CN102790746B

CN102790746B - Channel estimation method for OFDM (orthogonal frequency division multiplexing) system

Info

Publication number: CN102790746B
Application number: CN201210279402.0A
Authority: CN
Inventors: 左海; 雷霞; 宋亚敏; 刘吉; 张亘
Original assignee: University of Electronic Science and Technology of China
Current assignee: University of Electronic Science and Technology of China
Priority date: 2012-08-08
Filing date: 2012-08-08
Publication date: 2014-10-15
Anticipated expiration: 2032-08-08
Also published as: CN102790746A

Abstract

The invention discloses a channel estimation method for an OFDM (orthogonal frequency division multiplexing) system, which specifically comprises the steps of: initialization and reconstruction of received pilot signals, estimation of a mean value of channel gains, estimation of a slope of a channel, interference estimation and cyclic iteration. According to the method, channel impulse responses is modeled as a multi-path channel of a linear model with rapid time variation and large delay, inter-symbol interference elimination is performed at first for the received pilot data to complete cyclic convolution and reconstruction, and then the mean value of the channel gains and the slope of the channel are estimated by eliminating interference of iterative carriers. By adopting the channel estimation method, effects on the received pilot data caused by interference of carriers resulting from time variation and inter-symbol interference resulting from multi-path delay can be eliminated effectively, and the accuracy of channel estimation can be improved.

Description

Channel Estimation Method for OFDM System

技术领域 technical field

本发明属于移动通信技术领域，具体涉及一种在快时变信道下，利用块状导引代替循环前缀的正交频分复用(Orthogonal Frequency Division Multiplexing，OFDM)系统的信道估计方法。The invention belongs to the technical field of mobile communication, and in particular relates to a channel estimation method of an Orthogonal Frequency Division Multiplexing (OFDM) system using block-like steering instead of a cyclic prefix under fast time-varying channels.

背景技术 Background technique

在未来的无线通信中，要求通信系统能够提供高的传输速率，具有良好的移动性，能够有效的对抗大的多径时延，且具有高的频谱效率，然而，这样的传输环境通常会导致时间选择性和频率选择性衰落，产生符号间干扰(Inter-Symbol Interference，ISI)和载波间干扰(Inter-Carrier Interference，ICI)，破坏OFDM系统的传输特性，降低系统性能。因此，为了对抗快变信道对OFDM系统的影响，进行有效的信道估计，跟踪信道的变化，是不得不考虑的问题。In the future wireless communication, it is required that the communication system can provide high transmission rate, good mobility, be able to effectively resist large multipath delay, and have high spectral efficiency. However, such a transmission environment usually leads to Time-selective and frequency-selective fading produces inter-symbol interference (Inter-Symbol Interference, ISI) and inter-carrier interference (Inter-Carrier Interference, ICI), which destroys the transmission characteristics of the OFDM system and reduces system performance. Therefore, in order to counteract the impact of fast-changing channels on OFDM systems, effective channel estimation and tracking of channel changes are issues that have to be considered.

在多径恶劣的环境中，为提高系统的频谱效率，可采用块状训练序列来替代循环前缀的保护功能，提高频谱效率，同时完成信道估计。针对符号间干扰，提出了一些信道估计方法，具体可参考文献：Shigang Tang;Fang Yang;Kewu Peng,Iterative Channel Estimationfor Block Transmission with Known Symbol Padding-A New Look at TDS-OFDM,inTelecommun.Global Commun.Conf，vol.1,in Washington D.C,Nov. 26-30,2007,pp.4269–4273；及Jun Wang;Zhi-Xing Yang;Chang-Yong Pan;Jian Song;Lin Yang；IterativePadding Subtraction of the PN Sequence for the TDS-OFDM over Broadcast Channels，IEEETransactions on Consumer Electronics,Vol.51,No.4,pp.1148-1152,NOVEMBER2005。然而这些估计方法只能适用于慢变甚至准静态信道。In a harsh multipath environment, in order to improve the spectral efficiency of the system, a block training sequence can be used to replace the protection function of the cyclic prefix to improve the spectral efficiency and complete channel estimation at the same time. Aiming at inter-symbol interference, some channel estimation methods are proposed. For details, please refer to the literature: Shigang Tang; Fang Yang; Kewu Peng, Iterative Channel Estimation for Block Transmission with Known Symbol Padding-A New Look at TDS-OFDM, in Telecommun.Global Commun.Conf , vol.1, in Washington D.C, Nov. 26-30, 2007, pp.4269–4273; and Jun Wang; Zhi-Xing Yang; Chang-Yong Pan; Jian Song; Lin Yang; IterativePadding Subtraction of the PN Sequence for the TDS-OFDM over Broadcast Channels, IEEE Transactions on Consumer Electronics, Vol.51, No.4, pp.1148-1152, NOVEMBER2005. However, these estimation methods can only be applied to slowly varying or even quasi-stationary channels.

