CN100546234C - A kind of method of new reduction orthogonal frequency-division multiplex singal peak-to-average ratio - Google Patents

Abstract

The invention discloses a kind of method of new reduction orthogonal frequency-division multiplex singal peak-to-average ratio, it is to come the PAPR of signal is optimized by uniting a series of peak value optimal module.In the optimization process of i level peak value, predetermined thresholding is with m _iPAPR and this thresholding of bunch phase factor corresponding data compare one by one, are less than or equal to thresholding in case the PAPR of data occurs, think that promptly these data can satisfy the PAPR requirement of system, and the unnecessary optimization of carrying out back i+1 level.Packet mode difference in the processing at different levels, the number of clusters m of employed phase factor _iDifferent.Employed phase factor bunch

,

The value difference.This new method is compared with improving the PTS method, under the certain prerequisite of computation complexity, can obtain the effect of better reduction peak-to-average ratio.

Description

A kind of method of new reduction orthogonal frequency-division multiplex singal peak-to-average ratio

Technical field

The invention belongs to communication technical field, particularly ofdm system.

Background technology

As everyone knows, in the existing communication technology, OFDM (Orthogonal Frequency DivisionMultiplexing, abbreviation OFDM) technology declines and interference free performance with its very high availability of frequency spectrum, good anti-multipath, becomes one of main candidate technologies of following mobile multimedia communication.

Ofdm signal has peak power and average power ratio (Peak-to-average Power, the English PAPR that is called for short, Chinese is called for short peak-to-average ratio) higher shortcoming, to non-linear very sensitive, this has the very big range of linearity with regard to the high power amplifier that requires transmitter, but,, mean higher cost than the power device of great dynamic range.And, behind the linear unit of the ofdm signal that surpasses linear dynamic range, can produce significant inband distortion and out-of-band radiation, thereby cause the error rate to raise by power amplifier and so on.

Patent " a kind of new improvement PTS method " (number of patent application 02128076.2) has proposed a kind of improvement part transfer sequence (Partial Transmit Sequences, be called for short PTS) method be used to reduce the peak-to-average ratio of ofdm signal, its operation principle as shown in Figure 1: data point is divided into V group: A through grouping module 1 ⁽¹⁾, A ⁽²⁾A ⁽ⁱ⁾A ^(V)(i=1,2,3 ... V), each group is carried out IFFT2 respectively, and gained is respectively organized data and many bunches of phase factor α _j ⁽¹⁾α _j ⁽²⁾α _j ⁽ⁱ⁾α _j ^(V)(i=1,2,3 ... V, j=1,2,3 ... U) multiply each other by multiplier 3.Each group is by adder 4 stacks then, and above process is identical with traditional PTS method.Module 5 is innovative contents of this patent, relatively selects in 5 at thresholding, predetermined thresholding, and (m is the number of clusters of employed phase factor, m≤U as can be known with m ^V) PAPR and this thresholding of bunch phase factor corresponding data compare one by one, be less than or equal to thresholding in case the PAPR of data occurs, the phase factor of this data correspondence is the suboptimum phase factor so, and suboptimum phase factor respectively organize data corresponding with it are multiplied each other, superpose, export; Export the secondary information of relevant suboptimum phase factor simultaneously; If after m time is selected relatively, the data that PAPR is lower than thresholding do not occur, the cluster phase factor of so just selecting PAPR minimal data correspondence is the suboptimum phase factor, this bunch phase factor and the corresponding data of respectively organizing are multiplied each other, superpose, export, export the secondary information of relevant suboptimum phase factor simultaneously: processing method and the conventional P TS method of this moment are identical, and its workflow is seen Fig. 2.

The existing major advantage of improving the PTS method is: the computation complexity that has reduced conventional P TS method.But its shortcoming is: under the certain prerequisite of computation complexity, the control ability of peak-to-average ratio is still on the low side.

The present invention has proposed a kind of new method based on the above method, can obtain higher peak-to-average ratio control ability than former method under the certain situation of computation complexity.

Summary of the invention

Task of the present invention provides a kind of method of peak-to-average ratio of new reduction orthogonal frequency-division multiplex singal, adopts method of the present invention, can more effective reduction peak-to-average ratio.

Below introduce term used among the present invention earlier:

1) phase factor α _j(j=1,2,3 ... U) be a complex value constant.The set that bunch is defined as the one group of phase factor that multiplies each other respectively with grouped data of phase factor.

(j=1,2,3 ... U _i) represent among the present invention the value of employed phase factor in the optimization process of i level peak value.

2) grouping is meant that adopting certain way will import data is divided into different data blocks, and the present invention stipulates that the packet mode in the peak value optimization processes at different levels do not limit, but must distinguish to some extent between at different levels.

3) thresholding is meant the PAPR value of determining according to the concrete needs of system, when the PAPR of the signal of optimised processing is lower than this thresholding, thinks that promptly the PAPR of this signal can satisfy system requirements.

4) the optimum angle factor is meant the corresponding cluster phase factor of PAPR minimal data in the conventional P TS method.The suboptimum phase factor is meant the cluster phase factor of determining after thresholding is relatively selected in the present invention's resume module at different levels.

5) secondary information is meant the information that contains selected phase factor content in processing modules at different levels.

6) PAPR is the ratio of the peak power and the average power of signal, and the PAPR computational methods of ofdm signal are as follows:

Supposing the system has N subcarrier, and S (1), S (2) are arranged in the OFDM symbol ... the individual data point of S (N), and then this moment, PAPR was expressed as: $10\lg \frac{\underset{0\&le;k<N}{\max }{|S\left(k\right)|}^2}{\frac{1}{N}\underset{i=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{|S\left(i\right)|}^2}---\left(\mathrm{dB}\right)$

The present invention proposes a kind of method of new reduction orthogonal frequency-division multiplex singal peak-to-average ratio, the excellent processing module of its characteristic binding peak values at different levels comes signal is handled, as shown in Figure 3, the wherein workflow of i level peak value optimal module such as Fig. 4, shown in Figure 5 can be expressed as the following step:

Step 1, data point adopt i kind packet mode to be divided into V through grouping i module 7 _iGroup:

(k=1,2,3 ... V _i) (packet mode in the peak value optimization processes at different levels, the group of getting are counted V _iDifference, this is an innovation part of the present invention);

Step 2, the data of respectively organizing that step 1 is distributed are carried out IFFT2 respectively;

Step 3, constant d=1;

Step 4, d bunch of phase factor of selection;

Step 5, the data of respectively organizing in the step 2 are multiplied each other, superpose with selected d bunch of phase factor respectively, produce data;

Step 6, the peak-to-average ratio of the data that step 5 produced is compared with thresholding, if be less than or equal to thresholding, cluster phase factor that then will be corresponding with these data multiplies each other, superpose, directly exports with each group data, exports secondary information simultaneously, the processing end; If be higher than thresholding, then enter step 7;

Step 7, select before the PAPR minimal data of d bunch of phase factor correspondence;

Step 8, d=d+1;

Step 9, with total number of clusters m of d and the employed phase factor of i level _iCompare, if d＞m _i, then enter step 10; Otherwise enter step 4;

Step 10, will multiply each other, superpose, input to i+1 level module and handle with the data of respectively organizing of gained in corresponding cluster phase factor of gained data and the step 1.

Fundamental block diagram of the present invention comprises the 1st grade of peak value optimal module 6 as shown in Figure 3 ¹, the 2nd grade of peak value optimal module 6 ²... i level peak value optimal module 6 ⁱ... wherein the number of peak value optimal module can require and flexible design according to the concrete PAPR of system, and whenever by one-level peak value optimization process module, its PAPR can be reduced signal in various degree.In general module progression is many more, and the ability that reduces peak-to-average ratio is strong more, but computation complexity also can raise to some extent.

The operation principle of peak value optimal module at different levels will be introduced in the back.

Workflow of the present invention: as shown in Figure 3, the input data

After the 1st grade of peak value optimization process resume module, if the PAPR of signal satisfies system requirements after the 1st grade of peak value optimization, think that then it is desirable peak signal, thereby stop to handle, desirable peak signal 1 and corresponding secondary information 1 are directly exported; If the PAPR of signal thinks then that less than the pedal system requirement it is imperfect peak signal, is designated as after the 1st grade of peak value optimization

It is sent into the 2nd grade of peak value optimization process module continues to handle.And the like, when the input data

After i level peak value optimization process resume module, if the PAPR of signal satisfies system requirements after the optimization of i level peak value, think that then it is desirable peak signal, thereby stop to handle, desirable peak signal i and corresponding secondary information i are directly exported; If the PAPR of signal thinks then that less than the pedal system requirement it is imperfect peak signal, is designated as after the optimization of i level peak value

It is sent into i+1 level peak value optimization process module continues to handle.Here i＞1.

From the above, the similar patent of method of inside modules peak value at different levels optimization process " a kind of new improvement PTS method " (number of patent application 02128076.2) employed method, but the present invention is not the simple reprocessing of the described method of patent, but following difference is arranged on processing method:

1) the packet mode difference in the processing at different levels: wherein the i level is handled the i kind packet mode that adopts.The group of getting is counted V _iDifferent.

2) the number of clusters m of employed phase factor in the processing at different levels _iDifferent.

3) employed phase factor bunch in the processing at different levels

The value difference.

When satisfying above-mentioned condition between the processing at different levels, just have independence between the peak value optimization processes at different levels, can obtain best peak value like this and optimize effect.And if only patent " a kind of new improvement PTS method " (number of patent application 02128076.2) method is carried out simple reprocessing, when not possessing above-mentioned three innovative points, just do not have independence between the processing procedure that repeats or independence is very little, the effect that reduces peak-to-average ratio is just bad.So the stack of the scheme that the solution of the present invention can provide than patent " a kind of new improvement PTS method " (number of patent application 02128076.2) has better peak value optimization effect.

Another innovation of the present invention is: unite a series of peak value optimal module and come the PAPR of signal is optimized.In the optimization process of i level peak value, predetermined thresholding is with m _iPAPR and this thresholding of bunch phase factor corresponding data compare one by one, be less than or equal to thresholding in case the PAPR of data occurs, think that promptly these data can satisfy the PAPR requirement of system, and the unnecessary optimization of carrying out back i+1 level is done like this and can be simplified computation complexity.Thereby make the complexity of computation complexity of the present invention far below the inventive method simple superposition of patent " a kind of new improvement PTS method " (number of patent application 02128076.2).

Conclusion: in sum, the new method of this reduction peak-to-average ratio is optimized processing owing to united a plurality of peak value optimal module, therefore can better reduce peak-to-average ratio; Can reduce the computation complexity that the module Combined Treatment is introduced by simplifying to handle simultaneously.

One of innovative point of the present invention: existing improvement PTS method, at computation complexity one regularly, the ability that reduces peak-to-average ratio is on the low side.The present invention has proposed a kind of method of new reduction ofdm signal peak-to-average ratio to this, unite a series of peak value optimal module and come the PAPR of signal is optimized.In the optimization process of i level peak value, predetermined thresholding is with m _iPAPR and this thresholding of bunch phase factor corresponding data compare one by one, be less than or equal to thresholding in case the PAPR of data occurs, think that promptly these data can satisfy the PAPR requirement of system, and the unnecessary optimization of carrying out back i+1 level is done like this and can be simplified computation complexity.

Two of innovative point of the present invention: in the process of Combined Treatment, have following difference between the peak value optimization process modules at different levels:

1) the packet mode difference in the processing at different levels: wherein the i level is handled the i kind packet mode that adopts.

2) the number of clusters m of employed phase factor in the processing at different levels _iDifferent.

3) employed phase factor bunch in the processing at different levels

The value difference.

Project Realization of the present invention is compared with conventional P TS method, and the characteristics that have are: unite a series of peak value optimal module and come the PAPR of signal is optimized.In the optimization process of i level peak value, predetermined thresholding is with m _iPAPR and this thresholding of bunch phase factor corresponding data compare one by one, are less than or equal to thresholding in case the PAPR of data occurs, think that promptly these data can satisfy the PAPR requirement of system, and the unnecessary optimization of carrying out back i+1 level.Packet mode difference in the processing at different levels, the number of clusters m of employed phase factor _iDifferent.Employed phase factor bunch

The value difference.This new method is compared with improving the PTS method, under the certain prerequisite of computation complexity, can obtain the effect of better reduction peak-to-average ratio.

Description of drawings

Fig. 1 is the fundamental diagram that improves the PTS method:

Wherein, the 1st, grouping module, the 2nd, the IFFT module, the 3rd, multiplier module, the 4th, addition module, the 5th, thresholding is relatively selected module.

A ⁽¹⁾, A ⁽²⁾A ^(V)Be respectively 1,2 ... the set of data points of V group, α _j ⁽¹⁾, α _j ⁽²⁾α _j ^(V)Expression is corresponding to 1,2 ... the phase factor of V group, j=1 wherein, 2,3 ... U;

Expression through thresholding relatively select to obtain corresponding to 1,2 ... the suboptimum phase factor of V group, j=1 wherein, 2,3 ... U;

Fig. 2 is the workflow diagram that improves the PTS method;

Fig. 3 is the fundamental diagram of new method proposed by the invention

Wherein, 6 ¹Be the 1st grade of peak value optimal module, 6 ²Be the 2nd grade of peak value optimal module ... 6 ⁱBe i level peak value optimal module i,

Be the input data, Be respectively through the 1st, after the optimization process of 2...i level peak value,, need to continue the signal of handling still less than the pedal system requirement.

Fig. 4 is the fundamental diagram of i level peak value optimization process module:

Wherein, the 7th, grouping i module, the 2nd, the IFFT module, the 3rd, multiplier module, the 4th, addition module, the 5th, thresholding is relatively selected module.

Be respectively 1,2 ... V _iThe set of data points of group,

Expression is corresponding to 1,2 ... V _iThe phase factor of group, j=1 wherein, 2,3 ... U _i

Expression through thresholding relatively select to obtain corresponding to 1,2 ... V _iThe suboptimum phase factor of group, j=1 wherein, 2,3 ... U _i

Fig. 5 is the workflow diagram of i level peak value optimization process module.

Embodiment

Main innovation of the present invention is that a series of peak value optimal module of associating come the PAPR of signal is optimized.Main thought of the present invention is embodied in the fundamental diagram of improvement PTS method shown in Figure 1: wherein, the 1st, grouping module, the 2nd, the IFFT module, the 3rd, multiplier module, the 4th, addition module, the 5th, thresholding is relatively selected module; A ⁽¹⁾, A ⁽²⁾A ^(V)Be respectively 1,2 ... the set of data points of V group, α _j ⁽¹⁾, α _j ⁽²⁾α _j ^(V)Expression is corresponding to 1,2 ... the phase factor of V group, j=1 wherein, 2,3 ... U;

Expression through thresholding relatively select to obtain corresponding to 1,2 ... the suboptimum phase factor of V group, j=1 wherein, 2,3 ... U.

The present invention finishes concrete enforcement by multistage peak value optimal module, as shown in Figure 3: wherein, 6 ⁱBe the 1st grade of peak value optimal module, 6 ²Be the 2nd grade of peak value optimal module ... 6 ⁱBe i level peak value optimal module i;

Be the input data,

Be respectively through the 1st, after the optimization process of 2...i level peak value,, need to continue the signal of handling still less than the pedal system requirement.

In i level peak value optimization process shown in Figure 3, predetermined thresholding is with m _iPAPR and this thresholding of bunch phase factor corresponding data compare one by one, are less than or equal to thresholding in case the PAPR of data occurs, think that promptly these data can satisfy the PAPR requirement of system, and the unnecessary optimization of carrying out back i+1 level.The execution mode of i level peak value optimization process module is as shown in Figure 4: wherein, the 7th, grouping i module, the 2nd, the IFFT module, the 3rd, multiplier module, the 4th, addition module, the 5th, thresholding is relatively selected module;

Be respectively 1,2 ... V _iThe set of data points of group, Expression is corresponding to 1,2 ... V _iThe phase factor of group, j=1 wherein, 2,3 ... U _i

Expression through thresholding relatively select to obtain corresponding to 1,2 ... V _iThe suboptimum phase factor of group, j=1 wherein, 2,3 ... U _iIn the peak value optimization process, the packet mode difference in the processing at different levels, the number of clusters m of employed phase factor _iDifferent.Employed phase factor bunch

The value difference.

New method proposed by the invention is compared with improving the PTS method, under the certain prerequisite of computation complexity, can obtain more The effect of good reduction peak-to-average ratio. Each module in the system can realize by the method for programming, utilizes prior art passable Be made into corresponding hardware, then, unite other modules, form whole system of the present invention.

Claims (1)

1, flat the method in peak of a kind of new reduction orthogonal frequency-division multiplex singal is characterized in that adopting the following step:

Step 1, data point adopt i kind packet mode to be divided into V through grouping i module (7) _iGroup: A ⁽¹⁾, A ⁽²⁾A ^(k)

K=1,2,3 ... V _i

Step 2, the data of respectively organizing that step 1 is distributed are carried out IFFT respectively, are respectively organized dateout;

Step 3, constant d=1;

Step 4, d bunch of phase factor: α of selection _d ⁽¹⁾, α _d ⁽²⁾α _d ^(k)

Step 5, step 2 is handled the resulting selected d bunch of phase factor α in dateout and the step 4 that respectively organize by IFFT _d ⁽¹⁾, α _d ⁽²⁾α _d ^(k) Multiply each other k=1 wherein, 2,3 successively ... V _i, then the V that obtains _iIndividual results added produces data;

Step 6, predetermined peak-to-average ratio thresholding are compared the peak-to-average ratio of the data that step 5 produced with thresholding, if be less than or equal to thresholding, cluster phase factor that then will be corresponding with these data and step 2 are exported respectively organize data multiply each other successively, then V _iIndividual results added, directly output, the secondary information of the selected d bunch of phase factor of output representative simultaneously, processing finishes; If be higher than thresholding, then enter step 7;

Step 7, select before the PAPR minimal data of d bunch of phase factor correspondence;

Step 8, d=d+1;

Step 9, with total number of clusters m of d and the employed phase factor of i level _iCompare, if d＞m _i, then enter step 10; Otherwise enter step 4;

Step 10, will with gained in the corresponding cluster phase factor of minimum data of step 7 output and the step 1 respectively organize data A ⁽¹⁾, A ⁽²⁾A ^(k)

Multiply each other successively, then V _iIndividual results added inputs to i+1 level module and handles.

Images