CN104734754A

CN104734754A - Beamforming weight training method and base station and terminal

Info

Publication number: CN104734754A
Application number: CN201310712329.6A
Authority: CN
Inventors: 赵晶; 陈艺戬; 鲁照华; 郁光辉; 肖华华; 王瑜新
Original assignee: ZTE Corp
Current assignee: ZTE Corp
Priority date: 2013-12-20
Filing date: 2013-12-20
Publication date: 2015-06-24
Also published as: WO2015089894A1

Abstract

The invention discloses a beamforming weight training method and base station and terminal, and relates to the field of communications. The beamforming weight training method comprises the steps that the base station sends N types of beam pilot signals to the terminal, the terminal selects one type or multiple types of received beam pilot frequencies in the beam pilot signals, and gives feedback on beam information corresponding to the selected beam pilot frequencies to the base station, and the base station determines an initial training signal vector according to the beam information fed back by the terminal, and sends a channel vector to the terminal to conduct beam weight training, wherein N is a positive integer. The invention further discloses other beamforming weight training methods and base stations and terminals. According to the technical scheme, the beamforming weight training method and base station and terminal improve the efficiency of the beamforming weight training.

Description

A kind of beam shape-endowing weight value training method and base station, terminal

Technical field

The present invention relates to the communications field, particularly relate to a kind of beam shape-endowing weight value training method and base station, terminal.

Background technology

In wireless communication system, transmitting terminal and receiving terminal take the mode of spatial reuse to use many antennas to obtain higher speed.Relative to general method for spacial multiplex, a kind of technology be widely used be receiving terminal feedback channel information to transmitting terminal, transmitting terminal according to obtain channel information use some launch precoding techniques, improve transmission performance greatly.For in single user multiple-input and multiple-output (Multi-inputMulti-output, MIMO), channel characteristics Vector Message is directly used to carry out precoding; For in multiuser MIMO, need channel information more accurately.

In some technology of 4G as in LTE, 802.16m standard criterion, the feedback of channel information mainly utilizes the feedback method of better simply single code book, and the performance of the transmitting precoding technique of MIMO more depends on the accuracy of wherein code book feedback.Here the general principle channel information based on code book being quantized feedback is briefly described below:

Suppose that Limited Feedback channel capacity is Bbps/Hz, so can the number of code word be N=2 ^bindividual.The space of feature vectors of channel matrix is through quantizing to form codebook space this code book (sending and receiving end is identical) is preserved or produced in real time to transmitting terminal and receiving terminal jointly.According to receiving terminal obtain channel matrix H, receiving terminal according to certain criterion from middle selection code word of mating most with channel and code word sequence number i is fed back to transmitting terminal.Here, code word sequence is known as PMI(Precoding Matrix Indicator, precoding matrix indicators).Transmitting terminal finds corresponding pre-encoding codeword according to this sequence number i thus acquisition channel information, illustrate the characteristic vector information of channel.

Along with the high speed development of wireless communication technology, user radio application is more and more abundanter, has driven wireless data service to increase rapidly.It is predicted, between coming 10 years, data service increases with annual 1.6-2 times speed.This brings huge challenge to Radio Access Network.Multi-antenna technology is the key technology of reply wireless data service explosive growth challenge, and the multi-antenna technology supported in current 4G only supports the horizontal dimensions beamforming technique of maximum 8 ports, also has the further significantly capacity of larger potentiality.

Massive MIMO technology is the crucial enhancing technology of in next generation communication technology, Massive mimo system principal character is: base station side is configured with large-scale antenna array, such as 100 antennas, even more, when transfer of data, utilize MU-MIMO technology, simultaneously with the multiplexing multiple user of frequency, in general, number of antennas and users multiplexing number scale dimension are held in 5-10 doubly left and right.Can prove, it is no matter the strong correlation channel at view distance environment, or the non-correlation channel under rich scattering, coefficient correlation between the channel of any two users decays along with the increase exponentially form of number of antennas, such as when base station side is configured with 100 antennas, between the channel of any two users, coefficient correlation levels off to 0, in other words between multi-user's respective channels close to orthogonal.On the other hand, large array can bring considerable array gain and diversity gain.

For Massive MIMO, due to the introducing of a large amount of antenna, traditional method: every root antenna transmitting channel measurement pilot frequency CSI-RS, terminal detects CSI-RS and obtains channel matrix corresponding to each transfer resource by channel estimating, the best transmission number of plies information in each frequency domain sub-band precoding vectors and broadband in best base band is obtained according to channel matrix, then feed back based on the code book feedback technique introduced above, when this mode is applied in massive MIMO, there is larger problem.Be mainly reflected in, pilot-frequency expense can increase with Nt and increases, antenna number pilot-frequency expense of many times is very huge, in addition, need to comprise very many code words in code book due to use during feedback, the selection of code word is very difficult, and the complexity causing terminal a large amount of increases, almost cannot realize, or need to pay huge cost price.The expense of code book feedback is also very large, makes uplink overhead huge.Therefore in general, for massiveMIMO, better mode adopts beam shape-endowing weight value training method, effectively can reduce pilot tone and feedback overhead and make system obtain good precoding performance.

As shown in Figure 1, define and send training pilot tone by base station to terminal, the data received send to the process of base station to be called an iterative process to the general principle of wave beam training technique by terminal after treatment.After successive ignition, finishing iteration has been called a training process.

Lower mask body introduces wave beam training principle:

Step 1: base station sends training pilot tone to terminal;

Step 2: terminal receives training pilot tone and carries out receiving area reason;

Step 3: terminal is to the training pilot tone after the transmission processing of base station;

Step 4: the training pilot tone that base station receiving terminal sends is gone forward side by side row relax, it can be used as new training pilot tone after process;

Repeat step 1-4 until trained.

Formula below illustrates this wave beam and trains the beneficial effect that can bring:

H = σ_{1} u_{1} v_{1}^{H} + σ_{2} u_{2} v_{2}^{H} + . . . σ_{p} u_{p} v_{p}^{H}

Here σ ₁σ _pbe singular value, u1-up is left characteristic vector, and v1-vp is right characteristic vector.

Definition

Can find:

\lim_{m &RightArrow; \infty} H^{2 m} = δ_{1}^{2 m} v_{1} v_{1}^{H}

(projection matrix form)

\lim_{m &RightArrow; \infty} H^{2 m + 1} = δ_{1}^{2 m + 1} u_{1} v_{1}^{H}

Generally think and work as

δ_{1}^{2 m} > 100^{δ_{2}^{2 m}}

Can think

H^{2 m} \approx δ_{1}^{2 m} v_{1} v_{1}^{H}, H^{2 m + 1} \approx δ_{1}^{2 m + 1} u_{1} v_{1}^{H},

Above-mentioned effect can be reached by iteration for several times.After successive ignition, precoding vectors can well match with channel eigenvectors, thus makes terminal obtain precoding more accurately, improves the signal to noise ratio of system.

Although there is beam shape-endowing weight value training method in prior art, imperfect training mechanism, can cause training not reach and require or training waste resource, when the method is employed, can also carry out perfect.

