JP4371698B2 - Synchronize initial handshake in ADSL Annex C to TTR - Google Patents

Description

具体例において、上記x(n)は時間領域のＲＥＶＥＲＢ３３〜６４信号であり、y(n)はx(n)を送信機から送信し、チャンネルh(n)に通し、受信機で受信した信号である。したがって、y(n)はx(n)とh(n)の畳み込みであり、次式で定義される。
【００２７】
【数２】

式（２）のy(n)を式（１）に代入すると、次式が得られる。
【００２８】
r_xy(l) = h(l) l = 0,±1,±2,... （３）
式（３）はx(n)が広帯域のＰＲＢＳ（擬似ランダム２進シーケンス）であるという事実に基づいている。この種の信号の例として、G.992.1ドキュメントに記載されるＲＥＶＥＲＢタイプの信号を考慮する。さらに、x(n)の自己相関はデルタ関数で近似される。図３は本発明の１つの実施の形態に基づき、相関信号の例を示したものである。この相関信号は、４つの連続するＲＥＶＥＲＢ３３〜６４信号がＤＭＴ受信機の受信信号に存在することを表している。５kmの紙パルプループと受信機に付加される-１３０dBm/HzのＡＷＧＮを想定している。図示のように、ＤＭＴシンボル０〜３にＲＥＶＥＲＢ３３〜６４信号が存在することを示す、４つの明白なピークがある。
デシメーション
ナイキスト標本化定理によれば、最大帯域幅f_mＨｚのローパスアナログ信号は、サンプリングレートf₀をf_mの２倍以上とすると、サンプルによって表現され、サンプルから完全に復元することができる。最大帯域幅f_mＨｚのアナログ信号を２f_m以上、例えば、４f_mのサンプリングレートでサンプリングした信号は、情報損失なしに２でデシメーション処理する（１つ置きにサンプルを捨てる）ことができる。オーバーサンプリングしたアナログ信号が使用帯域を超えた不必要な信号を含む場合、デシメーション後に不必要な信号が使用帯域内に折り返すのを避けるため、デシメーションに先立ちフィルタリングが必要である。この処理はデシメーションフィルタリングと呼ばれる。
【００２９】
特殊定義ＴＴＲ表示信号がシンボル０〜３の間送信されるＲＥＶＥＲＢ３３〜６４である場合において、検出器モジュール２２５ｂは、シンボル０、１、２、３を受信したことを示す、４つの連続するＲＥＶＥＲＢ３３〜６４信号を検出するまで、受信サンプルに対する線形相互相関処理を実行し続ける。ＤＭＴシンボルの受信毎に、相互相関信号が計算される。検出器モジュール２２５ｂで解析するために、受信サンプルを（検出器モジュールに結合されるか、その内部に置かれる、ＲＡＭ、その他の記憶装置に）バッファリングし、又はその他の方法で保持する。当初、受信信号にＲＥＶＥＲＢ３３〜６４信号が現れた時点では、相互相関のピーク位置は、その受信シンボルについて計算した相関信号のサンプル内部における任意の位置であり得る。この理由は、受信機のフレーム境界の位置がまだ送信機のフレーム境界と合っていないからである。ここに、相互相関信号のピーク位置は、ハイパーフレームの位置合わせを可能にする、送信機と受信機のフレーム境界のオフセットを示すものである。この意味で、シンボル０〜３はハイパーフレーム合わせの、予め決められた証印を有する。
【００３０】
ここで、検出器モジュール２２５ｂは相互相関信号のピーク位置に基づき、フレーム境界の位置合わせを実行し、再度相関処理を行って次のハイパーフレームのＮ個のＲＥＶＥＲＢ３３〜６４信号を探索する。最初の４つの連続するＲＥＶＥＲＢ３３〜６４信号の検出に基づきフレーム境界の位置合わせが適正になされたら、相互相関のピークが相互相関信号の既知のサンプル（例えば、最初のサンプル）の位置で生じるとして（次のハイパーフレームにおいて）次の４つのＲＥＶＥＲＢ３３〜６４信号を検出する。このようにして、送信機と受信機のフレーム境界間の既知のオフセットが決まる。特に、相互相関信号の既知のサンプルにピーク位置をもつ、４番目に検出されるＲＥＶＥＲＢ３３〜６４信号はハイパーフレーム内のＤＭＴシンボル３の境界を表す。このようにして受信機はハイパーフレーム合わせを達成する。
【００３１】
ここに、現在の標準によるトーン４８（慣用のＴＴＲ表示信号）について、検出器モジュールの機能に影響を与えることなく、その位相トグル動作を従来と同様に実行することができる。さらに、トーン６４（慣用のパイロット信号）がＲＥＶＥＲＢｘ〜ｙ信号に含まれている。したがって、本発明は後方互換性を有する。
実現の詳細
実施形態において、プロセッサモジュール２２５は、本発明に基づき、ハイパーフレーム合わせ又は初期ハンドシェイク同期化を実行するようにプログラムされた、デジタル信号プロセッサ（ＤＳＰ）又はその他の適当な処理環境である。代替的な実施形態において、プロセッサモジュール２２５はＡＳＩＣやチップセット等の専用シリコンで実現される。プロセッサモジュール２２５を実現するためにＤＳＰとＡＳＩＣ技術の組合せを使用してもよい。したがって、本発明は、例えばハイパーフレーム合わせを実行する方法としてＡＤＳＬサービスプロバイダーが使用することができる。代わりに、本発明はモデムやプログラム可能なＤＳＰ又は専用チップ／チップセット等の装置に組み込むことができる。代わりに、本発明は、一方又は両方の送受信機が、後述する、ハイパーフレーム合わせ又は初期ハンドシェイク同期化を実行するように構成される送受信機対のようなシステムに組み込むことができる。代替として、本発明は、サーバやディスク等のコンピュータ読取可能な媒体に具体化したコンピュータプログラム製品等として組み込むことができる。
【００３２】
また、図示しないモジュールについても用途に応じてモデムに組み込むことができる。例えば、モデムは重複スペクトル用途で動作し、エコーキャンセラを備えてよい。他のモジュール（例えば、送信機と受信機のクロック）や構成については本開示から明らかである。さらに、プロセッサモジュール２２５はリンク初期化手続きの一環としてハイパーフレーム合わせ処理を自ら起動することができる。代わりに、イベント（例えば、位置合わせの検出失敗）やホストからの制御信号（例えば、ハイパーフレーム合わせソフトウェアの呼出又は論理イネーブルライン）により位置合わせ処理を起動することができる。同様に、リモートモデムによる、ハイパーフレーム合わせの起動を検出した後、プロセッサモジュール２２５はハイパーフレーム合わせ処理を開始することができる。代表的には、ＡＴＵ−Ｃモデムがハイパーフレーム合わせ処理を起動する。しかしながら、所望であれば、ＡＴＵ−Ｒがこの処理を起動するように構成することができる。
方法論
図４に、本発明の実施の形態に基づきハイパーフレーム合わせ処理を実行する方法のフローチャートを示す。この方法は、一般に伝送媒体を介して通信可能に結合される２つのモデムを備える通信システム（例えば、ＡＤＳＬアネックスＣ）において実施される。特に、この方法はシステムの特定の通信方向に関連する送信ノードにおいて実行される。送信（例えばローカル）と受信（例えばリモート）の両ノードのモデムは、例えば、図２で説明したように構成される。しかしながら、その他の構成も本開示から明らかである。
【００３３】
この方法はハイパーフレーム合わせ手続きを起動すること（ステップ４０５）から始まる。実施形態において、ハイパーフレーム合わせ手続きは通信リンクの初期化（例えばトレーニング）の期間中にＡＴＵ−Ｃにより起動される。しかしながら、代替的実施形態では、その他の時間（例えばサービスの中断後）でも位置合わせ手続きを起動することができる。さらに、方法は特別定義信号を発生し（ステップ４１０）、リモートノードにこの特殊定義信号をデータフレームにおける１以上の既知の区間中に送信する（ステップ４１５）。
【００３４】
実施形態において、データフレームはＤＭＴシンボル０〜３４４を有する、ＡＤＳＬダウンストリームのハイパーフレームであり、データフレームにおける１以上の既知の区間はハイパーフレームにおけるＤＭＴシンボル０〜３である。このデータフレームの最初の送信により送信機と受信機のシンボル境界の相対位置を識別することが上述したように可能になる。２回目（又はそれ以降）のフレーム送信によりハイパーフレーム合わせの確認が可能になる。代わりに、データフレームにおける１以上の既知の区間はハイパーフレームにおけるシンボルＮ₁〜Ｎ₂（例えば１４０〜１４４）及びハイパーフレームにおけるシンボルＮ₃〜Ｎ₄（例えば２３７〜２４１）である。このような実施形態では、シンボルＮ₁〜Ｎ₂の期間中に送信した特殊定義信号により送信機と受信機のシンボル境界のオフセットを識別することが可能となり、シンボルＮ₃〜Ｎ₄の間に送信した特殊定義信号によりハイパーフレーム合わせの確認が可能になる。
【００３５】
例えば、特殊定義信号は、疑似ランダムデータシーケンス又はその他の固有に識別可能な信号である。実施形態において、特殊定義信号は上述したＲＥＶＥＲＢタイプの信号に含まれるトーンのサブセット（例えば、ＲＥＶＥＲＢ６〜３１、ＲＥＶＥＲＢ１６〜３１、ＲＥＶＥＲＢ３３〜６４又はＲＥＶＥＲＢｘ〜ｙ（ここにｘ〜ｙは長いループの情報伝送で使用可能なトーンのセットを表す））である。このような特殊定義信号をデータフレームの既知の区間で送信することにより、長いループのリモート端で受信できる強固なＴＴＲ表示信号を提供する。
【００３６】
さらに、この方法はハイパーフレーム合わせの確認のため、リモートノードからの応答）を受信し（ステップ４２０）、その応答に基づき位置合わせが達成されたか判定する（ステップ４２５）。応答がハイパーフレーム位置の合っていることを確認している場合、手続きを終了する。ついで通信システムは他の初期化手続きを完了させ、データモード（例えばショウタイム）に移行する。しかし、応答がハイパーフレーム位置の不一致を確認している場合、ステップ４１５から４２０を繰り返す。上述したように、何回かの試み（例えば、Ｎ個のハイパーフレームの送信）の後、肯定応答を受信しなかった場合、送信側の送受信機はエラーとして認識するように構成される。
【００３７】
図５に、本発明のもう１つの実施の形態に基づきハイパーフレーム合わせ処理を実行する方法のフローチャートを示す。特に、この方法はシステムの特定の通信方向に関連する受信ノードにおいて実行される。
【００３８】
この方法はハイパーフレーム合わせ手続きを（例えば通信リンクの初期化におけるトレーニングの期間中にＡＴＵ−Ｃにより）起動すること（ステップ５０５）から始まる。さらに、方法はリモートノード送信機からデータフレームのデータを受信し（ステップ５１０）、データフレームにおける１以上の既知の区間で特殊定義信号を検出する（ステップ５１５）。特殊定義信号と既知のシンボルに関する上記説明はここでも適用される。方法はステップ５２０に進み、送信機と受信機の区間境界間のオフセットを識別し、ステップ５２５でハイパーフレーム合わせが達成されたか判定する。これらのステップを実施する特定の実施形態について以下説明する。
【００３９】
