CN1131614C - Ruike and equalization cascade receiving method under the code division multiple address low band-spreading ratio and its equipment - Google Patents

Abstract

The present invention relates to a cascade receiving method and equipment of Ruike and equalization under the code-division multi-address low frequency-spreading ratio, which is a honeycomb CDMA receiving method and equipment based on the mutual cascade of Ruike receiving and soft output equalization (SOVA). The cascade method and the equipment can merge and receive effective multi-path signals, and can balance out intersymbol interference caused by the multi-path signals. The present invention is especially suitable for the high-speed data transmission of CDMA systems under the low frequency ratio. The equipment essentially comprises a Ruike receiver and a soft output equalizer; firstly, parameters of a multi-path fading channel are estimated and received by the Ruike; then, the intersymbol interference ISI coefficient of equivalent time change is calculated according to the channel parameters and effective path positions; finally, a soft output viterbi equalization algorithm is carried out. Computer simulation and test result indicate that the method and the equipment can improve obviously improve the performance of high-speed data transmission of a third generation mobile communication system under the condition of only increasing a small amount of calculation.

Description

Rake and equalization cascade receiving method and device under the code division multiple address low band-spreading ratio

One, technical field:

The invention belongs to and realize in CDMA (code division multiple access) cellular communication system that rake receives and the field of the technology of soft output viterbi equalization algorithm.

Two, background technology:

The cdma cellular communication technology is simple with its frequency planning, power system capacity is big, ability of anti-multipath is strong, good communication quality, electromagnetic interference are little etc., and characteristics demonstrate huge development potentiality, are the mainstream technologys of following 3G (Third Generation) Moblie.The CDMA spread spectrum signal receiver is divided into two kinds of coherent receiver and noncoherent receivers.Coherent receiver need be known the phase information of received signal, and noncoherent receiver does not need to know the phase information of received signal, is the quadrature modulation mode but require to send signal.There is the multipath fading phenomenon in the mobile communication system, can causes serious multipath to disturb.In having adopted the cdma cellular mobile communication system of spread spectrum technique, have pilot tone (Pilot) signal of the information of knowing by reception, can estimate the amplitude and the phase information of multipath signal, thereby make rake and coherent reception become possibility.The relevant band spread receiver that carries out the diversity processing at multipath fading signal is called Rake (RAKE) receiver, it can carry out phasing and carry out high specific merging and handling to a plurality of identical information and separate single drive signals of fading characteristic of carrying, overcome multipath fading thereby reach, improve the purpose of received signal interference ratio.

When adopting the RAKE reception programme, because the existence of multipath has caused intersymbol interference.When spreading ratio was big, this interference was compared less with the energy of receiving symbol, can ignore.But,, must use the spreading ratio of less (low) in order to support higher data speed.At this moment, intersymbol interference has had a strong impact on the performance of RAKE receiver, even under the signal to noise ratio condition with higher, RAKE receiver also can't reach desirable effect.The conventional method that solves intersymbol interference is a balancing technique, as the channel equalization algorithm in the gsm system etc.Can how to use balancing technique to offset intersymbol interference in spread spectrum system be the key that support the high data rate business.

Three, summary of the invention:

(1) goal of the invention

The objective of the invention is to propose a kind of multipath signal that both can merge reception, can offset each footpath intersymbol interference in signal again, thus a kind of code division multiple address low band-spreading that can effectively overcome the system performance degradation that traditional RAKE receiver brought when having intersymbol interference than under Rake and soft output Viterbi algorithm's equalization cascade receiving method and device.

(2) technical scheme

Rake and soft output Viterbi algorithm's equalization cascade receiving method under a kind of code division multiple address low band-spreading ratio, at first estimate channel parameter, secondly carrying out Rake to received signal receives, utilize the local spread spectrum code sequence that generates in the position of each the footpath channel parameter estimate and effective diameter and the receiver then, the intersymbol interference ISI coefficient that becomes when calculating equivalence is realized soft output viterbi equalization algorithm at last.

Estimate that channel parameter is to utilize existing channel parameter estimation method,, multipath channel is estimated more accurately, obtain the parameters such as time delay, phase place and amplitude in each footpath, and from each receives directly, select some effective diameters assisting down of sequence of pilot symbols; Carrying out the Rake reception to received signal is to carry out the effective diameter signal of Rake reception to received signal to carry out the high specific merging, realizes that Rake receives the symbol sebolic addressing that obtains receiving.

Realize that output viterbi equalization algorithm is to utilize the time output symbol sequence that becomes equivalent ISI coefficient and remove the Rake receiver after sequence of pilot symbols disturbs that calculates, carry out soft output viterbi equalization algorithm, export one group of symbol probability sequence that value is relevant with log-likelihood metrics, offer subsequent communication channel decoding.

