JP3920814B2 - Control method for wireless network and wireless communication system - Google Patents

Description

【００６１】
ここで、相関係数η（Ｐ）は、ルート間結合の度合いを測定するために用いられる（例えば、非特許文献１１及び１６参照）。すなわち、相関係数θ（Ｐ）は当該パスＰ自体に含まれない他の通信可能なノード無線局との間の電波干渉性に関する結合の度合いを表す。
【００６２】
なお、定義６及び定義７によれば、アドホック無線ネットワークにおいて、検索された複数のパスのうちの１つのバスＰ上のノード無線局を通信可能なノード無線局として仮定して設定し、ＧＬＳテーブル内の当該１つのパスＰに含まれるノード無線局に対するＡＮＬテーブルのデータを通信可能な状態に設定し、ＧＬＳテーブルに基づいて、発信元ノード無線局と宛先ノード無線局以外の発信元ノード無線局と宛先ノード無線局間の通信可能なパスＰ’に対して、当該１つのパスＰ上の各ノード無線局のサービスエリアの送信ゾーン内に存在する他の通信可能なノード無線局の個数を計算し、計算された通信可能なノード無線局の個数を１つのパスＰ毎に加算することによって、発信元ノード無線局と宛先ノード無線局以外の発信元ノード無線局と宛先ノード無線局間の通信可能なパスＰ’に対する当該１つのパスＰの相関係数η（Ｐ）を計算している。
【００６３】
パスＰの相関係数η（Ｐ）＝０のとき、パスＰは他のすべての「アクティブなパス」とは「ゾーン間非結合」であるという。ここで、アクティブなパスは、その時刻において通信プロセスに参加（関与）しているパスのことである。これ以外であれば、パスＰは、他のアクティブなパスとは相関係数ηで関係している。
【００６４】
一例として、再び、従来技術のマルチパスルーティングを示す図２６を参照すると、まず、発信元無線局Ｓ_１はノード無線局Ｎ_１及びＮ_２を介して宛先無線局Ｄ_１と通信している。従って、ＡＮＬテーブル，ＡＮＬ（ｔ）はパス｛Ｓ_１，Ｎ_１，Ｎ_２，Ｄ_１｝を含む。ここで、発信元無線局Ｓ_２が宛先無線局Ｄ_２との通信を希望し、パスＰ＝｛Ｓ_２，Ｎ_５，Ｎ_６，Ｄ_２｝を選択する。
【００６５】
まず、図２６においてオムニアンテナを使用する場合について以下に検討する。発信元無線局Ｓ_１及びノード無線局Ｎ_１はともに、発信元無線局Ｓ_２の送信ゾーン（この場合は３６０度）内にある。従って、
【数４】
η^Ｓ２（Ｐ）＝２
である。発信元無線局Ｓ_１及びノード無線局Ｎ_１はノード無線局Ｎ_５の全方向性の送信ゾーン内にあるため、
【数５】
η^Ｎ５（Ｐ）＝２
である。同様に、
【数６】
η^Ｎ６（Ｐ）＝２であり、よって、オムニアンテナを使用する場合において、パスＰの相関係数η（Ｐ）は、
【数７】
η（Ｐ）＝６
となる。
【００６６】
一方、指向性アンテナを使用する場合は、図２６に示すように、発信元無線局Ｓ_２、ノード無線局Ｎ_５及びノード無線局Ｎ_６によって形成される送信ゾーンは、ＡＮＬテーブル，ＡＮＬ（ｔ）からのノード無線局を全く含まない。従って、指向性アンテナを使用すれば、パスＰの相関係数η（Ｐ）は、
【数８】
η（Ｐ）＝０
となる。
【００６７】
パスＰの相関係数η（Ｐ）が大きくなれば、いずれのパスでも平均エンド・ツー・エンドの遅延時間は大きくなることは証明されている（例えば、非特許文献１１参照）。これは、より大きな相関係数を有する２つのパスは、無線伝搬のブロードキャストの特徴に起因して互いの送信に干渉するより多くの機会を有するためである。本実施形態からは、アドホック無線ネットワークにおける効率的なルーティングは、複数のノード無線局間の相関係数に大きく左右されると結論づけることができる。
【００６８】
しかしながら、オムニアンテナを使用して低い相関係数のルートを得ることは困難である。図２６から明らかであるように、指向性アンテナを使用すれば複数のルートを非結合することが可能であり、これにより、オムニアンテナを使用する場合に比べて格段に低い相関係数のルートを獲得することができる。
【００６９】
次いで、本実施形態に係る無線通信システムにおける各ノード無線局１における「ネットワーク認識」について以下に詳細に説明する。
【００７０】
本実施形態では、各ノード無線局が隣接ノード無線局のみを認識しているだけではなく、当該アドホック無線ネットワーク全体のネットワーク認識でもあるような機構を有する無線通信システムを提案する。この「ネットワーク認識」は、後述する前向きな（積極的に率先して行う）ルーティング方法を実施する際に役立つ。当該アドホック無線ネットワークにおける各ノード無線局ｎは、次のようなネットワーク状態情報を示す５つのテーブルを有し、これらのテーブルは図２の各無線局内のデータベースメモリ１５４内に格納される。以下に、これら５つのテーブルの詳細について説明する。
【００７１】
（１）ＡＳテーブル：ＡＳテーブルは、その一例を図６に示すように、自局のサービスエリア内の隣接ノード無線局毎に、方位角と、信号強度レベルの情報を格納し、後述する図１３及び図１４のパケット送受信制御処理により作成更新される。
【００７２】
（２）ＮＬＳテーブル（ノード無線局ｎにおけるＮＬＳテーブルを、ＮＬＳＴｎと記載する。）：各ノード無線局ｎは、その隣接ノード無線局の方向を追跡するために、周期的にその隣接情報を「収集」してＮＬＳテーブルを形成する。時刻ｔにおけるノード無線局ｎのＮＬＳテーブル（ＮＬＳＴｎ（ｔ））は、隣接ノード無線局ｍ（ｍ∈Ｇ^ｎ）の各々について、ある特定方向でノード無線局ｎにより検出された無線結合の最大信号強度ＳＩＧＮＡＬ^θ _ｎ，ｍ（ｔ）を特定することができる。従って、最大信号強度ＳＩＧＮＡＬ^θ _ｎ，ｍ（ｔ）は、ノード無線局ｎにおいてその隣接ノード無線局ｍからノード無線局ｎに対する方位角θで受信され、かつノード無線局ｎにより任意の時刻で検出されたときの最大信号強度である。ノード無線局ｎのＮＬＳテーブルは、その任意の隣接ノード無線局との最も可能性の高い通信方向を決定する際の手助けとなる。ここで、各ノード無線局は、予め取得された各隣接ノード無線局についてのＳＩＮＲ値が最大となる方位角を選び、このＳＩＮＲ値を隣接ノード無線局との間の親和度とする。各ノード無線局はこの方位角と親和度の値を各隣接ノード無線局毎に取り出し、現在日時を更新日時として、ＮＬＳテーブルを生成して更新する。当該ＮＬＳテーブルには、その一例を図７に示すように、各隣接ノード無線局毎に、最大のＳＩＮＲ値に対応する方位角、最大のＳＩＮＲ値である親和度、更新日時が格納されている。
【００７３】
（３）ＮＡＮＬテーブル（ノード無線局ｎにおけるＮＡＮＬテーブルを、ＮＡＮＬテーブルｎと記載する。）：ノード無線局ｎにおけるＮＡＮＬテーブルは、その隣接ノードの通信アクティブの状態（通信状態のフラグ値で示す。フラグ値＝１は通信アクティブであって通信可能である。フラグ値＝０は通信非アクティブであって通信不可能である。）を含む。言い替えれば、ノード無線局ｎの任意の隣接ノードが通信プロセスにアクティブに参加（関与）し又はしていなくても、ノード無線局ｎは、時刻ｔにおけるＮＡＮＬテーブル，［ＮＡＮＬテーブル_ｎ（ｔ）］と呼ばれるリストテーブルにその情報を記録する。これは、各ノード無線局が隣接ノード無線局での通信を検出（認識）するために役立つ。ＮＡＮＬテーブルでは、図８のＡＮＬテーブルのデータのうち、隣接ノード無線局に関する当該情報のデータのみを格納する。
【００７４】
（４）ＡＮＬテーブル（ノード無線局ｎにおけるＡＮＬテーブルを、ＡＮＬｎと記載する。）：ＡＮＬテーブルは、当該アドホック無線ネットワークにおける通信アクティビティに関するノード無線局ｎの検出結果を含む。これは、時刻ｔにおいてノード無線局ｎが検出できたアドホック無線ネットワーク内のすべてのアクティブなノード無線局を含むノード無線局ｎに関するリストである。ＡＮＬテーブルは、その一例を図８に示すように、ノード無線局毎に、上述の通信状態のフラグ値とそのデータの更新回数（バージョンを示す）を格納する。ＡＮＬテーブルについては、図８を参照して詳細後述する。
【００７５】
（５）ＧＬＳテーブル（ノード無線局ｎにおけるＧＬＳテーブルを、ＧＬＳｎと記載する。）：ＧＬＳテーブルは、時刻ｔにおいてノード無線局ｎが検出できたアドホック無線ネットワーク内のトポロジー情報を含む。具体的には、ＧＬＳテーブルは、その一例を図９に示すように、各ノード無線局毎に、隣接ノード無線局とその方位角データと、そのデータの更新回数（バージョンを示す）とを格納する。ＧＬＳテーブルについては、図９を参照して詳細後述する。
【００７６】
本実施形態においては、各ノード無線局は、上述のＡＮＬテーブルを含むパケット信号及びＧＬＳテーブルを含むパケット信号（以下、前者をＡＮＬパケット信号といい、後者をＧＬＳパケット信号という。）をそれぞれ異なる所定の周期間隔でブロードキャストして、ＡＮＬテーブル及びＧＬＳテーブルの生成及び更新処理を行うことを特徴としている。以下、これについて説明する。
【００７７】
各ノード無線局ｎは、そのＡＮＬテーブルをある周期的間隔（ＡＮＬパケット信号の送信周期）Ｔ_Ａでブロードキャストを行う。ここで、ＡＮＬパケット信号のブロードキャストは、次の２つの目的を果たす。すなわち、ノード無線局ｎがそのすべての隣接ノード無線局（例えば、ｉ、ｊ及びｋ）からのＡＮＬテーブルを受信すると、以下の処理が実行される。
（１）ノード無線局ｎは、隣接ノード無線局ｉ、ｊ及びｋをその隣接ノード無線局として含むように、ＮＬＳテーブル（ＮＬＳＴｎ）を形成し、それらの隣接ノード無線局のうちの任意の隣接ノード無線局との通信方向として最良である可能性のあるものとして記録する。
（２）ノード無線局ｎはまた、ノード無線局ｉの通信状態フラグ値（通信のアクティビティ状態）を記録し、他の隣接ノードについても同様に記録して、当該ノード無線局ｎ自身のＮＡＮＬテーブルを形成し、次いで、それに対応するＡＮＬテーブルを更新する。
【００７８】
また、各ノード無線局ｎは、そのＧＬＳパケット信号をある周期的間隔（ＧＬＳパケット信号の送信周期）Ｔ_Ｇでブロードキャストを行う。ここで、あるノード無線局ｎが、その隣接ノード無線局からＧＬＳパケット信号を受信すると、それ自身のＧＬＳテーブルを更新するが、これについては詳細後述する。
【００７９】
本実施形態に係る無線通信システムのシミュレーションでは、各ノード無線局の移動がないとき（固定時）、ＡＮＬパケット信号の送信周期Ｔ_Ａ＝２秒、かつＧＬＳパケット信号の送信周期Ｔ_Ｇ＝１０秒に設定した。また、各ノード無線局の移動が有るとき（移動時）、ＡＮＬパケット信号の送信周期Ｔ_Ａ＝１秒、かつＧＬＳパケット信号の送信周期Ｔ_Ｇ＝５秒に設定した。２つのパケット信号（ＡＮＬパケット信号とＧＬＳパケット信号）を２種類の送信周期でブロードキャストを行うことには、次のような理由がある。すなわち、ＡＮＬテーブルは各ノード無線局の通信状態のフラグ値（通信アクティビティ状態）を捕捉しており、一旦通信が開始されるとすぐに、あるノード無線局の集合におけるＡＮＬテーブルのデータが影響を受ける（そのデータが最新データで更新される）。従って、ＡＮＬパケット信号は、ＧＬＳパケット信号より速く伝搬される必要がある。さらに、ＡＮＬパケット信号はあたかもビーコン信号として動作する。従って、ＡＮＬパケット信号のより速い伝搬に従って、アクティブな（通信可能な）ノード無線局の重要情報をより速く浸透させ得るだけでなく、正確な隣接ノード無線局の情報（方位角や信号強度レベル）を取得することが可能である。本実施形態において実施しようとしているのはフィッシュアイの概念であることから（例えば、非特許文献２０参照）、正確な隣接ノード無線局の情報がより速く要求される。
【００８０】
これに対して、ＧＬＳテーブルはすべてのノード無線局の結合性に関するグローバルな情報であり、ノード無線局の物理的な移動（信号の伝搬に比べて遙かに遅い）に関するトポロジーの変化を反映する。さらに、任意のノード無線局におけるＧＬＳテーブルは、さほど正確である必要がない。データ量が大きいＧＬＳパケット信号はその伝搬が遅く、ＧＬＳテーブルよりもデータ量が小さいＡＮＬパケット信号はその伝搬が速いのは、このためである。
【００８１】
図１０は図１のアドホック無線ネットワークにおいて用いられるトーン信号とパケット信号の送受信処理を示すタイミングチャートである。
【００８２】
例えば非特許文献２７において開示された従来技術に係るＭＡＣプロトコルの方法においては、ビーコン信号、ＡＮＬパケット信号、ＧＬＳパケット信号の３つのパケット信号の定期的な送信が必要であり、占有時間が比較的長く、スループットが落ちるという問題点があった。この問題点を解決するために、本実施形態に係るＭＡＣプロトコルでは、図１０に示すように、送信側の無線局は、無変調搬送波であるトーン信号に続いて、ＡＮＬテーブル、ＧＬＳテーブル、ＲＴＳ又はＣＴＳ、もしくは送信したいデータを含むパケット信号を送信する。これに対して、受信側の無線局は、トーン信号を検出した後、パケット信号を復号化して復号化したデータを取り込み、データ処理を実行する。なお、ＡＮＬテーブル又はＧＬＳテーブルを含むＡＮＬパケット信号又はＧＬＳパケット信号の送信は上述のように周期的に実行される。
【００８３】
図１１は図１のアドホック無線ネットワークにおいて用いられる各無線局での放射パターンの種類と無線通信プロトコルを示すタイミングチャートである。図１１では、本実施形態に係る４方向ハンドシェイクのアンテナモードの使用例を示している。適応制御パターンは移動中のノード無線局を追跡することはできるが、ビーム及びヌルはパケット信号が受信されなければ形成され得ない。従って、送信側のノード無線局からのトーン信号＋ＲＴＳの送信やトーン信号＋ＣＴＳの送信では、オムニパターンが使用される。これに対して、受信側のノード無線局におけるトーン信号＋ＲＴＳの受信やトーン信号＋ＣＴＳの受信では、その開始時に回転セクターパターンが使用された後、ＡＳテーブルに基づいた方位角に向けられたセクタパターンを使用し、当該セクタパターンを用いた通信中において適応制御のための準備処理を実行し、最後に適応制御用放射パターンに移行する。
【００８４】
次いで、ＮＬＳテーブル及びＮＡＮＬテーブルの形成方法について以下に説明する。
【００８５】
例えば任意のノード無線局ｎは、例えば任意のその隣接ノード無線局ｍからＡＮＬパケット信号を受信すると、そのノード無線局ｎのＮＬＳテーブル（ＮＬＳＴ_ｎ）を逐次そのデータをインクリメントしながら形成する。ノード無線局ｎは、その可変ビームアンテナ１０１をある特定の方位角の角度に設定してＡＮＬパケットを受信することから、その特定の方向でノード無線局ｎによって認知される隣接ノード無線局ｍとの通信方向として最良である可能性のある方位角、及び無線回線でのノード間結合が最大である信号強度ＳＩＧＮＡＬ^θ _ｎ，ｍ（ｔ）を認識している。本実施形態では、互いに隣接する２つのノード無線局間で対称性のリンクを想定している。
【００８６】
例えば任意のノード無線局ｎは、それ自身の通信状態のフラグ値（アクティビティ状態）によってまず、それ自身のＮＡＮＬテーブル（ＮＡＮＬｎ）を形成する。ノード無線局ｎは、それが通信を希望していることを指示するＲＴＳを発行する必要があればいつでも、それ自身を「アクティブなノード無線局」（通信状態のフラグ値＝１）として設定し、所定のしきい値期間の間にＲＴＳを発行しない場合はそれ自体を「非アクティブな（通信不可能な）ノード無線局」（通信状態のフラグ値＝０）として設定する。また、ノード無線局ｎは、例えばその隣接ノード無線局ｍからＲＴＳを受信すると、必ずそれ自身のＮＬＳテーブル（ＮＬＳＴｎ）にノード無線局ｍをアクティブノード無線局（通信状態のフラグ値＝１）として設定する。ノード無線局ｍがそれ自身を非アクティブにするとき（通信不可能状態とするとき）、この情報がノード無線局ｍからＡＮＬパケットの周期的なブロードキャストによりノード無線局ｎに到達する。これを受信すると、ノード無線局ｎはそれ自身のＮＡＮＬテーブル（ＮＡＮＬｎ）においてノード無線局ｍを非アクティブ（通信状態のフラグ値＝０）として設定する。
【００８７】
次いで、ＡＮＬテーブルの形成方法について以下に説明する。
【００８８】
各ノード無線局は、当該アドホック無線ネットワークにおける通信アクティビティに関するその認知を含むそのアクティブノードリストを含むＡＮＬテーブルを含むＡＮＬパケット信号を、周期的にブロードキャストする。複数の異なるノード無線局から周期的なＡＮＬパケットを受信すると、各ノード無線局はこれらの情報を取り入れて改訂更新されたＡＮＬテーブルを形成し、周期的間隔を待って更新されたＡＮＬパケット信号をその隣接ノード無線局に対してブロードキャストする。
【００８９】
各ノード無線局では、ＡＮＬパケットはまず、そのノード無線局のＮＡＮＬテーブルによって更新される。従って、初期状態では、アドホック無線ネットワークの開始時には、すべてのノード無線局は単にそれ自身の隣接ノード無線局のアクティビティ状態（通信可能状態）を認識しているだけであって、当該無線通信システム内の他のノード無線局に関しては「何も知らない状態」にある。各ノード無線局は、周期的にそれ自身のＡＮＬテーブルを更新して、更新したＡＮＬパケット信号をその（１つ又は複数の）隣接ノード無線局に対してブロードキャストを行う。その隣接ノード無線局からのそれらの隣接ノード無線局に関するこの周期的な更新メッセージ（ＡＮＬパケット信号）の送信により、各ノード無線局は次第に、他のノード無線局及びそれらの隣接ノード無線局に関するアクティビティ情報（ＡＮＬテーブルに含まれる情報）を取得していく。このように、各ノード無線局は、他のノード無線局から受信した更新メッセージ（ＡＮＬパケット信号）を基礎としてそれ自身のＡＮＬテーブルを更新することができる。
【００９０】
移動中のノード無線局へのネットワーク状態情報の浸透処理における大きな概念は、伝送される情報がある程度正確に認識されなければならないことにある。複数の異なるノード無線局からの更新の伝搬は非同期的であるため、情報の更新回数（バージョン）の概念を導入することが肝要になる（例えば、非特許文献２１及び２４参照）。例えば、２つのＡＮＬパケットＡ_１及びＡ_２がともにノード無線局ｎから複数ホップで離れているノード無線局ｍに関する情報を伝送しながらノード無線局ｎに到達するものとする。ノード無線局ｎにおいてノード無線局ｍに関する情報を更新するためには、ノード無線局ｍに関する最も新しい情報を伝送しているのはどれか、すなわち、「ＡＮＬパケットＡ_１であるか、もしくはＡＮＬパケットＡ_２であるか？」を発見するメカニズムが存在していなければならない。
【００９１】
これを実行するために、本発明者らは最新性トークンの同一概念（最も新しいかの指標であるバージョンを示す更新回数情報を有すること；例えば、非特許文献２４参照）及びこれを適正にインクリメントする機構を使用している。仮に２つの更新メッセージ（ＡＮＬパケット）が、例えばノード無線局ｎである同じノード無線局に関するデータ集合を保有していれば、ノード無線局ｎのより高い最新性の値（更新回数）を伝送している更新メッセージ（ＡＮＬパケット）の方がそれに関するより新しい情報を保有している。
【００９２】
ノード無線局ｎにおけるＡＮＬテーブルの一例を示す図８において、Ｒ_ｉはＮ個のノード無線局よりなるアドホック無線ネットワークにおけるノード無線局ｎ_ｉの最新性を更新回数であり、Ｓ_ｉは各ノード無線局の対応するアクティビティ状態を示す通信状態のフラグ値を示し、０（非アクティブ：通信不可能）又は１（アクティブ：通信可能）のいずれかである。
【００９３】
次いで、ＧＬＳテーブルの形成方法について以下に説明する。
【００９４】
各ノード無線局は、ネットワーク連結情報を捕捉するためのＧＬＳテーブルを保持している。各ノード無線局では、ＧＬＳテーブルはまず、そのノード無線局のＮＬＳテーブルによって更新される。従って、初期状態では、アドホック無線ネットワークの開始時には、すべてのノード無線局は単にそれ自身の隣接ノード無線局を認識しているだけであって、当該無線通信システム内の他のノード無線局に関しては「何も知らない状態」にある。各ノード無線局は、周期的にそのＧＬＳテーブルを更新して、更新したＧＬＳパケット信号を所定の送信周期でその隣接ノード無線局に対してブロードキャストする。その隣接ノード無線局からのそれらの各隣接ノード無線局に関するこの周期的な更新メッセージ（ＧＬＳパケット信号）により、ノード無線局は次第に、他のノード無線局及びそれらの各隣接ノード無線局に関するＧＬＳテーブル情報を取得していく。このように、各ノード無線局は、他のノード無線局から受信した更新メッセージ（ＧＬＳパケット信号）を基礎として、それ自身のＧＬＳテーブルを更新する。
【００９５】
ここで、更新の周期性を制御すれば、当該アドホック無線ネットワーク内の更新トラフィック及び各ノード無線局に記憶されたネットワーク状態情報の精度を制御できることは留意されるべき点である。例えば、更新メッセージ（ＧＬＳパケット信号）の伝搬が頻繁すぎると、制御トラフィックは増大するが、各ノード無線局に記憶されたネットワーク状態情報の精度も上がる。ここで、当該アドホック無線ネットワークが更新の伝搬でフラッドされることは決してない。任意の時点におけるアドホック無線ネットワーク内の最大パケット数は、常時、アドホック無線ネットワーク内のノード無線局数より少ない。この場合もやはり、ＡＮＬパケットの伝搬について上述したような最新性の概念（更新回数の情報保持）を実施する必要がある。これは、仮に２つのＧＬＳパケット信号の更新メッセージが、例えばノード無線局ｎである同じノード無線局に関するデータ集合を保有していれば、ノード無線局ｎのより高い最新性トークン値（より大きい更新回数）を伝送している更新メッセージ（ＧＬＳパケット信号）の方がそれに関するより新しい情報を保有することを意味している。
【００９６】
ノード無線局ｎは、他のノード無線局からＧＬＳパケット信号を受信すると、それ自身のＧＬＳテーブルを更新する。これを行うために、ノード無線局ｎのＧＬＳテーブルに記憶されたすべてのノードの最新性トークン値（更新回数）と、最近に到着した更新のためのＧＬＳパケット信号に記憶されたすべてのノードの最新性トークン値（更新回数）とが比較される。ノード無線局ｎのＧＬＳテーブルにおける任意のノード無線局、例えばノード無線局Ｘの最新性トークン（更新回数）の方が更新するＧＬＳパケット信号におけるそれより少ないことが分かれば、更新するＧＬＳパケット信号の方がノード無線局Ｘに関する最新の情報を伝送していることは明らかである。従って、ノード無線局ｎのＧＬＳテーブルにおけるノード無線局Ｘに関する全体情報は、受信された更新ＧＬＳパケット信号におけるノード無線局Ｘの情報によって上書きされる。このステップは、すべての更新ＧＬＳパケット信号について、それらのそのホストノード無線局ｎへの到着に伴って非同期的に実行される。このステップにより、更新ＧＬＳパケット信号から収集することのできるすべての最新情報のノード無線局ｎによる捕捉が促進される。
【００９７】
ここで、この機構は、各ノード無線局がアドホック無線ネットワークの正確な状態を認識することを保証するものでないことを留意する必要がある。各ノード無線局がアドホック無線ネットワークのおおよその状態を解明する際に手助けとなるものは、単なる「認識」でしかない。これは、ノード無線局に近接した周辺のノード無線局においてはより正確な状態情報の保持を促進するが、距離が大きくなると（離れてくると）次第にネットワーク情報の詳細の精度が下がる「フィッシュアイアプローチ」（例えば、非特許文献２０参照）に類似している。
【００９８】
任意のノード無線局ｎにおけるＧＬＳテーブルの構造を示す図９において、Ｒ_ｉはＮ個のノードよりなるアドホック無線ネットワークにおけるノード無線局ｎ_ｉの最新性トークン値（更新回数）であり、＜ｎ_ｊ，α（ｎ_ｉ，ｎ_ｊ）＞はノード無線局ｎ_ｊがノード無線局ｎ_ｉの隣接ノード無線局であることを示している。ここで、α（ｎ_ｉ，ｎ_ｊ）はノード無線局ｎ_ｉがノード無線局ｎ_ｊと最も良好に通信できる送信ビーム方位角αを示している。
【００９９】
次いで、各ノード無線局の位置の追跡とＭＡＣプロトコルについて以下に説明する。
【０１００】
通常、アドホック無線ネットワークでは、すべてのノード無線局にオムニアンテナが装備されている。しかしながら、オムニアンテナを使用するアドホック無線ネットワークは、無線媒体を広域に渡って保持することにネットワーク容量の大部分を浪費するＲＴＳ／ＣＴＳベースのフロア予約方法を使用している。従って、送信機及び受信機の隣接にある多くのノード無線局は、送信機と受信機との間のデータ通信が終わるのを待って、無為に存在することを余儀なくされる。この問題を緩和するため、当該技術分野の研究者達は、無線干渉を大幅に低減し、これにより無線媒体の利用及び必然的にネットワークのスループットを増進させる指向性アンテナの使用を提案している（例えば、非特許文献２乃至５参照）。
