better than the CD algorithm under the same antenna configuration.

Figure 9 shows the comparison of the interrupt capacity (interruption rate: 0.1) between the BD algorithm and the CD algorithm in the Nokia space-dependent MIM0 channel. The simulation results respectively show the interrupt capacity performance of the BD algorithm and the CD algorithm under the condition that the number of base station antennas is 6, 7, 8, the number of UEs is 3, and the number of antennas is 2. It can be seen that the BD algorithm is slightly better than the CD algorithm under the same antenna configuration.

Based on the simulation results of Figures 5 to 9, the performance of the CD (all diagonalization) algorithm and the BD (Partial Block Diagonalization) algorithm are consistent under the independent MB10 channel and the Nokia space-dependent MIM0 channel. The correlation between the antennas increases and the system performance deteriorates. The difference is that in both cases the performance of the CD algorithm is slightly lower than the BD algorithm. This is not difficult to understand. The BD algorithm performs optimally under the conditions it assumes, but this assumption is unrealistic and cannot be implemented. Although the CD algorithm has slightly lower performance than the BD algorithm, it is one of the feasible algorithms. Moreover, in the case where the number of users and the number of receiving antennas per user are constant, as the number of transmitting antennas increases, the performance difference between the CD algorithm and the BD algorithm is gradually reduced. Industrial applicability

The method of the invention can be applied to the transmission and reception of the downlink signals of the multi-user MIMO system in the space division multiple access communication system, eliminating the interference between the multi-user interference and the internal data streams of the user, and simplifying the receiver Designed with industrial applicability.

Claims

Claim

A pre-filtered spatial division multiple access transmission and reception method for signal transmission and reception of a downlink of a multi-user multiple input multiple input system, comprising the following steps:

(A) The space division multiple access transmission system calculates the real-time satisfaction. ".F" = 0 and HTM . = J's U-sub-blocks consist of a fully diagonalized pre-filtering matrix F, ... ], where / is the unit

Array, 0 is a zero matrix, is the channel characteristic fed back by the user terminal _M , H _£ "is a matrix obtained by deleting all the rows corresponding to the user M from the channel matrix H, u = l, H

At the same time, the transmitting system sends the source data streams C'~C ^U to the user, and the data stream sent to the user W is recorded as C", and the data stream is pre-filtered, that is, the fully diagonalized pre- Multiplying the filter matrix to obtain a transmitted signal consisting of the U group data stream C"E'', " = 1,2, .., [/ ; and transmitting the transmitted signal from multiple antennas simultaneously, through the air The channel reaches every user;

(B) after receiving the transmission signal, each user terminal directly determines the received signal, and recovers the data stream sent to itself, and at the same time, the user terminal a measures the measured sub-channel from the transmitting system to the terminal. The channel characteristics are fed back into the transmitting system, u = H., U.

2. The space division multiple access transmission and reception method according to claim 1,

It is characterized in that the sub-block F" = N"E", where ^" is ^" corresponds to the orthogonal basis of the zero space, and E" is calculated by the following formula:

In the formula, it means to find the Frobenius norm, P" is the transmission power for the user w, _W = 1, 2, ..., [/, " ( ·) ⁺ " represents the pseudo inverse of the matrix.

3. The method of transmitting and receiving a spatial division multiple access according to claim 1,

The method is characterized in that, when determining the transmission power of the user signal, the space division multiple access transmitting system uses the water injection method to perform power unequal distribution between users.

4. A pre-filtered spatial division multiple access transmitting and receiving system, comprising one containing U a space division multiple access transmitting system and a U user terminal, the transmitting system comprising U user source generators, multiple antennas, a signal transmission power allocation module, a pre-filter and a generating module thereof, each user terminal including a receiver , a wireless channel matrix measuring device and a plurality of receiving antennas, characterized in that:

The pre-filter is a fully diagonalized pre-filter, and the matrix F is composed of U sub-blocks, ie

_F - · · ], F" satisfies the block diagonalization condition" · = 0 and the complete diagonalization condition H ^u F" = 1 , where / is a unit matrix, O is a zero matrix, which is the corresponding user in the entire channel matrix The matrix formed by the rows of u, H is the matrix obtained after deleting all the rows corresponding to the user u from the channel matrix, u = l, 2, ---, U;

The pre-filter generating module is configured to calculate, in real time, the complete diagonalized pre-filtering matrix that satisfies a block diagonalization condition and a full diagonalization condition according to the received channel characteristics HTM measured by each user terminal.

