OFDM Transmitter and OFDM Receiver

Abstract

To suppress both PAPR and OOBE, realize an ideal error rate, and reduce a calculation amount.SOLUTION: The present invention comprises: a constellation modulation unit 11 for generating from information bit sequences a data symbol that is a vector composed of information symbols mapped to a constellation; a DFT precoding processing unit 12 for multiplying a DFT precoder to the data symbols outputted from the constellation modulation unit 11; a cyclic shift unit 13 for applying a cyclic shift to the data symbols outputted from the DFT precoding processing unit 12; a pulse shaping unit 14 for applying a pulse shaping process to the data symbols outputted from the cyclic shift unit 13; a precoding processing unit 15 for applying a precoding process to the data symbols outputted from the pulse shaping unit 14; and applying inverse Fourier transformation to the modulated symbols that have been precoded, thereby generating an OFDM signal.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an OFDM transmitter that transmits an orthogonal frequency division multiplexing (OFDM) signal and an OFDM receiver that receives an OFDM signal. In particular, the present invention relates to an OFDM transmitter and an OFDM receiver suitable for accommodating more wireless systems within a frequency band by suppressing high interference power generated by an OFDM signal in an adjacent band.

Background Art

[0002] As a method for transmitting a digital signal, a modulation method called an orthogonal frequency division multiplexing (OFDM) method has been conventionally applied. In this OFDM method, a large number of orthogonal subcarriers are provided within a transmission band, data is assigned to the amplitude and phase of each subcarrier, and digital modulation is performed by PSK (Phase Shift Keying) or QAM (Quadrature Amplitude Modulation).

[0003] Since this OFDM method divides the transmission band with a large number of subcarriers, the band per subcarrier becomes narrow and the modulation rate does not become slow, but the total transmission rate has the characteristic that it is the same as the conventional modulation method. In addition, the OFDM method has the characteristic that the symbol rate becomes slow because a large number of subcarriers are transmitted in parallel. For this reason, this OFDM method can shorten the relative multipath time length with respect to the symbol time length, and can enhance the resistance to a multipath transmission path. In addition, since data is assigned to a plurality of subcarriers in the OFDM method, the OFDM method also has the characteristic that a transmission and reception circuit can be configured by using an IFFT (Inverse Fast Fourier Transform) arithmetic circuit that performs an inverse Fourier transform during modulation and an FFT (Fast Fourier Transform) arithmetic circuit that performs a Fourier transform during demodulation.

[0004] Among them, the DFT-s-OFDM method has been adopted in 4G and 5G wireless communication standards. By suppressing the high PAPR (peak power), which is one of the problems of the OFDM approach, it enables transmission on low-cost communication terminals. On the other hand, the transmission signal of this DFT-s-OFDM method has the problem of interference with the adjacent band due to high OOBE (out-of-band emission), similar to the conventional OFDM method. In the Beyond5G era, in order to meet the expected high performance requirements, the research and development of a method for simultaneously suppressing the PAPR and OOBE of OFDM transmission signals becomes even more important.

[0005] When transmitting information using such an OFDM method, it is necessary to suppress the interference power to the OFDM signal in order to prevent crosstalk from adjacent channels. As such a technique for suppressing interference power, the conventional OP (Orthogonal Precoding) technique has been proposed (see, for example, Non-Patent Document 1). This OP technique has an advantage in terms of error rate compared to the method of suppressing interference power using filters and window functions, and is also an excellent technique having backward compatibility with conventionally implemented OFDM systems. When applying OP to the DFT-s-OFDM method, a method for minimizing the error with respect to the original DFT-s-OFDM transmission signal in the sense of mean square error has already been studied (see, for example, Non-Patent Document 2). On the other hand, a method for improving the PAPR suppression effect of DFT-s-OFDM by combining phase rotation and pulse shaping has also been proposed (see, for example, Non-Patent Document 3).

Prior Art Documents

Non-Patent Documents

[0006]

Non-Patent Document 1

[0007] In the method of applying OP to the conventional DFT-s-OFDM system, there is a problem in the representation of the complex matrix for the optimal OP for combination with the DFT-s-OFDM system, resulting in a large computational load, so there are problems in terms of implementability. Also, since the memory required for matrix storage is large, there is a problem that it becomes difficult to adapt to various environments by storing a plurality of matrices. Further, it is difficult to improve the PAPR suppression effect in combination with phase rotation and pulse shaping. Especially in the case of phase rotation, a large number of complex multiplications are required, so there are also problems from the viewpoint of computational load.