Wave beam forming training method described in document is affected by noise larger, if when each training starts, and initial t ⁰for the vector of any Ntx1, because p (p=Nt) individual characteristic vector is one group of base of Nt dimension space, therefore necessarily can by t ⁰be expressed as consider conclusion above

H^{2 m + 1} \approx δ_{1}^{2 m + 1} u_{1} v_{1}^{H},

So have:

H^{2 m} t^{0} = σ_{1}^{2 m} c_{1} v_{1}

H^{2 m + 1} t^{0} = σ_{1}^{2 m + 1} c_{1} u_{1}

It is normalized and can obtains characteristic vector, and this characteristic vector still can through different iteration after obtaining.Here Problems existing mainly t ⁰selection time may because of running into the very little situation of c1, now be subject to noise effect can be larger.If noise is comparatively large, the training signal sent can be submerged in noise, even if carried out repeatedly training later, also can not obtain the gain of anticipation, this is that the program needs one of perfect place.

If the direction of channel characteristics vector just falls into direction zero of training wave beam, as shown in Figure 2, R is channel characteristics direction vector, and wave beam 1 is training wave beam, then no matter train how many times, wave beam 1 all can not obtain projection on characteristic vector, and training all can not be successful.Therefore the selection of initial beam, for most important training, and not good solution in prior art.

In sum, suitable iterations is selected for most important the program, if iterations is selected improper, such as iterations is too much, not only can cause the waste of resource, and the training time can be made elongated, the training time is long, and to make training terminate channel information when rear channel information starts with training no longer identical, and the iterative information now fed back will lose efficacy.And the training time too shortly may cause current signal quality to be not enough to transfer of data, or better performance can be obtained after more times training, improve the transmittability of whole system.Therefore select suitable iterations most important when training.Do not have good solution in prior art, therefore may cause the irrational situation of iterations.

Summary of the invention

Technical problem to be solved by this invention is, provides a kind of beam shape-endowing weight value training method and base station, terminal, to improve beam shape-endowing weight value training effectiveness.

In order to solve the problems of the technologies described above, the invention discloses a kind of beam shape-endowing weight value training method, comprise: base station sends N class wave beam pilot signal to terminal, described terminal selects a class in the wave beam pilot signal received or multiclass wave beam pilot tone, and beam information corresponding for selected wave beam pilot tone is fed back to described base station, described base station is according to the beam information of described terminal feedback, determine initial training signal vector, and send described channel vector to described terminal and carry out wave beam weight training, wherein, N is positive integer.

Alternatively, in said method, described base station is the configurating terminal beam index number information that need report also.3, method as claimed in claim 2, is characterized in that, described in the beam index number information that reports at least comprise concrete index number and index number grade.

Alternatively, in said method, described terminal selects a class in the wave beam pilot signal received or multiclass wave beam pilot tone, and beam information corresponding for selected wave beam pilot tone is fed back to described base station refers to:

The class wave beam pilot tone that described terminal selects signal quality the strongest from N class wave beam pilot signal, and beam information corresponding for such wave beam pilot tone is fed back to described base station;

Wherein, described signal quality at least comprises signal to noise ratio, the dry ratio of letter and carries dry ratio.

The invention also discloses a kind of beam shape-endowing weight value training method, comprising:

Terminal is to base station pilot signal transmitted, and the pilot signal that described base station selected receiving terminal sends, calculates the channel condition information of up channel, and according to channel reciprocity, select wave beam and send to terminal, train according to described pilot signal.

Alternatively, in said method, pilot frequency locations corresponding for beam index is informed to described terminal by described base station, and described terminal sends wave beam pilot signal to described base station in the pilot frequency locations of agreement.

Alternatively, in said method, described terminal sends sounding pilot signal to described base station, and described base station calculates the channel condition information of up channel according to described pilot signal, and select a wave beam according to described channel condition information, utilize this wave beam to train.

Base station sends training pilot signal to terminal, described terminal is selected to receive described training pilot signal, and reception process is carried out to selected training pilot signal, the signal quality of selected training pilot signal and thresholding Threshold are contrasted, and to described base station feedback comparing result indication information.

Whether alternatively, said method also comprises: after described base station receives described comparing result indication information, determine to described terminal transmission data information according to this indication information;

Wherein, described signal quality at least comprises signal to noise ratio, Signal to Interference plus Noise Ratio and carries dry ratio of making an uproar.

Alternatively, said method also comprises: described base station sends signal quality thresholding Threshold configuration information to described terminal.

Alternatively, in said method, described signal quality thresholding Threshold configuration information comprises signal quality threshold value, or comprises signal quality threshold level index;

When described signal quality thresholding Threshold configuration information comprises signal quality threshold level index, described terminal is configured according to the training signal quality threshold of signal quality threshold level index to self received.

Alternatively, in said method, described base station is that each terminal configures lowest signal-to-noise thresholding Threshold information respectively;

Or the channel condition information of terminal to report and signal quality thresholding Threshold contrast by described base station, and Record Comparison result.

Base station sends training pilot signal to terminal, described terminal is selected to receive training pilot signal, and reception process is carried out to selected training pilot signal, described terminal obtains iterations information n, and n and preset times N is contrasted, comparing result indication information is sent to described base station by described terminal, and described base station judges whether to proceed iteration according to described comparing result indication information, wherein, N is positive integer.

Alternatively, in said method, according to described indication information, described base station judges whether that the process proceeding iteration comprises:

When the iterations of described comparing result indication information instruction base station and terminal reaches N, automatic termination of iterations.

Alternatively, said method also comprises: described base station sends the configuration information of N to described terminal.

Alternatively, in said method, the configuration information of described N comprises N hierarchy index or concrete N numerical value;

When the configuration information of described N comprises N hierarchy index, described terminal is configured according to the maximum iteration time of the configuration information of the N received to self.

Alternatively, in said method, described base station is each terminal configured number N information respectively.

Alternatively, said method also comprises:

Described terminal is to described base station feedback iterations N information; Or,

Described terminal strengthens to described base station feedback or reduces N number; Or,

Described terminal promotes to described base station feedback or reduces N class information.

The invention also discloses a kind of beam shape-endowing weight value training method, the method comprises:

Described base station sends to described terminal and stops training Rule of judgment information, and described terminal judges whether to stop training according to described terminations training Rule of judgment, and described termination is trained Rule of judgment to comprise to meet λ _ithresholding σ configuration information, or meet signal quality thresholding R configuration information.

Alternatively, said method also comprises:

Described λ _ithresholding σ configuration information comprises thresholding σ numerical value, or thresholding σ hierarchy index;

As described λ _iwhen thresholding σ configuration information comprises thresholding σ hierarchy index, described terminal is configured self thresholding σ according to receiving thresholding σ hierarchy index information;

Or described signal quality thresholding R configuration information comprises concrete thresholding R numerical value, or thresholding R hierarchy index information;

When described signal quality thresholding R configuration information comprises thresholding R hierarchy index information, described terminal is configured self thresholding after receiving described thresholding R hierarchy index.

Alternatively, in said method, described base station is that each terminal configures described termination training Rule of judgment information respectively.

The channel condition information that base station receiving terminal reports, and whether stop according to described channel condition information training of judgement.

Alternatively, in said method, whether described base station stops referring to according to the channel condition information training of judgement of terminal to report:

Described base station calculates signal quality and λ _i, judge whether iteration continues according to thresholding σ.

Alternatively, in said method, after described base station training of judgement stops, also training is stopped indication information and send to described terminal.

The invention also discloses a kind of base station, comprising: first module, send N class wave beam pilot signal to terminal, N is positive integer;

Second unit, according to the beam information of terminal feedback, sends training information to described terminal.

The invention also discloses a kind of terminal, comprising:

First module, selects to receive the class in the wave beam pilot signal of base station transmission or multiclass wave beam pilot tone, and beam information corresponding for selected wave beam pilot tone is fed back to described base station;

Second unit, receives the training information that described base station sends.

Technical scheme improves beam shape-endowing weight value training effectiveness.

Accompanying drawing explanation

Fig. 1 is the basic principle schematic of wave beam training technique;

Fig. 2 is that wave beam zero falls into schematic diagram;

Fig. 3 is the wave beam schematic diagram in example 1-5 of the present invention;

Fig. 4 is the wave beam schematic diagram in example 1-6 of the present invention.

Embodiment

For making the object, technical solutions and advantages of the present invention clearly understand, hereafter will be described in further detail technical solution of the present invention by reference to the accompanying drawings.It should be noted that, when not conflicting, the feature in the embodiment of the application and embodiment can combine arbitrarily mutually.

Embodiment 1

The present embodiment provides a kind of beam shape-endowing weight value training method, comprise following operation: base station sends N class wave beam pilot signal to terminal, N is positive integer, and UE selects a class in the wave beam pilot signal received or multiclass wave beam pilot tone, and the beam information of its correspondence is fed back to base station.The beam information that base station is fed back according to UE, sends training information to terminal.

Particularly, the class received or beam. position information corresponding to multiclass wave beam pilot signal or index are fed back to base station by terminal, and according to the beam information that UE feeds back, base station determines that training sends signal, and send described training signal to terminal.

And base station sends different beam information to terminal on sub-carriers, the subcarrier information of the class received or multiclass wave beam or index are fed back to base station by terminal, and the beam information that base station is fed back according to UE sends training information to terminal.

Wherein, base station can the configurating terminal beam index number information that need report;

The beam index number information reported includes but not limited to concrete index number, index number grade etc.

When base station sends N class wave beam to terminal, a class strongest signal quality beam information can be fed back to base station by terminal, and base station sends training information according to the beam information of terminal feedback to terminal.

Signal quality in the present embodiment includes but not limited to the dry ratio of signal to noise ratio, letter or carries dry ratio.

In addition, also can by terminal to base station pilot signal transmitted, the pilot signal that base station selected receiving terminal sends, and the channel condition information of up channel is calculated according to pilot signal, according to channel heterogeneite, base station is selected wave beam and is sent to terminal, trains.

Above-mentioned pilot signal includes but not limited to wave beam pilot signal.

Particularly, terminal can send wave beam pilot signal to base station in the pilot frequency locations of agreement.

Pilot frequency locations corresponding for beam index is informed to terminal by base station.

Such as, terminal sends souding pilot signal to base station, and base station calculates the channel condition information of up channel according to pilot signal, and selects a wave beam according to described channel condition information, utilizes this wave beam to carry out training.

Below in conjunction with embody rule, the specific implementation process of said method is described.

Be defined in same once training, adjacent twice iteration such as signal quality ratio between i-th time and the i-th-1 time iteration is λ _i, i be greater than 1 positive integer.Described signal quality includes but not limited to signal to noise ratio, believes dry ripple, carries dry ratio, and represents with R.

λ _i=R _i/R _i-1

Example 1-1:

Base station and terminal complete training as follows before beam selection.

(1) base station is in the running time-frequency resource of M agreement, and M is positive integer, and select a precoding vector to send signal to terminal successively, each symbol and a precoding vector index are bound.

(2) terminal selects to receive in the running time-frequency resource of M agreement the signal that base station sends, and calculates the channel quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive year dry make an uproar than etc.Wherein, the channel quality information that i-th time slot is corresponding is CQI _i, i=1 ..., M.

(3) terminal feedback m CQI _icorresponding beam index feeds back to base station, and m is the positive integer being less than or equal to M.

(4) index that sends of base station receiving terminal, and the wave beam therefrom selecting finally to be used for send data by the following method.

If the beam index set received all is S2, and the wave beam being finally used for sending out data is S3, and S3 is initialized as empty set.(4.1) complete finally determining of wave beam to step (4.4) as follows.

(4.1) in S2, select a beam index, and it is deleted from S2, be incorporated to S set 3.

(4.2) select a wave beam in S2, and it deleted from S2, any one wave beam in this wave beam and S3 is not in same group.Here wave beam grouping, refers to what base station and terminal had been divided in advance, and such as, often in group wave beam, be all the close wave beam of physically transmission direction, such as, first group is that { M, 1,2}, i-th group is that { i-1, i, i+1}, wherein, the numeral in bracket is beam index.

(4.3) repeated execution of steps (4.2) is until S set 2 is empty.

(4.4) determine that element number that S3 gathers is the wave beam number of final transmission data, the wave beam that the index in S3 is corresponding is the wave beam being finally used for sending data.

(5) base station is continuous or discontinuous at N _beamthe running time-frequency resource of individual agreement sends training pilot signal to terminal, N _beamfor positive integer; Base station send running time-frequency resource can for the cycle send or aperiodic transmission;

(6) terminal is selected at N _beamreceive the training pilot signal that base station sends in individual running time-frequency resource, and calculate the channel quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, the channel quality information that i-th OFDM symbol is corresponding is CQI _i, i=1 ..., N _beam;

(7) terminal obtains the frequency n when time iteration;

(8) terminal according to a preconcerted arrangement criterion judge whether iteration terminates; Agreement criterion can for limiting maximum times criterion or other criterions; If training continues after terminal judges this iteration, then continue iteration to base station feedback;

If training terminates after terminal judges this iteration, then need to judge that whether this training be successful, concrete mode is:

The signal quality of last iteration and signal quality thresholding Threshold contrast by terminal, when signal quality is greater than or equal to Threshold, train successfully, if when signal quality is less than Threshold, and failure to train;

Or, when signal quality is greater than Threshold, train successfully, if when signal quality is equal to or less than Threshold, failure to train;

Terminal sends to base station by after training signal normalization;

Terminal is by training end mark or train whether Success Flag feeds back to base station;

(11) the whether continuation iteration flag bit of base station selected receiving terminal feedback, determines whether proceed iteration by flag bit feedback information; If proceed iteration, then will receive the normalization of training pilot frequency information and obtain Y ₂, Y ₂for new training pilot tone;

Completed the selection course of wave beam by step (1)-(4) base station and terminal, base station sends data by the wave beam finally determined to this terminal.This process can be carried out at certain cycle, or is undertaken by after terminal where necessary trigger base station, also can be to carry out after base station is triggered according to current channel quality information.