Whole device is made up of Rake receiver and soft output viterbi equalizer two parts, Rake receiver and the input of input termination baseband sampling signal, the input of the soft output viterbi equalizer of output termination of Rake receiver; Rake receiver comprises the channel parameter estimator module, effectively arrive the footpath selector module, Rake demodulation and high specific merge module, local frequency expansion sequence generator module, when comprising equivalence, soft output viterbi equalizer becomes the ISI coefficients calculation block, the pilot channel interference cancellation module, soft output viterbi equalization module, channel parameter estimation module wherein, effectively arrive the footpath selector module, Rake demodulation and high specific merge module, the pilot channel interference cancellation module, soft output viterbi equalization module order successively is connected in series, the input that becomes the ISI coefficients calculation block during equivalence connects the channel parameter estimator module respectively, effectively arrive the footpath selector module, local frequency expansion sequence generator module becomes ISI coefficients calculation block and soft output viterbi equalization module during equivalence.

(3) technique effect

Because during low spreading factor, intersymbol interference only continues 2 to 5 mark spaces usually, the receiver apparatus based on above-mentioned principle that the present invention proposes is simple, and the RAKE receiver that amount of calculation is more traditional increases few, is very suitable for the high speed data transfer of 3-G (Generation Three mobile communication system).

RAKE proposed by the invention and soft output viterbi equalization are united the algorithm of reception, are applicable under the situation that spreading ratio is lower in the cdma system, utilize the RAKE reception technique to merge multipath signal, realize receive diversity; Utilize soft output viterbi equalization algorithm to offset to merge the intersymbol interference in each footpath simultaneously, thereby reach the purpose that improves receptivity.

Major advantage of the present invention is to calculate simple, can in traditional RAKE receiver, add three independently module finish, and can be used for any CDMA mobile communication system with continuous pilot, comprise 3GPP WCDMA and 3GPP2 cdma2000.When low spreading factor, the influence of intersymbol interference be can not ignore often, and the present invention can overcome traditional RAKE receiver effectively and produced under this condition system performance degradation satisfies the requirement of high-speed data service.

Four, description of drawings:

Fig. 1 is the device general diagram of RAKE and soft output viterbi equalization associating method of reseptance.Soft output viterbi equalizer B is wherein arranged, it comprises channel parameter estimator module B100, effectively arrives footpath selector module B101, Rake demodulation and high specific merging module B102, local frequency expansion sequence generator module B103, becomes ISI coefficients calculation block A100, pilot channel interference cancellation module A101, soft output viterbi equalization modules A 102, baseband sampling signal input IN when Rake receiver A comprises equivalence.

When being equivalence, Fig. 2 becomes ISI coefficient calculations schematic diagram.Comprising definite effective path position module C100, calculate each effective diameter ISI coefficient module C101, merge each effective diameter ISI coefficient module C102, calculate each effective diameter ISI coefficient module DC103, merge each effective diameter ISI coefficient module DC104.

Fig. 3 is the schematic illustration that becomes the ISI coefficient calculations in first footpath pairing when equivalence.

Fig. 4 is the performance illustration relatively of traditional RAKE receiver and RAKE proposed by the invention and soft output Viterbi balancing combine receiver.

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

Five, embodiment

Below to the invention in each composition discussed respectively.

1, channel parameter estimator module (B100):

Pilot tone in the cdma system (Pilot) channel is used to transmit the pilot frequency sequence that knows in advance, can be used for system regularly and the extraction of carrier wave, channel estimating, handover etc.If system launches the signal of several code channels that are shown below simultaneously, $x\left(t\right)=\underset{i}{\mathrm{\Σ}}\underset{n=-\∞}{\overset{+\∞}{\mathrm{\Σ}}}{d}_i\left(n\right)w_i(t-{\mathrm{nT}}_{s,i})\mathrm{pn}\left(t\right)\≡\underset{i}{\mathrm{\Σ}}{x}_i\left(t\right)$ [formula 1a] wherein, the i item is corresponding to the equivalent baseband signal of i Code Channel; d _i(n) be i the symbol sebolic addressing that code channel sent; w _i(t) be i the quadrature Walsh sequence that is numbered i that code channel is specified, it is worth only at 0＜t≤T _sGet nonzero value in the interval; Pn (t) is the pseudorandom spread spectrum sequence; T _{S, i}Sent mark space by i code channel.The subitem x of i=0 ₀(t) corresponding to the Pilot channel.Being without loss of generality, can supposing that the symbol that pilot frequency sequence sends is constantly equal to 1, also is d ₀And w (n)=1, ₀(t)=1,0＜t≤T _sThrough behind the multidiameter fading channel, the equivalent baseband receiving signals of receiving terminal can be expressed as: $r\left(t\right)=\underset{l}{\mathrm{\Σ}}{c}_l\·x(t-{\mathrm{lT}}_c)+z\left(t\right)$ T in [formula 1b] formula _cBe chip-spaced; Z (t) is the plural additive white Gaussian of zero-mean; c _lFading factor for channel l footpath; The purpose of channel parameter estimation is according to received signal r (t) and the pilot frequency sequence x that knows ₀(t) estimate channel fading factor c _l

