CN102404028A - Wave beam forming method - Google Patents

Abstract

The present invention provides a beamforming method, including: UE sends DMRS in each uplink data subframe and periodically sends SRS; eNodeB receives SRS signal according to the SRS sending cycle to perform channel estimation, and uses the channel The estimation result calculates and saves the latest autocorrelation matrix R _xx ; the eNodeB receives the DMRS signal of the UE in each uplink subframe for channel estimation, and uses the channel estimation result to calculate and save the latest instantaneous beamforming vector w using the GOB algorithm; When performing downlink beamforming, use the latest autocorrelation matrix R _xx and the latest instantaneous beamforming vector w to calculate the downlink beamforming weight coefficient W _BF =R _xx *w, and according to the downlink beamforming The shape weight coefficient performs beamforming on the downlink scheduled data.

Description

A Beamforming Method

technical field

The invention relates to a data transmission technology in a communication system, in particular to a beam forming method.

Background technique

The beamforming transmission scheme in the LTE TDD system takes advantage of the reciprocity of the uplink and downlink of the TDD system, and calculates the beamforming vector for downlink transmission through the channel estimation of the uplink SRS. In this way, beamforming performance is directly affected by channel reciprocity.

Factors affecting uplink and downlink reciprocity mainly include:

1. The SRS period is long, and the estimation of the uplink channel cannot quickly adapt to the change of the downlink channel

The SRS transmission period stipulated in the 3GPP protocol can be {2, 5, 10, 20, 40, 80, 160, 320} ms, but considering the time-varying channel and UE moving speed and other factors, too long an SRS period will seriously damage the upper and lower The reciprocity of the upstream channel will cause the deterioration of the beamforming performance. The specific simulation results are shown in Figure 1.

2. SRS listening bandwidth and downlink resource allocation are asymmetric

According to the stipulations of the 3GPP agreement, the SRS interception bandwidth can be wideband or narrowband. For example, when Table 1 is for 20MHz (100RB) system bandwidth, the maximum SRS interception bandwidth is 96RB, and the minimum is 4RB. In the case of limited resources, it is impossible for all UEs in the cell to send wideband SRS. In this way, the asymmetry between the SRS listening frequency band and the frequency band for downlink resource allocation is likely to occur.

Table 1

For the two problems raised above, if the beamforming transmission scheme is adopted, the enhancement technology of uplink and downlink channel reciprocity must be adopted.

Among the existing solutions, the patent CN200710175220.8 proposes the concept of instantaneous channel and long-term statistical channel. The specific operation method is: estimate the uplink channel information according to the pilot signal, calculate the instantaneous channel state Estimated downlink channel information He, and determine the weight coefficient α between He and Hf, so as to determine the downlink channel as Hd=(1-α)He+αHf. This method enhances the reciprocity of the uplink and downlink channels, but the pilot used in it is SRS, and the precondition that the SRS listening frequency band and downlink resource allocation must be consistent must be met.

Patent CN200510009643.3 solves the frequency offset problem of the uplink and downlink of the FDD system by introducing a frequency calibration matrix. The covariance matrix of the uplink channel is obtained by using the frequency calibration matrix to process the covariance matrix of the uplink channel, and the beamforming coefficient can be directly obtained. The invention can be used to solve the frequency offset problem caused by the asymmetry of SRS listening frequency band and downlink resource allocation, but does not consider the deterioration of uplink and downlink channel reciprocity caused by longer SRS period.

Contents of the invention

The invention provides a beam forming method, which can solve the problem of deterioration of uplink and downlink channel reciprocity caused by long SRS period.

To achieve the above object, the present invention adopts the following technical solutions:

A beamforming method comprising:

The UE sends DMRS in each uplink data subframe, and periodically sends SRS;

The eNodeB receives the SRS signal according to the transmission period of the SRS to perform channel estimation, and uses the channel estimation result to calculate and save the latest autocorrelation matrix R _xx ;

The eNodeB receives the UE's DMRS signal for channel estimation in each uplink data subframe, and uses the channel estimation result to calculate and save the latest instantaneous beamforming vector w;

When performing downlink beamforming, use the latest autocorrelation matrix R _xx and the latest instantaneous beamforming vector w to calculate the downlink beamforming weight coefficient, and adjust the downlink scheduling according to the downlink beamforming weight coefficient The data is beamformed.

