CN1200521C - Method for inserting pilot frequency and estimating parameters of channel of equilibrium in frequency doman - Google Patents

Method for inserting pilot frequency and estimating parameters of channel of equilibrium in frequency doman Download PDF

Info

Publication number
CN1200521C
CN1200521C CN 03137565 CN03137565A CN1200521C CN 1200521 C CN1200521 C CN 1200521C CN 03137565 CN03137565 CN 03137565 CN 03137565 A CN03137565 A CN 03137565A CN 1200521 C CN1200521 C CN 1200521C
Authority
CN
China
Prior art keywords
amp
pilot
data
frequency
channel
Prior art date
Application number
CN 03137565
Other languages
Chinese (zh)
Other versions
CN1463083A (en
Inventor
周世东
赵明
周春晖
王京
施婷婷
任晓东
张秀军
Original Assignee
清华大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 清华大学 filed Critical 清华大学
Priority to CN 03137565 priority Critical patent/CN1200521C/en
Publication of CN1463083A publication Critical patent/CN1463083A/en
Application granted granted Critical
Publication of CN1200521C publication Critical patent/CN1200521C/en

Links

Abstract

用于频域均衡的导频插入与信道参数估计的方法属于无线通信技术领域,其特征在于:在无线通信系统的发送端,编码调制映射后的每个数据信号序列与导频信号序列复接形成一个数据块,经正交变换和交织,使导频信息只出现在某些频点上,而数据则没有这些频率分量,再对每个数据块在发送端加循环前缀;在接收端,再去掉循环前缀,以消除多径时延扩展引起的数据块之间的干扰和频域信号间的干扰,然后利用提取出的频域导频信息在相应频点上作信道参数估计,再据此通过插值的方法在所有频点上作信道参数估计。 A method for the frequency domain equalization guide inserted pilot channel parameter estimation belongs wireless communications technology, wherein: each of the data signal sequence in the transmitting end of the guide a wireless communication system, a pilot signal sequence mapped coded modulation multiplexing forming a data block interleaved by orthogonal transformation and the pilot information appears only at some frequency, and the data do not have these frequency components, then each data block for adding a cyclic prefix at the transmitting side; at the receiving end, then remove the cyclic prefix to eliminate the interference between the interference and the frequency domain signal between the blocks caused by multipath delay spread, and then using the extracted frequency-domain pilot information for channel estimation in the corresponding frequency parameter points, and then according to this method of interpolation by the channel parameter estimation for all frequency. 它不仅算法和实现简单而且信道估计的精度也高。 It is not only simple but also algorithm and channel estimation accuracy is high.

Description

用于频域均衡的导频插入与信道参数估计的方法 A method for frequency domain equalization inserted pilot channel parameter estimation

技术领域 FIELD

用于频域均衡的导频插入与信道参数估计的方法属于无线通信技术领域。 A method for the frequency domain equalization guide inserted pilot channel parameter estimation belongs wireless communications technology.

背景技术 Background technique

一个实际的基带传输系统不可能完全满足理想的波形传输无失真条件,因而串扰几乎是不可避免的。 A real baseband transmission system can not fully meet the ideal transmission without distortion of the waveform conditions, and thus the crosstalk is almost inevitable. 当串扰造成严重影响时,必须对整个系统的传递函数进行校正,即均衡,使其接近无失真传输条件。 When crosstalk is serious impact, must transfer function of the whole system of correction, that is balanced, so close to lossless transmission conditions. 均衡就是接收端的均衡器产生与信道特性相反的特性,用来抵消信道的时变多径传播特性引起的干扰,可分为时域均衡和频域均衡两种。 Equilibrium is to receive the opposite end of the equalizer characteristics of the channel characteristics, multipath propagation interference characteristics change due to offset when the channel can be divided into time domain equalization and frequency domain equalization two kinds. 在均衡前,需要估计信道参数,通常利用导频信号进行信道估计。 Before equalization, channel parameters to be estimated, usually using a pilot signal for channel estimation.

通信系统中发送的信号主要可分为导频信号和数据,其中数据是实际发送的信息信号,导频信号是一个附加的已知的信息信号,用于在接收部分进行信道参数的估计从而求出均衡的加权系数。 Communication system signal transmitted can be divided into estimated pilot signal and data signal wherein the data information is actually transmitted, the pilot signal is a known additional information signal, for channel parameters in the receiving section so as to seek balanced weighting factors.

传统的方法如3G系统中通常采用导频占用一个正交码道的插入方法,采用多码道传输数据,当信道多径严重时,码道间有严重的串扰。 The conventional methods commonly employed in 3G system occupies a pilot channel orthogonal code insertion method, using multi-code transmission of the data channel, when the channel multipath serious, severe crosstalk between tracks. 在均衡前,导频码道与数据码道间的串扰使得根据导频信号估计出的信道参数不准确,依据所得的信道参数进行的时域均衡,收敛速度慢且横向滤波器的抽头数太多、实现太复杂。 Before equalization, the pilot crosstalk between the code channels and the data track such that the estimated pilot signal channel parameter is not accurate, the time domain according to channel parameters resulting equalization, convergence speed and the number of taps is too transversal filter much too complicated to achieve.

为此需要改进导频插入和信道估计的方法,消除导频信号与数据信号间的串扰,提高信道估计的准确度,同时降低算法和实现的复杂度。 This requires improved methods and inserting pilot channel estimation, to eliminate crosstalk between the pilot signal and the data signal, to improve the accuracy of channel estimation, while reducing the complexity of the algorithm and implementation.

发明内容 SUMMARY

本发明的目的是:为了获得较高的信道估计精度,提出了一种实现简单的用于频域均衡的导频插入与信道参数估计的方法。 Object of the present invention: To achieve higher channel estimation accuracy, a method is proposed for the pilot to achieve a simple frequency-domain equalization insertion channel parameter estimation.

本发明提出的用于频域均衡的导频插入与信道参数估计的方法,其特征在于:在无线通信系统中,在发送端,编码调制映射后的每个数据信号序列与导频信号序列复接形成一个数据块,经正交变换和交织后,再对每个数据块加循环前缀;在接收端,再去掉循环前缀,以消除多径时延扩展引起的数据块之间的干扰和频域信号间的干扰,以便利用提取出的频域导频信息作信道估计;具体而言,它依次含有如下步骤,发送端:(1)把编码调制映射后的数据信号序列与导频信号序列复接:把已知的随机信号序列X即时域的导频信号序列按顺序均匀的插入到一个编码调制后的数据序列D中,形成一个数据块s,设X长度为M,D长度为(M·NM),其中M和N为正整数,则数据块s长度为NB=M·N,记X=(x0,x1,ΛxM-1),D=(d0,d1,ΛdM·NM-1),s=(s0,s1,s2,s3,ΛsM·N-1),每两个导频 Turned by the present invention for the frequency domain equalization inserted pilot channel parameter estimation method, comprising: in a wireless communication system, each data after the transmitting side, the modulated coded signal sequence mapped to the multiplexed pilot signal sequence then forming a data block, and interleaved by orthogonal transformation, and then adding a cyclic prefix for each data block; at the receiving end, and then remove the cyclic prefix, in order to eliminate interference between data blocks due to multipath delay spread and frequency frequency-domain pilot interference between domain signal, in order to utilize the extracted pilot information for channel estimation; in particular, which in turn comprises the steps of transmitting side: (1) the data signal sequence and the pilot coded modulation mapping pilot signal sequence multiplexing: the known pilot signal sequence X random time domain pilot signal sequence is inserted into a uniform order of data sequence D after a code modulation, to form a data block of s, let X of length M, D is a length ( M · NM), where M and N are positive integers, then the data block s of length NB = M · N, denoted X = (x0, x1, ΛxM-1), D = (d0, d1, ΛdM · NM-1 ), s = (s0, s1, s2, s3, ΛsM · N-1), each of the two pilot 号之间有N-1个数据,设把第0个导频信号x0插在第L个位置,0≤L≤N-1,则有 There between the N-1 number of data, the set of pilot signals x0 0 is inserted in the L-th position, 0≤L≤N-1, there is