Suppose that mobile channel is a frequency selectivity slow fading channel model, then can think c in a channel estimating interval _lBe approximately constant.Can draw c thus _lEstimated value as follows: ${\stackrel{\‾}{c}}_l=\frac{1}{{\mathrm{NE}}_c}{\∫}_0^{{\mathrm{NT}}_c}r(t+l{T}_c)\·x_0^{*}\left(t\right)\mathrm{dt}=c_l+N_a+N_c+N_z$ N in [formula 2] formula _a, N _cAnd N _zBe that the not ideal enough multipath that causes of the correlation properties of spreading code disturbs, multiple access disturbs respectively and output that white noise produces after by correlator; T _cBe the time width of a chip, NT _cIntegrating range for channel estimating; E _cBe the transmission energy of pilot channel within a chip.

2, effectively arrive footpath selector module (B101)

In actual applications, the arrival of not all signal footpath all is effective.Should set suitable thresholding for this reason, the energy of each footpath signal in the window is adjudicated.If the channel estimating intensity level that is positioned on the same path position is double greater than thresholding, then be effective arriving signal footpath; Otherwise then be pure interference footpath (IOP).For avoiding mis-behave, computing all should not participated in all pure interference footpaths.Decision threshold choose the side lobe levels that should be slightly larger than pilot signal (PN sign indicating number) part cross-correlation (Partial Correlation) value.

3, RAKE demodulation and high specific merge module (B102):

After the channel parameter estimation value that has obtained each footpath, can carry out coherent demodulation to other code channel of bearer data transmission, for these need carry out despreading respectively with each footpath of other code channel, and the channel estimating parameter c of utilizing [formula 2] to obtain _lEach footpath despreading result is carried out amplitude weighting and phasing, enable with being combined to.This process is referred to as high specific and merges, and specifically can be described by following formula: $y_i\left(m\right)=\underset{l}{\mathrm{\Σ}}{\stackrel{\‾}{c}}_l^{*}{\∫}_{{\mathrm{mT}}_s}^{{(m+1)T}_s}r(t+{\mathrm{lT}}_c)\·w_i(t-{\mathrm{mT}}_s){\mathrm{pn}}^{*}\left(t\right)\mathrm{dt}$ Y in [formula 3] formula _i(m) be i circuit-switched data carrying m the sign estimation value that code channel transmitted; T _sLasting interval for these data; c _l ^*Be c _lConjugate operation.In the practical application, be not RAKE receiver can differentiate each the footpath on the useful signal component is all arranged, need be to c _lCarry out threshold judgement, and only need effective footpath that arrives is selected that the selected multipath component that is higher than thresholding that goes out merges in the module (B101).

4, local frequency expansion sequence generator module (B103)

For realizing the coherent reception of band spread receiver, need to go out the frequency expansion sequence synchronous with received signal at local recovery.Need to utilize the local pilot frequency sequence that takes place to realize the frequency expansion sequence initial synchronisation for this reason, realize the synchronously thick of local pilot frequency sequence, utilize door track loop sooner or later then or realize the synchronously thin of local pilot frequency sequence based on the track loop of multipath energy window gravity center.The generation of the local frequency expansion sequence of other code channel generally can obtain by compound (XOR) computing of quadrature Walsh sequence and local pilot frequency sequence.

