CN103188199B - A kind of broadband despicking method and device based on fraction delay CPG - Google Patents
A kind of broadband despicking method and device based on fraction delay CPG Download PDFInfo
- Publication number
- CN103188199B CN103188199B CN201310096532.5A CN201310096532A CN103188199B CN 103188199 B CN103188199 B CN 103188199B CN 201310096532 A CN201310096532 A CN 201310096532A CN 103188199 B CN103188199 B CN 103188199B
- Authority
- CN
- China
- Prior art keywords
- peak
- information
- phase
- value
- cpg
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
Abstract
A kind of broadband despicking method and device, methods described based on fraction delay CPG include：After the signal N times of pre- interpolation of progress of the low sampling rate of input, polar coordinates conversion is carried out, is converted into the amplitude information and phase information of polar domain；Wherein, N is positive integer；After carrying out peak value searching calculating to amplitude information, N times is carried out to the peak information searched and is extracted；The CPG filter pulses for being triggered CPG_LUTN successively using the peak information extracted are exported, and are extracted fractional phase information and are used for selecting to control CPG sequence of the CPG_LUTN outputs corresponding to the fractional phase；Phase bit extraction is carried out to phase information, and rectangular co-ordinate conversion is carried out to the phase after extraction and peak information, obtains peak clipping cancellation pulses.Described device includes：N times of pre- differential filtering module, polar coordinates modular converter, peak value searching computing module, peak information abstraction module, peak value scheduler module, phase abstraction module and rectangular co-ordinate modular converter.
Description
Technical field
The present invention relates to moving communicating field, more particularly to one kind is based on fraction delay CPG (peak clipping cancellation pulses extraction meters
Calculate, Cancel Pulse Generator) broadband despicking method and device.
Background technology
With the development of mobile communication technology and industry, mobile communication base station sender is from one-segment single system product to more
The gradual transition in direction of frequency range multi-standard fusion, useful signal occupied bandwidth reach more than 100MHz, the original peak clipping side of industry
The processing clock rate that method needs is original more than several times, and very big choose is brought to the disposal ability of current hardware and chip
War.
Existing GSM is all using higher QPSK (the Quadrature Phase Shift of the availability of frequency spectrum
Keying, QPSK signal) and QAM (Quadrature Amplitude Modulation, quadrature amplitude modulation) etc.
Modulation system is modulated to the phase and amplitude of carrier wave, and the instantaneous power of output signal has larger fluctuation, and generation has
Larger PAPR (Peak to Average Power Ratio, the referred to as ratio of peak power and average power, peak-to-average force ratio)
Non- constant enveloped modulation signal.In addition, every antenna on the transceiver of 3G commercialization base station products carries multiple carrier waves, overloading
Ripple signal carries out linear superposition merging in digital intermediate frequency, so be there is also in antenna transmitting terminal due to caused by the superposition of peak peak
The phenomenon of stronger signal peak-to-average ratio.Peak power is too big, it is easy to and power amplifier is operated in inelastic region, it is stronger so as to produce
ACLR (Adjacent Channel Leakage power Ratio, adjacent channel leakage power ratio), and then reduce systematic function.
To reduce non-linear distortion, the instantaneous power of transmission signal is necessarily less than 1dB compression points, and this requires the mean power of signal to drop
Must be very low, cause the efficiency of power amplifier to reduce, and reduce the coverage of base station.Also, high PAPR can cause D/A (digital-to-analogues
Conversion) dynamic range of converter diminishes, if using high-resolution D/A converter, cost will greatly improve, if made
With the relatively low D/A converter of resolution ratio, quantizing noise can increase.
The purpose of peak clipping is to reduce the peak-to-average force ratio of signal, by certain technology larger in non-constant enveloped modulation signal
Peak value is pruned, and is improved the efficiency of power amplifier and is not influenceed power amplifier linearization index.Once reduce signal peak-to-average ratio, it is possible to
The output mean power of power amplifier improves, and efficiency also just improves.When power raising amount is less than peak-to-average force ratio reduction amount, work(can be allowed
Put and be operated in linear zone, then the linear index ACLR of power amplifier is also just more preferable.
Existing clipping algorithm is all molded cancellation technology using peak pulse substantially, i.e.,：First according to peak clipping Threshold detection
Obtain the peak value of multi-carrier signal after multiplex, will be above peak clipping thresholding peak value subtract after peak value threshold value with respective phase information
Merging reverts to right angle domain, and molding filtration is carried out to IQ (In-phase Quadrature, the inphase quadrature) pulse value in right angle domain
Peak clipping cancellation pulses are obtained, the original signal after being alignd with delay is offset.The effect of molding filtration is by peak clipping cancellation pulses
Frequency spectrum be limited in the band of original signal to keep the preferable ACLR of signal after peak clipping.
Fig. 1 show the basic schematic diagram of peak clipping flow, and input I/Q signal is divided into two-way, enters line delay wait all the way,
Another way is used to ask for peak clipping cancellation pulses, two paths of signals alignment is subtracted each other can obtain peak clipping after export I/Q signal.
As can be seen that CPG is part most crucial during peak clipping.Fig. 2 show peak clipping the most frequently used at present and offsets arteries and veins
Punching produces flow chart.I/Q signal is amplitude and phase by Coordinate Conversion, by the slope variation for judging adjacent four amplitude points
Find peak information, peak information and peak clipping thresholding are subtracted each other can obtain needing the amplitude information of peak pulse offset and
It is corresponding to offset positional information.For the amplitude of peak pulse in the wide range that frequency domain is diffusion, the processing after this is exactly to pass through one
The very high digital fabrication wave filter of exponent number, it would be desirable to which the peak pulse of counteracting is limited in signal band.Because data are adopted during peak clipping
Sample speed height causes the exponent number of digital fabrication wave filter very high, therefore the characteristic use for being generally interrupted single appearance using peak value is more
Individual CPG_LUT (Look-Up-Table, the look-up table) pattern in parallel searched and multiplied accumulating realizes digital fabrication wave filter to save
Resource-saving.What is stored in CPG_LUT is filter coefficient, triggers CPU_LUT output filter coefficients successively according to peak
Pulse, coefficient pulse are multiplied and added up with the corresponding peak amplitude IQ information that right angle domain is transformed to by sentencing the accuser to the punishment facing the person he falsely accused mark, so i.e.
It can obtain last peak clipping cancellation pulses.On the one hand the effect of peak value scheduling is that certain screening, less peak are carried out to peak value
Value is typically directly removed to reduce amount of calculation, be on the other hand to trigger CPG_LUT positional information carry out some scheduling with
Congestion is prevented, the CPG_LUT of CPG filter pulses is being exported and can not be interrupted and export again, otherwise CPG filter pulses meeting
There is undesirable radio-frequency component.
As shown in figure 3, caused by being calculated in CPG filter coefficients in flow, protype_ci (wherein, i=1,2 ...,
N, n are positive integer) for the ptototype filter coefficient for meeting spectral emission template of corresponding each carrier wave, ptototype filter with it is corresponding
The frequency exp (j*wi*n) and power factor P_ci of carrier wave, which are multiplied and added up, can obtain CPG filter complex coefficients.
To obtain preferable peak clipping performance, current algorithm requires that data sample rates are not less than useful signal institute during peak clipping
Account for 5 times of bandwidth.If being less than this ratio, continue to upconvert to radio frequency letter according to signal after above-mentioned pulse cancellation algorithms peak clipping
PAPR has larger recovery after number, causes performance to meet to require substantially.
With explosive growth of the mobile subscriber to data traffic requirement, requirement of the mobile communications network to bandwidth is increasingly
Width, mobile communication is from 3G (3rd-generation, 3rd generation mobile communication technology) to LTE (the Long Term of wider band
Evolution, Long Term Evolution) transition.In addition, 2G GSM (Global System of Mobile communication, entirely
Ball mobile communcations system) network also possesses a large number of users, and it is in this case reduction mobile communication equipment hardware cost, it is necessary to base
Product of standing can support the mobile network of the different systems of multiple frequency ranges simultaneously, and the bandwidth for causing base station products needs to be supported is more next
It is wider.
Therefore, broadband is the most significant trend of current mobile communications network, under the conditions of 100MHz bandwidth applications, peak clipping
When data sample rates to reach 500MHz, this work schedule and power consumption to existing digital device etc. proposes very big choose
War.In addition, data rate is higher during peak clipping, the digital fabrication wave filter relative bandwidth being designed according to carrier bandwidths is smaller,
This for digital fabrication wave filter design and exploitation bring very big challenge, it is difficult to ensure that after peak clipping signal ACLR characteristics.
The content of the invention
The invention discloses a kind of broadband despicking method and device based on fraction delay CPG, to overcome existing peak clipping side
Method is too high to digital requirement on devices under the conditions of bandwidth applications the defects of.
To solve the above problems, the invention provides a kind of broadband despicking method based on fraction delay CPG, including：
After the signal N times of pre- interpolation of progress of the low sampling rate of input, polar coordinates conversion is carried out, is converted into polar domain
Amplitude information and phase information；Wherein, N is positive integer；
After carrying out peak value searching calculating to the amplitude information, N times is carried out to the peak information searched and is extracted；Using taking out
The peak information of taking-up triggers CPG_LUTN CPG filter pulses output successively, extracts fractional phase information and is used for selecting to control
CPG sequence of the CPG_LUTN outputs processed corresponding to the fractional phase；
Phase bit extraction is carried out to the phase information, and rectangular co-ordinate is carried out to the phase after extraction and peak information and turned
Change, obtain peak clipping cancellation pulses.
Further, methods described also includes：
After phase bit extraction is completed, Fs/4 shift frequency conversions are carried out to phase information；Wherein, Fs is the sample frequency of signal；
Before output offset pulse, the conversion of inverted-F s/4 shift frequencies is carried out to signal.
Further,
N value is any one in 2,3 and 4.
Further,
The progress polar coordinates conversion refers to use general Coordinate Rotation Digital iterative algorithm by the I/Q signal in right angle domain
It is transformed into the amplitude and phase of polar domain.
Further,
The amplitude information of described pair of output carries out peak value searching calculating, specifically includes：
Range value takes range value to subtract peak for current amplitude value if less than taking 0 if peak clipping thresholding if peak clipping thresholding is higher than
It is worth the difference of detection threshold；
In the range value obtained after treatment, the search of a minor peaks is carried out in per continuous 4 sampled points, continuous
In 4 sampled points A, B, C, D, if B ＞ A and C ＞=D, B and larger one peak value of conduct of C intermediate values are taken, determines the sampling
The value of point subtracts the difference after peak clipping thresholding to need peak value to be processed, and the value of other sampled points in this 4 sampled points is put
For 0.
Further,
If being unsatisfactory for B ＞ A and C ＞=D this condition, the value of this 4 sampled points is set to 0.
Further,
The progress peak information extraction, is specifically included：
Compare the range value of two points of high velocity of peak value correspondence low rate point, take the big position of range value to extract
Peak afterwards, and compare extract before and after peak value sampling point sampling interval obtain extract fractional phase information.
Further,
The phase information of described pair of output carries out phase bit extraction, specifically includes：
The peak point for taking phase value corresponding to peak value point to correspond to after extracting.
Further,
The CPG filter pulses output for being triggered CPG_LUTN successively using the peak information extracted, is specifically included：
It is described to be multiplied and tire out with the frequency exp (j*wi*n) and power factor P_ci of corresponding carrier wave using ptototype filter
After adding, using Farrow wave filters to CPG coefficients carry out fraction delay, fraction delay value be followed successively by 0,1/N, 2/N ..., N-1/
N。
Correspondingly, present invention also offers a kind of broadband peak cutting device based on fraction delay CPG, including：
N times of pre- differential filtering module, the signal for the low sampling rate to input carry out N times of pre- interpolation；Wherein, N is
Positive integer；
Polar coordinates modular converter, for carrying out polar coordinates conversion to the signal of N times of pre- differential filtering module output, it is converted into
The amplitude information and phase information of polar domain；
Peak value searching computing module, the amplitude information for being exported to the polar coordinates modular converter carry out peak value and searched
Rope calculates；
Peak information abstraction module, the peak information for being searched to the peak value searching computing module carry out N times and taken out
Take；
Peak value scheduler module, the peak information for being extracted using the peak information abstraction module trigger CPG_ successively
LUTN CPG filter pulses output, extract fractional phase information and be used for selecting to control CPG_LUTN outputs to correspond to the fraction
The CPG sequences of phase；
Phase abstraction module, the phase information for being exported to the polar coordinates modular converter carry out phase bit extraction；
Rectangular co-ordinate modular converter, for carrying out rectangular co-ordinate conversion to the phase after extraction and peak information, cut
Peak cancellation pulses.
Further,
The phase abstraction module is additionally operable to after phase bit extraction is completed, and Fs/4 shift frequency conversions are carried out to phase information；Its
In, Fs is the sample frequency of signal；
Also include in described device：
Inverted-F s/4 shift frequency modules, for before output offset pulse, the conversion of inverted-F s/4 shift frequencies to be carried out to signal.
Further,
N value is any one in 2,3 and 4.
Further,
The polar coordinates modular converter is used to use general Coordinate Rotation Digital iterative algorithm by the I/Q signal in right angle domain
It is transformed into the amplitude and phase of polar domain.
Further,
The peak value searching computing module is used to carry out peak to the amplitude information of polar coordinates modular converter output
Value search calculates, and specifically includes：
The peak value searching computing module is used to take 0 if judging that range value is less than peak clipping thresholding, if being higher than peak clipping door
It is the difference that current amplitude value subtracts peakvalue's checking thresholding that limit, which then takes range value,；
It is additionally operable in the range value for obtaining after treatment, a minor peaks is carried out in per continuous 4 sampled points
Search, in continuous 4 sampled points A, B, C, D, if B ＞ A and C ＞=D, takes B and larger one peak of conduct of C intermediate values
Value, determine that the value of the sampled point subtracts the difference after peak clipping thresholding to need peak value to be processed, and will be other in this 4 sampled points
The value of sampled point is set to 0.
Further,
The peak value searching computing module is additionally operable to when judging to be unsatisfactory for B ＞ A and this condition of C ＞=D, by this 4
The value of individual sampled point is set to 0.
Further,
Peak information abstraction module, the peak information for being searched to the peak value searching computing module carry out N times and taken out
Take, specifically include：
The peak information abstraction module, the amplitude of two points of high velocity for comparing peak value correspondence low rate point
Value, take the big position of range value as the peak after extracting, and compare extract before and after sampling interval of peak value sampling point obtain
Extract fractional phase information.
Further,
Phase abstraction module, the phase information for being exported to the polar coordinates modular converter carry out phase bit extraction,
Specifically include：
The phase abstraction module is used for the peak point for taking phase value corresponding to peak value point to correspond to after extracting.
Further,
The peak value scheduler module is used for the CPG wave filter arteries and veins for triggering CPG_LUTN successively using the peak information extracted
Punching output, is specifically included：
The peak value scheduler module is used for the frequency exp (j*wi*n) and power with corresponding carrier wave using ptototype filter
After factor P_ci is multiplied and added up, fraction delay is carried out to CPG coefficients using Farrow wave filters, fraction delay value is followed successively by 0,
1/N、2/N、...、N-1/N。
The present invention greatly reduces requirement of the peak clipping processing to digital hardware and chip processing speed.
Brief description of the drawings
Fig. 1 is peak clipping technology entire block diagram in the prior art；
Fig. 2 is peak clipping cancellation pulses extraction block diagram general at present；
Fig. 3 is that CPG filter coefficients general at present calculate schematic diagram；
Fig. 4 is the peak clipping cancellation pulses extraction added in the embodiment of the present invention after fraction delay process and the processing of Fs/4 shift frequencies
FB(flow block)；
Fig. 5 is that high speed rate of embodiment of the present invention peak value extracts and fractional phase delayed data produces schematic diagram；
Fig. 6 is the CPG filter coefficient meters added in the embodiment of the present invention after fraction filtering wave by prolonging time and the processing of Fs/4 shift frequencies
Calculate schematic flow sheet；
Fig. 7 is the amplitude of CPG coefficients and reciprocal fraction delayed data schematic diagram in the case of N=4 in the embodiment of the present invention；
Fig. 8 is the broadband peak cutting device structure chart based on fraction delay CPG in the embodiment of the present invention.
Embodiment
For the object, technical solutions and advantages of the present invention are more clearly understood, below in conjunction with accompanying drawing to the present invention
Embodiment be described in detail.It should be noted that in the case where not conflicting, in the embodiment and embodiment in the application
Feature can mutually be combined.
Peak clipping cancellation pulses extraction of the present invention is as shown in Figure 4 with producing flow.Here I/Q signal sampling speed is inputted
Rate is only 2.5 times of signal occupied bandwidth or so.Compared with the prior art shown in Fig. 2, mainly there is these points improvement
And difference：
1) in order to find accurate amplitude cancellation information, it is necessary to the signal of the low sampling rate of input is carried out N times it is pre-
Interpolation, interpolation filter is needed after interpolation and suppresses image signal；
2) phase bit extraction and Fs/4 shift frequencies modular converter are increased, it is necessary to which the phase of signal after interpolation n times is drawn into interpolation
In preceding speed, the conversion of Fs/4 shift frequencies is in order to which peak pulse signal is moved onto Fs/4 from zero-frequency；
3) increase peak information and extract processing module, the sampling rate of peak value is reduced by the high sampling rate after pre- interpolation
Onto the low sampling rate of input, and fractional phase information when extracting extraction；
4) bandpass center of CPG wave filters needs to move Fs/4 frequencies by zero-frequency, and its coefficient is no longer plural but real
Number, correspondingly behind it is multiply-add be calculated as that real number is multiply-add rather than plural multiply-add, different extraction fractional phases is corresponding using not
Same CPG coefficients, so CPG_LUTN is two-dimentional real number look-up table, including N groups correspond to the CPG real numbers of different extraction fractional phases
Coefficient；
5) centre frequency of cancellation pulses is Fs/4 after processing above, so needing increase one-level anti-last
The processing of Fs/4 shift frequencies, and filter out image signal using half-band filter.
In the present embodiment, it is relatively low to input the sampling rate of I/Q signal, however, to ensure that preferably peak clipping performance, peak value
Search is calculated to still need and carried out in higher sampling rate, generally requires the sampling speed at least 5 times signal occupied bandwidths
Carried out in rate, so needing to carry out N times of pre- interpolation to input I/Q signal before peak value searching.N takes 2,3 or 4, typically not
Need more than 4.Because interpolation is to extract peak clipping offseting signal, image-reject filter need not be according to transmitting chain
The mirror image of Digital Up Convert suppresses requirement and is designed, and mirror image, which suppresses requirement, to require drop than composing template as defined in standard agreement
Low 10dB to 20dB, therefore the hardware costs of this method interpolation is relatively small.
Signal after interpolation (COordinate Rotation DIgital Computer, is sat using existing CORDIC
Mark rotary digital computational methods) I/Q signal in right angle domain is transformed into the amplitude and phase of polar domain by iterative algorithm.Emulation and
It is sufficient that computational accuracy of test 8 CORDIC iteration of display.
After I/Q signal is converted into polar coordinates, amplitude and phase information are divided into two-way and are respectively processed.If range value
It is 0 then to take range value less than peak clipping thresholding, takes range value to subtract peakvalue's checking door for current amplitude value if peak clipping thresholding is higher than
The difference of limit；Wherein, the value of peakvalue's checking thresholding is more than or equal to the value of peak clipping thresholding.Then the amplitude to obtaining after treatment
In value, the search of a minor peaks is carried out in per continuous 4 sampled points, i.e., judges peak from every 4 continuous sampling point A, B, C, D
The position that value is likely to occur, if meet B ＞ A and C ＞=D, it is a peak value to take higher value in B, C, determines the sampled point
Value subtract the difference after peak clipping thresholding to need peak value to be processed, and other points in this 4 points are set to zero；If no
Meet B ＞ A and C ＞=D this condition, then this 4 sampled points are set to 0.Peak information is cut except signal envelope is more than
Remaining outer point of the peak point of peak thresholding is zero real-time stream.
Peak information needs N times of progress to be drawn into low sampling rate so as to subsequent treatment, as shown in fig. 5, it is assumed that during N=4,
The peak information point of corresponding 4 high-speeds of 1 low rate point, peak information extract 4 times of post-sampling rate reduction, and peak value extracts mould
The fractional phase information of high-speed data before block output increase is extracted, fractional phase information corresponding -3 to 3, i.e.-N+1 arrives N-1.Compare
Pre-and post-peaking corresponds to the range value of two high sampling rates of low rate point, takes the big position of range value as the peak value position after extraction
Put, and compare extract before and after peak value sampling point sampling interval obtain extract fractional phase information.Such as peak value P1 points in Fig. 5,
Its left and right two low rates sampled value A00 and A10 are compared in observation in high-speed range value, because A00 >=A10, after extracting
Peak point is placed in A00 positions, using A00 positions as zero point, compare extraction before and after peak obtain extract fraction delayed data be
2.Remaining several example points the like.From fig. 5, it can be seen that if peak intervals are less than 2*N under high sampling rate before extracting, that
Extraction operation can not continue or can produce two peak points continuously occurred, and such case needs to avoid the occurrence of, institute
To need to have certain processing to high sampling rate peak value before peak value extracts, remove relatively small peak value and cause between two peak values
Sampling interval be more than 2*N.
Phase information is also required to carry out extraction processing.It is similar with peak value extraction process, take phase value corresponding to peak value point
The peak point corresponded to after extracting, fractional phase information is extracted without repeating to record.The purpose of Fs/4 shift frequencies is by peak value
Information is moved on at Fs/4 from zero-frequency, and exp (j*Fs/4*2* π * n/Fs), i.e. exp are multiplied by equivalent to original signal in rectangular co-ordinate domain
(j* π * n/2), in polar domain, 0, pi/2, π and π * 3/2 is added successively for circulation according to 4 equivalent in phase information, is passed through
Some phase values can exceed 2* π after additional calculation, so the phase information for needing that circled addition is calculated is carried out to 2* π's
Complementation is handled to be normalized to 0 scope for arriving 2* π.
Peak amplitude information after extraction is transformed into the IQ pulses in right angle domain, IQ arteries and veins with phase information by CORDIC iteration
Punching obtains peak clipping cancellation pulses by CPG filter process.CPG handles equivalent to one digital FIR filter in dotted line frame in Fig. 4
Ripple device, is operated in that complex field is different from multiplication in Fig. 2 and addition, and multiplication here and additional calculation are in real number field, meter
Calculation amount reduces half.
The positional information for extracting amplitude is used for triggering CPG_LUTN CPG filter pulses output successively, extracts fraction
Phase information is used for selecting to control CPG sequence of the CPG_LUTN outputs corresponding to the fractional phase.
CPG_LUTN is a two-dimensional look-up table, and storage is N group CPG filter coefficients.Fractional phase value is from-(N-1)
To the common 2*N-1 groups of N-1, wherein during upper poor 1 sampling interval of opposite number CPG coefficients only integer delay as corresponding to -1 and+1 each other
Between, sequence of values value is identical, so the actual only N groups of corresponding CPG coefficients.Negative value fractional phase than accordingly on the occasion of
CPG coefficients corresponding to fractional phase propose previous sampled point triggering and produced.
The calculating of CPG coefficients unlike Fig. 3 after read group total as shown in fig. 6, add one-level Farrow wave filters
To CPG coefficients carry out fraction delay, fraction delay value be followed successively by 0,1/N, 2/N ..., N-1/N, reciprocal fraction phase is 0~N-
1.Fraction delay calculates generally to be realized using general Farrow wave filter technologies.Fs/4 is carried out again after fraction delay
Shift frequency, which handles and takes real part that center frequency point is located to the complex coefficient of zero-frequency, is converted to the real number that center frequency point is Fs/4.The same institute
State, exp (j* π * n/2) is the loop cycle of (1, j, -1,-j) this 4 values, takes Fs/4 processing to be converted into 4 to a plural number
Real imaginary part for the cycle exchanges and negates processing, needs guarantee wave filter middle coefficient to be counted with exp (j* π * 0/2) during processing
Calculate.Plural number after Fs/4 shift frequencies takes real part to produce CPG wave filter real number coefficients to the end.In the case of N=4 as shown in Figure 7
The CPG coefficient amplitudes figure and data peaks magnified partial view that certain application scenarios calculates, -3 and 3, -2 and 2, -1 and 1 distinguish
Corresponding CPG factor vs are the same, and simply sampling instant is advanced by a sampled point.CPG coefficients can be at general place
Manage to calculate using software program on device and complete, completion can also be calculated in real time using digital hardware circuits, the latter is usually applied to
The application scenario of real-time change occurs with power for carrier signal frequency.CPG coefficients are stored in CPG_LUTN storage tables after the completion of calculating
In lattice.CPG_LUTN storages form number has certain change with the difference of application scenarios, typically takes 6 to 10, input signal peak
CPG_LUTN than needing in the case of higher is more.
The cancellation pulses that peak I Q pulses are located at Fs/4 CPG real filters processing generation by bandpass center are located at
Fs/4 positions, there can be mirror image in-Fs/4 positions in addition, it is necessary to suppress the place of mirror image by inverted-F s/4 shift frequencies and half-band filter
Reason.Inverted-F s/4 processing needs to be multiplied by signal exp (- j*Fs/4*2* π * n/Fs), and this is 4 for circulation with (1,-j, -1, j)
Individual constant, i.e., it is the cycle to the real imaginary part exchange of complex data progress with 4 and negates processing.The mirror image after inverted-F s/4 processing
Position is located at Fs/2, and mirror image suppression is carried out using relatively simple general half-band filter.By this last inverted-F s/
4 shift frequencies and half band mirror image suppress, that is, have obtained the peak clipping cancellation pulses of last needs.
The above method make in two dimensions of amplitude and time counteracting shaping pulse alignd with original signal, this peak cutting device with
Only 2.5 times or so of the ratio of most important process part signal sampling speed and occupied bandwidth in method, i.e., than public at present
The method opened reduces by one times.In addition, be further to reduce requirement of the peak cutting device to hardware resource, signal is done Fs/4 shift frequencies and
Fs/4 backward shifts frequency, so as to which complex signal to be changed into the technology of real number signal, greatly reduce peak cutting device and hardware circuit is provided
The requirement in source.Here Fs is the sample frequency of signal, and the frequency domain for the complex digital signal that sample frequency is Fs arrives for-Fs/2
Fs/2, Fs/4 are the center of corresponding positive frequency.
A kind of digital hardware circuits device disclosed by the invention mainly applied to sender, it may be directly applied to intermediate frequency
Among ASIC or FPGA design.
Assuming that input signal occupied bandwidth is 60MHz, then baseband signal combining is upconverted into 2.5 times of occupied bandwidths
Sampling rate is input to this peak cutting device.
It is further assumed that the IQ sampling rates of the input present apparatus are such as 153.6MHz, and pre- interpolation multiple is 4.
4 times of the pre- interpolation of I/Q signal carries out polar coordinates conversion and peak value searching to 614.4MHz using CORDIC iterative algorithms.
Here pre- interpolation filter rents band mirror image suppression 60dBc, and 614.4MHz sampling rates are higher, more phase methods can be used to drop
Fast parallel computation.
Amplitude and phase are extracted according to Fig. 5 method, and extract fraction delayed data, and Fs/4 is carried out to phase
Shift frequency processing.
According to practical application scene prototype filter coefficient and Farrow filter coefficients, the stream shown in Fig. 6 is utilized
CPG filter coefficients are calculated in journey figure, and coefficient is configured to the CPG_LUTN shown in Fig. 4.CPG_ in the case of being normally applied
LUTN form numbers take 8 cans.
Peak amplitude and phase after extraction carry out rectangular co-ordinate using cordic algorithm and are converted to IQ peak pulses.
Peak and fractional phase delayed data trigger corresponding CPG filter pulses and are multiplied with IQ peak pulses successively
And add up and obtain cancellation pulses.This cancellation pulses suppresses mirror image by inverted-F s/4 shift frequencies and half-band filter and produces cutting to the end
Peak cancellation pulses.
Input I/Q signal after delay alignment subtracts peak value offset pulse and obtains peak clipping output signal.
As shown in figure 8, a kind of broadband peak cutting device based on fraction delay CPG, including：
N times of pre- differential filtering module, the signal for the low sampling rate to input carry out N times of pre- interpolation；Wherein, N is
Positive integer；
Polar coordinates modular converter, for carrying out polar coordinates conversion to the signal of N times of pre- differential filtering module output, it is converted into
The amplitude information and phase information of polar domain；
Peak value searching computing module, the amplitude information for being exported to the polar coordinates modular converter carry out peak value and searched
Rope calculates；
Peak information abstraction module, the peak information for being searched to the peak value searching computing module carry out N times and taken out
Take；
Peak value scheduler module, the peak information for being extracted using the peak information abstraction module trigger CPG_ successively
LUTN CPG filter pulses output, extract fractional phase information and be used for selecting to control CPG_LUTN outputs to correspond to the fraction
The CPG sequences of phase；
Phase abstraction module, the phase information for being exported to the polar coordinates modular converter carry out phase bit extraction；
Rectangular co-ordinate modular converter, for carrying out rectangular co-ordinate conversion to the phase after extraction and peak information, cut
Peak cancellation pulses.
It is preferred that
The phase abstraction module is additionally operable to after phase bit extraction is completed, and Fs/4 shift frequency conversions are carried out to phase information；Its
In, Fs is the sample frequency of signal；
Also include in described device：
Inverted-F s/4 shift frequency modules, for before output offset pulse, the conversion of inverted-F s/4 shift frequencies to be carried out to signal.
It is preferred that
N value is any one in 2,3 and 4.
It is preferred that
The polar coordinates modular converter is used to use general Coordinate Rotation Digital iterative algorithm by the I/Q signal in right angle domain
It is transformed into the amplitude and phase of polar domain.
It is preferred that
The peak value searching computing module is used to carry out peak to the amplitude information of polar coordinates modular converter output
Value search calculates, and specifically includes：
The peak value searching computing module is used to take 0 if judging that range value is less than peak clipping thresholding, if being higher than peak clipping door
It is the difference that current amplitude value subtracts peakvalue's checking thresholding that limit, which then takes range value,；
It is additionally operable in the range value for obtaining after treatment, a minor peaks is carried out in per continuous 4 sampled points
Search, in continuous 4 sampled points A, B, C, D, if B ＞ A and C ＞=D, takes B and larger one peak of conduct of C intermediate values
Value, determine that the value of the sampled point subtracts the difference after peak clipping thresholding to need peak value to be processed, and will be other in this 4 sampled points
The value of sampled point is set to 0.
It is preferred that
The peak value searching computing module is additionally operable to when judging to be unsatisfactory for B ＞ A and this condition of C ＞=D, by this 4
The value of individual sampled point is set to 0.
It is preferred that
Peak information abstraction module, the peak information for being searched to the peak value searching computing module carry out N times and taken out
Take, specifically include：
The peak information abstraction module, the amplitude of two points of high velocity for comparing peak value correspondence low rate point
Value, take the big position of range value as the peak after extracting, and compare extract before and after sampling interval of peak value sampling point obtain
Extract fractional phase information.
It is preferred that
Phase abstraction module, the phase information for being exported to the polar coordinates modular converter carry out phase bit extraction,
Specifically include：
The phase abstraction module is used for the peak point for taking phase value corresponding to peak value point to correspond to after extracting.
It is preferred that
The peak value scheduler module is used for the CPG wave filter arteries and veins for triggering CPG_LUTN successively using the peak information extracted
Punching output, is specifically included：
The peak value scheduler module is used for the frequency exp (j*wi*n) and power with corresponding carrier wave using ptototype filter
After factor P_ci is multiplied and added up, fraction delay is carried out to CPG coefficients using Farrow wave filters, fraction delay value is followed successively by 0,
1/N、2/N、...、N-1/N。
One of ordinary skill in the art will appreciate that all or part of step in the above method can be instructed by program
Related hardware is completed, and described program can be stored in computer-readable recording medium, such as read-only storage, disk or CD
Deng.Alternatively, all or part of step of above-described embodiment can also be realized using one or more integrated circuits.Accordingly
Ground, each module/unit in above-described embodiment can be realized in the form of hardware, can also use the shape of software function module
Formula is realized.The present invention is not restricted to the combination of the hardware and software of any particular form.
The preferred embodiments of the present invention are the foregoing is only, are not intended to limit the scope of the present invention.According to
The content of the invention of the present invention, can also there is other various embodiments, in the case of without departing substantially from spirit of the invention and its essence, be familiar with
Those skilled in the art works as can make various corresponding changes and deformation according to the present invention, all in the spirit and principles in the present invention
Within, any modification, equivalent substitution and improvements made etc., it should be included in the scope of the protection.
Claims (14)
1. a kind of broadband despicking method based on fraction delay CPG, including：
After the signal N times of pre- interpolation of progress of the low sampling rate of input, polar coordinates conversion is carried out, is converted into the width of polar domain
Spend information and phase information；Wherein, N 2,3 or 4；
After carrying out peak value searching calculating to the amplitude information, N times is carried out to the peak information searched and is extracted；Using extracting
Peak information trigger CPG_LUTN CPG filter pulses output successively, extract fractional phase information and be used for selecting to control
CPG sequence of the CPG_LUTN outputs corresponding to the fractional phase；
Phase bit extraction is carried out to the phase information, and rectangular co-ordinate conversion is carried out to the phase after extraction and peak information, is obtained
To peak clipping cancellation pulses；
Wherein, it is described that peak value searching calculating is carried out to the amplitude information, specifically include：
Range value takes range value to subtract peak value inspection for current amplitude value if less than taking 0 if peak clipping thresholding if peak clipping thresholding is higher than
Survey the difference of thresholding；
In the range value obtained after treatment, the search of a minor peaks is carried out in per continuous 4 sampled points, at continuous 4
In sampled point A, B, C, D, if B>A and C>=D, then B and larger one peak value of conduct of C intermediate values are taken, determines the sampled point
Value subtracts the difference after peak clipping thresholding to need peak value to be processed, and the value of other sampled points in this 4 sampled points is set into 0.
2. the method as described in claim 1, it is characterised in that also include：
After phase bit extraction is completed, Fs/4 shift frequency conversions are carried out to phase information；Wherein, Fs is the sample frequency of signal；
Before output offset pulse, the conversion of inverted-F s/4 shift frequencies is carried out to signal.
3. the method as described in claim 1, it is characterised in that：
The polar coordinates that carry out change the I/Q signal conversion referred to using general Coordinate Rotation Digital iterative algorithm by right angle domain
To the amplitude and phase of polar domain.
4. the method as described in claim 1, it is characterised in that：
If it is unsatisfactory for B>A and C>This condition of=D, then be set to 0 by the value of this 4 sampled points.
5. the method as described in claim 1, it is characterised in that：
The progress peak information extraction, is specifically included：
Compare the range value of two points of high velocity of peak value correspondence low rate point, after taking the big position of range value to extract
Peak, and compare extract before and after peak value sampling point sampling interval obtain extract fractional phase information.
6. the method as described in claim 1 or 5, it is characterised in that：
The phase information of described pair of output carries out phase bit extraction, specifically includes：
The peak point for taking phase value corresponding to peak value point to correspond to after extracting.
7. the method as described in claim 1, it is characterised in that：
The CPG filter pulses output for being triggered CPG_LUTN successively using the peak information extracted, is specifically included：
It is described to be multiplied and add up with the frequency exp (j*wi*n) and power factor P_ci of corresponding carrier wave using ptototype filter
Afterwards, using Farrow wave filters to CPG coefficients carry out fraction delay, fraction delay value be followed successively by 0,1/N, 2/N ..., N-1/N.
8. a kind of broadband peak cutting device based on fraction delay CPG, including：
N times of pre- differential filtering module, the signal for the low sampling rate to input carry out N times of pre- interpolation；Wherein, N 2,3 or
Person 4；
Polar coordinates modular converter, for carrying out polar coordinates conversion to the signal of N times of pre- differential filtering module output, conversion poling is sat
Mark the amplitude information and phase information in domain；
Peak value searching computing module, based on the amplitude information progress peak value searching exported to the polar coordinates modular converter
Calculate；
Peak information abstraction module, the peak information for being searched to the peak value searching computing module carry out N times and extracted；
Peak value scheduler module, the peak information for being extracted using the peak information abstraction module trigger CPG_LUTN successively
The output of CPG filter pulses, extract fractional phase information and be used for selecting to control CPG_LUTN outputs to correspond to the fractional phase
CPG sequences；
Phase abstraction module, the phase information for being exported to the polar coordinates modular converter carry out phase bit extraction；
Rectangular co-ordinate modular converter, for carrying out rectangular co-ordinate conversion to the phase after extraction and peak information, obtain peak clipping and support
Disappear pulse；
Wherein, the peak value searching computing module, the amplitude information for being exported to the polar coordinates modular converter are carried out
Peak value searching calculates, and specifically includes：
The peak value searching computing module is used to take 0 if judging that range value is less than peak clipping thresholding, if higher than if peak clipping thresholding
Range value is taken to subtract the difference of peakvalue's checking thresholding for current amplitude value；
It is additionally operable in the range value for obtaining after treatment, the search of a minor peaks is carried out in per continuous 4 sampled points,
In continuous 4 sampled points A, B, C, D, if B>A and C>=D, then B and larger one peak value of conduct of C intermediate values are taken, it is determined that should
The value of sampled point subtracts the difference after peak clipping thresholding to need peak value to be processed, and by other sampled points in this 4 sampled points
Value is set to 0.
9. device as claimed in claim 8, it is characterised in that：
The phase abstraction module is additionally operable to after phase bit extraction is completed, and Fs/4 shift frequency conversions are carried out to phase information；Wherein, Fs
For the sample frequency of signal；
Also include in described device：
Inverted-F s/4 shift frequency modules, for before output offset pulse, the conversion of inverted-F s/4 shift frequencies to be carried out to signal.
10. device as claimed in claim 8, it is characterised in that：
The polar coordinates modular converter is used to change the I/Q signal in right angle domain using general Coordinate Rotation Digital iterative algorithm
To the amplitude and phase of polar domain.
11. device as claimed in claim 8, it is characterised in that：
The peak value searching computing module is additionally operable to judging to be unsatisfactory for B>A and C>During this condition of=D, by this 4 samplings
The value of point is set to 0.
12. device as claimed in claim 8, it is characterised in that：
Peak information abstraction module, the peak information for being searched to the peak value searching computing module carry out N times and extracted, have
Body includes：
The peak information abstraction module, the range value of two points of high velocity for comparing peak value correspondence low rate point,
The big position of range value is taken as the peak after extracting, and compare extract before and after sampling interval of peak value sampling point extracted
Fractional phase information.
13. the device as described in claim 8 or 12, it is characterised in that：
Phase abstraction module, the phase information for being exported to the polar coordinates modular converter carry out phase bit extraction, specifically
Including：
The phase abstraction module is used for the peak point for taking phase value corresponding to peak value point to correspond to after extracting.
14. device as claimed in claim 8, it is characterised in that：
The CPG filter pulses that the peak value scheduler module is used to trigger CPG_LUTN successively using the peak information extracted are defeated
Go out, specifically include：
The peak value scheduler module is used for the frequency exp (j*wi*n) and power factor with corresponding carrier wave using ptototype filter
After P_ci is multiplied and added up, using Farrow wave filters to CPG coefficients carry out fraction delay, fraction delay value be followed successively by 0,1/N,
2/N、…、N-1/N。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310096532.5A CN103188199B (en) | 2013-03-22 | 2013-03-22 | A kind of broadband despicking method and device based on fraction delay CPG |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310096532.5A CN103188199B (en) | 2013-03-22 | 2013-03-22 | A kind of broadband despicking method and device based on fraction delay CPG |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103188199A CN103188199A (en) | 2013-07-03 |
| CN103188199B true CN103188199B (en) | 2018-02-16 |
Family
ID=48679168
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310096532.5A Active CN103188199B (en) | 2013-03-22 | 2013-03-22 | A kind of broadband despicking method and device based on fraction delay CPG |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103188199B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2786129C1 (en) * | 2019-05-30 | 2022-12-19 | Зте Корпорейшн | Method and device based on a polyphase structure to reduce peak load and a computer information carrier |
Families Citing this family (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9178740B1 (en) * | 2014-08-26 | 2015-11-03 | Ibiquity Digital Corporation | Peak-to-average power ratio reduction for QAM modulation with HD radio signals |
| CN105245480B (en) * | 2015-08-27 | 2019-01-04 | 中兴通讯股份有限公司 | digital signal processing method and device |
| CN105891701B (en) * | 2016-06-08 | 2018-09-21 | 中国电子科技集团公司第三十八研究所 | A kind of score filtering wave by prolonging time device delay performance test method and its test device |
| WO2018010154A1 (en) * | 2016-07-15 | 2018-01-18 | 张升泽 | Electronic chip voltage modulation method and system |
| WO2018010156A1 (en) * | 2016-07-15 | 2018-01-18 | 张升泽 | Electronic chip signal modulation method and system |
| CN106230406A (en) * | 2016-07-15 | 2016-12-14 | 张升泽 | The signal modulating method of electronic chip and system |
| WO2018014184A1 (en) * | 2016-07-19 | 2018-01-25 | 张升泽 | Power modulation method and system for electronic chip |
| US10454747B1 (en) * | 2018-09-07 | 2019-10-22 | Innophase, Inc. | Systems and methods for up-sampling a polar amplitude sample stream in a polar modulator |
| CN111107030B (en) * | 2018-10-25 | 2022-07-12 | 北京新岸线移动多媒体技术有限公司 | Method and device for reducing signal peak-to-average power ratio suitable for large bandwidth system |
| CN111107031B (en) * | 2018-10-25 | 2022-09-20 | 北京新岸线移动多媒体技术有限公司 | Method and device for reducing signal peak-to-average ratio |
| CN111343119B (en) * | 2018-12-18 | 2022-10-14 | 深圳市中兴微电子技术有限公司 | Data processing method and device and computer readable storage medium |
| CN111770042B (en) * | 2019-04-01 | 2021-12-07 | 天津工业大学 | Peak-to-average power ratio restraining method and device for generalized frequency division multiplexing system |
| CN112019471A (en) * | 2019-05-30 | 2020-12-01 | 中兴通讯股份有限公司 | Peak clipping method and device based on multiphase structure and computer storage medium |
| CN111107034B (en) * | 2019-12-26 | 2022-02-11 | 三维通信股份有限公司 | System, method and computer-readable storage medium for reducing signal peak-to-average ratio |
| CN113872621B (en) * | 2021-10-08 | 2022-04-12 | 星思连接(上海)半导体有限公司 | Signal processing apparatus, signal processing method, signal transmitter, and readable storage medium |
| CN114018468B (en) * | 2021-11-24 | 2024-04-12 | 华科电子股份有限公司 | Digital output atmospheric pressure sensor |
| CN114500211B (en) * | 2022-01-27 | 2023-03-24 | 重庆物奇微电子有限公司 | High-precision multiphase CFR system and method |
| CN117135644B (en) * | 2022-05-18 | 2024-06-18 | 中国移动通信有限公司研究院 | Peak cancellation processing method, device and computer readable storage medium |
| CN115551003B (en) * | 2022-12-05 | 2023-04-07 | 南京典格通信科技有限公司 | Resource optimization peak reduction CFR implementation system and implementation method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7423960B2 (en) * | 2003-12-27 | 2008-09-09 | Kuei-ann Wen | Orthogonal frequency division multiplexing (OFDM) equalizer |
| CN102075484A (en) * | 2011-01-18 | 2011-05-25 | 中兴通讯股份有限公司 | Method and device for reducing peak-to-average power ratio of signal |
| CN102170416A (en) * | 2010-12-27 | 2011-08-31 | 武汉邮电科学研究院 | Peak clipping structure and peak clipping method based on slip window peak detection |
-
2013
- 2013-03-22 CN CN201310096532.5A patent/CN103188199B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7423960B2 (en) * | 2003-12-27 | 2008-09-09 | Kuei-ann Wen | Orthogonal frequency division multiplexing (OFDM) equalizer |
| CN102170416A (en) * | 2010-12-27 | 2011-08-31 | 武汉邮电科学研究院 | Peak clipping structure and peak clipping method based on slip window peak detection |
| CN102075484A (en) * | 2011-01-18 | 2011-05-25 | 中兴通讯股份有限公司 | Method and device for reducing peak-to-average power ratio of signal |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2786129C1 (en) * | 2019-05-30 | 2022-12-19 | Зте Корпорейшн | Method and device based on a polyphase structure to reduce peak load and a computer information carrier |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103188199A (en) | 2013-07-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103188199B (en) | A kind of broadband despicking method and device based on fraction delay CPG | |
| CN101136890B (en) | Optimized multi-carrier signal slicing device and method therefor | |
| EP3345306B1 (en) | Interference phase estimate system and method | |
| CN102014090B (en) | Digital predistortion method and device | |
| CN105264781A (en) | Processing interference in a wireless network | |
| EP3968589A1 (en) | Peak clipping method and apparatus based on polyphase structure, and computer storage medium | |
| CN105553909B (en) | A kind of method that iteration limit filtration reduces ofdm signal peak-to-average force ratio | |
| CN201360268Y (en) | Peak clipping device based on peak detection | |
| CN103684570B (en) | Search GSM base station carrier frequencies method, device and repeater | |
| US9912357B1 (en) | Digital polar transmitter having a digital front end | |
| CN101370000B (en) | Multi-carrier implementing method and apparatus for TD-SCDMA system | |
| US8478325B2 (en) | Low sample rate peak power reduction | |
| Lin et al. | Efficient interpolation method for wireless communications and signal processing applications | |
| US8085890B2 (en) | Apparatus and method for base band filtering | |
| CN102223223B (en) | The method of reseptance and device of multi-carrier signal | |
| CN115299016B (en) | Multi-rate crest factor reduction | |
| CN102685055B (en) | Multi-data stream interpolation and extraction multiplexing device and method | |
| CN102104564B (en) | Channelized processing method for general ultra wideband reception | |
| EP1444814B1 (en) | Method and arrangement for limiting signal in radio transmitter | |
| Amulya et al. | Design and implementation of a reconfigurable digital down converter for 4G systems using MATLAB and FPGA-a review | |
| CN110650106B (en) | Airspace peak clipping device and method | |
| CN106603177A (en) | Channel fading model, channel signal generation method and simulation apparatus | |
| Li et al. | Statistical analysis of CDMA and 3G signal models | |
| US20150016573A1 (en) | System and method for passive wireless monitoring with efficient channelization | |
| CN113872621B (en) | Signal processing apparatus, signal processing method, signal transmitter, and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |