JP3940415B2 - Multi-carrier communication system and communication method - Google Patents

Description

The present invention relates to a multicarrier communication system and a communication method, and more particularly, to a multicarrier communication system that divides a band into a plurality of subbands that are independent narrow bands, and transmits and receives transmission data by frequency multiplexing for each subband. And a communication method.
The bit error rate (BER) of a communication system that employs a multi-carrier modulation method that cannot perform perfect reproduction such as filter bank modulation, DMT modulation, FMT modulation, etc., reduces the level of inter-channel interference (ICI). This can be improved. In a multicarrier communication system (unless the full regeneration condition is satisfied), interchannel interference is caused by leakage, which is called spectral energy leakage, sometimes called crosstalk between subchannels. The main factors of ICI in multi-carrier systems are
(A) loss of orthogonality between subchannels;
(B) a frequency offset between the transmitter and receiver;
(C) spectrum leakage,
It is. All these factors have similar physical properties. That is, these factors most often have spectral leakage to adjacent subchannels due to overlapping subchannel spectra. Due to the considerable level of overlap between the subchannel main lobes, some of the spectral energy leaks to the adjacent channels, which behaves like Gaussian noise and also leads to BER degradation.
The introduction of symbol repetition techniques in the sub-channel improves the spectral suppression level and also reduces the spectral main lobe frequency span, thus reducing the ICI level. In addition to this improvement in ICI performance, redundant data obtained by repetition of transmission symbols is used to improve BER performance (redundant data operates as an error correction code). In summary, repeated transmission of data is coupled in a natural way to FMT modulation techniques or similar filter bank modulation techniques. This combined system exhibits good immunity against frequency offsets associated with iterative code gain.

In a multi-carrier communication system that divides a band into a plurality of subbands that are independent narrow bands, and transmits and receives transmission data for each subband by frequency multiplexing, that is, filter bank modulation, DMT (Discrete Multitone) modulation, In multi-carrier communication systems such as FMT (Filtered Multitone) modulation, filter set selection has traditionally been performed under the constraint of completely eliminating inter-symbol interference (ISI) and inter-channel interference (ICI). . In an ideal transmission channel with no Doppler shift, no offset frequency between the transceivers and no signal distortion, this constraint ensures error-free recovery of the transmission symbols at the receiver. A frequency offset generated in each channel due to inaccurate tuning of the oscillator or Doppler shift causes BER degradation due to spectrum leakage or ICI. The only way to mitigate such BER degradation is to keep the frequency offset as small as possible, typically within 1% of the subcarrier frequency spacing. However, this method requires precise frequency offset estimation, and when receiving a multicarrier signal mixed with noise, if the noise level is large, the frequency offset estimation is impaired. Furthermore, this method does not work correctly in fast fading channels, i.e. in fast fading channels where the Doppler shift is not constant for the transmitted symbols and also varies with time.
As described above, an object of the present invention is to enable a highly robust multicarrier signal reception even when a frequency offset exists due to inaccurate tuning of an oscillator, Doppler shift, or the like.
Another object of the present invention is to improve BER by suppressing ICI in the subchannel.

In the multicarrier communication system of the present invention, the transmission side repeats transmission data for each subchannel a predetermined number of times, and the transmission data of each subchannel is frequency-multiplexed and transmitted. The reception side transmits each subchannel from the received signal. The transmission data is separated, and redundant data repeatedly inserted on the transmission side for each subchannel is removed. By repeatedly transmitting transmission data, it is possible to reduce ICI and reduce BER even when there is a frequency offset. Also, when the subchannel of interest for which BER is to be reduced is known, transmission data is repeatedly transmitted a predetermined number of times in the subchannel adjacent to the subchannel, and redundant data repeatedly inserted in the subchannel is received on the receiving side. Remove. Even in this way, the BER of the target subchannel can be reduced. Specifically, the transmission device of the multicarrier communication system repeatedly outputs transmission data for each subchannel a predetermined number of times. A data copying unit, and a frequency multiplexing unit that frequency-multiplexes transmission data of each subchannel and transmits the data, a receiving device that separates transmission data of each subchannel from the received signal, and a data copy on the transmission side A majority voting decoder or decimator for removing the redundant data introduced by the sub-channel for each sub-channel.

