CN101977098A - User selection and precoding method in wireless multi-user multi-input/output system - Google Patents

Abstract

The invention discloses a user selection and precoding method in a wireless multi-user multi-input/output system, belonging to the field of wireless communication technology. The method comprises the steps that: channel evaluation is implemented through pilot information which is sent by a user to a base station, and according to the resultant channel matrix and signal to noise ratio, channel direction information as well as corresponding code words and the ratios of channel signal to interference and to noise are respectively calculated and fed back to the base station, and the base station implements user selection respectively according to the channel direction information and implements precoding according to the ratios of channel signal to interference and to noise. By adopting a codebook based on discrete Fourier transform matrix, the method realizes the user selection and the precoding at the base station in the multi-user multi-input/output system, and has the characteristics of supporting multi-user transmission, high system handling capacity, small inter-user interference and convenient codebook storage.

Description

User in the radio multi-user multi-input multi-output system selects and method for precoding

Technical field

What the present invention relates to is a kind of method of wireless communication technology field, specifically is that the user in a kind of radio multi-user multi-input multi-output system selects and method for precoding.

Background technology

(Multiple Input Multiple Output, MIMO) technology are by utilizing the multi-antenna structure of base station and user side, realization diversity and space division multiplexing, thereby a kind of technology of increase throughput of system to traditional multiple-input and multiple-output.On the basis of MIMO technique, people have proposed multi-user's multiple-input and multiple-output (Multiple Users Multiple Input Multiple Output, MU-MIMO) technology again.Multi-user's MIMO technique is that with traditional the different of MIMO technique data can be transmitted for a plurality of users in the base station of adopting multi-user's MIMO technique in synchronization.This improvement makes that the spatial reuse rate of multiuser mimo system is better, has improved the total capacity of system greatly.But the shortcoming that multiuser mimo system brings is to exist between the user of transmitting simultaneously to disturb.Therefore, must design corresponding preprocess method,, and reach optimum throughput with the elimination inter-user interference at the characteristics of multiuser mimo system.Preprocess method in the multiuser mimo system mainly comprises: the user's selection operation of base station end and the pre-encode operation of base station end.

Through existing literature search is found, Taesang Yoo etc. are at " IEEE Journal on Selected Areas in Communications " (the U.S. electric periodical of communicating by letter with Electronic Engineering Association, September the 7th in 2007, volume was the 1478th to the 1491st page) on, delivered " Multi-Antenna Downlink Channels with Limited Feedback and User Selection " (" many antennas down channel research of adopting Limited Feedback and user to select "), this article has proposed, in many antennas multi-input multi-output system, adopt the limited feedback method of channel quality indication (CQI) and channel directional information (CDI), and semi-orthogonal user choosing method (Semi-orthogonal User Selection, SUS), can obtain more excellent throughput of system.Its method is that at first, the maximum user of selective channel quality indication in all users is as first dispatched users; Secondly, find out and first user's correlation other user, and select the maximum user of wherein channel quality indication as second dispatched users less than certain threshold value; Secondly, find out and first user's correlation other user, and select the maximum user of wherein channel quality indication as second dispatched users less than certain threshold value; Then, by that analogy, the dispatched users number equals antenna number, and perhaps semi-orthogonal user collection is empty set; At last, selected user is carried out ZF precoding (Zero-Forcing Precoding), and the transmission data.Its weak point is that the semi-orthogonal user choosing method can not obtain optimum dispatched users set; Adopt the pre-coding matrix at random in the literary composition can not obtain optimum throughput of system.

Find by retrieval again, Kaibin Huang etc. are at " IEEE Transaction on Vehicular Technology " (the vehicle-mounted technology journal of U.S. electric and Electronic Engineering Association, January the 58th in 2009, volume was the 152nd page to the 163rd page) on, delivered " Performance of Orthogonal Beamforming for SDMA With Limited Feedback " (" performance study of orthogonal beams excipient in the SDM system of employing Limited Feedback "), this article has proposed to adopt user choosing method (the Orthogonal User Selection based on the quadrature group, OUS) and the method for precoding of every user's unitary matrice rate controlled (Per-user UnitaryRate Control, PU ²RC), can obtain than the higher throughput of system of close-to zero beam shaping.Its method is at first, the user to be divided into groups according to the channel directional information of its feedback; Secondly, the maximum user of selective channel quality indication in each group is as the pre-scheduling user of this group and get rid of this and organize other user; Then, calculate all Orthogonal Composite among the pre-scheduling user and speed, and get with the speed maximum one group as the dispatched users group; At last, dispatched users is carried out the precoding of every user's unitary matrice rate controlled, and send data.Its weak point is that the method for precoding of every user's unitary matrice rate controlled requires the complete quadrature of channel between the user, has limited the use of this method in reality system; Completely orthogonal user choosing method has loss with respect to semi-perpendicular user choosing method on the throughput; The scheme that provides in the literary composition is not considered the design of precoding codebook.

