JP3839569B2 - Unique word detection method - Google Patents

Description

但し、・は内積演算を示す。
【０００６】
さらに、相関値ＣＯＲは複素相関値の電力で表わされるので、次式の計算も必要である。
【０００７】
【数２】
ＣＯＲ＝Ｃ_I ² ＋Ｃ_Q ² ……（２）
【０００８】
図６に示すように、この２つの式の演算を行なう複素相関演算部１０１は、４個の内積演算回路（ＦＩＲフィルタ）１０２〜１０５と、２個の加算器１０６，１１０と、減算器１０７と、２個の乗算器１０８，１０９が必要になり、回路規模は非常に大きい。
【０００９】
次に、複素相関を図８に示すようなシンボル配置をとるＢＰＳＫ信号に対して行なう場合について説明する。この場合、ユニークワードはＩ成分とＱ成分が等しいので、Ｕ_I ＝Ｕ_Q ＝Ｕ_IQとおけるので（１）式の計算は次式のように簡略なものとなる。
【００１０】
【数３】

【００１１】
（３）式において、Ａ＝Ｕ_IQ・Ｒ_I 、Ｂ＝Ｕ_IQ・Ｒ_Q とおくと相関値電力ＣＯＲは次式のように表わせる。
【００１２】
【数４】

【００１３】
この場合に必要となる回路は図９のようになり、複素相関演算部１２１は、２個の内積演算回路（ＦＩＲフィルタ）１２２，１２３と２個の乗算器１２４，１２５と１個の加算器１２６で構成することができるので、図６に示した方式の半分以下の回路規模ですむ。
【００１４】
このことを利用して、ＱＰＳＫや１６値ＱＡＭ等を用いたデジタル変調方式においても、ユニークワードだけＢＰＳＫを用いることによってユニークワード検出のための回路規模が縮小でき、しかも移動通信のような過酷な伝搬路状況においても充分な性能が得られることが知られている。
【００１５】
【発明が解決しようとする課題】
ところが、π／４シフトＱＰＳＫは図１０に示すスペースダイヤグラム上で○印のポイントにマッピングされたシンボルの系列Ａ，Ｂ，Ｃ，Ｄと●印のポイントにマッピングされたシンボルの系列ａ，ｂ，ｃ，ｄとを交互に繰り返すので、ＢＰＳＫの２点のみを用いると上記の簡略化方式をそのまま採用することができないという問題がある。
【００１６】
この発明はこのような課題を解決するためなされたもので、π／４シフトＱＰＳＫを用いたデジタル通信システムにおいてユニークワード検出回路の構成を簡略化することのできるユニークワード検出方式を提供することを目的とする。
【００１７】
【課題を解決するための手段】
前記課題を解決するためこの発明に係るユニークワード検出方式は、送信側において２ｎシンボルからなるユニークワードを１シンボル毎に相互に切り替え送出されるｎシンボルからなる第１系統ユニークワードと第２系統ユニークワードとの組み合わせで構成し、第１系統ユニークワードをπ／４シフトＱＰＳＫのコンステレーション上における相互に１８０度の位相差を有する１組のＢＰＳＫシンボル点にマッピングし、第２系統ユニークワードを前記コンステレーション上において前記第１系統ユニークワードと４５度または−４５度の角度差で交差するＢＰＳＫシンボル点にマッピングするとともに、受信側において受信ベースバンド信号のＩ相成分の過去２ｎシンボルの系列のうち１シンボルずつ交互にｎシンボル系列に２分した系列と第１系統ユニークワード，前記第２系統ユニークワードのそれぞれとの相関演算を実行しその結果を加算することによりＩ相側の相関値を求め、受信ベースバンド信号のＱ相成分の過去２ｎシンボルの系列のうち１シンボルずつ交互にｎシンボル系列に２分した系列と前記第１系統ユニークワード，第２系統ユニークワードのそれぞれとの相関演算を実行しその結果を加算することによりＱ相側の相関値を求め、Ｉ相側の相関値とＱ相側の相関値とを加算した結果が予め設定したしきい値よりも大きい点をユニークワードタイミング候補として出力することを特徴とする。
【００１８】
この発明に係るユニークワード検出方式は、送信側において２ｎシンボルからなるユニークワードを１シンボル毎に相互に切り替え送出されるｎシンボルからなる第１系統ユニークワードと第２系統ユニークワードとの組み合わせで構成し、前記第１系統ユニークワードをπ／４シフトＱＰＳＫのコンステレーション上における相互に１８０度の位相差を有する１組のＢＰＳＫシンボル点にマッピングし、前記第２系統ユニークワードを前記コンステレーション上において前記第１系統ユニークワードと４５度または−４５度の角度差で交差するＢＰＳＫシンボル点にマッピングするとともに、受信側においてＩ相成分およびＱ相成分それぞれの受信ベースバンド信号の極性符号を検出して１ビットのデータに変換し、Ｉ相成分の過去２ｎビットの系列のうち１ビットずつ交互にｎビット系列に２分した系列と第１系統ユニークワード，第２系統ユニークワードのそれぞれとの排他的論理和演算をそれぞれ実行してＩ相側の相関値を求め、Ｑ相成分の過去２ｎビットの系列のうち１ビットずつ交互にｎビット系列に２分した系列と第１系統ユニークワード，第２系統ユニークワードのそれぞれとの排他的論理和演算をそれぞれ実行してＱ相側の相関値を求め、Ｉ相側の相関値とＱ相側の相関値とを加算した結果が予め設定したしきい値よりも大きい点をユニークワードタイミング候補として出力することを特徴とする。
【００１９】
【発明の実施の形態】
以下、この発明の実施の形態について添付図面に基づいて説明する。本発明によるユニークワードの構成方法を図１に、本発明によるユニークワード検出回路の構成例を図２に、本発明によるユニークワード検出回路の他の構成例を図３に示す。
【００２０】
まず、送信側におけるユニークワードの構成方法を図１を参照に説明する。π／４シフトＱＰＳＫは、図１０に示したコンステレーション（データ点配置）上で○印のシンボル（Ａ，Ｂ，Ｃ，Ｄ）にマッピングされたシンボル系列と●印のシンボル（ａ，ｂ，ｃ，ｄ）にマッピングされたシンボル系列とを交互に繰り返す。従って、ｎシンボル（図１の例ではｎ＝４）からなる第１系統ユニークワードＵＷＡとｎシンボルからなる第２系統ユニークワードＵＷＢとの２つのＢＰＳＫ系列を交互に並べて２ｎシンボルからなるユニークワードＵＷを構成すれば（図８に示したスペースダイヤグラム上で、○印のシンボル（Ａ，Ｃ）に第１系統ユニークワードＵＷＡをマッピングし、●印のシンボル（ａ，ｃ）に第２系統ユニークワードＵＷＢをマッピングし、点線で示す軌跡ｘを通過するようにする。
【００２１】
例えば、ユニークワードＵＷが、第１系統ユニークワードＵＷＡ：１１００、第２系統ユニークワードＵＷＢ：１０１０で構成される場合、１を（Ａとａ）に割り当て、０を（Ｃとｃ）に割り当てると、ユニークワードＵＷは”ＡａＡｃＣａＣｃ”のようにマッピングされ、その軌跡は図１０で点線で示す軌跡ｘのように見えることになる。
【００２２】
実際の回路構成は従来と同様で、ユニークワードＵＷを上記のようなポイントにマッピングされるように与えるだけでよい。
【００２３】
次に、受信側での処理について説明する。図２はこの発明によるユニークワード検出回路のブロック構成図である。ユニークワードＵＷの長さは適用するデジタル通信システムによって様々で、一般に１０〜２０数シンボルで構成されるが、簡単のために以下の説明はユニークワードＵＷの長さが８シンボル（すなわちｎ＝４）の場合について記述する。
【００２４】
本発明では、受信側で直交検波して得られる受信ベースバンド信号のＩ相成分，Ｑ相成分のそれぞれをシンボルタイミングでサンプリングしたデータをユニークワード検出回路１の入力信号として、１シンボル毎にユニークワードを検出するための処理を行なう。
【００２５】
まず、Ｉ相成分の処理について説明する。ユニークワード候補としてシンボルタイミングでサンプリングされた受信ベースバンド信号のＩ相成分の過去２ｎシンボルずつ切り出したものＲ_i （ｔ）：ｔ＝０，１，２，…，７を１シンボルずつ交互にｎシンボルに２分した（Ｒ_i（ｔ）：ｔ＝０，２，４，６），（Ｒ_i（ｔ）：ｔ＝１，３，５，７）と第１系統ユニークワードＵＷＡ，第２系統ユニークワードＵＷＢとの相関値をそれぞれ計算する。
【００２６】
２つのＢＰＳＫ系列ＵＷＡ，ＵＷＢで構成されるユニークワードＵＷ（Ｉ成分＝Ｑ成分）が次式で表わされるとき、
【数５】
ＵＷ＝［ＵＷＡ（０），ＵＷＢ（０），ＵＷＡ（１），ＵＷＢ（１），ＵＷＡ（２），ＵＷＡ（３），ＵＷＢ（３）］ ……（５）
但し、ＵＷＡ＝［ＵＷＡ（０），ＵＷＡ（１），ＵＷＡ（２），ＵＷＡ（３）］
ＵＷＢ＝［ＵＷＢ（０），ＵＷＢ（１），ＵＷＢ（２），ＵＷＢ（３）］
【００２７】
ある時点におけるＩ相成分の相関値Ｃ_I は次式で表わされる。
【数６】
Ｃ_I ＝（ＵＷＡ（０）Ｒ_i（０）＋ＵＷＡ（１）Ｒ_i（２）＋ＵＷＡ（２）Ｒ_i（４）＋ＵＷＡ（３）Ｒ_i（６））² ＋（ＵＷＢ（０）Ｒ_i（１）＋ＵＷＢ（１）Ｒ_i（３）＋ＵＷＢ（２）Ｒ_i（５）＋ＵＷＢ（３）Ｒ_i（７））² …（６）