(2)把上述复接后的信号进行N维正交变换;(3)把上述正交变换得到的信号作交织操作以改变信号序列的顺序,即按列写入矩阵中,再按行读出;(4)对上述交织后的数据块加循环前缀;接收端:(3)去掉在发送端加的循环前缀;(4)信道参数估计:先对导频信息所有频点作信道参数估计,然后再通过插值对数据信息的频点作信道参数估计,以得到所有频点的信道频域特性;在上述方法中,所述的N维正交变换可以表示为:BN×1=W·AN×1,其中A是N维向量,由N-1个数据信号与一个导频信号组成的,B是正交变换后的信号向量,W是N维正交变换对应的N×N的矩阵,若导频信号位于向量A的第L个位置,则矩阵W第L列的N个元素满足:wn,L=1N·ej2πknN,n=0,1,2ΛN-1]]>的形式,其中wn,L是矩阵W的第n行第L列的元素,k是一个任意的整数;在上述方法中,所述的交织操作改 (2) The above-described complex signal then performing N-dimensional orthogonal transform; (3) the signal of the orthogonal transform operation to change the order of interleaving for the signal sequence, i.e., written by columns in a matrix, then the read line out; (4) adding a cyclic prefix of data blocks after the interleaving; receiving terminal: (3) removing the transmission side add cyclic prefix; (4) channel parameter estimation: the first pair of guide All frequencies for channel parameter frequency estimation then for channel parameter estimation by interpolation frequency data information to obtain all the frequency channels in the frequency domain characteristics; in the above method, the N-dimensional orthogonal transformation can be expressed as: BN × 1 = W · aN × 1, wherein a is an N-dimensional vector, by the N-1 data signal and a pilot signal component, B is the signal vector orthogonal transform, W is N × N matrix corresponding to the N-dimensional orthogonal transform If the pilot signal is located in the L position vector a, the matrix W of the L columns of N elements satisfy: wn, L = 1N & CenterDot; ej2 & pi; knN, n = 0,1,2 & Lambda; N-1]]> form, Wn of wherein, L is an element n-th row of matrix W L columns, k is an arbitrary integer; in the above method, the change of the interleaving operation 信号序列的顺序,先把数据顺序的按列写入矩阵中,每列N个数据,共写M列,再按行读出,设交织前后的数据块为q=(q0,q1,q2,q3,ΛqM·N-1)T和g=(g0,g1,g2,g3,ΛgM·N-1)T,其中T表示向量转置,则:gn·M+m=qm·N+n,(0≤m≤M-1,0≤n≤N-1);在上述方法中,所述的加循环前缀操作是对交织后的数据块g的操作,设循环前缀长度为Ng,将g的最后Ng个信号复制到数据块g的前端,交织后的数据块g的长度等于NB,则生成一个包含有循环前缀的长为Ns=NB+Ng的数据块f,即f=TCP·g其中TCP=ICPINB]]>用INB表示NB×NB的单位矩阵,用ICP表示INB的最后Ng行,TCP是由ICP和INB组合在一起而形成的矩阵。 Signal sequence order, the order of data by first writing a column matrix, N columns of data each, write a total of M columns, then row readout, data blocks before and after interleaving is set q = (q0, q1, q2, q3, ΛqM · N-1) T and g = (g0, g1, g2, g3, ΛgM · N-1) T, where T represents the vector transpose, then: gn · m + m = qm · N + n, (0≤m≤M-1,0≤n≤N-1); in the above method, the cyclic prefix adding operation is an operation on the data of the interleaved blocks g, Ng, provided the cyclic prefix length, the g the last Ng replication signal to the front end block g, a data block interleaving g of a length equal to NB, in turn generates a cyclic prefix length Ns = NB + Ng data block F, i.e., f = TCP · g wherein the TCP = ICPINB]]> represents a NB × NB matrix, represented by ICP INB last Ng rows, TCP and INB by the ICP matrix composition formed by INB together.

实验证明采用这种方法,能够以较低的复杂度,获得较高的信道估计和均衡的精度,从而提高基带传输的信噪比,有利于系统采用高进制调制达到高效的频谱利用率。 Experiments show that using this method, it is possible to lower complexity, to obtain higher channel estimation and equalization accuracy, thereby improving the signal to noise ratio of the baseband transmission, facilitate high-ary modulation system to achieve efficient spectrum utilization.

附图说明 BRIEF DESCRIPTION

图1表示本发明的发送端基带数字信号处理程序流程图。 1 shows the transmission of the baseband digital signal processing program of the present invention. FIG.

图2表示本发明的接收端基带数字信号处理程序流程图。 Figure 2 shows the receiving end the baseband digital signal processing program of the present invention. FIG.

图3表示交织算法数据读写的顺序。 Figure 3 shows the sequence of read and write data interleaving algorithm.

具体实施方式 Detailed ways

通信系统中利用多个相互隔离且相对独立的子载波,各载波结构基本相同,每个子载波的基带数字信号处理采用本发明所述的方法,即编码调制映射后的每个数据信号序列与导频信号序列复接形成一个数据块,通过正交变换和交织,使得导频信息只出现在某些频点上,而数据没有这些频率分量,再对每个数据块在发端加循环前缀,收端将其去掉,从而消除了多径时延扩展引起的数据块之间的干扰和频域信号间的干扰,保证了接收到的导频信息在频域不受数据信号的干扰,进而利用提取出的频域导频信息作信道估计和频域均衡,再解交织和解正交变换得到发送的数据信号。 Communication system utilizing a plurality of isolated and independent sub-carriers, each carrier structure is substantially the same, the baseband digital signal processing for each subcarrier using the method according to the present invention, i.e., each of the data signal sequence and the coded modulation mapping guide multiplexing the pilot signal sequence to form a data block and interleaved by orthogonal transform, such that the pilot information appears only at some frequency, and the frequency components of these data is not, then for each data block in the originating add cyclic prefix, received end which is removed, thereby eliminating the interference between the interference and the frequency domain signal between the blocks caused by multipath delay spread, it ensures that the information received pilot interference from the data signal in the frequency domain, and then using the extracted a frequency-domain pilot pilot information for channel estimation and frequency domain equalization, deinterleave reconciliation then orthogonally transformed data to obtain the transmission signal.

