CN101483627B - Method for reducing PAR of carrier signal - Google Patents

Abstract

The invention discloses a method of reducing the peak-to-mean ratio of multicarrier signals. The method comprises steps of generating hard despicking signals according to multicarrier combining signals and a despicking threshold; generating multicarrier noise peak signals according to time-delayed multicarrier combining signals and the hard despicking signals; estimating the baseband signal power of each carrier on the basis of multicarrier baseband signals, generating carrier power estimated value, compensating the delay caused by power estimation, and obtaining a carrier power estimation value for delay compensation; performing matched filtering to the multicarrier noise peak signals according to the carrier power estimation value for delay compensation to obtained filtered multicarrier noise peak signals; summing the filtered multicarrier noise peak signals and the multicarrier combining signals to obtain multicarrier despicking signals. By adopting the invention, the phenomenon that carriers with different powers can cause signal distortion and imbalance when using a same matched filter gain can be avoided.

Description

Reduce the method for multi-carrier signal peak equal ratio

Technical field

The present invention relates to the transmission technology of digital signal in a kind of wireless communication system, specifically, relate to a kind of method that reduces the multi-carrier signal peak equal ratio.

Background technology

In wireless communication system, utilize the peak clipping technology to reduce the signal peak-to-average ratio of ingoing power amplifier, to improve the efficient of power amplifier, but also can bring problems such as to a certain degree inband signaling distortion, out of band spectrum broadening this moment, and make transmission signal quality descend and then influence the demodulation performance of receiver.In CDMA access system (as WCDMA, cdma2000 and TD-SCDMA), the inband distortion degree of coming gauge signal with Error Vector Magnitude is leaked the out of band spectrum degree of expansion that recently characterizes transmitter with adjacent-channel power usually.

The peak clipping technology mainly is divided into digital baseband peak clipping and digital intermediate frequency peak clipping two big classes.Digital baseband peak clipping technology makes peak-to-average force ratio have raising to a certain degree because the signal after its peak clipping carries out the multilevel interpolation Filtering Processing, and the effect of its peak clipping is not fine.The digital intermediate frequency peak clipping is meant that baseband signal carries out peak clipping again after merging through the multilevel interpolation filter of digital intermediate frequency and handle, and commonly used have hard peak clipping, soft peak clipping and a matched filtering peak clipping etc.The advantage of the hard peak clipping technology of digital intermediate frequency is that hardware realizes comparatively simple, and shortcoming is that it directly carries out the hard peak clipping processing of digital intermediate frequency, can cause serious adjacent-channel power leakage phenomenon.Compare with the hard peak clipping of digital intermediate frequency, it is much smaller that the advantage of the soft peak clipping of digital intermediate frequency is that the video stretching of its generation or neighboring trace leak, but its peak clipping effect is still not ideal enough, realizes also complicated.But leak than performance index if will guarantee to satisfy adjacent-channel power, hard peak clipping of digital intermediate frequency and soft peak clipping technology all are difficult to reach peak-to-average force ratio performance preferably.

Chinese patent application number is that the peak clipping technology that 03804978.3 application for a patent for invention is used just belongs to digital intermediate frequency matched filtering peak clipping technology, this patent application designs at multi-carrier signal peak clipping situation, the baseband filter of a digital intermediate frequency correspondence of design earlier sample frequency, after the baseband filter shift frequency being arrived the carrier frequency of correspondence, again the filter addition behind the shift frequency is obtained closing path filter, the peak noise signal is carried out matched filtering with closing path filter.When input is closed the sample frequency of way word intermediate-freuqncy signal when big, for example 12 times or bigger, the tap number of filter is required very high, need to carry out complex filter in addition and handle, these have all increased the cost and the difficulty of hardware realization (FPGA or ASIC) greatly.Though when this technical scheme has been considered each carrier wave matched filter coefficient generation suitable adjustment is carried out in its gain, adjust more performance to obtain than only using scaler to unify gain, but explanation is not specifically how to the adjustment that gains of each carrier wave matched filter coefficient, and how in real time each carrier wave matched filter coefficient to be adjusted, especially how this adjusts each matched filter gain in system realizes.

Summary of the invention

The technical problem to be solved in the present invention provides a kind of method that reduces the multi-carrier signal peak equal ratio, the distorted signals energy imbalance of having avoided the different carrier wave of power to use the identical match filter gain to be brought.

