CN100433594C - Adaptive beamforming method in wideband code division multiple access system - Google Patents

Abstract

The present invention discloses a method and a device for forming self-adapting beams in a wideband code division multiple access system. The present invention provides self-adapting beams on the basis of approximate solution of minimum mean square error criterion; the self-adapting beams further realize channel estimation and compensation by scrambling and re-spreading pilot symbols so as to reduce operation amount, reduce a system structure and increase calculation speed; the self-adapting beams can trace user's quick travel in real time and can resist much interference. The method of the present invention comprises that spatial domain processing beams of array signals are generated; time domain matched filtering and pilot symbol descrambling and de-spreading assist Rake combination; reference signals are formed by re-spreading and scrambling feedback. The device of the present invention comprises a spatial domain beam forming module, a time domain matched filtering module and a re-spreading and scrambling feedback module. The device and the method of the present invention not only greatly reduce the calculation amount of algorithm but also increase calculation speed; simultaneously, negative influence brought about by a probable mistake in judgment is overcome to make weighted values more close to optimal weighted values; the system structure is simplified, the technical difficulty is reduced, and the project realization is simple.

Description

Adaptive beam in the broadband CDMA system forms apparatus and method

Technical field

The present invention relates to Wideband Code Division Multiple Access (WCDMA) (WCDMA) the system beam-forming technology of communication field, relate in particular to the adaptive beam formation method of smart antenna in the broadband CDMA system.

Technical background

Along with 3-G (Generation Three mobile communication system) developing rapidly in the world, also more and more obtained people's attention as the intelligent antenna technology of one of third generation mobile system core technology.Smart antenna mainly contains two kinds: adaptive beam antenna and switched-beam antenna, though in theory still on the engineering adaptive beam antenna implement all than the switched-beam antenna complex many, but, undoubtedly, the adaptive beam antenna is the best smart antenna of present performance, it can provide than the switched-beam antenna superior the performance of Duoing.In adaptive mode, each weights that corresponding spatial domain or sky, time domain are handled can be adjusted arbitrarily according to certain adaptive algorithm, and current transmission environment is carried out the maximum possible coupling, corresponding smart antenna received beam can be to point to arbitrarily.And each weights can only be selected from calculate good several class values in advance in switched-beam, the smart antenna mode of operation in a certain moment can only be selected from pre-designed several wave beams, not to point to arbitrarily, thereby only may carry out the part coupling to current transmission environment, say from point of theory not to be optimum.

Most crucial technology is exactly an adaptive beam formation method in the antenna system, and antenna system depends primarily on the performance of adaptive beam formation method to the improvement degree of communication system.There have been some adaptive beam formation methods at present, more influential conventional method is a cascade feedback wave beam formation method in cdma system, the user profile wave beam that this method arrives array received form with despreading after, carrying out Rake merges, user profile is adjudicated, with the information that rules out again after the spread spectrum as new reference signal, utilize minimum mean square error criterion to ask for best weight value then.But there is following shortcoming in this method:

(1) need matrix inversion, operand is big, is easy to generate the bigger error of inverting.This method is very high to the requirement of amount of calculation and computing time, and present processor hardware implements the bigger difficulty of existence, and cost is very high.

(2) when the user is in mobile status, when especially being in high-speed motion state, the arithmetic speed of self-adaptive processing is difficult to catch up with user's translational speed, thereby reduces the Communications service quality.

(3) this method is not considered the influence of scrambler, does not make full use of the advantage of frequency pilot sign yet, is not suitable for the practical application in the WCDMA system.