The receiver of the user terminal is configured to directly determine the received signal and resume the data stream sent to itself.

5. The space division multiple access transmitting and receiving system according to claim 4,

The signal transmission power distribution module uses a water injection method to perform power unequal distribution between users.

6. The space division multiple access transmitting and receiving system according to claim 4,

The sub-block " = ν" ", where N" is an orthogonal basis corresponding to the zero space, is a pseudo inverse of X N".

7. The space division multiple access transmitting and receiving system according to claim 6,

The pre-filter generating module is calculated in the following manner

Where ||· represents the Frobenius norm, P" is the transmit power for the user, u = l, 2, . , U, " ( · ) + " represents the pseudo inverse of the matrix.

8. A pre-filtered spatial division multiple access transmission system comprising a space division multiple access transmission system comprising U users, the system comprising U user source generators, multiple antennas, signal transmission power allocation modules, pre- Filter and its generation module,

It is characterized by:

The pre-filter is a fully diagonalized pre-filter, the matrix F of which consists of U sub-blocks, ie _F = [i^„ _F "], F "satisfies the block diagonalization condition ^ " = 0 and completely diagonal The condition H"F" = 1 , where / is a unit matrix, 0 is a zero matrix, which is a matrix of rows corresponding to the user _U in the entire channel matrix, "is to remove all rows corresponding to the user from the channel matrix Matrix, = l, 2, - - -, U;

The pre-filter generating module is configured to calculate, in real time, the fully diagonalized pre-filtering matrix F that satisfies a block diagonalization condition and a full diagonalization condition according to the received channel characteristics H" measured by each user terminal. .

9. The space division multiple access transmitting system of claim 7

The signal transmission power distribution module adopts a water injection method to perform power unequal distribution between users.

10. The space division multiple access transmission system of claim 7

It is characterized in that the sub-block '=N' E in the pre-filtering matrix where N" is the orthogonal basis of H corresponding to the zero space, and E" is the pseudo-inverse of XN".

11. The space division multiple access transmission system of claim 10,

The pre-filter generating module is calculated as follows:

Where ||| . || _F denotes the Frobenius norm, P" is the transmit power for the user _M , u = l, 2, -, U, "( · ) +" denotes a pseudo-inverse operation on the matrix.

12. A method for generating a fully diagonalized pre-filtering matrix in a multi-user multiple input multiple output system, comprising the following steps: (a) The space division multiple access transmitting system including U users receives the channel characteristics H" from the feedback measured by the U user terminals within a set time interval, and then vertically splices them to obtain a channel matrix H, where _M = l, 2,...,t/ _;

(b) For any _3⁄4 value in the set [lI/], remove all rows corresponding to the user from the matrix, and obtain a matrix H.";

(c) by arranging "the orthogonal basis i of its corresponding zero space respectively", and then splicing the U N'' horizontally to form a matrix N, N^N ¹ -^ ;

(d) According to the obtained "and N«, obtain a pseudo-inverse E" satisfying H"XN", u = \,2,-,U;

(e) Let F" = N" ", then splicing the C7 pre-filtering matrices F" horizontally to obtain complete diagonalization

The matrix of the pre-filter / ^[/^..^], when the next time interval comes, returns to step (a).

13. The method for generating a fully diagonalized pre-filtering matrix according to claim 10, wherein in the step (d), the following formula is used to calculate E'':

In the formula, it means to find the Frobenius norm, "is the transmission power for the user w, u = l, 2, -, U, " ( · ) + " represents the pseudo inverse of the matrix.

14. A fully diagonalized pre-filter in an incoming and outgoing system, comprising an N-matrix filter for eliminating inter-user interference, characterized in that it further comprises a method for eliminating interference between multiple data streams within the user. E matrix filter, where:

The matrix N of the N matrix filter is formed by horizontally splicing U sub-blocks.

N" is H: the orthogonal basis corresponding to the zero space, and H is the matrix obtained after deleting all the lines corresponding to the user from the channel matrix, " = 1,2,···, ί/ _;

The matrix E £·" of the unitary matrix filter is a pseudo of H"XN"

The matrix of the fully diagonalized pre-filter is F = NxE.