JP4134683B2 - Automatic steering device - Google Patents

Description

【００２８】
また、目標軌道の接線の傾きの変化量が道路曲率ρで定義したとき、単位時間当たりの目標軌道からの相対的な車両のヨー角変化、すなわち目標軌道に対する相対ヨーレートφ'_rは、
【００２９】
【数２】

【００３０】
であり、また、相対横速度ｙ'_crは、
【００３１】
【数３】

【００３２】
であるから、前記（１）式より横滑り角及び横滑り角速度を消去し、整理するとそれぞれ下記（４）式及び（５）式が得られる。
【００３３】
【数４】

【００３４】
【数５】

【００３５】
また、基準走行ラインに対する車両１の横方向の前方注視点横変位ｙ_srは、前方注視点位置の車両前方の距離をＬ_s[ｍ]とおいて、目標軌道（走行車線）からの相対横変位ｙ_cr 、相対ヨー角φ_r及び道路曲率ρを用いると、下記（６）式として記述できる。
【００３６】
【数６】

【００３７】
一方、前記図２に示す自動操舵装置における操舵系の運動方程式は次のように記述できる。
【００３８】
【数７】

【００３９】
ここで、ζ：車両のトレール長さ、Ｋ_T：操舵アクチュエータのトルク係数、Ｎ_s：操舵アクチュエータから操舵系へのトルク伝達比、Ｃ_s：操舵系の減衰係数そして、前記（４）式及び（５）式と同様に、横滑り角β及び横滑り角速度β'を消去し、整理すると下記（８）式のようになる。
【００４０】
【数８】

【００４１】
これら（４）式〜（６）式及び（８）式を新たに目標として道路曲率を状態変数に加えて行列表現することにより、曲線走行状態における車両モデルは下記（９）式のように記述できる。
【００４２】
【数９】

【００４３】
前記動的システムに対して、ある行列Ｌを用いて
【００４４】
【数１０】

【００４５】
なる演算を行って、状態変数ベクトルｚを算出する。ここで算出する状態変数ベクトルｚと、真の車両状態量ベクトルｘとの偏差をｅとすると、前記（９）式及び（１０）式から下記（１１）式が得られる。
【００４６】
【数１１】

【００４７】
前記遷移行列Ａ−ＬＣのすべての固有値の実部が負になるように行列Ｌを決定すると下記（１２）式のようになる。
【００４８】
【数１２】

【００４９】
これにより、任意の初期値ｅ（ｔ₀）の値によらず、十分に時間の経過後には先に算出された状態変数ベクトルｚと真の車両状態量ベクトルｘとの偏差ベクトルｅは０となり、状態変数ベクトルｚと真の車両状態量ベクトルｘと一致するので、車両の状態量を正しく検出することができる。
ここで、前記（１０）式はオブザーバを示すもので、このようなオブザーバを用いた解法は動的システムの状態推定に良く利用される手法である。
【００５０】
一方、システムへのプロセスノイズｗ及び観測ノイズ（センサーノイズ）ｎを考慮するとシステムの行列表現は下記（１３）式として記述できる。
【００５１】
【数１３】

【００５２】
そして、前述したオブザーバと同様に偏差ベクトルｅを下記（１４）式として得ることができる。
【００５３】
【数１４】

【００５４】
ここで、プロセスノイズｗとセンサーノイズｎとが正規性白色ノイズであると仮定した場合、推定誤差ｅを最小にするように行列Ｌを求めたものがカルマンフィルタとなり、動的システムの状態推定にしばしば利用される。
ところで、プロセスノイズｗには、横風や路面カント（片勾配）を原因とするヨー方向へのモーメント成分や横方向への横力成分のように、厳密には正規性白色ノイズではない持続的な外乱が作用したり、車両パラメータ誤差等のモデル化誤差が存在するため、状態推定値が真の値に収束しない場合がある。このような結果、前記図３に示すレギュレータ１２１だけの単純な状態フィードバック制御では定常的な位置ずれを生じる可能性がある。このような場合、車両が基準走行ラインからそれる方向へに舵制御されることになり、運転者は強い違和感を感じる可能性が高い。
【００５５】
これに対し、本実施の形態におけるコントローラ２０は、次のような構成を備えている。図５は、コントローラ２０を、基本的な機能構成（図３に示した構成）を含む本発明を適用した系全体をブロック図である。
すなわち、図５に示すコントローラ２０は、図３に示したコントローラ２０と同様に、曲線走行状態をモデルとして構成されたオブザーバ１１０と、レギュレータゲインＫ_s（図３では、Ｋ₀）とを含んだ構成になっている。そして、オブザーバ１１０についても、前述したような手法で決定される行列Ａ，Ｂ，Ｃ及びゲインＬを有するとともに、加算器１１１，１１２を含んで構成されている。ここで、オブザーバ１１０中に示す、ｓはラプラス演算子であり、ｚは状態量ｘの推定値、 Ｃｚは観測値ベクトルｙの推定値である。
【００５６】
ここで、本実施の形態におけるコントローラ２０の詳細をさらに図６を用いて説明すると、コントローラ２０は、車両１の前方注視点横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙを得る加算器１２３と、加算器１２３により得た偏差Δｙに基づいて車両１に発生すべき基準走行ライン追従トルクｕ₂を算出する基準走行ライン追従トルク発生部１３０と、安定化制御トルクｕ_sを算出する安定化制御トルク発生部（Ｋ_s）１２４と、基準走行ライン追従トルク発生部１３０が算出した前記基準走行ライン追従トルクｕ₂と安定化制御トルク発生部（Ｋ_s）１２４が算出した安定化制御トルクｕ_sとを加算して目標操舵トルクｕを得る加算器１２２とを含んで構成されている。
【００５７】
基準走行ライン追従トルク発生部１３０は、積算部１４０と、積算部１４０からの出力値にゲインＫ₂を掛ける乗算部１３１とを備えている。
積算部１４０は図７に示すように構成されている。積算部１４０は、前記加算部１２３からの出力値である、前方注視点横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙにサンプリング時間Ｔを乗じ、その値に当該積算部１４０で１サンプリング時間前に得た積算値を加算し、リミット処理部１４１で所定のリミット処理を行って最新の積算値を算出する。
【００５８】
なお、積算部１４０中のｑ^-1はシフト演算子であり、入力された信号を、ｑ^-1で１サンプリング時刻だけ遅れて出力する。また、リミット処理部１４１で用いる積算値の上限値及び下限値については、基準走行ライン追従トルクｕ₂が所定の上限値及び下限値の範囲内にとどまるように決定している。
基準走行ライン追従トルク発生部１３０は、このような積算部１４０により算出した積算値に、乗算部１３１でゲインＫ₂を乗じ、基準走行ライン追従トルクｕ₂を得る。なお、乗算部のゲインＫ₂の値の選定については後で詳述する。
【００５９】
安定化制御トルク発生部（Ｋ_s）１２４は、道路曲率ρによって与えられる基準走行ライン周りの車両の走行状態を安定化させるための安定化制御トルクｕ_sを算出するように構成されている。図８は、その具体的構成を示す。
安定化制御トルク発生部１２４にはオブザーバ１１０が推定する状態量ｚが入力されており、安定化制御トルク発生部１２４は、２種類のレギュレータゲインＫ₀，Ｋ₁それぞれをその状態量ｚに乗じ、それに対応して第１及び第２の安定化制御トルクｕ₀，ｕ₁を算出する。なお、レギュレータゲインＫ₀，Ｋ₁の値の選定については、レギュレータゲインＫ_sの値の選定として後で詳述する。
【００６０】
安定化制御トルク発生部１２４は、基準走行ライン追従トルクｕ₂がリミット処理されている場合とそうでない場合に応じて第１の安定化制御トルクｕ₀と第２の安定化制御トルクｕ₁とを切り換えて、第１の安定化制御トルクｕ₀と第２の安定化制御トルクｕ₁とのいずれか一方を安定化制御トルクｕ_sとして出力する。ここで、第１の安定化制御トルクｕ₀の方が第２の安定化制御トルクｕ₁よりも小さな値になっており、基準走行ライン追従トルクｕ₂のリミット処理作動時には、第１の安定化制御トルクｕ₀を安定化制御トルクｕ_sとして出力する。
【００６１】
このように安定化制御トルク発生部（Ｋ_s）１２４により安定化制御トルクｕ_sが得られ、この安定化制御トルクｕ_sは、道路曲率をρとする基準走行ライン周りの微小なずれを補正する値となる。そして、この安定化制御トルクｕ_sは前記加算器１２２に入力される。加算器１２２では、この安定化制御トルクｕ_sと前記基準走行ライン追従トルク発生部１３０からの基準走行ライン追従トルクｕ₂とを加算して目標操舵トルクｕを得る。
【００６２】
ここで、積算部１３０のゲインＫ₂やレギュレータゲインＫ_s（レギュレータゲインＫ₀，Ｋ₁）の値の決定について説明する。
ゲインＫ₂は、その値を大きくすると基準走行ラインへの追従速度が速くなるといった特性をもつゲインである。このようなことから、ゲインＫ₂は、前方注視点横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙの積分値がわずかに変化した場合であっても、基準走行ライン追従トルクｕ₂に大きな変化を与えてしまう。この結果、例えば前方注視点横変位ｙ_srの検出結果に含まれるノイズ成分に過敏に反応した基準走行ライン追従トルクｕ₂が算出されて、操舵系の動きがぎこちなくなる可能性がある。
【００６３】
また、レギュレータゲインＫ_sに比べてゲインＫ₂があまりに大きすぎると、基準走行ラインを基準に車両の安定化を図る安定化制御トルクｕ_sが相対的に低下してしまう。この結果、前方注視点横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙがそれほど大きくない場合であっても、基準走行ライン追従トルクｕ₂が支配的になり、基準走行ラインに対する行き過ぎ量が大きくなる。これにより、車両が基準走行ライン上を通過する瞬間には基準走行ラインから逸脱する方向への制御トルクが残ってしまい、車両の走行安定性といった面或いは運転者の操舵感といった面から好ましくない状態が発生する可能性がある。
【００６４】
逆に、ゲインＫ₂を小さくしすぎると、走行安定性は相対的に向上するが、追従速度は低下する。この結果、車両がまだ目標軌道に達していないにも拘わらず目標軌道へ向かっていくことに抵抗するような操舵トルクが演算されてしまい、運転者の操舵感といった面から好ましくない状態が発生する可能性がある。
以上のようなことから、ゲインＫ₂及びレギュレータゲインＫ_s（レギュレータＫ₀，Ｋ₁）をそれぞれ任意の値として単独で設定するのではなく、ゲインＫ₂とレギュレータゲインＫ_sとがうまく協調するような組み合わせで決定する。例えば、車両１に対して自動操舵制御を実行した状態での動特性を実験やシミュレーションで充分に確認し、その動特性を考慮しつつ適宜選定するようにする。例えば、ゲインＫ₂の値とレギュレータゲインＫ_sの値との最適な組み合わせは次のような手法により求める。
【００６５】
例えば、図３のような構成においてレギュレータゲインＫ₀を求める手法として、前記（９）式に対応した次の評価関数Ｊを最小にするレギュレータを最適レギュレータにする、といった手法が知られている。
【００６６】
【式１５】

【００６７】
ここで、行列Ｑ，Ｒを調整することにより系の応答を調整することができ、また、下記（１６）式により安定化制御トルクｕ₀を算出できる。
【００６８】
【式１６】

【００６９】
このような手法を適用して、図６に示した装置では、前方注視点横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙの積分値をｈとしたとき、下記（１７）式を得ることができる。
【００７０】
【式１７】

【００７１】
このような拡張系で下記（１８）式に示す評価関数Ｊを最小にする最適レギュレータを演算すればよい。
【００７２】
【式１８】

【００７３】
ここで、行列ｑ_h，Ｑ'，Ｒ'を調整することで系の応答を調整することができ、また、下記（１９）により目標操舵トルクｕを算出することができる。
【００７４】
【式１９】

【００７５】
なお、この場合でも、車両１に対して自動操舵制御を実行した状態での動特性を実験やシミュレーションで充分に確認し、その動特性を考慮しつつ適宜選定する必要がある。しかし、実際には基準走行ラインへの追従速度を行列ｑ_hで調整し、そのときの車両の安定度が不足気味なら行列Ｑ'の要素を徐々に大きい値へ修正していく程度で望ましい応答が得られる場合が多い。
【００７６】
図９は、コントローラ２０の処理手順を示す。コントローラ２０はこの処理を所定のサンプリング・クロックに同期して実行する。
自動操舵制御が実行されると、先ずステップＳ１において、コントローラ２０は、舵角センサ１４からの舵角検出信号δ_f、車速センサ１９からの車速検出信号Ｖ、及びＣＣＤカメラ１７から画像をそれぞれ読み込む。
【００７７】
なお、車速検出信号Ｖは、オブザーバ１１０の行列Ａ，Ｂ，Ｃの要素を決定するのに必要な情報であるため、車速Ｖとして行列Ａ，Ｂ，Ｃに取り込まれることになる。
続いてステップＳ２において、コントローラ２０は、前記ステップＳ１で読み込んだ画像に基づいて、前記前方注視点横変位ｙ_srを検出する。
【００７８】
続いてステップＳ３において、コントローラ２０は、オブザーバ１１０に基づいた演算（前記（１０）式）を実行することで、実際の車両状態量ｘを推定する演算を行い、車両状態推定ベクトルｚを出力する。
続いてステップＳ４において、コントローラ２０は、前記ステップＳ３で得た車両状態量推定ベクトルｚと、２種類の減衰特性の異なるＫ₀，Ｋ₁とに基づいて、２種類の減衰特性の異なる安定化制御トルクｕ₀，ｕ₁を得る（前記（１９式）により得る）。すなわち、
ｕ₀＝−Ｋ₀ｚ，ｕ₁＝−Ｋ₁ｚ
を得る。
【００７９】
続いてステップＳ５において、コントローラ２０は、前記ステップＳ２で検出した前記前方注視点横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｆを積算し、その積算値にゲインＫ₂を乗算し、基準走行ライン追従トルクｕ₂を算出する。
また、このステップＳ５において、コントローラ２０は、その基準走行ライン追従トルクｕ₂に対して所定の上限値又は下限値によるリミット処理を行うとともに、前記横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙの積算値についてもリミット処理を行う。そして、リミット時にリミットフラグをセットする。このようにそれぞれリミット処理することで、偏差積分値の絶対値がそれ以上増加しないようにする。すなわち、基準走行ライン追従トルクｕ₂がリミット処理された時点での前記横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙの積算値が上限値又は下限値になる。
【００８０】
続いてステップＳ６において、コントローラ２０は、前記ステップＳ５で得た基準走行ライン追従トルクｕ₂がリミット処理されているか否かの判定を行う。すなわち、前記リミットフラグがセットされているか否かを判定し、基準走行ライン追従トルクｕ₂がリミット処理されているか否かの判定を行う。ここで、コントローラ２０は、基準走行ライン追従トルクｕ₂がリミット処理されている場合、ステップＳ７に進み、基準走行ライン追従トルクｕ₂がリミット処理されていない場合、ステップＳ８に進む。
【００８１】
ステップＳ７では、コントローラ２０は、第１の安定化制御トルクｕ₀を安定化制御トルクｕ_sとして決定する。一方、コントローラ８では、コントローラ２０は、第２の安定化制御トルクｕ₁を安定化制御トルクｕ_sとして決定する。このようにステップＳ７及びステップＳ８により、コントローラ２０は、基準走行ライン追従トルクｕ₂がリミット処理されている場合とそうでない場合に応じて２種類の安定化制御トルクｕ₀，ｕ₁を切り換えて安定化制御トルクｕ_sを決定している。
【００８２】
続いてステップＳ９において、コントローラ２０は、この安定化制御トルクｕ_sと基準走行ライン追従トルクｕ₂とを加算して、目標操舵トルクｕを得る。ここでは、例えば、制御される操舵トルクが滑らかに推移するような処理を行う。
続いてステップＳ１０において、コントローラ２０は、ステップＳ９で得た目標操舵トルクｕの出力処理としてアクチュエータ駆動処理行う。すなわち、コントローラ２０は、自動操舵機構１０の電動モータ１０ａへの電流を制御してステアリングシャフト４に目標駆動トルクを入力することで、その目標操舵トルクとなるように車両の前輪３Ｌ，３Ｒに舵角を発生させる。ここでの制御として、例えば、自動操舵機構を構成している電動モータ１０ａに印加する電流をＨブリッジ回路を用いたＰＷＭ機構等によって制御し、目標操舵トルクｕを操舵系に入力する。
【００８３】
このステップＳ１０の処理が終了したら、今回の処理を終了する。その後は、所定のサンプリング時刻が経過するまで待機した後に、前記ステップＳ１から処理を再び実行する。
以上のような手順でコントローラ２０は処理を行う。このような処理により、車両は、基準走行ライン追従トルクの作用により曲率のある道路を追従して走行するとともに、安定化制御トルクの作用によりその曲率を有する道路に沿って安定して走行するようになる。
【００８４】
次に効果を説明する。
前述したように、基準走行ライン追従トルクｕ₂に対して所定の上限値又は下限値によるリミット処理を行うとともに、横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙの積算値についてもリミット処理を行うことで（前記ステップＳ５）、基準走行ライン追従トルクｕ₂の過剰な発生を抑え、これにより過剰な目標操舵トルクｕの発生を抑えることができる。
【００８５】
これにより、例えば、運転者の操舵介入が原因でなかなか基準走行ラインに到達しない状態から、運転者が急に操舵介入を中止した場合であっても、基準走行ライン追従トルクｕ₂の残留成分の影響で基準走行ラインを通り過ぎても未だ操舵トルクが入力され続けるような状態を防止できる。また同様に、強めの外乱が急に減少した場合や外乱の向きが急に変化した場合、さらには基準走行ラインの位置を大きく変更した場合等の状況において、余分な車両挙動を発生することなく車両１を走行させることができる。
【００８６】
リミット処理を行うことで以上のような効果がある一方で、リミット処理が作動している場合には、目標軌道が基準走行ライン追従トルクを増加させる方向へどんなに変化しても算出される基準走行ライン追従トルクが一定値になるので、等価的にオフセットトルクのフィードバックゲインが小さくなる。これは、本来は積分値が増大していく過程であるにも拘わらず基準走行ライン追従トルクが一定値で維持されるので、リミット処理が作動している間は、ゲインが小さくなっていくのと等価となるからである。そして、このような場合、制御系全体として減衰が過度に強い状態になってしまう。すなわち、状態フィードバックにより車両の姿勢を現状にとどめようとする作用が過剰に作用し、基準走行ラインへの移動を阻害する制御が行われてしまう。
【００８７】
このような事態に対応して、安定化制御トルクｕ_sを基準走行ライン追従系とは独立して設定可能とし、基準走行ライン追従トルクｕ₂の変化に応じて安定化制御トルクｕ_sを変化させることにより、基準走行ライン追従トルクｕ₂に関する制限が定常的に系に作用した場合に発生してしまう急激なトルク変化や走行位置の変化を抑止することができる。具体的には、基準走行ライン追従トルクｕ₂のリミット処理作動時には、安定化制御トルクｕ_sも小さくすることで、すなわちレギュレータゲインＫ_s（車両状態量のフィードバックゲイン）を小さい値に変更（前記ステップＳ７においてレギュレータゲインＫ_sとしてレギュレータゲインＫ₁を選択）することで、系の減衰力を相対的に低下させて、基準走行ライン追従系の過減衰状態を解消する。このようにすることで、車両１の移動速度を適切にすることができる。
【００８８】
以上のように、リミッタ処理の作動時に、車両状態を安定させるための状態フィードバック制御を変化させることで、目標軌道への追従速度を劣化させてしまうことを解消でき、目標軌道への追従性と走行安定性とを両立することができる。
また、安定化制御トルクｕ_sの調整或いは補正を、複数のレギュレータゲインＫ₀，Ｋ₁を用意するだけでよく、すなわち基準走行ライン追従トルクｕ₂のリミット処理の作動により発生する基準走行ライン追従系の過減衰状態が起きないようにレギュレータゲインＫ_sを設計しておくだけでよく、目標軌道への追従性と走行安定性とを両立することができるシステムを簡単な構成として構築することができる。
【００８９】
また、前述したようにオブザーバ１１０としてカルマンフィルタを用いているので、より好適な演算を実効することができる。すなわち、オブザーバゲインの要素数は出力数（センサの個数）と状態変数との組み合わせ（出力数（センサの個数）×状態変数）になるので、安定で（オブザーバの極の実部が全て負の値であって）、推定精度が良く、かつセンサのノイズ成分を有効に除去することが可能なオブザーバの設計は一般に易しくない。一方、カルマンフィルタの設計にあたっては、センサのノイズや車両に作用するプロセスノイズの分散値が設計のパラメータとして必要ではあるが、これらは予め調べることが可能であるため、ほとんど試行錯誤することなくオブザーバゲインを設計することが可能である。このようなことから、前述したようにオブザーバ１１０としてカルマンフィルタを用いることで、より好適な演算を実効することが可能になる。
【００９０】
以上、本発明の実施の形態について説明した。しかし、本発明は、前述の実施の形態として実現されることに限定されるものではない。
すなわち、前述の実施の形態では、ＣＣＤカメラ１７が撮影した画像に基づいて、車両１の基準走行ラインに対する相対位置（横変位）を検出しているが、これに限定されるものではい。例えば、公知の誘導ケーブル方式や公知の磁気マーカ方式によって横変位を検出するようにしてもよい。或いは、特開平３−２９３４１１号公報に開示される技術のように、走行路に沿って設置されたガードレールと車両との間の距離を超音波を利用して測定し、その測定した距離から基準走行ラインに対する車両の横方向位置を検出してもよい。さらには、複数の方式を組み合わせてもよい。
【００９１】
また、前述の実施の形態では、前輪３Ｌ，３Ｒを操舵するようにしているが、後輪を操舵するようにしてもよいし、或いは、前輪及び後輪の両方を操舵するようにしてもよい。
また、前述の実施の形態では、操舵角のみが自動制御できるようになっているが、これに限定されるものではない。例えば、４輪のブレーキや駆動を独立に制御することができるアクチュエータを設け、そのアクチュエータの状態をコントローラ２０から供給する制御信号によって制御可能とすることで、車両のヨーモーメントを任意に制御できるようにしてもよい。
【００９２】
また、前述の実施の形態では、リミッタ処理作動時に選択される第１の安定化制御トルクｕ₀は１種類に限るものではなく、複数種類であってもよい。具体的には、車両１の横方向の移動速度が大きくなるときに徐々に小さくなるような複数のレギュレータゲインを用意し、この複数のレギュレータゲインを用いて複数の第１の安定化制御トルクｕ₀を得るようにする。この場合、横方向の移動速度に合わせてレギュレータゲインを変更して、第１の安定化制御トルクｕ₀（安定化制御トルクｕ_s）を得ることで、多少移動速度は低下するが、より滑らかな車両挙動を得ることができる。すなわち、目標走行ライン追従トルクがリミット処理されている状態におけるレギュレータゲイン切り換え時には、ゲインの切り換えに伴う目標操舵トルクの変動や車両の横方向への移動速度を考慮した設定をきめ細く行うことが可能になり、これにより、なめらかな車両の走行を保つことが可能になる。
【００９３】
なお、ここでは、車両状態を車両１の横方向の移動速度とし、その移動速度に応じてレギュレータゲインを決定し、そして第１の安定化制御トルクｕ₀を変更しているが、他の車両状態に応じてレギュレータゲインし、第１の安定化制御トルクｕ₀を変更してもよい。
また、ゲイン切り換え後に車両が基準走行ライン上を交差する瞬間か、又は基準走行ライン上を交差することが予想されたとき、基準走行ライン追従トルクｕ_sを０にすると同時に基準走行ラインを修正する。具体的には、制御目標としている基準走行ラインを現在走行している走行ラインに修正する。これにより、車両の目標軌道に対する行き過ぎ量（オーバーシュート量）を極力減らし、より望ましい走行状態を保つことができる。
【００９４】
また、前述の実施の形態では、図７に積算部１４０の具体的な構成を示したが、これに限定されるものではない。例えば、積算部１４０を図１０に示すように構成してもよい。
この場合、積算部１４０は、前記加算部１２３からの前方注視点横変位ｙ_srと基準走行ラインｙ_refとの偏差Δｙと、１サンプリング時刻前の偏差Δｙとの移動平均を演算し、演算した値にサンプリング時間Ｔ／２を乗じる。そしてその値に、１サンプリング時間前の積算値を加算し、リミット処理部１４１で所定のリミット処理を行って最新の積算値を算出する。
【００９５】
なお、積算部１４０中のｑ^-1はシフト演算子であり、入力された信号をｑ^-1で１サンプリング時刻だけ遅れて出力する。また、リミット処理部１４１で用いる積算値の上限値及び下限値は、基準走行ライン追従トルクｕ₂が所定の上限値及び下限値の範囲内にとどまるように決定している。
そして、基準走行ライン追従トルク発生部１３０は、このような積算部１４０により算出した積算値に乗算部１３１でゲインＫ₂を乗じ、基準走行ライン追従トルクｕ₂を得る。
【００９６】
この図１０に示す積算部１４０と前記図７に示す積算部１４０とでは演算精度が違っている。すなわち、前記図７に示す積算部１４０は、サンプリング時点での値だけを考慮する零次ホールド回路による近似演算を用いた構成をなし、図１０に示す積算部１４０は、サンプリング区間の前後の値を考慮した台形公式による近似演算を用いた構成をなしている。図７に示す積算部１４０による演算と図１０に示す積算部１４０による演算とは実質的には同じ演算であるが、得られる精度に差がある。例えば、サンプリング時間Ｔをあまり小さくできない等の制約がある場合、台形公式を用いた方が演算精度が高く、また滑らかに変化する結果を得られるので、この場合には、図１０に示す積算部１４０を用いる方がよい。このように使用状況等に応じて使い分けるようにしてもよい。
【００９７】
なお、前述の実施の形態において、ＣＣＤカメラ１７及びステップＳ１の処理は、基準走行ラインに対する車両の横方向の相対位置を検出する相対位置検出手段を構成し、オブザーバ１１０、基準走行ライン追従トルク発生部１３０、安定化制御トルク発生部（Ｋ_s）１２４、及び加算器１２２，１２３は、相対位置検出手段の検出結果に応じて目標操舵トルクを演算する目標操舵トルク演算手段を構成し、自動操舵機構１０は、目標操舵トルクに基づいて車両の転舵輪に舵角を発生させる舵角発生手段を構成している。
【００９８】
また、目標操舵トルク演算手段を構成するオブザーバ１１０、基準走行ライン追従トルク発生部１３０、安定化制御トルク発生部（Ｋ_s）１２４、及び加算器１２２，１２３のうち、オブザーバ１１０は、車両が走行中の道路の曲率を推定する道路曲率推定手段と、相対位置検出手段の検出結果に基づいて車両の走行状態を推定する走行状態推定手段とを構成し、安定化制御トルク発生部（Ｋ_s）１２４は、走行状態推定手段が得た車両の走行状態に基づいて、道路曲率推定手段が得た道路曲率の基準走行ラインを基準にして車両を安定にする安定化制御トルクを算出する安定化制御トルク演算手段を構成し、基準走行ライン追従トルク発生部１３０及び加算器１２３は、相対位置検出手段が検出した相対位置に基づいて基準走行ラインに対する実際の走行ラインのずれ量を算出し、ずれ量に基づいて車両を基準走行ラインに追従走行させるための基準走行ライン追従トルクを算出する基準走行ライン追従トルク演算手段を構成し、加算器１２２は、安定化制御トルクと基準走行ライン追従トルクとを加算して目標操舵トルクを算出する加算手段を構成している。
【図面の簡単な説明】
【図１】本発明の実施の形態における車両の平面図である。
【図２】層だけいの一部を拡大した斜視図である。
【図３】車両の２輪モデル図である。
【図４】曲線走行状態をモデルとするオブザーバを用いた自動操舵系の構成を示すブロック図である。
【図５】本発明の実施の形態におけるコントローラの基本的な機能構成を示すブロック図である。
【図６】本発明の実施の形態における基準走行ライン追従トルク発生部の構成を示すブロック図である。
【図７】前記基準走行ライン追従トルク発生部の積算部の構成を示すブロック図である。
【図８】本発明の実施の形態における安定化制御トルク発生部の構成を示すブロック図である。
【図９】本発明の実施の形態におけるコントローラの処理手順を示すフローチャートである。
【図１０】前記基準走行ライン追従トルク発生部の積算部の他の構成を示すブロック図である。
【符号の説明】
１ 車両
２ ステアリングホイール
３Ｌ，３Ｒ 前輪（操舵輪）
１０ 自動操舵機構
１４ 舵角センサ
１７ ＣＣＤカメラ
１９ 車速センサ
２０ コントローラ
１１０ オブザーバ（カルマンフィルタ）
１２２，１２３ 加算器
１２４ 安定化制御トルク発生部
１３０ 基準走行ライン追従トルク発生部
１４０ 積算部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an automatic steering device that calculates a target steering angle according to a relative position of a vehicle in a lateral direction with respect to a reference travel line.
[0002]
[Prior art]
Some automatic steering devices calculate a target steering angle in accordance with the relative position of the vehicle in the lateral direction with respect to a reference travel line. In this automatic steering device, the state of the vehicle is estimated by an observer, and the target steering angle is calculated based on the estimated state quantity.
[0003]
However, in such an automatic steering device, a Kalman filter based on a straight traveling model is used as an observer. When traveling on a curved road using an observer based on a straight traveling model, a displacement of the traveling position or the like occurs due to an equilibrium point shift of the observer. In the automatic steering device disclosed in Patent Document 1, an equilibrium point steering angle (offset steering angle) estimator having an integral characteristic is provided in order to eliminate such a traveling position deviation caused by the observer's equilibrium point deviation. Provided. Specifically, the traveling position deviation caused by the observer's equilibrium point deviation is eliminated as follows.
[0004]
Calculate the equilibrium point steering angle as the value obtained by integrating the target steering angle until the target trajectory and the actual traveling trajectory match, or multiply the integrated value of the difference between the target trajectory and the actual traveling trajectory by a predetermined coefficient. An equilibrium point steering angle is calculated, and this equilibrium point steering angle is input to an observer based on the linear travel model. Here, the observer is configured to calculate a vehicle state quantity estimated from the yaw rate, yaw angle, lateral displacement, or the like.
[0005]
By inputting the virtually calculated equilibrium point steering angle to the observer in this way, even if the observer is designed as a straight running model, the equilibrium point deviation can be corrected and the correct vehicle state quantity can be estimated. Yes. Then, by performing state feedback control of the vehicle state quantity that is correctly estimated, it is realized that the vehicle travels along the reference travel line that is a curved road.
[0006]
As described above, on the assumption that the observer is built with a straight running model, the deviation of the running position due to modeling error that occurs when running on a curved road is corrected, and the standard running that is a curved road The vehicle is driven along the line.
[0007]
[Patent Document 1]
JP 10-297521 A
[0008]
[Problems to be solved by the invention]
In the above configuration, since the equilibrium point steering angle is integrated until the target track and the actual traveling track match, the design is performed in consideration of the influence of the steering operation due to the driver's intervention in consideration of mounting on the vehicle. There is a need. That is, when there is a steering operation by the driver's intervention, it is assumed that it does not coincide with the target trajectory for a long time, and in response to such a situation, a predetermined limiter means is provided in the integration processing unit, It must be configured so that the absolute value of the steering angle does not diverge.
[0009]
However, the limiter process is not limited to when the driver intervenes in steering. For example, the limiter process is also performed when the target trajectory is significantly changed. However, even when the target trajectory is greatly changed in this way, if the limiter process is performed in the integral calculation unit, the gain balance with the stabilization control unit will be lost, and the tracking speed to the target trajectory will be significantly reduced. Problem occurs.
[0010]
In the above-described configuration, since the state feedback element and the integral feedback gain are set independently, it is very difficult to properly operate the entire system, and it is necessary to repeat trial and error. . Furthermore, since the model itself is a straight traveling model, a large error occurs when the radius of curvature is small or the target trajectory is frequently switched, which may disturb traveling stability.
[0011]
Therefore, the present invention has been made in view of the above-described circumstances, and even when the limiter process is performed by the integration processing unit that causes the vehicle to travel along the reference travel line, the vehicle is stably supplied to the reference travel line. An object of the present invention is to provide an automatic steering device that can be followed.
[0012]
[Means for Solving the Problems]
In order to solve the above-described problem, an automatic steering apparatus according to the present invention includes a relative position detection unit that detects a relative position in a lateral direction of a vehicle with respect to a reference travel line, and a target according to a detection result of the relative position detection unit. In an automatic steering apparatus comprising target steering torque calculating means for calculating steering torque and steering angle generating means for generating a steering angle for a steered wheel of a vehicle based on the target steering torque, the target steering torque calculating means is The curvature of the road on which the vehicle is traveling is estimated by the road curvature estimating means, and the traveling state of the vehicle is estimated by the traveling state estimating means based on the detection result of the relative position detecting means, and the traveling state estimating means is obtained. Stabilization control torque calculating means for stabilizing the vehicle based on a road curvature reference travel line obtained by the road curvature estimation means based on the running state of the vehicle. To calculate a deviation amount of the actual travel line with respect to the reference travel line based on the relative position detected by the relative position detecting means, and to make the vehicle follow the reference travel line based on the deviation amount. The reference travel line follow-up torque is calculated by reference travel line follow-up torque calculating means, the stabilization control torque and the reference travel line follow-up torque are added, and the target steering torque is calculated by the adding means.
[0013]
That is, it is possible to individually set a reference traveling line follow-up torque for following a reference traveling line obtained as a predetermined road curvature and a stabilization control torque for stabilizing the vehicle with reference to the reference traveling line. . Specifically, the value of the stabilization control torque is set based on the change in the reference travel line following torque.
[0014]
【The invention's effect】
According to the present invention, by setting the reference travel line following torque and the stabilization control torque individually, it is possible to suppress a rapid torque change and a travel position change.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. 1 to 6 are diagrams showing an embodiment of the present invention.
FIG. 1 is a plan view showing a schematic configuration of a vehicle 1 on which an automatic steering device is mounted. First, the configuration shown in FIG. 1 will be described. Even in this vehicle 1, the driver steers the front wheels 3 </ b> L and 3 </ b> R as steered wheels by basically steering the steering wheel 2 in the same manner as a normal vehicle. A corner is generated. That is, as shown in an enlarged view in FIG. 2, the steering wheel 2 is integrally connected with the upper end portion of the steering shaft 4 in the rotational direction, and a relay means such as a universal joint (not shown) is used at the lower end of the steering shaft 4. To the input shaft to the power steering device 5. A pinion shaft is integrally formed with the input shaft by a power steering device 5, and the pinion shaft is engaged with a rack 6 that moves forward and backward in the vehicle lateral direction by the power steering device 5. Further, a torque sensor (not shown) for detecting, for example, torque generated in the steering system is provided between the input shaft and the pinion shaft to the power steering device 5, and the power steering device 5 is detected by the torque sensor. A steering assist torque is applied to the rack 6 by controlling a hydraulic actuator or the like so as to reduce the steering torque.
[0016]
The left and right tie rods 7L, 7R are connected to both ends of the rack 6, and the outer ends of the tie rods 7L, 7R are coupled to knuckles that rotatably support the front wheels 3L, 3R. Therefore, when the driver steers the steering wheel 2, the steering force is converted into the forward / backward force of the rack 6 via the steering shaft 4 and the power steering device 5, and is input to both the tie rods 7L, 7R, and is applied to the front wheels 3L, 3R. A rudder angle is generated. At this time, the steering assist torque is generated by the power steering device 5 so as to reduce the steering torque generated in the steering system, thereby reducing the burden on the driver.
[0017]
Further, the steering shaft 4 is provided with an automatic steering mechanism 10 that automatically generates a steering angle at the front wheels 3L and 3R. The automatic steering mechanism 10 includes an electric motor 10a, a reduction mechanism such as a worm gear, and an electromagnetic clutch that enables the power of the electric motor 10a to be intermittent in order to increase the rotational force of the electric motor 10a and transmit it to the steering shaft 4. Transmission mechanism 10b.
[0018]
As an operation of the automatic steering mechanism 10, when the driver selects manual steering, the electromagnetic clutch is disconnected so that the electric motor 10a does not become a load on the steering system. That is, when automatic steering is selected, the electromagnetic clutch is connected, and the rotational force of the electric motor 10a is transmitted to the steering shaft 4 via the transmission mechanism. A drive current is supplied to the electric motor 10a of the automatic steering mechanism 10 from a controller 20 mounted on the vehicle 1, and a rotational torque having a direction and a magnitude corresponding to the drive current is supplied to the electric motor. Since this occurs at 10a, the steering angle is generated at the front wheels 3L and 3R in the same manner as during manual steering.
[0019]
In this embodiment, since the torque sensor of the power steering device 5 is disposed closer to the rack 6 than the connection position of the transmission mechanism of the steering shaft 4, the rotational force of the electric motor 10a is small. Also, the steering angle can be generated in the front wheels 3L, 3R.
The steering shaft 4 is provided with a steering angle sensor 14 that measures the rotational displacement and detects the actual steering angle of the front wheels 3L and 3R. The steering angle sensor 14 is configured by combining a gear mechanism that accelerates the rotation of the steering shaft and an angle detection mechanism such as a rotary encoder or a potentiometer as a configuration that can accurately detect an actual steering angle. The output of the steering angle sensor 14 is supplied to the controller 20 as a steering angle detection signal.
[0020]
Further, the vehicle 1 is equipped with a CCD camera 17 for photographing a traveling road ahead of the vehicle 1. The CCD camera 17 is configured to be able to photograph a road surface several meters to several tens meters ahead of the vehicle 1. The image captured by the CCD camera 17 is supplied to the controller 20. The controller 20 detects the relative position in the lateral direction of the vehicle 1 with respect to the reference travel line along the travel path based on the image captured by the CCD camera 17.
[0021]
Specifically, the controller 20 extracts information (for example, a center line, a side line or a guide rail drawn on the road surface) that can be used as a reference travel line from the image captured by the CCD camera 17. The controller 20 assumes a reference travel line based on the extracted information (for example, the center portion of the travel lane), and the reference travel line and the reference line extending in the vehicle front-rear direction through the center of gravity of the vehicle 1. The distance (difference) from a predetermined position (for example, the position 10 m ahead of the vehicle) is the relative position in the lateral direction of the vehicle 1 with respect to the reference travel line at the forward gazing point (hereinafter, forward gazing point lateral displacement y). _sr That's it. ) To detect.
[0022]
Further, the vehicle 1 is provided with a vehicle speed sensor 19 that detects the vehicle speed by detecting, for example, the number of rotations on the output side of the transmission and the number of rotations of the wheels. A vehicle speed detection signal detected by the vehicle speed sensor 19 is supplied to the controller 20.
The controller 20 inputs a steering torque to the steering shaft 4 based on each detection signal supplied and an image taken by the CCD camera 17. That is, the controller 20 performs a predetermined calculation based on each supplied detection signal and the like to obtain a target steering torque necessary for the vehicle 1 to travel along the reference traveling line on the traveling road surface, and to perform automatic steering. The current supplied to the electric motor 10 a of the mechanism 10 is controlled and the target steering torque is input to the steering shaft 4. Thereby, a steering angle is generated in the front wheels 3L and 3R which are the steered wheels of the vehicle according to the target steering torque.
[0023]
Here, although not shown, the controller 20 includes an A / D converter, an interface circuit such as a D / A converter, a memory device such as a ROM or a RAM, or a microcomputer.
Next, the automatic steering system will be described in more detail. Here, an automatic steering system using an observer will be described. FIG. 3 shows the configuration of the control system of the automatic steering system (configuration in the controller 20), and basic processing and the like will be described with reference to FIG.
[0024]
In FIG. 3, 110 is an observer and 121 is a regulator gain.
The observer 110 is constructed as a Kalman filter. The observer 110 has matrices A, B, and C and a gain L, and includes adders 111 and 112. In addition, s shown in the observer 110 is a Laplace operator, and z is an estimated value of the state quantity x.
[0025]
With such a configuration, the controller 20 controls the steering torque u of the vehicle 1. That is, the observer 110 receives the steering torque u and the front gaze lateral displacement y. _sr The observer 110 is provided with the steering torque u and the forward gaze lateral displacement y. _sr The state quantity vector (state quantity) x is estimated based on Here, the state quantity vector x has a yaw rate γ and a relative yaw angle φ with respect to the traveling lane. _r , Relative lateral displacement y with respect to the driving lane _cr , Relative lateral speed y 'with respect to the driving lane _cr , Steering angle δ _f And steering angular velocity δ ′ _f It is.
[0026]
The matrices A, B, and C are determined from the model of the vehicle 1, and first, the vehicle model is represented by a general two-wheel model as shown in FIG. In addition, the meaning of each symbol in FIG. 4 is as follows.
β: body slip angle, C _f : Front wheel cornering power (for two wheels), C _r : Rear wheel cornering power (for two wheels), m: vehicle weight, I: yaw moment of inertia, δ _f : Front wheel actual steering angle, V: Vehicle speed, l _f : Distance between center of gravity and front wheel, l _r : Distance between the center of gravity and the rear wheel, F _f : Front wheel cornering force, F _r : Rear wheel cornering force
The equation of motion of this model is as shown in the following equation (1).
[0027]
[Expression 1]

[0028]
Further, when the amount of change in the tangential slope of the target track is defined by the road curvature ρ, the change in the yaw angle of the vehicle relative to the target track per unit time, that is, the relative yaw rate φ ′ with respect to the target track _r Is
[0029]
[Expression 2]

[0030]
And the relative lateral velocity y ′ _cr Is
[0031]
[Equation 3]

[0032]
Therefore, when the skid angle and the skid angular velocity are eliminated from the formula (1) and arranged, the following formulas (4) and (5) are obtained, respectively.
[0033]
[Expression 4]

[0034]
[Equation 5]

[0035]
Further, the lateral gaze lateral displacement y in the lateral direction of the vehicle 1 with respect to the reference travel line _sr Is the distance in front of the vehicle at the front gazing point position L _s In [m], relative lateral displacement y from the target track (traveling lane) _cr , Relative yaw angle φ _r When the road curvature ρ is used, the following equation (6) can be used.
[0036]
[Formula 6]

[0037]
On the other hand, the equation of motion of the steering system in the automatic steering apparatus shown in FIG. 2 can be described as follows.
[0038]
[Expression 7]

[0039]
Where ζ: trail length of the vehicle, K _T : Torque coefficient of steering actuator, N _s : Torque transmission ratio from steering actuator to steering system, C _s : Damping coefficient of steering system And, similarly to the above formulas (4) and (5), the side slip angle β and the side slip angular velocity β ′ are eliminated and rearranged, the following formula (8) is obtained.
[0040]
[Equation 8]

[0041]
By using these equations (4) to (6) and (8) as new targets and adding the road curvature to the state variables in a matrix representation, the vehicle model in the curved running state is described as the following equation (9) it can.
[0042]
[Equation 9]

[0043]
For the dynamic system, using a matrix L
[0044]
[Expression 10]

[0045]
The state variable vector z is calculated by performing the following calculation. Assuming that the deviation between the state variable vector z calculated here and the true vehicle state quantity vector x is e, the following equation (11) is obtained from the equations (9) and (10).
[0046]
## EQU11 ##

[0047]
When the matrix L is determined so that the real part of all eigenvalues of the transition matrix A-LC is negative, the following equation (12) is obtained.
[0048]
[Expression 12]

[0049]
Thereby, an arbitrary initial value e (t ₀ ), The deviation vector e between the previously calculated state variable vector z and the true vehicle state quantity vector x becomes 0 after a sufficient amount of time has elapsed, and the state variable vector z and the true vehicle state quantity Since it coincides with the vector x, the state quantity of the vehicle can be detected correctly.
Here, the equation (10) shows an observer, and a solution using such an observer is a technique often used for state estimation of a dynamic system.
[0050]
On the other hand, in consideration of process noise w and observation noise (sensor noise) n to the system, the matrix expression of the system can be described as the following equation (13).
[0051]
[Formula 13]

[0052]
Then, the deviation vector e can be obtained as the following equation (14), similarly to the observer described above.
[0053]
[Expression 14]

[0054]
Here, when it is assumed that the process noise w and the sensor noise n are normal white noises, the matrix L obtained so as to minimize the estimation error e is a Kalman filter, which is often used for state estimation of a dynamic system. Used.
By the way, strictly speaking, the process noise w is not normal white noise, such as a moment component in the yaw direction and a lateral force component in the lateral direction caused by a crosswind or a road surface cant (one slope). Due to disturbances and modeling errors such as vehicle parameter errors, the estimated state value may not converge to a true value. As a result, there is a possibility that a steady displacement occurs in the simple state feedback control using only the regulator 121 shown in FIG. In such a case, the vehicle is steered in a direction away from the reference travel line, and the driver is likely to feel a strong sense of discomfort.
[0055]
On the other hand, the controller 20 in the present embodiment has the following configuration. FIG. 5 is a block diagram of the entire system to which the present invention is applied, including the basic functional configuration (configuration shown in FIG. 3) of the controller 20.
That is, similarly to the controller 20 shown in FIG. 3, the controller 20 shown in FIG. 5 includes an observer 110 configured with a curved traveling state as a model, and a regulator gain K. _s (In FIG. 3, K ₀ ). The observer 110 also includes matrices A, B, and C and a gain L determined by the method described above, and includes adders 111 and 112. Here, s is a Laplace operator shown in the observer 110, z is an estimated value of the state quantity x, and Cz is an estimated value of the observed value vector y.
[0056]
Here, the controller 20 in the present embodiment will be described in detail with reference to FIG. 6. _sr And standard driving line y _ref And an adder 123 for obtaining a deviation Δy from the vehicle, and a reference travel line following torque u to be generated in the vehicle 1 based on the deviation Δy obtained by the adder 123 ₂ A reference running line follow-up torque generator 130 for calculating the _s Stabilization control torque generator (K _s ) 124 and the reference travel line following torque u calculated by the reference travel line following torque generating unit 130 ₂ And stabilization control torque generator (K _s ) Stabilization control torque u calculated by 124 _s And an adder 122 for obtaining the target steering torque u.
[0057]
The reference travel line follow-up torque generator 130 includes an accumulator 140 and an output value from the accumulator 140 with a gain K ₂ And a multiplication unit 131 for multiplying by.
The integrating unit 140 is configured as shown in FIG. The accumulating unit 140 is the output value from the adding unit 123, which is the forward gazing point lateral displacement y. _sr And standard driving line y _ref Is multiplied by the sampling time T, and the integrated value obtained by the sampling unit 140 one sampling time before is added to the value, and the limit processing unit 141 performs a predetermined limit process to calculate the latest integrated value. To do.
[0058]
In addition, q in the accumulating unit 140 ^-1 Is a shift operator, and the input signal is expressed as q ^-1 The output is delayed by one sampling time. Further, the upper limit value and the lower limit value of the integrated values used in the limit processing unit 141 are the reference travel line following torque u. ₂ Is determined to remain within the range of the predetermined upper limit value and lower limit value.
The reference travel line follow-up torque generator 130 adds the gain K calculated by the multiplier 131 to the integrated value calculated by the integrator 140. ₂ Multiplied by the reference travel line following torque u ₂ Get. The gain K of the multiplication unit ₂ The selection of the value of will be described in detail later.
[0059]
Stabilization control torque generator (K _s ) 124 is a stabilization control torque u for stabilizing the traveling state of the vehicle around the reference traveling line given by the road curvature ρ. _s Is calculated. FIG. 8 shows a specific configuration thereof.
A state quantity z estimated by the observer 110 is input to the stabilization control torque generation unit 124, and the stabilization control torque generation unit 124 has two types of regulator gains K. ₀ , K ₁ Each is multiplied by its state quantity z and correspondingly the first and second stabilization control torques u ₀ , U ₁ Is calculated. Regulator gain K ₀ , K ₁ For selection of the value of regulator gain K _s This will be described in detail later.
[0060]
The stabilization control torque generator 124 is configured to generate a reference travel line following torque u. ₂ The first stabilization control torque u depending on whether or not the limit is processed ₀ And the second stabilization control torque u ₁ And the first stabilization control torque u ₀ And the second stabilization control torque u ₁ And either one of the stabilization control torque u _s Output as. Here, the first stabilization control torque u ₀ Is the second stabilization control torque u ₁ The reference running line following torque u ₂ When the limit process is activated, the first stabilization control torque u ₀ Stabilizing control torque u _s Output as.
[0061]
In this way, the stabilization control torque generator (K _s ) 124 to stabilize the control torque u _s Is obtained, and this stabilization control torque u _s Is a value for correcting a small deviation around the reference travel line with ρ as the road curvature. And this stabilization control torque u _s Is input to the adder 122. In the adder 122, this stabilization control torque u _s And the reference travel line follow-up torque u from the reference travel line follow-up torque generator 130 ₂ Are added to obtain the target steering torque u.
[0062]
Here, gain K of integrating section 130 ₂ And regulator gain K _s (Regulator gain K ₀ , K ₁ ) Will be described.
Gain K ₂ Is a gain having a characteristic that, when the value is increased, the follow-up speed to the reference travel line is increased. Because of this, gain K ₂ Is the forward gaze lateral displacement y _sr And standard driving line y _ref Even if the integrated value of the deviation Δy is slightly changed, the reference running line following torque u ₂ Will give a big change. As a result, for example, the forward gaze lateral displacement y _sr Reference running line following torque u that reacts sensitively to noise components included in the detection result ₂ May be calculated and the movement of the steering system may become awkward.
[0063]
Regulator gain K _s Gain K compared to ₂ Is too large, the stabilization control torque u for stabilizing the vehicle based on the reference travel line _s Will fall relatively. As a result, the forward gaze lateral displacement y _sr And standard driving line y _ref Even if the deviation Δy is not so large, the reference travel line following torque u ₂ Becomes dominant, and the amount of overshoot with respect to the reference driving line increases. As a result, the control torque in the direction deviating from the reference travel line remains at the moment when the vehicle passes on the reference travel line, which is not preferable from the viewpoint of vehicle running stability or the driver's steering feeling. May occur.
[0064]
Conversely, gain K ₂ If the value is made too small, the running stability is relatively improved, but the follow-up speed is lowered. As a result, a steering torque that resists the vehicle from reaching the target track even though the vehicle has not yet reached the target track is calculated, and an undesirable state occurs in terms of the driver's steering feeling. there is a possibility.
From the above, gain K ₂ And regulator gain K _s (Regulator K ₀ , K ₁ ) Are not set individually as arbitrary values, but gain K ₂ And regulator gain K _s The combination is determined so as to cooperate well. For example, the dynamic characteristics in a state in which the automatic steering control is performed on the vehicle 1 are sufficiently confirmed by experiments and simulations, and are appropriately selected in consideration of the dynamic characteristics. For example, gain K ₂ Value and regulator gain K _s The optimal combination with the value of is obtained by the following method.
[0065]
For example, in the configuration as shown in FIG. ₀ As a method for obtaining the above, there is known a method in which a regulator that minimizes the next evaluation function J corresponding to the equation (9) is an optimum regulator.
[0066]
[Formula 15]

[0067]
Here, the response of the system can be adjusted by adjusting the matrices Q and R, and the stabilization control torque u is expressed by the following equation (16). ₀ Can be calculated.
[0068]
[Formula 16]

[0069]
By applying such a technique, the apparatus shown in FIG. _sr And standard driving line y _ref The following equation (17) can be obtained, where h is the integral value of the deviation Δy.
[0070]
[Formula 17]

[0071]
What is necessary is just to calculate the optimal regulator which minimizes the evaluation function J shown to following (18) Formula in such an extended system.
[0072]
[Formula 18]

[0073]
Where the matrix q _h , Q ′, R ′ can be adjusted to adjust the response of the system, and the target steering torque u can be calculated by the following (19).
[0074]
[Formula 19]

[0075]
Even in this case, it is necessary to sufficiently check the dynamic characteristics in a state where the automatic steering control is performed on the vehicle 1 through experiments and simulations, and to select them appropriately in consideration of the dynamic characteristics. However, in practice, the tracking speed to the reference travel line is expressed as a matrix q. _h If the vehicle stability at that time seems to be insufficient, a desired response can often be obtained by gradually correcting the elements of the matrix Q ′ to a larger value.
[0076]
FIG. 9 shows a processing procedure of the controller 20. The controller 20 executes this process in synchronization with a predetermined sampling clock.
When the automatic steering control is executed, first in step S1, the controller 20 detects the steering angle detection signal δ from the steering angle sensor 14. _f The vehicle speed detection signal V from the vehicle speed sensor 19 and the image from the CCD camera 17 are read.
[0077]
Note that the vehicle speed detection signal V is information necessary for determining the elements of the matrices A, B, and C of the observer 110, and thus is taken into the matrices A, B, and C as the vehicle speed V.
Subsequently, in step S2, the controller 20 determines the front gazing point lateral displacement y based on the image read in step S1. _sr Is detected.
[0078]
Subsequently, in step S3, the controller 20 performs a calculation for estimating the actual vehicle state quantity x by executing a calculation (equation (10)) based on the observer 110, and outputs a vehicle state estimation vector z. .
Subsequently, in step S4, the controller 20 determines that the vehicle state quantity estimation vector z obtained in step S3 is different from two types of attenuation characteristics. ₀ , K ₁ Based on the above, two kinds of stabilization control torques u having different damping characteristics ₀ , U ₁ (Obtained from the above (Equation 19)). That is,
u ₀ = -K ₀ z, u ₁ = -K ₁ z
Get.
[0079]
Subsequently, in step S5, the controller 20 determines that the forward gazing point lateral displacement y detected in step S2. _sr And standard driving line y _ref Is integrated with the difference Δf and gain K ₂ Multiplied by the reference travel line following torque u ₂ Is calculated.
Further, in this step S5, the controller 20 detects the reference travel line following torque u. ₂ Limit processing with a predetermined upper limit value or lower limit value, and the lateral displacement y _sr And standard driving line y _ref The limit process is also performed on the integrated value of the deviation Δy. And a limit flag is set at the time of a limit. Each limit process is performed in this way, so that the absolute value of the deviation integral value does not increase any more. That is, the reference travel line following torque u ₂ The lateral displacement y at the time when the limit is processed _sr And standard driving line y _ref The integrated value of the deviation Δy becomes an upper limit value or a lower limit value.
[0080]
Subsequently, in step S6, the controller 20 detects the reference travel line following torque u obtained in step S5. ₂ It is determined whether or not limit processing has been performed. That is, it is determined whether or not the limit flag is set, and the reference travel line following torque u is determined. ₂ It is determined whether or not limit processing has been performed. Here, the controller 20 determines the reference travel line following torque u. ₂ If the limit process is performed, the process proceeds to step S7, where the reference travel line following torque u ₂ If the limit process is not performed, the process proceeds to step S8.
[0081]
In step S7, the controller 20 performs the first stabilization control torque u. ₀ Stabilizing control torque u _s Determine as. On the other hand, in the controller 8, the controller 20 performs the second stabilization control torque u. ₁ Stabilizing control torque u _s Determine as. As described above, the controller 20 causes the reference travel line follow-up torque u by step S7 and step S8. ₂ Two types of stabilization control torque u depending on whether the limit is processed or not ₀ , U ₁ To stabilize control torque u _s Is determined.
[0082]
Subsequently, in step S9, the controller 20 controls the stabilization control torque u. _s And reference travel line following torque u ₂ Are added to obtain the target steering torque u. Here, for example, processing is performed in which the controlled steering torque changes smoothly.
Subsequently, in step S10, the controller 20 performs actuator drive processing as output processing of the target steering torque u obtained in step S9. That is, the controller 20 controls the current to the electric motor 10a of the automatic steering mechanism 10 and inputs the target drive torque to the steering shaft 4, thereby steering the front wheels 3L and 3R of the vehicle so as to obtain the target steering torque. Generate a corner. As the control here, for example, the current applied to the electric motor 10a constituting the automatic steering mechanism is controlled by a PWM mechanism using an H bridge circuit or the like, and the target steering torque u is input to the steering system.
[0083]
When the process of step S10 ends, the current process ends. Thereafter, after waiting for a predetermined sampling time to elapse, the process is executed again from step S1.
The controller 20 performs processing in the above procedure. By such processing, the vehicle follows the road with the curvature by the action of the reference running line follow-up torque, and stably runs along the road having the curvature by the action of the stabilization control torque. become.
[0084]
Next, the effect will be described.
As described above, the reference travel line following torque u ₂ Is subjected to limit processing with a predetermined upper limit value or lower limit value, and lateral displacement y _sr And standard driving line y _ref By performing the limit process also on the integrated value of the deviation Δy from the reference travel line follow-up torque u (step S5). ₂ Therefore, excessive generation of the target steering torque u can be suppressed.
[0085]
Thus, for example, even when the driver suddenly stops the steering intervention from a state in which the driver does not easily reach the reference driving line due to the steering intervention of the driver, the reference driving line following torque u ₂ Thus, it is possible to prevent the steering torque from being continuously input even after passing the reference travel line due to the influence of the residual component. Similarly, there is no excessive vehicle behavior in situations such as when a strong disturbance suddenly decreases, the direction of the disturbance changes suddenly, or when the position of the reference travel line is changed significantly. The vehicle 1 can be run.
[0086]
While the limit processing has the above effects, when the limit processing is activated, the reference travel is calculated no matter how much the target track changes in the direction that increases the reference travel line following torque. Since the line following torque becomes a constant value, the feedback gain of the offset torque becomes equivalently small. This is because the reference running line follow-up torque is maintained at a constant value in spite of the process of increasing the integral value, so the gain decreases while the limit process is operating. Is equivalent to. In such a case, the entire control system is attenuated excessively. That is, the action of trying to keep the posture of the vehicle as it is by the state feedback acts excessively, and the control for inhibiting the movement to the reference travel line is performed.
[0087]
In response to such a situation, the stabilization control torque u _s Can be set independently of the reference travel line following system, and the reference travel line following torque u ₂ Stabilization control torque u in response to changes in _s By changing the reference travel line following torque u ₂ It is possible to suppress a rapid torque change and a change in the travel position that occur when the restriction on the system steadily acts on the system. Specifically, reference running line following torque u ₂ When the limit process is activated, the stabilization control torque u _s Is reduced, that is, the regulator gain K _s (Vehicle state quantity feedback gain) is changed to a smaller value (regulator gain K in step S7). _s Regulator gain K as ₁ By selecting the), the damping force of the system is relatively lowered, and the overdamped state of the reference traveling line following system is eliminated. By doing in this way, the moving speed of the vehicle 1 can be made appropriate.
[0088]
As described above, by changing the state feedback control for stabilizing the vehicle state during the operation of the limiter process, it is possible to eliminate the deterioration of the follow-up speed to the target track, and to follow the target track. It is possible to achieve both running stability.
Also, the stabilization control torque u _s Adjustment or correction of multiple regulator gains K ₀ , K ₁ Need only be prepared, that is, the reference running line following torque u ₂ Regulator gain K so that the overdamped state of the reference running line following system that occurs due to the limit processing operation of the motor does not occur _s It is only necessary to design the system, and it is possible to construct a system that can achieve both the followability to the target track and the running stability as a simple configuration.
[0089]
In addition, since the Kalman filter is used as the observer 110 as described above, more suitable calculation can be performed. That is, the number of elements of the observer gain is a combination of the number of outputs (number of sensors) and state variables (number of outputs (number of sensors) x state variables), so it is stable (all real parts of the observer poles are all negative) In general, it is not easy to design an observer that has good estimation accuracy and can effectively remove the noise component of the sensor. On the other hand, when designing a Kalman filter, the variance of sensor noise and process noise acting on the vehicle is required as a design parameter. However, since these can be examined in advance, the observer gain is almost never trial and error. Can be designed. For this reason, as described above, it is possible to execute a more preferable calculation by using the Kalman filter as the observer 110.
[0090]
The embodiment of the present invention has been described above. However, the present invention is not limited to being realized as the above-described embodiment.
That is, in the above-described embodiment, the relative position (lateral displacement) of the vehicle 1 with respect to the reference travel line is detected based on the image taken by the CCD camera 17, but the present invention is not limited to this. For example, the lateral displacement may be detected by a known induction cable method or a known magnetic marker method. Alternatively, as in the technique disclosed in Japanese Patent Laid-Open No. 3-293411, the distance between the guardrail installed along the traveling road and the vehicle is measured using ultrasonic waves, and the reference is determined from the measured distance. The lateral position of the vehicle relative to the travel line may be detected. Furthermore, a plurality of methods may be combined.
[0091]
In the above-described embodiment, the front wheels 3L and 3R are steered. However, the rear wheels may be steered, or both the front wheels and the rear wheels may be steered. .
In the above-described embodiment, only the steering angle can be automatically controlled. However, the present invention is not limited to this. For example, by providing an actuator that can independently control the braking and driving of four wheels, and controlling the state of the actuator by a control signal supplied from the controller 20, the yaw moment of the vehicle can be arbitrarily controlled. It may be.
[0092]
Further, in the above-described embodiment, the first stabilization control torque u selected at the time of the limiter processing operation. ₀ Is not limited to one type, and may be a plurality of types. Specifically, a plurality of regulator gains that gradually decrease when the lateral movement speed of the vehicle 1 increases are prepared, and a plurality of first stabilization control torques u are prepared using the plurality of regulator gains. ₀ To get. In this case, the first stabilization control torque u is changed by changing the regulator gain in accordance with the moving speed in the lateral direction. ₀ (Stabilization control torque u _s ), The moving speed is somewhat reduced, but a smoother vehicle behavior can be obtained. In other words, when switching the regulator gain when the target travel line follow-up torque is in the limit process, it is possible to fine-tune settings that take into account fluctuations in the target steering torque that accompany gain switching and the moving speed of the vehicle in the lateral direction. Thus, it is possible to keep the vehicle running smoothly.
[0093]
Here, the vehicle state is the lateral movement speed of the vehicle 1, the regulator gain is determined according to the movement speed, and the first stabilization control torque u ₀ However, the gain of the regulator is increased according to other vehicle conditions, and the first stabilization control torque u is changed. ₀ May be changed.
In addition, when it is predicted that the vehicle will cross the reference travel line after gain switching, or when it is predicted that the vehicle will cross the reference travel line, the reference travel line following torque u _s At the same time, the reference running line is corrected. Specifically, the reference travel line as the control target is corrected to the travel line currently traveling. Thereby, the overshoot amount (overshoot amount) with respect to the target track of the vehicle can be reduced as much as possible, and a more desirable traveling state can be maintained.
[0094]
In the above-described embodiment, the specific configuration of the integrating unit 140 is shown in FIG. 7, but is not limited to this. For example, the integrating unit 140 may be configured as shown in FIG.
In this case, the accumulating unit 140 calculates the forward gazing point lateral displacement y from the adding unit 123. _sr And standard driving line y _ref The moving average of the deviation Δy and the deviation Δy one sampling time before is calculated, and the calculated value is multiplied by the sampling time T / 2. Then, the integrated value one sampling time before is added to the value, and the limit processing unit 141 performs a predetermined limit process to calculate the latest integrated value.
[0095]
In addition, q in the accumulating unit 140 ^-1 Is a shift operator, and the input signal q ^-1 The output is delayed by one sampling time. The upper limit value and the lower limit value of the integrated values used in the limit processing unit 141 are the reference travel line following torque u. ₂ Is determined to remain within the range of the predetermined upper limit value and lower limit value.
Then, the reference travel line follow-up torque generation unit 130 multiplies the integrated value calculated by the integration unit 140 with the gain K by the multiplication unit 131. ₂ Multiplied by the reference travel line following torque u ₂ Get.
[0096]
The accumulating unit 140 shown in FIG. 10 and the accumulating unit 140 shown in FIG. That is, the integration unit 140 shown in FIG. 7 has a configuration using an approximation calculation by a zero-order hold circuit that takes into account only the value at the time of sampling, and the integration unit 140 shown in FIG. This is a configuration that uses an approximate calculation based on a trapezoidal formula that takes into account. The calculation by the integration unit 140 shown in FIG. 7 and the calculation by the integration unit 140 shown in FIG. 10 are substantially the same calculation, but there is a difference in the obtained accuracy. For example, when there is a restriction such that the sampling time T cannot be made too small, the use of the trapezoidal formula provides higher calculation accuracy and results that change smoothly. In this case, the integration unit shown in FIG. It is better to use 140. In this way, it may be used properly according to the usage situation or the like.
[0097]
In the above-described embodiment, the CCD camera 17 and the processing in step S1 constitute relative position detecting means for detecting the relative position in the lateral direction of the vehicle with respect to the reference travel line, and the observer 110 generates the reference travel line following torque. Part 130, stabilization control torque generator (K _s ) 124 and adders 122 and 123 constitute target steering torque calculating means for calculating the target steering torque according to the detection result of the relative position detecting means, and the automatic steering mechanism 10 determines the vehicle based on the target steering torque. Steering angle generating means for generating a steering angle in the steered wheels is configured.
[0098]
Further, the observer 110 constituting the target steering torque calculation means, the reference travel line follow-up torque generator 130, and the stabilization control torque generator (K _s ) 124 and the adders 122 and 123, the observer 110 estimates the traveling state of the vehicle based on the road curvature estimating means for estimating the curvature of the road on which the vehicle is traveling and the detection result of the relative position detecting means. A running state estimating means, and a stabilization control torque generating section (K _s ) 124 is a stabilization that calculates a stabilization control torque that stabilizes the vehicle based on the road curvature reference travel line obtained by the road curvature estimation unit based on the vehicle travel state obtained by the travel state estimation unit. The control torque calculating means constitutes the reference running line follow-up torque generator 130 and the adder 123, which calculates the deviation of the actual running line with respect to the reference running line based on the relative position detected by the relative position detecting means. A reference travel line follow-up torque calculating means for calculating a reference travel line follow-up torque for causing the vehicle to follow the reference travel line based on the amount; and an adder 122 includes a stabilization control torque, a reference travel line follow-up torque, Are added to calculate the target steering torque.
[Brief description of the drawings]
FIG. 1 is a plan view of a vehicle in an embodiment of the present invention.
FIG. 2 is an enlarged perspective view of a part of the layer.
FIG. 3 is a two-wheel model diagram of a vehicle.
FIG. 4 is a block diagram showing a configuration of an automatic steering system using an observer modeled on a curved running state.
FIG. 5 is a block diagram showing a basic functional configuration of a controller in the embodiment of the present invention.
FIG. 6 is a block diagram showing a configuration of a reference travel line following torque generating unit in the embodiment of the present invention.
FIG. 7 is a block diagram illustrating a configuration of an integrating unit of the reference travel line following torque generating unit.
FIG. 8 is a block diagram showing a configuration of a stabilization control torque generating unit in the embodiment of the present invention.
FIG. 9 is a flowchart showing a processing procedure of the controller in the embodiment of the present invention.
FIG. 10 is a block diagram showing another configuration of the integrating unit of the reference travel line following torque generating unit.
[Explanation of symbols]
1 vehicle
2 Steering wheel
3L, 3R Front wheel (steering wheel)
10 Automatic steering mechanism
14 Rudder angle sensor
17 CCD camera
19 Vehicle speed sensor
20 controller
110 Observer (Kalman filter)
122,123 Adder
124 Stabilization control torque generator
130 Reference travel line following torque generator
140 Integration unit

Claims (6)

Based on the relative position detecting means for detecting the relative position in the lateral direction of the vehicle with respect to the reference travel line, the target steering torque calculating means for calculating the target steering torque according to the detection result of the relative position detecting means, and the target steering torque A steering angle generating means for generating a steering angle for the steered wheels of the vehicle,
The target steering torque calculating means includes a road curvature estimating means for estimating the curvature of a road on which the vehicle is running, a running state estimating means for estimating a running state of the vehicle based on a detection result of the relative position detecting means, Stabilization control torque calculation for calculating a stabilization control torque that stabilizes the vehicle based on a road curvature reference travel line obtained by the road curvature estimation means based on a vehicle travel state obtained by the travel state estimation means And an actual travel line deviation amount with respect to the reference travel line based on the relative position detected by the relative position detection means, and for causing the vehicle to follow the reference travel line based on the deviation amount. A reference travel line follow-up torque calculating means for calculating a reference travel line follow-up torque, and adding the stabilization control torque and the reference travel line follow-up torque to the target Automatic steering apparatus characterized by comprising an adding means for calculating the steering torque. 2. The stabilization control torque calculating means corrects the stabilization control torque in a decreasing direction when the reference running line following torque calculating means limits the reference running line following torque. Automatic steering device. The reference travel line follow-up torque calculating means calculates the reference travel line follow-up torque based on an integrated value of the deviation amount of the actual travel line with respect to the reference travel line,
When the integrated value of the deviation amount exceeds a predetermined value, at least a limit process in which the reference running line follow-up torque calculation means does not increase the integrated value of the deviation amount or a limit process that does not increase the reference running line follow-up torque. 3. The automatic steering apparatus according to claim 2, wherein the stabilization control torque calculating means corrects the stabilization control torque in a decreasing direction while performing one. The road curvature estimating means and the running state estimating means are configured as an observer for estimating a vehicle state quantity defined with reference to a reference running line that is a curved road,
The stabilization control torque calculation means calculates the stabilization control torque by multiplying a vehicle state quantity estimated by the observer by a regulator gain,
The automatic steering apparatus according to claim 2 or 3, wherein the stabilization control torque calculation unit corrects the stabilization control torque in a decreasing direction by changing the regulator gain. Until the reference running line follow-up torque calculating means limits the reference running line follow-up torque, the stabilization control torque calculating means applies a regulator gain to be multiplied by the vehicle state quantity estimated by the observer according to the vehicle state. The automatic steering apparatus according to claim 4, wherein the automatic steering apparatus is changed. The automatic steering apparatus according to claim 4, wherein the observer is a Kalman filter.

Images