[0008] Therefore, the present invention has been devised in view of the above-described problems. An object of the present invention is to provide an OFDM transmitter that transmits an orthogonal frequency division multiplexing (OFDM) signal and an OFDM receiver that receives an OFDM signal. In particular, under the DFT-s-OFDM system, an OFDM transmitter and an OFDM receiver that can suppress both PAPR and OOBE, realize an ideal error rate, and adapt to various environments by reducing the computational load are provided.

Means for Solving the Problems

[0009] The OFDM transmitter according to the first invention is an OFDM transmitter that generates and transmits an orthogonal frequency division multiplexing (OFDM) signal from a series of D information bits. In the OFDM transmitter, a constellation modulation unit generates a data symbol, which is a vector having information symbols mapped to a constellation from the series of information bits as elements; a DFT precoding processing unit multiplies the data symbol output from the constellation modulation unit by a DFT precoder; a cyclic shift unit performs a cyclic shift on the data symbol output from the DFT precoding processing unit; a pulse shaping unit performs pulse shaping processing on the data symbol output from the cyclic shift unit; a precoding processing unit performs precoding processing on the data symbol output from the pulse shaping unit; an inverse Fourier transform unit generates a baseband OFDM signal in which D information series are multiplexed in the frequency direction by performing an inverse Fourier transform on the modulated symbol output from the precoding processing unit; a serial-to-parallel conversion unit performs serial-to-parallel conversion on the baseband OFDM signal; and a transmission unit frequency-converts the serially-to-parallel converted OFDM signal into an RF band signal and transmits the signal.

[0010] The OFDM transmitter according to the second invention is the OFDM transmitter according to the first invention, wherein the precoding processing unit performs the precoding processing by multiplying by an orthogonal precoder P0 represented by the following formula. P0 = QGL l W AP0 = O Here, Q = I K -YY H is a unitary matrix I K : K-th order identity matrix Y ∈ C K×(K-D) : rectangular matrix G: orthogonal matrix for pulse shaping (sparse and orthogonal) L l ∈ {0, 1} D×D : lower cyclic shift matrix (sparse) l: cyclic shift amount that minimizes the expected value of PAPR W: D-th normalized DFT matrix (unitary) A ∈ C K×(K-D) : Any matrix describing OOBE suppression characteristics K: Number of subcarriers D: Number of information symbols

[0011] In the OFDM transmitter according to the third invention, in the first invention, the precoding processing unit multiplies a matrix P0, which minimizes the error with P, as an orthogonal precoder for PAPR and OOBE suppression, thereby performing the precoding processing. DFT Here, P = GL DFT W: Orthogonal precoder in redundant DFT-s-OFDM for PAPR suppression l P = P0 is the solution of the minimization problem min ||P P - P0|| DFT s.t. AP0 = O, P × P F = I H = I D of

[0012] The OFDM receiver according to the fourth invention is an OFDM receiver that receives an orthogonal frequency division multiplexing (OFDM) signal, and includes a receiving unit that frequency-converts a received RF-band signal into a baseband OFDM signal, a serial-to-parallel conversion unit that performs serial-to-parallel conversion on the OFDM signal frequency-converted by the receiving unit, a Fourier transform unit that performs Fourier transform processing on the serial-to-parallel converted OFDM signal, a post-decoding processing unit that performs post-decoding processing on the equalized symbol output from the Fourier transform unit, a pulse shaping unit that performs pulse shaping processing on the demodulated symbol output from the post-decoding processing unit, a cyclic shift unit that performs a cyclic shift on the demodulated symbol output from the pulse shaping unit, a DFT post-decoding processing unit that multiplies a DFT post-decoder by the demodulated symbol output from the cyclic shift unit, and a constellation demodulation unit that generates D information bit sequences from the demodulated symbol output from the post-decoding processing unit.

[0013] For the OFDM receiver according to the fifth invention, in the fourth invention, the post-decoding processing unit is an orthogonal post-decoder P0 consisting of the following formula H (where H represents conjugate transpose) is multiplied to perform the post-decoding processing. P0 = QGL l W AP0 = O Here, Q = I K -YY H a unitary matrix consisting of I K : K-th identity matrix Y ∈ C K×(K-D) : rectangular matrix G: orthogonal matrix for pulse shaping (sparse and orthogonal) L l ∈ {0, 1} D×D : lower cyclic shift matrix (sparse) l: cyclic shift amount that minimizes the expected value of PAPR W: D-th normalized DFT matrix (unitary) A ∈ C K×(K-D) : any matrix that describes the OOBE suppression characteristic K: number of subcarriers D: number of information symbols

[0014] For the OFDM receiver according to the sixth invention, in the fourth invention, the post-decoding processing unit is an orthogonal post-decoder P0 consisting of the following formula H (where H represents conjugate transpose) is multiplied to perform the post-decoding processing. Here, P DFT = GL l W: orthogonal precoder in redundant DFT-s-OFDM for PAPR suppression P = P0 is the solution of the minimization problem min P ||P DFT - P0|| F s.t. AP0 = O, P × P H = I D of

Advantages of the Invention

[0015] According to the present invention having the above-described configuration, in a practical DFT-s-OFDM system, additional low-computation circular shift processing of DFT inputs, pulse shaping processing of DFT outputs, etc. are added, and also by multiplying with the orthogonal precoder matrix P and the orthogonal post-decoder P H both PAPR and OOBE can be suppressed, an ideal error rate can be achieved, and by reducing the computational complexity, it becomes possible to adapt to various environments.

Brief Description of the Drawings

[0016]

Figure 1

Figure 2

Figure 3

Figure 4

Embodiments for Carrying Out the Invention

[0017] FIG. 1 is a block diagram showing the configuration of an OFDM (Orthogonal Frequency Division Multiplexing) transmitter 1 to which the present invention is applied.

[0018] This OFDM transmitter 1 has a constellation modulation unit 11, a DFT precoding processing unit 12, a circular shift unit 13, a pulse shaping unit 14, a precoding processing unit 15, a subcarrier mapping unit 1, an inverse Fourier transform unit 17, a CP insertion unit 18, a serial-parallel conversion unit 19, and a transmission unit 20 connected in sequence.

[0019] The constellation modulation unit 11 performs constellation modulation on the input D-bit information bit sequence according to a predetermined data modulation method such as QPSK modulation. As a result, in this constellation modulation unit 11, a data symbol d i is generated by mapping from the information bit sequence to the constellation. The constellation modulation unit 11 outputs this generated data symbol d i .

[0020] The DFT precoding processing unit 12 performs DFT precoding processing on the data symbol d i output from the constellation modulation unit 11. In this DFT precoding processing, a DFT precoder is multiplied.

[0021] The cyclic shift unit 13 performs a cyclic shift on the data symbol output from the DFT precoding processing unit 12.

[0022] The pulse shaping unit 14 performs pulse shaping processing on the data symbol output from the cyclic shift unit 13.

[0023] The precoding processing unit 15 multiplies the data symbol output from the pulse shaping unit 14 by the orthogonal precoder P0. Details of the orthogonal precoder matrix P0 to be multiplied will be described in detail later. The precoding processing unit 15 outputs the modulated symbol c i (= P0 × d i ) to the subcarrier mapping unit 16.

[0024] The subcarrier mapping unit 16 maps the precoded modulated symbol c i to the subcarriers used for transmission.

[0025] The inverse Fourier transform unit 17 performs IFFT (inverse Fourier transform) processing on the modulation symbols as sub-carrier samples input in parallel from the sub-carrier mapping unit 16 and synthesizes them. As a result, the inverse Fourier transform unit 17 can generate a baseband OFDM signal in which D information bit sequences are multiplexed in the frequency direction.

[0026] The CP insertion unit 18 inserts, as a cyclic prefix (CP), a length Tc from the end of the OFDM signal to the head of the OFDM signal so that the OFDM signal becomes cyclically longer by a predetermined length Tc.

[0027] The serial-to-parallel conversion unit 19 performs serial-to-parallel conversion on the OFDM signal output from the CP insertion unit 18.

[0028] The transmission unit 20 performs frequency conversion on the OFDM signal output from the serial-to-parallel conversion unit 19 to an RF band signal, amplifies it as necessary, and transmits it via an antenna.

[0029] FIG. 2 is a block diagram showing the configuration of the OFDM receiver 2 to which the present invention is applied.