Summary of the invention

The present invention is directed to the prior art above shortcomings, provide the user in a kind of radio multi-user multi-input multi-output system to select and method for precoding, adopted code book based on the discrete Fourier transform matrix, realized the user's selection operation and the precoding processing of base station end in the multiuser mimo system, has the multi-user of support transmission, the throughput of system height, inter-user interference is little, the characteristics of code book convenient storage.

The present invention is achieved by the following technical solutions, the present invention carries out channel estimating by the pilot frequency information that the user sends the base station, and according to obtain channel matrix and signal to noise ratio respectively calculating channel directional information and corresponding code word thereof and channel signal to disturb and noise ratio after feed back to the base station, carry out user's selection according to the channel directional information respectively by the base station, and precoding processing is carried out in interference and noise ratio according to channel signal.

Described channel estimating is meant:

1) adopt least square method to obtain the estimated value of channel matrix at the pilot frequency information place;

2) adopt the linear minimum mean-squared error algorithm at frequency domain, channel matrix is carried out the frequency domain channel estimated value that interpolation obtains the data message place.

3) adopt the linear minimum mean-squared error algorithm in time domain, channel matrix is carried out the time domain channel estimated value that interpolation obtains the data message place.

4) from the channel estimation value of time domain and frequency domain, extract channel matrix and channel signal to noise ratio.

Described calculating channel directional information is meant:

A) channel matrix is done normalized, obtain the channel matrix after the normalization;

B) utilization is based on the code book and the channel normalization matrix of discrete Fourier transform matrix, and the pairing code word of calculating channel directional information specifically is meant: the code book F based on the discrete Fourier transform matrix is:

$F={\left(\begin{array}{cccc}{f}_1^H& f_2^H& ...& f_N^H\end{array}\right)}^H=\frac{1}{\sqrt{M}}\left(\begin{array}{cccc}{e}^{j\frac{2\mathrm{\π}}{N}{u}_1\·0}& e^{j\frac{2\mathrm{\π}}{N}{u}_2\·0}& ...& e^{j\frac{2\mathrm{\π}}{N}{u}_M\·0}\\ e^{j\frac{2\mathrm{\π}}{N}{u}_1\·1}& e^{j\frac{2\mathrm{\π}}{N}{u}_2\·1}& ...& e^{j\frac{2\mathrm{\π}}{N}{u}_M\·1}\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ e^{j\frac{2\mathrm{\π}}{N}{u}_1\·(N-1)}& e^{j\frac{2\mathrm{\π}}{N}{u}_2\·(N-1)}& ...& e^{j\frac{2\mathrm{\π}}{N}{u}_M\·(N-1)}\end{array}\right)$

Wherein: F is the code book based on the discrete Fourier transform matrix, and M is the antenna for base station number, u=(u ₁u ₂... u _M) for generating vector, f _iBe i code word in the code book.By selecting to generate vectorial u, obtain the maximum F of orthogonal code group as code book.

C) utilization is based on code book, channel directional information and the channel signal to noise ratio ρ of discrete Fourier transform matrix, and the calculating channel signal is to disturbing and noise ratio.

Described user selects to be meant:

Wherein: S is a dispatched users, U _iBe all users' i kind transmission combination, R _iFor being combined as U the user _iThe time system total throughout.

Described precoding processing is meant: B=WX, and wherein: W is a pre-coding matrix, and X is the information matrix that all users information to be sent is formed among the S, and B is the transmission signal after the precoding.

Compared with prior art, the invention has the beneficial effects as follows: the code book convenient storage, feedback overhead is little; Throughput of system is higher, and inter-user interference is little; Computation complexity is littler, and the user selects more excellent.

Description of drawings

Fig. 1 is a flow chart of the present invention.

When Fig. 2 adopts 4 antennas and the transmission of 16 code word code books for base station end among the embodiment, the system throughput contrast schematic diagram under the different user system of selection.

When Fig. 3 adopts 4 antennas and the transmission of 16 code word code books for base station end among the embodiment, the system throughput contrast schematic diagram under the different precoding codebooks.

Embodiment

Below embodiments of the invention are elaborated, present embodiment is being to implement under the prerequisite with the technical solution of the present invention, provided detailed execution mode and concrete operating process, but protection scope of the present invention is not limited to following embodiment.

