WO2010048861A1 - 一种快速退出训练的方法、系统及装置 - Google Patents

一种快速退出训练的方法、系统及装置 Download PDF

Info

Publication number
WO2010048861A1
WO2010048861A1 PCT/CN2009/074543 CN2009074543W WO2010048861A1 WO 2010048861 A1 WO2010048861 A1 WO 2010048861A1 CN 2009074543 W CN2009074543 W CN 2009074543W WO 2010048861 A1 WO2010048861 A1 WO 2010048861A1
Authority
WO
WIPO (PCT)
Prior art keywords
filter
subscriber line
digital subscriber
convergence condition
training process
Prior art date
Application number
PCT/CN2009/074543
Other languages
English (en)
French (fr)
Inventor
方李明
辛德瑞那⋅拉斐尔
张朋瑞
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP09823053.5A priority Critical patent/EP2343832B1/en
Publication of WO2010048861A1 publication Critical patent/WO2010048861A1/zh
Priority to US13/097,950 priority patent/US8537882B2/en

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B3/00—Line transmission systems
    • H04B3/02—Details
    • H04B3/32—Reducing cross-talk, e.g. by compensating
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04M—TELEPHONIC COMMUNICATION
    • H04M11/00—Telephonic communication systems specially adapted for combination with other electrical systems
    • H04M11/06—Simultaneous speech and data transmission, e.g. telegraphic transmission over the same conductors
    • H04M11/062—Simultaneous speech and data transmission, e.g. telegraphic transmission over the same conductors using different frequency bands for speech and other data

