CN101414874A - Method for judging down space division multiple access mode of TDD system - Google Patents

Abstract

The invention discloses a method for judging the mode of down space division multiple access of a time division duplex system, which includes the steps: the operation of channel estimation is carried out to a special physical channel for up adjoint to obtain the channel impulse response matrix of each user; the normalized spatial correlation matrix of each user is calculated; the eigenvalue of each spatial channel correlation matrix is analyzed so as to gather out and take the corresponding eigenvector of each eigenvalue of maximum as the forming weight of each user; the spatial correlation measure of every two users is calculated according to the forming weight of each user; and any user and the user with the spatial correlation measure less than a preset value are classified into the same set, and the actuating codes of the users in the same set are multiplexed by the mode of space division multiple access. The method can judge whether the users can adopt the SDMA mode to carry out the operation of data down-load or not according to the space division properties of the users, and the TDD HSDPA system can effectively utilize spatial information to increase the throughput of data transmission of the system and can increase the capacity of the system.

Description

A kind of method for judging down space division multiple access mode of tdd systems

Technical field

The present invention relates to a kind of TDD (Time Division Duplexing, time division duplex) SDMA of communication system (Spatial Division Mutiple Access, space division multiple access) method, in particular, whether the user that judges can enter the SDMA pattern and carry out empty method of dividing grouping on this basis the outdoor environment of a kind of third generation mobile field HDSPA (High-Speed Downlink Packet Access, high-speed downlink packet insert) system.

Background technology

In TD-SCDMA HSDPA system, because (wave beam forms array antenna Beamforming, also claim wave beam forming) technology, make have that certain angle isolates the user have empty possibility of dividing, with respect to non-SDMA system, conventional system is relatively low to the utilance of sign indicating number resource, and the data transmission rate of sub-district is also lower, for this reason, how to judge matter of utmost importance when can the user adopt the SDMA mode to carry out that the sky demal is multiplexing to have become the SDMA system design.

For down link,, condition restriction realizes estimating in real time for the reliable of channel because being difficult in down link.In the prior art, adopt DOA (Direction of Arrival, direction of arrival) isolates the method for carrying out the SDMA judgement and go for some scene, under the very little condition of up-downgoing channel spacing time slot, can be similar to and think that channel does not change, thereby can use the estimated value of the empty time domain parameter of channel that obtains by upstream data at down link, even can directly use the data of uplink beam figuration.Promptly according to upstream data direction of arrival is estimated, but owing to there is bigger DOA evaluated error, and the antenna element number limited the performance of algorithm, therefore systematic function is relatively poor when practical application.

When in the simple open environment of multi-path environment, when existing a tangible master directly, this moment, the DOA method was effective.But when multi-path environment was complicated, there was obvious defects in the DOA method, and precision is difficult to satisfy the needs of real system, will seriously cause the reduction of systematic function.

Therefore, there is defective in prior art, needs to improve.

Summary of the invention

The object of the present invention is to provide a kind of method for judging down space division multiple access mode of tdd systems, the error of avoiding the DOA method to cause improves the reliable estimated accuracy in real time of channel, thereby improves the performance of system.

Technical scheme of the present invention is as follows:

A kind of method for judging down space division multiple access mode of tdd systems, it comprises step: A1, to up following special physical channel, carry out channel estimating, obtain each user's channel impulse response matrix; A2, calculate each user's normalized spatial correlation matrix; A3, respectively each space channel correlation matrix is carried out characteristic value decomposition, select the shape-endowing weight value of the pairing characteristic vector of each eigenvalue of maximum as each user; A4, according to each user's shape-endowing weight value, calculate each user spatial correlation between any two; A5, to arbitrary user, form set with this user and with its spatial correlation less than the user of preset value, it is multiplexing to adopt the space division multiple access mode that the user of identity set is carried out sign indicating number.

Described determination methods, wherein, steps A 5 is further comprising the steps of: store the set that whole users form separately, it is multiplexing to adopt the space division multiple access mode that the user of identity set is carried out sign indicating number.

Described determination methods, wherein, in the steps A 5, it is multiplexing that described employing space division multiple access mode is carried out sign indicating number to the user of identity set, specifically may further comprise the steps: the user to identity set divides into groups according to pre-set criteria, and it is multiplexing to adopt the space division multiple access mode to divide the user to carry out sign indicating number to same group sky.

Described determination methods, wherein, describedly divide into groups, specifically may further comprise the steps: some users that selection scheduling priority is the highest according to the user of pre-set criteria to identity set, transmission priority or accessible transmission quantity according to these users divide into groups to the user.

Described determination methods wherein, is describedly divided into groups to the user, is to divide at same group with a certain user and with another user of the highest or accessible transmission quantity sum of its transmission priority sum maximum.

Described determination methods wherein, is gone back execution in step: be that the dispatched users Resources allocation sends data after steps A 5.

Described determination methods, wherein, steps A 2 specifically may further comprise the steps: calculate each user's channel impulse response matrix and amassing of its associate matrix, as each user's normalized spatial correlation matrix.

Described determination methods, wherein, steps A 2 is further comprising the steps of: adopt forgetting factor that each spatial correlation matrix is carried out smoothing processing.

Described determination methods, wherein, steps A 4 is further comprising the steps of: adopt each user spatial correlation between any two, form the spatial correlation matrix.

Described determination methods wherein, is also carried out following steps: obtain the span of described preset value by link simulation, carry out test optimization then in the outfield and obtain described preset value before the steps A 5.

Adopt such scheme, the present invention can compatible DOA partition method, can judge the sky branch characteristic between the user effectively, for empty demal multiplex system provides support, thereby judge whether the user can adopt the SDMA pattern to carry out the downlink transfer of data, like this, TDD HSDPA system can effectively utilize spatial information to improve the transfer of data throughput of system, and improves power system capacity.

Description of drawings

Fig. 1 is the empty divisional processing decision flow chart of the user of the inventive method.

Embodiment

Below preferred embodiment of the present invention is described in detail.

As shown in Figure 1, the invention provides a kind of method for judging down space division multiple access mode of tdd systems, technical scheme thinking of the present invention is: the space channel impulse response matrix that the channel estimating of the up DPCH of following is obtained each user by NodeB, go out each user's self space channel correlation matrix then by this matrix computations, at last by drawing optimum Beamforming weights someway, by the space correlation factor between the weights calculating different user, calculate optimum empty demal users multiplexing combination according to certain criterion at last.The inventive method comprises following concrete steps.

A1, to up following special physical channel, carry out channel estimating, obtain each user's channel impulse response matrix; Promptly obtain many array element channel estimate matrix (being the channel impulse response matrix) of each user, for example, to user i, the channel impulse response matrix of the capable N row of the M shown in obtaining below, wherein, array antenna has M array element, and the channel estimating tap number on each array element is N.

$H\left(i\right)={\left[\begin{array}{c}{h}_{i,\mathrm{1,1}},h_{i,\mathrm{1,2}},...h_{i,1,N}\\ h_{i,\mathrm{2,1}},h_{i,\mathrm{2,2}},...h_{i,2,N}\\ \·\·\·\\ h_{i,M,1},h_{i,M,2},...h_{i,M,N}\end{array}\right]}_{M*N}$

A2, calculate each user's normalized spatial correlation matrix.For example, steps A 2 specifically may further comprise the steps: calculate each user's channel impulse response matrix and amassing of its associate matrix, as each user's normalized spatial correlation matrix, promptly obtain normalized channel space correlation matrix.

For example, if R (i) is the normalized spatial correlation matrix of user i, can adopt following method to obtain, R (i)=channel_correlation (H (i)), wherein, channel_correlation is user's normalization spatial correlation matrix computing function, and correlation matrix normalization is exactly the coefficient correlation normalizing with the channel impulse response of each user between different array antennas, for example adopts formula: R _i=H (i) * H (i) ^H, wherein, H (i) ^HAssociate matrix for H (i).

Be more preferably, steps A 2 is further comprising the steps of: adopt forgetting factor that each spatial correlation matrix is carried out smoothing processing, be used to prevent that instantaneous channel shake from causing matrix to change violent and make system's instability; When estimating correlation matrix, introduce forgetting factor and can reduce the sensitiveness of algorithm greatly, promptly reduce sensitivity instantaneous correlation matrix to the cycle frequency error.