【０１０１】
指向性アンテナの能力を十分に活用するためには、発信元無線局及び宛先無線局の隣接ノード無線局はすべて、他の方向で新たな通信を開始することができて発信元無線局と宛先無線局との間で現に行われているデータ通信との干渉を防止できるように、通信の方向を認識していなければならない。こうして、各ノード無線局においてその隣接ノード無線局の方向を追跡するメカニズムを保有することが不可欠になる。
【０１０２】
しかしながら、指向性アンテナを使用する無線アドホック無線ネットワークにおけるこの方向追跡メカニズムは、多大な制御オーバーヘッドを被ることから重大な問題である。
【０１０３】
例えば、非特許文献１では、方向の追跡は、トーン信号の集合を使用しかつアドホック無線ネットワーク内の各ノード無線局で広範なネットワーク状態情報を保持することによって実行されている。ここで、各ノード無線局が移動する動的状態の場合には、これは非現実的である。また、非特許文献４では、提案されたＭＡＣプロトコルは位置情報を認識する必要がなく、発信元無線局及び宛先無線局は全方向性のＲＴＳ−ＣＴＳ交換の間にお互いの方向をオンデマンドベースで同定する。このＲＴＳ−ＣＴＳの対話通信を聞いている発信元無線局ｓ及び宛先無線局ｄのすべての隣接ノード無線局は、この情報を使用して実行中のデータ送信との干渉を防止するように想定されている。しかしながら、ＲＴＳ及びＣＴＳパケットの全方向性の送信に起因して、このプロトコルは無線チャネルの空間再利用という恩典をもたらさない。
【０１０４】
さらに、例えば非特許文献２では、各ノード無線局の位置の追跡にＧＰＳ（Global Positioning System）の使用が提案されているが、情報交換の正確なメカニズム及び結果的に生じるオーバーヘッドについては論じられていない。また、例えば非特許文献３では、ノード無線局はその隣接ノード無線局へ指向的にアクセスするために送信方向を認識していることが想定されているが、位置の追跡機構は示されていない。さらに、非特許文献２及び３の方法はともに、各ユーザ端末におけるハードウェアの追加を必要としている。本発明者の１人ほかによる先の研究結果である非特許文献５及び１０では、各ノード無線局が方位角対ＳＩＮＲテーブル（ＡＳテーブル）の保全を介して隣接情報を動的に維持するＭＡＣプロトコルが提案されている。この方法では、ＡＳテーブルを作成するために各ノード無線局は、指向性ビーコン信号を周期的に指向性のブロードキャストの形式で全方向へ３０度間隔で順次送信し、３６０度の空間全体をカバーする。ここで、この従来技術の方法では、制御パケットに起因するオーバーヘッドが極めて高い。
【０１０５】
本実施形態における本発明者らにより発明されたＭＡＣプロトコルは、基本的に「受信機による指向、回転セクタベースの指向性のＭＡＣプロトコル」であり、これはまた位置追跡メカニズムとしても機能する。この場合、各ノード無線局はアイドルの間、全方向性の検出モードで待機する。これは、しきい値を超える何らかの信号を検出するたびに、「回転セクタ受信モード」に入る。「回転セクタ受信モード」では、ノード無線局ｎはその指向性アンテナを全方向へ４５度間隔で順次回転させて、各方向における順次指向性受信の形式で３６０度の空間全体をカバーし、各方向で受信された信号を検出する。一回転の後、これは、受信された最大の信号強度によって信号の受信方向として最良である可能性のあるものを決定する。次に、そのビームをその方向へ設定し、信号を受信する。
【０１０６】
ここで、受信された信号を復号する受信機を有効化するために、各制御パケットは、受信機の回転受信ビームを３６０度回転させる時間がトーン信号の持続時間（本発明者らのシミュレーションでのケースでは２００マイクロ秒）より少し短いような持続時間を有する先行トーン信号を伴って送信される。任意の制御パケット信号より前にこのトーン信号を送信する目的は、受信機が、信号の受信方向として最良である可能性のあるものを追跡できるようにすることにある。これがそのビームをその方向へ設定すると、トーン信号の目的は果たされ、続いて制御パケット信号が送信される。
【０１０７】
本発明者らにより提案した本実施形態に係るフレームワークでは、媒体アクセス制御のために、次の４タイプのブロードキャスト（全方向性）制御パケットを使用している。
（１）ＡＮＬテーブルを含むパケット信号（ＡＮＬパケット信号）、
（２）ＧＬＳテーブルを含むパケット信号（ＧＬＳパケット信号）、
（３）ＲＴＳ（送信要求）信号を含むパケット信号、及び
（４）ＣＴＳ（送信可）信号を含むパケット信号。
【０１０８】
このほか、制御パケットＡＣＫ信号も、指向性の制御パケットである。データを含むパケット信号は、ＲＴＳ／ＣＴＳのハンドシェイクが行われた後に方向を決めて送信される。ＡＮＬパケット信号及びＧＬＳパケット信号は上述のように、周期的な信号であり、各ノード無線局から予め決められた送信周期（送信間隔）で送信される。各周期的間隔において、例えば各ノード無線局ｍは、ＡＮＬパケットをその隣接ノード無線局へ、もしその無線媒体（無線チャンネル）が空き状態であればブロードキャストを行う。先に指摘した通り、ＡＮＬパケットは、受信機が信号の受信方向として最良である可能性のあるものを検出する際に役立つ先行のトーン信号を伴って送信される。次いで、各受信機はそのビームをその方向へ設定し、パケットを受信して復号する（図１０参照）。
【０１０９】
さらに、ノード無線局ｎは、例えばノード無線局ｊとのデータ通信の開始を希望する度に無線媒体（無線チャンネル）をチェックし、無線媒体が空き状態であれば全方向性のＲＴＳ信号を発行して送信する。宛先無線局ｊは、ＲＴＳ信号を受信すると全方向性のＣＴＳ信号を発行して送信する。この場合のＲＴＳ／ＣＴＳの目的は、（オムニアンテナを使用する場合のように）ノード無線局ｎ及びｊの隣接ノード無線局による送信又は受信を禁止することではなく、ノード無線局ｎ及びｊの隣接ノード無線局に、隣接ノード無線局ｊがノード無線局ｎからデータパケットを受信しようとしていることを知らせることにある。これはまた、通信のおおよその持続時間も特定する。ノード無線局ｎ及びｊのすべての隣接ノードは、それらの指向性ネットワーク配置ベクトル（Directional Network Allocation Vector（ＤＮＡＶ））をノード無線局ｎ及びｊの方向へ設定することにより、ノード無線局ｎとｊとの間の通信を追跡する。従って、ノード無線局ｎ及びｊの隣接に存在するノード無線局は、「ノード無線局ｎとノード無線局ｊとの間で行われている通信を妨害することなく」他の方向への通信を開始することができる。発信元無線局と宛先無線局は、指向性の受信モードでそれぞれ肯定応答信号及びデータパケット信号を待つ。
【０１１０】
図１２は、上述のＤＮＡＶを用いて放射パターンの制御を行う方法を示すものであって、既に他の無線通信が行われていることを知っている場合の各ノード無線局の配置及びアンテナ放射パターンを図示している。図１２において、ノード無線局Ｓとノード無線局Ｄが通信中にそれらの周辺にあるノード無線局Ｘとノード無線局Ｙがノード無線局Ｘを発信元として通信を行おうとしている場合を図１２に示す。また、ノード無線局Ｓからノード無線局Ｄへのセクタビームパターンは図１２のようになっている。ここで、ノード無線局Ｘとノード無線局Ｙはノード無線局Ｓやノード無線局ＤからのＲＴＳ信号及びＣＴＳ信号を既に受信しており、それを登録している。
【０１１１】
図１２において、θ_ｘｙをノード無線局Ｘから見たノード無線局Ｙへの方位角の値とすると、ノード無線局Ｘとノード無線局ＹはそれぞれのＡＳテーブル及びＮＬＳテーブルより、ノード無線局Ｘは方位角θ_ｘｓとθ_ｘｄの値、ノード無線局Ｙは方位角θ_ｙｓとθ_ｙｄの値を知ることができる。まず、ノード無線局Ｘからノード無線局ＹへのＲＴＳ信号の送信がノード無線局Ｓやノード無線局Ｄに影響を与えてしまうような場合、つまり方位角θ_ｘｙがθ_ｘｓやθ_ｘｄと重なってしまう場合には、ノード無線局Ｘは送信をすることが出来ず、アイドル状態でノード無線局Ｓとノード無線局Ｄの通信が終了するのを待つ必要がある。そうでない場合、ノード無線局ＸはＲＴＳ信号を送信することができる。この際のノード無線局Ｘのアンテナ放射パターンは前提条件から、方位角θ_ｘｓとθ_ｘｄの方向にヌル点を形成した排他的セクタパターンである。同様に、ノード無線局Ｙからノード無線局Ｘへの排他的セクタパターンがノード無線局Ｓやノード無線局Ｄを捉えていない場合、すなわち、方位角θ_ｙｘとθ_ｙｓやθ_ｙｄが重なっていない場合、ノード無線局Ｙはノード無線局ＸへＣＴＳ信号を送信することができる。以降、ノード無線局Ｘとノード無線局ＹはセクタビームパターンによりＤＡＴＡ信号及びＡＣＫ信号の送受信を行う。
【０１１２】
次いで、最大化されたゾーン間非結合最短パスによる適応型ルーティングプロトコルについて以下に説明する。
【０１１３】
アドホック無線ネットワークにおける従来技術のルーティングプロトコルは、「オムニアンテナ」の使用に依存している。ルーティングに関する指向性アンテナの優位点の活用は、上述したように未だ十分には研究されていない（例えば、非特許文献１０及び１９など参照）。本発明者らは、指向性アンテナの能力を、改善された媒体利用に向けて活用するためには、指向性のＭＡＣプロトコルとともに、「効果的な負荷のバランス化」を使用するルーティング方法を提供する必要があると考える。本発明者らは、負荷のバランス化を使用するルーティングのためのテーブル駆動のルーティングプロトコルを提案する。本発明者らは、相関係数ηを使用して、「最大化されたゾーン間非結合」を負荷のバランス化のためのルート選択基準として測定した。ここで、各ノード無線局におけるネットワーク認識は、「実際の」ネットワーク状態ではなくネットワーク状態に関する「認知」に過ぎないため、中間のノード無線局はそれぞれ、ルーティングの間にルーティング決定を適応的に補正して変更する。本発明者らは、最大化されたゾーン間非結合ルートを使用する効果的な負荷のバランス化に関して、下記のようなルーティング方法を実施している。
【０１１４】
アドホック無線ネットワーク内の各ノード無線局は、ネットワーク状態に関する現在の情報（近似トポロジー情報及び進行中の通信情報）を使用して、指定された宛先無線局へ到達するための「適切な次のホップ」を計算する。これは、新たなパスが何らかの通信に既に関与しているノード無線局への干渉を最小限に抑えるべく試行するような方法で、下記の方法を使用して発信元無線局及び宛先無線局ペア間の適切なパスを選択する。
【０１１５】
ステップＩ：所定のホップ最大値Ｈ_ｍａｘ（本実施例では、Ｈ_ｍａｘ＝６）より少ないホップ数Ｈを有する発信元無線局と宛先無線局ペア間のすべてのパスを発見する。
ステップII：ＡＮＬテーブルを調べて、その時点で進行中の通信に関与しているノード無線局を発見し、ルート相関係数ηを計算する。
ステップIII：もしＡＮＬテーブルが空であれば（すなわち、アドホック無線ネットワーク内で進行中の通信がなければ）、宛先無線局へ最小のホップで利用可能なパスのうちの任意の１つが選択される。
ステップIV：もしＡＮＬテーブルが空でなくＡＮＬテーブルに記録されているように、アドホック無線ネットワーク内に既に幾つかの通信が存在している場合、以下の処理を実行する。
（Ａ）発信元無線局と宛先無線局が多数のホップ数（２ホップを超える）で離隔されていれば、発信元無線局は、発信元無線局と宛先無線局ペア間のパスとして可能性のあるものすべての中からηが最小のパスを検索する。
（Ｂ）発信元無線局と宛先無線局が２ホップでしか離隔されていなければ、発信元無線局は、発信元無線局と宛先無線局間のペアの２ホップパスとして可能性のあるものすべての中からηが最小のパスを検索する。
【０１１６】
以上の処理の詳細を図１３及び図１４のフローチャートを参照して以下に説明する。図１３は図２の無線局１の管理制御部１０５によって実行されるパケット送受信制御処理を示すフローチャートであり、図１４は図１３のサブルーチンであるルーティング及び通信処理（ステップＳ１２）の処理を示すフローチャートである。
【０１１７】
図１３において、まず、ステップＳ１で回転セクターパターンで可変ビームアンテナ１０１を所定の方位角（例えば、３０度）毎に変化して回転走査するように制御して受信信号を受信し、ステップＳ２において所定のしきい値以上の信号強度レベルの受信信号を受信したか否かが判断され、ＹＥＳのときはステップＳ３に進む一方、ＮＯのときはステップＳ９に進む。ステップＳ９において送信すべきパケット信号があるか否かが判断され、ＹＥＳのときはステップＳ１０に進む一方、ＮＯのときはステップＳ１に戻る。ステップＳ１０において送信すべきパケット信号はＡＮＬパケット信号、もしくはＧＬＳパケットであるか否かが判断され、ＹＥＳのときはステップＳ１１に進む一方、ＮＯのときはステップＳ１２に進む。ここで、前者のＡＮＬパケットは、所定の周期期間Ｔ_Ａで周期的にそのイベントが発生され、後者のＧＬＳパケットは、所定の周期期間Ｔ_Ｇで周期的にそのイベントが発生され、このステップでＹＥＳとなる。ステップＳ１１においてオムニパターンで当該テーブルを含むパケット信号を送信した後、ステップＳ１に戻る。一方、ステップＳ１２において図１４のサブルーチンであるルーティング及び通信処理を実行した後、ステップＳ１に戻る。
【０１１８】
ステップＳ３では、回転セクターパターンを停止して、可変ビームアンテナ１０１の放射パターンを、停止した所定の方位角に向けるセクターパターンに設定する。次いで、ステップＳ４において適応制御パターンで受信信号を受信し、パケット情報を復号化し受信信号の信号強度レベルを測定し、ステップＳ５において受信信号ＲＳを判別して以下のようにステップＳ６，Ｓ７又はＳ８に分岐する。
【０１１９】
ここで、受信信号ＲＳ＝ＧＬＳパケット信号であるときは、ステップＳ６でＧＬＳテーブル等更新処理を実行した後、ステップＳ１に戻る。このＧＬＳテーブル等更新処理では、現在格納されているＧＬＳテーブルと、受信信号に含まれるＧＬＳテーブルとを比較して、更新回数がより多いバージョンが新しいデータのみをそれを用いてＧＬＳテーブルを更新し、新しいノード無線局であるときはそのノード無線局についての情報を追加する。また、ステップＳ５では、受信信号ＲＳ＝ＡＮＬパケット信号であるときは、ステップＳ７に進み、ＡＳテーブル等更新処理を実行した後、ステップＳ１に戻る。このＡＳテーブル等更新処理では、パケット信号内のノード無線局のＩＤと、当該受信信号を受信するときに検出された方位角と信号強度レベルに基づいてＡＳテーブルを更新するとともに、上記のＧＬＳテーブルの更新方法と同様に、現在格納されているＡＮＬテーブルと、受信信号に含まれるＡＮＬテーブルとを比較して、更新回数がより多いバージョンが新しいデータのみをそれを用いてＡＮＬテーブルを更新し、新しいノード無線局であるときはそのノード無線局についての情報を追加し、さらに、更新されたＡＮＬテーブルに基づいて、ＮＡＮＬテーブルを更新する。さらに、ステップＳ５で受信信号ＲＳ＝その他の信号であるときは、ステップＳ８でその他の信号の受信処理を実行した後、ステップＳ１に戻る。
【０１２０】
図１３の制御フローにおいては、ステップＳ１，Ｓ２において、可変ビームアンテナ１０１を回転走査して所定のしきい値以上の信号強度レベルの受信信号を受信したときに、その受信信号を検出しているが、本発明はこれに限らず、可変ビームアンテナ１０１を３６０度にわたって回転走査して、所定のしきい値以上の信号強度レベルの受信信号を受信しかつそのうちの最大の受信信号を、検出された受信信号としてもよい。
【０１２１】
次いで、サブルーチンであるルーティング及び通信処理を示す図１４のステップＳ２１では、まず、ＧＬＳテーブルを参照して、発信元無線局と宛先無線局の間において、ホップ数Ｈが所定の最大値Ｈｍａｘよりも小さい、すべてのパスｐ１，…，ｐｎを見つける。そして、ステップＳ２２において１つのパスｐｉをアクティブであるとみなし、ＧＬＳテーブル内のパスｐｉに含まれるノード無線局に対するＡＮＬテーブルの通信状態フラグ値を１にし、ステップＳ２３においてＧＬＳテーブルを参照し、無線局Ｓ及びＤ以外の発信元無線局及び宛先無線局間のアクティブなパスに対するパスｐｉの相関係数η（ｐｉ）とを計算し、パスｐｉのホップ数Ｈを乗算することにより、パスｐｉのルーティング基準指数γ（ｐｉ）を計算する。本実施形態では、あるパスｐｉの相関係数η（ｐｉ）とホップ数Ｈとの積を最小化することをルート選択の基準にしており、この積であるルーティング基準指数γ（ｐｉ）を最小化すると、ゾーン間分離が最大化された最短のパスを得ることができる。
【０１２２】
次いで、ステップＳ２４においてＧＬＳテーブル内のパスｐｉに含まれるノード無線局に対するＡＮＬテーブルの通信状態フラグ値を０にセットし、ステップＳ２５においてパスｐ１，…，ｐｎのうちのすべての組み合わせのパスの対に関してルーティング基準指数γ（ｐｉ）を計算したか否かについて判断し、ＮＯであれば、他のパスを選択して上述の処理を繰り返すためにステップＳ２２に戻る。一方、ステップＳ２５でＹＥＳであるときは、ステップＳ２６において最小のルーティング基準指数γ（ｐａ）に対応するパスｐａを選択し、ＧＬＳテーブル内のパスｐａに含まれるノード無線局に対する通信状態フラグ値を１にセットし、ステップＳ２７においてパスｐａに含まれるノード無線局に対するアクティブノードリスト情報を含むＡＮＬテーブルを、それを含むＡＮＬパケット信号により他のノード無線局に送信し、各ノード無線局のＡＮＬテーブルを更新させる。さらに、ステップＳ２８において発信元無線局はＧＬＳテーブル内のルーティング情報に基づいてパスｐａを介して宛先無線局と通信し、ステップＳ２９において通信終了後、ＧＬＳテーブル内のパスｐａに含まれるノード無線局に対するＡＮＬテーブルの通信状態フラグ値を０にし、パスｐａに含まれるノード無線局に対するアクティブノードリスト情報を含むＡＮＬテーブルを、それを含むＡＮＬパケット信号により他のノード無線局に送信し、各ノード無線局のＡＮＬテーブルを更新させ、この後、元のメインルーチンに戻る。
【０１２３】
以上の図１４の処理では、宛先無線局から遠く離れている場合には、中間の各ノード無線局が進行中の通信との干渉を低減するように、相関係数ηが最小のパスを選択することを保証する。しかしながら、中間のノード無線局がわずか２ホップでしか宛先無線局と離隔されていないことを発見した場合には、これは、より大きなホップ計数を有する最小の相関係数ηのパスよりも、低い相関係数ηを有する最小ホップのパスの方へ高い優先性を与える。
【０１２４】
このメカニズムは、各ノード無線局がアドホック無線ネットワークの正確な状態を認識することを保証するものではないため、中間のノード無線局は、そのルーティング決定を補正し、代替パスを採ってデータパケット信号を宛先無線局へとルーティングする。ここで、宛先無線局のより近くに存在するノード無線局は、宛先無線局及びその隣接の通信状態に関するより正確な情報を得ることになる。
図１５はこの点を表しており、すなわち、宛先無線局に到達するための中間ノード無線局による適応ルーティング選択処理を示している。図１５において、発信元無線局Ｓは、当初、宛先無線局Ｄへ至るおおよそのルート｛Ｓ−Ｘ−Ｙ−Ｚ−Ｄ｝を決定している。図１５中の点線は、この当初の宛先無線局Ｄの位置を示している。しかしながら、宛先無線局Ｄは移動によってその位置を変え、実線は宛先無線局Ｄの現在の位置を示している。この宛先無線局Ｄの位置変更の情報が中間のノード無線局Ｙに届くと、宛先無線局Ｄに関してより正確な情報を有する中間のノード無線局Ｙは、すぐに宛先無線局Ｄへのパスの変更を決定し、かつ宛先無線局ＤへのＰ及びＱを介するより優れたパスを決定することができる。こうしてパスは、多数の制御パケットを発生させることなく、利用可能な情報の正確さに依存して適応的に選択され、修正される。各ノード無線局はＧＬＳテーブル及びＡＮＬテーブルを保有するため、これもルーティング性能を向上させる。
【０１２５】
本発明者らのシミュレーションにおける１つの場合では、パス内の各ノード無線局ｎが宛先無線局へ到達するためのその「最良の次のホップ」を計算する。計算が終わると、ノード無線局ｎは、その情報がノード無線局ｎの場合と同じアンテナパターンで到達可能である限りにおいて、この次のホップをその特定の通信に使用する。言い替えれば、この次のホップが同じアンテナパターンでノード無線局ｎにアクセス可能でなければ、もしくはこの次のホップが到達不能なものであれば、ノード無線局ｎは、同じルート計算方法を使用して宛先無線局に到達する次のホップを計算し直す。
【０１２６】
【実施例】
次いで、本実施形態に係るアドホック無線ネットワークのための無線通信システムの性能を評価するためのシミュレーションとその結果について以下に説明する。
【０１２７】
このシミュレーションは、クアルネット（ＱｕａｌＮｅｔ）３．１を使用して実施した（例えば、非特許文献２２参照）。本発明者らの指向性アンテナは、方位角が４５度で離散的に配向することが可能であり、３６０度の範囲がカバーされる。セクタのビームパターンとしては、最大利得１５．６ｄＢｉのビームパターンを使用する。クアルネットのシミュレータでは、上述したＭＡＣプロトコルと、ルーティングプロトコルとを用いた。
【０１２８】
１０００×１０００ｍ^２の面積のエリアに、３０個のノード無線局をランダムに配置し、８個のノード無線局を、一定のビット伝送速度（ＣＢＲ；Constant Bit Rate）の発信元無線局としてタイムラグをおいて選択する。これらはそれぞれ、ランダムに選択した宛先無線局へ１０２４バイトのデータパケットを２乃至５００パケット／秒の速度で発生させる。従って、８個の発信元無線局のすべてが同時にデータ通信を開始するわけではない。１つの発信元無線局を選択すると、１５秒後に次の発信元無線局を選択する。ここで、通信はすべてシミュレーションの終わりまで継続する。こうした技術を使用したのは、ネットワークの性能を、開始時間を異にした同時的データ通信を行う多数の発信元無線局と宛先無線局とのペアによって試験するためである。表３に、使用したパラメータセットを示す。
【０１２９】
【表１】

【０１３０】
次いで、当該シミュレーションにおけるオーバーヘッドの影響について以下に説明する。
【０１３１】
ＧＬＳパケット信号又はＡＮＬパケット信号はともに周期的な更新パケットであり、かつこれらの伝搬は１ホップのブロードキャストに限定されているため、以下の解析が示すように、ＡＮＬテーブル又はＧＬＳテーブルを使用してアドホック無線ネットワークがフラッドすることは全くない。事実上本発明者らは、近似的なグローバルネットワーク状態情報及び非特許文献２０に記載されているフィッシュアイの概念に類する正確なローカル状態情報に依存している。従って、中間のノード無線局は、周囲のより正確なローカル情報を基礎としてルーティング決定を適応的に変更する。
【０１３２】
各更新パケットは、Ｔミリ秒の時間差で移動し、かつ１つのノード無線局から他のノード無線局への物理的移動にｔミリ秒を要するものとする。また、アドホックオペレーション用の有界領域をＡｍ^２、当該有界領域Ａ内のノード数をＮ及び各ノードの全方向性送信到達距離をＲとする。あるノード無線局がその隣接ノード無線局へ更新パケットをブロードキャストを行っているときは、そのノード無線局を取り巻く円形の送信ゾーン内にあるノード無線局は使用状態にあるが、有界領域のエリアＡにおける他の領域内にあるノード無線局はパケット信号のブロードキャストを実行できる。従って、ある平均的なケースでは、当該有界領域のエリアＡに渡ってトポロジーが均一に分散されていれば、当該有界領域のエリアＡにおいて複数の更新パケットが互いに干渉し合うことなくノード無線局間を同時に移動できるゾーンの数は、（Ａ／（πＲ^２））に等しい。ここで、ノード無線局は均一に分散されていることから、１つのゾーンに限定されて包含されるノード無線局の数（及び必然的に更新パケットの数Ｐ）は、次式で表される。
【０１３３】
【数９】

【０１３４】
