CN105790912A - Selection and distribution method of uplink piolet frequency in massive MIMO system in TDD mode - Google Patents

Abstract

The invention provides a selection and distribution method of uplink piolet frequency in a massive MIMO system in a TDD mode. The method comprises the following steps: each cell interacting related piolet frequency sequence use information of a local cell to an adjacent cell, each cell base station interacting the uplink piolet frequency sequence use information and a DTX mode of a user in the local cell to the adjacent base station; the base station determining a user using one unlink piolet frequency sequence in the cell and a main sub-space of a channel matrix between the cell base stations, and the user using the same uplink piolet frequency sequence in the adjacent cell and the main sub-space of the channel matrix between the cell base stations; judging whether the user in the local cell can use the same uplink piolet frequency sequence with the user in the adjacent cell according to the relation of two main sub-spaces. The selection and distribution of the uplink piolet frequency reduces the uplink piolet frequency pollution of the massive MIMO system in the TDD mode through the information interaction between the user and the base station and the information interaction between the base stations.

Description

The selection of ascending pilot frequency and distribution method in tdd mode massive mimo system

Technical field

The present invention relates to a kind of ascending pilot frequency Pollution restraint method in tdd mode multiple cell massiveMIMO in 5G system, particularly relate to a kind of TDD (TimeDivisionDuplexing, time division duplex) selection of ascending pilot frequency and distribution method in mode m assiveMIMO (massiveMultipleInputMultipleOutput, extensive multiple-input and multiple-output) system.

Background technology

In the downlink of the massiveMIMO system of tdd mode, in order to reduce the pilot-frequency expense of downlink, it is common to use the method for the multiple wave beam of pilot resources multiplexing, thus the ascending pilot frequency introduced in multiple cell MassiveMIMO system pollutes.

As it is shown in figure 1, the extensive MU-MIMO system of tdd mode multiple cell has L community.There are the base station (BaseState, BS) of a N antenna and K single-antenna subscriber in each community.Assuming that all L communities use K identical pilot frequency sequence, and represent with τ × K orthogonal dimension matrix Φ, it meets Φ^HΦ=τ I.Further it is assumed that the pilot transmission of different districts is to synchronize.So, the reception signal matrix of base station iIt is represented by:

$Y_i^p=\sqrt{{\mathrm{\ρ}}_p}\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{H}_{i,l}{\mathrm{\Φ}}^T+N_i^p---\left(1\right)$

In formula, H_i,l∈C^N×KChannel matrix is tieed up for N × K between K user and base station i in the l of community,For the noise matrix (wherein every for average 0, unit variance) at base station i place, ρ during pilot transmission_pFor pilot transmission power.

In order to estimate channel, base station i is received signal and is projected to Φ^*On, thus obtaining for estimating channel H_i,iStatistic.By formula (1) it can be seen that the channel matrix estimated is:

${\hat{H}}_{i,i}=\frac{1}{\sqrt{{\mathrm{\ρ}}_p\mathrm{\τ}}}{Y}_i^p{\mathrm{\Φ}}^{*}=H_{i,i}+\underset{l\≠i}{\mathrm{\Σ}}{H}_{i,l}+\frac{1}{\sqrt{{\mathrm{\υ}}_p\mathrm{\τ}}}{N}_i^p{\mathrm{\Φ}}^{*}---\left(2\right)$

In the derivation of above formula, make use of Φ^HΦ=τ I.Kth rowFor channel vector h_i,k,iEstimated value.By formula (2) it can be seen thatFor different districts uses those subscriber channels vector h of same pilot sequence_i,k,l(l=1 ..., L) linear combination.This phenomenon is called pilot pollution.Owing to pilot pollution causes channel estimation errors, not only make uplink signal detection performance reduce, and if adopt this channel matrix to carry out the precoding of downlink, then undesired user can be caused very big interference.Therefore, the method reducing pilot pollution has been carried out substantial amounts of research work by Many researchers, it is proposed that some solutions.

At present, it has been proposed that some suppress the scheme of pilot pollution.Several typical scheme is described below.