Example 1-2:

Base station and terminal also can complete training as follows before beam selection.

(1) base station is in M subcarrier, and M is positive integer, and select a precoding vector to send signal to terminal successively, each symbol and a precoding vector index are bound.

(2) terminal selects to receive in M subcarrier the signal that base station sends, and calculates the channel quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive year dry make an uproar than etc.Wherein, the channel quality information that i-th time slot is corresponding is CQI _i, i=1 ..., M.

A described M subcarrier also can be M ODFM symbol, or pilot frequency locations etc.;

(4) index that sends of base station receiving terminal, and the wave beam that is finally used for sending data is therefrom selected by the method similar with embodiment 1-1.

Completed the selection course of wave beam by step (1)-(4) base station and terminal, base station sends data by the wave beam finally determined to this terminal.This process can be carried out at certain cycle.Or being undertaken by after terminal where necessary trigger base station, also can be carry out after base station is triggered according to current channel quality information.

Example 1-3:

In the above embodiments, terminal according to a preconcerted arrangement, feeds back to base station to base station feedback m corresponding beam index, and m is 1 or multiple.Base station and terminal complete beam selection by following steps:

(1) base station notifies the beam index number m information that user needs to report on down control channel, and number m information includes but not limited to specifically need report beam index number, or appoint report number m grade;

(2) terminal is configured the index number that reports of self after receiving base station notice;

(3) base station is in M appointed position, and M is positive integer, and select a precoding vector to send signal to terminal successively, each symbol and a precoding vector index are bound;

(4) terminal selects to receive in appointed position the signal that base station sends, and calculates the channel quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc.Wherein, the channel quality information that i-th time slot is corresponding is CQI _i, i=1 ..., M;

(5) terminal feeds back m CQI according to a preconcerted arrangement _icorresponding beam index feeds back to base station;

When m is 1, terminal can travel through in M the appointed position received, and selects the best index of Received signal strength channel quality information to feed back, and channel quality information includes but not limited to received power, received signal to noise ratio, reception Signal to Interference plus Noise Ratio, receives and carry dry ratio of making an uproar.

When m be greater than 1 several time, terminal is selected to travel through on the running time-frequency resource of M the agreement received, and the K selecting received signal quality information best is individual, and K is the positive integer being more than or equal to m, and by the principle appointed, from K, pick out m beam index report base station;

The described election process appointed can be:

(5.1) terminal picks the best K of a received signal quality information beam index;

(5.2) terminal finds the larger m of a mutual chordal distance wave beam in K wave beam, and m beam index information is sent to base station;

(6) index of base station receiving terminal transmission, therefrom selects the wave beam being finally used for sending data.

Example 1-4:

Base station and terminal complete training as follows before beam selection.

(1) base station is in the running time-frequency resource of M agreement, and M is positive integer, such as M=8, and select a precoding vector to send signal to terminal successively, each symbol and a precoding vector index are bound.

Such as 8 precoding vectors are any row of matrix below, and its code book numbering is followed successively by 0-7:

(2) terminal selects to receive in the running time-frequency resources of 8 agreements the signal that base station sends, and calculates the channel quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc.Wherein, the channel quality information that i-th time slot is corresponding is CQI _i, i=1 ..., M.

(3) terminal feedback m CQI _icorresponding beam index feeds back to base station, and m is the positive integer being less than or equal to M.Such as m is 2, and beam index numbering 1 and 3 is fed back to base station by terminal.

(4) index 1 and 3 of base station receiving terminal transmission, and determine to adopt wave beam 3 to do wave beam training.

(5) base station is continuous or discontinuous at N _beamthe running time-frequency resource of individual agreement sends training pilot signal wave beam 3, N to terminal _beamfor positive integer; Base station send running time-frequency resource can for the cycle send or aperiodic transmission;

(7) terminal obtains the frequency n when time iteration;

Terminal sends to base station by after training signal normalization;

(10) terminal is by training end mark or train whether Success Flag feeds back to base station;

Example 1-5

Such as 8 precoding vectors be shown in Fig. 3 in any one wave beam, its code book numbering is followed successively by 0-7:

(3) terminal feedback m CQI _icorresponding beam index feeds back to base station, and m is the positive integer being less than or equal to M.Such as m is 1, and it is dry than maximum wave beam 3 that terminal selects letter, and index number 3 is fed back to base station.

(4) index 3 of base station receiving terminal transmission, and determine to adopt wave beam 3 to do wave beam training.

(7) terminal obtains the frequency n when time iteration;

(9) terminal sends to base station by after training signal normalization;

Example 1-6:

Base station and terminal can also complete training as follows before beam selection.

(1) terminal is in the running time-frequency resource of M agreement, and M is positive integer, such as M=16, and select a precoding vector to send signal to base station successively, each symbol and a precoding vector index are bound.

Such as 16 precoding vectors are any one wave beam in Fig. 4, and its code book numbering is followed successively by 0-15:

(2) the base station selected signal receiving base station and send in the running time-frequency resources of 16 agreements, and calculate the signal quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc.Wherein, the channel quality information that i-th time slot is corresponding is CQI _i, i=1 ..., M.

(3) base station selected wave beam 5 carries out wave beam training.

(4) base station is continuous or discontinuous at N _beamthe running time-frequency resource of individual agreement sends training pilot signal wave beam 5, N to terminal _beamfor positive integer; Base station send running time-frequency resource can for the cycle send or aperiodic transmission;

(5) terminal is selected at N _beamreceive the training pilot signal that base station sends in individual running time-frequency resource, and calculate the channel quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, the channel quality information that i-th OFDM symbol is corresponding is CQI _i, i=1 ..., N _beam;

(6) terminal obtains the frequency n when time iteration;

(7) terminal according to a preconcerted arrangement criterion judge whether iteration terminates; Agreement criterion can for limiting maximum times criterion or other criterions; If training continues after terminal judges this iteration, then continue iteration to base station feedback;

Or, when signal quality is greater than Threshold, train successfully, if signal quality is equal to or less than SNR ₀time, failure to train;

(8) terminal sends to base station by after training signal normalization;

(9) terminal is by training end mark or train whether Success Flag feeds back to base station;

(10) the whether continuation iteration flag bit of base station selected receiving terminal feedback, determines whether proceed iteration by flag bit feedback information; If proceed iteration, then will receive the normalization of training pilot frequency information and obtain Y ₂, Y ₂for new training pilot tone;

Example 1-7:

(1) terminal is in the running time-frequency resource of agreement, sends Souding pilot signal to base station.

(2) the base station selected pilot signal receiving base station and send in the running time-frequency resource of agreement, and the channel condition information of up channel is calculated according to described pilot signal.

(3) base station is according to described channel condition information and channel heterogeneite, selects to do training wave beam with the wave beam that up channel mates most from M wave beam.

(4) base station is continuous or discontinuous at N _beamthe running time-frequency resource of individual agreement sends training pilot signal wave beam to terminal, N _beamfor positive integer; Base station send running time-frequency resource can for the cycle send or aperiodic transmission;

(6) terminal obtains the frequency n when time iteration;

(8) terminal sends to base station by after training signal normalization;

Embodiment 2

The present embodiment provides a kind of beam shape-endowing weight value training method, comprising:

Base station sends training pilot signal to terminal, and UE selects to receive training pilot signal, and carries out reception process to training pilot signal, and UE is by the signal quality of training pilot signal and thresholding SNR ₀contrast, and to base station feedback comparing result indication information.

Based on said method, after base station receives the described comparing result indication information of terminal feedback, whether can determine to terminal transmission data information according to this indication information.

Wherein, signal quality includes but not limited to signal to noise ratio, Signal to Interference plus Noise Ratio or carries dry ratio of making an uproar.

In addition, base station can also send signal quality thresholding Threshold configuration information to terminal.

Particularly, signal quality thresholding Threshold configuration information can be signal quality threshold value.

Signal quality thresholding Threshold configuration information can also comprise signal quality threshold level index, and now, terminal is configured self training signal quality thresholding according to the index received.

Also be noted that base station can configure lowest signal-to-noise thresholding Threshold information respectively for each terminal.

Or the channel condition information of terminal to report and signal quality thresholding Threshold contrast by base station, and Record Comparison result.

Example 2-1:

By following steps, base station and terminal determine that whether training is successful:

(1) lowest signal qualities threshold information is sent to terminal by base station;

Lowest signal qualities threshold information includes but not limited to concrete threshold value, or threshold level etc.;

Described signal quality includes but not limited to the dry ratio of signal to noise ratio, letter or carries dry ratio;

(2) terminal is configured self minimum quality thresholding according to the configuration information of base station;

(3) as described in example 1-1, between base station and terminal, beam selection is completed;

(4) base station is continuous or discontinuous at N _beamthe running time-frequency resource of individual agreement sends training pilot signal to terminal, N _beamfor positive integer; Base station send running time-frequency resource can for the cycle send or aperiodic transmission;

(6) terminal obtains the number of times m when time iteration;

(7) terminal according to a preconcerted arrangement criterion judge whether iteration terminates; Agreement criterion can for limiting maximum times criterion or other criterions;

If training continues after terminal judges this iteration, then continue iteration to base station feedback;

(8) terminal sends to base station by after training signal normalization;

In embodiment, step (8) and step (9) can be carried out simultaneously; The minimum transmission thresholding configuration cycle of step (1)-(2) can equal or be longer than a training process of base station and terminal.

Example 2-2:

As described in example 2-1, base station is terminal configuration lowest signal qualities thresholding, and the environment of base station residing for terminal is different, can be the lowest signal qualities thresholding that terminal configuration is different.

Or base station can, according to the change of terminal channel environment, not in the same time, be the lowest signal qualities thresholding that terminal configuration is different.

Embodiment 3

Base station sends training pilot signal to terminal, terminal is selected to receive pilot tone training signal, and reception process is carried out to training pilot signal, terminal obtains iterations information n, and n and times N are contrasted, comparing result indication information is sent to base station by terminal, and base station judges whether to proceed iteration according to described indication information.

Wherein, when the iterations of base station and terminal reaches N, N is positive integer, automatic termination of iterations.

Based on the above method, base station can also send the configuration information of N to terminal.

Concrete, the configuration information of N can be N hierarchy index, and now, terminal is configured according to the maximum iteration time of the configuration information of the N received to self.

In addition, the configuration information of N can be concrete N numerical value.

Be noted that base station can configure iterations N information respectively for each terminal.

In addition, base station can send to terminal and stop training Rule of judgment information, and whether terminal judges stops training.

Wherein, base station is that each terminal configures termination training Rule of judgment respectively.

Concrete, stop training Rule of judgment for meeting λ _ithresholding σ configuration information.

And configuration information can be thresholding σ numerical value, or it is thresholding σ hierarchy index.When configuration information is thresholding σ hierarchy index, terminal is configured self thresholding σ according to receiving index information.

Stopping training Rule of judgment also can for meeting signal quality thresholding R configuration information.

Wherein, signal quality comprise signal to noise ratio or Signal to Interference plus Noise Ratio or carry dry make an uproar than etc.;

Configuration information can be then concrete thresholding R numerical value; Or

For thresholding R hierarchy index information;

When configuration information is thresholding R hierarchy index information, after terminal receives hierarchy index, self thresholding is configured.

Based on said method, terminal can also be fed back times N.The content of concrete feedback can be:

Terminal is to base station feedback iterations N information;

Or terminal strengthens to base station feedback or reduces N number;

Or terminal promotes to base station feedback or reduces N class information.

Also be noted that whether base station can also stop according to the channel condition information training of judgement of terminal to report.

Concrete, base station calculates signal quality and λ _i, judge whether iteration continues according to thresholding σ;

Wherein, described signal quality is signal to noise ratio or Signal to Interference plus Noise Ratio or carries dry ratio of making an uproar;

And after the termination of base station training of judgement, training can also be stopped indication information and send to terminal.

Example 3-1:

In the present embodiment, base station and terminal complete a training process as follows:

(1) base station is at N _beamin individual OFDM symbol, continuous or discontinuous transmission to terminal in the pilot frequency locations of agreement trains pilot signal, N _beamfor positive integer; Base station send OFDM symbol can for the cycle send or aperiodic transmission;

(2) terminal is selected at N _beamreceive the training pilot signal that base station sends in individual OFDM symbol, and calculate the channel quality information of corresponding Received signal strength, as received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, the channel quality information that i-th OFDM symbol is corresponding is CQI _i, i=1 ..., N _beam;

(3) terminal obtains the number of times m when time iteration; Terminal can obtain when secondary iterations by the pilot frequency locations of agreement;

(4) iterations m and maximum iteration time N contrasts by terminal, if m equals N, then on ascending control channel or data channel to base station feedback finishing iteration, training completes; If m is less than N, then continue to base station feedback iteration on ascending control channel or data channel;

(5) the training signal normalization that will receive of terminal, obtains Y ₁, at N _beam2in individual OFDM symbol, the continuous or discontinuous pilot frequency locations arranging sends normalized Received signal strength Y to base station ₁;

(6) base station selected at N _beam2the training pilot signal that in individual OFDM symbol, receiving terminal sends;

(7) the whether continuation iteration flag bit of base station selected receiving terminal feedback, determines whether proceed iteration by flag bit feedback information; If proceed iteration, then will receive the normalization of training pilot frequency information and obtain Y ₂, Y ₂for new training pilot tone;

Repeat step (1)-(7), until base station receives terminal feedback finishing iteration in step (7).