本发明所提供的导频插入与信道估计方法在发送端的基带数字信号处理过程如下:1、数据与导频复接首先把已知的随机信号序列X即时域的导频信号序列按顺序均匀的插入到一个编码调制后的数据序列D中,形成一个数据块s,设X长度为M,D长度为(M·NM),其中M和N为正整数,则数据块s长度为NB=M·N。 Guides of the present invention provides for inserting pilot and channel estimation method in the digital signal processing of the transmit end a baseband follows: 1, data and pilot multiplexing is first turned known random signal sequence X time domain pilot signal sequences even in the order of inserted into the data sequence D after a code modulation, to form a data block s, set X of length M, D the length of (M · NM), where M and N are positive integers, then the data block s of length NB = M · N. 记X=(x0,x1,ΛxM-1),D=(d0,d1,ΛdM·NM-1),s=(s0,s1,s2,s3,ΛsM·N-1),每两个导频信号之间有N-1个数据,设把第0个导频信号x0插在第L个位置,0≤L≤N-1,则s=(s0,s1,s2,s3,ΛsM·N-1)=(d0,ΛdL-1,x0,dL,Λ,dN+L-2,x1,dN+L-1,ΛdM·NM-1)用公式表示导频插入操作: Note X = (x0, x1, ΛxM-1), D = (d0, d1, ΛdM · NM-1), s = (s0, s1, s2, s3, ΛsM · N-1), each of the two pilot there between the N-1 data signals, the set of pilot signals x0 0 is inserted in the L-th position, 0≤L≤N-1, then s = (s0, s1, s2, s3, ΛsM · N- 1) = (d0, ΛdL-1, x0, dL, Λ, dN + L-2, x1, dN + L-1, ΛdM · NM-1) represented by the formula pilot inserting operation: 2、N维正交变换q=q0q1MqM·N-1=Wρ·s=W0Λ00WΛ0MMOM00ΛW·s0s1MsM·N-1.]]>其中W是N维正交变换矩阵, 2, N-dimensional orthogonal transformation q = q0q1MqM & CenterDot; N-1 = W & rho; & CenterDot; s = W0 & Lambda; 00W & Lambda; 0MMOM00 & Lambda; W & CenterDot; s0s1MsM & CenterDot;. N-1]]> where W is N-dimensional orthogonal transform matrix, 的对角线上有M个W矩阵,W的第L列的N个元素满足:wn,L=1N·ej2πknN,n=0,1,2ΛN-1]]> There are on the diagonal matrix W of M, N elements L columns of W satisfying: wn, L = 1N & CenterDot; ej2 & pi; knN, n = 0,1,2 & Lambda; N-1]]>

的形式,其中wn,L是矩阵W的第n行第L列的元素,k是一个任意的整数。 Form, Wn of wherein, L is an element n-th row of matrix W L columns, k is an arbitrary integer. 特别注意,作正交变换的矩阵乘法时,W第L列的元素与向量s中的导频信号相乘,其他的N-1列元素与数据信号相乘。 Special attention, as orthogonal transform matrix multiplication, the guide elements and the vector W s of the L columns in the frequency multiplied signal, multiplying the other elements of row N-1 data signal.

3、交织交织操作改变信号序列的顺序,先把数据顺序的按列写入矩阵中,每列N个数据,共写M列,再按行读出,设交织后的数据块为g=(g0,g1,g2,g3,ΛgNB-1)T,]]>用J表示交织操作,即g=J·q,用公式表示为:gn·M+m=qm·N+n,(0≤m≤M-1,0≤n≤N-1)。 3, the operation of changing the order of deinterleaving a signal sequence, the order of data written in first-column matrix, each row of N data, write a total of M columns, and then read out row, interleaved data block provided as g = ( g0, g1, g2, g3, & Lambda; gNB-1) T,]]> with J represents an interleaving operation, i.e. g = J · q, formulated as: gn · m + m = qm · N + n, ( 0≤m≤M-1,0≤n≤N-1).

4、加循环前缀将数据块的最后Ng个信号复制到其前端,若原数据块长为NB,则生成一个包含有循环前缀的长为Ns=NB+Ng的数据块f,即f=TCP·g=TCP·J·Wρ·s---(1)]]>其中TCP=ICPINB]]>TCP是由NB×NB的单位矩阵INB和INB的最后Ng行(用ICP表示)组合在一起而形成的矩阵。 4, adding the cyclic prefix to copy the last Ng signal data block to the front end thereof, if the original data block length NB, generating a cyclic prefix of length Ns = NB + Ng data block f comprises, i.e. f = TCP & CenterDot; g = TCP & CenterDot; J & CenterDot; W & rho; & CenterDot; s --- (1)]]> where TCP = ICPINB]]> TCP is (expressed as ICP) by the last Ng rows NB × NB unit matrix INB and INB are combined in together form a matrix.

下面用矩阵表示信号在多径信道下的数据块的传输,这里使用的是系统的离散等效模型,假定在接收端,包括载波、码元和采样时钟各种同步都是准确的,则当tm=mT时可得采样信号:v′(tm)=Σk=-∞+∞f(k)h(mT-kT,mT)+n(mT)---(2)]]>其中h(τ,t)=Σl=1Lhl(t)·δ(τ-τl)]]>为信道的时域冲击响应,n(t)为信道噪声。 The following matrix represents the transmission data block signals in the multipath channel, the system used here is the equivalent discrete model, it is assumed that the receiving side, comprising a carrier, a variety of symbols and the sampling clock synchronization is accurate, then when available tm mT when = the sampled signal: v & prime; (tm) = & Sigma; k = - & infin; + & infin; f (k) h (mT-kT, mT) + n (mT) --- (2)]] > where h (& tau;, t) = & Sigma; l = 1Lhl (t) & CenterDot; & delta; (& tau; - & tau; l)]]> impulse response is a time-domain channel, n (t) is the channel noise.

假定信道冲击响应是慢变化的,即在一个数据块内,假定信道是不变的,但从一个数据块到下一个数据块,信道是变化的。 Assuming that the channel is slowly varying impulse response, i.e. in a data block, it assumes that the channel is constant, but a data block to the next block of data, the channel is varied. 另外,信道冲击响应的采样序列在精度允许的范围内总可以被近似成具有有限长度。 Further, the channel impulse response sample sequence in the range of the total allowable accuracy can be approximated to have a finite length. 假定该序列的长度为K+1,则可以用向量hi=[hi(0),Λhi(K)]T来表示在第i个数据块内的整个信道冲击响应序列,其中hi(m)=h(mT,iTNs),其中T·Ns是包括循环前缀在内的整个数据块的长度。 Assumed length of the sequence of K + 1, you can use the vector hi = [hi (0), Λhi (K)] T to represent the entire channel impulse within the i-th data block in response to sequence, wherein hi (m) = h (mT, iTNs), wherein T · Ns is the length of the entire block, including including a cyclic prefix. 为了消除数据块之间的干扰,在系统设计中,循环前缀的长度总是要取得大于或等于信道的最大附加延时,即要满足Ng≥K。 In order to eliminate the interference between data blocks, in the system design, the length of the cyclic prefix to obtain the maximum additional delay is always greater than or equal to the channel, i.e. to meet Ng≥K.

use 表示发送的第i个数据块, It represents i-th data block is transmitted, 表示接收的第i个数据块,由(2)式,并利用信道冲击响应为有限长度的特性,可以得到发射块和接收块之间的关系如下: Indicates the i-th received data block by the formula (2), and using the channel impulse response as characteristic of finite length can be obtained the relationship between the transmit and receive blocks are as follows: 由于多径信道引入的数据块之间的干扰,使 Due to interference between the data blocks of the multipath introduced by the channel, so that 将同时依赖于 It will also depend on the with 式中 Where 是相应的噪声向量,C0和C1分别是具有如下形式的Ns×Ns矩阵,其对角线上的元素都相等:C0=hi(0)00Λ0Mhi(0)0Λ0hi(k)ΛOΛMMOMO00Λhi(K)Λhi(0)]]>C1=0Λhi(K)Λhi(1)MO0OM0ΛOΛhi(K)MOMO00Λ0Λ0]]>接收端的基带数字信号处理过程如下:1、去掉循环前缀 Is the corresponding noise vector, C0, and C1 are Ns × Ns matrix has a form, which elements on the diagonal are equal: C0 = hi (0) 00 & Lambda; 0Mhi (0) 0 & Lambda; 0hi (k) & Lambda; O & Lambda; MMOMO00 & Lambda; hi (K) & Lambda; hi (0)]]> C1 = 0 & Lambda; hi (K) & Lambda; hi MO0OM0 & Lambda (1); O & Lambda; hi (K) MOMO00 & Lambda; 0 & Lambda; 0]]> the receiving end the baseband digital signals are as follows: 1, the cyclic prefix is ​​removed 其中 among them 表示去除循环前缀操作, It represents a cyclic prefix removal operation, 为NB×Ng的零矩阵,INB为NB×NB的单位矩阵。 NB × Ng is the zero matrix, INB NB × NB matrix unit.

可见,当Ng≥K时,去除循环前缀操作也将消除块间干扰, Seen, when Ng≥K, cyclic prefix removal action will eliminate inter-block interference, 将只依赖于 We will only depend on the 此时,我们就可以对各个数据块进行单独解调,表示数据块序号的记号i将被省略。 In this case, we can individually demodulate each data block represents the block number i tokens will be omitted.

把(1)式代入(3)式可得: The (1) is substituted into the formula (3) can be obtained: 其中 among them 是噪声向量, Is the noise vector, 是复合信道冲击响应,是一个NB×NB的循环矩阵,具有以下特殊形式Cρ=hi(0)0Λ0hi(K)Λhi(1)Mhi(0)0ΛOOMMMOOΛOhi(K)hi(K)MMOOΛ00hi(K)MMOOMMOOMMO00Λ0hi(K)ΛΛhi(0)]]> A composite channel impulse response, is a cyclic matrix of a NB × NB has the following special form C & rho; = hi (0) 0 & Lambda; 0hi (K) & Lambda; hi (1) Mhi (0) 0 & Lambda; OOMMMOO & Lambda; Ohi (K) hi (K) MMOO & Lambda; 00hi (K) MMOOMMOOMMO00 & Lambda; 0hi (K) & Lambda; & Lambda; hi (0)]]>

也就是 That is 的第(k,l)个元素为hi((kl)mod NB),这里mod表示取模运算。 Of (k, l) th element hi ((kl) mod NB), where mod denotes a modulo operation.

利用循环矩阵的性质,可以分解成如下形式: The cyclic nature of the matrix can be decomposed into the following form: 其中F和F-1分别是NB×NB的傅立叶变换和傅立叶反变换矩阵,Λ是复合信道的传输函数矩阵,为一个对角阵。 Wherein F and F-1 are respectively NB × NB Fourier transform and inverse Fourier transform matrix, a transfer function Lambda composite channel matrix, a diagonal matrix. 若h=[h(0),Λh(K),0,Λ,0]T的NB点傅立叶变换为H=[H(0),ΛH(NB-1)]T,则Λ=H(0)0Λ00H(1)Λ0MMOM00ΛH(NB-1)NB×NB]]>把(5)式代入(4)式可得去掉循环前缀后的接收信号为: If h = [h (0), Λh (K), 0, Λ, 0] NB point T Fourier transform H = [H (0), ΛH (NB-1)] T, the & Lambda; = H ( 0) 0 & Lambda; 00H (1) & Lambda; 0MMOM00 & Lambda; H (NB-1) NB & times; NB]]> received signals to (5) into (4), we have to remove the cyclic prefix is: 2、信道估计通过傅立叶变换得到频域的接收信号: 2, the channel estimation of the reception signal obtained by Fourier transform in the frequency domain: 由(6)式可知 It is seen from (6) VmN+k=N·H(mN+k)·pm+nmN+k,0≤m≤M-1]]>其中 VmN + k = N & CenterDot; H (mN + k) & CenterDot; pm + nmN + k, 0 & le; m & le; M-1]]> wherein 是频域的噪声矢量,p0p1MpM-1=F·x0x1·e-j2πk/MNMxM-1·e-j2πk(M-1)/MN]]>是发送的频域导频信息。 It is a frequency noise vector domain, p0p1MpM-1 = F & CenterDot; x0x1 & CenterDot; e-j2 & pi; k / MNMxM-1 & CenterDot; e-j2 & pi; k (M-1) / MN]]> is the frequency domain pilot transmitted information.

F是傅立叶变换矩阵,(x0,x1,ΛxM-1)是与发送端相同的时域导频信号,xne-j2πk·n/MN中的k等于正交变换矩阵W中的参数k。 F is the Fourier transform matrix, (x0, x1, ΛxM-1) is the same as the transmission end time domain pilot signals, xne-j2πk · n / MN is equal to the parameter k W k in the orthogonal transform matrix.

接收的导频信息只出现在mN+k(m为整数)的频点上,数据不存在这些频率分量,提取出这些频点上的信号,{p^m|p^m=VmN+k,m=0,1,2,KM-1},]]>根据接收到的导频信息和发送的导频信息在频域上的变化来推断信道相应频点的频域特性,也就是用接收到的导频频域信息除以发送的导频频域信息,可以得到在这些频点上的信道频率响应H(mN+k)的最小二乘估计, Receiving pilot appears only mN + k (m is an integer) frequency data of these frequency components do not exist, the extracted signal, {p at these frequencies ^ m | p ^ m = VmN + k, m = 0,1,2, KM-1},]]> the frequency change information and pilot information transmitted in the frequency domain pilot received to infer the corresponding frequency-domain channel frequency characteristic, i.e. reception leads to the pilot frequency domain information transmitted by dividing the frequency domain information can be obtained at the channel frequency response frequency of these least-squares estimation H (mN + k), and

H^(mN+k)=p^mN·pm,0≤m≤M-1]]>为了得到其它频点的信道频域特性,需要进行频域插值,采用变换域插值的方法,详细地说就是再把M个信道估计值做M点的反傅立叶变换变换到时域,hn=1MΣm=0M-1H^(mN+k)·ej2π·m·nM,0≤n≤M-1,]]>在时域信道估计序列中插入M×(N-1)个零, H ^ (mN + k) = p ^ mN & CenterDot; pm, 0 & le; m & le; M-1]]> In order to obtain a channel other frequency points in the frequency domain characteristics, the need for frequency domain interpolation, the method transform domain interpolation, detail that is to then m channel estimation values ​​do m-point inverse Fourier transform to the time domain, hn = 1M & Sigma; m = 0M-1H ^ (mN + k) & CenterDot; ej2 & pi; & CenterDot; m & CenterDot; nM, 0 & le; n & le; M-1,]]> in the time domain channel estimation sequence inserted M × (N-1) zeros, 用公式表示: Represented by the formula: U=I000MM000IMN×M]]>其中单位阵I的大小为M2×M2,]]>同样大小的全零矩阵0共(4·N-2)个。 U = I000MM000IMN & times; M]]> where I is the identity matrix size M2 & times; M2,]]> all-zero matrix of the same size were 0 (4 · N-2) th.

