CN1829206B

CN1829206B - A Channel Estimation Method for Time Division Synchronous Code Division Multiple Access System

Info

Publication number: CN1829206B
Application number: CN2005101377504A
Authority: CN
Inventors: 李立华; 杨晓辉; 陶小峰; 张平; 王毅; 何丹丹
Original assignee: Beijing University of Posts and Telecommunications; LG Electronics Inc
Current assignee: Beijing University of Posts and Telecommunications; LG Electronics Inc
Priority date: 2005-12-19
Filing date: 2005-12-19
Publication date: 2010-06-09
Anticipated expiration: 2025-12-19
Also published as: CN1829206A

Abstract

本发明公开了一种用于时分同步码分多址系统的信道估计方法。根据本发明，将一个时隙等分为若干个时间段，假设信道特性在每个时间段内是恒定不变的，将整个导频序列也相应的划分为具有相同长度的若干段，改变中间码矩阵的排列方式，新的矩阵由分割后得到的导频序列小段构成，利用迫零算法或者最小均方误差算法，可以得到各个时间段内的信道冲激响应估计值。对多个连续时隙内导频部分的信道估计值进行拟合计算出非导频部分的信道特性，从而得到整个发送数据期间内的信道精确信息。本发明基于现有TD-SCDMA系统，以最小的复杂度代价换取信道估计精确度的极大提高，从而改善了系统在快速变化信道条件下的性能。The invention discloses a channel estimation method for time division synchronous code division multiple access system. According to the present invention, a time slot is equally divided into several time segments, assuming that the channel characteristics are constant in each time segment, the entire pilot sequence is correspondingly divided into several segments with the same length, and the intermediate The arrangement of the code matrix, the new matrix is composed of small segments of the pilot sequence obtained after division, and the estimated value of the channel impulse response in each time period can be obtained by using the zero-forcing algorithm or the minimum mean square error algorithm. The channel estimation value of the pilot part in multiple consecutive time slots is fitted to calculate the channel characteristics of the non-pilot part, so as to obtain accurate channel information during the entire data transmission period. Based on the existing TD-SCDMA system, the present invention exchanges the minimum complexity cost for greatly improving the accuracy of channel estimation, thereby improving the performance of the system under fast-changing channel conditions.

Description

A Channel Estimation Method for Time Division Synchronous Code Division Multiple Access System

技术领域technical field

本发明涉及一种用于时分同步码分多址(TD-SCDMA)系统的信道估计方法。The invention relates to a channel estimation method for a time division synchronous code division multiple access (TD-SCDMA) system.

背景技术Background technique

时分同步码分多址(TD-SCDMA)系统是基于时分双工(TDD)方式的第三代移动通信系统，它运用了上行同步、智能天线、联合检测等一系列关键技术，而这些技术的实现都依赖于对无线信道响应的快速而准确的估计。TD-SCDMA系统在突发(Burst)结构中设置了用来进行信道估计的导频序列部分，根据接收到的信号和已知的导频序列，可以估算出信道冲激响应。目前TD-SCDMA系统采用的信道估计方案假设信道在一个时隙内是恒定不变的。然而随着移动台速度的不断增大，一个时隙内信道衰落的包络相关性下降，TDD时隙用于信道估计的导频序列，无法准确反映这种信道变化，将此信道估计值用于整个时隙，则无疑带来系统性能的极大下降。Time division synchronous code division multiple access (TD-SCDMA) system is the third generation mobile communication system based on time division duplex (TDD), it uses a series of key technologies such as uplink synchronization, smart antenna, joint detection, and the Implementations all rely on fast and accurate estimation of the wireless channel response. The TD-SCDMA system sets the pilot sequence part used for channel estimation in the burst (Burst) structure, and the channel impulse response can be estimated according to the received signal and the known pilot sequence. The channel estimation scheme adopted by the TD-SCDMA system at present assumes that the channel is constant within a time slot. However, as the speed of the mobile station continues to increase, the envelope correlation of channel fading within a time slot decreases, and the pilot sequence used for channel estimation in TDD time slots cannot accurately reflect this channel change. The channel estimation value is used as For the entire time slot, it will undoubtedly bring about a great drop in system performance.

发明内容Contents of the invention

针对现有TD-SCDMA系统中对快速变化信道难以准确估计的缺点，本发明在不改变现有TD-SCDMA系统帧结构和导频数据的前提下，提出了一种用于TD-SCDMA系统的信道估计方法，以应对目前TDD系统中终端高速运动所带来的技术问题。Aiming at the shortcomings of the existing TD-SCDMA system that it is difficult to accurately estimate the fast-changing channel, the present invention proposes a method for the TD-SCDMA system without changing the frame structure and pilot data of the existing TD-SCDMA system. The channel estimation method is used to deal with the technical problems caused by the high-speed movement of terminals in the current TDD system.

根据本发明，提供了一种适用于时分同步码分多址系统的信道估计方法，包括步骤：According to the present invention, a kind of channel estimation method suitable for time division synchronous code division multiple access system is provided, comprising steps:

(1)将一个时隙等分为多个时间段，其中信道特性在每个时间段中保持不变；(1) Divide a time slot into multiple time segments, wherein the channel characteristics remain unchanged in each time segment;

(2)将导频部分划分为具有相同长度的多个段，构造出如下的中间码矩阵：(2) The pilot part is divided into a plurality of sections with the same length, and the following midamble matrix is constructed:

A_m ⁿ为第n个用户发送的第m段导频构成的中间码矩阵：A _m ⁿ is the midamble matrix formed by the m-th pilot sent by the n-th user:

n＝1，2；m＝1，...，Mn=1,2; m=1,...,M

其中P是导频序列的长度，p(n，k)(k＝1+(m-1)＊Q，...，Q+(m-1)＊Q)是第n个用户的第k个导频符号，M是导频分段的最大数量，W为信道径数，Q为每一段导频数据的长度，其中Q＝P/M；where P is the length of the pilot sequence, p(n,k)(k=1+(m-1)*Q,...,Q+(m-1)*Q) is the kth of the nth user Pilot symbols, M is the maximum number of pilot segments, W is the number of channel paths, and Q is the length of each pilot data, where Q=P/M;

(3)根据上述中间码矩阵估计各个时间段内的信道特性，得到导频部分的信道估计；(3) Estimate the channel characteristics in each time period according to the above-mentioned midamble matrix, and obtain the channel estimation of the pilot part;

(4)根据所述得到的导频部分的信道估计，计算出非导频部分的信道特性，从而得到整个发送数据期间内的信道估计。(4) Calculate the channel characteristics of the non-pilot part according to the obtained channel estimation of the pilot part, so as to obtain the channel estimation in the whole data transmission period.

根据本发明的一个实施例，根据信道特性以及系统要求设定导频分段的最大数量M。According to an embodiment of the present invention, the maximum number M of pilot segments is set according to channel characteristics and system requirements.

本发明的基本构思在于，改变传统信道估计算法对在传送整个导频期间信道不变的假设，提出将一个时隙等分为若干个时间段，假设信道特性在每个时间段内是恒定不变的，将整个导频序列也相应的划分为具有相同长度的若干段，改变中间码(midamble)矩阵的排列方式，新的矩阵由分割后得到的导频序列小段构成，利用迫零算法或者最小均方误差算法，可以得到各个时间段内的信道冲激响应估计值。对多个连续时隙内导频部分的信道估计值进行拟合计算出非导频部分的信道特性，从而得到整个发送数据期间内的信道精确信息。The basic idea of the present invention is to change the traditional channel estimation algorithm's assumption that the channel is constant during the transmission of the entire pilot, and propose to divide a time slot into several time segments, assuming that the channel characteristics are constant in each time segment. change, the entire pilot sequence is correspondingly divided into several segments with the same length, and the arrangement of the midamble matrix is changed. The new matrix is composed of small segments of the pilot sequence obtained after division, and the zero-forcing algorithm or The minimum mean square error algorithm can obtain the estimated value of the channel impulse response in each time period. The channel estimation value of the pilot part in multiple consecutive time slots is fitted to calculate the channel characteristics of the non-pilot part, so as to obtain accurate channel information during the entire data transmission period.

本发明基于现有TD-SCDMA系统，以最小的复杂度代价换取信道估计精确度的极大提高，从而改善了系统在快速变化信道条件下的性能。Based on the existing TD-SCDMA system, the present invention exchanges the minimum complexity cost for greatly improving the accuracy of channel estimation, thereby improving the performance of the system under fast-changing channel conditions.

附图说明Description of drawings

下面参照附图并结合实例来进一步描述本发明，其中：Further describe the present invention below with reference to accompanying drawing and in conjunction with example, wherein:

图1示出目前采用的TD-SCDMA系统的模型图；Fig. 1 shows the model diagram of the TD-SCDMA system adopted at present;

图2比较了在150km/h车速下，根据本发明的信道估计方法和现有的信道估计方法对连续四个时隙的信道特性进行估计得到的结果；Figure 2 compares the results obtained by estimating the channel characteristics of four consecutive time slots according to the channel estimation method of the present invention and the existing channel estimation method at a vehicle speed of 150km/h;

图3示出多天线系统的模型图；Figure 3 shows a model diagram of a multi-antenna system;

图4分别示出了采用本发明的信道估计方法的上行MIMO(多天线)-JD(联合检测)链路和下行MIMO-JT(联合发送)链路在不同信噪比条件下系统误比特率的性能曲线。为了便于比较，图中还同时示出了采用现有信道估计方法的MIMO-JD/JT性能。仿真中假设在TD-SCDMA常规时隙中，除TS0用于广播信息以外，TS1、TS2、TS3配置为上行业务时隙，TS4、TS5、TS6配置为下行业务时隙。3个上行时隙的信道特性根据本发明的信道估计方法得到，3个下行时隙的信道特性则利用上行时隙的信道估计结果预测得出。Fig. 4 respectively shows the system bit error rate under different SNR conditions of the uplink MIMO (multi-antenna)-JD (joint detection) link and the downlink MIMO-JT (joint transmission) link using the channel estimation method of the present invention performance curve. For the convenience of comparison, the figure also shows the performance of MIMO-JD/JT using the existing channel estimation method. In the simulation, it is assumed that in TD-SCDMA conventional time slots, except TS0 is used for broadcasting information, TS1, TS2, and TS3 are configured as uplink service time slots, and TS4, TS5, and TS6 are configured as downlink service time slots. The channel characteristics of the three uplink time slots are obtained according to the channel estimation method of the present invention, and the channel characteristics of the three downlink time slots are predicted by using the channel estimation results of the uplink time slots.

具体实施方式Detailed ways

本发明采用现有TD-SCDMA系统导频序列。为了提高信道估计的精度，修改传统估计器整个时隙信道不变的假设，将一个时隙等分为多个时间段，假设信道特性在每个时间段内保持不变的，设信道径数为W。首先，将整个导频序列划分为M个小段，利用各小段重构中间码矩阵，根据迫零或者最小均方误差方法估计导频部分各个时间段内的信道冲激响应。然后，利用相邻多个时隙导频部分的信道估计值进行LSM拟合计算出数据部分的信道特性，从而得到整个信道的响应特性。The present invention adopts the existing TD-SCDMA system pilot sequence. In order to improve the accuracy of channel estimation, the assumption that the channel of the traditional estimator is constant throughout the time slot is modified, and a time slot is divided into multiple time periods. Assuming that the channel characteristics remain unchanged in each time period, the number of channel paths is set for W. Firstly, the whole pilot sequence is divided into M small segments, the midamble matrix is reconstructed by using each small segment, and the channel impulse response in each time period of the pilot part is estimated according to the zero-forcing or minimum mean square error method. Then, the channel characteristics of the data part are calculated by LSM fitting using the channel estimation values of the pilot parts of multiple adjacent time slots, so as to obtain the response characteristics of the entire channel.

根据实际及性能要求，导频分段可以采用不同的大小。信道变化很慢时可采用较小的M；信道快速变化时采用比较大的M。According to the actual and performance requirements, the pilot segment can adopt different sizes. When the channel changes slowly, a smaller M can be used; when the channel changes quickly, a larger M can be used.

下面举例说明在TD-SCDMA系统中的应用。The following example illustrates the application in TD-SCDMA system.

不失一般性，下面的分析中以两个用户为例，即图1中M_t＝2。信道最大径数设为W＝8；导频分的段数为M＝8。每一段中间码码中数据的长度为：Q＝P/M，P＝128为导频序列的长度。Without loss of generality, the following analysis takes two users as an example, that is, M _t =2 in FIG. 1 . The maximum path number of the channel is set to W=8; the number of segments divided by the pilot frequency is M=8. The length of data in each midamble code is: Q=P/M, and P=128 is the length of the pilot sequence.

用户端1发送的导频序列为M(1)＝[p(1，0)，p(1，1)，..p(1，P-1)]^T，用户端2发送的导频序列为M(2)＝[p(2，0)，p(2，1)，...p(2，P-1)]^T，其中p(n，k)是第n个用户的第k个导频符号，总的导频符号数为P＝128。The pilot sequence sent by client 1 is M(1)=[p(1, 0), p(1, 1), ..p(1, P-1)] ^T , the pilot sequence sent by client 2 M(2)=[p(2,0),p(2,1),...p(2,P-1)] ^T , where p(n,k) is the kth pilot symbols, the total number of pilot symbols is P=128.

接收方接收的信号可以表示为：The signal received by the receiver can be expressed as:

R＝AH+nR=AH+n

其中R＝[r₁，r₂...，r_P+W-1]^T，r_k代表接收方所接收的第k个符号。Wherein R=[r ₁ , r _{2 .} . . , r _P+W-1 ] ^T , r _k represents the kth symbol received by the receiver.

信道矩阵为：The channel matrix is:

$H h = = [[{h h}_{1,1 1,1}^{11} {h h}_{2,1 2,1}^{11} \cdot \cdot \cdot \cdot \cdot \cdot {h h}_{W W,, 11}^{11} \cdot \cdot \cdot \cdot \cdot \cdot {h h}_{11,, M m}^{11} {h h}_{22,, M m}^{11} \cdot \cdot \cdot \cdot \cdot &Center Dot; {h h}_{W W,, M m}^{11},, {h h}_{1,1 1,1}^{22} {h h}_{2,1 2,1}^{22} \cdot \cdot \cdot \cdot \cdot \cdot {h h}_{W W,, 11}^{33} \cdot \cdot \cdot \cdot \cdot \cdot {h h}_{11,, M m}^{22} {h h}_{22,, M m}^{22} \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot; {h h}_{W W,, M m}^{22}]]$

h_w，m ⁿ表示第n个用户第m段导频序列第w条径的信道响应值。例如代表第一个用户在导频序列的第一段的信道冲击响应，假设在传送该段导频序列时信道在各个径上的值保持不变，h_1，1 ¹h_2，1 ¹…h_W，1 ¹分别是第一径，第二径一直到第W径。h _{w, m} ⁿ represents the channel response value of the wth path of the mth pilot sequence of the nth user. For example, it represents the channel impulse response of the first user in the first segment of the pilot sequence, assuming that the value of the channel on each path remains unchanged when transmitting the pilot sequence, h _1, ^{1 1} h _2, ^{1 1} ... h _{W, 1} ¹ are respectively the first path, the second path until the Wth path.

考虑到多径信道，接收到的每一段导频序列都会叠加到下一段导频序列中。Considering the multipath channel, each received pilot sequence will be superimposed on the next pilot sequence.

发送的中间码矩阵可以写成：The transmitted intermediate code matrix can be written as:

A＝[A¹A²]A＝[A ¹ A ² ]

其中：A¹、A²分别代表用户1、2的中间码矩阵Among them: A ¹ and A ² represent the intermediate code matrix of users 1 and 2 respectively

n＝1，2；m＝1，...，Mn=1,2; m=1,...,M

Q＝P/M，为每一段导频数据的长度。矩阵A_m ⁿ中，每一列非零数据相等，都是该段中间码导频数据。Q=P/M, which is the length of each piece of pilot data. In the matrix A _m ⁿ , each column of non-zero data is equal, which is the midamble pilot data of this segment.

根据迫零原则可以得到信道估计为：According to the zero-forcing principle, the channel estimate can be obtained as:

H＝A⁺RH = A ⁺ R

其中A⁺表示A的伪逆。where A ⁺ denotes the pseudo-inverse of A.

根据MMSE原则可得信道估计为：According to the MMSE principle, the channel estimation can be obtained as:

H＝A^T(AA^T+σ²I)RH＝ ^AT (AA ^T +σ ² I)R

其中A^T表示A的共轭转置，σ²为噪声方差，I为单位对角阵。where ^AT represents the conjugate transpose of A, ^σ2 is the noise variance, and I is the unit diagonal matrix.

应用该算法，可以得到多个连续上行时隙发送导频序列时间段的信道估计值，根据不同的原则对这些估计值进行拟合，得到发送导频序列之间数据时间段上的信道估计。By applying this algorithm, the channel estimation values of multiple consecutive uplink time slots for sending pilot sequences can be obtained, and these estimated values can be fitted according to different principles to obtain channel estimates for data periods between sending pilot sequences.

图2是对连续4个时隙进行信道估计的。Figure 2 performs channel estimation on four consecutive time slots.

信道估计的均方误差值可由下式计算：The mean square error value of channel estimation can be calculated by the following formula:

$E E. [[{| | e e ((τ τ)) | |}^{22}]] = = E E. [[{| | c c ((τ τ)) - - \overset{^^}{c c} ((τ τ)) | |}^{22}]]$

其中c(τ)表示实际信道值，表示信道估计值。经仿真计算可得，传统方法的均方估计误差为E_org＝0.0035，而新的信道估计方法的均方误差为E_est1＝3.1038e-004。可见，采用新的估计方法可使信道估计的准确性显著提高。where c(τ) represents the actual channel value, Indicates the channel estimate. It can be obtained through simulation calculation that the mean square estimation error of the traditional method is E_org=0.0035, while the mean square error of the new channel estimation method is E_est1=3.1038e-004. It can be seen that the accuracy of channel estimation can be significantly improved by adopting the new estimation method.

以上方法说明均针对两发一收系统，如图1所示。The above method descriptions are all for the two-transmit-one-receive system, as shown in Figure 1.

实际上该算法也可应用于MIMO系统，此时只需将基站单天线扩展为基站多天线，以两发两收系统(参见图3)为例：In fact, this algorithm can also be applied to the MIMO system. At this time, it is only necessary to expand the base station single antenna to the base station multi-antenna. Take the two-transmit and two-receive system (see Figure 3) as an example:

图中的两发两收系统可看作由两个两发一收系统构成，系统1和系统2分别如图中实线和虚线所示。两个系统分别按照上文说明进行信道估计，则可分别估得两个发射天线与第一个接收天线之间的信道响应H⁽¹⁾＝[H^(1，1)，H^(2，1)]和两个发射天线与第二个接收天线之间的信道响应H⁽²⁾＝[H^(1，2)，H^(2，2)]。The two-transmit and two-receive system in the figure can be regarded as composed of two two-transmit and one-receive systems. System 1 and system 2 are shown by the solid line and the dotted line in the figure, respectively. The two systems respectively perform channel estimation according to the above description, then the channel response between the two transmitting antennas and the first receiving antenna can be estimated separately H ⁽¹⁾ = [H ^{(1, 1)} , H ^{(2, 1 )} ] and the channel response H ⁽²⁾ = [H ^(1,2) , H ^(2,2) ] between the two transmit antennas and the second receive antenna.

在利用信道估计结果进行联合检测时，信道矩阵H可表示为When using channel estimation results for joint detection, the channel matrix H can be expressed as

$H h = = [\begin{matrix} {H h}^{((1,1 1,1))} & {H h}^{((2,1 2,1))} & . . . . . . & {H h}^{(({M m}_{t t},, 11))} \\ {H h}^{((1,2 1,2))} & {H h}^{((2,2 2,2))} & . . . . . . & {H h}^{(({M m}_{t t},, 22))} \\ . . . . . . & . . . . . . & . . . . . . & . . . . . . \\ {H h}^{((11,, {M m}_{r r}))} & {H h}^{((22,, {M m}_{r r}))} & . . . . . . & {H h}^{(({M m}_{t t},, {M m}_{r r}))} \end{matrix}]$

其中in

为其中的第n个单元，n＝1，...，N，m_t＝1，...，M_t，m_r＝1，...，M_r。N为每个时隙的符号个数，M_t与M_r分别表示发射天线数和接收天线数，K为每个用户的码道数，S为扩频因子。为若采用传统算法估得的结果，则有其中c_k为第k个码道的扩频序列，

为第k个码道第m_t个发射天线与第m_r个接收天线之间的信道冲激响应：

而如果用新的信道估计算法，则有

为第k个码道第m_t个发射天线与第m_r个接收天线之间第n个数据符号位置上的信道冲激响应：

这里n表示数据符号的位置，N为每个用户每个时隙传送的数据符号数。由此可见，传统算法假定信道冲激响应在一个时隙内不变，而在新算法中，一个时隙内各个数据符号位置上的信道冲激响应是不同的，故采用新算法得到的信道估计结果进行联合检测将更为精确；但是检测算法中的信道矩阵H的维数并没有改变，因此新的信道估计算法不会增加联合检测算法的复杂度。 is the nth unit therein, n=1, ..., N, m _t =1, ..., M _t , m _r =1, ..., M _r . N is the number of symbols in each time slot, M _t and Mr _r represent the number of transmitting antennas and receiving antennas respectively, K is the number of code channels for each user, and S is the spreading factor. For the result estimated by the traditional algorithm, there is where c _k is the spreading sequence of the kth code channel,

is the channel impulse response between the m _t transmit antenna of the k code channel and the m _r receive antenna:

And if the new channel estimation algorithm is used, then there is

is the channel impulse response at the position of the nth data symbol between the _mtth transmit antenna and the _mrth receive antenna of the kth code channel:

Here n represents the position of the data symbol, and N is the number of data symbols transmitted by each user in each time slot. It can be seen that the traditional algorithm assumes that the channel impulse response is constant in a time slot, but in the new algorithm, the channel impulse response at each data symbol position in a time slot is different, so the channel impulse response obtained by the new algorithm is The joint detection of the estimated results will be more accurate; but the dimension of the channel matrix H in the detection algorithm has not changed, so the new channel estimation algorithm will not increase the complexity of the joint detection algorithm.

本发明已结合具体实例作了说明。然而对于熟悉本领域的人员来说，显而易见可以在不偏离本发明的教导的情况下，将本发明的构思应用到不同的具体实例中去。这里的描述只是说明性的，而完全不应认为是限制性的。本发明的专利保护范围由所附权利要求书限定，而不是前面的说明。因此所有落在权利要求范围内的各种变型和等效形式都应属于本发明的保护范围之内。The invention has been described with reference to specific examples. However, it will be apparent to those skilled in the art that the concept of the present invention can be applied to different embodiments without departing from the teaching of the present invention. The descriptions herein are illustrative only and should not be considered limiting in any way. The scope of patent protection of the present invention is defined by the appended claims rather than the foregoing description. Therefore, all the various modifications and equivalent forms falling within the scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A channel estimation method for time division synchronous code division multiple access system, comprising steps:

(1) Divide a time slot into multiple time segments, wherein the channel characteristics remain unchanged in each time segment;

(2) The pilot part is divided into a plurality of sections with the same length, and the following midamble matrix is constructed:

A _m ⁿ is the midamble matrix formed by the m-th pilot sent by the n-th user:

where P is the length of the pilot sequence, p(n,k)(k=1+(m-1)*Q,...,Q+(m-1)*Q) is the kth of the nth user Pilot symbols, W is the number of channel paths, M is the maximum number of pilot segments, and Q=P/M is the length of each section of pilot data;

(3) Estimate the channel characteristics in each time period according to the above-mentioned midamble matrix, and obtain the channel estimation of the pilot part;

(4) Calculate the channel characteristics of the non-pilot part according to the obtained channel estimation of the pilot part, so as to obtain the channel estimation in the whole data transmission period.

2. method according to claim 1, wherein step (3) also comprises the step:

The channel characteristics in each time period are estimated according to the midamble matrix, and the channel impulse response of the pilot part is obtained by using the zero-forcing principle.

3. method according to claim 1, wherein step (3) also comprises the step:

The channel characteristics in each time period are estimated according to the midamble matrix, and the channel impulse response of the pilot part is obtained by using the minimum mean square error principle.

4. method according to claim 1, wherein step (4) also comprises the step:

Using the channel estimation of the pilot part in multiple consecutive time slots, multi-slot fitting is performed to obtain the channel characteristics of the non-pilot part.

5. The method according to claim 1, wherein the maximum number M of the pilot segments is set according to channel characteristics and system requirements.