Example 3-2:

Base station is that terminal configuration maximum iteration time process is as follows:

(1) base station sends maximum iteration time N information to terminal on down control channel, the maximum iteration time N information sent can be directly times N, or for representing the flag bit of N grade, base station can according to terminal channel environment or out of Memory, for different terminals configures identical or different maximum iteration time, identical or different maximum iteration time can not configured in the same time for same terminal yet;

Maximum iteration time N grade mark position can represent to be become:

Grade Index	Iterations
		0	5
1	10
		2	15
3	Retain

(2) terminal obtains maximum iteration time information on down control channel, and configures the maximum iteration time of self;

(3), during training, as embodiment 3-1 step (5), according to maximum iteration time, terminal judges whether iteration continues.

Example 3-3:

The present embodiment essentially describes base station and sends training pilot tone to terminal, starts to train rear terminal can stop training according to the judgement of some condition.Concrete implementation step is as follows:

(1) as shown in example 3-2, base station is terminal configuration maximum iteration time;

(2) base station is that terminal configuration terminates to train Rule of judgment in advance;

(4) base station is at N _beamin individual OFDM symbol, continuous or discontinuous transmission to terminal in the pilot frequency locations of agreement trains pilot signal, N _beamfor positive integer; Base station send OFDM symbol can for the cycle send or aperiodic transmission;

(5) terminal is selected at N _beamthe training pilot signal that base station sends is received in individual OFDM symbol, the line item and the channel quality information calculating corresponding Received signal strength is gone forward side by side, channel quality information can be received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, channel quality information corresponding to i-th OFDM symbol is CQI _i, i=1 ..., N _beam;

(6) terminal obtains the number of times m when time iteration; Terminal can obtain when secondary iterations by the pilot frequency locations of agreement;

(7) iterations m and maximum iteration time N contrasts by terminal, if m equals N, then on ascending control channel or data channel to base station feedback finishing iteration, training completes; If m is less than N, then enter into step (8);

(8) terminal judges whether finishing iteration in advance by the channel quality information of record, if can shift to an earlier date finishing iteration, then terminal terminates to base station feedback iteration; If need to continue training, then terminal terminates to base station feedback training; Terminal can judge whether to need finishing iteration in advance according to current C QI grade, also can judge whether finishing iteration in advance by CQI change of rank;

(9) if terminal judges iteration terminates, then as described in example 2-1, whether terminal judges this training be successful, and to base station feedback.

(10) the training signal normalization that will receive of terminal, obtains Y ₁, at N _beam2in individual OFDM symbol, the continuous or discontinuous pilot frequency locations arranging sends normalized Received signal strength Y to base station ₁

(11) base station selected at N _beam2the training pilot signal that in individual OFDM symbol, receiving terminal sends;

(12) the whether continuation iteration flag bit of base station selected receiving terminal feedback, determines whether proceed iteration by flag bit feedback information; If proceed iteration, then will receive the normalization of training pilot frequency information and obtain Y ₂, Y ₂for new training pilot tone;

Repeat step (4)-(12), until step (7) or (8) middle base station receive terminal feedback finishing iteration.

Step described in embodiment (1) or (2) cycle of training can be different from terminal, and the configuration maximum iteration time cycle is longer than or equals terminal cycle of training; Configuration terminates cycle of training in advance and is longer than or equals terminal cycle of training.

Example 3-4:

Base station is that terminal configuration terminates to train criterion process as follows in advance:

(1) base station sends to terminal and terminates training judgment criterion information in advance on down control channel, and the judgment criterion of transmission can be signal quality threshold information, or is λ _ithresholding δ information;

Described thresholding δ information can be concrete δ value, or δ grade;

(2) terminal obtains to terminate in advance to train and judges threshold information information on down control channel, and is configured the thresholding of self;

(3) start training, as described in embodiment 3-3, as embodiment 3-3 step (6) or (7), terminal is according to judging that thresholding judges whether iteration continues; And physically on row control channel feedback iteration whether terminate in advance.

Example 3-5:

The present embodiment essentially describes base station and sends training pilot tone to terminal, starts to train rear terminal can according to λ _ithresholding σ value judges premature termination training.Concrete implementation step is as follows:

(2) as shown in example 3-4, base station is that terminal configuration terminates to train criterion in advance;

(3) base station is at N _beamin individual OFDM symbol, continuous or discontinuous transmission to terminal in the pilot frequency locations of agreement trains pilot signal, N _beamfor positive integer; Base station send OFDM symbol can for the cycle send or aperiodic transmission;

(4) terminal is selected at N _beamthe training pilot signal that base station sends is received in individual OFDM symbol, the line item and the signal quality information calculating corresponding Received signal strength is gone forward side by side, channel quality information can be received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, channel quality information corresponding to i-th OFDM symbol is CQI _i, i=1, δ, N _beam;

(5) terminal is according to the signal quality information CQI of adjacent twice iteration _i-1, CQI _icalculate the λ when time iteration _i;

λ _i＝CQI _i/CQI _i-1

(8) terminal passes through λ _ijudge whether finishing iteration in advance, if λ _ibe greater than or equal to thresholding σ, then terminal judges proceeds iteration; If λ _ibe less than thresholding σ, then terminal judges finishing iteration, complete when time training;

Or judgment criterion is if λ _ibe greater than thresholding σ, then terminal judges proceeds iteration; If λ _ibe less than or equal to thresholding σ, then terminal judges finishing iteration, complete when time training.

(9) if shift to an earlier date finishing iteration, then terminal terminates to base station feedback iteration; If need to continue training, then terminal terminates to base station feedback training; If terminal judges iteration terminates, then as described in embodiment 2-1, whether terminal judges this training be successful, and to base station feedback training success or not.

(10) the training signal normalization that will receive of terminal, obtains Y ₁, at N _beam2in individual OFDM symbol, the continuous or discontinuous pilot frequency locations arranging sends normalized Received signal strength Y to base station ₁;

Repeat step (3)-(12), until step (7) or (8) middle base station receive terminal feedback finishing iteration.

Example 3-6:

The present embodiment essentially describes base station and sends training pilot tone to terminal, starts to train rear terminal can judge whether to stop training according to signal quality thresholding R.Concrete implementation step is as follows:

(4) terminal is selected at N _beamthe training pilot signal that base station sends is received in individual OFDM symbol, the line item and the signal quality information calculating corresponding Received signal strength is gone forward side by side, signal quality information can be received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, signal quality information corresponding to i-th OFDM symbol is CQI _i, i=1 ..., N _beam;

(5) terminal obtains the number of times m when time iteration; Terminal can obtain when secondary iterations by the pilot frequency locations of agreement;

(6) iterations m and maximum iteration time N contrasts by terminal, if m equals N, then on ascending control channel or data channel to base station feedback finishing iteration, training completes; If m is less than N, then enter into step (7);

(7) signal quality calculated and signal quality thresholding R contrast by terminal, and after judging this iteration, whether iteration terminates.If signal quality is greater than or equal to thresholding R, then iteration terminates; If signal quality is less than thresholding R, then iteration continues.

Or if signal quality is greater than thresholding R, then iteration terminates; If signal quality is less than or equal to thresholding R, then iteration continues.