FIG. 1 is a multi-carrier communication system using filter bank modulation.
FIG. 2 is an explanatory diagram of subchannel frequency intervals.
FIG. 3 is a subchannel frequency response of a filter bank modulation system (FMT based system) as a system response to a complex exponential function.
FIG. 4 is a subchannel frequency response of a filter bank modulation system as a system response to a random data modulated FMT signal.
FIG. 5 is a subchannel frequency response of a filter bank modulation system as a system response to random data modulated with a repetitive fill bank modulation signal (where the repetition factor is 2).
FIG. 6 is a subchannel frequency response of a filter bank modulation system as a system response to random data modulated with a repetitive fill bank modulation signal (where the repetition factor is 3).
FIG. 7 is a constellation of raw data and received signal transmitted using filter bank modulation and grouping with repetition factor = 2.
FIG. 8 is a BER plot of a filter bank modulation based system with raw data transmission.
FIG. 9 is a BER plot of a filter bank modulation-based system transmitting data with a repetition factor = 3.
FIG. 10 is a BER plot of a filter bank modulation-based system with data transmission with repetition factor = 3 and a receiver with a decoder based on the majority rule.
FIG. 11 is a configuration diagram of a multicarrier communication system that employs repeated transmission of data only in adjacent subchannels.
FIG. 12 is a BER plot of a subchannel of a filter bank modulation-based system in which data is repeatedly transmitted only in adjacent subchannels.
FIG. 13 shows an embodiment equivalent to the communication system of the first embodiment, which is a modified example in which DFT (discrete Fourier transformation) and IDFT (inverse DFT) are adopted.

素信号）であり、それぞれのシンボルレートは１／Ｔで、Ｎ個のサブチャンネル介して並列に入力される。
各サブチャンネルのコピー部１_０〜１_Ｎ−１はそれぞれ、１入力シンボルにつき予め設定されている個数のコピーシンボルを作成して出力し、アップサンプリング部２_０〜２_Ｎ−１は入力シンボルをファクタＫに応じてオーバサンプリングしてフィルタ３_０〜３_Ｎ−１に入力する。各サブチャネルのフィルタ３_０〜３_Ｎ−１は同一のベースバンド特性Ｈ_０（ｆ）を有し、該フィルタ特性に従って入力信号にフィルタリング処理を施す。周波数変換器４_０〜４_Ｎ−１は、各サブチャンネルのフィルタ出力に周波数ｆ_ｉ（ｉ＝０〜Ｎ−１）の信号を乗算して各サブチャネルの信号周波数を互いに異なる周波数にアップ変換する。合成部５は各周波数変換器４_０〜４_Ｎ−１から出力する信号を合成してマルチキャリア信号にして伝送路６に送出する。マルチキャリア信号には伝送中にノイズが重畳して受信装置に到達する。

もできる。すなわち、ユーザデータ列をＮサブチャネルデータに直列並列変換し、サブチャネル毎に、サブチャネルデータ列を１ビットづつ交互に同相成分（Ｉ成分：Ｉｎ−Ｐｈａｓｅ ｃｏｍｐｏｎｅｎｔ）と直交成分（Ｑ成分：Ｑｕａｄｒａｔｕｒｅ ｃｏｍｐｏｎｅｎｔ）に振り分

入力シンボルを発生する場合には、各サブチャネルのコピー部、アップサンプリング部、フィルタをＩ、Ｑ成分用に２系列設け、周波数変換器をＱＰＳＫ直交変調器とする。そして、ＱＰＳＫ直交変調器において、周波数ｆ_ｉ（ｉ＝０〜Ｎ−１）の正弦波、余弦波を各Ｉ、Ｑ成分に乗算して合成することにより直交変調して周波数変換する。
受信装置の解析セクション２０において，受信信号はＮ分岐されて各サブチャネルの周波数変換器１１_０〜１１_Ｎ−１に入力する。周波数変換器１１_０〜１１_Ｎ−１は、受信信号に周波数ｆ_ｉ（ｉ＝０〜Ｎ−１）の信号を乗算して各サブチャネル信号を分離すると共に、各サブチャネル信号の周波数をベースバンド周波数にダウン変換する。各サブチャネルのフィルタ１２_０〜１２_Ｎ−１は同一のベースバンド特性Ｈ_０ ^＊（ｆ）を有し、該フィルタ特性に従って入力信号にフィルタリング処理を施す。受信フィルタセット１２_０〜１２_Ｎ−１は送信装置の合成セクションにおけるフィルタ３_０〜３_Ｎ−１と整合しており、シンボル’^＊’は複素共役を意味している。
ダウンサンプリング部１３_０〜１３_Ｎ−１は入力信号をファクタＫに応じてダウンサンプリングして多数決デコーダあるいはデシメータ１４_０〜１４_Ｎ−１に入力する。多数決デコーダあるいはデシメータ１４_０〜１４_Ｎ−１は送信装置のコピー部で導入された冗長データを多数決によりあるいは所定の方法により、１つのシンボル