5, when equivalence becomes ISI coefficients calculation block (A100)

When multidiameter delay expansion and transmission symbol interval T _{S, i}When being close, intersymbol interference is calculated the intersymbol interference ISI that is brought by the multidiameter delay expansion earlier with the appreciable impact receiver performance for this reason in RAKE output, consider to adopt equalization algorithm then, eliminates the mis-behave that is brought of ISI.Can derive the following expression formula of RAKE output by [formula 1] to [formula 3]: $y_i\left(m\right)\≡y_i^s\left(m\right)+y_i^p\left(m\right)+y_i^q\left(m\right)$ $=\underset{n=-\∞}{\overset{+\∞}{\mathrm{\Σ}}}{d}_i\left(n\right)\·f_{m-n}^{i,j}(k-l)+\underset{n=-\∞}{\overset{+\∞}{\mathrm{\Σ}}}{f}_{m-n}^{i,0}(k-l)+\underset{n=-\∞}{\overset{+\∞}{\mathrm{\Σ}}}\underset{j\≠0,j\≠i}{\mathrm{\Σ}}{d}_j\left(n\right)\·f_{m-n}^{i,j}(k-l)+z^{\′}\left(m\right)$ [formula 4] wherein, first y in the equation left side _i ^s(m) intersymbol interference that brings for multipath channel; Second y _i ^p(m)) disturb between the code channel that pilot channel brought; The 3rd y _i ^q(m) be of the mutual interference of other code channel to current the reception code channel; Last z ' (m) is noise component(s) in the RAKE output; And definition $f_{m-n}^{i,j}(k-l)\≡\underset{k}{\mathrm{\Σ}}{f}_{m-n,k}^{i,j}(k-l)$ [formula 5a] $f_{m-n,k}^{i,j}(k-l)\≡\underset{l}{\mathrm{\Σ}}{\stackrel{\‾}{c}}_k^{*}{c}_l{\∫}_{{\mathrm{mT}}_{s,i}}^{{(m+1)T}_{s,j}}{w}_j(t-{\mathrm{nT}}_{s,j}-{\mathrm{lT}}_c+{\mathrm{kT}}_c)\mathrm{pn}(t-{\mathrm{lT}}_c+{\mathrm{kT}}_c)$ [formula 5b] $\·w_i(t-{\mathrm{mT}}_{s,i}){\mathrm{pn}}^{*}\left(t\right)\mathrm{dt}$

6, pilot channel interference cancellation module (A101) is considered y _i ^p(m) be the distracter that knows that pilot channel brought, therefore can from RAKE output, eliminate; And y _l ^q(m) be of the mutual interference of other code channel, relate to complicated multi-user association and detect problem, temporarily ignored current reception code channel; It is as follows to obtain revised RAKE output thus: ${\stackrel{\‾}{y}}_i\left(m\right)=y_i\left(m\right)-y_i^p\left(m\right)$ [formula 6]

7, soft output Viterbi balance module (A102)

At last, realize soft output Viterbi algorithm.The Viterbi equilibrium of standard is classical algorithm.Provide the method that under the QPSK modulation system standard Viterbi equilibrium is changed into soft output Viterbi equilibrium below.Suppose moment state s at k _iOn have four to retain path P ath ₀, Path ₁, Path ₂And Path ₃, Path _m(m=0,1,2,3) are for corresponding respectively to I _K-L=b _mThe path.The log-likelihood value of supposing them is respectively M ₀, M ₁, M ₂And M ₃, correspond to-1-j-1+j, 1-j, 1+j one of four states respectively.After knowing the probability of each state, just can the compute sign probability.First symbol of current state is that-1 probability is on the survivor path $e^{M_0}+{\mathrm{<figure-callout\; id="1"\; label="e\; M"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">e</figure-callout>}}^{M_{}},$ And be that 1 probability is ${\mathrm{<figure-callout\; id="2"\; label="e\; M"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">e</figure-callout>}}^{M_{}}+{\mathrm{<figure-callout\; id="3"\; label="e\; M"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">e</figure-callout>}}^{M_{}}.$ Corresponding to first symbol, preservation soft declared information and should be $\ln [(e^{M_0}+e^{M_1})/({\mathrm{<figure-callout\; id="2"\; label="e\; M"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">e</figure-callout>}}^{M_{}}+e^{M_3})].$ Adopt approximate formula e ^a+ e ^b≈ max{e ^a, e ^bAfter, preservation soft declared information and can be reduced to max{M ₀, M ₁}-max{M ₂, M ₃.In like manner corresponding to second symbol of current state, the simplification of preservation is soft, and to declare information be max{M ₀, M ₃}-max{M ₂, M ₄.

Compare with the Viterbi algorithm of standard, this soft-output algorithm does not increase on memory space, just will store to declare the result firmly and change the soft result of declaring into; On operand, calculate after the log-likelihood value at every turn, increase by twice and compare and twice subtraction.So, compare with the Viterbi algorithm of standard, on operand, increase also few.

The method according to this invention can be applied to the auxiliary low spreading factor CDMA mobile communication system of any pilot tone, comprises 3GPP WCDMA and 3GPP2 cdma2000.

Fig. 1 has provided the device detailed structure schematic diagram that this method is applied to the 3GPP2 cdma2000.Channel parameter estimator module (Bl00) is carried out related operation to receiving the baseband sampling signal, finishes the channel estimating computing shown in [formula 2].Effectively arrive the size of footpath selector (B101), select useful signal to arrive the footpath according to the channel parameter estimator amplitude.RAKE demodulation and high specific merge module (B102) and receive from the baseband sampling signal and according to the position that effectively arrives the place, footpath, carry out related operation, finish the RAKE demodulation computing shown in [formula 3].Local frequency expansion sequence generator (B103) is with generating the frequency expansion sequence synchronous with received signal.When equivalence becomes ISI coefficients calculation block (A100) receive channel parameter estimator, effectively arrives the output of three modules such as footpath selector, local frequency expansion sequence generator, change ISI coefficient when calculating equivalent as [formula 5a] and [formula 5b] shown in.When receiving equivalence, pilot channel interference cancellation module (A101) becomes the output of ISI coefficients calculation block, became the ISI coefficient when equivalent and from RAKE demodulation and high specific merge the output sequence of module, removed pilot channel produces, exported thereby obtain the modified R AKE demodulation shown in [formula 6].Soft output Viterbi equilibrium (A102) receives revised RAKE demodulation output, and the output that becomes the ISI coefficients calculation block during equivalence, carries out soft output Viterbi balancing operational, calculates soft output symbol sequence.It is 7 parts that concrete signal processing is divided into, and is described in detail as follows respectively:

1, at first, the base-band analog signal of receive after the A/D conversion, is become digital signal, enter channel parameter estimator module (B100) then, this module receives the digital baseband sampled signal and carries out related operation, estimates the channel parameter c in each footpath according to [formula 2] _i(n) (i=0,1, A, L-1) with corresponding time delay, its result becomes ISI coefficient calculation unit two modules such as (A100) when offering effective arrival footpath selector (B101) and equivalence simultaneously.

2, effectively arrive the channel parameter c that footpath selector (B101) receive channel estimator estimates _i(n) and corresponding each footpath time delay, select the double footpath of channel estimating intensity level, obtain the time delay and the channel estimation value c of these effective diameters greater than thresholding _lBecome ISI coefficient calculations two modules such as (A100) when (n), exporting to RAKE demodulation and high specific merging module (B102) and equivalence.

3, RAKE demodulation and high specific merge module (B102) and receive the line correlation computing of going forward side by side of baseband sampling signal, and carry out high specific by [formula 3] and merge, and obtain the output symbol sequences y of RAKE receiver _i(m).

4, local frequency expansion sequence generator (B103) is with generating the frequency expansion sequence synchronous with received signal.In order to realize the coherent reception of receiver, need to go out the frequency expansion sequence synchronous with received signal at local recovery.The local frequency expansion sequence of pilot tone is also directly obtained after synchronously by the generator polynomial generation of prior agreement.The generation of the local frequency expansion sequence of other code channel can be obtained (result that the local frequency expansion sequence of pilot tone also can be regarded No. 0 Walsh sign indicating number and local pilot frequency sequence XOR as) by the XOR of quadrature Walsh sequence and local pilot frequency sequence.

5, when equivalence becomes ISI coefficients calculation block (A100) and receives from the local frequency expansion sequence of pilot tone of local frequency expansion sequence generator (B103) and the local frequency expansion sequence w of other code channel ₁(t) each footpath channel parameter estimation value c of being provided of pn (t), channel parameter estimator module (B100) _i(n) and corresponding time delay, effectively arrive directly channel parameter estimation value c of effective arrivals that footpath selector (B101) provided _lN) with corresponding time delay etc., and change ISI coefficient when utilizing [formula 5a] and [formula 5b] to calculate equivalence, comprise the ISI coefficient of data channel and the ISI coefficient of pilot channel (being interference sections between the code channel that brings of pilot channel).

Become ISI coefficient f when becoming pilot channel equivalent of ISI coefficients calculation block (A100) output when 6, pilot channel interference cancellation module (A101) receives equivalence _M-n ^{I, 0}(k-l), utilize [formula 6] that it is merged the output sequence y of module (B102) from RAKE demodulation and high specific _i(m) deduct in.

7, soft output Viterbi balance module (A102) receives the sequences y of pilot channel interference cancellation module (A101) output _i(m) and when equivalence become the output f of ISI coefficients calculation block (A100) _M-n ^{I, i}(k-l),, utilize the algorithm of soft output Viterbi equilibrium to obtain soft output sequence, supply with the decoder of back as the input of soft output Viterbi equilibrium.

Below become the computational process of ISI coefficient when in detail introducing equivalence.Fig. 2 has provided the schematic diagram of this computational process, and being divided into is 5 parts, is described below respectively.

1, effective diameter position (time delay) determination module (C100) receives the output that effectively arrives footpath selector (B101) among Fig. 1 and determines that each effectively arrives the position (time delay) in footpath, so that become the ISI coefficient when calculating each footpath equivalent.

2, effective diameter ISI coefficient (data division) computing module (C101) is according to fixed effective position that arrives the footpath, in conjunction with the output c of channel parameter estimator module (B100) _i(n) with the local frequency expansion sequence w of local frequency expansion sequence generator (B103) output _i(t) pn (t) utilizes [formula 5b] to calculate and becomes the ISI coefficient when each effectively arrives footpath equivalent.Become the ISI coefficient during this equivalence and constitute, the corresponding respectively data symbol that arrives with different time by one group of numerical value.Fig. 3 has provided first footpath pairing when equivalence and has become ISI coefficient f _{M-n, l} ^{I, 0}(k-l) example calculation (back will be introduced in detail).

3, effective diameter ISI coefficient (data division) merges and to become ISI coefficient (data division) when module (C102) receives effective diameter ISI coefficient (data division) computing module (C101) output equivalent, and these coefficients are correspondingly added up, finally obtain one group and become ISI coefficient f during corresponding to different pieces of information symbol equivalent _M-n ^{I, i}(k-l).

4, effective diameter ISI coefficient (pilot portion) computing module (C103) is according to fixed effective position that arrives the footpath, in conjunction with the output c of channel parameter estimator module (B100) _i(n) with the local frequency expansion sequence w of local frequency expansion sequence generator (B103) output ₀(t) pn (t) (having used the Walsh sequence of No. 0 complete ' l ' here) calculates respectively and becomes the ISI coefficient when each effectively arrives footpath equivalent.Become the ISI coefficient during this equivalence and constitute, the corresponding respectively frequency pilot sign that arrives with different time by one group of numerical value.But the computational methods of detailed process comparable data part.

Become ISI coefficient (pilot portion) when 5, each effective diameter of effective diameter ISI coefficient (pilot portion) merging module (C104) reception effective diameter ISI coefficient (pilot portion) computing module (C103) output is equivalent, and these coefficients are correspondingly added up, finally obtain one group and become ISI coefficient f during corresponding to different frequency pilot signs equivalent _M-n ^{I, 0}(k-l).

Become ISI coefficient f when introducing the pairing equivalence in first footpath below in conjunction with Fig. 3 _{M-n, l} ^{I, i}(k-l) example calculation.Be without loss of generality, consider that here spreading ratio is 4 and suppose that maximum time delay expands to the situation of 4 chips.W among the figure _i(t) the synthetic multiple spread spectrum code sequence of local equivalence of pn (t) expression Walsh orthogonal code and pilot tone frequency expansion sequence, $a_n^h=w_i({\mathrm{nT}}_{s,i}-{\mathrm{hT}}_c)\mathrm{pn}({\mathrm{nT}}_{s,i}-{\mathrm{hT}}_c)$ Represent h chip of the corresponding spreading code of n receiving symbol value (value is+1/-1).First line display be synchronized with first the footpath local frequency expansion sequence, second the row, the 3rd, the 4th and fifth line represent respectively the time-delay be 1 chip, 2 chips, the local frequency expansion sequence of 3 chips and 4 chips.The right side one row's parameter c _l(l=0,1,2,3,4) are respectively the channel parameter corresponding estimated value.Last column has provided the local equivalent again conjugation w of spread spectrum code sequence synchronous with first footpath _i(t) pn ^*And corresponding channel parameter estimation c (t), ₀ ^*Value provides at its right-hand member.In order to show the difference of n symbol and n-1 the corresponding spreading code of symbol, the spreading code of n-1 symbol correspondence illustrates with shade.Like this, the integration of [formula 5b] (the n value is m and m+1, i=j, k=0, l=0,1,2,3,4) can be represented with following summation $f_0^{i,i}=[(a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{1*}*a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^1+a_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{2*}*a_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^2+a_{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{3*}*a_{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^3+a_{\mathrm{<figure-callout\; id="4"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{4*}*a_n^4){\stackrel{\&OverBar;}{c}}_0$ $+(a_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{2*}*a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^1+a_{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{3*}*a_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^2+a_{\mathrm{<figure-callout\; id="4"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{4*}*a_n^3){\stackrel{\&OverBar;}{c}}_1$ [formula 7] $+(a_{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{3*}*a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^1+a_{\mathrm{<figure-callout\; id="4"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{4*}*a_n^2){\stackrel{\&OverBar;}{c}}_2$ $+(a_{\mathrm{<figure-callout\; id="4"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{4*}*a_n^1){\stackrel{\&OverBar;}{c}}_3]{\stackrel{\&OverBar;}{c}}_0^{*}$ $f_l^{i,i}=[(a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{1*}*a_{n-1}^4){\stackrel{\&OverBar;}{c}}_1$ $+(a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{1*}*a_{n-1}^3+a_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{2*}*a_n^4){\stackrel{\&OverBar;}{c}}_2$ [formula 8] $+(a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{1*}*a_{n-1}^2+a_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{2*}*a_{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^3+a_{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{3*}*a_n^4){\stackrel{\&OverBar;}{c}}_3$ $+(a_{\mathrm{<figure-callout\; id="1"\; label="n"\; filenames="C0211269800141.png,C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{1*}*a_{n-1}^1+a_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{2*}*a_{n-1}^2+a_{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{3*}*a_{n-1}^3+a_{\mathrm{<figure-callout\; id="4"\; label="n"\; filenames="C0211269800152.png"\; state="\{\{state\}\}">n</figure-callout>}}^{4*}*a_{n-1}^4){\stackrel{\&OverBar;}{c}}_4]{\stackrel{\&OverBar;}{c}}_0^{*}$