(8) if shift to an earlier date finishing iteration, then terminal terminates to base station feedback iteration; If need to continue training, then terminal terminates to base station feedback training; If terminal judges iteration terminates, then as described in embodiment 2-1, whether terminal judges this training be successful, and to base station feedback.

(9) the training signal normalization that will receive of terminal, obtains Y ₁, at N _beam2in individual OFDM symbol, the continuous or discontinuous pilot frequency locations arranging sends normalized Received signal strength Y to base station ₁;

(10) base station selected at N _beam2the training pilot signal that in individual OFDM symbol, receiving terminal sends;

Repeat step (3)-(11), until step (6) or (7) middle base station receive terminal feedback finishing iteration.

Example 3-7

The present embodiment essentially describes terminal and can feed back maximum iteration time N information according to self training situation.Concrete implementation step is as follows:

(3) as described in example 3-3, terminal judges whether finishing iteration, until iteration terminates;

(4) whether terminal judges this training be successful, and to base station feedback;

(5) iterations of this training of terminal record, and preserved;

(6) whether this training of terminal record is successful, and preserved;

(7), after M training, M information from base station configures, and M time of preservation iterations is added up by terminal, and obtains best iterations;

The step of best iterations is asked to be: to be averaging by M iterations.

After M training, M information from base station configures, and terminal by training, whether add up by the probability of success;

(8) the best iterations of trying to achieve or the corresponding class information of best number of times are reported base station by terminal;

Or terminal, according to the statistics training probability of success, reports and promotes or reduce N grade, or feedback continues to continue to use and configured N grade last time.

(9) after base station receives reporting information by terminal, determine whether adopt terminal proposes, adjustment N or N grade.

Example 3-8:

Present embodiment describes the channel condition information of base station by terminal to report, determine whether training stops, and concrete steps are:

(2) terminal is selected at N _beamthe training pilot signal that base station sends is received in individual OFDM symbol, the line item and the signal quality information calculating corresponding Received signal strength is gone forward side by side, channel quality information can be received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, channel quality information corresponding to i-th OFDM symbol is CQI _i, i=1 ..., N _beam;

(3) the training signal normalization that will receive of terminal, obtains Y ₁, at N _beam2in individual OFDM symbol, the continuous or discontinuous pilot frequency locations arranging sends normalized Received signal strength Y to base station ₁;

(4) base station selected at N _beam2the training pilot signal that in individual OFDM symbol, receiving terminal sends;

(5) base station obtains the CQI information of terminal to report, according to the CQI of adjacent twice iteration _i-1, CQI _icalculate the λ when time iteration _i;

λ _i＝CQI _i/CQI _i-1

(6) base station obtains iterations, and is contrasted by iterations m and maximum iteration time N, if m equals N, then and termination of iterations, training completes; If m is less than N, then enter into step (8);

Wherein, base station can obtain iterations information by the position of training pilot tone.

(7) λ is passed through in base station _ijudge whether finishing iteration in advance, if λ _ibe greater than or equal to thresholding σ, then terminal judges proceeds iteration; If λ _ibe less than thresholding σ, then terminal judges finishing iteration, complete when time training;

(8) repeat step (1-7), until according to step (6) or (7), base station judges that iteration terminates, training completes.

(9) according to signal condition information gate limit value, base station judges that whether this time training be successful.

Example 3-9:

(7) signal quality calculated and signal quality thresholding R1 contrast by terminal, and after judging this iteration, whether iteration terminates.If signal quality is greater than or equal to thresholding R1, then continue iteration; If signal quality is less than thresholding R1, then think that signal is flooded by noise completely, then finishing iteration.

Or, if signal quality is greater than thresholding R1, then continue iteration; If signal quality is less than or equal to thresholding R, then finishing iteration.

Example 3-10:

λ _i＝CQI _i/CQI _i-1

(6) λ is passed through in base station _ijudge whether finishing iteration, if λ _ibe greater than or equal to thresholding σ, then terminal judges proceeds iteration; If λ _ibe less than thresholding σ, then terminal judges finishing iteration, complete when time training;

(7) repeat step (1-6), until according to step (6), base station judges that iteration terminates, training completes.

(8) according to signal condition information gate limit value, base station judges that whether this time training be successful.

Example 3-11:

(1) as shown in example 3-4, base station is that terminal configuration terminates to train criterion;

(2) base station is at N _beamin individual OFDM symbol, continuous or discontinuous transmission to terminal in the pilot frequency locations of agreement trains pilot signal, N _beamfor positive integer; Base station send OFDM symbol can for the cycle send or aperiodic transmission;

(3) terminal is selected at N _beamthe training pilot signal that base station sends is received in individual OFDM symbol, the line item and the signal quality information calculating corresponding Received signal strength is gone forward side by side, signal quality information can be received power, also can be received signal to noise ratio, receive Signal to Interference plus Noise Ratio, receive carry dry make an uproar than etc., wherein, signal quality information corresponding to i-th OFDM symbol is CQI _i, i=1 ..., N _beam;

(4) terminal obtains the number of times m when time iteration; Terminal can obtain when secondary iterations by the pilot frequency locations of agreement;

(5) signal quality calculated and signal quality thresholding R1 contrast by terminal, and after judging this iteration, whether iteration terminates.If signal quality is greater than or equal to thresholding R1, then continue iteration; If signal quality is less than thresholding R1, then think that signal is flooded by noise completely, then finishing iteration.

(6) if finishing iteration, then terminal terminates to base station feedback iteration; If need to continue training, then terminal terminates to base station feedback training; If terminal judges iteration terminates, then as described in embodiment 2-1, whether terminal judges this training be successful, and to base station feedback.

(7) the training signal normalization that will receive of terminal, obtains Y ₁, at N _beam2in individual OFDM symbol, the continuous or discontinuous pilot frequency locations arranging sends normalized Received signal strength Y to base station ₁;

(8) base station selected at N _beam2the training pilot signal that in individual OFDM symbol, receiving terminal sends;

(9) the whether continuation iteration flag bit of base station selected receiving terminal feedback, determines whether proceed iteration by flag bit feedback information; If proceed iteration, then will receive the normalization of training pilot frequency information and obtain Y ₂, Y ₂for new training pilot tone;

Repeat step (2)-(9), until base station receives terminal feedback finishing iteration in step (5).

Embodiment 4

Terminal receives the training pilot signal that base station sends, and sends to described base station after being normalized to described training pilot signal, and judges whether to proceed iteration according to training pilot signal current iteration frequency n, and judged result is fed back to described base station;

When described judged result is for continuing iteration, the pilot tone after the normalization of terminal feedback during the last iteration is sent to described terminal as new training pilot tone by described base station.

Wherein, before described terminal receives the training pilot signal of base station transmission, also include the selection operation of training pilot tone, this process the method also comprises:

Described base station sends the pilot signal of N wave beam to described terminal, and wherein, N is positive integer;

Described terminal selects a class or multiclass wave beam pilot signal from N class wave beam pilot signal, and beam information corresponding for selected wave beam pilot signal is fed back to described base station;

Described base station is determined to train wave beam from received beam information, and the pilot signal of determined training wave beam is sent to described terminal as training pilot signal.

And above-mentioned have training pilot tone selection operation also can be undertaken by base station side, now, this process then comprises:

Described terminal sends the pilot signal of up channel to described base station;

Described base station receives the pilot signal of described up channel, the channel condition information of up channel is calculated according to the pilot signal of described up channel, and select training wave beam according to described channel condition information and channel heterogeneite, pilot signal corresponding for selected training wave beam is sent to described terminal as training pilot signal.

Embodiment 5

The present embodiment provides a kind of base station, at least comprises: first module, and send N class wave beam pilot signal to terminal, N is positive integer;

Wherein, first module, can send beam information corresponding to different beams pilot signal to described terminal on sub-carriers.

On the basis of above-mentioned architecture of base station, can also comprise: dispensing unit, for the beam index number information that configurating terminal need report.

In some scheme, the selection training pilot tone then can be carried out in base station, and now base station can also comprise:

Unit the 3rd, the pilot signal selecting receiving terminal to send, calculates the channel condition information of up channel according to described pilot signal, and according to channel heterogeneite, selects wave beam and send to train to terminal.

Also be noted that above-mentioned base station also may carry out continuing training or the operation of deconditioning, now, base station also comprises:

Unit the 4th, whether the signal quality of training pilot signal of receiving terminal feedback and the comparing result indication information of thresholding Threshold, determine to described terminal transmission data information according to this indication information.

When base station includes above-mentioned Unit the 4th, dispensing unit, send signal quality thresholding Threshold configuration information can also to described terminal, described signal quality thresholding Threshold configuration information comprises signal quality threshold value, or signal quality threshold level index.

In addition, dispensing unit, also sends the configuration information of iterations N to terminal, and the configuration information of described N is N hierarchy index or concrete N numerical value.

Dispensing unit, also can configure described termination training Rule of judgment information respectively for each terminal, described termination training Rule of judgment is for meeting λ _ithresholding σ configuration information, or meet signal quality thresholding R configuration information.

Embodiment 6

The present embodiment introduces a kind of terminal, comprising:

Certainly, also may realize by base station side the selection training wave beam, now, terminal also comprises:

Unit the 3rd, to base station pilot signal transmitted;

Unit the 4th, receives the training wave beam of base station feedback, trains.

Also be noted that, terminal also may carry out judgement operation to be needed to continue training to help base station to determine whether, now terminal also comprises: Unit the 5th, select the training pilot signal receiving base station transmission, and reception process is carried out to selected training pilot signal, the signal quality of selected training pilot signal and thresholding Threshold are contrasted, and to described base station feedback comparing result indication information.

Or comprise Unit the 6th, obtain iterations information n, and n and preset times N is contrasted, and comparing result indication information is sent to base station, to make base station judge whether to proceed iteration according to described comparing result indication information, wherein, N is positive integer.

The all or part of step that one of ordinary skill in the art will appreciate that in said method is carried out instruction related hardware by program and is completed, and described program can be stored in computer-readable recording medium, as read-only memory, disk or CD etc.Alternatively, all or part of step of above-described embodiment also can use one or more integrated circuit to realize.Correspondingly, each module/unit in above-described embodiment can adopt the form of hardware to realize, and the form of software function module also can be adopted to realize.The application is not restricted to the combination of the hardware and software of any particular form.

The above, be only preferred embodiments of the present invention, be not intended to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any amendment made, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims

1. a beam shape-endowing weight value training method, it is characterized in that, comprise: base station sends N class wave beam pilot signal to terminal, described terminal selects a class in the wave beam pilot signal received or multiclass wave beam pilot tone, and beam information corresponding for selected wave beam pilot tone is fed back to described base station, and described base station is according to the beam information of described terminal feedback, determine initial training signal vector, and send described channel vector to described terminal and carry out wave beam weight training, wherein, N is positive integer.

2. the method for claim 1, is characterized in that,

Described base station is the configurating terminal beam index number information that need report also.

3. method as claimed in claim 2, is characterized in that, described in the beam index number information that reports at least comprise concrete index number and index number grade.

4. the method for claim 1, is characterized in that, described terminal selects a class in the wave beam pilot signal received or multiclass wave beam pilot tone, and beam information corresponding for selected wave beam pilot tone is fed back to described base station refers to:

5. a beam shape-endowing weight value training method, is characterized in that, comprising:

6. method as claimed in claim 5, is characterized in that,

Pilot frequency locations corresponding for beam index is informed to described terminal by described base station, and described terminal sends wave beam pilot signal to described base station in the pilot frequency locations of agreement.

7. the method as described in claim 5 or 6, is characterized in that,

Described terminal sends sounding pilot signal to described base station, and described base station calculates the channel condition information of up channel according to described pilot signal, and selects a wave beam according to described channel condition information, utilizes this wave beam to train.

8. a beam shape-endowing weight value training method, is characterized in that, comprising:

9. method as claimed in claim 8, whether it is characterized in that, the method also comprises: after described base station receives described comparing result indication information, determine to described terminal transmission data information according to this indication information;

10. method as claimed in claim 8 or 9, it is characterized in that, the method also comprises: described base station sends signal quality thresholding Threshold configuration information to described terminal.

11. methods as claimed in claim 10, is characterized in that, described signal quality thresholding Threshold configuration information comprises signal quality threshold value, or comprises signal quality threshold level index;

12. methods as claimed in claim 11, is characterized in that,

Described base station is that each terminal configures lowest signal-to-noise thresholding Threshold information respectively;

13. 1 kinds of beam shape-endowing weight value training methods, is characterized in that, comprising:

14. methods as claimed in claim 13, it is characterized in that, according to described indication information, described base station judges whether that the process proceeding iteration comprises:

15. methods as described in claim 13 or 14, it is characterized in that, described method also comprises:

Described base station sends the configuration information of N to described terminal.

16. methods as claimed in claim 15, is characterized in that, the configuration information of described N comprises N hierarchy index or concrete N numerical value;

17. methods as claimed in claim 16, is characterized in that,

Described base station is each terminal configured number N information respectively.

18. methods as claimed in claim 17, it is characterized in that, the method also comprises:

19. 1 kinds of beam shape-endowing weight value training methods, it is characterized in that, the method comprises:

20. methods as claimed in claim 19, it is characterized in that, the method also comprises:

21. methods as claimed in claim 20, is characterized in that,

Described base station is that each terminal configures described termination training Rule of judgment information respectively.

22. 1 kinds of beam shape-endowing weight value training methods, is characterized in that, comprising:

23. methods as claimed in claim 22, it is characterized in that, whether described base station stops referring to according to the channel condition information training of judgement of terminal to report:

24. methods as described in claim 22 or 23, is characterized in that,

After described base station training of judgement stops, also training is stopped indication information and send to described terminal.

25. 1 kinds of base stations, is characterized in that, comprising: first module, and send N class wave beam pilot signal to terminal, N is positive integer;

26. 1 kinds of terminals, is characterized in that, comprising: