JP4555429B2

JP4555429B2 - Predistortion type nonlinear distortion compensation circuit and digital transmitter using the same

Info

Publication number: JP4555429B2
Application number: JP2000101436A
Authority: JP
Inventors: 敏志西尾
Original assignee: Kenwood KK
Current assignee: Kenwood KK
Priority date: 2000-04-03
Filing date: 2000-04-03
Publication date: 2010-09-29
Anticipated expiration: 2020-04-03
Also published as: JP2001284980A

Description

【０００１】
【発明の属する技術分野】
本発明は、概してデジタル送信機に関し、特に、デジタル送信機の電力増幅器で発生する非線形歪みをプリディストーションにより補償する装置に関する。
【０００２】
【従来の技術】
一般に、電力増幅器では電力効率を向上させるため飽和点に近い動作点が設定されるため、入力振幅／出力振幅非線形（ＡＭ／ＡＭ）や入力振幅／出力位相非線形（ＡＭ／ＰＭ）などの特性を持つようになり、非線形歪みが発生する。このため、デジタル送信機などでは、電力増幅器のＡＭ／ＡＭ特性およびＡＭ／ＰＭ特性を基に、可能性のある入力シンボルの各振幅値に対する同相成分と直交成分の補償値の対を予め計算し、メモリに記憶しておく。そして、各入力シンボルをその振幅に対応する補償値の対で補償するようにしている。
しかし、電力増幅器のＡＭ／ＡＭ特性およびＡＭ／ＰＭ特性は、経年変化する可能性があるので、入力シンボルと電力増幅器の出力から得られる信号との誤差に応じて補償値を計算し、補償値を更新して歪み補償動作に適応性を持たせている。一般にこのような歪み補償回路をプリディストーション型歪み補償回路と称する。
【０００３】
【発明が解決しようとする課題】
補償値を更新する場合、べースバンド帯の元信号あるいはべースバンド帯の歪補償信号、べースバンド帯の復調信号の位相を求め、その位相情報を用いて歪補償更新値を求めている。しかしながら、このように従来の歪補償値を更新する方法では、信号（I，Q）の位相θを求めるために逆正接関数θ＝ｔａｎ^-1（Q／I）を用いる必要がある。ＦＰＬＡ（現地でプログラム可能な論理アレイField Programmable Logic Array）などで逆正接関数を計算する回路は、回路規模も大きくなり、演算時間も増えてしまうという問題点があった。
したがって、本発明は、信号の位相情報を用いることなく、即ち、逆正接関数計算回路を用いることなく更新歪補償値を計算する歪補償値の更新技術を提供することを１つの目的とする。
上述の補償値更新のために電力増幅器の出力から得る信号は、電力増幅器における利得や位相回転を相殺するように減衰器で調整した上で、歪み補償部の更新部分に入力する。この入力される信号を仮に「帰還信号」と呼び、また電力増幅器の出力から更新部分に至る経路を「帰還路」と呼ぶことにする。しかし、減衰器での振幅・位相の調整が正しく行われずに、前記の帰還信号に振幅、位相またはその両方の線形変動が加わると、入力信号との間に誤差が生じることになる。このため、歪補償値が幾ら適切で、電力増幅器の歪みが正しく補償されていても、その誤差のために正しい歪み補償値までも更新することになる。
したがって、本発明は、帰還路において発生する振幅・位相の線形変動に起因する歪補償値の更新を行わずに済む歪補償値の更新技術を提供することを１つの目的とする。
また、従来のプリディストーション型歪み補償回路では、補償値の更新を繰り返し行うことで、歪補償値が徐々に正しいものに近づいていく。このため更新回数が少ない間は歪補償効果が低いので、十分な効果を得るには更新を何回か行う必要がある。このような従来の歪補償値の更新手段が記載された文献としては、例えば特開平５−３０１４７号公報などがある。
本発明は、比較的少ない更新回数で歪補償値を正しい値に設定できる歪補償値更新技術を提供することを更なる目的とする。
【０００４】
【課題を解決するための手段】
以上の目的を達成するため、請求項１記載のプリディストーション型非線形歪み補償回路は、歪み特性を有する要素を含む回路の前段に配置され、入力信号の振幅に応じた歪み補償値で前記入力信号を歪ませた補償信号を前記回路に与えるプリディストーション型非線形歪み補償回路であり、予測される入力信号の振幅とこれに対応する歪み補償値とを関係付けて収容するテーブル手段と、本補償回路を制御する制御手段と、前記入力信号の受信に応じて前記制御手段から渡される歪み補償値を用いて前記入力信号を歪み補償する補償手段と、前記歪み特性を有する要素を含む回路の出力から帰還用の信号を取り出して前記入力信号に匹敵する帰還信号を出力する帰還手段と、前記帰還信号、前記入力信号および前記入力信号の歪み補償に用いた歪み補償値を基に前記帰還信号に対する新たな歪み補償値（以降、更新歪み補償値と称する）を逆三角関数を含まない計算式を用いて算出する算出手段と、前記テーブル手段の対応する歪み補償値を前記更新歪み補償値で更新する、前記制御手段に含まれる更新手段とを備え、前記計算式は、前記帰還手段が歪み特性を持たないものと仮定し、且つ前記更新歪み補償値による補償特性行列が前記歪み特性を有する要素を含む回路の伝達行列の逆行列に等しいと仮定して、前記入力信号と前記帰還信号との関係を表す関係式から導出され、前記算出手段が、前記算出に先立ち、前記入力信号と前記帰還信号の誤差を求め、前記誤差が許容範囲内である場合、前記算出をせず、この旨を前記制御手段に伝える手段を含むことを特徴とする。
請求項２記載の歪み補償回路は、歪み特性を有する要素を含む回路の前段に配置され、入力信号の振幅に応じた歪み補償値で前記入力信号を歪ませた補償信号を前記回路に与えるプリディストーション型非線形歪み補償回路であり、予測される入力信号の振幅とこれに対応する歪み補償値とを関係付けて収容するテーブル手段と、本補償回路を制御する制御手段と、前記入力信号の受信に応じて前記制御手段から渡される歪み補償値を用いて前記入力信号を歪み補償する補償手段と、前記歪み特性を有する要素を含む回路の出力から帰還用の信号を取り出して前記入力信号に匹敵する帰還信号を出力する帰還手段と、前記帰還信号、前記入力信号および前記入力信号の歪み補償に用いた歪み補償値を基に前記帰還信号に対する新たな歪み補償値（以降、更新歪み補償値と称する）を逆三角関数を含まない計算式を用いて算出する算出手段と、前記テーブル手段の対応する歪み補償値を前記更新歪み補償値で更新する、前記制御手段に含まれる更新手段とを備え、前記計算式は、前記帰還手段が歪み特性を持たないものと仮定し、且つ前記更新歪み補償値による補償特性行列が前記歪み特性を有する要素を含む回路の伝達行列の逆行列に等しいと仮定して、前記入力信号と前記帰還信号との関係を表す関係式から導出され、前記更新手段は、前記の対応する歪み補償値として、前記テーブル手段における前記帰還信号の振幅値に対応する歪み補償値を更新することを特徴とする。
請求項３記載の歪み補償回路は、歪み特性を有する要素を含む回路の前段に配置され、入力信号の振幅に応じた歪み補償値で前記入力信号を歪ませた補償信号を前記回路に与えるプリディストーション型非線形歪み補償回路であり、予測される入力信号の振幅とこれに対応する歪み補償値とを関係付けて収容するテーブル手段と、本補償回路を制御する制御手段と、前記入力信号の受信に応じて前記制御手段から渡される歪み補償値を用いて前記入力信号を歪み補償する補償手段と、前記歪み特性を有する要素を含む回路の出力から帰還用の信号を取り出して前記入力信号に匹敵する帰還信号を出力する帰還手段と、前記帰還信号、前記入力信号および前記入力信号の歪み補償に用いた歪み補償値を基に前記帰還信号に対する新たな歪み補償値（以降、更新歪み補償値と称する）を逆三角関数を含まない計算式を用いて算出する算出手段と、前記テーブル手段の対応する歪み補償値を前記更新歪み補償値で更新する、前記制御手段に含まれる更新手段とを備え、前記計算式は、前記帰還手段が歪み特性を持たないものと仮定し、且つ前記更新歪み補償値による補償特性行列が前記歪み特性を有する要素を含む回路の伝達行列の逆行列に等しいと仮定して、前記入力信号と前記帰還信号との関係を表す関係式から導出され、前記更新手段は、前記の対応する歪み補償値として、前記テーブル手段において前記入力信号の振幅値に対応する歪み補償値を更新することを特徴とする。
【０００５】
請求項４記載のプリディストーション型非線形歪み補償回路は、歪み特性を有する要素を含む回路の前段に配置され、入力信号の振幅に応じた歪み補償値で前記入力信号を歪ませた補償信号を前記回路に与えるプリディストーション型非線形歪み補償回路であり、予測される入力信号の振幅とこれに対応する歪み補償値とを関係付けて収容するテーブル手段と、本補償回路を制御する制御手段と、前記入力信号の受信に応じて前記制御手段から渡される歪み補償値を用いて前記入力信号を歪み補償する補償手段と、前記歪み特性を有する要素を含む回路の出力から帰還用の信号を取り出して前記入力信号に匹敵する帰還信号を出力する帰還手段と、前記帰還信号、前記入力信号および前記入力信号の歪み補償に用いた歪み補償値を基に前記帰還信号に対する新たな歪み補償値（以降、更新歪み補償値と称する）を逆三角関数を含まない計算式を用いて算出する算出手段と、前記テーブル手段の対応する歪み補償値を前記更新歪み補償値で更新する、前記制御手段に含まれる更新手段とを備え、前記計算式は、前記帰還手段が歪み特性を持たないものと仮定し、且つ前記更新歪み補償値による補償特性行列が前記歪み特性を有する要素を含む回路の伝達行列の逆行列に等しいと仮定して、前記入力信号と前記帰還信号との関係を表す関係式から導出され、前記帰還手段は、前記の帰還用の信号を取り出した以降の経路で発生する振幅・位相の変動を補償する手段を該帰還手段の出力端に備えることを特徴とする。
【０００６】
請求項５記載のプリディストーション型非線形歪み補償回路は、請求項１乃至４の何れかに記載の歪み補償回路において、前記制御手段が、前記帰還信号の元の前記入力信号の振幅値が前記の予測される入力信号の中の最大値である場合、前記テーブル手段において前記帰還信号の前記振幅値より大きい振幅値に対応する歪み補償値をすべて前記更新歪み補償値で更新する同一値更新手段を含むことを特徴とする。
請求項６記載の歪み補償回路は、請求項５記載の歪み補償回路において、前記制御手段が、前記帰還信号の元の前記入力信号の振幅値が前記の予測される入力信号の中の最大値である場合、前記テーブル手段において前記帰還信号の前記振幅値より大きい振幅値に対応する全ての歪み補償値を前記更新歪み補償値を含む既に更新された歪み補償値に基づいて外挿する外挿更新手段を、前記同一値更新手段の代わりに含むことを特徴とする。
請求項７記載の歪み補償回路は、請求項６記載の歪み補償回路において、前記外挿更新手段が、前記更新歪み補償値とこれと空間的に最も近い歪み補償値とを用いて比例計算により外挿することを特徴とする。
請求項８記載の歪み補償回路は、請求項５乃至７の何れか記載の歪み補償回路において、前記制御手段が、更新の回数が所定回数に達した場合、前記テーブル手段の更新されていない歪み補償値を更新済みの歪み補償値に基づいて全て補完する手段を含むことを特徴とする。
【０００７】
請求項９記載のディジタル送信機は、非線形歪み特性を有する電力増幅器とこれによる非線形歪みを補償する歪み補償手段を備えたディジタル送信機であり、前記歪み補償手段が、予測される入力信号の振幅とこれに対応する歪み補償値とを関係付けて収容するテーブル手段と、前記歪み補償手段を制御する制御手段と、前記入力信号の受信に応じて前記制御手段から渡される歪み補償値を用いて前記入力信号を予め歪ませる手段と、前記電力増幅器の出力から帰還用の信号を取り出して前記入力信号に匹敵する帰還信号を出力する帰還手段と、前記帰還信号、前記入力信号および前記入力信号の歪み補償に用いた歪み補償値を基に前記帰還信号に対する新たな歪み補償値（以降、更新歪み補償値と称する）を逆三角関数を含まない計算式を用いて算出する算出手段と、前記テーブル手段に於いて対応する歪み補償値を前記更新歪み補償値で更新する、前記制御手段に含まれる更新手段とを備え、前記計算式は、前記帰還手段が歪み特性を持たないものと仮定し、且つ前記更新歪み補償値による補償特性行列が前記非線形歪み特性を有する電力増幅器の伝達行列の逆行列に等しいと仮定して、前記入力信号と前記帰還信号との関係を表す関係式から導出され、前記算出手段が、前記算出に先立ち、前記入力信号と前記帰還信号の誤差を求め、前記誤差が許容範囲内である場合、前記算出をせず、この旨を前記制御手段に伝える手段を含むことを特徴とする。
請求項１０記載のディジタル送信機は、非線形歪み特性を有する電力増幅器とこれによる非線形歪みを補償する歪み補償手段を備えたディジタル送信機であり、前記歪み補償手段が、予測される入力信号の振幅とこれに対応する歪み補償値とを関係付けて収容するテーブル手段と、前記歪み補償手段を制御する制御手段と、前記入力信号の受信に応じて前記制御手段から渡される歪み補償値を用いて前記入力信号を予め歪ませる手段と、前記電力増幅器の出力から帰還用の信号を取り出して前記入力信号に匹敵する帰還信号を出力する帰還手段と、前記帰還信号、前記入力信号および前記入力信号の歪み補償に用いた歪み補償値を基に前記帰還信号に対する新たな歪み補償値（以降、更新歪み補償値と称する）を逆三角関数を含まない計算式を用いて算出する算出手段と、前記テーブル手段に於いて対応する歪み補償値を前記更新歪み補償値で更新する、前記制御手段に含まれる更新手段とを備え、前記計算式は、前記帰還手段が歪み特性を持たないものと仮定し、且つ前記更新歪み補償値による補償特性行列が前記非線形歪み特性を有する電力増幅器の伝達行列の逆行列に等しいと仮定して、前記入力信号と前記帰還信号との関係を表す関係式から導出され、前記制御手段が、前記帰還信号の元の前記入力信号の振幅値が前記の予測される入力信号の中の最大値である場合、前記テーブル手段において前記帰還信号の前記振幅値より大きい振幅値に対応する歪み補償値をすべて前記更新歪み補償値で更新する同一値更新手段を含むことを特徴とする。
請求項１１記載のディジタル送信機は、請求項９又は１０記載のディジタル送信機において、前記帰還手段が、前記の帰還用の信号を取り出した以降の経路で発生する振幅・位相の変動を補償する手段を出力端に備えることにより、前記帰還手段における前記の振幅・位相の変動に起因する不要な歪み補償値更新動作を避けることを特徴とする。
請求項１２記載のディジタル送信機は、請求項１０記載のディジタル送信機において、前記制御手段が、前記帰還信号の元の前記入力信号の振幅値が前記の予測される入力信号の中の最大値である場合、前記テーブル手段において前記帰還信号の前記振幅値より大きい振幅値に対応する全ての歪み補償値を前記更新歪み補償値を含む既に更新された歪み補償値に基づいて外挿する外挿更新手段を、前記同一値更新手段の代わりに含むことを特徴とする。
【０００８】
【発明の実施の形態】
以下、本発明の実施の形態例と添付図面により本発明を詳細に説明する。
なお、複数の図面に同じ要素を示す場合には同一の参照符号を付ける。
図１は、本発明の一実施形態によるプリディストーション型非線形歪み補償器を備えたデジタル無線送信機の一部を示す略ブロック図である。図１に於いて、デジタル無線送信機１の入力端Ｔｉ、Ｔｑには、送信するべき情報シンボルを供給する情報源（図示せず）が接続されているものとする。図示しない情報源から供給される各シンボルは、同相成分Ｓｉと直交成分ＳｑからなるベクトルＳと考え、（Ｓｉ，Ｓｑ）（＝Ｓ）と表す。２０より大きい要素番号を付けた要素が、本発明のプリディストーション型非線形歪み補償器を構成する。送信機１は、基本的には、送信すべきデジタル信号ＳｉおよびＳｑをそれぞれアナログ信号に変換する２つのデジタル／アナログ（Ｄ／Ａ）変換器１１、局部発振器１２、局部発振器からの発振信号を用いて各Ｄ／Ａ変換器１１の出力を直交変調する直交変調器１３、直交変調器１３の出力を電力増幅する電力増幅器１４および電力増幅器１４の出力を電波として送信するアンテナ１５からなる。
前述の図示しない情報源からの入力シンボルを受信する端子ＴｉおよびＴｑと２つのＤ／Ａ変換器１１との間には、本発明による歪み補償器１００の補償値乗算器４０が挿入され、電力増幅器１４とアンテナ１５の間には、電力増幅器１４の出力を取り出すカプラ５１が挿入される。歪み補償器１００は、さらに受信端子ＴｉおよびＴｑから入力される信号（Ｓｉ、Ｓｑ）の振幅Ｘを計算する振幅計算器２０、および振幅Ｘに対応する歪み補償値を歪み補償乗算器４０に供給する歪み補償制御部３０を備える。カプラ５１の結合出力には、減衰器５３、直交復調器５５および一対のアナログ／ディジタル（Ａ／Ｄ）変換器５７が直列に接続され、電力増幅器１４の出力信号の帰還路を形成する。さらに、歪み補償器１００は、歪み補償制御部３０から渡される入力信号に基づきＡ／Ｄ変換器５７の出力（Ｄｉ′，Ｄｑ′）の振幅・位相を補償した信号（Ｄｉ，Ｄｑ）を出力する線形補償器６０、線形補償器６０の出力およびこれに対応する歪み補償制御部３０からの歪み補償値（Ｃｉ，Ｃｑ）および入力信号を基にひずみ補償値の更新値を計算する歪み補償値計算器７０、および信号（Ｄｉ′，Ｄｑ′）の振幅ｙを計算する振幅計算器８０を備える。
【０００９】
歪み補償制御部３０は、例えばプログラムを格納した読出し専用記憶（図示せず）およびランダムアクセスメモリ（図示せず）などを備えた周知のマイクロコンピュータで構成する。歪み補償制御部３０には、歪み補償に使用する補償値（Ｃｉ，Ｃｑ）を予測される入力信号の振幅Ｘに関係付けて記憶した補償値テーブル３１を書き換え可能なメモリに格納している。入力信号（Ｓｉ、Ｓｑ）の振幅Ｘは、正規化により最大でも１を超えないように表現されているものとする。補償値テーブル３１には、電力増幅器１４のＡＭ／ＡＭ特性およびＡＭ／ＰＭ特性から予め求めた標準的な初期値を設定するか、または振幅Ｘの値に関わり無く、例えばＳｉ＝１、Ｓｑ＝０と一律に設定しておき、徐々に学習させるようにしてもよい。
さらに、歪み補償制御部３０は、入力信号Ｓとその歪み補償に用いた歪み補償値Ｃを処理に必要な期間だけ保存するためのデータバッファ３２領域を図示しないランダムアクセスメモリに確保している。図２にこのデータバッファの構造を示す。このバッファ３２は、先入れ先出し（ＦＩＦＯ）方式で書き込み読み出しを行う循環型である。図２に示した例では、入力信号と歪み補償値が次のように保存されている。

ただし、ｎ＜ｍである。バッファ３２中の最新のデータ（この例では、Ｓ_m-1とＣ_m-1）と最古のデータ（ＳｎとＣｎ）のアドレスはそれぞれ専用のポインタで管理されている。この場合は、これから入力する信号はＳｍである。また、後述のように、線形補償器６０がこれから使用する信号はＳｎであり、補償値計算器７０がこれから使用する入力信号と歪み補償値はそれぞれＳｎとＣｎである。
【００１０】
実際の送信動作において、ある時点ｎにおけるベースバンド帯の入力信号を列ベクトルＳｎ＝｛Ｓｎｉ、Ｓｎｑ｝（ｎは入力シンボルに付けた自然数で時間と共に増加するとする）とすると、振幅計算器２０は、次式に従って振幅値Ｘｎを計算する。
【数１】

（以降の説明において、｛Ａ，Ｂ｝はＡ、Ｂを要素とする列ベクトルを表すものとする。）
歪み補償制御部３０は、この振幅値Ｘnに対応する補償値（Ｃｉ，Ｃｑ）を補償値テーブル３１から取り出して歪み補償乗算器４０に渡す。これに応じて、歪み補償乗算器４０は、次式に従って、歪み補償された信号Ｌｎ＝｛Ｌｎｉ，Ｌｎｑ｝を計算し出力する。
【数２】

なお、この時、歪み補償制御部３０は、入力信号Ｓｎと歪み補償に用いた補償値とを後で歪み補償値の更新値の計算に使用できるように、図示しないメモリ内のバッファ領域３２に格納する。
信号Ｌｎは、以降、Ｄ／Ａ変換器１１、直交変調器１３および電力増幅器１４で周知のように処理されて送信信号Ｔｎとなり、アンテナ１５から送信される。
電力増幅器１４における利得・位相の変移を次式の行列Ｐで表す。
【数３】

一方、電力増幅器１４の出力に設けられたカプラ５１で取り出された信号も、周知のように減衰器５３、直交復調器５５およびＤ／Ａ変換器５７からなる帰還路を通って帰還する。この時の帰還路における利得・位相の変移を次式の行列Δで表す。
【数４】

入力信号ＳｎのＡ／Ｄ変換器５７の出力、即ち、線形補償器６０への入力を列ベクトルＤｎ´＝｛Ｄｎｉ´，Ｄｎｑ´｝とすると、
【数５】

となる。
【００１１】
＊線形補償器６０の補償動作
このような帰還信号Ｄｎ´に対し、線形補償器６０は、帰還路の利得・位相の変移を補償するように動作する。ここで、電力増幅器１４の入力信号Ｃ・Ｓｎ（＝Ｌｎ）が小さい場合、電力増幅器１４を線形と見なせば、帰還信号Ｄｎ´は次のように書くことができる。
【数６】

したがって、帰還路の振幅・位相変動の逆行列Δ^-1をＤｎ´に掛ければ、振幅・位相変動を補正することができる。ここで逆行列は、
【数７】

で求めることができる。ここで、
【数８】

である。そして振幅・位相補正をした信号Ｄｎ＝｛Ｄｎｉ，Ｄｎｑ｝は、
【数９】

と計算できる。したがって、線形補償器６０は、式（３）を用いて式（４）を計算してＤｎを求めて出力する。
このように、本発明によれば、線形補償器６０により、帰還路で発生しうる振幅・位相の線形変動がほぼ除去されるので、これに起因する歪み補償値計算器７０の不要な更新値計算動作を避けることができる。
なお、電力増幅器１４の入力信号Ｃ・Ｓｎが小さいときには、入力ベースバンド信号Ｓも小さいと考えられるので、Ｃの項も省いて計算しても良い。
【００１２】
＊歪み補償値の更新
図３は、図１の歪み補償値計算器７０が帰還信号Ｄｎを受け取る度に行う動作の例を示すフローチャートである。前段のように線形補償された帰還信号（ベースバンド復調信号）Ｄｎを受け取ると、歪み補償値計算器７０は、図３の処理を開始する。まずデータバッファ３２に格納されている最も古いデータ、即ちＤｎに対応する元の信号Ｓｎとその歪み補償に用いられた歪み補償値Ｃｎを入手し（ステップ１０１）、元の信号Ｓｎと帰還信号Ｄｎとの誤差ｅを計算する（ステップ１０２）。次に、誤差ｅが所定の許容範囲以内かどうか判断する（ステップ１０３）。誤差ｅが所定の許容範囲以内ならば、補正値Ｃｎは適切な値であると判断し、この判断が歪み補償制御部３０に伝わるようにし（ステップ１０５）て帰還信号Ｄｎに対する処理を終了する。この判断を伝える方法としては、例えば専用のフラグを用いたり、後述の歪み補償値の更新値（ＲＣｉ，ＲＣｑ）としてあり得ない所定の値を設定するなど種々考えられる。ステップ１０３において、誤差ｅが許容範囲を超える場合、誤差ｅが０となるように、帰還信号Ｄｎ、元の信号Ｓｎおよび歪み補償値Ｃｎを用いて歪み補償値の更新値（ＲＣｉ，ＲＣｑ）を算出し（１０４）、帰還信号Ｄｎに対する処理を終了する。
以下、本発明による歪み補償値の更新値（ＲＣｉ，ＲＣｑ）の算出方法を説明する。
【００１３】
まず、更新値行列ＲＣを次式で定義する。
【数１０】

ここで、帰還路において発生する振幅・位相の変動が線形補償器６０の線形補償作用により十分補償されているものとすれば、歪み補償乗算器４０の入力Ｓｎと線形補償器６０の出力Ｄｎの関係を表す式は、式（３）および（４）から、
Ｄｎ＝Ｐ・Ｃ・Ｓｎ・・・（６）
と書くことができる。
誤差ｅが０でない場合、Ｓｎ≠Ｄｎ、且つＣ≠Ｐ^-1であるから、（数１０）の更新値行列がＲＣ＝Ｐ^-1となるように更新値（ＲＣｉ，ＲＣｑ）を算出する。
換言すれば、本発明によれば、歪み補償値の更新値（ＲＣｉ，ＲＣｑ）による伝達行列ＲＣは電力増幅器１４の伝達行列の逆行列に等しいと仮定して、即ち、ＲＣ＝Ｐ^-1として式（６）からＲＣを算出する。
ここで、Ｐ・Ｃ＝Ｒとおくと、
【数１１】

ここで、ＲＣ＝Ｐ^-1とすれば、

となる。Ｒを求めるため、式（６）を行列Ｒを用いて表すと、

となる。行列Ｒを
【数１２】

と定義すると、式（８）は、次のように表せる。
【数１３】

式（９）から得られる２元連立方程式をＲｉ、Ｒｑについて解くことにより、行列Ｒは、次のように表される。
【数１４】

したがって、更新値（ＲＣｉ、ＲＣｑ）は、式（７）および（１０）から次のように計算される。
【数１５】

…（１１）
ただし、式（１１）において、ＳｉとＳｑ、ＤｉとＤｑは、ｎ番目の入力信号ＳｎおよびこのＳｎの帰還（または復調）信号であり、添字ｎを省略したものである。
このように、歪み補償値の更新値（ＲＣ１，ＲＣｑ）をべースバンドの入力信号（Ｓｎｉ，Ｓｎｑ）、べースバンド復調信号（Ｄｎｉ，Ｄｎｑ）および歪補償値（Ｃｉ，Ｃｑ）から求めることができる。
【００１４】
式（１１）から分かるように、本発明の補償値計算器７０によれば、逆正接関数θ＝ｔａｎ^-1（Q／I）を用いることなく、歪補償値の更新値（ＲＣｉ，ＲＣｑ）を算出することができる。
このとき更新歪補償値（ＲＣｉ，ＲＣｑ）に対応する振幅は、帰還信号Ｄｎの振幅Ｙ＝（Ｄｎｉ２＋Ｄｎｑ２）１／２とする。したがって、歪み補償制御部３０は、歪み補償値テーブル３１のＸの値が振幅（Ｙ）計算器８０の出力値Ｙに等しいレコードの歪み補償値（Ｃｉ，Ｃｑ）を補償値計算器７０から得た更新歪み補償値（ＲＣｉ，ＲＣｑ）で置き換える。
このように歪み補償値テーブル３１を必要に応じて更新することにより適応的な動作が可能となる。しかし、動作開始後、歪み補償値テーブル３１が適切な値に設定されるまで、ある程度時間を要する。この間、歪みの補償は適切に行われない可能性がある。そこで、本発明の歪み補償制御部３０は、比較的少ない更新回数で歪み補償値テーブル３１に適切な値が設定できるよう、次のような特徴的動作を行う。即ち、振幅Ｘｎが歪み補償値テーブル３１の振幅（Ｘ）欄の最大値に等しい場合、振幅欄が振幅（Ｙ）計算器８０の出力値（即ち、Ｄｎの振幅）Ｙに等しいレコードの歪み補償値（Ｃｉ，Ｃｑ）を更新歪み補償値（ＲＣｉ，ＲＣｑ）で更新したのち、それより大きい全ての振幅値Ｘに対する歪み補償値を、少なくとも最後に更新した歪み補償値に基づいて得られる値で全て更新する（図５（ｂ））。それ以外は、普通に振幅値Ｙに等しいレコードの歪み補償値（Ｃｉ，Ｃｑ）を更新歪み補償値（ＲＣｉ，ＲＣｑ）で更新する。
【００１５】
図５は、歪み補償値テーブル３１を更新するようすを示す図である。同図に示すように、元の信号の振幅Ｘと帰還信号の振幅Ｙを比較すると、電力増幅器の歪によりＹの方が小さい場合がある。このような状況下では、振幅値Ｘが最大値Ｘｍａｘ（＝１）の場合、この時の振幅値ＹをＹｍａｘとすると、ＹｍａｘはＸｍａｘより小さいため、歪み補償値テーブル３１の振幅欄（即ち、Ｘフィールド）がＹｍａｘより大きいレコードの歪補償値は更新することができない。
そこで、図５（ｂ）に示すように、信号Ｓｎの振幅ＸがＸｍａｘ、即ち１に等しい場合（振幅Ｙを観察するだけでは、そのＹが最大値であるかどうかは分からない）、振幅Ｙに対応する歪み補償値を更新歪み補償値（図の例では、Ｉ４とＱ４）で更新し、振幅Ｘ欄がＹより大きいレコードの歪み補償値を、少なくとも最後に更新した歪み補償値（図５（ｂ）の例では（Ｉ４，Ｑ４））に基づいて得られる値で全て更新する。
振幅欄がＹより大きいレコードの更新方法としては、例えば、振幅Ｙに対する更新値（図の例では（Ｉ４，Ｑ４））で以降の歪み補償値を置換する方法が最も簡単である。
一般に、変数ｘと相関を持つ（が、ｘの関数として容易に表すことができないような）量ｙがあるとき、ｘ１およびｘ２（ｘ１＜ｘ２）に対する量ｙ１およびｙ２が既知である場合、これらの値を基にｘ１＜Ｘ＜ｘ２なる値Ｘに対するｙを求めることは、内挿として知られる補間法の１つである。これを拡張して、Ｘ＜ｘ１＜ｘ２またはｘ１＜ｘ２＜Ｘであるような値Ｘに対応するｙの値を求めることを外挿と称する。
そこで、振幅Ｘ欄がＹより大きいレコードの歪み補償値を、それまで更新した更新値を用いて外挿してもよい。種々の外挿方法が知られているが、最も簡単なのは、振幅Ｙに対する更新値（図の例では（Ｉ４，Ｑ４））とこれに最も近い更新値（図の例では（Ｉ２，Ｑ２））を用いて比例計算を行うことである。
【００１６】
比例計算の場合、一般に、Ｘｍａｘに対応するＹの値をＹｍａｘ、Ｙｍａｘに対する更新値を（Ｉｍａｘ、Ｑｍａｘ）、後続の更新値を（Ｉｍａｘ＋ｊ、Ｑｍａｘ＋ｊ）（ただし、ｊ＝１，２，．．）、Ｙｍａｘに空間的に最も近い更新値を（Ｉｍａ，Ｑｍａ）、その間の隔たりをＤレコードとすると、
Ｉｍａｘ＋ｊ＝Ｉｍａｘ＋ｊ・（Ｉｍａｘ−Ｉｍａ）／Ｄ
Ｑｍａｘ＋ｊ＝Ｑｍａｘ＋ｊ・（Ｑｍａｘ−Ｑｍａ）／Ｄ
が成立する。これを図５（ｂ）に適用すると、
Ｉｍａｘ＋ｊ＝Ｉ４＋ｊ・（Ｉ４−Ｉ２）／２
Ｑｍａｘ＋ｊ＝Ｑ４＋ｊ・（Ｑ４−Ｑ２）／２
となる。ただし、ｊ＝１、２、３である。
また、更新回数をある程度重ねても一度も更新されない補償値が残る場合も考えられる。このため、更新回数が所定の回数（例えば１０回）に達した場合、一度も更新されていない歪み補償値を補間により更新する。このため、更新済みか否かを判断できるように、歪み補償値テーブル３１には、歪み補償値の他に、更新済みかどうかを示す更新フラグのフィールドを設けることが好ましい。
以上の歪み補償値更新原理を実現する歪み補償制御部３０の動作を説明する。
動作説明に先立ち、本発明のディジタル送信装置１は、電源投入時の初期設定において、次の設定を行うものとする。即ち、歪み補償値テーブル３１の更新フラグのクリア、データバッファ３２の最新のデータと最古のデータを指し示すポインタ（図示せず）の初期設定、および更新動作の回数を数える更新カウンタ（図示せず）の初期設定である。
【００１７】
図６は、本発明の一実施例による歪み補償制御部３０が歪み補償値テーブル３１を更新する動作を表すフローチャートである。帰還値Ｄｎに対する更新歪み補償値ＲＣ＝（ＲＣｉ，ＲＣｑ）の算出が終了した旨の知らせを歪み補償値計算器７０から、例えば割り込み信号などで受け取ると、歪み補償制御部３０は、これに応じて図６の処理を開始する。まず、ステップ１１１において、歪み補償値計算器７０による所定の動作（フラグを立てる、または更新歪み補償値ＲＣに特殊な値を設定するなど）に基づき更新が必要か否か判断する。不要な場合、ステップ１１２において、データバッファ３２の最古のデータを指し示すポインタ（図示せず）の値を１単位進める。必要な場合、ステップ１１３において、振幅Ｙに対応する歪み補償値を更新歪み補償値（ＲＣｉ，ＲＣｑ）で更新し、対応する更新フラグをセットする。次に、判断ステップ１１５において、歪み補償値テーブル３１の歪み補償値はすべて一度は更新したかどうかを判断する。すべて更新済みならば、前記の１１２に進み前述の動作を行う。
ステップ１１５において、更新していない歪み補償値がある場合、ステップ１１７において、更新カウンタ（図示せず）をインクリメントする。次に、判断ステップ１１９において、更新カウンタの値が所定の値に達したかどうか判断する。達していない場合、ステップ１２１において、データバッファ３２の最古のデータの振幅値Ｘが最大値、即ち１であるかどうか判断する。１でない場合、前述のステップ１１２に進み、前述の処理を行う。１の場合、ステップ１２３において、すでに述べたように歪み補償値テーブル３１において、振幅Ｘ欄が振幅Ｙより大きいレコードの歪み補償値を、少なくとも最後に更新した歪み補償値を基に得られる値で全て更新し、更新したレコードの更新フラグをセットする。ここでの更新は、段落番号１５または１６で説明した方法で行う。このステップ終了後は、上述のステップ１１２に進み上述の動作を行う。なお、判断ステップ１１９において、更新カウンタの値が所定の値に達した場合、ステップ１２５において、更新していない歪み補償値をすべて補完する。この場合、内挿できるものは内挿し、できないものは外挿する。その後、ステップ１１２に進み上述の動作を行って、更新処理を終了する。
【００１８】
このようにして、比較的少ない更新回数で歪み補償値テーブル３１の初期更新を行うことができる。参考までに、ステップ１２３の外挿を行わない場合、最大振幅Ｙｍａｘに対する更新歪み補償値で外挿した場合、比例計算で外挿した場合の入・出力特性を表すグラフを図７〜９に示す。図７〜９において、各グラフに描かれた５本の曲線Ａ〜Ｃ及びＤ１、Ｄ２、Ｄ３は、次のとおりである。何れの場合も、横軸は正規化された入力信号レベルを示す。
Ａ：電力増幅器１４の入・出力特性を表す曲線
Ｂ：理想的な電力増幅器の入・出力特性を表す直線
Ｃ：電力増幅器と歪み補償回路の全体の特性を直線Ｂとするために歪み補償回路が持つべき入・出力特性を表す曲線
Ｄ１：歪み補償回路の歪み補償値テーブルを１回更新した後の歪み補償回路の入・出力特性を表す曲線
Ｄ２：歪み補償回路の歪み補償値テーブルを４回更新した後の歪み補償回路の入・出力特性を表す曲線
Ｄ３：歪み補償回路の歪み補償値テーブルを２回更新した後の歪み補償回路の入・出力特性を表す曲線
前段の説明から、曲線ＡとＣは直線Ｂに関して対称な関係になることが分かる。
しかし、実際の個々の電力増幅器がどのような入・出力特性を持つかを常に正確に把握することはできないので、即ち曲線Ａが不明なので、歪み補償値テーブルの初期値として、歪み補償回路の特性が直線Ｂとなるように設定する。これは、歪み補償回路が無補償の状態となることを意味する。
この状態で動作させると、曲線Ａと直線Ｂとの差に相当する誤差が生じるので、この誤差に応じて、１回目の歪み補償値テーブルの更新が行われる。この時の歪み補償回路の入・出力特性がＤ１である。しかし、依然として理想的な補償特性である曲線Ｃとは一致していないので、誤差が生じる。この誤差に基づいて補償値の更新を４回繰り返した場合の、歪み補償回路の入・出力特性がＤ２である。
図７は、図６のステップ１２３の動作を行わないので、４回の更新の後も理想的な特性（曲線Ｃ）と離れている。これに対し、図８および９では、図７に比べて曲線Ｃへの近づき方が早くなっている。即ち、少ない更新回数で曲線Ｃに近づくことが分かる
なお、上述の実施の形態では、帰還信号Ｄｎの振幅Ｙに対応する歪み補正値を更新したが、バッファ３２の最古のデータに含まれる振幅値Ｘｎ（即ち、元の入力信号Ｓｎの振幅Ｘ）に対応する歪み補正値を更新するようにしてもよい。
以上は、本発明の説明のために実施の形態の例を掲げたに過ぎない。したがって、本発明の技術思想または原理に沿って上述の実施の形態に種々の変更、修正または追加を行うことは、当業者には容易である。故に、本発明は、以上述べた実施の形態に捕らわれることなく、ただ特許請求の範囲の記載に従って解釈するべきである。
【００１９】
【発明の効果】
本発明によれば、帰還路の歪み補償部の手前に線形補償器を挿入することにより、帰還路で発生しうる振幅・位相の線形変動がほぼ除去されるので、これに起因する歪み補償値計算器の不要な更新値計算動作を避けることができる。
歪み補正値の更新値の計算に逆三角関数を用いないので、回路が簡単になり且つ動作時間も短縮できる。
また、歪み補償値テーブルにおいて、入力信号の最大振幅値（＝１）に相当する帰還信号の振幅に対応する歪み補償値を更新した場合、少なくともその時の更新値に基ずく値で後続の歪み補償値をすべて外挿することにより、比較的少ない更新回数で歪み補償値テーブルを適切な値に近付けることができる。
【図面の簡単な説明】
【図１】本発明の一実施形態によるプリディストーション型非線形歪み補償器を備えたデジタル無線送信機の一部を示す略ブロック図である。
【図２】図１の歪み補償制御部３０の書き換え可能なメモリ（図示せず）に設けられるデータバッファのデータ構成例を示す図である。
【図３】図１の歪み補償値計算器７０が帰還信号Ｄｎを受け取る度に行う動作の例を示すフローチャートである。
【図４】図２のデータバッファに代わる更に好ましいバッファのデータ構成例を示す図である。
【図５】本発明により歪み補償値テーブル３１を更新するようすを示す図である。
【図６】本発明の一実施例による歪み補償制御部３０が歪み補償値テーブル３１を更新する動作を表すフローチャートである。
【図７】図６のステップ１２３に示した外挿を行わない場合の入・出力特性を示すグラフである。
【図８】図６のステップ１２３において最大振幅Ｙｍａｘに対する更新歪み補償値で一律に外挿した場合の入・出力特性を示すグラフである。
【図９】図６のステップ１２３において比例計算で外挿した場合の入・出力特性を示すグラフである。
【符号の説明】
１本発明のデジタル無線送信機
１１Ｄ／Ａ変換器
１２局部発振器
１３直交変調器
５１カプラ
１５アンテナ
２０振幅計算器
３０歪み補償制御部
３１歪み補償値テーブル
３２バッファ
４０歪み補償乗算器
５３減衰器
５５直交復調器
５７Ａ／Ｄ変換器
６０線形補償器
７０歪み補償値計算器
８０振幅計算器
１００本発明のプリディストーション型非線形歪み補償器[0001]
BACKGROUND OF THE INVENTION
The present invention relates generally to digital transmitters, and more particularly to an apparatus for compensating for non-linear distortion generated in a power amplifier of a digital transmitter by predistortion.
[0002]
[Prior art]
In general, since an operating point close to a saturation point is set in order to improve power efficiency in a power amplifier, characteristics such as input amplitude / output amplitude nonlinearity (AM / AM) and input amplitude / output phase nonlinearity (AM / PM) are obtained. As a result, nonlinear distortion occurs. For this reason, in a digital transmitter or the like, based on the AM / AM characteristics and AM / PM characteristics of the power amplifier, a pair of in-phase component and quadrature component compensation values for each amplitude value of a possible input symbol is calculated in advance. Store in memory. Each input symbol is compensated with a pair of compensation values corresponding to its amplitude.
However, since the AM / AM characteristic and AM / PM characteristic of the power amplifier may change over time, the compensation value is calculated according to the error between the input symbol and the signal obtained from the output of the power amplifier. To make the distortion compensation operation adaptable. Such a distortion compensation circuit is generally called a predistortion type distortion compensation circuit.
[0003]
[Problems to be solved by the invention]
When the compensation value is updated, the phase of the baseband original signal or the baseband distortion compensation signal and the baseband demodulation signal is obtained, and the distortion compensation update value is obtained using the phase information. However, in the conventional method of updating the distortion compensation value as described above, the arctangent function θ = tan is used in order to obtain the phase θ of the signal (I, Q).^-1(Q / I) must be used. A circuit that calculates an arc tangent function using an FPLA (Field Programmable Logic Array) has a problem that the circuit scale increases and the calculation time also increases.
Accordingly, an object of the present invention is to provide a distortion compensation value updating technique for calculating an updated distortion compensation value without using signal phase information, that is, without using an arctangent function calculation circuit.
A signal obtained from the output of the power amplifier for updating the compensation value is adjusted by an attenuator so as to cancel the gain and phase rotation in the power amplifier, and then input to the update portion of the distortion compensation unit. This input signal is temporarily called a “feedback signal”, and a path from the output of the power amplifier to the update portion is called a “feedback path”. However, if the amplitude and / or phase are not adjusted correctly in the attenuator and linear fluctuations in amplitude and / or phase are added to the feedback signal, an error occurs between the input signal and the input signal. Therefore, even if the distortion compensation value is appropriate and the distortion of the power amplifier is correctly compensated, the distortion compensation value is updated due to the error.
Accordingly, an object of the present invention is to provide a distortion compensation value updating technique that does not require updating of a distortion compensation value caused by a linear variation in amplitude and phase that occurs in a feedback path.
In the conventional predistortion type distortion compensation circuit, the distortion compensation value gradually approaches the correct one by repeatedly updating the compensation value. For this reason, since the distortion compensation effect is low while the number of updates is small, it is necessary to perform several updates to obtain a sufficient effect. As a document describing such a conventional means for updating a distortion compensation value, there is, for example, JP-A-5-30147.
It is a further object of the present invention to provide a distortion compensation value update technique capable of setting a distortion compensation value to a correct value with a relatively small number of updates.
[0004]
[Means for Solving the Problems]
  In order to achieve the above object, a predistortion type nonlinear distortion compensation circuit according to claim 1 is arranged in a preceding stage of a circuit including an element having a distortion characteristic, and the input signal has a distortion compensation value corresponding to the amplitude of the input signal. Is a predistortion type non-linear distortion compensation circuit for giving a compensation signal distorted to the circuit, table means for storing the amplitude of the predicted input signal and the corresponding distortion compensation value in association with each other, and the present compensation circuit Control means for controlling the input signal, compensation means for compensating for distortion of the input signal using a distortion compensation value passed from the control means in response to reception of the input signal, and output of a circuit including an element having the distortion characteristics A feedback means for extracting a feedback signal and outputting a feedback signal comparable to the input signal; and for use in distortion compensation of the feedback signal, the input signal, and the input signal. A calculation unit that calculates a new distortion compensation value (hereinafter referred to as an updated distortion compensation value) for the feedback signal based on the distortion compensation value using a calculation formula that does not include an inverse trigonometric function; Updating means included in the control means for updating a distortion compensation value with the updated distortion compensation value, and the calculation formula assumes that the feedback means does not have a distortion characteristic, and the updated distortion compensation value. A circuit in which the compensation characteristic matrix according to the above includes an element having the distortion characteristicThe legend ofIt is derived from the relational expression representing the relationship between the input signal and the feedback signal, assuming that it is equal to the inverse of the arrival matrix.,in frontThe calculation means includes means for obtaining an error between the input signal and the feedback signal prior to the calculation, and notifying the control means when the error is within an allowable range, and notifying the control means of the fact. It is characterized by.
  Claim2The distortion compensation circuit described isA predistortion type non-linear distortion compensation circuit that is arranged in front of a circuit including an element having a distortion characteristic, and that gives a compensation signal obtained by distorting the input signal with a distortion compensation value corresponding to the amplitude of the input signal to the circuit. Table means for storing the amplitude of the input signal and the distortion compensation value corresponding to the amplitude, control means for controlling the compensation circuit, and distortion passed from the control means in response to reception of the input signal Compensation means for compensating for distortion of the input signal using a compensation value; feedback means for extracting a feedback signal from an output of a circuit including an element having the distortion characteristics and outputting a feedback signal comparable to the input signal; A new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) based on the feedback signal, the input signal, and a distortion compensation value used for distortion compensation of the input signal. ) Using a calculation formula that does not include an inverse trigonometric function, and an update means included in the control means for updating the corresponding distortion compensation value of the table means with the updated distortion compensation value, The calculation formula assumes that the feedback means does not have a distortion characteristic, and assumes that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the circuit including the element having the distortion characteristic. And derived from a relational expression representing the relation between the input signal and the feedback signal,The updating means updates the distortion compensation value corresponding to the amplitude value of the feedback signal in the table means as the corresponding distortion compensation value.
  Claim3The distortion compensation circuit described isA predistortion type non-linear distortion compensation circuit that is arranged in front of a circuit including an element having a distortion characteristic, and that gives a compensation signal obtained by distorting the input signal with a distortion compensation value corresponding to the amplitude of the input signal to the circuit. Table means for storing the amplitude of the input signal and the distortion compensation value corresponding to the amplitude, control means for controlling the compensation circuit, and distortion passed from the control means in response to reception of the input signal Compensation means for compensating for distortion of the input signal using a compensation value; feedback means for extracting a feedback signal from an output of a circuit including an element having the distortion characteristics and outputting a feedback signal comparable to the input signal; A new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) based on the feedback signal, the input signal, and a distortion compensation value used for distortion compensation of the input signal. ) Using a calculation formula that does not include an inverse trigonometric function, and an update means included in the control means for updating the corresponding distortion compensation value of the table means with the updated distortion compensation value, The calculation formula assumes that the feedback means does not have a distortion characteristic, and assumes that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the circuit including the element having the distortion characteristic. And derived from a relational expression representing the relation between the input signal and the feedback signal,The updating means updates the distortion compensation value corresponding to the amplitude value of the input signal in the table means as the corresponding distortion compensation value.
[0005]
  Claim4The predistortion type nonlinear distortion compensation circuit described isA predistortion type non-linear distortion compensation circuit that is arranged in front of a circuit including an element having a distortion characteristic, and that gives a compensation signal obtained by distorting the input signal with a distortion compensation value corresponding to the amplitude of the input signal to the circuit. Table means for storing the amplitude of the input signal and the distortion compensation value corresponding to the amplitude, control means for controlling the compensation circuit, and distortion passed from the control means in response to reception of the input signal Compensation means for compensating for distortion of the input signal using a compensation value; feedback means for extracting a feedback signal from an output of a circuit including an element having the distortion characteristics and outputting a feedback signal comparable to the input signal; A new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) based on the feedback signal, the input signal, and a distortion compensation value used for distortion compensation of the input signal. ) Using a calculation formula that does not include an inverse trigonometric function, and an update means included in the control means for updating the corresponding distortion compensation value of the table means with the updated distortion compensation value, The calculation formula assumes that the feedback means does not have a distortion characteristic, and assumes that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the circuit including the element having the distortion characteristic. And derived from a relational expression representing the relation between the input signal and the feedback signal,The feedback means is characterized in that means for compensating for fluctuations in amplitude and phase generated in a path after the feedback signal is taken out is provided at the output terminal of the feedback means.
[0006]
  Claim5The predistortion type non-linear distortion compensation circuit according to any one of claims 1 to4In the distortion compensation circuit according to any one of the above, when the control means has an amplitude value of the original input signal of the feedback signal that is a maximum value in the predicted input signal, the table means It is characterized by comprising the same value updating means for updating all distortion compensation values corresponding to amplitude values larger than the amplitude value of the feedback signal with the updated distortion compensation values.
  Claim6The distortion compensation circuit according to claim5In the distortion compensation circuit described above, when the amplitude value of the original input signal of the feedback signal is the maximum value among the predicted input signals, the control unit is configured to output the feedback signal in the table unit. Extrapolation updating means for extrapolating all distortion compensation values corresponding to amplitude values larger than the amplitude value based on already updated distortion compensation values including the updated distortion compensation value is included instead of the same value updating means. It is characterized by that.
  Claim7The distortion compensation circuit according to claim6In the distortion compensation circuit described above, the extrapolation updating means extrapolates by proportional calculation using the updated distortion compensation value and the distortion compensation value spatially closest thereto.
  Claim8The distortion compensation circuit according to claim5Thru7In any one of the distortion compensation circuits described above, when the number of updates reaches a predetermined number, the control means complements all the distortion compensation values not updated in the table means based on the updated distortion compensation values. Means.
[0007]
  Claim9The digital transmitter described is a digital transmitter including a power amplifier having nonlinear distortion characteristics and distortion compensation means for compensating for nonlinear distortion caused by the power amplifier. Table means for storing corresponding distortion compensation values in association with each other, control means for controlling the distortion compensation means, and the input signal using the distortion compensation value passed from the control means in response to reception of the input signal Predistortion means, feedback means for extracting a feedback signal from the output of the power amplifier and outputting a feedback signal comparable to the input signal, distortion compensation for the feedback signal, the input signal, and the input signal Based on the distortion compensation value used, a new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) is calculated using a calculation formula that does not include an inverse trigonometric function. Calculating means for updating, and updating means included in the control means for updating a corresponding distortion compensation value in the table means with the updated distortion compensation value, wherein the feedback means has a distortion characteristic. A power amplifier that has a nonlinear distortion characteristic in which the compensation characteristic matrix based on the updated distortion compensation value is assumed to have no nonlinear distortion characteristicThe legend ofIt is derived from the relational expression representing the relationship between the input signal and the feedback signal, assuming that it is equal to the inverse of the arrival matrix.,in frontThe calculation means includes means for obtaining an error between the input signal and the feedback signal prior to the calculation, and notifying the control means when the error is within an allowable range, and notifying the control means of the fact. It is characterized by.
  Claim10The digital transmitter described isA digital transmitter comprising a power amplifier having nonlinear distortion characteristics and distortion compensation means for compensating for the nonlinear distortion caused by the power amplifier, wherein the distortion compensation means calculates an amplitude of a predicted input signal and a distortion compensation value corresponding thereto. Table means for storing in association; control means for controlling the distortion compensation means; means for predistorting the input signal using a distortion compensation value passed from the control means in response to reception of the input signal; A feedback means for extracting a feedback signal from the output of the power amplifier and outputting a feedback signal comparable to the input signal; and the distortion compensation value used for distortion compensation of the feedback signal, the input signal, and the input signal. A calculation means for calculating a new distortion compensation value (hereinafter referred to as an updated distortion compensation value) for the feedback signal using a calculation formula not including an inverse trigonometric function; Update means included in the control means for updating the corresponding distortion compensation value with the updated distortion compensation value in the control means, and the calculation formula assumes that the feedback means does not have distortion characteristics. Assuming that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the power amplifier having the nonlinear distortion characteristic, it is derived from the relational expression representing the relation between the input signal and the feedback signal. ,When the amplitude value of the original input signal of the feedback signal is the maximum value among the predicted input signals, the control means sets the amplitude value to be larger than the amplitude value of the feedback signal in the table means. It includes the same value updating means for updating all the corresponding distortion compensation values with the updated distortion compensation value.
  Claim11The described digital transmitter is claimed.9 or 10In the digital transmitter described above, the feedback means includes, on an output end, means for compensating for fluctuations in amplitude and phase generated in a path after the feedback signal is extracted, so that the feedback means It is characterized by avoiding an unnecessary distortion compensation value update operation caused by fluctuations in amplitude and phase.
  Claim12The described digital transmitter is claimed.10In the digital transmitter according to claim 1, when the amplitude value of the original input signal of the feedback signal is a maximum value among the predicted input signals, the control means is configured to output the feedback signal in the table means. Extrapolation updating means for extrapolating all distortion compensation values corresponding to amplitude values larger than the amplitude value based on already updated distortion compensation values including the updated distortion compensation value is included instead of the same value updating means. It is characterized by that.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, the present invention will be described in detail with reference to embodiments of the present invention and the accompanying drawings.
In addition, when showing the same element in several drawing, the same referential mark is attached | subjected.
FIG. 1 is a schematic block diagram illustrating a part of a digital radio transmitter including a predistortion type nonlinear distortion compensator according to an embodiment of the present invention. In FIG. 1, it is assumed that information sources (not shown) for supplying information symbols to be transmitted are connected to the input ends Ti and Tq of the digital wireless transmitter 1. Each symbol supplied from an information source (not shown) is considered as a vector S composed of an in-phase component Si and a quadrature component Sq, and is represented as (Si, Sq) (= S). Elements with an element number greater than 20 constitute the predistortion type nonlinear distortion compensator of the present invention. The transmitter 1 basically includes two digital / analog (D / A) converters 11 that convert digital signals Si and Sq to be transmitted into analog signals, a local oscillator 12, and oscillation signals from the local oscillator. It comprises an orthogonal modulator 13 that orthogonally modulates the output of each D / A converter 11, a power amplifier 14 that amplifies the output of the orthogonal modulator 13, and an antenna 15 that transmits the output of the power amplifier 14 as a radio wave.
A compensation value multiplier 40 of the distortion compensator 100 according to the present invention is inserted between the terminals Ti and Tq for receiving input symbols from the information source (not shown) and the two D / A converters 11, A coupler 51 for taking out the output of the power amplifier 14 is inserted between the amplifier 14 and the antenna 15. The distortion compensator 100 further supplies an amplitude calculator 20 for calculating the amplitude X of the signals (Si, Sq) input from the receiving terminals Ti and Tq, and a distortion compensation value corresponding to the amplitude X to the distortion compensation multiplier 40. The distortion compensation control unit 30 is provided. An attenuator 53, a quadrature demodulator 55, and a pair of analog / digital (A / D) converters 57 are connected in series to the combined output of the coupler 51 to form a feedback path for the output signal of the power amplifier 14. Further, the distortion compensator 100 outputs a signal (Di, Dq) in which the amplitude / phase of the output (Di ′, Dq ′) of the A / D converter 57 is compensated based on the input signal passed from the distortion compensation control unit 30. Distortion compensator 60, a distortion compensation value for calculating an updated value of the distortion compensation value based on the output of the linear compensator 60, the distortion compensation value (Ci, Cq) from the distortion compensation control unit 30 corresponding thereto, and the input signal A calculator 70 and an amplitude calculator 80 for calculating the amplitude y of the signals (Di ′, Dq ′) are provided.
[0009]
The distortion compensation control unit 30 is configured by a known microcomputer including, for example, a read-only memory (not shown) storing a program and a random access memory (not shown). The distortion compensation control unit 30 stores a compensation value table 31 in which compensation values (Ci, Cq) used for distortion compensation are stored in association with the predicted amplitude X of the input signal in a rewritable memory. It is assumed that the amplitude X of the input signal (Si, Sq) is expressed so as not to exceed 1 at the maximum by normalization. In the compensation value table 31, standard initial values obtained in advance from the AM / AM characteristics and AM / PM characteristics of the power amplifier 14 are set, or regardless of the value of the amplitude X, for example, Si = 1, Sq = It may be set uniformly to 0 and learned gradually.
Further, the distortion compensation control unit 30 secures a data buffer 32 area in a random access memory (not shown) for storing the input signal S and the distortion compensation value C used for the distortion compensation only for a period required for processing. FIG. 2 shows the structure of this data buffer. The buffer 32 is a circular type that performs writing and reading by a first-in first-out (FIFO) method. In the example shown in FIG. 2, the input signal and the distortion compensation value are stored as follows.

However, n <m. The latest data in the buffer 32 (in this example, S_m-1And C_m-1) And the addresses of the oldest data (Sn and Cn) are managed by dedicated pointers. In this case, the signal to be input is Sm. Further, as will be described later, the signal that the linear compensator 60 will use from now on is Sn, and the input signal and the distortion compensation value that the compensation value calculator 70 will use from now on are Sn and Cn, respectively.
[0010]
In an actual transmission operation, if an input signal in a baseband at a certain time point n is a column vector Sn = {Sni, Snq} (n is a natural number attached to an input symbol and increases with time), the amplitude calculator 20 The amplitude value Xn is calculated according to the following equation.
[Expression 1]

(In the following description, {A, B} represents a column vector having A and B as elements.)
The distortion compensation control unit 30 extracts the compensation value (Ci, Cq) corresponding to the amplitude value Xn from the compensation value table 31 and passes it to the distortion compensation multiplier 40. In response to this, the distortion compensation multiplier 40 calculates and outputs a distortion-compensated signal Ln = {Lni, Lnq} according to the following equation.
[Expression 2]

At this time, the distortion compensation control unit 30 stores the input signal Sn and the compensation value used for the distortion compensation in a buffer area 32 in a memory (not shown) so that the update value of the distortion compensation value can be used later. Store.
Thereafter, the signal Ln is processed in a known manner by the D / A converter 11, the quadrature modulator 13, and the power amplifier 14 to become a transmission signal Tn, which is transmitted from the antenna 15.
The gain / phase shift in the power amplifier 14 is represented by the following matrix P.
[Equation 3]

On the other hand, the signal taken out by the coupler 51 provided at the output of the power amplifier 14 is also fed back through a feedback path including the attenuator 53, the quadrature demodulator 55, and the D / A converter 57 as is well known. The gain / phase transition in the feedback path at this time is represented by the following matrix Δ.
[Expression 4]

If the output of the input signal Sn from the A / D converter 57, that is, the input to the linear compensator 60 is a column vector Dn ′ = {Dni ′, Dnq ′},
[Equation 5]

It becomes.
[0011]
* Compensation operation of linear compensator 60
For such a feedback signal Dn ′, the linear compensator 60 operates so as to compensate for a gain / phase shift of the feedback path. If the input signal C · Sn (= Ln) of the power amplifier 14 is small, the feedback signal Dn ′ can be written as follows if the power amplifier 14 is regarded as linear.
[Formula 6]

Therefore, the inverse matrix Δ of the amplitude and phase variation of the feedback path^-1Is multiplied by Dn ′, amplitude and phase fluctuations can be corrected. Where the inverse matrix is
[Expression 7]

Can be obtained. here,
[Equation 8]

It is. And the signal Dn = {Dni, Dnq} subjected to amplitude / phase correction is
[Equation 9]

Can be calculated. Accordingly, the linear compensator 60 calculates the equation (4) using the equation (3) to obtain and output Dn.
  As described above, according to the present invention, the linear compensator 60 almost eliminates the linear fluctuation of the amplitude and phase that may occur in the feedback path, and therefore, an unnecessary update value of the distortion compensation value calculator 70 caused by the fluctuation. Calculation operations can be avoided.
  Note that when the input signal C · Sn of the power amplifier 14 is small, the input baseband signal S is also considered to be small.
[0012]
* Update of distortion compensation value
  FIG. 3 is a flowchart showing an example of an operation performed each time the distortion compensation value calculator 70 of FIG. 1 receives the feedback signal Dn. When the linearly compensated feedback signal (baseband demodulated signal) Dn is received as in the previous stage, the distortion compensation value calculator 70 starts the process of FIG. First, the oldest data stored in the data buffer 32, that is, the original signal Sn corresponding to Dn and the distortion compensation value Cn used for the distortion compensation are obtained (step 101), and the original signal Sn and feedback signal Dn are obtained. (Step 102). Next, it is determined whether or not the error e is within a predetermined allowable range (step 103). If the error e is within a predetermined allowable range, it is determined that the correction value Cn is an appropriate value, this determination is transmitted to the distortion compensation control unit 30 (step 105), and the processing for the feedback signal Dn is terminated. As a method for conveying this determination, for example, a dedicated flag is used, or a distortion compensation value described later is used.Update valueVarious settings are possible, such as setting a predetermined value that is not possible as (RCi, RCq). In step 103, when the error e exceeds the allowable range, the update value (RCi, RCq) of the distortion compensation value is obtained using the feedback signal Dn, the original signal Sn, and the distortion compensation value Cn so that the error e becomes zero. Calculation (104), and the processing for the feedback signal Dn is terminated.
  Hereinafter, a method for calculating the update value (RCi, RCq) of the distortion compensation value according to the present invention will be described.
[0013]
  First, the update value matrix RC is defined by the following equation.
[Expression 10]

  Here, if the amplitude and phase fluctuations generated in the feedback path are sufficiently compensated by the linear compensation function of the linear compensator 60, the input Sn of the distortion compensation multiplier 40 and the output Dn of the linear compensator 60 are Expressions representing the relationship are the expressions (3) and (4From)
  Dn = P · C · Sn (6)
Can be written.
  When the error e is not 0, Sn ≠ Dn and C ≠ P^-1So (number10) Update value matrix is RC = P^-1Update values (RCi, RCq) are calculated so that
  In other words, according to the present invention, it is assumed that the transfer matrix RC with the updated distortion compensation values (RCi, RCq) is equal to the inverse matrix of the transfer matrix of the power amplifier 14, that is, RC = P^-1RC is calculated from the equation (6) as follows.
  Here, if P · C = R,
## EQU11 ##

Where RC = P^-1given that,

It becomes. In order to obtain R, when Expression (6) is expressed using a matrix R,

It becomes. Matrix R
[Expression 12]

Equation (8) can be expressed as follows.
[Formula 13]

By solving the binary simultaneous equations obtained from Equation (9) for Ri and Rq, the matrix R is expressed as follows.
[Expression 14]

Therefore, the update value (RCi, RCq) is calculated from the equations (7) and (10) as follows.
[Expression 15]

                                                  ... (11)
  In Equation (11), Si and Sq, Di and Dq are the nth input signal Sn and the feedback (or demodulated) signal of this Sn, and the subscript n is omitted.
  As described above, the update value (RC1, RCq) of the distortion compensation value can be obtained from the baseband input signal (Sni, Snq), the baseband demodulation signal (Dni, Dnq), and the distortion compensation value (Ci, Cq). .
[0014]
As can be seen from equation (11), according to the compensation value calculator 70 of the present invention, the arc tangent function θ = tan.^-1The update value (RCi, RCq) of the distortion compensation value can be calculated without using (Q / I).
At this time, the amplitude corresponding to the update distortion compensation value (RCi, RCq) is set to the amplitude Y of the feedback signal Dn = (Dni2 + Dnq2) 1/2. Therefore, the distortion compensation control unit 30 obtains, from the compensation value calculator 70, the distortion compensation value (Ci, Cq) of the record in which the X value of the distortion compensation value table 31 is equal to the output value Y of the amplitude (Y) calculator 80. The updated distortion compensation values (RCi, RCq) are replaced.
In this way, an adaptive operation can be performed by updating the distortion compensation value table 31 as necessary. However, it takes some time until the distortion compensation value table 31 is set to an appropriate value after the operation starts. During this time, distortion compensation may not be performed properly. Therefore, the distortion compensation control unit 30 of the present invention performs the following characteristic operation so that an appropriate value can be set in the distortion compensation value table 31 with a relatively small number of updates. That is, when the amplitude Xn is equal to the maximum value in the amplitude (X) column of the distortion compensation value table 31, the distortion compensation of the record in which the amplitude column is equal to the output value (ie, the amplitude of Dn) Y of the amplitude (Y) calculator 80. After updating the values (Ci, Cq) with the updated distortion compensation values (RCi, RCq), the distortion compensation values for all larger amplitude values X are values obtained based on at least the last updated distortion compensation value. All are updated (FIG. 5B). In other cases, the distortion compensation value (Ci, Cq) of the record that is normally equal to the amplitude value Y is updated with the updated distortion compensation value (RCi, RCq).
[0015]
FIG. 5 is a diagram showing how the distortion compensation value table 31 is updated. As shown in the figure, when comparing the amplitude X of the original signal and the amplitude Y of the feedback signal, Y may be smaller due to distortion of the power amplifier. Under such circumstances, when the amplitude value X is the maximum value Xmax (= 1), if the amplitude value Y at this time is Ymax, since Ymax is smaller than Xmax, the amplitude column of the distortion compensation value table 31 (that is, The distortion compensation value of the record whose X field) is larger than Ymax cannot be updated.
Therefore, as shown in FIG. 5 (b), when the amplitude X of the signal Sn is equal to Xmax, that is, 1 (when only the amplitude Y is observed, it is not known whether Y is the maximum value), the amplitude Y Is updated with updated distortion compensation values (I4 and Q4 in the example in the figure), and the distortion compensation value of the record whose amplitude X column is larger than Y is at least the distortion compensation value updated last (FIG. 5). In the example of (b), all the values obtained based on (I4, Q4)) are updated.
The simplest method for updating a record whose amplitude column is larger than Y is, for example, a method of replacing a subsequent distortion compensation value with an update value for amplitude Y ((I4, Q4 in the example in the figure)).
In general, when there are quantities y that are correlated with the variable x (but cannot be easily expressed as a function of x), the quantities y1 and y2 for x1 and x2 (x1 <x2) are known Obtaining y for a value X such that x1 <X <x2 based on the value of is one of interpolation methods known as interpolation. Extending this and obtaining a value of y corresponding to a value X such that X <x1 <x2 or x1 <x2 <X is referred to as extrapolation.
Therefore, the distortion compensation value of the record whose amplitude X column is larger than Y may be extrapolated using the updated value updated so far. Various extrapolation methods are known, but the simplest is the update value for the amplitude Y ((I4, Q4) in the example in the figure) and the update value closest to this ((I2, Q2) in the example in the figure) Is to perform a proportional calculation using.
[0016]
In the case of proportional calculation, generally, the value of Y corresponding to Xmax is Ymax, the update value for Ymax is (Imax, Qmax), and the subsequent update value is (Imax + j, Qmax + j) (where j = 1, 2,...). , Ymax is an update value spatially closest to (Ima, Qma), and a gap between them is a D record.
Imax + j = Imax + j · (Imax−Ima) / D
Qmax + j = Qmax + j · (Qmax−Qma) / D
Is established. When this is applied to FIG.
Imax + j = I4 + j · (I4-I2) / 2
Qmax + j = Q4 + j · (Q4-Q2) / 2
It becomes. However, j = 1, 2, and 3.
In addition, there may be a case where a compensation value that has not been updated remains even if the number of updates is repeated to some extent. For this reason, when the number of updates reaches a predetermined number (for example, 10 times), a distortion compensation value that has never been updated is updated by interpolation. For this reason, it is preferable to provide an update flag field indicating whether or not updated in the distortion compensation value table 31 in addition to the distortion compensation value so that it can be determined whether or not it has been updated.
An operation of the distortion compensation control unit 30 that realizes the above-described distortion compensation value update principle will be described.
Prior to the description of the operation, the digital transmitter 1 of the present invention performs the following settings in the initial setting when the power is turned on. That is, the update flag of the distortion compensation value table 31 is cleared, the initial setting of the pointer (not shown) indicating the latest data and the oldest data in the data buffer 32, and the update counter (not shown) for counting the number of update operations. ) Is the initial setting.
[0017]
  FIG. 6 is a flowchart showing an operation of updating the distortion compensation value table 31 by the distortion compensation control unit 30 according to an embodiment of the present invention. When the notification that the calculation of the updated distortion compensation value RC = (RCi, RCq) for the feedback value Dn has been completed is received from the distortion compensation value calculator 70, for example, by an interrupt signal, the distortion compensation control unit 30 responds accordingly. Then, the process of FIG. 6 is started. First, in step 111, it is determined whether or not an update is necessary based on a predetermined operation (setting a flag or setting a special value for the updated distortion compensation value RC) by the distortion compensation value calculator 70. If unnecessary, in step 112, the value of a pointer (not shown) pointing to the oldest data in the data buffer 32 is advanced by one unit. If necessary, in step 113, the distortion compensation value corresponding to the amplitude Y is updated with the updated distortion compensation value (RCi, RCq), and the corresponding update flag is set. Next, in determination step 115, it is determined whether or not all the distortion compensation values in the distortion compensation value table 31 have been updated once. If all have been updated, the process proceeds to 112 and the above-described operation is performed.
  If there is a distortion compensation value that has not been updated in step 115, an update counter (not shown) is incremented in step 117. Next, in a determination step 119, it is determined whether or not the value of the update counter has reached a predetermined value. If not, in step 121, it is determined whether the amplitude value X of the oldest data in the data buffer 32 is the maximum value, that is, 1. If it is not 1, the process proceeds to the above-described step 112 and the above-described processing is performed. In the case of 1, in step 123, as already described, in the distortion compensation value table 31, the distortion compensation value of the record whose amplitude X column is larger than the amplitude Y is a value obtained based on at least the distortion compensation value last updated. Update all and set the update flag for the updated record.hereUpdate the paragraph number15Or16This is done using the method described in. After completion of this step, the process proceeds to step 112 described above to perform the above-described operation. Note that when the value of the update counter reaches a predetermined value in the determination step 119, all distortion compensation values that have not been updated are supplemented in step 125. In this case, those that can be interpolated are interpolated, and those that cannot be interpolated. Then, it progresses to step 112, performs the above-mentioned operation | movement, and complete | finishes an update process.
[0018]
  Thus, the initial update of the distortion compensation value table 31 can be performed with a relatively small number of updates. For reference, FIGS. 7 to 9 are graphs showing input / output characteristics when the extrapolation at step 123 is not performed, when the extrapolated distortion compensation value is extrapolated with respect to the maximum amplitude Ymax, and when extrapolated by proportional calculation. . 7-9, five curves A to C and D1 drawn in each graph.,D2 and D3 are as follows. In either case, the horizontal axis represents the normalized input signal level.
A: Curve representing input / output characteristics of the power amplifier 14
B: A straight line representing the input / output characteristics of an ideal power amplifier
C: a curve representing the input / output characteristics that the distortion compensation circuit should have in order to make the entire characteristic of the power amplifier and the distortion compensation circuit a straight line B
D1: Curve representing the input / output characteristics of the distortion compensation circuit after updating the distortion compensation value table of the distortion compensation circuit once
D2: a curve representing the input / output characteristics of the distortion compensation circuit after the distortion compensation value table of the distortion compensation circuit has been updated four times
D3: a curve representing the input / output characteristics of the distortion compensation circuit after updating the distortion compensation value table of the distortion compensation circuit twice
  From the explanation of the previous stage, it can be seen that the curves A and C are symmetrical with respect to the straight line B.
  However, since it is not always possible to accurately grasp what input / output characteristics each actual power amplifier has, that is, since the curve A is unknown, the initial value of the distortion compensation value table is used as the initial value of the distortion compensation circuit. The characteristic is set to be a straight line B. This is because the distortion compensation circuitUncompensatedIt means that it becomes the state of.
  When operated in this state, an error corresponding to the difference between the curve A and the straight line B occurs, and the first distortion compensation value table is updated according to this error. The input / output characteristic of the distortion compensation circuit at this time is D1. However, since it still does not match the curve C, which is an ideal compensation characteristic, an error occurs. The input / output characteristic of the distortion compensation circuit when the update of the compensation value is repeated four times based on this error is D2.
  Since FIG. 7 does not perform the operation of step 123 of FIG. 6, it is separated from the ideal characteristic (curve C) even after four updates. On the other hand, in FIGS. 8 and 9, the approach to the curve C is faster than in FIG. That is, it can be seen that the curve C approaches with a small number of updates.
  In the above-described embodiment, the distortion correction value corresponding to the amplitude Y of the feedback signal Dn is updated. However, the amplitude value Xn (that is, the amplitude X of the original input signal Sn) included in the oldest data in the buffer 32 is updated. ) May be updated.
  The above is merely an example of an embodiment for explaining the present invention. Accordingly, it is easy for those skilled in the art to make various changes, modifications, or additions to the above-described embodiments in accordance with the technical idea or principle of the present invention. Therefore, the present invention should not be construed as the embodiments described above, but should be construed according to the description of the claims.
[0019]
【The invention's effect】
According to the present invention, by inserting a linear compensator in front of the distortion compensation unit in the feedback path, linear variations in amplitude and phase that can occur in the feedback path are almost eliminated. An unnecessary update value calculation operation of the calculator can be avoided.
Since the inverse trigonometric function is not used to calculate the update value of the distortion correction value, the circuit becomes simple and the operation time can be shortened.
Further, in the distortion compensation value table, when the distortion compensation value corresponding to the amplitude of the feedback signal corresponding to the maximum amplitude value (= 1) of the input signal is updated, the subsequent distortion compensation is performed with a value based on the updated value at that time. By extrapolating all the values, the distortion compensation value table can be brought close to an appropriate value with a relatively small number of updates.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram illustrating a part of a digital radio transmitter including a predistortion type nonlinear distortion compensator according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a data configuration example of a data buffer provided in a rewritable memory (not shown) of the distortion compensation control unit 30 in FIG. 1;
FIG. 3 is a flowchart showing an example of an operation performed each time the distortion compensation value calculator 70 of FIG. 1 receives a feedback signal Dn.
FIG. 4 is a diagram showing an example of a data structure of a more preferable buffer instead of the data buffer of FIG. 2;
FIG. 5 is a diagram showing how a distortion compensation value table 31 is updated according to the present invention.
FIG. 6 is a flowchart illustrating an operation in which the distortion compensation control unit 30 updates the distortion compensation value table 31 according to an embodiment of the present invention.
7 is a graph showing input / output characteristics when the extrapolation shown in step 123 of FIG. 6 is not performed. FIG.
8 is a graph showing input / output characteristics when the extrapolated distortion compensation value for the maximum amplitude Ymax is uniformly extrapolated in step 123 of FIG.
9 is a graph showing input / output characteristics when extrapolated by proportional calculation in step 123 of FIG. 6;
[Explanation of symbols]
    1 Digital wireless transmitter of the present invention
  11 D / A converter
  12 Local oscillator
  13 Quadrature modulator
  51  Coupler
  15 Antenna
  20 Amplitude calculator
  30 Distortion compensation controller
  31 Distortion compensation value table
  32 buffers
  40 Distortion compensation multiplier
  53 Attenuator
  55 Quadrature demodulator
  57 A / D converter
  60 linear compensator
  70 Distortion compensation value calculator
  80 Amplitude calculator
100 Predistortion type nonlinear distortion compensator of the present invention

Claims

A predistortion type non-linear distortion compensation circuit that is arranged in front of a circuit including an element having a distortion characteristic, and that gives a compensation signal obtained by distorting the input signal with a distortion compensation value corresponding to the amplitude of the input signal to the circuit. Table means for storing the amplitude of the input signal and the distortion compensation value corresponding to the amplitude, control means for controlling the compensation circuit, and distortion passed from the control means in response to reception of the input signal Compensation means for compensating for distortion of the input signal using a compensation value; feedback means for extracting a feedback signal from an output of a circuit including an element having the distortion characteristics and outputting a feedback signal comparable to the input signal; A new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) based on the feedback signal, the input signal, and a distortion compensation value used for distortion compensation of the input signal. ) Using a calculation formula that does not include an inverse trigonometric function, and an update means included in the control means for updating the corresponding distortion compensation value of the table means with the updated distortion compensation value, The calculation formula assumes that the feedback means does not have a distortion characteristic, and assumes that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the circuit including the element having the distortion characteristic. The calculation means is derived from a relational expression representing the relationship between the input signal and the feedback signal, and the calculation means obtains an error between the input signal and the feedback signal prior to the calculation, and the error is within an allowable range. A predistortion type non-linear distortion compensation circuit including means for notifying the control and notifying the control means to the effect.

A predistortion type non-linear distortion compensation circuit that is arranged in front of a circuit including an element having a distortion characteristic, and that gives a compensation signal obtained by distorting the input signal with a distortion compensation value corresponding to the amplitude of the input signal to the circuit. Table means for storing the amplitude of the input signal and the distortion compensation value corresponding to the amplitude, control means for controlling the compensation circuit, and distortion passed from the control means in response to reception of the input signal Compensation means for compensating for distortion of the input signal using a compensation value; feedback means for extracting a feedback signal from an output of a circuit including an element having the distortion characteristics and outputting a feedback signal comparable to the input signal; A new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) based on the feedback signal, the input signal, and a distortion compensation value used for distortion compensation of the input signal. ) Using a calculation formula that does not include an inverse trigonometric function, and an update means included in the control means for updating the corresponding distortion compensation value of the table means with the updated distortion compensation value, The calculation formula assumes that the feedback means does not have a distortion characteristic, and assumes that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the circuit including the element having the distortion characteristic. The update means is derived from a relational expression representing the relation between the input signal and the feedback signal, and the updating means uses the distortion compensation value corresponding to the amplitude value of the feedback signal in the table means as the corresponding distortion compensation value. Is a predistortion type nonlinear distortion compensation circuit.

A predistortion type non-linear distortion compensation circuit that is arranged in front of a circuit including an element having a distortion characteristic, and that gives a compensation signal obtained by distorting the input signal with a distortion compensation value corresponding to the amplitude of the input signal to the circuit. Table means for storing the amplitude of the input signal and the distortion compensation value corresponding to the amplitude, control means for controlling the compensation circuit, and distortion passed from the control means in response to reception of the input signal Compensation means for compensating for distortion of the input signal using a compensation value; feedback means for extracting a feedback signal from an output of a circuit including an element having the distortion characteristics and outputting a feedback signal comparable to the input signal; A new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) based on the feedback signal, the input signal, and a distortion compensation value used for distortion compensation of the input signal. ) Using a calculation formula that does not include an inverse trigonometric function, and an update means included in the control means for updating the corresponding distortion compensation value of the table means with the updated distortion compensation value, The calculation formula assumes that the feedback means does not have a distortion characteristic, and assumes that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the circuit including the element having the distortion characteristic. The update means is derived from a relational expression representing the relation between the input signal and the feedback signal, and the updating means uses the distortion compensation value corresponding to the amplitude value of the input signal in the table means as the corresponding distortion compensation value. Is a predistortion type nonlinear distortion compensation circuit.

A predistortion type non-linear distortion compensation circuit that is arranged in front of a circuit including an element having a distortion characteristic, and that gives a compensation signal obtained by distorting the input signal with a distortion compensation value corresponding to the amplitude of the input signal to the circuit. Table means for storing the amplitude of the input signal and the distortion compensation value corresponding to the amplitude, control means for controlling the compensation circuit, and distortion passed from the control means in response to reception of the input signal Compensation means for compensating for distortion of the input signal using a compensation value; feedback means for extracting a feedback signal from an output of a circuit including an element having the distortion characteristics and outputting a feedback signal comparable to the input signal; A new distortion compensation value for the feedback signal (hereinafter referred to as an updated distortion compensation value) based on the feedback signal, the input signal, and a distortion compensation value used for distortion compensation of the input signal. ) Using a calculation formula that does not include an inverse trigonometric function, and an update means included in the control means for updating the corresponding distortion compensation value of the table means with the updated distortion compensation value, The calculation formula assumes that the feedback means does not have a distortion characteristic, and assumes that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the circuit including the element having the distortion characteristic. The feedback means is derived from a relational expression representing the relation between the input signal and the feedback signal, and the feedback means compensates for amplitude and phase fluctuations that occur in the path after the feedback signal is extracted. A predistortion type nonlinear distortion compensation circuit comprising an output terminal of the feedback means.

When the amplitude value of the original input signal of the feedback signal is the maximum value among the predicted input signals, the control means sets the amplitude value to be larger than the amplitude value of the feedback signal in the table means. corresponding predistortion type nonlinear distortion compensating circuit according to any of claims 1 to 4, characterized in that it comprises a same value updating means for all the distortion compensation value is updated by the update distortion compensation value.

When the amplitude value of the original input signal of the feedback signal is the maximum value among the predicted input signals, the control means sets the amplitude value to be larger than the amplitude value of the feedback signal in the table means. The extrapolation updating means for extrapolating all corresponding distortion compensation values based on the already updated distortion compensation values including the updated distortion compensation values is included instead of the same value updating means. 5. A predistortion type nonlinear distortion compensation circuit according to 5.

7. The predistortion type nonlinear distortion compensation circuit according to claim 6 , wherein the extrapolation updating means extrapolates by proportional calculation using the updated distortion compensation value and a distortion compensation value spatially closest thereto. .

The control means includes means for complementing all of the non-updated distortion compensation values of the table means based on the updated distortion compensation values when the number of updates reaches a predetermined number. The predistortion type nonlinear distortion compensation circuit according to any one of 5 to 7 .

A digital transmitter comprising a power amplifier having nonlinear distortion characteristics and distortion compensation means for compensating for the nonlinear distortion caused by the power amplifier, wherein the distortion compensation means calculates an amplitude of a predicted input signal and a distortion compensation value corresponding thereto. Table means for storing in association; control means for controlling the distortion compensation means; means for predistorting the input signal using a distortion compensation value passed from the control means in response to reception of the input signal; A feedback means for extracting a feedback signal from the output of the power amplifier and outputting a feedback signal comparable to the input signal; and the distortion compensation value used for distortion compensation of the feedback signal, the input signal, and the input signal. A calculation means for calculating a new distortion compensation value (hereinafter referred to as an updated distortion compensation value) for the feedback signal using a calculation formula not including an inverse trigonometric function; Update means included in the control means for updating the corresponding distortion compensation value with the updated distortion compensation value in the control means, and the calculation formula assumes that the feedback means does not have distortion characteristics. Assuming that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the power amplifier having the nonlinear distortion characteristic, it is derived from the relational expression representing the relation between the input signal and the feedback signal. The calculation means includes means for obtaining an error between the input signal and the feedback signal prior to the calculation and notifying the control means when the error is within an allowable range and notifying the control means of the fact. A digital transmitter characterized by the above.

A digital transmitter comprising a power amplifier having nonlinear distortion characteristics and distortion compensation means for compensating for the nonlinear distortion caused by the power amplifier, wherein the distortion compensation means calculates an amplitude of a predicted input signal and a distortion compensation value corresponding thereto. Table means for storing in association; control means for controlling the distortion compensation means; means for predistorting the input signal using a distortion compensation value passed from the control means in response to reception of the input signal; A feedback means for extracting a feedback signal from the output of the power amplifier and outputting a feedback signal comparable to the input signal; and the distortion compensation value used for distortion compensation of the feedback signal, the input signal, and the input signal. A calculation means for calculating a new distortion compensation value (hereinafter referred to as an updated distortion compensation value) for the feedback signal using a calculation formula not including an inverse trigonometric function; Update means included in the control means for updating the corresponding distortion compensation value with the updated distortion compensation value in the control means, and the calculation formula assumes that the feedback means does not have distortion characteristics. Assuming that the compensation characteristic matrix by the updated distortion compensation value is equal to the inverse matrix of the transfer matrix of the power amplifier having the nonlinear distortion characteristic, it is derived from the relational expression representing the relation between the input signal and the feedback signal. The control means has an amplitude value greater than the amplitude value of the feedback signal in the table means when the amplitude value of the original input signal of the feedback signal is a maximum value in the predicted input signal; A digital transmitter comprising the same value updating means for updating all of the distortion compensation values corresponding to 1 with the updated distortion compensation value.

The feedback means is provided with means for compensating for amplitude / phase fluctuations occurring in the path after the feedback signal is taken out at the output end, thereby causing the amplitude / phase fluctuations in the feedback means. 11. The digital transmitter according to claim 9, wherein an unnecessary distortion compensation value update operation is avoided.

When the amplitude value of the original input signal of the feedback signal is the maximum value among the predicted input signals, the control means sets the amplitude value to be larger than the amplitude value of the feedback signal in the table means. The extrapolation updating means for extrapolating all corresponding distortion compensation values based on the already updated distortion compensation values including the updated distortion compensation values is included instead of the same value updating means. 10. The digital transmitter according to 10 .