[0030] This OFDM receiver 2 has a receiving unit 21, a parallel-to-serial conversion unit 22, a CP insertion unit 23, a Fourier transform unit 24, a sub-carrier demapping unit 25, a post-decoding processing unit 26, a pulse shaping unit 27, a cyclic shift unit 28, a DFT post-decoding processing unit 29, and a constellation demodulation unit 30 connected in sequence.

[0031] The receiving unit 21 frequency-converts the RF band signal received via the antenna into a baseband OFDM signal.

[0032] The parallel-to-serial conversion unit 22 performs parallel-to-serial conversion on the OFDM signal output from the receiving unit 21.

[0033] The CP removal unit 23 removes the CP inserted into the OFDM signal.

[0034] The Fourier transform unit 24 performs a Fourier transform process on the OFDM signal output from the CP removal unit 23. At this time, the Fourier transform unit 24 may further perform channel equivalent processing on this OFDM signal.

[0035] The subcarrier demapping unit 25 performs a process of demapping subcarriers on the equalized symbol on which the Fourier transform has been performed, thereby obtaining the equivalent symbol q i to obtain.

[0036] The post-decoding processing unit 26 multiplies the equalized symbol q i output from the subcarrier demapping unit 25 by the orthogonal post-decoder matrix P H . The post-decoding processing unit 26 outputs the demodulated symbol r i (= P H × q i ) to the pulse shaping unit 27.

[0037] The pulse shaping unit 27 performs a pulse shaping process on the demodulated symbol r i output from the post-decoding processing unit 26.

[0038] The cyclic shift unit 28 performs a cyclic shift on the demodulated symbol output from the pulse shaping unit 27.

[0039] The DFT post-decoding processing unit 29 performs DFT post-decoding processing on the demodulated symbol output from the cyclic shift unit 28.

[0040] The constellation demodulation unit 30 performs constellation demodulation on the demodulated symbol output from the DFT post-decoding processing unit 29 according to a predetermined data modulation method such as QPSK modulation. Through this constellation demodulation, the constellation demodulation unit 30 generates a series of D information bits from the demodulated symbol r i .

[0041] The OFDM transmitter 1 and the OFDM receiver 2 having the above-described configuration can be similarly realized only by newly adding a precoding processing unit 12 and a post-decoding processing unit 26 to a conventional OFDM system that performs CP, frequency-domain equivalence, etc.

[0042] Next, in the precoding processing unit 15, an orthogonal precoder matrix P0 that multiplies the data symbol d i will be described.

[0043] The precoding processing unit 15 multiplies by an orthogonal precoder matrix P0 that is established by the following equation. Also, the post-decoding processing unit 26 multiplies by an orthogonal post-decoder matrix P H (where H represents the complex conjugate transpose). P0 = QGL l W AP0 = O Here, Q = I K -YY H is a unitary matrix I K : K-th order identity matrix Y ∈ C K×(K-D) : rectangular matrix G: orthogonal matrix for pulse shaping (sparse and orthogonal) L l ∈ {0, 1} D×D : lower cyclic shift matrix (sparse) l: cyclic shift amount that minimizes the expected value of PAPR W: D-th order normalized DFT matrix (unitary) A ∈ C K×(K-D) : any matrix that describes the OOBE suppression characteristic K: number of subcarriers D: number of information symbols

[0044] The number of subcarriers K mentioned here is the number of subcarriers that are mapped in the subcarrier mapping unit 16 or demapped in the subcarrier demapping unit 23.

[0045] In the present invention having the above-described configuration, the precoding processing unit 15 is not limited to multiplying by the orthogonal precoder matrix P0 that satisfies the above equation.

[0046] For example, for the data symbol d i output from the constellation modulation unit 11, in performing DFT precoding processing, by multiplying through the following equation by a D-th normalized DFT matrix (unitary), d i ´ (= W × d i ) is calculated.

[0047] Next, in the cyclic shift unit 13, for the data symbol d i ´, by multiplying by a lower cyclic shift matrix (sparse) L l , c i ´ (= L l × d i ´) is calculated.

[0048] Next, in the pulse shaping unit 14, for the data symbol c i ´ output from the cyclic shift unit 13, by multiplying by a pulse shaping orthogonal matrix (sparse and orthogonal) G, g i (= G × c i ´) is calculated.

[0049] Next, in the precoding processing unit 15, for the data symbol g i output from the pulse shaping unit 14, by multiplying by a rectangular matrix Y, c i (= Y × g i ) is calculated.

[0050] Also, in the present invention having the above-described configuration, on the OFDM receiver 2 side as well, it is not limited to multiplying by the orthogonal post-decoder matrix P0 that satisfies the above equation. By dividing by Y, G, L1, and W respectively through the post-decoding processing unit 26, the pulse shaping unit 27, the cyclic shift unit 28, and the DFT post-decoding processing unit 29, r i = P H×q i may also perform the calculation of.

[0051] On the OFDM receiver 2 side, perform post-decoding that is reversible with respect to the precoding executed on the OFDM transmitter 1 side. That is, the OFDM receiver 2 can achieve ideal error rate characteristics by performing an inverse operation with low computational complexity.

[0052] Any orthogonal precoder matrix P and orthogonal post-decoder matrix P that satisfy the above-mentioned conditions H may be used.

[0053] At this time, the precoding processing unit 15 may perform the above precoding processing by multiplying a matrix P0 that minimizes the error with P as an orthogonal precoder for PAPR and OOBE suppression. DFT

[0054] Here, P DFT = GL l W: Orthogonal precoder in redundant DFT-s-OFDM for PAPR suppression P = P0 is the solution of the minimization problem min P ||P DFT - P0|| F s.t. AP0 = O, P × P H = I D

[0055] The post-decoding processing unit 26 may perform post-decoding processing by multiplying a matrix P0 that minimizes the error with P as an orthogonal post-decoder P DFT for PAPR and OOBE suppression. H

[0056] According to the OFDM transmitter 1 and OFDM receiver 2 having the above-described configuration, in a practical DFT-s-OFDM system, additional low-complexity processing such as circular shift processing of DFT inputs and pulse shaping processing of DFT outputs is added, and orthogonal precoder matrix P and orthogonal post-decoder PH By multiplying, both the PAPR and the OOBE can be suppressed, an ideal error rate can be achieved, and by reducing the computational complexity, it becomes possible to adapt to various environments.

[0057] Figure 3 shows the suppression effect of the OOBE (interference power). In both cases, the horizontal axis is the offset frequency (MHz), and the vertical axis is the power spectral density (dBm / Hz). CP-OFDM is the normal OFDM using a cyclic prefix. In this example of Figure 3, the present invention is compared with the conventional CP-OFDM. The dotted line A is a 55 dB spectral mask provided at both ends of the transmission band based on the average power density within the system band.

[0058] As shown in Figure 3, it can be seen that through the present invention, the interference power in the targeted band where notches are formed can be effectively suppressed, and an improvement effect of 30 dB or more is obtained. By appropriately generating and selecting the matrix (precoder P0) used in the precoding process, effective interference power suppression is possible even for wireless systems with different bandwidths.

[0059] Figure 4 shows the suppression effect of the PAPR (peak power). The horizontal axis is the PAPR threshold, and the vertical axis is the CCDF exceeding the PAPR threshold. The effects of a general conventional CP-OFDM, a general DFT-s-OFDM, and the present invention are respectively verified. As shown in Figure 4, according to the present invention having the above-described configuration, it is shown that the PAPR can be significantly reduced. That is, it is shown that the OOBE suppression method generally degrades the PAPR suppression performance of DFT-s-OFDM, but the degradation of the present invention is guaranteed to be minimal. For example, it can be seen that it is the least squares method of the error with respect to the original DFT-s-OFDM signal.

[0060] Hereinafter, the calculation basis of the orthogonal precoder matrix P and the orthogonal post-decoder matrix P described above in the present invention H will be described.

[0061] In the digital OFDM modulation method, the OFDM transmission signal in the discrete-time domain complex equivalent baseband is represented in the form of block transmission as follows.

[0062] [Number] (1)

[0063] Here, τ is the discrete time, S i [τ] is the i-th OFDM symbol, N is the OFDM symbol sample period, and L is the guard interval (GI) sample length. When inserting a cyclic prefix (CP) into the GI, S i [τ] is represented in the form of the inverse discrete Fourier transform (IDFT) of the following equation.

[0064] [Number] (2)

[0065] Here, Σ is the summation operator, k ∈ {k0, k1,..., kK-1} and K is the number of subcarriers, c i,k ∈ C is the modulation symbol, e is the Napier number, j is the imaginary unit, π is the pi, and I[τ] is a rectangular window function such that I[τ] = 1 for τ = -L, -L + 1,..., N - 1 and I[τ] = 1 and I[τ] = 0 for other τ, without impairing the effective length of the CP. The calculation of equation (2) can be efficiently implemented by the FFT algorithm, and the number of FFT points coincides with N. The power spectral density (PSD) of equation (1) is represented as follows.

[0066] [Number] (3)

[0067] Here, taking f as the frequency and T as the OFDM symbol period, the angular frequency ω = 2πfT / N and Ns = N + L. Also, Pi(ω) is the i-th OFDM symbol S iIt is the power spectrum of [τ] and is given by the following equation (4).

[0068]

Number

[0069] Also, the following equation (5) is the Fourier transform of S i [τ], where F is the Fourier transform operator.

[0070]

Number

[0071] By substituting equation (1) into equation (5), the following equation (6) can be obtained.

[0072]

Number

Number

Number

[0073]

Number

[0074] Also, equation (7) can be transformed as follows.

[0075]

Number

[0076] Here, Ns = N + L. Equation (8) can be summarized by the following equations (9) and (10).

[0077]

Number

[0078] Here, the following equation (10) represents the cardinal sign.

Number

[0079] Next, consider the pre-coding represented by the following equation (11).

[0080]

Number

[0081] Here, P ∈ C K×D is a matrix called a pre-coder, and d i = [d i,0 d i,1 ... d i , D-1 T is the data symbol for transmitting D information symbols {d i,l}. D-1 l=0 Also, the process of restoring d i from c i is defined as post-decoding. When the orthogonality of the following equation is satisfied, equation (11) is called orthogonal pre-coding, and P is called an orthogonal pre-coder.

[0082]

Number

[0083] Here, (·) H is an operator representing the complex conjugate transpose of (·), and I m ∈ {0} m×m is the identity matrix. From equations (11) and (12), the following equation holds.

[0084] [Number] (13)

[0085] From equation (13), for the orthogonal precoding represented by equation (11), the multiplication by P H becomes post - decoding. Therefore, at the receiving side, for the symbol q i ∈ C K×1 , post - decoding is applied. That is, the transmitted data d i can be restored by calculating the following equation (14).

[0086] [Number] (14)

[0087] By calculating this (14), the transmitted data d i can be restored. By appropriately determining P, the characteristics of the OFDM signal can be improved. In particular, the method of using P to suppress the out - of - band emission (OOBE) of the OFDM signal and improve the spectral characteristics is called spectral precoding. Assuming that the constraint conditions for c i to suppress OOBE are represented by an M - th order linear equation as follows. At this time, equation (13) is rewritten as the following equation.

[0088] [Number] (15)

[0089] Here, 0 m ∈ {0} m×1 represents a zero vector. For equation (15) to hold for any d i by the orthogonal precoding represented by equation (11), the following equation being satisfied is a necessary and sufficient condition.

[0090] [Number] (16)

Number

[0091] Define the implementation of redundant DFT-s-OFDM with cyclic shift, which is more implementable than the conventional redundant DFT-s-OFDM with phase rotation. The DFT precoding is expressed by the following equation.

[0092]

Number

[0093] Here, W ∈ C D×D is a normalized DFT matrix and is defined by the following equation (18) with l and n being the indices of rows and columns respectively.

[0094]

Number

[0095] Define a cyclic shift extension consisting of the following equation (19) for the above equation (17).

[0096]

Number

[0097] Here, T M is a matrix for adding a cyclic suffix of length M by multiplying from the left as shown in the following equation (20).

[0098]

Number

[0099] L lis called a lower cyclic shift matrix and is defined by the following equation.

[0100] [Number] (21)

[0101] For equation (19), pulse shaping of equation (22) is performed.

[0102] [Number] (22)

[0103] Here, G K is represented by the following equation (23).

[0104] [Number] (23)

[0105] Equation (19) is equivalent to the phase rotation for each element of d i . Generally, in the case of an arbitrary phase rotation for a complex vector, there is a problem that 4D real multiplications are required. However, since the cyclic extension with shift shown in equation (19) is a calculation-free operation, it is excellent in terms of implementation. The value of l0 is appropriately determined so that the expected value of the maximum peak of the OFDM transmission signal is minimized.

[0106] From equations (17), (19), and (22), the pre-coder of the cyclic shift redundant DFT-s-OFDM can be expressed as the following equation (24).

[0107] [Number] (24)

[0108] Here, G = G K T M ∈ R K×D .

[0109] P shown in equation (24)DFT In order to make G an orthogonal precoder, it is necessary to define G as a semi-orthogonal matrix. In the following, to satisfy this condition, {g l} is assumed to be a square root-raised cosine function as in (25).

[0110]

number

[0111] In this case, {g l Considering the non-zero elements of} and symmetry, only M real-type storage units are required to store G. Similarly, G is a sparse matrix with only 2M non-zero and non-one elements, and multiplying a complex vector by G requires only 4M real-type multiplications.

[0112] For the orthogonal precoders defined by equations (24) and (25), when K=D and l0=0, G=I D KatsuL l0 =I D Since this holds, then P DFT =W, which is the same as the DFT precoder W used in the normal DFT-s-OFDM. In the present invention, P DFT The calculations of (11) and (14) are implemented using the quadrature spectral precoder of (26) that minimizes the error for

[0113]

number

[0114] Here, Q is a unitary matrix as shown in the following equation (27) and Y∈C K×M is.

[0115]

number

[0116] Q in equation (26) is determined such that P0 is the solution of the following optimization problem.

[0117]

Equation

[0118] This equation (28) means that when P = P0, it is to search for the solution of the minimization problem min P ||P DFT - P0|| F s.t. AP0 = O.

[0119] (1) Since the OFDM modulation shown in the equation is a linear operation, the error of the transmitted signal in the time domain when comparing the cases using P and P DFT is also minimized in the sense of the Euclidean norm when P = P0.

[0120] From equation (26), by multiplying Q to the output vector of the redundant DFT-s-OFDM with cyclic shift, the pre-coding process and the post-decoding process of the present invention for simultaneously suppressing OOBE and PAPR can be implemented. The storage of real number type required for storing Q in equation (27) is only 2KM for storing Y. This is extremely smaller than the K2 required when directly storing.

[0121] The pre-coding computational complexity of the redundant DFT-s-OFDM only adds the computational complexity related to the multiplication of GK for the complex vector compared with the normal DFT-s-OFDM. This additional computational complexity is negligibly small compared with the DFT pre-coding computational complexity.

[0122] The pre-coding computational complexity in the present invention, compared with redundant DFT-s-OFDM, only adds the computational complexity related to the multiplication of Q for the complex vector because the cyclic extension with shift in Equation (19) does not require calculation. The number of real multiplications of Q for the complex vector can be achieved with 4MK = 4M(D + M) by performing the multiplications in the right-associative order using the matrix decomposition shown in Equation (27). This additional computational complexity is comparable to the DFT pre-coding computational complexity or the IDFT computational complexity for OFDM modulation. Since DFT pre-coding is adopted in the uplink of 3GPP's 4G and 5G standards, it can be confirmed that the present invention is practical from the perspective of computational complexity.

Explanation of Signs

[0123] 1 Transmitter 2 Receiver 11 Constellation Modulation Unit 12 Pre-coding Processing Unit 13 Cyclic Shift Unit 14 Pulse Shaping Unit 15 Pre-coding Processing Unit 16 Sub-carrier Mapping Unit 17 Inverse Fourier Transform Unit 18 CP Insertion Unit 19 Serial-to-Parallel Conversion Unit 20 Transmitting Unit 21 Receiving Unit 22 Post-decoding Processing Unit 23 CP Removal Unit 24 Fourier Transform Unit 25 Sub-carrier Demapping Unit 26 Post-decoding Processing Unit 27 Pulse Shaping Unit 28 Cyclic Shift Unit 29 Post-decoding Processing Unit 30 Constellation Demodulation Unit

Claims

1. In an OFDM transmitter that generates and transmits an orthogonal frequency division multiplexing (OFDM) signal from D information bit sequences, a constellation modulation unit that generates a data symbol, which is a vector having information symbols mapped to a constellation from the information bit sequences as elements; a DFT precoding processing unit that multiplies a DFT precoder to the data symbol output from the constellation modulation unit; a cyclic shift unit that performs a cyclic shift on the data symbol output from the DFT precoding processing unit; a pulse shaping unit that performs pulse shaping processing on the data symbol output from the cyclic shift unit; a precoding processing unit that performs precoding processing on the data symbol output from the pulse shaping unit; an inverse Fourier transform unit that generates a baseband OFDM signal in which D information sequences are multiplexed in the frequency direction by performing an inverse Fourier transform on the modulated symbol output from the precoding processing unit; a serial-to-parallel conversion unit that performs serial-to-parallel conversion on the baseband OFDM signal; and a transmission unit that frequency-converts the serially-to-parallel converted OFDM signal into an RF band signal and transmits it. The OFDM transmitter is characterized by comprising these components.

2. The precoding processing unit performs the precoding processing by multiplying an orthogonal precoder P represented by the following equation ０ ​ The OFDM transmitter according to claim 1, characterized in that: P ０ = QGL ｌ W AP ０ = O Here, Q = I Ｋ - YY Ｈ is a unitary matrix consisting of I Ｋ : Identity matrix of order K Y ∈ C Ｋ×(Ｋ-Ｄ) : rectangular matrix G: Orthogonal matrix for pulse shaping (sparse and orthogonal) L l ∈ {0, 1} Ｄ×Ｄ : Lower-circulant matrix (sparse) l: Cyclic shift amount that minimizes the expected value of PAPR W: D-th normalized DFT matrix (unitary) A ∈ C Ｋ×(Ｋ-Ｄ) : Any matrix that describes the OOBE suppression characteristic K: Number of subcarriers D: Number of information symbols

3. The pre-coding processing unit multiplies a matrix P ＤＦＴ with the minimum error with ０ to perform the pre-coding processing by multiplying it as an orthogonal pre-coder for PAPR and OOBE simultaneous suppression. The OFDM transmitter according to claim 1, characterized in that: Here, P ＤＦＴ = GL ｌ W: Orthogonal precoder in redundant DFT-s-OFDM for PAPR suppression P = P ０ is the solution of the minimization problem min Ｐ ||P ＤＦＴ - P ０ || Ｆ s.t. AP ０ = O, P × P Ｈ = I Ｄ of

4. In an OFDM receiver that receives an orthogonal frequency division multiplexing (OFDM) signal, a receiving unit that frequency-converts the received RF band signal into a baseband OFDM signal; a parallel-to-serial conversion unit that performs parallel-to-serial conversion on the OFDM signal frequency-converted by the receiving unit; a Fourier transform unit that performs Fourier transform processing on the parallel-to-serial converted OFDM signal; a post-decoding processing unit that performs post-decoding processing on the equalization symbol output from the Fourier transform unit; a pulse shaping unit that performs pulse shaping processing on the demodulated symbol output from the post-decoding processing unit; a cyclic shift unit that performs a cyclic shift on the demodulated symbol output from the pulse shaping unit; A DFT post-decoding processing unit that multiplies a DFT post-decoder to the demodulated symbol output from the above-mentioned cyclic shift unit, and a constellation demodulation unit that generates D information bit sequences from the demodulated symbol output from the above-mentioned post-decoding processing unit. An OFDM receiver characterized by the above.

5. The post-decoding processing unit performs the post-decoding processing by multiplying the orthogonal post-decoder P having the following formula ０ Ｈ (where H represents the complex conjugate transpose). The OFDM receiver according to claim 4, characterized by the above. P ０ = QGL ｌ W AP ０ = O Here, Q = I Ｋ - YY Ｈ is a unitary matrix consisting of I Ｋ : Identity matrix of order K Y ∈ C Ｋ×(Ｋ-Ｄ) : rectangular matrix G: Orthogonal matrix for pulse shaping (sparse and orthogonal) L l ∈ {0, 1} Ｄ×Ｄ : Lower-circulant matrix (sparse) l: Cyclic shift amount that minimizes the expected value of PAPR W: D-th normalized DFT matrix (unitary) A ∈ C Ｋ×(Ｋ-Ｄ) : Any matrix describing the OOBE suppression characteristic K: Number of subcarriers D: Number of information symbols

6. The post-decoding processing unit performs the post-decoding processing by multiplying with an orthogonal post-decoder P represented by the following formula ０ Ｈ (where H represents conjugate transpose). The OFDM receiver according to claim 4, characterized by the above. Here, P ＤＦＴ = GL ｌ W: Orthogonal precoder in redundant DFT-s-OFDM for PAPR reduction P = P ０ is the minimization problem min Ｐ ||P ＤＦＴ - P ０ || Ｆ s.t. AP ０ = O, P × P Ｈ = I Ｄ Solution of

Images