LOW-COMPLEXITY SELECTED MAPPING METHOD USING CYCLICAL REDUNDANCY CHECK

Abstract

A low-complexity selective mapping method using cyclic redundancy check is provided. In performing coding, a transmitter adds a check bit to information bits to be transmitted to obtain modulated data. Demodulation is performed on an M-order modulation symbol received by a receiver to obtain a decoding result of a coding polynomial of the modulation symbol and bit information received by the receiver. A modulo-2 division result of the decoding result of the coding polynomial and a generation polynomial is calculated. In a case that a remainder of the modulo-2 division result is equal to zero, if the modulated data corresponding to the same index value of the receiver and the transmitter are identical, a current iteration is stopped, and a current value is outputted as a phase rotation sequence index recovery value. Finally, the receiver obtains a decoded signal.

Claims

1. A low-complexity selective mapping method using cyclic redundancy check, comprising: adding, by a transmitter in performing coding, a check bit and an auxiliary check bit to information bits to be transmitted, and then performing modulation, by the transmitter, to obtain modulated data A(n); multiplying, by the transmitter, the modulated data A(n) to be transmitted by Q groups of different phase sequences to perform phase rotation to obtain a modulation symbol A′(n), and transmitting, by the transmitter, a sequence with a smallest PAPR; performing demodulation on an M-order modulation symbol Â′.sub.Q(n) received by a receiver to obtain a decoding result of a coding polynomial of the modulation symbol and bit information received by the receiver; calculating a modulo-2 division result of the decoding result of the coding polynomial and a generation polynomial; in a case that a remainder of the modulo-2 division result is not equal to zero, performing a next iteration until the remainder of the modulo-2 division result is equal to zero or the number of iterations is equal to Q; in a case that a remainder of the modulo-2 division result is equal to zero, if modulation data corresponding to an index directory s(m) is consistent with a decoding result of auxiliary check information in an i-th iteration corresponding to the directory, that is, A(s(m))=Â.sub.i(s(m)) and auxiliary verification is successful, stopping the i-th iteration, and outputting a current value of i as a phase rotation sequence index recovery value; and if (s(m))≠(s(m)), and auxiliary verification fails, determining whether the number of iterations is equal to Q, performing a next iteration in a case that the number of iterations is not equal to Q, and performing an ML algorithm on Â.sub.i(s(m)) to obtain a phase rotation sequence index recovery value in a case that the number of iterations is equal to Q; and obtaining a decoded signal by the receiver based on the phase rotation sequence index recovery value and a received modulation symbol Y′(n′) after the check bit is removed.

2. The low-complexity selective mapping method using cyclic redundancy check according to claim 1, wherein the information bit after being adding with the check bit by the transmitter is expressed as: $A\left(n\right)=\{\begin{array}{ccc}0& n=\left(2s\left(m\right)-1,2s\left(m\right)\right)& 1\le m\le \frac{n}{2}\\ A\left(n\right)& n\ne \left(2s\left(m\right)-1,2s\left(m\right)\right)& n=\left(1,2.Math.n^{\prime }\right)\\ R\left(l\right)& l=\left(1,2,.Math.r\right)& n=\left(n^{\prime }+1,n^{\prime }+2.Math.n^{\prime }+r\right)\end{array}$ where A(n) represents the information bit after being added with the check bit by the transmitter; one symbol comprises n bits, comprising r check bits, m auxiliary check bits, and n′ data bits to be transmitted; s(m) represents an m-th element in an index vector of a QPSK modulation symbol of A(n); and R(l) represents an l-th check bit.

3. The low-complexity selective mapping method using cyclic redundancy check according to claim 1, wherein the M-order modulation symbol Â′.sub.Q(n) received by the receiver is expressed as:
Â′.sub.Q(n)=U.sub.QY′(n) where U.sub.Q represents a phase rotation matrix formed by the Q groups of different phase sequences, Y′(n) represents a modulation symbol obtained by performing IFFT transformation on a received signal, Ĥ(n) represents a channel response estimation value, and y(n) represents the received signal received by the receiver.

4. The low-complexity selective mapping method using cyclic redundancy check according to claim 1, wherein the coding polynomial is expressed as:
C(x)=A(x)×x.sup.r+R(x)=Q(x)G(x) where C(x) represents the coding polynomial, A(x) represents an information polynomial, R(x) represents a remainder polynomial, Q(x) represents a quotient polynomial, and G(x) represents the generation polynomial.

5. The low-complexity selective mapping method using cyclic redundancy check according to claim 4, wherein the modulo-2 division result of the decoding result of the coding polynomial and the generation polynomial is expressed as: $\frac{{\hat{R}}_i\left(x\right)}{G\left(x\right)}=\frac{{\hat{C}}_i\left(x\right)}{G\left(x\right)}-\widehat{Q\left(x\right)}$ where Ĉ.sub.i(x) represents a decoding result of the coding polynomial in an i-th iteration, {circumflex over (Q)}(x) represents a decoding result of the quotient polynomial, and {circumflex over (R)}.sub.i(x) represents a decoding result of the remainder polynomial in the i-th iteration.

6. The low-complexity selective mapping method using cyclic redundancy check according to claim 1, wherein the ML calculation algorithm is performed on Â.sub.i(s(m)) to obtain the phase rotation sequence index recovery value by using the following equation: $\hat{q}=\underset{i\in \left(1,Q\right)}{\min }\underset{m=1}{\overset{m}{.Math.}}{.Math.{\hat{A}}_i^{\prime }\left(s\left(m\right)\right)-A^{\prime }\left(s\left(m\right)\right).Math.}^2$ where m represents the number of the auxiliary check bits.

7. The low-complexity selective mapping method using cyclic redundancy check according to claim 1, wherein the receiver obtains the decoded signal based on the phase rotation sequence index recovery value and the received modulation symbol Y′(n′) after the check bit is removed by using the following equation:
Â′(n′)=Y′(n′).Math.U.sub.{circumflex over (q)} where Â′(n′) represents the decoded signal, and U.sub.{circumflex over (q)} represents a rotation phase corresponding to the phase rotation sequence index recovery value.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0023] FIG. 1 is a system block diagram of an SLM solution based on a comb-type pilot OFDM system according to the conventional technology;

[0024] FIG. 2 is a system block diagram of a low-complexity selective mapping method using cyclic redundancy check according to the present disclosure;

[0025] FIG. 3 is a schematic diagram showing a comparison of suppression performances of PAPR corresponding to different parameters in an SLM algorithm;

[0026] FIG. 4 shows phase index error rates corresponding to different solutions in a case of recovering a phase sequence by using a Rayleigh channel;

[0027] FIG. 5 shows BERs corresponding to different channels according to different solutions; and

[0028] FIG. 6 shows a comparison of computation complexity ratios of different algorithms.

DETAILED DESCRIPTION OF EMBODIMENTS

[0029] The technical solutions in the embodiments of the present disclosure are described clearly and completely in conjunction with the drawings in the embodiments of the present disclosure hereinafter. It is apparent that the described embodiments are only some embodiments of the present disclosure, rather than all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present disclosure without any creative work fall within the protection scope of the present disclosure.

[0030] A low-complexity selective mapping method using cyclic redundancy check is provided according to the present disclosure. The method includes the following operations.

[0031] A transmitter in performing coding adds a check bit to information bits to be transmitted to obtain modulated data A(n).

[0032] The transmitter multiplies the modulated data A(n) to be transmitted by Q groups of different phase sequences to perform phase rotation to obtain a modulation symbol A′(n). That is, A′(n) is an M-order modulation symbol corresponding to A(n). The transmitter transmits a sequence with a smallest PAPR.

[0033] Demodulation is performed on an M-order modulation symbol Â′.sub.Q(n) received by a receiver to obtain a decoding result of a coding polynomial of the modulation symbol and bit information received by the receiver.

[0034] A modulo-2 division result of the decoding result of the coding polynomial and a generation polynomial is calculated. In a case that a remainder of the modulo-2 division result is not equal to zero, a next iteration is performed until the remainder of the modulo-2 division result is equal to zero or the number of iterations is equal to Q.

[0035] In a case that a remainder of the modulo-2 division result is equal to zero, if A(s(m))=Â.sub.i(s(m)), it indicates that auxiliary verification is successful, a current iteration is stopped, and a current value of i is outputted as a phase rotation sequence index recovery value.

[0036] If (s(m))≠(s(m)), it indicates that auxiliary verification fails, and it is determined whether the number of iterations is equal to Q. In a case that the number of iterations is not equal to Q, a next iteration is performed. In a case that the number of iterations is equal to Q, an ML algorithm is performed on Â.sub.i(s(m)) to obtain a phase rotation sequence index recovery value.

[0037] The receiver obtains a decoded signal based on the phase rotation sequence index recovery value and Y′(n′).

[0038] In the SLM, the transmitter multiplies the modulated data A(n) to be transmitted by Q groups of different phase sequences U.sub.Q=(U.sub.1, U.sub.2, . . . , U.sub.q, . . . U.sub.Q), U.sub.q=[u.sub.q,1, u.sub.q,2 . . . u.sub.q,N−1].sup.T, q∈(1, Q), u.sub.q,n=e.sup.jϕ.sup.q,n, to perform phase rotation to obtain A.sub.Q(n). ϕ.sub.q,n represents a uniform random number ranging from 0 and 2π. Usually, for facilitating calculation,

[00002]${\phi }_{q,n}=\left\{\pm \pi \right\}\mathrm{or}{\phi }_{q,n}=\left\{\pm \pi ,\pm \frac{\pi }{2}\right\}.$

N represents the number of subcarriers. IFFT is calculated, and a PAPR of each of the sequences is calculated. A sequence with a smallest PAPR is transmitted. The transmitted sequence is expressed as:

[00003]$a_q\left(n\right)=\underset{q\in \left(1,Q\right)}{\min }\mathrm{PAPR}\left(a\left(n\right)•U_q\right)$

[0039] where a(n) represents a time-domain symbol after IFFT transformation, and a.sub.q(n) represents the q-th group of time-domain symbol after IFFT transformation that is determined as data to be transmitted.

[0040] FIG. 1 is a system block diagram of an SLM solution based on a comb-type pilot OFDM system according to the conventional technology. An input signal bit A(n) may be expressed as:

[00004]$A\left(n\right)=A\left(\mathrm{mL}+l\right)=\{\begin{array}{cc}{A}_{\mathrm{pilot}}\left(m\right)& l=0\\ A_{\mathrm{data}}\left(\mathrm{mL}+l\right)& l=1,2.Math.L-1\end{array}$

where A.sub.pilot(m) represents a pilot bit, L represents an interval between two adjacent pilots, and A.sub.data(mL+1) represents a data bit.

[0041] The phase sequence U.sub.q includes a pilot phase sequence U.sub.q(P) and a data phase U.sub.q(d). The phase sequence is expressed as:

[00005]$u_{q,p}\left(m\right)=\{\begin{array}{cc}1& m=2r_0\\ e^{j{\varphi }_{q,m}}& m=2r_0+1\end{array}$

[0042] where

[00006]$0\le r_0\le .Math.\frac{N_P}{2}-1.Math.,$

N.sub.p represents the number of pilot subcarriers, and └ ┘ represents a round-down operation.

[0043] At the receiver, channel estimation is performed by using the pilot signal A.sub.pilot(m) to obtain a channel response estimation value Ĥ.sub.pilot(k). Then, interpolation operation is performed on Ĥ.sub.pilot(2r) and Ĥ.sub.pilot(2r+2) to obtain Ĥ.sub.pilot(2r).

[0044] The phase rotation sequence index recovery value q is obtained by using the following equation:

[00007]$\hat{q}=\underset{q\in \left(1,Q\right)}{\min }\underset{r=0}{\overset{.Math.\left(N_p-1\right)/2-1.Math.}{.Math.}}{.Math.Y_{\mathrm{pilot}}\left(2r+1\right)-{\hat{H}}_{\mathrm{pilot}}\left(2r+1\right)U_{q,2r+1}{A}_{\mathrm{pilot}}\left(2r+1\right).Math.}^2$

First Embodiment

[0045] With the SLM algorithm, phase recovery is performed at the receiver. Based on the ML-SLM solution with high-complexity calculation, a low-complexity method for performing phase recovery using CRC is provided according to the present disclosure, stopping calculation when a check value is equal to 0. The basic idea of the present disclosure is that in an ideal condition, if a phase is recovered correctly, a CRC value is equal to 0. Using known information for auxiliary verification can effectively reduce the possibility of false detection.

[0046] It is assumed that A(x) represents an information polynomial, G(x) represents a generation polynomial, Q(x) represents a quotient polynomial, R(x) represents a remainder polynomial, the order of A(x) is k, and the order of R(x) is r, it indicates that the length of the information bit is k and the length of the check bit is r.

[0047] CRC may be expressed as:

[00008]$\frac{A\left(x\right)\times x^r}{G\left(x\right)}=Q\left(x\right)+\frac{R\left(x\right)}{G\left(x\right)}$

[0048] The coding polynomial C(x) may be expressed as:

C(x)=A(x)×x.sup.r+R(x)=Q(x)G(x)

[0049] At the receiver, it is determined whether a bit error occurs based on a modulo-2 division result of a decoding result Ĉ(x) of C(x) and G(x). The modulo-2 division result is expressed as:

[00009]$\frac{\hat{C}\left(x\right)}{G\left(x\right)}=\hat{Q}\left(x\right)+\frac{\hat{R}\left(x\right)}{G\left(x\right)}$

[0050] In a case that Ĉ(x)=C(x), {circumflex over (R)}(x) is equal to 0. In a case that {circumflex over (R)}(x)=0, Ĉ(x) may be not equal to C(x) due to false detection P=1/2.sup.r.

[0051] As shown in FIG. 2, a coefficient of a polynomial, that is, a codeword, corresponds to a random binary sequence. A codeword C(g,r), g=k+r may be obtained based on C(x).

[0052] According to the above CRC check, a check bit may be added to the information bits to obtain A(n). In FIG. 2, A′(n) is a QPSK modulation symbol of A(n). After the check bit is added, A(n) may be expressed as:

[00010]$A\left(n\right)=\{\begin{array}{ccc}0& n=\left(2s\left(m\right)-1,2s\left(m\right)\right)& 1\le m\le -\frac{n}{2}\\ A\left(n\right)& n\ne \left(2s\left(m\right)-1,2s\left(m\right)\right)& n=\left(1,2.Math.n^{\prime }\right)\\ R\left(l\right)& l=\left(1,2,.Math.r\right)& n=\left(n^{\prime }+1,n^{\prime }+2.Math.n^{\prime }+r\right)\end{array}$

[0053] where an integer vector S={s(1), s(2) . . . s(m)} represents an index in A′(n); R(l) represents a coefficient corresponding to the polynomial R(x), that is, the number of the check bit; and one symbol includes n bits, including r check bits, m auxiliary check bits, n′=n−r data bits to be transmitted, and n−r−m data bits.

[0054] At the receiver, a received signal is expressed as:

y(n)=H(n).Math.a.sub.q(n)+n.sub.0

where n.sub.0 represents added Gaussian white noise.

[0055] The channel response estimation value Ĥ(n) may be obtained based on the pilot:

[00011]$\hat{H}\left(n\right)=\frac{y_p\left(n\right)}{x_p\left(n\right)}$

[0056] where y.sub.p(n) represents a received pilot signal, and x.sub.p(n) represents a pilot signal.

[0057] Y(n) is obtained by performing FFT operation on the received signal y(n). φ.sub.g,n={±π} represents a phase factor in a phase rotation matrix U.sub.Q, and U.sub.Q=U.sub.Q.sup.•. An M-order modulation symbol A′(n) is expressed as:

Â′(n′)=Y′(n′)®U.sub.{circumflex over (q)}

[0058] Ĉ(x) and Â.sub.i(s(m)) may be obtained by performing demodulation on Â′.sub.Q(n). A CRC check may be performed to determine whether to stop the iteration, which includes the following operations.

[0059] In a case that {circumflex over (R)}.sub.i(x)≠0, it indicates Ĉ.sub.i(x)≠C(x) and Ui≠Uq, a next iteration is performed until R.sub.i(x)=0 or i=Q. The process is expressed as:

[00012]$\begin{array}{c}\frac{{\hat{R}}_i\left(x\right)}{G\left(x\right)}=\frac{{\hat{C}}_i\left(x\right)}{G\left(x\right)}-\hat{Q}\left(x\right)\ne 0\\ A\left(s\left(m\right)\right)-{\hat{A}}_i\left(s\left(m\right)\right)=0\end{array}\}.Math.i=Q$

[0060] In a case that U.sub.i=U.sub.q, {circumflex over (R)}.sub.i(x)≠0 due to the bit error Ĉ.sub.i(g,r), the iteration in the algorithm is performed until the last time.

[0061] In a case that {circumflex over (R)}.sub.i(x)=0, in order to avoid misjudgment caused by missed detection and bit error, it is required to perform auxiliary verification on Ĉ.sub.i(g,r). In a case that an auxiliary verification is performed as the following expression:

[00013]$\begin{array}{c}\frac{{\hat{R}}_i\left(x\right)}{G\left(x\right)}=\frac{{\hat{C}}_i\left(x\right)}{G\left(x\right)}-\hat{Q}\left(x\right)=0\\ A\left(s\left(m\right)\right)-{\hat{A}}_i\left(s\left(m\right)\right)\ne 0\end{array}\}.Math.i\ne q$

Ĉ.sub.i(g,r)=C(g,r)+αG(k)≠C(g,r) due to a wrong phase sequence. Although {circumflex over (R)}.sub.i(x)=0, A(s(m))−Â.sub.i(s(m))≠0, and the auxiliary verification fails. Thus, it is required to perform a next iteration.

[0062] In a case that an auxiliary verification is performed as the following expression:

[00014]$\begin{array}{c}\frac{{\hat{R}}_i\left(x\right)}{G\left(x\right)}=\frac{{\hat{C}}_i\left(x\right)}{G\left(x\right)}-\hat{Q}\left(x\right)=0\\ A\left(s\left(m\right)\right)-{\hat{A}}_i\left(s\left(m\right)\right)=0\end{array}\}.Math.i=\hat{q}=q$

A(s(m))=Â.sub.i(s(m)), Ĉ.sub.i(g,r)=C(g,r), and i={circumflex over (q)}=q<Q, where i represents the number of iterations, q represents a phase rotation sequence index value, and q represents a phase rotation sequence index recovery value. It indicates that the phase recovery calculation in the method according to the present disclosure may be stopped, and it is unnecessary to perform ML calculation.

[0063] The case that i=Q is caused by determining a correct phase as a wrong phase (that is, {circumflex over (R)}.sub.i(x)=0) or the selected phase sequence being equal to Q. In the case that i=Q, it is required to perform ML solution according to the method of the present disclosure, and ML calculation is performed on Â.sub.Q (s(m)) by using the following equation:

[00015]$\hat{q}=\underset{i\in \left(1,Q\right)}{\min }\underset{m=1}{\overset{m}{.Math.}}{.Math.{\hat{A}}_i^{\prime }\left(s\left(m\right)\right)-A^{\prime }\left(s\left(m\right)\right).Math.}^2$

[0064] Finally, the phase rotation sequence index recovery value {circumflex over (q)} is expressed as:

[00016]$\hat{q}=\{\begin{array}{cc}i& i<Q,A\left(s\left(m\right)\right)-{\hat{A}}_i\left(s\left(m\right)\right)=0\\ \underset{q\in \left(1,Q\right)}{\min }\underset{1}{\overset{m}{.Math.}}{.Math.{\hat{A}}_q^{\prime }(\left(s\left(m\right)\right)-A^{\prime }\left(s\left(m\right)\right).Math.}^2& i=Q\end{array}$

[0065] At the receiver, the decoded signal Â′(n′) is expressed as:

Â′(n′)=Y′(n′).Math.U.sub.{circumflex over (q)}

Second Embodiment

[0066] In the embodiment, BERs and computation complexities corresponding to different channels according to different SLM solutions are compared, and the influence of the number of the phase sequence groups on the BER and the computation complexity is further analyzed. Tables 1, 2 and 3 show simulation parameters in the method. In addition, message bit sequences as input signals in different solutions are the same.

TABLE-US-00001 TABLE 1 OFDM system parameters Names Parameters Number of subcarriers N = 256 Channel AWGN and Ralyrnd Modulation mode QPSK Auxiliary verification information index S = {5, 30, 55, 80, 105} Frame length N.sub.sym = 8

TABLE-US-00002 TABLE 2 CRC verification parameters Names Parameters Parameter values CRC-4 x.sup.4 + x + 1 10011 Number of check bits r 4 CRC codeword C (g, r) A (g), g ∈ (n′ − k + 1, n′) Number of CRC g 16 codewords

TABLE-US-00003 TABLE 3 SLM parameter Names Parameters Parameter values Number of phase Q 4 or 8 sequences

[0067] Suppression performance on a PAPR is evaluated by using a complementary cumulative distribution function (CCDF) of the PAPR. The function represents a probability of a PAPR exceeding a threshold PAPR0. FIG. 3 shows suppression performances on the PAPR corresponding to different solutions. The phase factor φ.sub.q,n={±π} (corresponding to the solution with P=2 in FIG. 3) or the phase factor

[00017]${\phi }_{q,n}=\left\{\pm \pi ,\pm \frac{\pi }{2}\right\}$

(corresponding to the solution with P=4 in FIG. 3). It can be clearly seen from FIG. 3 that the suppression effect on the peak-to-average power ratio with the method according to the present disclosure is similar to the suppression effect on the peak-to-average power ratio with the P-NSI-SLM algorithm.

[0068] FIG. 4 shows phase index error rates (IERs) corresponding to different solutions in a case of recovering a phase sequence by using a Rayleigh channel. The IER is positively correlated with the number of the phase sequence groups. Compared with the SI-SLM solution, the influence of noise on the phase index decoding can be effectively reduced in the CRC-NSI-SLM solution and the P-NSI-SLM solution that adopt ML. In addition, with the CRC-NSI-SLM solution and the P-NSI-SLM solution, a good phase sequence recovery result can be achieved with high SNR.

[0069] The recovery of the phase sequence index is affected by noise, and false recovery of the phase sequence index results in an increase in the BER of the SLM algorithm. FIG. 5 shows BERs corresponding to different channels according to different solutions. The SI-SLM solution has a higher BER than the other two solutions due to that the SI-SLM solution has a higher IER. Since different phase sequences may have a same phase factor at a same subcarrier position, the BER of the SI-SLM solution is not as high as expected.

[0070] That is, an incorrect phase sequence index does not cause the message bits of the entire OFDM symbol to be incorrectly decoded. Apparently, with the method according to the present disclosure, the phase sequence can be effectively recovered, and a BER similar to the BER of the P-NSI-SLM solution can be achieved.

[0071] In the present disclosure, the complexity of the solution in the document of Low Complexity Data Decoding for SLM-based OFDM Systems without Side Information (abbreviated as CP-NSI-SLM) is compared with the complexity of the present disclosure (abbreviated as CRC-NSI-SLM). The complexity of determining q is shown in Table 4.

TABLE-US-00004 TABLE 4 Computation complexity comparison Operation mode CP-NSI-SLM CRC-NSI-SLM Complex multiplication [00018]$3Q.Math.\left[\frac{N_p-1}{2}\right]$ [00019]$m+[E\left(i\right).Math.g+\underset{1}{\overset{Q-1}{.Math.}}i\left(\frac{1}{2^r}+\mathrm{BER}\right).Math.2m)+2m]$ Addition [00020]$Q*.Math.\frac{N_p-1}{2}.Math.$ [00021]$3m+\left(\frac{1}{2^r}\mathrm{BER}\right)\mathrm{Qm}$

[0072] FIG. 6 shows a comparison of computation complexity ratios. In a case that the number of symbols is large enough, a phase sequence q is equally distributed in (1, 2 . . . Q). Due to bit errors and missed detections, an expectation E(i) of a phase sequence i may be expressed as:

[00022]$E\left(i\right)\approx \left[\frac{\underset{q=1}{\stackrel{Q}{.Math.}}q}{Q}+\underset{q=1}{\overset{Q-1}{.Math.}}q\left(\frac{1}{2^r}+\mathrm{BER}\right)\right]$

[0073] It can be seen from the Table 4 that Q and the IER have different weights for the complexity of multiplication and the complexity of addition.

[0074] Table 5 shows actual CCRR values corresponding to the CRC-NSI-SLM solution and P-NSI-SLM solution. CCRR is defined as:

[00023]$CCRR=\left(1-\frac{\mathrm{complexity}\mathrm{of}\mathrm{CRC}-NSI-SLM}{\mathrm{complexity}\mathrm{of}P-\mathrm{NSI}-\mathrm{SLM}}\right)\times 100\%$

[0075] By analyzing the data in Table 5, it can be seen that the complexity of the multiplication is affected by Q and BER. The IER increases with the increase of BER. The false recovery of the phase sequence index results in the increase of the number of the CRC calculations, thereby reducing the system performance. In addition, a larger Q indicates a more obvious effect of stopping calculation early. The complexity of the addition is mainly caused by calculating the phase rotation sequence index recovery value {circumflex over (q)} by using the ML algorithm in the present disclosure, and is greatly affected by Q. In addition, since the weight of the number N.sub.p of subcarriers is greater than m and g in a case that Q is constant, the CCRRs of the complexity of the multiplication and the complexity of the addition increases with the number of the subcarriers.

TABLE-US-00005 TABLE 5 Computation complexity of {circumflex over (q)} Type of SLM Multiplication addition CRC4, Q = 8, WAGN 59.81% 75.40% CRC4, Q = 4, WAGN 45.75% 50.10% CRC4, Q = 8, Ralyrnd 48.54% 75.25% CRC4, Q = 4, Ralyrnd 36.40% 50.04%

[0076] By comparing the IERs, BERs and computation complexities of the SI-SLM solution, the P-NSI-SLM solution and the CRC-NSI-SLM solution in different cases, it is shown that PAPR can be effectively suppressed with the method according to the present disclosure. Compared with the P-SNI-SLM solution, a greater computation gain can be achieved with the CRC-NSI-SLM solution without increasing the BER. In addition, it can be seen from the analysis of the calculation equations that CCRR is positively proportional to the number of subcarriers and the number of the phase sequence groups, and is inversely proportional to the BER.

[0077] Although the embodiments of the present disclosure are shown and described, those skilled in the art can understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principle and purpose of the present disclosure, and the scope of the present disclosure is defined by the claims and their equivalents.