サブチャネルの中心周波数ｆ_ｉは
ｆ_ｉ＝（Ｋ／Ｎ・Ｔ）・ｉ，ｉ＝０，１，．．．（Ｎ−１）
であり、連続する２つのサブチャネルの中心周波数間隔はＫ／Ｎ・Ｔとなる。従って、Ｋ＞Ｎのフィルタバンクシステム（ｎｏｎｃｒｉｔｉｃａｌｌｙ ｓａｍｐｌｅｄ ｆｉｌｔｅｒ ｂａｎｋ）において、連続する２つのサブチャネルの中心周波数間隔Ｋ／Ｎ・Ｔは、図２に示すように、シンボルレート１／Ｔより大きくなり、サブチャンネルの帯域幅もシンボルレート１／Ｔより大きくなる。なお、Ｋ＝Ｎのフィルタバンクシステム（ｃｒｉｔｉｃａｌｌｙ ｓａｍｐｌｅｄ ｆｉｌｔｅｒ ｂａｎｋ）において、連続する２つのサブチャネルの中心周波数間隔Ｋ／Ｎ・Ｔはシンボルレート１／Ｔに等しくなる。
（ｂ）本発明のマルチキャリアシステムのパフォーマンス
マルチキャリアシステムのパフォーマンスを、図１のブロック図を参照して説明する。ただし、Ｎ＝６４であり、かつ、Ｈ_０（ｆ）は、超過帯域幅α_０＝０．１２５、パルス応答長γ＝１０シンボルのナイキストフィルタ（ｓｑｕａｒｅ−ｒｏｏｔ ｒａｉｓｅｄ Ｎｙｑｕｉｓｔ ｆｉｌｔｅｒ）であり、アップサンプリングファクタはＫ＝（１＋α_０）・Ｍであるとする。Ｈ_０（ｆ）は、次式

で与えられる。
図３は、複素指数関数で表現される入力信号に対するシステム応答であるＦＭＴベースシステムのサブチャネル周波数応答を示し、縦軸は振幅特性（ｄＢ）、横軸はｆＴ／Ｎである。ＦＭＴ変調により達成されるサブチャンネル間の低いスペクトルの重なりと、高レベルのサブチャネルスペクトル抑圧は、周波数オフセットや局部発振器の位相ノイズによるＩＣＩを緩和するキーであり、これによりＢＥＲパフォーマンスを改善したりシステム動作を簡単化する。
図３に示される周波数応答は、複素指数信号が入力したときにシステム応答として得られた。ＱＰＳＫ変調を使用したランダムデータ送信において、ＦＭＴベースシステムのサブチャネル信号は、送受信装置内部の過渡プロセスに起因する振幅変調（ＡＭ）成分を有する。この振幅変調（ＡＭ）は、サブチャネルで送信されるデータが長い単一符号情報シンボル列であるならば、すなわち、送信側フィルタの応答長γより大きな長さを有する単一符号情報シンボル列であるならば、絶対に発生しない。しかし、現実の状況では、サブチャネルデータはランダムに送信される。そのようなランダム送信状況下のＦＭＴベースシステムにおいて、１つのサブチャネルで送信される信号はもはや複素指数関数であるとみなすことはできない。替わって、入力信号は、図３に示すサブチャネル周波数応答に整合するスペクトル持つように考慮されねばならない。
複素指数関数とは異なった入力信号に対するＦＭＴベースシステムの周波数応答は、図４に示す自己畳み込みサブチャネルフィルタ周波数応答となる。この図４はランダム変調されたＦＭＴ信号に対するシステム応答として得られた連続サブチャネルの周波数応答である。
図４では、隣接サブチャネルのスペクトルはかなり重なっている。すなわち、サイドローブレベルは６０ｄＢ以上抑圧されているけれど、メインローブは広がり、このメインローブ同士のスペクトル重なりによりＩＣＩが現れる。このＩＣＩレベルは隣接サブチャネル間で重なっているスペクトルの面積に依存し、ＩＣＩは各サブキャリアを異なった態様で、ガウシャンノイズによる妨害と同じように妨害する。このようなノイズと同様の妨害は、ＦＭＴベースシステムのＢＥＲを劣化する。
望ましくないＡＭを減少させる解決策は、送信サブチャネルフィルタ内の過渡プロセス数を最小化することである。この送信サブチャネルフィルタ内の過渡プロセス数の最小化は、データの符号が変化する位置の数を減少させることと同じである。これを実現する最も容易な方法は、繰り返しデータ送信手法を採用することである。すなわち、データを連続的に繰り返し送信し、且つ、繰り返しグループ内の全データの符号を同一にする。例えば、データの２回繰り返し（デュープレット（ｄｕｐｌｅｔｓ））、３回繰り返し（トリプレット（ｔｒｉｐｌｅｔｓ））、４回繰り返し（クアドルプレット（ｑｕａｄｒｕｐｌｅｔｓ））等によりグループ化して送信し、各繰り返しグループ内の全データの符号を同一にする。
図５及び図６は繰り返しファクタが２（デュープレット）及び３（トリプレット）のＦＭＴサブチャネル周波数応答をそれぞれ示している。図２は特別のケースとして特にγを越える、非常に長い繰り返しファクタを有するＦＭＴベースシステムのサブチャネル周波数応答と見なすことができる。
図５及び図６より明らかなように、隣接サブチャネル間のスペクトルの重なりに起因するＩＣＩのレベルは、繰り返しファクタの増大に従って減少している。図４では着目チャネルの中心周波数における隣接チャネルの振幅は−３０ｄＢであるが、図５の繰り返しファクタが２の場合には−３９ｄＢとなり、図６の繰り返しファクタが３の場合には−４３ｄＢとなり減少している。
図７は、生データ（繰り返し無し）の受信信号のコンステレーション及び繰り返しファクタ＝２のグループ化を行ってＦＭＴ変調された受信信号のコンステレーションを示す。なお、それぞれのサブチャネル周波数応答は図４及び図５に示されている。各コンステレーション上の濃い暗点部（ダークポイント）は周波数オフセットが零の場合にプロットされる位置であり、参照のためにそれぞれの図中に示されている。また、濃い暗点部周辺の薄い暗点部はＩＣＩの影響で該ＩＣＩのレベルに応じて濃い暗点部から離れてプロットされていることを示す。すなわち、プロット点が濃い暗点位置から離れれば離れるほどＩＣＩの影響が大きいことを示している。又、αは周波数オフセットが０の場合における隣接チャネル間の帯域幅を基準値とし、該基準値と実際の帯域幅の差を基準値で除算した値である。
図７より明らかなように繰り返しファクタ＝２のグループ化がなされた受信信号の方が、薄い暗点部の広がりは小さく、ＩＣＩの影響が小さいことが判る。又、αが大きいほど帯域幅が広がってスペクトルの重なりが大きくなり、この結果、着目チャネルの中心周波数における隣接チャネルの振幅レベルが大きくなってＩＣＩが大きくなることがわかる。
図８及び図９はＱＰＳＫ変調によるサブチャネル数Ｎ＝６４のＦＭＴベースシステムのＢＥＲシミューレーション結果を示す。すなわち、図８は、繰り返し無しの生データ送信において、周波数オフセットをパラメータとしたＢＥＲシミューレーション結果であり、図９は繰り返しあり（繰り返しファクタ＝３）の送信において、周波数オフセットをパラメータとしたＢＥＲシミューレーション結果である。図８〜図９においてＤｅｌｋと記された周波数オフセット値は、周波数オフセットのサブチャネル間隔に対する正規化値である。図９のシミューレーションでは、ファクタ＝３の繰り返しが送信装置のＦＭＴシステムにおいて選択され、また、ファクタ＝３の簡単なデシメーションが受信装置側で用いられる。図８〜図９のシミューレーション結果は、周波数オフセットが小さいほどＢＥＲが良好であることを示している。又、このシミューレーション結果は、周波数オフセットが存在するＦＭＴシステムにおけるＢＥＲパフォーマンスが、データの繰り返しによりかなり改善されていることを示している。
データの繰り返しにより、送信データに冗長が付加される。この冗長は、ＢＥＲパフォーマンスを改善するために使用され（冗長データはエラー訂正コードとして動作する）、同時に、周波数オフセットが存在する際のＢＥＲパフォーマンスを改善するために使用される。冗長データは周知の繰り返しコードを表わしている。受信側では冗長データを除去する必要があり、このため、本発明では、多数決ルールに基いた繰り返しコードのデコードを、サブチャネルにおいて実施する。
図１０は繰り返しファクタ＝３のデータ伝送で、かつ、受信装置に多数決ルールに基いたデコーダを備えた送受信において、周波数オフセットをパラメータとしたＢＥＲシミューレーション結果である。このシミューレーション結果より、多数決ルールに基いたデコーダを設けることによりＢＥＲが改善されていることが判る（図９、図１０を比較）。
（Ｂ）第２実施例のマルチキャリア通信システム
図１１は第２実施例のマルチキャリア通信システムの構成図であり、図１の第１実施例と同一部分には同一符号を付している。異なる点は、データコピーを着目チャネル（Ｍチャネル）に隣接する２つのチャネル（Ｍ−１チャネル、Ｍ＋１チャネル）でのみ行っている点である。送信装置において全サブチャネルでコピーを実施する必要はない。コピーは、ＩＣＩの改善が要求される着目サブチャネルに隣接する２つのサブチャネルで実施しなければならない。そのため、もし中間のサブチャネルが第Ｍサブチャネルであるとすれば、第Ｍ−１サブチャネルと第Ｍ＋１サブチャネルにおけるコピーの実施が第ＭサブチャネルのＩＣＩレベルを改善し、それゆえ、ＢＥＲを改善する。
図１２は、隣接サブチャネルにおいて、▲１▼データ繰り返しが無いときの着目サブチャネルのＢＥＲ（１Ｃｈａｎと記す）と、▲２▼隣接サブチャネルにおいてファクタ＝３のデータ繰り返しをしたときの着目サブチャネルのＢＥＲ（１Ｃｈａｎ ３と記す）と、▲３▼ファクタ＝５のデータ繰り返しをしたときの着目サブチャネルのＢＥＲ（１Ｃｈａｎ ５と記す）を示す。明らかなように、隣接サブチャネルのみのデータ繰り返しであっても、ＢＥＲパフォーマンスを改善し、繰り返しデータは着目サブチャネルに対して従来のエラー訂正コードのように動作する。
（Ｃ）第３実施例
図１３は第１実施例のマルチキャリア通信システムと等価な実施態様であり、ＤＦＴ（ｄｉｓｃｒｅｔｅ Ｆｏｕｒｉｅｒ ｔｒａｎｓｆｏｒｍａｔｉｏｎ）とＩＤＦＴ（ｉｎｖｅｒｓｅ ＤＦＴ）を採用し、かつ、Ｋ＝Ｎとした場合である。第３実施例において複素指数処理はＤＦＴ１６とＩＤＦＴ７に集約されている。又、送信装置のフィルタ３’_０，３’_１，．．．，３’_Ｎ−１の特性Ｈ^（０）（ｆ），Ｈ^（１）（ｆ），．．．Ｈ^{（Ｍ−１）}（ｆ）は、プロトタイプのフィルタ特性Ｈ_０（ｆ）を周波数シフトした特性を有している。すなわち、Ｈ^（ｉ）（ｆ）は次式

