CN1139192C - Method and receiver for uplink receiving adaptive array in wireless communication system - Google Patents

Abstract

The present invention relates to a method and a receiver for up receiving an adaptive array in a radio communication system. Analog signals received by a receiving array are converted into digital signals by analog-to-digital conversion; the digital signals are dispread to obtain array signal data flow; weighting vectors are formed by using the array signals before the dispreading and the array signals after the dispreading, weighting summation forms a wave beam, and the wave beam is conveyed to the corresponding track of a Rake receiver to carry out the maximum ratio combination. The present invention is characterized in that the wave beam of the receiver is formed by using the weighting summation method of an optimized sample array inversion algorithm structure. The present invention can effectively suppress interference and noise, and enhance the capacity and the quality of the radio communication system.

Description

Method and receiver for uplink receiving adaptive array in wireless communication system

The invention is applied to wireless communication systems such as wireless cellular products and wireless local loops, and relates to an adaptive method for an uplink receiving array of a wireless communication system and a receiver thereof.

For a long time, wireless communication systems have always been faced with the contradiction between limited available spectrum resources and rapidly growing user demands. Therefore, people began to take advantage of the spatial characteristics of the channel, adopt diversity, sectorization, and recently proposed technologies such as smart antennas using array antennas, which can improve the communication quality of the wireless communication system to varying degrees and increase the system capacity.

Diversity mainly utilizes the fact that the signals received by different antennas with a distance greater than 10 carrier wavelengths are not correlated with each other. The signals received by each antenna are combined by maximum ratio to improve the anti-multipath fading performance of the system.

The sectorization method is to divide the cell into 3, 6, 9 or 12 sectors, and each sector has its own supporting antenna and preset spectrum range. Sectorization reduces the co-channel interference to a certain extent, thereby improving the communication quality of the system.

Smart antenna technology forms beams in the signal direction by adjusting the phase and amplitude of signals on multiple antenna elements to improve signal quality. The smart antenna method can be mainly divided into two categories, one is the switch multi-beam method. This type of method forms fixed beams in different directions to cover the entire cell. The base station detects the signal quality of the desired signal in each beam and selects the best beam for reception.

Another important smart antenna method is the adaptive array, which adaptively weights and combines the signals received by each antenna element according to a certain criterion, enhances the signal, suppresses interference and noise, and thus improves the overall performance of the wireless system. Common adaptive methods include Least Mean Square (LMS), Recursive Least Squares (RLS), Sampling Matrix Inversion (SMI) based on the minimum mean square error criterion, and code filtering based on the maximum output signal-to-interference ratio criterion. Isoblind method.

The diversity method requires a large distance between the antennas (generally greater than 10 wavelengths), so the space occupied by the antennas is relatively large. In addition, although the diversity method using the maximum ratio combination has the effect of anti-multipath fading, it cannot effectively suppress the interference signal.

The common sectorization method is to use 3 sectors or 6 sectors. The reason why more sectors are not used is because the more sectors are divided, the less spectrum resources each sector can use, which reduces the Inheritance efficiency, and frequent switching is required, reducing system efficiency.

The switched multi-beam method is a sub-optimal receiving method, which has poor ability to suppress colored noise. In addition, the switch matrix of the existing switch multi-beam system is implemented by radio frequency switch devices, which increases the hardware cost of the system.

Adaptive antenna array can make the system performance optimal under certain criteria in theory. Commonly used criteria are mostly the minimum mean square error criterion and the maximum output signal-to-noise ratio criterion, and the two criteria are equivalent under certain conditions. In actual application, SMI, RLS and most of the blind methods require a large amount of calculation, which is difficult to implement with existing digital processing chips. Although the calculation amount of the LMS method is small, the convergence speed and stability of the method are seriously affected by factors such as the signal environment, the selection of the reference signal, and the size of the calibration step.

The purpose of the present invention is to overcome the above-mentioned shortcomings of numerical instability of direct matrix inversion under low signal-to-noise conditions in the prior art, and propose an adaptive array receiver with low computational complexity and excellent performance in a wireless communication system. The method and its receiver, by optimizing the method structure of the standard SMI method, the method has good performance and strong achievability through analysis and simulation.

The technical solution for realizing the object of the present invention is: the method for uplink receiving array of the wireless communication system, the analog signal received by each antenna unit of the receiving array is converted into array digital signal through analog/digital conversion, and the array digital signal is obtained after despreading through despreading The array signal data flow after despreading is used to form the weighted vector of the array by using the despreaded array data signal, and the generated weighted vector is used to weight and sum the despreaded array signal to form a beam, and the beamformed signal is sent to the rake receiver for corresponding The path is carried out by the rake receiver to the maximum ratio combination of signals in the time domain; it is characterized in that, the beam is formed by a weighted summation method that optimizes the sampling matrix inversion structure, and its steps are:

(1) Starting from the first time slot, the digital beamformer sequentially reads the data before despreading and the pilot sequence or training signal received by the array of the current time slot;

(2) Use the signal on an array element of the current time slot to correlate with the pilot sequence or the training signal to calculate the average, and estimate the channel response of the current time slot;

(3) generate the reference signal after the channel response compensation with the channel response of the current time slot estimated in step (2) and the pilot sequence or training signal of the current time slot;

(4) The data matrix in step (1) is extracted with a certain sampling interval by the data extraction module to form a data matrix with fewer samples, and then use this data matrix to generate the incremental matrix of the autocorrelation matrix of the current time slot. The autocorrelation matrix of a time slot introduces an appropriate forgetting factor and accumulates with the increment matrix to obtain the autocorrelation matrix of the current time slot;

(5) The data matrix in the step (1) is passed through the data extraction module to form a data matrix with fewer samples at a certain sampling interval, and at the same interval the reference signal in (3) is extracted to obtain a smaller number of samples For the reference signal, use the extracted related data matrix and the reference signal to accumulate to obtain the incremental vector of the correlation vector of the current time slot, and then use the correlation vector of the previous time slot to introduce an appropriate forgetting factor and accumulate it with the incremental vector of the correlation vector Get the autocorrelation vector of the current time slot;

(6) Apply the autocorrelation matrix of step (4) and the correlation vector of step (5), perform Cholesky decomposition on the autocorrelation matrix, and update the weight of the current time slot by simple previous and back generation methods vector.

(7) Normalize the updated weight vector and send it to the weighting module;

(8) Repeat steps (1) to (7) for updating the weight vector of the next slot array.

In the above-mentioned method for uplink receiving arrays in a wireless communication system, the sampling interval extracted in step (5) may be the same as or different from the sampling interval extracted in step (4).

In the method for the uplink receiving array of the wireless communication system, the value range of the forgetting factor in the method for estimating the autocorrelation array and the correlation vector is between 0.5 and 1.

In the method for the uplink receiving array of the wireless communication system, the estimated channel response refers to the channel fading response estimated by using the delay alignment signal.

An antenna array receiver, comprising an antenna array, an array digital signal generation module, a digital beam forming module, and a rake receiver; the array digital signal generation module includes a receiving unit and an analog-to-digital conversion unit; the antenna array input multiple path signal to the array digital signal generation module, and the digital beamforming module forms beams for different multipath signals of a channel at the same time. The signal is combined with the maximum ratio; it is characterized in that: the digital beamforming module includes a multipath search unit, a delay alignment unit, a channel response estimation unit, a reference signal generation unit, an autocorrelation matrix estimation unit, a weighting vector generator, a multiplication device, adder, despreading unit, and regenerated pilot symbol or training sequence unit; the array digital signal generation module converts the analog signal received by each antenna into a digital signal, which is divided into two channels after being sent to the digital beamforming module: one channel Enter the autocorrelation matrix estimation unit to estimate the autocorrelation matrix, and send it to the weight vector generator in all paths in all channels; the other path enters the multipath search and tracking cable and tracking module to use one or more for each channel The digital signal of the antenna performs multi-path search and tracking processing, and provides the multi-path delay information to the despreading unit; the delay unit aligns the digital signals received by the array according to the specified time delay, and the aligned m-channel signals are divided into two channels , all the way into the weighted vector generator, all the way into the multiplier as the weighted data, ready to be multiplied with the weighted vector output by the weighted vector generator, and finally combined by the adder to obtain the output data of the array beam, which is resolved by the despreading unit The output of each path is formed after expansion and connected to the corresponding path of the rake receiver.

Above-mentioned a kind of antenna array receiver, wherein, described weight vector generator comprises extraction module, correlation vector calculation module, matrix triangular decomposition module and solution triangular equations module; The output of array digital signal generation module is sent into delay alignment module At the same time, after being extracted by the data extraction module, it is sent to the autocorrelation matrix calculation module; similarly, the array digital signal and reference signal are sent to the correlation vector calculation module after being extracted by the data extraction module; the output of the autocorrelation matrix calculation module is decomposed by the matrix triangular decomposition module into two triangular matrices, and the outputs of the matrix triangular decomposition module and the correlation vector calculation module are simultaneously sent to the module for solving triangular equations to generate weighted vectors.

The aforementioned antenna array receiver, wherein the autocorrelation matrix estimation module includes m data extraction modules and an autocorrelation matrix calculation module connected thereto; the m-channel signals output by the array digital signal generation module are output to the The autocorrelation matrix calculation module is output to the weighted vector generator.

The aforementioned antenna array receiver, wherein the channel response estimation module correlates the digital signal of an array element output by the delay alignment unit with the known pilot symbol in each time slot, and obtains the difference of a channel A rough estimate of channel fading by multipath in each slot.

The above-mentioned antenna array receiver, wherein the reference signal generation module uses the channel response estimation module in each time slot to obtain the estimated value and the spread spectrum sequence of the regenerated pilot symbol or training sequence to generate the weight vector generator required the reference signal.

The aforementioned antenna array receiver, wherein the weight vector generator updates the weight vector of the array antenna in each time slot, and provides the updated weight vector in each time slot to the multiplier for beam weighting.

Since the present invention adopts the above technical scheme, it overcomes the numerical instability of direct matrix inversion under low signal noise, and the amount of calculation is obviously smaller than that of the traditional SMI method, so that the amount of calculation of each weight estimation is reduced to about 4SaKm2+8SaK′m+4/3m2+m times of floating-point addition/multiplication; more importantly, each path only needs to be despread once, so the total calculation amount and hardware overhead of the system are reduced, and the existing general-purpose The digital signal processor (DSP) chip can be completed, which is highly achievable, and can effectively suppress interference and noise signals at the same time, and improve the capacity and communication quality of the wireless communication system.

The characteristic properties of the present invention are further described in detail by the following examples and accompanying drawings.

FIG. 1 is a schematic diagram of an array receiver currently used.

FIG. 2 is a schematic diagram of an SMI digital beamforming module corresponding to a Rake path (finger) of a multipath signal according to the present invention.

Fig. 3 is a schematic diagram of the autocorrelation matrix estimation module of the present invention.

Fig. 4 is a schematic diagram of the weight vector generator of the present invention.

Fig. 5 is a graph of the uplink simulation results of the present invention for the WCDMA system.

Please refer to Figure 1 and Figure 2. FIG. 1 is a schematic diagram of an array receiver currently used. The array antenna receiver mainly includes an antenna array 101 , an array digital signal generating module 102 , a digital beam forming module 103 and a Rake receiver 104 . Within the scope of the serving cell, the distance between the antenna elements can be selected to be half of the wavelength corresponding to the center frequency of the carrier. The array digital signal generation module 102 includes a receiving unit 105 and an analog-to-digital conversion unit 106, whose main function is to convert the analog signal received by the antenna array into an array digital signal that can be processed digitally, and the subsequent signal processing is performed in the digital domain. The digital beamforming module 103 simultaneously forms beams for different multipath signals of a channel, and the beamformed multipath signals 108 are sent to the corresponding paths of the Rake receiver 104, and the Rake receiver 104 performs maximum ratio combining on the multipath signals.

Taking the CDMA system as an example below, the principle of the method of the present invention will be described.

Suppose a channel array receives a digital signal expressed as: $x=\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\mathrm{\α}\left({\mathrm{\θ}}_l\right)h_l\left(t\right)C_{\mathrm{the\; s}}(t-{\mathrm{\τ}}_l)\mathrm{the\; s}(t-{\mathrm{\τ}}_l)+\mathrm{no}\left(t\right)---\left(1\right)$

Where l=1, 2,..., L is the number of multipaths, α(θ _l ) is the array response of the lth multipath signal, and _θl is the direction of arrival DOA of the lth multipath; h _l ( t) is the fading experienced by the l multipath signal; s(t) is the desired signal transmitted, C _s (t) is the spreading code corresponding to the desired signal, and τ _l is the time delay of the l multipath signal; n(t) is the array interference and noise signal, and the noise on each antenna is regarded as independent and uncorrelated zero-mean additive white Gaussian noise.

In the optimized SMI method of the present invention, the reference signal is taken as

d(t)＝h _d (t)C _s (t-τ _d )s(t-τ _d ) (2)

Correlating the array signal with the reference signal yields:

R _xd ＝E[X(t)d*(t)]＝α(θ _d )E[h _d (t)s(t)*s(t)]+E[n′(t)d*(t )] (3)

Let s(t)s*(t)=‖s(t)‖ ² =1, we can get: $R_{\mathrm{xd}}=\mathrm{\α}\left({\mathrm{\θ}}_d\right)\mathrm{E.}\[{\left|\right|h_d\left(t\right)\left|\right|}^2\]+\mathrm{E.}\[h_d^{*}\left(t\right)\]\mathrm{E.}\[{\mathrm{no}}^{\′}\left(t\right)\mathrm{Cs}*\left(t\right)\mathrm{the\; s}*\left(t\right)\]---\left(4\right)$

Among them: θ _d , τ _d are the direction of arrival and time delay of the desired user signal.

It can be seen from formula (4) that the reference signal can be selected according to formula (2) to ensure the coherent accumulation of the signal and at the same time reduce the influence of the noise term on the estimation of Rxd, which can improve the estimation accuracy of Rxd, so the performance of the SMI method can be been greatly improved. According to the minimum mean square error criterion, the SMI method converges to the Wiener solution: $w_{\mathrm{opl}}=R_{\mathrm{XX}}^{-1}*R_{\mathrm{xd}}---\left(5\right)$

In practical applications, both the autocorrelation matrix Rxd and the correlation vector Rxd are obtained through the likelihood estimation of a large number of sampling points, so the calculation amount of the method is large, and it is difficult to realize it with the existing DSP chip.

FIG. 2 is a schematic diagram of an SMI digital beamforming module 210_1 corresponding to a Rake path (finger) of a multipath signal according to the present invention. It mainly includes a multipath search unit 201, a delay alignment unit 202, a channel response estimation unit 204, a reference signal generation unit 206, a weight vector generator 207, a multiplier 208, an adder 209, and a despreading unit 211. The array digital signal generating module 102 converts the analog signal received by each antenna 101 into a digital signal, which is divided into two channels after being sent to the digital beam forming module 103 . One way enters the autocorrelation matrix estimating unit 212 to estimate the autocorrelation matrix in (5), and sends it to the weight vector generator 207 in all paths in all channels, and the other way enters the multipath search and tracking module 201 for each channel The digital signals of one or more antennas are used to perform multipath search and tracking processing, and the multipath time delay information is provided to the despreading unit 202 . The delay unit aligns the digital signals 107 received by the array according to specified time delays. The m-channel signals after alignment are respectively divided into two channels, one of which enters the weighted vector generator 207, and the other enters the multiplier 208 as the weighted data, and is prepared to be multiplied with W1~Wm output by the weighted vector generator 207 and finally passes through the adder 209 After combining, the output data of the array beams is obtained, and the data is despread by the despreading unit 211 to form the output of each path, and connected to the corresponding path 108 of the Rake receiver 104 . The channel response estimation unit 204 estimates the fading response of the channel by using the delay alignment signal on a certain antenna (the despread signal 203_1 of the first antenna is marked in the figure), and the reference signal generation unit 206 uses the channel response estimated by 204 and the regeneration The spread spectrum sequence of pilot symbols or training sequence generates the reference signal (2) of the SMI method in the weight vector generator.

Figure 3 is a schematic diagram of the autocorrelation matrix estimation module. It includes n data extraction modules 302 and an autocorrelation matrix calculation module 301 . The n signals output by the array digital signal generation module are respectively output to the autocorrelation matrix calculation module through the extraction module, and then output to the weighted vector generator.

Figure 4 is a schematic diagram of the weight vector generator. It mainly includes an extraction module 402 , a correlation vector calculation module 401 , a matrix triangular decomposition module 403 and a triangular equation solving module 404 . The output 107 of the array digital signal generation module is sent to the autocorrelation matrix calculation module 301 after being extracted by the data extraction module 302 while being sent to the delay alignment module. Similarly, the array digital signal 203 and the reference signal 213 are extracted by the data extraction module 402 and then sent to the correlation vector calculation module 401 . Since data processing is basically processed before despreading, the data extraction module can significantly reduce the computational requirements of the method. The output 303 of the autocorrelation matrix calculation module is decomposed into two triangular matrices through the matrix triangular decomposition module 403, and the outputs of the matrix triangular decomposition module and the correlation vector calculation module are simultaneously sent to the triangular equation solving module 404 to generate weighted vectors W1-Wm.

The channel response estimation module correlates the digital signal of an array element output by the delay alignment unit with the known pilot symbol in each time slot, and obtains the channel fading in each time slot of different multipaths of a channel rough estimate of .

The reference signal generation module uses the channel response estimation module to obtain the estimated value and the spread spectrum sequence of the regenerated pilot symbol or training sequence in each time slot to generate the reference signal required by the weight vector generator.

The weight vector generator updates the weight vector of the array antenna in each time slot, and provides the weight vector updated in each time slot to the multiplier for beam weighting.

The autocorrelation matrix calculation module 301 of the present invention is similar to the structure of the correlation vector calculation module 401, and the operations they complete respectively are as follows:

The outputs of the extraction modules 302 and 402 are expressed in matrix form:

d(n)=[d(n)d(n+1)...(n+Sak′)]

Among them, K and K' are respectively the number of data samples of the digital signal generating module and the number of samples of the reference signal within the weight update time, and Sα is the sampling interval. First calculate in the autocorrelation matrix calculation module 301 and the correlation vector calculation module 401: $R_{\mathrm{xx}}^{\′}\left(\mathrm{no}\right)=\mathrm{\Σ}{x}^{\′}\left(\mathrm{no}\right)x^{\′\′}\left(\mathrm{no}\right)$ $R_{\mathrm{xx}}^{\′}\left(\mathrm{no}\right)=\mathrm{\Σ}\tilde{x}\left(\mathrm{no}\right)d^h\left(\mathrm{no}\right)---\left(7\right)$

Then use the following formula to recursively estimate the autocorrelation matrix:

R _xx (n)=αR _xx (n-1)+R′ _xx (n)

R _xd (n) = αR _xd (n-1) + R′ _xd (n) (8)

Where α is a positive real number less than 1, called the forgetting factor. According to the fading situation of the channel, α should take a value between 0.5 and 1. In the SMI method of the present invention, the usual numerically unstable matrix direct inversion operation is avoided and two triangular equations are solved instead. The matrix factorization module is identical for all multipaths for all channels, so it only needs to be calculated once. In addition, estimating the autocorrelation matrix Rxx generally requires a large number of samples, so the performance of the SMI method after despreading is very poor. The amount of computation for each weight estimation of the SMI method of the present invention is about 4SaKm ² +8SaK'm+4/3m ³ +12m ² +m times of floating point addition/multiplication. More importantly, each path only needs to be despread once, so the total calculation amount and hardware overhead of the system are relatively small.

The method for SMI digital beamforming of the present invention comprises the following steps (note that the signal data stream before despreading received by the array is X(n), and simultaneously utilize the estimated channel response and pilot symbols or training sequences to generate reference signal d(n ):

(1) Starting from the first time slot, the data before despreading and the reference signal received by the array of one time slot are read in sequence, forming the following matrix form:

d(n)=[d(n)d(n+1)…d(n+N′*SF) (10)

Where N represents the symbol length in a time slot, N' represents the length of the pilot symbol, and SF is the length of the spreading code used for spreading.

(2) Form an autocorrelation matrix according to the process shown in Figure 3 in the patent disclosure book: first, extract one every Sα from the data matrix in (1) to form a data matrix

K is the number of samples after extraction, and then calculate the autocorrelation matrix of the current time slot

_R'xx (n)=∑X'(n)X(n) (12)

_Rxx (n)＝ _αRxx (n-1)+R′xx(n) (13)

Where α is the forgetting factor, and R _xx (n-1) is the autocorrelation matrix calculated and saved in the previous time slot.

(3) Form the relevant vector according to the process shown in Fig. 3 in the patent disclosure book: firstly, the data matrix in (1) is extracted every Sα by column to form a data matrix,

At the same time, the reference signal is extracted at the same interval to obtain $\tilde{d}\left(\mathrm{no}\right)=\[d\left(\mathrm{no}\right)d(\mathrm{no}+1)\&Center\; Dot;\&Center\; Dot;\·d(\mathrm{no}+S^{\′}\mathrm{\α}{K}^{\′})\]---\left(15\right)$

K' is the number of samples after the extraction module in Figure 4 in the patent disclosure book, and then calculate the autocorrelation vector of the current time slot $R_{\mathrm{xd}}^{\′}\left(\mathrm{no}\right)=\mathrm{\Σ}\tilde{x}\left(\mathrm{no}\right){\tilde{d}}^h\left(\mathrm{no}\right)---\left(16\right)$

R _xd (n) = αR _xd (n-1) + R′ _xd (n) (17)

R _xd (n-1) is the correlation vector calculated in the previous time slot. Note that the sampling interval Sα' in (6), (7) is not necessarily the same as the sampling interval Sα in (3).

(4) Perform Cholesky decomposition R _xx (n)=L _xx (n)*L ^H _xx (n) on the autocorrelation matrix R _xx (n) to obtain a lower triangular matrix L _xx (n).

(5) Solve the lower triangular equations L _xx (n)*u=R _xd (n) and the upper triangular equations respectively with previous and back substitution methods $L_{\mathrm{xx}}^h\left(\mathrm{no}\right)W\left(\mathrm{no}\right)=u,$ In this way, the updated weight vector W(n) is obtained;

(6) Normalize the updated weight vector W(n) sent to the weighting module;

(7) Repeat steps (1) to (6) for the next time slot array weight vector update.

Using the above SMI digital beamforming method to realize array reception can effectively suppress interference and noise signals, and improve the capacity and communication quality of the wireless communication system. The system signal-to-noise ratio gain is close to the optimal value received by the array, effectively resisting the path fading effect, and the calculation amount is lower than the blind calculation methods such as the RLS method and code filtering, and is also significantly smaller than the traditional SMI method (assuming that the number of RAKE paths is 6, when When the sampling interval is 32, the present invention can reduce the amount of calculation to about 1/32 of the original), and the method can be completed by using the existing general-purpose DSP chip, which has strong realizability.

Fig. 5 is the uplink simulation result for WCDMA system. The multipath environment parameters are shown in Table 1:

Table 1 Simulation parameters multipath 6 Multipath delay (seconds) 0 310e-9 710e-9 1090e-9 1730e-9 2510e-9 Multipath average fading factor (dB) 0dB -1dB -9dB -10dB -15dB -20dB Multipath DOA Benchmark DOA130° DOA228° DOA333° DOA426° DOA555° DOA615° Spreading factor 256 mobile speed 120km/h

表示SMI算法，抽样间隔(1，1)； 表示SMI算法，抽样间隔(32，2)；

表示标准的SMI算法。In the picture: Indicates the single-antenna result;

Indicates the SMI algorithm, sampling interval (1, 1); Indicates the SMI algorithm, sampling interval (32, 2);

Indicates the standard SMI algorithm.

It can be seen from the above that the system performance is close to optimal when the present invention is adopted for a 4-element array receiver.

Claims (10)

1, a kind of method that is used for the up reception adaptive array of wireless communication system, the analog signal that every antenna element of receiving array receives is the array digital signal through analog/digital conversion, the despreading of array digital signal process obtains the array signal data flow after the despreading, utilize before the despreading and despreading after the array data signal form the weighing vector of array, and form wave beam with the array signal weighted sum of the weighing vector that generates after to despreading, signal behind the beam weighting is delivered to the corresponding footpath of Rake receiver, in time domain signal is carried out high specific by Rake receiver and merges; It is characterized in that described wave beam adopt to be optimized the invert weighted sum method of structure of sampling matrix by Beam-former and formed, and the steps include:

(1) from first time slot, digital beam forms data and pilot frequency sequence or training signal before the despreading that array received that device reads in current time slots successively arrives;

(2) ask average with the signal on array element of current time slots is relevant with pilot frequency sequence or training signal, estimate the channel response of current time slots;

(3) pilot frequency sequence of the channel response of the current time slots of usefulness step (2) estimation and current time slots or training signal generate the reference signal after channel response compensates;

(4) data matrix in the step (1) is extracted a formation sample less data matrix by row with certain sampling interval through data extraction module, utilize this data matrix to generate the Increment Matrix of the autocorrelation matrix of current time slots then, the autocorrelation matrix of previous time slot is introduced suitable forgetting factor and is accumulated the autocorrelation matrix that obtains current time slots mutually with Increment Matrix;

(5) data matrix in the step (1) is formed sample less data matrix by row with certain sampling interval through data extraction module, simultaneously the reference signal in (3) is extracted the reference signal that obtains less number of samples with same interval, utilize the data matrix of being correlated with after extracting to accumulate the incremental vector that obtains the current time slots associated vector mutually, introduce suitable forgetting factor and accumulate the auto-correlation vector that obtains current time slots mutually with the associated vector of previous time slot then with the incremental vector of associated vector with reference signal;

(6) associated vector of the autocorrelation matrix of applying step (4) and step (5) is decomposed by autocorrelation matrix being made Qiao Lisiji, and adopts the weight vector of the method renewal current time slots of simple former generation and back substitution.

(7) the weight vector normalization after will upgrading is delivered in the weighting block;

(8) renewal to next time slot array weight vectors repeats (1)～(7) step.

2. the method for the up receiving array of wireless communication system according to claim 1 is characterized in that, the sampling interval of extracting in the described step (5) can be the same with the sampling interval of extraction in the step (4), also can be different.

3. the method for the up receiving array of wireless communication system according to claim 1 is characterized in that, the span of the forgetting factor of introducing in step (4), (5) is: between 0.5～1.

4. the method for the up receiving array of wireless communication system according to claim 1 is characterized in that, described estimated channel response is meant by utilizing time-delay aligned signal and known emission frequency pilot sign/training sequence estimated channel decline response.

5. an aerial array receiver comprises that aerial array, array digital signal generation module, digital beam form module and Rake receiver; Described array digital signal generation module comprises receiving element and analog to digital converting unit; Aerial array input multipath signal is to array digital signal generation module, digital beam forms module and simultaneously the different multipath signals of a channel is formed wave beam respectively, multipath signal after wave beam forms is delivered to the corresponding footpath of Rake receiver, and Rake receiver carries out high specific to multipath signal and merges; It is characterized in that: described digital beam forms module and comprises Multipath searching unit, time-delay alignment unit, channel response estimation unit, reference signal generation unit, auto-correlation battle array estimation unit, weighing vector maker, multiplier, adder, despread unit and regeneration frequency pilot sign or training sequence unit; Array digital signal generation module becomes digital signal with the analog signal conversion that each antenna receives, be divided into two-way after digital beam forms module sending into: the one tunnel enters auto-correlation battle array estimation unit estimates the auto-correlation battle array, and sends into the weighing vector maker in all footpaths in all passages; Another road then enters Multipath searching and tracking module carries out Multipath searching and follows the tracks of processing the digital signal of each one or more antenna of channel usage, and multidiameter delay information is offered despread unit; Delay unit aligns the array received digital signal respectively by the time delay of appointment, m road signal after the alignment respectively is divided into two-way, one the tunnel enters the weighing vector maker, one the tunnel enters multiplier as being weighted data, prepare to carry out multiplying and finally after adder merges, obtain the array beams dateout with the weighing vector of weighing vector maker output, form the output in each footpath after this data process despread unit despreading, and be connected to the corresponding footpath of Rake receiver.

6. a kind of aerial array receiver according to claim 5 is characterized in that, described weighing vector maker comprises abstraction module, associated vector computing module, matrix triangle decomposition module reconciliation trigonometric equation pack module; The output of array digital signal generation module is sent in the autocorrelation matrix computing module after the process data extraction module extracts when sending into the time-delay alignment module; Equally, array digital signal and reference signal are sent in the associated vector computing module after extracting through data extraction module; The output of autocorrelation matrix computing module becomes two triangular matrixes through matrix triangle decomposition decomposition module, and the output of matrix triangle decomposition module and associated vector computing module is sent into simultaneously and separated trigonometric equation pack module generation weighing vector.

7. a kind of aerial array receiver according to claim 5 is characterized in that, described autocorrelation matrix estimation module comprises m data abstraction module and connected autocorrelation matrix computing module; The m road signal of array digital signal generation module output outputs to the autocorrelation matrix computing module by abstraction module respectively, outputs to the weighing vector maker again.

8. a kind of aerial array receiver according to claim 5, it is characterized in that, the digital signal of an array element of described channel response estimation module utilization time-delay alignment unit output is relevant in each time slot with known frequency pilot sign, obtains the rough estimate of different multipaths channel fading in each time slot of a channel.

9. a kind of aerial array receiver according to claim 5, it is characterized in that the spread spectrum sequence that described reference signal generation module utilizes the channel response estimation module to obtain estimated value and regeneration frequency pilot sign or training sequence produces the required reference signal of weighing vector maker in each time slot.

10. a kind of aerial array receiver according to claim 5 is characterized in that, described weighing vector maker upgrades the weight vector of array antenna at each time slot, and the weight vector that each time slot upgrades is offered multiplier carries out beam weighting.

Images