For example, normalization spatial correlation matrix R (i) is carried out smoothing processing, R (i)=p*R according to following formula _i+ (1-p) R _iWherein the current channel space correlation matrix of each user is designated as R _i, history value is designated as R _i, wherein, history value can be initialized as a certain matrix that presets: R _i=R ₀, such as R ₀Can be the unit diagonal matrix of M * M dimension; The forgetting factor of correlation matrix is designated as p, and the span of forgetting factor is decided according to actual conditions between 0 to 1; In general the high more forgetting factor of user's translational speed is big more, and translational speed is low more more little, can optimize forgetting factor by engineering test or emulation.

A3, respectively each space channel correlation matrix is carried out characteristic value decomposition, select the shape-endowing weight value of the pairing characteristic vector of each eigenvalue of maximum, also claim the figuration vector as each user.So far, calculate each user's wave beam forming vector.

Be more preferably, when having adopted forgetting factor that each spatial correlation matrix is carried out smoothing processing in the steps A 2, in the steps A 3 each the space channel correlation matrix after level and smooth carried out characteristic value decomposition.

Concerning user i, the correlation matrix after level and smooth is carried out characteristic value decomposition, select the shape-endowing weight value W of the characteristic vector of corresponding eigenvalue of maximum as user i _iFor example, the correlation matrix after utilizing smoothly calculates user's Beamforming shape-endowing weight value vector w _i: w _i=beamforming_gen (R _i), generally can decompose acquisition w with eigenvalue of maximum _i, just find eigenvalue of maximum characteristic of correspondence vector as shape-endowing weight value.For example, adopt power method to calculate eigenvalue of maximum, thereby select the characteristic of correspondence vector.

A4, according to each user's shape-endowing weight value, calculate each user spatial correlation between any two, i.e. the space correlation factor (being the space channel correlation factor); This step is the space channel correlation factor of calculating between the user, and the described space correlation factor is meant two users in the space channel sizes related of a period of time, and this is worth high more explanation, two users' sky divides isolation more little, be not suitable for carrying out SDMA more, otherwise, then be fit to carry out SDMA.And steps A 4 can also may further comprise the steps: adopt each user spatial correlation between any two, form the spatial correlation matrix, each element of this matrix characterizes the degree of correlation in space between two respective user.For example,

User_Spatial_Correlation_Matrix(i，j)＝correlation_factor _ij＝corr_calc(R _i，R _j)

Wherein, K _aExpression array antenna element number of array, correlation_factor _IjThe space correlation factor of expression user i and user j $\mathrm{correlation}\_{\mathrm{factor}}_{\mathrm{ij}}=\frac{1}{K_a}{W}_i*W_j^H,$ Corr_calc is a kind of computational methods, and this method is used to calculate user i, the spatial correlation between j is mutual; The input of corr_calc function is two users' a channel space correlation matrix, and it is used for calculating the space correlation degree between the user, computational methods $\mathrm{correlation}\_{\mathrm{factor}}_{\mathrm{ij}}=\frac{1}{K_a}{W}_i*W_j^H$ Be a kind of special circumstances, obviously, can also other computational methods obtain the space correlation degree between the user, the present invention does not make any additional limits to this.

A5, to arbitrary user, form set with this user and with its spatial correlation less than the user of preset value, it is multiplexing to adopt the space division multiple access mode that the user of identity set is carried out sign indicating number.

Be more preferably, can also carry out following steps before the steps A 5: obtain the span of described preset value by link simulation, carry out test optimization then in the outfield and obtain described preset value.

For example, steps A 5 is further comprising the steps of: store the set that whole users form separately, it is multiplexing to adopt the space division multiple access mode that the user of identity set is carried out sign indicating number.Wherein, it is multiplexing that described employing space division multiple access mode is carried out sign indicating number to the user of identity set, and specifically can may further comprise the steps: the user to identity set divides into groups according to pre-set criteria, and it is multiplexing to adopt the space division multiple access mode to divide the user to carry out sign indicating number to same group sky.

In the steps A 5, can also obtain user's transmission priority information; These precedence informations belong to the input parameter of realizing module, are generally provided by basic scheduler.Specifically, for example, describedly divide into groups, specifically can may further comprise the steps: some users that selection scheduling priority is the highest according to the user of pre-set criteria to identity set, transmission priority or accessible transmission quantity according to these users divide into groups to the user.Wherein, described the user being divided into groups, can be to divide at same group with a certain user and with another user of the highest or accessible transmission quantity sum of its transmission priority sum maximum.

Carry out the search of sky branch best packet, satisfying the sky branch grouping of finding out the best under certain criterion under the situation of SDMA, for example, select the highest user of transmission priority and be designated as i, find out then and the user set of user i spatial correlation less than predetermined threshold value, this threshold value is generally by emulation or actual measurement calibration, it be one less than 1 number, such as 0.1, be designated as candidate_user_set, correlation_factor _Ij＜Γ, j ∈ candidate_user_set.

Then,, in candidate_user_set, find out optimum user and user and carry out the grouping of sky branch, be such as but not limited to two following criterions according to certain optiaml ciriterion:

Criterion a: reach maximum throughput of transmissions, promptly

$j=\mathrm{MAX}(\underset{\underset{j\∈\mathrm{candidate}\_\mathrm{user}\_\mathrm{set}}{j\≠i}}{\arg }\mathrm{Transmisstion}\_\mathrm{Capacity}(i,j))$

In candidate_user_set, find out a user j and satisfy two conditions, 1) j is not equal to i, 2) transmission capacity maximum when j and i transmit together.(i, j) i is calculated in expression to Transmisstion_Capacity in the following formula, the empty a certain method of dividing capacity when transmitting of j.Tansmission_capacity is a kind of computing function, can calculate the transmission capacity when i and j are empty to divide transmission, in general is the transmission block size sum when calculating both and transmitting together, the non-emphasis of the present invention of specific algorithm.

Criterion b: empty branch user grouping has maximum transmission priority, promptly

$j=\mathrm{MAX}(\underset{\underset{j\∈\mathrm{candidate}\_\mathrm{user}\_\mathrm{set}}{j\≠i}}{\arg }\mathrm{transmisstion}\_\mathrm{priority}\left(j\right))$

The implication of this formula is: find out a user j in candidate_user_set, it satisfies a condition, be that transmission priority is the highest, the function of the dispatching priority of user j is found out in transmisstion_priority (j) expression, finds the transmission priority of j correspondence by the user priority information from basic scheduler.

Distribute rational resource to carry out downlink data transmission for the dispatched users of selecting then.Need to prove, the invention is not restricted to a, these two kinds of criterions of b, on the basis of adopting said method, other any one criterions can be included in the method frame of the present invention.

After steps A 5, can also comprise follow-up processing, for example, carry out following steps:, can distribute rational resource to carry out data for dispatched users like this and send for the dispatched users Resources allocation sends data or the like.

Below specify method of the present invention more for example.

Suppose 1: the up DPCH of following is continuous transmission, and just each TTI (time interval) all has the up DPCH of following signal to receive and corresponding channel estimating;

Suppose 2: suppose to have N user, and only selecting one group of grouping that comprises two users, to carry out the sky demal multiplexing;

Suppose 3: suppose that array antenna has 8 array elements, the channel estimating tap number on each array element is 16, and then each user's channel estimate matrix is $H\left(i\right)={\left[\begin{array}{c}{h}_{i,\mathrm{1,1}},h_{i,\mathrm{1,2}},...h_{i,1,16}\\ h_{i,\mathrm{2,1}},h_{i,\mathrm{2,2}},...h_{i,2,16}\\ \·\·\·\\ h_{i,8,1},h_{i,8,2},...h_{i,8,16}\end{array}\right]}_{8*16};$

Suppose 4: the current channel space correlation matrix of each user is designated as R _i, history value is R _i, the forgetting factor of correlation matrix is designated as p;

Suppose 5: each user's shape-endowing weight value is w _i

Suppose 6: user i, the space correlation coefficient between the j is designated as correlation_factor _Ij

Step 1: the channel impulse response matrix H (i) that obtains each user;

Step 2: calculate each user's channel space correlation matrix, R _i=H (i) * H (i) ^H, the correlation matrix after then level and smooth is R (i)=p*R _i+ (1-p) R _i

Step 3: the correlation matrix after adopting smoothly calculates each user's shape-endowing weight value:

Correlation matrix is carried out characteristic value decomposition, and promptly [V, D]=SVD (R (i)) selects the shape-endowing weight value w of the characteristic vector of corresponding eigenvalue of maximum as user i _i

Step 4: calculate user's space correlation factor between any two, $\mathrm{correlation}\_{\mathrm{factor}}_{\mathrm{ij}}=\frac{1}{8}{W}_i*W_j^H,$ Illustrate: Dui Ying corr_calc function is to select user i here, and the eigenvalue of maximum of j carries out relevant treatment then;

Step 5: obtain MAC-hs (Media Access Control-high speed, the high speed media access control) each user's dispatching priority, and hypothesis user's 1 transmission priority is the highest, the user who then selects energy and user's 1 empty branch transmission gathers user_candidate_set, this gathers any one user j, correlation_factor _Ij＜Γ; Wherein, Γ is empty thresholding, the promptly described preset value of dividing.

