JP3576065B2 - Antenna optimum design method, recording medium storing antenna optimum design program, and antenna device - Google Patents

Description

【００１５】
また、ＭＳ−ＭＰＹＡは上記『モノポール八木・宇田アンテナを用いた無線ＬＡＮ用小型マルチセクタアンテナの解析と設計』に示されるように、セクタ配置を行い、隣接アレーを有効利用することによりビーム幅を小さくして利得を上げ特性を向上させているが、これを一般化した無給電素子の最適配置にまでは到っていなかった。また下記［表２］に示すように設計条件の種類が多い場合、一つ一つについて特性を調べ最適化を計っていた。このため、多大な労力と時間を必要としていた。
【００１６】
【表２】

【００１７】
図２１は従来の移動通信用基地局アンテナのための二周波共用アンテナの例である（鈴木、鹿子嶋 「任意ビーム幅２周波数帯共用コーナリフレクタアンテナ」、Ｖｏｌ．Ｊ７５−Ｂｌｌ，ｐｐ．９５０−９５６，Ｎｏ．１２，１９９２年１２月参照）。従来の設計法においては、はじめにリフレクタの構造を決定する。たとえば、図２１のようなコーナリフレクタや、丸山 珠美、鹿子嶋 憲一：「等ビーム２周波共用コーナリフレクタアンテナ」、ＮＴＴＲ＆Ｄ，Ｖｏｌ．４２ Ｎｏ．９ １９９３．， ｐｐ．１１３５−１１４６，１９９３年９月に記載されているような円筒型のリフレクタなどとする。そののち、コーナリフレクタアンテナのコーナ角、コーナ長、などについてパラメータスタディを行って設計していた。
【００１８】
これも前述の従来例と同様に、限定された条件の中での最適化に多大な労力と時間を要していた。
【００１９】
【発明が解決しようとする課題】
以上示したように、従来のＧＡを用いた最適化方法、また、セクタアンテナ設計法はともに、構造をはじめに限定してからパラメータスタディをしているに過ぎず、構造は設計者の決定にゆだねられていた。また、従来のＧＡを用いた八木・宇田アンテナの最適化方法では、素子数を可変にすることができず、無給電素子の配置にも自由度がなかった。さらに従来の手法では、セクタアンテナ小型化のための最適化の手段について、なんら工夫がされていなかったという欠点があった。
【００２０】
【課題を解決するための手段】
上記課題を解決することを目的として、本発明は以下のように構成される。
【００２１】
本発明は、染色体（たとえば、染色体に擬せられ、遺伝的処理の対象となるデータ構造体）を用いてアンテナ構造を設計するためのアンテナ最適設計方法であって、位置及び素子長を含む少なくとも一つの素子情報からなる染色体を予め定めた個数だけ生成する第１工程（Ｓ１）と、該予め定めた個数の染色体のそれぞれに含まれる素子情報のうち、その素子長が所定値以下のものを削除し、かつその位置情報が重複している二以上の素子情報のうち、一つの素子情報を残して他を削除する第２工程（Ｓ２）と、 該第２工程によって得られた染色体のそれぞれにもどづいて、１セクタ分の素情報を構成する第３工程（Ｓ３）と、該第３工程によって得られた１セクタ分の素子情報のそれぞれに基づいてアンテナの素子情報を生成する第４工程（Ｓ４）と、該アンテナの素子情報のそれぞれに基づいて、アンテナ特性を計算する第５工程と（Ｓ５）、該アンテナの素子情報のそれぞれを対象関数（Ｏ（ｘ））にもとづいて評価することにより、前記予め定めた数の染色体それぞれにランク付けを行う第６工程（Ｓ６）と、該予め定められた染色体につけられたランクに基づいて、特性ごとに優れたものを優先的に保存しつつ遺伝的操作を行い、新たな世代の遺伝子群を生成する第７工程（Ｓ７）と、該新たな世代の遺伝子群が所定の収束条件を満たしているかを判定し、該所定の収束条件が満たされていればアンテナ設計を終了し、該所定の収束条件が満たされていなければ前記新たな世代の遺伝子群に基づいて前記第１工程から第７工程を繰り返す、第８工程（Ｓ８）と、を具備することを特徴とする。
【００２２】
本発明はまた、上記第１工程から第８工程を実施するためのプログラムを格納した記録媒体としても実施可能である。
【００２３】
本発明の実施の態様においては、前記第１工程において素子の位置情報を円筒座標（ｒｉ，θｉ，）を用いて構成し、前記第４工程において、該アンテナの素子情報を前記１セクタ分の素子情報の鏡像を用いて生成する、ように構成しても良い。
【００２４】
本発明の別の実施の態様においては、前記第７工程において、該予め定められた数の染色体のうち、ランクの高いものから所定数個を選別し、該選別された所定数個の染色体を次世代に必ず残ように遺伝的操作を行うように構成しても良い。
【００２５】
本発明のさらに別の実施の態様においては、前記対象関数として
Ｏ（ｘ）＝ａＧ（ｘ）−ｂ｜Ａ−Ｒｅ（Ｚ（ｘ））｜−ｃ｜Ｉｍ（Ｚ（ｘ））｜−ｄ｜Ｂ−ＨＢＷ｜−ｅＡＬ、
（ただし、Ｇ（ｘ）は該アンテナの指向性利得、Ｚ（ｘ）は該アンテナのインピーダンス、ＨＢＷは該アンテナの水平面内放射指向性の半値幅を、ＡＬはアレー長（ＡＬ＝ｒ_ｍａｘ）を、Ｔ（ｘ）はビーム割れが発生しないように設けた補助関数を示し、ｌ（ｘ）は素子の重なりや素子長の短いものを削除した後の素子数、ａ，ｂ，ｃ，ｄ，ｅ，ｆ及びｇは、それぞれの項目の重要度を表す係数であって実数にかぎられない、ＡおよびＢは所望のインピーダンス値および半値幅）である対象関数を用いても良い。
【００２６】
【実施例】
添付の図面を参照しながら、本発明の実施例について詳細に説明する。
【００２７】
［第１の実施例］
まず、本発明の第１の実施例を図１、図２、図３、図４、図５、図６、図７及び図２３を参照しながら説明する。
【００２８】
図１は、セクタアンテナ小型化のために、円筒座標を持ちいる場合のモデルを示している。図１において、記号「○」は給電素子＃１の位置を示し、極座標の角度φ＝０゜の軸上、原点からの任意の距離ｒの位置に配置される。記号「●」は、無給電素子＃２から＃ｊを示している。
【００２９】
各素子は、パラメータｈ_ｉ、ｒ_ｉ、φ_ｉで定義される。ここでｈ_ｉはｉ番目の素子の素子長、ｒ_ｉはｉ番目の素子の原点からの距離、φ_ｉはｉ番目の素子の極座標の角度を表す（ｉ＝＃１〜＃ｊ）。
【００３０】
また、セクタアンテナを表すＧＡアルゴリズムの未知数列である染色体（遺伝的処理の対象となるデータ構造体）を、図２に示すように構成する。すなわち該染色体は、素子長ｈ_＃１と原点からの距離ｒ_＃１で構成された給電素子部（給電素子＃１に相当する）と、それぞれが素子長ｈ_ｉと、原点からの距離ｒ_＃ｉと、極座標の角度φ_＃ｉとからなる（ｊ−１）個の組み合わせで構成される無給電素子部（無給電素子＃２〜＃ｊに相当する）とで構成される。
【００３１】
ここで、図１に示すようにアンテナをセクタ化し、複数のセクタに分割する。セクタの数は予め決定しておく。セクタ数をＳとし、アンテナ特性の解析に用いるφの最大値φｍａｘを１８０／Ｓ゜とする。また、アンテナ素子の原点からの距離ｒ_＃ｉについては、予めその最大値ｒ_ｍａｘ及び最小値ｒ_ｍｉｎを設定しておく。素子長ｈ_＃ｉについても、その最大値ｈ_ｍａｘを０．４λ（λ：波長）、最小値ｈ_ｍｉｎを０．０７λとして制限を設ける。これによりアンテナの励振条件が決定される。
【００３２】
このように、アンテナの構造を極座標で表し、セクタ化し、その励振条件を決定して、その未知数列である染色体それぞれの各要素値（ｈ_＃ｉ，ｒ_＃ｉ，φ_＃ｉ）を乱数を用いて決定し、図１に示すような染色体を作成する。このようにして作成される染色体を予め定めた個体数だけ用意する（図２３：ステップＳ１）。
【００３３】
なお、以下の説明においては説明の便宜上、セクタ数：Ｓ＝１２、アンテナ素子の原点からの距離の最小値：ｒ_ｍｉｎ＝０［λ］（λ：波長）、同最大値：ｒ_ｍａｘ＝５［λ］、染色体の個体数Ｎを１２０個とする条件下で、アンテナの最適設計を行うものとして説明するが、本発明はかかる条件下に限定されるものではなく、所望のアンテナ特性等に応じてこれらの条件を任意に設定して実行することが可能である。
【００３４】
次に、図２３のステップＳ２に移る。ここでは、前記ステップＳ１で準備した１２０個の染色体それぞれに対して、次に定義する条件に従い、その素子数を決定する。
【００３５】
まず、素子長ｈ_＃ｉが０．１λ未満のものについてはその素子を削除するとともに、素子の一部が重なった場合は重なった素子を除去する。この除去の後に残った素子がその染色体の素子数となる。この素子除去処理を各染色体について行う。このようにして各染色体毎に素子数を決定するということは、各染色体に対応するアンテナにおいて、φ＝０゜を中心とした１セクタの半分の素子位置が定められたことになる。
【００３６】
このように染色体毎に乱数により未知数を設定して得られるアンテナ構造においては、無給電素子の素子長や取り得る座標位置に応じて、該無給電素子が導波器としても反射器としても動作する。したがって、上記ステップ２において得られた、染色体に対応するアンテナ構造は、コーナリフレクタアンテナのような反射板付きのアンテナと八木・宇田アレーアンテナのような無給電素子（導波器）を有するアンテナの融合モデルとしてとらえることができる。
【００３７】
次に、図２３のステップＳ３について説明する。このステップにおいては、前記ステップＳ２において定められたφ＝０゜を中心とした１セクタの半分の素子位置等に基づいて、φ＝０゜を中心とした素子の鏡像を作成することで１セクタ他の半分の素子配置を決定し、この両者により１セクタ全体の素子位置を決定する。なお、鏡像を作成する際に、セクタの境界線上に素子の一部が重なる場合には、その素子の中心をセクタの境界線上に移動させる処理を行う。上記処理を１２０個の染色体すべてについて行うことにより、１セクタ分の素子配置が１２０個分作成されることになる。
【００３８】
次に、前記ステップＳ３で作成した１セクタの素子配置を元にして、残りのセクタ（１１セクタ）の素子配置を極座標の原点を中心として２×φ_ｍａｘ（＝３６０゜／Ｓ）毎に一次変換、あるいは回転移動させることによって、１つのセクタアンテナ全体の構造を設定することができる。該一次変換あるいは回転移動による処理を１２０個の染色体それぞれについて行うことによって、１２０個分のセクタアンテナ構造を作成する。これがステップＳ４である。ステップＳ４の処理が完了すると、素子数が可変で設定されたセクタアンテナ解析モデルが１２０個決定される。
【００３９】
次に、前記ステップＳ４で決定された１２０個のセクタアンテナ解析モデルのそれぞれについて、Ｉｍｐｒｏｖｅｄ Ｃｉｒｃｕｉｔ Ｔｈｅｏｒｙ （Ｉ．Ｃ．Ｔ法）を用いてアンテナ特性の計算を行う（ステップＳ５）。Ｉ．Ｃ．Ｔ法の詳細については、例えば、文献「稲垣、関口：“線状アンテナを素子とするアレイの指向性利得を最大にする厳密な設計”、信学論（Ｂ），５３−Ｂ，Ｎｏ．１１，ｐｐ６８７−６９２，１９７０」や「Ｎ． Ｉｎａｇａｋｉ：“Ａｎ ｉｍｐｒｏｖｅｄ ｃｉｒｃｕｉｔ ｔｈｅｏｒｙ ｏｆ ａ ｍｕｌｔｉ−ｅｌｅｍｅｎｔ ａｎｔｅｎｎａ．”，ＩＥＥＥ Ｔｒａｎｓ． ｏｎ ＡＰ，Ｖｏｌ．１７，Ｎｏ．２，ｐ．１２０，Ｍａｒｃｈ，１９６９」に述べられている。
【００４０】
前記ステップＳ５により全モデルのアンテナ特性を計算した後で、以下の（８）式及び（９）式で定義される対象関数Ｏ（ｘ）並びに適応度関数ｆ（ｘ）を用いて、１２０個の全染色体のランク付けを行う（ステップＳ６）。
【００４１】
Ｏ（ｘ）＝ａＧ（ｘ）−ｂ｜５０−Ｒｅ（Ｚ（ｘ））｜−ｃ｜Ｉｍ（Ｚ（ｘ））｜−ｄ｜３０−ＨＢＷ｜−ｅＡＬ ＋ｆＴ（ｘ）−ｇｌ（ｘ） ・・・・・・（８）
ｆ （ｘ）＝Ｏ（ｘ）−（Ｏ_ａｖｇ−ασ） ・・・・・・（９）
ここで（８）式のＧ（ｘ）は指向性利得を、Ｚ（ｘ）はインピーダンスを、ＨＢＷは水平面内放射指向性の半値幅を、ＡＬはアレー長（ＡＬ＝ｒ_ｍａｘ）を、Ｔ（ｘ）はビーム割れが発生しないように設けた補助関数を示し、ｌ（ｘ）は素子の重なりや素子長の短いものを削除した後の素子数を示している。また、ａ，ｂ，ｃ，ｄ，ｅ，ｆ及びｇは、それぞれの項目の重要度を表す係数である。
【００４２】