説明の便宜上、データフレームにおける既知の区間で送信される特殊定義信号は、ハイパーフレームにおけるシンボル０〜３の間送信されるＲＥＶＥＲＢｘ〜ｙ信号であるとする。実施形態において、送信機と受信機の区間境界間のオフセットは、４つの連続するＲＥＶＥＲＢｘ〜ｙ信号が検出される（これによりシンボル０〜３が示される）まで受信機のサンプルに対し線形相互相関を実行し続けることにより識別される。ここに、送信機と受信機のフレーム境界のオフセットはＲＥＶＥＲＶｘ〜ｙ信号の相互相関信号ピークに基づいて識別される。線形相互相関処理は後続のハイパーフレーム（例えば次のハイパーフレーム）で繰り返され、次の４つの連続するＲＥＶＥＲＢｘ〜ｙ信号が検出される。相互相関のピークが相互相関信号の既知のサンプル（例えば最初のサンプル）に位置する状態で次の４つの連続するＲＥＶＥＲＢｘ〜ｙ信号を検出することによりハイパーフレーム合わせが達成される。
【００４０】
もう１つの実施形態において、送信機と受信機の区間境界間のオフセットは、ＤＭＴシンボルの受信毎に相互相関信号を計算し、相互相関信号中に予め決められた、ハイパーフレーム合わせの証印を有する、Ｎ個の連続するＤＭＴシンボルを識別することによって識別される。この予め決まられた証印は、例えば、疑似ランダムデータシーケンスにトーンのセット（ｘ〜ｙ）（例えば、ＲＥＶＥＲＢ信号のトーン３３〜６４）を含む。フレーム境界位置合わせはこの証印を有する相互相関信号のピーク位置に基づいて行われる。２回目の相互相関処理を実行してＮ個の連続するＤＭＴシンボルの第２のセットを識別する。予め決められた証印の２度目の検出によりフレーム境界位置合わせを確認して、ハイパーフレーム合わせが達成されたことを示す。なお、２回目の相互相関処理は同一のハイパーフレーム又は後続のハイパーフレームのいずれで実行してもよい。さらに、プロセスを３回、４回、５回等の回数繰り返し実行して適切な位置合わせを確保するようにしてもよい。
【００４１】
ハイパーフレーム合わせが達成されない場合、方法はリモートノードにハイパーフレーム合わせが達成されていないことを示す否定の応答を送信する（ステップ５３０）。このような応答は例えば、特別に定義した信号あるいは単なる無音の期間である。ここで、受信状態にある送受信機は、複数回の試み（例えば複数回のハイパーフレームの送信）の後、肯定応答が受信されない場合、エラーとして認知するように構成することができる。一方、ハイパーフレーム合わせを達成した場合、方法はリモートノードにハイパーフレーム位置の一致を確認する応答を送信する（ステップ５３５）。ついで、通信システムはその他の初期化手続き（例えばチャンネル解析や交換）を完了し、データモード（例えばショータイム）に移行する。
初期化プロセスの全体概要
図６に、本発明の１つの実施の形態に基づく、初期化プロセスの概要を表すタイム図を示す。初期化プロセスは、撚り銅線対又はその他の伝送媒体を介して通信可能に結合される１以上のＡＤＳＬ通信のペアを備える通信システムにおいて実行されるものと想定する。
【００４２】
各送受信機について示す初期化プロセスは、初期ハンドシェイク手続きの後、送受信機のトレーニングとチャンネル解析の手続きが後続する構成である。その後、トレーニングと解析の手続きで習得した情報が交換される。ハンドシェイク手続きにより、一般に、新たに結合された送受信機は機能（capabilities）を交換し、共通の動作モードを選択することが可能になる。この初期ハンドシェイク手続きは、代表的には、ＩＴＵ−Ｔ勧告G.994.1（G.hs）又はその他の適当なハンドシェイク手続きに基づき実行される。
【００４３】
しかしながら、G.994.1は、上述したＴＴＲに基づく時分割二重方式を適用する、ＴＣＭ−ＩＳＤＮには同期しない。したがって、G.994.1信号はＴＣＭ−ＩＳＤＮ又はアネックスＣのＦＢＭのＮＥＸＴと、特に長距離ループ（例えば５km以上）で干渉する。その結果ハンドシェイク手続きは失敗に帰すおそれがある。したがって、G.992のアネックスＣモデムのためにG.994.1の信頼性を高め強力なものにするニーズがある。
【００４４】
本発明の１つの実施の形態によりG.994.1をＴＣＭ−ＩＳＤＮに同期させることが可能となり、同期化はハンドシェイク手続きの前部期間において達成される。これによりG.994.1信号を長距離ループ上で例えばＦＥＸＴタイムの期間で交換することができ、ＮＥＸＴタイムの期間での送信を回避することができる。したがってより強力で信頼性の高いハンドシェイク手続きが実現される。
アネックスＣ G.994.1 のＴＴＲへの同期化
図７は本発明の１つの実施の形態に基づく、複式の初期化手続きのフローチャートである。特に、G.994.1のハンドシェイク手続きをＴＴＲに同期させる。いったんハンドシェイク手続きがＴＴＲに同期すると、ＮＥＸＴとＦＥＸＴの期間は既知になる。これにより意図的にG.hs信号（所望であればトレーニング信号等の他の初期化信号も可能である）を他方のモデムのＦＥＸＴ期間中に送信して信頼の置ける交換を確保することができる。
【００４５】
ハンドシェイク手続きの間、通信モデムはそれぞれＨＳＴＵ−Ｃ、ＨＳＴＵ−Ｒと呼ばれる。既存のＨＳＴＵ−Ｒの起動信号（Ｒ−ＴＯＮＥＳ−ＲＥＱ）は１６ｍｓ毎に位相反転し、ＴＴＲは５×１６ｍｓである８０ｍｓの周期を有する。本発明の１つの実施の形態に基づき、ＨＳＴＵ−ＣのＴＴＲ表示信号は、位相反転がＴＴＲに同期してなくて１６ｍｓという短い周期毎の正確な受信を妨げるような信号である、Ｒ−ＴＯＮＥＳ−ＲＥＱを送信しているＨＳＴＵ−Ｒにより確実に受信されるように定められる。
ハンドシェイク手続きのＨＳＴＵ−Ｃ起動方式
図７を参照すると、ＨＳＴＵ−ＣはＣ−ＳＩＬＥＮＴ１からの移行により開始し、ＴＴＲハイパーフレームの開始時にＣ−ＩＮＩＴと呼ばれる特別に定義されたハンドシェイク信号を送信する。例えば、Ｃ−ＩＮＩＴはＲＥＶＥＲＢ１６〜３１又はその他の上述したような特殊定義信号である。実施形態において、Ｃ−ＩＮＩＴはＤＭＴのＦＥＸＴ_Rシンボル０〜３、３３〜３６、１０８〜１１１及び１４０〜１４４の間、送信される。この特殊定義信号によりＴＴＲとの同期が可能となり、これにより同期化G.hs手続きが可能となり、その手続きのなかでハンドシェイク信号は図７に示すようにＦＥＸＴシンボルの期間のみ送信される。所望であれば、後続する初期化手続きであるトレーニングやチャンネル解析や交換においてＴＴＲとの同期を保持してよい。代わりに、後続する初期化手続きのなかで初期化を再度確立するようにしてもよい。
【００４６】
ここに、このＤＭＴシンボルのシーケンスは、ＴＴＲ期間の任意の位相において１６ｍｓ間隔の位相反転が現れる状況下で、このシーケンスが堅牢さを有することから選択された。上述したように、その他のデータパターン及びシンボル組合せを本発明の原理に基づきここで使用して信頼性の高い送信と受信を達成することもできる。
【００４７】
Ｃ−ＩＮＩＴを検出すると（例えば上述したＴＴＲ検出器）、ＨＳＴＵ−ＲはそのＰＧＡをセットし、シンボル位置合わせとタイミングの回復を実行する。さらに、ＨＳＴＵ−Ｒは、シンボル６〜９から始まり、シンボル１６〜１９、２７〜３０、・・・２１０〜２１４、・・・３４０〜３４３と続く（図１ｂに示す）、ＦＥＸＴ_Cシンボルの間、Ｒ−ＴＯＮＥ１_Fを送信する。実施形態において、ＨＳＴＵ−ＲがＣ−ＴＯＮＥＳ（慣用ハンドシェイク信号）を検出したら、ＨＳＴＵ−Ｒは適切に応答する前に、タイムアウトＴ１を待機してＨＳＴＵ−Ｃが本発明の原理に基づくG.hs手続きをサポートするかを決定する。
【００４８】
Ｒ−ＴＯＮＥ１_Fを検出した後、ＨＳＴＵ−Ｃは、ＦＥＸＴ_Rシンボルの間のみ、ＴＴＲハイパーフレームの開始時に始まるＣ−ＧＡＬＦ１_Fを送信する。各ビットはサブフレーム（０〜３、１１〜１４、２２〜２５、３３〜３６、・・・１４０〜１４３、・・・３３５〜３３８）の最初の４つのＦＥＸＴ_Rシンボルで送信される。３２ビットが８０ｍｓのハイパーフレーム内で送信される。上述したようにG.992ハイパーフレームは３４５個のＤＭＴシンボルから成る。ハイパーフレームには３２個のサブフレームがあり、各サブフレームはＴＴＲタイミングに依存する、１０〜１１個の連続するＤＭＴシンボルである。
【００４９】
Ｃ−ＧＡＬＦ１_Fの受信後、ＨＳＴＵ−Ｒはシンボル６〜９から始まる（Ｒ−ＴＯＮＥ１_Fの場合と同様）、ＦＥＸＴ_Cシンボルの間Ｒ−ＦＬＡＧ１_Fを送信する。Ｒ−ＦＬＡＧ１_Fを受信すると、ＨＳＴＵ−Ｃは、任意のＴＴＲサブフレームから始まる、サブフレームにおける最初の４つのＦＥＸＴ_Rシンボルの間、Ｃ−ＦＬＡＧ１_Fを送信する。ＨＳＴＵ−ＲでＣ−ＦＬＡＧ１_Fを受信すると、ＨＳＴＵ−Ｒは任意のＴＴＲサブフレームから始まる、最初のハンドシェイクトランザクションを開始することができる。
ハンドシェイク手続きのＨＳＴＵ−Ｒ起動方式
ＨＳＴＵ−Ｒが最初からＴＴＲにロックされるのでＨＳＴＵ−Ｃがハンドシェイク手続きを起動（開始）することは有益である。これによりＮＥＸＴ期間中に信号は送信されない。しかしながら、Ｒ−ＳＩＬＥＮＴ０から移行してＲ−ＩＮＩＴを送信することによりＨＳＴＵ−Ｒがハンドシェイク手続きを起動することができる。実施形態において、Ｒ−ＩＮＩＴはＲＥＶＥＲＢ６〜１５であるが、上述した信号やその他の特別定義ＴＴＲ表示信号をここで使用することができる。Ｒ−ＩＮＴの受信後、ＨＳＴＵ−Ｃは、ＨＳＴＵ−Ｃ起動方式で述べたのと同様にして、ＴＴＲハイパーフレームの開始時に始まるＣ−ＩＮＩＴを送信する。
【００５０】
ここに、Ｒ−ＩＮＩＴは、ＴＴＲにロックされるＨＳＴＵ−Ｃから信号をＨＳＴＵ−Ｒが受信する前に送信される。隣接する対のモデムに影響を及ぼすＮＥＸＴを回避するために、ＨＳＴＵ−Ｒはノイズパターンを検出し、高ノイズレベル（ＦＥＸＴ_C期間）のときのみＲ−ＩＮＩＴを送信するように構成することができる。
フォールバック手続き／後方互換性
上述したように、本発明は慣用の初期化手続きとの間で後方互換性を有する。例えば、ＨＳＴＵ−ＣがＣ−ＩＮＩＴを送ってハンドシェイク手続きを起動したがＨＳＴＵ−Ｒが応答しない場合を想定してみる。応答のない理由は、例えば、ＨＳＴＵ−Ｒが接続されていない、それが本発明の原理に基づく初期化手続きをサポートしていない、あるいはそれが与えられたループについて慣用のG.994.1初期化手続きの方がベターであると決定した、等である。
【００５１】
タイムアウト後、フォールバック手続きに入ることができる。例えば、ＨＳＴＵ−ＣはＣ−ＴＯＮＥＳを送るように構成される。ここでＨＳＴＵ−ＲがＲ−ＴＯＮＥ１で応答するならば、送受信機対は慣用のG.994.1初期化手続きにフォールバックする。しかしながら、ＨＳＴＵ−ＲがＣ−ＴＯＮＥＳに応答しない場合、ＨＳＴＵ−Ｃはタイムアウト後、再びＣ−ＩＮＩＴを送るように構成される。この交互動作はＨＳＴＵ−Ｒが肯定的に応答するか、あるいは長いタイムアウトが経過するまで何回か続けられる。このような交互動作により、既存のG.994.1装置における相互動作性の問題を少なくすることができる。
【００５２】
交互動作の長さは予め又はその他の方法で定められ、これによりＨＳＴＵ−Ｒは、ＨＳＴＵ−Ｃが本発明の原理に基づいて構成されているかどうかの決定を下すため、どのくらい長く受信信号をモニターすべきか知っている。これにしたがってＨＳＴＵ−Ｒは応答する。なお、本発明の原理に基づくＴＴＲ表示信号を既存のG.994.1起動信号と並列して送ることもできる。
【００５３】
次に、ＨＳＴＵ−ＲがＲ−ＩＮＩＴを送ってハンドシェイク手続きを起動したが（何らかの理由で）ＨＳＴＵ−Ｃが応答しない場合を想定してみる。Ｔ２のタイムアウト後、ＨＳＳＴＵ−ＲはＲ−ＴＯＮＥＳ−ＲＥＱ（慣用のハンドシェイク信号）を送るように構成される。ＨＳＴＵ−ＣがＣ−ＴＯＮＥＳで応答したならば、送受信機対は慣用のG.994.1及び初期化手続きにフォールバックする。しかしながら、ＨＳＴＵ−ＣがＲ−ＩＮＩＴに対しＣ−ＩＮＩＴで応答し、ＨＳＴＵ−Ｒがその後（例えば、ＨＳＴＵ−Ｒがループはさほど長くなく慣用のG.994.1が好ましいと決定したために）慣用のG.994.1にフォールバックすると決定した場合、ＨＳＴＵ−Ｒは無音を返すことができる。これによりＨＳＴＵ−Ｃはタイムアウトし、慣用のG.994.1に交互に後退して、Ｃ−ＴＯＮＥＳを送る。
フラグとメッセージ
慣用のG.994.1は８つの連続するＤＭＴシンボル内で同じ情報ビットを繰り返す必要がある。（図１aに示すように）各ＴＴＲサブフレームには、代表的に４〜５個のＦＥＸＴ_Rシンボルと６〜７個のＮＥＸＴ_Rシンボルがある。本発明の１つの実施の形態において、１サブフレーム内の最初の４個のＦＥＸＴ_Rを用いて１ビットを送信する。したがって各情報ビットは４個の連続するＦＥＸＴ_Rシンボル内で送信される。もう一つの実施の形態ではサブフレームにおける全てのＦＥＸＴシンボルを用いて１ビットを送信する。したがって各情報ビットは４個又は５個の連続するＦＥＸＴ_Rシンボル内で送信される。このような構成は３４５個のシンボルを用いて３２ビットの情報を送信し、G.hsビット当たり平均１０．７８個のＤＭＴシンボルとなる。モデムは、ＡＤＳＬアネックスＣのＦＢＭモードと同様にＦＥＸＴシンボルのみを送信する。エコーが存在しないので受信機のアルゴリズムは簡単になる。
【００５４】
本発明の代替的実施の形態において、ＦＥＸＴ又はＮＥＸＴシンボルに関わらず、１ＴＴＲ周期内の全シンボルを使用して１ビット情報を送信する。受信ノードを、一方はＦＥＸＴシンボル用、他方はＮＥＸＴシンボル用の、２つの受信機で構成する。２つの受信機の検出結果を比較し、（例えばＣＲＣ検査に基づき）信頼性の高い方を使用する。ＴＣＭ−ＩＳＤＮのノイズ下で、ＦＥＸＴ検出器はより信頼性のある結果を出す。しかしながら、他のノイズ状況下では、信頼性を良くするのにＮＥＸＴ検出器が役立つ可能性がある。このような環境は受信機の構成を複雑にし、半二重モードの利点を失うことになる（例えば、エコーキャンセレーションが必要になる）。
【００５５】
なお、上記ハンドシェイク信号（例えば、Ｃ−ＩＮＩＴ、Ｒ−ＴＯＮＥ１_F、Ｒ−ＩＮＩＴ）とハンドシェイク用のメッセージ長は、８０ｍｓの長さであるＴＴＲハイパーフレーム又は２．５ｍｓの長さであるサブフレームの倍数である。G.hsからアネックスＣ動作モードへの移行時におけるＴＴＲとの同期は保持、又は保持されない。アネックスＣは、ＴＴＲと同期しない既存のG.994.1手続きとともに動作する必要があるので、いずれにせよアネックスＣ自身によるＴＴＲとの同期を実行する必要がある。
【００５６】
初期同期化の後、ＨＳＴＵ−ＣとＨＳＴＵ−ＲはＴＴＲとの同期を維持する必要がある。ＴＴＲクロックはＨＳＴＵ−Ｃで利用できるので、ＨＳＴＵ−Ｃはその送信クロックをＴＴＲにロックすることができる。ついでＨＳＴＵ−ＲはＨＳＴＵ−Ｃから受信した信号からタイミングの回復を実行する。実施形態において、ＨＳＴＵ−Ｒにおけるタイミング回復はループバックタイミング方式のなかで受信ハンドシェイク信号に基づき実行される。例えば、ループバックタイミングの実行により、ＨＳＴＵ−Ｒ送信機のクロックはＨＳＴＵ−Ｒ受信機のクロックにロックされ、ＨＳＴＵ−Ｒ受信機のクロックはＨＳＴＵ−Ｃ送信機のクロックにロックされ、ＨＳＴＵ−Ｃ送信機のクロックはＴＴＲにロックされる。もう１つの実施形態において、タイミング回復を可能にするために専用のパイロットトーンがＨＳＴＵ−ＣからＨＳＴＵ−Ｒに送信される。このパイロットトーンは全てのＤＭＴシンボルのなかで送信することができる。例えば、トーン１６をパイロットトーンとして使用できる。
混合ＳＮＲ環境用並列復号化
上述したように、G.hs受信機はＮＥＸＴとＦＥＸＴビットマップにおける異なるノイズ状況下で受信情報を復号する。さらに、任意の単一ＴＴＲ周期内においてＮＥＸＴからＦＥＸＴノイズへの移行中に１〜２個のシンボルがある。このような単一ＴＴＲ周期中の混合ＳＮＲ環境に対し、G.hs情報ストリームの復号を最適化する必要がある。Ｎ個の連続するシンボルに亘り単一の情報ビットが送られるので、本発明の１つの実施の形態では各ビット時間中に最大Ｎ個の受信機を並列に動作させる（例えばＮ＝８）。G.hsフレームの受信後に、各受信機についてエラーチェック（例えばＣＲＣ）を計算する。正しいＣＲＣをもつ受信機からのビットストリームを使用する。
【００５７】
本発明の実施の形態に関する上記の説明は例示と説明のために提示したものである。説明は包括的ではなく、本発明を開示した形態そのものに限定することを意図していない。本書の開示に照らして多くの変形、変更形態が可能である。
【００５８】
例えば、本書で説明した特定の応用では、ハイパーフレーム合わせを確立するため、又はG.hs手続きをＴＴＲにロックするために特殊定義ＴＴＲ表示信号を使用する。しかしながら、本発明の原理に基づく特殊定義証印信号は、一般に特定の手続き又はイベントを与えられた基準に同期させるのに使用することができる。いったん同期が確立すれば、同期関係に関連する既知の特性を使用することができる。本書における一例ではG.hs信号がＴＴＲに同期する。この同期に関連する既知の特性として識別可能なＦＥＸＴとＮＥＸＴタイムの期間が含まれる。この与えられた知識に基づき、強力で信頼性の高い通信を行うことができる。このような同期関係を有するその他の応用についても等しく本発明の恩恵を受ける。
【００５９】
本発明の範囲は、特許請求の範囲によるべきであり、詳細な説明により限定されないことを意図している。
【図面の簡単な説明】
【図１ａ】 ダウンストリームのＡＤＳＬアネックスＣハイパーフレーム構造を示す図である。
【図１ｂ】 上り方向のＡＤＳＬアネックスＣハイパーフレーム構造を示す図である。
【図２】 本発明の実施の形態に基づき、ハイパーフレーム合わせを実行するように構成されるＡＤＳＬモデムのブロック図である。
【図３】 本発明の実施の形態に基づく、相互相関信号の例を示す図である。
【図４】 本発明の実施の形態に基づき、ハイパーフレーム合わせを実行する方法を示すフローチャートである。
【図５】 本発明のもう１つの実施形態に基づき、ハイパーフレーム合わせを実行する方法を示すフローチャートである。
【図６】 本発明の実施の形態に基づく、初期化処理の概要を示すタイム図である。
【図７】 本発明の実施の形態に基づく、初期化処理を示すフロー図である。[0001]
BACKGROUND OF THE INVENTION
The present invention relates to communication, and more particularly, to TTR indication technology in an ADSL Annex C communication system, whereby communication of an ASDL Annex C transceiver is made into a timing reference (TTR) signal of a communication network. It is possible to synchronize.
[0002]
[Prior art]
The Telecommunication Standardization Department (ITU-T) of the International Telecommunication Union makes recommendations to facilitate interoperability of communication networks. Two of these recommendations are G.992.1 and G.992.2, which define asynchronous digital subscriber line (ADSL) transceiver technology.
[0003]
Both the G.992.1 and G.992.2 standards use a multicarrier modulation scheme called the Discrete Multitone (DMT) modulation technique. The DMT modulation scheme uses a number of carriers or “tones” for both upstream and downstream communications. For example, G.992.1 downlink communication uses indexes up to 255, while G.992.2 downlink communication uses only indexes up to 127. By using multiple carriers rather than a single carrier, the available channel capacity is maximized and transmission bandwidth performance is optimized.
[0004]
Both G.992.1 and G.992.2 have annex C that provides special recommendations for ASDL transceivers in the noise environment of the time compression multiplexed / integrated digital communication service network (TCM-ISDN). Reference to “Annex C” is either or both of G.992.1 and G.992.2. TCM-ISDN is defined in Appendix III of ITU-T recommendation G.961. Appendix G of Recommendation G.961 describes a ping-pong method in which data transmission (and reception) between two transceivers is performed under the control of a 400 Hz clock called TCM-ISDN timing reference (TTR). In Annex C of G.992.1 and G.992.2, the central office transceiver transmits a TTR indication signal during training of the transceiver, and the customer transceiver can receive it, and the receiver and transmitter are connected to the TTR clock. Lock to. This TTR indication signal is tone 48 with phase reversal in the current versions of G.992.1 and G.992.2. The central office transmitter (ATU-C) basically transmits the data stream in the first half of the TTR cycle, and the customer transceiver (ATU-R) basically transmits in the second half of the TTR cycle. This type of ping-pong method is used in communication channel environments where the crosstalk interference level is high due to low-quality insulators (eg pulp insulation) in cable bundles where simultaneous transmission between the central office and the customer's transceiver is difficult. Especially useful.
Hyperframe alignment TTR based on TTR detection is used to lock the ATU-C local clock frequency that controls the ATU-C A / D and D / A sampling rates and the transmitter and receiver symbol rates. . The ATU-C transmitter checks the phase of the TTR and locks its hyperframe window to the TTR. On the ATU-R side, the receiver tracks the signal received from the ATU-C transmitter and locks the local clock to the ATU-C clock frequency. The ATU-R further detects a hyperframe pattern from the TTR display signal received from the ATU-C, and matches the symbol counter with the hyperframe pattern (generally called hyperframe alignment). The symbol counter is used to keep track of the symbol index and is incremented by 1 for each symbol. This counter is reset to zero when it reaches 345 (there are 345 symbols in the hyperframe). In this way, the transmitter and receiver are synchronized to the hyperframe. This alignment process is performed during transceiver training of the communication link between the two ADSL modems.
[0005]
Annex C defines a dual bitmap mode (DBM) encoding technique that provides a dual bitmap that is switched in synchronism with a hyperframe pattern that is synchronized to TTR to provide a data stream with a dual bit rate. This technique is based on the observation that for short to medium distance local loops (eg about 3 km), the signal-to-noise ratio (SNR) of the channel is high enough during the NEXT interference period to transmit data at a low bit rate. ing. Thus, under certain conditions, the DBM allows the TCM-ISDN transceiver to operate at full duplex by using different bit rates under NEXT and FEXT interference. In this sense, the communication channels operating under the DBM in the TCM-ISDN environment are substantially two communication channels, one being the FEXT channel and the other being the NEXT channel. A single bitmap mode (SBM, especially called FEXT bitmap or FBM) encoding technique is also defined. In this case, the central office and the remote transmitter / receiver transmit data only in FEXT time, and do not transmit data at the same time (half-duplex mode).
[0006]
In DBM encoding, the bit rate can be changed by switching the bitmap used to encode the symbols to be transmitted. As will be appreciated by those skilled in the art, a “bitmap” determines the number of bits of a symbol that can be encoded into each subchannel. A “symbol” is a basic unit of information transmitted by a transceiver. The number of bits encoded in each subchannel of the symbol is limited by the quality of the communication channel. The quality of the communication channel can be expressed by its SNR. Thus, a system using DBM uses two bitmaps with different data rates, one bitmap for NEXT time and one bitmap for FEXT time. In a system using the FBM, data is not transmitted at the NEXT time, so only one bitmap (FEXT bitmap) is used. Since each bitmap used in the DBM corresponds to a different bit rate, a bitmap transition from one to the other must be detected in order to ensure a compatible transmitter / receiver pair. Since the standardized frame rate is not a multiple of the 400 Hz TTR signal, it is not easy to identify the boundary between the NEXT channel and the FEXT channel.
[0007]
According to Annex C of G.992.1 and G.992.2, ATU-C indicates the transition between the NEXT bitmap and the FEXT bitmap with a single tone (tone 48) phase change. Specifically, the bitmap transition is displayed by changing the phase of this single tone by 90 degrees on the transmitter side. Here, the ATU-R receiver needs to detect this phase change, and then detects and aligns the 345 symbol patterns shown in FIG. 1a with the hyperframe of the transmitter.
[0008]
However, when the loop is long and noisy, it is very difficult to detect the phase change of tone 48, and TTR detection is a bottleneck for long reach. Furthermore, since the TTR indication signal with tone 48 only indicates a bitmap transition, the receiver needs to search the pattern inherent in FIG. 1a in order to identify and align the hyperframe boundaries. . This complicates hyperframe alignment processing on the receiver side and increases the possibility of failure in a long loop.
[0009]
[Problems to be solved by the invention]
Therefore, what is needed is a TTR display signal that can be detected on a long, noisy loop. Generally speaking, there is a need for signal synchronization or procedures tailored to TTR.
[0010]
[Means for Solving the Problems]
One embodiment of the present invention provides a synchronization method for synchronizing an initialization procedure to a timing reference signal in an ADSL Annex C communication system. In this synchronization method, the special definition signal is transmitted during one or more known symbols related to the output hyperframe. When received at the receiving node, synchronization with the timing reference signal is established by the special definition signal. This allows a synchronous initialization procedure. In an alternative embodiment, a special definition signal is received during one or more sets of known FEXT symbols for the output hyperframe to establish synchronization with the timing reference, thereby performing a synchronization initialization procedure. enable.
[0011]
Another embodiment of the present invention provides a synchronizer configured to synchronize transmission by an ADSL Annex C transceiver with a timing reference signal. Such a synchronizer comprises a signal generation module configured to generate a specially defined signal that is transmitted during one or more known DMT symbols included in an output hyperframe. When received at the receiving node, the special definition signal establishes synchronization with the timing reference signal to enable a synchronization procedure. An alternative embodiment comprises a detector module configured to detect specially defined signals during one or more known DMT symbols included in the input hyperframe. This detection establishes synchronization with the timing reference signal and enables the synchronization procedure.
[0012]
Another embodiment of the present invention provides a signal used in an ADSL Annex C communication system. This signal is, for example, a set of tones that are transmitted during a plurality of known DMT symbols contained in a hyperframe. By detecting this signal, it is possible to synchronize the transmission performed by the transceiver included in the communication system with the timing reference signal, thereby enabling hyperframe alignment and / or synchronization initialization procedures.
[0013]
The features and advantages described herein are not all inclusive, that is, many additional features and advantages will be apparent to those skilled in the art from the drawings and detailed description. Further, the language used herein is selected for the purposes of readability and explanation and is not intended to limit the scope of the subject matter of the present invention.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
In general, the present invention is implemented in a communication system having a transceiver pair that is communicatively coupled by a transmission medium (eg, a stranded copper wire pair). Thus, there is a pair of transceivers in each communication direction (up and down). A specially defined TTR display signal is generated by the transmitter at the transmitting node and transmitted to the receiving node. The special definition signal is reliably detected by the receiver, which allows the transceiver to synchronize the transmission with the TTR. Therefore, hyperframe alignment can be performed together with the synchronization initialization procedure.
Instead of using a 90 degree phase reversal with a single tone to mark the bitmap boundaries of the specially defined TTR display signals FEXT and NEXT, one embodiment of the present invention can be used in a given hyperframe . REVERV type tones 33-64 are transmitted only during the periods of symbols 0, 1, 2, and 3. This special definition signal is hereinafter referred to as REVERB 33-64. As will be apparent from the disclosure herein, other specially defined TTR indication signals may also be used (eg, inverted REVERBx˜y, where xy represents a set of tones).
[0015]
For example, the processor module included in the ATU-C is programmed to supply REVERB 33-64 in a plurality of transmission hyperframes, or configured in other ways. The processor module included in the ATU-R can be configured to detect known patterns of REVERB 33-64 at symbols 0-3. Once the pattern is detected from four consecutive symbols, the positions of known symbols 0-3 within the hyperframe are identified. The ATU-R processor module can confirm these identification positions by detecting the REVERB 33-64 signal again from four consecutive symbols in the following hyperframe. Specifically, when four consecutive symbols with a known pattern are detected for the second time, the receiver knows that it is receiving symbol 3, and thus the hyperframe boundary is known.
[0016]
To ensure robust communication, the ATU-C transmitter should continue to transmit REVERB 33-64 signals until the ATU-R responds by transmitting R-REVBRB1 (or other response signal or flag). Can be configured. Accordingly, the REVERB 33-64 signals are transmitted as many hyperframes as necessary until the receiving node returns an alignment confirmation response. If no acknowledgment is received after several attempts (eg 5), the transmitting transmitter is configured to recognize the error. Appropriate processing is then performed, eg, restarting the start or loop diagnostic procedure to check for problems with the current communication link or switch to a new link.
Alternative TTR Indication Signal The present invention is not intended to be limited to the embodiments described herein. Rather, the principles of the invention are applicable to many specific configurations. For example, DMT spectrum tones 33-64 were selected because they represent low frequency downstream tones as defined in G.992.1. Such low frequency tones can be transmitted over long distances, improving the “reach” of the present invention. However, other tones can be used, for example any tone or capable of transmitting detectable information over a subset of tones 33-64 (eg 34-61) or long distance loops (eg 5km or more) Tone groups (tones 6-31 or a subset thereof in overlapping spectrum applications where tones 6-31 are also used in the downstream direction) can be used. Similarly, data patterns other than the REVERB type can be used to provide uniqueness to the transmitted TTR signal. In short, any signal that can be uniquely detected at the receiving end of a long distance loop can be used in accordance with the principles of the present invention.
[0017]
Furthermore, symbols other than 0-3 can be used. For example, symbols 341-344 may be used, or any group of uniquely identifiable symbols or portions of frame data may be used to identify transmitter and receiver frame boundary offsets. In one alternative embodiment, a known pattern (eg, REVERB 33-64 or REVERB 6-31) is transmitted during symbols N ₁ -N ₂ (eg, symbols 140, 141, 142, 143, 144); Again, during the period of symbols N _{3 to} N ₄ (eg, 237, 238, 239, 240, 241). In this case, the special definition signal is transmitted twice over a group symbol within one hyperframe. According to such a specific example, a hyperframe alignment can be identified within _one hyperframe (while receiving symbols N _{1 to} N ₂ ) rather than using two or more hyperframes as described above, This can be confirmed (while receiving symbols N ₃ -N ₄ ).
[0018]
The more symbols that are associated with a specially defined signal, the lower the chance of signal detection errors at the receiver. Thus, transmitting a special definition signal during a plurality of symbols (eg, two or more symbols) provides a robust TTR indication. However, by transmitting a specially defined signal during only one known symbol period (when this signal can be accurately detected at the receiving node due to the noise and interference characteristics of the communication link), a hypertext based on the principles of the present invention. Frame alignment can be achieved.
Transceiver Unit FIG. 2 is a block diagram of an ADSL modem configured to perform hyperframe alignment, according to one embodiment of the present invention. This type of modem pair is communicatively coupled, with one modem used at the local node (eg, CO (central office)) and the other modem used at the remote node (eg, CPE (customer terminal)). The
[0019]
The illustrated embodiment includes a transformer 205, a hybrid 210, a line driver 215, an analog front end (AFE) 220, and a processor module 225. The system interface is coupled to a network that operates, for example, in an asynchronous transmission mode or a synchronous transmission mode. The network operator or manager can communicate with the modem via the system interface. Alternatively, the system interface may be coupled to an application running on a device such as a customer personal computer. The transmission medium is typically a copper wire loop or stranded wire pair. However, the invention also works with other types of transmission media.
[0020]
Transformer 205 couples the transmission medium to the modem's circuitry and provides electrical isolation and a balanced interface. The hybrid 210 performs 2 / 4-wire conversion and converts a bidirectional 2-wire signal from the transmission medium into two pairs of one-way transmissions. One pair is for reception and the other pair is for transmission. The AFE 220 typically includes a digital / analog (D / A) converter 220a and an analog / digital (A / D) converter 220b, and further includes a gain adjustment module that optimizes a signal supplied to the processor module 225. Including. A signal received via the AFE 220 from the four-wire interface of the hybrid 210 is converted from analog to digital by the A / D converter module 220b. The digital signal is then supplied to the processor module 225.
[0021]
Regarding the transmission direction, the data received from the system interface is processed by the processor module 225. The digital output of the processor module 225 is converted into a corresponding analog by the D / A converter 220a. The analog output of the AFE 220 is provided to a line driver 25 that is operatively coupled to the four-wire interface of the hybrid 210. The transmission signal is supplied to the communication medium via the transformer 205. In an alternative configuration, it is not necessary to include all the elements shown in FIG. 2, for example a transformerless design.
[0022]
The processor module 225 is programmed or otherwise configured to perform hyperframe alignment processing in accordance with the principles of the present invention. In the illustrated embodiment, the processor module 225 includes a signal generation module 225a and a detector module 225b. The signal generation module 225a is configured to generate a specially defined TTR display signal for transmission from the transmitting node to the receiving node, and the detector module 225b is configured to detect the specially defined TTR display signal. The processor module 225 performs a number of other functions such as modulation / demodulation, coding / decoding, scrambling / descrambling, error detection (eg, CRC checking), framing / deframing and other algorithmic functions May be.
[0023]
For convenience of explanation, it is assumed that a modem having a configuration as shown in FIG. 2 is used at both a transmission node and a reception node that define one direction of a communication link. The special definition TTR display signal is transmitted from the transmitting node and received by the receiving node.
TTR indication signal generator During initialization, the processor module 225 generates a specially defined TTR indication signal, and a uniquely identifiable signal is transmitted to the transmission medium during one or more known symbol periods. The signal generation module 225a may be integrated into the processor module 225 as shown, or may be an independent module that is communicatively coupled to the processor module 225. For example, the signal generation module 225a is a programmed, independent module, activated by a request from the module 225, supplying a specially defined data pattern, scrambled or otherwise processed by the module 225, and a predetermined number of Can be encoded into known DMT symbols. In a specific example, the characteristics of the specially defined TTR display signal are programmed into the signal generation module 225a. For example, the programmed characteristics specify a subset of tones contained in a REVERB type data pattern.
[0024]
In another specific example, the characteristics of the specially defined TTR display signal are held in a storage unit (eg, EEPROM or other programmable storage device). When the hyperframe matching process is started, the signal generation module 225a accesses the signal characteristics from the storage unit and generates a corresponding TTR signal. In this type of implementation, the local host can access storage characteristics to dynamically update or change the specially defined TTR indication signal. For example, the special definition TTR display signal is changed from REVERB 33 to 64 (for example, used in the non-overlapping spectrum method) to REVERB 6 to 31 (for example, used in the overlapping spectrum method) or vice versa. In any case, the specially defined TTR indication signal is transmitted during one or more symbol periods.
[0025]
At the receiving node, the specially defined TTR indication signal is separated from the transmission medium by the transformer 205 and supplied to the processor module 225 via the hybrid 210 and the AFE 220. Based on the known characteristics of the specially defined TTR display signal, the detector module 225a of the receiver can detect the signal and perform hyperframe alignment.
TTR detector The detector module 225a at the receiving node detects a specially defined TTR indication signal for a predetermined period of time (e.g., DMT symbols 0-3). In a specific example, during this detection, detector module 225a uses linear cross-correlation. In addition, the detector module 225a can perform decimation on the received data before performing cross-correlation to reduce computational complexity.
Cross-correlation In general, the cross-correlation between two signal sequences x (n) and y (n) is a sequence r _xy (l) defined by the following equation.
[0026]
[Expression 1]

In a specific example, x (n) is a REVERB 33-64 signal in the time domain, and y (n) is a signal received by the receiver through x (n) transmitted from the transmitter and passed through channel h (n). It is. Therefore, y (n) is a convolution of x (n) and h (n) and is defined by the following equation.
[0027]
[Expression 2]

Substituting y (n) in equation (2) into equation (1) yields:
[0028]
r _xy (l) = h (l) l = 0, ± 1, ± 2, ... (3)
Equation (3) is based on the fact that x (n) is a broadband PRBS (pseudo-random binary sequence). As an example of this type of signal, consider the REVERB type signal described in the G.992.1 document. Furthermore, the autocorrelation of x (n) is approximated by a delta function. FIG. 3 shows an example of a correlation signal based on one embodiment of the present invention. This correlation signal indicates that four consecutive REVERB 33-64 signals are present in the received signal of the DMT receiver. A -130 dBm / Hz AWGN added to a 5 km paper pulp loop and receiver is assumed. As shown, there are four distinct peaks that indicate the presence of REVERB 33-64 signals in DMT symbols 0-3.
Decimation According to the Nyquist sampling theorem, a low-pass analog signal with a maximum bandwidth of f _m Hz is represented by a sample and fully recovered from the sample, given a sampling rate f ₀ greater than or equal to f _m be able to. A signal obtained by sampling an analog signal having a maximum bandwidth f _m Hz at a sampling rate of 2 f _m or more, for example, 4 f _m , can be decimated (discarded every other sample) by 2 without any information loss. When the oversampled analog signal includes an unnecessary signal exceeding the use band, filtering is necessary prior to the decimation in order to avoid unnecessary signals from being turned back into the use band after decimation. This process is called decimation filtering.
[0029]
In the case where the specially defined TTR indication signal is REVERB 33-64 transmitted during symbols 0-3, detector module 225b indicates that four consecutive REVERBs 33- Continue to perform linear cross-correlation processing on received samples until 64 signals are detected. Each time a DMT symbol is received, a cross-correlation signal is calculated. The received samples are buffered (or stored in RAM, other storage devices coupled to or within the detector module) or otherwise retained for analysis by the detector module 225b. Initially, when the REVERB 33-64 signal appears in the received signal, the cross-correlation peak position can be any position within the sample of the correlation signal calculated for that received symbol. This is because the frame boundary position of the receiver is not yet aligned with the frame boundary of the transmitter. Here, the peak position of the cross-correlation signal indicates the offset of the frame boundary between the transmitter and the receiver that enables hyperframe alignment. In this sense, the symbols 0 to 3 have a predetermined indicia for hyperframe alignment.
[0030]
Here, the detector module 225b performs frame boundary alignment based on the peak position of the cross-correlation signal, and performs correlation processing again to search for N REVERB 33 to 64 signals of the next hyperframe. If the frame boundaries are properly aligned based on the detection of the first four consecutive REVERB 33-64 signals, a cross-correlation peak will occur at the position of a known sample (eg, the first sample) of the cross-correlation signal ( The next four REVERB 33-64 signals are detected (in the next hyperframe). In this way, a known offset between the transmitter and receiver frame boundaries is determined. In particular, the fourth detected REVERB 33-64 signal having a peak position at a known sample of the cross-correlation signal represents the boundary of the DMT symbol 3 in the hyperframe. In this way, the receiver achieves hyperframe alignment.
[0031]
Here, the phase toggle operation of the tone 48 according to the current standard (conventional TTR display signal) can be performed as before without affecting the function of the detector module. Further, a tone 64 (a conventional pilot signal) is included in the REVERBx to y signals. Thus, the present invention is backward compatible.
Implementation Details In an embodiment, the processor module 225 is a digital signal processor (DSP) or other suitable device programmed to perform hyperframe alignment or initial handshake synchronization in accordance with the present invention. It is a processing environment. In an alternative embodiment, the processor module 225 is implemented with dedicated silicon such as an ASIC or chipset. A combination of DSP and ASIC technology may be used to implement the processor module 225. Thus, the present invention can be used by ADSL service providers as a method for performing hyperframe alignment, for example. Alternatively, the present invention can be incorporated into devices such as modems, programmable DSPs or dedicated chips / chipsets. Alternatively, the present invention can be incorporated into a system such as a transceiver pair in which one or both transceivers are configured to perform hyperframe alignment or initial handshake synchronization, described below. Alternatively, the present invention can be incorporated as a computer program product or the like embodied on a computer readable medium such as a server or disk.
[0032]
Further, a module (not shown) can be incorporated into the modem according to the application. For example, a modem may operate for overlapping spectrum applications and may include an echo canceller. Other modules (eg, transmitter and receiver clocks) and configurations will be apparent from this disclosure. Further, the processor module 225 can start the hyperframe matching process as part of the link initialization procedure. Alternatively, the alignment process may be triggered by an event (eg, alignment detection failure) or a control signal from the host (eg, hyperframe alignment software call or logic enable line). Similarly, after detecting activation of hyperframe alignment by the remote modem, the processor module 225 can start hyperframe alignment processing. Typically, the ATU-C modem starts hyperframe alignment processing. However, if desired, the ATU-R can be configured to initiate this process.
Methodology FIG. 4 shows a flowchart of a method for performing hyperframe alignment processing according to an embodiment of the present invention. This method is typically implemented in a communication system (eg, ADSL Annex C) comprising two modems that are communicatively coupled via a transmission medium. In particular, this method is performed at the transmitting node associated with a particular communication direction of the system. The modems of both the transmitting (for example, local) and receiving (for example, remote) nodes are configured as described with reference to FIG. However, other configurations are apparent from the present disclosure.
[0033]
The method begins by invoking a hyperframe alignment procedure (step 405). In an embodiment, the hyperframe alignment procedure is initiated by the ATU-C during communication link initialization (eg, training). However, in alternative embodiments, the alignment procedure can be triggered at other times (eg, after a service interruption). Further, the method generates a special definition signal (step 410) and transmits the special definition signal to the remote node during one or more known intervals in the data frame (step 415).
[0034]
In an embodiment, the data frame is an ADSL downstream hyperframe with DMT symbols 0-344, and one or more known intervals in the data frame are DMT symbols 0-3 in the hyperframe. The first transmission of this data frame makes it possible to identify the relative positions of the transmitter and receiver symbol boundaries as described above. The hyperframe alignment can be confirmed by the second (or subsequent) frame transmission. Instead, one or more known intervals in the data frame are symbols N ₁ -N ₂ (eg, 140-144) in the hyperframe and symbols N ₃ -N ₄ (eg, 237-241) in the hyperframe. In such an embodiment, it becomes possible to identify the offset of the symbol boundary between the transmitter and the receiver by the special definition signal transmitted during the period of symbols N _{1 to} N ₂ , and between symbols N _{3 to} N ₄ . The hyperframe alignment can be confirmed by the transmitted special definition signal.
[0035]
For example, a specially defined signal is a pseudo-random data sequence or other uniquely identifiable signal. In an embodiment, the special definition signal is a subset of tones (eg, REVERB 6 to 31, REVERB 16 to 31, REVERB 33 to 64, or REVERB x to y (where x to y are information transmissions of a long loop) included in the REVERB type signal described above. Represents a set of tones that can be used in By transmitting such a special definition signal in a known section of a data frame, a strong TTR display signal that can be received at the remote end of a long loop is provided.
[0036]
Further, the method receives a response from the remote node) for confirmation of hyperframe alignment (step 420), and determines whether alignment has been achieved based on the response (step 425). If the response confirms that the hyperframe is aligned, the procedure ends. The communication system then completes other initialization procedures and transitions to a data mode (eg, show time). However, if the response confirms a hyperframe position mismatch, steps 415 through 420 are repeated. As described above, if no acknowledgment is received after several attempts (eg, transmission of N hyperframes), the transmitting transceiver is configured to recognize as an error.
[0037]
FIG. 5 shows a flowchart of a method for performing hyperframe alignment processing according to another embodiment of the present invention. In particular, this method is performed at a receiving node associated with a particular communication direction of the system.
[0038]
The method begins by invoking a hyperframe alignment procedure (eg, by ATU-C during training during communication link initialization) (step 505). Further, the method receives data frame data from the remote node transmitter (step 510) and detects special definition signals in one or more known intervals in the data frame (step 515). The above description regarding specially defined signals and known symbols also applies here. The method proceeds to step 520 where the offset between the transmitter and receiver interval boundaries is identified and it is determined in step 525 whether hyperframe alignment has been achieved. Specific embodiments that implement these steps are described below.
[0039]
For convenience of explanation, it is assumed that the special definition signal transmitted in a known section in the data frame is a REVERBx to y signal transmitted during symbols 0 to 3 in the hyper frame. In an embodiment, the offset between the transmitter and receiver interval boundaries is a linear cross-correlation for the receiver samples until four consecutive REVERBx-y signals are detected (thus indicating symbols 0-3). Is identified by continuing to execute. Here, the frame boundary offset between the transmitter and the receiver is identified based on the cross-correlation signal peak of the REVERVx to y signals. The linear cross-correlation process is repeated in subsequent hyperframes (eg, the next hyperframe), and the next four consecutive REVERBx-y signals are detected. Hyperframe alignment is achieved by detecting the next four consecutive REVERBx-y signals with the cross-correlation peak located at a known sample (eg, the first sample) of the cross-correlation signal.
[0040]
In another embodiment, the offset between the transmitter and receiver interval boundaries calculates a cross-correlation signal for each DMT symbol reception and has a predetermined hyperframe alignment indicia in the cross-correlation signal. , N by identifying N consecutive DMT symbols. This predetermined indicium includes, for example, a set of tones (xy) (eg, tones 33-64 of the REVERB signal) in a pseudo-random data sequence. Frame boundary alignment is performed based on the peak position of the cross-correlation signal having this indicia. A second cross-correlation process is performed to identify a second set of N consecutive DMT symbols. Frame boundary alignment is confirmed by a second detection of a predetermined indicium to indicate that hyperframe alignment has been achieved. Note that the second cross-correlation process may be executed in either the same hyperframe or the subsequent hyperframe. Furthermore, the process may be repeated three times, four times, five times, etc. to ensure proper alignment.
[0041]
If hyperframe alignment is not achieved, the method sends a negative response to the remote node indicating that hyperframe alignment has not been achieved (step 530). Such a response is, for example, a specially defined signal or just a period of silence. Here, the transceiver in the receiving state can be configured to recognize as an error if no acknowledgment is received after multiple attempts (eg, multiple hyperframe transmissions). On the other hand, if hyperframe alignment is achieved, the method sends a response confirming the hyperframe position match to the remote node (step 535). Next, the communication system completes other initialization procedures (for example, channel analysis and exchange) and shifts to a data mode (for example, show time).
Overall Overview of Initialization Process FIG. 6 shows a time diagram outlining the initialization process according to one embodiment of the present invention. It is assumed that the initialization process is performed in a communication system comprising one or more pairs of ADSL communications that are communicatively coupled via twisted copper wire pairs or other transmission media.
[0042]
The initialization process shown for each transmitter / receiver has a configuration in which a transmitter / receiver training and a channel analysis procedure follow after an initial handshake procedure. Then, the information acquired in the training and analysis procedures is exchanged. The handshake procedure generally allows newly combined transceivers to exchange capabilities and select a common operating mode. This initial handshake procedure is typically performed based on ITU-T Recommendation G.994.1 (G.hs) or other suitable handshake procedure.
[0043]
However, G.994.1 is not synchronized with TCM-ISDN, which applies the time division duplex scheme based on TTR described above. Thus, the G.994.1 signal interferes with TCM-ISDN or Annex C FBM NEXT, especially in long distance loops (eg, 5 km or more). As a result, the handshake procedure may be unsuccessful. Therefore, there is a need to make G.994.1 reliable and powerful for G.992 Annex C modems.
[0044]
One embodiment of the present invention allows G.994.1 to be synchronized with TCM-ISDN, which is achieved in the front period of the handshake procedure. As a result, the G.994.1 signal can be exchanged on the long distance loop, for example, during the FEXT time period, and transmission during the NEXT time period can be avoided. Therefore, a more powerful and reliable handshake procedure is realized.
Synchronization of Annex C G.994.1 to TTR FIG. 7 is a flowchart of a dual initialization procedure according to one embodiment of the present invention. In particular, the G.994.1 handshake procedure is synchronized with the TTR. Once the handshake procedure is synchronized to TTR, the NEXT and FEXT periods are known. This can intentionally send a G.hs signal (other initialization signals, such as a training signal if possible) during the FEXT period of the other modem to ensure a reliable exchange. .
[0045]
During the handshake procedure, the communication modems are called HSTU-C and HSTU-R, respectively. The existing HSTU-R activation signal (R-TONES-REQ) is phase-inverted every 16 ms, and TTR has a period of 80 ms which is 5 × 16 ms. In accordance with one embodiment of the present invention, the HSTU-C TTR indication signal is a signal such that the phase inversion is not synchronized with the TTR and prevents accurate reception every 16 ms. -It is determined to be surely received by the HSTU-R that is transmitting the REQ.
HSTU-C activation scheme of handshake procedure Referring to Figure 7, HSTU-C starts with a transition from C-SILENT1 and is a specially defined hand called C-INIT at the start of the TTR hyperframe. Send a shake signal. For example, C-INIT is REVERB 16-31 or other specially defined signal as described above. In an embodiment, the C-INIT is transmitted during DMT FEXT _R symbols 0-3, 33-36, 108-111 and 140-144. This special definition signal enables synchronization with the TTR, thereby enabling a synchronized G.hs procedure, in which the handshake signal is transmitted only during the FEXT symbol period as shown in FIG. If desired, synchronization with the TTR may be maintained in subsequent initialization procedures such as training, channel analysis and exchange. Alternatively, initialization may be re-established in a subsequent initialization procedure.
[0046]
Here, this sequence of DMT symbols was chosen because this sequence is robust under the circumstances where phase inversions of 16 ms intervals appear in any phase of the TTR period. As mentioned above, other data patterns and symbol combinations may also be used here in accordance with the principles of the present invention to achieve reliable transmission and reception.
[0047]
When C-INIT is detected (eg, the TTR detector described above), the HSTU-R sets its PGA and performs symbol alignment and timing recovery. Furthermore, HSTU-R begins with symbols 6-9, continues with symbols 16-19, 27-30, ... 210-214, ... 340-343 (shown in Fig. 1b), between FEXT _C symbols. , R-TONE1 _F is transmitted. In an embodiment, if the HSTU-R detects C-TONES (conventional handshake signal), the HSTU-R waits for a timeout T1 before responding appropriately, and the HSTU-C is based on the principles of the present invention. Determine whether to support hs procedures.
[0048]
After detecting R-TONE1 _F , HSTU-C transmits C-GALF1 _F starting at the start of the TTR hyperframe only during the FEXT _R symbol. Each bit is transmitted in the first four FEXT _R symbols of a subframe (0-3, 11-14, 22-25, 33-36, ... 140-143, ... 335-338). 32 bits are transmitted in an 80 ms hyperframe. As mentioned above, the G.992 hyperframe consists of 345 DMT symbols. There are 32 subframes in a hyperframe, and each subframe is 10 to 11 consecutive DMT symbols depending on TTR timing.
[0049]
After receiving the C-GALF1 _F, (as in the case of R-TONE1 _F) HSTU-R begins symbols 6-9, it transmits between the FEXT _C symbol R-FLAG1 _F. Upon receipt of R-FLAG1 _F , HSTU-C transmits C-FLAG1 _F during the first four FEXT _R symbols in the subframe, starting from any TTR subframe. Upon receiving C-FLAG1 _F at the HSTU-R, the HSTU-R can begin the first handshake transaction starting from any TTR subframe.
HSTU-R activation method of handshake procedure Since HSTU-R is locked to TTR from the beginning, it is beneficial for HSTU-C to activate (start) the handshake procedure. As a result, no signal is transmitted during the NEXT period. However, the HSTU-R can start the handshake procedure by shifting from R-SILENT0 and transmitting R-INIT. In an embodiment, R-INIT is REVERB 6-15, but the signals described above and other specially defined TTR indication signals can be used here. After receiving the R-INT, the HSTU-C transmits a C-INIT starting at the start of the TTR hyperframe, as described in the HSTU-C activation method.
[0050]
Here, R-INIT is transmitted before the HSTU-R receives a signal from the HSTU-C locked to the TTR. To avoid NEXT affecting adjacent pairs of modems, the HSTU-R can detect the noise pattern and can be configured to send R-INIT only at high noise levels (FEXT _C period). .
Fallback procedure / backward compatibility As noted above, the present invention is backward compatible with conventional initialization procedures. For example, assume that HSTU-C sends a C-INIT to initiate a handshake procedure, but HSTU-R does not respond. The reason for no response is, for example, that the HSTU-R is not connected, does not support an initialization procedure according to the principles of the present invention, or is a conventional G.994.1 initialization procedure for the loop for which it is provided. Has decided that is better.
[0051]
After the timeout, the fallback procedure can be entered. For example, HSTU-C is configured to send C-TONES. If the HSTU-R now responds with R-TONE1, the transceiver pair falls back to the conventional G.994.1 initialization procedure. However, if the HSTU-R does not respond to C-TONES, the HSTU-C is configured to send C-INIT again after a timeout. This alternating operation continues several times until the HSTU-R responds positively or a long timeout has elapsed. Such alternating operation can reduce the interoperability problems in existing G.994.1 devices.
[0052]
The length of the alternation is predetermined or otherwise determined so that the HSTU-R monitors how long the received signal is to determine whether the HSTU-C is configured according to the principles of the present invention. I know what to do. The HSTU-R responds accordingly. It is also possible to send a TTR display signal based on the principles of the present invention in parallel with an existing G.994.1 activation signal.
[0053]
Next, suppose that HSTU-R sends R-INIT to initiate a handshake procedure (for some reason) HSTU-C does not respond. After a T2 timeout, the HSSTU-R is configured to send an R-TONES-REQ (a conventional handshake signal). If the HSTU-C responds with C-TONES, the transceiver pair falls back to conventional G.994.1 and initialization procedures. However, HSTU-C responds to R-INIT with C-INIT, and HSTU-R then uses conventional G (for example, because HSTU-R has determined that the loop is not very long and conventional G.994.1 is preferred). If it decides to fall back to .994.1, the HSTU-R can return silence. This causes the HSTU-C to time out, retreat back to conventional G.994.1 and send C-TONES.
Flags and messages Conventional G.994.1 requires the same information bits to be repeated in eight consecutive DMT symbols. In each TTR subframe (as shown in FIG. 1a) there are typically 4-5 FEXT _R symbols and 6-7 NEXT _R symbols. In one embodiment of the present invention, one bit is transmitted using the first four FEXT _R in one subframe. Thus, each information bit is transmitted in 4 consecutive FEXT _R symbols. In another embodiment, one bit is transmitted using all FEXT symbols in a subframe. Thus, each information bit is transmitted in 4 or 5 consecutive FEXT _R symbols. Such a configuration transmits 32 bits of information using 345 symbols, resulting in an average of 10.78 DMT symbols per G.hs bit. The modem transmits only FEXT symbols, similar to the ADSL Annex C FBM mode. The receiver algorithm is simple because there is no echo.
[0054]
In an alternative embodiment of the invention, 1-bit information is transmitted using all symbols within 1 TTR period, regardless of FEXT or NEXT symbols. The receiving node is composed of two receivers, one for FEXT symbols and the other for NEXT symbols. Compare the detection results of the two receivers and use the more reliable one (eg, based on CRC check). Under the noise of TCM-ISDN, the FEXT detector gives more reliable results. However, under other noise conditions, a NEXT detector may help to improve reliability. Such an environment complicates the receiver configuration and loses the benefits of half-duplex mode (eg, echo cancellation is required).
[0055]
The handshake signals (for example, C-INIT, R-TONE1 _F , R-INIT) and the message length for handshake are TTR hyperframes having a length of 80 ms or subs having a length of 2.5 ms. It is a multiple of the frame. Synchronization with TTR at the time of transition from G.hs to Annex C operating mode is maintained or not maintained. Annex C needs to work with existing G.994.1 procedures that are not synchronized with TTR, so anyway, Annex C itself needs to perform synchronization with TTR.
[0056]
After initial synchronization, HSTU-C and HSTU-R need to maintain synchronization with TTR. Since the TTR clock is available in the HSTU-C, the HSTU-C can lock its transmission clock to the TTR. The HSTU-R then performs timing recovery from the signal received from the HSTU-C. In the embodiment, the timing recovery in HSTU-R is performed based on the received handshake signal in a loopback timing scheme. For example, by executing loopback timing, the HSTU-R transmitter clock is locked to the HSTU-R receiver clock, the HSTU-R receiver clock is locked to the HSTU-C transmitter clock, and the HSTU-C The transmitter clock is locked to TTR. In another embodiment, a dedicated pilot tone is transmitted from the HSTU-C to the HSTU-R to allow timing recovery. This pilot tone can be transmitted in all DMT symbols. For example, tone 16 can be used as a pilot tone.
Parallel decoding for mixed SNR environments As described above, the G.hs receiver decodes received information under different noise conditions in the NEXT and FEXT bitmaps. In addition, there are 1-2 symbols during the transition from NEXT to FEXT noise within any single TTR period. It is necessary to optimize the decoding of the G.hs information stream for such a mixed SNR environment during a single TTR period. Since a single information bit is sent over N consecutive symbols, one embodiment of the invention operates up to N receivers in parallel during each bit time (eg, N = 8). After receiving the G.hs frame, an error check (eg CRC) is calculated for each receiver. Use a bitstream from a receiver with the correct CRC.
[0057]
The above description of the embodiments of the present invention has been presented for purposes of illustration and description. The description is not exhaustive and is not intended to limit the invention to the precise form disclosed. Many variations and modifications are possible in light of the disclosure herein.
[0058]
For example, the specific application described herein uses a specially defined TTR indication signal to establish hyperframe alignment or to lock a G.hs procedure to TTR. However, specially defined indicia signals based on the principles of the present invention can generally be used to synchronize a particular procedure or event to a given reference. Once synchronization is established, known properties associated with the synchronization relationship can be used. In one example in this document, the G.hs signal is synchronized to TTR. Identified periods of FEXT and NEXT time are included as known characteristics associated with this synchronization. Based on this given knowledge, powerful and reliable communication can be performed. Other applications having such a synchronization relationship will equally benefit from the present invention.
[0059]
The scope of the invention should be determined by the following claims and is not intended to be limited by the detailed description.
[Brief description of the drawings]
FIG. 1a shows a downstream ADSL Annex C hyperframe structure.
FIG. 1b is a diagram illustrating an ADSL Annex C hyperframe structure in the upstream direction.
FIG. 2 is a block diagram of an ADSL modem configured to perform hyperframe alignment according to an embodiment of the present invention.
FIG. 3 is a diagram showing an example of a cross-correlation signal based on the embodiment of the present invention.
FIG. 4 is a flowchart illustrating a method for performing hyperframe alignment according to an embodiment of the present invention.
FIG. 5 is a flow chart illustrating a method for performing hyperframe alignment according to another embodiment of the present invention.
FIG. 6 is a time chart showing an overview of initialization processing based on the embodiment of the present invention.
FIG. 7 is a flowchart showing an initialization process based on the embodiment of the present invention.

Claims (14)

A synchronization method for synchronizing an initialization procedure with a TTR in an ADSL Annex C communication system having a transmitting node and a receiving node that are communicably coupled,
Transmitting a specially defined signal that directly identifies a hyperframe boundary only within a selected symbol of at least one selected output hyperframe;
Detecting the special definition signal at the receiving node ;
Identifying an offset between hyperframe boundaries between a transmitter and a receiver in response to the detecting action;
Aligning the receiver hyperframe with the transmitter hyperframe in response to the identifying operation;
Including a synchronization method. The synchronization method of claim 1, wherein the selected symbols having the special definition signal include FEXT symbols of a selected subset of selected hyperframes . The synchronization method of claim 1 , wherein the specially defined signal includes a selected subset of tones within each of the selected symbols . The synchronization method according to claim 1 , wherein the at least one output hyperframe has a start, and the selected symbol includes only a symbol at the start of the first output hyperframe. Said at least one output hyperframe includes symbols 0 to 344, the selected symbol having the specially defined signal comprises FEXT symbols 0～3,33～36,108～111, and 140-144 only, wherein Item 2. The synchronization method according to Item 1. The synchronization method according to claim 1, wherein the specially defined signal includes a selected subset of tones modulated by a REVERB type signal in each of the selected symbols. The identifying operation further includes
Correlating the received signal with a specially defined signal to determine its cross-correlation peak;
Determining an offset between a hyperframe boundary of a transmitter and a receiver based on the correlating operation;
The synchronization method according to claim 1, comprising: A system for synchronizing an initialization procedure to a TTR in an ADSL Annex C communication system having a transmitting node and a receiving node that are communicatively coupled,
A signal generator of a transmitting node configured to transmit a specially defined signal identifying a hyperframe boundary within a selected symbol of at least one output hyperframe;
A reception configured to detect the special definition signal, identify an offset between hyperframe boundaries between a transmitter and a receiver in response to the detecting operation, and align the hyperframes of the transmitting node and the receiving node A node detection module;
including,
system. The system of claim 8, wherein the selected symbols of the at least one output hyperframe include a selected subset of FEXT symbols. The system of claim 8, wherein the specially defined signal includes a selected subset of tones within each of the selected symbols. 9. The system of claim 8, wherein the at least one selected output hyperframe has a start, and the selected symbol includes only a symbol at the start of the first output hyperframe. The at least one output hyperframe includes symbols 0-344, and the selected symbol with the specially defined signal includes only FET symbols 0-3, 33-36, 108-111, and 140-144. Item 9. The system according to Item 8. 9. The system of claim 8, wherein the specially defined signal includes a selected subset of tones modulated by a REVERB type signal within each selected symbol. The detector module further correlates the received signal with a specially defined signal to determine a peak of its cross-correlation, and determines an offset between the hyperframe boundaries of the transmitter and the receiver based on the correlating operation. 9. The system of claim 8, wherein the system is configured as follows.

Images