Can select a user in the empty user of the branch set, make this user satisfy a or one of them condition of b:

In user_candidate_set, select the highest user of transmission priority and user 1 and carry out sky and divide transmission, be assumed to be user 3, then empty divide transmission be grouped into 1,3};

The user who selects throughput maximum when transmitting simultaneously with user 1 in user_candidate_set carries out sky and divides transmission, is assumed to be user 6, then empty divide transmission be grouped into 1,6}.

The present invention proposes the SDMA decision method of the TDD system under a kind of comparatively general outdoor environment, can effectively judge sky branch characteristic between the user by the inventive method, can judge effectively whether the user can adopt the SDMA pattern to carry out the downlink transfer of data, thereby provide support for empty demal multiplex system, the present invention has simultaneously provided the basic norm of empty minute user grouping on this basis, satisfied the primary demand of space division system design, by adopting the inventive method, TDD HSDPA system can effectively utilize spatial information to improve the transfer of data throughput of system, improves power system capacity.

Should be understood that, for those of ordinary skills, can be improved according to the above description or conversion, and all these improvement and conversion all should belong to the protection range of claims of the present invention.

Claims (10)

1, a kind of method for judging down space division multiple access mode of tdd systems, it comprises step:

A1, to up following special physical channel, carry out channel estimating, obtain each user's channel impulse response matrix;

A2, calculate each user's normalized spatial correlation matrix;

A3, respectively each space channel correlation matrix is carried out characteristic value decomposition, select the shape-endowing weight value of the pairing characteristic vector of each eigenvalue of maximum as each user;

A4, according to each user's shape-endowing weight value, calculate each user spatial correlation between any two;

A5, to arbitrary user, form set with this user and with its spatial correlation less than the user of preset value, it is multiplexing to adopt the space division multiple access mode that the user of identity set is carried out sign indicating number.

2, determination methods according to claim 1 is characterized in that, steps A 5 is further comprising the steps of: store the set that whole users form separately, it is multiplexing to adopt the space division multiple access mode that the user of identity set is carried out sign indicating number.

3, determination methods according to claim 1, it is characterized in that, in the steps A 5, it is multiplexing that described employing space division multiple access mode is carried out sign indicating number to the user of identity set, specifically may further comprise the steps: the user to identity set divides into groups according to pre-set criteria, and it is multiplexing to adopt the space division multiple access mode to divide the user to carry out sign indicating number to same group sky.

4, determination methods according to claim 3, it is characterized in that, describedly divide into groups according to the user of pre-set criteria to identity set, specifically may further comprise the steps: some users that selection scheduling priority is the highest, transmission priority or accessible transmission quantity according to these users divide into groups to the user.

5, determination methods according to claim 4 is characterized in that, described the user is divided into groups, and is to divide at same group with a certain user and with another user of the highest or accessible transmission quantity sum of its transmission priority sum maximum.

6, determination methods according to claim 1 is characterized in that, goes back execution in step after steps A 5: be that the dispatched users Resources allocation sends data.

7, determination methods according to claim 1 is characterized in that, steps A 2 specifically may further comprise the steps: calculate each user's channel impulse response matrix and amassing of its associate matrix, as each user's normalized spatial correlation matrix.

8, determination methods according to claim 1 is characterized in that, steps A 2 is further comprising the steps of: adopt forgetting factor that each spatial correlation matrix is carried out smoothing processing.

9, determination methods according to claim 1 is characterized in that, steps A 4 is further comprising the steps of: adopt each user spatial correlation between any two, form the spatial correlation matrix.

10, determination methods according to claim 1 is characterized in that, also carries out following steps before the steps A 5: obtain the span of described preset value by link simulation, carry out test optimization then in the outfield and obtain described preset value.

Images