CN102594761B

CN102594761B - Method for estimating noise power of orthogonal frequency division multiplexing (OFDM) signals

Info

Publication number: CN102594761B
Application number: CN201110394209.7A
Authority: CN
Inventors: 邵怀宗; 杨帆; 吴迪
Original assignee: University of Electronic Science and Technology of China
Current assignee: University of Electronic Science and Technology of China
Priority date: 2011-11-22
Filing date: 2011-11-22
Publication date: 2014-10-15
Anticipated expiration: 2031-11-22
Also published as: CN102594761A

Abstract

Provided is a method for estimating noise power of orthogonal frequency division multiplexing (OFDM) signals, which includes conducting fast fourier transform (FFT) algorithm of energy normalization after removing cyclic prefixes of the received signals; then extracting useful messages of pilot frequency or synchronous head from the FFT algorithm, and also extracting useful messages from the pilot frequency or synchronous head messages of the sending end data; and finally, calculating the noise power of signals by the method of using four points of adjacent subcarriers of the same or the adjacent two OFDM signs for offsetting the subcarrier channel frequency response error value, and combining with the useful messages extracted before. The method has good performance of resistance to frequency selective fading and time selective fading; simultaneously, the method can utilize both the pilot frequency and the synchronous head for estimating the noise power, while used for estimating the noise power, the method is applicable for situations of a plurality of synchronous heads, single symmetric synchronous head and single asymmetric synchronous head, and has good adaptability.

Description

A method for estimating the noise power of OFDM signal

技术领域 technical field

本发明涉及无线移动系统参数估计领域，具体涉及一种估算OFDM信号的噪声功率的方法。 The invention relates to the field of wireless mobile system parameter estimation, in particular to a method for estimating the noise power of OFDM signals. the

背景技术 Background technique

正交频分复用(OFDM)是一种多载波调制方式，它将串行的高速比特流转化为低速的并行比特流，运用简化的快速傅里叶逆变换(IFFT)，将需要传输的信息符号调制到各个相互正交的子信道上，使宽带的信道转变成较平坦的窄带信道，从而抑制了码间干扰。由于能够抵抗频率选择性衰落信道，并且具有频谱效率高和均衡简单等优点，因此很多通信标准和系统都采用OFDM调制，并被作为第四代移动通信的关键技术。噪声功率作为度量信道质量的一种指标，是OFDM系统的一个重要的参数。目前研究OFDM系统噪声功率估计的文献比较多，主要可以分为两大类，一种是是基于同步头或者导频的估计算法，另一种是盲估计法。盲估计虽然不需要已知数据，但是其算法需要较大计算量，且收敛速度慢，所以大多数研究都是基于同步头或者导频。 Orthogonal Frequency Division Multiplexing (OFDM) is a multi-carrier modulation method, which converts a serial high-speed bit stream into a low-speed parallel bit stream, and uses a simplified inverse fast Fourier transform (IFFT) to convert the The information symbols are modulated onto each mutually orthogonal sub-channel, so that the wideband channel is transformed into a relatively flat narrowband channel, thereby suppressing intersymbol interference. Because of its ability to resist frequency selective fading channels, and has the advantages of high spectrum efficiency and simple equalization, many communication standards and systems use OFDM modulation, and it is regarded as the key technology of the fourth generation mobile communication. Noise power, as an index to measure channel quality, is an important parameter of OFDM system. At present, there are many literatures on noise power estimation of OFDM systems, which can be mainly divided into two categories, one is the estimation algorithm based on the synchronization header or pilot frequency, and the other is the blind estimation method. Although blind estimation does not require known data, its algorithm requires a large amount of calculation and has a slow convergence speed, so most researches are based on synchronization headers or pilots. the

OFDM同步头，是存在于每个OFDM帧中前面的一个或者几个对发送和接收双方都已知的OFDM符号，它主要用于同步和信道估计，也可以用来估计噪声功率。工程实现时，OFDM符号两边为若干个0，用作保护带，防止相邻频带干扰。而导频则是分布在OFDM数据符号内的某些特定位置的子载波，在发送端这些点上的数据是已知的。导频分为梳状导频和块状导频；而同步头也分为单同步头和多同步头，在单个同步头的情况下，还会分为时域对称和时域不对称情况。 The OFDM synchronization header is one or several OFDM symbols known to both the sender and receiver in each OFDM frame. It is mainly used for synchronization and channel estimation, and can also be used to estimate noise power. When the project is implemented, there are several 0s on both sides of the OFDM symbol, which are used as guard bands to prevent interference from adjacent frequency bands. The pilot frequency is a subcarrier distributed at some specific positions within the OFDM data symbol, and the data at these points are known at the sending end. Pilots are divided into comb pilots and block pilots; and sync headers are also divided into single sync headers and multiple sync headers. In the case of a single sync header, they can also be divided into time domain symmetry and time domain asymmetry. the

在OFDM系统中，发送端的一帧的数据在频域表示如下： In the OFDM system, the data of one frame at the sending end is represented in the frequency domain as follows:

其中，N为OFDM系统子载波的个数，也是FFT、IFFT的点数；M为一帧内含有的OFDM符号数；c_m，n为第m个OFDM符号、第n个子载波上的数据。 Among them, N is the number of subcarriers of the OFDM system, which is also the number of points of FFT and IFFT; M is the number of OFDM symbols contained in a frame; c _m,n is the data on the mth OFDM symbol and the nth subcarrier.

在所加循环前缀足够长和系统完全同步的前提下，信号在进过了多径信道再加上加性高斯白噪声，则接收信号在频域可以表示为： Under the premise that the added cyclic prefix is long enough and the system is fully synchronized, the signal has passed through the multipath channel and added Gaussian white noise, then the received signal can be expressed in the frequency domain as:

Y_m，n＝H_m，nc_m，n+η_m，n Y _{m, n} = H _{m, n} c _{m, n} + η _{m, n}

其中，H_m，n为信道的频域响应，η_m，n为加性高斯白噪声，c_m，n为能量归一化的信号。 Among them, H _{m, n} is the frequency domain response of the channel, η _{m, n} is the additive Gaussian white noise, c _{m, n} is the energy normalized signal.

信道的频域响应，由如下式子得到： The frequency domain response of the channel is obtained by the following formula:

${H h}_{m m,, n no} = = {Σ Σ}_{p p = = 11}^{p p} {h h}_{p p} (({mT mT}_{s the s})) {e e}^{- - j j 22 π π \frac{n no {τ τ}_{p p}}{{NT NT}_{s the s}}}$

其中，h_p(mT_s)为第p条路径的增益，τ_p为路径延迟，T_s为OFDM符号周期。 Among them, h _p (mT _s ) is the gain of the p-th path, τ _p is the path delay, and T _s is the OFDM symbol period.

现有技术中，主要是在Boumard算法和REN算法中提及了如何基于同步头和导频结构来估算OFDM系统的噪声功率。 In the prior art, it is mainly mentioned in the Boumard algorithm and the REN algorithm how to estimate the noise power of the OFDM system based on the synchronization header and the pilot frequency structure. the

Boumard算法在作者是S.Boumard，论文题目是″Novelnoise variance and SNR estimation algorithm for wireless MIMO OFDM systems，″发表于Global Telecommunications Conference，2003.GLOBECOM′03.IEEE，2003，pp.1330-1334vol.3的论文上面，以及作者是任光亮、罗美玲、常义林，论文题目是《OFDM系统信噪比估计新方法》，发表于2007年10月(34卷，第五期)的西安电子科技大学学报(自然科学版)的论文上面，均有详细记载，包括以下几个步骤： The author of the Boumard algorithm is S.Boumard, the title of the paper is "Novelnoise variance and SNR estimation algorithm for wireless MIMO OFDM systems," published in Global Telecommunications Conference, 2003.GLOBECOM'03.IEEE, 2003, pp.1330-1334vol.3 The paper above, and the authors are Ren Guangliang, Luo Meiling, and Chang Yilin. The title of the paper is "A New Method for Signal-to-Noise Ratio Estimation in OFDM Systems", which was published in the Journal of Xidian University ( Natural Science Edition) above papers, all have detailed records, including the following steps:

步骤1：对接收端接收信息的预处理，主要包括以下步骤： Step 1: Preprocessing the information received by the receiving end, mainly includes the following steps:

步骤1-1：对每个OFDM符号去掉序列前端长度为N_cp循环前缀(CP)，得到时域信息，用矩阵表示如下： Step 1-1: For each OFDM symbol, the length of the front end of the sequence is removed as N _cp cyclic prefix (CP), and the time domain information is obtained, which is represented by a matrix as follows:

$X x = = [\begin{matrix} {x x}_{0,0 0,0} & . . & . . & . . & {x x}_{00,, n no} & . . & . . & . . & {x x}_{00,, N N - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {x x}_{m m,, 00} & . . & . . & . . & {x x}_{m m,, n no} & . . & . . & . . & {x x}_{m m,, N N - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {x x}_{M m - - 1,0 1,0} & . . & . . & . . & {x x}_{M m - - 11,, n no} & . . & . . & . . & {x x}_{M m - - 11,, N N - - 11} \end{matrix}]$

步骤1-2：将由步骤1-1得到的时域信息进行N点FFT运算，N为FFT点数，得到的频域信息用矩阵表示如下： Step 1-2: Perform N-point FFT operation on the time domain information obtained in step 1-1, where N is the number of FFT points, and the obtained frequency domain information is expressed in a matrix as follows:

步骤1-3：从步骤1-2得到的频域信息中，取出同步头内的信息，分为以下2个步骤： Step 1-3: From the frequency domain information obtained in step 1-2, take out the information in the synchronization header, which is divided into the following two steps:

步骤1-3-1：从步骤1-2得到的矩阵中，取出属于同步头的N_p个OFDM符号： Step 1-3-1: From the matrix obtained in step 1-2, take out N _p OFDM symbols belonging to the synchronization header:

步骤1-3-2：分别从步骤1-3-1的N_p行序列中，去掉左边保护带N_b1个点、右边保护带N_b2个点和Y_i，N/2(0频点)，到一组新的序列，表示为如下矩阵形式： Step 1-3-2: Remove N _b1 points in the left guard band, N _b2 points in the right guard band and Y _{i, N/2} (0 frequency point) from the N _p row sequence in step 1-3-1 respectively , to a new set of sequences expressed in the following matrix form:

$Z Z = = [\begin{matrix} {y the y}_{{i i}_{00},, {N N}_{b b 11}} & . . & . . & . . & {y the y}_{{i i}_{00},, N N,, 22 - - 11} & {y the y}_{{i i}_{00},, N N,, 22 + + 11} & . . & . . & . . & {y the y}_{{i i}_{00},, N N - - {N N}_{b b 22}} \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ {y the y}_{{i i}_{m m},, {N N}_{b b 11}} & {y the y}_{{i i}_{m m},, N N / / 22 - - 11} & {y the y}_{{i i}_{m m},, N N / / 22 + + 11} & {y the y}_{{i i}_{m m},, N N - - {N N}_{b b 22}} \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ {y the y}_{{i i}_{{N N}_{p p} - - 11,, {N N}_{b b 11}}} & . . & . . & . . & {y the y}_{{i i}_{{N N}_{p p} - - 11,, N N,, 22 - - 11}} & {y the y}_{{i i}_{{N N}_{p p} - - 11,, N N,, 22 + + 11}} & . . & . . & . . & {y the y}_{{i i}_{{N N}_{p p} - - 11,, N N - - {N N}_{b b 22}}} \end{matrix}] = = \begin{matrix} \end{matrix} [\begin{matrix} {z z}_{0,0 0,0} & . . & . . & . . & {z z}_{00,, j j} & . . & . . & . . & {z z}_{00,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {z z}_{i i,, 00} & . . & . . & . . & {z z}_{i i,, j j} & . . & . . & . . & {z z}_{i i,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {z z}_{{N N}_{p p} - - 1,0 1,0} & . . & . . & . . & {z z}_{{N N}_{p p} - - 11,, j j} & . . & . . & . . & {z z}_{{N N}_{p p} - - 11,, J J - - 11} \end{matrix}]$

等式右边Z矩阵为新定义符号，J为每个同步头有用数据的长度， J＝N-N_b1-N_b2-1。 The Z matrix on the right side of the equation is the newly defined symbol, J is the length of the useful data of each sync header, J=NN _b1 -N _b2 -1.

步骤1-4：根据步骤1-3的取样方式，从已知的发送端数据取出同步头中的有用序列，表示如下矩阵： Step 1-4: According to the sampling method of step 1-3, take out the useful sequence in the synchronization header from the known sender data, and express the following matrix:

$D D. = = [\begin{matrix} {d d}_{0,0 0,0} & . . & . . & . . & {d d}_{00,, j j} & . . & . . & . . & {d d}_{00,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {d d}_{i i,, 00} & . . & . . & . . & {d d}_{i i,, j j} & . . & . . & . . & {d d}_{i i,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {d d}_{{N N}_{p p} - - 1,0 1,0} & . . & . . & . . & {d d}_{{N N}_{p p} - - 11,, j j} & . . & . . & . . & {d d}_{{N N}_{p p} - - 11,, J J - - 11} \end{matrix}] = = \begin{matrix} \end{matrix} [\begin{matrix} {c c}_{{i i}_{00},, {N N}_{b b 11}} & . . & . . & . . & {c c}_{{i i}_{00},, N N,, 22 - - 11} & {c c}_{{i i}_{00},, N N,, 22 + + 11} & . . & . . & . . & {c c}_{{i i}_{00},, N N - - {N N}_{B B 22}} \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ {c c}_{{i i}_{m m},, {N N}_{b b 11}} & . . & . . & . . & {c c}_{{i i}_{m m},, N N / / 22 - - 11} & {c c}_{{i i}_{m m},, N N / / 22 + + 11} & . . & . . & . . & {c c}_{{i i}_{m m},, N N - - {N N}_{b b 22}} \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ . . & . . & . . & . . \\ {c c}_{{i i}_{{N N}_{p p} - - 11,,} {N N}_{b b 11}} & . . & . . & . . & {c c}_{{i i}_{{N N}_{p p} - - 11,,},, N N,, 22 - - 11} & {c c}_{{i i}_{{N N}_{p p} - - 11,,},, N N,, 22 - - 11} & . . & . . & . . & {c c}_{{i i}_{{N N}_{p p} - - 11,,} N N - - {N N}_{b b 22}} \end{matrix}]$

D矩阵为新定义符号，表示发送端的有用数据。 The D matrix is a newly defined symbol, which represents useful data at the sending end. the

步骤2：通过步骤1得到的2个新序列，计算系统的噪声功率： Step 2: Calculate the noise power of the system through the two new sequences obtained in step 1:

$\overset{^^}{W W} = = \frac{11}{44 {N N}_{p p} ((J J - - 11))} {Σ Σ}_{i i = = 00}^{{N N}_{p p} - - 11} {Σ Σ}_{j j = = 00}^{J J - - 11} {| | {z z}_{i i,, j j - - 11} {d d}_{i i,, j j} - - {z z}_{i i,, j j} {d d}_{i i,, j j - - 11} | |}^{22}$

Boumard算法的上述噪声功率计算公式，运用了假设条件：相邻子载波信道频率响应近似相等，即H_m，N-1≈H_m，n，它成立的前提是时延扩展较小。但是很多无线信道的多径数多，时延扩展较大，这时不能得到H_m，n-1≈H_m，n，这会导致Boumard算法性能急剧下降。 The above noise power calculation formula of the Boumard algorithm uses the assumption that the channel frequency responses of adjacent subcarriers are approximately equal, that is, H _{m, N-1} ≈ _{H m, n} , which is established on the premise that the delay spread is small. However, many wireless channels have a large number of multipaths and a large delay spread. At this time, H _m,n-1 ≈H _m,n cannot be obtained, which will lead to a sharp decline in the performance of the Boumard algorithm.

REN算法在作者是任光亮、常义林，论文题目是″SNR estimation algorithm based on the preamble for OFDM systems in frequency selective channels，"，发表于Communications，IEEE Transactions on，Vol.57，pp.2230-2234，2009.的论文上面有详细记载，包括以下几个步骤： The authors of the REN algorithm are Ren Guangliang and Chang Yilin. The title of the paper is "SNR estimation algorithm based on the preamble for OFDM systems in frequency selective channels," published in Communications, IEEE Transactions on, Vol.57, pp.2230-2234 , 2009. The paper has detailed records above, including the following steps:

步骤1：对接收端接收信息的预处理，与前述的Boumard算法步骤1相同； Step 1: The preprocessing of the information received by the receiving end is the same as the aforementioned step 1 of the Boumard algorithm;

步骤2：从步骤1得到的两个新矩阵Z和D中，选取其中第一行与第二行的数据，即令i＝0、1，计算系统的噪声功率： Step 2: From the two new matrices Z and D obtained in step 1, select the data of the first row and the second row, that is, set i=0, 1, and calculate the noise power of the system:

其中，当d_i，j≠0时，；当d_i，j＝0时，表示对d_i，j取共轭。 Among them, when d _i,j ≠0, ; When d _{i, j} = 0, Indicates to take the conjugate of d _{i, j} .