(1) based on the pilot pollution Restrain measurement of agreement

A kind of the most direct method is to use the user of non-orthogonal pilot sequence to lower pilot pollution effect by channeling or minimizing.But, it is however generally that, channeling can not utilize the advantage of massiveMIMO energy concurrent multiplexing a large number of users.Although this is because the method improves the signal to noise ratio SINR of specific user, but the number of users serviced decreases simultaneously.In the above-mentioned methods, all users send pilot frequency sequence simultaneously, and this arises that pilot pollution.In order to reduce pilot pollution, prior art proposes a kind of scheme based on time shift (asynchronous) host-host protocol, and its basic thought is to divide the cell into some groups of A₁,…,A_Γ, and in each group, using time shift host-host protocol, Fig. 2 is the example of Γ=3.As group A₁When middle user launches pilot tone, organize A₂In base station then send the data signal of downlink.Thus avoid group A₁With group A₂Pilot pollution between middle user.Meanwhile, group A₁Middle base station needs in existence from organizing A₂With group A₃Estimation group A during the downlink transmitted signal of middle base station₁The channel matrix of middle user.Transmitting power ρ due to downlink_dIt is typically much deeper than the power ρ of pilot frequency sequence_p, therefore it be not immediately clear the method is provided that how many gains.

(2) based on the method for precoding

Method for precoding based on multi-cell cooperating can suppress pilot pollution effect.But, the mutual expense of information needed between base station increases along with the increase of antenna number.Therefore, these methods are only applicable to the mimo system with finite antenna number.For the mutual expense of advantage and restricted information obtaining cooperation, prior art proposes a kind of pilot pollution precoding (PCP) scheme.PCP scheme is based on two hypothesis: (a) each base station can obtain the source signal of all users in all communities；B () all base stations can obtain large scale fading coefficient d_{I, k, l}.Due to pilot pollution, channel estimation vectorFor different districts uses those subscriber channels vector h of same pilot sequence_i,k,l(l=1 ..., L) linear combination.In PCP scheme, each base station is not the interference suppressing to be caused by the channel estimation value polluted, but utilizes pilot pollution to send information for users all in network.But, there are some researches show only as antenna number N > 10⁶Time, PCP scheme just begins with performance gain.

(3) based on the pilot pollution suppressing method arriving angle (AOA)

There are some researches show, in actual channel model, for having identical or non-orthogonal pilot sequence certain user, it is likely to each other and is absent from interference.According to the multipath channel models about even linear array (ULA), from the l of community, the channel matrix of user k to base station i has following form:

$h_{i,k,l}=\frac{1}{\mathrm{\Θ}}\underset{\mathrm{\θ}=1}{\overset{\mathrm{\Θ}}{\mathrm{\Σ}}}a\left(\mathrm{\θ}\right){\mathrm{\η}}_{i,k,l}---\left(3\right)$

In formula, Θ is multipath number,Independent of path sequence number θ,For the average path loss of user, the direction vector that a (θ) is array.For ULA, direction vector is represented by:

$a\left(\mathrm{\θ}\right)=\left[\begin{array}{c}1\\ e^{-j2\mathrm{\π}\frac{D}{\mathrm{\λ}}\cos \left(\mathrm{\θ}\right)}\\ .\\ .\\ .\\ e^{-j2\mathrm{\π}\frac{(N-1)D}{\mathrm{\λ}}\cos \left(\mathrm{\θ}\right)}\end{array}\right]---\left(4\right)$

In formula, D is antenna distance, and λ is the wavelength of carrier wave, and θ arrives angle (AOA) at random for what have probability density function (PDF) f (θ).Verified, even if user uses identical pilot frequency sequence, but as long as mutually having the PDF of nonoverlapping AOA, then its pilot tone is difficult to interfere.According to this thought, prior art also teaches a kind of scheme reducing pilot pollution, distribute to that class user of the PDF mutually with not overlapping AOA by identical pilot frequency sequence.The method requires that the correlation matrix between each user of each base station meets its dominant subspace this condition not overlapping, and requires that the correlation matrix between itself and all users is known in each base station.If it is to say, there be L base station, each base station having K user, then require that each base station can obtain LK correlation matrix between itself and LK user.

Therefore, the method has the disadvantage that

A, the method require that the correlation matrix between each user of each base station meets its dominant subspace this condition not overlapping, and require that the correlation matrix between itself and all users is known in each base station.If it is to say, there be L base station, each base station having K user, then require that each base station can obtain LK correlation matrix between itself and LK user.But, actually base station is difficult to obtain the information of LK correlation matrix.

B, when in different districts, 2 users have close AOA to a certain base station, then this base station cannot be distinguished by the two user, thus causing that ascending pilot frequency pollutes.