In order to adapt to wireless communications environment, various countries have compared many research to intelligent antenna technology.Both comprised and be applied to Frequency Division Duplexing (FDD) (Frequency Division Duplex, the intelligent antenna technology of TDMA wireless communication system FDD), as United States Patent (USP) U.S.5, the TDMA wireless communication system of the applying intelligent antenna that 887,262 (Smart antenna backwardscompatibility in digital cellular systems) introduce; U.S.6, the structure of the base station receiver of a kind of applying intelligent antenna that 2523548 (Transceiver arrangement for a smart antennasystem in a communication base station) introduce.Also have and much be applied to adopt time division duplex (Time Division Duplex, TDD) intelligent antenna technology in the cdma wireless communication system, as Chinese patent CN1339888A (receiver structure of applying intelligent antenna and joint-detection and algorithm thereof in the wireless communication system), related to the time-division synchronization CDMA wireless communication system of applying intelligent antenna.But above-mentioned patented technology does not all fundamentally solve the correlation technique of the problem that exists in the above-mentioned conventional method.

Summary of the invention

Core concept of the present invention is to propose a kind of approximate solution based on minimum mean square error criterion, avoids matrix inversion; Utilize decision-feedback, heavily expand the adjustment that reference signal that scrambling obtains is controlled the self adaptation weight vector; And by frequency pilot sign realization channel estimating and compensation; Thereby reduce operand and simplied system structure, improve arithmetic speed, fast moving that can the real-time tracking user and can resist the adaptive beam formation method of a plurality of interference, the operand that fundamentally overcomes prior art is big, realize difficulty height, the shortcoming that speed is slow and system configuration is complicated.

It is to realize with following technical scheme that cascade feedback adaptive wave beam formation method is heavily expanded in the auxiliary despreading of the pilot bit of the WCDMA of being applicable to system up-link provided by the invention:

For achieving the above object, the adaptive beam that the present invention has constructed in a kind of broadband CDMA system forms device, it is characterized in that, comprise three modules: the spatial domain wave beam forms module (21), time domain matched filtering module (22) and heavily expands scrambling feedback module (23).

Described spatial domain wave beam forms module (21) and further comprises multiplier (211), adder (212) and weights estimation module (213);

Described time domain matched filtering module (22) comprises that further descrambling module (221), despreading module (222), the auxiliary Rake of frequency pilot sign merge module (223), data bit judging module (224), Signal to Interference plus Noise Ratio calculating and judging module (225);

Described heavy expansion scrambling feedback module (23) further comprises spread spectrum module (231) and scrambling module (232);

Described spatial domain wave beam forms module (21) and utilizes the output of weights estimation module (213) to carry out wave beam formation by multiplier (211) and adder (212) to the array signal that receives.When carrying out spatial domain wave beam formation, utilize and heavily expand scrambling feedback module (23) output formation reference signal (233), and and new array acknowledge(ment) signal obtain cross-correlation matrix, utilize approximate solution that cross-correlation matrix obtains according to general minimum mean square error criterion as new weights then.Data flow after wave beam forms becomes data bit through descrambling module (221) despreading module (222), then utilize frequency pilot sign to carry out channel estimating and compensation and Rake and merge (223), and the data bit adjudicated (224), calculate Signal to Interference plus Noise Ratio then and carry out Signal to Interference plus Noise Ratio threshold judgement (225), carry out subsequent treatment for the data bit that meets the demands; Otherwise, give up this data.At last to the data bit and known pilot symbols spread spectrum (231) scrambling (232) of judgement output, and with this stream of chips as new reference signal (233).

The present invention also provides the formation method of the adaptive beam in a kind of broadband CDMA system, it is characterized in that, may further comprise the steps:

Step 1: as the reference signal, ask weights with known Dedicated Control Channel frequency pilot sign;

Step 2: the signal to array received carries out the spatial domain processing;

Step 3: descrambling and de-spreading judgement output;

Step 4: estimate the Signal to Interference plus Noise Ratio of judgement back data bit, and compare with the thresholding of setting;

Step 5:, heavily expand after the scrambling as new reference signal, and ask for the optimization weights to known frequency pilot sign and the non-frequency pilot sign that rules out;

Step 6: utilize and optimize weights, handle dedicated data channel information.