REN算法的运用有两个限制条件，其一，它需要用到两个结构相同的同步头，因为它需要用不同周期的相同子载波上的Y′来相减，即，以消去信道频率响应；其二，不同OFDM符号相同子载波处的信道频率响应要近似相等，即H_m，n≈H_m+，ln≈H_n。但是，在实际应用时两个同步头的结构可能不相同；此外，REN算法有两个不足之处：其一，随着信道最大多普勒频率的增加，信道的时间选择性衰落变大，不同OFDM符号相同子载波的信道系数会发生变化，这时他们不能近似相等；其二，如果两个同步头间隔其他的OFDM符号，则它们的信道系数之间仍然会存在较大差异，也会影响REN算法的性能。 The application of the REN algorithm has two restrictions. First, it needs to use two synchronization headers with the same structure, because it needs to subtract Y′ on the same subcarrier with different periods, that is , to eliminate the channel frequency response; secondly, the channel frequency response at the same subcarrier of different OFDM symbols should be approximately equal, that is, H _{m, n} ≈ _{H m+, ln} ≈ H _n . However, the structures of the two sync heads may be different in practical applications; in addition, the REN algorithm has two shortcomings: first, with the increase of the maximum Doppler frequency of the channel, the time-selective fading of the channel becomes larger, The channel coefficients of the same subcarrier of different OFDM symbols will change, and they cannot be approximately equal at this time; secondly, if two synchronization headers are separated by other OFDM symbols, there will still be a large difference between their channel coefficients, and there will be Affect the performance of the REN algorithm.

我们经过大量实验发现：不同OFDM符号间隔相同子载波的两个信道频率响应的差值较为接近，而且相邻OFDM符号的相邻子载波的差值最小，即(H_m，n-H_m，n+1)-(H_m+1，n-H_m+1，n+1)≈0，这样我们就可以用同一个或者相邻两个OFDM符号的相邻子载波这样4个点来抵消子载波信道频率响应的误差值，它所带来的误差更小。 After a lot of experiments, we found that the difference between the two channel frequency responses of the same subcarriers with different OFDM symbol intervals is relatively close, and the difference between adjacent subcarriers of adjacent OFDM symbols is the smallest, that is, (H _{m, n} -H _{m, n+1} )-(H _{m+1, n} -H _{m+1, n+1} )≈0, so we can use the same or adjacent subcarriers of two adjacent OFDM symbols to offset 4 points The error value of the subcarrier channel frequency response, which brings smaller errors.

发明内容 Contents of the invention

为了解决现有的Boumard算法和REN算法存在的上述技术问题，本发明提出的一种用于估算OFDM信号的噪声功率的方法，其特征是包括如下步骤： In order to solve the above-mentioned technical problems that existing Boumard algorithm and REN algorithm exist, a kind of method for estimating the noise power of OFDM signal that the present invention proposes is characterized in comprising the following steps:

步骤1：设定系统参数，包括：根据用户的选择决定采用下面四种信息之一来估算噪声功率：导频、多个同步头、单个对称同步头、单个非对称同步头；设定左边保护带点数N_b1，右边保护带点数N_b2，循环前缀长度N_cp，同步头个数N_p，导频符号个数N_pl，发送端的一帧数据在频域的表示为矩阵： Step 1: Set system parameters, including: According to the user's choice, one of the following four information is used to estimate the noise power: pilot frequency, multiple synchronization heads, single symmetrical synchronization head, and single asymmetric synchronization head; set the left guard The number of band points N _b1 , the number of right guard band points N _b2 , the length of the cyclic prefix N _cp , the number of synchronization headers N _p , the number of pilot symbols N _pl , and a frame of data at the sending end in the frequency domain are expressed as a matrix:

，N为OFDM系统子载波的个数，M为一帧内含有的OFDM符号数； , N is the number of OFDM system subcarriers, and M is the number of OFDM symbols contained in a frame;

步骤2：对接收端接收到的信息进行预处理，包括： Step 2: Preprocess the information received by the receiver, including:

步骤2-1：将接收到的OFDM符号去掉序列前端长度为N_cp的循环前缀(CP)，得到的时域信息，用矩阵表示如下： Step 2-1: Remove the cyclic prefix (CP) whose sequence front length is N _cp from the received OFDM symbol, and obtain the time domain information, expressed as follows with a matrix:

步骤2-2：将步骤2-1的时域信号X转换为频域信号：如果采用单个对称同步头来估算噪声功率，则提取出步骤2-1得到的矩阵X的第一行数据，将它分为前后两个部分，即：[x_0，0…x_0，N2-1]和[x_0，N/2…x_0，N-1]，分别对他们进行N/2点的能量归一化的FFT，得到矩阵： Step 2-2: Convert the time-domain signal X in step 2-1 into a frequency-domain signal: If a single symmetrical sync head is used to estimate the noise power, then extract the first row of data in the matrix X obtained in step 2-1, and convert It is divided into two parts, namely: [x _{0, 0} ... x _{0, N2-1} ] and [x _{0, N/2} ... x _{0, N-1} ], respectively perform N/2 points of energy on them Normalized FFT, resulting in matrix:

否则，对步骤2-1得到的矩阵X每一行都进行N点的能量归一化FFT，得到矩阵： Otherwise, perform N-point energy normalized FFT on each row of the matrix X obtained in step 2-1 to obtain the matrix:

步骤3：从接收端提取出有用数据： Step 3: Extract useful data from the receiving end:

(1)如果采用导频来估算噪声功率，则根据该OFDM系统规范的定义，从定义的结构中找出在一帧中哪些OFDM符号具有导频信息，并提取出这些符号；再从他们中，确定导频子载波的位置，将该子载波处的接收信号数据提取出来，组成一个新的导频矩阵： (1) If the pilot frequency is used to estimate the noise power, then according to the definition of the OFDM system specification, find out which OFDM symbols in a frame have pilot information from the defined structure, and extract these symbols; then from them , determine the position of the pilot subcarrier, extract the received signal data at the subcarrier, and form a new pilot matrix:

其中，N_pl为一帧内OFDM导频符号的个数，J为一个导频符号内导频的点数； Wherein, N _pl is the number of OFDM pilot symbols in a frame, and J is the number of pilot points in a pilot symbol;

(2)如果采用多个同步头来估算噪声功率，则对步骤2得到的矩阵Y中的N_p个同步头序列所在的每一行，都去掉左边保护带的N_b1个点、右边保护带的N_b2个点、0频点和对应发送数据为0的子载波上的点，得到有用数据组成的矩阵： (2) If multiple sync heads are used to estimate the noise power, for each row where the N _p sync head sequences in the matrix Y obtained in step 2 are located, the N _b1 points of the left guard band and the right guard band are removed. N _b2 points, 0 frequency point and the point on the subcarrier corresponding to the transmitted data is 0, and the matrix composed of useful data is obtained:

$Z Z = = [\begin{matrix} {y the y}_{{i i}_{00},, {j j}_{00}} & . . & . . & . . & {y the y}_{{i i}_{00},, {j j}_{k k}} & . . & . . & . . & {y the y}_{{i i}_{00},, {j j}_{J J - - 11}} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {y the y}_{{i i}_{m m},, {j j}_{00}} & . . & . . & . . & {y the y}_{{i i}_{m m},, {j j}_{k k}} & . . & . . & . . & {y the y}_{{i i}_{m m},, {j j}_{J J - - 11}} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {y the y}_{{i i}_{{N N}_{p p} - - 11,,},, {j j}_{00}} & . . & . . & . . & {y the y}_{{i i}_{{N N}_{p p} - - 11,, {j j}_{k k}}} & . . & . . & . . & {y the y}_{{i i}_{{N N}_{p p} - - 11,, {j j}_{J J - - 11}}} \end{matrix}] = = [\begin{matrix} {z z}_{00,, 00} & . . & . . & . . & {z z}_{00,, j j} & . . & . . & . . & {z z}_{00,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {z z}_{i i,, 00} & . . & . . & . . & {z z}_{00,, j j} & . . & . . & . . & {z z}_{i i,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {Z Z}_{{N N}_{p p} - - 1,0 1,0} & . . & . . & . . & {z z}_{{N N}_{p p} - - 11,, j j} & . . & . . & . . & {z z}_{{N N}_{p p} - - 11,, J J - - 11} \end{matrix}]$

其中J为同步头符号内发送数据不为0的点的个数，j_k的取值范围在N_b1到N/2-1之间，或者N/2+1到N-N_b2； Where J is the number of points where the data sent in the sync header symbol is not 0, and the value range of j _k is between N _b1 to N/2-1, or N/2+1 to NN _b2 ;

(3)如果采用单个非对称同步头来估算噪声功率，则对步骤2得到的矩阵Y中第一行的同步头序列，去掉左边保护带N_b1个点、右边保护带N_b2个点、0频点和对应发送数据为0的子载波上的点，得到有用数据矩阵 (3) If a single asymmetric sync header is used to estimate the noise power, then for the sync header sequence in the first row of the matrix Y obtained in step 2, remove N _b1 points in the left guard band, N _b2 points in the right guard band, and 0 The frequency point and the point on the subcarrier corresponding to the transmitted data is 0, and the useful data matrix is obtained

(4)如果采用单个对称同步头来估算噪声功率，则有用数据组成的矩阵 (4) If a single symmetrical synchronous head is used to estimate the noise power, the matrix composed of useful data

J＝N/2； J=N/2;

步骤4：从发送端提取出有用数据： Step 4: Extract useful data from the sender:

(1)如果采用导频来估算噪声功率，则在发送端频域信号矩阵C中，先确定哪些OFDM符号具有导频信息，提取出C中相应的行序列；再从这些行序列中，确定导频子载波的位置，将这些子载波处的发送信息提取出来，组成一个新的导频矩阵： (1) If pilots are used to estimate the noise power, in the frequency domain signal matrix C at the sending end, first determine which OFDM symbols have pilot information, and extract the corresponding row sequences in C; then from these row sequences, determine The positions of the pilot subcarriers, the transmission information at these subcarriers are extracted to form a new pilot matrix:

$D D. = = [\begin{matrix} {d d}_{0,0 0,0} & . . & . . & . . & {d d}_{00,, j j} & . . & . . & . . & {d d}_{00,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {d d}_{i i,, 00} & . . & . . & . . & {d d}_{i i,, j j} & . . & . . & . . & {d d}_{i i,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {d d}_{{N N}_{pl pl} - - 1,0 1,0} & . . & . . & . . & {d d}_{{N N}_{pl pl} - - 11,, j j} & . . & . . & . . & {d d}_{{N N}_{pl pl} - - 11,, J J - - 11} \end{matrix}]$

J为一个符号内导频的点数； J is the number of pilot points in a symbol;

(2)如果采用多个同步头来估算噪声功率，则在矩阵C的N_p个同步头序列所在行中，去掉左边保护带N_b1个点、右边保护带N_b2个点、0频点和数据为0的子载波上的点，得到有用数据组成的矩阵： (2) If multiple sync heads are used to estimate the noise power, in the row where the N _p sync head sequences of the matrix C are located, N _b1 points in the left guard band, N _b2 points in the right guard band, 0 frequency points and Points on the subcarrier whose data is 0, get a matrix composed of useful data:

$D D. = = [\begin{matrix} {d d}_{0,0 0,0} & . . & . . & . . & {d d}_{00,, j j} & . . & . . & . . & {d d}_{00,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {d d}_{i i,, 00} & . . & . . & . . & {d d}_{i i,, j j} & . . & . . & . . & {d d}_{i i,, J J - - 11} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {d d}_{{N N}_{p p} - - 1,0 1,0} & . . & . . & . . & {d d}_{{N N}_{p p} - - 11,, j j} & . . & . . & . . & {d d}_{{N N}_{p p} - - 11,, J J - - 11} \end{matrix}] = = [\begin{matrix} {c c}_{{i i}_{00},, {j j}_{00}} & . . & . . & . . & {c c}_{{i i}_{00},, {j j}_{k k}} & . . & . . & . . & {c c}_{{i i}_{00},, {j j}_{J J - - 11}} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {c c}_{{i i}_{m m},, {j j}_{00}} & . . & . . & . . & {c c}_{{i i}_{m m},, {j j}_{k k}} & . . & . . & . . & {c c}_{{i i}_{m m},, {j j}_{J J - - 11}} \\ . . & . . & . . \\ . . & . . & . . \\ . . & . . & . . \\ {c c}_{{i i}_{{N N}_{p p} - - 11},, {j j}_{00}} & . . & . . & . . & {c c}_{{i i}_{{N N}_{F f} - - 11,,} - - 22,, JK JK} & . . & . . & . . & {c c}_{{i i}_{{N N}_{F f} - - 11,,},, JJ JJ - - 11} \end{matrix}]$

其中，J为同步头符号内发送数据不为0的点的个数，j_k的取值范围在N_b1到N/2-1之间，或者N/2+1到N-N_b2； Among them, J is the number of points where the data sent in the sync header symbol is not 0, and the value range of j _k is between N _b1 to N/2-1, or N/2+1 to NN _b2 ;

(3)如果采用单个非对称同步头来估算噪声功率，则将矩阵C中第一行数据，去掉左边保护带N_b1个点、右边保护带N_b2个点、0频点和对应发送数据为0的子载波上的点，得到有用数据组成的矩阵： (3) If a single asymmetric sync head is used to estimate the noise power, then the first row of data in the matrix C, remove the left guard band N _b1 points, the right guard band N _b2 points, 0 frequency point and the corresponding transmission data is Points on the subcarrier of 0, get a matrix composed of useful data:

$D D. = = [\begin{matrix} {d d}_{0,0 0,0} & . . & . . & . . & {d d}_{00,, j j} & . . & . . & . . & {d d}_{00,, J J - - 11} \end{matrix}] = = [\begin{matrix} {c c}_{00,, {j j}_{00}} & . . & . . & . . & {c c}_{00,, {j j}_{k k}} & . . & . . & . . & {c c}_{00,, {j j}_{J J - - 11}} \end{matrix}]$

J为同步头符号内发送数据不为0的点的个数，j_k的取值范围在N_b1到N/2-1之间，或者N/2+1到N-N_b2； J is the number of points where the data sent in the sync header symbol is not 0, and the value range of j _k is between N _b1 and N/2-1, or N/2+1 to NN _b2 ;

(4)如果采用单个对称同步头来估算噪声功率，则： (4) If a single symmetrical sync head is used to estimate the noise power, then:

其中，J＝N/2； Among them, J=N/2;

步骤5：利用前面两步得到的发送端和接收端的有用序列计算系统的噪声功率：如果采用单个非对称同步头来估算噪声功率，则噪声功率 Step 5: Calculate the noise power of the system using the useful sequences of the transmitter and receiver obtained in the previous two steps: If a single asymmetric synchronization head is used to estimate the noise power, then the noise power

；否则，分别从矩阵Z和D中选出任意两行用于计算噪声功率 ; Otherwise, select any two rows from matrices Z and D to calculate the noise power

，其中，i，i′表示矩阵Z和D的任意两行的行号；d^*表示对d取共轭。 , where i, i' represent the row numbers of any two rows of matrices Z and D; d ^* represents taking the conjugate of d.

所述能量归一化的FFT，是指在原有FFT的表达式下除以常数因子，其中N为FFT点数。 The energy-normalized FFT refers to dividing by a constant factor under the expression of the original FFT , where N is the number of FFT points.

所述步骤3分别从矩阵Z和D中选出的任意两行，是矩阵中的相邻两行。 The arbitrary two rows respectively selected from the matrices Z and D in the step 3 are two adjacent rows in the matrix. the

从上面的描述可以看出，本发明提出的技术方案，用一个或者两个相邻的OFDM符号的相邻子载波这样四个点来抵消子载波信道频率响应的误差值，这就使得本噪声功率估计方法具有良好的抗频率选择性衰落和时间选择性衰落的特性。同时，本方法既可以采用导频来估算噪声功率，也可以采用同步头来估算噪声功率，用同步头来估算噪声功率时，对于多个同步头、单个对称同步头和单个不对称同步头的情况都能适用，具有良好的适应性。 As can be seen from the above description, the technical solution proposed by the present invention uses four points such as adjacent subcarriers of one or two adjacent OFDM symbols to offset the error value of the subcarrier channel frequency response, which makes the noise The power estimation method has good characteristics of resisting frequency selective fading and time selective fading. At the same time, this method can not only use the pilot frequency to estimate the noise power, but also use the synchronization head to estimate the noise power. When using the synchronization head to estimate the noise power, for multiple synchronization heads, a single symmetrical synchronization head and a single asymmetric synchronization head It can be applied to any situation and has good adaptability. the

附图说明 Description of drawings

附图1为本发明提出的一种用于估算OFDM信号的噪声功率的方法的总体流程图； Accompanying drawing 1 is the overall flowchart of a kind of method for estimating the noise power of OFDM signal that the present invention proposes;

附图2为本发明具体实施例一用到的导频结构图； Accompanying drawing 2 is the pilot frequency structure figure that specific embodiment one of the present invention is used;

附图3为本发明具体实施例一用到的系统帧结构图； Accompanying drawing 3 is the system frame structure diagram that specific embodiment one of the present invention uses;

附图4为本发明具体实施例二用到的同步头结构图； Accompanying drawing 4 is the synchronous head structural diagram that specific embodiment two of the present invention is used;

附图5为本发明具体实施例三和实施例四用到的同步头结构图； Accompanying drawing 5 is the synchronous head structural diagram that embodiment three and embodiment four of the present invention are used;

附图6为本发明具体实施例四用到的同步头有用序列结构图。 Accompanying drawing 6 is the useful sequence structure diagram of the sync header used in Embodiment 4 of the present invention. the

具体实施方式 Detailed ways

下面具体说明本发明技术方案的实施方式。 The implementation manner of the technical solution of the present invention will be described in detail below. the

所有实施例都选用MBSFN(Multicast Broadcast Single Frequency Network)系统，即组播广播单频网络。MBSFN系统下行链路中，移动终端接收到各个基站发来的数据是相同的，类似于用户接收到从一个基站发来的经过多径衰落的数据，该信道为典型的频率选择性衰落信道。 All embodiments select the MBSFN (Multicast Broadcast Single Frequency Network) system, that is, the multicast broadcast single frequency network. In the downlink of the MBSFN system, the mobile terminal receives the same data from each base station, which is similar to the user receiving multipath fading data from a base station. This channel is a typical frequency selective fading channel. the

下面各个实施例中提及的能量归一化的FFT，是指在原有FFT的表达式下除以常数因子，其中N为FFT点数。 The energy-normalized FFT mentioned in the following embodiments refers to dividing by a constant factor under the expression of the original FFT , where N is the number of FFT points.

实施例一：用导频信息估算噪声功率 Embodiment 1: Estimating Noise Power Using Pilot Information

本实施例以3GPP标准的导频结构为例，该导频结构采用块状导频，每个OFDM导频符号间隔3个OFDM数据符号，第二个OFDM导频符号与1，3错开1位子载波，导频采用QPSK调制。该导频结构及系统帧结构如图2和图3所示。此外，本实施例假定OFDM子载波数N为512，循环前缀长度N_cp为N/4，有用数据的子载波数为300，左边保护带个数N_b1＝106，右边保护带个数N_b2＝105，0频点为0，发送端的一帧数据在频域的表示为矩阵 This embodiment takes the pilot structure of the 3GPP standard as an example. The pilot structure adopts block pilots. Each OFDM pilot symbol is spaced by 3 OFDM data symbols, and the second OFDM pilot symbol is staggered from 1, 3 by 1 bit. Carrier and pilot are modulated by QPSK. The pilot frequency structure and the system frame structure are shown in Fig. 2 and Fig. 3 . In addition, this embodiment assumes that the number N of OFDM subcarriers is 512, the length of the cyclic prefix _Ncp is N/4, the number of subcarriers of useful data is 300, the number of left guard bands N _b1 =106, and the number of right guard bands N _b2 =105, 0 frequency point is 0, a frame of data at the sending end is represented as a matrix in the frequency domain

步骤1：将接收到的OFDM符号去掉序列前端长度为N_cp的循环前缀(CP)，然后做N点能量归一化的FFT运算，转换为频域信号； Step 1: Remove the cyclic prefix (CP) that the sequence front length is N _cp from the received OFDM symbol, then do the FFT operation of N point energy normalization, and convert it into a frequency domain signal;

步骤2：从接收数据中取出导频数据：按照附图2所示的导频结构图和附图3所示的系统帧结构图，对于步骤1得到的数据，先去掉左右保护带和0频点，再取出导频点上的数据：对于偶数块，取出第3个OFDM符号，在此符号中取出奇数点数据；对于奇数块，只取第1个OFDM符号，在此符号中取出偶数点数据。组成新矩阵： Step 2: Take out the pilot data from the received data: According to the pilot structure diagram shown in Figure 2 and the system frame structure diagram shown in Figure 3, for the data obtained in Step 1, first remove the left and right guard bands and the zero frequency point, and then take out the data on the pilot point: for an even block, take out the third OFDM symbol, and take out the odd point data in this symbol; for an odd block, take only the first OFDM symbol, and take out the even point in this symbol data. Form a new matrix:

再按上述方式，从发送端的数据矩阵C中取出对应导频点的数据，组成新的矩阵： According to the above method, the data corresponding to the pilot point is taken out from the data matrix C of the sending end to form a new matrix:

其中，J＝150。 where J=150. the

步骤3：利用步骤2得到的发送端和接收端的导频点矩阵，计算系统噪声功率：，其中d^*表示对d取共轭，J＝150。 Step 3: Calculate the system noise power by using the pilot point matrix of the transmitting end and the receiving end obtained in step 2: , where d ^* means taking the conjugate of d, J=150.

需要说明的是，任何的OFDM系统的导频结构和格式，对于本领域技术人员而言，是可以通过公开的技术规范查询到的信息。本实施例以3GPP标准的导频结构为例子，说明如何提取出导频中的数据，对于其他的采用了导频的OFDM系统，在提取导频中的数据时，只要根据相应的技术规范的定义，找出导频数据，并去除其中的保护带和0频点(如果有的话)数据，就可以组成本实施例中的矩阵Z。 It should be noted that, for those skilled in the art, the pilot structure and format of any OFDM system can be found through public technical specifications. This embodiment takes the pilot structure of the 3GPP standard as an example to illustrate how to extract the data in the pilot. For other OFDM systems that use the pilot, when extracting the data in the pilot, as long as the corresponding technical specification Define, find out the pilot data, and remove the guard band and 0 frequency point (if any) data, the matrix Z in this embodiment can be formed. the

实施例二：用两个同步头信息估算噪声功率。 Embodiment 2: Estimating noise power by using two synchronization header information. the

本实施例中，采用两个相同的同步头，同步头符号间隔一个OFDM数据符号， OFDM子载波数N为128，循环前缀长度N_cp为N/4，有数据的子载波数为116，为BPSK信号和0间隔的形式，左边保护带个数N_b1＝6，右边保护带个数N_b2＝5，0频点为0。同步头整体结构如图4所示。假定图4中同步头左边数据 In this embodiment, two identical sync headers are used, the sync header symbols are separated by one OFDM data symbol, the number of OFDM subcarriers N is 128, the length of the cyclic prefix _Ncp is N/4, and the number of subcarriers with data is 116, which is In the form of BPSK signal and 0 interval, the number of guard bands on the left is N _b1 =6, the number of guard bands on the right is N _b2 =5, and the 0 frequency point is 0. The overall structure of the synchronization head is shown in Figure 4. Assume that the data on the left side of the sync header in Figure 4

A＝[1，0，1，0，-1，0，-1，0，1，0，1，0，-1，0，-1，0，1，0，1，0，1，0，1，0，-1，0，-1，0，-1，0，-1，0，1，0，-1，0，-1，0，1，0，-1，0，-1，0，-1，0，-10，-1，0，-1，0，-1，0，-1，0，-1，0]； A=[1,0,1,0,-1,0,-1,0,1,0,1,0,-1,0,-1,0,1,0,1,0,1,0 ,1,0,-1,0,-1,0,-1,0,-1,0,1,0,-1,0,-1,0,1,0,-1,0,-1 ,0,-1,0,-10,-1,0,-1,0,-1,0,-1,0,-1,0];

同步头右边数据 Synchronize the data on the right side of the head

B=[0，-1，0，-1，0，-1，0，1，0，-1，0，1，0，-1，0，-1，0，-1，0，-1，0，-1，0，-1，0，-1，0，1，0，1，0，-1，0，1，0，1，0，-1，0，-1，0，1，0，-1，0，-1，0，-1，0，-1，0，-1，0，-1，0，-1，0，-1]。 B=[0, -1, 0, -1, 0, -1, 0, 1, 0, -1, 0, 1, 0, -1, 0, -1, 0, -1, 0, -1 ,0,-1,0,-1,0,-1,0,1,0,1,0,-1,0,1,0,1,0,-1,0,-1,0,1 ,0,-1,0,-1,0,-1,0,-1,0,-1,0,-1,0,-1,0,-1]. the

步骤1：将接收到的OFDM符号去掉序列前端长度为N_cp的循环前缀(CP)，然后做N点能量归一化FFT运算，将接收到的时域信号转换为频域信号； Step 1: Remove the cyclic prefix (CP) with a sequence front length of N _cp from the received OFDM symbol, and then perform N-point energy normalized FFT operations to convert the received time-domain signal into a frequency-domain signal;

步骤2：从接收数据中取出同步头中的有用数据：按照图4所示的同步头结构，从步骤1得到的数据中，先取第一行和第三行的序列，这两个就是OFDM同步头；再去掉左右保护带和0频点，即取出接收端数据中对应图4中A和B位置的数据；之后，取出A中的奇数点和B中的偶数点，即A与B中BPSK数据所在点对应的接收端频域数据组成新矩阵： Step 2: Take out the useful data in the synchronization header from the received data: According to the synchronization header structure shown in Figure 4, from the data obtained in step 1, first take the sequence of the first row and the third row, these two are OFDM synchronization Then remove the left and right guard bands and 0 frequency points, that is, take out the data corresponding to the positions of A and B in Figure 4 in the receiving end data; after that, take out the odd points in A and the even points in B, that is, BPSK in A and B The frequency domain data of the receiving end corresponding to the data point forms a new matrix:

其中，J＝58。 Among them, J=58. the

从发送端数据中取出有用数据： Get useful data from the sender data:

矩阵D的第一行由A中BPSK数据和B中BPSK数据组成，第二行序列和第一行相同，记为： The first row of matrix D is composed of BPSK data in A and BPSK data in B, and the sequence of the second row is the same as the first row, recorded as:

A₁＝[1，1，-1，-1，1，1，-1，-1，1，1，1，1，-1，-1，-1，-1，1，-1，-1，1，-1，-1，-1，-1-1，-1，-1，-1，-1]； A ₁ =[1, 1, -1, -1, 1, 1, -1, -1, 1, 1, 1, 1, -1, -1, -1, -1, 1, -1, - 1,1,-1,-1,-1,-1-1,-1,-1,-1,-1];

B₁＝[1，-1，-1，1，-1，1，-1，-1，-1，-1，-1，-1，-1，1，1，-1，1，1，-1，-1，1，-1，-1，-1，-1，-1，-1，-1，-1]。 B ₁ =[1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, 1, -1, 1, 1 ,-1,-1,1,-1,-1,-1,-1,-1,-1,-1,-1].

步骤3：利用步骤2得到的发送端和接收端的到频点矩阵，计算系统噪声功率：，其中d^*表示对d取共轭，J＝58。 Step 3: Calculate the system noise power by using the arrival-frequency matrix of the transmitter and receiver obtained in Step 2: , where d ^* means taking the conjugate of d, J=58.

实施例三：用单个对称同步头信息估算噪声功率。 Embodiment 3: Estimate noise power by using single symmetric synchronization header information. the

实施例二所用同步头为一种对称的同步头，实施例三只采用一个同步头结构其他条件和实施例二相同。同步头结构如图5所示。 The synchronous head used in the second embodiment is a symmetrical synchronous head, and only one synchronous head structure is used in the third embodiment, and the other conditions are the same as those in the second embodiment. The synchronous head structure is shown in Fig. 5. the

OFDM子载波数N为128，循环前缀长度N_cp为N/4，有数据的子载波数为116，为BPSK信号和0间隔的形式，左边保护带个数N_b1＝6，右边保护带个数N_b2＝5，0频点为0。 The number of OFDM subcarriers N is 128, the length of the cyclic prefix N _cp is N/4, the number of subcarriers with data is 116, in the form of BPSK signals and 0 intervals, the number of guard bands on the left is N _b1 = 6, and the number of guard bands on the right is 1 Number N _b2 =5, 0 frequency point is 0.

步骤1：将接收到的信号去循环前缀，得到长度为N的序列X，再将它分为[x_0，0…x_0，N2-1]和[x_0，N/2…x_0，N-1]两个部分，都进行N/2点能量归一化FFT运算，分别作为Z矩阵的第一行和第二行，得到矩阵Y： Step 1: Remove the cyclic prefix from the received signal to obtain a sequence X of length N, and then divide it into [x _{0, 0} ... x _{0, N2-1} ] and [x _{0, N/2} ... x _{0, N-1} ] two parts, N/2 point energy normalized FFT operation is performed, respectively as the first row and second row of the Z matrix, and the matrix Y is obtained:

步骤2：直接将Y矩阵变换符号，统一变成Z矩阵，即： Step 2: Directly transform the symbol of the Y matrix into a Z matrix, namely:

而矩阵D为全1矩阵 And the matrix D is all 1 matrix

$D D. = = [\begin{matrix} {d d}_{0,0 0,0} & . . & . . & . . & {d d}_{00,, j j} & . . & . . & . . & {d d}_{00,, J J - - 11} \\ {d d}_{1,0 1,0} & . . & . . & . . & {d d}_{11,, J J} & . . & . . & . . & {d d}_{11,, J J - - 11} \end{matrix}] = = [\begin{matrix} 11 & . . & . . & . . & 11 \\ 11 & . . & . . & . . & 11 \end{matrix}]$

其中，J＝N/2＝64。 Wherein, J=N/2=64. the

步骤3：利用步骤2得到的发送端和接收端的矩阵，计算系统噪声功率：，其中d^*表示对d取共轭，J＝64。 Step 3: Using the matrices of the transmitter and receiver obtained in Step 2, calculate the system noise power: , where d ^* means taking the conjugate of d, J=64.

实施例四：用单个非对称同步头信息估算噪声功率。 Embodiment 4: Estimating noise power by using single asymmetric synchronization header information. the

CAZAC是一种非对称的同步头，本实施例就以CAZAC序列为例来说明。组帧格式采用图5结构，OFDM子载波数N为128，循环前缀长度N_cp为N/4，有用数据的子载波数为116，为BPSK信号和0间隔的形式，左边保护带个数N_b1＝6，右边保护带个数N_b2＝5，0频点为0。 CAZAC is an asymmetric synchronization header, and this embodiment uses the CAZAC sequence as an example for illustration. The framing format adopts the structure shown in Figure 5, the number of OFDM subcarriers N is 128, the length of the cyclic prefix _Ncp is N/4, the number of subcarriers of useful data is 116, in the form of BPSK signals and 0 intervals, and the number of guard bands on the left is N _b1 =6, the number of right guard bands N _b2 =5, and the frequency point of 0 is 0.

有用数据序列为0与CAZAC序列的点相互间隔的形式，具体形式如图6所示。CAZAC序列由如下公式产生： The useful data sequence is in the form of intervals between 0 and the points of the CAZAC sequence, and the specific form is shown in FIG. 6 . The CAZAC sequence is generated by the following formula:

CA(k)＝exp(-j2πk²u/(2N_u)) CA(k)＝exp(-j2πk ² u/(2N _u ))

其中，k＝0：N_u-1，N_u为CAZAC序列长度，N_u＝58，u的取值范围和k一样，用于生产某一种特定的序列，此例中u＝1。将有用序列分成等长的A，B序列，CAZAC序列点分布在A，B序列的偶数点上，再放入图6中，生成同步头。 Wherein, k=0: _Nu -1, _Nu is the length of the CAZAC sequence, _Nu = 58, the value range of u is the same as k, and is used to produce a specific sequence, in this example, u=1. Divide the useful sequence into A and B sequences of equal length, and the CAZAC sequence points are distributed on the even points of the A and B sequences, and put them into Figure 6 to generate the synchronization header.

步骤1：将接收到的信号去循环前缀，得到长度为N的序列X，对它进行N点能量归一化FFT运算，得到频域序列Y： Step 1: Remove the cyclic prefix from the received signal to obtain a sequence X with a length of N, and perform N-point energy normalized FFT operations on it to obtain a frequency domain sequence Y:

Y＝[y_0，0…y_0，j…y_0，N-1] Y=[y _{0, 0} ... y _{0, j} ... y _{0, N-1} ]

步骤2：从接收数据中取出同步头中的有用数据：按照图5所示的同步头结构，从步骤1得到的数据中，先取第一行的OFDM符号序列，这就是OFDM同步头；再去掉左右保护带和0频点，即取出接收端数据中对应图4中A和B位置的数据；之后，取出A中的偶数点和B中的偶数点，即A与B中CAZAC序列所在点对应的接收端频域数据组成新矩阵： Step 2: Take out the useful data in the synchronization header from the received data: according to the synchronization header structure shown in Figure 5, from the data obtained in step 1, first take the OFDM symbol sequence of the first row, which is the OFDM synchronization header; then remove The left and right guard bands and 0 frequency points, that is, take out the data corresponding to the positions of A and B in Figure 4 in the receiving end data; after that, take out the even points in A and the even points in B, that is, A corresponds to the point where the CAZAC sequence in B is located The frequency domain data of the receiving end form a new matrix:

z＝[z_0，0…z_0，j…z_0，J-1] z=[z _0,0 ...z _0,j ...z _0,J-1 ]

D＝[d_0，0…d_0，j…d_0，J-1] D=[d _0,0 ...d _0,j ...d _0,J-1 ]

其中，J＝58。 Among them, J=58. the

步骤3：利用步骤2得到的发送端和接收端的矩阵，计算系统噪声功率：，其中d^*表示对d取共轭，J＝58。 Step 3: Using the matrices of the transmitter and receiver obtained in Step 2, calculate the system noise power: , where d ^* means taking the conjugate of d, J=58.

Claims

1. A method for estimating the noise power of an OFDM signal, characterized in comprising the steps:

Step 1: Set system parameters, including: According to the user's choice, one of the following four information is used to estimate the noise power: pilot frequency, multiple synchronization heads, single symmetrical synchronization head, and single asymmetric synchronization head; set the left guard The number of band points N _b1 , the number of right guard band points N _b2 , the length of the cyclic prefix N _cp , the number of synchronization headers N _p , the number of pilot symbols N _pl , and a frame of data at the sending end in the frequency domain are expressed as a matrix:

N is the number of subcarriers in the OFDM system, and M is the number of OFDM symbols contained in one frame;

Step 2: Preprocessing the information received by the receiving end, including:

Step 2-1: Remove the cyclic prefix (CP) whose sequence front length is N _cp from the received OFDM symbol, and obtain the time domain information, expressed as follows with a matrix:

X x = = [\begin{matrix} {x x}_{0,0 0,0} & \cdot &Center Dot; \cdot \cdot \cdot \cdot & {x x}_{00,, n no} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {x x}_{00,, N N - - 11} \\ \cdot &Center Dot; & \cdot \cdot & \cdot &Center Dot; \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ \cdot \cdot & \cdot &Center Dot; & \cdot &Center Dot; \\ {x x}_{m m,, 00} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {x x}_{m m,, n no} & \cdot &Center Dot; \cdot \cdot \cdot \cdot & {x x}_{m m,, N N - - 11} \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot \cdot \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ {x x}_{M m - - 1,0 1,0} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {x x}_{M m - - 11,, n no} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {x x}_{M m - - 11,, N N - - 11} \end{matrix}]

Step 2-2: Convert the time-domain signal X in step 2-1 into a frequency-domain signal: If a single symmetrical sync head is used to estimate the noise power, then extract the first row of data in the matrix X obtained in step 2-1, and convert It is divided into two parts before and after, namely: [x _{0, 0} ... x _{0, N/2-1} ] and [x _{0, N/2} ... x _{0, N-1} ], and perform N/2 points on them respectively The energy-normalized FFT of , yields the matrix:

[\begin{matrix} {y the y}_{0,0 0,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {y the y}_{00,, j j} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {y the y}_{00,, N N / / 22 - - 11} \\ {y the y}_{1,0 1,0} & \cdot &Center Dot; \cdot \cdot \cdot \cdot & {y the y}_{11,, j j} & \cdot \cdot \cdot \cdot \cdot \cdot & {y the y}_{11,, N N / / 22 - - 11} \end{matrix}]

Otherwise, perform N-point energy normalized FFT on each row of the matrix X obtained in step 2-1 to obtain the matrix:

Y Y = = [\begin{matrix} {y the y}_{0,0 0,0} & \cdot \cdot \cdot \cdot \cdot \cdot & {y the y}_{00,, n no} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {y the y}_{00,, N N - - 11} \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot \cdot \\ {y the y}_{m m,, 00} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {y the y}_{m m,, n no} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {y the y}_{m m,, N N - - 11} \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot \cdot \\ {y the y}_{M m - - 1,0 1,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {y the y}_{M m - - 11,, n no} & \cdot \cdot \cdot \cdot \cdot \cdot & {y the y}_{M m - - 11,, N N - - 11} \end{matrix}]

Step 3: Extract useful data from the receiving end:

(1) If the pilot frequency is used to estimate the noise power, then according to the definition of the OFDM system specification, find out which OFDM symbols in a frame have pilot information from the defined structure, and extract these symbols; then from them , determine the position of the pilot subcarrier, extract the received signal data at the subcarrier, and form a new pilot matrix:

Z Z = = [\begin{matrix} {z z}_{0,0 0,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {z z}_{00,, j j} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {z z}_{00,, J J - - 11} \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot \cdot & \cdot \cdot & \cdot \cdot \\ {z z}_{i i,, 00} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {z z}_{i i,, j j} & \cdot &Center Dot; \cdot \cdot \cdot \cdot & {z z}_{i i,, J J - - 11} \\ \cdot \cdot & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot \cdot \\ {z z}_{{N N}_{pl pl} - - 1,0 1,0} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {z z}_{{N N}_{pl pl} - - 11,, j j} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {z z}_{{N N}_{pl pl} - - 11,, J J - - 11} \end{matrix}]

Wherein, N _pl is the number of OFDM pilot symbols in a frame, and J is the number of pilot points in a pilot symbol;

(2) If multiple sync heads are used to estimate the noise power, then for each row where the N _p sync head sequences in the matrix Y obtained in step 2 are located, the N _b1 points of the left guard band and the right guard band N _b2 points, 0 frequency point and the point on the subcarrier corresponding to the transmitted data is 0, and the matrix composed of useful data is obtained:

Z Z = = [\begin{matrix} {y the y}_{{i i}_{00},, {j j}_{00}} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {y the y}_{{i i}_{00},, {j j}_{k k}} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {y the y}_{{i i}_{00},, {j j}_{J J - - 11}} \\ \cdot \cdot & \cdot &Center Dot; & \cdot &Center Dot; \\ \cdot \cdot & \cdot &Center Dot; & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot \cdot & \cdot &Center Dot; \\ {y the y}_{{i i}_{m m},, {j j}_{00}} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {y the y}_{{i i}_{m m},, {j j}_{k k}} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {y the y}_{{i i}_{m m},, {j j}_{J J - - 11}} \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ {y the y}_{{i i}_{{N N}_{p p} - - 11},, {j j}_{00}} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {y the y}_{{i i}_{{N N}_{p p} - - 11},, {j j}_{k k}} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {y the y}_{{i i}_{{N N}_{p p} - - 11},, {j j}_{J J - - 11}} \end{matrix}] = = \begin{matrix} [\begin{matrix} {z z}_{0,0 0,0} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {z z}_{00,, j j} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {z z}_{00,, J J - - 11} \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ {z z}_{i i,, 00} & \cdot \cdot \cdot \cdot \cdot \cdot & {z z}_{i i,, j j} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {z z}_{i i,, J J - - 11} \\ \cdot \cdot & \cdot \cdot & \cdot \cdot \\ \cdot \cdot & \cdot \cdot & \cdot \cdot \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ {z z}_{{N N}_{p p} - - 1,0 1,0} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {z z}_{{N N}_{p p} - - 11,, j j} & \cdot \cdot \cdot \cdot \cdot \cdot & {z z}_{{N N}_{p p} - - 11,, J J - - 11} \end{matrix}] \end{matrix}

Where J is the number of points where the data sent in the sync header symbol is not 0, and the value range of j _k is between N _b1 to N/2-1, or N/2+1 to NN _b2 ;

(3) If a single asymmetric sync header is used to estimate the noise power, then for the sync header sequence in the first row of the matrix Y obtained in step 2, remove N _b1 points in the left guard band, N _b2 points in the right guard band, and 0 The frequency point and the point on the subcarrier corresponding to the transmitted data is 0, and the useful data matrix is obtained

Z Z = = [\begin{matrix} {y the y}_{00,, {j j}_{00}} & \cdot \cdot \cdot \cdot \cdot \cdot & {y the y}_{00,, {j j}_{k k}} & \cdot \cdot \cdot \cdot \cdot \cdot & {y the y}_{00,, {j j}_{J J - - 11}} \end{matrix}] = = [\begin{matrix} {z z}_{0,0 0,0} & \cdot \cdot \cdot \cdot \cdot \cdot & {z z}_{00,, j j} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {z z}_{00,, J J - - 11} \end{matrix}]

(4) If a single symmetrical sync head is used to estimate the noise power, the matrix composed of useful data

Z Z = = [\begin{matrix} {y the y}_{0,0 0,0} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {y the y}_{00,, j j} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {y the y}_{00,, N N / / 22 - - 11} \\ {y the y}_{1,0 1,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {y the y}_{11,, j j} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {y the y}_{11,, N N / / 22 - - 11} \end{matrix}] = = [\begin{matrix} {z z}_{0,0 0,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {z z}_{00,, j j} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {z z}_{00,, J J - - 11} \\ {z z}_{1,0 1,0} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {z z}_{11,, j j} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {z z}_{11,, J J - - 11} \end{matrix}]

J=N/2;

Step 4: Extract useful data from the sender:

(1) If pilots are used to estimate the noise power, in the frequency domain signal matrix C at the sending end, first determine which OFDM symbols have pilot information, and extract the corresponding row sequences in C; then from these row sequences, determine The positions of the pilot subcarriers, the transmission information at these subcarriers are extracted to form a new pilot matrix:

D D. = = [\begin{matrix} {d d}_{0,0 0,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{00,, j j} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {d d}_{00,, J J - - 11} \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot &Center Dot; \\ {d d}_{i i,, 00} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {d d}_{i i,, j j} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {d d}_{i i,, J J - - 11} \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot \cdot & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ {d d}_{{N N}_{pl pl} - - 1,0 1,0} & \cdot \cdot \cdot &Center Dot; \cdot \cdot & {d d}_{{N N}_{pl pl} - - 11,, j j} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{{N N}_{pl pl} - - 11,, J J - - 11} \end{matrix}]

J is the number of pilot points in a symbol;

(2) If multiple sync heads are used to estimate the noise power, in the row where the N _p sync head sequences of the matrix C are located, N _b1 points in the left guard band, N _b2 points in the right guard band, 0 frequency points and Points on the subcarrier whose data is 0, get a matrix composed of useful data:

D D. = = [\begin{matrix} {d d}_{0,0 0,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{00,, j j} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {d d}_{00,, J J - - 11} \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot \cdot & \cdot \cdot & \cdot \cdot \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ {d d}_{i i,, 00} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{i i,, j j} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {d d}_{i i,, J J - - 11} \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot \cdot \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot &Center Dot; & \cdot \cdot \\ {d d}_{{N N}_{p p} - - 1,0 1,0} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {d d}_{{N N}_{p p} - - 11,, j j} & \cdot \cdot \cdot \cdot \cdot \cdot & {d d}_{{N N}_{p p} - - 11,, J J - - 11} \end{matrix}] = = [\begin{matrix} {c c}_{{i i}_{00},, {j j}_{00}} & \cdot \cdot \cdot \cdot \cdot \cdot & {c c}_{{i i}_{00},, {j j}_{k k}} & \cdot \cdot \cdot \cdot \cdot \cdot & {c c}_{{i i}_{00},, {j j}_{J J - - 11}} \\ \cdot \cdot & \cdot &Center Dot; & \cdot \cdot \\ \cdot \cdot & \cdot &Center Dot; & \cdot \cdot \\ \cdot \cdot & \cdot &Center Dot; & \cdot \cdot \\ {c c}_{{i i}_{m m},, {j j}_{00}} & \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot & {c c}_{{i i}_{m m},, {j j}_{k k}} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {c c}_{{i i}_{m m},, {j j}_{J J - - 11}} \\ \cdot \cdot & \cdot \cdot & \cdot &Center Dot; \\ \cdot &Center Dot; & \cdot \cdot & \cdot \cdot \\ \cdot &Center Dot; & \cdot \cdot & \cdot &Center Dot; \\ {c c}_{{i i}_{{N N}_{p p} - - 11},, {j j}_{00}} & \cdot \cdot \cdot &Center Dot; \cdot &Center Dot; & {c c}_{{i i}_{{N N}_{p p} - - 11} {j j}_{k k}} & \cdot \cdot \cdot \cdot \cdot \cdot & {c c}_{{i i}_{{N N}_{p p} - - 11},, {j j}_{J J - - 11}} \end{matrix}]

Among them, J is the number of points where the data sent in the sync header symbol is not 0, and the value range of j _k is between N _b1 to N/2-1, or N/2+1 to NN _b2 ;

(3) If a single asymmetric sync head is used to estimate the noise power, the first row of data in matrix C, remove the left guard band N _b1 points, the right guard band N _b2 points, 0 frequency point and the corresponding transmission data is Points on the subcarrier of 0, get a matrix composed of useful data:

D D. = = [\begin{matrix} {d d}_{0,0 0,0} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {d d}_{00,, j j} & \cdot \cdot \cdot \cdot \cdot &Center Dot; & {d d}_{00,, J J - - 11} \end{matrix}] = = [\begin{matrix} {c c}_{00,, {j j}_{00}} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {c c}_{00,, {j j}_{k k}} & \cdot &Center Dot; \cdot \cdot \cdot &Center Dot; & {c c}_{00,, {j j}_{J J - - 11}} \end{matrix}]

J is the number of points where the data sent in the sync header symbol is not 0, and the value range of j _k is between N _b1 and N/2-1, or N/2+1 to NN _b2 ;

(4) If a single symmetrical sync head is used to estimate the noise power, then:

D D. = = [\begin{matrix} {d d}_{0,0 0,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{00,, j j} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{00,, J J - - 11} \\ {d d}_{1,0 1,0} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{11,, j j} & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & {d d}_{11,, J J - - 11} \end{matrix}] = = [\begin{matrix} 11 & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & 11 \\ 11 & \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; & 11 \end{matrix}]

Among them, J=N/2;

Step 5: Calculate the noise power of the system using the useful sequences of the transmitter and receiver obtained in the previous two steps: If a single asymmetric sync head is used to estimate the noise power, then the noise power

\hat{W} = \frac{1}{4 (J - 3)} Σ_{j = 0}^{J - 4} {| (z_{0, j} d_{0, j}^{*} - z_{0, j + 2} d_{0, j + 2}^{*}) - (z_{0, j + 1} d_{0, j + 1}^{*} - z_{0, j + 3} d_{0, j + 3}^{*}) |}^{2};

Otherwise, select any two rows from matrices Z and D to calculate the noise power

\hat{W} = \frac{1}{4 (J - 1)} Σ_{j = 0}^{J - 2} {| (z_{i, j} d_{i, j}^{*} - z_{i, j + 1} d_{i, j + 1}^{*}) - (z_{i^{'}, j} d_{i^{'}, j}^{*} - z_{i^{'}, j + 1} d_{i^{'}, j + 1}^{*}) |}^{2},

Among them, i, i' represent the row numbers of any two rows of matrices Z and D; d ^* represents taking the conjugate of d.

2. a kind of method for estimating the noise power of OFDM signal according to claim 1 is characterized in that: the FFT of described energy normalization refers to dividing by constant factor under the expression of original FFT Where N is the number of FFT points.

3. A kind of method for estimating the noise power of OFDM signal according to claim 1, it is characterized in that: described step 5 selects any two rows from matrix Z and D respectively, is the adjacent in the matrix two lines.