CN107979454B - Method for rapidly extracting OFDMA distributed mode downlink comb spectrum - Google Patents

Method for rapidly extracting OFDMA distributed mode downlink comb spectrum Download PDF

Info

Publication number
CN107979454B
CN107979454B CN201711170826.2A CN201711170826A CN107979454B CN 107979454 B CN107979454 B CN 107979454B CN 201711170826 A CN201711170826 A CN 201711170826A CN 107979454 B CN107979454 B CN 107979454B
Authority
CN
China
Prior art keywords
block
split
radix
data
level
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201711170826.2A
Other languages
Chinese (zh)
Other versions
CN107979454A (en
Inventor
余华
裴欣月
邹游
季飞
陈芳炯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
South China University of Technology SCUT
Original Assignee
South China University of Technology SCUT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by South China University of Technology SCUT filed Critical South China University of Technology SCUT
Priority to CN201711170826.2A priority Critical patent/CN107979454B/en
Publication of CN107979454A publication Critical patent/CN107979454A/en
Application granted granted Critical
Publication of CN107979454B publication Critical patent/CN107979454B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/0001—Arrangements for dividing the transmission path
    • H04L5/0003—Two-dimensional division
    • H04L5/0005—Time-frequency
    • H04L5/0007—Time-frequency the frequencies being orthogonal, e.g. OFDM(A), DMT
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0037—Inter-user or inter-terminal allocation
    • H04L5/0041—Frequency-non-contiguous
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/0091—Signaling for the administration of the divided path
    • H04L5/0094—Indication of how sub-channels of the path are allocated

Abstract

The invention provides a method for quickly extracting an OFDMA distributed mode downlink comb spectrum. Aiming at N time domain receiving signals of a system, the invention adopts discrete Fourier transform to extract L multiplied by M comb-shaped frequency points, namely the problem of quick realizationLStarting at the same position in each segment of the sub-bandMExtracting the data carried on continuous sub-carriers, and decomposing the data by using a transform decomposition methodNAn input data is decomposed intoLHas a length ofPThe data block (2) is simplified by performing a Discrete Fourier Transform (DFT) operation on each block by using a split-radix clipping method based on time extractionLThe DFT operation of the block data disperses the extra rotation operation brought by the output frequency shift to the rotation factors of each stage of operation and performs the tailored butterfly operation on some stages, thereby realizing the rapid extraction of all comb-shaped frequency points allocated to the user at the user terminal and greatly reducing the operation complexity. The method can be applied to the fields of signal processing or image processing and the like, and realizes the extraction of the discrete Fourier transform comb-shaped frequency points.

Description

Method for rapidly extracting OFDMA distributed mode downlink comb spectrum
Technical Field
The invention relates to the field of communication signal processing, in particular to a method for quickly extracting comb spectrums when a downlink user terminal receives and demodulates in a distributed subcarrier allocation mode aiming at an Orthogonal Frequency Division Multiple Access (OFDMA) communication system.
Background
Orthogonal Frequency Division Multiple Access (OFDMA) is the mainstream multiple access scheme for the downlink of mobile communication systems such as LTE, 4G, and IEEE 802.16 m. OFDMA divides the entire frequency band into many subcarriers, converts a frequency selective fading channel into several flat fading subchannels, and thus can effectively resist frequency selective fading in a wireless mobile environment. By allocating different sub-carriers to different users, the sub-carriers between the users meet the requirement of mutual orthogonality, and the users can access the system at the same time. OFDMA may employ two subcarrier allocation patterns: centralized and distributed. Several continuous sub-carriers are distributed to one user in a centralized mode, and sub-carriers distributed to one user in a distributed mode are dispersed to the whole frequency band, so that a comb-shaped frequency spectrum is formed. The difficulty of channel estimation can be reduced in a centralized manner, but the frequency diversity gain obtained by the method is small, and the average performance of users is slightly poor; and the distributed mode can make the frequency selective fading correlation between the subcarriers weaker by utilizing the farther interval between the subcarriers, thereby obtaining stronger diversity gain and better user average performance, but the channel estimation is more complex in the mode. The two modes can be flexibly selected according to actual conditions. The patent mainly researches the extraction problem of comb spectrum when a user terminal receives and demodulates in a distributed subcarrier allocation mode downlink of an OFDMA communication system. The current main method is to adopt the conventional FFT methods such as radix-2 or split radix to extract the data carried on all subcarriers, and the invention cuts the conventional FFT method and realizes the rapid extraction of the data on the comb-shaped frequency point.
Disclosure of Invention
The present invention is directed to overcome the drawbacks and disadvantages of the prior art, and to provide a method for rapidly extracting a comb spectrum when a downlink user terminal receives and demodulates a distributed subcarrier allocation pattern for an Orthogonal Frequency Division Multiple Access (OFDMA) communication system.
Assuming that the total number of subcarriers in the system is N, all subcarriers are divided into subcarriers of length P (P ═ 2)m) L blocks of (i.e., N ═ L × P), each user starts occupying M (M ═ 2) contiguous at the same position in P carriers of each blockd<P) sub-carriers, user signals are comb-shaped frequency spectrum, the user terminal only needs to extract the data loaded on the L multiplied by M sub-carriers allocated to the user terminal for receiving and processing, and the data on the other sub-carriers are irrelevant to the user and do not need to be extracted. The conventional Fast Fourier Transform (FFT) methods such as a radix-2 or a split radix are adopted to extract data borne by all subcarriers in the conventional method, and the conventional FFT method is cut, so that the fast extraction of the data on the required frequency point is realized; the case where both P and M are powers of 2 is limited. The invention realizes DFT comb shape based on transform decomposition and butterfly graph cuttingAnd (4) rapidly solving the spectrum. The specific method is that N input data are decomposed into L blocks with the length of P points by a transform decomposition method to be respectively subjected to DFT operation, the L DFT blocks with the length of P are simply calculated by a split-based cutting method based on time extraction DIT, extra rotation operation brought by output frequency shift is dispersed on twiddle factors of each stage of operation, and butterfly operation of cutting is carried out on some stages, so that a fast operation method is realized, all comb frequency points allocated to a user are quickly extracted at a user terminal, and operation complexity is greatly reduced.
Specifically, for N input data [ x (0), x (1), x (2), …, x (N-1)]Calculating the comb output of the discrete Fourier transform; note that N points DFT of N input data are [ X (0), X (1), X (2), …, X (N-1)]Dividing the output N data into L blocks with the length of P, and setting M continuous frequency points which are occupied by a user from the Kth position to the K + M-1 position in each block; for the user, the total L multiplied by M output data of the L blocks are recorded as [ X (K), X (K +1), …, X (K + M-1); x (P + K), X (P + K +1), …, X (P + K + M-1); …, respectively; x ((L-1) P + K), X ((L-1) P + K +1), …, X ((L-1) P + K + M-1)](ii) a Particularly, in the present invention, the arbitrary continuous M data in each block of output data also includes the case that the M data are continuous in a circular loop, that is, the M data connected end to end in each block of output data also belong to the case to be considered; in the technical scheme, input data is equally divided into L blocks with the length of P by using a transformation decomposition mode, and the r-th block is marked as [ x ]r-1(0),xr-1(1),…,xr-1(P-1)]R is 1, …, L; the calculation of each block is realized in a hierarchical mode, and each block of input data has m levels, namely P is 2mWhere d-level operations cannot be clipped, i.e. M2d(ii) a The output of the t-th stage computation of the r-th block is denoted as [ x ](r-1),t(0),x(r-1),t(1),…,x(r-1),t(P-1)]The output of the t stage is the input of the t +1 stage, and the first M continuous data output by the mth stage of the mth block is the operation result of the block;
any continuous M data in each block of output data are solved, and the data are shifted to the left by cyclic shift and K bits to be changed to the 0 th bit to the M-1 th bit of the output end; cycling of the output according to the nature of the discrete Fourier transformLeft shifting by K bits is equivalent to rotating each data at the input of each block, i.e. multiplying the nth input data by a frequency shift twiddle factor
Figure GDA0002749165130000021
Shifting the frequency by a rotation factor
Figure GDA0002749165130000022
Dispersing the data to original twiddle factors of each level of a butterfly diagram of a split-radix method, and quickly solving the segmented DFT comb spectrum output, thereby realizing the quick extraction of all comb frequency points allocated to users, and specifically comprising the following steps:
step 1, dividing N time-lapse input data into L blocks with the length of P by using a transformation decomposition mode;