Pairing first footpath of pilot portion becomes ISI coefficient f when equivalent _{M-n, l} ^{I, 0}(k-l) calculating principle is the same, as long as change the multiple spread spectrum code sequence of the equivalence among Fig. 2 into local pilot tone spread spectrum code sequence.This computational process is corresponding to effective diameter ISI coefficient (pilot portion) computing module (C103) shown in Fig. 2.

The present invention is on the basis of traditional RAKE receiver, increases independently that three modules realize.Analyze the amount of calculation of several modules of increase below, because local frequency expansion sequence and channel parameter estimation value are plural number, following estimation is all carried out with plural form.

1, when equivalence becomes the estimation of ISI coefficient operand

We only discuss intersymbol interference becomes the ISI coefficient during (maximum delay expands to 8 inclined to one side intervals of sign indicating number) equivalent in two symbols maximum operand.The solution procedure that becomes ISI coefficient (data division) during by equivalence for the coefficient calculations of current sign, needs to calculate 7 time delay spacings relevant (when calculating time-delay when being the ISI coefficient of 4 effective diameter) of the multiple frequency expansion sequence of equivalence that is 1 chip at most as can be known.Because the formation of the multiple frequency expansion sequence of equivalence all is+and the sequence of 1/-1, just multiplying each other can be with the realization of phase Calais.The calculating of this coefficient needs 25 single-bit additions altogether, and 7 times less than 4 integer (2 bit) multiplication with channel parameter (many bits), bit addition more than 7 times, bit multiplication more than 1 time.For the coefficient calculations of previous symbol, need to calculate 4 time delay spacings the relevant of the multiple frequency expansion sequence of equivalence that is 1 chip at most.The calculating of this coefficient needs 16 single-bit additions altogether, and 4 times less than 4 integer (2 bit) multiplication with channel parameter (many bits), bit addition more than 4 times, bit multiplication more than 1 time.By above analysis as can be known, two coefficients that calculate each footpath need at most 41 single-bit additions, and 11 times less than 4 integer (2 bit) multiplication with channel parameter (many bits), bit addition more than 11 times, bit multiplication more than 2 times.Suppose the RAKE receiver that we use the strongest 3 effective diameters to merge, concerning each symbol, six coefficients that calculate these 3 footpaths need 123 single-bit additions at most, 33 times less than 4 integer (2 bit) multiplication with channel parameter (many bits), bit addition more than 33 times, bit multiplication more than 6 times.And three footpath computings simultaneously.At last, three footpath coefficients are merged the ISI coefficient that obtains final two symbols and need bit addition more than 4 times again.

The amount of calculation that becomes ISI coefficient (data division) when the amount of calculation of change ISI coefficient (pilot portion) is with equivalence during equivalence is identical.

2, pilot channel interference eliminated

By [formula 6] as can be known, each symbol only needs once many bit subtraction just can realize.

3, soft output Viterbi equilibrium

Compare with the Viterbi algorithm of standard, calculate after the log-likelihood value of retaining the path on each state at every turn, only need to increase relatively size and twice subtraction twice.Because during high-speed data communication, travelling carriage is nearer apart from the base station usually, the multidiameter delay expansion is also less, only continues 2 to 3 symbols usually, and therefore the increase of the amount of calculation of being brought by soft output Viterbi equilibrium is not remarkable.

Fig. 4 has provided the performance comparison result of traditional RAKE receiver and receiver proposed by the invention.Stipulated the minimum transmitted data rates under the varying environment in the standard of 3G (Third Generation) Moblie.For 307.2K bps the situation that we discuss, should use indoor channel model to outdoor or walking environment.Table 1 has provided the M.1225 indoor various parameters that arrive the channel model under outdoor or the walking environment that IMT-2000 recommends.This paper has used M.1225 channel B model, and the speed of travelling carriage is 10 kilometers/hour.

Concrete simulated conditions is as follows: unique user, spreading rate are 1.2288Mcs, and spreading gain is 4, and forward channel is the QPSK modulation system, and the channel estimating interval is 256 symbols, and code rate is 1/2, and constraint length is 9, and the use size is 12288 block interleaving.Adopt the strongest three footpaths to merge in the RAKE receiver.

Table 1, the more used channel model of performance

Tap	Channel A		Channel B		Doppler frequency spectrum
						Relative time delay (ns)	Average power (dB)	Relative time delay (ns)	Average power (dB)
	1	0	0.0	0						0	Classic
2					110	-9.7	200	-0.9	Classic
	3	190	-19.2	800						-4.9	Classic
4					410	-22.8	1200	-8.0	Classic
	5	-	-	2300						-7.8	Classic
6					-	-	3700	-23.9	Classic

As seen from Figure 4, when signal to noise ratio was higher, because the influence of intersymbol interference, the performance of RAKE receiver did not improve along with the continuation of signal to noise ratio and has clear improvement.The RAKE that this paper proposes receives receiver and the balanced composition algorithm of soft output Viterbi had both utilized RAKE receiver to realize receive diversity, utilizes soft output Viterbi equilibrium to eliminate intersymbol interference again, can greatly improve the performance of receiver.

Claims (5)

1, Rake and equalization cascade receiving method under a kind of code division multiple address low band-spreading ratio, it is characterized in that at first estimating channel parameter, secondly carrying out Rake to received signal receives, utilize the local spread spectrum code sequence that generates in the position of each the footpath channel parameter estimate and effective diameter and the receiver then, the intersymbol interference coefficient that becomes when calculating equivalence, realize soft output viterbi equalization algorithm at last, realize that output viterbi equalization algorithm is to utilize the time output symbol sequence that becomes equivalent ISI coefficient and remove the Rake receiver after sequence of pilot symbols disturbs that calculates, carry out soft output viterbi equalization algorithm, export one group of symbol probability sequence that value is relevant with log-likelihood metrics, offer subsequent communication channel decoding.

2, code division multiple address low band-spreading according to claim 1 is than Rake and equalization cascade receiving method down, and the intersymbol interference coefficient that becomes when it is characterized in that calculating equivalence is to utilize the local spread spectrum code sequence that generates in the position of each the footpath channel parameter that estimates and effective diameter and the receiver to determine that end restrains the equivalent intersymbol interference time-varying coefficient that receiver is comprised in exporting.

3, Rake and the equalization cascade receiving method under a kind of code division multiple address low band-spreading ratio according to claim 1 and 2, it is characterized in that disturbing the method and the interference elimination method of time-varying coefficient to determine the pilot signal interference time-varying coefficient, and from the output of end gram receiver, offset with calculating the frequency pilot sign that is comprised in the output of end gram receiver.

4, Rake and the equalization cascade receiving system under a kind of code division multiple address low band-spreading ratio, it is characterized in that this device is made up of Rake receiver (A) and soft output viterbi equalizer (B) two parts, Rake receiver (A) and the input of input termination baseband sampling signal, the input of the soft output viterbi equalizer of the output termination of Rake receiver (A) (B).

5, Rake and equalization cascade receiving system under a kind of code division multiple address low band-spreading ratio according to claim 5, it is characterized in that Rake receiver comprises channel parameter estimator module (B100), effectively arrive footpath selector module (B101), Rake demodulation and high specific merge module (B102), local frequency expansion sequence generator module (B103), when comprising equivalence, soft output viterbi equalizer becomes ISI coefficients calculation block (A100), pilot channel interference cancellation module (A101), soft output viterbi equalization module (A102), channel parameter estimation module (B100) wherein, effectively arrive through selector module (B101), Rake demodulation and high specific merge module (B102), pilot channel interference cancellation module (A101), soft output viterbi equalization module (A102) order successively is connected in series, the input that becomes ISI coefficients calculation block (A100) during equivalence connects channel parameter estimator module (B100) respectively, effectively arrive through selector module (B101), local frequency expansion sequence generator module (B103) becomes ISI coefficients calculation block (A101) and soft output viterbi equalization module (A102) during equivalence.

Images