Described step 1 further comprises following process:

(1) with the signal after the pilot signal spectrum-spreading and scrambling as reference signal d, determine the pilot signal X of array received _pCross-correlation matrix r with reference signal d _Xd=E[X _pd ^*], wherein * represents conjugate operation;

(2) approximate solution of obtaining according to minimum mean square error criterion (is r _Xd=E[X _pd ^*]) as new adaptive weight W=r _Xd

Described step 2 further comprises following process:

(1) the pilot tone time period signal of array received is done wave beam and form Y _p=W ^HX _p

(2) to the non-pilot tone time period signal X of array received _tDo wave beam and form Y _t=W ^HX _t, wherein H represents the transposition computing.

Described step 3 further comprises following process:

(1) to Dedicated Control Channel signal descrambling $P_s\left(k\right)=Y_p\left(k\right)S_{\mathrm{dpch}}^{*}\left(k\right);$

(2) to the despreading of Dedicated Control Channel signal $Z\left(k\right)=\frac{1}{\mathrm{SF}}\underset{i=(k-1)\·\mathrm{SF}+1}{\overset{k\·\mathrm{SF}}{\mathrm{\Σ}}}{P}_s\left(i\right)c_{\mathrm{cch}}\left(i\right),$ Which information symbol k represents in the formula, and SF is the spreading factor of Dedicated Control Channel, S _Dpch() is scrambler, c _Cch() is the spreading code of Dedicated Control Channel;

(3) estimate the complex gain of the l bar multipath that Rake merges with the result of Dedicated Control Channel signal descrambling and de-spreading $G_c\left(l\right)=\frac{1}{q}\underset{k=1}{\overset{q}{\mathrm{\Σ}}}Z\left(k\right),$ Q is the pilot tone figure place in the formula;

(4) carrying out channel compensation and the user sent control information adjudicates $d_c=\mathrm{sign}\left(\mathrm{imag}\left(Y_t\right)G_c^{*}\left(l\right)\right),$ Sign represents the symbol judgement computing in the formula, and imag represents to get imaginary-part operation.

Described step 4 further comprises following process:

(1) calculates Signal to Interference plus Noise Ratio and carry out the Signal to Interference plus Noise Ratio threshold judgement;

(2) if Signal to Interference plus Noise Ratio meets the demands, the data bit after then will adjudicating continues following step;

(3) if Signal to Interference plus Noise Ratio does not meet the demands, then give up this data, return step 1.

Described step 5 further comprises following process:

(1) utilize frequency pilot sign to replace the court verdict symbol heavily to expand scrambling in the time in pilot tone, in non-pilot tone in the time, with the control information spread spectrum d that rules out _d=d _cC _Cch, with control information scrambling d _s=d _dS _Dpch

(2) form new reference signal d ₁=[d d _s], new array received signal X ₁=[X _pX _t], ask the cross-correlation matrix of two signals $r_{\mathrm{xd}}=E\left[X_1{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="C0213688300141.png"\; state="\{\{state\}\}">d</figure-callout>}}_1^{*}\right];$

(3) approximate solution of obtaining according to minimum mean square error criterion is as optimizing weights W _Opt=r _XdDescribed step 6 further comprises following process:

(1) weights that utilize Dedicated Control Channel to try to achieve carry out respective handling to dedicated data channel;

(2) multipath is carried out Rake and merge, obtain the data bit information of desired user.

