CN104065610A - Method for reducing PAPR of OFDM system through improved SLM algorithm - Google Patents

Abstract

The invention discloses a method for reducing the PAPR of an OFDM system through an improved SLM algorithm and relates to an OFDM technology. The method for reducing the PAPR of the OFDM system through the improved SLM algorithm aims to solve the problem that the system error rate of an existing non-sideband transmission SLM algorithm is high, and includes the first step of determining a frequency domain data vector of an OFDM signal, the second step of processing the input vector X through the improved SLM algorithm, the third step of carrying out IFFT on obtained D paths of independent output vectors S(d) to obtain D paths of time-domain signals, the fourth step of selecting and sending the path of signal with the minimum PAPR value, the fifth step of carrying out FFT on the received D paths of signals by a receiving end, the sixth step of carrying out module picking on the obtained D paths of signals after FFT, the seventh step of estimating out which path of signal is emitted by an emitting end and the eighth step of recovering original signal information. The method is applied to the field of wireless communications.

Description

A kind of modified model SLM algorithm reduces the method for ofdm system PAPR

Technical field

The present invention relates to OFDM (OFDM) technology, be specifically related to adopt balanceable Gold sequence and select mapping algorithm (SLM), combine and reduce ofdm system peak-to-average power ratio (PAPR), and then OFDM technology is applied in the radio communication of High Speed Modulation and transmission.

Background technology

OFDM (OFDM) is a kind of wireless multi-carrier transmission technology, relies on its higher availability of frequency spectrum and effectively to advantages such as anti-multipath fadings, becomes one of key technology of radio communication.Yet the in-phase signal stack of a plurality of subchannel outputs of ofdm system can produce very large instantaneous peak value, and the acute variation characteristic of sort signal envelope can be weighed with peak power ratio (PAPR).

In practice, the modulation of the parallel subcarrier in ofdm system Zhong Ge road can be completed by IDFT, and demodulation can be completed by DFT, and from the knowledge of Digital Signal Processing, IDFT and DFT can adopt efficient FFT fast algorithm to realize.

For discrete signal, the equivalent IFFT expression of OFDM base band is:

$s\left(n\right)=\underset{k=0}{\overset{N-1}{\mathrm{\Σ}}}{C}_k{e}^{j\frac{2\mathrm{\π}}{N}\mathrm{nk}},n=\mathrm{0,1,2},...,N-1$

Wherein s (n) represents n sample value in this OFDM symbol, and N represents ofdm system subcarrier number, and peak-to-average power ratio PAPR can be defined as:

$\mathrm{PAPR}=10\lg \left\{\frac{\max {\left|s\left(n\right)\right|}^2}{E\left[{\left|s\left(n\right)\right|}^2\right]}\right\}\left(\mathrm{dB}\right),(0\≤n\≤N-1)$

E[|s (n) wherein | ²] represent the mathematic expectaion of signal power.

Conventionally, can with CCDF (CCDF), carry out the distribution situation of gauge signal PAPR, cumulative distribution function F (z) can be expressed as:

$F\left(z\right)=\underset{0}{\overset{z}{\∫}}\exp (-z)\mathrm{dz}=1-e^{-z}$

During without sampling, suppose between each sampling point it is incoherent mutually, the PAPR of ofdm signal is greater than a certain threshold value PAPR so ₀probability, CCDF (CCDF) can be expressed as:

$\begin{array}{c}P\left\{{\mathrm{PAPR}}_0\right\}=P_r\{\mathrm{PAPR}>{\mathrm{PAPR}}_0\}\\ =1-F{\left({\mathrm{PAPR}}_0\right)}^N=1-{(1-e^{-{\mathrm{PAPR}}_0})}^N\end{array}$

The ofdm signal of high peak-to-average force ratio is easy to produce nonlinear distortion, produces between larger subcarrier and disturbs (ICI), causes systematic function severe exacerbation.The several method that reduces at present PAPR mainly contains amplitude limit class, coding class and probability class methods.Wherein application is selectivity mapping (SLM) method in probability class methods more widely.

The basic thought of SLM method is to utilize D to add up the identical information of independently vector representation, selects its time-domain symbol to have minimum PAPR Zhi mono-tunnel for transmission.The separate vector in D road of transmission is by D the random phase shift vector A diverse, length is N ^(d)(1≤d≤D) produces.The general A that sets ^(d)∈ ± 1, and ± j}, the execution of this factor does not need to carry out multiplying, can greatly reduce operand, reduces the hardware complexity of real system simultaneously.After initial data sends over, its corresponding time-domain signal s of all D circuit-switched data parallel computation ^(d), and select to there is minimum PAPR Zhi mono-tunnel and transmit.

For SLM method, at receiving terminal, must carry out the operation contrary with transmitting terminal, to recover the raw information of transmission.What therefore, receiving terminal must be known transmitting terminal selection is which signal transmits.In reality, receiving terminal all D the random phase shift sequences that prestore, which signal what therefore only need to know transmission is.Common solution is as side information, to send together the branch road sequence number d of selection to receiving terminal, for D road SLM transmitter, need to transmit log ₂(D-1) side information of bit, when D is larger, the transmission of extra side information will reduce the band efficiency of system.

For SLM method, need to transmit the shortcoming of side information, existing without the proposition of sideband transmission SLM method at present.Similar with SLM method principle without sideband transmission SLM method basic principle, just without the random phase shift vector of sideband transmission SLM method, take the mode of amplitude mark, thereby omitted the step of side information transmission.Random phase shift vector is all that mould value is 1 sequence in SLM method, and be greater than 1 and be less than 1 constant without the mould value of ad-hoc location in every group of random phase shift sequence being made as respectively in the SLM method of sideband transmission, with this, carry out mark, once the variation of ad-hoc location mould value be detected at receiving terminal, which sequence what just can judge transmission is.Emulation experiment shows, although do not need to transmit side information without sideband transmission SLM method, can make the error rate of system raise, and it exchanges band efficiency for the sacrificial system error rate as seen.

Gold sequence is by two preferably right m sequences mould two and obtaining by turn, has good correlation properties.Because its auto-correlation function secondary lobe is lower, and the auto-correlation function of time domain and the power spectrum of frequency domain are a pair of Fourier transform pairss, so its envelope can be comparatively smooth.Therefore, can consider, by multiplying each other with the data flow of input after Gold sequence deformation process, to produce comparatively smooth frequency spectrum, thereby reach the object that reduces signal amplitude peak value.

Summary of the invention

The present invention will solve existingly without the high problem of the sideband transmission SLM algorithmic system error rate, and provides a kind of modified model SLM algorithm to reduce the method for ofdm system PAPR.

A kind of method that modified model SLM algorithm reduces ofdm system PAPR realizes according to the following steps:

Step 1, transmitting terminal, according to ofdm system subcarrier number, are determined the frequency domain data vector of ofdm signal in modified model SLM algorithm, are expressed as X=[X ₁, X ₂..., X _n] ^t, wherein N represents sub-carrier number, the transposition of T representing matrix;

Step 2, employing modified model SLM algorithm are processed the frequency domain data vector X of the ofdm signal of input:

Described modified model SLM algorithm is the frequency domain data vector X=[X of the ofdm signal of transmitting terminal input ₁, X ₂..., X _n] ^trespectively with the D road length random phase shift vector A that is N ^(d)balanceable Gold sequence phase weighting with D road after amplitude set obtains the separate output vector S in D road ^(d); S wherein ^(d)for frequency domain data vector, 1≤d≤D;

Random phase shift vector A ^(d)choosing method be:

i=1,2 ..., N; D=1,2 ..., D, obey being uniformly distributed between [0,2 π], j represents imaginary unit;

Step 3, the output vector S separate to resulting D road ^(d)carry out respectively IFFT computing, obtain D road time-domain signal s ^(d)=IFFT (S ^(d));

Step 4, to the D road time-domain signal s obtaining in step 3 ^(d)=IFFT (S ^(d)) calculate respectively its PAPR value, and select and there is minimum PAPR Zhi mono-road signal and send;

Step 5, receiving terminal carry out FFT computing to the D road signal receiving;

Step 6, after FFT computing to the D road signal delivery obtaining, and mould value be wherein greater than to 1 part be set to P, mould value is less than 1 part and is set to Q;

Step 7, ask in step 6 the sequence after set and pre-stored in receiving terminal the Hamming distance between improved local balanceable Gold sequence, obtain with set after sequence there is one group of smallest hamming distance local balanceable Gold sequence, which signal what estimate transmitting terminal transmitting is;

Step 8, according to the sequence number of the signal transmitted of estimating in step 7 to obtain, carry out to received signal the operation contrary with transmitting terminal, recover original signal information.

Invention effect:

Modified model SLM algorithm can not only reduce the PAPR value of system, and than reduced the error rate of system without sideband transmission SLM method.

The present invention is based on balanceable Gold sequence and SLM method and combine reduction ofdm system PAPR value, can when effectively reducing ofdm system PAPR value, reduce the error rate of system, realize reliable communication.

From Fig. 3 and Fig. 4, all can obviously find out, modified model SLM scheme has the effect of obvious inhibition peak-to-average force ratio, although further do not reduce peak-to-average force ratio on the SLM scheme basis without sideband transmission, the curve of the two almost overlaps, and illustrates that they are similar to the inhibition degree of PAPR.But in modified model scheme, owing to having added balanceable Gold sequence, randomness strengthens, and the Euclidean distance between phase sequence increases, and this will make new departure improve aspect the error rate.

In the situation that not increasing system complexity, the use of new method makes the error rate obtain obvious improvement.

In modified model SLM algorithm, random phase shift vector has also adopted the mode of amplitude mark, owing to having added balanceable Gold sequence, than common amplitude mark mode, the randomness of the random phase shift sequence of new method strengthens, make the interference reduction between different subchannels, can effectively reduce the error rate of system, improve error performance.

Accompanying drawing explanation

Fig. 1 is modified model SLM Method And Principle figure;

Fig. 2 is that modified model SLM method reduces PAPR flow chart;

Fig. 3 is emulation experiment sub-carriers while counting Num=32, the CCDF comparison diagram of three kinds of schemes;

Fig. 4 is emulation experiment sub-carriers while counting Num=128, the CCDF comparison diagram of three kinds of schemes;

Fig. 5 represents in emulation experiment that the set of modified model SLM method amplitude gets error ratio characteristic figure corresponding to different P values;

Fig. 6 is three kinds of scheme error ratio characteristic curve comparison figure while representing sub-carrier number Num=32 in emulation experiment;

Fig. 7 is three kinds of scheme error ratio characteristic curve comparison figure while representing sub-carrier number Num=64 in emulation experiment.

Embodiment

Embodiment one: the method that a kind of modified model SLM algorithm of present embodiment reduces ofdm system PAPR realizes according to the following steps:

Step 1, transmitting terminal, according to ofdm system subcarrier number, are determined the frequency domain data vector of ofdm signal in modified model SLM algorithm, are expressed as X=[X ₁, X ₂..., X _n] ^t, wherein N represents sub-carrier number, the transposition of T representing matrix;

Step 2, employing modified model SLM algorithm are processed the frequency domain data vector X of the ofdm signal of input:

Described modified model SLM algorithm is the frequency domain data vector X=[X of the ofdm signal of transmitting terminal input ₁, X ₂..., X _n] ^trespectively with the D road length random phase shift vector A that is N ^(d)balanceable Gold sequence phase weighting with D road after amplitude set obtains the separate output vector S in D road ^(d); S wherein ^(d)for frequency domain data vector, 1≤d≤D;

Random phase shift vector A ^(d)choosing method be:

i=1,2 ..., N; D=1,2 ..., D, obey being uniformly distributed between [0,2 π], j represents imaginary unit;

Step 3, the output vector S separate to resulting D road ^(d)carry out respectively IFFT computing, obtain D road time-domain signal s ^(d)=IFFT (S ^(d));

Step 4, to the D road time-domain signal s obtaining in step 3 ^(d)=IFFT (S ^(d)) calculate respectively its PAPR value, and select and there is minimum PAPR Zhi mono-road signal and send;

Step 5, receiving terminal carry out FFT computing to the D road signal receiving;

Step 6, after FFT computing to the D road signal delivery obtaining, and mould value be wherein greater than to 1 part be set to P, mould value is less than 1 part and is set to Q;

Step 7, ask in step 6 the sequence after set and pre-stored in receiving terminal the Hamming distance between improved local balanceable Gold sequence, obtain with set after sequence there is one group of smallest hamming distance local balanceable Gold sequence, which signal what estimate transmitting terminal transmitting is;

Step 8, according to the sequence number of the signal transmitted of estimating in step 7 to obtain, carry out to received signal the operation contrary with transmitting terminal, recover original signal information.

In this described mode, modified model SLM algorithm is mainly to improve the producing method of its random phase shift vector.

Embodiment two: present embodiment is different from embodiment one: the r level shift register generation cycle is N=2 ^r-1 balanceable Gold sequence, due to N=2 ^r-1 is odd number, and ofdm system subcarrier number is generally even number, in order to guarantee that cycle and the OFDM subcarrier number of balanceable Gold sequence equates, one zero of the last benefit of balanceable Gold sequence, then " 1 " locational element in sequence is set to the constant P that is greater than 1, " 0 " locational element is set to the constant Q that is less than 1, by the balanceable Gold sequence after set more respectively with the random phase shift sequence phase dot product that produced before, the balanceable Gold sequence after amplitude set.

In this described mode, balanceable Gold sequence is that according to spread spectrum and coding theory, two r level shift registers have 2 by two preferably right m sequences mould two and obtaining by turn ^r-1individual different relative shift, adds two original m sequence preferences pair, and level shift register can produce 2 ^r+1individual Gold sequence, Gold sequence quantity is much more a lot of than m sequence quantity, at a Gold sequence period N=2 ^rin-1, wherein, the length that r is shift register, if element 1 is respectively 2 with the number of times that element 0 occurs ^r-1inferior and 2 ^r-1-1 time, element 1 than element 0, occur often 1, be called balanceable Gold sequence.

Consider to be applied in the design of random phase shift sequence, 1 and 0 position in the balanceable Gold sequence of generation can be set to respectively and be greater than 1 and be less than 1 constant, then multiply each other with random phase shift sequence.Cycle due to Gold sequence is odd number in addition, during design, only need after the sequence producing, mend one 0.

Other step and parameter are identical with embodiment one.

Embodiment three: present embodiment is different from embodiment one or two: in described step 6 P and Q to choose P and the Q used with transmitting terminal set identical.

Other step and parameter are identical with embodiment one or two.

Embodiment four: present embodiment is different from one of embodiment one to three: in described step 6, choosing of P and Q will guarantee ofdm system gross energy conservation.

Other step and parameter are identical with one of embodiment one to three.

Emulation experiment:

Emulation platform is MATLAB software, supposes that baseband signal adopts QPSK modulation, and oversample factor is 4, and the number of sub carrier wave of ofdm system is Num=32, and random phase shift sequence number is D=4, and the span of random phase shift vector is chosen for channel model is Gaussian white noise channel.In without sideband transmission SLM scheme and modified model SLM scheme, the P value of amplitude mark is all taken as 1.2, PAPR threshold value from 2dB to 10dB, contrasts the CCDF figure of several schemes, and simulation result as shown in Figure 3.Sub-carrier number is extended to Num=64, and random phase shift sequence number is D=8, and other parameters are the same, and simulation result as shown in Figure 4.From Fig. 3 and Fig. 4, all can obviously find out, modified model SLM scheme has the effect of obvious inhibition peak-to-average force ratio, although further do not reduce peak-to-average force ratio on the SLM scheme basis without sideband transmission, the curve of the two almost overlaps, and illustrates that they are similar to the inhibition degree of PAPR.But in modified model scheme, owing to having added balanceable Gold sequence, randomness strengthens, and the Euclidean distance between phase sequence increases, and this will make new departure improve aspect the error rate.

Simulated conditions is the same, and to modified model SLM method, the P value that balances the set of Gold sequence is taken as respectively 1.1,1.15,1.2,1.25,1.3, according to above-mentioned principle, carries out emulation, obtains simulation result as shown in Figure 5.As can be seen from Figure 5, when P value is less, as get 1.1 o'clock, error rate of system is higher.This is because P departs from 1 degree hour, and it is also less that corresponding Q departs from 1 degree, and signal, by ofdm system, under the impact of noise, is easily judged by accident the estimation of random phase shift sequence during receiving terminal decoding, causes the error rate to raise.When P is greater than certain numerical value, as P in figure gets 1.2,1.25 and at 1.3 o'clock, in the situation that signal to noise ratio is constant, along with the increase of P value, error rate of system also raises.This is because the P value when set departs from 1 degree when too large, it is also very large that corresponding Q value departs from 1 degree, can make the signal to noise ratio skewness weighing apparatus on subchannel, Euclidean distance between random phase shift sequence diminishes, and then produce larger error while causing receiving terminal to random phase shift sequencal estimation, finally cause the rising of the error rate.Meanwhile, when P value is too high, also too high to the threshold requirement of system, increased the complexity of system.Therefore, the value of P will consider various key elements, and compromise is chosen, and it is 1.2 to carry out emulation that analysis below will be chosen P value.

The impact of modified model SLM scheme on the bit error rate performance of ofdm system is discussed below.During emulation, still take sub-carrier number Num=32 as example, and carrier modulation adopts QPSK, and oversample factor is 4, and the amplitude P that balances the set of Gold sequence is taken as 1.2, without the amplitude P of sideband SLM method set, is also taken as 1.2, and simulation result as shown in Figure 6.From simulation result Fig. 6, can find out, when two schemes is got identical set value P, modified model SLM scheme ber curve is obviously at the downside without side information SLM method.This shows in the situation that not increasing system complexity, and the use of new method makes the error rate obtain obvious improvement.

Sub-carrier number is extended to Num=64, the amplitude of balanceable Gold sequence set and be still taken as 1.2 without the amplitude P of sideband SLM method set, simulation result is as shown in Figure 7.As seen from Figure 7, when sub-carrier number is increased to 64, modified model SLM method ber curve is below without sideband SLM scheme curve, and along with the increase of signal to noise ratio, modified model SLM method is than more obvious without sideband SLM method error performance advantage.

From Fig. 7 and Fig. 6, contrasted, to be 64 o'clock new departures be better than subcarrier to the improvement degree of the error rate to subcarrier is the situation of 32 o'clock.The introducing of visible balanceable Gold sequence has made up well without the high deficiency of the sideband transmission SLM scheme error rate, and sub-carrier number is larger, and the effect of improvement is more obvious without the situation of sideband, and this and theory are coincide.

Claims (4)

1. modified model SLM algorithm reduces a method of ofdm system PAPR, it is characterized in that the method for modified model SLM algorithm reduction ofdm system PAPR a kind of realizes according to the following steps:

Step 1, transmitting terminal, according to ofdm system subcarrier number, are determined the frequency domain data vector of ofdm signal in modified model SLM algorithm, are expressed as X=[X ₁, X ₂..., X _n] ^t, wherein N represents sub-carrier number, the transposition of T representing matrix;

Step 2, employing modified model SLM algorithm are processed the frequency domain data vector X of the ofdm signal of input:

Described modified model SLM algorithm is the frequency domain data vector X=[X of the ofdm signal of transmitting terminal input ₁, X ₂..., X _n] ^trespectively with the D road length random phase shift vector A that is N ^(d)balanceable Gold sequence phase weighting with D road after amplitude set obtains the separate output vector S in D road ^(d); S wherein ^(d)for frequency domain data vector, 1≤d≤D;

Random phase shift vector A ^(d)choosing method be:

i=1,2 ..., N; D=1,2 ..., D, obey being uniformly distributed between [0,2 π], j represents imaginary unit;

Step 3, the output vector S separate to resulting D road ^(d)carry out respectively IFFT computing, obtain D road time-domain signal s ^(d)=IFFT (S ^(d));

Step 4, to the D road time-domain signal s obtaining in step 3 ^(d)=IFFT (S ^(d)) calculate respectively its PAPR value, and select and there is minimum PAPR Zhi mono-road signal and send;

Step 5, receiving terminal carry out FFT computing to the D road signal receiving;

Step 6, after FFT computing to the D road signal delivery obtaining, and mould value be wherein greater than to 1 part be set to P, mould value is less than 1 part and is set to Q;

Step 7, ask in step 6 the sequence after set and pre-stored in receiving terminal the Hamming distance between improved local balanceable Gold sequence, obtain with set after sequence there is one group of smallest hamming distance local balanceable Gold sequence, which signal what estimate transmitting terminal transmitting is;

Step 8, according to the sequence number of the signal transmitted of estimating in step 7 to obtain, carry out to received signal the operation contrary with transmitting terminal, recover original signal information.

2. a kind of modified model SLM algorithm according to claim 1 reduces the method for ofdm system PAPR, it is characterized in that in described step 2, the balanceable Gold sequence after amplitude set is specially:

The r level shift register generation cycle is N=2 ^r-1 balanceable Gold sequence, one zero of the last benefit of balanceable Gold sequence, then " 1 " locational element in sequence is set to the constant P that is greater than 1, " 0 " locational element is set to the constant Q that is less than 1, by the balanceable Gold sequence after set more respectively with the random phase shift sequence phase dot product that produced before, be improved the random phase shift vector of type SLM algorithm.

3. a kind of modified model SLM algorithm according to claim 2 reduces the method for ofdm system PAPR, it is characterized in that P and Q in described step 6 to choose the P used with transmitting terminal set identical with Q.

4. a kind of modified model SLM algorithm according to claim 3 reduces the method for ofdm system PAPR, it is characterized in that in described step 6 that P and choosing of Q will guarantee ofdm system gross energy conservation.