KR20200126663A - SVD-based F-OFDM System and Transmitting/Receiving Method Thereof - Google Patents

Abstract

The present invention can provide a filtered-orthogonal frequency division multiplexing (F-OFDM) system and a transmission/reception method thereof, which can perform pre- and post-processing at the time of transmission and reception of the F-OFDM system according to a singular value decomposition (SVD) method according to a pre-processing matrix and a post-processing matrix, thereby greatly reducing computational complexity by reducing the filter length of the F-OFDM system requiring high computational complexity, setting the filter lengths of the transmitter and receiver to be different from each other, and offering the low out-of-band emission, low bit error rate, and low peak-to-average power ratio.

Description

Singular value decomposition-based filter orthogonal frequency division multiplexing system and its transmission/reception method {SVD-based F-OFDM System and Transmitting/Receiving Method Thereof}

The present invention relates to a filter orthogonal frequency division multiplexing system and a transmission/reception method thereof, and to a filter orthogonal frequency division multiplexing system based on singular value decomposition and a transmission/reception method thereof.

In the current wireless communication system, a cyclic prefix-orthogonal frequency division multiplexing (CP-OFDM) is used. However, in recent years, as a cyclic prefix (hereinafter referred to as CP) is added, various techniques have been studied to compensate for the shortcomings occurring in CP-OFDM.

As one of them, a filter having a very large frequency attenuation characteristic is added to the CP-OFDM, so that the out-of-band (OOB) emission is much lower than that of the CP-OFDM. Orthogonal frequency division multiplexing (Filtered- There is an Orthogonal Frequency Division Multiplexing (F-OFDM) scheme. Since F-OFDM has a much lower OOB emission than CP-OFDM, even when signals are asynchronously received in adjacent subbands, there is an advantage in that there is little interference between subbands. In other words, F-OFDM allows a combination of a synchronous/asynchronous system in which interference management between adjacent subbands is a major issue.

However, F-OFDM has a limitation in that it is difficult to apply to a wireless communication system due to problems such as high computational complexity of an added filter and high peak to average power ratio (PAPR).

To compensate for the shortcomings of F-OFDM, previously proposed Resource Block Filtered-Orthogonal Frequency Division Multiplexing (RB-F-OFDM), Fast Convolution Filtered Fast Convolution Filtered -Orthogonal Frequency Division Multiplexing: FC-F-OFDM), etc. have been proposed.

However, although the RB-F-OFDM divides the bandwidth into a plurality of resource blocks and uses small IFFT, FFT, and filters to reduce complexity, there is a problem in that error performance is degraded due to a short symbol length. In addition, FC-F-OFDM reduces complexity by converting the filtering window in the time domain into the filtering window in the frequency domain, but there is a problem that backward compatibility cannot be provided due to the overlap structure of the added IFFT and FFT modules.

Korean Patent Registration No. 10-0996535 (registered on November 18, 2010)

An object of the present invention is to provide a filter orthogonal frequency division multiplexing system having a low computational complexity through singular value decomposition and a transmission/reception method thereof.

Another object of the present invention is to provide a filter orthogonal frequency division multiplexing system in which a filter length is not restricted, and a transmission/reception method thereof.

Another object of the present invention is to provide a filter orthogonal frequency division multiplexing system capable of having filters of different lengths between a transmitter and a receiver, and a transmission/reception method thereof.

In order to achieve the above object, a transmitter of a filter orthogonal frequency division multiplexing system according to an embodiment of the present invention includes: a preprocessor for receiving transmission data and preprocessing according to a preset preprocessing matrix to obtain preprocessed transmission data; And transmitting an SF-OFDM symbol obtained by obtaining a CP-OFDM symbol from the preprocessed transmission data in a known manner and filtering the CP-OFDM symbol using a transmission filter having a predetermined filter length (L _F ). A transmitting unit; Including, the pre-processing unit in the F-OFDM system, corresponding to the pre-processing unit and the pre-processing unit, except for the post-processing unit included in the receiver, a system matrix (C _k ) corresponding to all processing performed by a singular value The transmission data may be decomposed into a plurality of matrices according to a decomposition (hereinafter, referred to as SVD) technique, and the transmission data may be preprocessed according to a preprocessing matrix obtained using at least one of the decomposed plurality of matrices.

로 획득될 수 있다.The preprocessing matrix decomposes the system matrix (C _k ) into first and second singular matrix vectors (U _k , V ^H _k ) and diagonal matrix (Λ _k ), which are complex identity matrices, and the decomposed second singular matrix vector by using the inverse number (1 / Λ _k) of the Hermitian matrix (V _k) and diagonal matrix (Λ _k) of (V ^H _k),

Can be obtained with

The transmitting unit includes an IDFT unit that converts the preprocessed transmission data into OFDM symbols in a designated manner; A CP inserting unit that inserts a periodic prefix of a predetermined length into the OFDM symbol and serializes it to obtain the CP-OFDM symbol; And a transmission filter unit configured to obtain and transmit the SF-OFDM symbol by filtering the CP-OFDM symbol using the transmission filter. It may include.

A receiver of a filter orthogonal frequency division multiplexing system according to another embodiment of the present invention to achieve the above object receives the SF-OFDM symbol transmitted from the transmitter and filters using a reception filter having a predetermined filter length (L _F ). A receiving unit that obtains a CP-OFDM symbol and obtains received data from the CP-OFDM symbol in a known manner; And a post-processing unit receiving the received data and performing post-processing according to a preset post-processing matrix to obtain post-processed received data. Including, the post-processing unit in the F-OFDM system corresponding to the post-processing unit and the post-processing unit, except for the pre-processing unit included in the transmitter, a system matrix (C _k ) corresponding to all processing performed by a singular value The received data may be decomposed into a plurality of matrices according to a decomposition (hereinafter SVD) technique, and the received data may be post-processed according to a post-processing matrix obtained using at least one of the decomposed plurality of matrices.

In order to achieve the above object, a transmission method of a filter orthogonal frequency division multiplexing system according to another embodiment of the present invention includes: a preprocessing step of acquiring preprocessed transmission data by receiving transmission data and preprocessing according to a preset preprocessing matrix; And transmitting an SF-OFDM symbol obtained by obtaining a CP-OFDM symbol from the preprocessed transmission data in a known manner and filtering the CP-OFDM symbol using a transmission filter having a predetermined filter length (L _F ). Transmitting step; Including, the pre-processing step is a system matrix (C _k ) obtained by analyzing and integrating all processes except for the pre-processing step and the post-processing step performed at the time of reception corresponding to the pre-processing in the F-OFDM system. The transmission data may be decomposed into a plurality of matrices according to a value decomposition (hereinafter, referred to as SVD) technique, and the transmission data may be preprocessed according to a preprocessing matrix obtained using at least one of the decomposed plurality of matrices.

The receiving method of a filter orthogonal frequency division multiplexing system according to another embodiment of the present invention to achieve the above object is to receive a SF-OFDM symbol and filter using a reception filter having a predetermined filter length (L _F ) A receiving step of obtaining an OFDM symbol and obtaining received data from the CP-OFDM symbol in a predetermined manner; And a post-processing step of obtaining post-processed received data by receiving the received data and performing post-processing according to a preset post-processing matrix. Including, the post-processing step is a system matrix (C _k ) obtained by analyzing and integrating all processes except for the post-processing step and the pre-processing step performed at the time of transmission corresponding to the post-processing in the F-OFDM system The received data may be decomposed into a plurality of matrices according to a singular value decomposition (hereinafter SVD) technique, and the received data may be post-processed according to a post-processing matrix obtained using at least one of the decomposed plurality of matrices.

Therefore, the filter orthogonal frequency division multiplexing system according to an embodiment of the present invention and its transmission/reception method can greatly reduce computational complexity by reducing the filter length of the filter orthogonal frequency division multiplexing system that requires high computational complexity. Make it possible to set the filter lengths differently from each other. It can also provide low out-of-band emission, low bit error rate and low peak-to-average power ratio.

1 shows a schematic structure of an F-OFDM system.
2 and 3 show filter response characteristics in the time and frequency domains according to the filter length.
4 and 5 show schematic structures of a transmitter and a receiver of an SF-OFDM system according to an embodiment of the present invention.
6 and 7 illustrate a method of transmitting and receiving an SF-OFDM system according to an embodiment of the present invention.
8 to 14 show simulation results of the SF-OFDM system according to the present embodiment.

In order to fully understand the present invention, operational advantages of the present invention, and objects achieved by the implementation of the present invention, reference should be made to the accompanying drawings illustrating preferred embodiments of the present invention and the contents described in the accompanying drawings.

Hereinafter, the present invention will be described in detail by describing a preferred embodiment of the present invention with reference to the accompanying drawings. However, the present invention may be implemented in various different forms, and is not limited to the described embodiments. In addition, in order to clearly describe the present invention, parts irrelevant to the description are omitted, and the same reference numerals in the drawings indicate the same members.

Throughout the specification, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components unless specifically stated to the contrary. In addition, terms such as "... unit", "... group", "module", and "block" described in the specification mean units that process at least one function or operation, which is hardware, software, or hardware. And software.

In the present specification, a ^* denotes a complex conjugate of the scalar a, A _{a ×b} denotes a matrix of dimension a×b, and [A] _ij denotes an element of row i and column j of matrix A. And A ^T denotes a transpose matrix in which the elements in row i and j are [A] _ji , and A ^H is Hermit, which is a complex conjugate matrix in which elements in row i and j are [A] ^* _ji . Represents a Hermitian matrix. In addition, I _a represents the identity matrix of a × a dimension (identity matrix), 0 _{a × b} denotes a dimension of a × b matrix 0 (zero matrix).

Hereinafter, prior to describing the SVD-based F-OFDM system of the present embodiment, the F-OFDM system will be described first.

1 shows a schematic structure of an F-OFDM system.

Referring to FIG. 1, the F-OFDM system includes a transmitter 100 and a receiver 200. In addition, the transmitter 100 includes an IDFT unit 110, a CP insertion unit 120, and a transmission filter unit 130, and the receiver 200 includes a reception filter unit 210, a CP removal unit 220, and a DFT unit. It may include 230.

Here, assume an F-OFDM system in which each of the K subbands has M subcarriers and has a total of N (N≦KM) subcarriers. And the modulated transmission data (s _i,k ) of the k-th subband in the i-th symbol interval is s _i,k = [s _i,k (0),s _i,k (1), ..., s _{i ,k} (M-1)] ^T can be expressed as a vector.

The IDFT unit 110 outputs an OFDM symbol by performing an inverse discrete Fourier transform on the transmission data s _{i, k} to be transmitted for each subband. The CP insertion unit 120 receives and serializes the OFDM symbol, inserts the CP, and outputs a CP-OFDM symbol. The transmission filter unit 130 transmits an F-OFDM symbol (x _i,k ) by filtering the CP-OFDM symbol. Here, the transmission filter unit 130 is a transmission filter matrix (F) that is implemented by multiplying a sync filter for the k-th subband and a window function having a length L _w having a non-zero sample. By filtering the CP-OFDM symbol by _k ), an F-OFDM symbol (x _i,k ) can be obtained.

The F-OFDM symbol (x _i,k ) transmitted in the k-th subband in the i-th symbol period in the transmitter 100 may be expressed by Equation 1.

인 k번째 서브 밴드에 대한 IDFT 행렬로서, IDFT부(110)의 변환 행렬이다. 그리고 R_t는 CP 삽입부(120)가 OFDM 심볼에 지정된 길이(L_CP)의 CP를 추가한 행렬로서,

이다. F_k는 송신 필터부(130)의 송신 필터 행렬로서 [f_k(0), 0_1×(N+LCP-1)]^T 의 제1 열과 [f_k(0), 0_1×(N+LCP-1)]의 제1 행을 갖는 퇴플리츠 행렬(Toeplitz matrix)이며, 여기서 f_k 는 L_F 길이를 갖는 k 번째 서브 밴드에 대한 송신 필터의 임펄스 응답으로 f_k = [f_k(0), f_k(1), ..., f_k(L_F-1)]이다.Where W ^H _K is the element in row a and column b ([W ^H _K ] _ab )

An IDFT matrix for the k-th subband, which is a transform matrix of the IDFT unit 110. And R _t is a matrix in which the CP inserting unit 120 adds a CP having a length (L _CP ) designated to an OFDM symbol,

to be. F _k is the transmission filter matrix of the transmission filter unit 130 and the first column of [f _k (0), 0 _1×(N+LCP-1) ] ^T and [f _k (0), 0 _{1×(N+) LCP-1)} ] is a Toeplitz matrix with the first row, where f _k is the impulse response of the transmission filter for the k-th subband of length L _F , where f _k = [f _k (0) , f _k (1), ..., f _k (L _F -1)].

In Equation 1, A _k is a transmission processing matrix for the entire process for transmitting the F-OFMD symbol corresponding to the transmission data (s _i,k ) in the k-th subband _, as A _k = F _k R _t W ^H _k Is defined.

And after overlapping the filter tail, the F-OFDM symbol (x _i,k ) is transmitted in the kth subband.

Meanwhile, in the receiver 200, the reception filter unit 210 filters the received F-OFDM symbols in order to restore data (s _i,k ) of the k-th subband. The reception filter unit 210 filters the received F-OFDM symbols with a reception filter matrix G _k configured to correspond to the transmission filter matrix F _k of the transmission filter unit 130. The impulse response of the receiving filter (g _k ) for the k-th subband is L _{G with} length g _k = [g _k (0), g _k (1), ..., g _k (L _G -1)] Can be assumed.

The CP removal unit 220 removes the filter tail together with the CP from the CP-OFDM symbol filtered and output by the reception filter unit 210, and parallelizes the CP-OFDM symbol to output an OFDM symbol. The DFT unit 230 obtains received data y by performing discrete Fourier transform of N samples from the OFDM symbol output from the CP removing unit 220. Since the received signal y includes inter symbol interference (hereinafter referred to as ISI) and noise, the received data y _{i, k} of the k-th subband in the i-th symbol period is expressed by Equation 2.

는 각각 k번째 서브밴드에서의 ISI 성분과 노이즈 성분이다.Here, y _DES,k is the requested data to be acquired, y _ISI,k and

Is an ISI component and a noise component in the k-th subband, respectively.

In Equation 2, the requested data (y _DES,k ) includes transmission data (s _i,k ), and may be expressed by Equation 3.

이고, 제1 행이

인 퇴플리츠 행렬이며, h_k 는 최대 지역 확산 길이(L_CH)를 갖는 k 번째 서브 밴드에 다중 경로 채널의 임펄스 응답으로 h_k = [h_k(0), h_k(1), ..., h_k(L_CH-1)]이다. 그리고 G_k 는 송신 필터부(210)의 수신 필터 행렬로서 제1 열이

이고, 제1 행이

인 퇴플리츠 행렬이다. 한편, R_r은 CP 제거부(220)가 CP와 필터 꼬리를 제거하기 위한 행렬로

이다.In Equation 3, H _k is a channel matrix and the first column is

And the first row is

In Toeplitz matrix, h _k is the impulse response of the multipath channel in the kth subband with the maximum local spreading length (L _CH ), h _k = [h _k (0), h _k (1), ... , h _k (L _CH -1)]. In addition, G _k is the reception filter matrix of the transmission filter unit 210 and the first column is

And the first row is

It is the Intouplitz procession. Meanwhile, R _r is a matrix for the CP removal unit 220 to remove CP and the filter tail.

to be.

W _k is a DFT matrix for the k-th subband, and is a transform matrix of the DFT unit 230.

In Equation 3, B _k is a reception processing matrix for the entire process for obtaining requested data (y _DES,k ) from the F-OFDM signal received in the k-th subband _, and is defined as B _k = W _k R _r G _k do.

Meanwhile, the multiplication of the channel matrix (H _k ) in Equation 3 is a channel coefficient matrix (diag{h _m }) that is the product of the diagonal matrix and the frequency response of the channel defined as h _m (m ∈ [0, M-1]). It can be approximated, and thus Equation 3 can be expressed again as Equation 4 with reference to Equation 1.

According to Equation 4, C _k = B _k A _k may be defined as a system matrix for the entire signal processing process for the k-th subband in the F-OFDM system. In addition, if the system matrix C _k is an identity matrix, the received signal can be restored without distortion. However, since F-OFDM generally uses a finite impulse response (FIR), the system matrix C _k cannot be an exact identity matrix. However, the system matrix C _k has a diagonal element of equal power and a non-diagonal element of power similar to zero. This means that channel interference (ICI) can be received with almost no inter-channel interference (ICI) with equal power within the subband.

Meanwhile, the ISI component (y _ISI,k ) in Equation 2 is the intersymbol interference caused by the delay spread of the multipath channel and the filter tail of the (i-1)-th and (i + 1)-th received F-OFDM symbols. As a component, it may be expressed as Equation 5 according to Equation 1 and Equations 3 and 4.

는 이전 전송된 심볼(x_i-1,k)의 마지막 (L_F-1)개의 샘플이고,

는 다음 전송되는 심볼(x_i+1,k)의 첫 번째 (L_F-1)개의 샘플로서 각각 수학식 6으로 표현된다.In Equation 5

Is the last (L _F -1) samples of the previously transmitted symbol (x _i-1,k ),

Is the first (L _F -1) samples of the next transmitted symbol (x _i+1,k ) and is represented by Equation 6, respectively.

In addition, considering the maximum delay spread of the multipath channel and the length L _G of the reception filter g _k , the ISI can be rewritten as Equation 7.

와

는 각각 수신 필터(G_k)가 필터링하고, CP와 필터 꼬리를 제거한 이후 시간 도메인에서의 나머지 이전 심볼(x_i-1,k) 및 다음 심볼(x_i+1,k)을 의미한다.here

Wow

Denotes the remaining previous symbols (x _i-1,k ) and next symbols (x _i+1,k ) in the time domain after the reception filter (G _k ) filters and removes the CP and the filter tail, respectively.

In general, in F-OFDM, the filter length (L _F , L _G ) of each of the transmit filter (f _k ) and the receive filter (g _k ) is equal to N/2+1 (OFDM) for the total number of subcarriers (N). 1/2 of the symbol length), so each of the ISI caused by the previous symbol (x _i-1,k ) and the next symbol (x _i+1,k ) is L _p and non-zero according to Equation 8 It has the number of L _n samples, that is, the length.

Referring to Equation 8, the length of the ISI (L _p , L _n ) by each of the previous symbol (x _i-1,k ) and the next symbol (x _i+1,k ) is a transmission filter (f _k ) and a reception It depends on the length (L _F , L _G ) of the filter (f _k ), but the length of the ISI (L _p , L _n ) also does not change because the filter length (L _F , L _G ) is fixed.

In the F-OFDM system described above, filtering is performed at both sides of the transmitter 100 and the receiver 200, and filtering is performed by a convolution operation or the like, and thus has a high computational complexity. Based on the number of complex multiplications (CM) that require a high amount of computation, the computational complexity of the transmitter 100 (CM _TX ) and the computational complexity of the receiver 200 (CM _RX ) can be calculated as shown in Equation 9. I can.

According to Equation 9, the computational complexity of the transmitter 100 in the F-OFDM system depends on the length L _F of the transmission filter F _k , but the computational complexity of the receiver 200 is of the reception filter G _k Not only the length (L _G ), but also the length (L _F ) of the transmission filter (F _k ) is affected.

As described above, in general, in the F-OFDM system, since the filter lengths (L _F , L _G ) of the transmit filter (F _k ) and the receive filter (G _k ) are equally limited to N/2+1, a long filter The amount of computation increases according to the length (L _F , L _G ), resulting in high complexity, which is an obstacle in applying the F-OFDM method to a wireless communication system.

Therefore, if the filter lengths L _F and L _{G of the} transmission filter F _k and the reception filter G _k can be reduced, the computational complexity can be greatly reduced. In addition, if the filter length (L _F , L _G ) of each of the transmission filter (F _k ) and the reception filter (G _k ) can be set differently, the size or manufacturing of the transmitter 100 or the receiver 200 is designed and implemented. By overcoming limitations due to cost, etc., the F-OMDM method can be easily applied to a wireless communication system.

However, if the lengths (L _F , L _G ) of the filters (f _k , g _k ) become shorter than the generally designated length (N/2+1), distortion may occur in the received data.

2 and 3 show filter response characteristics in the time and frequency domains according to the filter length.

In FIGS. 2 and 3, the total number of subcarriers (N) is 1024 (N=1024), and three resource blocks (RBs) using 36 (M = 36) subcarriers each are shown. I did.

2 and 3 are sync impulse responses of a transmit filter (F _k ) and a receive filter (G _k ) implemented by multiplying a sync filter with a window function of a length L _w having non-zero samples. It appears as a square in the frequency domain.

When the length (L _F , L _G ) of the filter (f _k , g _k ) becomes short, that is, when the length (L _w ) with non-zero samples in the window function is shortened, the filter as shown in FIG. The tail becomes shorter in the filter tail in the time domain. As shown in FIG. 3, this causes more changes because it does not have a square-shaped response in the frequency domain.

Therefore, in the case of using the short-length (L _F , L _G ) filters (f _k , g _k ), the system matrix (C _k ) in Equation 4 has diagonal elements having unequal power. Therefore, the signal-to-noise ratio (hereinafter, SNR) of the received signal is not constant in the sub-carriers in the sub-band, thereby deteriorating the error performance.

On the other hand the transmitting filter (F _k) and the reception filter (G _k), each of the filter length (L _F, L _G) is different from (L _F ≠ L _G) if a transmission filter (F _k) and the reception filter (G _k) with each other When the filter lengths (L _F , L _G ) of are the same, the SNR maximization effect generated by matched filtering cannot be used, and as in the case where the filter lengths (L _F , L _G ) are shortened, the system matrix (C _k ) This uneven power leads to a problem of deteriorating error performance.

Therefore, in this embodiment, the filter lengths (L _F , L _G ) of each of the transmit filter (F _k ) and the receive filter (G _k ) are less than N/2+1 compared to the total number of carriers (N) (L _F <N/2 + 1, L _G <N/2 + 1) or when the filter lengths (L _F , L _G ) are different (L _F ≠ L _G ), it is possible to solve the data distortion based on SVD. -It is possible to greatly reduce the computational complexity of the OFDM system. Hereinafter, the Singular Value Decomposition (SVD) based F-OFDM system according to the present embodiment is referred to as an SF-OFDM system.

4 and 5 show schematic structures of a transmitter and a receiver of an SF-OFDM system according to an embodiment of the present invention.

4 and 5, the SF-OFDM system according to the present embodiment includes a transmitter and a receiver. The transmitter may include a preprocessor 300 and a transmitter 400. In addition, the transmission unit 400 includes an IDFT unit 410, a CP insertion unit 420, and a transmission filter unit 430, and the receiver may include a receiving unit 500 and a post-processing unit 600, and the receiving unit 500 ) May include a receiving filter unit 510, a CP removing unit 520, and a DFT unit 530.

That is, the SF-OFDM system according to the present embodiment illustrated in FIG. 4 further includes a preprocessor 300 in the transmitter and a post-processor 600 in the receiver, compared to the SF-OFDM system illustrated in FIG. 1. .

Here, the pre-processing unit 300 and the post-processing unit 600 are the transmission filter (F _k ) and the reception filter (G _k ) of the transmission filter unit 430 and the reception filter unit 510, respectively, the filter length (L _F , L _G) ) Is less than 1/2 of the OFDM symbol length (L _F <N/2 + 1, L _G <N/2 + 1) or the filter lengths (L _F , L _G ) are different (L _F ≠ L _G ) Also, it is to compensate for data distortion through pre- and post-processing.

In the SF-OFDM system according to the present embodiment, the pre-processing unit 300 and the post-processing unit 600 include the system matrix C _k and the filter lengths L _F and L of the transmission filter F _k and the reception filter G _k . _By making the identity matrix independent of _G ), the problem of data distortion caused by the filter lengths (L _F and L _G ) is solved.

In FIG. 4, the IDFT unit 410 of the transmission unit outputs OFDM symbols by performing inverse discrete Fourier transform on the transmission data _{s'i, k} preprocessed by the preprocessing unit 300 for each subband. In addition, each of the CP insertion unit 420 and the transmission filter unit 430 inserts a CP into the OFDM symbol as in the CP insertion unit 120 and the transmission filter unit 130 of FIG. 1 (a) to generate a CP-OFDM symbol. It outputs, filters the CP-OFDM symbol, and transmits the SF-OFDM symbol (x' _i,k ).

Meanwhile, in FIG. 5, the receiving filter unit 510, the CP removing unit 520 and the DFT unit 530 of the receiving unit are the receiving filter unit 210, the CP removing unit 220 and the DFT unit of Fig. 1(b). Like (230), the received SF-OFDM symbol is filtered, the received SF-OFDM symbol is filtered with the reception filter matrix (G _k ), and the filter tail along with the CP is removed from the filtered output CP-OFDM symbol. , The OFDM symbol is obtained by paralleling, and received data (y _i,k ) is obtained by performing discrete Fourier transform of N samples in the OFDM symbol.

Here, the filter length (L _G) of the receive filter (G _k) of the receive filter 510 is the filter length (L _F) of the transmission filter (F _k) of the transmitting filter 430, as shown in Figure 5 It can be different from (L _F ≠ L _G ), and the filter length (L _G ) of the transmit filter (F _k ) and the receive filter (G _k ) is less than 1/2 of the OFDM symbol length (L _F <N/2 + 1, L _G <N/2 + 1).

After the processing unit 600 outputs the received data (y _{i, k)} is performed a process to compensate the received data (y _{'i, k)} after about.

In this embodiment, the preprocessor 300 and the postprocessor 600 use the SVD technique mainly used to provide a uniform subchannel gain in a multiple-input multiple-output (MIMO) system, and the filter length (L _F , L _G ) to compensate for data distortion.

Accordingly, the system matrix (C _k ) of Equation 4 is used as a diagonal element using complex unitary matrices (U _k , V ^H _k ) and non-negative real numbers (λ _m ) as shown in Equation 10 according to the SVD technique. It is decomposed into a diagonal matrix (Λ _k ).

Here, U _k is a first singular matrix vector consisting of C _k C ^H _k eigenvectors, and V ^H _k is a second singular matrix vector consisting of C ^H _k C _k eigenvectors.

A pre-processing unit 300 El of the transmitted data (s _{i, k)} modulated by the k-th subband diagonal matrix (Λ _k), the inverse (1 / Λ _k) and a second specific matrix vector (V ^H _k) of the contact By preprocessing as in Equation 11 using the matrix V _k , the preprocessed transmission data _{s'i, k} is output.

를 전처리부(300)가 송신 데이터(s_i,k)에 대해 전처리를 수행하는 전처리 행렬로 볼 수 있다. 이에 수학식 1을 참조하면, 송신기에서 전송되는 SF-OFDM 심볼(x'_i,k)은 수학식 12와 표현될 수 있다.That is, in Equation 11

May be regarded as a preprocessing matrix in which the preprocessor 300 performs preprocessing on the transmission data (s _i,k ). Accordingly, referring to Equation 1, the SF-OFDM symbol (x′ _{i, k} ) transmitted from the transmitter may be expressed as Equation 12.

)의 곱으로 정의되는 송신 처리를 수행하여, SF-OFDM 심볼(x'_i,k)을 획득할 수 있다.According to Equation 12, a transmitter transmission processing matrix (A _k) and the pre-matrix for the data transmission (s _{i, k)} (

By performing transmission processing defined as the product of ), an SF-OFDM symbol (x' _i,k ) can be obtained.

Meanwhile, in the receiver, the post-processing unit 600 post-processes the received data (y _i,k ) using the Hermit matrix (U ^H _k ) of the first singular matrix vector (U _k ) as shown in Equation 13, The post-processed received data (y' _i,k ) is output.

는 후처리 이후의 수신 노이즈로

이다.here

Is the received noise after post processing

to be.

In Equation 13, U ^H _k may be viewed as a post-processing matrix in which the post-processing unit 600 performs post-processing on the received data (y _i,k ).

And assigns the transmission data (s _{i, k)} and F-OFDM symbols (x _{i, k)} pretreated transmitted instead of the data (s _{'i, k)} and SF-OFDM symbols (x' _{i, k)} in equation (4) , Referring to Equation 11, the requested data ( _y'DES,k ) compensated in Equation 13 may be expressed again as Equation 14.

That is, the compensated request data (y' _DES,k ) is recovered in the same form as the transmission data (s _i,k ) multiplied by the channel coefficient (h _m ). This means that data can be recovered without distortion. This is because the pre-processing of the SVD-based pre-processing unit 300 and the post-processing of the post-processing unit 600 convert the entire system matrix C _k into a unit matrix.

As a result, the filter length (L _F , L _G ) of the transmit filter (F _k ) and the receive filter (G _k ) is less than 1/2 of the OFDM symbol length (L _F <N/2 + 1, L _G <N/2 It can be seen that transmission data (s _{i, k} ) can be recovered without data distortion through pre-processing and post-processing even if they are + 1) or are different from each other (L _F ≠ L _G ).

Meanwhile, referring to Equations 5 and 7, the compensated ISI component (y' _ISI ) by the previously transmitted symbol (x' _i-1,k ) and the next transmitted symbol (x' _i+1,k ) _,k ) can be expressed as Equation 15.

와

는 이전 전송된 심볼(x'_i-1,k)의 마지막 (L_F-1)개의 샘플과 다음 전송되는 심볼(x'_i+1,k)의 첫 번째 (L_F-1)개의 샘플이다.here

Wow

Was previously transmitted symbol _{(x 'i-1, k} ) the last _(F L -1) samples and then a transmission symbol (x is a "is the first one second _(F L -1) of a sample _{i + 1, k)} .

The filter length (L _F , L _G ) of the transmit filter (F _k ) and the receive filter (G _k ) is less than 1/2 of the OFDM symbol length (L _F <N/2 + 1, L _G <N/2 + 1 ), or may be different from each other (L _F ≠ L _G ) _, ISI caused by the previous symbol (x' _i-1,k ) and the next symbol ('x _i+1,k ) with reference to Equation 8 Each has a non-zero number of L' _p and L' _n samples, that is, a length according to Equation 16.

Referring to Equation 16, the compensated ISI component (y' _ISI,k ) also considers the length of the filter length (L _F , L _G ) of the transmission filter (F _k ) and the reception filter (G _k ), but the filter length If (L _F , L _G ) is shortened, the number of samples (L' _p , L' _n ) of the ISI component decreases, so that ISI can be rather reduced. In addition, ISI can be reduced by adjusting the filter lengths L _F and L _G differently according to the performance of the transmitter and the receiver.

That is, in the SF-OFDM system according to the present embodiment, the system matrix (C _k ) of the F-OFDM system is combined with two complex identity matrices (U _k , V ^H _k ) and a diagonal matrix (Λ _k ) according to the SVD technique. Decomposition (C _k = U _k Λ _k V ^H _k ), and the Hermitian matrix (V _k ) of the second singular matrix vector (V ^H _k ) of the two decomposed complex identity matrices (U _k , V ^H _k ) Using a diagonal matrix (Λ _k ), the transmitter preprocesses the transmission data (s _i,k ) according to Equation 11, and the receiver _calculates the Hermitian matrix (U ^H _k ) of the first singular matrix vector (U _k ). By post-processing the received data (y _i,k ) according to Equation 13 using, regardless of the filter length (L _F , L _G ) of the transmission filter (F _k ) and the reception filter (G _k ), without data distortion Make it possible to transmit data.

According to Equation 9, the computational complexity of the transmitter 100 in the F-OFDM system depends on the length L _F of the transmission filter F _k , but the computational complexity of the receiver 200 is of the reception filter G _k Not only the length (L _G ), but also the length (L _F ) of the transmission filter (F _k ) is affected.

6 and 7 illustrate a method of transmitting and receiving an SF-OFDM system according to an embodiment of the present invention.

Referring to FIGS. 4 and 5, the transmission method of the SF-OFDM system of FIG. 6 will be described. First, filter lengths L _F and L _{G of the} transmission filter F _k and the reception filter G _k are set. Do (S11). Here, the filter length (L _F , L _G ) of each of the transmit filter (F _k ) and the receive filter (G _k ) is less than 1/2 of the OFDM symbol length (L _F <N/2 + 1, L _G <N/2 + 1) or the filter lengths (L _F , L _G ) can be set to be different from each other (L _F ≠ L _G ).

And when the filter lengths (L _F , L _G ) of the transmission filter (F _k ) and the reception filter (G _k ) are set, the F-OFMD system according to the set filter length (L _F , L _G ) is analyzed and transmitted. The system matrix C _k is obtained from the matrix A _k , the reception processing matrix B _k , and the transmission processing matrix A _k and the reception processing matrix B _k (S12).

)을 획득한다(S14).When the system matrix (C _k ) is obtained, the obtained system matrix is converted to the first and second singular matrix vectors (U _k , V ^H _k ) and non-negative real numbers (λ _m ) as shown in Equation 10 according to the SVD technique. It is decomposed into a diagonal matrix (Λ _k ) having a diagonal element (S13). When the system matrix (C _k ) is decomposed into first and second singular matrix vectors (U _k , V ^H _k ) and diagonal matrix (Λ _k ) according to the SVD technique, the er of the second singular matrix vector (V ^H _k ) Using the meat matrix (V _k ) and the reciprocal of the diagonal matrix (Λ _k ) (1/Λ _k ), the preprocessing matrix (

) Is obtained (S14).

)을 획득하는 단계(S14)까지는 송신기를 설정하는 단계로서, 송신기는 전처리 행렬(

)이 획득되면, 이후 SF-OFDM 방식으로 신호를 전송할 수 있다. 즉 송신기를 설정하는 단계는 송신기의 구성 시에 적용되며, 이후 송신 데이터를 송신하는 경우에는 반복적으로 적용될 필요가 없다.Here, from the step (S11) of setting the filter length (L _F , L _G ), the preprocessing matrix (

) Until the step (S14) of acquiring a transmitter is a step of setting the transmitter, and the transmitter is a preprocessing matrix (

) Is obtained, then the signal can be transmitted in the SF-OFDM scheme. That is, the step of setting the transmitter is applied when configuring the transmitter, and does not need to be repeatedly applied when transmitting data afterwards.

)에 따라 수학식 11과 같이 k번째 서브밴드에서 변조된 송신 데이터(s_i,k)를 전처리하여 전처리된 송신 데이터(s'_i,k)를 획득한다(S15).After the step of setting the transmitter, in the step of transmitting the transmission data, the obtained preprocessing matrix (

) According to Equation 11, the transmitted data s _i,k modulated in the k-th subband is pre-processed to obtain pre-processed transmission data _{s'i, k} (S15).

Then, the preprocessed transmission data _{s'i, k} is converted into OFDM symbols (S16). In this case, as an example, the transmitter may perform IDFT on the preprocessed transmission data (s' _{i, k} ) and convert it into OFDM symbols. Thereafter, a CP is inserted into an OFDM symbol and serialized to obtain a CP-OFDM symbol (S17). When the CP-OFDM symbol is obtained, the SF-OFDM symbol (x′ _i,k ) is obtained by filtering the CP-OFDM symbol using a transmission filter (F _k ) having a predetermined filter length (L _F ) ( S18). Then, the obtained SF-OFDM symbol (x' _i,k ) is transmitted through at least one antenna (S19).

On the other hand, looking at the reception method of the SF-OFDM system of FIG. 7, similar to FIG. 6, first, filter lengths L _F and L _{G of} each of the transmission filter F _k and the reception filter G _k are set ( S21). And by analyzing the F-OFMD system according to the set filter length (L _F , L _G ), the transmission processing matrix (A _k ), the reception processing matrix (B _k ), the transmission processing matrix (A _k ), and the reception processing matrix (B _A system matrix (C _k ) is obtained from _k ). In this case, a channel coefficient matrix (diag{h _m }) according to the frequency response of the channel may be obtained together.

When the system matrix (C _k ) is obtained, the obtained system matrix is converted to the first and second singular matrix vectors (U _k , V ^H _k ) and non-negative real numbers (λ _m ) as shown in Equation 10 according to the SVD technique. It is decomposed into a diagonal matrix (Λ _k ) having a diagonal element (S23). Then, the Hermitian matrix U ^H _{k of} the first singular matrix vector U _k is obtained as a post-processing matrix (S24).

Here, from the step of setting the filter length (L _F , L _G ) (S21) to the step of acquiring the post-processing matrix (U ^H _k ) (S34), the step of setting a receiver corresponding to the step of setting the transmitter of FIG. When the post-processing matrix (U ^H _k ) is obtained, the receiver may recover data by receiving the SF-OFDM symbol (x′ _{i, k} ) transmitted in the SF-OFDM scheme. The step of configuring the receiver is applied when configuring the receiver, and does not need to be repeatedly applied when the SF-OFDM symbol (x' _i,k ) is received thereafter.

When the receiver is configured by the step of setting the receiver, the SF-OFDM symbol (x' _i,k ) is received (S25). When the SF-OFDM symbol (x' _i,k ) is received, the SF-OFDM symbol (x' _{i, k} ) is filtered using a transmission filter (G _k ) having a predetermined filter length (L _G ), thereby CP Acquire the -OFDM symbol (S26). Then, the CP is removed from the CP-OFDM symbol and parallelized to obtain an OFDM symbol (S27). When the OFDM symbol is obtained, the OFDM symbol is converted in a known manner to obtain received data (y _i,k ) (S28). Here, the received data (y _i,k ) may be obtained by performing DFT on an OFDM symbol, for example.

The received data y _i,k is post-processed according to the obtained post-processing matrix U ^H _{k as} shown in Equation 13 to obtain post-processed received data _{y'i, k} . Here, the post-processed received data (y' _i,k ) includes compensated request data (y' _DES,k ), and compensated using the channel coefficient matrix (diag{h _m }) as shown in Equation 14. Transmission data (s _i,k ) can be recovered from the requested data (y' _DES,k ).

8 to 14 show simulation results of the SF-OFDM system according to the present embodiment.

Hereinafter, the simulation was performed by setting the total number of subcarriers (N) to be 1024 (N = 1024) and each resource block (RB) to have the number of subcarriers of 36 (M = 36). The length of the _CP (L _CP ) was designated as 72, and the transmission filter (F _k ) and the reception filter (G _k ) were generated using a Hamming window.

8 is a result of simulating the computational complexity of the SF-OFDM system according to the present embodiment. When there are 6 resource blocks RB, the filter length L of the transmission filter F _k and the reception filter G _{k The} computational complexity according to _F and L _G ) is shown.

In FIG. 8, the x-axis represents the filter length (L _F ) of the transmission filter (F _k ), the y-axis represents the filter length (L _G ) of the reception filter (G _k ), and the level on the right represents the overall complexity.

)을 곱하고, 수신 처리 행렬(B_k)에 후처리 행렬(U^H _k)을 곱하는 추가 단계를 가지고 있다.For the pre-treatment and post-treatment SF-OFDM system, compared to F-OFDM in accordance with the present embodiment pre-processing matrices to the transmission processing matrix (A _k) (

) And multiplying the received processing matrix (B _k ) by the post processing matrix (U ^H _k ).

Accordingly, in the SF-OFDM system according to the present embodiment, the computational complexity of the transmitter (CM _{SF, TX} ) and the computational complexity of the receiver (CM _{SF, RX} ) are each of the subcarriers for each subband based on the number of complex multiplications (CM) When the number is M, it is calculated as in Equation 17.

Referring to FIG. 8 simulated according to Equation 17, SF-OFDM as at least one of the filter length (L _F ) of the transmission filter (F _k ) or the filter length (L _G ) of the reception filter (G _k ) decreases. The computational complexity (CM) of the system is reduced. If the SF-OFDM system of this embodiment uses the filter length of 513 (L _F = L _G = N/2 + 1 = 513) in the same way as the F-OFDM system, the computational complexity of the F-OFDM system is 1.397e + It requires a computational complexity of 1.408e + 06 CM, which is slightly larger than 06 CM.

However, the SF-OFDM system of this embodiment may use a filter having a shorter length than that of the F-OFDM system. For example, when using a filter length of 257 (L _F = L _G = N/8 + 1 = 257), SF -OFDM system has 53.50% less complex multiplication operation (CM) than F-OFDM system. Also, when using a filter length of 33 (L _F = L _G = N/32 + 1 = 33), the SF-OFDM system has 93.27% less complex multiplication operations than the F-OFDM system.

9 and 10 are an SF-OFDM system, an F-OFDM system, and a CP-OFDM system according to a filter length (L _F ) of a transmission filter (F _k ) and a filter length (L _G ) of a reception filter (G _k ), respectively. This is a simulation result by comparing the power spectral density of (hereinafter referred to as PSD).

In FIG. 9, it is assumed that the SF-OFDM system uses L _G = N/2+1 as in the F-OFDM system in which the filter length L _G of the reception filter G _k is used. The SF-OFDM system has almost the same low OOB emission when the filter length (L _F ) of the transmit filter (F _k ) is L _F = N/2+1 like the F-OFDM system. However, the shorter the filter length L _F of the transmission filter F _k , the higher OOB emission occurs. However, it can be seen that even when the filter length (L _F ) of the transmission filter (F _k ) is very low as L _F = N/32+1, it has a much lower OOB emission than that of CP-OFDM.

In FIG. 10, it is assumed that the SF-OFDM system and the F-OFDM system have a filter length (L _F ) of a transmission filter (F _k ) of L _G = N/2+1. It can be seen that the greater the filter length L _G of the reception filter G _k , the greater the power of the edge subcarrier relative to the center subcarrier in the subband. However, it can be seen that even if the filter length L _G is changed, the power of the subcarrier outside the subband is not significantly affected. That is, it can be seen that the preprocessing affects the power of the subcarriers in the subband, but has little effect on the OOB emission.

Therefore, the SF-OFDM system of this embodiment has a low OOB emission like the F-OFDM system due to the effect of transmission filtering, and a smaller transmission filter length (L _F ) increases the OOB emission of the SF-OFDM system, but still CP -It has low OOB emission of OFDM. In addition, it can be seen that the OOB emission of the SF-OFDM system is not highly dependent on the transmit filter length (L _G ).

As a result, the SF-OFDM system also retains the advantages of the F-OFDM system, which is much less susceptible to interference from adjacent subbands than the CP-OFDM system.

11 and 12 show bit error rates of an SF-OFDM system, an F-OFDM system, and a CP-OFDM system with modulations of 16 QAM and 64 QAM, respectively, in a TDL-A channel having a very long delay spread. Rate: BER).

The BER of FIGS. 11 and 12 was evaluated by simulating a case in which 14 symbols are transmitted in each of the three subbands according to the parameters of Table 1, and the BER of the middle subband was evaluated.

In FIG. 11, filter lengths (L _F , L _G ) of the F-OFDM system as well as the SF-OFDM system according to the present embodiment are variously modified and compared. As shown in FIG. 11, in the case of the F-OFDM system, when the filter length (L _F , L _G ) is shortened, the BER performance is greatly degraded, whereas the SF-OFDM system of the present embodiment is preprocessed and after the SVD It can be seen that almost the same BER performance is maintained as processing is performed. That is, the SF-OFDM system of the present embodiment can prevent data distortion from occurring even when filters F _k and G _k of a short length or different lengths are used. In addition, signals having the same power can be received in all subcarriers in the subband.

13 shows BER in consideration of multiple users in 16QAM modulation.

Here, assuming that signals are received asynchronously in the first and third subbands, a time offset (τ) of each subband is randomly set to [0, M/4]. Asynchronously received signals cause interference to adjacent subbands, thus degrading BER performance. The BER degradation of the CP-OFDM system from the asynchronous system is greater than that of the F-OFDM system. This is because the high OOB emission of the CP-OFDM system causes more interference to adjacent subbands than the F-OFDM system. In contrast, the SF-OFDM system of this embodiment uses a transmission filter (F _k ) having a short filter length (L _F ), but it can be seen that the BER does not significantly decrease even in the case of asynchronous multi-users. This is because even if the SF-OFDM system uses a transmission filter (F _k ) of a short filter length (L _F ), OOB emission is still lower than that of CP-OFDM in FIG. 9.

In conclusion, the SF-OFDM system of this embodiment has a BER level similar to that of the CP-OFDM system and the F-OFDM system regardless of the modulation order. In particular, in the case of multiple users, the SF-OFDM system has a lower BER than the CP-OFDM system. In addition, the SF-OFDM system has the advantage of F-OFDM with low interference between adjacent subbands due to low OOB emission.

14 shows a complementary cumulative distribution function (CCDF) of a peak to average power ratio (PAPR) of an SF-OFDM system, an F-OFDM system, and a CP-OFDM system. In FIG. 14, in the SF-OFDM system, not only the filter lengths L _F and L _G are changed, but also a case where the filter lengths L _F and L _G are changed and set differently.

PAPR can be calculated by Equation 18.

Where |x _peak | is the peak amplitude of the transmitted signal, and x _rms is the root mean square (RMS) value of the power of the transmitted signal.

Referring to FIG. 14, the F-OFDM system has a high PAPR due to transmission filtering, which is because the average power (x _rms ) of the filtered signal, which is the denominator of PAPR, is lowered by having a very small transmission filter tail. However, as the transmission filter length (L _F ) decreases, the average power (x _rms ) increases and the PAPR decreases.

The SF-OFDM system has a low PAPR when using a short transmission filter length (L _F ). This is because the data distortion problem caused by the short filter has been solved. When using the transmission filter F _k whose transmission filter length L _F is L _F = N / 32 +1, the SF-OFDM of this embodiment has a PAPR similar to that of the CP-OFDM. In addition, it can be seen that the reception filter length (L _G ) affects preprocessing in transmission, but does not affect PAPR.

In addition to the high computational complexity, the main drawback of the F-OFDM system is that transmit filtering increases PAPR. However, the SF-OFDM system of this embodiment can achieve a low PAPR similar to that of the CP-OFDM system by enabling the use of a transmission filter of a short length.

In other words, the SF-OFDM system allows the use of transmission and reception filters (F _k , G _k ) of short length (L _F , L _G ) without causing data distortion, which is the main drawback of F-OFDM. Complexity and high PAPR can be solved.

The method according to the present invention may be implemented as a computer program stored in a medium for execution on a computer. Here, the computer-readable medium may be any available medium that can be accessed by a computer, and may also include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, and ROM (Read Dedicated memory), RAM (random access memory), CD (compact disk)-ROM, DVD (digital video disk)-ROM, magnetic tape, floppy disk, optical data storage device, and the like.

The present invention has been described with reference to the embodiments shown in the drawings, but these are merely exemplary, and those of ordinary skill in the art will appreciate that various modifications and other equivalent embodiments are possible therefrom.

Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

300: preprocessing unit 410: IDFT unit
420: CP insertion unit 430: transmission filter unit
510: receiving filter unit 520: CP removing unit
530: DFT unit 600: post-processing unit

Claims (12)

In the transmitter of the F-OFDM (Filtered-Orthogonal Frequency Division Multiplexing) system,
A preprocessor for receiving transmission data and preprocessing according to a preset preprocessing matrix to obtain preprocessed transmission data; And
Transmitting an SF-OFDM symbol obtained by obtaining a CP-OFDM symbol from the preprocessed transmission data in a known manner and filtering the CP-OFDM symbol using a transmission filter having a predetermined filter length (L _F ) Transmission unit; Including,
The pretreatment unit
In the F-OFDM system, a system matrix (C _k ) corresponding to the preprocessor and the preprocessor, corresponding to the preprocessor and the system matrix (C _k ) corresponding to all processing performed by the other components except for the postprocessor included in the receiver, is performed using a singular value decomposition (hereinafter, SVD) technique A transmitter of an F-OFDM system that decomposes into a plurality of matrices accordingly and preprocesses the transmission data according to a preprocessing matrix obtained using at least one of the decomposed plurality of matrices. The method of claim 1, wherein the preprocessing matrix is
The system matrix (C _k ) is decomposed into first and second singular matrix vectors (U _k , V ^H _k ) and diagonal matrix (Λ _k ), which are complex identity matrices, and the decomposed second singular matrix vector (V ^H _k) Using the Hermit matrix (V _k ) of) and the reciprocal of the diagonal matrix (Λ _k ) (1/Λ _k ),

Transmitter of the F-OFDM system obtained as a. The method of claim 1, wherein the transmission unit
An IDFT unit converting the preprocessed transmission data into OFDM symbols in a designated manner;
A CP inserting unit that inserts a periodic prefix of a predetermined length into the OFDM symbol and serializes it to obtain the CP-OFDM symbol; And
A transmission filter unit configured to obtain and transmit the SF-OFDM symbol by filtering the CP-OFDM symbol using the transmission filter; Transmitter of the F-OFDM system comprising a. In the receiver of the F-OFDM system,
The SF-OFDM symbol transmitted from the transmitter is received and filtered using a reception filter having a predetermined filter length (L _F ) to obtain a CP-OFDM symbol, and received data is obtained from the CP-OFDM symbol in a known manner. A receiving unit; And
A post-processing unit receiving the received data and performing post-processing according to a preset post-processing matrix to obtain post-processed received data; Including,
The post-processing unit
In the F-OFDM system, in response to the post-processing unit and the post-processing unit, a system matrix (C _k ) corresponding to all processing performed by the remaining components except for the pre-processing unit included in the transmitter is decomposed according to the singular value decomposition (hereinafter, SVD) method. A receiver of an F-OFDM system that decomposes into a plurality of matrices and post-processes the received data according to a post-processing matrix obtained by using at least one of the decomposed plurality of matrices. The method of claim 4, wherein the post-processing matrix is
The system matrix (C _k ) is decomposed into first and second singular matrix vectors (U _k , V ^H _k ) and diagonal matrix (Λ _k ), which are complex identity matrices, and the decomposed first singular matrix vector (U _k ) The receiver of the F-OFDM system obtained by the Hermit matrix (U ^H _k ) of. The method of claim 4, wherein the receiving unit
A reception filter unit for obtaining a CP-OFDM symbol by filtering the SF-OFDM symbol using the reception filter;
A CP removal unit that removes a periodic prefix inserted in the CP-OFDM symbol with a predetermined length and performs parallelization to obtain an OFDM symbol; And
A DFT unit converting the OFDM symbol into the received data in a specified manner; Receiver of the F-OFDM system comprising a. In the transmission method of the F-OFDM system,
A preprocessing step of receiving transmission data and preprocessing according to a preset preprocessing matrix to obtain preprocessed transmission data; And
Transmitting an SF-OFDM symbol obtained by obtaining a CP-OFDM symbol from the preprocessed transmission data in a known manner and filtering the CP-OFDM symbol using a transmission filter having a predetermined filter length (L _F ) Sending step; Including,
The pretreatment step
In the F-OFDM system, a system matrix (C _k ) obtained by analyzing and integrating all processes except for the pre-processing step and the post-processing step performed at the time of reception corresponding to the pre-processing is analyzed and integrated into a singular value decomposition (hereinafter, referred to as SVD) technique. A transmission method of an F-OFDM system in which the transmission data is decomposed into a plurality of matrices accordingly and the transmission data is pre-processed according to a preprocessing matrix obtained using at least one of the decomposed plurality of matrices. The method of claim 7, wherein the preprocessing matrix is
The system matrix (C _k ) is decomposed into first and second singular matrix vectors (U _k , V ^H _k ) and diagonal matrix (Λ _k ), which are complex identity matrices, and the decomposed second singular matrix vector (V ^H _k) ) Of the Hermitian matrix (V _k ) and the diagonal matrix (Λ _k ) inverse (1/Λ _k )

Transmission method of the F-OFDM system obtained as. The method of claim 7, wherein the transmitting step
Converting the preprocessed transmission data into OFDM symbols in a designated manner;
Inserting a periodic prefix of a predetermined length into the OFDM symbol and serializing it to obtain the CP-OFDM symbol; And
Filtering the CP-OFDM symbol using the transmission filter to obtain and transmit the SF-OFDM symbol; Transmission method of the F-OFDM system comprising a. In the receiving method of the F-OFDM system,
A receiving step of obtaining a CP-OFDM symbol by receiving an SF-OFDM symbol and filtering using a reception filter having a predetermined filter length (L _F ), and obtaining received data from the CP-OFDM symbol in a known manner; And
A post-processing step of obtaining post-processed received data by receiving the received data and performing post-processing according to a preset post-processing matrix; Including,
The post-treatment step
Singular value decomposition (hereinafter referred to as SVD) method of a system matrix (C _k ) obtained by analyzing and integrating all processes except for the post-processing step and the pre-processing step performed at the time of transmission in response to the post-processing in the F-OFDM system A reception method of an F-OFDM system for decomposing into a plurality of matrices according to the method and post-processing the received data according to a post-processing matrix obtained by using at least one of the decomposed plurality of matrices. The method of claim 10, wherein the post-processing matrix is
The system matrix (C _k ) is decomposed into first and second singular matrix vectors (U _k , V ^H _k ) and diagonal matrix (Λ _k ), which are complex identity matrices, and the decomposed first singular matrix vector (U _k ) The reception method of the F-OFDM system obtained by the Hermit matrix (U ^H _k ) of. The method of claim 10, wherein the receiving step
Filtering the SF-OFDM symbol using the reception filter to obtain a CP-OFDM symbol;
Removing a periodic prefix inserted into the CP-OFDM symbol with a predetermined length and performing parallelization to obtain an OFDM symbol; And
Converting the OFDM symbol into the received data in a designated manner; F-OFDM system receiving method comprising a.

Images