Present embodiment is applicable to multiuser mimo system, and 4 dual polarized antenna transmission can be adopted in the base station.Each user antenna number is 1, and the data fluxion is 1.Channel adopts flat fading, quasi-static Rayleigh channel, the element of channel matrix is the multiple Gaussian random variable of independent identically distributed zero-mean, unit variance, noise is that average is zero, covariance matrix is the additive white Gaussian noise of unit matrix, and the channel of each user's experience is separate.

As shown in Figure 1, present embodiment may further comprise the steps:

The first step, the base station sends pilot frequency information to the user.

In second step, the user receives pilot frequency information, and channel is carried out channel estimating, obtains channel matrix and signal to noise ratio.

Described channel estimating may further comprise the steps:

1) (Least Square LS) obtains the estimated value of channel matrix to adopt least square method at the pilot frequency information place

2) (Linear Minimum Mean Square Error, LMMSE) algorithm carries out the frequency domain channel estimated value that interpolation obtains the data message place to channel matrix to adopt linear minimum mean-squared error at frequency domain.

3) adopt the linear minimum mean-squared error algorithm in time domain, channel matrix is carried out the time domain channel estimated value that interpolation obtains the data message place.

4) from the channel estimation value of time domain and frequency domain, extract channel matrix h and channel signal to noise ratio ρ.

The 3rd step, the user is according to channel matrix h and code book C, respectively the calculating channel directional information (Channel Direction Information, CDI) pairing code word and channel signal to disturb and noise ratio (Signal to Interference and Noise Ratio, SINR).

The calculating of described channel directional information may further comprise the steps:

1) channel matrix h is done normalized, obtain the channel directional information

${\hat{h}}_K=h_K/\left|\right|h_K\left|\right|$

Wherein: h _KThe channel matrix of representing K user, The channel directional information of representing K user.

2) utilization is based on the code book C and the channel method information of discrete Fourier transform matrix

The pairing code word of calculating channel directional information:

$d_k=\arg \underset{f_j\∈C}{\max }\left|{\hat{h}}_k{f}_j^H\right|$

Wherein: d _kRepresent K the pairing code word of subscriber channel directional information, f _jJ code word in the middle of the expression code book C.

3) utilization is based on code book C, channel matrix h, the channel directional information of discrete Fourier transform matrix

Code word d and channel signal to noise ratio ρ, the calculating channel signal is to disturbing and noise ratio:

${\mathrm{SINR}}_k=\frac{\mathrm{\ρ}{\left|\right|h_k\left|\right|}^2{\left|{\hat{h}}_k{d}_k^H\right|}^2}{1+\mathrm{\ρ}{\left|\right|h_k\left|\right|}^2\underset{f_j\≠w_k}{\mathrm{\Σ}}{\left|{\hat{h}}_k{f}_j^H\right|}^2}$

Wherein: SINR _kThe channel signal that is K user is to disturbing and noise ratio, and ρ is the channel signal to noise ratio.

In the present embodiment, code book C is based on the code book of discrete Fourier transform matrix.

When 4 antenna transmission were adopted in the base station, the code book of its 8 code word was:

$\left[\begin{array}{cccc}0.5000& 0.5000& 0.5000& 0.5000\\ 0.3536+0.3536i& -0.3536+0.3536i& -0.3536-0.3536i& 0.3536-0.3536i\\ 0.5000i& -0.5000i& 0.5000i& -0.5000i\\ -0.3536+0.3536i& 0.3536+0.3536i& 0.3536-0.3536i& -0.3536-0.3536i\\ -0.5000& -0.5000& -0.5000& -0.5000\\ -0.3536-0.3536i& 0.3536-0.3536i& 0.3536+0.3536i& -0.3536+0.3536i\\ -0.5000i& 0.5000i& -0.5000i& 0.5000i\\ 0.3536-0.3536i& -0.3536-0.3536i& -0.3536+0.3536i& 0.3536+0.3536i\end{array}\right]$

When 4 antenna transmission were adopted in the base station, the code book of its 16 code word was:

$\left[\begin{array}{cccc}0.5000& 0.5000& 0.5000& 0.5000\\ 0.4619+0.1913i& 0.1913+0.4619i& -0.4619-0.1913i& -0.1913-0.4619i\\ 0.3536+0.3536i& -0.3536+0.3536i& 0.3536+0.3536i& -0.3536+0.3536i\\ 0.1913+0.4619i& -0.4619-0.1913i& -0.1913-0.4619i& 0.4619+0.1913i\\ 0.5000i& 0.5000i& 0.5000i& -0.5000i\\ -0.1913+0.4619i& 0.4619-0.1913i& 0.1913-0.4619i& -0.4619+0.1913i\\ -0.3536+0.3536i& 0.3536+0.3536i& -0.3536+0.3536i& 0.3536+0.3536i\\ -0.4619+0.1913i& -0.1913+0.4619i& 0.4619-0.1913i& 0.1913-0.4619i\\ -0.5000& -0.5000& -0.5000& -0.5000\\ -0.4619-0.1913i& -0.1913-0.4619i& 0.4619+0.1913i& 0.1913+0.4619i\\ -0.3536-0.3536i& 0.3536-0.3536i& -0.3536-0.3536i& 0.3536-0.3536i\\ -0.1913-0.4619i& 0.4619+0.1913i& 0.1913+0.4619i& -0.4619-0.1913i\\ -05000i& 0.5000i& -0.5000i& 0.5000i\\ 0.1913-0.4619i& -0.4619+0.1913i& -0.1913+0.4619i& 0.4619-0.1913i\\ 0.3536-0.3536i& -0.3536-0.3536i& 0.3536-0.3536i& -0.3536-0.3536i\\ 0.4619-0.1913i& 0.1913-0.4619i& -0.4619+0.1913i& -0.1913+0.4619i\end{array}\right]$

In the 4th step, the user feeds back to the base station with the code word and the channel signal of channel directional information correspondence to interference and noise ratio, and user's selection is carried out according to feedback information in the base station.

Described user selects, and may further comprise the steps:

1) according to the code word of feedback the user is divided into groups, the user who feeds back same code word is divided into one group:

T _j＝{u _k|d _k＝f _j，f _j∈C}

Wherein: T _iUser's group of representing j code word among the corresponding code book C, u _kRepresent K user.

2) from the pairing user of each code word group, take out channel signal for disturbing the representative of consumer as this code word, and abandon this and organize other user with the user of noise ratio maximum:

${\mathrm{\ξ}}_i=\arg \underset{u_k\∈T_i}{\max }{\mathrm{SINR}}_k$

Wherein: ξ _iThe expression user organizes T _iRepresentative of consumer.

3), calculate the total throughout of all code word representative of consumer in this group, and one group of the total throughout maximum as dispatched users to semi-orthogonal Codeword Sets all in the code book:

$S=\arg \underset{O_i}{\max }\underset{{\mathrm{\ξ}}_k\∈O_i}{\mathrm{\Σ}}\log (1+{\mathrm{SINR}}_k)$

Wherein: S is the dispatched users group, O _iBe i semi-orthogonal Codeword Sets among the code book C.

In the 5th step, the ZF precoding processing is done to all users in the dispatched users group, concurrent feed signals according to feedback code in the base station.

Described ZF precoding processing is:

B＝WX＝H ^H(HH ^H) ^-1X

Wherein: W is a pre-coding matrix, and H is the channel matrix that all user's code words are formed among the S, and X is the information matrix that all users information to be sent is formed among the S, and B is the transmission signal after the precoding.

The scheme of present embodiment is designated as scheme 1; Adopt the scheme of the method for precoding of traditional semi-orthogonal user choosing method and present embodiment to be designated as scheme 2; Adopt the scheme of the method for precoding of traditional semi-orthogonal user choosing method and 3GPP LTE (Long Term Evolution) system to be designated as scheme 3.

In multiuser mimo system, adopt scheme 1 and scheme 2 to carry out the user respectively and select and precoding, the system throughput that obtains compares schematic diagram as shown in Figure 2.As shown in Figure 2: be respectively 0dB in the channel signal to noise ratio, under the situation of 10dB and 20dB, the system throughput of scheme 1 is higher than the system throughput of scheme 2, and along with the increase of number of users, the advantage of scheme 1 is more obvious.In view of two schemes all adopt identical method for precoding, therefore as can be known, under identical complexity condition, adopt the user choosing method of present embodiment more thorough to the search of user's space, selected user's combination is more excellent, thereby system throughput gets a promotion.

In multiuser mimo system, adopt scheme 2 and scheme 3 to carry out the user respectively and select and precoding, the system throughput that obtains compares schematic diagram as shown in Figure 3.As shown in Figure 3: be respectively 0dB in the channel signal to noise ratio, under the situation of 10dB and 20dB, number of users was less than 45 o'clock, and the system throughput of scheme 2 all is higher than the system throughput of scheme 3; Greater than 45 o'clock, the system throughput of two schemes all reached system's upper limit at number of users.Simultaneously, the system throughput of scheme 2 at number of users greater than reaching system's upper limit at 25 o'clock, the system throughput of scheme 3 then at number of users greater than just reaching system's upper limit at 40 o'clock.In view of two schemes all adopt identical user choosing method, therefore as can be known, under identical complexity condition, when the method for precoding of employing present embodiment and precoding codebook, orthogonality between the scheduled user is better, inter-user interference is littler, and system throughput is higher, and system throughput can be issued to the upper limit in the lower situation of number of users.

Therefore, it is higher that this method has a system throughput, system throughput rises faster with number of users, orthogonality is better between dispatched users, the characteristics that inter-user interference is littler, go for various movements and wireless communication system, particularly adopt the wireless communication system of multi-user's MIMO technique and precoding technique.

Claims (6)

1. the user in the radio multi-user multi-input multi-output system selects and method for precoding, it is characterized in that, the pilot frequency information that the base station is sent by the user carries out channel estimating, and according to obtain channel matrix and signal to noise ratio respectively calculating channel directional information and corresponding code word thereof and channel signal to disturb and noise ratio after feed back to the base station, carry out user's selection according to the channel directional information respectively by the base station, and precoding processing is carried out in interference and noise ratio according to channel signal.

2. the user in the radio multi-user multi-input multi-output system according to claim 1 selects and method for precoding, it is characterized in that described channel estimating is meant:

1) adopt least square method to obtain the estimated value of channel matrix at the pilot frequency information place;

2) adopt the linear minimum mean-squared error algorithm at frequency domain, channel matrix is carried out the frequency domain channel estimated value that interpolation obtains the data message place.

3) adopt the linear minimum mean-squared error algorithm in time domain, channel matrix is carried out the time domain channel estimated value that interpolation obtains the data message place.

4) from the channel estimation value of time domain and frequency domain, extract channel matrix and channel signal to noise ratio.

3. the user in the radio multi-user multi-input multi-output system according to claim 1 selects and method for precoding, it is characterized in that described calculating channel directional information is meant:

A) channel matrix is done normalized, obtain the channel matrix after the normalization;

B) utilization is based on the code book and the channel normalization matrix of discrete Fourier transform matrix, the pairing code word of calculating channel directional information;

C) utilization is based on code book, channel directional information and the channel signal to noise ratio ρ of discrete Fourier transform matrix, and the calculating channel signal is to disturbing and noise ratio.

4. the user in the radio multi-user multi-input multi-output system according to claim 3 selects and method for precoding, and it is characterized in that described step b) specifically is meant: the code book F based on the discrete Fourier transform matrix is:

$F={\left(\begin{array}{cccc}{f}_1^H& f_2^H& ...& f_N^H\end{array}\right)}^H=\frac{1}{\sqrt{M}}\left(\begin{array}{cccc}{e}^{j\frac{2\mathrm{\π}}{N}{u}_1\·0}& e^{j\frac{2\mathrm{\π}}{N}{u}_2\·0}& ...& e^{j\frac{2\mathrm{\π}}{N}{u}_M\·0}\\ e^{j\frac{2\mathrm{\π}}{N}{u}_1\·1}& e^{j\frac{2\mathrm{\π}}{N}{u}_2\·1}& ...& e^{j\frac{2\mathrm{\π}}{N}{u}_M\·1}\\ .& .& .& .\\ .& .& .& .\\ .& .& .& .\\ e^{j\frac{2\mathrm{\π}}{N}{u}_1\·(N-1)}& e^{j\frac{2\mathrm{\π}}{N}{u}_2\·(N-1)}& ...& e^{j\frac{2\mathrm{\π}}{N}{u}_M\·(N-1)}\end{array}\right)$

Wherein: F is the code book based on the discrete Fourier transform matrix, and M is the antenna for base station number, u=(u ₁u ₂... u _M) for generating vector, f _iBe i code word in the code book.By selecting to generate vectorial u, obtain the maximum F of orthogonal code group as code book.

5. the user in the radio multi-user multi-input multi-output system according to claim 1 selects and method for precoding, it is characterized in that, described user selects to be meant:

Wherein: S is a dispatched users, U _iBe all users' i kind transmission combination, R _iFor being combined as U the user _iThe time system total throughout.

6. the user in the radio multi-user multi-input multi-output system according to claim 3 selects and method for precoding, it is characterized in that, described precoding processing is meant: B=WX, wherein: W is a pre-coding matrix, X is the information matrix that all users information to be sent is formed among the S, and B is the transmission signal after the precoding.

Images