JP2001102972A - System and method for training echo canceler - Google Patents

System and method for training echo canceler

Info

Publication number
JP2001102972A
JP2001102972A JP27783699A JP27783699A JP2001102972A JP 2001102972 A JP2001102972 A JP 2001102972A JP 27783699 A JP27783699 A JP 27783699A JP 27783699 A JP27783699 A JP 27783699A JP 2001102972 A JP2001102972 A JP 2001102972A
Authority
JP
Japan
Prior art keywords
echo canceller
training
transmission
data
band
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
JP27783699A
Other languages
Japanese (ja)
Other versions
JP3395894B2 (en
Inventor
Morihiko Ota
守彦 大田
Katsutoshi Seki
克敏 関
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
NEC Corp
Original Assignee
NEC Corp
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 NEC Corp filed Critical NEC Corp
Priority to JP27783699A priority Critical patent/JP3395894B2/en
Publication of JP2001102972A publication Critical patent/JP2001102972A/en
Application granted granted Critical
Publication of JP3395894B2 publication Critical patent/JP3395894B2/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Landscapes

  • Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)

Abstract

PROBLEM TO BE SOLVED: To decrease an operation quantity by shortening the training time of an echo canceler of a transmission system like ADSL. SOLUTION: The training system for an echo canceler for multicarrier transmission whose down transmission band is wider than the up reception band is equipped with the echo canceler 200 which generates a cancel signal for line echo according to a signal having data arranged in a transmission carrier and subtracts it from received data and a training data generation part 110 which divides the transmission band by the width of the reception band, outputs a signal having training data arranged in a transmission carrier for each divided transmission band to the echo canceler, and trains the echo canceler.

Description

【発明の詳細な説明】DETAILED DESCRIPTION OF THE INVENTION

【0001】[0001]

【産業上の利用分野】本発明はエコーキャンセラに関す
る。特に、本発明は、ADSL(Asymmetric
Digital Subscriber Line)
のような伝送方式のエコーキャンセラ(EC)のトレー
ニング時間を短縮し、演算量を削減するエコーキャンセ
ラのトレーニングシステム及び方法に関する。
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an echo canceller. In particular, the present invention relates to ADSL (Asymmetric
Digital Subscriber Line)
The present invention relates to a training system and method for an echo canceller (EC) using a transmission system such as the one described above, which shortens the training time and reduces the amount of calculation.

【0002】[0002]

【従来の技術】上記ADSL方式のような下り使用帯域
が上り使用帯域より広いマルチキャリア伝送方式を例に
して、従来用いられたきたエコーキャンセラと、その問
題点とを説明する。ADSL方式は、マルチキャリア伝
送方式の一種であるDMT(Discrete Mul
tiTone)方式が採用されている。
2. Description of the Related Art A conventional echo canceller and its problems will be described with reference to a multicarrier transmission system in which a downstream band such as the ADSL system is wider than an upstream band. The ADSL system is a type of multi-carrier transmission system, which is a DMT (Discrete Mull).
tiTone) system.

【0003】DMT方式の詳細な説明は、「秒読み開
始、光の国のADSL(日経コミュニケーション、N
o.283、pp.92−111、1998.1.
4)」に示されている。ADSL方式の送信データは、
数ビット毎にまとめられ、各キャリアに振幅、位相の情
報として配置される。
[0003] The detailed description of the DMT method is described in "Starting Countdown, ADSL in the Country of Light (Nikkei Communication, N.K.
o. 283 pp. 92-111, 1998.1.
4) ". ADSL transmission data is:
The information is grouped every several bits and is arranged as amplitude and phase information on each carrier.

【0004】周波数域に配置されたデータは、IFFT
(逆高速フーリエ変換)により、時間域のデータとして
復調される。ADSL方式の受信データは、FFT(高
速フーリエ変換)により、周波数域のデータに変調され
る。ADSL方式において、回線エコーをキャンセルす
る手段として、従来、「METHOD AND APP
ARATUS FOR ECHO CANCELLAT
ION WITH DISCRETE MULTITO
NE MODULATION,United Stat
e Patent(5317596)」で示されるよう
な周波数―時間域エコーキャンセラが用いられていた。
[0004] Data arranged in the frequency domain is IFFT.
(Inverse fast Fourier transform) is demodulated as time domain data. ADSL received data is modulated into data in a frequency range by FFT (Fast Fourier Transform). In the ADSL system, as a means for canceling the line echo, conventionally, "METHOD AND APP"
ARATUS FOR ECHO CANCELLAT
ION WITH DISCRETE MULTITO
NE Modulation, United Stat
e-Patent (5317596) ".

【0005】図5はADSL方式の使用帯域を示す概略
図である。本図に示すように、キャリアは、4.312
5kHz毎に配置される。ADSLシステムの局側装置
では、送信帯域と比較して受信帯域が狭く、送信は0〜
255までのキャリアを使用し、受信は0〜31までの
キャリアを使用する。
FIG. 5 is a schematic diagram showing a band used in the ADSL system. As shown in this figure, the carrier is 4.312.
It is arranged every 5 kHz. In the station side device of the ADSL system, the reception band is narrower than the transmission band.
Up to 255 carriers are used, and reception uses 0 to 31 carriers.

【0006】図6は本発明の前提となるエコーキャンセ
ラのトレーニングシステムを用いたADSL送受信部の
構成を示すブロック図である。本図に示すように、AD
SL送受信部の送信部100はトレーニングデータを生
成するトレーニングデータ生成器110と、生成された
トレーニングデータを時間域のデータに復調するIFF
T(逆高速フーリエ変換)120と、復調されたデータ
を直列変換するパラレルシリアル変換部130と、直列
変換されたデータをディジタルからアナログに変換する
ディジタルアナログ変換部140と、アナログに変換さ
れた信号から不要な高周波成分を除去するローパスフィ
ルタ150とにより構成されている。
FIG. 6 is a block diagram showing the configuration of an ADSL transmission / reception unit using a training system for an echo canceller, which is a premise of the present invention. As shown in FIG.
The transmission unit 100 of the SL transmission / reception unit includes a training data generator 110 that generates training data, and an IFF that demodulates the generated training data into time-domain data.
T (Inverse Fast Fourier Transform) 120, a parallel-to-serial converter 130 for serializing demodulated data, a digital-to-analog converter 140 for converting serialized data from digital to analog, and a signal converted to analog And a low-pass filter 150 for removing unnecessary high-frequency components from the signal.

【0007】トレーニングデータ生成器110は、既知
のトレーニングデータXk,nを、前述のように、IF
FT120に出力するが、トレーニングデータXk,n
は、以下のようにして、作成される。まず、512ビッ
トの既知のランダムバイナリーデータを作成し、先頭か
ら2ビット毎に以下の変換を行い、256個の複素デー
タSk,nを作成する。
The training data generator 110 converts the known training data X k, n into IF
Output to the FT 120, the training data X k, n
Is created as follows. First, 512-bit known random binary data is created, and the following conversion is performed every two bits from the beginning to create 256 complex data S k, n .

【0008】 1ビット目 2ビット目 実部 虚部 0 0 → 1 1 0 1 → −1 1 1 0 → −1 −1 1 1 → 1 −11st bit 2nd bit real part imaginary part 00 → 1101 → -11 11 → -1 11 1 → 1 -1

【0009】次に、下記式(8)を用いて、トレーニン
グデータXk,nを生成する。 Xk,n=0 (n=0、N/2) Xk,n=Sk,n (0<n<N/2) Xk,n=Sk,N−n (n>N/2) …(8) ここで、kはトレーニング番号、nはキャリア番号、N
/2は送信キャリア数(ADSL方式では、256)で
ある。
Next, training data X k, n is generated using the following equation (8). X k, n = 0 (n = 0, N / 2) X k, n = S k, n (0 <n <N / 2) X k, n = S k, N−n (n> N / 2) ) (8) where k is the training number, n is the carrier number, and N
/ 2 is the number of transmission carriers (256 in the ADSL system).

【0010】トレーニングデータは、ハミルトニアン対
称性を持つため、IFFTの結果はNサンプルの実数の
みが残る。トレーニングデータ生成器110により生成
されたトレーニングデータXk, は、IFFT120
により時間域のデータに変換され、パラレルシリアル変
換部130、ディジタルアナログ変換部140、ローパ
スフィルタ150を経由し、さらに、2線−4線の変換
を行うハイブリッドトランス300を通り、経路400
に出力される。
Since the training data has Hamiltonian symmetry, only a real number of N samples remains in the IFFT result. The training data X k, n generated by the training data generator 110 is
Is converted into data in the time domain by the parallel-serial conversion unit 130, the digital-analog conversion unit 140, the low-pass filter 150, and further passes through the hybrid transformer 300 that performs 2-wire to 4-wire conversion.
Is output to

【0011】次に、ADSL送受信部の受信部1000
はハイブリッドトランス300に接続され不要な高周波
成分を除去するローパスフィルタ1100と、高周波成
分が除去されたアナログ信号をディジタル信号に変換す
るアナログディジタル変換部1200と、アナログディ
ジタル変換部1200に接続される減算器1300と、
減算器1300の出力データを並列変換するシリアルパ
ラレル変換部1400と、並列変換されたデータを周波
数域のデータに復調するFFT1500と、復調された
データを減算処理する複素減算器1600とにより構成
される。
Next, the receiving section 1000 of the ADSL transmitting / receiving section
Is a low-pass filter 1100 connected to the hybrid transformer 300 for removing unnecessary high-frequency components, an analog-to-digital converter 1200 for converting an analog signal from which high-frequency components have been removed to a digital signal, and a subtractor connected to the analog-to-digital converter 1200 Vessel 1300;
It is composed of a serial-parallel conversion unit 1400 that converts the output data of the subtractor 1300 in parallel, an FFT 1500 that demodulates the parallel-converted data into frequency-domain data, and a complex subtractor 1600 that performs a subtraction process on the demodulated data. .

【0012】受信データはハイブリッドトランス30
0、ローパスフィルタ1100、1200、シリアルパ
ラレル変換部1400、FFT1500を経由して周波
数域のデータに復調される。次に、ADSL送受信部の
周波数域−時間域エコーキャンセラ200は、受信部1
000で減算処理するために、送信部100からの送信
データを基にしてハイブリッドトランス300で発生す
る回線エコーをキャンセルするキャンセルデータを形成
する。
The received data is transmitted to the hybrid transformer 30
0, and is demodulated to frequency band data via low-pass filters 1100 and 1200, serial-parallel converter 1400, and FFT 1500. Next, the frequency domain-time domain echo canceller 200 of the ADSL transmission / reception section includes the reception section 1
In order to perform the subtraction process by 000, cancellation data for canceling the line echo generated in the hybrid transformer 300 is formed based on the transmission data from the transmission unit 100.

【0013】このため、ADSL送受信部の周波数域−
時間域エコーキャンセラ200は、トレーニングデータ
生成部110からのトレーニングデータ、複素減算器1
600からの受信データを入力し、複素減算器1600
に出力を行なう周波数域エコーキャンセラ210と、周
波数域エコーキャンセラ(EC)210の出力データを
時間域のデータに復調するIFFT220と、IFFT
220により復調されたデータ、IFFT120により
復調されたデータを入力し、減算器1300に出力を行
なう時間域エコーキャンセラ(EC)230とにより構
成される。
Therefore, the frequency range of the ADSL transmitting / receiving section
The time domain echo canceller 200 includes the training data from the training data generator 110 and the complex subtractor 1.
Receiving the received data from the complex subtractor 1600
Frequency echo canceller 210 for outputting the data to the frequency domain echo canceller (EC) 210, IFFT 220 for demodulating the output data of the frequency domain echo canceller 210 to time domain data,
A time domain echo canceller (EC) 230 that inputs data demodulated by 220 and data demodulated by IFFT 120 and outputs the data to a subtractor 1300.

【0014】このように、上記周波数域−時間域エコー
キャンセラ200では周波数域エコーキャンセラ21
0、時間域エコーキャンセラ230が併用され、トレー
ニング生成器110からのトレーニングデータが用いら
れ、回線エコーがキャンセルされる最適な条件設定を行
うためのトレーニングが行われる。周波数域エコーキャ
ンセラ210は、各キャリア毎に用意される複素数の係
数を持つ1タップのフィルタにより構成される。
As described above, the frequency-domain / time-domain echo canceller 200 uses the frequency-domain echo canceller 21.
0, the time domain echo canceller 230 is used in combination, the training data from the training generator 110 is used, and the training for setting the optimal condition for canceling the line echo is performed. The frequency band echo canceller 210 is configured by a one-tap filter having a complex coefficient prepared for each carrier.

【0015】時間域エコーキャンセラ230は、FIR
(Finite ImpulseResponse)フ
ィルタにより構成される。周波数域エコーキャンセラ2
10は、下記式(9)により、トレーニングデータX
k,nの各キャリアに対応したキャリアのタップ係数C
k,nをかけてデータY eck,nを算出する。
The time-domain echo canceller 230 has an FIR
(Finite ImpulseResponse) filter. Frequency range echo canceller 2
10 is the training data X by the following equation (9).
Tap coefficient C of carrier corresponding to each carrier of k and n
Data Y multiplied by k and n ec k, n is calculated.

【0016】 Y eck,n=Ck,n×Xk,n …(9) Y eck,n:シンボルkにおけるキャリアnの周波
数域エコーキャンセラ210の内部データ Xk,n:シンボルkにおけるキャリアnのトレーニン
グデータ Ck,n:シンボルkにおけるキャリアnの周波数域エ
コーキャンセラ210のタップ係数
[0016] Y ec k, n = C k, n × X k, n ... (9) Y ec k, n : internal data of carrier n frequency band echo canceller 210 in symbol k X k, n : training data of carrier n in symbol k C k, n : tap of carrier n frequency band echo canceller 210 in symbol k coefficient

【0017】周波数域エコーキャンセラ210は、送信
キャリアに対応したデータY ec k,nを下記式(1
0)により受信使用帯域まで圧縮し、複素減算器160
0に出力する。 Y ec’k,n=Y eck,n+Y eck,64+n+ Y eck,128+n+Y eck,192+n+Y eck,256+n+ Y eck,320+n+Y eck,384+n+Y eck,448+n (0≦n<M) …(10)
The frequency band echo canceller 210 transmits
Data Y corresponding to carrier ec k, nTo the following equation (1
0) to compress the signal to the reception use band, and
Output to 0. Y ec 'k, n= Y eck, n+ Y eck, 64 + n+ Y eck, 128 + n+ Y eck, 192 + n+ Y eck, 256 + n+ Y eck, 320 + n+ Y eck, 384 + n+ Y eck, 448 + n (0 ≦ n <M) (10)

【0018】M:受信のキャリア数(=64) Y ec’k,n:周波数域エコーキャンセラ210の
出力データ 周波数域エコーキャンセラ210のタップ係数は、下記
式(11)に示すLMS(Least Mean Sq
uare)アルゴリズムによって更新される。
M: the number of receiving carriers (= 64) Y ec ' k, n : output data of the frequency band echo canceller 210 The tap coefficient of the frequency band echo canceller 210 is represented by LMS (Least Mean Sq) shown in the following equation (11).
ua) algorithm.

【0019】 Ck+1,n k,n+μEk,n k,n (0≦n<512) …(11) Ek,n=E’k,mod(n,64) …(12) μ:係数更新のステップサイズ X k,n:トレーニングデータXk,nの複素共役 E’k,n:識別誤差Ck + 1, n= Ck, n+ ΜEk, nX* k, n (0 ≦ n <512) (11) Ek, n= E 'k, mod (n, 64) ... (12) μ: Step size of coefficient update X* k, n: Training data Xk, nComplex conjugate of E 'k, n: Identification error

【0020】半2重通信時には、対向側装置は送信を停
止しているので、複素減算器1600の出力は識別誤差
E’k,nのみとなる。時間域エコーキャンセラ230
のタップ係数は、周波数域エコーキャンセラ210で求
めたタップ係数をIFFT220で逆高速フーリエ変換
して算出される。
At the time of half-duplex communication, the output of the complex subtracter 1600 is only the identification error E'k, n since the opposite device has stopped transmitting. Time-domain echo canceller 230
Are calculated by performing an inverse fast Fourier transform on the tap coefficients obtained by the frequency domain echo canceller 210 by the IFFT 220.

【0021】時間域エコーキャンセラ230は2つの前
処理を行う。第一に、周波数域エコーキャンセラ210
では消去できない過去のシンボルによって生じるシンボ
ル間干渉を消去する。第二に、現在の送信シンボルによ
って、次のシンボル区間に生じるシンボル間干渉を、現
在のシンボルの先頭に複写する。
The time domain echo canceller 230 performs two preprocessings. First, the frequency domain echo canceller 210
Cancels the inter-symbol interference caused by past symbols that cannot be eliminated by the above. Second, the inter-symbol interference caused by the current transmission symbol in the next symbol section is copied to the head of the current symbol.

【0022】この2つの処理により、FFT区間には、
現在のシンボルによって生じるエコーだけが、周期性を
持って残る。このエコーは、周波数域エコーキャンセラ
210で消去される。時間域エコーキャンセラ230の
出力Y ecは、送受信でサンプリングレートが異な
るので、下記式(13)で示されるように、8サンプル
毎に1つづつ取り出され、64サンプルのデータに変換
される。
By these two processes, in the FFT section,
Only echoes caused by the current symbol remain with periodicity. This echo is canceled by the frequency band echo canceller 210. Output Y of time domain echo canceller 230 Since ec n has a different sampling rate between transmission and reception, as shown in the following equation (13), ec n is extracted one by one every eight samples and converted into data of 64 samples.

【0023】 Y ec’n’=Y ec8n …(13) ここで、Y ec’n’は、間引きされた後の時間域エ
コーキャンセラ230の出力データである。図7は式
(13)によるデータ間引きの様子を示す模式図であ
る。本図に示すように、送信シンボル512のデータが
データ間引き後64データになる。
Y ec 'n' = Y ec 8n (13) where Y ec ′ n ′ is output data of the time-domain echo canceller 230 after being decimated. FIG. 7 is a schematic diagram showing a state of data thinning-out according to equation (13). As shown in the drawing, the data of the transmission symbol 512 becomes 64 data after the data is decimated.

【0024】全二重通信時の場合も半二重通信時と同様
に、タップ係数の更新、エコーの消去が行われる。この
ように、エコーキャンセラのトレーニングシステムは、
最適なエコーキャンセラのデータの設定を行う。
In the case of full-duplex communication, the update of tap coefficients and the elimination of echo are performed in the same manner as in the case of half-duplex communication. Thus, the training system of the echo canceller is
Set the optimal echo canceller data.

【0025】[0025]

【発明が解決しょうとする課題】ところで、上記エコー
キャンセラのトレーニングシステムでは、全キャリアに
データを配置してトレーニングデータを送信していたた
め、1つの受信キャリアに、複数の送信キャリアのエコ
ーが重畳されていた。このため、十分なエコー抑圧特性
を得るためには、周波数域エコーキャンセラ210のス
テップサイズを小さくする必要があった。
By the way, in the training system of the echo canceller, since the training data is transmitted by arranging the data on all the carriers, the echoes of a plurality of transmitting carriers are superimposed on one receiving carrier. I was Therefore, in order to obtain sufficient echo suppression characteristics, it is necessary to reduce the step size of the frequency band echo canceller 210.

【0026】しかしながら、ステップサイズを小さくす
ると、周波数域エコーキャンセラ210の収束時間が長
くなり、結果的に、エコーキャンセラ全体のトレーニン
グ時間が長期化するという問題がある。さらに、逆高速
フーリエ変換を含んだトレーニングを繰り返して、時間
域エコーキャンセラ230のタップ係数を算出している
ため、トレーニング時の演算量が多くなるという問題が
ある。
However, when the step size is reduced, the convergence time of the frequency band echo canceller 210 becomes longer, and as a result, there is a problem that the training time of the entire echo canceller becomes longer. Further, since the training including the inverse fast Fourier transform is repeated to calculate the tap coefficients of the time-domain echo canceller 230, there is a problem that the amount of calculation during training increases.

【0027】したがって、本発明は上記問題点に鑑み
て、トレーニング時間の短縮、演算量の削減を可能にす
るエコーキャンセラのトレーニングシステム及び方法を
提供することを目的とする。
Accordingly, it is an object of the present invention to provide a training system and method for an echo canceller, which can reduce the training time and the amount of calculation in view of the above problems.

【0028】[0028]

【課題を解決するための手段】本発明は前記問題点を解
決するために、下りの送信帯域が上りの受信帯域よりも
広いマルチキャリア伝送のエコーキャンセラのトレーニ
ングシステムにおいて、データが送信キャリアに配置さ
れた信号に基づいて、回線エコーのキャンセル信号を形
成して受信データに減算処理するためのエコーキャンセ
ラと、送信帯域を受信帯域の幅に分割し、分割された1
つの送信帯域ごとの送信キャリアにトレーニングデータ
が配置される信号を前記エコーキャンセラに出力して前
記エコーキャンセラのトレーニングを行うためのトレー
ニングデータ生成部とを備えることを特徴とするエコー
キャンセラのトレーニングシステムを提供する。
SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a training system for a multicarrier transmission echo canceller in which a downstream transmission band is wider than an upstream reception band, wherein data is allocated to transmission carriers. An echo canceller for forming a line echo cancel signal based on the signal thus obtained and subtracting the same from the received data, and dividing the transmission band into the width of the reception band,
A training data generating unit for outputting a signal in which training data is arranged on a transmission carrier for each transmission band to the echo canceller and performing training of the echo canceller. provide.

【0029】この手段により、送信帯域を受信帯域幅で
分割し、分割した1つの送信帯域毎にデータをのせて送
信することにより、従来のように複数の送信キャリアの
エコーは重畳されなくなった。これにより、LMS法の
ような漸近的にタップ係数を求める方法ではなく、周波
数領域において受信データをトレーニングデータで割る
ことによりエコーキャンセラのタップ係数を算出するこ
とができ、トレーニング時間を短縮することができる。
By this means, the transmission band is divided by the reception bandwidth, and data is transmitted for each divided transmission band, so that echoes of a plurality of transmission carriers are no longer superimposed as in the prior art. Thus, the tap coefficient of the echo canceller can be calculated by dividing the received data by the training data in the frequency domain, instead of the method of asymptotically obtaining the tap coefficient as in the LMS method, and the training time can be reduced. it can.

【0030】好ましくは、前記マルチキャリア伝送の方
式はADSL方式である。この手段により、ADSL方
式のようなマルチキャリア伝送方式に本発明の適用が可
能になる。好ましくは、前記エコーキャンセラは、前記
トレーニング生成器から周波数域の前記信号を入力し各
キャリア毎に用意される複素数の係数をもつ1タップの
フィルタで構成される周波数域エコーキャンセラと、前
記トレーニング生成器から時間域の前記信号を入力しF
IRフィルタで構成される時間域エコーキャンセラとか
らなり、前記時間域エコーキャンセラのタップ係数は前
記周波数域エコーキャンセラで求められたタップ係数を
逆高速フーリエ変換して算出される。
Preferably, the multi-carrier transmission system is an ADSL system. By this means, the present invention can be applied to a multi-carrier transmission system such as the ADSL system. Preferably, the echo canceller receives the signal in the frequency domain from the training generator, and includes a one-tap filter having a complex coefficient prepared for each carrier. Input the signal in the time domain from the
A time-domain echo canceller composed of an IR filter, and the tap coefficient of the time-domain echo canceller is calculated by performing an inverse fast Fourier transform on the tap coefficient obtained by the frequency-domain echo canceller.

【0031】この手段により、周波数域エコーキャンセ
ラでは消去できない過去のシンボルによって生じるシン
ボル間干渉が消去される。現在の送信シンボルによっ
て、次のシンボル区間に生じるシンボル間干渉が、現在
のシンボルの先頭に複写される。FFT区間には、現在
のシンボルによって生じるエコーだけが、周期性を持っ
て残る。このエコーは、周波数域エコーキャンセラで消
去される。
By this means, inter-symbol interference caused by past symbols that cannot be canceled by the frequency band echo canceller is canceled. The inter-symbol interference generated in the next symbol section by the current transmission symbol is copied to the head of the current symbol. In the FFT section, only the echo generated by the current symbol remains with periodicity. This echo is canceled by the frequency band echo canceller.

【0032】好ましくは、前記エコーキャンセラに初期
値算出回路が設けられ、前記初期値算出回路は、前記周
波数域エコーキャンセラのタップ係数の初期値を算出す
る。この手段により、初期値算出回路で周波数域エコー
キャンセラのタップ係数を全て算出した後、逆高速フー
リエ変換を1回行うだけで時間域エコーキャンセラのタ
ップ係数を求めることができ、トレーニング時の演算量
を削減することができる。
Preferably, the echo canceller is provided with an initial value calculating circuit, and the initial value calculating circuit calculates an initial value of a tap coefficient of the frequency band echo canceller. By this means, after calculating all the tap coefficients of the frequency domain echo canceller by the initial value calculation circuit, the tap coefficient of the time domain echo canceller can be obtained only by performing the inverse fast Fourier transform once, and the amount of computation during training Can be reduced.

【0033】好ましくは、前記周波数域エコーキャンセ
ラと前記時間域エコーキャンセラの間に時間域補正部を
設け、前記時間域補正部は、分割された送信帯の送信キ
ャリアがナイキストの受信キャリアに対応することに起
因してタップ係数の補正を行う。この手段により、回線
エコー抑圧特性の劣化を改善することが可能になる。
Preferably, a time-domain correction unit is provided between the frequency-domain echo canceller and the time-domain echo canceller, and the time-domain correction unit corresponds to a Nyquist reception carrier whose transmission carrier in the divided transmission band corresponds to Nyquist. Therefore, the tap coefficient is corrected. By this means, it is possible to improve the deterioration of the line echo suppression characteristic.

【0034】さらに、本発明は、下りの送信帯域が上り
の受信帯域よりも広いマルチキャリア伝送のエコーキャ
ンセラのトレーニング方法において、データが送信キャ
リアに配置された信号に基づいて、回線エコーのキャン
セル信号を形成して受信データに減算処理する工程と、
送信帯域を受信帯域の幅に分割し、分割された1つの送
信帯域ごとの送信キャリアにトレーニングデータが配置
される信号を前記エコーキャンセラに出力して前記エコ
ーキャンセラのトレーニングを行う工程とを備えること
を特徴とするエコーキャンセラのトレーニング方法を提
供する。
Further, the present invention provides a training method of an echo canceller for multicarrier transmission in which a downstream transmission band is wider than an upstream reception band, wherein a line echo cancel signal is provided based on a signal in which data is arranged on a transmission carrier. And forming a subtraction process on the received data;
Dividing the transmission band into the width of the reception band, outputting a signal in which training data is allocated to the divided transmission carriers for each transmission band to the echo canceller, and training the echo canceller. And a training method for an echo canceller characterized by the following.

【0035】この手段により、上記発明と同様に、送信
帯域を受信帯域幅で分割し、分割した1つの送信帯域毎
にデータをのせて送信することにより、従来のように複
数の送信キャリアのエコーは重畳されなくなった。これ
により、LMS法のような漸近的にタップ係数を求める
方法ではなく、周波数領域において受信データをトレー
ニングデータで割ることによりエコーキャンセラのタッ
プ係数を算出することができ、トレーニング時間を短縮
することができる。
By this means, the transmission band is divided by the reception bandwidth, and data is placed for each of the divided transmission bands and transmitted in the same manner as in the above-mentioned invention. Is no longer superimposed. Thus, the tap coefficient of the echo canceller can be calculated by dividing the received data by the training data in the frequency domain, instead of the method of asymptotically obtaining the tap coefficient as in the LMS method, and the training time can be reduced. it can.

【0036】[0036]

【発明の実施の形態】以下、本発明の実施の形態につい
て図面を参照して説明する。図1は本発明に係るエコー
キャンセラのトレーニングシステムを用いたADSL送
受信部の構成を示すブロック図である。本図に示すよう
に、図6の構成と異なるものは、周波数域−時間域エコ
ーキャンセラ200の初期値算出回路240である。
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an ADSL transmission / reception unit using a training system for an echo canceller according to the present invention. As shown in the figure, what differs from the configuration of FIG. 6 is an initial value calculation circuit 240 of the frequency-time-domain echo canceller 200.

【0037】初期値算出回路240はFFT1500の
出力を入力し、周波数域エコーキャンセラ210に出力
を行う。図2は初期値算出回路240の詳細を説明する
ブロック図である。本図に示すように、初期値算出回路
240は、トレーニングデータ生成器243と、受信部
1000のFFT1500の出力を入力とする帯域拡張
部241と、帯域拡張部241の出力を入力とする平均
化回路242と、トレーニングデータ生成器243と平
均化回路242との出力を入力とし周波数域エコーキャ
ンセラ210に出力を行う複素割り算器244を有す
る。
The initial value calculation circuit 240 receives the output of the FFT 1500 and outputs the output to the frequency domain echo canceller 210. FIG. 2 is a block diagram illustrating details of the initial value calculation circuit 240. As shown in the figure, the initial value calculation circuit 240 includes a training data generator 243, a band extension unit 241 that receives an output of the FFT 1500 of the reception unit 1000 as an input, and an averaging unit that receives an output of the band extension unit 241 as an input. It has a circuit 242 and a complex divider 244 that receives the outputs of the training data generator 243 and the averaging circuit 242 as input and outputs to the frequency domain echo canceller 210.

【0038】周波数域−時間域エコーキャンセラ200
のトレーニングは、以下の手順〜で行われる。な
お、初期値算出回路240のトレーニングデータ生成器
243は、送信部100のトレーニングデータ生成器1
10で出力されるトレーニングデータXk,nと同じト
レーニングデータX’k,nを出力する。また、Nは送
信のキャリア数の2倍、Mは受信のキャリア数の2倍を
表す(N、Mは2のべき乗、N>M)。
Frequency domain-time domain echo canceller 200
Training is performed by the following procedures. The training data generator 243 of the initial value calculation circuit 240 is the same as the training data generator 1 of the transmitting unit 100.
The same training data X ′ k, n as the training data X k, n output at 10 is output. N represents twice the number of transmission carriers and M represents twice the number of reception carriers (N and M are powers of 2 and N> M).

【0039】 トレーニングデータ生成器110は、
kM/2から((k+1)M/2)までのキャリアと
(N−(k+1)M/2+1)から(N−kM/2)ま
で(但し、k=0の場合は(N−M/2+1)から(N
−1)までである)のキャリアにのみパワーを持つトレ
ーニングデータXk,nを生成し、Lシンボル繰り返し
て送信する。
The training data generator 110
carrier from kM / 2 to ((k + 1) M / 2) and from (N- (k + 1) M / 2 + 1) to (N-kM / 2) (however, when k = 0, (NM-2 + 1) ) To (N
The training data X k, n having power only in the carriers up to -1) is generated, and L symbols are repeatedly transmitted.

【0040】ここで、kは初期値算出のトレーニング番
号を示し、初期値は0で、N/M−1までの値を取る。
図3は送信するトレーニングデータXk,nを説明する
模式(ここではk=0〜7)を示す図である。本図に示
すように、トレーニングデータ生成部110により生成
されたトレーニングデータXk,nは、k=0の場合に
は0番目から31番目、481番目から511番目のキ
ャリアにデータが配置され、k=1の場合には32番目
から63番目、449番目から480番目のキャリアに
データが配置され、…、k=7の場合には224番目か
ら288番目のキャリアにデータが配置される。
Here, k indicates a training number for calculating an initial value, and the initial value is 0 and takes a value up to N / M-1.
FIG. 3 is a diagram illustrating a model (here, k = 0 to 7) for explaining the training data X k, n to be transmitted. As shown in the figure, the training data X k, n generated by the training data generating unit 110 is arranged in the 0th to 31st and 481th to 511th carriers when k = 0, When k = 1, data is allocated to the 32nd to 63rd, 449th to 480th carriers,..., when k = 7, data is allocated to the 224th to 288th carriers.

【0041】従来のエコーキャンセラでは、前述したよ
うに、全キャリアにデータを配置してトレーニングデー
タを送信していたため、1つの受信キャリアに、複数の
送信キャリアのエコーが重畳されていた。このため、十
分なエコー抑圧特性を得るためにに、周波数域エコーキ
ャンセラ210のステップサイズを小さくする必要があ
った。
In the conventional echo canceller, as described above, the training data is transmitted by arranging the data on all the carriers. Therefore, echoes of a plurality of transmission carriers are superimposed on one reception carrier. Therefore, in order to obtain sufficient echo suppression characteristics, it is necessary to reduce the step size of the frequency band echo canceller 210.

【0042】ステップサイズを小さくすると、周波数域
エコーキャンセラ210の収束時間が長くなり、結果的
に、エコーキャンセラ全体のトレーニング時間が長期化
していた。これに対して、本発明のエコーキャンセラで
は、送信帯域を受信帯域幅で分割し、分割した1つの帯
域毎にデータをのせて送信することにより、複数の送信
キャリアのエコーは重畳されない。
When the step size is reduced, the convergence time of the frequency band echo canceller 210 becomes longer, and as a result, the training time of the entire echo canceller becomes longer. On the other hand, in the echo canceller of the present invention, the transmission band is divided by the reception bandwidth, and data is transmitted for each divided band so that echoes of a plurality of transmission carriers are not superimposed.

【0043】これにより、LMS法のような漸近的にタ
ップ係数を求める方法ではなく、周波数領域において受
信データをトレーニングデータで割ることによりエコー
キャンセラのタップ係数を算出することができ、トレー
ニング時間を短縮することができる。 帯域拡張部241は、トレーニングデータXk,n
送信時におけるFFT1500の出力Yk,nの帯域
を、下記式(1)によって拡張する。
Thus, the tap coefficient of the echo canceller can be calculated by dividing the received data by the training data in the frequency domain, instead of the method of asymptotically obtaining the tap coefficient as in the LMS method, thereby shortening the training time. can do. The band extending unit 241 provides the training data X k, n
The band of the output Y k, n of the FFT 1500 at the time of transmission is extended by the following equation (1).

【0044】 Y’k, n=Yk,mod(n,M) …(1) Yk,mod(n,M):FFT1500の出力 Y’k, n:帯域拡張部241の出力(キャリア番号n
=0〜N) ここで、mod(n、M)はnをMで割った余りを表
す。平均化回路242は、帯域拡張部241の出力デー
タをLシンボル間平均化し、データY’’k, nを複素
割り算器244に出力する。
Y ′ k, n = Y k, mod (n, M) (1) Y k, mod (n, M) : output of FFT 1500 Y ′ k, n : output of band extender 241 (carrier number) n
= 0 to N) Here, mod (n, M) represents a remainder obtained by dividing n by M. The averaging circuit 242 averages the output data of the band extending section 241 between L symbols, and outputs the data Y ″ k, n to the complex divider 244.

【0045】複素割り算器244では、トレーニングデ
ータX’k,nのパワーが存在するキャリアに対し、下
記式(2)を用いて、周波数域エコーキャンセラ210
のタップ係数初期値Ck,nを出力する。 Ck,n=Y’’k, n/X’k,n …(2) とをk=1からN/M−1まで繰り返す。 すべてのタップ係数初期値が求まったら、複素割り
算器244は、周波数域エコーキャンセラ210に値を
出力し、タップ係数初期値の設定を行う。
The complex divider 244 uses the frequency domain echo canceller 210 for the carrier in which the power of the training data X'k, n exists, using the following equation (2).
Is output as the tap coefficient initial value C k, n . C k, n = Y ″ k, n / X ′ k, n (2) are repeated from k = 1 to N / M−1. When all tap coefficient initial values are obtained, the complex divider 244 outputs a value to the frequency-domain echo canceller 210, and sets tap coefficient initial values.

【0046】周波数域−時間域エコーキャンセラ200
は、周波数域エコーキャンセラ210のタップ係数が設
定されると、従来方式と同様に動作を行う。以下に本発
明におけるエコーキャンセラのトレーニングシステムの
動作について説明を行う。ADSL方式の局側装置を例
として、送信は512キャリア、受信は64キャリアで
行われる。
Frequency-Time Echo Canceller 200
When the tap coefficient of the frequency-domain echo canceller 210 is set, the same operation is performed as in the conventional method. The operation of the echo canceller training system according to the present invention will be described below. Taking an ADSL station apparatus as an example, transmission is performed with 512 carriers and reception is performed with 64 carriers.

【0047】トレーニングデータ生成器110は、既知
のトレーニングデータXk,nを生成する。ここで、k
はトレーニング番号(k=0〜N/M−1)、Nは送信
キャリア数の2倍(ここでは、512)、Mは受信キャ
リアの数の2倍(ここでは、64)、nはキャリア番号
(n=0〜N−1)を表す。トレーニングデータX
k,nは、例えば以下のように、生成される。
The training data generator 110 generates known training data X k, n . Where k
Is the training number (k = 0 to N / M-1), N is twice the number of transmission carriers (here, 512), M is twice the number of reception carriers (here, 64), and n is the carrier number (N = 0 to N-1). Training data X
k and n are generated, for example, as follows.

【0048】まず、既知の擬似ランダムバイナリデータ
を64ビット生成し、先頭から2ビットずつ以下の変換
を行い、32個の複素データSを作成する。 1ビット目 2ビット目 実部 虚部 0 0 → 1 1 0 1 → −1 1 1 0 → −1 −1 1 1 → 1 −1
Firstly, the known pseudo-random binary data 64-bit product, two bits from the head performs the following transformation to generate 32 complex data S n. 1st bit 2nd bit real part imaginary part 0 0 → 1 1 0 1 → -1 1 1 1 0 → -1 1 1 1 → 1 -1

【0049】次に、下記式(3)を用いて、トレーニン
グデータXk,nを生成する。下記式(3)により生成
されたトレーニングデータXk,nは、ハミルトニアン
対称と呼ばれ、このハミルトニアン対称であるトレーニ
ングデータXk,nを、IFFTすると、実数のみのデ
ータとなる性質を持っている。 Xk,n=0 (n=0、N/2)
Next, the training data X k, n is generated by using the following equation (3). Training data X k, n generated by the following equation (3) is called a Hamiltonian symmetry, training data X k is the Hamiltonian symmetry, the n, IFFT result, has a property as a real-only data . X k, n = 0 (n = 0, N / 2)

【0050】 Xk,n=Sn−kM/2 (kM/2≦n≦(k+1)M/2−1) Xk,n=S N−kM/2−n (N−(k+1)M/2+1≦n≦ N−kM/2)、但し、k=0の時は、 (N−(k +1)M/2+1≦n≦ N−kM/2−1である。 Xk,n=0 (その他) …(3)X k, n = S n−kM / 2 (kM / 2 ≦ n ≦ (k + 1) M / 2-1) X k, n = S * N−kM / 2−n (N− (k + 1) M / 2 + 1 ≦ n ≦ N−kM / 2) However, when k = 0, (N− (k + 1) M / 2 + 1 ≦ n ≦ N−kM / 2-1. X k, n = 0 (Other) ... (3)

【0051】トレーニングデータ生成器110は、k=
0からトレーニングデータをLシンボルづつk=7まで
出力する。次に、初期値算出回路240の動作を説明す
る。半2重通信時には、対向側の装置は、送信を停止し
ているため、FFT1500により、復調されたY
k,nは、送信データのエコー成分のみである。
The training data generator 110 calculates k =
The training data is output from 0 to k = 7 for each L symbol. Next, the operation of the initial value calculation circuit 240 will be described. At the time of the half-duplex communication, the device on the opposite side has stopped transmitting, so that the YFT demodulated by the FFT 1500 is used.
k and n are only the echo components of the transmission data.

【0052】初期値算出回路240では、その周波数域
のエコー成分をトレーニングデータで割ることによっ
て、エコーパスの周波数特性が導出される。まず、帯域
拡張部241は、FFT1500の出力データYk,n
について、下記式(4)にしたがい、帯域拡張を行う。 Y’k,n=Yk,mod(n,64) …(4)
The initial value calculation circuit 240 derives the frequency characteristics of the echo path by dividing the echo component in the frequency range by the training data. First, the band extending unit 241 outputs the output data Y k, n of the FFT 1500.
, Band expansion is performed according to the following equation (4). Y ′ k, n = Y k, mod (n, 64) (4)

【0053】平均化回路242は、帯域拡張されたY’
k,nをLシンボル間平均し、データY’’k,nを出力
する。トレーニングデータ生成器243は、トレーニン
グデータ生成器110で出力されるデータXk,nと同
じトレーニングデータX’k,nを出力する。複素割り
算器244は、データを配置したキャリアに対して、下
記式(5)を用いて、周波数域エコーキャンセラ210
のタップ係数の初期値Ck,nを算出する。
The averaging circuit 242 calculates the band-extended Y ′
k, n are averaged between L symbols, and data Y ″ k, n is output. The training data generator 243 outputs the same training data X ′ k, n as the data X k, n output by the training data generator 110. The complex divider 244 calculates the frequency domain echo canceller 210 for the carrier on which the data is allocated, using the following equation (5).
Is calculated as an initial value C k, n of the tap coefficient.

【0054】 Ck,n=Y’’k,n/X’k,n …(5) 例えば、k=0の場合、0番目から31番目、481番
目から511番目のキャリアにデータが配置されている
ので、複素割り算器244は、平均化回路242の出力
Y’’0,nとトレーニングデータX’0,nを用い、下
記式(6)にしたがいタップ係数の初期値C0,nを算
出する。
C k, n = Y ″ k, n / X ′ k, n (5) For example, when k = 0, data is allocated to the 0th to 31st and 481st to 511st carriers. Therefore, the complex divider 244 uses the output Y ″ 0, n of the averaging circuit 242 and the training data X ′ 0, n to calculate the initial value C 0, n of the tap coefficient according to the following equation (6). calculate.

【0055】 C0,n=Y’’0,n/X’0,n (n=0〜31) …(6) 複素割り算器244は、求めた周波数域エコーキャンセ
ラ210のタップ係数の値を保存しておく。全てのキャ
リアに対するタップ係数の初期値が求まったら、複素割
り算器244は、周波数域エコーキャンセラ210へタ
ップ係数の初期値を出力する。
C 0, n = Y ″ 0, n / X ′ 0, n (n = 0 to 31) (6) The complex divider 244 calculates the value of the tap coefficient of the frequency-domain echo canceller 210. Save it. When the initial values of the tap coefficients for all the carriers are obtained, the complex divider 244 outputs the initial values of the tap coefficients to the frequency-domain echo canceller 210.

【0056】周波数域エコーキャンセラ210は、初期
値算出回路240によって、タップ係数の初期値が設定
されると、従来のエコーキャンセラと同様に、エコーレ
プリカの特性と、タップ係数の更新を行う。従来のエコ
ーキャンセラでは、前述したように、逆高速フーリエ変
換を含んだトレーニングを繰り返して時間域エコーキャ
ンセラ230のタップ係数を求めていた。
When the initial value of the tap coefficient is set by the initial value calculation circuit 240, the frequency band echo canceller 210 updates the echo replica characteristics and the tap coefficient as in the conventional echo canceller. In the conventional echo canceller, as described above, the training including the inverse fast Fourier transform is repeated to obtain the tap coefficient of the time-domain echo canceller 230.

【0057】これに対して、本発明のエコーキャンセラ
では、初期値算出回路240を設け、周波数域エコーキ
ャンセラ210のタップ係数をすべて算出した後、逆高
速フーリエ変換を行い時間域エコーキャンセラ230の
タップ係数を求めている。このため、逆高速フーリエ変
換を1回行うだけで時間域エコーキャンセラ230のタ
ップ係数を求めることができ、トレーニング時の演算量
を削減することができる。
On the other hand, in the echo canceller of the present invention, the initial value calculation circuit 240 is provided, and after calculating all the tap coefficients of the frequency domain echo canceller 210, the inverse fast Fourier transform is performed and the tap of the time domain echo canceller 230 is performed. Find the coefficient. Therefore, the tap coefficient of the time-domain echo canceller 230 can be obtained only by performing the inverse fast Fourier transform once, and the amount of calculation at the time of training can be reduced.

【0058】次に32×i(i=1、2、…、15)番
目の送信キャリアは、ナイキストの受信キャリアに対応
する。32×i番目のキャリアのエコーは、虚数が0と
なり、実数のみのデータになっているため、正しい周波
数域エコーキャンセラ210のタップ係数を求めること
ができない。
Next, the 32 × i (i = 1, 2,..., 15) -th transmission carrier corresponds to the Nyquist reception carrier. Since the imaginary number of the echo of the 32 × i-th carrier is 0 and the data is only the real number, the tap coefficient of the correct frequency band echo canceller 210 cannot be obtained.

【0059】そのため、正しい時間域エコーキャンセラ
230のタップ係数も求めることができず、エコー抑圧
特性が劣化するので、以下のように、時間域エコーキャ
ンセラ230のタップ係数の補正を行う。図4は本発明
に係るエコーキャンセラのトレーニングシステムを用い
たADSL送受信部の他の構成を示すブロック図であ
る。
For this reason, the correct tap coefficient of the time-domain echo canceller 230 cannot be obtained, and the echo suppression characteristic deteriorates. Therefore, the tap coefficient of the time-domain echo canceller 230 is corrected as follows. FIG. 4 is a block diagram showing another configuration of the ADSL transmitting / receiving unit using the training system of the echo canceller according to the present invention.

【0060】本図に示すように、周波数域−時間域エコ
ーキャンセラ200において、IFFT220と時間域
エコーキャンセラ230の間に時間域補正部250が設
けられる。時間域補正部250は、下記式(7)にした
がい、時間域エコーキャンセラ230のタップ係数の補
正を行う。
As shown in the figure, in the frequency domain-time domain echo canceller 200, a time domain correction section 250 is provided between the IFFT 220 and the time domain echo canceller 230. The time domain correction unit 250 corrects the tap coefficient of the time domain echo canceller 230 according to the following equation (7).

【0061】 C’=C−Amod(n,64) …(7) A=Cm+448 C’:補正後の時間域エコーキャンセラ230のタッ
プ係数 C:補正前の時間域エコーキャンセラ230のタップ
係数 A:補正前の時間域エコーキャンセラ230のタップ
係数に関する後半64タップの係数
C ′n= Cn-Amod (n, 64) … (7) Am= Cm + 448  C 'n: Touch of time-domain echo canceller 230 after correction
Coefficient Cn: Tap of time domain echo canceller 230 before correction
Coefficient Am: Tap of time domain echo canceller 230 before correction
Coefficient of the latter 64 taps regarding the coefficient

【0062】n:時間域エコーキャンセラ230のタッ
プ係数のタップ番号(n=0〜511) m:保存した成分のタップ番号(m=0〜63) mod(n,64):nを64で割った余り 上記式(7)の補正により、実際のエコーパスに、より
近い時間域エコーキャンセラ230のタップ係数を算出
することができ、エコー抑圧量を改善することができ
る。
N: tap number of tap coefficient of time domain echo canceller 230 (n = 0 to 511) m: tap number of stored component (m = 0 to 63) mod (n, 64): n is divided by 64 Remainder By the correction of the above equation (7), the tap coefficient of the time-domain echo canceller 230 closer to the actual echo path can be calculated, and the echo suppression amount can be improved.

【0063】[0063]

【発明の効果】以上説明したように、本発明によれば、
送信帯域を受信帯域幅で分割し、分割した1つの帯域毎
にデータをのせて送信することにより、従来のように複
数の送信キャリアのエコーは重畳されなくなった。これ
により、LMS法のような漸近的にタップ係数を求める
方法ではなく、周波数領域において受信データをトレー
ニングデータで割ることによりエコーキャンセラのタッ
プ係数を算出することができ、トレーニング時間を短縮
することができる。
As described above, according to the present invention,
The transmission band is divided by the reception bandwidth, and data is transmitted for each divided band so that echoes of a plurality of transmission carriers are not superimposed as in the related art. Thus, the tap coefficient of the echo canceller can be calculated by dividing the received data by the training data in the frequency domain, instead of the method of asymptotically obtaining the tap coefficient as in the LMS method, and the training time can be reduced. it can.

【0064】初期値算出回路240を設け、周波数域エ
コーキャンセラ210のタップ係数を全て算出した後、
逆高速フーリエ変換を1回行うだけで時間域エコーキャ
ンセラ230のタップ係数を求めることができ、トレー
ニング時の演算量を削減することができる。
An initial value calculation circuit 240 is provided, and after calculating all tap coefficients of the frequency band echo canceller 210,
The tap coefficient of the time-domain echo canceller 230 can be obtained by performing the inverse fast Fourier transform only once, and the amount of calculation during training can be reduced.

【図面の簡単な説明】[Brief description of the drawings]

【図1】本発明に係るエコーキャンセラのトレーニング
システムを用いたADSL送受信部の構成を示すブロッ
ク図である。
FIG. 1 is a block diagram illustrating a configuration of an ADSL transmission / reception unit using a training system for an echo canceller according to the present invention.

【図2】初期値算出回路240の詳細を説明するブロッ
ク図である。
FIG. 2 is a block diagram illustrating details of an initial value calculation circuit 240.

【図3】送信するトレーニングデータXk,nを説明す
る模式(ここではk=0〜7)を示す図である。
FIG. 3 is a diagram illustrating a model (here, k = 0 to 7) for explaining training data X k, n to be transmitted.

【図4】本発明に係るエコーキャンセラのトレーニング
システムを用いたADSL送受信部の他の構成を示すブ
ロック図である。
FIG. 4 is a block diagram showing another configuration of the ADSL transmission / reception unit using the training system for the echo canceller according to the present invention.

【図5】ADSL方式の使用帯域を示す概略図である。FIG. 5 is a schematic diagram showing a use band of the ADSL system.

【図6】本発明の前提となるエコーキャンセラのトレー
ニングシステムを用いたADSL送受信部の構成を示す
ブロック図である。
FIG. 6 is a block diagram showing a configuration of an ADSL transmission / reception unit using a training system for an echo canceller, which is a premise of the present invention.

【図7】式(13)によるデータ間引きの様子を示す模
式図である。
FIG. 7 is a schematic diagram illustrating a state of data thinning-out according to Expression (13).

【符号の説明】[Explanation of symbols]

100…送信部 110…トレーニングデータ生成器 120…IFFT 130…パラレルシリアル変換部 140…ディジタルアナログ変換部 150…ローパスフィルタ 200…周波数−時間域エコーキャンセラ 210…周波数域エコーキャンセラ 220…IFFT 230…時間域エコーキャンセラ 240…初期値算出回路 241…帯域拡張部 242…平均化回路 243…トレーニングデータ生成器 250…時間域補正部 300…ハイブリッドトランス 400…経路 1000…受信部 1100…ローパスフィルタ 1200…アナログディジタル変換部 1300…減算器 1400…シリアルパラレル変換部 1500…FFT 1600…複素減算器 REFERENCE SIGNS LIST 100 transmission unit 110 training data generator 120 IFFT 130 parallel-to-serial conversion unit 140 digital-to-analog conversion unit 150 low-pass filter 200 frequency-time echo canceller 210 frequency echo canceller 220 IFFT 230 time domain Echo canceller 240 ... Initial value calculation circuit 241 ... Band expansion unit 242 ... Averaging circuit 243 ... Training data generator 250 ... Time domain correction unit 300 ... Hybrid transformer 400 ... Path 1000 ... Receiving unit 1100 ... Low pass filter 1200 ... Analog-to-digital conversion Unit 1300 subtractor 1400 serial-parallel converter 1500 FFT 1600 complex subtractor

───────────────────────────────────────────────────── フロントページの続き Fターム(参考) 5K022 AA07 AA11 AA28 AA42 5K046 AA00 AA01 BA06 BB05 EE06 EF00 EF01 EF07 EF11 EF15 HH01 HH11  ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5K022 AA07 AA11 AA28 AA42 5K046 AA00 AA01 BA06 BB05 EE06 EF00 EF01 EF07 EF11 EF15 HH01 HH11

Claims (6)

【特許請求の範囲】[Claims] 【請求項1】 下りの送信帯域が上りの受信帯域よりも
広いマルチキャリア伝送のエコーキャンセラのトレーニ
ングシステムにおいて、 データが送信キャリアに配置された信号に基づいて、回
線エコーのキャンセル信号を形成して受信データに減算
処理するためのエコーキャンセラと、 送信帯域を受信帯域の幅に分割し、分割された1つの送
信帯域ごとの送信キャリアにトレーニングデータが配置
される信号を前記エコーキャンセラに出力して前記エコ
ーキャンセラのトレーニングを行うためのトレーニング
データ生成部とを備えることを特徴とするエコーキャン
セラのトレーニングシステム。
In a training system for an echo canceller for multicarrier transmission in which a downstream transmission band is wider than an upstream reception band, data is used to form a line echo cancel signal based on a signal arranged on a transmission carrier. An echo canceller for performing subtraction processing on the reception data; and a transmission band divided into reception band widths, and a signal in which training data is arranged on a transmission carrier for each divided transmission band is output to the echo canceller. A training system for an echo canceller, comprising: a training data generator for training the echo canceller.
【請求項2】 前記マルチキャリア伝送の方式はADS
L方式であることを特徴とする、請求項1に記載のエコ
ーキャンセラのトレーニングシステム。
2. The method of multi-carrier transmission is ADS.
The echo canceller training system according to claim 1, wherein the training system is an L system.
【請求項3】 前記エコーキャンセラは、前記トレーニ
ング生成器から周波数域の前記信号を入力し各キャリア
毎に用意される複素数の係数をもつ1タップのフィルタ
で構成される周波数域エコーキャンセラと、前記トレー
ニング生成器から時間域の前記信号を入力しFIRフィ
ルタで構成される時間域エコーキャンセラとからなり、
前記時間域エコーキャンセラのタップ係数は前記周波数
域エコーキャンセラで求められたタップ係数を逆高速フ
ーリエ変換して算出されることを特徴とする、請求項1
に記載のエコーキャンセラのトレーニングシステム。
3. The echo canceller according to claim 1, wherein the echo canceller receives the signal in the frequency domain from the training generator, and includes a 1-tap filter having a complex coefficient prepared for each carrier. A time-domain echo canceller configured to receive the time-domain signal from the training generator and configured by an FIR filter,
The tap coefficient of the time-domain echo canceller is calculated by performing an inverse fast Fourier transform on the tap coefficient obtained by the frequency-domain echo canceller.
The training system for an echo canceller described in 1.
【請求項4】 前記エコーキャンセラに初期値算出回路
が設けられ、前記初期値算出回路は、前記周波数域エコ
ーキャンセラのタップ係数の初期値を算出することを特
徴とする、請求項3に記載のエコーキャンセラのトレー
ニングシステム。
4. The echo canceller according to claim 3, wherein an initial value calculating circuit is provided in the echo canceller, and the initial value calculating circuit calculates an initial value of a tap coefficient of the frequency domain echo canceller. Echo canceller training system.
【請求項5】 前記周波数域エコーキャンセラと前記時
間域エコーキャンセラの間に時間域補正部を設け、前記
時間域補正部は、分割された送信帯の送信キャリアがナ
イキストの受信キャリアに対応することに起因してタッ
プ係数の補正を行うことを特徴とする、請求項3に記載
のエコーキャンセラのトレーニングシステム。
5. A time domain correction unit is provided between the frequency domain echo canceller and the time domain echo canceller, wherein the time domain correction unit corresponds to a Nyquist reception carrier in a transmission carrier of a divided transmission band. 4. The echo canceller training system according to claim 3, wherein the tap coefficient is corrected due to the following.
【請求項6】 下りの送信帯域が上りの受信帯域よりも
広いマルチキャリア伝送のエコーキャンセラのトレーニ
ング方法において、 データが送信キャリアに配置された信号に基づいて、回
線エコーのキャンセル信号を形成して受信データに減算
処理する工程と、 送信帯域を受信帯域の幅に分割し、分割された1つの送
信帯域ごとの送信キャリアにトレーニングデータが配置
される信号を前記エコーキャンセラに出力して前記エコ
ーキャンセラのトレーニングを行う工程とを備えること
を特徴とするエコーキャンセラのトレーニング方法。
6. A training method of an echo canceller for multicarrier transmission in which a downstream transmission band is wider than an upstream reception band, wherein a data echo cancellation signal is formed based on a signal arranged on a transmission carrier. Subtracting the reception data from the received data; dividing the transmission band into the width of the reception band; outputting a signal in which training data is allocated to transmission carriers for each of the divided transmission bands to the echo canceller; Training the echo canceller.
JP27783699A 1999-09-30 1999-09-30 Echo canceller training system and method Expired - Fee Related JP3395894B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP27783699A JP3395894B2 (en) 1999-09-30 1999-09-30 Echo canceller training system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP27783699A JP3395894B2 (en) 1999-09-30 1999-09-30 Echo canceller training system and method

Publications (2)

Publication Number Publication Date
JP2001102972A true JP2001102972A (en) 2001-04-13
JP3395894B2 JP3395894B2 (en) 2003-04-14

Family

ID=17588953

Family Applications (1)

Application Number Title Priority Date Filing Date
JP27783699A Expired - Fee Related JP3395894B2 (en) 1999-09-30 1999-09-30 Echo canceller training system and method

Country Status (1)

Country Link
JP (1) JP3395894B2 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100421957B1 (en) * 2001-10-30 2004-03-11 엘지전자 주식회사 Method for acquirement of echo impulse response of echo canceller in DMT
KR100443009B1 (en) * 2001-10-05 2004-08-04 엘지전자 주식회사 Apparatus and method for time domain echo canceller of DMT
WO2005027386A1 (en) * 2003-09-10 2005-03-24 Japan Industrial Technology Association Digital data transmission device
JP2007096469A (en) * 2005-09-27 2007-04-12 Matsushita Electric Ind Co Ltd Multicarrier communication apparatus, multicarrier communication, system, and multicarrier communication method
CN109961797A (en) * 2017-12-25 2019-07-02 阿里巴巴集团控股有限公司 A kind of echo cancel method, device and electronic equipment

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100443009B1 (en) * 2001-10-05 2004-08-04 엘지전자 주식회사 Apparatus and method for time domain echo canceller of DMT
KR100421957B1 (en) * 2001-10-30 2004-03-11 엘지전자 주식회사 Method for acquirement of echo impulse response of echo canceller in DMT
WO2005027386A1 (en) * 2003-09-10 2005-03-24 Japan Industrial Technology Association Digital data transmission device
JP2007096469A (en) * 2005-09-27 2007-04-12 Matsushita Electric Ind Co Ltd Multicarrier communication apparatus, multicarrier communication, system, and multicarrier communication method
JP4749815B2 (en) * 2005-09-27 2011-08-17 パナソニック株式会社 Multi-carrier communication apparatus, multi-carrier communication system, and multi-carrier communication method
CN109961797A (en) * 2017-12-25 2019-07-02 阿里巴巴集团控股有限公司 A kind of echo cancel method, device and electronic equipment

Also Published As

Publication number Publication date
JP3395894B2 (en) 2003-04-14

Similar Documents

Publication Publication Date Title
US7801233B1 (en) Method and system for channel equalization and crosstalk estimation in a multicarrier data transmission system
US6674810B1 (en) Method and apparatus for reducing peak-to-average power ratio in a discrete multi-tone signal
US5623513A (en) Mitigating clipping and quantization effects in digital transmission systems
US7577084B2 (en) ISDN crosstalk cancellation in a DSL system
US20020067824A1 (en) Two-step algorithm for training an echo cancellation filter
JP2007195218A (en) Digital radio frequency interference canceller
WO2010042350A1 (en) Adaptive frequency-domain reference noise canceller for multicarrier communications systems
JPH04354212A (en) Method and device for denying echo of multiple carrier wave signal
US7672447B1 (en) Frequency domain echo canceller
US20100124289A1 (en) Transmission suppression
WO1999043123A1 (en) Improvements in, or relating to, vdsl
JP3717363B2 (en) xDSL transceiver
JP3672825B2 (en) DMT nonlinear echo compensation
US8064501B2 (en) Method and apparatus for generating a periodic training signal
JP2001102972A (en) System and method for training echo canceler
TW558890B (en) Echo channel estimation mechanism applied in ADSL
US7852869B2 (en) Method and apparatus for the transmission of data
US7899127B2 (en) Method and device for reducing the crest factor of a signal
US7697619B2 (en) Training sequence for channel estimation in a data transmission system
US6683859B1 (en) Method and apparatus for echo cancellation updates in a multicarrier transceiver system
US20010036225A1 (en) Coefficient update method and receive method of time domain equalizer of DMT system, DMT system and DMT modem
US6810076B1 (en) Efficient echo canceller for ADSL applications
JP2001523914A (en) Improvements in or related to subscriber line transmission systems
US20050069045A1 (en) Method and apparatus for reducing a crest factor of a multi-tone data signal
CN113904906A (en) Method for realizing frequency domain nonlinear continuous interference suppression

Legal Events

Date Code Title Description
LAPS Cancellation because of no payment of annual fees