CN105207628A - Frequency domain linearization technique for power amplifier in OFDM (orthogonal frequency division multiplexing) system - Google Patents
Frequency domain linearization technique for power amplifier in OFDM (orthogonal frequency division multiplexing) system Download PDFInfo
- Publication number
- CN105207628A CN105207628A CN201510566527.5A CN201510566527A CN105207628A CN 105207628 A CN105207628 A CN 105207628A CN 201510566527 A CN201510566527 A CN 201510566527A CN 105207628 A CN105207628 A CN 105207628A
- Authority
- CN
- China
- Prior art keywords
- signal
- symbol
- carrier
- count
- serial
- 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.)
- Granted
Links
Abstract
The invention relates to a frequency domain linearization technique for a power amplifier in an OFDM (orthogonal frequency division multiplexing) system. The technique comprises the following steps: S1, sending a random test signal, and obtaining an initial value of a two-dimensional query table through feedback iteration; S2, sending an actual signal and obtaining an output signal. The frequency domain linearization technique has the advantages as follows: pre-distortion is performed on the power amplifier in a frequency domain of the OFDM system to realize linearization of the power amplifier, the defect that the power amplifier is required to be subjected to repeated pre-distortion including linear distortion and memory type distortion through a low pass filter in the prior art is overcome, only one-time pre-distortion is required, the structure is simple, implementation is easy, the stability of the OFDM system is improved greatly, and the application field of the OFDM system is expanded.
Description
Technical field
The present invention relates to wireless communication technology field, be specifically related to a kind of frequency-domain linear method of ofdm system intermediate power amplifier.
Background technology
OFDM (OrthogonalFrequencyDivisionMultiplexing) i.e. orthogonal frequency division multiplexi is a kind of multi-carrier modulation technology.Used in 4GLTE technology, be one of large key technology of LTE tri-, estimate at 5G still as main modulation system.OFDM general principle is N number of subcarrier by signal segmentation, then modulates N number of mutually orthogonal subcarrier respectively with N number of subsignal.Because the frequency spectrum of subcarrier is overlapped, thus higher spectrum efficiency can be obtained.Ofdm system belongs to multi-carrier modulation technology, and peak-to-average power ratio (PAPR) is higher, makes the linear requirements of transmitting terminal to power amplifier high.Therefore in an ofdm system, due to the unsteadiness of signal envelope, make this system to non-linear very sensitive, if do not improve the measure of non-linear sensitivity, the application of ofdm system will be reduced greatly.More popular for the power amplifier linearization research in ofdm system, but great majority carry out predistortion research in time domain.Need to carry out twice predistortion to the linear distortion of power amplifier and Memorability distortion, and need low pass filter.
At present, a kind of frequency-domain linear method that is simple about structure, that be easy to the ofdm system intermediate power amplifier realized yet there are no relevant report.
Summary of the invention
The object of the invention is for deficiency of the prior art, a kind of frequency-domain linear method that be applied to wireless communication field, that be easy to the ofdm system intermediate power amplifier realized is provided.
For achieving the above object, the invention discloses following technical scheme:
A frequency-domain linear method for ofdm system intermediate power amplifier, comprises the steps:
S1: send random test signal, obtained the initial value of two-dimensional polling list by feedback iteration:
S1.1 sets up two-dimentional predistortion table, and initialization:
Ha(m,n)＝1
Hp(m,n)＝0
Wherein Ha represents that signal is through the distortion of power amplifier amplitude convergent-divergent, and Hp represents signal phase rotating distortion after power amplifier; M ∈ [1, carrier_count], carrier_count=512 represent OFDM modulation sub-carriers number, correspond to different transmission signal frequency;
N ∈ [1,16], respectively 16 different constellation point in corresponding 16QAM modulation constellation;
S1.2 produces input random test Bitstream signal bit_x (n), is 1*N serial data:
N=carrier_count*symbols_per_carrier*bits_per_symbol is one group of bitstream length, wherein carrier_count=512 is sub-carrier number, symbols_per_carrier=1 is OFDM symbol number in each subcarrier, bits_per_symbol=4 is the bit number of each modulation symbol, and 16QAM modulates every symbol and contains 4 bits;
The binary signal bit_x obtained in S1.3 step S1.2 is converted into hexadecimal signal bit16_x:
1*N serial binary signal bit_x in step 2 is converted to the hexadecimal signal bit16_x of 1*carrier_count, as 16QAM modulating input signal in step 4;
The signal bit16_x that S1.4 step S1.3 obtains, carries out 16QAM modulation:
mo＝modem.qammod(16)；
symbol_x＝modulate(mo,bit16_x)
Symbol_x is the 1*carrier_count serial signal after 16QAM modulation;
According to 16 different points on planisphere, the amplitude label of the signal after modulation is saved in index (n), and index scope, from 1 to 16, is the ordinate of bivariate table Ha and Hp, i.e. the second dimension coordinate;
The signal symbol_x that S1.5 step S1.4 obtains carries out back-off:
Symbol_x2 (n)=symbol_x (n)/num is 1*carrier_count string character;
Wherein num=sqrt (10*10^ (ibo/10)), IBO=6dB, is converted into amplitude dB, and 16QAM power normalization;
The signal symbol_x2 that S1.6 step S1.5 obtains carries out serial/parallel conversion:
Symbol_x2F=reshape (symbol_x2, length (symbol_x2), 1), string character symbol_x2 (n) of 1*carrier_count is converted to parallel signal symbol_x2F (n) of carrier_count*1;
The signal symbol_x2F that S1.7 step S1.6 obtains carries out frequency domain pre-distortion:
xTa(n)＝Ha(n,index(n))*xa(n)；
xTp(n)＝xp(n)+Hp(n,index(n))；n∈[1,carrier_count],
Wherein xa (n) and xp (n) represents amplitude and the phase place of signal symbol_x2F (n) that step S1.6 obtains respectively; Amplitude and the phase place of signal symbolx_xF is obtained after xTa (n) and xTp (n) represents pre-distortion;
The signal symbolx_xF that S1.8 step S1.7 obtains carries out IFFT conversion, i.e. OFDM modulation:
symbol_xt(n)＝ifft(symbol_xF(n),IFFT_bin_length)；
N ∈ [1, carrier_count], IFFT_bin_length=carrier_count are IFFT transform length;
Frequency-region signal symbol_xF (n), transform to time domain
The parallel signal of symbol_xt (n), carrier_count*1;
The parallel/serial conversion of signal symbol_xt that S1.9 step S1.8 obtains:
symbol_xt2＝reshape(symbol_xt,1,length(symbol_xt))；
The signal symbol_xt of parallel carrier_count*1 is converted into the signal symbol_xt2 of serial 1*carrier_count;
The signal symbol_xt2 that S1.10 step S1.9 obtains inserts Cyclic Prefix:
CP=PrefixRatio*IFFT_bin_length=128; PrefixRatio is the ratio of protection interval and OFDM length;
Namely the rear cp position of signal symbol_xt2 being added to before signal, become and add Cyclic Prefix signal symbol_xt3 (n), is 1* (carrier_count+cp) serial signal;
The signal symbol_xt2 that S1.11 step S1.10 obtains by power amplifier HPA, namely signal excessively equivalence wiener model:
(1) Memorability FIR filter is crossed:
b＝[0.7692,0.1538,0.0769]；
a＝1；
p＝filter(b,a,symbol_xt3)；
A, b are the parameter of Memorability filter, and symbol_xt3 is input signal, and p is Memorability distortion output signal;
(2) nonlinear s aleh power amplifier model is crossed;
Pa=abs (p); Pa is the amplitude of signal p;
Pp=angle (p); Pp is the phase place of signal p;
ya＝(2.1587*pa)./(1.1517*pa.^2+1)；
yp＝4.0033*((pa.^2)./(9.1040*pa.^2+1))+pp；
Ya, yp are respectively signal p by the amplitude of the output signal symbol_xt4 of nonlinear s aleh power amplifier model and phase place;
Symbol_xt4=ya*exp (1i*yp) is by the output signal after power amplifier, belongs to the data of 1* (carrier_count+cp) serial;
S1.12 feedback signal Feedback=symbol_xt4:
Amplified by power amplifier and signal symbol_xt4 after distortion as feedback signal Feedback, belong to the data of 1* (carrier_count+cp) serial;
The feedback signal Feedback that S1.13 step S1.12 obtains removes Cyclic Prefix:
The feedback signal Feedback of 1* (carrier_count+cp) serial removes CP symbol above, and the feedback signal Feedback2 of prefix is removed in the serial obtaining 1*carrier_count;
The signal Feedback2 serial/parallel conversion that S1.14 step S1.13 obtains:
Feedback3＝reshape(Feedback2,length(Feedback2),1)；
The serial signal Feedback2 of 1*carrier_count is converted into the parallel signal Feedback3 of carrier_count*1;
S1.15 step S1.14 obtains signal Feedback3 and carries out FFT conversion:
Feedback_F＝fft(Feedback3,IFFT_bin_length)；
The parallel signal Feedback3 of carrier_count*1 is carried out the FFT conversion of IFFT_bin_length point; Signal is transformed to the parallel signal Feedback_F of frequency domain carrier_count*1 by time domain;
Symbol_x2F signal mean square error before the predistortion that in S1.16 calculation procedure S1.15, feedback signal Feedback_F and step S1.6 obtains
The degree that algorithm corrects power amplification distortion is observed by signal mean square error;
S1.17Rascal algorithm upgrades two-dimensional polling list:
Ha(n,index(n))＝Ha(n,index(n))-uu*(Fa(n)-xa(n))；
Hp(n,index(n))＝Hp(n,index(n))-uu*(Fp(n)-xp(n))；
Wherein, ua=0.1 is amplitude iteration step length, and up=0.1 is phase place iteration step length;
Xa and xp is respectively amplitude and the phase place of the signal symbol_x2F that step S1.6 obtains,
Fa and Fp is respectively amplitude and the phase place of the signal Feedback_F that step S1.15 obtains;
Iteration step S1.2 is to step S1.17loop_test time, and loop_test=10000, namely sends loop_test group random test signal, upgrades the initial value of the signal amplitude convergent-divergent table Ha that obtains and signal phase rotation table Hp as Ha and Hp of transmission signal;
S2 sends actual signal, is outputed signal:
The loop_test group random test signal that S2.1 sends step S1.1 to S1.17, continuous iteration upgrades Ha and Hp obtained, and is set to send the amplitude convergent-divergent table Ha of signal and the initial value of signal phase rotation table Hp:
S2.2 sends actual bit stream signal bit_x (n), is 1*N serial data;
N=carrier_count*symbols_per_carrier*bits_per_symbol is one group of bitstream length; Wherein carrier_count=512 is sub-carrier number, and symbols_per_carrier=1 is OFDM symbol number in each subcarrier, and bits_per_symbol=4 is the bit number of each modulation symbol, and 16QAM modulates every symbol and contains 4 bits;
Obtain binary signal bit_x in S2.3 step S2.2 and be transformed into hexadecimal:
1*N serial binary signal bit_x in step S2.2 is converted to the hexadecimal signal bit16_x of 1*carrier_count, as 16QAM input signal in step S1.4;
S2.4 repeats step S1.3 to S1.17loop_signal to actual signal, wherein loop_signal=100;
S2.5 crosses channel, plus noise:
After actual transmission signal iteration loop_signal time, obtain the transmission signal symblo_xt4 of 1* (carrier_count+cp) serial, cross channel, plus noise obtains Received signal strength Yreceive;
Yreceive＝awgn(symbol_xt4,snr)；
Scalar snr specifies the ratio of each sampled point signal and noise, i.e. channel SNRs, and unit is dB;
The Received signal strength Yreceive obtained in S2.6 step S2.5 removes Cyclic Prefix:
The Received signal strength Yreceive of 1* (carrier_count+cp) serial removes CP symbol above, and the Received signal strength Yreceive2 after prefix is removed in the serial obtaining 1*carrier_count;
The signal Yreceive2 serial/parallel conversion that S2.7 step S2.6 obtains:
Yreceive3＝reshape(Yreceive2,length(Yreceive2),1)；
The serial signal Yreceive2 of 1*carrier_count is converted into the parallel signal Yreceive3 of carrier_count*1;
The signal Yreceive3FFT that S2.8 step S2.7 obtains converts:
Yreceive_F＝fft(Yreceive3,IFFT_bin_length)；
The parallel signal Yreceive3 of carrier_count*1 is carried out the FFT of IFFT_bin_length point; Signal is transformed to the parallel signal Yreceive_F of frequency domain carrier_count*1 by time domain;
The parallel/serial conversion of signal Yreceive_F that S2.9 step S2.8 obtains:
Yreceive_F2＝reshape(Yreceive_F,length(Yreceive2),1)；
The parallel signal Yreceive_F of carrier_count*1 is converted into the serial signal Yreceive_F2 of 1*carrier_count;
The signal Yreceive_F2 power restore that S2.10 step S2.9 obtains:
Yreceive_F3＝Yreceive_F2*num；
The power restore sending rollback;
The signal 16QAM demodulation that S2.11 step S2.10 obtains:
de＝modem.qamdemod(16)；
out＝demodulate(de,Yreceive_F3)；
Out is the serial signal of the 1*carrier_count after 16QAM demodulation;
The hexadecimal signal out that S2.12 step S2.11 obtains is transformed into binary system:
The serial signal out of 1*carrier_count after 16QAM demodulation, be converted to the signal bit_y of binary one * (carrier_count*symbols_per_carrier);
Can obtain sending the output signal bit_y of signal bit_x after system to step S2.12.
Further, in described step S1.10, PrefixRatio value is 1/6 ~ 1/4.
Further, in described step S2.5, in order to detect the rectification of frequency-domain linear technology to power amplifier distortion, channel SNRs snr span is snr >=18dB.
Further, described channel SNRs snr value is 20dB.
The frequency-domain linear method of a kind of ofdm system intermediate power amplifier disclosed by the invention, has following beneficial effect:
The present invention carries out predistortion at the frequency domain of ofdm system to power amplifier, makes its linearisation.Avoid in prior art the shortcoming needing by low pass filter, the linear distortion of power amplifier and Memorability distortion to be carried out repeatedly to predistortion, only need a predistortion, its structure is simple, is easy to realize, greatly improve the stability of ofdm system, add the application of ofdm system.
Accompanying drawing explanation
Fig. 1 is system block diagram of the present invention;
Fig. 2 is the equivalent model wiener model schematic of power amplifier HPA;
Fig. 3 is the 16QAM planisphere only having Memorability distortion;
Fig. 4 is the 16QAM planisphere only having nonlinear distortion;
Fig. 5 is the 16QAM planisphere simultaneously having Memorability and nonlinear distortion;
Fig. 6 is the 16QAM planisphere of power amplification distortion after loop_test=10000 group random test data iteration;
Fig. 7 is one group and sends signal iteration loop_signal=50 rear 16QAM planisphere;
Fig. 8 is one group and sends signal iteration loop_signal=100 rear 16QAM planisphere;
Fig. 9 is one group and sends signal iteration loop_signal=150 rear 16QAM planisphere;
Figure 10 is the bit error rate figure of Received signal strength after loop_test=5000 group random test data iteration;
Figure 11 is the bit error rate figure of Received signal strength after loop_test=1000 group random test data iteration;
Embodiment
Below in conjunction with embodiment, also with reference to accompanying drawing, the invention will be further described.
Refer to Fig. 1, Fig. 2.A frequency-domain linear method for ofdm system intermediate power amplifier, comprises the steps:
S1: send random test signal, obtained the initial value of two-dimensional polling list by feedback iteration:
S1.1 sets up two-dimentional predistortion table, and initialization:
Ha(m,n)＝1
Hp(m,n)＝0
Wherein Ha represents that signal is through the distortion of power amplifier amplitude convergent-divergent, and Hp represents signal phase rotating distortion after power amplifier; M ∈ [1, carrier_count], carrier_count=512 represent OFDM modulation sub-carriers number, correspond to different transmission signal frequency;
N ∈ [1,16], respectively 16 different constellation point in corresponding 16QAM modulation constellation;
S1.2 produces input random test Bitstream signal bit_x (n), is 1*N serial data:
N=carrier_count*symbols_per_carrier*bits_per_symbol is one group of bitstream length, wherein carrier_count=512 is sub-carrier number, symbols_per_carrier=1 is OFDM symbol number in each subcarrier, bits_per_symbol=4 is the bit number of each modulation symbol, and 16QAM modulates every symbol and contains 4 bits;
The binary signal bit_x obtained in S1.3 step S1.2 is converted into hexadecimal signal bit16_x:
1*N serial binary signal bit_x in step 2 is converted to 1*carrier_count
Hexadecimal signal bit16_x, as 16QAM modulating input signal in step 4;
The signal bit16_x that S1.4 step S1.3 obtains, carries out 16QAM modulation:
mo＝modem.qammod(16)；
symbol_x＝modulate(mo,bit16_x)
Symbol_x is the 1*carrier_count serial signal after 16QAM modulation;
According to 16 different points on planisphere, the amplitude label of the signal after modulation is saved in index (n), and index scope, from 1 to 16, is the ordinate of bivariate table Ha and Hp, i.e. the second dimension coordinate;
The signal symbol_x that S1.5 step S1.4 obtains carries out back-off:
Symbol_x2 (n)=symbol_x (n)/num is 1*carrier_count string character;
Wherein num=sqrt (10*10^ (ibo/10)), IBO=6dB, is converted into amplitude dB, and 16QAM power normalization;
The signal symbol_x2 that S1.6 step S1.5 obtains carries out serial/parallel conversion:
Symbol_x2F=reshape (symbol_x2, length (symbol_x2), 1), string character symbol_x2 (n) of 1*carrier_count is converted to parallel signal symbol_x2F (n) of carrier_count*1;
The signal symbol_x2F that S1.7 step S1.6 obtains carries out frequency domain pre-distortion:
xTa(n)＝Ha(n,index(n))*xa(n)；
XTp (n)=xp (n)+Hp (n, index (n)); N ∈ [1, carrier_count], wherein xa (n) and xp (n) represents amplitude and the phase place of signal symbol_x2F (n) that step S1.6 obtains respectively; Amplitude and the phase place of signal symbolx_xF is obtained after xTa (n) and xTp (n) represents pre-distortion;
The signal symbolx_xF that S1.8 step S1.7 obtains carries out IFFT conversion, i.e. OFDM modulation:
symbol_xt(n)＝ifft(symbol_xF(n),IFFT_bin_length)；
N ∈ [1, carrier_count], IFFT_bin_length=carrier_count are IFFT transform length;
Frequency-region signal symbol_xF (n), transform to time domain
The parallel signal of symbol_xt (n), carrier_count*1;
The parallel/serial conversion of signal symbol_xt that S1.9 step S1.8 obtains:
symbol_xt2＝reshape(symbol_xt,1,length(symbol_xt))；
The signal symbol_xt of parallel carrier_count*1 is converted into the signal symbol_xt2 of serial 1*carrier_count;
The signal symbol_xt2 that S1.10 step S1.9 obtains inserts Cyclic Prefix:
CP=PrefixRatio*IFFT_bin_length=128; PrefixRatio is the ratio of protection interval and OFDM length;
Namely the rear cp position of signal symbol_xt2 being added to before signal, become and add Cyclic Prefix signal symbol_xt3 (n), is 1* (carrier_count+cp) serial signal;
The signal symbol_xt2 that S1.11 step S1.10 obtains by power amplifier HPA, namely signal excessively equivalence wiener model--wiener system (i.e. linear time invariant system (FIR) connect memoryless nonlinear model) can well simulate the high power amplifier model of memory:
(1) Memorability FIR filter is crossed:
b＝[0.7692,0.1538,0.0769]；
a＝1；
p＝filter(b,a,symbol_xt3)；
A, b are the parameter of Memorability filter, and symbol_xt3 is input signal, and p is Memorability distortion output signal;
(2) nonlinear s aleh power amplifier model is crossed;
Pa=abs (p); Pa is the amplitude of signal p;
Pp=angle (p); Pp is the phase place of signal p;
ya＝(2.1587*pa)./(1.1517*pa.^2+1)；
yp＝4.0033*((pa.^2)./(9.1040*pa.^2+1))+pp；
Ya, yp are respectively signal p by the amplitude of the output signal symbol_xt4 of nonlinear s aleh power amplifier model and phase place;
Symbol_xt4=ya*exp (1i*yp) is by the output signal after power amplifier, belongs to the data of 1* (carrier_count+cp) serial;
S1.12 feedback signal Feedback=symbol_xt4: amplified by power amplifier and signal symbol_xt4 after distortion as feedback signal Feedback, belong to the data of 1* (carrier_count+cp) serial;
The feedback signal Feedback that S1.13 step S1.12 obtains removes Cyclic Prefix:
The feedback signal Feedback of 1* (carrier_count+cp) serial removes CP symbol above, and the feedback signal Feedback2 of prefix is removed in the serial obtaining 1*carrier_count;
The signal Feedback2 serial/parallel conversion that S1.14 step S1.13 obtains:
Feedback3＝reshape(Feedback2,length(Feedback2),1)；
The serial signal Feedback2 of 1*carrier_count is converted into the parallel signal Feedback3 of carrier_count*1;
S1.15 step S1.14 obtains signal Feedback3 and carries out FFT conversion:
Feedback_F＝fft(Feedback3,IFFT_bin_length)；
The parallel signal Feedback3 of carrier_count*1 is carried out the FFT conversion of IFFT_bin_length point; Signal is transformed to the parallel signal Feedback_F of frequency domain carrier_count*1 by time domain;
Symbol_x2F signal mean square error before the predistortion that in S1.16 calculation procedure S1.15, feedback signal Feedback_F and step S1.6 obtains
The degree that algorithm corrects power amplification distortion is observed by signal mean square error;
S1.17Rascal algorithm upgrades two-dimensional polling list:
Ha(n,index(n))＝Ha(n,index(n))-uu*(Fa(n)-xa(n))；
Hp(n,index(n))＝Hp(n,index(n))-uu*(Fp(n)-xp(n))；
Wherein, ua=0.1 is amplitude iteration step length, and up=0.1 is phase place iteration step length;
Xa and xp is respectively amplitude and the phase place of the signal symbol_x2F that step S1.6 obtains,
Fa and Fp is respectively amplitude and the phase place of the signal Feedback_F that step S1.15 obtains;
Iteration step S1.2 is to step S1.17loop_test time, and loop_test=10000, namely sends loop_test group random test signal, upgrades the initial value of the signal amplitude convergent-divergent table Ha that obtains and signal phase rotation table Hp as Ha and Hp of transmission signal;
S2 sends actual signal, is outputed signal:
The loop_test group random test signal that S2.1 sends step S1.1 to S1.17, continuous iteration upgrades Ha and Hp obtained, and is set to send the amplitude convergent-divergent table Ha of signal and the initial value of signal phase rotation table Hp:
S2.2 sends actual bit stream signal bit_x (n), is 1*N serial data;
N=carrier_count*symbols_per_carrier*bits_per_symbol is one group of bitstream length; Wherein carrier_count=512 is sub-carrier number, and symbols_per_carrier=1 is OFDM symbol number in each subcarrier, and bits_per_symbol=4 is the bit number of each modulation symbol, and 16QAM modulates every symbol and contains 4 bits;
Obtain binary signal bit_x in S2.3 step S2.2 and be transformed into hexadecimal:
1*N serial binary signal bit_x in step S2.2 is converted to the hexadecimal signal bit16_x of 1*carrier_count, as 16QAM input signal in step S1.4;
S2.4 repeats step S1.3 to S1.17loop_signal to actual signal, wherein loop_signal=100;
S2.5 crosses channel, plus noise:
After actual transmission signal iteration loop_signal time, obtain the transmission signal symblo_xt4 of 1* (carrier_count+cp) serial, cross channel, plus noise obtains Received signal strength Yreceive;
Yreceive＝awgn(symbol_xt4,snr)；
Scalar snr specifies the ratio of each sampled point signal and noise, i.e. channel SNRs, and unit is dB;
The Received signal strength Yreceive obtained in S2.6 step S2.5 removes Cyclic Prefix:
The Received signal strength Yreceive of 1* (carrier_count+cp) serial removes CP symbol above, and the Received signal strength Yreceive2 after prefix is removed in the serial obtaining 1*carrier_count;
The signal Yreceive2 serial/parallel conversion that S2.7 step S2.6 obtains:
Yreceive3＝reshape(Yreceive2,length(Yreceive2),1)；
The serial signal Yreceive2 of 1*carrier_count is converted into the parallel signal Yreceive3 of carrier_count*1;
The signal Yreceive3FFT that S2.8 step S2.7 obtains converts:
Yreceive_F＝fft(Yreceive3,IFFT_bin_length)；
The parallel signal Yreceive3 of carrier_count*1 is carried out the FFT of IFFT_bin_length point; Signal is transformed to the parallel signal Yreceive_F of frequency domain carrier_count*1 by time domain;
The parallel/serial conversion of signal Yreceive_F that S2.9 step S2.8 obtains:
Yreceive_F2＝reshape(Yreceive_F,length(Yreceive2),1)；
The parallel signal Yreceive_F of carrier_count*1 is converted into the serial signal Yreceive_F2 of 1*carrier_count;
The signal Yreceive_F2 power restore that S2.10 step S2.9 obtains:
Yreceive_F3＝Yreceive_F2*num；
The power restore sending rollback;
The signal 16QAM demodulation that S2.11 step S2.10 obtains:
de＝modem.qamdemod(16)；
out＝demodulate(de,Yreceive_F3)；
Out is the serial signal of the 1*carrier_count after 16QAM demodulation;
The hexadecimal signal out that S2.12 step S2.11 obtains is transformed into binary system:
The serial signal out of 1*carrier_count after 16QAM demodulation, be converted to the signal bit_y of binary one * (carrier_count*symbols_per_carrier);
Can obtain sending the output signal bit_y of signal bit_x after system to step S2.12.
In the present invention, in step S1.10, PrefixRatio value is 1/6 ~ 1/4.
In step S2.5, in order to detect the rectification of frequency-domain linear technology to power amplifier distortion, channel SNRs snr span is snr >=18dB, preferred 20dB.
It should be noted that, in step S2.5, in order to detect the rectification of frequency-domain linear technology to power amplifier distortion, channel SNRs snr should arrange larger value, or does not add noise.
Planisphere
See Fig. 3, Fig. 3 does not add the 16QAM planisphere that predistortion only has Memorability distortion.The distortion of power amplifier Memorability is different to the signal amplification factor of different frequency, and OFDM modulation has the subcarrier of different frequency, so there is Memorability distortion by power amplifier.
See Fig. 4, Fig. 4 does not add the 16QAM planisphere that predistortion only has nonlinear distortion.Nonlinear distortion is that power amplifier is different to the signal amplification factor of different amplitude, and adopt 16QAM to modulate in OFDM modulation, signal amplitude is different, so there is nonlinear distortion by power amplifier.
See Fig. 5, Fig. 5 does not add the 16QAM planisphere that predistortion exists Memorability and nonlinear distortion simultaneously.There is Memorability distortion and nonlinear distortion by power amplifier in ofdm signal, by frequency-domain linear technology, corrects this two kinds of distortions.
See Fig. 6, Ha and Hp produced by loop_test=10000 group random test data iteration, as the initial value of Ha and Hp of transmission signal.
See Fig. 7,8,9.From Fig. 7,8,9, when test data iterations loop_signal=150 time, the non-linear and Memorability distortion of power amplifier can be offset completely, can linearisation amplify.
The bit error rate that the power amplification distortion obtained sending loop_test group different random test data brings.
See Figure 10, when channel does not add noise, the bit error rate of Received signal strength after loop_test=5000 group random test data iteration;
See Figure 11, when channel does not add noise, the bit error rate of Received signal strength after loop_test=1000 group random test data iteration.
Known by error rate Figure 10 and Figure 11: when not adding noise, the linear and Memorability distortion of a power amplifier, when sending 1000 groups of test datas, it is 0 that the error rate can reach, and well can correct the distortion that power amplifier brings.So when sending signal, predistortion iteration calibration number of times loop_signal=1, the error code that power amplifier brings can be corrected.But planisphere is more discrete, in order to make planisphere better, the iterations loop_signal sending signal can be increased.When about loop_signal=100 time, planisphere is more satisfactory.But increase iterations, can bring delay.So according to system to the susceptibility postponed, the predistortion iteration often organizing actual transmission signal can be arranged and corrects number of times.
The above is only the preferred embodiment of the present invention, it should be pointed out that for those of ordinary skill in the art; under the premise of not departing from the present invention; the some improvement can also made the present invention and supplement, these improve and supplement, also should be considered as protection scope of the present invention.
Claims (4)
1. a frequency-domain linear method for ofdm system intermediate power amplifier, is characterized in that, comprise the steps:
S1: send random test signal, obtained the initial value of two-dimensional polling list by feedback iteration:
S1.1 sets up two-dimentional predistortion table, and initialization:
Ha(m,n)＝1
Hp(m,n)＝0
Wherein Ha represents that signal is through the distortion of power amplifier amplitude convergent-divergent, and Hp represents signal phase rotating distortion after power amplifier; M ∈ [1, carrier_count], carrier_count=512 represent OFDM modulation sub-carriers number, correspond to different transmission signal frequency; N ∈ [1,16], respectively 16 different constellation point in corresponding 16QAM modulation constellation;
S1.2 produces input random test Bitstream signal bit_x (n), is 1*N serial data:
N=carrier_count*symbols_per_carrier*bits_per_symbol is one group of bitstream length, wherein carrier_count=512 is sub-carrier number, symbols_per_carrier=1 is OFDM symbol number in each subcarrier, bits_per_symbol=4 is the bit number of each modulation symbol, and 16QAM modulates every symbol and contains 4 bits;
The binary signal bit_x obtained in S1.3 step S1.2 is converted into hexadecimal signal bit16_x:
1*N serial binary signal bit_x in step 2 is converted to the hexadecimal signal bit16_x of 1*carrier_count, as 16QAM modulating input signal in step 4;
The signal bit16_x that S1.4 step S1.3 obtains, carries out 16QAM modulation:
mo＝modem.qammod(16)；
symbol_x＝modulate(mo,bit16_x)
Symbol_x is the 1*carrier_count serial signal after 16QAM modulation;
According to 16 different points on planisphere, the amplitude label of the signal after modulation is saved in index (n), and index scope, from 1 to 16, is the ordinate of bivariate table Ha and Hp, i.e. the second dimension coordinate;
The signal symbol_x that S1.5 step S1.4 obtains carries out back-off:
Symbol_x2 (n)=symbol_x (n)/num is 1*carrier_count string character;
Wherein num=sqrt (10*10^ (ibo/10)), IBO=6dB, is converted into amplitude dB, and 16QAM power normalization;
The signal symbol_x2 that S1.6 step S1.5 obtains carries out serial/parallel conversion:
symbol_x2F＝reshape(symbol_x2,length(symbol_x2),1)，
String character symbol_x2 (n) of 1*carrier_count is converted to parallel signal symbol_x2F (n) of carrier_count*1;
The signal symbol_x2F that S1.7 step S1.6 obtains carries out frequency domain pre-distortion:
xTa(n)＝Ha(n,index(n))*xa(n)；
xTp(n)＝xp(n)+Hp(n,index(n))；n∈[1,carrier_count],
Wherein xa (n) and xp (n) represents amplitude and the phase place of signal symbol_x2F (n) that step S1.6 obtains respectively; Amplitude and the phase place of signal symbolx_xF is obtained after xTa (n) and xTp (n) represents pre-distortion;
The signal symbolx_xF that S1.8 step S1.7 obtains carries out IFFT conversion, i.e. OFDM modulation:
symbol_xt(n)＝ifft(symbol_xF(n),IFFT_bin_length)；
N ∈ [1, carrier_count], IFFT_bin_length=carrier_count are IFFT transform length;
Frequency-region signal symbol_xF (n), transform to time domain symbol_xt (n), the parallel signal of carrier_count*1;
The parallel/serial conversion of signal symbol_xt that S1.9 step S1.8 obtains:
symbol_xt2＝reshape(symbol_xt,1,length(symbol_xt))；
The signal symbol_xt of parallel carrier_count*1 is converted into the signal symbol_xt2 of serial 1*carrier_count;
The signal symbol_xt2 that S1.10 step S1.9 obtains inserts Cyclic Prefix:
CP=PrefixRatio*IFFT_bin_length=128; PrefixRatio is the ratio of protection interval and OFDM length;
Namely the rear cp position of signal symbol_xt2 being added to before signal, become and add Cyclic Prefix signal symbol_xt3 (n), is 1* (carrier_count+cp) serial signal;
The signal symbol_xt2 that S1.11 step S1.10 obtains by power amplifier HPA, namely signal excessively equivalence wiener model:
(1) Memorability FIR filter is crossed:
b＝[0.7692,0.1538,0.0769]；
a＝1；
p＝filter(b,a,symbol_xt3)；
A, b are the parameter of Memorability filter, and symbol_xt3 is input signal, and p is Memorability distortion output signal;
(2) nonlinear s aleh power amplifier model is crossed;
Pa=abs (p); Pa is the amplitude of signal p;
Pp=angle (p); Pp is the phase place of signal p;
ya＝(2.1587*pa)./(1.1517*pa.^2+1)；
yp＝4.0033*((pa.^2)./(9.1040*pa.^2+1))+pp；
Ya, yp are respectively signal p by the amplitude of the output signal symbol_xt4 of nonlinear s aleh power amplifier model and phase place;
Symbol_xt4=ya*exp (1i*yp) is by the output signal after power amplifier, belongs to the data of 1* (carrier_count+cp) serial;
S1.12 feedback signal Feedback=symbol_xt4:
Amplified by power amplifier and signal symbol_xt4 after distortion as feedback signal Feedback, belong to the data of 1* (carrier_count+cp) serial;
The feedback signal Feedback that S1.13 step S1.12 obtains removes Cyclic Prefix:
The feedback signal Feedback of 1* (carrier_count+cp) serial removes CP symbol above, and the feedback signal Feedback2 of prefix is removed in the serial obtaining 1*carrier_count; The signal Feedback2 serial/parallel conversion that S1.14 step S1.13 obtains:
Feedback3＝reshape(Feedback2,length(Feedback2),1)；
The serial signal Feedback2 of 1*carrier_count is converted into the parallel signal Feedback3 of carrier_count*1;
S1.15 step S1.14 obtains signal Feedback3 and carries out FFT conversion:
Feedback_F＝fft(Feedback3,IFFT_bin_length)；
The parallel signal Feedback3 of carrier_count*1 is carried out the FFT conversion of IFFT_bin_length point; Signal is transformed to the parallel signal Feedback_F of frequency domain carrier_count*1 by time domain;
Symbol_x2F signal mean square error before the predistortion that in S1.16 calculation procedure S1.15, feedback signal Feedback_F and step S1.6 obtains
The degree that algorithm corrects power amplification distortion is observed by signal mean square error;
S1.17Rascal algorithm upgrades two-dimensional polling list:
Ha(n,index(n))＝Ha(n,index(n))-uu*(Fa(n)-xa(n))；
Hp(n,index(n))＝Hp(n,index(n))-uu*(Fp(n)-xp(n))；
Wherein, ua=0.1 is amplitude iteration step length, and up=0.1 is phase place iteration step length;
Xa and xp is respectively amplitude and the phase place of the signal symbol_x2F that step S1.6 obtains,
Fa and Fp is respectively amplitude and the phase place of the signal Feedback_F that step S1.15 obtains;
Iteration step S1.2 is to step S1.17loop_test time, and loop_test=10000, namely sends loop_test group random test signal, upgrades the initial value of the signal amplitude convergent-divergent table Ha that obtains and signal phase rotation table Hp as Ha and Hp of transmission signal;
S2 sends actual signal, is outputed signal:
The loop_test group random test signal that S2.1 sends step S1.1 to S1.17, continuous iteration upgrades Ha and Hp obtained, and is set to send the amplitude convergent-divergent table Ha of signal and the initial value of signal phase rotation table Hp:
S2.2 sends actual bit stream signal bit_x (n), is 1*N serial data;
N=carrier_count*symbols_per_carrier*bits_per_symbol is one group of bitstream length; Wherein carrier_count=512 is sub-carrier number,
Symbols_per_carrier=1 is OFDM symbol number in each subcarrier, and bits_per_symbol=4 is the bit number of each modulation symbol, and 16QAM modulates every symbol and contains 4 bits;
Obtain binary signal bit_x in S2.3 step S2.2 and be transformed into hexadecimal:
1*N serial binary signal bit_x in step S2.2 is converted to the hexadecimal signal bit16_x of 1*carrier_count, as 16QAM input signal in step S1.4;
S2.4 repeats step S1.3 to S1.17loop_signal to actual signal, wherein loop_signal=100;
S2.5 crosses channel, plus noise:
After actual transmission signal iteration loop_signal time, obtain the transmission signal symblo_xt4 of 1* (carrier_count+cp) serial, cross channel, plus noise obtains Received signal strength Yreceive;
Yreceive＝awgn(symbol_xt4,snr)；
Scalar snr specifies the ratio of each sampled point signal and noise, i.e. channel SNRs, and unit is dB;
The Received signal strength Yreceive obtained in S2.6 step S2.5 removes Cyclic Prefix:
The Received signal strength Yreceive of 1* (carrier_count+cp) serial removes CP symbol above, and the Received signal strength Yreceive2 after prefix is removed in the serial obtaining 1*carrier_count;
The signal Yreceive2 serial/parallel conversion that S2.7 step S2.6 obtains:
Yreceive3＝reshape(Yreceive2,length(Yreceive2),1)；
The serial signal Yreceive2 of 1*carrier_count is converted into the parallel signal Yreceive3 of carrier_count*1;
The signal Yreceive3FFT that S2.8 step S2.7 obtains converts:
Yreceive_F＝fft(Yreceive3,IFFT_bin_length)；
The parallel signal Yreceive3 of carrier_count*1 is carried out the FFT of IFFT_bin_length point; Signal is transformed to the parallel signal Yreceive_F of frequency domain carrier_count*1 by time domain;
The parallel/serial conversion of signal Yreceive_F that S2.9 step S2.8 obtains:
Yreceive_F2＝reshape(Yreceive_F,length(Yreceive2),1)；
The parallel signal Yreceive_F of carrier_count*1 is converted into the serial signal Yreceive_F2 of 1*carrier_count;
The signal Yreceive_F2 power restore that S2.10 step S2.9 obtains:
Yreceive_F3＝Yreceive_F2*num；
The power restore sending rollback;
The signal 16QAM demodulation that S2.11 step S2.10 obtains:
de＝modem.qamdemod(16)；
out＝demodulate(de,Yreceive_F3)；
Out is the serial signal of the 1*carrier_count after 16QAM demodulation;
The hexadecimal signal out that S2.12 step S2.11 obtains is transformed into binary system:
The serial signal out of 1*carrier_count after 16QAM demodulation, be converted to the signal bit_y of binary one * (carrier_count*symbols_per_carrier);
Can obtain sending the output signal bit_y of signal bit_x after system to step S2.12.
2. the frequency-domain linear method of a kind of ofdm system intermediate power amplifier according to claim 1, is characterized in that, in described step S1.10, PrefixRatio value is 1/6 ~ 1/4.
3. the frequency-domain linear method of a kind of ofdm system intermediate power amplifier according to claim 1, it is characterized in that, in described step S2.5, in order to detect the rectification of frequency-domain linear technology to power amplifier distortion, channel SNRs snr span is snr >=18dB.
4. the frequency-domain linear method of a kind of ofdm system intermediate power amplifier according to claim 3, is characterized in that, described channel SNRs snr value is 20dB.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510566527.5A CN105207628B (en) | 2015-09-08 | 2015-09-08 | A kind of frequency-domain linear method of ofdm system intermediate power amplifier |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510566527.5A CN105207628B (en) | 2015-09-08 | 2015-09-08 | A kind of frequency-domain linear method of ofdm system intermediate power amplifier |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105207628A true CN105207628A (en) | 2015-12-30 |
| CN105207628B CN105207628B (en) | 2018-06-29 |
Family
ID=54955111
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510566527.5A Active CN105207628B (en) | 2015-09-08 | 2015-09-08 | A kind of frequency-domain linear method of ofdm system intermediate power amplifier |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105207628B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108964681A (en) * | 2018-09-15 | 2018-12-07 | 陕西师范大学 | The method and apparatus that non-centrosymmetry is modulated in communication system |
| JP2019083510A (en) * | 2017-10-31 | 2019-05-30 | 富士通株式会社 | Distortion compensating device and distortion compensation method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103297378A (en) * | 2013-06-24 | 2013-09-11 | 苏州工业园区昶辰通信科技有限公司 | Lookup table pre-distortion method based on orthogonal frequency division multiplexing signal amplitude characteristics |
-
2015
- 2015-09-08 CN CN201510566527.5A patent/CN105207628B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103297378A (en) * | 2013-06-24 | 2013-09-11 | 苏州工业园区昶辰通信科技有限公司 | Lookup table pre-distortion method based on orthogonal frequency division multiplexing signal amplitude characteristics |
Non-Patent Citations (2)
| Title |
|---|
| 任智源 等: "OFDM系统中记忆非线性放大器的高效预失真技术", 《西安电子科技大学学报（自然科学版）》 * |
| 任智源: "高峰均比系统中放大器的高效高线性化技术", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2019083510A (en) * | 2017-10-31 | 2019-05-30 | 富士通株式会社 | Distortion compensating device and distortion compensation method |
| JP7206719B2 (en) | 2017-10-31 | 2023-01-18 | 富士通株式会社 | Distortion compensation device and distortion compensation method |
| CN108964681A (en) * | 2018-09-15 | 2018-12-07 | 陕西师范大学 | The method and apparatus that non-centrosymmetry is modulated in communication system |
| CN108964681B (en) * | 2018-09-15 | 2020-09-11 | 陕西师范大学 | Method and apparatus for non-centrosymmetric modulation in a communication system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105207628B (en) | 2018-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101753500A (en) | Method and system for efficiently restraining large peak-to-average power ratio (PAPR) of OFDM system based on companding | |
| CN102497350B (en) | OFDM (Orthogonal Frequency Division Multiplexing) peak-to-average power ratio lowering method based on constellation linear expansion | |
| Shin et al. | A weighted OFDM signal scheme for peak-to-average power ratio reduction of OFDM signals | |
| CN100499626C (en) | Block clipping method for resolving nonlinear distortion question in OFDM system | |
| CN102404273B (en) | Method for transmitting OFDM signals based on new companding transform | |
| CN105207628A (en) | Frequency domain linearization technique for power amplifier in OFDM (orthogonal frequency division multiplexing) system | |
| Zhu et al. | Reducing the peak-to-average power ratio using unitary matrix transformation | |
| CN103825855A (en) | Method reducing peak average power ratio in OFDM system based on amplitude limiting technology | |
| CN106161320B (en) | Method for reducing PAPR of OFDM system by self-adaptive companding technology | |
| CN104994051B (en) | Based on instantaneously the signal transmitting with the linear-scale conversion of the raising mixed carrier system PA efficiency of average power ratio and method of reseptance | |
| CN102404274B (en) | Method for hyperbolic tangent companding transform capable of decreasing peak-to-average power ratio (PAPR) of OFDM (orthogonal frequency division multiplexing) signal | |
| CN103259758A (en) | OFDM communication system and method based on companding | |
| CN105721056B (en) | One kind reduces visible light communication system light source led non-linear method and system | |
| CN105634651A (en) | Method for reducing PAPR of OFDM system based on pre-coding and exponential companding technologies | |
| Hu et al. | PAPR Reduction scheme in SFBC MIMO-OFDM systems without side information | |
| Guerreiro et al. | Optimum performance and spectral characterization of CE-OFDM signals | |
| Jayati et al. | The Analysis of the High Power Amplifier Distortion on the MIMO-GFDM System | |
| Deepa et al. | A joint clipping and logarithmic based companding for the reduction of peak-to-average power ratio in OFDM system | |
| Kim et al. | An SLM-based real-time PAPR reduction method using dummy sequence insertion in the OFDM communication | |
| CN102523190B (en) | Compression and expansion transform method based on threshold and hyperbolic tangent function | |
| Bansode et al. | Hardware Implementation of OFDM system to reduce PAPR using Selective Level Mapping on FPGA | |
| CN108632195B (en) | Segmented companding algorithm for reducing peak-to-average power ratio of OFDM (orthogonal frequency division multiplexing) signal | |
| Guerreiro et al. | Optimum Asymptotic Performance for Nonlinearly Amplified OFDM Signals | |
| Upadhyay et al. | A Novel Technique of Peak-To-Average Power Ratio (PAPR) Reduction in OFDM Signal | |
| Salini et al. | A Hybrid SLM Technique for PAPR Reduction in OFDM |
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 |