The present invention has a lot of advantages owing to having adopted the auxiliary despreading of pilot bit heavily to expand the adaptive beam formation method of cascade feedback, having compared with existing adaptive beam formation method.Separate because adopt batch processing to calculate and substitute Wei Na, avoid that loaded down with trivial details iterative computation and large-scale matrix multiplies each other, matrix inversion, thereby reduced the amount of calculation of algorithm greatly with the approximate solution of least mean-square error.In addition, the present invention utilizes frequency pilot sign to replace the court verdict symbol heavily to expand, and has overcome the negative influence that possible mistake in judgment brings, and makes weights more approach best weight value; Also realize channel estimating and compensation by frequency pilot sign, avoided the influence of channel to reference signal, the variation that makes algorithm can effectively resist channel obtains accurate more result, has simplified system configuration simultaneously.And method and apparatus of the present invention can be followed the tracks of user's direction of arrival, adaptive adjusting weight vector; And can form the maximum sensing of wave beam at the direction of arrival of desired user, and form zero in the interference user direction and fall into, effectively suppress the influence of interference, improve the performance of desired signal.

Description of drawings

Fig. 1 method flow diagram of the present invention;

Cascade feedback adaptive beamforming algorithm structure chart is heavily expanded in the auxiliary despreading of Fig. 2 pilot bit of the present invention.

Embodiment

The present invention is described in further detail below in conjunction with accompanying drawing.

Fig. 1 is a method flow diagram of the present invention.In the adaptive mode smart antenna, the weights that corresponding spatial domain or sky, time domain are handled can be adjusted arbitrarily according to certain adaptive algorithm, current transmission environment is carried out the maximum possible coupling, corresponding smart antenna received beam can be to point to arbitrarily.Because algorithm is usually very complicated, implement comparatively difficulty in practice.Therefore, optimization system structure, minimizing algorithm computation amount are starting points of the present invention.Method set forth in the present invention can realize by following several steps:

The first step (102) as the reference signal, is asked weights with known Dedicated Control Channel frequency pilot sign; Base station side in the WCDMA system, user's scrambler, spreading code and frequency pilot sign all are known.In general, the multidiameter delay of signal can be obtained by the Rake receiver, and therefore the hypothesis multidiameter delay is known in the process below, does not consider time delay influence under the prerequisite that does not influence algorithm; Signal after the pilot signal spectrum-spreading and scrambling as reference signal d, is asked the pilot signal X of array received _pCross-correlation matrix r with reference signal d _Xd=E[X _pd ^*]; The approximate solution of obtaining according to minimum mean square error criterion is as adaptive weight W=r _Xd

Second step (103), the signal of array received to be carried out the spatial domain handle, the multipath information that obtains according to search is carried out wave beam respectively to each big time-delay multipath component of user and is formed, and the pilot tone time period signal of array received is done wave beam form Y _p=W ^HX _p, the non-pilot tone time period signal of array received is done wave beam forms Y _t=W ^HX _t, then to the pilot descrambling of control section $P_s\left(k\right)=Y_p\left(k\right)S_{\mathrm{dpch}}^{*}\left(k\right),$ To the pilot despreading $Z\left(k\right)=\frac{1}{\mathrm{SF}}\underset{i=(k-1)*\mathrm{SF}+1}{\overset{k*\mathrm{SF}}{\mathrm{\&Sigma;}}}{P}_s\left(i\right)c_{\mathrm{cch}}\left(i\right),$ In DPCH, control section and data division are by code division multiplexed transmission, and when transmitting as if many yards that do not consider DPCH, wherein the spreading factor of control section is fixed as SF=256;

The 3rd step (104), descrambling and de-spreading judgement output, result with pilot information descrambling and de-spreading estimates that (so-called Rake merging just is to use the correlation receiver group in the Rake merging, a correlation receiver is used in each path, each correlation receiver is relevant with a delay form of same desired signal, the output of these correlation receivers is weighted according to their relative intensity then, and each the road output addition after the weighting, synthetic output, the selection principle of weight coefficient are that to make output signal-to-noise ratio be maximum.) the complex gain of l bar multipath $G_c\left(l\right)=\frac{1}{q}\underset{k=1}{\overset{q}{\mathrm{\&Sigma;}}}Z\left(k\right),$ The non-pilot information of control section is adjudicated $d_c=\mathrm{sign}\left(\mathrm{imag}\left(Y_t\right)G_c^{*}\left(l\right)\right);$