(4) based on the blind suppressing method of the pilot pollution of Subspace partition technology

Method based on Subspace partition technology can suppress pilot pollution equally.In the prior art, it is proposed that feature based value decomposes the channel estimating of (EVD) and the channel vector method of estimation of iterative projection least square.The channel vector of the channel estimation methods requirement different user that feature based value decomposes (EVD) is orthogonal, and this hypothesis makes base station that the statistic receiving data can be utilized to estimate channel vector.The property the taken advantage of scalar fuzziness of this kind of method inherence may utilize mutually orthogonal district pilots sequence and solves.

The method may be summarized as follows: in the ul transmissions stage, and user first sends data signal, then sends district pilots sequence.

Estimation procedure comprises the steps:

Step 1: estimate the covariance matrix receiving signal:

$R_{y_l^u}=E\left[y_l^u{\left(y_l^u\right)}^H\right]\≈{\hat{R}}_{y_l^u}=\frac{1}{T}\underset{t=1}{\overset{T}{\mathrm{\Σ}}}{y}_l^u\left[t\right]{\left(y_l^u\right[t\left]\right)}^H---\left(5\right)$

In formula,It it is the lth base station reception signal at moment t.

Step 2: solve N × K and tie up matrix U_l, its kth is classified asCorrespond to close to eigenvalue N ρ_ud_l,l,kCharacteristic vector.And, by means of pilot frequency sequence, the property taken advantage of matrix Ξ can be obtained_lEstimated value

Step 3: calculate estimated value ${\hat{G}}_{l,l}=U_l{\widehat{\mathrm{\Ξ}}}_l.---\left(6\right)$

But, if using two users of same pilot sequence to be all located at edge and when nearer in adjacent two communities, two users are most likely in identical subspace, then EVD method is difficult to by the subspace of two users separately.Now, will be unable to effectively reduce pilot pollution based on the blind method of EVD.

Summary of the invention

The shortcoming of prior art in view of the above, it is an object of the invention to provide the selection of ascending pilot frequency in a kind of tdd mode massiveMIMO system and distribution method, information between and base station mutual by the information between user and base station is mutual, significantly reduces the ascending pilot frequency in tdd mode massiveMIMO system and pollutes.

For achieving the above object and other relevant purposes, the present invention provides selection and the distribution method of ascending pilot frequency in a kind of tdd mode massiveMIMO system, assume that the massiveMIMO system of tdd mode has L community, each base station to have M root antenna, be up to K bar uplink pilot sequence pilot frequency sequence a_k, k=1,2 ..., K, each community has K user；Each base station is by K bar uplink pilot sequence a_kDistributing to K user in community, user k uses pilot frequency sequence a_k；The selection of ascending pilot frequency and distribution method comprise the following steps:

Step S1: assume use pilot frequency sequence a in the l of community_kThe DTX pattern of user k be p_l,k, the associated pilot of this community is used information with form { the ID:(user k uplink pilot sequence a of community l by community l_kSequence number, user k DTX pattern p_l,k, k=1,2 ..., K) } mutual give its adjacent (L-1) individual community；

Step S2: the associated pilot received from adjacent (L-1) individual community when the base station of community l uses information { the ID:(user k uplink pilot sequence a of community j_kSequence number, the DTX pattern p of user k_j,k, k=1,2 ..., K), j=1 ..., l-1, l+1 ..., after L}, according to this community and neighbor cell use pilot frequency sequence a_kThe DTX pattern of user, calculate use pilot frequency sequence a in adjacent (L-1) individual community_kUser send the union during signal, obtain use pilot frequency sequence a in adjacent (L-1) individual community_kTotal DTX pattern, and calculate total DTX patternSupplementary set

Assume to receive T in the base station of total inspection time small area l_pIndividual reception sample of signal, whereinTransmission signal during in have T_aIndividual reception sample of signal, andTransmission signal during in have T_b=(T_p-T_a) individual reception sample of signal；

According to T_aAnd T_bRelation determine in this community use pilot frequency sequence a_kUser k whether can with adjacent (L-1) individual community uses pilot frequency sequence a_kUser use identical uplink pilot sequence a_k。

According to the selection of ascending pilot frequency in above-mentioned tdd mode massiveMIMO system and distribution method, wherein: in described step S2, the uplink received signal of the base station of community lFor

$Y_l^p=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{H}_{l,l}{a}_k+w_l^p$

Wherein, ρ_pRepresent the transmitting power of ascending pilot frequency, H_l,jRepresent and community j uses pilot frequency sequence a_kUser and the base station of community l between channel matrix,For additive white Gaussian noise.