言い替えれば、Ｐ個の更新パケットは、１つのノード無線局から他のノード無線局へ順次移動しなければならない。各更新パケットはＴミリ秒の時間差で移動し、かつそれにｔミリ秒（移動所要時間）を要するため、本媒体は、更新トラフィックにその時間の［ｔ×Ｐ×１００／Ｔ］％を占有される。例えばＧＬＳテーブルの場合、オペレーションの有界領域が１０００×１０００ｍ^２であって、送信ゾーンの到達距離（半径）Ｒが３００ｍであり、３０個のノード無線局を含むアドホック無線ネットワーク及び時間差Ｔ＝２ミリ秒に対するＧＬＳパケット信号の送信周期Ｔ_Ｇが５秒であれば、１つのゾーンに限定されて包含されるノード無線局の数（及び必然的に更新パケットの数）ＰはＰ＝８．４８となる。従って、本媒体は、ＧＬＳパケット信号を送信するトラフィックによってその時間のわずか０．３４％を占有されるだけである。ＡＮＬパケット信号の送信の場合、その送信周期Ｔ_Ａは１秒であり、かつ移動所要ｔ＝１ミリ秒である。従って、本媒体は、ＡＮＬパケット信号の送信によるトラフィックによってその時間のわずか０．８５％を占有されるだけである。従って、本媒体は、その時間の９８．８％を更新パケットから解放される。言い替えれば、「本媒体が更新トラフィックによって拘束されるのはその合計時間のわずか１．１９％であり、９８．８％はデータ通信プロセスに自由に使用することができる」。この方法の実際上の利得はこの点にあり、従来技術リンク状態ルーティングに取って代わる唯一の重要なモチベーションともなっている。
【０１３５】
本発明者らは、オーバーヘッドの解析のためのシミュレーションを、ランダムに配置された３０個のノード無線局よりなる静的トポロジー上に静的ルートを有するクアルネットに対して実行した。静的ルートを使用しているのは、ルーティング方法に関わりなく、制御トラフィックのオーバーヘッドの有無による性能を調べるためである。
【０１３６】
図１６は本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、平均スループットにおけるオーバーヘッドの影響を示すグラフであり、図１７は本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、平均エンド・ツー・エンドの遅延時間におけるオーバーヘッドの影響を示すグラフである。すなわち、図１６及び図１７におけるシミュレーションの結果は、更新パケットに起因するオーバーヘッドの影響がさほど重要ではないことを示している。当該シミュレーションにおける第１の実験は、オーバーヘッドの全くない状態で実行し、第２の実験は、ＡＮＬパケット信号の送信周期＝２秒かつＧＬＳパケット信号の送信周期＝１０秒（Ｔ_Ａ＝２，Ｔ_Ｇ＝１０）によるオーバーヘッドで実行し、第３の実験は、ＡＮＬパケット信号の送信周期＝１秒かつＧＬＳパケット信号の送信周期＝５秒（Ｔ_Ａ＝１，Ｔ_Ｇ＝５）によるオーバーヘッドを用いて実行している。
【０１３７】
本発明者らは、ＩＥＥＥ８０２．１１に準拠するＤＳＲ（例えば、非特許文献２３参照）及びそのＭＡＣプロトコルを本発明者らの提案の性能を比較評価するためのベンチマークとして使用した。本発明者らが行った評価は、４つの基準、すなわち「平均スループット」、「平均エンド・ツー・エンド遅延時間」、「ＡＣＫタイムアウトに起因する平均パケット再送信数」及び「再送信限界に起因する平均パケットドロップ数」に基づくものである。
【０１３８】
まず、２０個の静的スナップショットを取り上げ、その性能をこれらの４つの基準に照らしてＤＳＲと比較観察した。本発明者らが行った観察による平均値を図１８乃至図２１に図示し、パケットサイズ１０２４バイトのときのＣＢＲパケット到来速度２パケット／秒乃至５００パケット／秒で示されている。本実施形態に係るメカニズムには、電子制御導波器アレーアンテナ装置（ＥＳＰ）の見出しが付いている。高いデータレートでは、平均スループットは５００Ｋｂｐｓであって、これはＤＳＲの場合の５倍であり、平均エンド・ツー・エンドの遅延時間はＤＳＲに比べて３．５倍少ない１秒である。ＡＣＫタイムアウトに起因するパケット再送信数は、ＤＳＲでは１７５であるのに対して本実施形態のケースでは微々たるものである。同じく、平均パケットドロップ数も、本実施形態のケースはＤＳＲに比べて格段に少ない。
【０１３９】
多くの発信元無線局及び宛先無線局が高いデータレートで一度に通信する場合には、指向性アンテナを使用する媒体の利用が大幅に増加する可能性がある。ここで、これとともに、最大化されたゾーン間非結合パスを選択すれば、ルート間の共有する媒体へのアクセス獲得競争が大幅に軽減されるため、アドホック無線ネットワーク内のすべてのノードに渡ってネットワーク負荷が平衡化されるというシナリオを得ることができる。これらの２つの局面による効果が組み合わさると、最終的にはシステムの性能が劇的に向上し、図１８乃至図２１が示すように、スループットは増大し、エンド・ツー・エンドの遅延時間は低減する。
【０１４０】
図２２乃至図２５では、２００パケット／秒のデータレートにおける５ｍ／秒という低い移動度の下でＥＳＰの性能が評価されている。図２２乃至図２５から明らかなように、移動に対応するためには、ネットワーク情報の浸透をより速く行う必要があり、従って、ＡＮＬパケット信号の送信周期Ｔ_Ａ及びＧＬＳパケット信号の送信周期Ｔ_Ｇがそれぞれ２秒から１秒へ、かつ１０秒から５秒へ変更されている。制御トラフィックの増大に起因して性能に幾分かの低下が観察されるが、重要なものではない。
【０１４１】
以上説明したように、アドホック無線ネットワークにおける指向性アンテナの使用は、ルーティングの問題を適切な指向性ＭＡＣプロトコルに負荷のバランス化を加えて検討すれば、システムの性能を劇的に向上させることができる。本実施形態においては、最大化されたゾーン間非結合ルートが選択されたパス間のルート間結合の低減を促進し、これにより、エンド・ツー・エンドの遅延時間及びスループットが改善される。アドホック無線ネットワークにおけるＧＬＳパケット信号及びＡＮＬパケット信号の周期的伝搬に起因して生じる制御オーバーヘッドにも関わらず、その性能は、全方向性のＭＡＣプロトコルを使用する従来のリアクティブルーティングより遙かに優れている。提案している本実施形態に係る最大化されたゾーン間非結合によるテーブル駆動適応型ルーティングは、指向性アンテナの優位点を活用し、システム性能を向上させる。また、本発明者らの提案するルーティング方法はテーブル駆動式であるため、その成果は情報の浸透に依存するが、各パケット信号の送信周期Ｔ_Ａ及びＴ_Ｇの値によって決定された最新の制御オーバーヘッドは、高い移動度に十分に対応するものであると考えられる。
【０１４２】
【発明の効果】
以上詳述したように本発明によれば、複数の無線局を備え、各無線局間で無線通信を行う無線ネットワークのための制御方法又は無線通信システムにおいて、上記複数の無線局が通信可能な状態であるか否かを示すデータを上記各無線局毎に格納する第１のテーブルと、上記各無線局から無線通信可能な隣接無線局と当該各無線局から当該隣接無線局への方位角のデータを上記各無線局毎に格納する第２のテーブルとを上記各無線局の記憶手段に保存し、上記第１のテーブルを含む第１の無線信号を所定の第１の送信周期で周期的に送信するとともに、上記第２のテーブルを含む第２の無線信号を所定の第２の送信周期で周期的に送信し、上記送信された第１の無線信号を受信したとき、自局の記憶手段に保存された第１のテーブルのデータを更新するとともに、上記送信された第２の無線信号を受信したとき、自局の記憶手段に保存された第２のテーブルのデータを更新する。すなわち、上記第１のテーブルと上記第２のテーブルとに分割して、無線送信するので、従来技術に比較してオーバーヘッドを大幅に軽減でき、しかも上記２つのテーブルのデータの状況を迅速に当該無線ネットワーク内の各無線局に対して通知できる。
【０１４３】
ここで、上記第１の送信周期は上記第２の送信周期よりも短い。これにより、上記第１のテーブルにより示される、ある無線局における隣接無線局の状態をより早く把握できる一方、上記第２のテーブルにより示される第２のテーブルのデータについてはあまり迅速性は要求されず、概略の無線ネットワークのトポロジーを把握できればよい。また、上記第１のテーブルのデータは上記第２のテーブルのデータよりも小さいので当該無線ネットワークのトラフィックに与える影響も少なく、データ通信を効率的に実行できる。
【０１４４】
また、上記送信するときに、所定のトーン信号に続いて上記第１の無線信号を送信するとともに、上記トーン信号に続いて上記第２の無線信号を送信する。すなわち、従来技術のビーコン信号を使用せずに、迅速な伝搬が必要な上記第１のテーブルのデータに基づいて方向推定を行うことができ、各無線局の状態を高速で把握できる。
【０１４５】
さらに、最小のルーティング基準指数を有するパスを選択し、上記第２のテーブル内の当該選択されたパスに含まれる無線局を通信可能な無線局として設定し、上記選択されたパスに含まれる無線局に対する第１のテーブルのデータを第１の無線信号を用いて他の無線局に送信して上記他の無線局の第１のテーブルを更新した後、発信元無線局はその第２のテーブル内のデータに基づいて上記選択されたパスを介して宛先無線局と無線通信する。従って、当該無線ネットワーク全体の負荷を平衡化したシングルパスルーティングを実現できる。
【図面の簡単な説明】
【図１】 本発明に係る実施形態であるアドホック無線ネットワークを構成する複数の無線局１−１乃至１−９の平面配置図である。
【図２】 図１の各無線局１の内部構成を示すブロック図である。
【図３】 図１の可変ビームアンテナ１０１のセクタビームパターンの一例を示す図である。
【図４】 図１のアドホック無線ネットワークにおいて用いられるパケットデータのフォーマットを示す図である。
【図５】 本実施形態に係る無線通信システムの説明において用いる各ノード無線局からの送信セクタ放射パターンである送信ゾーンＢｎ（α，β，Ｒ）を示す平面図である。
【図６】 図２のデータベースメモリ１５４において格納される方位角及び信号強度レベルテーブル（ＡＳテーブル）の一例を示す表である。
【図７】 図２のデータベースメモリ１５４に格納される隣接リンク状態テーブル（ＮＬＳテーブル）の一例を示す図である。
【図８】 図２のデータベースメモリ１５４に格納されるアクティブノード無線局テーブル（ＡＮＬテーブル）の一例を示す図である。
【図９】 図２のデータベースメモリ１５４に格納されるグローバルリンク状態テーブル（ＧＬＳテーブル）の一例を示す図である。
【図１０】 図１のアドホック無線ネットワークにおいて用いられるトーン信号とパケット信号の送受信処理を示すタイミングチャートである。
【図１１】 図１のアドホック無線ネットワークにおいて用いられる各無線局での放射パターンの種類と無線通信プロトコルを示すタイミングチャートである。
【図１２】 図１のアドホック無線ネットワークにおいて用いられるＤＮＡＶを用いた方法を示す無線局の平面図である。
【図１３】 図２の無線局１の管理制御部１０５によって実行されるパケット送受信制御処理を示すフローチャートである。
【図１４】 図１３のサブルーチンであるルーティング及び通信処理（ステップＳ１２）の処理を示すフローチャートである。
【図１５】 本実施形態で用いる、宛先無線局に到達するための中間ノード無線局による適応ルーティング選択処理を示すノード無線局の平面図である。
【図１６】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、平均スループットにおけるオーバーヘッドの影響を示すグラフである。
【図１７】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、平均エンド・ツー・エンド遅延時間におけるオーバーヘッドの影響を示すグラフである。
【図１８】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケットサイズが１０２４バイトのときの異なるパケット到着レート（１秒当たりのパケット数）に対する、ＤＳＲとＥＳＰの場合における平均スループットを示すグラフである。
【図１９】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケットサイズが１０２４バイトのときの異なるパケット到着レート（１秒当たりのパケット数）に対する、ＤＳＲとＥＳＰの場合における平均エンド・ツー・エンド遅延時間を示すグラフである。
【図２０】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケットサイズが１０２４バイトのときの異なるパケット到着レート（１秒当たりのパケット数）に対する、ＤＳＲとＥＳＰの場合における平均パケット再送信数を示すグラフである。
【図２１】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケットサイズが１０２４バイトのときの異なるパケット到着レート（１秒当たりのパケット数）に対する、ＤＳＲとＥＳＰの場合における平均パケットドロップ数を示すグラフである。
【図２２】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケット到着レートが２００パケット／秒のときの、ＥＳＰを有するノード無線局が固定したときと移動したときにおける平均スループットを示すグラフである。
【図２３】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケット到着レートが２００パケット／秒のときの、ＥＳＰを有するノード無線局が固定したときと移動したときにおける平均エンド・ツー・エンド遅延時間を示すグラフである。
【図２４】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケット到着レートが２００パケット／秒のときの、ＥＳＰを有するノード無線局が固定したときと移動したときにおける、ＡＣＫ信号のタイムアウトによる平均パケット再送信数を示すグラフである。
【図２５】 本実施形態に係るアドホック無線ネットワークのシミュレーション結果であって、パケット到着レートが２００パケット／秒のときの、ＥＳＰを有するノード無線局が固定したときと移動したときにおける、再送信制限による平均パケットドロップ数を示すグラフである。
【図２６】 従来技術に係る、指向性アンテナを用いた無線局Ｓ_１−Ｄ_１間及び無線局Ｓ_２−Ｄ_２間の２つのゾーン非結合通信を示す平面図である。
【符号の説明】
１，１−１乃至１−９，１−ｉ，１−ｊ，１−ｋ，１−ｌ，Ｎ_１乃至Ｎ_６…ノード無線局、
ＢＳ_１，ＢＳ_２，ＢＮ_１乃至ＢＮ_６…送信ゾーン、
Ｓ，Ｓ_１，Ｓ_２…発信元無線局、
Ｄ，Ｄ_１，Ｄ_２…宛先無線局、
１０１…可変ビームアンテナ、
１０２…サーキュレータ、
１０３…指向制御部、
１０４…パケット送受信部、
１０５…トラヒックモニタ部、
１０６…回線制御部、
１０７…上位レイヤ処理装置、
１３０…パケット受信部、
１３１…高周波受信機、
１３２…復調器、
１３３…受信バッファメモリ、
１４０…パケット送信部、
１４１…送信タイミング制御部、
１４２…送信バッファメモリ、
１４３…変調器、
１４４…高周波送信機、
１５１…管理制御部、
１５２…検索エンジン、
１５３…更新エンジン、
１５４…データベースメモリ、
１５５…クロック回路、
１６０…拡散符号発生器。[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a control method and a wireless communication system for a wireless network such as an ad hoc wireless network that performs packet communication in a wireless network such as a wireless LAN provided with a plurality of wireless stations.
[0002]
[Prior art]
In an ad hoc wireless network that supports communication between an unspecified number of people temporarily gathered in a specific area using a wireless line, there is no infrastructure such as a router device of the Internet. Users need to relay and route packets in cooperation.
[0003]
As a routing of an ad hoc wireless network, for example, Non-Patent Document 25 proposes a method of obtaining route information by transmitting a route search packet. However, in this method, since an omnidirectional antenna is used, co-channel interference is likely to occur, the bit error rate (BER) increases, power consumption during packet transmission / reception increases, and the battery of the terminal device is increased. There was a problem that the load became high.
[0004]
In order to solve this problem, in Patent Document 2, “in a routing method for a wireless network that performs packet communication between a plurality of wireless stations, each wireless station in a service area among the plurality of wireless stations is described. The signal power-to-noise power ratio (SINR) for each predetermined azimuth angle is measured in advance and stored in the storage device as an SINR table, and when the packet is transmitted to the destination wireless station, the first hop is based on the SINR table. A routing method for routing a packet signal by transmitting a packet by controlling a variable beam antenna so as to form a beam for the determined first hop wireless station. Has been.
[0005]
In this routing method, in order to realize effective routing not only by starting communication but also by using directivity and adaptive control type antennas, a radio station effectively sets its transmission direction and transmits a packet to it. You must know how to send to your neighbors. Accordingly, each wireless station periodically collects its neighbor information and forms the above-described SINR table. At the same time, in order to form this SINR table, the following steps are performed periodically and asynchronously by any wireless station. Need to be executed.
[0006]
(I) Whenever a radio channel is available, the radio station transmits a setup packet signal in an omni pattern to neighboring radio stations.
(Ii) The radio station scans its directional antenna and sequentially transmits an azimuth angle request (RQ: ReQuest) packet signal in each azimuth direction in a directional broadcast format in all directions. For example, if this is performed at 30 degree intervals and the entire 360 degree space is sequentially covered, 12 directional RQ packet signals are generated. The RQ packet signal includes a radio station ID and broadcast azimuth information.
(Iii) Upon receiving the setup packet signal from the transmission source radio station, each radio station adjacent to the transmission source radio station waits in a reception mode for a predetermined time, and Confirm the end of direction-oriented broadcast (sequential RQ packet signal). Thereby, at that time, it is ensured that each adjacent radio station has received all RQ packet signals for all directions generated by the radio station of the transmission source. In other words, each adjacent radio station can accumulate all the columns of each adjacent radio station in the SINR table indicating the SINR with respect to the azimuth angle of the source radio station, and the directionality from the source radio station Based on the RQ packet signal, the signal strength of the source radio station in that direction is determined.
(Iv) Each adjacent wireless station transmits this information as a response (RE: REply) packet signal to the transmitting wireless station. The RE packet signal is processed as a data packet. When this information is received from all radio stations adjacent to the transmission source radio station, the SINR table is completed.
[0007]