In the 4th step (105), ask the Signal to Interference plus Noise Ratio of adjudicating the back data bit, and compare with the thresholding of setting, if Signal to Interference plus Noise Ratio is less than set point, the possibility that error code occurs is bigger, and this can bring negative effect to right value update, therefore give up this data, do not carry out right value update.If Signal to Interference plus Noise Ratio greater than set point, then continues follow-up flow process;

The 5th step (106), heavily expand scrambling and form new weights, the front has 3 to 8 to be known frequency pilot sign in 10 sign bits of the every time slot of control section, utilize frequency pilot sign to replace the court verdict symbol heavily to expand scrambling in time at pilot, thereby overcome the negative influence that mistake in judgment brings, make weights more approach best weight value.At non-pilot in the time, with the non-pilot information spread spectrum d that rules out _d=d _cC _Cch, scrambling d again _s=d _dS _DpchForm new reference signal d then ₁=[d d _s].Because control section is always used ovsf code c _Cch=C _{Ch, 256,0}Spread spectrum, 256 data of this spread spectrum code sequence are 1 entirely, this computational complexity that algorithm is heavily expanded in reduction despreading is very favourable.At last, utilize new reference signal and new array received signal X ₁=[X _pX _t], and calculate correlation matrix $r_{\mathrm{xd}}=E\left[X_1{\mathrm{<figure-callout\; id="1"\; label="d"\; filenames="C0213688300141.png"\; state="\{\{state\}\}">d</figure-callout>}}_1^{*}\right],$ Try to achieve optimization weights W according to minimum mean square error criterion _Opt=r _Xd

The 6th step (107), dedicated data channel is handled, because the airborne spread path of control section and data division is identical, it is suitable equally to upgrade the weights data portion that obtains by control section like this, so try to achieve corresponding data information according to the weights that obtain.

Fig. 2 is that cascade feedback algorithm structure chart is heavily expanded in the auxiliary despreading of pilot bit provided by the present invention.Base station side in the WCDMA system, user's scrambler, spreading code and frequency pilot sign all are known.In general, the multidiameter delay of signal can be obtained by the Rake receiver, supposes that here multidiameter delay is known, does not consider time delay influence under the prerequisite that does not influence algorithm.In addition, signal is carried out discretization in the baseband-converted module, subsequent treatment all is a Digital Signal Processing.According to the 3GPP agreement, in three kinds of up channels (Random Access Channel, DPCH, CPCH) control section and data division are arranged all, they pass through code division multiplexed transmission, wherein the spreading factor of control section is fixed as bigger by 256, and the spreading factor of data division gets 4,8,16 by certain rule,, the number in 256.Generally control section can provide reasonable spreading gain, therefore utilize the data bit of control section to carry out despreading and heavily expand reasonable algorithm robustness can be provided, the auxiliary despreading of pilot bit is heavily expanded the cascade feedback algorithm and is just utilized the control section data message to carry out right value update behind descrambling and de-spreading.In the method, at first each the big time-delay multipath component to user m carries out wave beam formation respectively in spatial domain wave beam formation module (21), and the reception weight vector of establishing the l bar multipath component of user m is $w_{m,l}={[w_{m,l}^{\left(1\right)},w_{m,l}^{\left(2\right)},...,w_{m,l}^{\left(N\right)}]}^T,$ Be the output of weights estimation module (213), again in conjunction with multiplier (211), adder (212), then the wave beam of the l bar multipath of user m forms and is output as:

$y_{m,l}\left(k\right)=\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}{x}^{\left(j\right)}\left(k\right)w_{m,l}^{{\left(j\right)}^{*}}\left(k\right)$

In time domain matched filtering module (22), the wave beam of the l bar multipath of user m is formed output carries out descrambling (221) despreading (222):