【００２８】
この計算は２つのＦＩＲフィルタの出力を２乗して加算することに他ならない。図２に示したＩ相成分内積演算回路２にあてはめると、ＵＷＡレジスタ３の中身とＲｉ（ｔ）：ｔ＝０，２，４，６との乗累算が１つめのＦＩＲフィルタを構成し、その結果の２乗がＣ_IAである。一方、ＵＷＢレジスタ４の中身とＲ_i（ｔ）：ｔ＝１，３，５，７との乗累算が２つめのＦＩＲフィルタを構成し、その結果の２乗がＣ_IBである。つまり（６）式のはじめの２乗の項がＣ_IAに相当し、あとの２乗の項がＣ_IBに相当する。Ｉ相成分の相関値Ｃ_I の求め方は以上である。
【００２９】
なお、図２に示したＩ相成分内積演算回路２は、ＵＷＡレジスタ３と、ＵＷＢレジスタ４と、直列接続された７個の遅延回路５ａ〜５ｇと、Ｒ_i（７）とＵＷＢ（３）との乗算を行なう乗算器６ａと、Ｒ_i（６）とＵＷＡ（３）との乗算を行なう乗算器６ｂと、Ｒ_i（５） とＵＷＢ（２）との乗算を行なう乗算器６ｃと、Ｒ_i（４）とＵＷＡ（２）との乗算を行なう乗算器６ｄと、Ｒ_i（５）とＵＷＢ（２）との乗算を行なう乗算器６ｃと、Ｒ_i（４）とＵＷＡ（２）との乗算を行なう乗算器６ｄと、Ｒ_i（３）とＵＷＢ（２）との乗算を行なう乗算器６ｅと、Ｒ_i（３）とＵＷＢ（１）との乗算を行なう乗算器６ｅと、Ｒ_i（２）とＵＷＡ（１）との乗算を行なう乗算器６ｆと、Ｒ_i（１）とＵＷＢ（０）との乗算を行なう乗算器６ｇと、Ｒ_i（０）とＵＷＡ（０）との乗算を行なう乗算器６ｈと、４個の乗算器６ｂ，６ｄ，６ｆ，６ｈの各乗算結果を加算（累算）する加算器７ａと、他の４個の乗算器６ａ，６ｃ，６ｅ，６ｇの各乗算結果を加算（累算）する加算器７ｂと、加算器７ａの加算結果を２乗する乗算器８ａと、加算器７ｂの加算結果を２乗する乗算器８ｂと、各乗算器８ａ，８ｂの乗算結果を加算する加算器９とから構成している。
【００３０】
各遅延回路５ａ〜５ｇは１シンボル分の時間を遅延させる。各加算器６ｂ，６ｄ，６ｆ，６ｈと加算器７ａとで前述の１つめのＦＩＲフィルタを構成している。各加算器６ａ，６ｃ，６ｅ，６ｇと加算器７ｂとで前述の２つめのＦＩＲフィルタを構成している。
【００３１】
同時にＱ相成分に対する処理は、Ｑ相成分内積演算回路１０においてＩ相成分と全く同様に行なわれ、次式に示すようにＱ相成分の相関値Ｃ_Q が求められる。
【００３２】
【数７】
Ｃ_Q ＝（ＵＷＡ（０）Ｒ_q（０）＋ＵＷＡ（１）Ｒ_q（２）＋ＵＷＡ（２）Ｒ_q（４）＋ＵＷＡ（３）Ｒ_q（６））² ＋（ＵＷＢ（０）Ｒ_q（１）＋ＵＷＢ（１）Ｒ_q（３）＋ＵＷＢ（２）Ｒ_q（５）＋ＵＷＢ（３）Ｒ_q（７））² …（７）
【００３３】
このようにして求められたＩ相成分の相関値Ｃ_IとＱ相成分の相関値Ｃ_Q とを加算器１１で加算した結果が相関値Ｃ_IQとして相関値ピーク検出回路１２へ供給される。相関値ピーク検出回路１２では従来方式同様に相関値が設定したしきい値を越えた点をユニークワードタイミングとみなしユニークワードタイミング候補として出力する。
【００３４】
次に、上記方式をさらに簡略化した方式を図３を参照して説明する。図３に示す方式では、図２に示した方式における相関演算の演算語長を１ビットに制限することによって回路規模のさらなる縮小を図るものである。
【００３５】
図３に示すユニークワード検出回路２１は、Ｉ相側符号検出回路２２と、Ｉ相成分ＥＸＯＲ演算回路３０と、Ｑ相側符号検出回路２３と、Ｑ相成分ＥＸＯＲ演算回路４０と、加算器２４と、相関値ピーク検出回路２５とからなる。
【００３６】
Ｉ相成分ＥＸＯＲ演算回路３０は、ＵＷＡレジスタ３１と、ＵＷＢレジスタ３２と、７個の遅延回路３３ａ〜３３ｇと、８個の乗算器３４ａ〜３４ｈと、２個の１ビット相関演算回路３５，３６と、加算器３７とを備える。
【００３７】
まずシンボルタイミングでサンプリングされた受信ベースバンド信号は０を中心に正負の値をとるものとする。このときシンボルタイミングでサンプリングされた受信ベースバンド信号のＩＱ成分それぞれを符号検出回路２２，２３にて信号の極性符号に対応したデータ｛１，０｝の１ビットデータへ変換する。ここでは、負の信号のときは１、正の信号のときは０のように出力するものとするが、逆に規定しても全体の処理としては等化になるのでかまわない。
【００３８】
次に、各符号検出回路２２，２３から出力される１ビットデータの過去２ｎビットずつを切り出したユニークワード候補のＩ相成分（Ｒ_i（ｔ）：ｔ＝０，１，２，…，７）およびＱ相成分（Ｒ_q（ｔ）：ｔ＝０，１，２，…，７）との相関値をそれぞれ計算する。
【００３９】
図２に示した方式ではここで乗累算により相関値を計算するのだが、図３に示した方式では符号同士の乗算だけを行なえばよいので、乗算をＥＸＯＲ（排他的論理和）演算で置き換えることができる。ＥＸＯＲ演算と実数乗算の等価性を図４に示す。
【００４０】
まずユニークワード候補のＩ成分Ｒ_i（ｔ）：ｔ＝０，１，２，…，７に対する処理について説明する。過去２ｎビットからなるＲ_i（ｔ）：ｔ＝０，１，２，…，７を１シンボルずつ交互にｎシンボルに２分したうちのＲ_i（ｎ）：ｎ＝０，２，４，６のデータはＵＷＡレジスタ３１の中身とそれぞれのビットのＥＸＯＲをとり、その結果を一方の１ビット相関演算回路３５へ供給する。ｎシンボルに２分した残りのＲ_i（ｎ）：ｎ＝１，３，５，７のデータはＵＷＢレジスタ３２の中身とそれぞれのビットのＥＸＯＲをとり、その結果を他方の１ビット相関演算回路３６へ供給する。なお、ＵＷＡレジスタ３１とＵＷＢレジスタ３２の中身もそれぞれが｛０，１｝の１ビットのデータになっているものとする。
【００４１】
各１ビット相関演算回路３５，３６では、ＥＸＯＲ演算の結果が”１”の時は”−１”、”０”のときは”＋１”として、すべてのＥＸＯＲ演算結果を加算した後に絶対値をとる。この計算はＲＯＭなどを用いて構成すれば簡単である。例えば、図３の例の場合ならば、図５に示すような１ビット相関演算用テーブルを作成し、ＲＯＭなどの記憶装置に書き込んでおき、入力に応じた出力が得られるようにしておけばよい。以上の操作により、Ｉ成分の相関値Ｃ_I が求められる。
【００４２】
前述のＩ相成分に対する処理と平行して、ユニークワード候補のＱ成分Ｒ_q （ｎ）：ｎ＝０，１，２，…，７に対する処理が行なわれる。Ｑ相成分ＥＸＯＲ演算回路４０においてＩ成分と全く同様な手順によりＱ成分の相関値Ｃ_Q が求められる。
【００４３】
このようにして求められたＩ成分の相関値Ｃ_I とＱ成分の相関値Ｃ_Q を加算器２４で加算した結果が相関値Ｃ_IQとして相関値ピーク検出回路２５へ供給される。相関値ピーク検出回路２５では、従来方式同様に相関値が設定したしきい値を越えた点をユニークワードタイミングとみなしユニークワード候補として出力する。
【００４４】
以上述べたように、図３に示す方式によれば乗算器なしでユニークワード検出が可能となる。
【００４５】
【発明の効果】
以上説明したようにこの発明に係るユニークワード検出方式を実施することにより回路規模の小さなユニークワード検出回路を実現することができる。
【００４６】
また、ユニークワード検出回路の演算語長を１ビットに制限することにより乗算器が全く不要になり一層回路規模の小さなユニークワード検出回路を実現することができる。
【図面の簡単な説明】
【図１】本発明によるユニークワードの構成方法の説明図である。
【図２】本発明によるユニークワード検出回路のブロック構成図である。
【図３】本発明による他のユニークワード検出回路のブロック構成図である。
【図４】ＥＸＯＲと実数演算の等価性を示す説明図である。
【図５】１ビット相関演算用テーブルの一例を示す説明図である。
【図６】従来のユニークワード検出回路のブロック構成図である。
【図７】受信系列と相関値の関係を示す説明図である。
【図８】ＢＰＳＫのスペースダイヤグラムである。
【図９】従来の他のユニークワード検出回路のブロック構成図である。
【図１０】π／４シフトＱＰＳＫのスペースダイヤグラムである。
【符号の説明】
１，２１ ユニークワード検出回路
２ Ｉ相成分内積演算回路
１０ Ｑ相成分内積演算回路
１２，２５ 相関値ピーク検出回路
２２ Ｉ相側符号検出回路
２３ Ｑ相側符号検出回路
３０ Ｉ相成分ＥＸＯＲ演算回路
４０ Ｑ相成分ＥＸＯＲ演算回路[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a unique word detection method in a digital communication system using π / 4 shift QPSK as a modulation method.
[0002]
[Prior art]
A configuration example of the unique word detection circuit in the conventional digital communication is shown in FIG. 6, and the relationship between the reception sequence and the correlation value is shown in FIG. As a unique word detection method, as shown in FIG. 6, a method in which a complex correlation between a received sequence and a unique word and a peak of a correlation value is detected is used as a unique word timing candidate. Is the method. In general, since a sequence having excellent correlation characteristics (for example, M sequence) is used for the unique word, the correlation value of the unique word is clearly larger than the correlation value of the data portion. Therefore, as shown in FIG. 7, the point where the correlation value exceeds the set threshold value can be regarded as a unique word timing candidate.
[0003]
Here, it is described as a unique word timing candidate instead of a unique word timing. In a regular communication channel, the unique word timing is almost equal to the unique word timing candidate, but it is transmitted in a harsh environment such as mobile communication. This is because the correlation value may increase even at times other than the unique word timing due to the influence of path fluctuations, or the correlation value may not increase at the unique word timing, and the two do not necessarily match. However, the unique word timing can be easily extracted by applying a known synchronization protection technique based on the unique word timing candidates.
[0004]
However, the circuit required for complex correlation is large. For example, the unique word candidate complex series R = R _I + jR _Q cut from the received signal as, if the unique word detecting and U = U _I + jU _Q, the complex correlation value C = C _I + jC _Q represented by the following formula .
[0005]
[Expression 1]

However, · indicates an inner product operation.
[0006]
Further, since the correlation value COR is represented by the power of the complex correlation value, it is also necessary to calculate the following equation.
[0007]
[Expression 2]
COR = C _I ² + C _Q ² (2)
[0008]
As shown in FIG. 6, the complex correlation calculation unit 101 that performs calculation of these two equations includes four inner product calculation circuits (FIR filters) 102 to 105, two adders 106 and 110, and a subtractor 107. The two multipliers 108 and 109 are required, and the circuit scale is very large.
[0009]
Next, the case where complex correlation is performed on a BPSK signal having a symbol arrangement as shown in FIG. 8 will be described. In this case, since the I component and the Q component of the unique word are equal, U _I = U _Q = U _IQ , so the calculation of equation (1) is simplified as the following equation.
[0010]
[Equation 3]

[0011]
In the equation (3), if A = U _IQ · R _I and B = U _IQ · R _Q , the correlation value power COR can be expressed as the following equation.
[0012]
[Expression 4]

[0013]
The circuit required in this case is as shown in FIG. 9, and the complex correlation calculation unit 121 includes two inner product calculation circuits (FIR filters) 122 and 123, two multipliers 124 and 125, and one adder. 126, the circuit scale can be less than half that of the system shown in FIG.
[0014]
By utilizing this, even in a digital modulation system using QPSK, 16-value QAM, etc., the circuit scale for detecting a unique word can be reduced by using only BPSK for a unique word, and it is also a severe condition such as mobile communication. It is known that sufficient performance can be obtained even in propagation path conditions.
[0015]
[Problems to be solved by the invention]
However, the π / 4 shift QPSK is a series of symbols A, B, C, D mapped to the points marked with ○ on the space diagram shown in FIG. 10 and a series of symbols a, b, Since c and d are alternately repeated, if only two points of BPSK are used, there is a problem that the above simplified method cannot be employed as it is.
[0016]
The present invention has been made to solve such a problem, and provides a unique word detection system capable of simplifying the configuration of a unique word detection circuit in a digital communication system using π / 4 shift QPSK. Objective.
[0017]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, the unique word detection method according to the present invention provides a first-system unique word and a second-system unique word composed of n symbols which are transmitted and switched between each unique symbol composed of 2n symbols on the transmission side. The first system unique word is mapped to a set of BPSK symbol points having a phase difference of 180 degrees on the π / 4 shift QPSK constellation, and the second system unique word is The constellation is mapped to a BPSK symbol point that intersects with the first system unique word at an angle difference of 45 degrees or -45 degrees, and on the receiving side, among the past 2n symbol sequences of the I-phase component of the received baseband signal Divided into n-symbol series alternately every 1 symbol A correlation operation is performed between the column and each of the first system unique word and the second system unique word, and the result is added to obtain a correlation value on the I phase side, and the past 2n of the Q phase components of the received baseband signal The Q-phase side is obtained by performing a correlation operation between a series of symbols alternately divided into n-symbols for each symbol and the first system unique word and the second system unique word and adding the results. And a point where the result of adding the correlation value on the I-phase side and the correlation value on the Q-phase side is larger than a preset threshold value is output as a unique word timing candidate.
[0018]
The unique word detection method according to the present invention is composed of a combination of a first system unique word composed of n symbols and a second system unique word composed of 2 symbols that are switched and sent to each other on a transmission side. The first system unique word is mapped to a set of BPSK symbol points having a phase difference of 180 degrees from each other on the π / 4 shift QPSK constellation, and the second system unique word is mapped to the constellation. Mapping to the BPSK symbol point intersecting with the first system unique word at an angle difference of 45 degrees or -45 degrees, and detecting the polarity code of the received baseband signal of each of the I-phase component and the Q-phase component on the receiving side Converted to 1-bit data and the past 2n I-phase components I-phase side correlation value by executing exclusive OR operation on the 1st system unique word and the 2nd system unique word, respectively, and the 1st system unique word and each of the 2nd system unique word. And the exclusive OR operation of each of the 1st system unique word and the 2nd system unique word with each of the 1st system unique word and the 2nd system unique word, each of which is alternately divided into n bits in the past 2n bits of the Q phase component. To obtain a correlation value on the Q-phase side and output a point where the result of adding the correlation value on the I-phase side and the correlation value on the Q-phase side is larger than a preset threshold value as a unique word timing candidate It is characterized by.
[0019]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the accompanying drawings. A unique word configuration method according to the present invention is shown in FIG. 1, a configuration example of a unique word detection circuit according to the present invention is shown in FIG. 2, and another configuration example of a unique word detection circuit according to the present invention is shown in FIG.
[0020]
First, a method of forming a unique word on the transmission side will be described with reference to FIG. The π / 4 shift QPSK is a sequence of symbols mapped to symbols (A, B, C, D) marked with ○ on the constellation (data point arrangement) shown in FIG. 10 and symbols (a, b, The symbol series mapped to c, d) are repeated alternately. Accordingly, the unique word UW consisting of 2n symbols by alternately arranging two BPSK sequences of the first unique word UWA consisting of n symbols (n = 4 in the example of FIG. 1) and the second unique word UWB consisting of n symbols. (In the space diagram shown in FIG. 8, the first system unique word UWA is mapped to the symbol (A, C) marked with ○, and the second system unique word is mapped to the symbol (a, c) marked with ●. UWB is mapped so as to pass the locus x indicated by the dotted line.
[0021]
For example, when the unique word UW is composed of the first system unique word UWA: 1100 and the second system unique word UWB: 1010, 1 is assigned to (A and a) and 0 is assigned to (C and c). The unique word UW is mapped as “AaAcCaCc”, and its trajectory looks like a trajectory x indicated by a dotted line in FIG.
[0022]
The actual circuit configuration is the same as in the prior art, and it is only necessary to provide the unique word UW so as to be mapped to the above points.
[0023]
Next, processing on the receiving side will be described. FIG. 2 is a block diagram of a unique word detection circuit according to the present invention. The length of the unique word UW varies depending on the digital communication system to be applied, and is generally composed of 10 to 20 or more symbols. However, for the sake of simplicity, the length of the unique word UW is 8 symbols (that is, n = 4). ) Is described.
[0024]
In the present invention, data obtained by sampling each of the I-phase component and Q-phase component of the received baseband signal obtained by quadrature detection on the receiving side at the symbol timing is used as an input signal to the unique word detection circuit 1 and is unique for each symbol. Processing for detecting a word is performed.
[0025]
First, processing of the I phase component will be described. R _i (t) obtained by cutting out the past 2n symbols of the I-phase component of the received baseband signal sampled at the symbol timing as a unique word candidate n is alternately n symbols one by one at t = 0, 1, 2,. The symbols are divided into two (R _i (t): t = 0, 2, 4, 6), (R _i (t): t = 1, 3, 5, 7) and the first unique word UWA, second Correlation values with the system unique word UWB are calculated.
[0026]
When a unique word UW (I component = Q component) composed of two BPSK sequences UWA and UWB is expressed by the following equation:
[Equation 5]
UW = [UWA (0), UWB (0), UWA (1), UWB (1), UWA (2), UWA (3), UWB (3)] (5)
However, UWA = [UWA (0), UWA (1), UWA (2), UWA (3)]
UWB = [UWB (0), UWB (1), UWB (2), UWB (3)]
[0027]
The correlation value C _I of the I-phase component at a certain time is expressed by the following equation.
[Formula 6]
C _I = (UWA (0) R _i (0) + UWA (1) R _i (2) + UWA (2) R _i (4) + UWA (3) R _i (6)) ² + (UWB (0) R _i (1) + UWB (1) R _i (3) + UWB (2) R _i (5) + UWB (3) R _i (7)) ² (6)
[0028]
This calculation is nothing but the squared addition of the outputs of the two FIR filters. When applied to the I-phase component dot product calculation circuit 2 shown in FIG. 2, the first FIR filter is constructed by multiplying the contents of the UWA register 3 and Ri (t): t = 0, 2, 4, 6 with the first. The square of the result is _CIA . On the other hand, the contents of the UWB register 4 and R _i (t): t = 1, 3, 5, and 7 constitute a second FIR filter, and the square of the result is C _IB . That is, the first square term in the equation (6) corresponds to C _IA and the second square term corresponds to C _IB . Determination of the correlation values C _I of the I-phase component is greater.
[0029]
Incidentally, I phase component inner product arithmetic circuit 2 shown in FIG. 2, the UWA register 3, the UWB register 4, and seven delay circuits 5a~5g connected in series, R _i (7) and UWB (3) A multiplier 6a that performs multiplication with R _i (6) and UWA (3), a multiplier 6b that performs multiplication between R _i (5) and UWB (2), A multiplier 6d for multiplying R _i (4) and UWA (2), a multiplier 6c for multiplying R _i (5) and UWB (2), R _i (4) and UWA (2) A multiplier 6d that multiplies R _i (3) and UWB (2), a multiplier 6e that multiplies R _i (3) and UWB (1), A multiplier 6f for multiplying R _i (2) and UWA (1); a multiplier 6g for multiplying R _i (1) and UWB (0); A multiplier 6h for multiplying R _i (0) and UWA (0), an adder 7a for adding (accumulating) the multiplication results of the four multipliers 6b, 6d, 6f, and 6h; An adder 7b that adds (accumulates) the multiplication results of the four multipliers 6a, 6c, 6e, and 6g, a multiplier 8a that squares the addition result of the adder 7a, and an addition result of the adder 7b. It comprises a multiplier 8b that squares and an adder 9 that adds the multiplication results of the multipliers 8a and 8b.
[0030]
Each delay circuit 5a-5g delays the time for one symbol. Each adder 6b, 6d, 6f, 6h and adder 7a constitute the first FIR filter described above. The adders 6a, 6c, 6e, 6g and the adder 7b constitute the second FIR filter described above.
[0031]
Processing for Q-phase component at the same time, the Q-phase component inner product computation circuit 10 is carried out exactly as I-phase component, the correlation value C _Q Q-phase component is determined as shown in the following equation.
[0032]
[Expression 7]
C _Q = (UWA (0) R _q (0) + UWA (1) R _q (2) + UWA (2) R _q (4) + UWA (3) R _q (6)) ² + (UWB (0) R _q (1) + UWB (1) R _q (3) + UWB (2) R _q (5) + UWB (3) R _q (7)) ² (7)
[0033]
Thus the correlation value C _Q correlation values C _I and Q phase components I-phase component obtained by the result of adding by the adder 11 is supplied to the correlation value peak detection circuit 12 as the correlation value C _IQ. The correlation value peak detection circuit 12 regards the point where the correlation value exceeds the set threshold value as in the conventional method as a unique word timing and outputs it as a unique word timing candidate.
[0034]
Next, a method obtained by further simplifying the above method will be described with reference to FIG. In the system shown in FIG. 3, the circuit scale is further reduced by limiting the operation word length of the correlation operation to 1 bit in the system shown in FIG.
[0035]
The unique word detection circuit 21 shown in FIG. 3 includes an I-phase side code detection circuit 22, an I-phase component EXOR operation circuit 30, a Q-phase side code detection circuit 23, a Q-phase component EXOR operation circuit 40, and an adder 24. And a correlation value peak detection circuit 25.
[0036]
The I-phase component EXOR operation circuit 30 includes a UWA register 31, a UWB register 32, seven delay circuits 33a to 33g, eight multipliers 34a to 34h, and two 1-bit correlation operation circuits 35 and 36. And an adder 37.
[0037]
First, it is assumed that the received baseband signal sampled at the symbol timing takes positive and negative values centering on 0. At this time, each IQ component of the received baseband signal sampled at the symbol timing is converted into 1-bit data of data {1, 0} corresponding to the polarity code of the signal by the code detection circuits 22 and 23. Here, 1 is output for a negative signal, and 0 is output for a positive signal. However, even if it is defined in reverse, the overall processing may be equalized.
[0038]
Next, the I-phase components (R _i (t): t = 0, 1, 2,..., 7) of the unique word candidates obtained by cutting out the past 2n bits of 1-bit data output from the code detection circuits 22 and 23. ) And Q-phase components (R _q (t): t = 0, 1, 2,..., 7), respectively.
[0039]
In the method shown in FIG. 2, the correlation value is calculated by multiplication and accumulation. However, in the method shown in FIG. 3, it is only necessary to perform multiplication between the codes, so the multiplication is performed by an EXOR (exclusive OR) operation. Can be replaced. FIG. 4 shows the equivalence between the EXOR operation and real number multiplication.
[0040]
First, processing for the unique word candidate I component R _i (t): t = 0, 1, 2,..., 7 will be described. R _i (t) consisting of the past 2n bits: t = 0, 1, 2,..., 7 alternately divided into n symbols one by one R _i (n): n = 0, 2, 4, 6 data is EXORed with the contents of the UWA register 31 and each bit, and the result is supplied to one 1-bit correlation operation circuit 35. The remaining R _i (n) divided into n symbols: n = 1, 3, 5, and 7 are subjected to EXOR of the contents of the UWB register 32 and the respective bits, and the result is obtained as the other 1-bit correlation operation circuit. 36. It is assumed that the contents of the UWA register 31 and the UWB register 32 are also 1-bit data of {0, 1}.
[0041]
In each of the 1-bit correlation operation circuits 35 and 36, when the result of the EXOR operation is “1”, it is “−1”, and when it is “0”, it is “+1”. After adding all the EXOR operation results, the absolute value is obtained. Take. This calculation is simple if it is configured using a ROM or the like. For example, in the case of the example of FIG. 3, a 1-bit correlation calculation table as shown in FIG. 5 is created and written in a storage device such as a ROM so that an output corresponding to the input can be obtained. Good. By the above operation, it is required correlation value C _I of the I component.
[0042]
In parallel with the above-described processing for the I-phase component, processing for the Q component R _q (n): n = 0, 1, 2,. In Q-phase component EXOR operation circuit 40 the correlation value C _Q of Q components by exactly the same procedure as I-component is obtained.
[0043]
Thus the correlation value C _Q correlation values C _I and Q components of the I component obtained by the result of adding by the adder 24 is supplied to the correlation value peak detection circuit 25 as the correlation value C _IQ. The correlation value peak detection circuit 25 regards a point where the correlation value exceeds a set threshold value as in the conventional method as a unique word timing and outputs it as a unique word candidate.
[0044]
As described above, according to the method shown in FIG. 3, it is possible to detect a unique word without a multiplier.
[0045]
【The invention's effect】
As described above, a unique word detection circuit with a small circuit scale can be realized by implementing the unique word detection system according to the present invention.
[0046]
Further, by limiting the operation word length of the unique word detection circuit to 1 bit, a multiplier is completely unnecessary, and a unique word detection circuit with a smaller circuit scale can be realized.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram of a unique word construction method according to the present invention.
FIG. 2 is a block diagram of a unique word detection circuit according to the present invention.
FIG. 3 is a block diagram of another unique word detection circuit according to the present invention.
FIG. 4 is an explanatory diagram showing equivalence between EXOR and real number calculation;
FIG. 5 is an explanatory diagram showing an example of a 1-bit correlation calculation table.
FIG. 6 is a block diagram of a conventional unique word detection circuit.
FIG. 7 is an explanatory diagram showing a relationship between a reception sequence and a correlation value.
FIG. 8 is a space diagram of BPSK.
FIG. 9 is a block diagram of another conventional unique word detection circuit.
FIG. 10 is a space diagram of π / 4 shift QPSK.
[Explanation of symbols]
1, 21 Unique word detection circuit 2 I-phase component inner product operation circuit 10 Q-phase component inner product operation circuit 12, 25 Correlation value peak detection circuit 22 I-phase side code detection circuit 23 Q-phase side code detection circuit 30 I-phase component EXOR operation circuit 40 Q-phase component EXOR operation circuit

Claims (2)

A unique word composed of 2n symbols on the transmission side is composed of a combination of a first system unique word composed of n symbols and a second system unique word which are transmitted and switched mutually for each symbol, and the first system unique word is represented by π / 4 shift QPSK is mapped to a set of BPSK symbol points having a phase difference of 180 degrees on the constellation, and the second system unique word is 45 degrees or the first system unique word on the constellation. Mapping to intersecting BPSK symbol points with an angular difference of -45 degrees,
On the receiving side, a sequence obtained by alternately dividing the past 2n symbol sequences of the I-phase component of the received baseband signal into n symbol sequences one by one and each of the first system unique word and the second system unique word The correlation value on the I-phase side is obtained by executing the correlation operation and adding the results,
Correlation between a series of the 2nd symbol of the Q-phase component of the received baseband signal that is alternately divided into 1 symbol by 1 symbol and each of the first system unique word and the second system unique word is performed. The correlation value on the Q phase side is obtained by executing and adding the results,
A unique word detection system that outputs a point where a result of adding the correlation value on the I-phase side and the correlation value on the Q-phase side is larger than a preset threshold value as a unique word timing candidate. A unique word composed of 2n symbols on the transmission side is composed of a combination of a first system unique word composed of n symbols and a second system unique word which are transmitted and switched mutually for each symbol, and the first system unique word is represented by π / 4 shift QPSK on the constellation is mapped to a set of BPSK symbol points having a phase difference of 180 degrees, and the second system unique word is 45 degrees or the first system unique word on the constellation. Mapping to intersecting BPSK symbol points with an angular difference of -45 degrees,
At the receiving side, the polarity code of the received baseband signal of each of the I-phase component and the Q-phase component is detected and converted into 1-bit data,
Executes exclusive OR operation on the first system unique word and the second system unique word, each of which is divided into two n-bit sequences alternately in the past 2n bits of the I-phase component. To obtain the correlation value on the I-phase side,
Executes exclusive OR operation of each of the first system unique word and the second system unique word, and a series of the Q phase component of the past 2n bits that are alternately divided into 2 bits into n bits. To obtain the correlation value on the Q-phase side,
A unique word detection system that outputs a point where a result of adding the correlation value on the I-phase side and the correlation value on the Q-phase side is larger than a preset threshold value as a unique word timing candidate.

Images