また、（９）式のＯ_ａｖｇはＯ（ｘ）の平均値を示し、σはＯ（ｘ）の標準偏差を示している。更に（９）式のαは、σに対して重み付けを行う係数であって、１≦α≦３の範囲の値を取る。
【００４３】
（８）式を見ると、利得が高いもの、インピーダンスが５０Ωに近いもの、水平面内放射指向性の半値幅が３０゜となるもの、アレー長が短いものの優先順位を高くすることを目的とした関数であることが分かる。
【００４４】
今、アレーアンテナ特性の解析を行う場合の目標を、水平面内放射指向性の半値幅が３０゜を満たし、アレー長を最小とする条件を満足するものを得ることとする。上記（８）式の係数の選び方によっては、別の項目が優先されてしまう場合があるため係数の選択には注意が必要である。ここでは、説明の便宜上（８）式の係数ａを２００とし、ｂ及びｃを２、ｄを１００、ｅを５、ｇを１として解析したものとして説明するが、本発明において、係数の選び方はこれに限定されるものではなく、各係数の値は任意に設定することが可能である。
【００４５】
次に、ステップＳ７について説明する。前記ステップＳ６においてランク付けされた１２０個の染色体の中から、アンテナ特性の優れたものを優先的に保存しつつ、次に示す遺伝的操作を実施する。
【００４６】
ここで、この遺伝的操作を実施するに当たり、ルーレット保存法とエリート保存法の２種類の方法がある。これらの方法の詳細については、参考文献「Ｄ．Ｅ．Ｇｏｌｄｂｅｒｇ，”Ｇｅｎｅｔｉｃ ａｌｇｏｒｉｔｈｍｓ ｉｎ ｓｅｒａｃｈ，ｏｐｔｉｍｉｚａｔｉｏｎ ａｎｄ ｍａｃｈｉｎｅ ｌｅａｒｎｉｎｇ”，Ｒｅａｄｉｎｇ，Ｍａｓｓ．：Ａｄｄｉｓｏｎ−Ｗｅｓｌｅｙ，１９８９．」に書かれている。以下にその概要を説明する。
【００４７】
まず、ルーレット保存法について説明する。
【００４８】
ランク付けされた染色体の中からランクの高いものが選ばれる確率が高くなるように構成されたルーレットを用意する。このルーレットを回すと一つの染色体が選択される。かかるルーレットによる染色体の選択を１２０回繰り返すことにより、新たに１２０個の染色体を選び出す操作を行う。該選択操作においては、ランクの高い染色体がより多く選ばれる可能性が高く、またランクの高い染色体は複数回選ばれる可能性もある。一方、ランクの低い染色体は該ルーレットにより選択される可能性が低く、そのため徐々に消滅していくことになる。その結果ランクの高い染色体が優先して選択される。
【００４９】
こうして選択された複数の染色体の中から、予め設定した交配率で交配（染色体の半分の要素を互いに交換すること）の親を選ぶ。交配率は任意に設定可能であるが、例えば交配率を０．５とした場合、該ルーレットにより選択された１２０個の染色体から６０個の染色体が選出される。この親を選ぶ方法には、１２０個の染色体からランダムに選んだり、最初に選択された６０個の染色体としたり、いろいろとその選択方法が考えられるが、いずれの方法を採用しても良い。
【００５０】
このようにして交配の親となる６０個の染色体を選んだら、次にこれらの親を２個ずつ対にして、各々の染色体を半分に分けて対同士で交配させる。
【００５１】
更にこうして得られた１２０個の染色体の中から予め定める確率で突然変異の染色体を作成する。この突然変異は選択された染色体の要素の一部を適当に変更することで実施する。
【００５２】
以上の遺伝的操作処理によって新たな世代の染色体グループを生成する。
【００５３】
これが、ステップＳ７における新しい世代の染色体を作る遺伝的操作である。
【００５４】
上記の通り、ステップＳ７においては、アンテナ特性の優秀なものほど選択されて残される確率が高くなるように染色体に優先的ランク付けを行っているので、新たな世代の染色体の中に特性の優れたものが残る確率が高くなる。
【００５５】
次に、エリート保存法について説明する。
【００５６】
前述の通り、前記ステップＳ６の処理により、１２０個の全染色体のそれぞれにランク付けを行い、このランク付けされた１２０個の染色体の中から、先ず、ランクの高い順に、予め定めた個数の染色体を選んでこれをそのまま変化させずに保存する染色体として選択する。例えば上位２０個の染色体を選択するものとする。遺伝的操作を行うための交配の親となる染色体を１２０個用意するために選択すべき残りの個数は１２０−２０＝１００個となる。ここで、前述のルーレット保存法において使用されたルーレットと同様の、ランク付けされた１２０個の染色体の中からランクの高いものが選ばれる確率が高くなるようなルーレットを用意して、これを１００回繰り返して回し、新たに１００個の染色体を選び出す操作を行う。
【００５７】
次に、この１００個の中から予め設定した交配率で交配（染色体の半分の要素を互いに交換すること）の親を選ぶ。例えば交配率を０．５とすれば５０個の染色体を選ぶことになる。この親を選ぶ場合には、ランダムに選んだり、最初に選択された５０個の染色体としたり、いろいろとその選択方法が考えられるが、いずれの方法によって選出しても良い。
【００５８】
このようにして交配の親となる５０個の染色体を選んだら、次にこれらの親となる染色体を２個ずつ対にして、各々の染色体を半分に分けて対同士で交配させる。
【００５９】
更にこうして得られた１００個の染色体の中から予め定める確率で突然変異の染色体を作成する。この突然変異は選択された染色体の要素の一部を適当に変更することで実施する。
【００６０】
このようにして、最初に優先的に確実に残す染色体として選んだ２０個の染色体と、その後選んで遺伝的操作を加えた１００個の染色体を合わせて、１２０個の新たな染色体が新世代の染色体グループとして誕生する。これがエリート保存法を用いた場合の新しい世代の染色体の形成例である。
【００６１】
この他にも、前述の遺伝的操作の中で、より良い結果が得られるように、更に条件を設定したりする事も考えられるが、本発明の本質ではないため、それらの詳細については前述した参考文献を参照されたい。
【００６２】
上記操作が図２３のステップＳ７で行われる。
【００６３】
次に、ステップＳ８でその染色体グループの収束度を判定して、予め設定した収束度を満足すれば、ステップＳ９に移って操作を終了する。一方、収束度を満足していなければステップＳ２に戻って、前述の操作を繰り返す。収束度の判定には、前記Ｏ_ａｖｇが収束状態にあるかどうかチェックしたり、前記σが小さくなる（０に近づく）ことで判断したりすることができる。
【００６４】
また、収束状態にあるかどうかを判定する基準として、アンテナ特性が所望の値に収束しているかを種々のパラメータに対して判断する判定条件を加えることにより、さらに良い特性のアレーアンテナを設計することが可能となる。
【００６５】
以下に、ルーレット保存法を用いた場合の本実施例に係るアンテナ最適設計方法の実施結果を示す。なお、この実施においては上記ルーレット保存法の設定パラメータとして、交配率を０．５とし、突然変異率を０．００１とし、また前述の（８）式の係数ａを２００、ｂ及びｃを２、ｄを１００、ｅを５、ｇを１とし、本実施例にかかるアンテナ最適設計方法に従って第１世代から第９０世代の染色体を形成して解析している。
【００６６】
図３は世代毎の対象関数Ｏ（ｘ）の平均値を示したグラフであり、世代を重ねる毎に対象関数の平均値が収束していく様子が分かる。
【００６７】
図４は世代毎の水平面内放射指向性の半値幅の値を示しており、所望の設計値である３０゜に収束していることが分かる。
【００６８】
図５は入力インピーダンスの世代毎の平均値がどのように収束するかを示しており、リアクタンス成分Ｘは所望の値０［Ω］に近いところで収束しているが、レジスタンス成分Ｒは１００［Ω］近辺と所望の５０［Ω］からはやや離れたところで収束している。これは、（８）式において係数ｂの値が他よりも低いためにこのレジスタンス成分の項目の重要度が下がり、本要素が軽視されたためと考えられる。
【００６９】
また、図２３に示す本発明のアルゴリズムにより得られた無給電素子の配置と各素子の素子長の関係を図６に示す。この図６で示されたアンテナ構造は、従来の八木・宇田アンテナ、コーナリフレクタアンテナをセクタ配置してパラメータスタディを行う手法では導くことのできない形状である。このアンテナ構造の水平面内放射指向性を図７に示す。図７からわかるように、メインビームの方向にずれもなく、また、水平面内放射指向性の半値幅３０゜を実現する結果が得られている。
【００７０】
このように、本発明の方法を用いることで、Ｇｅｎｅｔｉｃ Ａｌｇｏｒｉｔｈｍｓの適用により、従来の方法では困難であったアンテナ構造までを含むセクタアンテナの小型化に対するアンテナ設計が容易となる。
【００７１】
［第二の実施例］
次に、本発明の第二の実施例について説明する。
【００７２】
かかる第二の実施例は、Ｓｔｅｐ６において使用する対象関数が異なる点を除き、前記第一の実施例と同じである。
【００７３】
第二の実施例における対象関数は前記（８）式に、サイドローブを低減する項＋ｋＳＬ（ｋは係数、ＳＬはサイドローブレベル）とバックローブを低減する項＋ｈＢＬ（ｈは係数、ＢＬはバックローブ）を追加し、ビーム割れが発生しないように設けた関数ｆＴ（ｘ）と素子数に制約を与える項−ｇＬ（ｘ）を削除して得られる以下の（１０）式である。
【００７４】
O(x)＝aG(x)−b｜A−Re(Z(x))｜−c｜Im(Z(x))｜−d｜B−HBW｜−eAL
＋kSL＋hBL …（１０）
【００７５】
本発明の第二の実施例にかかるアンテナ最適設計方法を実行して、最適化されれたアンテナ構造を得た。該最適化されたアンテナ構造における無給電素子の配置例を図８に、その水平面内放射指向性を図９に示す。図９からわかるように（１０）式で追加したサイドローブとバックローブの低減項により、サイドローブとバックローブが低減されていることがわかる。
【００７６】
ここで、本実施例では、サイドローブは４５゜＜φ＜１５０゜の範囲で−１０ｄＢ以下となることを目標とした。かかる目標を達成するために、サイドローブレベルが−１０ｄＢ以下を満たすとき評価値を高くし（この実施例ではｋＳＬ＝１０００．０に設定した）、−１０ｄＢ以下を満さないときには評価値を下げる（この実施例ではｋＳＬ＝−１００．０に設定した）ように規定した。また、バックローブは１５０゜＜φ＜１８０゜の範囲でそのレベルが−１５ｂＢ以下となることを目標とし、バックローブレベルが−１５ｂＢ以下を満たす場合に評価値を高くし、それ以外の場合には評価値を低くするように規定した。このように、本発明においては対象関数Ｏ（ｘ） の中に離散関数が含まれていても良い。かかる手法は、連続的な関数では表現できない問題を解く場合において他の手法に比べて有利である。
【００７７】
なお、本実施例では、対象関数Ｏ（ｘ） に素子数に関する項ｌ（ｘ）を設けなかったために、各セクタ毎の素子数は２２とやや多くなっている。
【００７８】
［第三の実施例］
本発明の第一および第二の実施例においては各素子の位置を円筒座標系を用いて規定したが、円筒座標系の代わりに直交座標系を用いて各素子の位置を規定する場合にも本発明は実施可能である。かかる直交座標系を用いた場合を第三の実施例として、以下に説明する。
【００７９】
直交座標を用いた本実施例においては、図２３に示した本発明のフローに対し、ステップＳ１における染色体の定義が以下に述べるように異なり、かつ、セクタ化の概念を除去した点（ステップＳ３，Ｓ４の処理が変わる）でそのアンテナの素子配置を設定する処理が変わるが、アルゴリズムとしては、前述の第１、及び第２の実施例と同様である。
【００８０】
まず、本実施例において用いられる染色体の構成について説明する。図１０は直交座標を用いた場合のアンテナ構成例である。該アンテナには、原点（座標（０，０））に固定された素子長Ｈ_＃１を有する給電素子＃１と、それぞれの座標（ｘ、ｙ）および素子長Ｈを未知数とする無給電反射器＃２から＃ｍと、それぞれの座標（ｘ、ｙ）および素子長Ｈを未知数とする無給電導波器＃ｍ＋１から＃ｎとが設けられる。ここでｍは２以上の自然数、ｎは３以上の自然数である。なお、図１０に示すアンテナ構成例においては、無給電反射器は＃２のみ （ｍ＝２）、無給電導波器は＃３から＃６ （ｎ＝６）である。
【００８１】
図１０において記号「○」は給電素子＃１の位置を示しており、原点に配置する。＃２〜＃ｍは給電素子の背面、あるいは側面などに配置される無給電素子で構成される反射器である。反射器は素子間隔が０．１λ以下のとき反射板と同一視できる（Ｋｒａｕｓ Ｊ．Ｄ：Ａｎｔｅｎｎａｓ Ｓｅｃｏｎｄ Ｅｄｉｔｉｏｎ，ｐ．５５６，ＭｃＧＲＡＷ−ＨＩＬＬ，１９８８．）。記号●は無給電素子で構成される導波器＃ｍ＋１〜＃ｎである。
【００８２】
上記のような給電素子＃１、無給電反射器＃２から＃ｍ、および無給電導波器＃ｍ＋１から＃ｎより構成されるアンテナに対応する染色体を、図１１に示す。該染色体は、給電素子部と、反射器部と、導波器部とから構成される。給電素子部はその給電素子の素子長Ｈ_＃１を格納している。反射器部は、無給電反射器＃２から＃ｍのそれぞれのｘ座標値、ｙ座標値、素子長（ＲＸ_ｉ，ＲＹ_ｉ，Ｈ_ｉ）を格納している（ｉは＃２から＃ｍ）。また、導波器部は無給電導波器＃ｍ＋１から＃ｎのそれぞれのｘ座標値、ｙ座標値、素子長（ＤＸ_ｉ，ＤＹ_ｉ，Ｈ_ｉ）を格納している（ｉは＃ｍ＋１から＃ｎ）。なお、上記例においては反射器と導波器部をそれぞれ独立に設けているが、反射器と導波器それぞれの最大値、最小値、ビット数などに与えられる条件を分ける必要がない場合はこのように分離せず一種類の無給電素子部として統合しても良い。
【００８３】
上記染色体において、各素子の特性は（ＲＸ_ｉ，ＲＹ_ｉ，Ｈ_ｉ）あるいは（ＤＸ_ｉ，ＤＹ_ｉ，Ｈ_ｉ）によって決定する。各未知数ＲＸ_ｉ，ＲＹ_ｉ，ＤＸ_ｉ，ＤＹ_ｉ，Ｈ_ｉの範囲をそれぞれ以下のように定め、励振条件を決定する。
【００８４】
ＲＸｉ：最小値ＲＸ_ｍｉｎ ＝ ０；最大値ＲＸ_ｍａｘ ＝ −０．２５，
ＲＹｉ：最小値ＲＹ_ｍｉｎ ＝ ０；最大値ＲＹ_ｍａｘ ＝ １．０，
ＤＸｉ：最小値ＤＸ_ｍｉｎ ＝ ０；最大値ＤＸ_ｍａｘ ＝ ５．０，
ＤＹｉ：最小値ＤＹ_ｍｉｎ ＝ ０；最大値ＤＹ_ｍａｘ ＝ １．０，
Ｈｉ ：最小値Ｈ_ｍｉｎ ＝ ０ ； 最大値Ｈ_ｍａｘ ＝ ０．３５，
以上のように、アンテナの構造を直交座標で表し、その励振条件を決定して、未知数列である染色体それぞれの各要素値（ＲＸ_ｉ，ＲＹ_ｉ，Ｈ_ｉ）および（ＤＸ_ｉ，ＤＹ_ｉ，Ｈ_ｉ）を乱数を用いて決定し、染色体を作成する。このようにして作成される染色体を予め定めた個体数だけ用意する（ステップＳ１）。
【００８５】
つぎに、各染色体における素子数の決定を行う（ステップＳ２）。それぞれの染色体において、異なる素子に含まれる要素値が同一である場合、すなわち素子の構造の一部が重なっている場合、あるいは素子長Ｈ_ｉがあらかじめ定められた値よりも小さい場合、その染色体からその素子を除去する。
【００８６】
このようにして得られた素子群を、水平面内放射指向性が正面方向に対して対象となるよう、Ｘ軸に対して鏡像を作成する。この際Ｘ軸上にある素子は削除される。この処理を前記予め定められた個体数の染色体それぞれについておこない、素子数が可変で設定されたアンテナ解析モデルが予め定められた個数だけ決定される（ステップＳ３’）。
【００８７】
つぎに、前記アンテナ解析モデルの全てについて、Ｉｍｐｒｏｖｅｄ Ｃｉｒｃｕｉｔ Ｔｈｅｏｒｙ （Ｉ．Ｃ．Ｔ法）を用いてアンテナ特性の計算を行い（ステップＳ５）、該アンテナ特性の計算に基づいて、対象関数Ｏ（ｘ）並びに適応度関数ｆ（ｘ）を用いて、全染色体のランク付けを行い、（ステップＳ６）、該ランク付けされた染色体の中から、アンテナ特性の優れたものを優先的に保存しつつ、遺伝的操作を実施して新しい世代の染色体を作り（ステップＳ７）、最後に該新しい世代の染色体の収束度を判定して、予め設定した収束度を満足すれば、処理を終了し、該収束度を満足していなければ前述の処理（ステップＳ２からステップＳ７）を繰り返す。
【００８８】
次にこの直交座標による染色体を用いたときの具体的実施結果を示す。
【００８９】
この実施においては対象関数Ｏ（ｘ） として以下の（１１）式を用いた。
【００９０】
Ｏ（ｘ）＝ａＧ（ｘ）−ｂ｜５０−Ｒｅ（Ｚ（ｘ））｜−ｃ｜Ｉｍ（Ｚ（ｘ））｜−ｄ｜３０−ＨＢＷ｜−ｅＡＬ＋ｆＴ（ｘ） ・・・（１１）
上記（１１）を用いた第三の実施例を実施した結果として得られた無給電素子の配置を図１２に示す。これは、「与えられた空間内に、無給電素子をばらまいて得られる構造のうち、水平面内放射指向性のビーム幅が３０゜、インピーダンスが５０［Ω］に近い値がとれるものでアレー長を小さくできる配置は何か」という問題を、直交座標ＩＣＴによる計算とＧＡで解いた結果である。該得られた無給電素子の配置を有するアンテナ構造に対する水平面内放射指向性を図１３に示す。この図からわかるように、該アンテナ構造においてはビーム幅３０゜を満たしている。
【００９１】
以上示したように、本発明の第三の実施例のような直交座標ＧＡ−ＩＣＴも所望特性を得るための素子の配置を考えるための有効な手段となる。ただし、セクタ化したとき小型となる配置を考える際には、直交座標の場合、Ｙ座標の範囲の設定が困難になるため、本発明の第１の実施例である、円筒座標を用いる方が望ましい。また、図１３は、サイドローブ、バックローブが−１０ｄＢほど発生していることを示しているが、これは対象関数Ｏ（ｘ） にこれらを低減するための規定（前記第二の実施例におけるｋＳＬ， ｈＢＬ）を入れていないからである。
【００９２】
［第四の実施例］
次に本発明の第四の実施例を示す。本実施例は、反射器と導波器を分離せず単に無給電素子として統合している点を除いて前記第三の実施と同一である。
【００９３】
本実施例の染色体の構成を図１４に示す。該染色体は給電素子部と、無給電素子部で構成されている。給電素子部はその給電素子の素子長Ｈ_＃１を格納しており、無給電素子部は無給電素子＃２から＃ｎのそれぞれのｘ座標値、ｙ座標値、素子長（ＤＸ_ｉ，ＤＹ_ｉ，Ｈ_ｉ）を格納している（ｉは＃２から＃ｎ）。
【００９４】
該第四の実施例にかかるアンテナ最適化設計方法を実行した結果を以下に示す。ここで、ＤＸ_ｉ，ＤＹ_ｉ，Ｈ_ｉに与えられる条件は前述の第３の実施例と同様とし、対象関数Ｏ（ｘ）として（１１）式の代わりに以下の（１２）式を用いた。
【００９５】
Ｏ（Ｘ）＝ａＧ（ｘ）−ｂ｜５０−Ｒｅ（Ｚ（ｘ））｜−ｃ｜Ｉｍ（Ｚ（ｘ））｜ ・・・（１２）
この実施例におけるアンテナの最適化は、「大きさの決められた空間内に、素子数の最大値だけを決めて無給電素子をばらまいたときに、どのような素子長、素子配置にすれば高い利得が得られるアンテナとして動作するか」という問題を解くことに他ならない。本実施例の最適化法を用いて計算した結果を図１５、図１６に示す。図１５は、該最適化方法を実施した結果得られたアンテナ構造における無給電素子の配置例である。該無給電素子の配置を有するアンテナの水平面内放射指向性は、図１６に示すとおりである。図１６よりわかるように、最適化されたアンテナにおいては半値幅２２゜、利得１２ｄＢが得られた。アレー長５λにしては利得がやや低いのは、給電素子より後ろには何も素子をおかない構造になっているためである。
【００９６】
本発明の第一及び第二の実施例の特徴は、図２２で示した従来の設計法を示すと比較した場合、（１）座標系を用いて染色体を一般化し、従来の素子間隔を素子長といった特定の構造に限らない点、（２）円筒座標を採用している点、（３）セクタ間相互結合を考慮し、鏡像と一次変換を駆使して計算量を減らしている点、（４）対象関数にセクタアンテナ小型化を目的とした、一定値に近いビーム幅とアンテナサイズが小さいものを優先する項を採用している点にある。
【００９７】
本発明の第三及び第四の実施例の特徴は、図２２で示した従来の設計法を示すと比較した場合、（１）座標系を用いて染色体を一般化し、従来の素子間隔を素子長といった特定の構造に限らない点、（２）鏡像を駆使して計算量を減らしている点、を採用している点にある。
【００９８】
尚、本発明で給電素子を複数にしても、あるいは、全ての素子を無給電素子から給電素子に変更しても、必要に応じて、染色体に位相と振幅の未知数の項を素子数分追加することで、本明細書に示したのと同様の効果が得られるのはいうまでもない。
【００９９】
【発明の効果】
以上説明したように、本発明によれば、セクタアンテナ、特に、無給電素子のアレーを用いる八木・宇田アンテナや、反射板を無給電素子の列で近似できるコーナリフレクタタイプのアンテナの素子の配置を、所望特性を満たすように、自由度の高いところで最適化することが可能となる。特に、本発明によれば染色体の構造は固定長のまま素子数を未知数にすることが可能である。本発明を用いて、任意のセクタアンテナの小型化問題、等ビーム多周波共用問題の最適化を行うことが可能であり、無線ＬＡＮ用のアンテナの小型化最適化、移動通信基地局用等ビーム多周波共用アンテナの小型化最適化に有効な手段である。
【０１００】
また本発明によれば、セクタ間に生じる相互結合の影響を含めた解析を容易に実現できる。
【図面の簡単な説明】
【図１】円筒座標を用いたアンテナ素子は位置の構成図である。
【図２】円筒座標を用いた場合の染色体の構成例を示す図である。
【図３】対象関数の世代毎の収束の様子を示す図である。
【図４】ビーム幅の世代毎の収束の様子を示す図である
【図５】入力インピーダンスの世代毎の収束の様子を示す図である。
【図６】本発明の実施の結果得られたアンテナ構造における無給電素子の配置を示す図である。
【図７】本発明の実施の結果得られたアンテナ構造における水平面内放射指向性を示す図である。
【図８】最適化されたアンテナ構造における無給電素子の配置を示す図である。
【図９】最適化されたアンテナ構造における水平面内放射指向性を示す図である。
【図１０】直交座標を用いたアンテナ構成例を示す図である。
【図１１】直交座標のときの染色体の構成例を示す図である。
【図１２】第三の実施例を実施した結果として得られた無給電素子の配置を示す図である。
【図１３】第三の実施例を実施した結果として得られたアンテナ構造の水平面内放射指向性を示す図である。
【図１４】第四の実施例における直交座標のときの染色体の構成例を示す図である。
【図１５】第四の実施例における最適化方法を実施した結果得られたアンテナ構造における無給電素子の配置を示す図である。
【図１６】第四の実施例における最適化方法を実施した結果得られたアンテナ構造における水平面内放射指向性
【図１７】従来の４素子八木・宇田アンテナの構造を示す図である。
【図１８】従来の４素子八木・宇田アンテナの染色体の構成を示す図である。
【図１９】従来のＣｒｏｏｋｅｄ−Ｗｉｒｅ ｇｅｎｅｔｉｃ ａｎｔｅｎｎａを示す図である。
【図２０】従来のマルチセクタモノポール八木・宇田アンテナの構造を示す図である。
【図２１】従来の移動通信基地局のための等ビーム２周波共用アンテナを示す図である。
【図２２】従来の従来のＧＡを用いた八木・宇田アンテナの設計方法を示すフローチャートである。
【図２３】本発明の第１の実施例に係るアンテナ最適設計方法を示すフローチャートである。
【符号の説明】
ｈｉ：素子長
（ｘｉ，ｙ，ｚｉ）：素子の直交座標、
（ｒｉ，θｉ）：素子の円筒座標
φｉ：位相
Ｏ（ｘ）：対象関数
Ｓ：セクタ数[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an antenna design method effective for the design of miniaturization of a sector antenna and a storage medium storing a program for implementing the method, and more particularly, to miniaturization optimization of a high-speed wireless LAN sector antenna, and mobile communication. The present invention relates to an antenna design method effective as a method for optimizing the miniaturization of a base station dual frequency or triple frequency shared antenna and a storage medium storing a program for implementing the method.