Preferably, the calculating and storing the latest instantaneous beamforming vector w is: calculating the latest instantaneous beamforming vector w by using a GOB algorithm.

Preferably, the calculation of the downlink beamforming coefficient is: W _BF =R _xx *w.

It can be seen from the above technical solution that in the present invention, on the one hand, the eNodeB uses the periodically received SRS signal to perform channel estimation, and calculates and saves the autocorrelation matrix of the channel as long-term CSI; on the other hand, the eNodeB uses the periodically received SRS signal The DMRS signal is used for channel estimation, and the estimation result is used as short-term CSI, and the short-term CSI is used to calculate and save the instantaneous beamforming vector; then, the downlink beamforming weight is calculated by combining the above long-term CSI and short-term CSI Coefficient W _BF =R _xx *w, using the downlink beamforming weight coefficient to perform beamforming of downlink scheduling data. In this way, the combination of long-term CSI and short-term CSI is used to calculate downlink beamforming weight coefficients, which can solve the problem of deterioration of uplink and downlink channel reciprocity caused by a long SRS period. Furthermore, the GOB algorithm can be used when calculating the instantaneous beamforming vector, so as to solve the frequency offset problem caused by the asymmetry of the uplink and downlink resources.

Description of drawings

Figure 1 is a schematic diagram of beamforming performance comparison corresponding to different SRS transmission periods in the existing beamforming method;

Fig. 2 is the specific flowchart of the beamforming method in the present invention;

Figure 3 is a schematic diagram of the relationship between the signal-to-noise ratio and the bit error rate after only using the GOB algorithm for beamforming under different uplink and downlink frequency differences;

4 is a schematic diagram of the relationship between the signal-to-noise ratio and the bit error rate after the beamforming method of the present invention is used to perform beamforming under different uplink and downlink frequency differences;

FIG. 5 is a first schematic diagram of the performance comparison between the existing beamforming method and the beamforming method of the present invention;

FIG. 6 is a second schematic diagram of the performance comparison between the existing beamforming method and the beamforming method of the present invention;

FIG. 7 is a schematic diagram 3 of performance comparison between the existing beamforming method and the beamforming method of the present invention;

FIG. 8 is a fourth schematic diagram of performance comparison between the existing beamforming method and the beamforming method of the present invention.

Detailed ways

In order to make the purpose, technical means and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings.

The basic idea of the present invention is: to solve the problem of long SRS transmission period, use DMRS to estimate short-term uplink channel, use the result of SRS channel estimation as long-term CSI, and then combine short-term CSI when performing beamforming term uplink channel and long-term CSI to determine the current beamforming weight coefficient, so as to solve the problem of deterioration of uplink and downlink channel reciprocity caused by longer SRS period; further, for the frequency offset caused by the asymmetry of uplink and downlink resources The problem is to use the GOB algorithm to calculate the instantaneous shaping vector w according to the short-term uplink channel.

The specific implementation of the present invention will be described below through specific embodiments.

Fig. 2 is the specific flowchart of the beamforming method in the present invention, as shown in the figure, the method includes:

In step 201, the UE sends a DMRS in each uplink data subframe, and periodically sends an SRS.

Wherein, SRS and DMRS are existing signals. The DMRS is sent every time the UE is scheduled for uplink. Therefore, when the SRS sending period is long, the UE will send the DMRS relatively frequently within the interval of the SRS period, so the DMRS can be used to update the uplink channel.

In step 202, the eNodeB receives the SRS signal according to the transmission period of the SRS to perform channel estimation, and uses the channel estimation result to calculate and save the latest autocorrelation matrix R _xx .

In this step, when each SRS transmission period arrives, the eNodeB receives the SRS signal and performs uplink channel estimation, uses the channel estimation result to calculate the autocorrelation matrix of the channel, and always saves the latest autocorrelation matrix. The autocorrelation matrix reflects the long-term statistical characteristics of the uplink channel, that is, long-term CSI.

The specific processing method in this step is the same as that of the prior art, and will not be repeated here.

In step 203, the eNodeB receives the DMRS signal of the UE for channel estimation in each uplink data subframe, and uses the channel estimation result to calculate and save the latest instantaneous beamforming vector w.

In this step, the UE sends a DMRS in each uplink subframe, and the eNodeB receives the DMRS signal and performs uplink channel estimation. Since DMRS is sent frequently, the channel estimated from this signal reflects the instantaneous characteristics of the uplink channel, that is, short-term CSI.

For short-term CSI DMRS, its transmission frequency band may be inconsistent with downlink resource allocation. Preferably, this embodiment uses the GOB algorithm to solve this problem. Compared with the frequency calibration algorithm provided by CN200510009643.3, both of them use the basic principle that the wireless propagation path keeps DOA constant within a certain frequency band. Specifically, by using Wireless Insite software to simulate wireless transmission paths and DOA at different frequencies, the results shown in Table 2 can be obtained.

Table 2

It can be seen from the data in Table 2 that the frequency has no influence on the wireless propagation path, and the DOA of different frequency points are the same. Although the propagation power of each path at different frequency points is different, the main path remains unchanged within the simulated 100MHz bandwidth. Therefore, the DOA information of the uplink channel can be used to perform beamforming of the downlink channel, even if the frequency of the uplink and downlink channels is offset.

Based on the above analysis, in this step, preferably, after each time the uplink channel estimation result is obtained according to the DMRS signal, the instantaneous beamforming vector w is calculated using the uplink channel estimation result according to the GOB algorithm, and the current latest instantaneous beamforming is always saved vector. Wherein, the specific method of performing channel estimation according to the received DMRS and calculating the beamforming vector according to the GOB algorithm is the same as the existing method, and will not be repeated here.

In addition, since the sending periods of DMRS and SRS are different, the above steps 202 and 203 are not performed in a fixed order, but may be performed interspersed. Generally, the DMRS is sent frequently, therefore, within the period of the SRS, it may be necessary to perform DMRS signal reception, channel estimation, and calculation and update of the instantaneous beamforming vector many times. Generally, each uplink data subframe of the UE needs to perform the processing of this step, so that the most recent instantaneous shaping coefficient of the uplink channel is used in the downlink shaping.

Step 204, when performing downlink beamforming, use the latest autocorrelation matrix R _xx and the latest instantaneous beamforming vector w to calculate the downlink beamforming weight coefficient W _BF =R _xx *w.

Through the processing of the first two steps, the long-term CSI and short-term CSI of the uplink channel are obtained. In this step, when the downlink beamforming is performed, the current downlink beamforming weight coefficient is calculated according to the two contents. Specifically, the downlink beamforming weight coefficient is: W _BF =R _xx *w. Wherein, R _xx is the latest autocorrelation matrix saved by the eNodeB, and w is the latest instantaneous beamforming vector saved by the eNodeB.

In the process of calculating downlink beamforming weight coefficients, compared with CN200510009643.3, the present invention does not need to correct the channel autocorrelation matrix before performing operations such as eigenvalue decomposition, so the implementation complexity is greatly reduced.

Step 205 : Perform beamforming on the downlink scheduled data according to the downlink beamforming weight coefficient calculated in step 204 .

In the aforementioned process, on the one hand, the GOB algorithm is used to calibrate the frequency offset, and on the other hand, the long-term CSI+short-term CSI is used to overcome the deterioration of channel reciprocity caused by the long period of SRS. Therefore, using the downlink beam When the shaping weight coefficient is used for downlink beamforming, it can not only solve the problem of long SRS period, but also solve the problem of asymmetry between uplink and downlink frequency bands.

The specific processing of this step is the same as the existing method, and will not be repeated here.

