JP2004049523A - Method and equipment for controlling running vehicle - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To relax an impact force caused by an impact, or the like, with an obstacle, or the like, and to obtain desired operation characteristics even when the situation of a running surface changes. <P>SOLUTION: An output of a joystick 2 is converted into a power command F<SB>hum</SB>in a virtual translation direction and a power command N<SB>hum</SB>in a rotatory direction and they are inputted to a translation direction control system 10 and a rotational direction control system 20. The power commands F<SB>hum</SB>and N<SB>hum</SB>are inputted to compliance control means 12 and 22, while high-frequency components of external forces F<SB>anv</SB>and N<SB>anv</SB>applied to a wheelchair 1 and estimated by a counter force estimating means 7 are inputted to compliance control means 13 and 23, and a compliance control is conducted on a human input and the external forces. Outputs of the control means 11 and 21 are subjected to coordinate transformation, converted further into current signals and given to the running vehicle 1. The low-frequency components of the counter force are inputted to an adaptive control means 18 and a conversion coefficient for the power command F<SB>hum</SB>is varied by a value corresponding to the friction force between the wheelchair 1 and the running surface. <P>COPYRIGHT: (C)2004,JPO

Description

【０００８】
上記仮想的な力指令Ｆ_ｈｕｍとトルク指令Ｎ_ｈｕｍに対応する駆動信号は、最終的に車椅子１の左右両輪１ｂを駆動するモータに与えられ、図２に示すように車椅子１の並進方向への駆動力Ｆ_ｈｕｍ、回転方向への駆動力Ｎ_ｈｕｍを発生させる。
【０００９】
上記力指令Ｆ_ｈｕｍとトルク指令Ｎ_ｈｕｍは、それぞれ図１に示す並進方向制御系１０、回転方向制御系２０に与えられる。
並進方向制御系１０は、制御手段１１、コンプライアンス制御手段１２，１３、ハイパスフィルタ１４、座標変換手段１５，１６を備え、車椅子１からフィードバックされる速度信号θ’^ｒｅｓ、反力推定手段７が出力する走行面との摩擦、障害物との衝突等（以下これらを合わせて環境という）により生ずる反力τ_ｅｎｖ、逐次更新される車椅子１の現在位置、速度を表す位置信号Ｘ^ｃｍｄ、速度信号Ｘ’^ｃｍｄ、および、上記操作量→力変換手段３が出力する力指令Ｆ_ｈｕｍに基づき、並進方向の加速度指令Ｘ”^ｒｅｆを出力する。
一方、回転方向制御系２０は、制御手段２１、コンプライアンス制御手段２２，２３、ハイパスフィルタ２４、座標変換手段２５，２６を備え、車椅子１からフィードバックされる速度信号θ’^ｒｅｓ、反力推定手段７が出力する外部環境により生ずる反力τ_ｅｎｖ、逐次更新される車椅子１の現在の回転角度、回転速度を表す回転角度信号φ^ｃｍｄ、回転速度信号φ’^ｃｍｄ、および、上記操作量→力変換手段３が出力するトルク指令Ｎ_ｈｕｍに基づき、回転方向の加速度指令φ”^ｒｅｆを出力する。
なお、図中では、上記速度、加速度等の微分信号を、上にドット１つあるいはドット２つを付して示すが、文中では、上記のように「’」，「”」を付して示す。
【００１０】
上記並進方向制御系１０、回転方向制御系２０が出力する加速度指令Ｘ”^ｒｅｆ、φ”^ｒｅｆは座標変換手段４に与えられ、車椅子１の両輪の角加速度指令θ”^ｒｅｆに変換され、さらに、上記角加速度指令θ”^ｒｅｆは、θ”→Ｉ変換手段５により電流指令Ｉ^ｒｅｆに変換される。
一方、外乱推定手段６により、車椅子１に加わる外乱が推定され、外乱推定手段６から外乱量に応じた電流信号Ｉ^ｃｍｐが出力される。上記θ”→Ｉ変換手段５が出力する電流指令Ｉ^ｒｅｆと、外乱推定手段６が出力する電流信号Ｉ^ｃｍｐは、加算され車椅子１の両輪を駆動するモータ制御装置（図示せず）に与えられ、車椅子１が駆動される。
また、反力推定手段７は、外部環境により生ずる外力を推定し、推定外力τ_ｅｎｖを出力する。反力推定手段７が出力する推定外力τ_ｅｎｖは、並進方向制御系１０、回転方向制御系２０の座標変換手段１５，２５によりそれぞれ並進方向の成分であるＦ_ｅｎｖ成分、回転方向の成分であるＮ_ｅｎｖ成分に変換され、ハイパスフィルタ１４，２４により高周波成分が取り出され、並進方向制御系１０、回転方向制御系２０のコンプライアンス制御手段１３，２３に与えられる。
【００１１】
コンプライアンス制御は、外界から力を受けた場合に外力に倣うように軌道補正を行なうという制御手段であり、これにより環境との安定な接触を実現することができ、車椅子の走行に柔軟性を持たせることが可能となる。
本実施例では、ジョイスティック２による人間の入力と反力推定手段７によって推定された環境からの外力にコンプライアンス制御を適用している。
上記コンプライアンス制御手段１３，２３およびコンプライアンス制御手段１２，２２は、後述するように、模式的に質量とバネとダンパーで表される２次遅れ要素からなる緩衝制御系に相当するものであり、入力信号である推定外力τ_ｅｎｖのＦ_ｅｎｖ成分、Ｎ_ｅｎｖ成分をゆっくりと変化する出力信号に変換する。
したがって、例えば車椅子が障害物に衝突した際に、急激に変化する外力が生じても、並進方向制御系１０、回転方向制御系２０の制御手段１１，２１にフィードバックされる信号は、ゆっくりと変化した信号となり、制御系に柔らかい応答をさせることができる。
同様に、ジョイスティック２の操作力をコンプライアンス制御手段２２，２３に入力することにより、ジョイスティック２の出力をゆっくり変化する信号に変換することができ、ジョイスティック２を急激に操作しても、制御系に柔らかい応答をさせることができる。
【００１２】
さらに、上記反力推定手段７が出力する推定外力τ_ｅｎｖのＦ_ｅｎｖ成分は、ローパスフィルタ１７に入力され、低周波分が取り出され、適応制御手段１８に入力される。
適応制御手段１８は後述するように、Ｆ_ｅｎｖの低周波成分から走行面の摩擦力を求め、該摩擦力に応じて、操作量→力変換手段３における力指令Ｆ_ｈｕｍへの変換係数〔前記（１）式におけるＫ_Ｆ〕（以下ゲインともいう）を変える。
例えば、走行面の摩擦係数が大きくなった場合には、上記ゲインを大きくして、ジョイスティックの操作量が同じ量でも、車椅子に与える駆動力を大きくする。これにより、例えば搭乗者や走行面の状況が変化した場合にも所望の操作特性が得ることができる。
【００１３】
図４、図５は上記並進方向制御系１０、回転方向制御系２０の詳細ブロック図である。同図は、図１の並進方向制御系と、回転方向制御系を分けて２枚の図で表したものであり、図４は並進方向制御系１０の構成を示し、図５は回転方向制御系２０の構成を示しており、図１と同一のものには同一の符号が付されている。
なお、図４、図５では、ジョイスティック２、操作量→力変換手段３、座標変換手段４、θ”→Ｉ変換手段５、車椅子１、外乱推定手段６、反力推定手段７をそれぞれ示しているが、これらは、図１に示したように共通部分である。
また、図４、図５では、車椅子１の車輪の回転角度θ^ｒｅｓを、回転速度θ’^ｒｅｓに変換するブロック１ｃ〔（ｓｇ／ｓ＋ｇ）のブロック〕を示しているが、図１では、車椅子１とブロック１ｃをまとめて車椅子１とし、車椅子１の出力を回転速度θ’^ｒｅｓとしている。
【００１４】
以下、図４、図５の各部について説明する。
（１）外乱推定手段６