で表現される。又、受信装置のフィルタ１２’_０，１２’_１，．．．，１２’_Ｎ−１の特性Ｈ^（０）＊（ｆ），Ｈ^（１）＊（ｆ），．．．Ｈ^{（Ｍ−１）＊}（ｆ）は、特性Ｈ^（０）（ｆ），Ｈ^（１）（ｆ），．．．Ｈ^{（Ｍ−１）}（ｆ）と複素共役の関係にある。
送信装置において、各サブチャネルのコピー部１_０〜１_Ｎ−１はそれぞれシンボル

のコピーシンボルを作成し、サブキャリア信号としてＩＦＤＴ部７に入力する。ＩＦＤＴ部７は、Ｎ個のサブキャリア信号に対してＩＤＦＴ処理を行い、時間に関して離散的なデータ_ａｋ ^（０），_ａｋ ^（１），．．．_ａｋ ^{（Ｍ−１）}を出力する。フィルタ３’_０，３’_１，．．．，３’_Ｎ−１は上記特性を備えており、ＩＤＦＴ部７から出力する時間離散信号に該フィルタ特性を作用させて出力する。並列直列変換部８は各フィルタ３’_０，３’_１，．．．，３’_Ｎ−１から入力する信号を並列直列変換により合成して伝送速度Ｎ／Ｔで伝送路６に送出する。
受信装置において、Ｎ／Ｔの速度でサンプルされた受信信号は直列並列変換器１５において直列並列変換される。受信装置のＭ個のフィルタ１２’_０，１２’_１，．．．，１２’_Ｎ−１は、ＤＦＴ１６で最終的に処理する前に各サンプル信号に所定のフィルタ特性を作用させる。ＤＦＴ１６は各フィルタ１２’_０，１２’_１，．．．，１２’_Ｎ−１から入力される信号に対してＤＦＴ処理を施して、Ｎ個のサブキャリア信号を発生し、各サブキャリア信号を多数決デコーダあるいはデシメータ１４_０，１４_１，．．．，１４_Ｎ−１に入力する。多数決デコーダあるいはデシメータ１４_０〜１４_Ｎ−１は送信装置のコピー部で導入された冗長データを多数決によりあるいは所定の方法により、１つのシンボル

以上、要約すれば、データの繰り返しを備えた送信は、ＦＭＴ又は類似のフィルタバンク変調技術と自然な方法で結合される。その結合システムは繰り返しコードのコードゲインと関連して周波数オフセットに対して免疫を示す。
なお、実施例において、データの繰り返しによりＢＥＲは改善するが、伝送速度が遅くなる。このため、許容される伝送速度とＢＥＲの改善率とのトレードオフにより適切な繰り返し数が決定される。(A) First Embodiment (a) Multi-carrier Communication System of First Embodiment FIG. 1 is a block diagram of a multi-carrier communication system having an N subchannel filter bank configuration (FMT base configuration). A synthesis section 20 is an analysis section in the receiving apparatus.

Each symbol rate is 1 / T and is input in parallel through N subchannels.
Each sub-channel copy unit 1 ₀ to 1 _N-1 generates and outputs a preset number of copy symbols for each input symbol, and up-sampling units 2 ₀ to 2 _N-1 receive input symbols. and oversampling is input to the filter ₃ 0 _{~3 N-1} in accordance with the factor K. Each of the sub-channel filters 3 _{0 to} 3 _N−1 has the same baseband characteristic H ₀ (f), and performs a filtering process on the input signal according to the filter characteristic. Frequency converters 4 ₀ to 4 _N−1 multiply the filter output of each subchannel by a signal of frequency f _i (i = ₀ to _N−1 ), and up-convert the signal frequency of each subchannel to a different frequency. To do. The synthesizer 5 synthesizes the signals output from the frequency converters 4 ₀ to 4 _N-1 and sends them to the transmission line 6 as multicarrier signals. Noise is superimposed on the multicarrier signal during transmission and reaches the receiver.

You can also. That is, the user data sequence is serially parallel-converted into N subchannel data, and the subchannel data sequence is alternately changed by 1 bit for each subchannel, in-phase component (I component: In-Phase component) and quadrature component (Q component: Quadrature). distributed to component)

When generating an input symbol, two sub-channel copy units, up-sampling units, and filters are provided for I and Q components, and the frequency converter is a QPSK quadrature modulator. Then, in the QPSK quadrature modulator, the I and Q components are multiplied and synthesized by a sine wave and a cosine wave having a frequency f _i (i = 0 to N−1) to perform quadrature modulation and frequency conversion.
In the analysis section 20 of the receiving device, the received signal is branched into _N and input to the frequency converters 11 _{0 to} 11 _N−1 of each subchannel. The frequency converters 11 _{0 to} 11 _N−1 multiply the received signal by a signal having a frequency f _i (i = ₀ to _N−1 ) to separate each subchannel signal, and based on the frequency of each subchannel signal. Down-convert to band frequency. The filters 12 _{0 to} 12 _N-1 of each subchannel have the same baseband characteristic H ₀ ^* (f), and perform filtering processing on the input signal according to the filter characteristic. The reception filter sets 12 _{0 to} 12 _N−1 are matched with the filters 3 _{0 to} 3 _N−1 in the synthesis section of the transmission device, and the symbol “ ^* ” means complex conjugate.
The downsampling units 13 _{0 to} 13 _N-1 downsample the input signal according to the factor K and input the result to the majority decoder or decimator 14 _{0 to} 14 _N-1 . The majority decoder or decimator 14 _{0 to} 14 _N-1 uses the redundant data introduced by the copy unit of the transmission device as one symbol by majority or by a predetermined method.

The center frequency f _i of the subchannel is f _i = (K / N · T) · i, i = 0, 1,. . . (N-1)
The center frequency interval between two consecutive subchannels is K / N · T. Therefore, in a filter bank system with K> N (non-cyclically sampled filter bank), the center frequency interval K / N · T between two consecutive subchannels is larger than the symbol rate 1 / T, as shown in FIG. The bandwidth of the subchannel is also larger than the symbol rate 1 / T. Note that, in a K = N filter bank system, the center frequency interval K / N · T between two consecutive subchannels is equal to the symbol rate 1 / T.
(B) Performance of the multi-carrier system of the present invention The performance of the multi-carrier system will be described with reference to the block diagram of FIG. However, N = 64 and H ₀ (f) is a Nyquist filter (square-root raised Nyquist filter) having an excess bandwidth α ₀ = 0.125 and a pulse response length γ = 10 symbols. It is assumed that the factor is K = (1 + α ₀ ) · M. H ₀ (f) is given by

Given in.
FIG. 3 shows a subchannel frequency response of an FMT-based system, which is a system response to an input signal expressed by a complex exponential function, where the vertical axis represents amplitude characteristics (dB) and the horizontal axis represents fT / N. The low spectral overlap between subchannels achieved by FMT modulation and the high level of subchannel spectral suppression are key to mitigating ICI due to frequency offset and local oscillator phase noise, thereby improving BER performance. Simplify system operation.
The frequency response shown in FIG. 3 was obtained as a system response when a complex exponential signal was input. In random data transmission using QPSK modulation, the subchannel signal of the FMT-based system has an amplitude modulation (AM) component due to a transient process inside the transceiver. This amplitude modulation (AM) is performed if the data transmitted in the subchannel is a long single code information symbol sequence, that is, a single code information symbol sequence having a length larger than the response length γ of the transmitting filter. If there is, it never occurs. However, in real situations, subchannel data is transmitted randomly. In an FMT-based system under such a random transmission situation, the signal transmitted on one subchannel can no longer be considered a complex exponential function. Instead, the input signal must be considered to have a spectrum that matches the subchannel frequency response shown in FIG.
The frequency response of the FMT-based system for an input signal different from the complex exponential function is the self-convolution subchannel filter frequency response shown in FIG. FIG. 4 shows the frequency response of continuous subchannels obtained as a system response to a randomly modulated FMT signal.
In FIG. 4, the spectrum of adjacent subchannels overlaps considerably. That is, the side lobe level is suppressed by 60 dB or more, but the main lobe spreads, and ICI appears due to the spectral overlap between the main lobes. This ICI level depends on the area of the spectrum that overlaps between adjacent subchannels, and ICI blocks each subcarrier in a different manner, as well as Gaussian noise. Such interference, as well as noise, degrades the BER of FMT-based systems.
A solution to reduce undesirable AM is to minimize the number of transient processes in the transmit subchannel filter. Minimizing the number of transient processes in this transmit subchannel filter is the same as reducing the number of positions where the sign of the data changes. The easiest way to achieve this is to employ a repetitive data transmission technique. That is, data is transmitted continuously and repeatedly, and all data in the repetition group have the same sign. For example, data is grouped and transmitted twice (duples), three times (triplets), four times (quadruplets), etc. Use the same data sign.
FIGS. 5 and 6 show FMT subchannel frequency responses with repetition factors of 2 (duplex) and 3 (triplet), respectively. FIG. 2 can be regarded as a subchannel frequency response of an FMT-based system with a very long repetition factor, especially exceeding γ, as a special case.
As is apparent from FIGS. 5 and 6, the level of ICI due to spectral overlap between adjacent subchannels decreases with increasing repetition factor. In FIG. 4, the amplitude of the adjacent channel at the center frequency of the channel of interest is −30 dB. However, when the repetition factor in FIG. 5 is 2, it becomes −39 dB, and when the repetition factor in FIG. is doing.
FIG. 7 shows a constellation of a received signal that has been subjected to FMT modulation by performing a constellation of a reception signal of raw data (without repetition) and a grouping with a repetition factor = 2. Each subchannel frequency response is shown in FIGS. The dark dark spot on each constellation is the position plotted when the frequency offset is zero and is shown in the respective figure for reference. Further, it is shown that the thin dark spot portion around the dark dark spot portion is plotted away from the dark dark spot portion according to the ICI level due to the influence of ICI. In other words, it shows that the effect of ICI increases as the plot point moves away from the dark spot position. Α is a value obtained by dividing the difference between the reference value and the actual bandwidth by the reference value with the bandwidth between adjacent channels when the frequency offset is 0 as the reference value.
As can be seen from FIG. 7, the received signal that has been grouped with a repetition factor of 2 has a smaller spread of the thin dark spot and is less affected by ICI. Further, it can be seen that the larger α is, the wider the bandwidth is, and the spectrum overlap is larger. As a result, the amplitude level of the adjacent channel at the center frequency of the channel of interest increases and ICI increases.
8 and 9 show BER simulation results of an FMT-based system with N = 64 subchannels by QPSK modulation. That is, FIG. 8 shows a BER simulation result with a frequency offset as a parameter in raw data transmission without repetition, and FIG. 9 shows a BER with a frequency offset as a parameter in transmission with repetition (repetition factor = 3). It is a simulation result. 8 to 9, the frequency offset value indicated as Delk is a normalized value for the subchannel interval of the frequency offset. In the simulation of FIG. 9, repetition of factor = 3 is selected in the FMT system of the transmitting device, and simple decimation of factor = 3 is used on the receiving device side. The simulation results in FIGS. 8 to 9 indicate that the smaller the frequency offset, the better the BER. This simulation result also shows that the BER performance in the FMT system where there is a frequency offset is considerably improved by the repetition of data.
Redundancy is added to the transmission data by repeating the data. This redundancy is used to improve BER performance (redundant data acts as an error correction code) and at the same time is used to improve BER performance in the presence of frequency offsets. Redundant data represents a known repetitive code. Redundant data needs to be removed on the receiving side. Therefore, in the present invention, decoding of the repetition code based on the majority rule is performed in the subchannel.
FIG. 10 shows a BER simulation result using a frequency offset as a parameter in transmission / reception in which data is transmitted with repetition factor = 3 and the receiving apparatus includes a decoder based on the majority rule. From this simulation result, it can be seen that the BER is improved by providing a decoder based on the majority rule (compare FIGS. 9 and 10).
(B) Multi-carrier communication system of the second embodiment FIG. 11 is a block diagram of the multi-carrier communication system of the second embodiment. Components identical with those of the first embodiment of FIG. The difference is that data copying is performed only on two channels (M−1 channel, M + 1 channel) adjacent to the channel of interest (M channel). It is not necessary to perform copying on all subchannels in the transmitting device. Copying must be performed on two subchannels adjacent to the subchannel of interest where ICI improvement is required. Therefore, if the intermediate subchannel is the Mth subchannel, the implementation of copying in the M-1 and M + 1 subchannels improves the ICI level of the Mth subchannel, and hence the BER Improve.
FIG. 12 shows (1) the BER of the target subchannel when there is no data repetition in the adjacent subchannel (denoted as 1 Chan), and (2) the target subchannel when data repetition of factor = 3 is performed in the adjacent subchannel. BER (denoted as 1 Chan 3) and {circle around (3)} BER (denoted as 1 Chan 5) of the target sub-channel when data of factor = 5 is repeated. As can be seen, BER performance is improved even with data repetition of only adjacent subchannels, and the repeated data behaves like a conventional error correction code for the target subchannel.
(C) Third Example FIG. 13 is an embodiment equivalent to the multicarrier communication system of the first example, adopting DFT (discrete Fourier transformation) and IDFT (inverse DFT), and K = N Is the case. In the third embodiment, complex exponential processing is integrated into DFT 16 and IDFT 7. Further, the filters 3 ′ ₀ , 3 ′ ₁ ,. . . , 3 ′ _N−1 characteristics H ⁽⁰⁾ (f), H ⁽¹⁾ (f),. . . H ^(M-1) (f) has a characteristic obtained by frequency shifting the prototype filter characteristic H ₀ (f). That is, H ⁽ⁱ⁾ (f) is

It is expressed by The filters 12 ′ ₀ , 12 ′ ₁ ,. . . , 12 ′ _N−1 characteristics H ^{(0) *} (f), H ^{(1) *} (f),. . . H ^{(M-1) *} (f) is the characteristic H ⁽⁰⁾ (f), H ⁽¹⁾ (f),. . . H ^(M-1) (f) and complex conjugate.
In the transmission apparatus, the copy units 1 ₀ to 1 _N−1 of each subchannel are symbols.

Copy symbols are generated and input to the IFDT unit 7 as subcarrier signals. The IFDT unit 7 performs IDFT processing on the N subcarrier signals and performs discrete data _ak ⁽⁰⁾ , _ak ⁽¹⁾ _,. . . _ak ^(M-1) is output. Filters 3 ′ ₀ , 3 ′ ₁ ,. . . , 3 ′ _N−1 have the above _- described characteristics, and outputs the time discrete signals output from the IDFT unit 7 by applying the filter characteristics. The parallel-serial converter 8 includes filters 3 ′ ₀ , 3 ′ ₁ ,. . . , 3 ′ _N−1 are combined by parallel-serial conversion and sent to the transmission line 6 at a transmission rate N / T.
In the receiving apparatus, the received signal sampled at a rate of N / T is serial-parallel converted by the serial-parallel converter 15. M filters 12 ′ ₀ , 12 ′ ₁ ,. . . , 12 ′ _N−1 apply a predetermined filter characteristic to each sample signal before final processing by the DFT 16. The DFT 16 includes filters 12 ' ₀ , 12' ₁ ,. . . , 12 ′ _N−1 is subjected to DFT processing to generate N subcarrier signals, and each subcarrier signal is divided into a majority decoder or decimator 14 ₀ , 14 ₁ ,. . . , 14 _N-1 . The majority decoder or decimator 14 _{0 to} 14 _N-1 uses the redundant data introduced by the copy unit of the transmission device as one symbol by majority or by a predetermined method.

In summary, transmissions with data repetition are combined in a natural way with FMT or similar filter bank modulation techniques. The combined system is immune to frequency offset in relation to the code gain of the repetitive code.
In the embodiment, the BER is improved by repeating the data, but the transmission speed is lowered. For this reason, an appropriate number of repetitions is determined by a trade-off between an allowable transmission rate and a BER improvement rate.

Claims (3)

In a multi-carrier communication system including a transmission device and a reception device,
The transmission device includes, for each subchannel, a data copy unit that repeatedly outputs transmission data for a preset number of times, and a frequency multiplexing unit that frequency-multiplexes and transmits transmission data of each subchannel,
The receiving device includes a data separation unit that separates transmission data of each subchannel from the received signal, and a majority decoder or decimator that removes redundant data introduced by the data copy unit on the transmission side for each subchannel,
The frequency multiplexing unit of the transmission device,
For each subchannel, an upsampling unit that upsamples the output data of the data copy unit, a transmission filter that passes the subchannel signal output from each upsampling unit, and a frequency conversion that increases the frequency of each subchannel signal to a different frequency A combining unit that combines signals output from the frequency conversion unit of each subchannel,
The data separation unit of the receiving device includes:
A branching unit that branches a received signal, a frequency converter that converts the frequency of the branched received signal for each subchannel and outputs transmission data of the subchannel, and a reception that passes through a subchannel signal input from the frequency converter A downsampling unit that downsamples the subchannel signal output from the filter and the reception filter;
A multi-carrier communication system comprising: In a multi-carrier communication system including a transmission device and a reception device,
The transmission device includes, for each subchannel, a data copy unit that repeatedly outputs transmission data for a preset number of times, and a frequency multiplexing unit that frequency-multiplexes and transmits transmission data of each subchannel,
The receiving device includes a data separation unit that separates transmission data of each subchannel from the received signal, and a majority decoder or decimator that removes redundant data introduced by the data copy unit on the transmission side for each subchannel,
The frequency multiplexing unit of the transmission device,
An IDFT unit that performs IDFT processing by inputting output data of a data copy unit provided for each subchannel as a subcarrier signal,
A filter having a characteristic obtained by sequentially frequency-shifting a predetermined filter characteristic, to which a time discrete signal output from the IDFT unit is input,
A conversion unit for parallel-to-serial conversion of signals input from the filters;
The data separation unit of the receiving device includes:
A conversion unit for serial-parallel conversion of the received signal,
A filter having a transmission-side filter characteristic and a complex conjugate characteristic, to which the serial-parallel converted reception signal is input,
A DFT unit that receives an output signal of each filter and inputs a subcarrier signal obtained by performing DFT processing to the majority decoder or decimator,
A multi-carrier communication system comprising: Only in the subchannel adjacent to the subchannel whose BER is to be reduced, the transmitting device includes the data copy unit, and the receiving device includes the majority decoder or decimator.
The multicarrier communication system according to claim 1, wherein the multicarrier communication system is a multicarrier communication system.

Images