$z_{m,l}\left(k\right)=\frac{1}{\mathrm{SF}}\underset{n=(k-1)\&CenterDot;\mathrm{SF}+1}{\overset{k\&CenterDot;\mathrm{SF}}{\mathrm{\&Sigma;}}}{y}_{m,l}\left(n\right)S_{\mathrm{dpch}}^{*}\left(n\right)c_m\left(n\right)$

Auxiliary Rake merges in the module (223) at frequency pilot sign, utilize pilot signal to estimate the complex amplitude of every multipath, suppose to contain in the time slot q frequency pilot sign, at this moment the multipath signal power of desired user is far longer than interference plus noise, adopt frequency pilot sign to obtain the simple average estimation technique that the non-pilot data channel of same time slot is estimated, the channel estimating of the l bar multipath component of user m is:

$g_{m,l}=\frac{1}{q}\underset{k=1}{\overset{q}{\mathrm{\&Sigma;}}}{z}_{m,l}\left(k\right)$

At last according to the channel estimating of each multipath component by the Rake merging that be concerned with of high specific criterion, can obtain the output of user m:

$b_m\left(n\right)=\mathrm{sign}\left(\underset{k=1}{\overset{q}{\mathrm{\&Sigma;}}}\mathrm{imag}\left(z_{m,l}\left(k\right)\right)g_{m,l}^{*}\right)$

Because the airborne spread path of control section and data division is identical, it is suitable equally to upgrade the weights data portion that obtains like this.The front has 3 to 8 to be known frequency pilot sign in 10 sign bits of the every time slot of control section, can utilize frequency pilot sign to replace the court verdict symbol heavily to expand in time at pilot, can overcome the negative influence that mistake in judgment brings like this, make weights more approach best weight value.At non-pilot in the time, if n the bit b (n) of certain user's control section by correct detection, b (n) is the output of detector here, this user is at time interval [(n-1) T so _b, nT _b] (T _bBe bit period, n is a positive integer) signal waveform can be by with this user's control section spreading code c _Cch(k) heavily expanding detected data bit b (n) obtains.

Data bit after the judgement is input to and heavily expands scrambling feedback module (23), to the data bit after the judgement and known frequency pilot sign again spread spectrum (231) scrambling (232) in order to form new reference signal (233), and this reference signal is input to the spatial domain wave beam forms module (21), thereby the weighted value of adjusting this user is carried out beam shaping.

In a word, adopt method provided by the invention to realize that adaptive beam forms, not only reduced the amount of calculation of algorithm greatly, improved arithmetic speed, overcome the negative influence that possible mistake in judgment brings simultaneously, made weights more approach best weight value, simplified system configuration, reduce technical difficulty, be easy to Project Realization.

Claims (2)

1, the adaptive beam formation method in a kind of broadband CDMA system is characterized in that, may further comprise the steps:

Step 1: generate array signal spatial domain processing-wave beam;

Comprise the steps:

(1) with the signal after the pilot signal spectrum-spreading and scrambling as reference signal d, determine the pilot signal X of array received _pCross-correlation matrix r with reference signal d _Xd=E[X _pd ^*], wherein ^*Represent conjugate operation;

(2) the approximate solution r that obtains according to minimum mean square error criterion _Xd=E[X _pd ^*] as new adaptive weight W=R _Xd

(3) the pilot tone time period signal of array received is done wave beam and form Y _p=W ^HX _p, to the non-pilot tone time period signal X of array received _tDo wave beam and form Y _t=W ^HX _t, wherein H represents the transposition computing;

Step 2: the auxiliary Rake of time domain matched filtering-descrambling and de-spreading and frequency pilot sign merges;

Comprise the steps:

(1) to the pilot tone descrambling $P_s\left(k\right)=Y_p\left(k\right)S_{\mathrm{dpch}}^{*}\left(k\right),$ To pilot despread $Z\left(k\right)=\frac{1}{\mathrm{SF}}\underset{i=(k-1)\&CenterDot;\mathrm{SF}+1}{\overset{k\&CenterDot;\mathrm{SF}}{\mathrm{\&Sigma;}}}{P}_s\left(i\right)c_{\mathrm{cch}}\left(i\right),$ Which information symbol k represents in the formula, and SF is the spreading factor of Dedicated Control Channel, S _Dpch() is scrambler, c _Cch() is the spreading code of Dedicated Control Channel;

(2) estimate the complex gain of the l bar multipath that Rake merges with the result of pilot tone descrambling and de-spreading $G_c\left(l\right)=\frac{1}{q}\underset{k=1}{\overset{q}{\mathrm{\&Sigma;}}}Z\left(k\right),$ Q is the pilot tone figure place in the formula;

(3) carry out channel compensation and the control information that the user sends adjudicated $d_c=\mathrm{sign}\left(\mathrm{imag}\left(Y_t\right)G_c^{*}\left(l\right)\right),$ Sign represents the symbol judgement computing in the formula, and imag represents to get imaginary-part operation;

(4) calculate Signal to Interference plus Noise Ratio and carry out the Signal to Interference plus Noise Ratio threshold judgement, if dried the making an uproar of letter meets the demands, the data bit after then will adjudicating is input to and heavily expands the scrambling feedback module, otherwise gives up this data;

Step 3: heavily expand the scrambling feedback and form reference signal;

Comprise the steps:

(1) utilize frequency pilot sign to replace the court verdict symbol heavily to expand scrambling in the time in pilot tone, in non-pilot tone in the time, with the control information spread spectrum d that rules out _d=d _cC _Cch, with control information scrambling d _s=d _dS _Dpch(k);

(2) form new reference signal d ₁=[d d _s], new array received signal X ₁=[X _pX _t], ask the cross-correlation matrix of two signals $r_{\mathrm{xd}}=E\left[X_1{d}_1^{*}\right];$

(3) approximate solution of obtaining according to minimum mean square error criterion is as optimizing adaptive weight W _Opt=r _Xd

(4) weights that utilize Dedicated Control Channel to try to achieve carry out respective handling to dedicated data channel.

2, the adaptive beam in a kind of broadband CDMA system forms device, it is characterized in that, comprises three modules: the spatial domain wave beam forms module (21), time domain matched filtering module (22) and heavily expands scrambling feedback module (23);

Described spatial domain wave beam forms module (21) and further comprises multiplier (211), adder (212) and weights estimation module (213);

Described time domain matched filtering module (22) comprises that further descrambling module (221), despreading module (222), the auxiliary Rake of frequency pilot sign merge module (223), data bit judging module (224), Signal to Interference plus Noise Ratio calculating and judging module (225);

Described heavy expansion scrambling feedback module (23) further comprises spread spectrum module (231) and scrambling module (232);

Described spatial domain wave beam forms module (21) and utilizes the output of weights estimation module (213) to carry out wave beam formation by multiplier (211) and adder (212) to the array signal that receives; When carrying out spatial domain wave beam formation, utilize and heavily expand scrambling feedback module (23) output formation reference signal (233), and and new array acknowledge(ment) signal obtain cross-correlation matrix, utilize approximate solution that cross-correlation matrix obtains according to general minimum mean square error criterion as new weights then; Data flow after wave beam forms becomes data bit through descrambling module (221) and despreading module (222), then merge module (223) and carry out channel estimating and compensation and Rake merging by the auxiliary Rake of frequency pilot sign, and utilize data bit judging module (224) that the data bit is adjudicated, calculate Signal to Interference plus Noise Ratio and carry out the Signal to Interference plus Noise Ratio threshold judgement by Signal to Interference plus Noise Ratio calculating and judging module (225) then, carry out subsequent treatment for the data bit that meets the demands; Otherwise, give up this data; Data bit and the known pilot symbols spread spectrum and the scrambling of judgement being exported at last by spread spectrum module (231) and scrambling module (232), and with this stream of chips as new reference signal (233).

Images