针对快时变环境，为了消除载波间干扰，也提出了一些基于基扩展模型或线性模型的信道估计方法，具体可参考文献：Mostofi,Y.;Cox,D.C.ICI Mitigation for Pilot-Aided OFDMMobile Systems，IEEE Transactions on Wiress Communications,Vol.4,pp.765-774,March2005，及Zijian Tang;Cannizzaro,R.C.;Leus,G.;Banelli,P.,Pilot-Assisted Time-VaryingChannel Estimation for OFDM Systems，IEEE Transactions on Signal Processing；Vol.55,No.5,pp.2226-2238,MAY 2007。但是在已有的信道估计方法中，都仅仅消除了符号间干扰或载波间干扰对信道估计的影响，但在快时变大时延环境中，同时存在的符号间干扰和载波间干扰将恶化系统的接收性能，降低现有信道估计方法的估计精度。For fast time-varying environments, in order to eliminate inter-carrier interference, some channel estimation methods based on base extension models or linear models have also been proposed. For details, please refer to the literature: Mostofi, Y.; Cox, D.C.ICI Mitigation for Pilot-Aided OFDM Mobile Systems, IEEE Transactions on Wiress Communications, Vol.4, pp.765-774, March2005, and Zijian Tang; Cannizzaro, R.C.; Leus, G.; Banelli, P., Pilot-Assisted Time-VaryingChannel Estimation for OFDM Systems, IEEE Transactions on Signal Processing; Vol.55, No.5, pp.2226-2238, MAY 2007. However, in the existing channel estimation methods, only the influence of inter-symbol interference or inter-carrier interference on channel estimation is eliminated, but in the environment of fast time and large delay, the simultaneous existence of inter-symbol interference and inter-carrier interference will deteriorate The receiving performance of the system reduces the estimation accuracy of existing channel estimation methods.

发明内容 Contents of the invention

本发明的目的是针对在快时变大时延的多径信道中现有的信道估计方法都仅仅消除了符号间干扰或载波间干扰对信道估计的影响造成的估计精度较低的问题，提出了一种OFDM系统的信道估计方法。The purpose of the present invention is to solve the problem that the existing channel estimation methods only eliminate the influence of inter-symbol interference or inter-carrier interference on channel estimation in multi-path channels with fast time and large delay, and propose A channel estimation method for OFDM system is proposed.

为了实现上述目的，本发明的技术方案是：一种OFDM系统的信道估计方法，具体包括如下步骤：In order to achieve the above object, the technical solution of the present invention is: a channel estimation method of an OFDM system, specifically comprising the following steps:

S0.初始化：设置迭代次数为I，迭代变量j=0，符号间干扰频域响应等于零；载波间干扰等于零，其中，i表示估计的导引块的标号，表示第(i-1)个OFDM数据单元对第i个导引的符号间干扰的频域响应，表示重构循环卷积时引入的符号间干扰的频域响应；S0. Initialization: set the number of iterations to I, iteration variable j=0, inter-symbol interference frequency domain response Equal to zero; intercarrier interference is equal to zero, where i represents the label of the estimated steering block, Represents the (i-1)th OFDM data unit The frequency-domain response to inter-symbol interference for the i-th pilot, Represents the frequency domain response of the inter-symbol interference introduced when reconstructing the circular convolution;

S1.重构接收的导引信号：利用接收的导引信号减去符号间干扰频域响应重构导引与信道冲激响应的圆周卷积的频域响应 S1. Reconstructing the received pilot signal: using the received pilot signal Subtract Intersymbol Interference Frequency Domain Response Frequency Domain Response of Circular Convolution of Reconstruction Guidance and Channel Impulse Response

S2.估计信道增益均值：利用S1重构的频域响应减去载波间干扰得到信道增益均值与导引的数据响应利用已知的导引数据计算信道增益均值表示第l条径的信道增益均值，L表示归一化的最大信道时延；S2. Estimating the mean channel gain: Using the frequency domain response reconstructed by S1 minus intercarrier interference Get mean channel gain and steered data response Utilize known pilot data Calculate channel gain mean Indicates the average channel gain of the lth path, and L indicates the normalized maximum channel delay;

S3.估计信道斜率：利用步骤S2估计的信道增益均值和第i块导引数据更新步骤S2得到的数据响应利用得到载波间干扰然后计算得到信道斜率其中，μ_l表示第l条径的信道斜率；S3. Estimating the channel slope: using the average channel gain estimated in step S2 and the i-th block of bootstrap data Update the data response obtained in step S2 use intercarrier interference Then calculate the channel slope Among them, μ _l represents the channel slope of the lth path;

S4.干扰估计；利用信道状态信息参数和第(i-1)段OFDM数据更新载波间干扰符号间干扰频域响应和 S4. Interference estimation; using channel state information parameters and paragraph (i-1) OFDM data update intercarrier interference Intersymbol Interference Frequency Domain Response and

S5.迭代变量j=j+1，如果j<I，则执行步骤S1至S4，否则，循环结束，利用估计的信道状态参数和完成对第i块导引数据持续时间的信道估计。S5. iteration variable j=j+1, if j<1, then execute steps S1 to S4, otherwise, loop ends, utilizes the channel state parameter of estimation and Complete the i-th block guide data Duration of channel estimation.

本发明的有益效果：本发明将信道冲激响应建模为线性模型的快时变大时延多径信道，提出了一种基于迭代符号间干扰和载波间干扰消除的循环卷积重构的信道估计方法。首先，对接收的导引数据进行符号间干扰消除，完成循环卷积重构，然后采用迭代载波间干扰消除的方式估计信道增益均值和信道斜率。本发明提出的信道估计方法，可以有效的消除时变性引起载波间干扰和多径时延导致的符号间干扰对接收的导引数据的影响，提高信道估计的精度。Beneficial effects of the present invention: the present invention models the channel impulse response as a linear model of fast-time-varying large-delay multipath channels, and proposes a circular convolution reconstruction based on iterative inter-symbol interference and inter-carrier interference elimination channel estimation method. Firstly, inter-symbol interference elimination is performed on the received pilot data, and the circular convolution reconstruction is completed, and then the mean channel gain and channel slope are estimated by iterative inter-carrier interference elimination. The channel estimation method proposed by the invention can effectively eliminate the influence of inter-carrier interference caused by time-varying and inter-symbol interference caused by multipath time delay on received pilot data, and improve the accuracy of channel estimation.

附图说明 Description of drawings

图1为本发明的采用的信道估计方法数据结构模型图。Fig. 1 is a data structure model diagram of the channel estimation method adopted in the present invention.

图2为在COST207TU信道下的信道估计性能比较示意图。Fig. 2 is a schematic diagram of channel estimation performance comparison under COST207TU channel.

具体实施方式 Detailed ways

下面给出本发明的具体实施实例。需要说明的是：实例中的参数并不影响本发明的一般性。The specific implementation examples of the present invention are given below. It should be noted that the parameters in the examples do not affect the generality of the present invention.

在本发明中，发送端发送的数据结构如图1(a)所示。在发送端，采用时域块状已知序列代替循环前缀作为保护间隔，且导引长度大于信道归一化最大时延。由于填加的导引数据和发送的OFDM数据前都没有循环前缀的保护，因此接收的导引信号和OFDM数据将分别受到来自相邻OFDM数据或导引数据的符号间干扰，如图1(b)阴影部分所示。针对图1的数据模型，本发明提出了一种有效的信道估计方法。In the present invention, the data structure sent by the sending end is shown in Fig. 1(a). At the sending end, the time-domain block known sequence is used instead of the cyclic prefix as the guard interval, and the pilot length is greater than the channel normalized maximum delay. Since there is no cyclic prefix protection before the added pilot data and the transmitted OFDM data, the received pilot signal and OFDM data will be subject to intersymbol interference from adjacent OFDM data or pilot data respectively, as shown in Figure 1( b) Shown in the shaded part. Aiming at the data model in Fig. 1, the present invention proposes an effective channel estimation method.

在本发明的系统模型中，假设具有L+1径的多径信道h=[h₀,h₁,h₂,...,h_L]^T，导引块的数据长度是M，且M>L；OFDM符号长度N；设第i块导引数据令N₂=N+M，定义第i个OFDM数据单元表示为：In the system model of the present invention, assuming a multipath channel h=[h ₀ , h ₁ ,h ₂ ,...,h _L ] ^T with L+1 paths, the data length of the pilot block is M, and M >L; OFDM symbol length N; set the i-th block of pilot data Let N ₂ =N+M, define the ith OFDM data unit Expressed as:

${x x}_{n no}^{i i} = = \{\begin{matrix} {s the s}_{n no}^{i i},, 00 \leq \leq n no < < N N - - 11 \\ {p p}_{n no}^{i i + + 11},, N N \leq \leq n no < < {N N}_{22} \end{matrix} - - - - - - ((11))$

定义长度为N₃=N+2M的OFDM扩展数据块它由第i段OFDM数据和其前后两段导引数据组成：Define an OFDM extended data block with a length of N ₃ =N+2M It consists of the i-th piece of OFDM data and two pieces of guidance data before and after it:

${\overset{&OverBar; &OverBar;}{x x}}_{n no}^{i i} = = \{\begin{matrix} {x x}_{n no}^{i i},, 00 \leq \leq n no < < {N N}_{22} - - 11 \\ {p p}_{n no}^{i i},, - - M m \leq \leq n no < < 00 \end{matrix} - - - - - - ((22))$

其中，n表示数据采样时刻，i表示定义的响应数据块的标号。按照图1(a)的数据结构，如果满足则导引数据pⁱ可以作为OFDM数据单元的循环前缀。Among them, n represents the data sampling time, and i represents the label of the defined response data block. According to the data structure in Figure 1(a), if it satisfies Then the pilot data p ⁱ can be used as OFDM data unit cyclic prefix.

基于接收导引的信号响应分解：Signal response decomposition based on reception guidance:

假设在OFDM数据块持续时间内，信道的归一化多普勒扩展值小于20%，则信道冲激响应近似满足线性变化准则。定义导引数据持续时间内第l条径中间时刻的信道增益为则持续时间内的信道状态信息描述为：Assume that in the OFDM data block If the normalized Doppler spread value of the channel is less than 20% within the duration, the channel impulse response approximately satisfies the linear variation criterion. Define bootstrap data Channel gain at the middle moment of the lth path in the duration for but The channel state information during the duration is described as:

${h h}_{n no,, l l} = = {h h}_{l l}^{ave ave} + + {μ μ}_{l l} [[n no - - \frac{{N N}_{p p}}{22}]],, 00 \leq \leq l l \leq \leq L L,, 00 \leq \leq n no < < {N N}_{22} + + {N N}_{p p} - - - - - - ((33))$

其中：h_n，l表示第l条径第n时刻的信道增益；μ_l表示第l条径的信道斜率。从图1知，接收的导引的数据响应 Among them: h _{n, l} represents the channel gain of the lth path at the nth moment; _μl represents the channel slope of the lth path. From Figure 1, the received guide data response for

${r r}_{__pilot pilot}^{i i} ((n no)) = = \{\begin{matrix} {Σ Σ}_{l l = = 00}^{n no} {h h}_{n no,, l l} {p p}_{< < n no - - l l > > {N N}_{p p}}^{i i} + + {Σ Σ}_{l l = = n no + + 11}^{L L} {h h}_{n no,, l l} {s the s}_{< < n no - - l l > > N N}^{i i - - 11} + + {w w}_{n no}^{i i},, 00 \leq \leq n no < < L L \\ {Σ Σ}_{l l = = 00}^{L L} {h h}_{n no,, l l} {p p}_{< < n no - - 11 > > {N N}_{p p}}^{i i} + + {w w}_{n no}^{i i},, L L \leq \leq n no < < {N N}_{p p} \end{matrix} - - - - - - ((44))$

由于没有循环前缀的保护，接收的导引数据前L个值将受到ISI的干扰，为了重构导引数据与信道响应的圆周卷积，对接收的导引进行数据重构，对的前L个样本修改为：Since there is no cyclic prefix protection, the first L values of the received pilot data will be interfered by ISI. In order to reconstruct the circular convolution of the pilot data and the channel response, data reconstruction is performed on the received pilot data. The first L samples of are modified as:

${r r}_{__pilot pilot}^{i i} ((n no)) = = {Σ Σ}_{l l = = 00}^{L L} {h h}_{n no,, l l} {p p}_{< < n no - - l l > > {N N}_{p p}}^{i i} + + {Σ Σ}_{l l = = n no + + 11}^{L L} {h h}_{n no,, l l} {s the s}_{< < n no - - l l > > N N}^{i i - - 11} - - {Σ Σ}_{l l = = n no + + 11}^{L L} {h h}_{n no,, l l} {p p}_{< < n no - - l l > > {N N}_{p p}}^{i i} + + {w w}_{n no}^{i i},, 00 \leq \leq n no < < L L - - - - - - ((55))$

因此，接收的导引信号等效为：Therefore, the received pilot signal is equivalent to:

${r r}_{__pilot pilot}^{i i} = = {r r}_{__pilot pilot}^{11,, i i} + + {r r}_{__pilot pilot}^{22,, i i} + + {r r}_{__pilot pilot}^{33,, i i} + + {w w}^{i i} - - - - - - ((66))$

其中，满足信道冲激响应与导引数据的圆周卷积，是来自OFDM数据的符号间干扰，是重构数据时引入的符号间干扰。其中，是N_p×N_p的信道矩阵，0≤k,n<N_p，如式(7)描述；表示导引数据持续时间内，第k个时刻在第<k-n〉N_p条径上的信道冲激响应，<k〉n表示k除以n的余数。是相邻的第(i-1)段OFDM数据对接收的导引信号的符号间干扰，其中是N_p×N的矩阵，如式(8)所示，且0≤k<N_p，0≤n<N。是为了构造引入的修复数据；其中是N_p×N_p的矩阵，如式(9)所示，且0≤k,n<N_p。in, Satisfy the circular convolution of the channel impulse response and the guidance data, is the intersymbol interference from OFDM data, is the intersymbol interference introduced when reconstructing the data. in, is the N _p ×N _p channel matrix, 0≤k,n<N _p , as described in formula (7); Represents bootstrap data Within the duration, the channel impulse response at the kth moment on the <kn>N _p path, where <k>n represents the remainder of dividing k by n. is the adjacent (i-1) section OFDM data Intersymbol interference to the received pilot signal, in is a matrix of N _p ×N, as shown in formula (8), and 0≤k<N _p , 0≤n<N. is to construct imported repair data; in is a matrix of N _p ×N _p , as shown in formula (9), and 0≤k, n<N _p .

${H h}_{CIR CIR}^{i i} = = {[\begin{matrix} {h h}_{00,, 00} & 00 & . . . . . . & {h h}_{00,, L L} & . . . . . . & {h h}_{0,1 0,1} \\ {h h}_{1,1 1,1} & {h h}_{1,0 1,0} & 00 & . . . . . . & {h h}_{1,3 1,3} & {h h}_{1,2 1,2} \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . \\ {h h}_{L L,, L L} & . . . . . . & {h h}_{L L,, 00} & . . . . . . & . . . . . . & 00 \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . \\ 00 & . . . . . . & 00 & {h h}_{{N N}_{p p} - - 11,, L L} & . . . . . . & {h h}_{{N N}_{p p} - - 1,0 1,0} \end{matrix}]}_{{N N}_{p p} \times \times {N N}_{p p}} - - - - - - ((77))$

${A A}_{ISI ISI}^{i i} = = {[\begin{matrix} 00 & . . . . . . & 00 & {h h}_{00,, L L} & {h h}_{00,, L L - - 11} & . . . . . . & {h h}_{0,1 0,1} \\ 00 & . . . . . . & . . . . . . & 00 & {h h}_{11,, L L} & . . . . . . & {h h}_{1,2 1,2} \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . \\ 00 & . . . . . . & . . . . . . & . . . . . . & . . . . . . & 00 & {h h}_{L L - - 11 . . L L} \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & 00 & 00 \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . \\ 00 & . . . . . . & 00 & . . . . . . & . . . . . . & . . . . . . & 00 \end{matrix}]}_{{N N}_{p p} \times \times N N} - - - - - - ((88))$

${B B}_{ISI ISI}^{i i} = = {[\begin{matrix} 00 & . . . . . . & 00 & {h h}_{00,, L L} & {h h}_{00,, L L - - 11} & . . . . . . & {h h}_{0,1 0,1} \\ 00 & . . . . . . & . . . . . . & 00 & {h h}_{11,, L L} & . . . . . . & {h h}_{1,2 1,2} \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . \\ 00 & . . . . . . & . . . . . . & . . . . . . & . . . . . . & 00 & {h h}_{L L - - 11,, L L} \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . & . . . . . . \\ 00 & . . . . . . & 00 & . . . . . . & . . . . . . & . . . . . . & 00 \end{matrix}]}_{{N N}_{p p} \times \times {N N}_{p p}} - - - - - - ((99))$

将式(3)带入式(7)、(8)、(9)，得到每个式子将由两部分组成，分别是信道均值部分，即时不变部分和时变部分，其中， 0≤k,n<N_p，0≤k,n<N_p；M是对角矩阵，M(k,k)=-N_p/2+k,0≤n<N_p。类似的，分别利用信道增益的均值矩阵和斜率矩阵替代，对做N_p点的FFT变换：Putting formula (3) into formulas (7), (8), and (9), each formula will be composed of two parts, which are the channel mean part, the instant invariant part and the time-varying part. Among them, 0≤k,n<N _p , 0≤k,n<N _p ; M is a diagonal matrix, M(k,k)=-N _p /2+k,0≤n<N _p . akin, Use the mean matrix and slope matrix of the channel gain to replace, right Do the FFT transformation of N _p points:

${R R}_{__pilot pilot}^{i i} = = {R R}_{__pilot pilot}^{11,, i i} + + {R R}_{__pilot pilot}^{22,, i i} + + {R R}_{__pilot pilot}^{33,, i i} + + {W W}^{i i} - - - - - - ((1010))$

其中，Wⁱ是高斯噪声wⁱ的N_p点FFT变换where W ⁱ is the N _p- point FFT transformation of Gaussian noise w ⁱ

$\{\begin{matrix} {R R}_{__pilot pilot}^{i i} = = {F f}_{{N N}_{p p}} \times \times {r r}_{__pilot pilot}^{i i} \\ {R R}_{__pilot pilot}^{11,, i i} = = {F f}_{{N N}_{p p}} \times \times {r r}_{__pilot pilot}^{11,, i i} = = {G G}^{11,, i i} \times \times {P P}^{i i} \\ {R R}_{__pilot pilot}^{22,, i i} = = {F f}_{{N N}_{p p}} \times \times {r r}_{__pilot pilot}^{22,, i i} = = {G G}^{22,, i i} \times \times {S S}^{i i - - 11} \\ {R R}_{__pilot pilot}^{33,, i i} = = {F f}_{{N N}_{p p}} \times \times {r r}_{__pilot pilot}^{33,, i i} = = {G G}^{33,, i i} \times \times {P P}^{i i} \end{matrix} - - - - - - ((1111))$

将式(7)带入式(11)的第二项，得到:Put formula (7) into the second term of formula (11), get:

${G G}^{11,, i i} = = {F f}_{{N N}_{p p}} (({H h}_{CIR CIR}^{ave ave,, i i} + + M m \times \times {H h}_{CIR CIR}^{var var,, i i})) {F f}_{{N N}_{p p}}^{H h}$

$= = {F f}_{{N N}_{p p}} {H h}_{CIR CIR}^{ave ave,, i i} {F f}_{{N N}_{p p}}^{H h} + + {F f}_{{N N}_{p p}} ((M m \times \times {H h}_{CIR CIR}^{var var,, i i})) {F f}_{{N N}_{p p}}^{H h} - - - - - - ((1212))$

$= = {G G}_{11}^{11,, i i} + + {G G}_{22}^{11,, i i}$

且是对角矩阵，and is a diagonal matrix,

${G G}_{11}^{11,, i i} ((k k,, k k)) = = {Σ Σ}_{l l = = 00}^{L L} {h h}_{l l}^{ave ave} {e e}^{- - j j \frac{22 π π}{{N N}_{p p}} kl kl},, 00 \leq \leq k k \leq \leq {N N}_{p p} - - 11 - - - - - - ((1313))$

${G G}_{22}^{11,, i i} ((k k,, m m)) = = \{\begin{matrix} {Σ Σ}_{l l = = 00}^{L L} \frac{{μ μ}_{l l}}{{N N}_{p p}} {e e}^{- - j j \frac{22 π π}{{N N}_{p p}} ml ml} {Σ Σ}_{n no = = 00}^{{N N}_{p p} - - 11} ((n no - - \frac{{N N}_{p p}}{22})) {e e}^{j j \frac{22 π π}{{N N}_{p p}} ((m m - - k k)) n no},, 00 \leq \leq k k,, m m < < {N N}_{p p},, k k &NotEqual; &NotEqual; m m \\ 00,, k k = = m m \end{matrix} - - - - - - ((1414))$

因此，得到：Therefore, get:

${R R}_{__pilot pilot}^{11,, i i} = = {R R}_{__pilot pilot,, 11}^{11,, i i} + + {R R}_{__pilot pilot,, 22}^{11,, i i} - - - - - - ((1515))$

即：Right now:

${R R}_{__pilot pilot,, 11}^{11,, i i} = = {G G}_{11}^{11,, i i} \times \times {P P}^{i i} - - - - - - ((1616))$

${R R}_{__pilot pilot,, 22}^{11,, i i} = = {G G}_{22}^{11,, i i} \times \times {P P}^{i i} - - - - - - ((1717))$

利用式(17)，简化为：Using formula (17), Simplifies to:

${R R}_{__pilot pilot,, 22}^{11,, i i} ((k k)) = = \frac{11}{{N N}_{p p}} C C \times \times \overset{&RightArrow; &Right Arrow;}{μ μ} - - - - - - ((1818))$

其中，C是N_p×(L+1)的矩阵，μ是由信道斜率构成的列向量；C=E×F，且分别满足：Among them, C is a matrix of N _p ×(L+1), μ is a column vector composed of channel slope; C=E×F, and respectively satisfy:

$E E. ((k k,, m m)) = = {Σ Σ}_{n no = = 00}^{{N N}_{p p} - - 11} ((n no - - \frac{{N N}_{p p}}{22})) {e e}^{j j \frac{22 π π}{{N N}_{p p}} ((m m - - k k)) n no},, 00 \leq \leq k k,, m m < < {N N}_{p p}$

$F f ((m m,, l l)) = = {P P}_{k k}^{i i} \times \times {e e}^{- - j j \frac{22 π π}{{N N}_{p p}} ml ml},, 00 \leq \leq l l \leq \leq L L,, 00 \leq \leq m m < < {N N}_{p p}$

类似的，得到：Similarly, get:

${G G}^{22,, i i} = = {F f}_{{N N}_{p p}} {A A}_{ISI ISI}^{i i} {F f}_{N N}^{H h} = = {F f}_{{N N}_{p p}} {A A}_{ISI ISI}^{ave ave,, i i} {F f}_{N N}^{H h} + + {F f}_{{N N}_{p p}} ((M m \times \times {A A}_{ISI ISI}^{var var,, i i})) {F f}_{N N}^{H h} - - - - - - ((1919))$

${G G}^{33,, i i} = = - - {F f}_{{N N}_{p p}} {B B}_{ISI ISI}^{i i} {F f}_{{N N}_{p p}}^{H h} = = - - [[{F f}_{{N N}_{p p}} {B B}_{ISI ISI}^{ave ave,, i i} {F f}_{{N N}_{p p}}^{H h} + + {F f}_{{N N}_{p p}} ((M m \times \times {B B}_{ISI ISI}^{var var,, i i})) {F f}_{{N N}_{p p}}^{H h}]] - - - - - - ((2020))$

因此，可以得到：Therefore, one can get:

${R R}_{__pilot pilot}^{i i} ((k k)) = = {R R}_{__pilot pilot,, 11}^{11,, i i} ((k k)) + + {R R}_{__pilot pilot,, 22}^{22,, i i} ((k k)) + + {R R}_{__pilot pilot}^{22,, i i} ((k k)) + + {R R}_{__pilot pilot}^{33,, i i} ((k k)) + + {W W}^{i i} ((k k)),, 00 \leq \leq k k < < {N N}_{p p} - - - - - - ((21 twenty one))$

其中in

${R R}_{__pilot pilot}^{22,, i i} ((k k)) = = {Σ Σ}_{m m = = 00}^{N N - - 11} {G G}^{22,, i i} ((k k,, m m)) {S S}_{m m}^{i i - - 11} - - - - - - ((22 twenty two))$

${R R}_{__pilot pilot}^{33,, i i} ((k k)) = = {Σ Σ}_{m m = = 00}^{{N N}_{p p} - - 11} {G G}^{33,, i i} ((k k,, m m)) {P P}_{m m}^{i i} - - - - - - ((23 twenty three))$

因此，接收的导引信号的频域响应中，是未受到ISI和ICI干扰的频域响应，它是信道状态信息的均值与导引数据的频域响应；是由于信道的时变性导致的频域ICI分量，它与信道变化的斜率有关；是由于多径时延引起的ISI频域分量，是重构数据响应引入符号间干扰分量。在利用接收的导引数据进行信道估计时，必须首先消除符号间干扰分量和完成循环卷积重构，然后分别利用完成对信道增益均值和斜率的估计，获取信道状态信息。Therefore, the frequency domain response of the received pilot signal middle, is the frequency domain response without ISI and ICI interference, which is the mean value of channel state information and the frequency domain response of the pilot data; is the frequency-domain ICI component caused by the time-varying nature of the channel, which is related to the slope of the channel change; is the ISI frequency domain component due to multipath delay, is the reconstruction of the data response to introduce an inter-symbol interference component. When using the received pilot data for channel estimation, the inter-symbol interference component must be eliminated first and Complete the circular convolution reconstruction, and then use the Complete the estimation of channel gain mean value and slope, and obtain channel state information.

本发明的信道估计方法中，对接收的导引数据首先消除符号间干扰，完成数据重构，然后利用重构的导引数据估计信道线性模型的信道增益均值和斜率，完成信道估计。具体信道估计步骤描述为：In the channel estimation method of the present invention, inter-symbol interference is firstly eliminated for the received pilot data, and data reconstruction is completed, and then the channel gain average and slope of the channel linear model are estimated by using the reconstructed pilot data, and channel estimation is completed. The specific channel estimation steps are described as:

S0.初始化：设置迭代次数I，迭代变量j=0，符号间干扰频域响应等于零；载波间干扰等于零。其中，i表示估计的导引块的编号，表示第(i-1)个OFDM数据单元对第i个导引的符号间干扰频域响应，表示重构循环卷积时引入的符号间干扰的频域响应；S0. Initialization: set the number of iterations I, iteration variable j=0, inter-symbol interference frequency domain response Equal to zero; intercarrier interference is equal to zero. where i represents the number of the estimated pilot block, Represents the (i-1)th OFDM data unit Intersymbol interference frequency domain response to the ith pilot, Represents the frequency domain response of the inter-symbol interference introduced when reconstructing the circular convolution;

S2.估计信道增益均值h^ave：利用S1重构的频域响应减去载波间干扰得到信道增益均值与导引的数据响应利用已知的导引数据计算信道增益均值表示第l条径的信道增益均值，L表示归一化的最大信道时延；S2. Estimated channel gain mean value h ^ave : Frequency domain response reconstructed using S1 minus intercarrier interference Get mean channel gain and steered data response Utilize known pilot data Calculate channel gain mean Indicates the average channel gain of the lth path, and L indicates the normalized maximum channel delay;

S3.估计信道斜率μ：利用步骤S2估计的信道增益均值和第i块导引数据更新步骤S2得到的数据响应利用得到载波间干扰然后计算得到信道斜率其中，μ_l表示第l条径的信道斜率;S3. Estimated channel slope μ: use the mean channel gain estimated in step S2 and the i-th block of bootstrap data Update the data response obtained in step S2 use intercarrier interference Then calculate the channel slope Wherein, μ _l represents the channel slope of the lth path;

下面结合具体的参数对上述步骤进行展开说明：The above steps are described below with specific parameters:

本实例中的OFDM系统参数设置如下：设OFDM信号调制方式采用QPSK调制，子载波总数N＝512，子载波序号为[0,1,2,....,511]。设块状导引数据长度M＝400，p=[p₀,p₁，...,p₃₉₉]^T，且各段时域块状导引序列数据一样。本实例中选择的信道参数如下：采用COST207TU信道，多径时延分别为[0，0.2，0.5，1.6，2.3，5]us，多径衰落功率分别是[3，0，5，6，8，10]dB，归一化多普勒扩展值是0.1，设系统采样周期T＝1.0e-7s。则归一化的信道最大多径时延L=5.0e-6/T=50。The OFDM system parameters in this example are set as follows: it is assumed that the OFDM signal modulation mode adopts QPSK modulation, the total number of subcarriers is N=512, and the subcarrier numbers are [0, 1, 2, ... , 511]. It is assumed that _the length of the block pilot data is M=400, p=[p ₀ , p ₁ , ^. The channel parameters selected in this example are as follows: COST207TU channel is used, the multipath delay is [0, 0.2, 0.5, 1.6, 2.3, 5]us, and the multipath fading power is [3, 0, 5, 6, 8] , 10] dB, the normalized Doppler spread value is 0.1, and the system sampling period is T=1.0e-7s. Then the normalized channel maximum multipath delay L=5.0e-6/T=50.

本实例中发送端信号处理过程为：二进制比特信号先经QPSK调制，然后经过IFFT变换组成长度为N的时域OFDM数据。将长度为M的相同的导引数据加载到相邻的、长度为N的时域OFDM数据间，组成发送的数据模型。这样，由时域块状导引数据和OFDM数据组成的数据信息由发射机进行发射。In this example, the signal processing process at the sending end is as follows: the binary bit signal is first modulated by QPSK, and then transformed by IFFT to form OFDM data in the time domain with a length of N. The same pilot data with a length of M is loaded between adjacent OFDM data in the time domain with a length of N to form a transmitted data model. In this way, data information composed of time domain block pilot data and OFDM data is transmitted by the transmitter.

在接收端，根据图1(b)，以接收的第i个时域OFDM块的信号响应为例，采用本发明所述的方法对第i块导引数据进行信道估计时，本发明提出的信道估计算法具体描述为：At the receiving end, according to Figure 1(b), the i-th time-domain OFDM block received The signal response of the present invention is taken as an example, when adopting the method described in the present invention to carry out channel estimation to i-th block pilot data, the channel estimation algorithm that the present invention proposes is specifically described as:

S0.初始化：设置迭代次数I=3，迭代变量j=0，符号间干扰频域响应等于零；载波间干扰等于零。其中，i表示估计的导引块的编号，表示第(i-1)个OFDM数据单元对第i个导引的符号间干扰的频域响应，表示重构循环卷积时引入的符号间干扰的频域响应；S0. Initialization: Set the number of iterations I=3, the iteration variable j=0, and the frequency domain response of inter-symbol interference Equal to zero; intercarrier interference is equal to zero. where i represents the number of the estimated pilot block, Represents the (i-1)th OFDM data unit The frequency-domain response to inter-symbol interference for the i-th pilot, Represents the frequency domain response of the inter-symbol interference introduced when reconstructing the circular convolution;

需要说明的是：当迭代次数I达到一定的次数时，性能基本不变，收敛稳定；反之，继续增加次数会增加复杂度，但性能却提高不大，迭代次数I一般可以根据经验或仿真结果确定。It should be noted that: when the number of iterations I reaches a certain number, the performance is basically unchanged and the convergence is stable; on the contrary, increasing the number of times will increase the complexity, but the performance will not improve much. The number of iterations I can generally be determined based on experience or simulation results. Sure.

S2.估计信道增益均值利用S1重构的频域响应减去载波间干扰得到数据响应利用已知的导引数据得到时域信道增益均值表示估计的第l条径的信道增益均值，L表示归一化的最大信道时延；S2. Estimate the average channel gain Frequency Domain Response Using S1 Reconstruction minus intercarrier interference get data response Utilize known pilot data Get the mean channel gain in the time domain Represents the estimated mean channel gain of the lth path, L represents the normalized maximum channel delay;

S3.估计信道斜率利用步骤S2估计的信道均值和预先得到的导引更新步骤S2得到的数据响应得到载波间干扰进一步估计得到信道斜率 S3. Estimate channel slope Using the channel mean estimated by step S2 and pre-received guidance Update the data response obtained in step S2 intercarrier interference Further estimate the channel slope

S4.干扰估计；利用信道状态信息参数和第(i-1)段OFDM数据更新载波间干扰符号间干扰频域响应和 ${{\hat{R}}_{_pilot}^{3, i} (k)}_{k = 0}^{399};$ S4. Interference estimation; using channel state information parameters and paragraph (i-1) OFDM data update intercarrier interference Intersymbol Interference Frequency Domain Response and ${{\hat{R}}_{_pilot}^{3, i} (k)}_{k = 0}^{399};$

图2给出了该实例的信道估计的归一化均方误差的比较示意图，其中Num表示不同的迭代次数。Fig. 2 shows a schematic diagram of comparison of normalized mean square error of channel estimation in this example, where Num represents different iteration times.

其中，归一化均方误差表示为：Among them, the normalized mean square error is expressed as:

$MMSE MMSE = = \frac{11}{M m} {Σ Σ}_{i i = = 00}^{M m} {{\frac{{Σ Σ}_{n no = = 00}^{{N N}_{p p} - - 11} {Σ Σ}_{l l = = o o}^{L L} {| | h h ((n no,, l l)) - - \overset{^^}{h h} ((n no,, l l)) | |}^{22}}{{Σ Σ}_{n no = = 00}^{{N N}_{p p} - - 11} {Σ Σ}_{l l = = o o}^{L L} {| | h h ((n no,, l l)) | |}^{22}}}}$

其中，表示第n时刻第l条径的信道增益h(n,l)的估计值，M表示完成信道估计的导引数目。仿真结果表明，当迭代次数I=3时，SNR=24dB时，MMSE能降低到5e-3。对比进行ISI干扰消除的信道估计，本发明提出的信道估计算法能显著提高信道估计性能。in, Indicates the estimated value of the channel gain h(n,l) of the lth path at the nth moment, and M indicates the number of pilots for which the channel estimation is completed. The simulation results show that when the number of iterations I=3 and the SNR=24dB, the MMSE can be reduced to 5e-3. Compared with the channel estimation of ISI interference elimination, the channel estimation algorithm proposed by the present invention can significantly improve the channel estimation performance.

以上实例仅为本发明的较佳例子而已，本发明的使用并不局限于该实例，凡在本发明的精神和原则之内，所做的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。The above example is only a preferred example of the present invention, and the use of the present invention is not limited to this example. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the within the protection scope of the present invention.

Claims

1. A channel estimation method of an OFDM system, specifically comprising the steps of:

S0. Initialization: set the number of iterations to I, iteration variable j=0, inter-symbol interference frequency domain response Equal to zero; intercarrier interference is equal to zero, where i represents the label of the estimated steering block, Represents the (i-1)th OFDM data unit The frequency-domain response to inter-symbol interference for the i-th pilot, Represents the frequency domain response of the inter-symbol interference introduced when reconstructing the circular convolution;

S1. Reconstructing the received pilot signal: using the received pilot signal Subtract Intersymbol Interference Frequency Domain Response Frequency Domain Response of Circular Convolution of Reconstruction Guidance and Channel Impulse Response

S2. Estimating the mean channel gain: Using the frequency domain response reconstructed by S1 minus intercarrier interference get data response Utilize known pilot data Calculate channel gain mean Indicates the average channel gain of the lth path, and L indicates the normalized maximum channel delay;

S3. Estimating the channel slope: using the average channel gain estimated in step S2 and the i-th block of bootstrap data Update the data response obtained in step S2 use intercarrier interference Then calculate the channel slope Among them, μ _l represents the channel slope of the lth path;

S4. Interference estimation; using channel state information parameters and paragraph (i-1) OFDM data update intercarrier interference Intersymbol Interference Frequency Domain Response and

S5. iteration variable j=j+1, if j<1, then carry out steps S1 to S4, otherwise, loop ends, utilizes the channel state parameter of estimation and Complete the i-th block guide data Duration of channel estimation.