According to the selection of ascending pilot frequency in above-mentioned tdd mode massiveMIMO system and distribution method, wherein: if T_bLength can calculate in the l of community use pilot frequency sequence a_kUser uplink base station place receive signal spatial correlation matrix, then the base station of community l calculates respectivelyTransmission signal during andSend in during signal the spatial correlation matrix receiving signal；

?Transmission signal during, base station receives only the pilot signal transmitted of user in its adjacent (L-1) individual community, receives signal and is expressed as

$\begin{array}{cc}{y}_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1,j\≠l}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p& t\∈{\tilde{p}}_{l,k}\end{array}$

?To send the spatial correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\underset{t\∈{\tilde{p}}_k}{\mathrm{\Σ}}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beTo obtain the feature matrix corresponding to dominant eigenvalue

?Transmission signal during, base station receives only the pilot signal sent of user k in its this community, namely receives signal and is expressed as

$\begin{array}{cc}{y}_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}{H}_{l,j}{a}_k+w_l^p& t\∈{\tilde{p}}_{l,k}^c\end{array}$

?To send the spatial correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\underset{t\∈{\tilde{p}}_{l,k}^c}{\mathrm{\Σ}}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beTo obtain the feature matrix corresponding to dominant eigenvalue

The base station of community l according toWithRelation, it is determined that in the l of community user k can with neighbor cell in use pilot frequency sequence a_kUser use identical pilot frequency sequence a_k。

Further, according to the selection of ascending pilot frequency in above-mentioned tdd mode massiveMIMO system and distribution method, wherein: the base station of community l according toWithRelation, it is determined that in the l of community user k can with neighbor cell in use pilot frequency sequence a_kUser use identical pilot frequency sequence a_kTime,

IfCommunity l uses pilot frequency sequence a_kUser can not with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k；

IfCommunity l uses pilot frequency sequence a_kUser k can with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k。

According to the selection of ascending pilot frequency in above-mentioned tdd mode massiveMIMO system and distribution method, wherein: if adjacent (L-1) individual community of community l uses pilot tone a_kThe user of sequence is all without using DTX transmission mode, then now the base station of community l is use pilot tone a in this community_kThe user k of sequence configures the DTX transmission mode p of a short-term_l,k, and notify that this user enters DTX transmission mode p_l,k；After receiving this configuration information, this user enters this DTX transmission mode p_l,k；

In the l of community, base station calculates p respectively_l,kTransmission signal period p '_l,kNot pilot signal transmitted period p "_l,kThe spatial correlation matrix of middle reception signal, and determine use pilot frequency sequence a in neighbor cell and this community respectively_kThe feature matrix of dominant eigenvalue of mobile subscriber；

Whether the user k that the relation between feature matrix according to two dominant eigenvalues is determined in this community in user k and neighbor cell can use identical pilot frequency sequence a_k。

Further, according to the selection of ascending pilot frequency in above-mentioned tdd mode massiveMIMO system and distribution method, wherein: according to whether the user k that the relation between the feature matrix of two dominant eigenvalues is determined in this community in user k and neighbor cell can use identical pilot frequency sequence a_kTime,

The transmission pilot frequency sequence a of user k in the l of community_kPeriod p '_l,k, serving BS is simultaneously received the pilot frequency sequence a that this user sends_kPilot frequency sequence a with adjacent cell user transmission_k, receive signal and be expressed as:

$y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p$

p_kTo send the correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\mathrm{\Σ}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beAnd obtain the feature matrix corresponding to its dominant eigenvalue

In the l of community, user k's does not send pilot frequency sequence a_kPeriod p "_l,k, serving BS receives only the pilot frequency sequence a that its adjacent (L-1) community user sends_k；Reception signal is expressed as:

$Y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1,j\≠1}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p$

The correlation matrix receiving signal is:

${\hat{R}}_{b,k}=\mathrm{\Σ}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beAnd obtain the feature matrix corresponding to its dominant eigenvalue

The base station of community l according toWithDimension relation, it is determined that in its this community, whether the user k in user k and adjacent (L-1) individual community can use identical pilot frequency sequence a_k。

Further, according to the selection of ascending pilot frequency in above-mentioned tdd mode massiveMIMO system and distribution method, wherein: the base station of community l according toWithDimension relation, it is determined that in its this community, whether the user k in user k and adjacent (L-1) individual community can use identical pilot frequency sequence a_kTime,

IfDimension withDimension identical, community l uses pilot frequency sequence a_kUser can not with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k；

IfDimension more thanDimension, community l uses pilot frequency sequence a_kUser can with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k。

As it has been described above, the selection of ascending pilot frequency and distribution method in the tdd mode massiveMIMO system of the present invention, have the advantages that

(1) information between mutual by the information between user and base station and base station is mutual, significantly reduces the ascending pilot frequency in tdd mode massiveMIMO system and pollutes；

(2) these information are likely to be standardized in following massiveMIMO system alternately.

Accompanying drawing explanation

Fig. 1 is shown as the schematic diagram of the extensive MU-MIMO system in prior art small area；

The schematic diagram that Fig. 2 is shown as in prior art based on the pilot schemes of time shift；

Fig. 3 is shown as in the tdd mode massive-MIMO system of the present invention selection of ascending pilot frequency and the flow chart of distribution method.

Detailed description of the invention

Below by way of specific instantiation, embodiments of the present invention being described, those skilled in the art the content disclosed by this specification can understand other advantages and effect of the present invention easily.The present invention can also be carried out by additionally different detailed description of the invention or apply, and the every details in this specification based on different viewpoints and application, can also carry out various modification or change under the spirit without departing from the present invention.

It should be noted that, the diagram provided in the present embodiment only illustrates the basic conception of the present invention in a schematic way, then assembly that in graphic, only display is relevant with the present invention but not component count when implementing according to reality, shape and size drafting, during its actual enforcement, the kenel of each assembly, quantity and ratio can be a kind of random change, and its assembly layout kenel is likely to increasingly complex.

In the tdd mode massive-MIMO system of the present invention in the selection of ascending pilot frequency and distribution method, first, the associated pilot sequence of this community is used information alternately to neighbor cell by each community, and the uplink pilot sequence of user in community is used information and DTX pattern alternately to adjacent each base station by each cell base station；Then, base station uses the interactive information of information according to the associated pilot sequence from neighbor cell base station, determine the dominant subspace of channel matrix between the user using a certain uplink pilot sequence in this community and this cell base station, and neighbor cell uses the dominant subspace of channel matrix between the user of identical uplink pilot sequence and this cell base station；Finally, the relation of two dominant subspace of gained judge whether the user in this community can use identical uplink pilot sequence with the user in neighbor cell.

Assume that the massiveMIMO system of tdd mode has L community, each base station to have M root antenna.Further it is assumed that be up to K bar uplink pilot sequence pilot frequency sequence a_k(k=1,2 ..., K), and each community has K user.Moreover, it is assumed that each base station is by K bar uplink pilot sequence a_kDistribute to K user in community, and assume that user k uses pilot frequency sequence a_k。

With reference to Fig. 3, in the tdd mode massive-MIMO system of the present invention, the selection of ascending pilot frequency and distribution method comprise the following steps:

Step S1: assume use pilot frequency sequence a in the l of community_kDTX (the user discontinuously propagates) pattern of user k be p_l,k, the associated pilot of this community is used information with form { the ID:(user k uplink pilot sequence a of community l by community l_kSequence number, user k DTX pattern p_l,k, k=1,2 ..., K) } mutual give its adjacent (L-1) individual community.

Step S2: the associated pilot received from adjacent (L-1) individual community when the base station of community l uses information { the ID:(user k uplink pilot sequence a of community j_kSequence number, the DTX pattern p of user k_j,k, k=1,2 ..., K), j=1 ..., l-1, l+1 ..., after L}, according to this community and neighbor cell use pilot frequency sequence a_kThe DTX pattern of user, calculate use pilot frequency sequence a in adjacent (L-1) individual community_kUser send the union during signal, obtain use pilot frequency sequence a in adjacent (L-1) individual community_kTotal DTX pattern, and calculate total DTX patternSupplementary set

Assume to receive T in the base station of total inspection time small area l_pIndividual reception sample of signal, whereinTransmission signal during in have T_aIndividual reception sample of signal, andTransmission signal during in have T_b=(T_p-T_a) individual reception sample of signal.The uplink received signal of the base station of community lThere is following form:

$Y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p---\left(7\right)$

In formula, ρ_pRepresent the transmitting power of ascending pilot frequency, H_l,jRepresent and community j uses pilot frequency sequence a_kUser and the base station of community l between channel matrix,For additive white Gaussian noise.

According to T_aAnd T_bRelation, the base station of community l divides the following two kinds situation (when for other situation, adopt similar processing method) to determine in this community use pilot frequency sequence a_kUser k whether can with adjacent (L-1) individual community uses pilot frequency sequence a_kUser use identical uplink pilot sequence a_k:

A, situation 1:

If T_bLength be enough to calculate in the l of community use pilot frequency sequence a_kUser uplink base station place receive signal spatial correlation matrix, then the base station of community l calculates respectivelyTransmission signal during andSend in during signal the spatial correlation matrix receiving signal.

BecauseTransmission signal during, base station receives only the pilot signal transmitted of user in its adjacent (L-1) individual community, namely receive signal be represented by

$\begin{array}{cc}{y}_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1,j\≠l}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p& t\∈{\tilde{p}}_{l,k}\end{array}---\left(8\right)$

Calculate?To send the spatial correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\underset{t\∈{\tilde{p}}_k}{\mathrm{\Σ}}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H---\left(9\right)$

Then, calculateEigenvalues DecompositionAnd thus obtain the feature matrix corresponding to its dominant eigenvalue

On the other hand, existTransmission signal during, base station receives only the pilot signal sent of user k in its this community, and namely it receives signal and is represented by

$\begin{array}{cc}{y}_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}{H}_{l,l}{a}_k+w_l^p& t\∈{\tilde{p}}_{l,k}^c\end{array}---\left(10\right)$

Then, calculate?To send the spatial correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\underset{t\∈{\tilde{p}}_{l,k}^c}{\mathrm{\Σ}}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H---\left(11\right)$

Then, calculateEigenvalues DecompositionAnd thus obtain the feature matrix corresponding to its dominant eigenvalue

Finally, the base station of community l according toWithRelation, it is determined that in the l of community user k can with neighbor cell in use pilot frequency sequence a_kThose users use identical pilot frequency sequence a_k.Specific as follows:

If a)Show use pilot frequency sequence a in the l of community_kUser k and in the channel matrix between l base station, community and adjacent (L-1) individual community use pilot frequency sequence a_kThose users and l base station, community between channel matrix between there is very strong dependency.Now, community l uses pilot frequency sequence a_kUser and its adjacent cell user ascending pilot frequency between there is strong jamming.So, community l uses pilot frequency sequence a_kUser can not use a_k, it is necessary to distribute other uplink pilot sequence for this user；

If b)Then community l uses pilot frequency sequence a_kUser k and its adjacent cell user ascending pilot frequency between be absent from strong jamming.Now, community l uses pilot frequency sequence a_kUser k and its adjacent cell in those users can use identical uplink pilot sequence a_k。

B, situation 2:

If adjacent (L-1) individual community of community l uses pilot tone a_kThose users of sequence are all without using DTX transmission mode (i.e. T_b=0), then now the base station of community l is use pilot tone a in this community_kThe user k of sequence configures the DTX transmission mode p of a short-term_l,k(without notifying its adjacent base station), and notify that this user enters DTX transmission mode p_l,k.After receiving this configuration information, this user enters this DTX transmission mode p_l,k。

Then, in the l of community, base station BS calculates p respectively_l,kTransmission signal period p '_l,kNot pilot signal transmitted period p "_l,kIn it receives the spatial correlation matrix of signal, and determine respectively neighbor cell and this community use pilot frequency sequence a_kThe feature matrix of dominant eigenvalue of mobile subscriber.

Then, whether the user k determined in this community in user k and adjacent cell according to the relation between the feature matrix of two dominant eigenvalues can use identical pilot frequency sequence a_k。

Concrete defining method is as follows:

The transmission pilot frequency sequence a of user k in the l of community_kPeriod p '_l,k, serving BS is simultaneously received the pilot frequency sequence a that this user sends_kPilot frequency sequence a with its adjacent cell user transmission_k.Therefore, reception signal now is represented by:

$y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p---\left(12\right)$

Then, p is calculated_kTo send the correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\mathrm{\Σ}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H---\left(13\right)$

Then, calculateEigenvalues DecompositionAnd thus obtain the feature matrix corresponding to its dominant eigenvalue

On the other hand, in the l of community user k do not send pilot frequency sequence a_kPeriod p "_l,k, serving BS receives only the pilot frequency sequence a that its adjacent (L-1) community user sends_k.Therefore, reception signal now is expressed as:

$Y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1,j\≠1}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p$

Similarly, the correlation matrix receiving signal is calculated:

${\hat{R}}_{b,k}=\mathrm{\Σ}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H---\left(14\right)$

Similarly, calculateEigenvalues DecompositionAnd thus obtain the feature matrix corresponding to its dominant eigenvalue

Finally, the base station of community l according toWithDimension relation, it is determined that in its this community, whether the user k in user k and adjacent (L-1) individual community can use identical pilot frequency sequence a_k, specific as follows:

If a)Dimension withDimension identical, then show that following two channels are height correlations: channel matrix between user k and its serving BS in the l of community, and in its adjacent (L-1) community user k to the channel matrix of this base station.Now, community l uses pilot frequency sequence a_kUser k be adjacent in (L-1) community and will there is strong jamming between the ascending pilot frequency of user.So, the user k in the l of community can not use uplink pilot sequence a_k, it is necessary to distribute other uplink pilot sequence for this user.

If b)Dimension more thanDimension, then show that the dependency of following two channels is poor: channel matrix between user k and its serving BS in the l of community, and in its adjacent (L-1) community user k to the channel matrix of this base station.Now, community l uses pilot frequency sequence a_kUser k be adjacent in (L-1) community and be absent from strong jamming between the ascending pilot frequency of user.So, now the user k in the l of community can use uplink pilot sequence a_k。

In sum, in the tdd mode massiveMIMO system of the present invention, the selection of ascending pilot frequency and distribution method be mutual by the information between user and base station and information between base station is mutual, significantly reduces the ascending pilot frequency in tdd mode massiveMIMO system and pollutes；These information are likely to be standardized in following massiveMIMO system alternately.So, the present invention effectively overcomes various shortcoming of the prior art and has high industrial utilization.

Above-described embodiment is illustrative principles of the invention and effect thereof only, not for the restriction present invention.Above-described embodiment all under the spirit and category of the present invention, can be modified or change by any those skilled in the art.Therefore, art has usually intellectual such as modifying without departing from all equivalences completed under disclosed spirit and technological thought or change, must be contained by the claim of the present invention.

Claims (7)

1. the selection of ascending pilot frequency and distribution method in a tdd mode massiveMIMO system, it is characterised in that: assume that the massiveMIMO system of tdd mode has L community, each base station to have M root antenna, be up to K bar uplink pilot sequence pilot frequency sequence a_k, k=1,2 ..., K, each community has K user；Each base station is by K bar uplink pilot sequence a_kDistributing to K user in community, user k uses pilot frequency sequence a_k；

The selection of ascending pilot frequency and distribution method comprise the following steps:

Step S1: assume use pilot frequency sequence a in the l of community_kThe DTX pattern of user k be p_l,k, the associated pilot of this community is used information with form { the ID:(user k uplink pilot sequence a of community l by community l_kSequence number, user k DTX pattern p_l,k, k=1,2 ..., K) } mutual give adjacent (L-1) individual community；

Step S2: the associated pilot received from adjacent (L-1) individual community when the base station of community l uses information { the ID:(user k uplink pilot sequence a of community j_kSequence number, the DTX pattern p of user k_j,k, k=1,2 ..., K), j=1 ..., l-1, l+1 ..., after L}, according to this community and neighbor cell use pilot frequency sequence a_kThe DTX pattern of user, calculate use pilot frequency sequence a in adjacent (L-1) individual community_kUser send the union during signal, obtain use pilot frequency sequence a in adjacent (L-1) individual community_kTotal DTX pattern, and calculate total DTX patternSupplementary set

Assume to receive T in the base station of total inspection time small area l_pIndividual reception sample of signal, whereinTransmission signal during in have T_aIndividual reception sample of signal, andTransmission signal during in have T_b=(T_p-T_a) individual reception sample of signal；

According to T_aAnd T_bRelation determine in this community use pilot frequency sequence a_kUser k whether can with adjacent (L-1) individual community uses pilot frequency sequence a_kUser use identical uplink pilot sequence a_k。

2. the selection of ascending pilot frequency and distribution method in tdd mode massiveMIMO system according to claim 1, it is characterised in that: in described step S2, the uplink received signal of the base station of community lFor

$y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p$

Wherein, ρ_pRepresent the transmitting power of ascending pilot frequency, H_l,jRepresent and community j uses pilot frequency sequence a_kUser and the base station of community l between channel matrix,For additive white Gaussian noise.

3. the selection of ascending pilot frequency and distribution method in tdd mode massiveMIMO system according to claim 1, it is characterised in that: if T_bLength can calculate in the l of community use pilot frequency sequence a_kUser uplink base station place receive signal spatial correlation matrix, then the base station of community l calculates respectivelyTransmission signal during andSend in during signal the spatial correlation matrix receiving signal；

?Transmission signal during, base station receives only the pilot signal transmitted of user in its adjacent (L-1) individual community, receives signal and is expressed as

$\begin{array}{cc}{y}_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1,j\≠l}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p& t\∈{\tilde{p}}_{l,k}\end{array}$

?To send the spatial correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\underset{t\∈{\tilde{p}}_k}{\mathrm{\Σ}}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beTo obtain the feature matrix corresponding to dominant eigenvalue

?Transmission signal during, base station receives only the pilot signal sent of user k in its this community, namely receives signal and is expressed as

$\begin{array}{cc}{y}_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}{H}_{l,l}{a}_k+w_l^p& t\∈{\tilde{p}}_{l,k}^c\end{array}$

?To send the spatial correlation matrix during signal as follows:

${\hat{R}}_{b,k}=\underset{t\∈{\tilde{p}}_{l,k}^c}{\mathrm{\Σ}}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beTo obtain the feature matrix corresponding to dominant eigenvalue

The base station of community l according toWithRelation, it is determined that in the l of community user k can with neighbor cell in use pilot frequency sequence a_kUser use identical pilot frequency sequence a_k。

4. the selection of ascending pilot frequency and distribution method in tdd mode massiveMIMO system according to claim 3, it is characterised in that: the base station of community l according toWithRelation, it is determined that in the l of community user k can with neighbor cell in use pilot frequency sequence a_kUser use identical pilot frequency sequence a_kTime,

IfCommunity l uses pilot frequency sequence a_kUser can not with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k；

IfCommunity l uses pilot frequency sequence a_kUser k can with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k。

5. the selection of ascending pilot frequency and distribution method in tdd mode massiveMIMO system according to claim 1, it is characterised in that: if adjacent (L-1) individual community of community l uses pilot tone a_kThe user of sequence is all without using DTX transmission mode, then now the base station of community l is use pilot tone a in this community_kThe user k of sequence configures the DTX transmission mode p of a short-term_l,k, and notify that this user enters DTX transmission mode p_l,k；After receiving this configuration information, this user enters this DTX transmission mode p_l,k；

In the l of community, base station calculates p respectively_l,kTransmission signal period p '_l,kNot pilot signal transmitted period p "_l,kThe spatial correlation matrix of middle reception signal, and determine use pilot frequency sequence a in neighbor cell and this community respectively_kThe feature matrix of dominant eigenvalue of mobile subscriber；

Whether the user k that the relation between feature matrix according to two dominant eigenvalues is determined in this community in user k and neighbor cell can use identical pilot frequency sequence a_k。

6. the selection of ascending pilot frequency and distribution method in tdd mode massiveMIMO system according to claim 5, it is characterised in that: according to whether the user k that the relation between the feature matrix of two dominant eigenvalues is determined in this community in user k and neighbor cell can use identical pilot frequency sequence a_kTime,

The transmission pilot frequency sequence a of user k in the l of community_kPeriod p '_l,k, serving BS is simultaneously received the pilot frequency sequence a that this user sends_kPilot frequency sequence a with adjacent cell user transmission_k, receive signal and be expressed as:

$y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1,}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p$

p_kTo send the correlation matrix during signal as follows:

${\hat{R}}_{a,k}=\mathrm{\Σ}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beAnd obtain the feature matrix corresponding to its dominant eigenvalue

In the l of community, user k's does not send pilot frequency sequence a_kPeriod p "_l,k, serving BS receives only the pilot frequency sequence a that its adjacent (L-1) community user sends_k；Reception signal is expressed as:

$y_l^p\left(t\right)=\sqrt{{\mathrm{\ρ}}_p}\underset{j=1,j\≠l}{\overset{L}{\mathrm{\Σ}}}{H}_{l,j}{a}_k+w_l^p$

The correlation matrix receiving signal is:

${\hat{R}}_{b,k}=\mathrm{\Σ}{y}_l^p\left(t\right){\left[y_l^p\left(t\right)\right]}^H$

Eigenvalues Decomposition beAnd obtain the feature matrix corresponding to its dominant eigenvalue

The base station of community l according toWithDimension relation, it is determined that in its this community, whether the user k in user k and adjacent (L-1) individual community can use identical pilot frequency sequence a_k。

7. the selection of ascending pilot frequency and distribution method in tdd mode massiveMIMO system according to claim 6, it is characterised in that: the base station of community l according toWithDimension relation, it is determined that in its this community, whether the user k in user k and adjacent (L-1) individual community can use identical pilot frequency sequence a_kTime,

IfDimension withDimension identical, community l uses pilot frequency sequence a_kUser can not with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k；

IfDimension more thanDimension, community l uses pilot frequency sequence a_kUser can with neighbor cell in use pilot frequency sequence a_kUser use identical uplink pilot sequence a_k。