So far, the flow of the beamforming method in the present invention ends. The above-mentioned method of the present invention is not only applicable to scenarios with a long SRS cycle, but also applicable to traditional application scenarios. That is, when the SRS period is shorter than the channel correlation time, if the UE sends a narrowband SRS, the method of the present invention can also be used to enhance the reciprocity of the uplink and downlink channels, thereby obtaining performance gains.

Below, a performance simulation comparison between the beamforming method of the present invention and the existing beamforming method is given. The simulation parameters are shown in Table 3.

parameter name value Antenna configuration 8x2 UE antenna spacing 0.5λ eNB antenna spacing 0.5λ Antenna Polarization BS cross polarization，UE co-polarization carrier frequency 2.6GHz system bandwidth 20MHz CP type Normal CP Modulation QPSK Downlink channel estimation LS SRS channel estimation LS Equilibrium algorithm MMSE Encoding turbo Channel model SCME Communication scene Urban Micro UE movement speed 3km/h

table 3

In the simulation, we set two parameters, SFoffset and Tsrs, where SFoffset represents the offset of the uplink and downlink subframes, which is set to 5ms in the simulation, that is to say, the update period of DMRS is 10ms; Tsrs is the SRS period, which is greater than that of DMRS cycle.

First, the influence of uplink and downlink frequency difference on beamforming performance is illustrated through simulation results. Fig. 3 is a schematic diagram of the relationship between the signal-to-noise ratio and the bit error rate after only using the GOB algorithm for beamforming under different uplink-downlink frequency differences. Wherein, the uplink-downlink frequency difference is the frequency difference between the uplink DMRS frequency band resource and the downlink scheduling frequency band. In the simulation, the PRB index of the fixed downlink resource allocation is 50, and different curves in the simulation results represent different PRB indexes corresponding to the uplink DMRS. It can be seen from Fig. 3 that the uplink-downlink frequency difference will have a considerable impact on the beamforming performance.

Next, another simulation is performed using the beamforming method of the present invention, and the simulation result is shown in FIG. 4 . It can be seen that by using the SRS correlation matrix Rxx to correct the w calculated by the DMRS, not only the beamforming performance is significantly improved, but also the performance difference caused by the frequency difference is reduced.

Finally, the effectiveness of the present invention is verified by simulating the performance of different beamforming processing methods with different SFoffset and Tsrs parameters. The specific simulation results are shown in Figures 5, 6, 7, and 8. Among them, the traditional method means that only the SRS channel estimation is used for beamforming; the DMRS-GOB means that only the DMRS-based GOB algorithm is used for beamforming. As can be seen from the simulation results shown in Figures 5, 6, 7, and 8, under the longer SRS transmission period, the algorithm of the present invention can obtain a performance gain of 2 to 3 dB compared to the traditional algorithm (BLER= ^{10- 1} ).

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the present invention. within the scope of protection.

Claims (3)

1. a beam form-endowing method is characterized in that, this method comprises:

UE sends DMRS in each upstream data subframe, and periodically carries out the transmission of SRS;

ENodeB carries out channel estimating according to the transmission cycle reception SRS signal of said SRS, and utilizes this channel estimation results to calculate and preserve up-to-date autocorrelation matrix R _Xx

ENodeB carries out channel estimating at the DMRS signal that each upstream data subframe receives UE, and utilizes this channel estimation results to calculate and preserve up-to-date instantaneous wave beam forming vector w;

When carrying out down beam shaping, utilize said up-to-date autocorrelation matrix R _XxCalculate the down beam shaping weight coefficient with said up-to-date instantaneous wave beam forming vector w, and the data of descending scheduling are carried out wave beam forming according to this down beam shaping weight coefficient.

2. method according to claim 1 is characterized in that, said calculating is also preserved up-to-date instantaneous wave beam forming vector w and is: adopt the said up-to-date instantaneous wave beam forming vector w of GOB algorithm computation.

3. method according to claim 1 and 2 is characterized in that, said calculating down beam shaping coefficient is: W _BF=R _Xx* w.

Images