step 2, calculating the total series m of the split-radix butterfly diagram of each block of data and the series d which does not need to be cut for L blocks with the length of P, and carrying out bit reverse order on the serial number of each block of input data;
step 3, calculating all time extraction radix-2 butterflies in the level 1 of each block for L blocks with the length of P;
step 4, for L blocks with the length of P, calculating all the time from the 2 nd level to the d-th level of each block to extract split-radix butterflies;
step 5, calculating all the cut split-radix butterflies from the d +1 th level to the m th level of the last level of each block with L blocks with the length of P according to a split-radix cutting method;
and 6, recombining the L output blocks with the length of P obtained in the step to obtain the required comb-shaped output.
Further, in step 1, for the input of N points, the sequence { x (Ln)1)},n1P-1 maps to a first block of length P, denoted x respectively, …0(0),x0(1),…,x0(P-1); will sequence { x (Ln)1+1)},n1P-1 maps to a second block of length P, denoted x respectively, …1(0),x1(1),…,x1(P-1); …, respectively; will sequence { x (Ln)1+L-1)},n1P-1 maps to the lth block of length P, denoted as P, 0, …xL-1(0),xL-1(1),…,xL-1(P-1)。
In step 2, the total stage number of the DIT split radix butterfly diagram for the input of the P point in each block
Figure GDA0002749165130000035
M-point arbitrary continuous output, no need of cutting stage number
Figure GDA0002749165130000036
The number of stages to be clipped is m-d. Input to the r-th block { xr-1(0),xr-1(1),xr-1(2),…,xr-1The serial number of (P-1) is subjected to bit reversal, specifically, P serial numbers of 0,1, … and P-1 are represented by m-level binary, and then the binary numbers are inverted and represented as new decimal serial numbers. Reordering the order of elements in the input sequence by the new decimal sequence number and noting as { x(r-1)0(0),x(r-1)0(1),…,x(r-1)0(P-1), which will be the input to the 1 st stage operation of the r-th block.
In step 3, all time extraction radix-2 butterflies in level 1 of each of the L blocks with the length of P are respectively calculated. For the r-th block, all of the time-decimated radix-2 butterflies in stage 1 are performed. The formula for the time-decimated radix-2 butterfly may be expressed as:
Figure GDA0002749165130000031
where p and q +1 denote the serial numbers of the upper and lower nodes operated by the butterfly unit. Only p is 2 × 4a-2+b×4a+1The radix-2 butterfly at (a) needs to be computed, where a is 0,1,2, …; b is 0,1,2, …; so that P is less than P-1.
In step 4, all time extraction split-radix butterflies from the 2 nd level to the d-th level of each block are calculated respectively. For the r-th block, stages 2 through d perform a time-decimating split-radix butterfly operation. T in the split-radix butterfly denotes the t-th level, NiDenotes the size of the twiddle factor block in which the split base is located, where Ni＝2tAnd t is more than or equal to 2 and less than or equal to m. Input terminalFrequency shift twiddle factor
Figure GDA0002749165130000032
The twiddle factors distributed after the intrinsic twiddle factors of the 2 nd to d nd stages are
Figure GDA0002749165130000033
Wherein
q(u)＝(u-K)·2(m-t)，u＝0,1,…,2t-2-1, t ═ 2, …, m equation (3)
In particular, when u is 0, phi (u) and phi (3u) are the twiddle factors of the first split-base butterfly of the twiddle factor block in which the t-th split-base is located, respectively; when u is 1, phi (u) and phi (3u) are respectively twiddle factors of the second split-base butterfly of the twiddle factor block where the t-th split base is located, and so on. The formula for the computation of the time-decimated split-radix butterfly can be expressed as:
Figure GDA0002749165130000034
where j is the unit of an imaginary number,
n＝Ni×4a+1-Ni+2b×Ni×4a+1+ u formula (5)
Wherein a is-1, 0,1,2, …; b is 0,1,2, …; u is 0,1, …,2t-2-1, such that N < P-3Ni/4。
And extracting the split radix butterfly from the 2 nd stage to the d-th stage at all times to perform complete split radix butterfly operation according to the formula.
In step 5, all the clipped split radix butterflies from the d +1 th level to the m-th level of the last level of each block are calculated respectively. For the r block, according to a time-extracted split-radix clipping method, butterfly operations for clipping are performed on the split bases from the (d +1) th level to the m-th level of the last level in an output clipping manner, and the specific clipping manner is as follows:
in the d +1 th stage, only the first two of the time-decimated split-radix butterflies are computedOutputting, cutting time extracting the last two outputs of the split-radix butterfly, wherein the first two outputs are xt(n) and
Figure GDA0002749165130000041
t in the split-radix butterfly denotes the t-th level, NiDenotes the size of the twiddle factor block in which the split base is located, where Ni＝2d+1Frequency shift twiddle factor
Figure GDA0002749165130000042
The twiddle factors φ (u) and φ (3u) dispersed to the d +1 th stage original twiddle factors are calculated according to the equations (2) and (3). n is calculated according to equation (5). The formula for computing the split-radix butterfly in this step can be expressed as:
Figure GDA0002749165130000043
in each of the stages from the (d + 2) th stage to the mth stage of the last stage, the first M split radix butterflies are calculated for each rotation factor block of each stage, and each split radix butterfly calculates only the first output. Wherein N isi＝2tD +2, …, m, frequency shift twiddle factor
Figure GDA0002749165130000044
The twiddle factors phi (u) and phi (3u) distributed after the original twiddle factors from the d +2 th stage to the M-th stage of the last stage are calculated according to the formula (2) and the formula (3), except that the value range of u is 0,1, …, and M-1. The formula for computing the split-radix butterfly in this step can be expressed as:
Figure GDA0002749165130000045
wherein n is calculated according to formula (5), except that u is in the range of 0,1, …, and M-1.
In step 6, L blocks with length P are recombined to obtain comb-shaped output. In L blocks with the length of P, recombining output required by the user by using the first M data in each block, and calculating comb spectrum output [ X (K), X (K +1), … and X (K + M-1) required finally by using the following formula; x (P + K), X (P + K +1), …, X (P + K + M-1); …, respectively; x ((L-1) P + K), X ((L-1) P + K +1), …, X ((L-1) P + K + M-1) ].
Figure GDA0002749165130000046
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the invention provides a method for quickly extracting comb spectrum when a distributed subcarrier allocation mode downlink user terminal of an Orthogonal Frequency Division Multiple Access (OFDMA) communication system receives and demodulates, which can quickly evaluate the continuous comb spectrum with the output of DFT being equal space, and greatly reduces time complexity compared with the conventional FFT methods such as radix-2 and split radix which must completely solve all point values.
2. The invention provides a method for decomposing N-point DFT into smaller blocks by means of transform decomposition to respectively carry out DFT operation, and in addition, the rotation of the frequency shift of the DFT operation output end of each block to the input end is ingeniously dispersed to the original twiddle factors of each level of a time extraction DIT butterfly diagram, and the output cutting of a split basis is combined, so that the operation complexity is greatly reduced, the value of DFT at a comb spectrum output point is required in the fields of signal processing or image processing and the like, and when the requirement on the time complexity is higher, the method can be used for replacing the traditional FFT operation.
Drawings
Fig. 1 is a downlink user signal comb spectrum of an OFDMA distributed subcarrier allocation pattern according to the present invention.
FIG. 2 is a graph of the temporal decimation base 2 and the split-base butterfly according to the present invention.
FIG. 3 is a two cropping butterflies of the time-decimated split-radix of the present invention.
FIG. 4 is a 16-point time-decimated DIT split-radix butterfly diagram of the present invention.
Fig. 5 is a schematic flow chart of comb spectrum output extraction according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but the present invention is not limited thereto.
Assuming that the total number of subcarriers of an OFDMA communication system is N, all subcarriers are divided into L blocks with a length of P, each user occupies M consecutive data from the K-th position to the K + M-1-th position in each P carriers of each block, and a user signal has a comb-shaped spectrum, as shown in fig. 1. The user terminal only needs to extract the data loaded on the L multiplied by M sub-carriers allocated to the user terminal, and the data on the rest sub-carriers are irrelevant to the user terminal and do not need to be extracted.
Specifically, a method for rapidly extracting a comb spectrum when a downlink user terminal receives and demodulates in a distributed subcarrier allocation mode in an OFDMA communication system is described by taking an example where N is 48 points, a decomposition factor L is 3, a block length P is 16, and each block continuously outputs M is 4 points from a K-6 point, including the steps of:
step 1, for the input of N points, mapping the N points to L blocks with the length of P by using a transformation decomposition mode.
Further, in step 1, for the input of N points, the sequence { x (Ln)1)},n1P-1 maps to a first block of length P, denoted x, 0, …, respectively0(0),x0(1),…,x0(P-1); will sequence { x (Ln)1+1)},n1P-1 maps to a second block of length P, denoted x respectively, 0, …1(0),x1(1),…,x1(P-1); …, respectively; will sequence { x (Ln)1+L-1)},n1P-1 maps to the lth block of length P, denoted x, 0, …, respectivelyL-1(0),xL-1(1),…,xL-1(P-1)。
Further, in step 1, as shown in the process from (a) to (b) in fig. 5, x (0), x (3), x (6), x (9), …, x (45) are mapped to the first block with length of 16, which is denoted as x0(0),x0(1),x0(2),x0(3),…,x0(15) (ii) a Mapping x (1), x (4), x (7), x (10), …, x (46) into a second block of length 16, denoted x1(0),x1(1),x1(2),x1(3),…,x1(15) (ii) a Mapping x (2), x (5), x (8), x (11), …, x (47) into a third block of length 16, denoted x2(0),x2(1),x2(2),x2(3),…,x2(15)。
And 2, calculating the total series of each split radix butterfly graph and the series required to be cut for L blocks with the length of P, and performing bit reverse order on the input point of each block.
Further, in step 2, the total number of stages of the input DIT split radix butterfly map for the P point in each block is
Figure GDA0002749165130000062
M points are output randomly and continuously, and the stage number without cutting is
Figure GDA0002749165130000063
The number of stages to be clipped is m-d. Input to the r-th block { xr-1(0),xr-1(1),xr-1(2),…,xr-1The serial number of (P-1) is subjected to bit reversal, specifically, P serial numbers of 0,1, … and P-1 are represented by m-level binary, and then the binary numbers are inverted and represented as new decimal serial numbers. Reordering the order of elements in the input sequence by the new decimal sequence number and noting as { x(r-1)0(0),x(r-1)0(1),…,x(r-1)0(P-1), which will be the input to the 1 st stage operation of the r-th block.
Further, in step 2, the data of 16 points in the first block is used for explanation, and as shown in fig. 4, the 16-point input M is a time-extracted DIT butterfly map which is output at 4 points arbitrarily and continuously. The 16-point DIT butterfly has 4 levels, the levels which do not need to be clipped are the 1 st level and the 2 nd level, and the levels which need to be clipped are the 3 rd level and the 4 th level. The 16 input points are subjected to bit reversal sequence, and then the input values are adjusted according to the new decimal serial numbers to obtain corresponding new decimal serial numbers x respectively00(0)(＝x0(0)),x00(1)(＝x0(8)),x00(2)(＝x0(4)),x00(3)(＝x0(12)),x00(4)(＝x0(2)),x00(5)(＝x0(10)),x00(6)(＝x0(6)),x00(7)(＝x0(14)),x00(8)(＝x0(1)),x00(9)(＝x0(9)),x00(10)(＝x0(5)),x00(11)(＝x0(13)),x00(12)(＝x0(3)),x00(13)(＝x013(11)),x00(14)(＝x0(7)),x00(15)(＝x0(15))。
And 3, calculating all time extraction radix-2 butterflies in the level 1 of each block for L blocks with the length of P.
Further, in step 3, all the time-decimated radix-2 butterflies in level 1 of each block are computed separately, and for the r-th block, all the time-decimated radix-2 butterflies in level 1 are performed. The formula for the time-decimated radix-2 butterfly may be expressed as:
Figure GDA0002749165130000061
where p and q +1 denote the serial numbers of the upper and lower nodes operated by the butterfly unit. Only p is 2 × 4a-2+b×4a+1The radix-2 butterfly at (a) needs to be computed, where a is 0,1,2, …; b is 0,1, 2; so that P is less than P-1.
Further, in step 3, all the time-decimated radix-2 butterflies in level 1 of each block are computed separately, for block 1, according to the formula for the computation of p, x in FIG. 400(0) And x00(1),x00(4) And x00(5),x00(6) And x00(7)、x00(8) And x00(9)、x00(12) And x00(13) Corresponding to the two input terminals of the radix-2 butterfly, respectively. When K is 6, the twiddle factor of radix-2 butterfly combined with offset coefficient is (-1)6The output value of each radix-2 butterfly of the first stage can be calculated according to equation (1) as 1.
And 4, for L blocks with the length of P, calculating all the time from the 2 nd stage to the d-th stage of each block to extract the split radix butterfly.
Further, in step 4, all time-decimated split-radix butterflies in level 2 through level d are computed for each block, respectively, as for the r-th block, level 2 through level dThe d-th stage performs a time-decimation split-radix butterfly operation. T in the split-radix butterfly denotes the t-th level, NiDenotes the size of the twiddle factor block in which the split base is located, where Ni＝2tAnd t is more than or equal to 2 and less than or equal to m. Input frequency shift twiddle factor
Figure GDA0002749165130000071
Twiddle factors distributed after intrinsic twiddle factors of 2 nd to d nd stages
Figure GDA0002749165130000072
Wherein
q(u)＝(u-K)·2(m-t)，u＝0,1,…,2t-2-1, t ═ 2, …, m equation (3)
In particular, when u is 0, phi (u) and phi (3u) are the twiddle factors of the first split-base butterfly of the twiddle factor block in which the t-th split-base is located, respectively; when u is 1, phi (u) and phi (3u) are respectively twiddle factors of the second split-base butterfly of the twiddle factor block where the t-th split base is located, and so on. The formula for the computation of the time-decimated split-radix butterfly can be expressed as:
Figure GDA0002749165130000073
where j is the unit of an imaginary number,
n＝Ni×4a+1-Ni+2b×Ni×4a+1+ u formula (5)
Wherein a is-1, 0,1,2, …; b is 0,1,2, …; u is 0,1, …,2t-2-1, such that N < P-3Ni/4。
The split radix butterfly is extracted from the 2 nd stage to the d nd stage at all times, and the complete split radix butterfly operation is carried out according to the formula.
Further, in step 4, computing the extracted split-radix butterfly at all times in each of the 2 nd to d-th stages, respectively, for the 1 st block, 16 points input the DIT split-radix butterfly with 4 points output in any sequence,only the second stage is a complete time-decimated radix-butterfly, the size N of the twiddle factor block in which the second stage's split bases are locatediThe values of n are 0, 8, 12, and u is 0 according to equation (5). When n is 0, 8, 12, the corresponding twiddle factors are expressed according to the formula (2) and the formula (3)
Figure GDA0002749165130000074
And
Figure GDA0002749165130000075
the output of each split-radix butterfly of the second stage can now be calculated according to equation (4).
And 5, calculating all the cut split-radix butterflies from the (d +1) th level to the (m) th level of the last level of each block for the L blocks with the length of P according to a split-radix cutting method.
Further, in step 5, all the tailored split-radix butterflies from the d +1 th level to the m th level of the last level of each block are respectively calculated, for the r-th block, the butterfly operation of tailoring is performed on the split bases from the d +1 th level to the m th level of the last level of the block in an output tailoring manner according to a split-radix tailoring method extracted in time, and the specific tailoring manner is as follows: in the d +1 stage, only the first two outputs of the time-decimated split-radix butterfly are calculated, and the second two outputs of the time-decimated split-radix butterfly are clipped, where the first two outputs are xt(n) and
Figure GDA0002749165130000076
t in the split-radix butterfly denotes the t-th level, NiDenotes the size of the twiddle factor block in which the split base is located, where Ni＝2d+1Frequency shift twiddle factor
Figure GDA0002749165130000081
The twiddle factors φ (u) and φ (3u) dispersed to the d +1 th stage original twiddle factors are calculated according to the equations (2) and (3). n is calculated according to equation (5). The formula for computing the split-radix butterfly in this step can be expressed as:
Figure GDA0002749165130000082
in each of the stages from the (d + 2) th stage to the mth stage of the last stage, the first M split radix butterflies are calculated for each rotation factor block of each stage, and each split radix butterfly calculates only the first output. Wherein N isi＝2tD +2, …, m, frequency shift twiddle factor
Figure GDA0002749165130000083
The twiddle factors phi (u) and phi (3u) distributed after the original twiddle factors from the d +2 th stage to the M-th stage of the last stage are calculated according to the formula (2) and the formula (3), except that the value range of u is 0,1, …, and M-1. The formula for computing the split-radix butterfly in this step can be expressed as:
Figure GDA0002749165130000084
wherein n is calculated according to formula (5), except that u is in the range of 0,1, …, and M-1.
Furthermore, in step 5, all clipped split-radix butterflies from the d +1 th level to the m-th level of the last level of each block are calculated according to the split-radix clipping method. For the first block, the 3 rd and 4 th split bases are subjected to butterfly operation of clipping by adopting an output clipping mode, and the specific mode is as follows: in stage 3, only the first two outputs of the time-decimated split-radix butterfly are computed. The size of the twiddle factor block where the 3 rd level splitting base is positioned is NiWhen u is 0, n is 0; when u is 1, n is 1. Thus, the two clipped outputs of the split-radix butterfly are x03(0) And x03(2)，x03(1) And x03(3). The output is x03(0) And x03(2) Has a rotation factor of
Figure GDA0002749165130000085
And
Figure GDA0002749165130000086
the output is x03(1) And x03(3) Butterfly shape ofThe twiddle factor is expressed by the following formula (2) and formula (3)
Figure GDA0002749165130000087
And
Figure GDA0002749165130000088
the first two outputs of the level 3 split radix butterfly can be calculated using equation (6). While at stage 4, only the first output of the time-decimated split-radix butterfly is computed. Wherein the size N of the twiddle factor block in which the 4 th level of the splitting base is locatediIn the 4 th stage, the value of n is obtained according to the formula (5), and when u is 0, n is 0, and the corresponding output is x04(0) (ii) a When u is 1, n is 1, corresponding to an output of x04(1) (ii) a When u is 2, n is 2, corresponding to an output of x04(2) (ii) a When u is 3, n is 3, corresponding to output x04(3). According to the formula (2) and the formula (3), the output is x04(0) Has a rotation factor of
Figure GDA0002749165130000089
And
Figure GDA00027491651300000810
the output is x04(1) Has a rotation factor of
Figure GDA00027491651300000811
And
Figure GDA00027491651300000812
the output is x04(2) Has a rotation factor of
Figure GDA00027491651300000813
And
Figure GDA00027491651300000814
the output is x04(3) Has a rotation factor of
Figure GDA00027491651300000815
And
Figure GDA00027491651300000816
the first output of the 4 th stage, 4 split radix butterfly, can be calculated using equation (7). All other outputs are not calculated.
And 6, carrying out recombination operation on the L blocks with the length of P to obtain the comb-shaped output required by us.
Further, in step 6, L blocks with length P are recombined to obtain the comb-like output we need. In L blocks with the length of P, recombining output required by the user by using the first M data in each block, and calculating comb spectrum output [ X (K), X (K +1), … and X (K + M-1) required finally by using the following formula; x (P + K), X (P + K +1), …, X (P + K + M-1); …, respectively; x ((L-1) P + K), X ((L-1) P + K +1), …, X ((L-1) P + K + M-1) ].
Figure GDA0002749165130000091
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.

Claims (5)

  1. A method for rapidly extracting an OFDMA distributed mode downlink comb spectrum is characterized by comprising the following steps: in a downlink of a distributed subcarrier allocation mode of an OFDMA communication system, assuming that the total number of subcarriers of the system is N, all the subcarriers are divided into L blocks with the length of P, each user starts to occupy continuous M subcarriers at the same position in each P blocks of the subcarriers, a user signal forms a comb-shaped frequency spectrum, a user terminal only needs to extract data loaded on L multiplied by M subcarriers allocated to the user terminal in receiving and processing, and the data on the rest subcarriers are irrelevant to the user and do not need to be extracted; the conventional FFT method is cut, so that the rapid extraction of data on required frequency points is realized, and the condition that P and M are both powers of 2 is limited;
    for N input data [ x (0), x (1), x (2), …,x(N-1)]Calculating the comb output of the discrete Fourier transform; note that N points DFT of N input data are [ X (0), X (1), X (2), …, X (N-1)]Dividing the output N data into L blocks with the length of P, and setting M continuous frequency points which are occupied by a user from the Kth position to the K + M-1 position in each block; for the user, the total L multiplied by M output data of the L blocks are recorded as [ X (K), X (K +1), …, X (K + M-1); x (P + K), X (P + K +1), …, X (P + K + M-1); …, respectively; x ((L-1) P + K), X ((L-1) P + K +1), …, X ((L-1) P + K + M-1)](ii) a The arbitrary continuous M data in each block of output data also includes the condition that the M data are circularly continuous in a circle, namely the M data which are connected end to end in each block of output data also belong to the condition to be considered; the input data is divided into L blocks with length P by using a transformation decomposition mode, and the r-th block is marked as [ x ]r-1(0),xr-1(1),…,xr-1(P-1)]R is 1, …, L; the calculation of each block is realized in a hierarchical mode, and each block of input data has m levels, namely P is 2mWhere d-level operations cannot be clipped, i.e. M2d(ii) a The output of the t-th stage computation of the r-th block is denoted as [ x ](r-1),t(0),x(r-1),t(1),…,x(r-1),t(P-1)]The output of the t stage is the input of the t +1 stage, and the first M continuous data output by the mth stage of the mth block is the operation result of the block;
    any continuous M data in each block of output data are solved, and the data are shifted to the left by cyclic shift and K bits to be changed to the 0 th bit to the M-1 th bit of the output end; according to the nature of discrete fourier transform, a cyclic left shift of K bits at the output is equivalent to rotating each data at the input of each block, i.e. multiplying the nth input data by a frequency shift twiddle factor
    Figure FDA0002749165120000011
    Shifting the frequency by a rotation factor
    Figure FDA0002749165120000012
    Dispersing the data to original twiddle factors of each level of a butterfly diagram of a split-radix method, and quickly solving the segmented DFT comb spectrum output, thereby realizing the quick extraction of all comb frequency points allocated to users, and specifically comprising the following steps:
    step 1, dividing N time input data into L blocks with length P by using a transformation decomposition mode: for N input data [ x (0), x (1), x (2), …, x (N-1)]The sequence { x (Ln)1)},n10, …, P-1, mapped into a first block of length P, denoted x respectively0(0),x0(1),…,x0(P-1); will sequence { x (Ln)1+1)},n10, …, P-1, into a second block of length P, denoted x respectively1(0),x1(1),…,x1(P-1); … …, respectively; will sequence { x (Ln)1+L-1)},n10, …, P-1, mapping into the lth block of length P, denoted x respectivelyL-1(0),xL-1(1),…,xL-1(P-1)；
    Step 2, calculating the total series m of the split-radix butterfly diagram of each block of data and the series d which does not need to be cut for L blocks with the length of P, and carrying out bit reverse order on the serial number of each block of input data: for each data block, the total number of stages of the DIT split radix butterfly is
    Figure FDA0002749165120000013
    M points are output randomly and continuously, and the stage number without cutting is
    Figure FDA0002749165120000014
    The stage number needing to be cut is m-d; input to the r-th block { xr-1(0),xr-1(1),xr-1(2),…,xr-1The serial number of (P-1) is subjected to bit reversal, specifically, P serial numbers of 0,1, … and P-1 are represented by m-level binary system, and then the binary system codes are inverted and represented as new decimal serial numbers; reordering the order of elements in the input sequence by the new decimal sequence number and noting as { x(r-1)0(0),x(r-1)0(1),…,x(r-1)0(P-1), which will be the input to the 1 st stage operation of the r-th block;
    step 3, calculating all time extraction radix-2 butterflies in the level 1 of each block for L blocks with the length of P;
    step 4, for L blocks with the length of P, calculating all the time from the 2 nd level to the d-th level of each block to extract split-radix butterflies;
    step 5, calculating all the cut split-radix butterflies from the d +1 th level to the m th level of the last level of each block with L blocks with the length of P according to a split-radix cutting method;
    and 6, recombining the L output blocks with the length of P obtained in the step to obtain the required comb-shaped output.
  2. 2. The method of claim 1 for fast extraction of OFDMA distributed mode downlink comb spectrum characterized by: in step 3, respectively calculating all time extraction radix-2 butterflies in the level 1 of each block in the L blocks with the length of P; for the r-th block, performing all of the time-decimated radix-2 butterflies in stage 1; the formula for the time-decimated radix-2 butterfly is given by:
    Figure FDA0002749165120000021
    wherein p and q are p +1 to represent the serial numbers of the upper and lower nodes operated by the butterfly unit; only p is 2 × 4a-2+b×4a+1The radix-2 butterfly at (a) needs to be computed, where a is 0,1,2, …; b is 0,1,2, …; so that P is less than P-1.
  3. 3. The method of claim 1 for fast extraction of OFDMA distributed mode downlink comb spectrum characterized by: in step 4, extracting the split radix butterfly at all times from the 2 nd level to the d-th level of each block respectively; for the r block, the 2 nd to d stages perform a time-decimation split-radix butterfly operation; t in the split-radix butterfly denotes the t-th level, NiDenotes the size of the twiddle factor block in which the split base is located, where Ni＝2tT is more than or equal to 2 and less than or equal to m; input frequency shift twiddle factor
    Figure FDA0002749165120000022
    The new twiddle factors after the twiddle factors distributed to the 2 nd to the d-th stages are
    Figure FDA0002749165120000023
    Wherein
    q(u)＝(u-K)·2(m-t)，u＝0,1,…,2t-2-1, t ═ 2, …, m equation (3)
    In particular, when u is 0, phi (u) and phi (3u) are the twiddle factors of the first split-base butterfly of the twiddle factor block in which the t-th split-base is located, respectively; when u is 1, phi (u) and phi (3u) are respectively twiddle factors of a second split-base butterfly of the twiddle factor block where the t-th split base is located, and so on; the formula for the computation of the time-decimated split-radix butterfly can be expressed as:
    Figure FDA0002749165120000031
    where j is the unit of an imaginary number,
    n＝Ni×4a+1-Ni+2b×Ni×4a+1+ u formula (5)
    Wherein a is-1, 0,1,2, …; b is 0,1,2, …; u is 0,1, …,2t-2-1, such that N < P-3Ni/4；
    And extracting the split radix butterfly from the 2 nd stage to the d-th stage at all times to perform complete split radix butterfly operation according to the formula.
  4. 4. The method of claim 1 for fast extraction of OFDMA distributed mode downlink comb spectrum characterized by: in step 5, calculating all the cut split radix butterflies from the d +1 th level to the m-th level of the last level respectively; for the r block, according to a time-extracted split-radix clipping method, butterfly operations for clipping are performed on the split bases from the (d +1) th level to the m-th level of the last level in an output clipping manner, and the specific clipping manner is as follows:
    in the d +1 stage, only the first two outputs of the time-decimated split-radix butterfly are calculated, and the second two outputs of the time-decimated split-radix butterfly are clipped, where the first two outputs are xt(n) and
    Figure FDA0002749165120000032
    t in the split-radix butterfly denotes the t-th level, NiDenotes the size of the twiddle factor block in which the split base is located, where Ni＝2d+1Frequency shift twiddle factor
    Figure FDA0002749165120000033
    New twiddle factors phi (u) and phi (3u) dispersed to the original twiddle factor of the (d +1) th level are calculated according to the formula (2) and the formula (3), and n is calculated according to the formula (5); the formula for computing the split-radix butterfly in this step can be expressed as:
    Figure FDA0002749165120000034
    in each stage from the (d + 2) th stage to the mth stage of the last stage, calculating the first M split radix butterflies for each rotation factor block of each stage, and calculating only the first output of each split radix butterfly; wherein N isi＝2tD +2, …, m, frequency shift twiddle factor
    Figure FDA0002749165120000035
    Twiddle factors phi (u) and phi (3u) which are dispersed after the original twiddle factors from the d +2 th stage to the M-th stage of the last stage are calculated according to the formula (2) and the formula (3), and the value range of u is only 0,1, … and M-1; the formula for computing the split-radix butterfly in this step is expressed as:
    Figure FDA0002749165120000036
    wherein n is calculated according to formula (5), except that u is in the range of 0,1, …, and M-1.
  5. 5. The method of claim 1 for fast extraction of OFDMA distributed mode downlink comb spectrum characterized by: in step 6, carrying out recombination operation on L blocks with the length of P to obtain the required comb-shaped output; in L blocks with the length of P, reconstructing required output by using the first M data in each block, and calculating finally required comb spectrum output [ X (K), X (K +1), … and X (K + M-1) by using the following formula; x (P + K), X (P + K +1), …, X (P + K + M-1); …, respectively; x ((L-1) P + K), X ((L-1) P + K +1), …, X ((L-1) P + K + M-1) ];
    Figure FDA0002749165120000041
CN201711170826.2A 2017-11-22 2017-11-22 Method for rapidly extracting OFDMA distributed mode downlink comb spectrum Active CN107979454B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711170826.2A CN107979454B (en) 2017-11-22 2017-11-22 Method for rapidly extracting OFDMA distributed mode downlink comb spectrum

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711170826.2A CN107979454B (en) 2017-11-22 2017-11-22 Method for rapidly extracting OFDMA distributed mode downlink comb spectrum

Publications (2)

Publication Number Publication Date
CN107979454A CN107979454A (en) 2018-05-01
CN107979454B true CN107979454B (en) 2021-01-19

Family

ID=62010789

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711170826.2A Active CN107979454B (en) 2017-11-22 2017-11-22 Method for rapidly extracting OFDMA distributed mode downlink comb spectrum

Country Status (1)

Country Link
CN (1) CN107979454B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1766798A4 (en) * 2004-06-30 2008-11-12 Motorola Inc Frequency-hopped ifdma communication system
CN102036135A (en) * 2011-01-25 2011-04-27 上海交通大学 Orthogonal frequency division multiple access (OFDMA) passive optical network transmission system
CN103402147A (en) * 2013-07-30 2013-11-20 电子科技大学 Orthogonal frequency division multiple access (OFDMA) metropolitan area access fusion network compatible with optical virtual private network (OVPN) communication, and communication method thereof
CN104780128A (en) * 2015-04-14 2015-07-15 哈尔滨工程大学 Sparse channel estimation and pilot optimization method for underwater sound OFDMA uplink communication
US9294218B2 (en) * 2004-07-16 2016-03-22 Qualcomm Incorporated Rate prediction in fractional reuse systems
CN106528495A (en) * 2016-11-23 2017-03-22 华南理工大学 Method for obtaining output value of FFT in any continuous interval

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1766798A4 (en) * 2004-06-30 2008-11-12 Motorola Inc Frequency-hopped ifdma communication system
US9294218B2 (en) * 2004-07-16 2016-03-22 Qualcomm Incorporated Rate prediction in fractional reuse systems
CN102036135A (en) * 2011-01-25 2011-04-27 上海交通大学 Orthogonal frequency division multiple access (OFDMA) passive optical network transmission system
CN103402147A (en) * 2013-07-30 2013-11-20 电子科技大学 Orthogonal frequency division multiple access (OFDMA) metropolitan area access fusion network compatible with optical virtual private network (OVPN) communication, and communication method thereof
CN104780128A (en) * 2015-04-14 2015-07-15 哈尔滨工程大学 Sparse channel estimation and pilot optimization method for underwater sound OFDMA uplink communication
CN106528495A (en) * 2016-11-23 2017-03-22 华南理工大学 Method for obtaining output value of FFT in any continuous interval

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"a general resource allocation algorithm with fairness for for SDMA/MISO/OFDMA systems";weishan lu,fei ji,hua yu;《IEEE》;20110908;全文 *
"lagrange multiplier based fair resource allocation for mimo-ofdma system";dou ni,fangjiong chen,cunmei ji;《IEEE》;20090505;全文 *

Also Published As

Publication number Publication date
CN107979454A (en) 2018-05-01

Similar Documents

Publication Publication Date Title
He et al. Designing pipeline FFT processor for OFDM (de) modulation
EP1554831B1 (en) Frequency hopping ofdma method using symbols of comb pattern
KR101984151B1 (en) Fbmc signal transmitting method, receiving method, transmitter and receiver
KR20020096833A (en) Apparatus for minimizing peak to average power ratio in orthogonal frequency division multiplexing system and method thereof
CN1507727A (en) Channel estimation for orthogonal frequency division multiplexed (OFDM) signals
KR20110014695A (en) Enhanced frequency division multiple access for wireless communication
Antonov et al. Reduced complexity tone reservation peak-to-average power ratio reduction algorithm for SEFDM signals
US8351525B2 (en) Orthogonal frequency division multiplexing transmitting and receiving device
Zavjalov et al. Efficiency of coherent detection algorithms nonorthogonal multifrequency signals based on modified decision diagram
WO2010116397A1 (en) Radio communication system, transmission device, reception device, and radio communication method in radio communication system
CN107979454B (en) Method for rapidly extracting OFDMA distributed mode downlink comb spectrum
Dali et al. Efficient FPGA implementation of high-throughput mixed radix multipath delay commutator FFT processor for MIMO-OFDM
CN106789820B (en) combined orthogonal transformation and non-orthogonal efficient frequency division multiplexing transmission peak-to-average ratio restraining method
US20140365547A1 (en) Mixed-radix pipelined fft processor and fft processing method using the same
CN110519006B (en) Single carrier interleaving type frequency division multiple access signal transmission method based on symbol shortening
Adiono et al. Low latency parallel-pipelined configurable FFT-IFFT 128/256/512/1024/2048 for LTE
Dinh et al. An area-efficient multimode FFT circuit for IEEE 802.11 ax WLAN devices
Orozco-Galvan et al. Reconfigurable architecture based on FPGA for OFDM transmitter
Jang et al. Study on the latency efficient IFFT design method for low latency communication systems
KR20080013315A (en) Fast fourier transform processor, and othogonal frequency division multiplexing receiver having the same
CN111291315A (en) Data processing method, device and equipment
Ajami et al. KLT-based PAPR reduction technique for LTE-Advanced uplink with carrier aggregation
Jang et al. Low latency IFFT design for OFDM systems supporting full-duplex FDD
Maršálek On the reduced complexity interleaving method for OFDM PAPR reduction
TWI792954B (en) Communication device and method of handling papr

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant