CN103227769A - Novel method for reducing peak-to-average ratio of STBC MIMO-OFDM system - Google Patents

Abstract

The invention discloses a novel method for reducing a peak-to-average ratio of an STBC (Space Time Block Code) MIMO-OFDM (Multiple Input Multiple Output-Orthogonal Frequency Division Multiplexing) system, and provides a self-adaption cross-antenna rotation and inversion method based on STBC. A PAPR (Peak-to-Average Power Ratio) of an MIMO-OFDM technology can be reduced. The method comprises the steps that the size of a weight factor is adjusted continuously according to a PAPR variation trend and the variation quantity; the operation probability of a signal subblock executing rotation and inversion is obtained, and then optimized by repeated iterations; an optimal sequence is further generated to perform subblock rotation and inversion; and the PAPR of a signal is reduced. If a threshold of the PAPR is preset for the system, the computation complexity can be further reduced.

Description

A kind of method of new reduction STBC MIMO-OFDM system peak-to-average ratio

Technical field

The present invention relates to the technical field in multi-I/O OFDM (MIMO-OFDM) system of broadband wireless communication signal improved, particularly relate to and a kind ofly reduce STBC MIMO-OFDM system peak-to-average ratio by the rotation of self adaptation crossed antenna and the method for negating.

Background technology

OFDM (OFDM) is considered to be in the important technology of realizing high speed data transfer in the wireless time varying channel as a kind of multi carrier modulation scheme efficiently.In addition, multiple-input and multiple-output (MIMO) is another attracting technology, by dispose a plurality of antennas at receiving-transmitting sides, can improve channel transmission rate and obtain bigger space diversity gain, thereby improve power system capacity and performance.MIMO is combined with OFDM, make this two kinds of technical advantage complementations, thereby have the availability of frequency spectrum and good numerous advantages such as anti-multipath decline performance efficiently.But because MIMO-OFDM uses the OFDM modulation system, it also exists signal to have higher peak-to-average power ratio problems such as (PAPR) when succession OFDM modulates numerous advantages inevitably.

Usually, in the MIMO-OFDM system, can use existing efficient algorithm to the data on the antenna.For example partial transmission sequence (partial transmit sequence, PTS) algorithm or select mapping (selected mapping, SLM) algorithm, thereby reduce the PAPR of signal on the every antenna.But, owing to not only adopt the OFDM modulating system on each antenna, can produce high PAPR, and the interference of a plurality of antennas also can further produce high PAPR in mimo system.Document " Mizhou Tan; Zoran Latinovi ' c; and Yeheskel Bar-Ness; " STBC MIMO-OFDM Peak-to-Average power ratio reduction by Cross-Antenna rotation and inversion; " IEEE Communications Letters, vol.9, no.7, pp.592-594,2005 " a kind of method that is called crossed antenna rotation and negate (CARI) has been proposed; this method is to produce the higher degree of freedom to carrying out on whole antennas by sub-piece rotation and complementary operation, to offset the possibility that meets with the bad sequence with high PAPR.Yet this method number of permutations is too big, though proposed more actual successive suboptimal CARI(SS-CARI) and random suboptimal CARI(RS-CARI) solution, its computation complexity has reduced, and its PAPR reduction has obtained restriction.Document " Yi-Sheng Su; Tsung-Cheng Wu; Chung-Hsuan Wang; and Min-Kuan Chang; " A Low-Complexity Cross-Antenna Rotation and Inversion Scheme for PAPR Reduction of STBC MIMO-OFDM Systems "; IEEE16th International Workshop on Computer Aided Modeling and Design of Communication Links and Networks, pp.31-35,2011 " a kind of cross entropy CARI(CE-CARI has been proposed) solution; though this scheme has reduced certain complexity and PAPR, the setting of its parameter is a fixed value.

The present invention proposes a kind of self adaptation CARI method, can further be optimized the traditional crossed antenna rotation and (CARI) method of negating.Self adaptation CARI algorithm is a kind of algorithm based on STBC MIMO-OFDM system, it is according to the variation tendency of PAPR and changes size, come to carry out on the sub-piece of signal calculated the evolutionary operator probability and the corresponding weights factor of rotating and negating, pass through repeatedly iteration optimization evolutionary operator probability again, the generation optimal sequence carries out sub-piece rotation and negates, thereby reduces the PAPR value of signal.Compare with conventional method under identical condition, the method can more effectively reduce the PAPR of system.If give the thresholding Th of systemic presupposition PAPR, can further reduce the complexity of calculating.

Summary of the invention

For more effectively overcoming the above-mentioned defective that exists in the MIMO-OFDM system, the object of the invention provides and a kind ofly can reduce peak-to-average power ratio in the MIMO-OFDM system, and can more effectively be applied to the method in the practical communication system.The method is with the sub-piece execution rotation of signal and negates, obtains the evolutionary operator probability of each sub-piece by certain data processing; The size of trend that changes according to signal PAPR and variable quantity is constantly adjusted the size of weight factor again, further optimizes the evolutionary operator probability of corresponding sub-piece; The operation that comes seizing signal to carry out the rotation of sub-piece and negate by this probability produces optimal sequence through iteration repeatedly and carries out sub-piece rotation and negate, thereby reduces the PAPR value of signal.In iterative process, reach default PAPR thresholding Th or iterations up to the PAPR of signal value and reach default maximum iteration time T, then the termination of iterations process.

The size that innovation part of the present invention has been to propose a kind of trend that changes according to signal PAPR and variable quantity is constantly adjusted the size of weight factor, the change according to the signal actual conditions of the size of weight factor is changed, carry out crossed antenna rotation thus adaptively and negate, reduce the PAPR value of signal through iteration repeatedly effectively.

Innovation part of the present invention is self adaptation crossed antenna rotation that is proposed and the method for negating, and can need according to system, sets the thresholding of PAPR, reduces iterations, and then has reduced the calculation of complex amount.

The present invention is a kind of method of new reduction STBC MIMO-OFDM system peak-to-average ratio.Described method detailed process may further comprise the steps:

The signal that step 1 input is original, the complex base band signal of MIMO-OFDM system can be expressed as:

$x_i\left(t\right)=\frac{1}{\sqrt{N}}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{X}_n{e}^{{j2\mathrm{\πf}}_{n}t}$ 0≤t≤NT (1)

N represents sub-carrier number in the formula (1), and i represents transmitting antenna, f _n=n Δ f is the frequency of n subcarrier on the i root antenna, Δ f=1/ (NT), and T represents mark space, X _nThe transmission symbol of representing n subcarrier on the i root antenna;

Step 2 pair signal carries out over-sampling;

Step 3 is with the signal X on each transmitting antenna _iBe divided into M equal-sized sub-piece X _i=[X _{I, 1}, X _{I, 2}..., X _{I, M}], wherein r represents the radical of transmitting antenna, i=1, and 2 ..., r; When the iteration first time, promptly during t=1, with probability P ¹=[p ¹ _{M, j}] _{1≤m≤M, 1≤j≤4}In each element be set to

Step 4 is with P ^tFor probability produces Q 1 * M dimension sequence at random

Q=1,2 ... Q, each element in this sequence go up the random integers that distribute in [1,4], wherein 1,2,3 and 4 represent that respectively signal is not carried out any operation, the sub-piece rotation of execution on a certain sub-piece, the sub-piece of execution is negated and carry out sub-piece rotation and negate signal X _iAccording to sequence

In data carry out the rotation of corresponding sub block and negate, obtain signal

I=1 wherein, 2 ..., r;

Step 5 is calculated and is worked as q=1, and 2 ..., the maximum PAPR of signal during Q is expressed as it:

$F\left(V_q^t\right)=\max \{\mathrm{PAPR}\left(X_1^q\right),\mathrm{PAPR}\left(X_2^q\right),\·\·\·,\mathrm{PAPR}\left(X_r^q\right)\}---\left(2\right)$

And will $F\left(V_q^t\right)$ Arrange from small to large, promptly $F\left(V_1^t\right)\≤F\left(V_2^t\right)\≤\·\·\·F\left(V_Q^t\right);$

In step 6 determining step 5

Value whether less than default PAPR thresholding, if judged result is

Value be less than or equal to thresholding Th, then forward step 12 to; If judged result is

Value greater than thresholding Th, then forward step 7 to;

Step 7 is established

Here hithermost integer is got in [*] expression, to rounding greatly;

Step 8 Probability p _{M, j}Be that signal is on m sub-piece, when carrying out the operation of j kind Less than a ^tTotal value under the condition and execution all operations Less than a ^tThe ratio of the total value under the condition, that is:

$p_{m,j}^{t+1}=\frac{\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\left\{\underset{V_{q,m}^t=1}{\overset{4}{\mathrm{\Σ}}}\right\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}\}}{\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}}---\left(3\right)$

Σ { * } expression summation operation in the formula (3), and m ∈ (1, M), j ∈ (1,4);

When calculating the t time iteration, step 9 carries out all operations Less than a ^tTotal value s under the condition (t), that is:

$s\left(t\right)=\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}---\left(4\right)$

Reduction degree according to PAPR in twice iterative process in front and back is adjusted the probability P proportion, and promptly weight factor λ is:

Parameter b ∈ (0,1) in the formula (5), λ ∈ (0,1);

Step 10 is utilized probability P that formula (3) and formula (5) obtain and the value of λ, the probability P that will use when promptly size that changes according to signal PAPR and trend obtain next iteration:

P ^t+1=λP ^t+(1-λ)P ^t+1 （6）

Here be that probability matrix P further is optimized;

Step 11 judges whether t reaches default maximum iteration time T, if do not reach then iterations t is increased by 1, and turns to step 2; If reach, then turn to step 12;

Step 12 finds the minimum PAPR of whole constant series, through optimal sequence V _OptCarry out the rotation of sub-piece and negate, resulting signal is carried out the STBC coding.

Beneficial effect of the present invention is that this method is according to system's needs, the thresholding of setting PAPR has reduced the calculation of complex amount, and according to the trend of PAPR variation and the size of variable quantity, come the sub-piece of seizing signal to be rotated and negate by certain data processing, reduce the PAPR value of signal through iteration repeatedly.Compare with traditional method, new method can more effectively reduce the PAPR of system under identical condition.

Description of drawings

The FB(flow block) of a kind of new reduction STBC MIMO-OFDM system peak-to-average ratio method of Fig. 1;

Fig. 2 adopts the curve chart of different PAPR threshold T h to the PAPR of system performance impact;

Fig. 3 adopts the curve chart of different iterationses to the PAPR of system performance impact;

Fig. 4 adopts the curve chart of different parameter ρ to the PAPR of system performance impact;

Fig. 5 adopts the curve chart of different parameter b to the PAPR of system performance impact;

Fig. 6 adopts different sub-pieces to count M the PAPR of system Effect on Performance is reached and other algorithm performance curve chart relatively;

Fig. 7 adopts different sequence number Q that the PAPR of system Effect on Performance is reached and other algorithm performance curve chart relatively, wherein among Fig. 6 and Fig. 7 successive suboptimal method and random suboptimal method be document " Mizhou Tan; Zoran Latinovi ' c; and Yeheskel Bar-Ness; " STBC MIMO-OFDM Peak-to-Average power ratio reduction by Cross-Antenna rotation and inversion; " IEEE Communications Letters, vol.9, no.7, pp.592-594,2005 " propose; cross-entropy method is document " Yi-Sheng Su, Tsung-Cheng Wu, Chung-Hsuan Wang, and Min-Kuan Chang, " A Low-Complexity Cross-Antenna Rotation and Inversion Scheme for PAPR Reduction of STBC MIMO-OFDM Systems ", IEEE16th International Workshop on Computer Aided Modeling and Design of Communication Links and Networks, pp.31-35,2011 " propose.

Embodiment

Provide the specific implementation method of this patent below:

The original signal of rapid 1 input, the complex base band signal of MIMO-OFDM system can be expressed as:

$x_i\left(t\right)=\frac{1}{\sqrt{N}}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{X}_n{e}^{{j2\mathrm{\πf}}_{n}t}$ 0≤t≤NT (1)

N represents sub-carrier number in the formula (1), and i represents transmitting antenna, f _n=n Δ f is the frequency of n subcarrier on the i root antenna, Δ f=1/ (NT), and T represents mark space, X _nThe transmission symbol of representing n subcarrier on the i root antenna;

Step 2 pair signal carries out over-sampling, and when the over-sampling coefficient L of system 〉=4, the PAPR of discrete signal just can represent the PAPR of continuous signal well;

Step 3 is with the signal X on each transmitting antenna _iBe divided into M equal-sized sub-piece X _i=[X _{I, 1}, X _{I, 2}..., X _{I, M}], wherein r represents the radical of transmitting antenna, i=1, and 2 ..., r; When the iteration first time, promptly during t=1, with probability P ¹=[p ¹ _{M, j}] _{1≤m≤M, 1≤j≤4}In each element be set to

Step 4 is if use two antennas, to m (after the height piece of 1≤m≤M) carries out the crossed antenna rotation and negate, can obtain 4 different OFDM constant series set, they are respectively:

Original collection: X ₁=[X _1,1..., X _{1, m}..., X _{1, M}] and X ₂=[X _2,1..., X _{2, m}..., X _{2, M}],

M the original collection that sub-piece is negated:

X ₁=[X _1,1... ,-X _{1, m}..., X _{1, M}] and X ₂=[X _2,1... ,-X _{2, m}..., X _{2, M}],

The original collection of m sub-piece exchange:

X ₁=[X _1,1..., X _{2, m}..., X _{1, M}] and X ₂=[X _2,1..., X _{1, m}..., X _{2, M}],

M sub-piece exchange and the original collection of negating:

X ₁=[X _1,1... ,-X _{2, m}..., X _{1, M}] and X ₂=[X _2,1... ,-X _{1, m}..., X _{2, M}],

By that analogy, other sub-piece is carried out identical operations; With P ^tFor probability produces Q 1 * M dimension sequence at random

, q=1,2,, Q, each element in this sequence is [1,4] go up the random integers that distribute, wherein 1,2,3 and 4 represent that respectively signal is not carried out any operation, the sub-piece rotation of execution on a certain sub-piece, the sub-piece of execution is negated and carry out sub-piece rotation and negate signal X _iAccording to sequence

In data carry out the rotation of corresponding sub block and negate, obtain signal

I=1 wherein, 2 ..., r;

Step 5 is calculated and is worked as q=1, and 2 ..., the maximum PAPR of signal during Q is expressed as it:

$F\left(V_q^t\right)=\max \{\mathrm{PAPR}\left(X_1^q\right),\mathrm{PAPR}\left(X_2^q\right),\·\·\·,\mathrm{PAPR}\left(X_r^q\right)\}---\left(2\right)$

And will $F\left(V_q^t\right)$ Arrange from small to large, promptly $F\left(V_1^t\right)\≤F\left(V_2^t\right)\≤\·\·\·F\left(V_Q^t\right);$

In step 6 determining step 5

Value whether less than default PAPR thresholding, if judged result is

Value be less than or equal to thresholding Th, then forward step 12 to; If judged result is

Value greater than thresholding Th, then forward step 7 to;

Step 7 is established

Here hithermost integer is got in [*] expression, to rounding greatly;

Step 8 Probability p _{M, j}Be that signal is on m sub-piece, when carrying out the operation of j kind

Less than a ^tTotal value under the condition and execution all operations

Less than a ^tThe ratio of the total value under the condition, that is:

$p_{m,j}^{t+1}=\frac{\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\left\{\underset{V_{q,m}^t=1}{\overset{4}{\mathrm{\Σ}}}\right\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}\}}{\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}}---\left(3\right)$

Σ { * } expression summation operation in the formula (3), and m ∈ (1, M), j ∈ (1,4);

When calculating the t time iteration, step 9 carries out all operations

Less than a ^tTotal value s under the condition (t), that is:

$s\left(t\right)=\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}---\left(4\right)$

Reduction degree according to PAPR in twice iterative process in front and back is adjusted the probability P proportion, and promptly weight factor λ is:

Parameter b ∈ (0,1) in the formula (5), λ ∈ (0,1);

Step 10 is utilized probability P that formula (3) and formula (5) obtain and the value of λ, the probability P that will use when promptly size that changes according to signal PAPR and trend obtain next iteration:

P ^t+1=λP ^t+(1-λ)P ^t+1 （6）

Here be that probability matrix P further is optimized;

Step 11 judges whether t reaches default maximum iteration time T, if do not reach then iterations t is increased by 1, and turns to step 2; If reach, then turn to step 12;

Step 12 finds the minimum PAPR of whole constant series, through optimal sequence V _OptCarry out the rotation of sub-piece and negate, resulting signal is carried out the STBC coding; Signal X _i(i=1,2 ..., r), become signal through after sub-piece rotation and negating Because

With

Have identical PAPR characteristic, therefore, first symbol period is carried out the reduction of PAPR, the mode that re-uses STBC is encoded, and resultant signal can not influence the size of whole PAPR; For the STBC coding, be example to use two antennas, with two symbols

With

Send into encoder, send from two slave antennas respectively through the symbol behind the coding: at first symbol period, symbol

With

On transmitting antenna 1 and transmitting antenna 2, send simultaneously respectively; At second symbol period, symbol

With

On transmitting antenna 1 and transmitting antenna 2, send wherein () simultaneously respectively ^*The expression conjugation.

Effect of the present invention further specifies by emulation, has used 2 transmitting antennas, and ofdm system subchannel number N is 128, and the over-sampling coefficient L of system is 4, in each subchannel data is carried out the QPSK modulation.

Fig. 2 is for adopting different PAPR threshold T h to the PAPR of system Effect on Performance.The sub-piece that parameter ρ=0.6, parameter b=0.6, maximum iteration time T=10, signal is divided into is set counts M=16 and sequence number Q=32.As can be seen from the figure, under the situation of other parameter constant, PAPR thresholding Th value is more little, and simulated effect is good more, but the calculation of complex amount is big more; And when thresholding Th value was reduced to certain value, the PAPR of signal no longer reduced, and this is because algorithm can only drop to the PAPR of signal to a certain degree, and the calculation of complex amount also no longer becomes big.So this threshold value can be done balance between the reduction degree of the PAPR of system and amount of calculation, set according to the situation of the reality of system.

Fig. 3 is for adopting different iterationses to the PAPR of system Effect on Performance.The sub-piece that parameter ρ=0.6, parameter b=0.6, PAPR threshold T h=6dB, signal is divided into is set counts M=16 and sequence number Q=32.As can be seen from the figure, under the situation of other parameter constant, iterations is many more, and it is many more that the PAPR of system reduces.But so the increase that amount of calculation also can be at double when improving iterations is the value that maximum iteration time can be suitable.

Fig. 4 is for adopting different parameter ρ to the PAPR of system Effect on Performance.The sub-piece that parameter b=0.6, PAPR threshold T h=6dB, maximum iteration time T=10, signal is divided into is set counts M=16 and sequence number Q=32.As can be seen from the figure, under the situation of other parameter constant, along with parameter ρ increases, the value of system PAPR is to reduce earlier afterwards to increase.This is because when parameter ρ increases, it is many more to relate to the displacement kind that produces less PAPR, finds optimal sequence easily, but when parameter ρ when continue becoming big, in the constant series that can produce big PAPR be included in, so parameter ρ can be suitable in 0.6 left and right sides value.

Fig. 5 is for adopting different parameter b to the PAPR of system Effect on Performance.The sub-piece that parameter ρ=0.6, PAPR threshold T h=6dB, maximum iteration time T=10, signal is divided into is set counts M=16 and sequence number Q=32.As can be seen from the figure, under the situation of other parameter constant, variation along with parameter b, system PAPR overlaps basically, this be since parameter b influence be that weight factor changes degree slowly, less to the weight factor influence, and the size of the variable quantity of PAPR is bigger to the weight factor influence, so parameter b value in (0,1) scope all can.

M reaches the PAPR of system Effect on Performance Fig. 6 and the comparison of other algorithm performance for the different sub-piece of employing is counted.Parameter ρ=0.6, parameter b=0.6, PAPR threshold T h=6dB, maximum iteration time T=10 and sequence number Q=32 are set.As can be seen from the figure, it is big more that the sub-piece of these four kinds of methods is counted M, and the PAPR performance of system is good more.When all parameter constants, new method is identical with the constant series number of random suboptimal method, and under the situation about lacking than the constant series number of other method, the PAPR performance of new method is best.This is to change because of the variation of the important parameter of new method in iterative process with signal PAPR, makes these parameters more meet the actual conditions of signal on the direction of amount that changes and variation.

Fig. 7 for adopt different sequence number Q to the PAPR of system Effect on Performance and with the comparison of other algorithm performance.The sub-piece that parameter ρ=0.6, parameter b=0.6, PAPR threshold T h=6dB, maximum iteration time T=10 and signal be divided into is set counts M=16.As can be seen from the figure, under the situation of other parameter constant, it doesn't matter owing to successive suboptimal method and parameter Q, and the PAPR performance of system does not change; Other method is that constant series are many more because of parameter Q is big more, and the PAPR performance of system is good more, but parameter Q can not be excessive, and too conference repeats the constant series of generation, and amount of calculation also can increase accordingly simultaneously.

Claims (2)

1. the method for a new reduction STBC MIMO-OFDM system peak-to-average ratio is characterized in that:

The signal that step 1 input is original, the complex base band signal of MIMO-OFDM system can be expressed as:

$x_i\left(t\right)=\frac{1}{\sqrt{N}}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{X}_n{e}^{j2\mathrm{\π}{f}_{n}t}$ 0≤t≤NT (1)

N represents sub-carrier number in the formula (1), and i represents transmitting antenna, f _n=n Δ f is the frequency of n subcarrier on the i root antenna, Δ f=1/ (NT), and T represents mark space, X _nThe transmission symbol of representing n subcarrier on the i root antenna;

Step 2 pair signal carries out over-sampling;

Step 3 is with the signal X on each transmitting antenna _iBe divided into M equal-sized sub-piece X _i=[X _{I, 1}, X _{I, 2}..., X _{I, M}], wherein r represents the radical of transmitting antenna, i=1, and 2 ..., r; When the iteration first time, promptly during t=1, with probability P ¹=[p ¹ _{M, j}] _{1≤m≤M, 1≤j≤4}In each element be set to

Step 4 is with P ^tFor probability produces Q 1 * M dimension sequence at random

Each element in this sequence is to go up the random integers that distribute in [1,4], and wherein 1,2,3 and 4 represent that respectively signal is not carried out any operation, the sub-piece rotation of execution on a certain sub-piece, the sub-piece of execution is negated and carry out sub-piece rotation and negate signal X _iAccording to sequence

In data carry out the rotation of corresponding sub block and negate, obtain signal

I=1 wherein, 2 ..., r;

Step 5 is calculated and is worked as q=1, and 2 ..., the maximum PAPR of signal during Q is expressed as it:

$F\left(V_q^t\right)=\max \{\mathrm{PAPR}\left(X_1^q\right),\mathrm{PAPR}\left(X_2^q\right),\·\·\·,\mathrm{PAPR}\left(X_r^q\right)\}---\left(2\right)$

And will $F\left(V_q^t\right)$ Arrange from small to large, promptly $F\left(V_1^t\right)\≤F\left(V_2^t\right)\≤\·\·\·\≤F\left(V_Q^t\right);$

In step 6 determining step 5

Value whether less than default PAPR thresholding, if judged result is

Value be less than or equal to thresholding Th, then forward step 12 to; If judged result is

Value greater than thresholding Th, then forward step 7 to;

Step 7 is established

Here hithermost integer is got in [*] expression, to rounding greatly;

Step 8 Probability p _{M, j}Be that signal is on m sub-piece, when carrying out the operation of j kind

Less than a ^tTotal value under the condition and execution all operations

Less than a ^tThe ratio of the total value under the condition, that is:

$p_{m,j}^{t+1}=\frac{\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\left\{\underset{V_{q,m}^t=1}{\overset{4}{\mathrm{\Σ}}}\right\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}\}}{\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}}---\left(3\right)$

Σ { * } expression summation operation in the formula (3), and m ∈ (1, M), j ∈ (1,4);

When calculating the t time iteration, step 9 carries out all operations

Less than a ^tTotal value s under the condition (t), that is:

$s\left(t\right)=\underset{q=1}{\overset{Q}{\mathrm{\Σ}}}\{F\left(V_q^t\right)|F\left(V_q^t\right)\≤a^t\}---\left(4\right)$

Reduction degree according to PAPR in twice iterative process in front and back is adjusted the probability P proportion, and promptly weight factor λ is:

Parameter b ∈ (0,1) in the formula (5), λ ∈ (0,1);

Step 10 is utilized probability P that formula (3) and formula (5) obtain and the value of λ, the probability P that will use when promptly size that changes according to signal PAPR and trend obtain next iteration:

P ^t+1=λP ^t+(1-λ)P ^t+1 （6）

Here be that probability matrix P further is optimized;

Step 11 judges whether t reaches default maximum iteration time T, if do not reach then iterations t is increased by 1, and turns to step 2; If reach, then turn to step 12;

Step 12 finds the minimum PAPR of whole constant series, through optimal sequence V _OptCarry out the rotation of sub-piece and negate, resulting signal is carried out the STBC coding.

2. the method for a kind of new reduction STBC MIMO-OFDM system peak-to-average ratio according to claim 1 is characterized in that in the step 9 according to system's actual needs, further reduces the complexity of calculating for the thresholding Th of systemic presupposition PAPR.

Images