Technical scheme is as follows:

A kind of method that reduces the multi-carrier signal peak equal ratio, step comprises:

(1) produce hard peak clipping signal according to multi-transceiver combination signal and peak clipping thresholding, comprising:

(11) adopt rotation of coordinate compute vectors pattern to calculate to the multi-transceiver combination signal, obtain amplitude, phase place and the rotation gain of each input sample signal;

(12) product with amplitude and peak clipping thresholding and rotation gain compares; If amplitude is not more than the product of peak clipping thresholding and rotation gain, then hard peak clipping signal is input multi-transceiver combination signal, otherwise it is that initial parameters obtains hard peak clipping signal that rotation of coordinate compute vectors pattern adopts peak clipping thresholding and the ratio and the phase place of rotation gain;

(2) according to the time multi-transceiver combination signal delayed and hard peak clipping signal produce multicarrier peak noise signal;

(3) estimate the baseband signal power of each carrier wave according to the multicarrier baseband signal, generate the carrier power estimated value, compensation power is estimated the delay that causes, is compensated the carrier power estimated value of delay;

(4) according to the carrier power estimated value of described compensating delay described multicarrier peak noise signal is carried out matched filtering, obtain filtered multicarrier peak noise signal;

(5) described filtered peak noise signal and the summation of described multi-transceiver combination signal are added up, obtain the multi-carrier peak-clipping signal.

Further, in the step (5), described filtered peak noise signal and described multi-transceiver combination signal are carried out multistage peak clipping iterative processing, obtain the multi-carrier peak-clipping signal.

Further, in the step (1), hard peak clipping unit carries out hard peak clipping according to the peak clipping thresholding to the multi-transceiver combination signal of importing to be handled, to produce hard peak clipping signal; Described hard peak clipping unit comprises amplitude computing unit, decision unit and amplitude limit unit, wherein, described amplitude computing unit calculates the range value of input multi-transceiver combination signal, decision unit compares described range value and peak clipping thresholding, if described range value is less than or equal to the peak clipping thresholding, described amplitude limit unit adopts current multi-transceiver combination signal as hard peak clipping signal; If described range value is greater than described peak clipping thresholding, described amplitude limit unit adopts the amplitude of peak clipping thresholding as current demand signal, and the inhibit signal phase invariant carries out hard peak clipping to the multi-transceiver combination signal of input and handles, and obtains the hard peak clipping signal after the hard-limiting.

Further, step (11) comprising:

(111) obtain the amplitude of each input sample signal with rotation of coordinate compute vectors pattern, preserve the rotation parameter that per step calculates;

(112) realize that with rotation of coordinate compute vectors pattern cartesian coordinate is to polar conversion;

(113), obtain the amplitude and the rotation gain of each input sample signal after the rotation of employing rotation of coordinate compute vectors pattern according to given multi-transceiver combination signal.

Further, step (12) comprising: the amplitude of described sampled signal and the product of peak clipping thresholding and rotation gain are compared; If described amplitude is not more than the product of described peak clipping thresholding and described rotation gain, then described hard peak clipping signal is input multi-transceiver combination signal; Otherwise, ratio and 0 with described peak clipping thresholding and described rotation gain is initial parameters, utilize the despining of rotation of coordinate compute vectors pattern, and the rotation parameter of preserving during according to rotation of coordinate compute vectors mode computation calculates hard peak clipping signal, wherein, the rotation parameter of every grade of use of described rotation of coordinate compute vectors pattern despining aligns with rotation of coordinate compute vectors pattern.

Further, in the step (4), described multicarrier peak noise signal obtains as follows:

(41) the peak noise signal is carried out f _s/ 4 Digital Up Converts conversion obtains with f _s/ 4 is the peak noise signal of centre frequency;

(42) calculate according to each carrier power estimated value and refresh corresponding carrier wave shift frequency to f _s/ 4 filter real coefficient obtains closing the filter real coefficient on road with its addition, and carries out matched filter processing, obtains filtered peak noise signal;

(43) the peak noise signal after the matched filtering is carried out-f _s/ 4 Digital Down Convert conversion obtains the zero-frequency peak noise signal after the matched filtering;

(44) the peak noise signal to down-conversion carries out the half-band filter processing, gets the multicarrier peak noise signal before offsetting to the end.

The present invention according to each carrier wave power of baseband signal estimated value to the adjustment that gains of the bank of filters tunable filter coefficient of corresponding carriers, multi-carrier signal behind the involutory road carries out digital intermediate frequency matched filtering peak clipping to be handled, and has avoided the distorted signals energy imbalance that uses the identical match filter gain to be brought to the different carrier wave of power.The present invention can obtain better peak clipping effect satisfying under the condition that same error amplitude of the vector and adjacent-channel power leak ratio, also can make the multi-transceiver combination signal after the peak clipping have lower peak-to-average force ratio.

The hard despicking method that the present invention adopts is different with traditional implementation method, and rotation of coordinate numerical calculation (CORDIC) pattern is applied in the hard peak clipping processing, has reduced hard-wired complexity and resource consumption; The peak noise processing unit is by arriving f with peak noise signal transformation _s/ 4 handle, thereby reach the purpose of real number filtering, have reduced traditional matched filtering plural number largely and have handled multiplier and the logical resource that needs.

When Digital IF Processing, if will be when obtaining higher sampling rate, consider the saving of hardware resource again, can use the cic filter of two-stage semi-band filter or higher interpolation multiple, in addition, design matched filter banks according to concrete if sampling speed, if, wish to obtain lower peak-to-average force ratio performance satisfying the leakage of Error Vector Magnitude and adjacent-channel power than under the condition.

The suitably used multi-stage iteration peak clipping of the present invention is handled, and its Digital Signal Processing adopts the sequential processes mode, does not relate to any feedback processing unit, easily adopts streamline to realize in the actual hardware system.

Description of drawings

Fig. 1 is the formation of multi-transceiver combination signal among the present invention and the schematic diagram of present position, multi-carrier peak-clipping unit;

Fig. 2 will compensate the amplitude-frequency response schematic diagram of the CRRC after CIC gains among the present invention in the specific implementation structure shown in Figure 1;

Fig. 3 is the structural representation of multi-carrier peak-clipping unit among the present invention;

Fig. 4 (a) and (b) are to realize hard peak clipping process schematic diagram in the structure shown in Figure 3;

Fig. 5 is that structure shown in Figure 3 realizes peak noise processing procedure schematic diagram;

Fig. 6 is that matched filter shown in Figure 5 is according to the adjusted amplitude-frequency response schematic diagram of carrier power estimated value;

Fig. 7 is the peak noise signal amplitude-frequency response schematic diagram after the matched filter matched filtering shown in Figure 5, behind the semi-band filtering;

Fig. 8 reduces multi-carrier signal peak equal than the structural representation of realizing multi-carrier peak-clipping by N level iteration among the present invention;

Fig. 9 is the CCDF curve chart before and after the multi-transceiver combination signal process multi-carrier peak-clipping among the present invention.

Embodiment

In the base station transmitter of wireless communication system, the input signal that peak-to-average force ratio is big has reduced the efficient of power amplifier, for the multi-transceiver combination signal, and CDMA access system for example, it is more outstanding that problem seems.Use peak clipping technology reduces the peak-to-average force ratio of the multi-transceiver combination signal of ingoing power amplifier, to improve the efficient of power amplifier.

As shown in Figure 1, be the formation of multi-carrier signal among the present invention and the schematic diagram of present position, multi-carrier peak-clipping unit.Multi-carrier digital frequency up-converted (DUC) unit is as the digital intermediate frequency filter, comprise root raised cosine interpolation filter (CRRC), semi-band filter (HBF) and stacked integration pectination interpolation filter (CIC) that multichannel is connected in series mutually, every roadbed band signal is carried out digital medium-frequency signal interpolation and Filtering Processing, and form the output of multi-transceiver combination signal.For every road carrier wave, its baseband signal (baseband signal 1, baseband signal 2 or baseband signal 3) is carried out pulse-shaping filtering by the root raised cosine interpolation filter (CRRC) of compensation, improve the sampling rate of digital medium-frequency signal through semi-band filter (HBF), strengthen by stacked integration pectination interpolation filter (CIC) again the mirror image of signal in frequency domain suppressed, and carry out mixing formation carrier wave digital medium-frequency signal with digital controlled oscillator and export, at last each carrier wave digital medium-frequency signal is merged to produce the multi-transceiver combination signal.

Realize the length of storage list in order to reduce hardware, the frequency form of channel grid integral multiple correspondence that can be by the design agreement regulation is realized.Further, if need obtain higher sampling rate, can adopt the CIC of two-stage semi-band filter or the more senior interpolation of design in Digital IF Processing.

In digital medium-frequency signal interpolation and Filtering Processing, realize cost and good mirror image rejection is provided though use CIC to reduce hardware, also can cause the unevenness of gain in the band simultaneously.In order to make whole digital intermediate frequency interpolation and Filtering Processing have preferably gain flatness in the band, it is uneven that the CRRC by design compensation compensates in the band of CIC gain.

The multi-carrier peak-clipping unit receives the multi-transceiver combination signal, according to the peak clipping thresholding that is provided with the multi-transceiver combination signal is carried out peak clipping and handles, and what reduce the ingoing power amplifier closes the road signal peak-to-average ratio, to improve the efficient of power amplifier.

As shown in Figure 2, be amplitude-frequency response schematic diagram according to the CRRC of Fig. 1 example design.Provided the amplitude-frequency response characteristic of compensation front and back CRRC and whole digital intermediate frequency filter (comprising CRRC, HBF and CIC) convolution results among Fig. 2, as can be seen if CRRC does not compensate the interior gain loss of the band of CIC, the inband flatness poor-performing of then whole digital intermediate frequency filter convolution.

As shown in Figure 3, be the structural representation of the multi-carrier peak-clipping unit that is used to reduce the multi-carrier signal peak equal ratio among the present invention.Multi-carrier peak-clipping unit among Fig. 3 comprises that a hard peak clipping unit, three time delay unit (comprising the first time delay unit, the second time delay unit and the 3rd time delay unit), carrier power estimation unit, peak noise processing unit and two summations unit (comprise first summation add up unit and second summation unit that adds up) that adds up forms.

Hard peak clipping unit carries out hard peak clipping according to the peak clipping thresholding to the multi-transceiver combination signal of input to be handled, and producing hard peak clipping signal, and sends to first summation unit that adds up.The first time delay unit postpones the multi-transceiver combination signal, to compensate the time delay that hard peak clipping unit and peak noise processing unit cause and to guarantee time unifying, sends to second summation unit that adds up.First summation unit that adds up deducts hard peak clipping signal through the multi-transceiver combination signal after the second time delay unit time delay processing (the second time delay unit is in order to the processing delay that compensates hard peak clipping unit and introduce and guarantee time unifying), produces multicarrier peak noise signal and sends to the peak noise processing unit.The peak noise processing unit carries out matched filtering to multicarrier peak noise signal, obtains filtered multicarrier peak noise signal, sends to the second summation unit that adds up.The carrier power estimation unit is used to estimate the baseband signal power of each carrier wave, and the carrier power estimated value is sent to the 3rd time delay unit.The delay that the estimation of the 3rd time delay unit compensation power causes is with the time complete matching that guarantees that peak noise is handled.Suing for peace to the multi-transceiver combination signal after the first time delay cell processing and filtered multicarrier peak noise signal and add up in second summation unit that adds up, produces the multi-carrier peak-clipping signal.

The decision unit that hard peak clipping unit comprises the amplitude computing unit of multi-transceiver combination (IQ) signal, compares with the peak clipping thresholding, and the amplitude limit unit that produces hard peak clipping signal according to the peak clipping thresholding.The amplitude computing unit calculates the range value of input IQ signal, is used for comparing with the peak clipping thresholding.Decision unit compares the range value and the peak clipping thresholding of IQ signal, if the range value that calculates is less than or equal to the peak clipping thresholding, then this IQ signal is not done any processing, and the amplitude limit unit adopts the current I Q signal as hard peak clipping signal; If the range value that calculates is greater than the peak clipping thresholding, the amplitude limit unit adopts the amplitude of peak clipping thresholding as current demand signal, and the inhibit signal phase invariant carries out hard peak clipping to the IQ signal of input and handles, and obtains the hard peak clipping signal after the hard-limiting.

The method of reduction multi-carrier signal peak equal ratio of the present invention comprises the steps:

1, produces hard peak clipping signal according to multi-transceiver combination signal and peak clipping threshold parameter;

2, according to the time multi-transceiver combination signal delayed and hard peak clipping signal produce the peak noise signal;

3, estimate the baseband signal power of each carrier wave according to the multicarrier baseband signal, generate the carrier power estimated value and send to the 3rd time delay unit, the delay that the estimation of the 3rd time delay unit compensation power causes, with the time complete matching that guarantees that peak noise is handled, the carrier power estimated value of compensating delay is sent to the peak noise processing unit;

4, according to the carrier power estimated value of compensating delay multicarrier peak noise signal is carried out matched filtering, obtain filtered multicarrier peak noise signal, send filtered multicarrier peak noise signal to the second summation unit that adds up;

5, sue for peace filtered peak noise signal and multi-transceiver combination signal and add up in second summation unit that adds up, and obtains the multi-carrier peak-clipping signal.

As Fig. 4 (a) and Fig. 4 (b) is hard despicking method implementation procedure in the multi-carrier peak-clipping unit.

Shown in Fig. 4 (a), be to use the process of rotation of coordinate numerical calculation (CORDIC) mode computation multi-transceiver combination signal amplitude, obtain signal amplitude $K*\sqrt{{\left(I_k^{\mathrm{sum}}\right)}^2+{\left(Q_k^{\mathrm{sum}}\right)}^2}$ With the direction of rotation parameter d _n, this amplitude result is a signal amplitude $\sqrt{{\left(I_k^{\mathrm{sum}}\right)}^2+{\left(Q_k^{\mathrm{sum}}\right)}^2}$ K doubly (K comes from the processing gain of cordic algorithm), d here _nDepend on y _nSymbol (if y _n＜0 o'clock d _nGet 1, otherwise d _nGet-1), every grade the anglec of rotation is respectively d _n* arctan 2 ^-n

Shown in figure (b), be d according to Thr/K and preservation _nParameter is carried out the CORDIC despining and is produced hard peak clipping output signal (Thr*cos φ _kWith Thr*sin φ _k, ${\mathrm{\φ}}_k=\arctan \frac{Q_k^{\mathrm{sum}}}{I_k^{\mathrm{sum}}}$ ) process, every grade the anglec of rotation is respectively-d _n* arctan 2 ^-n, every grade d during despining _nBe rotated in the forward identical.

As shown in Figure 5, be that structure shown in Figure 3 realizes peak noise processing procedure schematic diagram.

The peak noise processing unit comprises: f _s/ 4 shift frequency unit, matched filter processing unit ,-f _s/ 4 shift frequency unit, half-band filter unit.

Peak noise is handled and is comprised the steps:

1, f _s/ 4 shift frequency unit carry out f to the peak noise signal _s/ 4 Digital Up Converts conversions (DUC) obtain with f _s/ 4 is the peak noise signal of centre frequency, and this peak noise signal is sent to the matched filter processing unit;

2, the matched filter processing unit calculates according to each carrier power estimated value and refreshes corresponding carrier wave shift frequency to f _s/ 4 filter real coefficient obtains closing the filter real coefficient on road with its addition, and carries out matched filter processing, and the peak noise signal after the matched filtering is sent to-f _s/ 4 shift frequency unit;

3 ,-f _sThe peak noise signal of/4 shift frequency unit after to matched filtering carries out-f _s/ 4 Digital Down Convert conversion (DDC) obtains the zero-frequency peak noise signal after the matched filtering, and this zero-frequency peak noise signal is sent to the half-band filter unit;

4, the half-band filter unit carries out the semi-band filtering processing to the zero peak noise signal of down-conversion, gets the multicarrier peak noise signal before offsetting to the end.

As shown in Figure 6, be that matched filter shown in Figure 5 is according to the adjusted amplitude-frequency response schematic diagram of carrier power estimated value.

Solid line is the amplitude-frequency response of the matched filter real coefficient calculated when being full power (being normalized to 0dB) of three carrier waves among Fig. 6, among Fig. 6 dotted line then be three carrier powers for-2dB ,-8dB and-amplitude-frequency response of the matched filter coefficient that obtains during 5dB.This method has been avoided the distorted signals energy imbalance that uses the identical match filter gain to bring to the different carrier wave of power, thereby has improved the effect of peak clipping.In addition, suitably finely tune by the gain of each carrier wave respective filter coefficient, the EVM that can reach each carrier wave compares balance, to reach the further effect of improving the peak clipping performance.

As shown in Figure 7, it is the peak noise signal amplitude-frequency response schematic diagram after the matched filter matched filtering shown in Figure 5, behind the semi-band filtering, wherein, shown in Fig. 7 (a) example of the peak noise signal amplitude-frequency response after the matched filtering shown in Figure 5, effectively the peak noise frequency spectrum is f _sNear/4 part; Fig. 7 (b) then is the schematic diagram of the peak noise signal amplitude-frequency response behind the semi-band filtering, and the purpose that increases this half-band filter is that the outer useless spectrum signal of band is removed in filtering.

Consider the Mathematical Modeling of three carrier communication systems among the present invention.For convenience of description and expression, use different mathematic signs to represent the signal of each stage or unit below, represent carrier index, baseband sampling index and if sampling index with i, j and k respectively.

A, baseband I Q signal: I _{I, j} ^Bb, Q _{I, j} ^Bb

The intermediate-freuqncy signal of each carrier wave correspondence: I behind b, the filtering interpolation _{I, k} ^Dif, Q _{I, k} ^Dif

C, multi-transceiver combination signal: I _k ^Sum, Q _k ^Sum

E, peak clipping threshold parameter: Thr

F, hard peak clipping signal: I _k ^Hard_clip, Q _k ^Hard_clip

G, multicarrier peak noise signal: I _k ^Peak_noise, Q _k ^Peak_noise

H, f _sPeak noise signal after/4 up-conversions: I _{I, k} ^Fs/4, Q _{I, k} ^Fs/4

The peak noise signal of each carrier wave correspondence: I after i, the matched filtering _{I, k} ^Mf, Q _{I, k} ^Mf

J, matched filter coefficient and half-band filter coefficient: matched_filtering_coeff, hbf_coeff

K, filtered multicarrier peak noise signal: I _k ^{Final_peak_noise}, Q _k ^{Final_peak_noise}

L, multi-carrier peak-clipping signal: I _k ^Clipped, Q _k ^Clipped

Each carrier wave baseband I Q signal produces digital intermediate frequency sampled signal I through multilevel interpolation filtering (CRRC, HBF and CIC) back _{I, k} ^DifAnd Q _{I, k} ^Dif, produce multi-transceiver combination signal I with corresponding carrier frequency mixing and addition _k ^SumAnd Q _k ^SumThree carrier waves before the three carrier wave peak clipping unit close the road signal and can be expressed as:

$I_k^{\mathrm{sum}}+j*Q_k^{\mathrm{sum}}=\underset{i=1}{\overset{4}{\mathrm{\&Sigma;}}}(I_{i,k}^{\mathrm{dif}}+j*Q_{i,k}^{\mathrm{dif}})*{\mathrm{<figure-callout\; id="2"\; label="e\; j"\; filenames="H200810000335801E00022.png,H200810000335801E00031.png"\; state="\{\{state\}\}">e</figure-callout>}}^j$ (formula 1)

Hard peak clipping among Fig. 3 of the present invention can adopt cordic algorithm to realize.At first to each input sample signal I _k ^SumAnd Q _k ^SumAdopt the CORDIC arrow pattern to obtain amplitude A _k $(A_k=K*\sqrt{{\left(I_k^{\mathrm{sum}}\right)}^2+{\left(Q_k^{\mathrm{sum}}\right)}^2})$ And phase _kWith A _kCompare with Thr*K; If A _kBe less than or equal to Thr*K, then hard peak clipping signal is input multi-transceiver combination signal; If A _kGreater than Thr*K, then use the CORDIC rotary mode with Thr/K and φ _k(being calculated by the CORDIC arrow pattern) obtains hard peak clipping signal for initial parameters.

Hard peak clipping among Fig. 3 of the present invention also can adopt the cordic algorithm of correction to realize, can save 1/3 hardware circuit resource, to reduce hardware implementation cost.By Fig. 4 and in conjunction with following description, can be readily appreciated that hard peak clipping implementation method of the present invention.At first calculate the amplitude A of each input sample signal with the CORDIC arrow pattern _k, preserve the d of per step calculating _nFor the CORDIC despining usefulness of back, and do not need to calculate phase term φ _kThe CORDIC arrow pattern realizes that cartesian coordinate to polar conversion, is formulated as:

$\left\{\begin{array}{c}{x}_{n+1}=x_n-d_n\&CenterDot;y_n\&CenterDot;2^{-n}\\ y_{n+1}=y_n+d_n\&CenterDot;x_n\&CenterDot;2^{-n}\end{array}\right.n\&GreaterEqual;0$ (formula 2)

D in the formula 2 _nDepend on y _nSymbol:

$d_n=\left\{\begin{array}{cc}1& y_n<0\\ -1& y_n\&GreaterEqual;0\end{array}\right.$ (formula 3)

Given $x_0=I_k^{\mathrm{sum}}$ With ${\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="H200810000335801E00012.png,H200810000335801E00041.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=Q_k^{\mathrm{sum}},$ Through obtaining after the CORDIC rotation:

$x_{n+1}=K\&CenterDot;\sqrt{{x_0}^2+{{\mathrm{<figure-callout\; id="0"\; label="y"\; filenames="H200810000335801E00012.png,H200810000335801E00041.png"\; state="\{\{state\}\}">y</figure-callout>}}_0}^2}$

y _N+1=0 (formula 4)

$K=\underset{n=0}{\overset{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="H200810000335801E00012.png,H200810000335801E00041.png"\; state="\{\{state\}\}">N</figure-callout>}}{\mathrm{\&Pi;}}}\sqrt{1+2^{-2n}}$

In the formula 4, N is the progression (N gets 8 and just can obtain excellent precision usually) of CORDIC rotation, x _N+1Be input sample signal I _k ^SumAnd Q _k ^SumThe amplitude of calculating behind CORDIC is with A _kCompare with Thr*K (K is the gain of CORDIC rotation); If A _kBe less than or equal to Thr*K, then hard peak clipping signal is input multi-transceiver combination signal; If A _kGreater than Thr*K, x ₀And y ₀Be that initial parameters is carried out the CORDIC despining with Thr/K and 0 respectively, and the d that preserves when calculating according to the CORDIC arrow pattern _nParameter is calculated hard peak clipping signal.The d of every grade of use of CORDIC despining _nMust align with the CORDIC arrow pattern.

$\left\{\begin{array}{c}{x}_{n+1}=x_n+d_n\&CenterDot;y_n\&CenterDot;2^{-n}\\ y_{n+1}=y_n-d_n\&CenterDot;x_n\&CenterDot;2^{-n}\end{array}\right.n\&GreaterEqual;0$ (formula 5)

Given x ₀=Thr/K and y ₀=0, then have:

x _n+1＝Thr·cosφ _k

y _n+1＝Thr·sinφ _k

$K=\underset{n=0}{\overset{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="H200810000335801E00012.png,H200810000335801E00041.png"\; state="\{\{state\}\}">N</figure-callout>}}{\mathrm{\&Pi;}}}\sqrt{1+2^{-2n}}$ (formula 6)

In the formula 6, x _N+1And y _N+1For corresponding carrier wave closes road signal I _k ^Sum, Q _k ^SumHard peak clipping signal I _k ^Hard_clipAnd Q _k ^Hard_clip, φ _kBe I _k ^Sum, Q _k ^SumPhase place.

The multi-transceiver combination signal of delaying when hard peak clipping signal deducts then obtains multicarrier peak noise signal: $\left\{\begin{array}{c}{I}_k^{\mathrm{peak}\_\mathrm{noise}}=I_k^{\mathrm{hard}\_\mathrm{clip}}-I_k^{\mathrm{sum}}\\ Q_k^{\mathrm{peak}\_\mathrm{noise}}=Q_k^{\mathrm{hard}\_\mathrm{clip}}-Q_k^{\mathrm{sum}}\end{array}\right.$ (formula 7)

Carrier power estimation unit among Fig. 3, the power of baseband signal of every carrier wave correspondence calculates with formula 8:

$P_i^{\mathrm{est}}=\frac{1}{N_{\mathrm{samples}}}\underset{n=1}{\overset{N_{\mathrm{samples}}}{\mathrm{\&Sigma;}}}({I_{i,\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H200810000335801E00022.png,H200810000335801E00031.png"\; state="\{\{state\}\}">n</figure-callout>}}}^2+{Q_{i,n}}^2)i=\mathrm{1,2,3},$ (formula 8)

In the formula 8, N _SamplesFor being used for the baseband sampling number that power is estimated, according to P _i ^EstTable look-up (LUT) obtain corresponding amplitude gain A _i ^Est

$A_i^{\mathrm{est}}=\sqrt{P_i^{\mathrm{est}}}i=\mathrm{1,2,3}$ (formula 9)

The peak noise signal is carried out f _s/ 4 up-conversions obtain:

$I_{i,k}^{f_s/4}+j*Q_{i,k}^{f_s/4}=(I_k^{\mathrm{peak}\_\mathrm{noise}}+j*Q_k^{\mathrm{peak}\_\mathrm{noise}})*e^{j*\mathrm{\&pi;k}/2}$ (formula 10)

Peak noise signal after the up-conversion obtains through matched filter processing:

$I_{i,k}^{\mathrm{mf}}+j*Q_{i,k}^{\mathrm{mf}}=\mathrm{filter}(I_{i,k}^{f_s/4}+j*Q_{i,k}^{f_s/4},\mathrm{matched}\_\mathrm{filtering}\_\mathrm{coeff})$ (formula 11)

Peak noise signal after the matched filtering is carried out-f _sThe peak noise signal that/4 down-conversions and semi-band filtering are handled to the end:

$I_{i,k}^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}+j*Q_{i,k}^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}=\mathrm{filter}((I_{i,k}^{\mathrm{mf}}+j*Q_{i,k}^{\mathrm{mf}})*e^{-j*\mathrm{\&pi;k}/2},\mathrm{hbf}\_\mathrm{coeff})$ (formula 12)

Shi Yanhou multi-transceiver combination signal and filtered multicarrier peak noise signal plus obtain the multi-carrier peak-clipping signal:

$\left\{\begin{array}{c}{I}_k^{\mathrm{clipped}}=I_k^{\mathrm{sum}}+I_k^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}\\ Q_k^{\mathrm{clipped}}=Q_k^{\mathrm{sum}}+Q_k^{\mathrm{final}\_\mathrm{peak}\_\mathrm{noise}}\end{array}\right.$ (formula 13)

As shown in Figure 8, be to reduce multi-carrier signal peak equal than the structural representation of realizing multi-carrier peak-clipping by N level iteration among the present invention.The specific implementation of Fig. 8 and Fig. 3 difference are, can require and the hardware resource situation according to systematic function during system design, the multi-transceiver combination signal is carried out the multi-stage iteration peak clipping handle operation, make and satisfying under systematic error amplitude of the vector and adjacent-channel power leak than index condition, obtain more excellent peak clipping effect, guarantee that promptly the multi-carrier peak-clipping signal has littler peak-to-average force ratio.Among Fig. 8, multi-carrier peak-clippings at different levels unit also can use different peak clipping threshold parameters as required, to obtain better peak clipping effect.

As shown in Figure 9, be the multi-transceiver combination signal through CCDF (CCDF) curve chart before and after the multi-carrier peak-clipping, reflected three carrier wave WCDMA digital intermediate frequencies close the road signal through peak clipping of the present invention unit after the contrast situation of signal CCDF curve.As can be seen from Figure 9, by the 9.8dB (@0.01% before the peak clipping) reduce to 5.6dB, after the peak clipping not only peak-to-average force ratio have significantly and reduce.

Compare with similar peak clipping technology, the multi-carrier peak-clipping technology of the present invention's sampling, guaranteeing to satisfy under systematic error amplitude of the vector and adjacent-channel power leak than index condition, use hardware resource still less and obtain lower peak-to-average force ratio performance, obtain higher power amplification efficiency easilier.

The present invention can carry out suitable modification and cutting according to hardware resource and the performance requirement in the practical communication system design.During Digital IF Processing,, can use the cic filter of two-stage semi-band filter or higher interpolation multiple if will when obtaining higher sampling rate, consider the saving of hardware resource again.If satisfying the leakage of Error Vector Magnitude and adjacent-channel power, wish to obtain lower peak-to-average force ratio performance, can suitably use multi-stage iteration peak clipping processing method to realize than under the condition.Scheme provided by the invention, its Digital Signal Processing adopts the sequential processes mode, does not relate to any feedback processing unit, thereby easily realizes in the actual hardware system.

Here the present invention is described in detail by specific embodiment, provide the description of the foregoing description to make or be suitable for the present invention in order to make those skilled in the art, the various modifications of these embodiment are to understand easily for a person skilled in the art.The present invention is not only applicable to code division multiple access systems such as WCDMA, cdma2000 and TD-SCDMA, also is applicable to global mobile communication (GSM) system, enhanced data rates for gsm evolution technology (EDGE) system.

Claims (5)

1. method that reduces the multi-carrier signal peak equal ratio, step comprises:

(1) produce hard peak clipping signal according to multi-transceiver combination signal and peak clipping thresholding, comprising:

(11) adopt rotation of coordinate compute vectors pattern to calculate to the multi-transceiver combination signal, obtain amplitude, phase place and the rotation gain of each input sample signal;

(12) product with amplitude and peak clipping thresholding and rotation gain compares; If amplitude is not more than the product of peak clipping thresholding and rotation gain, then hard peak clipping signal is input multi-transceiver combination signal, otherwise it is that initial parameters obtains hard peak clipping signal that rotation of coordinate compute vectors pattern adopts peak clipping thresholding and the ratio and the phase place of rotation gain;

(2) according to the time multi-transceiver combination signal delayed and hard peak clipping signal produce multicarrier peak noise signal;

(3) estimate the baseband signal power of each carrier wave according to the multicarrier baseband signal, generate the carrier power estimated value, compensation power is estimated the delay that causes, is compensated the carrier power estimated value of delay;

(4) according to the carrier power estimated value of described compensating delay described multicarrier peak noise signal is carried out matched filtering, obtain filtered multicarrier peak noise signal;

(5) described filtered peak noise signal and the summation of described multi-transceiver combination signal are added up, obtain the multi-carrier peak-clipping signal.

2. the method for claim 1 is characterized in that, in the step (5), described filtered peak noise signal and described multi-transceiver combination signal is carried out multistage peak clipping iterative processing, obtains the multi-carrier peak-clipping signal.

3. the method for claim 1 is characterized in that, in the step (1), hard peak clipping unit carries out hard peak clipping according to the peak clipping thresholding to the multi-transceiver combination signal of importing to be handled, to produce hard peak clipping signal; Described hard peak clipping unit comprises amplitude computing unit, decision unit and amplitude limit unit, wherein, described amplitude computing unit calculates the range value of input multi-transceiver combination signal, decision unit compares described range value and peak clipping thresholding, if described range value is less than or equal to the peak clipping thresholding, described amplitude limit unit adopts current multi-transceiver combination signal as hard peak clipping signal; If described range value is greater than described peak clipping thresholding, described amplitude limit unit adopts the amplitude of peak clipping thresholding as current demand signal, and the inhibit signal phase invariant carries out hard peak clipping to the multi-transceiver combination signal of input and handles, and obtains the hard peak clipping signal after the hard-limiting.

4. the method for claim 1 is characterized in that, step (11) comprising:

(111) obtain the amplitude of each input sample signal with rotation of coordinate compute vectors pattern, preserve the rotation parameter that per step calculates;

(112) realize that with rotation of coordinate compute vectors pattern cartesian coordinate is to polar conversion;

(113), obtain the amplitude and the rotation gain of each input sample signal after the rotation of employing rotation of coordinate compute vectors pattern according to given multi-transceiver combination signal;

Step (12) comprising: the amplitude of described sampled signal and the product of peak clipping thresholding and rotation gain are compared; If described amplitude is not more than the product of described peak clipping thresholding and described rotation gain, then described hard peak clipping signal is input multi-transceiver combination signal; Otherwise, ratio and 0 with described peak clipping thresholding and described rotation gain is initial parameters, utilize the despining of rotation of coordinate compute vectors pattern, and the rotation parameter of preserving during according to rotation of coordinate compute vectors mode computation calculates hard peak clipping signal, wherein, the rotation parameter of every grade of use of described rotation of coordinate compute vectors pattern despining aligns with rotation of coordinate compute vectors pattern.

5. the method for claim 1 is characterized in that, in the step (4), described multicarrier peak noise signal obtains as follows:

(41) the peak noise signal is carried out f _s/ 4 Digital Up Converts conversion obtains with f _s/ 4 is the peak noise signal of centre frequency;

(42) calculate according to each carrier power estimated value and refresh corresponding carrier wave shift frequency to f _s/ 4 filter real coefficient obtains closing the filter real coefficient on road with its addition, and carries out matched filter processing, obtains filtered peak noise signal;

(43) the peak noise signal after the matched filtering is carried out-f _s/ 4 Digital Down Convert conversion obtains the zero-frequency peak noise signal after the matched filtering;

(44) the peak noise signal to down-conversion carries out the half-band filter processing, gets the multicarrier peak noise signal before offsetting to the end.

Images