[0002]
[Prior art]
A conventional method of designing a Yagi-Uda antenna using Genetic Algorithms (hereinafter, referred to as “GA”) is described in, for example, Reference 1: “Eric A. Jones, William T. Joines:“ Design of Yagi-Uda Antennas Generating Electronics, "IEEE Trans. AP Vol. 45, No. 9 Sept. 1997".
[0003]
FIG. 17 is a configuration diagram showing a configuration example of a four-element Yagi-Uda antenna designed using a conventional method. In the conventional method of optimizing the Yagi-Uda antenna by GA, the number of elements is determined in advance (the number of elements = 4 in the case of this configuration example), and then the element length of each element (L1, L2, L3 in this configuration example) , L4) and element intervals (S2, S3, S4 in this configuration example) are unknown, and these unknowns are arranged as an unknown number sequence to form a chromosome.
[0004]
FIG. 18 shows a configuration example of a chromosome constituted by unknown numbers in this configuration example. As shown in FIG. 18, the element lengths (L1,..., L4) and the element intervals (S2,..., S4) are each composed of 6-bit binary data, and the element lengths L1, L2, L3, L4 are The chromosome of the antenna configuration example is formed by alternately arranging four 6-bit binary data representing the three-bit binary data representing the element intervals S2, S3, and S4.
[0005]
FIG. 22 is a flowchart showing a method for designing a Yagi-Uda antenna using the conventional GA. First, an unknown number sequence (= chromosome) composed of element intervals and element lengths as shown in FIG. 18 is generated by a predetermined number using random numbers, and the generated chromosomes are set as initial values ( Step 1).
[0006]
Next, using the chromosomes generated as these initial values, the antenna characteristics of each chromosome are obtained by simulation using the moment method (Step 2).
[0007]
Then the following formula
O (x) = aG (x) -b | 50-Re (Z (x)) | -c | Im (Z (x)) |
The chromosomes are ranked using the target function O (x) given by (Step 3). Here, G (x) is directional gain, Z (x) is impedance, Re (x) represents a real part of x, Im (x) represents an imaginary part of x, and a, b, c Is a coefficient representing the importance of each item.
[0008]
Based on the ranking of each chromosome obtained by the target function, a genetic operation for creating a new generation chromosome from the ranked chromosome is performed (Step 4).
[0009]
Finally, it is determined whether or not the chromosome of the new generation satisfies the convergence condition (Step 5). If the convergence condition is satisfied, the process is terminated. To Step 5, and the same processing is repeated until the convergence condition is satisfied.
[0010]
In such a conventional antenna design method using GA, when a fixed-length chromosome is used, the number of elements must be fixed. Therefore, there is a limit in the flexibility of the antenna structure that can be input in the antenna design method using the conventional GA. Further, there has been no device for optimizing a structure having rotational symmetry and line symmetry. Further, there is a disadvantage that the optimization problem of a constant beam width and the optimization problem of miniaturization cannot be solved.
[0011]
Such a conventional method also has another problem. That is, in such a conventional design method, first, it is necessary to determine the number of elements and the basic configuration of the antenna (for example, whether to use a Yagi-Uda antenna or a corner reflector antenna), and the optimization is decided. It is only performed as a parameter study in the range of the number of elements and the structure. In particular, in the case of the Yagi-Uda antenna, the phase and amplitude of the current induced in the director and the reflector affect the antenna characteristics (CG Publishing: Antenna Handbook, pp. 347-354). This is a value determined by a structure such as an element interval and an element length. On the other hand, the gain of the Yagi-Uda antenna tends to increase as the array length increases. For this reason, if the antenna is optimized under the condition that the selectable range of the number of elements and the element interval is limited, the result is that the gain is increased as a result of simply increasing the array length of the antenna. There is a problem that the current distribution on the device is not always the optimum result, that is, the result is that the amplitude is maximum and the complex component of the phase is small.
[0012]
Another example of an antenna design method using a GA is shown in FIG. FIG. 19 is called Crooked wire Genetic Antenna (Edward E. Altshuller, Derek S. Linden: “Wire-Antenna Designs Using Using Genetic Algorithms, see IE Energy and Genes, IE 39A. ), And a monopole antenna composed of wires formed by connecting arbitrary points in a design space composed of a 0.5λ (λ: wavelength) cube. In this antenna design method, there is a certain degree of freedom in the antenna structure that can be selected, and it is possible to optimize the structure as well as the parameter study. However, the number of wires is determined in advance and cannot be applied to optimization of a sector antenna.
[0013]
Next, a conventional method of designing a sector antenna without using a GA will be described. FIG. 20 is a diagram showing a multi-sector monopole Yagi-Uda antenna (MS-MPYA) proposed as a small sector antenna for a 19 GHz band wireless LAN. MS-MPYA has many design parameters as shown in the following [Table 1]. Conventionally, "Maruyama, Uehara, Kagoshima: IEICE B-ll," Using Monopole Yagi-Uda Antenna " Analysis and Design of Small Multi-Sector Antenna for Wireless LAN, Vol. J80-Bll, pp. 424-433, No. 5, May 1997 "and" 1996 IEICE General Conference: B-105 Maruyama, Uehara, Kagoshima: Analysis of sectorized monopole Yagi array with metal fins installed on a finite ground plane ”, etc., the array length for each sector, the radius of the cylindrical reflector, the element length, and the For the height and the length of the metal fins, parameter studies were separately performed to optimize them.
[0014]
[Table 1]

[0015]
Further, as shown in the above “Analysis and Design of Small Multi-Sector Antenna for Wireless LAN Using Monopole Yagi-Uda Antenna”, MS-MPYA arranges sectors and makes effective use of adjacent arrays to achieve beam width. , The gain is increased and the characteristics are improved, but the optimum arrangement of the parasitic element, which generalizes this, has not yet been achieved. When there are many types of design conditions as shown in the following [Table 2], characteristics were examined for each one and optimization was performed. For this reason, a lot of labor and time were required.
[0016]
[Table 2]

[0017]
FIG. 21 shows an example of a conventional dual-frequency antenna for a mobile communication base station antenna (Suzuki, Kagoshima "Corner reflector antenna with arbitrary beam width and two frequency bands", Vol. J75-Bll, pp. 950-). 956, No. 12, December 1992). In the conventional design method, first, the structure of the reflector is determined. For example, a corner reflector as shown in FIG. 21 or Tamami Maruyama and Kenichi Kagoshima: “Corner reflector antenna for dual use with equal beams”, NTTR & D, Vol. 42 No. 9 1993. Pp. 1135-1146, a cylindrical reflector as described in September 1993, and the like. After that, it was designed by conducting parameter studies on the corner angle, corner length, etc. of the corner reflector antenna.
[0018]
This also requires a great deal of labor and time for optimization under limited conditions, as in the above-described conventional example.
[0019]
[Problems to be solved by the invention]
As described above, both the optimization method using the conventional GA and the sector antenna design method are limited to the structure first, and then are subjected to parameter studies, and the structure is left to the decision of the designer. Had been. Further, in the conventional method for optimizing the Yagi-Uda antenna using the GA, the number of elements cannot be made variable, and the arrangement of the parasitic element has no flexibility. Further, in the conventional method, there is a drawback that no means has been devised for the optimization means for miniaturizing the sector antenna.
[0020]
[Means for Solving the Problems]
The present invention is configured as follows for the purpose of solving the above problems.
[0021]
The present invention relates to an antenna optimum design method for designing an antenna structure using chromosomes (for example, data structures imitated by chromosomes and subjected to genetic processing), and includes at least a position and an element length. A first step (S1) of generating a predetermined number of chromosomes consisting of one element information, and, among element information included in each of the predetermined number of chromosomes, an element whose element length is equal to or less than a predetermined value. A second step (S2) of deleting and removing one of two or more pieces of element information having overlapping position information while leaving one element information; and a chromosome obtained by the second step. The third step (S3) of forming elementary information for one sector and the fourth step of generating element information of the antenna based on each of the element information for one sector obtained in the third step. Engineering (S4), a fifth step of calculating antenna characteristics based on each of the element information of the antenna, and (S5), and evaluating each of the element information of the antenna based on the objective function (O (x)). Thus, a sixth step (S6) of ranking each of the predetermined number of chromosomes, and preferentially storing excellent ones for each characteristic based on the rank assigned to the predetermined chromosome Performing a genetic operation while generating a new generation of gene group (S7), and determining whether the new generation of gene group satisfies a predetermined convergence condition. If the condition is satisfied, the antenna design is terminated, and if the predetermined convergence condition is not satisfied, the first to seventh steps are repeated based on the new generation of gene groups, an eighth step (S8). With It is characterized by that.
[0022]
The present invention can also be implemented as a recording medium that stores a program for performing the first to eighth steps.
[0023]
In an embodiment of the present invention, the position information of the element is configured using cylindrical coordinates (ri, θi,) in the first step, and in the fourth step, the element information of the antenna is stored for one sector. The information may be generated using a mirror image of the element information.
[0024]
In another embodiment of the present invention, in the seventh step, among the predetermined number of chromosomes, a predetermined number of chromosomes are selected from those having a high rank, and the selected predetermined number of chromosomes are selected. The genetic operation may be performed so as to always remain in the next generation.
[0025]
In still another embodiment of the present invention, the objective function
O (x) = aG (x) -b | A-Re (Z (x)) | -c | Im (Z (x)) | -d | B-HBW | -eAL,
(Where G (x) is the directional gain of the antenna, Z (x) is the impedance of the antenna, HBW is the half-width of the radiation directivity in the horizontal plane of the antenna, AL is the array length (AL = r_max), T (x) indicates an auxiliary function provided so as not to cause a beam crack, and l (x) indicates the number of elements after removing elements having overlapping or short elements, a, b, c, and d, e, f, and g are coefficients representing the importance of each item and are not limited to real numbers. A and B may be target functions having desired impedance values and half-value widths.
[0026]
【Example】
Embodiments of the present invention will be described in detail with reference to the accompanying drawings.
[0027]
[First Embodiment]
First, a first embodiment of the present invention will be described with reference to FIGS. 1, 2, 3, 4, 5, 6, 7, and 23. FIG.
[0028]
FIG. 1 shows a model having cylindrical coordinates for miniaturizing a sector antenna. In FIG. 1, the symbol “示 し” indicates the position of the feed element # 1 and is located at an arbitrary distance r from the origin on the axis of the polar coordinate angle φ = 0 °. The symbol "●" indicates parasitic elements # 2 to #j.
[0029]
Each element has a parameter h_i, R_i,φ_iIs defined by Where h_iIs the element length of the i-th element, r_iIs the distance of the i-th element from the origin, φ_iRepresents the angle of the polar coordinate of the i-th element (i = # 1 to #j).
[0030]
In addition, a chromosome (a data structure to be subjected to genetic processing), which is an unknown number sequence of a GA algorithm representing a sector antenna, is configured as shown in FIG. That is, the chromosome has an element length h_{# 1}And the distance r from the origin_{# 1}And the element length h (each corresponds to the element # 1)._iAnd the distance r from the origin_#IAnd the angle φ of the polar coordinates_#IAnd a parasitic element portion (corresponding to the parasitic elements # 2 to #j) composed of (j-1) combinations of
[0031]
Here, the antenna is divided into a plurality of sectors as shown in FIG. The number of sectors is determined in advance. The number of sectors is S, and the maximum value φmax of φ used for analysis of antenna characteristics is 180 / S ／. Also, the distance r from the origin of the antenna element_#IAbout the maximum value r_maxAnd the minimum value r_minIs set. Element length h_#IAlso the maximum value h_maxIs 0.4λ (λ: wavelength), the minimum value h_minTo 0.07λ. Thereby, the excitation condition of the antenna is determined.
[0032]
As described above, the structure of the antenna is expressed in polar coordinates, sectorized, its excitation conditions are determined, and the element values (h_#I, R_#I,φ_#I) Is determined using random numbers, and a chromosome as shown in FIG. 1 is created. The chromosomes created in this way are prepared for a predetermined number of individuals (FIG. 23: step S1).
[0033]
In the following description, for convenience of explanation, the number of sectors: S = 12, the minimum value of the distance from the origin of the antenna element: r_min= 0 [λ] (λ: wavelength), maximum value: r_max= 5 [λ] and the optimum design of the antenna is described under the condition that the number N of chromosomes is 120, but the present invention is not limited to such a condition, and the desired antenna characteristics are not limited. It is possible to set and execute these conditions arbitrarily according to the conditions.
[0034]
Next, the process proceeds to step S2 in FIG. Here, the number of elements is determined for each of the 120 chromosomes prepared in step S1 according to the conditions defined below.
[0035]
First, the element length h_#IIs smaller than 0.1λ, the element is deleted, and when a part of the elements is overlapped, the overlapped element is removed. The elements remaining after this removal are the number of elements on the chromosome. This element removal processing is performed for each chromosome. Determining the number of elements for each chromosome in this manner means that, for the antenna corresponding to each chromosome, half the element position of one sector centered on φ = 0 °.
[0036]
In this way, in an antenna structure obtained by setting an unknown number by a random number for each chromosome, the parasitic element operates both as a director and a reflector according to the element length of the parasitic element and a possible coordinate position. I do. Therefore, the antenna structure corresponding to the chromosome obtained in the above step 2 is composed of an antenna with a reflector such as a corner reflector antenna and an antenna having a parasitic element (wave director) such as a Yagi-Uda array antenna. It can be regarded as a fusion model.
[0037]
Next, step S3 in FIG. 23 will be described. In this step, a mirror image of the element centered on φ = 0 ° is created based on the element position and the like of half of one sector centered on φ = 0 ° determined in step S2, thereby forming one sector. The element arrangement of the other half is determined, and the element positions of the entire one sector are determined by both of them. If a part of an element overlaps a sector boundary when a mirror image is created, a process of moving the center of the element to the sector boundary is performed. By performing the above processing for all 120 chromosomes, 120 element arrangements for one sector are created.
[0038]
Next, based on the element arrangement of one sector created in step S3, the element arrangement of the remaining sector (11 sectors) is calculated as 2 × φ around the origin of the polar coordinates._max(= 360 ° / S), the structure of one sector antenna as a whole can be set by performing primary conversion or rotational movement. By performing the processing by the primary conversion or the rotational movement for each of 120 chromosomes, a sector antenna structure for 120 pieces is created. This is step S4. When the process of step S4 is completed, 120 sector antenna analysis models with variable numbers of elements are determined.
[0039]
Next, for each of the 120 sector antenna analysis models determined in step S4, the antenna characteristics are calculated using the Implemented Circuit Theory (ICT method) (step S5). I. C. For details of the T method, see, for example, a document “Inagaki, Sekiguchi:“ Strict design to maximize the directional gain of an array using a linear antenna as an element ””, IEICE (B), 53-B, No. 11, pp 687-692, 1970 "and" N. Inagaki: "An improved circuit theory of a multi-element antenna. ", IEEE Trans. On AP, Vol. 17, No. 2, p. 120, March, 1969".
[0040]
After the antenna characteristics of all the models are calculated in step S5, 120 antenna characteristics are calculated using the target function O (x) and the fitness function f (x) defined by the following equations (8) and (9). Are ranked (step S6).
[0041]
O (x) = aG (x) -b | 50-Re (Z (x)) | -c | Im (Z (x)) | -d | 30-HBW | -eAL + fT (x) -gl (x ) ・ ・ ・ ・ ・ ・ (8)
f (x) = O (x)-(O_avg−ασ) (9)
Here, G (x) in the equation (8) is the directivity gain, Z (x) is the impedance, HBW is the half value width of the radiation directivity in the horizontal plane, AL is the array length (AL = r_max), T (x) indicates an auxiliary function provided so as not to cause a beam crack, and l (x) indicates the number of elements after overlapping or short elements are deleted. A, b, c, d, e, f, and g are coefficients representing the importance of each item.
[0042]
In addition, O in equation (9)_avgIndicates the average value of O (x), and σ indicates the standard deviation of O (x). Further, α in the expression (9) is a coefficient for weighting σ and takes a value in a range of 1 ≦ α ≦ 3.
[0043]
Looking at equation (8), it was aimed to increase the priority of those having a high gain, those having an impedance close to 50Ω, those having a half-width of radiation directivity in a horizontal plane of 30 °, and those having a short array length. You can see that it is a function.
[0044]
Now, it is assumed that a target in the analysis of the array antenna characteristics is one in which the half value width of the radiation directivity in the horizontal plane satisfies 30 ° and satisfies the condition of minimizing the array length. Care must be taken in the selection of the coefficient because another item may be prioritized depending on how to select the coefficient in equation (8). Here, for convenience of explanation, it is assumed that the coefficient a of the equation (8) is set to 200, b and c are set to 2, d is set to 100, e is set to 5, and g is set to 1; Is not limited to this, and the value of each coefficient can be set arbitrarily.
[0045]
Next, step S7 will be described. From the 120 chromosomes ranked in step S6, a genetic operation described below is performed while preferentially preserving one having excellent antenna characteristics.
[0046]
Here, in carrying out this genetic manipulation, there are two methods, a roulette preservation method and an elite preservation method. Details of these methods are described in the reference "DE Goldberg," Genetic algorithms in search, optimization and machine learning, "Reading, Mass .: Addison-Wesley, 1989." The outline is described below.
[0047]
First, the roulette storage method will be described.
[0048]
A roulette that is configured to increase the probability that a higher-ranked chromosome is selected from the ranked chromosomes is prepared. Turning this roulette selects one chromosome. By repeating the selection of chromosomes by roulette 120 times, an operation of newly selecting 120 chromosomes is performed. In the selection operation, there is a high possibility that a chromosome with a higher rank is selected more frequently, and a chromosome with a higher rank may be selected a plurality of times. On the other hand, chromosomes with lower ranks are less likely to be selected by the roulette, and will gradually disappear. As a result, chromosomes with a higher rank are preferentially selected.
[0049]
From the plurality of chromosomes thus selected, a parent of crossing (exchanging half elements of the chromosome with each other) is selected at a predetermined mating rate. The mating rate can be set arbitrarily. For example, when the mating rate is set to 0.5, 60 chromosomes are selected from the 120 chromosomes selected by the roulette. As a method for selecting the parent, there are various selection methods such as randomly selecting from 120 chromosomes, 60 chromosomes selected first, and any method may be adopted.
[0050]
After selecting 60 chromosomes to be mating parents in this way, these parents are then paired two by two, each chromosome is divided in half, and mated with each other.
[0051]
Further, a mutant chromosome is created from the 120 chromosomes thus obtained with a predetermined probability. This mutation is performed by appropriately altering some of the selected chromosomal elements.
[0052]
A new generation of chromosome groups is generated by the above-described genetic operation processing.
[0053]
This is the genetic operation for creating a new generation of chromosomes in step S7.
[0054]
As described above, in step S7, chromosomes are preferentially ranked such that the better the antenna characteristics, the higher the probability of being selected and left, so that the chromosomes of the new generation have excellent characteristics. Is more likely to remain.
[0055]
Next, the elite preservation method will be described.
[0056]
As described above, by performing the processing in step S6, each of the 120 chromosomes is ranked. From the ranked 120 chromosomes, a predetermined number of chromosomes are first sorted in descending order of rank. And select it as the chromosome to be preserved without change. For example, assume that the top 20 chromosomes are selected. The remaining number to be selected to prepare 120 chromosomes to be the parent of the cross for performing the genetic operation is 120−20 = 100. Here, similar to the roulette used in the above-described roulette preservation method, a roulette having a high probability of selecting a higher-ranked chromosome from the 120 ranked chromosomes is prepared, and this is set to 100 The operation is repeated to select 100 new chromosomes.
[0057]
Next, a parent of the mating (exchanging half elements of the chromosome with each other) is selected from the 100 matings at a predetermined mating rate. For example, if the mating rate is 0.5, 50 chromosomes will be selected. When selecting this parent, a random selection, the first selected 50 chromosomes, and various selection methods are conceivable, but any method may be used.
[0058]
After the fifty chromosomes to be selected as mating parents are selected in this way, these parent chromosomes are paired two by two, and each chromosome is divided in half and crossed into pairs.
[0059]
Further, a mutant chromosome is created with a predetermined probability from the 100 chromosomes thus obtained. This mutation is performed by appropriately altering some of the selected chromosomal elements.
[0060]
In this way, a total of 120 new chromosomes, including the 20 chromosomes initially selected as the chromosomes to be left with certainty and the 100 chromosomes selected and genetically manipulated thereafter, constitute 120 new chromosomes of the new generation. Born as a chromosome group. This is an example of the formation of a new generation of chromosomes when using the elite preservation method.
[0061]
In addition to the above, in the genetic manipulation described above, it is conceivable to further set conditions so as to obtain a better result. However, since it is not the essence of the present invention, the details thereof are described above. See the referenced references.
[0062]
The above operation is performed in step S7 of FIG.
[0063]
Next, the convergence degree of the chromosome group is determined in step S8, and if the convergence degree set in advance is satisfied, the process proceeds to step S9 to end the operation. On the other hand, if the convergence degree is not satisfied, the process returns to step S2, and the above-described operation is repeated. To determine the degree of convergence, the O_avgCan be checked whether or not is in a convergent state, or it can be determined by decreasing the σ (to approach 0).
[0064]
In addition, as a criterion for determining whether or not the convergence state is established, an array antenna having better characteristics is designed by adding a determination condition for determining whether or not the antenna characteristics have converged to a desired value with respect to various parameters. It becomes possible.
[0065]
Hereinafter, the results of implementing the antenna optimum design method according to the present embodiment when the roulette preservation method is used will be described. In this embodiment, the mating rate was set to 0.5, the mutation rate was set to 0.001, and the coefficient a in the above-mentioned equation (8) was set to 200, and b and c were set to 2 as setting parameters of the roulette preservation method. , D is 100, e is 5, and g is 1, and chromosomes of the first to 90th generations are formed and analyzed according to the antenna optimum design method according to the present embodiment.
[0066]
FIG. 3 is a graph showing the average value of the target function O (x) for each generation. It can be seen that the average value of the target function converges with each generation.
[0067]
FIG. 4 shows the value of the half width of the radiation directivity in the horizontal plane for each generation, and it can be seen that the values converge to the desired design value of 30 °.
[0068]
FIG. 5 shows how the average value of the input impedance for each generation converges. The reactance component X converges near a desired value of 0 [Ω], but the resistance component R becomes 100 [Ω]. ] And converges slightly away from the desired 50 [Ω]. This is presumably because the value of the coefficient b is lower than the others in the equation (8), so that the importance of the resistance component item is reduced, and this element is neglected.
[0069]
FIG. 6 shows the relationship between the arrangement of parasitic elements obtained by the algorithm of the present invention shown in FIG. 23 and the element length of each element. The antenna structure shown in FIG. 6 has a shape that cannot be derived by a conventional method of performing a parameter study by arranging sectors of a Yagi-Uda antenna and a corner reflector antenna. FIG. 7 shows the radiation directivity in the horizontal plane of this antenna structure. As can be seen from FIG. 7, a result is obtained in which there is no deviation in the direction of the main beam, and a half-width of radiation directivity in the horizontal plane of 30 ° is realized.
[0070]
As described above, by using the method of the present invention, the application of Genetic Algorithms facilitates antenna design for miniaturization of a sector antenna including an antenna structure that is difficult with the conventional method.
[0071]
[Second embodiment]
Next, a second embodiment of the present invention will be described.
[0072]
The second embodiment is the same as the first embodiment except that the target function used in Step 6 is different.
[0073]
The target function in the second embodiment is expressed by the above equation (8) by using a term + kSL (k is a coefficient, SL is a side lobe level) for reducing the side lobe and a term + hBL (h is a coefficient, and BL is a back side lobe) for reducing the back lobe. The following equation (10) is obtained by adding a function fT (x) and a term -gL (x) that restricts the number of elements, which is provided so as not to cause a beam crack.
[0074]
O (x) = aG (x) −b | A−Re (Z (x)) | −c | Im (Z (x)) | −d | B−HBW | −eAL
+ KSL + hBL ... (10)
[0075]
The optimized antenna structure was obtained by executing the antenna optimum design method according to the second embodiment of the present invention. FIG. 8 shows an arrangement example of the parasitic element in the optimized antenna structure, and FIG. 9 shows its radiation directivity in a horizontal plane. As can be seen from FIG. 9, the side lobes and back lobes are reduced by the side lobe and back lobe reduction terms added in equation (10).
[0076]
Here, in the present embodiment, the target is to set the side lobe to -10 dB or less in the range of 45 ° <φ <150 °. In order to achieve such a goal, the evaluation value is increased when the side lobe level satisfies −10 dB or less (kSL = 1000.0 in this embodiment), and the evaluation value is decreased when the side lobe level does not satisfy −10 dB or less. (In this example, kSL was set to −100.0). In addition, the back lobe is targeted to have a level of −15 bB or less in the range of 150 ° <φ <180 °. When the back lobe level satisfies −15 bB or less, the evaluation value is increased. Stipulated that the evaluation value be lowered. Thus, in the present invention, a discrete function may be included in the objective function O (x). Such a method is more advantageous than other methods when solving a problem that cannot be expressed by a continuous function.
[0077]
In the present embodiment, since the term l (x) relating to the number of elements is not provided in the target function O (x), the number of elements for each sector is slightly increased to 22.
[0078]
[Third embodiment]
In the first and second embodiments of the present invention, the position of each element is defined using a cylindrical coordinate system, but the position of each element is also defined using a rectangular coordinate system instead of the cylindrical coordinate system. The present invention can be implemented. A case using such an orthogonal coordinate system will be described below as a third embodiment.
[0079]
In the present embodiment using the rectangular coordinates, the definition of the chromosome in step S1 differs from the flow of the present invention shown in FIG. 23 as described below, and the concept of sectorization is removed (step S3). , S4 change), the process of setting the element arrangement of the antenna changes, but the algorithm is the same as in the first and second embodiments.
[0080]
First, the configuration of the chromosome used in this example will be described. FIG. 10 is an example of the antenna configuration when the rectangular coordinates are used. The antenna has an element length H fixed at the origin (coordinates (0, 0))._{# 1}, And the parasitic reflectors # 2 to #m having unknown coordinates (x, y) and element length H, and unknown coordinates (x, y) and element length H. Passive waveguides # m + 1 to #n are provided. Here, m is a natural number of 2 or more, and n is a natural number of 3 or more. In the antenna configuration example shown in FIG. 10, the parasitic reflector is only # 2 (m = 2), and the passive directors are # 3 to # 6 (n = 6).
[0081]
In FIG. 10, the symbol “示 し” indicates the position of the feed element # 1 and is arranged at the origin. # 2 to #m are reflectors each formed of a parasitic element disposed on the back or side of the feed element. A reflector can be identified with a reflector when the element spacing is 0.1λ or less (Kraus JD: Antennas Second Edition, p. 556, McGRAW-HILL, 1988.). Symbols ● indicate directors # m + 1 to #n constituted by parasitic elements.
[0082]
FIG. 11 shows a chromosome corresponding to an antenna composed of the feed element # 1, the parasitic reflectors # 2 to #m, and the parasitic directors # m + 1 to #n as described above. The chromosome is composed of a feed element unit, a reflector unit, and a director unit. The feed element section has an element length H of the feed element._{# 1}Is stored. The reflector section includes an x-coordinate value, a y-coordinate value, and an element length (RX) of each of the parasitic reflectors # 2 to #m._i, RY_i, H_i) Is stored (i is # 2 to #m). Further, the director section has the x-coordinate value, y-coordinate value, and element length (DX) of each of the parasitic directors # m + 1 to #n._i, DY_i, H_i) Is stored (i is # m + 1 to #n). In the above example, the reflector and the director are provided independently of each other, but when it is not necessary to divide the conditions given to the maximum value, the minimum value, the number of bits, etc. of the reflector and the director, respectively. Instead of being separated as described above, they may be integrated as one kind of parasitic element portion.
[0083]
In the above chromosome, the characteristics of each element are (RX_i, RY_i, H_i) Or (DX_i, DY_i, H_i). Each unknown RX_i, RY_i, DX_i, DY_i, H_iAre determined as follows, and the excitation conditions are determined.
[0084]
RXi: minimum value RX_min= 0: maximum value RX_max  = -0.25
RYi: minimum value RY_min= 0; maximum value RY_max  = 1.0,
DXi: minimum value DX_min= 0: maximum value DX_max  = 5.0,
DYi: minimum value DY_min= 0; maximum value DY_max  = 1.0,
Hi: minimum value H_min= 0; maximum value H_max  = 0.35
As described above, the structure of the antenna is represented by rectangular coordinates, the excitation conditions are determined, and the element values (RX_i, RY_i, H_i) And (DX_i, DY_i, H_i) Is determined using random numbers, and a chromosome is created. The chromosomes created in this way are prepared for a predetermined number of individuals (step S1).
[0085]
Next, the number of elements in each chromosome is determined (step S2). In each chromosome, when the element values included in different elements are the same, that is, when the element structures partially overlap, or when the element length H_iIf is smaller than a predetermined value, the element is removed from the chromosome.
[0086]
A mirror image is created of the element group obtained in this manner with respect to the X axis so that the radiation directivity in the horizontal plane is targeted in the front direction. At this time, the elements on the X axis are deleted. This process is performed for each chromosome of the predetermined number of individuals, and a predetermined number of antenna analysis models with variable numbers of elements are determined (step S3 ').
[0087]
Next, for all of the antenna analysis models, the antenna characteristics are calculated using the Implemented Circuit Theory (ICT method) (step S5), and based on the calculation of the antenna characteristics, the target function O (x ) And the fitness function f (x) are used to rank all chromosomes (step S6), and among the ranked chromosomes, those having excellent antenna characteristics are preferentially stored, Genetic operations are performed to create a new generation of chromosomes (step S7). Finally, the degree of convergence of the new generation chromosomes is determined. If the degree is not satisfied, the above-described processing (steps S2 to S7) is repeated.
[0088]
Next, the specific results of using the chromosomes based on the orthogonal coordinates will be described.
[0089]
In this embodiment, the following equation (11) was used as the target function O (x).
[0090]
O (x) = aG (x) -b | 50-Re (Z (x)) | -c | Im (Z (x)) | -d | 30-HBW | -eAL + fT (x) (11) )
FIG. 12 shows an arrangement of the parasitic element obtained as a result of performing the third embodiment using the above (11). This is because, among the structures obtained by dispersing parasitic elements in a given space, those having a beam width of radiation directivity in a horizontal plane of 30 ° and an impedance close to 50 [Ω] can be used as an array length. Is the result of solving the problem of "What is the arrangement that can reduce the size?" FIG. 13 shows the radiation directivity in the horizontal plane with respect to the obtained antenna structure having the arrangement of the parasitic elements. As can be seen from this figure, the antenna structure satisfies a beam width of 30 °.
[0091]
As described above, the orthogonal coordinate system GA-ICT as in the third embodiment of the present invention is also an effective means for considering the arrangement of elements for obtaining desired characteristics. However, when considering an arrangement that is small when divided into sectors, it is difficult to set the range of the Y coordinate in the case of rectangular coordinates. Therefore, it is preferable to use cylindrical coordinates, which is the first embodiment of the present invention. desirable. FIG. 13 shows that side lobes and back lobes are generated by about −10 dB, which is defined in the objective function O (x) by a rule for reducing them (in the second embodiment described above). kSL, hBL).
[0092]
[Fourth embodiment]
Next, a fourth embodiment of the present invention will be described. This embodiment is the same as the third embodiment except that the reflector and the director are not separated but are simply integrated as a parasitic element.
[0093]
FIG. 14 shows the structure of the chromosome of this example. The chromosome is composed of a feed element section and a parasitic element section. The feed element section has an element length H of the feed element._{# 1}And the parasitic element section stores the x-coordinate value, y-coordinate value, and element length (DX) of each of the parasitic elements # 2 to #n._i, DY_i, H_i) Is stored (i is # 2 to #n).
[0094]
The result of executing the antenna optimization design method according to the fourth embodiment is shown below. Where DX_i, DY_i, H_iAre the same as those in the third embodiment described above, and the following equation (12) is used instead of equation (11) as the target function O (x).
[0095]
O (X) = aG (x) -b | 50-Re (Z (x)) | -c | Im (Z (x)) | (12)
The optimization of the antenna in this embodiment is based on the following description. Does it operate as an antenna that can provide high gain? " FIGS. 15 and 16 show the results calculated using the optimization method of the present embodiment. FIG. 15 is an arrangement example of a parasitic element in an antenna structure obtained as a result of performing the optimization method. The radiation directivity in the horizontal plane of the antenna having the parasitic element arrangement is as shown in FIG. As can be seen from FIG. 16, in the optimized antenna, a half value width of 22 ° and a gain of 12 dB were obtained. The reason why the gain is slightly low for the array length of 5λ is that no element is provided behind the feed element.
[0096]
The features of the first and second embodiments of the present invention are that, when compared with the conventional design method shown in FIG. 22, (1) generalize the chromosome using the coordinate system and Not limited to a specific structure such as length, (2) adopting cylindrical coordinates, (3) considering the mutual coupling between sectors, reducing the amount of calculation by making full use of mirror images and linear transformation, ( 4) The objective function is to adopt a term for reducing the size of the sector antenna, which prioritizes a beam width close to a certain value and a small antenna size.
[0097]
The features of the third and fourth embodiments of the present invention are that, when compared with the conventional design method shown in FIG. 22, (1) the chromosome is generalized using the coordinate system and the conventional The point is not limited to a specific structure such as length, and (2) the amount of calculation is reduced by making full use of a mirror image.
[0098]
In addition, even if a plurality of feed elements are used in the present invention, or all the elements are changed from passive elements to feed elements, if necessary, unknown terms of phase and amplitude are added to the chromosome by the number of elements. By doing so, it goes without saying that the same effects as those shown in this specification can be obtained.
[0099]
【The invention's effect】
As described above, according to the present invention, the arrangement of elements of a sector antenna, particularly a Yagi-Uda antenna using an array of parasitic elements or a corner reflector type antenna in which a reflector can be approximated by a row of parasitic elements Can be optimized at a high degree of freedom so as to satisfy desired characteristics. In particular, according to the present invention, the number of elements can be changed to an unknown number while the chromosome structure remains fixed. By using the present invention, it is possible to optimize the problem of miniaturization of an arbitrary sector antenna and the problem of sharing an equal beam multi-frequency. This is an effective means for optimizing miniaturization of a multi-frequency antenna.
[0100]
Further, according to the present invention, analysis including the influence of mutual coupling occurring between sectors can be easily realized.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a position of an antenna element using cylindrical coordinates.
FIG. 2 is a diagram illustrating a configuration example of a chromosome when cylindrical coordinates are used.
FIG. 3 is a diagram illustrating a state of convergence of a target function for each generation.
FIG. 4 is a diagram showing a state of convergence of a beam width for each generation.
FIG. 5 is a diagram showing a state of convergence of an input impedance for each generation.
FIG. 6 is a diagram showing an arrangement of a parasitic element in the antenna structure obtained as a result of the embodiment of the present invention.
FIG. 7 is a diagram showing radiation directivity in a horizontal plane in the antenna structure obtained as a result of the embodiment of the present invention.
FIG. 8 is a diagram showing an arrangement of a parasitic element in an optimized antenna structure.
FIG. 9 is a diagram illustrating radiation directivity in a horizontal plane in an optimized antenna structure.
FIG. 10 is a diagram illustrating an antenna configuration example using rectangular coordinates.
FIG. 11 is a diagram showing a configuration example of a chromosome at the time of rectangular coordinates.
FIG. 12 is a diagram illustrating an arrangement of parasitic elements obtained as a result of performing the third embodiment.
FIG. 13 is a diagram showing the radiation directivity in the horizontal plane of the antenna structure obtained as a result of implementing the third embodiment.
FIG. 14 is a diagram showing a configuration example of a chromosome at the time of rectangular coordinates in the fourth embodiment.
FIG. 15 is a diagram showing an arrangement of parasitic elements in an antenna structure obtained as a result of performing the optimization method in the fourth embodiment.
FIG. 16 shows radiation directivity in a horizontal plane in an antenna structure obtained as a result of performing the optimization method in the fourth embodiment.
FIG. 17 is a diagram showing the structure of a conventional four-element Yagi-Uda antenna.
FIG. 18 is a diagram showing the structure of a chromosome of a conventional four-element Yagi-Uda antenna.
FIG. 19 is a diagram showing a conventional Crooked-Wire genetic antenna.
FIG. 20 is a diagram showing the structure of a conventional multi-sector monopole Yagi-Uda antenna.
FIG. 21 is a diagram showing an equal beam dual frequency shared antenna for a conventional mobile communication base station.
FIG. 22 is a flowchart showing a conventional method for designing a Yagi-Uda antenna using a GA.
FIG. 23 is a flowchart showing an antenna optimum design method according to the first embodiment of the present invention.
[Explanation of symbols]
hi: element length
(Xi, y, zi): rectangular coordinates of the element,
(Ri, θi): cylindrical coordinates of the element
φi: phase
O (x): target function
S: Number of sectors

Claims (21)

Feeding element, or antenna structure having a feeding element and a parasitic element, an antenna optimal design method of designing using a genetic algorithm,
The element is defined by element information including a position value and an element length represented by coordinate values without defining an initial structure of the antenna, and a first chromosome that generates a plurality of chromosomes including at least one element information as an unknown sequence Generation process,
For each of the plurality of generated chromosomes, if there is two or more element information whose position information overlaps in the element information included in the chromosome, leave one of the two or more element information overlapping. A deleting step of deleting other things by
A process of determining the antenna structure based on the element information remaining as a result of the deletion, a determining step of performing for each of the plurality of generated chromosomes,
A calculating step of calculating antenna characteristics for each of the plurality of determined antenna structures,
By evaluating each of the calculated plurality of antenna characteristics based on the target function, an imparting step of assigning a rank to the generated plurality of chromosomes,
A determination step of determining whether or not the plurality of chromosomes to which the rank is assigned satisfies a convergence condition,
A second generation step of performing a genetic operation on the plurality of chromosomes to which the rank has been assigned by the genetic algorithm based on the assigned rank, thereby generating a plurality of chromosomes of a new generation. Prepare
First, for the plurality of chromosomes generated by the first generation step, the deletion step, the determination step, the calculation step, the execution step and the determination step are sequentially executed,
When it is determined that the plurality of chromosomes do not satisfy the convergence condition by the determination step, the plurality of chromosomes of a new generation generated by the second generation step are determined until the convergence condition is determined. On the other hand, an antenna optimum design method characterized by repeatedly executing the deletion step, the determination step, the calculation step, the provision step, and the determination step. In the first generation step, the element is defined by element information including a position information and an element length represented by a distance from an origin of cylindrical coordinates without defining an initial structure of the antenna, and the element information is defined as at least one. The method according to claim 1, wherein a plurality of chromosomes are generated. In the first generation step, the element is represented by element information including a position coordinate and an element length represented by coordinate values of rectangular coordinates without defining an initial structure of the antenna, and a plurality of elements including at least one of the element information are included. 2. The method according to claim 1, wherein said chromosome is generated. The deletion step further includes, for each of the plurality of chromosomes generated in the first generation step, elements having element lengths equal to or less than a predetermined value among element information included in the chromosomes. Item 4. An antenna optimum design method according to any one of Items 1 to 3. When the antenna is configured by a plurality of regions having line symmetry or rotational symmetry,
The determining step includes:
Creating a mirror image of the element information remaining as a result of the deletion,
5. The method according to claim 1, further comprising: determining the antenna structure for each area from the created mirror image and the element information from which the mirror image is created. 6. The antenna optimal design method described in Crab. When the antenna is configured by a plurality of regions having line symmetry or rotational symmetry,
The determining step includes:
Primary conversion or rotational movement of the element information remaining as a result of the deletion,
4. The method according to claim 1, further comprising: determining the antenna structure for each area from the element information before the primary conversion or the rotational movement and the element information generated by the primary conversion or the rotational movement. 5. The antenna optimum design method according to any one of 5. The determining step includes:
The process of creating the mirror image, or the result of the primary conversion or rotational movement process, a step of determining whether there is an overlap of elements at the boundaries of the plurality of regions,
7. The antenna optimum design method according to claim 5, further comprising, when it is determined that the elements overlap, deleting one of the overlapping elements. The giving step of giving the rank,
G (x) is the directional gain of the antenna, Z (x) is the impedance of the antenna, HBW is the half-width of the radiation directivity in the horizontal plane of the antenna, AL is the array length, and T (x) is the beam splitting. When the set auxiliary function, l (x) is the number of elements after deletion, and a, b, c, d, e, f, and g are coefficients representing the importance of each item,
O (x) = aG (x) −b | A−Re (Z (x)) | −c | Im (Z (x)) | −d | B−HBW | −eAL
+ FT (x) -gl (x)
Prepare the target function O (x) represented by
8. The antenna optimum design method according to claim 1, wherein each of the calculated antenna characteristics is evaluated based on the objective function O (x). The giving step of giving the rank,
G (x) is the directional gain of the antenna, Z (x) is the impedance of the antenna, HBW is the half-width of the radiation directivity in the horizontal plane of the antenna, AL is the array length, and T (x) is the beam splitting. The set auxiliary function, l (x) is the number of elements after deletion, SL is the sidelobe level, BL is the backlobe level, and a, b, c, d, e, h, and k represent the importance of each item. When it is a coefficient,
O (x) = aG (x) −b | A−Re (Z (x)) | −c | Im (Z (x)) | −d | B−HBW | −eAL
+ KSL + hBL
Prepare the target function O (x) represented by
8. The antenna optimum design method according to claim 1, wherein each of the calculated antenna characteristics is evaluated based on the objective function O (x). The second generation step includes:
A step of selecting a first number of chromosomes in descending order of the assigned rank among the plurality of chromosomes to which the rank has been assigned, and saving the selected first number of chromosomes without change; ,
Generating a genetic operation by the genetic algorithm based on the assigned rank to the plurality of chromosomes to which the rank has been assigned, thereby generating the remaining second number of chromosomes not selected. Process and
10. The method according to claim 1, further comprising a step of constructing a plurality of chromosomes of a new generation from the stored first number of chromosomes and the generated second number of chromosomes. The optimal antenna design method according to any one of the above. A recording medium storing an antenna optimum design program for causing a computer to execute a process of designing a structure of an antenna having a feed element or a feed element and a parasitic element using a genetic algorithm,
The element is defined by element information including a position value and an element length represented by coordinate values without defining an initial structure of the antenna, and a first chromosome that generates a plurality of chromosomes including at least one element information as an unknown sequence Generation processing,
For each of the plurality of generated chromosomes, if there is two or more element information whose position information overlaps in the element information included in the chromosome, leave one of the two or more element information overlapping. Delete processing to delete other things
A process for determining the antenna structure based on the element information remaining as a result of the deletion process, a determination process for each of the generated chromosomes,
A calculation process of calculating antenna characteristics for each of the determined plurality of antenna structures,
By evaluating each of the calculated plurality of antenna characteristics based on the target function, an assigning process of assigning a rank to the plurality of generated chromosomes,
A determination process of determining whether or not the plurality of chromosomes to which the rank has been applied satisfies a convergence condition,
A second generation process of performing a genetic operation by the genetic algorithm on the plurality of chromosomes to which the rank has been assigned based on the assigned rank, thereby generating a plurality of chromosomes of a new generation. Prepare
First, the plurality of chromosomes generated by the first generation processing, the deletion processing, determination processing, calculation processing, giving processing and determination processing in order to cause the computer,
When it is determined that the plurality of chromosomes do not satisfy the convergence condition by the determination process, the plurality of chromosomes of the new generation generated by the second generation process until the convergence condition is determined to be satisfied. On the other hand, a recording medium storing an antenna optimum design program characterized by causing the computer to repeatedly execute the deletion processing, the determination processing, the calculation processing, the assignment processing, and the determination processing. The first generation processing defines the element by element information including a position information and an element length represented by a distance from an origin of cylindrical coordinates without defining an initial structure of the antenna, and defines the element information by at least one. A storage medium storing the antenna optimum design program according to claim 11, wherein a plurality of chromosomes including one are generated. In the first generation processing, the element is represented by element information including an element length and a position coordinate represented by a coordinate value of rectangular coordinates without defining an initial structure of the antenna, and a plurality of elements including at least one of the element information are included. A recording medium storing the antenna optimum design program according to claim 11, wherein the chromosome is generated. The deletion process further includes, for each of a plurality of chromosomes generated by the first generation process, among device information included in the chromosome, those having a device length equal to or less than a predetermined value are further deleted. Item 14. A recording medium storing the antenna optimum design program according to any one of Items 11 to 13. When the antenna is configured by a plurality of regions having line symmetry or rotational symmetry,
The determination process includes:
Processing to create a mirror image of the element information remaining as a result of the deletion,
The method according to any one of claims 11 to 14, further comprising: a process of determining the antenna structure on a region-by-region basis from the created mirror image and the element information from which the mirror image was created. A recording medium storing the antenna optimum design program according to any of the above. When the antenna is configured by a plurality of regions having line symmetry or rotational symmetry,
The determination process includes:
Primary conversion or rotational movement of the element information remaining by the deletion processing;
12. The method according to claim 11, further comprising: a process of determining the antenna structure for each area from the element information before the primary conversion or the rotational movement and the element information generated by the primary conversion or the rotational movement. 15. A recording medium storing the antenna optimum design program according to any one of 15. The determination process includes:
The process of creating the mirror image, or the result of the primary conversion or the rotational movement process, a process of determining whether there is an overlap of elements at the boundaries of the plurality of regions,
17. The recording medium according to claim 15, further comprising: a process of deleting one of the overlapping elements when it is determined that the elements overlap. The giving process of giving the rank,
G (x) is the directional gain of the antenna, Z (x) is the impedance of the antenna, HBW is the half-width of the radiation directivity in the horizontal plane of the antenna, AL is the array length, and T (x) is the beam splitting. When the set auxiliary function, l (x) is the number of elements after deletion, and a, b, c, d, e, f, and g are coefficients representing the importance of each item,
O (x) = aG (x) −b | A−Re (Z (x)) | −c | Im (Z (x)) | −d | B−HBW | −eAL
+ FT (x) -gl (x)
Prepare the target function O (x) represented by
18. The recording medium storing the antenna optimum design program according to claim 11, wherein each of the calculated antenna characteristics is evaluated based on the objective function O (x). The giving process of giving the rank,
G (x) is the directional gain of the antenna, Z (x) is the impedance of the antenna, HBW is the half-width of the radiation directivity in the horizontal plane of the antenna, AL is the array length, and T (x) is the beam splitting. The set auxiliary function, l (x) is the number of elements after deletion, SL is the sidelobe level, BL is the backlobe level, and a, b, c, d, e, h, and k represent the importance of each item. When it is a coefficient,
O (x) = aG (x) −b | A−Re (Z (x)) | −c | Im (Z (x)) | −d | B−HBW | −eAL
+ KSL + hBL
Prepare the target function O (x) represented by
18. The recording medium storing the antenna optimum design program according to claim 11, wherein each of the calculated antenna characteristics is evaluated based on the objective function O (x). The second generation processing includes:
A process of selecting a first number of chromosomes in descending order of the assigned rank among the plurality of chromosomes to which the rank has been assigned, and storing the selected first number of chromosomes without change; ,
Generating a genetic operation by the genetic algorithm based on the assigned rank to the plurality of chromosomes to which the rank has been assigned, thereby generating the remaining second number of chromosomes not selected. Processing,
20. A process for configuring a plurality of chromosomes of a new generation by using the stored first number of chromosomes and the generated second number of chromosomes. A recording medium storing the antenna optimum design program according to any one of the above. An antenna device designed by the antenna optimum design method according to any one of claims 1 to 10.

Images