这样得到了NB点的时域信道估计 This gave the NB point time domain channel estimation 再把它作傅立叶变换变换到频域,就得到了所有频点的信道估计值,H^i=Σn=0NB-1h′n·e-j2π·i·nNB,0≤i≤NB-1.]]>3、频域均衡利用所得的信道估计,并把接收到的信号进行傅立叶变换,变换到频域,再根据设计要求,选择一种现有的频域均衡算法,即可实现接收信号的频域均衡。 Then it is the Fourier transformed into the frequency domain, to obtain all the frequencies of a channel estimate, H ^ i = & Sigma; n = 0NB-1h & prime; n & CenterDot; e-j2 & pi; & CenterDot; i & CenterDot; nNB, 0 & le; i & le; NB-1.]]> 3, the resulting frequency domain equalization using the channel estimation and the received signal is Fourier transformed to the frequency domain, and then depending on design requirements, to select a conventional frequency domain equalization algorithm, the received signal may be implemented in the frequency domain equalization.

4、提取数据信号先对均衡后的数据流y=(y0,y1,y2,y3,ΛyM·N-1)T做解交织操作,完成与发送端交织操作相反的处理,先把数据按行写入矩阵中,每行M个数据,共写N行,再按列读出,解交织后得到z=(z0,z1,z2,z3,ΛzM·N-1)T,用公式表示为:zm·N+n=yn·M+m,(0≤m≤M-1,0≤n≤N-1)。 4, extracts a data signal to data stream equalized y = (y0, y1, y2, y3, ΛyM · N-1) T do deinterleaving operation, and performs the reverse operation of the transmitting side processing interleaving, first data row writing the matrix, each row of M data, write a total of N lines, then the column read out after deinterleaving to obtain z = (z0, z1, z2, z3, ΛzM · N-1) T, as represented by the formula: zm · N + n = yn · m + m, (0≤m≤M-1,0≤n≤N-1).

解交织后再作解正交变换,用矩阵 Solutions for deinterleaving after orthogonal transform matrix 的逆矩阵(也就是 The inverse matrix (that is, 的共轭转置 The conjugate transpose )乘以数据块向量。 ) Multiplied by the data block vector.

b=b0b1MbM·N-1=WρH·z=WH0Λ00WHΛ0MMOM00ΛWH·z0z1MzM·N-1.]]>最后去除导频信号,也就是去掉b中N的整倍数加L即mN+L(0≤m≤M-1位置上的信号,得到原发送数据信号的估计值D^=(d^0,d^1,Λd^MNM-1),]]>其中 b = b0b1MbM & CenterDot; N-1 = W & rho; H & CenterDot; z = WH0 & Lambda; 00WH & Lambda; 0MMOM00 & Lambda; WH & CenterDot; z0z1MzM & CenterDot;. N-1]]> Finally, removal of the pilot signal, i.e. to remove b, N is an integer multiple plus L i.e. mN + L (a position signal 0≤m≤M-1, to give estimates of the original transmitted data signal D ^ = (d ^ 0, d ^ 1, & Lambda; d ^ MNM-1),]]> wherein 把所得的数据信号输出给解映射和解码器,从而完成基带数字信号的接收处理。 The resulting data signal is output to the demapper and the decoder, thereby completing the reception process base band digital signals.

下面结合附图和实施例进一步说明本发明。 The present invention is further described below in conjunction with the accompanying drawings and embodiments.

在实施例中,正交变换矩阵采用哈达码(Hadamar)矩阵WN,WN是2n×2n维的,要求N=2n;元素只有1和-1,从1阶到2n阶的哈达码矩阵可利用矩阵的克罗内克积运算递推得到,即WN=W2⊗WN2,]]>其中:W2=12111-1.]]>哈达码变换是正交变换参数k=0、L=0的一个特例,其中k=0使得信道估计操作中xnej2πk·n/MN导频信号的加权系数均等于1,实现最简单。 In an embodiment, orthogonal transform matrix using Hadamard (Hadamar) matrix WN, WN is 2n × 2n dimensions, requires N = 2n; element has only 1 and -1, the Hadamard matrix of order 1 to order may utilize 2n recursive operation matrix Kronecker product obtained, i.e. WN = W2 & CircleTimes; WN2,]]> where:. W2 = 12111-1]]> Hadamard transform is orthogonal transform parameter k = 0, L = 0, a Specific examples, where k = 0 so that the channel estimation weighting coefficient xnej2πk · n / MN pilot signal operations are equal to 1, to achieve the most simple.

实施例中采用正交变换的维数N=4,4阶哈达码矩阵为W4=W2⊗W2=1411111-11-111-1-11-1-11.]]>数据块长度NB=M·N=128,其中M=32,循环前缀长度Ng=26。 Dimension Example orthogonal transform N = 4,4-order Hadamard matrix is ​​W4 = W2 & CircleTimes;. W2 = 1411111-11-111-1-11-1-11]]> data block length NB = M · N = 128, where M = 32, the cyclic prefix length Ng = 26.

具体的导频插入与信道估计方法的主要过程如下:1、数据与导频复接首先把已知的随机信号序列X=(x0,x1,Λx31)按顺序均匀的插入到编码调制后的数据序列D=(d0,d1,Λd95)中,形成一个数据块s,s=(s0,s1,s2,s3,Λs127)。 Specific pilot insertion of the main process estimation channel as follows: 1, data and pilot multiplexed data after the first of the known random signal sequence X = (x0, x1, Λx31) uniformly inserted in order into the code modulated sequence D = (d0, d1, Λd95) to form a data block s, s = (s0, s1, s2, s3, Λs127). 其中把第0个导频信号x0插在第0个位置则s=(s0,s1,s2,s3,Λs127)=(x0,d0,d1,d2,x1,d3,d4,d5,x2,d6,Λx31,d93,d94,d95)。 Wherein the 0-th pilot signal x0 is inserted in the 0th position is s = (s0, s1, s2, s3, Λs127) = (x0, d0, d1, d2, x1, d3, d4, d5, x2, d6 , Λx31, d93, d94, d95).

2、N维正交变换对数据块做N维正交变换,具体采用的是4阶哈达码变换。 2, N-dimensional orthogonal transform for N-dimensional orthogonal transformation made to the data block uses a specific fourth-order Hadamard transform.

q=q0q1Mq127=Wρ·sW40Λ00W4Λ0MMOM00ΛW4·s0s1Ms127.]]>其中W4=1211111-11-111-1-11-1-11]]>3、交织交织操作改变信号序列的顺序,如图3所示,先把数据顺序的按列写入矩阵中,每列4个数据,共写32列,再按行读出,交织后的数据块:g=(g0,g1,g2,g3,Λg127)T=(q0,q4,q8,q12,Λq124,q1,q5,q9,q13,Λq125,q2,q6,q10,Λq126,q3,q7,q11,Λq127)T,其中gn·32+m=qm·4+n,(0≤m≤31,0≤n≤3)。 q = q0q1Mq127 = W & rho; & CenterDot; sW40 & Lambda; 00W4 & Lambda; 0MMOM00 & Lambda; W4 & CenterDot;. s0s1Ms127]]> wherein W4 = 1211111-11-111-1-11-1-11]]> 3, the operation of changing the interleaving interleaved order signal sequence , 3, the first data is sequentially written in the column of the matrix, each column of four data, write a total of 32, then the readout row, the interleaved data block: g = (g0, g1, g2 , g3, Λg127) T = (q0, q4, q8, q12, Λq124, q1, q5, q9, q13, Λq125, q2, q6, q10, Λq126, q3, q7, q11, Λq127) T, where gn · 32 + m = qm · 4 + n, (0≤m≤31,0≤n≤3).

4、加循环前缀将数据块的最后Ng=26个数据复制到其前端,生成一个包含有循环前缀的长为Ns=NB+Ng=154的数据块f,即f=(f0,f1,f2,f3,Λf153)T=(g102,g103,Λg127,g0,g1,g2,g3,Λg127)T。 4, adding the cyclic prefix to copy the last Ng = 26 data blocks to its front end, in turn generates a cyclic prefix length Ns = NB + Ng = 154 data block F, i.e., f = (f0, f1, f2 , f3, Λf153) T = (g102, g103, Λg127, g0, g1, g2, g3, Λg127) T.

所得的基带数字信号经过成形滤波和上变频,经过信道发射到接收端,接收端采用下变频、匹配滤波和采样,假定在接收端,包括载波、码元和采样时钟各种同步都是准确的,则可以得到对应的基带数字接收信号v′=(v′0,v′1,v′2,v′3,Λv′153)T。 The resulting baseband digital signal through shaping filters, and upconverts, channel-transmitted to the receiving end, receiving end down-conversion, matched filtering and sampling, it is assumed that the receiving side, comprising a carrier, symbols, and the sampling clock various synchronization is accurate , can be obtained corresponding to the baseband digital reception signal v '= (v'0, v'1, v'2, v'3, Λv'153) T.

接收信号序列处理过程如下:1、去掉循环前缀vρ=(v0,v1,v2Λv127)T=(v′26,v′26,v′28,Λv′153)T.]]>2、信道估计通过傅立叶变换得到频域信号: Received signal sequence processing is as follows: 1, remove the cyclic prefix v & rho; = (v0, v1, v2 & Lambda; v127) T = (v & prime; 26, v & prime; 26, v & prime; 28, & Lambda; v & prime; 153) T]]>. 2, the estimated channel frequency-domain signals by Fourier transform: Vi=Σm=0127vm·e-j2π·m·i/128,0≤i≤127.]]>因为k=0,所以发送的频域导频信息: Vi = & Sigma; m = 0127vm & CenterDot; e-j2 & pi; & CenterDot; m & CenterDot; i / 128,0 & le; i & le; 127]]> as k = 0, so the frequency domain pilot transmitted information:

p0p1Mp31=F·x0x1·e-j2πk/128Mx31·e-j2πk31/128F·x0x1Mx31,]]>pi=Σm=031xm·e-j2π·m·i/32,0≤i≤31]]>接收的导频信息只出现在mN+k=4m(m为整数,0≤m≤31)的频点上,数据不存在这些频率分量,提取出这些频点上的导频信息,{p^m|p^m=V4m,m=0,1,2,KM-1},]]>利用除法可以得到在这些频点上的信道频率响应的最小二乘估计,H^(4m)=p^m2·pm,0≤m≤31]]>把信道估计值做32点的反傅立叶变换变换到时域,h0=132Σm=031H^(4m)·ej2π·m·nM,0≤n≤31,]]>在时域信道估计序列中间插入M×(N-1)=96个零, p0p1Mp31 = F & CenterDot; x0x1 & CenterDot; e-j2 & pi; k / 128Mx31 & CenterDot; e-j2 & pi; k31 / 128F & CenterDot; x0x1Mx31,]]> pi = & Sigma; m = 031xm & CenterDot; e-j2 & pi; & CenterDot; m & CenterDot; i / 32,0 & le; i & le ; 31]]> the received pilot information to appear only on frequency mN + k = 4m (m is an integer, 0≤m≤31), the frequency components of these data do not exist, on the extracted pilot frequency frequency least squares estimation,]]> division can be obtained by using these frequency channel frequency response, H ^ | information, {p ^ m = V4m, m = 0,1,2, KM-1 p ^ m} (4m) = p ^ m2 & CenterDot; pm, 0 & le; m & le; 31]]> the channel estimation values ​​do inverse Fourier transform 32 points to the time domain, h0 = 132 & Sigma; m = 031H ^ (4m) & CenterDot; ej2 & pi; & CenterDot ; m & CenterDot; nM, 0 & le; n & le; 31,]]> in the time domain channel estimation sequence intervening m × (N-1) = 96 zeros, 这样得到了128点的时域信息 Thus obtained time domain information 128 points 再把它作傅立叶变换变换到频域,就得到了所有频点的信道估计值,H^i=Σn=0127h′n·e-j2π·i·n128,0≤i≤128.]]>3、频域均衡均衡可以采用迫零均衡或最小均方误差均衡等。 Then it is the Fourier transformed into the frequency domain, to obtain all channel estimation values ​​in the frequency point, H ^ i = & Sigma; n = 0127h & prime; n & CenterDot; e-j2 & pi; & CenterDot; i & CenterDot; n128,0 & le; i & le; 128. ]]> 3, frequency domain equalization may be zero forcing equalization equalization or minimum mean square error equalization. 以迫零均衡为例,加权系数等于信道传输函数的倒数,把接收的信号进行傅立叶变换到频域,进行频域加权,再反变换到时域,以实现对整个系统的传递函数进行校正,使传输接近无失真条件。 In a zero forcing equalizer for example, the weighting factor equal to the inverse channel transfer function, the received signal is Fourier transformed to the frequency domain, frequency domain weighting, then inverse transformed to the time domain in order to achieve the transfer function of the whole system is corrected, close the transmission distortion-free conditions.

接收的去掉循环前缀后的数据块作傅立叶变换得到频域接收信号: Block frequency domain the Fourier transform of the received signal after removing the cyclic prefix received: Vi=Σm=0127vm·e-j2π·m·i/128,0≤i≤127.]]>频域加权操作:ηi=ViH^i,0≤i≤127]]>反傅立叶变换到时域: Vi = & Sigma; m = 0127vm & CenterDot; e-j2 & pi; & CenterDot; m & CenterDot; i / 128,0 & le; i & le; 127]]> frequency-domain weighting operation:. & Eta; i = ViH ^ i, 0 & le; i & le; 127]]> inverse Fourier transform to the time domain:

yn=1128Σn=0127ηi·ej2π·i·n128,0≤n≤127]]>4、提取数据信号先对均衡后的数据流做解交织操作,完成与发送端交织相反的处理,先把数据按行写入矩阵中,每行32个数据,共写4行,再按列读出,解交织后得到z=(z0,z1,z2,z3,Λz127)T=(y0,y32,y64,y96,y1,y33,y65,y97,Λy31,y63,y95,y127)T,其中z4m+n=y32n+m,(0≤m≤31,0≤n≤3)。 yn = 1128 & Sigma; n = 0127 & eta; i & CenterDot; ej2 & pi; & CenterDot; i & CenterDot; n128,0 & le; n & le; 127]]> 4, extracts a data signal to the data equalized flows do deinterleaving operation to complete the transmitting side interleaver opposite processing, data are first written row matrix, each row of data 32, write a total of 4 lines, then the column read out after deinterleaving to obtain z = (z0, z1, z2, z3, Λz127) T = (y0 , y32, y64, y96, y1, y33, y65, y97, Λy31, y63, y95, y127) T, where z4m + n = y32n + m, (0≤m≤31,0≤n≤3).

解交织后再作解正交变换,用矩阵 Solutions for deinterleaving after orthogonal transform matrix 的逆矩阵(也就是 The inverse matrix (that is, 的共轭转置 The conjugate transpose )乘以数据块向量。 ) Multiplied by the data block vector.

b=b0b1Mb127=WρH·z=W4H0Λ00W4HΛ0MMOM00ΛW4H·z0z1Mz127.]]>哈达码矩阵是对称实数矩阵,共轭转置就等于它本身,W4H=1211111-11-111-1-11-1-11.]]>最后去除导频信号,也就是去掉b中4的整倍数位置上的信号,得到原发送数据信号的估计值D^=(d^0,d^1,Λd^95)=(b1,b2,b3,b5,b6,b7,b9,b10,Λb127),]]>其中d^3m+n=b4m+n+1,]]>其中0≤m≤31 0≤n≤2。 b = b0b1Mb127 = W & rho; H & CenterDot; z = W4H0 & Lambda; 00W4H & Lambda; 0MMOM00 & Lambda; W4H & CenterDot;. z0z1Mz127]]> Hadamard matrix is ​​symmetric real matrix, conjugate transpose is equal to itself, W4H = 1211111-11-111-1- 11-1-11.]]> Finally remove pilot signals, i.e. signals on the position b is an integer multiple of 4 is removed, to give estimates of the original transmitted data signal D ^ = (d ^ 0, d ^ 1, & Lambda ; d ^ 95) = (b1, b2, b3, b5, b6, b7, b9, b10, & Lambda; b127),]]> where d ^ 3m + n = b4m + n + 1,]]> where 0≤ m≤31 0≤n≤2.

最后把所得的数据信号输出给解映射和解码器,从而完成基带数字信号的接收处理。 Finally, the resultant output data signal to the demapper and the decoder, thereby completing the reception process base band digital signals.

此算法可以用计算机程序仿真,也可在各种可编程逻辑器件或专用集成电路中实施。 This algorithm may be simulated by a computer program, it may also be implemented in a variety of programmable logic devices or application-specific integrated circuits.

现有技术和本发明中的导频插入算法分别采用导频信息占用一个正交码道和占用频域上某些频点的方法,在多径信道传输时,前者的导频信号会与数据发生串扰,导致根据导频信号进行的信道估计和均衡的精度较低,后者克服了信道的多径效应,导频信息和数据不会发生串扰,获得了较高的信道估计的精度。 In the prior art and the present invention respectively algorithm pilot insertion method pilot information occupies occupies a certain orthogonal code and frequency channel on the frequency domain, in the multipath channel transmission, the former pilot signal and the data will be crosstalk occurs, resulting in lower accuracy of estimation and equalization based on the channel for the pilot signal, which overcomes the multipath effects of the channel, and pilot information data crosstalk does not occur, to obtain a high accuracy of channel estimation.

Claims (4)

1.用于频域均衡的导频插入与信道参数估计的方法,其特征在于:在无线通信系统中,在发送端,编码调制映射后的每个数据信号序列与导频信号序列复接形成一个数据块,经正交变换和交织后,再对每个数据块加循环前缀;在接收端,再去掉循环前缀,以消除多径时延扩展引起的数据块之间的干扰和频域信号间的干扰,以便利用提取出的频域导频信息作信道估计;具体而言,它依次含有如下步骤,发送端:(1)把编码调制映射后的数据信号序列与导频信号序列复接:把已知的随机信号序列X即时域的导频信号序列按顺序均匀的插入到一个编码调制后的数据序列D中,形成一个数据块s,设X长度为M,D长度为(M·NM),其中M和N为正整数,则数据块s长度为NB=M·N,记X=(x0,x1,ΛxM-1),D=(d0,d1,ΛdM·NM-1),s=(s0,s1,s2,s3,ΛsM·N-1),每两个导频信号之间有N- 1. A method for the frequency domain equalization guide for inserting pilot channel parameter estimation, wherein: in a wireless communication system, each data after the transmitting side, the modulated coded signal sequence mapped to the multiplexing pilot signal sequence is formed a data block, and interleaved by orthogonal transformation, and then adding a cyclic prefix for each data block; at the receiving end, and then remove the cyclic prefix, and to eliminate the interference between the frequency-domain signal block due to multipath delay spread interference among order to using the extracted frequency-domain pilot information for channel estimation; in particular, which in turn comprises the steps of transmitting side: (1) the data signal sequence and the pilot coded modulation mapping pilot signal sequence multiplexing : the known pilot signal sequence X random time domain pilot signal sequence is inserted into a uniform order of data sequence D after a code modulation, to form a data block of s, let X of length M, D is a length (M · the NM), where M and N are positive integers, then the data block s of length NB = M · N, denoted X = (x0, x1, ΛxM-1), D = (d0, d1, ΛdM · NM-1), N- there between s = (s0, s1, s2, s3, ΛsM · N-1), each of the two pilot signals 1个数据,设把第0个导频信号x0插在第L个位置,0≤L≤N-1,则有 1 data, the set of pilot signals x0 0 is inserted in the L-th position, 0≤L≤N-1, there is (2)把上述复接后的信号进行N维正交变换;(3)把上述正交变换得到的信号作交织操作以改变信号序列的顺序,即按列写入矩阵中,再按行读出;(4)对上述交织后的数据块加循环前缀;接收端:(1)去掉在发送端加的循环前缀;(2)信道参数估计:先对导频信息所有频点作信道参数估计,然后再通过插值对数据信息的频点作信道参数估计,以得到所有频点的信道频域特性。 (2) The above-described complex signal then performing N-dimensional orthogonal transform; (3) the signal of the orthogonal transform operation to change the order of interleaving for the signal sequence, i.e., written by columns in a matrix, then the read line out; (4) adding a cyclic prefix of data blocks after the interleaving; receiving end: (1) removing the transmission side add cyclic prefix; (2) the channel parameter estimation: the first pair of guide All frequencies for channel parameter frequency estimation , then the parameter for channel estimation by interpolation frequency information to the data, all the frequency points to obtain a frequency domain characteristic of the channel.
2.根据权利要求1所述的用于频域均衡的导频插入与信道参数估计的方法,其特征在于:所述的N维正交变换可以表示为:BN×1=W·AN×1,其中A是N维向量,由N-1个数据信号与一个导频信号组成的,B是正交变换后的信号向量,W是N维正交变换对应的N×N的矩阵,若导频信号位于向量A的第L个位置,则矩阵W第L列的N个元素满足:wn,L=1N·ej2πknN,n=0,1,2ΛN-1]]>的形式,其中wn,L是矩阵W的第n行第L列的元素,k是一个任意的整数。 2. The method of inserting the pilot channel parameter estimation for the pilot frequency domain equalization according to claim 1, wherein: said N-dimensional orthogonal transformation can be expressed as: BN × 1 = W · AN × 1 wherein a is a N-dimensional vector, by the N-1 data signal and a pilot signal component, B is the signal vector orthogonal transform, W is N × N matrix corresponding to the N-dimensional orthogonal transformation, if the guide pilot signals are located vector a L-th position, the matrix W N elements of the first L columns satisfying: wn, L = 1N & CenterDot; ej2 & pi; knN, n = 0,1,2 & Lambda; N-1]]> form, wherein wn, L is an element of the n-th row of matrix W L columns, k is an arbitrary integer.
3.根据权利要求1所述的用于频域均衡的导频插入与信道参数估计的方法,其特征在于:所述的交织操作改变信号序列的顺序,先把数据顺序的按列写入矩阵中,每列N个数据,共写M列,再按行读出,设交织前后的数据块为q=(q0,q1,q2,q3,ΛqM·N-1)T和g=(g0,g1,g2,g3,ΛgM·N-1)T,其中T表示向量转置,则:gn·M+m=qm·N+n,(0≤m≤M-1,0≤n≤N-1)。 3. The method of inserting the pilot channel parameter estimation for the pilot frequency domain equalization according to claim 1, wherein: the operation of changing the order of the interleaving of the signal sequence, the order of data by first writing a column matrix , each row of N data, write a total of M columns, then row readout, data blocks before and after interleaving is set q = (q0, q1, q2, q3, ΛqM · N-1) T and g = (g0, g1, g2, g3, ΛgM · N-1) T, where T represents the vector transpose, then: gn · m + m = qm · N + n, (0≤m≤M-1,0≤n≤N- 1).
4.根据权利要求1所述的用于频域均衡的导频插入与信道参数估计的方法,其特征在于:所述的加循环前缀操作是对交织后的数据块g的操作,设循环前缀长度为Ng,将g的最后Ng个信号复制到数据块g的前端,交织后的数据块g的长度等于NB,则生成一个包含有循环前缀的长为Ns=NB+Ng的数据块f,即f=TCP·g其中TCP=ICPINB]]>用INB表示NB×NB的单位矩阵,用ICP表示INB的最后Ng行,TCP是由ICP和INB组合在一起而形成的矩阵。 4. The method of inserting the pilot channel parameter estimation for the pilot frequency domain equalization according to claim 1, wherein: the operation is a cyclic prefix is ​​added to the data blocks g interleaving operation, set the cyclic prefix length is Ng, the copy g last Ng signals to the front end block g, the data block interleave g of a length equal to NB, it in turn generates a cyclic prefix length Ns = NB + Ng data block F, i.e. f = TCP · g where TCP = ICPINB]]> INB represented by a NB × NB matrix, represented by ICP INB last Ng rows, TCP and INB by the ICP compositions together to form a matrix.
CN 03137565 2003-06-18 2003-06-18 Method for inserting pilot frequency and estimating parameters of channel of equilibrium in frequency doman CN1200521C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 03137565 CN1200521C (en) 2003-06-18 2003-06-18 Method for inserting pilot frequency and estimating parameters of channel of equilibrium in frequency doman

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 03137565 CN1200521C (en) 2003-06-18 2003-06-18 Method for inserting pilot frequency and estimating parameters of channel of equilibrium in frequency doman

Publications (2)

Publication Number Publication Date
CN1463083A CN1463083A (en) 2003-12-24
CN1200521C true CN1200521C (en) 2005-05-04

Family

ID=29748549

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 03137565 CN1200521C (en) 2003-06-18 2003-06-18 Method for inserting pilot frequency and estimating parameters of channel of equilibrium in frequency doman

Country Status (1)

Country Link
CN (1) CN1200521C (en)

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100896684B1 (en) * 2004-01-27 2009-05-14 삼성전자주식회사 Digital broadcasting transmission/reception capable of improving receiving performance and signal processing method thereof
US8553822B2 (en) * 2004-01-28 2013-10-08 Qualcomm Incorporated Time filtering for excess delay mitigation in OFDM systems
CN100481826C (en) * 2005-10-20 2009-04-22 清华大学 Transmitting and receiving method for signal of orthogonal frequency division multiplexing system
US7813421B2 (en) * 2006-01-17 2010-10-12 Marvell World Trade Ltd. Order recursive computation for a MIMO equalizer
CN100493056C (en) * 2006-08-18 2009-05-27 清华大学 Frequency domain channel estimation method of crossing frequency division multiplexing system with time-domain enveloping weighting
CN100553247C (en) 2006-09-08 2009-10-21 清华大学 Accurate method for estimating frequency deviation of inseting constant pilot carrier in orthogonal frequency division multiplex system
CN101222459B (en) * 2007-01-12 2010-11-10 展讯通信(上海)有限公司 Pilot frequency insertion and channel estimation method of frequency domain equalization system
CN101179549B (en) * 2007-12-14 2011-03-30 清华大学 Communication signal carrier frequency estimation method employing three-point weight interpolation algorithm
CN101364840B (en) * 2008-09-28 2012-05-30 中国科学院声学研究所 Equalizing system and method for changing block length of batch data block
CN101394385B (en) * 2008-11-10 2011-06-22 哈尔滨工业大学 Method for promoting OFDM system based on time domain processing combined channel estimation
CN102457447A (en) * 2010-10-21 2012-05-16 财团法人工业技术研究院 Channel for wireless communication system and method for estimating delay expansion time of channel
CN103427873B (en) * 2012-05-23 2017-12-12 江苏智光创业投资有限公司 The removing method and device of interference
CN102833060B (en) * 2012-08-22 2015-05-13 国网电力科学研究院 Delay equalization method for realization of service code stream irrelevance by automatic mark insertion
CN103338171B (en) * 2013-06-27 2016-06-15 北京大学 A kind of receiving terminal equalization methods based on channel estimation in frequency domain and system
CN106533997B (en) * 2015-09-09 2019-08-16 扬智科技股份有限公司 The calculation method and its decoding apparatus of system signal noise ratio

Also Published As

Publication number Publication date
CN1463083A (en) 2003-12-24

Similar Documents

Publication Publication Date Title
Farhang-Boroujeny OFDM versus filter bank multicarrier
Stitz et al. Pilot-based synchronization and equalization in filter bank multicarrier communications
Wang et al. Wireless multicarrier communications
Van Nee OFDM codes for peak-to-average power reduction and error correction
Koga et al. High-speed power line communication system based on wavelet OFDM
Fischer Precoding and signal shaping for digital transmission
US5367540A (en) Transversal filter for use in a digital subscriber line transmission interface
Jamin et al. Wavelet packet modulation for wireless communications
JP3044899B2 (en) Apparatus for coherent demodulation of the multiplexed digital data to the time-frequency region with the evaluation and limitations determination of the frequency response of the communication channel
JP4099191B2 (en) Channel estimation apparatus and channel estimation method
KR100484050B1 (en) A receiver
JP3796221B2 (en) Data communication apparatus and method based on orthogonal frequency division multiple access
CN100558095C (en) Device and method of estimating multiple channels
Strohmer Approximation of dual Gabor frames, window decay, and wireless communications
US7082159B2 (en) Methods and arrangements in a telecommunications system
CN101326740B (en) Method of transmitting OFDM signal and transmitter and receiver thereof
CN1224228C (en) Multi-carrier transmission system using multiplication by leakage matrix with complex reduction
CN1221083C (en) Receiver for multiuser detection of CDMA signals
FI110349B (en) Digital transmission system using subband coding of a digital signal
CN1224198C (en) Multi-user detection method using finite-impulse-response matrix filter
JP4287777B2 (en) Transmitting apparatus and receiving apparatus
Tsatsanis et al. Blind estimation of direct sequence spread spectrum signals in multipath
EP1665517B1 (en) Method and apparatus for reducing interference within a communication system
US5416767A (en) Method of transmitting a data stream, transmitter and receiver
US7502425B2 (en) Efficient reduced complexity windowed optimal time domain equalizer for discrete multitone-based DSL modems

Legal Events

Date Code Title Description
C06 Publication
C10 Request of examination as to substance
C14 Granted
C56 Change in the name or address of the patentee

Owner name: QINGHUA UNIVERSITY; ZHANXUN COMMUNICATIONS (SHANGH

Free format text: FORMER NAME OR ADDRESS: QINGHUA UNIVERSITY