According to this routing method, it is possible to periodically collect information on adjacent wireless stations indoors and outdoors, and a communication method having directivity and an adaptive antenna can be more effectively utilized using this information. It has the advantage of making things.
[0008]
By the way, the routing method of the ad hoc wireless network usually adopts routing by a single path through wireless stations of a plurality of user terminals. However, for example, as disclosed in Non-Patent Document 26, once a set consisting of a plurality of paths is found between a source radio station and a destination radio station, the total capacity of data is divided into a plurality of individual capacities. An end-to-end delay time (between two terminals, i.e., source radio station) by dividing into blocks and transmitting these from the source radio station to the destination radio station via the selected paths. It may be possible to improve the transmission time of packets from the destination wireless station to the destination wireless station, which ultimately reduces network congestion and end-to-end delay time. Such routing is called “multipath routing”.
[0009]
Further, if a directional antenna is used in each node radio station of an ad hoc radio network, radio interference can be greatly reduced, and thus the use of a radio medium and inevitably the performance of the network can be improved. Documents 1 to 10 have been demonstrated.
[0010]
[Patent Document 1]
JP 2001-024431A.
[Patent Document 2]
JP 2001-244983 A
[Non-Patent Document 1]
T. S. Yum et al., “Design algorithms for multihop packet radio networks with multiple directional antennas stations”, IEEE Transactions on communications, Vol. 40, No. 11, pp. 1716-1724, 1992.
[Non-Patent Document 2]
Y. B. Ko et al., Medium access control protocols using directional antennas in ad hoc networks ", 'Proceeding of the IEEE INFOCOM 2000, March 2000.
[Non-Patent Document 3]
Romit Roy Choudhury et al., "Media Access Control for Ad Hoc Networks: Using directional antennas for medium access control in ad hoc networks", Proceedings of the eighth annual international conference on Mobile computing and networking, September 2002.
[Non-Patent Document 4]
Nasipuri, S. Ye, J. You et al., "A MAC Protocol for Mobile Ad Hoc Networks Using Directional Antennas", Proceeding of the IEEE WCNC 2000, 2000.
[Non-Patent Document 5]
S. Bandyopadhyay et al., "An Adaptive MAC Protocol for Wireless Ad Hoc Community Network (WACNet) Using Electronically Steerable Passive Array Radiator Antenna", Proc. Of the GLOBECOM 2001, San Antonio, Texas, USA, November 25-29, 2001 .
[Non-Patent Document 6]
R. Ramanathan, “On the Performance of Ad Hoc Networks with Beamforming Antennas”, ACM MOBIHOC, October 2001.
[Non-Patent Document 7]
M. Takai et al., “Directional Virtual Carrier Sensing for Directional Antennas in Mobile Ad Hoc Networks”, ACM MOBIHOC, June 2002.
[Non-Patent Document 8]
Lichun Bao et al., "Transmission scheduling in ad hoc networks with directional antennas", Proceedings of the eighth annual international conference on Mobile computing and networking, Atlanta, Georgia, U.S.A., 2002.
[Non-patent document 9]
Asis Nasipuri et al., “Power Consumption and throughput in Mobile Ad Hoc Networks using Directional Antennas” in Proceedings of the IEEE International Conference on Computer Communication and Networks (ICCCN 2002), Miami, Florida, U.S.A., October 14-16, 2002.
[Non-Patent Document 10]
S. Bandyopadhyay et al., "An Adaptive MAC and Directional Routing Protocol for Ad Hoc Wireless Network Using Directional ESPAR Antenna", Proceeding of the ACM Symposium on Mobile Ad Hoc Networking & Computing 2001 (MOBIHOC 2001), Long Beach, California, USA , 4-5 October 2001.
[Non-Patent Document 11]
Kui Wu et al., “On-Demand Multipath Routing for Mobile Ad Hoc Networks”, EPMCC 2001, Vienna, Austria, 20th 22nd February 2001.
[Non-Patent Document 12]
M. R. Pearlman et al., “On the Impact of Alternate Path Routing for Load Balancing in Mobile Ad Hoc Networks”, MOBIHOC 2000, pp. 150, 3-10, 2000.
[Non-Patent Document 13]
Hossam Hassanein et al., "Routing with load balancing in wireless Ad hoc networks", Proceedings of the 4th ACM international workshop on Modeling, analysis and simulation of wireless and mobile systems, Rome, Italy, 2001.
[Non-Patent Document 14]
Sung-Ju Lee et al., "Dynamic Load-Aware Routing in Ad hoc Networks", Proceedings of IEEE ICC 2001, Helsinki, Finland, June 2001.
[Non-Patent Document 15]
S.J. Lee et al., "Split Multi-path Routing with Maximally Disjoint Paths in Ad Hoc Networks", ICC 2001, 2001.
[Non-Patent Document 16]
Somprakash Bandyopadhyay et al., “Multipath Routing in Ad Hoc Networks with Directional Antenna”, Proceeding of the IFIP TC6 / WG6.8 Conference on Personal Wireless Communications (PWC 2002), Singapore, October 2002.
[Non-Patent Document 17]
T. Ohira et al., “Electronically Steerable Passive Array Radiator (ESPAR) Antennas for Low-cost Adaptive Beam forming”, IEEE International Conference on Phased Array Systems, Dana Point, California, U.S.A., May 2000.
[Non-Patent Document 18]
K. Gyoda et al., “Beam and Null Steering Capability of ESPAR Antennas”, Proceedings of the IEEE AP-S International Symposium, July 2000.
[Non-Patent Document 19]
Asis Nasipuri et al., “On-Demand Routing using Directional Antennas in Mobile Ad Hoc Networks, IEEE ICCCN 2000, 2000.
[Non-Patent Document 20]
Guangyu Pei et al., “Fisheye State Routing: A Routing Scheme for Ad Hoc Wireless Networks”, in Proceeding of the IEEE International Conference on Communication, New Orleans, Louisiana, June 2000.
[Non-patent document 21]
Romit Roy Choudhury et al., "A Distributed Mechanism for Topology Discovery in Ad hoc Wireless Networks using Mobile Agents", Proceedings of the First Annual Workshop On Mobile Ad Hoc Networking & Computing (MOBIHOC 2000), Boston, Massachusetts, USA, August 11 , 2000.
[Non-Patent Document 22]
QualNet Simulator Version 3.1, Scalable Network Technologies, URL: //www.scalable-networks.com, May 20, 2003.
[Non-Patent Document 23]
J. Broch et al., "A Performance Comparison of Multi-Hop Wireless Ad Hoc Network Routing Protocols", Proceedings of ACM / IEEE Mobile Computer and Network, Dallas, Texas, U.S.A., October 1998.
[Non-patent Document 24]
Romit Roy Choudhury et al., "An Agent-based Connection Management Protocol for Ad Hoc Wireless Networks", Journal of Network and System Management, December 2002.
[Non-Patent Document 25]
D. B. Johnson, et al., "Dynamic Source Routing in Ad Hoc Wireless Networks", in book on "Mobile Computing", Chapter 5, pp.153-181, Kluwer Academic Publishers, 1996.
[Non-Patent Document 26]
S. K. Das et al., “Improving Quality-of-Service in Ad hoc Wireless Networks with Adaptive Multi-path Routing”, Proceedings of the GLOBECOM 2000, San Francisco, California, U.S.A., November 27-December 1, 2000.
[Non-Patent Document 27]
Tetsuro Ueda et al., "An Approach towards improving Quality of Service in ad hoc Networks with ESPAR Antenna", Proceedings of The 16th International Workshop on Communications Quality & Reliability, CQR 2002, May 14-17, 2002.
[0011]
[Problems to be solved by the invention]
However, at the same time, it is difficult for each node radio station to find a way to set and control the direction of these antennas so as to achieve the expected performance improvement in an ad hoc wireless network multi-hop communication environment. There was a point. The difficulty of this problem is mainly due to the lack of mobility and centralized management in ad hoc wireless networks. Therefore, it is a challenge to develop proper MAC and routing protocols in ad hoc wireless networks that take advantage of directional antennas to improve overall performance.
[0012]
Recently, several Media Access Control (MAC) protocols using directional antennas have been used in wireless communication systems for ad hoc wireless networks to increase the number of simultaneous communications and improve media utilization. Proposed. However, even with an efficient directional MAC protocol, excellent system performance cannot be guaranteed by itself unless it has an appropriate routing method that uses the advantages of directional antennas.
[0013]
Therefore, the first object of the present invention is to utilize the advantages of directional antennas in ad hoc wireless networks in addition to the MAC protocol, greatly reducing overhead and greatly reducing the end-to-end delay time. It is an object of the present invention to provide a control method including a single path routing method that can be reduced and a wireless communication system using the same.
[0014]
FIG. 26 shows a conventional radio station S using a directional antenna.₁-D₁And radio station S₂-D₂It is a top view which shows non-bonded communication between two zones between. In FIG. 26, BS₁, BS₂, BN₁To BN₆Indicates a transmission zone of a sector radiation pattern provided in each wireless station. Here, the source radio station S₁Node radio station N₁And N₂Destination radio station D via₁When considering communication with other source wireless station S₂Is the destination radio station D₂Want to communicate with. Possible paths include the path {S₂, N₁, N₂, D₂}, Path {S₂, N₃, N₄, D₂} And path {S₂, N₅, N₆, D₂} Three paths exist. Source radio station S₂Is the source radio station S₁If the first path that overlaps the path used by is used, the routing performance cannot be improved simply by using a directional antenna. Source radio station S₂However, if the second path is used, routing performance deteriorates due to a phenomenon known as inter-route coupling (see, for example, Non-Patent Documents 11, 12, and 16).
[0015]
Inter-route coupling occurs when two routes are in close physical proximity to interfere with each other during data communication. As a result, each node radio station in these two routes is constantly in competition for the medium they share. In the case of FIG. 26, even if a directional antenna is used as shown in FIG. 26, since the node radio stations belonging to these two routes exist in each other's transmission zone, these two communications are simultaneously performed. Cannot occur. That is, the node radio station N₁And N₃Is the source radio station S₁And S₂Cannot receive data at the same time. Similarly, the node radio station N₂And N₄Each node wireless station N₁And N₃Cannot receive data simultaneously.
[0016]
Therefore, the routing performance between any source radio station and destination radio station does not depend only on the congestion characteristics of the nodes in the path. Communication patterns in adjacent areas will also affect this delay time. This is a phenomenon known as inter-route coupling. Thus, the path {S₁, N₁, N₂, D₁}, Path {S₂, N₃, N₄, D₂} Is non-node-to-node, and even if a directional antenna is used, routing performance is degraded in the wireless communication system for this ad hoc wireless network.
[0017]
The influence of the directional antenna on the routing is as follows.₂Is the third path, ie the path {S₂, N₅, N₆, D₂} Will be apparent when selecting. These two routes, the path {S₁, N₁, N₂, D₁} And path {S₂, N₅, N₆, D₂} Are coupled to each other when an omni antenna is used (see the dotted line in FIG. 26). However, as FIG. 26 shows, using a “directional antenna” is “fully decoupled”. These two routes are called “inter-zone uncoupled” because the data communication on one path does not interfere with the data communication on the other path.
[0018]
Therefore, in order to utilize the capability of a directional antenna to improve medium utilization, it is essential to always equip with a routing method having “effective load balancing”. Recently, a routing method having load balancing in an ad hoc wireless network using an omni antenna has been developed (see, for example, Non-Patent Documents 13 and 14). In this routing method, the routing load of intermediate node radio stations or the activity information of all node radio stations is regarded as a main route selection criterion.
[0019]
Applications of multipath routing technology in mobile ad hoc wireless networks are also being considered to reduce end-to-end delay times and perform load balancing. M.R.Perlman et al. Have demonstrated in recent papers such as Non-Patent Document 12 that multipath routing can balance network load. In addition, split multipath routing (SMR) proposed in Non-Patent Document 15 focuses on constructing and holding a maximized non-coupled multipath. However, none of these proposals considered the inter-route coupling phenomenon for effective load balancing.
[0020]
Distributing the routing process evenly across the network has two major advantages in this wireless communication system. First, it is a commonly used node radio station to prevent the load from concentrating on one node radio station set and to spread it evenly among other node radio stations. The possibility of power depletion (a phenomenon in which the amount of transmission power used is large and energy in the battery is immediately lost) is reduced. Second, it distributes traffic throughout, reducing congestion and improving end-to-end delay time. Most of the current proposals for load balancing in this context are to promote overall traffic distribution, so the first advantage mentioned above can be achieved.
[0021]
However, due to the inter-route coupling in the wireless medium, as shown in FIG. 26, the end-to-end delay time improvement cannot be guaranteed only by traffic distribution. As FIG. 26 shows, the path {S₁, N₁, N₂, D₁} And path {S₂, N₃, N₄, D₂} Is “node uncoupled” and therefore meets the criteria for load balancing. However, since they are coupled together, the end-to-end delay time is increased. As the degree of coupling increases, the average end-to-end delay time of both paths increases (see, for example, Non-Patent Document 11). This is because the two paths have more opportunities to interfere with each other's transmission due to the radio propagation broadcast characteristics. This is the reason why the discovery of “inter-zone uncoupled routes” for “effective load balancing” is important.
[0022]
However, even using an omni antenna to obtain an inter-zone non-coupled path, or a part thereof, is difficult due to the wide transmission zone. In the case of an omni antenna, the area of the transmission zone of each node radio station is πR²Can be expressed as Here, the beam angle θ = 360 ° and the transmission reach distance (radius) is R. If the beam angle θ is controlled using a directional antenna (θ <360 °), the effective range of each node radio station can be reduced to the beam θ (transmission zone area = R²/ 2). In the example set by the present inventors, a two-route path {S₁, N₁, N₂, D₁} And path {S₂, N₅, N₆, D₂} Becomes “inter-zone uncoupled” only when a directional antenna is used. Using directional antennas instead of omni antennas with radio stations of each user terminal forming an ad hoc wireless network makes it much easier and inevitably to obtain an “inter-zone uncoupled” route. It has been proven that the effects of inter-route coupling can be dramatically reduced (see, for example, Non-Patent Document 16).
[0023]
Accordingly, a second object of the present invention is to provide a control method including a routing method of single path routing that can execute effective load balancing in an ad hoc wireless network, and a wireless communication system using the same.
[0024]
[Means for Solving the Problems]
A control method for a wireless network according to the first invention is a control method for a wireless network that includes a plurality of wireless stations and performs wireless communication between the wireless stations.
A first table storing data indicating whether or not the plurality of radio stations are in a communicable state for each radio station; an adjacent radio station capable of radio communication from each radio station; and each radio station Storing in the storage means of each radio station a second table storing azimuth angle data from the radio station to the adjacent radio station for each radio station;
The first radio signal including the first table is periodically transmitted at a predetermined first transmission cycle, and the second radio signal including the second table is periodically transmitted at a predetermined second transmission cycle. Sending automatically,
When the transmitted first wireless signal is received, the data of the first table stored in the storage means of the own station is updated, and when the transmitted second wireless signal is received, the own station And updating the data in the second table stored in the storage means.
[0025]
In the control method for the wireless network, the first transmission cycle is shorter than the second transmission cycle.
[0026]
Further, in the control method for the wireless network, the transmitting step transmits the first wireless signal following a predetermined tone signal, and transmits the second wireless signal following the tone signal. It is characterized by doing.
[0027]
Further, in the control method for the wireless network, all hop counts H are smaller than a predetermined maximum value Hmax between the source wireless station and the destination wireless station based on the second table. A first step of searching for a path;
In the wireless network, a wireless station on one bus P of the plurality of searched paths is assumed to be a communicable wireless station, and is included in the one path P in the second table. The data of the first table for the wireless station to be transmitted is set in a communicable state, and based on the second table, between the source wireless station and the destination wireless station other than the source wireless station and the destination wireless station The number of other communicable radio stations existing in the transmission zone of the service area of each radio station on the one path P is calculated for the communicable path P ′, and the calculated communication By adding the number of possible radio stations for each of the one path P, the 1 for the path P ′ that can be communicated between the source radio station and the destination radio station other than the source radio station and the destination radio station. Relationship of two paths P By calculating η (P) and multiplying the calculated correlation coefficient η (P) by the number of hops H of the one path P, the multiplication result is obtained as the routing reference index γ (P ), And setting and returning the radio station included in the one path P in the second table as a radio station incapable of communication in the first table;
A third step of calculating a routing criterion index γ (P) for all of the searched paths by repeating the processing of the second step for all of the searched paths;
Select the path corresponding to the smallest routing criterion index γ (P) among the calculated routing criterion indices γ (P) for all the searched paths, and the selected path in the second table. Is set as a communicable radio station, and data of the first table for the radio station included in the selected path is transmitted to another radio station using the first radio signal. After updating the first table of the other radio station, the source radio station wirelessly communicates with the destination radio station via the selected path based on the data in the second table; Is further included.
[0028]
A wireless communication system for a wireless network according to a second invention is a wireless communication system for a wireless network that includes a plurality of wireless stations and performs wireless communication between the wireless stations.
A first table storing data indicating whether or not the plurality of radio stations are in a communicable state for each radio station; an adjacent radio station capable of radio communication from each radio station; and each radio station Storage means for storing for each radio station a second table for storing azimuth angle data from the radio station to the adjacent radio station for each radio station;
The first radio signal including the first table is periodically transmitted at a predetermined first transmission cycle, and the second radio signal including the second table is periodically transmitted at a predetermined second transmission cycle. Means for transmitting automatically,
When the transmitted first wireless signal is received, the data of the first table stored in the storage means of the own station is updated, and when the transmitted second wireless signal is received, the own station And a control means for updating data of the second table stored in the storage means.
[0029]
In the wireless communication system for the wireless network, the first transmission period is shorter than the second transmission period.
[0030]
In the wireless communication system for the wireless network, the transmission means transmits the first wireless signal following a predetermined tone signal, and transmits the second wireless signal following the tone signal. It is characterized by doing.
[0031]
Further, in the radio communication system for the radio network, all the hop count H is smaller than a predetermined maximum value Hmax between the source radio station and the destination radio station based on the second table. First control means for searching for a path of
In the wireless network, a wireless station on one bus P of the plurality of searched paths is assumed to be a communicable wireless station, and is included in the one path P in the second table. The data of the first table for the wireless station to be transmitted is set in a communicable state, and based on the second table, between the source wireless station and the destination wireless station other than the source wireless station and the destination wireless station The number of other communicable radio stations existing in the transmission zone of the service area of each radio station on the one path P is calculated for the communicable path P ′, and the calculated communication By adding the number of possible radio stations for each of the one path P, the 1 for the path P ′ that can be communicated between the source radio station and the destination radio station other than the source radio station and the destination radio station. Relationship of two paths P By calculating η (P) and multiplying the calculated correlation coefficient η (P) by the number of hops H of the one path P, the multiplication result is obtained as the routing reference index γ (P ), And a second control unit that sets and returns the wireless station included in the one path P in the second table as a wireless station that cannot communicate in the first table;
Third control means for calculating a routing reference index γ (P) for all the searched paths by repeating the processing of the second step for all the searched paths;
Select the path corresponding to the smallest routing criterion index γ (P) among the calculated routing criterion indices γ (P) for all the searched paths, and the selected path in the second table. Is set as a communicable radio station, and data of the first table for the radio station included in the selected path is transmitted to another radio station using the first radio signal. After updating the first table of the other radio station, the source radio station performs radio communication with the destination radio station via the selected path based on the data in the second table. And further comprising.
[0032]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments according to the present invention will be described below with reference to the drawings.
[0033]
FIG. 1 is a plan layout diagram of a plurality of radio stations 1-1 to 1-9 (generally referred to by reference numeral 1) showing the configuration of an ad hoc radio network according to an embodiment of the present invention. These are block diagrams which show the structure of each radio station 1 of FIG.
[0034]
In the wireless communication system of this embodiment, as shown in FIG. 1, a plurality of wireless stations 1 are present in a plane and each wireless station 1 has a gain, transmission power, and reception sensitivity of the variable beam antenna 101. A predetermined service area determined by parameters such as, and can perform packet communication within the service area. When performing packet communication with the wireless station 1 outside the service area, the wireless station 1 within the service area Is used as a relay station to relay packet data, thereby transmitting the packet data to a desired destination wireless station 1. That is, each wireless station 1 has a router function for routing packets and operates as a source wireless station, a relay station, or a destination wireless station.
[0035]
The wireless communication system of this embodiment is applied to a packet communication system of an ad hoc wireless network such as a wireless LAN, for example, and includes an omni pattern that is an omnidirectional radiation pattern and a predetermined plane in a horizontal plane centered on the own station. A variable beam antenna 101 capable of selectively switching between a sector beam pattern capable of selectively changing a sector-shaped main beam for each azimuth angle and an exclusive sector pattern capable of forming a null point for each azimuth angle. The following tables (a) to (e) are stored in the database memory 154, and packet signals are routed while controlling the radiation pattern of the variable beam antenna 101 based on these tables. .
[0036]
(A) Adjacent node wireless station that can be measured when receiving a radio signal from an adjacent node radio station (a node radio station capable of radio communication from the own station) 1 in a service area centered on the own station An azimuth and signal strength level table (hereinafter referred to as an AS table) that stores an azimuth angle and a signal strength level with respect to the radio station 1.
(B) In a horizontal plane centered on the own station, acquired in advance based on the radio signal from the adjacent node radio station including the measured value of SINR when the adjacent node radio station receives the radio signal from the own station. Based on the SINR viewed from the radio station 1 in the service area for each predetermined azimuth angle, the maximum SINR is selected for each adjacent node radio station and set as the affinity with each adjacent node radio station. An adjacent link state table (Neighbor Link-State Table) including the corresponding azimuth angle and the update time of the data (hereinafter referred to as NLS table).
(C) Among the node radio stations in the ad hoc radio network, an active node list table (Active node list) including a list of node radio stations participating in the ad hoc radio network and in a communicable state (hereinafter referred to as an active state) Node List Table) (hereinafter referred to as ANL table).
(D) Among the adjacent node radio stations in the NLS table, a neighboring active node list table (Neighbor Active Node List Table) including a list of adjacent node radio stations that are active in the ad hoc radio network , Referred to as a NANL table).
(E) By exchanging topology information of the NLS table (referred to as path information indicating the link state between all radio stations in the ad hoc radio network, the same applies hereinafter) between the radio stations, the ad hoc radio network NLS table topology information related to all wireless stations 1 in the network is collected, this topology information (including the link state between each wireless station 1), and the number of updates (version version) of the topology information for each wireless station in this topology information And a communication state flag value indicating the wireless station 1 involved in an arbitrary communication process at that time (Global Link State Table) (hereinafter referred to as a GLS table).
[0037]
In the present embodiment, the MAC and routing protocol of an ad hoc wireless network using a directional antenna for performing effective load balancing is selected using a route selection process that maximizes “inter-zone non-coupling”. Show. In the MAC protocol presented by the present inventors, each node radio station 1 dynamically retains predetermined adjacency information, and thus each node radio station 1 has “the adjacency and communication occurring in that adjacency at that time”. "Recognize". This facilitates each node radio station 1 to avoid interference by tracking other neighboring node radio stations 1 in its neighbor at that time. At the same time, it also tracks the directional access information of its neighboring node radio station 1. This is useful when each node radio station 1 determines the communication direction that seems to be the best with any neighboring node radio station 1. This information from each node radio station 1 is periodically propagated to the adjacent node radio station 1, and the adjacent node radio station 1 takes in this information, and the adjacent node radio station 1 further in a cycle interval (transmission cycle). Propagate to. In this way, information penetrates the entire network and facilitates each node radio station 1 to periodically capture the proper network state without generating much control traffic. Therefore, each node is substantially in a state of “recognizing topology” and “recognizing communication occurring in the network”.
[0038]
The inventors have proposed a table-driven routing protocol for load balancing routing, defining and developing criteria for measuring the degree of maximization of “inter-zone uncoupling”, and This was used as a route selection criterion. However, since the network awareness in each node radio station 1 is merely a “knowledge” of the network state, not the “real” state of the network, Modify and modify routing decisions adaptively. The performance of the framework of the wireless communication system according to the present embodiment proposed in the present invention is, as will be described in detail later, dynamic source (source wireless station) routing (performed by a known Qualnet). Evaluation is performed based on a Qualnet network simulator (for example, see non-patent document 22) using Dynamic Source Routing (DSR) (for example, refer to non-patent document 23) as a benchmark.
[0039]
Next, the device configuration of each radio station 1 will be described with reference to FIG. In FIG. 2, a radio station 1 includes a variable beam antenna 101, a directivity control unit 103 for controlling the directivity thereof, a circulator 102, a data packet transmission / reception unit having a data packet transmission unit 140 and a data packet reception unit 130. 104, a traffic monitor unit 105, a line control unit 106, and an upper layer processing apparatus 107.
[0040]
Transmission signal data for communication in packet format generated by the upper layer processing device 107 that processes data to be transmitted / received is input to the modulator 143 via the transmission buffer memory 142, and the modulator 143 has a predetermined radio frequency. The carrier wave signal is subjected to spread spectrum modulation in accordance with the input communication transmission signal data using a predetermined communication channel spreading code generated by the spread code generator 160 by the CDMA method, and the modulated transmission signal is transmitted at a high frequency. Output to the machine 144. The high-frequency transmitter 144 performs processing such as amplification on the input transmission signal, and then transmits the variable signal from the variable beam antenna 101 to another wireless station 1 via the circulator 102. On the other hand, the communication signal received in the packet format received by the variable beam antenna 101 is input to the high-frequency receiver 131 via the circulator 102, and the high-frequency receiver 131 performs low noise amplification on the input received signal. After executing the above process, the data is output to the demodulator 132. The demodulator 132 demodulates the input received signal by spectrum despreading using the communication channel spreading code generated by the spreading code generator 160 by the CDMA method, and receives the demodulated received signal data as a receiving buffer. The data is output to the upper layer processing apparatus 107 via the memory 133 and also output to the traffic monitor unit 105 for traffic monitoring.
[0041]
In the present embodiment, the variable beam antenna 101 that is a directional antenna is connected to a plurality of antenna elements and a control unit 103 that controls the directivity,
(A) an omni pattern which is an omnidirectional radiation pattern;
(B) For example, as shown in FIG. 3, a sector beam pattern capable of selectively changing a sector-shaped main beam for each predetermined azimuth angle in a horizontal plane centered on the own station;
(C) an exclusive sector pattern capable of forming a null point for each azimuth angle;
Is an antenna that can be selectively switched by electrical control. The variable beam antenna 101 may be, for example, a known phased array antenna device, or an electronically controlled waveguide array antenna device (Electronically disclosed in Patent Document 1, Non-Patent Documents 17 and 18). A variable beam antenna that is a Steerable Passive Array Radiator Array Antenna Apparatus) may be used.
[0042]
The traffic monitor unit 105 includes a search engine 152, an update engine 153, a database memory 154, and a clock circuit 155, and performs routing and communication processing described later, and the wireless station 1 communicates with other wireless stations 1. The communication channel to be used in the packet communication is determined, and the spread code generator 160 sends the spread code designation data corresponding to the determined communication channel to the spread code generator 160 via the line control unit 106. The transmission timing control unit is controlled by generating a spreading code corresponding to the designated data and sending the designated data of the time slot corresponding to the determined communication channel to the transmission timing control unit 141 via the line control unit 106. 141 is a communication channel transmission signal by the transmission buffer memory 142 It controls to transmit the signal for the communication channel is transmitted in the corresponding time slot by controlling the writing and reading of over data. Note that the clock circuit 155 measures the current date and time and outputs the information to the management control unit 151 as necessary.
[0043]
The search engine 152 of the traffic monitor unit 105 searches the data in the database memory 154 under the control of the management control unit 151 and returns the searched data to the management control unit 151. The update engine 153 updates data in the database memory 154 under the control of the management control unit 151. Further, the database memory 154 stores an AS table, an NLS table, an ANL table, a NANL table, and a GLS table that is a table for routing.
[0044]
In this embodiment, the effective transmission beam width of the sector beam pattern for changing the directivity so that the gain in the direction of a single communication partner is maximized is set to 30 °, and the variable beam antenna 101 is used. The azimuth angle can be selectively changed every 30 °. The beam width and azimuthal angle of change may be 60 ° or other angles. Further, the packet data used in the packet communication system of the present embodiment has a format of the format shown in FIG. That is, the packet data includes the ID of the destination wireless station, the packet type (tone, GLS table, ANL table, RTS (Request To Send), CTS (Clear To Send), DATA, etc.), the ID of the local station, data (Including data at higher layers). Further, as shown in FIG. 6, the AS table stored in the database memory 154 stores azimuth and signal strength level information for each adjacent node radio station in the service area of the own station. 13 and the packet transmission / reception control process of FIG. Further, as shown in FIG. 9, the GLS table stored in the database memory 154 includes, for each node radio station in the ad hoc radio network, the adjacent node radio station, its azimuth angle data, and the number of updates (version) of the data. Are created and updated by packet transmission / reception control processing described later.
[0045]
Next, the MAC communication protocol used in this embodiment will be described below. In the wireless communication network according to the present embodiment, it is assumed that a set of wireless stations 1 that perform wireless communication with each other move around in a two-dimensional enclosed space and share a common wireless communication channel. Each radio station 1 includes a variable beam antenna 101 having the above-described four radiation patterns, for example, an electronically controlled waveguide array antenna device. Each radio station 1 can perform either transmission or reception at a time, but a single radio station 1 cannot perform a plurality of transmissions / receptions.
[0046]
According to the IEEE802.11 MAC protocol standard, high-reliability data communication is guaranteed using the RTS / CTS / DATA / ACK access control method. However, in the method of this embodiment, the access control method is used as a base. As will be described later in detail with reference to FIGS. 10 and 11, a packet signal including an ANL table, a GLS table, and other data is transmitted / received using a control signal using a tone signal + packet signal. Therefore, data communication is performed periodically between the generation and update phases of the ANL table and the generation and update phases of the GLS table. In addition, a training sequence is added to each frame to enable control of the beam and null by the transmission / reception antenna and transition to the adaptive control mode.
[0047]
Next, details of a wireless communication system for an ad hoc wireless network using the wireless station 1 of FIG. 1 will be described below.
[0048]
As described above, as the variable beam antenna 101 shown in FIG. 1, an electronically controlled waveguide array antenna device is preferably used. A typical adaptive array antenna is usually a digital beamforming antenna. In contrast, electronically controlled waveguide array antenna devices rely on high-frequency beamforming, which significantly reduces circuit complexity. An electronically controlled waveguide array antenna device includes a central excitation element connected to a radio transmitter / receiver of a transmission source radio station, and a plurality of non-circular elements provided on a predetermined radius around the excitation element and surrounded by a circular shape. And an excitation element (typically 4 to 6). A variable reactance element is connected to each non-excitation element, and by adjusting its reactance value, each non-excitation element forms the radiation pattern of the array antenna device in a different shape. The electronically controlled waveguide array antenna device is characterized by the control of the beam direction, the formation of multiple beams with the same frequency, the control of a movable beam (capable of 360-degree sweep scanning) and null steering. The advantage of using the electronically controlled waveguide array antenna device as a generalized switching beam antenna is that it can be continuously tracked with a small number of antenna elements and can have a variable number of beam patterns. It is in. Since the electronically controlled waveguide array antenna device becomes a low-cost, low-power small antenna, reduction of power consumption of the node radio station as a user terminal is promoted, and all advantages of the switching beam antenna are derived. Is possible.
[0049]
In the present embodiment, each node wireless station 1 includes a variable beam antenna 101 that is a directional antenna, and an ad hoc wireless network including a plurality of N node wireless stations 1 distributed in a two-dimensional area space A is assumed. Each node radio station 1-n (hereinafter, the sign of the node radio station 1-n is abbreviated as n, where nεN) has a unique identifier of the node radio station n. Each node radio station n has a transmission reach distance (distance (radius) of a range where a transmission signal reaches) R and a beam width β assumed to be the same for all the node radio stations n in this embodiment. Have. If the beam width β is set to 360 degrees, it operates in an omnidirectional mode. In the simulation environment of the present inventors, it is assumed that the beam width β related to directivity is 45 degrees. A random way point model was set as the mobile node radio station model. That is, the node radio station n randomly selects a point of the destination radio station and designates the designated v_maxThru v_minIt moves to the destination point at a constant speed v uniformly selected from the set of speeds in between. When the destination is reached, the node radio station stops for a certain time called the pause time, and then the process is repeated.
[0050]
Next, some important definitions in the wireless communication system according to the present embodiment will be described below.
[0051]
Definition 1: When a node radio station n forms a transmission beam with an azimuth angle α and a beam width β at a transmission reach distance R, its effective range from the node radio station n at the azimuth angle α is as shown in FIG. Transmission zone B of node radio station n_nIt is defined as (α, β, R). This is because the node radio station m (mεN)_nIf it is within (α, β, R) and the node wireless station m is in the reception mode, the node wireless station n transmits the message wirelessly at the transmission azimuth angle α, the beam width β, and the transmission reachable distance R with respect to the node wireless station n. This means that each time a signal is transmitted, the radio signal of the message is always received by the node radio station m. Node radio station m is in transmission zone B_nWhen moving outside (α, β, R), the connectivity between the node radio station n and the node radio station m is lost. In the present embodiment, the transmission beam width β and the transmission arrival distance R are constant._n(Α, β, R) in transmission zone B_nCalled (α).
[0052]
Definition 2: Node radio station n (Gⁿ) (N (GⁿAn adjacent node radio station of) ∈N) is defined as a set of node radio stations within the omnidirectional transmission reach R of the node radio station n.
[0053]
Definition 3: GⁿA subset of Gⁿ _α(Gⁿ _α∈Gⁿ) Is defined as an adjacent node radio station in the direction of node radio station n. Where Gⁿ _αThe node radio stations in the transmission zone B_nIt exists in (α).
[0054]
Definition 4: An ANL table, [ANL (t)] is a set of node radio stations in an ad hoc radio network that is actively participating (participating) in an arbitrary communication process at time t.
[0055]
Definition 5: Transmission zone B_nThe active directivity adjacent node radio station of node radio station n in (α), [ActGⁿ _α(T)] is a transmission zone B that is actively participating (involved) in any communication process at time t_nIt is a set of node radio stations in (α) (that is, it belongs to ANL (t) at time t). Therefore, it can be described as:
[0056]
[Expression 1]
ActGⁿ _α(T) = Gⁿ _α(T) ∩ANL (t)
[0057]
Definition 6: Node wireless station n in path P_iCorrelation coefficient of [ηⁿⁱ(P)] is the transmission zone Bn_i(Α (n_i→ n_jNode radio station n))_iDefined as the number of active directional neighbor node radio stations. Where node radio station n_jIs the node radio station n in the path P_iNode radio station in the next hop of α (n_i→ n_j) Is a node radio station n_jTo communicate with the node radio station n_iNode wireless station n_jIt is a transmission zone formed toward Therefore, it can be described as:
[0058]
[Expression 2]
ηⁿⁱ(P) = (| ActGⁿⁱ _{α (ni → nj)}(T) |)
[0059]
Definition 7: The correlation coefficient η, [η (P)] of the path P is defined as the sum of the correlation coefficients of all the node radio stations in the path P as in the following equation.
[0060]
[Equation 3]

[0061]
Here, the correlation coefficient η (P) is used to measure the degree of coupling between routes (see, for example, Non-Patent Documents 11 and 16). That is, the correlation coefficient θ (P) represents the degree of coupling with respect to radio wave interference with other communicable node radio stations that are not included in the path P itself.
[0062]
According to definitions 6 and 7, in the ad hoc wireless network, a node wireless station on one bus P among a plurality of searched paths is assumed to be a communicable node wireless station, and a GLS table is set. The data of the ANL table for the node radio stations included in the one path P is set in a communicable state, and based on the GLS table, the source node radio station other than the source node radio station and the destination node radio station The number of other communicable node radio stations existing in the transmission zone of the service area of each node radio station on the one path P is calculated for the communicable path P ′ between the destination node radio station and the destination node radio station Then, by adding the calculated number of communicable node radio stations for each path P, source nodes other than the source node radio station and the destination node radio station It is calculating correlation coefficient of the one path P for the communication possible paths P 'between the lines station and the destination node radio station eta (P).
[0063]
When the correlation coefficient η (P) = 0 of the path P, the path P is said to be “inter-zone uncoupled” with all other “active paths”. Here, an active path is a path that participates (involves) in the communication process at that time. Otherwise, the path P is related to other active paths by a correlation coefficient η.
[0064]
As an example, referring again to FIG. 26 showing prior art multipath routing, first, the source radio station S₁Is the node radio station N₁And N₂Destination radio station D via₁Communicating with. Therefore, the ANL table, ANL (t) has the path {S₁, N₁, N₂, D₁}including. Here, the source radio station S₂Is the destination radio station D₂Wish to communicate with the path P = {S₂, N₅, N₆, D₂} Is selected.
[0065]
First, the case where an omni antenna is used in FIG. 26 will be discussed below. Source radio station S₁And node radio station N₁Are both source radio stations S₂Within the transmission zone (360 degrees in this case). Therefore,
[Expression 4]
η^S2(P) = 2
It is. Source radio station S₁And node radio station N₁Is the node radio station N₅In the omnidirectional transmission zone of
[Equation 5]
η^N5(P) = 2
It is. Similarly,
[Formula 6]
η^N6(P) = 2, and thus when using an omni antenna, the correlation coefficient η (P) of the path P is
[Expression 7]
η (P) = 6
It becomes.
[0066]
On the other hand, when a directional antenna is used, as shown in FIG.₂, Node radio station N₅And node radio station N₆The transmission zone formed by does not include any node radio stations from the ANL table, ANL (t). Therefore, if a directional antenna is used, the correlation coefficient η (P) of the path P is
[Equation 8]
η (P) = 0
It becomes.
[0067]
It has been proved that the average end-to-end delay time increases in any path if the correlation coefficient η (P) of the path P increases (see, for example, Non-Patent Document 11). This is because two paths with larger correlation coefficients have more opportunities to interfere with each other's transmissions due to the broadcast characteristics of radio propagation. From this embodiment, it can be concluded that efficient routing in an ad hoc wireless network is greatly influenced by a correlation coefficient between a plurality of node wireless stations.
[0068]
However, it is difficult to obtain a low correlation coefficient route using an omni antenna. As is clear from FIG. 26, it is possible to decouple a plurality of routes by using a directional antenna, and thereby, a route having a correlation coefficient much lower than that in the case of using an omni antenna. Can be earned.
[0069]
Next, “network recognition” in each node radio station 1 in the radio communication system according to the present embodiment will be described in detail below.
[0070]
In the present embodiment, a radio communication system is proposed that has a mechanism in which each node radio station not only recognizes adjacent node radio stations but also recognizes the entire ad hoc radio network. This “network recognition” is useful when implementing a positive (actively active) routing method described later. Each node radio station n in the ad hoc radio network has five tables indicating the following network status information, and these tables are stored in the database memory 154 in each radio station in FIG. Details of these five tables will be described below.
[0071]
(1) AS table: As shown in FIG. 6, the AS table stores information on the azimuth angle and the signal strength level for each adjacent node radio station in the service area of its own station. 13 and the packet transmission / reception control process of FIG.
[0072]
(2) NLS table (NLS table in node radio station n is described as NLSTn): Each node radio station n periodically records its neighbor information in order to track the direction of its neighbor node radio station. Collect "to form an NLS table. The NLS table (NLSTn (t)) of the node radio station n at time t is the adjacent node radio station m (mεGⁿ) For each radio link maximum signal strength SIGNAL detected by node radio station n in a particular direction.^θ _{n, m}(T) can be specified. Therefore, the maximum signal strength SIGNAL^θ _{n, m}(T) is the maximum signal intensity at the node radio station n received from the adjacent node radio station m at the azimuth angle θ with respect to the node radio station n and detected at any time by the node radio station n. The NLS table of node radio station n helps in determining the most likely communication direction with any adjacent node radio station. Here, each node radio station selects an azimuth angle that maximizes the SINR value for each adjacent node radio station acquired in advance, and uses this SINR value as the affinity with the adjacent node radio station. Each node radio station extracts the azimuth angle and affinity value for each adjacent node radio station, and generates and updates the NLS table with the current date and time as the update date and time. In the NLS table, as shown in FIG. 7, an azimuth corresponding to the maximum SINR value, the affinity that is the maximum SINR value, and the update date and time are stored for each adjacent node radio station. .
[0073]
(3) NANL table (NANL table in node radio station n is referred to as NANL table n): The NANL table in node radio station n is indicated by the communication active state (communication state flag value) of the adjacent node. Flag value = 1 indicates that communication is active and communication is possible, and flag value = 0 indicates that communication is inactive and communication is not possible. In other words, even if any adjacent node of the node radio station n is not actively participating (participating) or participating in the communication process, the node radio station n has the NANL table, [NANL table at time t._nThe information is recorded in a list table called (t)]. This is useful for each node radio station to detect (recognize) communication with adjacent node radio stations. In the NANL table, only the data of the information related to the adjacent node radio station among the data of the ANL table of FIG. 8 is stored.
[0074]
(4) ANL table (ANL table in node radio station n is described as ANLn): The ANL table includes the detection result of node radio station n regarding the communication activity in the ad hoc radio network. This is a list for node radio station n including all active node radio stations in the ad hoc radio network that node radio station n could detect at time t. As shown in FIG. 8 as an example, the ANL table stores the flag value of the above-described communication state and the number of data updates (indicating version) for each node radio station. The ANL table will be described in detail later with reference to FIG.
[0075]
(5) GLS table (GLS table in the node radio station n is described as GLSn): The GLS table includes topology information in the ad hoc radio network that can be detected by the node radio station n at time t. Specifically, as shown in FIG. 9, the GLS table stores, for each node radio station, adjacent node radio stations, their azimuth angle data, and the number of times the data is updated (indicating version). To do. Details of the GLS table will be described later with reference to FIG.
[0076]
In the present embodiment, each node radio station has a different predetermined packet signal including the ANL table and a packet signal including the GLS table (hereinafter, the former is referred to as an ANL packet signal and the latter is referred to as a GLS packet signal). The ANL table and the GLS table are generated and updated by broadcasting at periodic intervals. This will be described below.
[0077]
Each node radio station n stores its ANL table at a certain periodic interval (ANL packet signal transmission cycle) T._ABroadcast on. Here, the broadcast of the ANL packet signal serves the following two purposes. That is, when the node radio station n receives ANL tables from all its adjacent node radio stations (for example, i, j, and k), the following processing is executed.
(1) The node wireless station n forms an NLS table (NLSTn) so as to include the adjacent node wireless stations i, j, and k as its adjacent node wireless stations, and any adjacent one of the adjacent node wireless stations. Record as the best possible communication direction with the node radio station.
(2) The node radio station n also records the communication status flag value (communication activity status) of the node radio station i, records the other adjacent nodes in the same manner, and the NANL table of the node radio station n itself And then the corresponding ANL table is updated.
[0078]
In addition, each node radio station n transmits its GLS packet signal at a certain periodic interval (GLS packet signal transmission cycle) T._GBroadcast on. Here, when a certain node radio station n receives a GLS packet signal from its adjacent node radio station, it updates its own GLS table, which will be described in detail later.
[0079]
In the simulation of the wireless communication system according to the present embodiment, when each node wireless station does not move (when fixed), the transmission period T of the ANL packet signal_A= 2 seconds and GLS packet signal transmission period T_G= 10 seconds. In addition, when each node radio station moves (when moving), the transmission period T of the ANL packet signal_A= 1 second and GLS packet signal transmission period T_G= 5 seconds. There are the following reasons for broadcasting two packet signals (ANL packet signal and GLS packet signal) in two types of transmission cycles. That is, the ANL table captures the flag value (communication activity state) of the communication status of each node radio station, and as soon as communication is started, the data in the ANL table in a certain set of node radio stations has an effect. Receive (the data is updated with the latest data). Therefore, the ANL packet signal needs to be propagated faster than the GLS packet signal. Further, the ANL packet signal operates as a beacon signal. Therefore, as the ANL packet signal propagates faster, not only can the active (communicatable) node radio station important information penetrate faster, but also accurate adjacent node radio station information (azimuth and signal strength level) Is possible to get. Since it is the concept of fisheye that is going to be implemented in the present embodiment (see, for example, Non-Patent Document 20), accurate information on the adjacent node radio station is required more quickly.
[0080]
On the other hand, the GLS table is global information regarding the connectivity of all the node radio stations, and reflects the change in the topology related to the physical movement of the node radio stations (much slower than signal propagation). . Furthermore, the GLS table at any node radio station need not be very accurate. This is why a GLS packet signal with a large amount of data has a slow propagation, and an ANL packet signal with a smaller amount of data than the GLS table has a fast propagation.
[0081]
FIG. 10 is a timing chart showing tone signal and packet signal transmission / reception processing used in the ad hoc wireless network of FIG.
[0082]
For example, in the MAC protocol method according to the prior art disclosed in Non-Patent Document 27, it is necessary to periodically transmit three packet signals: a beacon signal, an ANL packet signal, and a GLS packet signal. There was a problem that throughput was lowered for a long time. In order to solve this problem, in the MAC protocol according to the present embodiment, as shown in FIG. 10, the radio station on the transmission side follows the tone signal that is an unmodulated carrier, followed by the ANL table, the GLS table, and the RTS. Alternatively, CTS or a packet signal including data to be transmitted is transmitted. On the other hand, after receiving the tone signal, the receiving radio station decodes the packet signal, takes in the decoded data, and executes data processing. Note that transmission of an ANL packet signal or GLS packet signal including an ANL table or a GLS table is periodically executed as described above.
[0083]
FIG. 11 is a timing chart showing the types of radiation patterns and radio communication protocols at each radio station used in the ad hoc radio network of FIG. In FIG. 11, the usage example of the antenna mode of the four-way handshake which concerns on this embodiment is shown. Although adaptive control patterns can track moving node radio stations, beams and nulls cannot be formed unless a packet signal is received. Therefore, the omni pattern is used in the transmission of the tone signal + RTS and the transmission of the tone signal + CTS from the transmitting node radio station. On the other hand, in the reception of the tone signal + RTS and the reception of the tone signal + CTS in the receiving node radio station, the sector pattern directed to the azimuth angle based on the AS table is used after the rotating sector pattern is used at the start. Is used to execute preparation processing for adaptive control during communication using the sector pattern, and finally shift to a radiation pattern for adaptive control.
[0084]
Next, a method for forming the NLS table and the NANL table will be described below.
[0085]
For example, when an arbitrary node radio station n receives an ANL packet signal from an arbitrary adjacent node radio station m, for example, an NLS table (NLST) of the node radio station n is received._n) Are sequentially formed while incrementing the data. Since the node radio station n receives the ANL packet with the variable beam antenna 101 set to a certain azimuth angle, the adjacent node radio station m recognized by the node radio station n in the specific direction Azimuth that may be the best communication direction for the signal, and signal strength SIGNAL with maximum inter-node coupling on the radio link^θ _{n, m}(T) is recognized. In the present embodiment, a symmetric link is assumed between two adjacent node radio stations.
[0086]
For example, an arbitrary node radio station n first forms its own NANL table (NANLn) based on its own communication state flag value (activity state). Node radio station n sets itself as an “active node radio station” (communication state flag value = 1) whenever it needs to issue an RTS indicating that it wishes to communicate. If the RTS is not issued during the predetermined threshold period, it sets itself as an “inactive (incommunicable) node radio station” (communication state flag value = 0). For example, when the node radio station n receives an RTS from the adjacent node radio station m, for example, the node radio station m is always set as an active node radio station (communication state flag value = 1) in its own NLS table (NLSTn). Set. When the node radio station m deactivates itself (when communication is disabled), this information reaches the node radio station n from the node radio station m by periodic broadcast of ANL packets. When receiving this, the node radio station n sets the node radio station m as inactive (flag value of communication state = 0) in its own NANL table (NANLn).
[0087]
Next, a method for forming an ANL table will be described below.
[0088]
Each node radio station periodically broadcasts an ANL packet signal that includes an ANL table that includes its active node list that includes its perception of communication activity in the ad hoc radio network. When periodic ANL packets are received from a plurality of different node radio stations, each node radio station incorporates this information to form a revised ANL table, and waits for a periodic interval to update the ANL packet signal. Broadcast to the adjacent node radio station.
[0089]
In each node radio station, the ANL packet is first updated by the NANL table of the node radio station. Therefore, in the initial state, at the start of the ad hoc wireless network, all the node wireless stations are merely aware of the activity state (communication possible state) of their adjacent node wireless stations, The other node radio stations are in a state of “knowing nothing”. Each node radio station periodically updates its own ANL table and broadcasts the updated ANL packet signal to its neighbor node radio station (s). By sending this periodic update message (ANL packet signal) regarding the neighboring node radio stations from the neighboring node radio stations, each node radio station gradually becomes active with respect to other node radio stations and their neighboring node radio stations. Information (information included in the ANL table) is acquired. Thus, each node radio station can update its own ANL table based on the update message (ANL packet signal) received from the other node radio stations.
[0090]
A major concept in the penetration process of network state information into a moving node radio station is that the information to be transmitted must be recognized to some extent accurately. Since propagation of updates from a plurality of different node radio stations is asynchronous, it is important to introduce the concept of information update count (version) (for example, see Non-Patent Documents 21 and 24). For example, two ANL packets A₁And A₂Are assumed to reach the node radio station n while transmitting information on the node radio station m that is separated from the node radio station n by a plurality of hops. In order to update the information regarding the node radio station m in the node radio station n, which of the latest information regarding the node radio station m is transmitted, ie, “ANL packet A₁Or ANL packet A₂Is it? There must be a mechanism to discover
[0091]
In order to do this, we have the same concept of a recency token (having update count information indicating the version that is the most recent indicator; see, for example, Non-Patent Document 24) and increment this appropriately. The mechanism to use is used. If two update messages (ANL packets) have a data set related to the same node radio station, for example, the node radio station n, a higher freshness value (update count) of the node radio station n is transmitted. Update messages (ANL packets) that have newer information about them.
[0092]
In FIG. 8 showing an example of the ANL table in the node radio station n, R_iIs a node radio station n in an ad hoc radio network composed of N node radio stations_iIs the number of updates, and S_iIndicates a flag value of a communication state indicating a corresponding activity state of each node wireless station, and is either 0 (inactive: communication is impossible) or 1 (active: communication is possible).
[0093]
Next, a method for forming the GLS table will be described below.
[0094]
Each node radio station holds a GLS table for capturing network connection information. In each node radio station, the GLS table is first updated with the NLS table of that node radio station. Thus, in the initial state, at the start of an ad hoc radio network, all node radio stations are only aware of their neighbor node radio stations, and with respect to other node radio stations in the radio communication system. "I don't know anything". Each node radio station periodically updates its GLS table, and broadcasts the updated GLS packet signal to its adjacent node radio station at a predetermined transmission cycle. With this periodic update message (GLS packet signal) for each of those neighboring node radio stations from that neighboring node radio station, the node radio stations gradually become GLS tables for other node radio stations and their respective neighboring node radio stations. Get information. In this way, each node radio station updates its own GLS table on the basis of an update message (GLS packet signal) received from another node radio station.
[0095]
Here, it should be noted that if the periodicity of the update is controlled, the update traffic in the ad hoc wireless network and the accuracy of the network state information stored in each node wireless station can be controlled. For example, if the update message (GLS packet signal) is propagated too frequently, the control traffic increases, but the accuracy of the network state information stored in each node radio station also increases. Here, the ad hoc wireless network is never flooded with update propagation. The maximum number of packets in the ad hoc wireless network at any time is always less than the number of node wireless stations in the ad hoc wireless network. In this case as well, it is necessary to implement the concept of up-to-dateness (information retention of the number of updates) as described above for propagation of ANL packets. This is because if two GLS packet signal update messages have a data set relating to the same node radio station, for example, node radio station n, the higher freshness token value of node radio station n (greater update This means that the update message (GLS packet signal) carrying the number of times has newer information about it.
[0096]
When the node radio station n receives a GLS packet signal from another node radio station, the node radio station n updates its own GLS table. In order to do this, the freshness token values (number of updates) of all nodes stored in the GLS table of the node radio station n, and all of the nodes stored in the GLS packet signal for the recently arrived update The freshness token value (update count) is compared. If it is known that the freshness token (update count) of an arbitrary node radio station in the GLS table of the node radio station n, for example, the node radio station X is less than that in the GLS packet signal to be updated, the update of the GLS packet signal to be updated It is clear that the latest information regarding the node radio station X is transmitted. Accordingly, the entire information regarding the node wireless station X in the GLS table of the node wireless station n is overwritten by the information of the node wireless station X in the received updated GLS packet signal. This step is performed asynchronously for all updated GLS packet signals with their arrival at the host node radio station n. This step facilitates the acquisition by the node radio station n of all the latest information that can be collected from the updated GLS packet signal.
[0097]
It has to be noted here that this mechanism does not guarantee that each node radio station recognizes the exact state of the ad hoc radio network. It is only "recognition" that helps each node radio station to elucidate the approximate state of the ad hoc radio network. This promotes the maintenance of more accurate status information in neighboring node radio stations close to the node radio station, but the accuracy of the details of the network information gradually decreases as the distance increases. It is similar to “Approach” (for example, see Non-Patent Document 20).
[0098]
In FIG. 9 showing the structure of the GLS table in an arbitrary node radio station n, R_iIs a node radio station n in an ad hoc radio network consisting of N nodes_iIs the freshness token value (number of updates) of <n_j, Α (n_i, N_j)> Indicates the node radio station n_jNode wireless station n_iIt is shown that it is an adjacent node radio station. Where α (n_i, N_j) Is a node radio station n_iNode wireless station n_jThe transmission beam azimuth angle α that can be communicated with each other best is shown.
[0099]
Next, the tracking of the position of each node radio station and the MAC protocol will be described below.
[0100]
Usually, in an ad hoc wireless network, all node wireless stations are equipped with an omni antenna. However, ad hoc wireless networks using omni antennas use RTS / CTS-based floor reservation methods that waste most of the network capacity in maintaining wireless media over a wide area. Therefore, many node radio stations adjacent to the transmitter and the receiver are forced to exist after waiting for the data communication between the transmitter and the receiver to end. To alleviate this problem, researchers in the art have proposed the use of directional antennas that significantly reduce radio interference, thereby increasing the use of the radio medium and inevitably increasing network throughput. (For example, refer nonpatent literature 2 thru | or 5.).
[0101]
In order to fully utilize the capabilities of the directional antenna, all the neighboring node radio stations of the source radio station and the destination radio station can start a new communication in the other direction, and the source radio station and the destination radio station The direction of communication must be known so that interference with data communication currently being performed with a radio station can be prevented. Thus, it is essential for each node radio station to have a mechanism for tracking the direction of its neighboring node radio stations.
[0102]
However, this direction tracking mechanism in wireless ad hoc wireless networks using directional antennas is a serious problem because it incurs significant control overhead.
[0103]
For example, in Non-Patent Document 1, direction tracking is performed by using a set of tone signals and maintaining extensive network state information at each node radio station in an ad hoc radio network. Here, in the case of a dynamic state in which each node radio station moves, this is impractical. Further, in Non-Patent Document 4, the proposed MAC protocol does not need to recognize location information, and the source radio station and the destination radio station change their directions on an on-demand basis during an omnidirectional RTS-CTS exchange. To identify. All neighboring node radio stations of the source radio station s and the destination radio station d listening to this RTS-CTS dialogue communication are assumed to use this information to prevent interference with the ongoing data transmission. Has been. However, due to the omnidirectional transmission of RTS and CTS packets, this protocol does not offer the benefit of spatial reuse of the radio channel.
[0104]
Furthermore, for example, Non-Patent Document 2 proposes the use of GPS (Global Positioning System) for tracking the position of each node radio station, but the precise mechanism of information exchange and the resulting overhead are discussed. Absent. Further, for example, in Non-Patent Document 3, it is assumed that the node radio station recognizes the transmission direction in order to directionally access the adjacent node radio station, but the position tracking mechanism is not shown. . Furthermore, both the methods of Non-Patent Documents 2 and 3 require the addition of hardware in each user terminal. In Non-Patent Documents 5 and 10, which are the results of previous research by one of the inventors, the MAC in which each node wireless station dynamically maintains neighbor information through maintenance of an azimuth pair SINR table (AS table). A protocol has been proposed. In this method, in order to create an AS table, each node radio station sequentially transmits directional beacon signals in 30-degree intervals in all directions in the form of directional broadcasts, covering the entire 360-degree space. To do. Here, with this prior art method, the overhead due to the control packet is extremely high.
[0105]
The MAC protocol invented by the present inventors in this embodiment is basically a “director-based, rotating sector-based directional MAC protocol by receiver”, which also functions as a position tracking mechanism. In this case, each node radio station stands by in an omnidirectional detection mode while idling. This enters “Rotating Sector Receive Mode” whenever it detects any signal that exceeds the threshold. In the “rotating sector reception mode”, the node radio station n sequentially rotates its directional antennas at 45 degree intervals in all directions to cover the entire 360 degree space in the form of sequential directional reception in each direction, Detect the signal received in the direction. After one revolution, this determines what may be the best signal reception direction by the maximum received signal strength. The beam is then set in that direction and a signal is received.
[0106]
Here, in order to enable the receiver to decode the received signal, each control packet has a time to rotate the receiver's rotating receive beam by 360 degrees, the duration of the tone signal (in our simulation). In this case, it is transmitted with a preceding tone signal having a duration slightly shorter than 200 microseconds). The purpose of transmitting this tone signal prior to any control packet signal is to allow the receiver to track what may be the best receiving direction of the signal. When this sets the beam in that direction, the purpose of the tone signal is fulfilled and a control packet signal is subsequently transmitted.
[0107]
In the framework according to this embodiment proposed by the present inventors, the following four types of broadcast (omnidirectional) control packets are used for medium access control.
(1) Packet signal including an ANL table (ANL packet signal),
(2) a packet signal (GLS packet signal) including a GLS table;
(3) a packet signal including an RTS (transmission request) signal, and
(4) A packet signal including a CTS (transmittable) signal.
[0108]
In addition, the control packet ACK signal is also a directivity control packet. The packet signal including data is transmitted after the direction is determined after the RTS / CTS handshake is performed. As described above, the ANL packet signal and the GLS packet signal are periodic signals, and are transmitted from each node radio station at a predetermined transmission cycle (transmission interval). At each periodic interval, for example, each node radio station m broadcasts an ANL packet to its adjacent node radio station if its radio medium (wireless channel) is idle. As pointed out above, ANL packets are transmitted with a preceding tone signal that helps the receiver to detect what may be the best receiving direction of the signal. Each receiver then sets its beam in that direction and receives and decodes the packet (see FIG. 10).
[0109]
Further, the node radio station n checks the radio medium (radio channel) every time it wishes to start data communication with the node radio station j, for example, and issues an omnidirectional RTS signal if the radio medium is free. Then send. Upon receiving the RTS signal, the destination radio station j issues and transmits an omnidirectional CTS signal. The purpose of RTS / CTS in this case is not to prohibit transmission or reception by the adjacent node radio stations of the node radio stations n and j (as in the case of using an omni antenna). This is to inform the adjacent node radio station that the adjacent node radio station j is about to receive a data packet from the node radio station n. This also specifies the approximate duration of the communication. All adjacent nodes of the node radio stations n and j set their directional network allocation vectors (Directional Network Allocation Vectors (DNAV)) in the direction of the node radio stations n and j, so that the node radio stations n and j Keep track of communications between and. Therefore, the node radio station existing adjacent to the node radio stations n and j can communicate in the other direction “without interfering with the communication performed between the node radio station n and the node radio station j”. Can start. The source radio station and the destination radio station wait for an acknowledgment signal and a data packet signal, respectively, in a directional reception mode.
[0110]
FIG. 12 shows a method of controlling the radiation pattern using the above-described DNAV, and the arrangement of each node radio station and the antenna radiation when it is known that another radio communication is already performed. The pattern is illustrated. In FIG. 12, a case where the node radio station X and the node radio station Y in the vicinity of the node radio station S and the node radio station D are communicating with each other while the node radio station X is trying to communicate with the node radio station X as a transmission source in FIG. Shown in The sector beam pattern from the node radio station S to the node radio station D is as shown in FIG. Here, the node radio station X and the node radio station Y have already received the RTS signal and the CTS signal from the node radio station S and the node radio station D, and have registered them.
[0111]
In FIG. 12, θ_xyIs the value of the azimuth angle from the node radio station X to the node radio station Y, the node radio station X and the node radio station Y are based on the AS table and the NLS table, respectively._xsAnd θ_xdValue of node wireless station Y is azimuth angle θ_ysAnd θ_ydYou can know the value of. First, when the transmission of the RTS signal from the node radio station X to the node radio station Y affects the node radio station S or the node radio station D, that is, the azimuth angle θ_xyIs θ_xsAnd θ_xdNode wireless station X cannot transmit, and it is necessary to wait for communication between node wireless station S and node wireless station D to end in an idle state. Otherwise, the node radio station X can transmit an RTS signal. At this time, the antenna radiation pattern of the node radio station X is determined from the preconditions as follows:_xsAnd θ_xdThis is an exclusive sector pattern in which null points are formed in the direction of. Similarly, when the exclusive sector pattern from the node radio station Y to the node radio station X does not capture the node radio station S or the node radio station D, that is, the azimuth angle θ_yxAnd θ_ysAnd θ_ydIf the two do not overlap, the node radio station Y can transmit a CTS signal to the node radio station X. Thereafter, the node radio station X and the node radio station Y transmit / receive a DATA signal and an ACK signal according to the sector beam pattern.
[0112]
Next, the adaptive routing protocol with the maximized inter-zone non-joining shortest path will be described below.
[0113]
Prior art routing protocols in ad hoc wireless networks rely on the use of “omni antennas”. Utilization of the advantages of directional antennas for routing has not been sufficiently studied as described above (for example, see Non-Patent Documents 10 and 19). We provide a routing method that uses "effective load balancing" along with a directional MAC protocol to take advantage of the capabilities of directional antennas for improved media utilization. I think it is necessary to do. We propose a table-driven routing protocol for routing that uses load balancing. We used the correlation coefficient η to measure “maximized interzone decoupling” as a route selection criterion for load balancing. Here, the network awareness at each node radio station is merely “knowledge” about the network status, not the “real” network status, so each intermediate node radio station adaptively corrects the routing decision during routing And change. We have implemented the following routing method for effective load balancing using maximized inter-zone non-bonded routes.
[0114]
Each node radio station in the ad hoc radio network uses the current information about the network state (approximate topology information and ongoing communication information) to “appropriate next hop to reach the specified destination radio station Is calculated. This is a method in which a new path is tried to minimize interference to a node radio station that is already involved in some communication, and the source radio station and destination radio station pair using the following method: Choose the appropriate path between.
[0115]
Step I: Predetermined hop maximum value H_max(In this example, H_max= 6) Find all paths between source and destination radio station pairs with fewer hops H.
Step II: Examine the ANL table, find the node radio station involved in the ongoing communication, and calculate the root correlation coefficient η.
Step III: If the ANL table is empty (ie, there is no ongoing communication in the ad hoc wireless network), any one of the paths available with the minimum hop to the destination wireless station is selected. .
Step IV: If some communications already exist in the ad hoc wireless network so that the ANL table is not empty and is recorded in the ANL table, the following processing is executed.
(A) If the source radio station and the destination radio station are separated by a large number of hops (exceeding two hops), the source radio station may be a path between the source radio station and the destination radio station pair. Search for the path with the smallest η from all of the.
(B) If the source radio station and the destination radio station are separated by only two hops, the source radio station will be able to identify all possible two-hop paths of the pair between the source radio station and the destination radio station. The path with the smallest η is searched from among them.
[0116]
Details of the above processing will be described below with reference to the flowcharts of FIGS. 13 and 14. 13 is a flowchart showing packet transmission / reception control processing executed by the management control unit 105 of the wireless station 1 of FIG. 2. FIG. 14 is a flowchart showing processing of routing and communication processing (step S12) which is a subroutine of FIG. It is.
[0117]
In FIG. 13, first, in step S1, the variable beam antenna 101 is controlled so as to be rotated and scanned every predetermined azimuth angle (for example, 30 degrees) with a rotating sector pattern, and a received signal is received. It is determined whether or not a received signal having a signal strength level equal to or higher than a predetermined threshold value is received. If YES, the process proceeds to step S3. If NO, the process proceeds to step S9. In step S9, it is determined whether there is a packet signal to be transmitted. If YES, the process proceeds to step S10. If NO, the process returns to step S1. In step S10, it is determined whether the packet signal to be transmitted is an ANL packet signal or a GLS packet. If YES, the process proceeds to step S11. If NO, the process proceeds to step S12. Here, the former ANL packet has a predetermined period T_AThe event is generated periodically, and the latter GLS packet has a predetermined period T_GThe event is generated periodically, and YES in this step. In step S11, after transmitting a packet signal including the table in an omni pattern, the process returns to step S1. On the other hand, after executing the routing and communication process which is the subroutine of FIG. 14 in step S12, the process returns to step S1.
[0118]
In step S3, the rotating sector pattern is stopped, and the radiation pattern of the variable beam antenna 101 is set to a sector pattern directed to the predetermined azimuth angle stopped. Next, in step S4, the received signal is received with the adaptive control pattern, the packet information is decoded, the signal strength level of the received signal is measured, the received signal RS is determined in step S5, and steps S6, S7 or S8 are performed as follows. Branch to
[0119]
If the received signal RS = GLS packet signal, the GLS table update process is executed in step S6, and the process returns to step S1. In this GLS table update process, the currently stored GLS table is compared with the GLS table included in the received signal, and the GLS table is updated using only the newer data that has a higher update count. If it is a new node radio station, information about the node radio station is added. In step S5, if the received signal RS is an ANL packet signal, the process proceeds to step S7, performs an AS table update process, and then returns to step S1. In the AS table update process, the AS table is updated based on the ID of the node radio station in the packet signal, the azimuth angle detected when the received signal is received, and the signal strength level. In the same way as the update method of, the currently stored ANL table is compared with the ANL table included in the received signal, and the ANL table is updated using only new data whose version is updated more frequently, When the node wireless station is a new node wireless station, information about the node wireless station is added, and the NANL table is updated based on the updated ANL table. Further, when the received signal RS is other signal in step S5, the other signal reception process is executed in step S8, and then the process returns to step S1.
[0120]
In the control flow of FIG. 13, in steps S1 and S2, when the variable beam antenna 101 is rotationally scanned and a reception signal having a signal intensity level equal to or higher than a predetermined threshold is received, the reception signal is detected. However, the present invention is not limited to this, and the variable beam antenna 101 is rotationally scanned over 360 degrees to receive a received signal having a signal intensity level equal to or higher than a predetermined threshold value, and the maximum received signal is detected. It may be a received signal.
[0121]
Next, in step S21 of FIG. 14 showing the routing and communication processing as a subroutine, first, referring to the GLS table, the hop count H is larger than a predetermined maximum value Hmax between the source radio station and the destination radio station. Find all small paths p1, ..., pn. Then, in step S22, one path pi is regarded as active, the communication state flag value of the ANL table for the node radio station included in the path pi in the GLS table is set to 1, and in step S23, the GLS table is referred to By calculating the correlation coefficient η (pi) of the path pi for the active path between the source radio station and the destination radio station other than the stations S and D, and multiplying by the hop count H of the path pi, A routing criterion index γ (pi) is calculated. In this embodiment, the route selection criterion is to minimize the product of the correlation coefficient η (pi) of a certain path pi and the number of hops H, and the routing criterion index γ (pi), which is this product, is minimized. In this case, it is possible to obtain the shortest path in which the separation between zones is maximized.
[0122]
Next, in step S24, the communication state flag value of the ANL table for the node radio stations included in the path pi in the GLS table is set to 0. In step S25, all combinations of path pairs of the paths p1,. Whether or not the routing reference index γ (pi) has been calculated is determined, and if NO, the process returns to step S22 to select another path and repeat the above-described processing. On the other hand, if YES in step S25, the path pa corresponding to the minimum routing reference index γ (pa) is selected in step S26, and the communication status flag value for the node radio station included in the path pa in the GLS table is set. In step S27, the ANL table including the active node list information for the node radio station included in the path pa is transmitted to other node radio stations using the ANL packet signal including the ANL table, and the ANL table of each node radio station is set. To update. Further, in step S28, the source radio station communicates with the destination radio station via the path pa based on the routing information in the GLS table, and after the communication ends in step S29, the node radio station included in the path pa in the GLS table. The communication state flag value of the ANL table for A is set to 0, and the ANL table including the active node list information for the node radio station included in the path pa is transmitted to other node radio stations using the ANL packet signal including the ANL table. The station ANL table is updated, and then the original main routine is returned.
[0123]
In the processing of FIG. 14 above, when the remote node is far away from the destination wireless station, the path with the minimum correlation coefficient η is selected so that each intermediate node wireless station reduces interference with ongoing communication. Guarantee to do. However, if it finds that the intermediate node radio station is only 2 hops away from the destination radio station, this is lower than the path of the smallest correlation coefficient η with a larger hop count A higher priority is given to the path of the minimum hop having the correlation coefficient η.
[0124]
This mechanism does not guarantee that each node radio station is aware of the exact state of the ad hoc radio network, so the intermediate node radio station will correct its routing decision and take an alternative path to the data packet signal. To the destination wireless station. Here, the node radio station existing closer to the destination radio station obtains more accurate information regarding the destination radio station and the adjacent communication state.
FIG. 15 illustrates this point, that is, the adaptive routing selection process by the intermediate node radio station to reach the destination radio station. In FIG. 15, the source radio station S initially determines an approximate route {S—X—Y—ZD} to the destination radio station D. The dotted line in FIG. 15 indicates the position of the initial destination wireless station D. However, the destination wireless station D changes its position by movement, and the solid line indicates the current position of the destination wireless station D. When the location change information of the destination radio station D reaches the intermediate node radio station Y, the intermediate node radio station Y having more accurate information regarding the destination radio station D immediately passes the path to the destination radio station D. A change can be determined and a better path through P and Q to the destination radio station D can be determined. Thus, the path is adaptively selected and modified depending on the accuracy of the available information without generating a large number of control packets. Since each node radio station has a GLS table and an ANL table, this also improves routing performance.
[0125]
In one case in our simulation, each node radio station n in the path calculates its “best next hop” for reaching the destination radio station. When the calculation is finished, the node radio station n uses this next hop for its specific communication as long as the information can be reached with the same antenna pattern as the node radio station n. In other words, if this next hop is not accessible to node radio station n with the same antenna pattern, or if this next hop is unreachable, node radio station n uses the same route calculation method. To recalculate the next hop to reach the destination radio station.
[0126]
【Example】
Next, a simulation for evaluating the performance of the wireless communication system for the ad hoc wireless network according to the present embodiment and the result thereof will be described below.
[0127]
This simulation was performed using QualNet 3.1 (see, for example, Non-Patent Document 22). Our directional antenna can be oriented discretely with an azimuth angle of 45 degrees and covers a range of 360 degrees. A beam pattern with a maximum gain of 15.6 dBi is used as the sector beam pattern. The Qualnet simulator uses the MAC protocol and routing protocol described above.
[0128]
1000x1000m²Thirty node radio stations are randomly arranged in an area of, and eight node radio stations are selected with a time lag as a source radio station having a constant bit transmission rate (CBR). . Each of these generates a 1024 byte data packet at a rate of 2 to 500 packets / second to a randomly selected destination wireless station. Therefore, not all eight source wireless stations start data communication at the same time. When one source radio station is selected, the next source radio station is selected after 15 seconds. Here, all communication continues until the end of the simulation. The reason for using such a technique is to test the performance of the network with a number of pairs of source and destination radio stations that perform simultaneous data communication with different start times. Table 3 shows the parameter set used.
[0129]
[Table 1]

[0130]
Next, the influence of overhead in the simulation will be described below.
[0131]
Since both GLS packet signals or ANL packet signals are periodic update packets, and their propagation is limited to one-hop broadcasts, the following analysis shows that an ANL table or GLS table is used. Ad hoc wireless networks never flood. In effect, we rely on approximate global network state information and accurate local state information similar to the fisheye concept described in Non-Patent Document 20. Therefore, the intermediate node radio station adaptively changes the routing decision based on the surrounding more accurate local information.
[0132]
Each update packet moves with a time difference of T milliseconds, and it takes t milliseconds for physical movement from one node radio station to another node radio station. Also, the bounded area for ad hoc operation², N is the number of nodes in the bounded area A, and R is the omnidirectional transmission reach of each node. When a node radio station broadcasts an update packet to its adjacent node radio station, the node radio station in the circular transmission zone surrounding the node radio station is in use, but the area of the bounded area Node radio stations in other areas in A can broadcast packet signals. Therefore, in a certain average case, if the topology is uniformly distributed over the area A of the bounded area, a plurality of update packets do not interfere with each other in the area A of the bounded area. The number of zones that can move between stations simultaneously is (A / (πR²))be equivalent to. Here, since the node radio stations are uniformly distributed, the number of node radio stations included in one zone (and necessarily the number P of update packets) is expressed by the following equation. .
[0133]
[Equation 9]

[0134]
In other words, P update packets must move sequentially from one node radio station to another node radio station. Since each update packet moves with a time difference of T milliseconds and requires t milliseconds (travel time), this medium is occupied by [t × P × 100 / T]% of the time for update traffic. The For example, in the case of the GLS table, the bounded area of the operation is 1000 × 1000 m²And an ad hoc radio network including 30 node radio stations and a transmission period T of a GLS packet signal with respect to a time difference T = 2 msec._GIs 5 seconds, the number of node radio stations included in one zone (and necessarily the number of update packets) P is P = 8.48. Thus, this medium only occupies 0.34% of its time by traffic transmitting GLS packet signals. When transmitting an ANL packet signal, its transmission cycle T_AIs 1 second and the required travel t = 1 ms. Thus, this medium only occupies 0.85% of its time by traffic due to the transmission of ANL packet signals. Thus, the media is freed from update packets for 98.8% of that time. In other words, “This media is only 1.19% of its total time that is bound by update traffic, and 98.8% is free to use in the data communication process.” The practical gain of this method is in this respect and is also the only significant motivation to replace prior art link state routing.
[0135]
The inventors performed a simulation for overhead analysis on a qualnet having a static route on a static topology composed of 30 node radio stations arranged at random. The reason for using the static route is to examine the performance depending on the presence or absence of control traffic overhead regardless of the routing method.
[0136]
FIG. 16 is a simulation result of the ad hoc wireless network according to the present embodiment, and is a graph showing the influence of overhead on the average throughput. FIG. 17 is a simulation result of the ad hoc wireless network according to the present embodiment, and the average end -It is a graph which shows the influence of the overhead in a two-end delay time. That is, the simulation results in FIGS. 16 and 17 indicate that the influence of overhead caused by the update packet is not so important. The first experiment in the simulation is performed without any overhead, and the second experiment is that the transmission period of the ANL packet signal = 2 seconds and the transmission period of the GLS packet signal = 10 seconds (T_A= 2, T_G= 10), and the third experiment shows that the transmission period of the ANL packet signal = 1 second and the transmission period of the GLS packet signal = 5 seconds (T_A= 1, T_G= 5) is executed using the overhead.
[0137]
The present inventors used DSR (for example, see Non-Patent Document 23) compliant with IEEE802.11 and its MAC protocol as a benchmark for comparing and evaluating the performance of the present inventors. The evaluations we have made are due to four criteria: “average throughput”, “average end-to-end delay time”, “average number of packet retransmissions due to ACK timeout” and “retransmission limit” Based on the “average number of dropped packets”.
[0138]
First, 20 static snapshots were taken and their performance was compared with DSR in the light of these four criteria. The average values observed by the inventors are shown in FIG. 18 to FIG. 21, and are shown at a CBR packet arrival rate of 2 packets / second to 500 packets / second when the packet size is 1024 bytes. The mechanism according to the present embodiment includes a heading of an electronically controlled waveguide array antenna device (ESP). At high data rates, the average throughput is 500 Kbps, which is five times that of DSR, and the average end-to-end delay time is one second that is 3.5 times less than DSR. The number of packet retransmissions due to the ACK timeout is 175 in the case of DSR, but insignificant in the case of this embodiment. Similarly, the number of average packet drops is much smaller in the case of this embodiment than in DSR.
[0139]
When many source and destination radio stations communicate at a high data rate at one time, the use of media using directional antennas can increase significantly. Here, along with this, choosing a maximized inter-zone non-joining path significantly reduces the competition for gaining access to the shared media between routes, so that it spans all nodes in the ad hoc wireless network. A scenario can be obtained in which the network load is balanced. The combined effect of these two aspects ultimately results in a dramatic improvement in system performance, with increased throughput and end-to-end delay times as shown in FIGS. Reduce.
[0140]
22 to 25, the performance of ESP is evaluated under a low mobility of 5 m / sec at a data rate of 200 packets / sec. As apparent from FIGS. 22 to 25, in order to cope with the movement, it is necessary to infiltrate the network information faster, and accordingly, the transmission period T of the ANL packet signal_AAnd GLS packet signal transmission period T_GAre changed from 2 seconds to 1 second and from 10 seconds to 5 seconds, respectively. Although some degradation in performance is observed due to increased control traffic, it is not significant.
[0141]
As explained above, the use of directional antennas in ad hoc wireless networks can dramatically improve system performance if routing issues are examined with an appropriate directional MAC protocol with load balancing. it can. In this embodiment, the maximized inter-zone non-joining route facilitates reducing inter-route coupling between selected paths, thereby improving end-to-end delay time and throughput. Despite the control overhead caused by the periodic propagation of GLS and ANL packet signals in ad hoc wireless networks, its performance is far superior to traditional reactive routing using omnidirectional MAC protocols. ing. The proposed table driven adaptive routing with maximized inter-zone decoupling according to this embodiment takes advantage of directional antennas and improves system performance. In addition, since the routing method proposed by the present inventors is table driven, the result depends on the penetration of information, but the transmission period T of each packet signal._AAnd T_GThe latest control overhead determined by the value of is considered to correspond well to high mobility.
[0142]
【The invention's effect】
As described above in detail, according to the present invention, in a control method or a wireless communication system for a wireless network that includes a plurality of wireless stations and performs wireless communication between the wireless stations, the plurality of wireless stations can communicate with each other. A first table that stores data indicating whether or not a state is present for each radio station, an adjacent radio station that can perform radio communication from each radio station, and an azimuth angle from each radio station to the adjacent radio station And a second table for storing the data for each wireless station in the storage means of each wireless station, and the first wireless signal including the first table is cycled at a predetermined first transmission cycle. When the second radio signal including the second table is periodically transmitted at a predetermined second transmission period and the transmitted first radio signal is received, Data of the first table stored in the storage means And updates, when receiving the second radio signal the transmission, to update the data of the second table stored in the storage means of its own station. That is, since the first table and the second table are divided and wirelessly transmitted, the overhead can be greatly reduced as compared with the prior art, and the data status of the two tables can be quickly determined. Notification can be made to each wireless station in the wireless network.
[0143]
Here, the first transmission cycle is shorter than the second transmission cycle. As a result, the state of an adjacent radio station in a certain radio station indicated by the first table can be grasped more quickly, while the data of the second table indicated by the second table is required to be very quick. However, it is only necessary to be able to grasp the schematic wireless network topology. Further, since the data in the first table is smaller than the data in the second table, there is little influence on the traffic of the wireless network, and data communication can be executed efficiently.
[0144]
When transmitting, the first radio signal is transmitted following a predetermined tone signal, and the second radio signal is transmitted following the tone signal. That is, it is possible to perform direction estimation based on the data in the first table that requires quick propagation without using a beacon signal of the prior art, and it is possible to grasp the state of each wireless station at high speed.
[0145]
Further, a path having the smallest routing reference index is selected, a radio station included in the selected path in the second table is set as a communicable radio station, and a radio included in the selected path After the data of the first table for the station is transmitted to the other radio station using the first radio signal and the first table of the other radio station is updated, the source radio station uses the second table. Wireless communication is performed with the destination wireless station via the selected path based on the data in the network. Therefore, single path routing in which the load of the entire wireless network is balanced can be realized.
[Brief description of the drawings]
FIG. 1 is a plan layout view of a plurality of radio stations 1-1 to 1-9 constituting an ad hoc radio network according to an embodiment of the present invention.
2 is a block diagram showing an internal configuration of each radio station 1 in FIG. 1. FIG.
3 is a diagram illustrating an example of a sector beam pattern of the variable beam antenna 101 of FIG. 1. FIG.
FIG. 4 is a diagram showing a format of packet data used in the ad hoc wireless network of FIG. 1;
FIG. 5 is a plan view showing a transmission zone Bn (α, β, R) which is a transmission sector radiation pattern from each node radio station used in the description of the radio communication system according to the present embodiment.
6 is a table showing an example of an azimuth angle and signal intensity level table (AS table) stored in the database memory 154 of FIG. 2; FIG.
7 is a diagram showing an example of an adjacent link state table (NLS table) stored in the database memory 154 of FIG. 2; FIG.
8 is a diagram showing an example of an active node radio station table (ANL table) stored in the database memory 154 of FIG.
9 is a diagram showing an example of a global link state table (GLS table) stored in the database memory 154 of FIG. 2. FIG.
10 is a timing chart showing transmission / reception processing of tone signals and packet signals used in the ad hoc wireless network of FIG.
11 is a timing chart showing types of radiation patterns and radio communication protocols at each radio station used in the ad hoc radio network of FIG. 1; FIG.
12 is a plan view of a wireless station showing a method using DNAV used in the ad hoc wireless network of FIG. 1. FIG.
13 is a flowchart showing packet transmission / reception control processing executed by the management control unit 105 of the wireless station 1 of FIG.
FIG. 14 is a flowchart showing processing of routing and communication processing (step S12) which is a subroutine of FIG.
FIG. 15 is a plan view of a node radio station showing adaptive routing selection processing by an intermediate node radio station to reach a destination radio station used in the present embodiment.
FIG. 16 is a graph showing a simulation result of the ad hoc wireless network according to the present embodiment and showing the influence of overhead on the average throughput.
FIG. 17 is a graph showing the effect of overhead on the average end-to-end delay time, which is a simulation result of the ad hoc wireless network according to the present embodiment.
FIG. 18 is a simulation result of the ad hoc wireless network according to the present embodiment, and shows the average throughput in the case of DSR and ESP for different packet arrival rates (number of packets per second) when the packet size is 1024 bytes. It is a graph to show.
FIG. 19 is a simulation result of the ad hoc wireless network according to the present embodiment, in which the average end times in the case of DSR and ESP for different packet arrival rates (number of packets per second) when the packet size is 1024 bytes; It is a graph which shows a two end delay time.
FIG. 20 is a simulation result of the ad hoc wireless network according to the present embodiment, in which average packet retransmission in the case of DSR and ESP for different packet arrival rates (number of packets per second) when the packet size is 1024 bytes; It is a graph which shows the number of transmission.
FIG. 21 is a simulation result of the ad hoc wireless network according to the present embodiment, and average packet drops in the case of DSR and ESP for different packet arrival rates (number of packets per second) when the packet size is 1024 bytes; It is a graph which shows a number.
FIG. 22 is a simulation result of the ad hoc wireless network according to the present embodiment and shows the average throughput when the node wireless station having ESP is fixed and moved when the packet arrival rate is 200 packets / second; It is a graph.
FIG. 23 is a simulation result of the ad hoc wireless network according to the present embodiment, and shows an average end-to-end when the node wireless station having ESP is fixed and moved when the packet arrival rate is 200 packets / second. -It is a graph which shows end delay time.
FIG. 24 is a simulation result of the ad hoc wireless network according to the present embodiment, and shows an ACK signal when the node wireless station having ESP is fixed and moved when the packet arrival rate is 200 packets / second. It is a graph which shows the average packet retransmission number by timeout.
FIG. 25 is a simulation result of the ad hoc wireless network according to the present embodiment, and the retransmission restriction when the node wireless station having ESP is fixed and moved when the packet arrival rate is 200 packets / second; It is a graph which shows the average packet drop number by.
FIG. 26 shows a radio station S using a directional antenna according to the prior art.₁-D₁And radio station S₂-D₂It is a top view which shows two zone non-combining communication between.
[Explanation of symbols]
1, 1-1 to 1-9, 1-i, 1-j, 1-k, 1-l, N₁To N₆... node radio stations,
BS₁, BS₂, BN₁To BN₆… Send zone,
S, S₁, S₂... source radio station,
D, D₁, D₂... destination radio station,
101 ... Variable beam antenna,
102 ... circulator,
103 ... Direction control unit,
104 ... packet transmission / reception unit,
105 ... Traffic monitor section,
106 ... line control unit,
107 ... upper layer processing apparatus,
130: Packet receiver,
131 ... high frequency receiver,
132: demodulator,
133: Receive buffer memory,
140 ... packet transmitter,
141. Transmission timing control unit,
142 ... transmission buffer memory,
143 ... modulator,
144 ... high frequency transmitter,
151... Management control unit,
152 ... Search engine,
153 ... Update engine,
154 ... Database memory,
155 ... Clock circuit,
160. Spread code generator.

Claims (6)

In a control method for a wireless network including a plurality of wireless stations and performing wireless communication between the wireless stations,
A first table storing data indicating whether or not the plurality of radio stations are in a communicable state for each radio station; an adjacent radio station capable of radio communication from each radio station; and each radio station Storing in the storage means of each radio station a second table storing azimuth angle data from the radio station to the adjacent radio station for each radio station;
The first radio signal including the first table is periodically transmitted at a predetermined first transmission cycle, and the second radio signal including the second table is periodically transmitted at a predetermined second transmission cycle. Sending automatically,
When the transmitted first wireless signal is received, the data of the first table stored in the storage means of the own station is updated, and when the transmitted second wireless signal is received, the own station look including the step of updating the data of the second table stored in the storage means,
The above control method is
Based on the second table, a first step of searching all paths between the source radio station and the destination radio station where the hop count H is smaller than a predetermined maximum value Hmax;
In the wireless network, a wireless station on one bus P of the plurality of searched paths is assumed to be a communicable wireless station, and is included in the one path P in the second table. The data of the first table for the wireless station to be transmitted is set in a communicable state, and based on the second table, between the source wireless station and the destination wireless station other than the source wireless station and the destination wireless station The number of other communicable radio stations existing in the transmission zone of the service area of each radio station on the one path P is calculated for the communicable path P ′, and the calculated communication By adding the number of possible radio stations for each of the one path P, the 1 for the path P ′ that can be communicated between the source radio station and the destination radio station other than the source radio station and the destination radio station. Correlation coefficient η of two paths P (P) is calculated, and the calculated correlation coefficient η (P) is multiplied by the number of hops H of the one path P, whereby the multiplication result is obtained as the routing reference index γ (P) of the one path P. A second step of calculating and returning a wireless station included in the one path P in the second table as a wireless station that cannot communicate in the first table;
A third step of calculating a routing criterion index γ (P) for all of the searched paths by repeating the processing of the second step for all of the searched paths;
Select the path corresponding to the smallest routing criterion index γ (P) among the calculated routing criterion indices γ (P) for all the searched paths, and the selected path in the second table. Is set as a communicable radio station, and data of the first table for the radio station included in the selected path is transmitted to another radio station using the first radio signal. After updating the first table of the other radio station, the source radio station wirelessly communicates with the destination radio station via the selected path based on the data in the second table; A control method for a wireless network , further comprising :   The control method for a wireless network according to claim 1, wherein the first transmission period is shorter than the second transmission period.   3. The transmitting step of transmitting the first radio signal following a predetermined tone signal and transmitting the second radio signal following the tone signal. Control method for wireless networks in the world. In a wireless communication system for a wireless network comprising a plurality of wireless stations and performing wireless communication between wireless stations,
A first table storing data indicating whether or not the plurality of radio stations are in a communicable state for each radio station; an adjacent radio station capable of radio communication from each radio station; and each radio station Storage means for storing for each radio station a second table for storing azimuth angle data from the radio station to the adjacent radio station for each radio station;
The first radio signal including the first table is periodically transmitted at a predetermined first transmission cycle, and the second radio signal including the second table is periodically transmitted at a predetermined second transmission cycle. Means for transmitting automatically,
When the transmitted first wireless signal is received, the data of the first table stored in the storage means of the own station is updated, and when the transmitted second wireless signal is received, the own station Bei example and a control means for updating the data of the second table stored in the storage means,
The wireless communication system is
Based on the second table, first control means for searching all paths between the source radio station and the destination radio station where the number of hops H is smaller than a predetermined maximum value Hmax;
In the wireless network, a wireless station on one bus P of the plurality of searched paths is assumed to be a communicable wireless station, and is included in the one path P in the second table. The data of the first table for the wireless station to be transmitted is set in a communicable state, and based on the second table, between the source wireless station and the destination wireless station other than the source wireless station and the destination wireless station The number of other communicable radio stations existing in the transmission zone of the service area of each radio station on the one path P is calculated for the communicable path P ′, and the calculated communication By adding the number of possible radio stations for each of the one path P, the 1 for the path P ′ that can be communicated between the source radio station and the destination radio station other than the source radio station and the destination radio station. Correlation coefficient η of two paths P (P) is calculated, and the calculated correlation coefficient η (P) is multiplied by the number of hops H of the one path P, whereby the multiplication result is obtained as the routing reference index γ (P) of the one path P. A second control unit that calculates and returns a radio station included in the one path P in the second table as a radio station that cannot communicate in the first table;
Third control means for calculating a routing reference index γ (P) for all the searched paths by repeating the processing of the second step for all the searched paths;
Select the path corresponding to the smallest routing criterion index γ (P) among the calculated routing criterion indices γ (P) for all the searched paths, and the selected path in the second table. Is set as a communicable radio station, and data of the first table for the radio station included in the selected path is transmitted to another radio station using the first radio signal. After updating the first table of the other radio station, the source radio station performs radio communication with the destination radio station via the selected path based on the data in the second table. A wireless communication system for a wireless network , further comprising: The wireless communication system for a wireless network according to claim 4, wherein the first transmission period is shorter than the second transmission period. Said transmission means transmits the first radio signals following the predetermined tone signal, subsequent to the tone signal according to claim 4 or 5, wherein the transmitting the second radio signal A wireless communication system for a wireless network.

Images