Definitions

  • the present invention relates to vector DSL (Vectored-DSL) technology in the field of communications, and in particular to a method, system and apparatus for quick exit training. Background technique
  • xDSL Digital Subscriber Line
  • DSLAM Digital Subscriber Line Access
  • Multiplexer digital subscriber line access multiplexer
  • the joint transceiving is performed by using a precoder or canceller at the DSLAM end, and the crosstalk interference is cancelled by the signal processing method, and finally there is no crosstalk interference in each signal.
  • Embodiments of the present invention provide a method, system, and apparatus for quickly exiting training, and exiting the training process in time to reduce interference to neighboring users.
  • an embodiment of the present invention provides a method for quickly exiting training, including:
  • the embodiment of the invention also provides a system for quickly exiting training, including:
  • the digital subscriber line access multiplexer connects the transceivers of the plurality of subscribers through the digital subscriber line, wherein
  • the filter of the digital subscriber line of the digital subscriber line access multiplexer transmits two orthogonal sequences during training with the transceivers of the plurality of subscribers;
  • the digital subscriber line access multiplexer is configured to determine whether the filter satisfies a convergence condition during a filter training process of the vector digital subscriber line, and exit when the filter satisfies the convergence condition The filter training process.
  • the embodiment of the invention further provides a digital subscriber line access multiplexer, comprising:
  • a vector digital subscriber line filter for processing data between the digital subscriber line access multiplexer and a transceiver at the subscriber end to suppress crosstalk
  • a judging module configured to compare a feedback error value of the filter and/or a swing of the filter coefficient with a preset threshold value during a filter training process of the vector digital subscriber line, according to the comparison As a result, it is determined whether the filter satisfies an convergence condition
  • an exiting module configured to exit the filter training process when the determining module determines that the filter satisfies the convergence condition.
  • Embodiments of the invention have the following advantages: By exiting the training process in time during the training process of the Vectored-DSL precoder or canceller, the time spent in the training process is shortened, and the interference of the precoder or canceller to other users during the training process is reduced. Small other users drop the chance. DRAWINGS
  • FIG. 1 is a flowchart of a method for quickly exiting training according to a first embodiment of the present invention
  • FIG. 2 is a flowchart of a method for quickly exiting training according to a second embodiment of the present invention
  • Schematic diagram of user-to-office transmission
  • FIG. 4 is a schematic diagram of a UE-to-office offset implementation according to Embodiment 2 of the present invention
  • FIG. 5 is a schematic diagram of a UE-to-user transmission according to Embodiment 2 of the present invention
  • FIG. 6 is a schematic diagram of a method for implementing local-to-user cancellation according to Embodiment 2 of the present invention
  • FIG. 7 is a schematic diagram of determining whether a filter is converged according to Embodiment 2 of the present invention
  • FIG. 8 is a schematic diagram of another method for determining whether a filter is converged according to Embodiment 2 of the present invention
  • FIG. 9 is a schematic flowchart of downlink filter training according to an embodiment of the present invention.
  • FIG. 10 is a schematic flowchart of uplink filter training according to an embodiment of the present invention.
  • FIG. 11 is a structural diagram of an apparatus for quickly exiting training according to an embodiment of the present invention. detailed description
  • a method for quickly exiting training according to the first embodiment of the present invention includes: Step S101: determining, in a filter training process of a vector digital subscriber line, whether the filter satisfies an convergence condition;
  • Step S102 when the filter satisfies the convergence condition, exiting the filter training process.
  • the training process is promptly exited during the training process of the Vectored-DSL precoder or canceller, so that the time required for the training process is shortened, and the precoder or canceller is reduced during the training process. Interference with other users, reducing the chances of other users dropping.
  • a method for quickly exiting training according to the second embodiment of the present invention includes: Step S201: After the filter training starts, the DSLAM initializes the coefficients of the filter.
  • the implementation is divided into an uplink direction and a downlink direction.
  • the uplink transmission refers to the transmission from the UE to the central office.
  • the uplink transmission is the joint reception processing of the received data of each user at the receiving end, and the crosstalk canceler (Canceller) is used for the crosstalk at the receiving end.
  • the signal is cancelled.
  • the specific implementation manner is shown in FIG. 4, and the receiving end performs joint crosstalk cancellation processing on the receiving data to ensure that the received data of the user has no crosstalk interference.
  • the filter designed at the receiving end is a canceller whose coefficient is W. At the time of initialization, the W can be a preset default value or 0, and the W value can be changed with time.
  • the downlink transmission refers to the transmission from the central office to the user end.
  • precoding is performed on the transmitting end during downlink transmission to suppress crosstalk in the line.
  • the specific implementation is shown in Figure 6.
  • the transmission data is pre-processed at the transmitting end to ensure that the data transmitted to each user has no crosstalk interference.
  • the filter designed at the transmitting end is a precoder whose coefficient is P. At the time of initialization, the P can be a preset default value or 0, and the P value can be changed with time.
  • Step S202 the DSLAM initializes the size of the transmission power.
  • the size of the transmit power is selected according to an empirical value, usually a smaller value or It is a conservative value, and the above steps S201 and S202 have no sequential relationship.
  • Step S203 The transceiver sends a training sequence.
  • all users will send two orthogonal sequences through their respective transceivers, wherein the two orthogonal sequences may be selected in a Hadamard matrix, and the DSLAM is received according to the received
  • the sequence calculates a feedback error value, and calculates a crosstalk channel matrix according to the feedback error value; in the downlink direction, the DSLAM sends a pair of orthogonal sequences to the user, wherein the two orthogonal sequences can be in a Hadamard matrix.
  • the user calculates a feedback error value according to the received sequence, and sends feedback error information to the DSLAM.
  • the DSLAM obtains a feedback error value according to the feedback error information, and calculates a crosstalk channel matrix according to the feedback error value. Further, the feedback The error information is obtained by scaling the feedback error value.
  • the DSLAM can calculate the crosstalk channel matrix according to the following formula (1):
  • T is the total number of symbols
  • X is the transmitted sequence (Hadamard)
  • E is the feedback error value
  • t is the t-th symbol
  • H is the crosstalk channel coefficient
  • Step S204 The DSLAM determines whether the training time of the filter has timed out. If the timeout expires, the process proceeds to step S206. If there is no timeout, the process proceeds to step S205. Whether the training time is timeout or not is specifically set to a preset time, and the preset time may be selected according to an empirical value. When the training time of the filter exceeds the preset time, the timeout is determined.
  • Step S205 The DSLAM determines whether the filter satisfies the convergence condition according to the feedback error value and/or the coefficient of the filter. When the filter satisfies the convergence condition, go to step 206, and when the filter does not satisfy the convergence condition, go to step 203.
  • the condition for the convergence of the filter includes, but is not limited to: the feedback error value is not greater than a predetermined threshold value ⁇ , and the threshold value ⁇ is confirmed according to historical experience and the like, and may be changed.
  • the feedback error value is not greater than a preset threshold value ⁇
  • the convergence condition is satisfied, and the filter is determined to converge.
  • the filter coefficients P and w tend to be stable, that is, the swing is smaller than a predetermined threshold e th , and the threshold e; 3 ⁇ 4 is based on historical experience, etc.
  • the factors confirmed can be changed. As shown in Figure 8, find the difference in swing during the observation period (tl ⁇ tn):
  • Emm m[ P mn t, ), P (k, t 2 ... P (k, t n ))
  • a e ⁇ max ⁇ min When Ae is less than, the convergence condition is satisfied, and when Ae is not less than, the convergence condition is not satisfied.
  • ⁇ ⁇ denotes the coefficient of the precoder of the kth subcarrier of the n- m user at time t. That is, the swing of the filter coefficient is less than a predetermined threshold to satisfy the convergence condition, and it is determined that the filter converges. When the filter satisfies the above convergence condition, it is not necessary to continue training and the training is suspended.
  • Step S206 the message of ending the training filter is sent to the opposite end, and the filter training of the current stage is ended.
  • the training process is exited in time during the training process of the Vectored-DSL precoder or canceller, so that the time required for the training process is shortened, and the precoder or canceller is reduced during the training process.
  • the interference to other users reduces the chance of other users dropping.
  • the embodiment of the present invention provides a method for quickly exiting training, which is applied in the training process of the downlink filter, as shown in FIG. 9, and includes:
  • Step S901 The sending end sends a training sequence.
  • the training column sequence may be a sequence of two orthogonal codes. Specifically, the two orthogonal sequences may be selected in a Hadamard matrix.
  • Step S902 the receiving end starts to receive the training sequence.
  • step S901 and step S902 are performed simultaneously, that is, the transmitting end starts to receive the training sequence while the transmitting end sends the training sequence.
  • Step S903 the receiving end sends the feedback error information to the sending end.
  • the receiving end feeds back the error information to the transmitting end through the Back channel.
  • the Back channel can be the underlying SOC channel or the upper layer (2 layer) feedback channel channel.
  • the receiver directly calculates the error and processes it, without feedback.
  • Step S904 at the transmitting end, obtaining a feedback error value according to the feedback error information, and determining whether the feedback error value is less than a predetermined threshold e , and determining that the feedback error value is less than a preset threshold ⁇ , then the filter is determined to be converged, and the message of the suspension training is sent to the receiving end to end the training in advance, and the process goes to step S905.
  • the filter does not converge and continues to transmit the training sequence, and proceeds to step S901 until the end of training.
  • the feedback error value before the receiving end sends the feedback error value to the transmitting end, the feedback error value may be scaled to generate feedback error information, so as to facilitate transmission or improve transmission accuracy; after receiving the feedback error information, the transmitting end may The feedback error information is inversely processed to obtain a feedback error value. For example: the receiving end can reduce the feedback error value error by a times (error/a) to generate feedback error information and send it to the transmitting end. After receiving the feedback error information (error/a), the transmitting end expands the feedback error information by a factor of (error/axa) Get the feedback error value error.
  • Step S905 The sending end sends a message to the receiving end to notify the receiving end to exit the filter training process at the same time.
  • the sending end sends a message to stop the training to the receiving end and ends the training.
  • the training process is exited in time during the training process of the Vectored-DSL precoder or canceller, so that the time required for the training process is shortened, and the precoder or canceller is reduced during the training process.
  • the interference to other users reduces the chance of other users dropping.
  • the embodiment of the present invention provides a method for quickly exiting training, which is applied in the training process of the uplink filter, as shown in FIG. 10, and includes:
  • Step S1001 The sending end sends a training sequence.
  • the training column sequence may be a sequence of two orthogonal codes, as in the above step 203. Specifically, the two orthogonal sequences may be selected in a Hadamard matrix.
  • Step S1002 The receiving end starts to receive the training sequence.
  • step S1001 and step S1002 are performed simultaneously, that is, the transmitting end starts to receive the training sequence while the transmitting end sends the training sequence.
  • Step S1003 The receiving end acquires a feedback error value, and determines whether the feedback error value is less than a predetermined threshold value e *, and when determining that the feedback error value is less than a preset threshold value e , determining The filter converges, and sends a message of the suspension training to the receiving end to end the training in advance, and proceeds to step S1004.
  • a preset threshold value e it is determined that the filter is not Convergence, and continue to send the training sequence, go to step S1001, - until the training is finished.
  • Step S1004 The receiving end sends a message to the sending end to notify the sending end to exit the filter training process at the same time.
  • the receiving end sends a message of aborting training to the transmitting end and ends the training.
  • the training process is exited in time during the training process of the Vectored-DSL precoder or canceller, so that the time required for the training process is shortened, and the precoder or canceller is reduced during the training process.
  • the interference to other users reduces the chance of other users dropping.
  • the embodiment of the present invention further provides a system for quickly exiting training, including a digital subscriber line access multiplexer, and a transceiver connected to multiple subscriber ends through a digital subscriber line, where
  • the filter of the digital subscriber line of the digital subscriber line access multiplexer transmits two orthogonal sequences during training with the transceivers of the plurality of subscribers;
  • the digital subscriber line access multiplexer is configured to determine whether the filter satisfies a convergence condition during a filter training process of the vector digital subscriber line, and exit when the filter satisfies the convergence condition The filter training process.
  • the digital subscriber line access multiplexer is configured to obtain a feedback error of the filter. Poor, determining that the feedback error is not greater than a preset threshold e , determining that the filter satisfies the convergence condition and exits the filter training process.
  • the digital subscriber line access multiplexer is configured to determine that the filter satisfies the convergence condition when the swing of the filter coefficient is determined to be less than a preset threshold value The filter training process.
  • the training process is exited in time during the training process of the Vectored-DSL precoder or canceller, so that the time required for the training process is shortened, and the precoder or canceller is reduced during the training process.
  • the interference to other users reduces the chance of other users dropping.
  • the embodiment of the present invention further provides a digital subscriber line access multiplexer, as shown in FIG. 11, comprising: a vector digital subscriber line filter 110, configured to access the multiplexer and the subscriber end of the digital subscriber line Data between transceivers is processed to suppress crosstalk;
  • a vector digital subscriber line filter 110 configured to access the multiplexer and the subscriber end of the digital subscriber line Data between transceivers is processed to suppress crosstalk;
  • the determining module 111 is configured to: during the training of the filter 110 of the vector digital subscriber line, the feedback error value of the filter 110 and/or the swing of the filter 110 coefficient and a preset gate The limit values are compared, and it is judged according to the comparison result whether the filter satisfies the convergence condition;
  • the exiting module 112 is configured to exit the filter 110 training process when the determining module 111 determines that the filter 110 satisfies the convergence condition.
  • the determining module 111 includes:
  • the determining unit 1110 is configured to: when determining that the feedback error value of the filter is not greater than a preset threshold value ⁇ , determine that the filter satisfies the convergence condition; and/or, when determining that the filter coefficient has a smaller swing When the threshold value is set in advance, the determination filter satisfies the convergence condition.
  • the training process is performed in the Vectored-DSL precoder or canceller.
  • the training process is exited in time, which makes the training process shorter and shorter, reduces the interference of the precoder or canceller to other users during the training process, and reduces the probability of other users dropping the line.
  • the present invention can be implemented by hardware, or can be implemented by means of software plus a necessary general hardware platform.
  • the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (may be a CD-ROM, a U disk, a mobile hard disk, a ROM/RAM, a magnetic device).
  • a disc or optical disc, etc. includes instructions for causing a computer device (which may be a personal computer, server, or network device, etc.) to perform the methods described in various embodiments of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
  • Telephonic Communication Services (AREA)
  • Filters That Use Time-Delay Elements (AREA)

Description

一种快速退出训练的方法、 系统及装置 本申请要求于 2008 年 10 月 31 日提交中国专利局、 申请号为 200810172309.3、 发明名称为"一种快速退出训练的方法、 系统及装置"的中国 专利申请的优先权， 其全部内容通过引用结合在本申请中。
技术领域
本发明涉及通信领域中的向量 DSL ( Vectored-DSL )技术， 特别是涉及 一种快速退出训练的方法、 系统及装置。 背景技术
xDSL ( Digital Subscriber Line, 数字用户线 )是一种在电话双绞线传输的 高速数据传输技术， 随着 xDSL使用频带的提高， 高频段串扰的问题变得日益 突出。 当一捆电缆中有多路用户需要开通 xDSL业务时，会因为串扰问题使得 一些线路的速率变低， 性能变得不稳定， 甚至会出现不能开通的情况， 从而 最终导致 DSLAM ( Digital Subscriber Line Access Multiplexer, 数字用户线接 入复用器） 的出线率比较低。 在用户的电缆上有着多对双绞线， 各个双绞线 上可能运行多种不同的业务，各种类型的 xDSL同时工作时相互之间会产生串 扰， 其中某些线路会因为串扰问题使得性能急剧下降， 当线路比较长时， 某 些双绞线的线路根本不能开通任何形式的 DSL业务。
现有技术中， 通过在 DSLAM端使用预编码器或抵消器进行联合收发， 并通过信号处理的方法来抵消串扰干扰， 最终使每一路信号中均不存在串扰 干扰。
预编码器或 ·ί氏消器在训练过程中每次训练时间为固定值， 在训练过程中 耗时较长， 使得在训练过程中对相邻用户的干扰较大， 严重时会导致其它用 户掉线。 发明内容
本发明实施例提供一种快速退出训练的方法、 系统及装置， 以及时退出 训练过程， 减少对相邻用户的干扰。
为了达到上述目的， 本发明实施例提出了一种快速退出训练的方法， 包 括：
在向量数字用户线的滤波器训练过程中， 将所述滤波器的反馈误差值和 / 或所述滤波器系数的摆幅与预先设定的门限值进行比较， 根据比较结果判断 所述滤波器是否满足收敛条件；
当所述滤波器满足所述收敛条件时， 退出所述滤波器训练过程。
本发明实施例还提出了一种快速退出训练的系统， 包括：
数字用户线接入复用器， 通过数字用户线连接多个用户端的收发器， 其 中，
所述数字用户线接入复用器的向量数字用户线的滤波器在与所述多个用 户端的收发器在进行训练的过程中传输两两正交的序列；
所述数字用户线接入复用器， 用于在所述向量数字用户线的滤波器训练 过程中， 判断所述滤波器是否满足收敛条件， 当所述滤波器满足所述收敛条 件时， 退出所述滤波器训练过程。
本发明实施例还提出了一种数字用户线接入复用器， 包括：
向量数字用户线的滤波器， 用于对所述数字用户线接入复用器和用户端 的收发器间的数据进行处理以抑制串扰；
判断模块， 用于在向量数字用户线的滤波器训练过程中， 将所述滤波器 的反馈误差值和 /或所述滤波器系数的摆幅与预先设定的门限值进行比较， 根 据比较结果判断所述滤波器是否满足收敛条件；
退出模块， 用于当所述判断模块判断出所述滤波器满足所述收敛条件时， 退出所述滤波器训练过程。
本发明实施例具有以下优点： 通过在 Vectored-DSL预编码器或抵消器的训练过程中， 及时退出训练过 程， 使得训练过程的耗时变短， 减小了预编码器或抵消器在训练过程中对其 它用户的干扰， 减小其它用户掉线几率。 附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案， 下面将对实 施例或现有技术描述中所需要使用的附图作简单地介绍， 显而易见地， 下面 描述中的附图仅仅是本发明的一些实施例， 对于本领域普通技术人员来讲， 在不付出创造性劳动的前提下， 还可以根据这些附图获得其他的附图。
图 1为本发明实施例一提出的一种快速退出训练的方法流程图； 图 2为本发明实施例二提出的一种快速退出训练的方法流程图； 图 3为本发明实施例二提出的用户端到局端传输的示意图；
图 4为本发明实施例二提出的用户端到局端抵消实现方式示意图； 图 5为本发明实施例二提出的局端到用户端传输的示意图；
图 6为本发明实施例二提出的局端到用户端抵消实现方式示意图； 图 7为本发明实施例二提出的判断滤波器是否收敛示意图；
图 8为本发明实施例二提出的另一种判断滤波器是否收敛示意图 图 9为本发明实施例提出的下行滤波器训练的流程示意图；
图 10为本发明实施例提出的上行滤波器训练的流程示意图；
图 11为本发明实施例提出的一种快速退出训练的装置结构图。 具体实施方式
下面将结合本发明实施例中的附图， 对本发明实施例中的技术方案进行 清楚、 完整地描述， 显然， 所描述的实施例仅仅是本发明一部分实施例， 而 不是全部的实施例。 基于本发明中的实施例， 本领域普通技术人员在没有作 出创造性劳动前提下所获得的所有其他实施例 , 都属于本发明保护的范围。 本发明实施例一提出的一种快速退出训练的方法， 如图 1所示， 包括： 步骤 S101 , 在向量数字用户线的滤波器训练过程中， 判断所述滤波器是 否满足收敛条件；
步骤 S102, 当所述滤波器满足所述收敛条件时， 退出所述滤波器训练过 程。
可见， 本发明实施例中， 通过在 Vectored-DSL预编码器或抵消器的训练 过程中， 及时退出训练过程， 使得训练过程的耗时变短， 减小了预编码器或 抵消器在训练过程中对其它用户的干扰， 减小其它用户掉线几率。
本发明实施例二提出的一种快速退出训练的方法， 如图 2所示， 包括： 步骤 S201 , 在滤波器训练开始后， DSLAM初始化滤波器的系数。
具体的， 在 Vectored-DSL的串扰抵消时， 实现上分为上行方向和下行方 向。 上行方向传输是指从用户端到局端的传输， 如图 3 所示， 上行传输是在 接收端对接收到的各用户的数据进行联合接收处理， 在接收端使用串扰抵消 器（Canceller )对串扰信号进行抵消。 其具体的实现方式如图 4所示， 在接收 端通过设计滤波器， 对接收数据做联合串扰抵消处理， 以保证接收到的各用 户的数据没有串扰干扰。 在该接收端设计的滤波器为抵消器， 该抵消器的系 数为 W, 在初始化的时候， 该 W可以为预先设置的默认值或是 0, 该 W值可 以随时间变化。
同样的， 下行方向传输是指从局端到用户端的传输， 如图 5 所示， 下行 传输时在发送端做预编码（Precoding ) 以抑制线路中的串扰。 其具体的实现 方式如图 6 所示， 在发送端通过设计滤波器， 对传输数据做预处理， 以保证 传输到各用户的数据没有串扰干扰。 在发送端设计的滤波器为预编码器， 该 预编码器的系数为 P,在初始化的时候，该 P可以为预先设置的默认值或是 0, 该 P值可以随时间变化。
步骤 S202, DSLAM初始化发送功率的大小。
具体的， 该发送功率的大小是根据经验值选取的， 通常是一个较小值或 是某个保守值， 上述步骤 S201和步骤 S202并没有先后的顺序关系。
步骤 S203， 收发器发送训练序列。
具体的， 在上行方向， 所有的用户将通过各自的收发器发送两两正交的 序列， 其中， 所述两两正交的序列可以在哈达玛 （Hadamard )矩阵中选取， DSLAM根据接收到的序列计算反馈误差值，根据该反馈误差值计算出串扰信 道矩阵； 在下行方向， DSLAM向用户发送两两正交的序列， 其中， 所述两两 正交的序列可以在哈达玛 （Hadamard )矩阵中选取， 用户根据接收到的序列 计算反馈误差值， 并向 DSLAM发送反馈误差信息， DSLAM根据该反馈误差 信息获取反馈误差值， 根据该反馈误差值计算出串扰信道矩阵， 进一步地， 所述反馈误差信息由所述反馈误差值经缩放处理后获得。
在本实施例中， DSLAM可以根据如下公式（1 )计算出串扰信道矩阵：
Figure imgf000007_0001
其中， "为功率系数， T为 symbol的总数， X为发送的序列（Hadamard) , E为反馈误差值， t为第 t个 symbols, H为串扰信道系数。
步骤 S204, DSLAM判断滤波器的训练时间是否超时， 如果超时则进入 步骤 S206, 如果没有超时， 则进入步骤 S205。 上述训练时间是否超时具体为 预先设定一个时间， 该预先设定的时间可以根据经验值进行选取， 当滤波器 的训练时间超过该预先设定的时间时， 则判定为超时。
步骤 S205 , DSLAM根据反馈误差值和 /或滤波器的系数， 判断该滤波器 是否满足收敛条件。 当滤波器满足收敛条件时， 转到步骤 206, 当滤波器不满 足收敛条件时， 则转到步骤 203。
具体的， 该滤波器收敛的条件包括但不限于： 反馈误差值不大于某一预 先设定的门限值^ , 该门限值^是根据历史经验等因素确认的， 是可以发生 变化的。 如图 7所示， 在 tl时刻， 反馈误差值不大于预先设定的门限值^ , 满足收敛条件， 判定该滤波器收敛。 或 滤波器系数 P和 w在观察的时间段内， 其自身的变化量趋于平稳， 即摆 幅小于某一预先设定的门限值 eth ,该门限值 e;¾是根据历史经验等因素确认的， 是可以发生变化的。 如图 8所示， 在观察时间段内 （tl~tn ) ,找出摆幅的差值：
Figure imgf000008_0001
emm = m[ Pmn t, ), P (k, t2 ...P (k, tn ))
A e = Θ m a x一 Θ m i n 当 Ae小于 时， 则满足收敛条件， 当 Ae不小于 时， 则不满足收敛条件。 其中^ ^ 表示 n对 m用户的第 k个子载波在 t时刻预编码器的系数。 即滤波器系数的摆幅小于某一预先设定的门限值 满足收敛条件， 判 定该滤波器收敛。 当该滤波器满足该上述的收敛条件时， 则不需要继续训练， 并中止训练。
步骤 S206, 向对端发送结束训练滤波器的消息， 并且结束当前阶段的滤 波器训练。
可见， 本实施例中， 通过在 Vectored-DSL预编码器或抵消器的训练过程 中， 及时退出训练过程， 使得训练过程的耗时变短， 减小了预编码器或抵消 器在训练过程中对其它用户的干扰， 减小其它用户掉线几率。
本发明实施例提供一种快速退出训练的方法， 应用在下行滤波器的训练 过程中， 如图 9所示， 包括：
步骤 S901 , 发送端发送训练序列。
其中， 与上述步骤 S203—样， 该训练列序列可以为两两正交的序列， 具 体的， 所述两两正交的序列可以在哈达玛（Hadamard )矩阵中选取。
步骤 S902, 接收端开始接收训练序列。
具体的， 上述步骤 S901和步骤 S902是同时进行的， 即发送端发送训练 序列的同时接收端开始接收训练序列。
步骤 S903 , 接收端将反馈误差信息发送给发送端。 对于下行方向， 接收端通过 Back channel反馈这个误差信息给发送端。 Back channel 可以是底层的 SOC通道， 或者是上层 (2层)反馈信道信道。对于 上行方向， 接收端直接计算出误差并进行处理， 没有反馈这个过程。
步骤 S904 , 在发送端， 根据反馈误差信息获取反馈误差值， 并判断该反 馈误差值是否小于某一预先设定的门限值 e , 当判断出该反馈误差值小于预 先设定的门限值^时， 则判定滤波器收敛， 并向接收端发送中止训练的消息， 以提前结束训练， 转到步骤 S905 , 当判断出该反馈误差值不小于预先设定的 门限值 e 时， 则判定滤波器不收敛， 并继续发送训练序列， 转到步骤 S901 , 一直到结束训练为止。
在本实施例中， 接收端向发送端发送反馈误差值之前， 可以对该反馈误 差值进行缩放处理， 生成反馈误差信息， 以便于传输或者提高传输精度； 发 送端接收到反馈误差信息后， 可以对反馈误差信息进行反处理获得反馈误差 值。 例如： 接收端可以将反馈误差值 error缩小 a倍（error/a )生成反馈误差 信息发送给发送端， 发送端接收到该反馈误差信息（error/a )后， 将该反馈误 差信息扩大 a倍（ error/a x a )获得反馈误差值 error。
步骤 S905 , 发送端向接收端发送消息， 以通知接收端同时退出所述滤波 器训练过程。
具体的， 当训练的时间超出预先设定的门限时， 发送端向接收端发送中 止训练的消息并结束训练。
可见， 本实施例中， 通过在 Vectored-DSL预编码器或抵消器的训练过程 中， 及时退出训练过程， 使得训练过程的耗时变短， 减小了预编码器或抵消 器在训练过程中对其它用户的干扰， 减小其它用户掉线几率。
同样的， 本发明实施例提供一种快速退出训练的方法， 应用在上行滤波 器的训练过程中， 如图 10所示， 包括：
步骤 S1001 , 发送端发送训练序列。 其中， 与上述步骤 203 —样， 该训练列序列可以为两两正交的序列， 具 体的， 所述两两正交的序列可以在哈达玛（Hadamard )矩阵中选取。
步骤 S1002 , 接收端开始接收训练序列。
具体的， 上述步骤 S1001和步骤 S1002是同时进行的， 即发送端发送训 练序列的同时接收端开始接收训练序列。
步骤 S1003 ,接收端获取反馈误差值， 并判断该反馈误差值是否小于某一 预先设定的门限值 e* , 当判断出该反馈误差值小于预先设定的门限值 e 时， 则判定该滤波器收敛， 并向接收端发送中止训练的消息， 以提前结束训练， 转到步骤 S1004 , 当判断出该反馈误差值不小于预先设定的门限值 e 时，则判 定该滤波器不收敛， 并继续发送训练序列， 转到步骤 S1001 , —直到结束训练 为止。
步骤 S1004 ,接收端向发送端发送消息， 以通知发送端同时退出所述滤波 器训练过程。
具体的， 当训练的时间超出预先设定的门限值时， 接收端向发送端发送 中止训练的消息并结束训练。
可见， 本实施例中， 通过在 Vectored-DSL预编码器或抵消器的训练过程 中， 及时退出训练过程， 使得训练过程的耗时变短， 减小了预编码器或抵消 器在训练过程中对其它用户的干扰， 减小其它用户掉线几率。
本发明实施例还提出了一种快速退出训练的系统， 包括数字用户线接入 复用器， 通过数字用户线连接多个用户端的收发器， 其中，
所述数字用户线接入复用器的向量数字用户线的滤波器在与所述多个用 户端的收发器在进行训练的过程中传输两两正交的序列；
所述数字用户线接入复用器， 用于在所述向量数字用户线的滤波器训练 过程中， 判断所述滤波器是否满足收敛条件， 当所述滤波器满足所述收敛条 件时， 退出所述滤波器训练过程。
进一步地， 所述数字用户线接入复用器， 用于获得所述滤波器的反馈误 差， 确定所述反馈误差不大于预先设定的门限值 e 时， 判定所述滤波器满足 所述收敛条件， 退出所述滤波器训练过程。
进一步地， 所述数字用户线接入复用器， 用于在确定所述滤波器系数的 摆幅小于预先设定的门限值 ^时， 判定所述滤波器满足所述收敛条件， 退出 所述滤波器训练过程。
本发明实施例提供的快速退出训练的系统的具体实现方法可以参见本发 明实施例提供的快速退出训练的方法所述， 此处不再赘述。
可见， 本实施例中， 通过在 Vectored-DSL预编码器或抵消器的训练过程 中， 及时退出训练过程， 使得训练过程的耗时变短， 减小了预编码器或抵消 器在训练过程中对其它用户的干扰， 减小其它用户掉线几率。
本发明实施例还提出一种数字用户线接入复用器， 如图 11所示， 包括： 向量数字用户线的滤波器 110 ,用于对所述数字用户线接入复用器和用户 端的收发器间的数据进行处理以抑制串扰；
判断模块 111 ,用于在向量数字用户线的滤波器 110训练过程中，过程中， 将所述滤波器 110的反馈误差值和 /或所述滤波器 110系数的摆幅与预先设定 的门限值进行比较， 根据比较结果判断滤波器是否满足收敛条件；
退出模块 112, 用于判断模块 111判断出该滤波器 110满足所述收敛条件 时， 退出滤波器 110训练过程。
进一步的， 该判断模块 111包括：
判断单元 1110, 用于当确定滤波器的反馈误差值不大于预先设定的门限 值^时， 判定滤波器满足所述收敛条件； 和 /或， 用于当确定滤波器系数的摆 幅小于预先设定的门限值^时， 判定滤波器满足所述收敛条件。
本发明实施例提供的数字用户线接入复用器的具体实现方法可以参见本 发明实施例提供快速退出训练的方法所述， 此处不做赘述。
可见， 本实施例中， 通过在 Vectored-DSL预编码器或抵消器的训练过程 中， 及时退出训练过程， 使得训练过程的耗时变短， 减小了预编码器或抵消 器在训练过程中对其它用户的干扰， 减小其它用户掉线几率。
通过以上的实施方式的描述， 本领域的技术人员可以清楚地了解到本发 明可以通过硬件实现， 也可以可借助软件加必要的通用硬件平台的方式来实 现。 基于这样的理解， 本发明的技术方案可以以软件产品的形式体现出来， 该软件产品可以存储在一个非易失性存储介质 （可以是 CD-ROM, U盘， 移 动硬盘， ROM/RAM, 磁碟或光盘等）中， 包括若干指令用以使得一台计算机 设备（可以是个人计算机， 服务器， 或者网络设备等）执行本发明各个实施 例所述的方法。
以上所述， 仅为本发明的具体实施方式， 但本发明的保护范围并不局限 于此， 任何熟悉本技术领域的技术人员在本发明揭露的技术范围内， 可轻易 想到变化或替换， 都应涵盖在本发明的保护范围之内。 因此， 本发明的保护 范围应所述以权利要求的保护范围为准。

Claims

权 利 要求 书
1、 一种快速退出训练的方法， 其特征在于， 包括：
在向量数字用户线的滤波器训练过程中， 将所述滤波器的反馈误差值和 /或 所述滤波器系数的摆幅与预先设定的门限值进行比较， 根据比较结果判断所述 滤波器是否满足收敛条件；
当所述滤波器满足所述收敛条件时， 退出所述滤波器训练过程。
2、 如权利要求 1所述的方法， 其特征在于， 所述将所述滤波器的反馈误差 值和 /或所述滤波器系数的摆幅与预先设定的门限值进行比较， 根据比较结果判 断滤波器是否满足收敛条件包括：
当所述滤波器的反馈误差值不大于预先设定的门限值 e 时， 判定所述滤波 器满足所述收敛条件； 和 /或， 当所述滤波器系数的摆幅小于预先设定的门限值^时， 判定所述滤波器满 足所述收敛条件。
3、 如权利要求 1或 2所述的方法， 其特征在于， 所述反馈误差值的获取方 式包括：
根据接收端返回的反馈误差信息获取所述反馈误差值； 或，
从本地获取所述反馈误差值。
4、 如权利要求 1或 2所述的方法， 其特征在于， 所述滤波器系数的摆幅的 获取方式包括：
获取设定时间内所述滤波器系数的最大值和所述滤波器系数的最小值； 获取所述滤波器系数的摆幅， 所述滤波器系数的摆幅为所述滤波器系数的 最大值与所述滤波器系数的最小值的差值。
5、 如权利要求 1所述的方法， 其特征在于， 在所述将所述滤波器的反馈误 差值和 /或所述滤波器系数的摆幅与预先设定的门限值进行比较， 根据比较结果 判断滤波器是否满足收敛条件之前， 还包括：
当训练的时间超出预先设定的门限值时， 退出所述训练过程。
6、 如权利要求 1所述的方法， 其特征在于， 所述滤波器包括： 氏消器和 /或预编码器。
7、 一种快速退出训练的系统， 其特征在于， 包括：
数字用户线接入复用器， 通过数字用户线连接多个用户端的收发器， 其中， 所述数字用户线接入复用器的向量数字用户线的滤波器在与所述多个用户 端的收发器在进行训练的过程中传输两两正交的序列；
所述数字用户线接入复用器， 用于在所述向量数字用户线的滤波器训练过 程中， 判断所述滤波器是否满足收敛条件， 当所述滤波器满足所述收敛条件时， 退出所述滤波器训练过程。
8、 如权利要求 7所述的系统， 其特征在于，
所述数字用户线接入复用器， 用于获得所述滤波器的反馈误差， 确定所述 反馈误差不大于预先设定的门限值 e 时， 判定所述滤波器满足所述收敛条件， 退出所述滤波器训练过程。
9、 如权利要求 8所述的系统， 其特征在于，
所述数字用户线接入复用器， 用于在确定所述滤波器系数的摆幅小于预先 设定的门限值 ^时， 判定所述滤波器满足所述收敛条件， 退出所述滤波器训练 过程。
10、 一种数字用户线接入复用器， 其特征在于， 包括：
向量数字用户线的滤波器， 用于对所述数字用户线接入复用器和用户端的 收发器间的数据进行处理以抑制串扰；
判断模块， 用于在向量数字用户线的滤波器训练过程中， 将所述滤波器的 反馈误差值和 /或所述滤波器系数的摆幅与预先设定的门限值进行比较， 根据比 较结果判断所述滤波器是否满足收敛条件；
退出模块， 用于当所述判断模块判断出所述滤波器满足所述收敛条件时， 退出所述滤波器训练过程。
11、 如权利要求 10所述的装置， 其特征在于， 所述判断模块包括： 判断单元， 用于当确定所述滤波器的反馈误差值不大于预先设定的门限值
^时， 判定所述滤波器满足所述收敛条件； 和 /或，
用于当确定所述滤波器系数的摆幅小于预先设定的门限值 ^时， 判定所述 滤波器满足所述收敛条件。
PCT/CN2009/074543 2008-10-31 2009-10-21 一种快速退出训练的方法、系统及装置 WO2010048861A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP09823053.5A EP2343832B1 (en) 2008-10-31 2009-10-21 Method, system and device of exiting training rapidly
US13/097,950 US8537882B2 (en) 2008-10-31 2011-04-29 Method, system, and device for rapidly exiting training

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200810172309.3A CN101729459B (zh) 2008-10-31 2008-10-31 一种快速退出训练的方法、系统及装置
CN200810172309.3 2008-10-31

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US13/097,950 Continuation US8537882B2 (en) 2008-10-31 2011-04-29 Method, system, and device for rapidly exiting training

Publications (1)

Publication Number Publication Date
WO2010048861A1 true WO2010048861A1 (zh) 2010-05-06

Family

ID=42128249

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2009/074543 WO2010048861A1 (zh) 2008-10-31 2009-10-21 一种快速退出训练的方法、系统及装置

Country Status (4)

Country Link
US (1) US8537882B2 (zh)
EP (1) EP2343832B1 (zh)
CN (1) CN101729459B (zh)
WO (1) WO2010048861A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8537882B2 (en) 2008-10-31 2013-09-17 Huawei Technologies Co., Ltd. Method, system, and device for rapidly exiting training

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2867845C (en) * 2012-03-19 2018-05-29 Adaptive Spectrum And Signal Alignment, Inc. System for diagnosing and optimizing vectored dsl lines
CN103378927B (zh) * 2012-04-20 2019-01-11 中兴通讯股份有限公司 数据发送方法及装置
EP2775696A1 (en) 2013-03-05 2014-09-10 British Telecommunications public limited company Communications network

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1904942A (zh) * 2006-08-01 2007-01-31 北京航空航天大学 基于自适应判决反馈均衡技术的盲图像复原方法
CN1918811A (zh) * 2003-05-16 2007-02-21 索拉尔弗拉雷通讯公司 用于均衡和串扰减轻的方法和装置
JP2007274714A (ja) * 2007-05-07 2007-10-18 Matsushita Electric Works Ltd エコーキャンセラ
CN101197592A (zh) * 2006-12-07 2008-06-11 华为技术有限公司 远端串扰抵消方法、装置及信号发送装置和信号处理系统
CN101213826A (zh) * 2005-06-02 2008-07-02 适应性频谱和信号校正股份有限公司 Dsl系统训练

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030112861A1 (en) * 2001-12-18 2003-06-19 Erdogan Alper Tunga Method and system for adaptively training time domain equalizers
US20060029148A1 (en) * 2004-08-06 2006-02-09 Tsatsanis Michail K Method and apparatus for training using variable transmit signal power levels
US7817745B2 (en) 2005-06-02 2010-10-19 Adaptive Spectrum And Signal Alignment, Inc. Tonal precoding
WO2008024967A2 (en) * 2006-08-25 2008-02-28 Conexant Systems, Inc. Systems and methods for mimo precoding in an xdsl system
CN101729459B (zh) 2008-10-31 2016-01-13 华为技术有限公司 一种快速退出训练的方法、系统及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1918811A (zh) * 2003-05-16 2007-02-21 索拉尔弗拉雷通讯公司 用于均衡和串扰减轻的方法和装置
CN101213826A (zh) * 2005-06-02 2008-07-02 适应性频谱和信号校正股份有限公司 Dsl系统训练
CN1904942A (zh) * 2006-08-01 2007-01-31 北京航空航天大学 基于自适应判决反馈均衡技术的盲图像复原方法
CN101197592A (zh) * 2006-12-07 2008-06-11 华为技术有限公司 远端串扰抵消方法、装置及信号发送装置和信号处理系统
JP2007274714A (ja) * 2007-05-07 2007-10-18 Matsushita Electric Works Ltd エコーキャンセラ

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP2343832A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8537882B2 (en) 2008-10-31 2013-09-17 Huawei Technologies Co., Ltd. Method, system, and device for rapidly exiting training

Also Published As

Publication number Publication date
CN101729459B (zh) 2016-01-13
EP2343832A4 (en) 2012-08-01
EP2343832A1 (en) 2011-07-13
EP2343832B1 (en) 2013-09-18
US8537882B2 (en) 2013-09-17
CN101729459A (zh) 2010-06-09
US20110206102A1 (en) 2011-08-25

Similar Documents

Publication Publication Date Title
EP3560109B1 (en) Bandwidth reduction with beamforming and data compression
US7075967B2 (en) Wireless communication system using block filtering and fast equalization-demodulation and method of operation
TWI376113B (en) Communication system, transmission device, reception device and communication method
TWI630812B (zh) 實行全雙工通訊的資料傳輸方法及實行全雙工通訊的基地台
EP2039026B1 (en) Method and apparatus for interference reduction
WO2013087034A1 (zh) 一种干扰协调方法、基站及通信系统
WO2012149791A1 (zh) 一种串扰信道估计方法、装置及系统
KR102404719B1 (ko) 레거시 호환성을 갖는 저전력 모드
US20180234144A1 (en) Multi-User Multiple-Input And Multiple-Output For Digital Subscriber Line
WO2014100991A1 (zh) 一种线路的初始化方法及设备
CN103959755A (zh) 数字用户线收发器的高效功率管理的符号门控非连续模式
WO2010048861A1 (zh) 一种快速退出训练的方法、系统及装置
WO2010072053A1 (zh) 天线协作组确定方法和装置、基站协作组确定方法和装置
WO2015042802A1 (zh) 基带处理系统、基带信号处理方法和基站
WO2010048822A1 (zh) 一种功率调整的方法、系统及装置
CN101867978A (zh) 半静态调度周期管理方法、用户设备、基站及通信系统
CN106488578B (zh) 数据的传输方法、ap和用户节点
CN109792426B (zh) 使空循环前缀适合于频域空单载波通信系统的方法
CN103782579B (zh) 为xDSL传输系统启动非标准模式的方法及使用该方法的常驻网关
WO2017079946A1 (zh) 用于数字用户线初始化的方法和装置
CN106209164B (zh) 功率控制方法及装置
WO2019128374A1 (zh) 初始化信号的处理方法和装置
WO2019100808A1 (zh) 捆绑大小确定方法、用户终端和网络侧设备
CN104125043B (zh) 一种k用户系统的干扰对齐方法及设备
WO2014094273A1 (zh) 一种在线重配置的执行方法、装置和系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09823053

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2009823053

Country of ref document: EP