本実施例では、車椅子に作用する外乱を推定する外乱推定手段６を用いる。外乱推定手段６は、パラメータ変動も含め、車椅子に作用する外乱の総和を推定し、推定された外乱をフィードバックすることにより、外乱を相殺する。これによりロバストな制御系を構成することができる。
上記外乱推定手段６は、例えば、村上俊之，　中村亮，　郁方銘，　大西公平：　”　反作用力推定オブザーバに基づいた多自由度ロボットの力センサレスコンプライアンス制御”　日本ロボット学会誌，　Ｖｏｌ．１１，　Ｎｏ．　５，　ｐｐ．７６５−７６８，　（１９９３）　（以下論文１という）、西川直樹，　藤本康孝，　村上俊之，　大西公平：　”　環境変動を考慮した３　次元２　足歩行ロボットの可変コンプライアンス制御”　電気学会産業応用部門誌，　Ｖｏｌ．１１９−Ｄ，Ｎｏ．１２，　ｐｐ．１５０７−１５１３，　（１９９９）　（以下論文２という）に記載されており、本実施例においても、上記論文１、論文２に記載されるものを使用することができる。
図６に外乱推定手段６のブロック図を示す。同図に示すように、外乱推定手段６は、車椅子の車輪の回転速度θ’^ｒｅｓと、入力電流Ｉ_ａを取り込み、車椅子１に作用する外乱を推定し、Ｉ_ｃｍｐを出力する。
推定された外乱は次の（３）式のようになる。
【００１５】
【数２】

【００１６】
ここで、ｇ_ｄｉｓは外乱推定手段６のゲインであり、外乱推定手段６によって推定された外乱τ_ｄｉｓは、電流信号Ｉ_ｃｍｐとして、前記電流指令Ｉ^ｒｅｆと加算され、車椅子１に入力される。
【００１７】
（２）反力推定手段７
環境により生ずる外力を検出するために、本実施例では反力推定手段７を用いる。上記反力推定手段７は、前記した論文１、論文２に記載されており、本実施例においても、上記論文１、論文２に記載されるものを使用することができる。図７に、反力推定手段７のブロック図を示す。反力推定手段７は、τ_ｉｎｉｔを差し引くようにした点を除き、上記外乱推定手段６と同様の構成であり、車椅子の車輪の回転速度θ’^ｒｅｓと、入力電流Ｉ_ａを取り込み、車椅子１に作用する外力を推定し、τ_ｅｎｖを出力する。
また、推定された外力は、前記（４）式のようになる。
ここで、ｇ_ｒｅａｃは反力推定手段７のゲインであり、τ_ｉｎｉｔは慣性行列のノミナル値との差である。反力推定手段７によって推定された外力τ_ｅｎｖは、座標変換手段１５，２５により、並進方向の成分であるＦ_ｅｎｖ、回転方向の成分であるＮ_ｅｎｖに変換される。
上記変換は、次の（５）式により行われる。
【００１８】
【数３】

【００１９】
上記座標変換手段１５，２５により変換されたＦ_ｅｎｖ成分、Ｎ_ｅｎｖ成分は、ハイパスフィルタ１４，２４に入力され、高周波成分のみが取り出され、コンプライアンス制御手段１３，２３に入力される。したがって、コンプライアンス制御手段１４，２４は車椅子１に作用する反力のうちの高周波成分にみに応答し、走行面の摩擦力、搭乗者の重量等に影響されない。
【００２０】
（３）コンプライアンス制御手段１２，１３，２２，２３
前記したように、本実施例では、ジョイスティック２による人間の入力と、反力推定手段７によって推定された環境からの外力に対して、コンプライアンス制御を適用している。
まず、人間の入力について仮想インピーダンス特性行列Ｍｃ　_ｈｕｍ，Ｄｃ　_ｈｕｍ，Ｋｃ　_ｈｕｍを設定すると、仮想インピーダンスモデルは次の（６）式で与えられる。
【００２１】
【数４】

【００２２】
ここで、Ａｃ　_ｈｕｍは人間の入力の増幅ゲインである。（６）式をブロック図で示すと、図８のように２次遅れ要素となる。
また、同様にして環境からの外力についても仮想インピーダンスモデルを設定すると、（７）式のようになる。
【００２３】
【数５】

【００２４】
ここで、Ａｃ　_ｅｎｖは外力のフィードバックゲインである。（７）式をブロックで示すると図９のように２次遅れ要素となる。
図４に示すコンプライアンス制御手段１２，１３は、位置、速度、加速度指令値Ｘｃ　_ｈｕｍ，Ｘ’ｃ　_ｈｕｍ，Ｘ”ｃ　_ｈｕｍ、および、位置、速度、加速度フィードバック信号Ｘｃ　_ｅｎｖ，Ｘ’ｃ　_ｅｎｖ，Ｘ”ｃ　_ｅｎｖが出力されるように、図８、図９に示すブロック図〔（６）（７）式〕を変形したものであり、上記位置、速度、加速度指令値および位置、速度、加速度フィードバック信号を制御手段１１に出力し、コンプライアンス動作を実現する。
上記では、並進方向のＦ_ｈｕｍ，Ｆ_ｅｎｖ成分について説明したが、コンプライアンス制御手段２２，２３においても、上記と同様であり、上記（６）（７）式、図８、図９の、Ｆ_ｈｕｍ，Ｆ_ｅｎｖをそれぞれＮ_ｈｕｍ，Ｎ_ｅｎｖに置き換え、Ｘをφに置き換えればよい。
【００２５】
（４）制御手段１１，２１
制御手段１１には、現在位置、速度を示す位置信号Ｘ^ｃｍｄ、速度信号Ｘ’^ｃｍｄおよび上記コンプライアンス制御手段１２が出力する位置、速度、加速度指令値Ｘｃ　_ｈｕｍ，Ｘ’ｃ　_ｈｕｍ，Ｘ”ｃ　_ｈｕｍ、および、コンプライアンス制御手段１３が出力する位置、速度、加速度フィードバック信号Ｘｃ　_ｅｎｖ，Ｘ’ｃ　_ｅｎｖ，Ｘ”ｃ　_ｅｎｖおよび、車椅子１の車輪の角速度θ’^ｒｅｓを座標変換手段１６により変換した速度フィードバック信号Ｘ’^ｒｅｓが入力される。
制御手段１１は、図４に示すように、現在位置指令Ｘ^ｃｍｄと、座標変換手段１６が出力するＸ’^ｒｅｓを積分した位置フィードバック信号Ｘ^ｒｅｓとの偏差からコンプラインアンス制御手段１２，１３が出力する位置信号Ｘｃ　_ｈｕｍ，Ｘｃ　_ｅｎｖを減じて係数Ｋｐを乗じて位置偏差分を求め、また、上記速度フィードバック信号Ｘ’^ｒｅｓと、現在位置に対応した速度指令Ｘ’^ｃｍｄとの偏差から、コンプラインアンス制御手段１２，１３が出力する速度信号Ｘ’ｃ　_ｈｕｍ，Ｘ’ｃ　_ｅｎｖを減じて係数Ｋｖを乗じて速度偏差分を求め、さらに上記位置偏差分と速度偏差分の和からさらに、コンプラインアンス制御手段１２，１３が出力する加速度信号Ｘ”ｃ　_ｈｕｍ，Ｘ”ｃ　_ｅｎｖを減じることにより、並進方向の加速度指令Ｘ”^ｒｅｆを求める。
すなわち、上記制御手段１１は、次の（８）式により、上記並進方向の加速度指令Ｘ”^ｒｅｆを出力する。これにより、コンプライアンス動作を実現することができる。
【００２６】
【数６】

【００２７】
上記現在位置、速度を示す位置信号Ｘ^ｃｍｄ、速度信号Ｘ’^ｃｍｄは、ジョイスティック２の操作指令により車椅子が移動し、車椅子の位置、速度が更新される毎に逐次更新される。
なお、本実施例では、位置信号Ｘ^ｃｍｄ、速度信号Ｘ’^ｃｍｄとして現在位置、速度を与えており、人間がジョイスティック２を操作しないときは、車椅子１が停止するようにしているが、上記位置信号Ｘ^ｃｍｄ、速度信号Ｘ’^ｃｍｄとして予め定められた軌道に沿って変化する信号を与えることにより、車椅子を軌道に沿って移動させることができ、ジョイスティック２の操作に応じて、車椅子を上記軌道から変位させることができる。
上記では、並進方向のＦ_ｈｕｍ，Ｆ_ｅｎｖ成分について説明したが、制御手段２１においても、上記と同様にして、回転方向の加速度指令φ”^ｒｅｆを出力する。なお、制御手段２１の場合には、上記（８）式のＸをφに置き換えればよい。
【００２８】
本実施例では、上記のように、ジョイスティック２による人間の入力と、反力推定手段７によって推定された環境からの外力に対して、それぞれ、コンプライアンス制御を適用しているので、位置指令、回転方向指令に対する位置・回転方向応答特性と、人間の力に対する位置・回転方向応答性、そして環境からの外力に対する位置・回転方向応答特性を独立に設計することができる。
そのため、位置・回転方向制御性能を維持しつつ所望のコンプライアンス特性を実現することが可能である。
また、並進方向、回転方向の環境のモードに対してもコンプライアンス特性を独立に設計することが可能である。例えば、車椅子の一方の車輪が小石につまずいてしまったような場合、並進方向に関しては柔らかくし、回転方向に関しては堅い制御をすると、石を乗り越えるようなこともできる。
したがって、コンプライアンス特性を変えることで、目的に応じて人間が使いやすいように車椅子の走行特性を変化させることが可能である。
【００２９】
（５）座標変換手段１６，２６、座標変換手段４、およびθ”→Ｉ変換手段５
上記車椅子１を、図１０に示すモデルで表し、車椅子１のパラメータを以下のように定義する。
ｘ_０　：世界座標系における参照点のｘ座標
ｙ_０　：世界座標系における参照点のｙ座標
φ　　：車椅子の方向角
Ｒ　　：駆動輪の半径
Ｗ　　：トレッド
θｒ　’：右輪の回転角
θｌ’：左輪の回転角
上記車椅子１の駆動輪１ｂの中点Ｐｏ（ｘ_０，ｙ_０）を制御参照点とすると、車椅子１の位置および姿勢は次の（９）式で与えられる。
また、上記車椅子１の運動学は次の（１０）式で与えられ、（１１）式に時間微分を施すと、次の（１２）式が得られる。
これより、以下の（１２）式に示す世界座標系における加速度指令Ｘ”^ｒｅｆ，加速度指令φ”^ｒｅｆから、角加速度指令θ”^ｒｅｆを求めることができる。
【００３０】
【数７】

【００３１】
座標変換手段１６，２６は、車椅子１の回転加速度θ’^ｒｅｓを上記（１０）式により、速度フィードバック信号Ｘ’^ｒｅｓ，φ^’ｒｅｓに変換する。
また、座標変換手段４は制御手段１１，２１が出力する加速度指令Ｘ”^ｒｅｆ，φ”^ｒｅｆを上記（１２）式により車椅子の両輪の回転加速度指令θ”^ｒｅｆに変換する。
【００３２】
また、車椅子の重心位置が、図１０に示すＰ_０であるとすると、運動エネルギー関数Ｋは、（１３）式で定義される。
これより、Ｌａｇｒａｎｇｅ方程式を解くことにより、動力学関係式が次の（１４）（１５）式のように定義される。ここで、Ｍθは等価慣性行列である。
【００３３】
【数８】

【００３４】
上記θ”→Ｉ変換手段５は、上記（１４）式により回転加速度指令θ”^ｒｅｆをトルク指令に変換し、トルク指令を車椅子の両輪を駆動するモータへの電流指令Ｉ^ｒｅｆに変換する。
【００３５】
（６）適応制御手段１８
適応制御手段１８は、前記したように、Ｆ_ｅｎｖの低周波成分から走行面の摩擦力を求め、該摩擦力に応じて、操作量→力変換手段３における力指令Ｆ_ｈｕｍへの変換係数〔前記（１）式におけるＫ_Ｆ〕を変える。
図４に示すように、反力推定手段７で推定した環境より車椅子１にかかる外力Ｆ_ｅｎｖをローパスフィルタ１７に入力する。
Ｆ_ｅｎｖのうち、ローパスフィルタ１７が出力する周波数の低い成分は定常的な外乱であるので、車椅子１と床の摩擦による成分とみなすことができる。
この摩擦力を車椅子の速度Ｘ’^ｒｅｓで割ることで、環境の摩擦係数μ_ｅｎｖを同定することができる。
環境の摩擦係数μ_ｅｎｖは搭乗者の重量や走行面の摩擦などをすべて含んでいる。したがって、それらの変動に適応することが可能である。
そして、この摩擦係数μ_ｅｎｖを所望の環境の摩擦係数μ_０で割り、操作量→力変換手段３における係数Ｋ_Ｆを求める。
このように、係数Ｋ_Ｆを可変にすることで、搭乗者や走行面の状況が変化した場合にも所望の操作特性を得ることができる。すなわち、搭乗者や走行面の状況に応じて、アシスト比を可変にすることができ、人間は一定の力を加えるだけで走行面の状況が変わったときでも同じ操作量を得ることができる。
【００３６】
以上のように、本実施例においては、環境からの反力を反力推定手段により推定し、推定した反力をハイパスフィルタ１４，２４と、ローパスフィルタ１７とにより、高周波成分と低周波成分とに分け、高周波成分は障害物との衝突による成分であるので、衝撃力を和らげるコンプライアンス制御を行い、また、低周波成分は、車輪と走行面との間の摩擦力であるので、摩擦力に応じて上記のように操作量→力変換手段３における係数Ｋ_Ｆ変えている。
このため、障害物との衝突等により生ずる急激な反力に対しては、柔らかい制御を行うことができ、また、走行面の摩擦係数の変化に対しては、適切なアシスト比に設定することができる。
すなわち、路面がジャリ道になったり、車椅子の重さが重くなった場合でも、人間は同じ力を加えることで同じ速度を得ることができる。よって大幅な操作性の向上が期待できる。また、操作を誤って障害物に衝突した場合は、衝撃力の緩和を行っているので安全性も高めることが可能である。
【００３７】
なお、上記実施例では、ジョイスティックを用いる場合について説明したが、図１１に示すように、車椅子１に２個のトルクセンサ３１ａ，３１ｂを取り付け、トルクセンサ３１ａ，３１ｂの出力を並進方向の力と回転方向の力に変換するように構成してもよい。
このように構成すれば、前記ジョイスティックを用いる場合と同様、人が上記トルクセンサ３１ａ，３１ｂに加える力に応じて、車椅子１を任意の方向に動かすことができる。
また、ハンドルを人が押しながら前進させる電動台車、ゴルフカート等において、このハンドル部分に上記トルクセンサ３１ａ，３１ｂを取り付け、本実施例で示した制御装置により、電動台車、ゴルフカート等を駆動するように構成すれば、比較的小さな力で電動台車、ゴルフカート等を任意の方向に動かすことが可能となる。
トルクセンサ３１ａ，３１ｂを用いる場合、前記操作量→力変換手段３は、以下の（１６）式により並進方向の力Ｆ^Ｈ、回転方向のトルクＮ^Ｈに変換する。
トルクセンサ３１ａ，３１ｂを用いる場合にも、（１６）式における係数の値を適応制御手段１８の出力により変えることにより、ジョイスティックを用いる場合と同様に、搭乗者や走行面の状況が変化した場合にも所望の操作特性を得ることができる。
【００３８】
【数９】

【００３９】
本発明の効果を検証するため、人間が本実施例の車椅子１を操作し、途中で物体を押す実験を行った。
前記ゲインＫ_Ｆを一定としたときの実験結果を図１２〜図１４に示す。この実験は、ジョイスティックを一杯に倒して、途中で物に当てた場合（外力を増やした）　の車椅子の位置（図１２）、仮想的な力指令値（図１３）、ローパスフィルタを通した環境からの外力（図１４）を求めたものであり、図１２〜図１４の横軸は時間であり、矢印は物に当たった時点を示す。
図１２〜図１４に示すように、前記ゲインＫ_Ｆが一定の場合でも、環境からの外力にコンプライアンス制御を行っているので、物体との接触時における衝撃力を緩和し、物体と安定な接触を行うことができる。
しかし、前記ゲインＫ_Ｆが一定であるので、物体を押すことによる床との摩擦力の増加によって車椅子の速度は遅くなり、止まってしまうことがわかる。
【００４０】
また、前記適応制御手段１８を設けて前記ゲインＫ_Ｆを環境に応じて可変にしたときの実験結果を図１５〜図１７に示す。
図１２〜図１４と同様、ジョイスティックを一杯に倒して、途中で物に当てた場合（外力を増やした）　の車椅子の位置（図１５）、仮想的な力指令値（図１６）、ローパスフィルタを通した環境からの外力（図１７）を求めたものであり、横軸は時間であり、矢印は物に当たった時点を示す。
適応制御手段１８を設けた場合には、図１５〜図１７に示すように、環境の摩擦力が増えた場合においても、ゲインＫ_Ｆが可変になっているために、物体による摩擦力の影響を受けずに所望の動作特性が実現できていることが分かる。
また、車椅子が動き始める際に大きな静止摩擦が働くが、この例では滑らかに発進することが可能となっている。
上記実験により、環境の摩擦係数を同定し、その値に応じてゲインＫ_Ｆを可変にすることで、環境の変化に適応して所望の動作特性が実現できることを確認することができた。
【００４１】
【発明の効果】
以上説明したように、本発明においては、以下の効果を得ることができる。
（１）走行車を操作するためのジョイスティック、あるいは、トルクセンサの出力を仮想的な並進方向の力指令と、回転方向の力指令に変換し、反力推定手段等により上記走行車に加わる外力を求め、上記ジョイスティックやトルクセンサによる人間の入力と、上記環境からの外力に対して、コンプライアンス制御を行っているので、障害物に衝突した場合等、環境からの外力が急激に変化することにより生ずる衝撃を緩和し、環境からの外力に対しても柔らかい制御を行うことができる。
（２）特に、人間の入力と、上記環境からの外力のそれぞれに対して、コンプライアンス制御を行っているので、人間の力に対する位置応答性、外力に対する位置応答性を独立に設計することができ、位置制御性能を維持しつつ、外力に対して所望のコンプライアンス特性を確保することができる。
（３）走行車に加わる外力の低周波成分から、走行車と走行面との間の摩擦力を求め、上記摩擦力に応じた値で、上記ジョイスティック、あるいは、トルクセンサのゲインを変化させることにより、走行車の重量の変動や走行面の状況が変化しても、所望の操作特性を得ることができる。
【図面の簡単な説明】
【図１】本発明の実施例の制御装置の全体構成を示す図である。
【図２】本実施例の適用対象である電動車椅子の概念構成を示す図である。
【図３】ジョイスティックをモデリングした図である。
【図４】並進方向制御系の詳細ブロック図である。
【図５】回転方向制御系の詳細ブロック図である。
【図６】外乱推定手段のブロック図である。
【図７】反力推定手段のブロック図である。
【図８】人間の入力から位置指令への変換のブロック図である。
【図９】外力から位置指令への変換のブロック図である。
【図１０】車椅子をモデル化した図である。
【図１１】トルクセンサによる入力を説明する図である。
【図１２】ゲインを一定にした場合の車椅子の位置変化を示す図である。
【図１３】ゲインを一定にした場合の仮想的な力指令値の変化を示す図である。
【図１４】ゲインを一定にした場合のローパスフィルタを通した外力の変動を示す図である。
【図１５】可変ゲインとした場合の車椅子の位置変化を示す図である。
【図１６】可変ゲインとした場合の仮想的な力指令値の変化を示す図である。
【図１７】可変ゲインとした場合のローパスフィルタを通した外力の変動を示す図である。
【符号の説明】
１　　　　　　　電動車椅子
２　　　　　　　ジョイスティック
３　　　　　　　操作信号→力変換手段
４　　　　　　　座標変換手段
５　　　　　　　θ”→Ｉ変換手段
６　　　　　　　外乱推定手段
７　　　　　　　反力推定手段
１０　　　　　　並進方向制御系
１１　　　　　　制御手段
１２，１３　　　コンプライアンス制御手段
１４　　　　　　ハイパスフィルタ
１５，１６　　　座標変換手段
１７　　　　　　ローパスフィルタ
１８　　　　　　適応制御手段
２０　　　　　　回転方向制御系
２１　　　　　　制御手段
２２，２３　　　コンプライアンス制御手段
２４　　　　　　ハイパスフィルタ
２５，２６　　　座標変換手段
３１ａ，３１ｂ　トルクセンサ[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a control device for a traveling vehicle that can be applied to an electric wheelchair, an electric cart, an electric bicycle, a mobile manipulator operated by radio, and the like.
[0002]
[Prior art]
In the face of an aging society with fewer children, the role of robots as human partners is expected, and various devices including assistive robots have been developed.
There is an electric wheelchair provided with a power assist function as a representative example of a movement support device used as an assistance device. This type of electric wheelchair is provided with an operating rod such as a handle for operating the wheelchair with the hand of the rider, and the wheelchair is driven by driving a pair of left and right main wheels according to the operating force applied to the operating rod. It is intended to run.
Similarly, an electric bicycle, an electric cart, and the like are known as driving a motor with a relatively small operating force by driving the electric motor with a force according to a human operating force.
An example of the electric wheelchair is disclosed in Japanese Patent Laid-Open No. 9-248319. In the above publication, even if the running resistance changes, the auxiliary force by the electric motor is controlled based on the operating force applied to the steering wheel, the vehicle speed, and the running resistance in order to perform appropriate control without excess or deficiency. Thus, an electric wheelchair with good running feeling is provided.
[0003]
[Problems to be solved by the invention]
What is disclosed in the above publication can perform appropriate control without excess or deficiency with respect to the running load, but performs control with a predetermined coefficient, and can flexibly cope with a complicated environment. I can't.
On the other hand, when a conventional electric wheelchair collides with an obstacle, the impact force is reduced by mechanically attaching a cushion or the like. However, no matter how much a cushion or the like is used, if the wheels continue to drive, the meaning will be lost.
In addition, many force sensors are used to detect external force, but the external force is detected only at the part where the force sensor is attached, that is, the part where the external force is applied is assumed in advance. Can not cope with.
Furthermore, in vehicles such as conventional electric wheelchairs that move with the assistance of human power, the assist ratio (driving force by the motor / driving force by the human power) is constant, and the human must change the force according to changes in the environment. I had to.
The present invention has been made to solve the above-mentioned problems of the prior art, and the first object of the present invention is to flexibly cope with changes in the environment, such as obstacles and the like. An object of the present invention is to provide a control device for a traveling vehicle that can alleviate an impact force caused by a collision or a change in a situation of a traveling surface and can perform a soft operation in response to a human command.
A second object of the present invention is a traveling vehicle control device that can obtain desired operation characteristics even when the condition of the occupant or the traveling surface changes by changing the assist ratio and allowing a human to apply a certain force. Is to provide.
[0004]
[Means for Solving the Problems]
The above problems are solved in the present invention as follows.
The output of the joystick for operating the traveling vehicle or the torque sensor is converted into a force command in the virtual translation direction and a force command in the rotation direction.
Then, an external force applied to the traveling vehicle is obtained by a reaction force estimating means or the like, and compliance control is performed on a human input from the joystick or torque sensor and an external force from the environment.
That is, a force command is input to the first secondary delay element to obtain a command signal, an external force applied to the traveling vehicle is obtained, and this external force is input to the second secondary delay element to provide feedback of the external force. Driving the vehicle based on the feedback signal from the traveling vehicle, the command signal output by the first secondary delay element, and the feedback signal of the external force output by the second secondary delay element. A signal is obtained and the traveling of the traveling vehicle is controlled by the driving means.
Thereby, it is possible to mitigate an impact such as when it collides with an obstacle, and to perform soft control against external force from the environment. In particular, compliance control is performed for both human input and external force from the above environment, so position responsiveness to human force and position responsiveness to external force can be designed independently. Desirable compliance characteristics with respect to external force can be ensured while maintaining performance.
Further, the external force applied to the traveling vehicle is input to a low-pass filter that cuts a high-frequency component, and the frictional force between the traveling vehicle and the traveling surface is obtained based on the output of the low-pass filter, and a value corresponding to the frictional force is obtained. By changing the conversion coefficient for converting the operation amount into a force command, it is possible to obtain desired operation characteristics even if the weight of the traveling vehicle changes or the state of the traveling surface changes.
Further, by inputting the external force applied to the traveling vehicle to the second second-order lag element through a high-pass filter that cuts low frequency components, the influence of fluctuations in the frictional force on the traveling surface, the weight of the traveling vehicle, etc. Therefore, a soft response characteristic can be obtained with respect to a sudden change in external force caused by a collision with an obstacle, and an impact due to a collision with the obstacle can be mitigated.
[0005]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 is a diagram illustrating an overall configuration of a control device according to an embodiment of the present invention, and FIG. 2 is a diagram illustrating a conceptual configuration of an electric wheelchair to which the present embodiment is applied. In the following embodiments, the case where the present invention is applied to an electric wheelchair will be described. However, as described above, the present invention is applied to human operations such as an electric cart, an electric bicycle, and a mobile manipulator operated by radio. It can be applied to various traveling vehicles that move by the above.
In FIG. 1, reference numeral 1 denotes an electric wheelchair. The electric wheelchair 1 includes a main body 1a and a pair of wheels 1b as shown in FIG. And a battery for supplying power to the control device, and a joystick 2 for operating the electric wheelchair 1. Moreover, the control apparatus shown in FIG. 1 is normally mounted on the main body 1a of the electric wheelchair 1.
[0006]
A human gives a command to the control device of the electric wheelchair 1 by operating the joystick 2 mounted on the electric wheelchair 1. In this embodiment, a joystick 2 having two axes in the X direction and the Y direction is used.
FIG. 3 is a diagram modeling the joystick 2, and the joystick 2 shows the x shown in FIG. _joy , Y _joy The operation signal is output.
The operation signal → force conversion means 3 shown in FIG. 1 is calculated by the following equations (1) and (2). _joy , Y _joy Virtual force command F _hum And torque command N _hum Convert to Note that K in equations (1) and (2) _F , K _N And the like are values determined when designing the control device in the wheelchair movement space. As will be described later, the conversion coefficient K _F By making the variable, desired operation characteristics can be obtained even when the conditions of the passenger and the running surface change.
[0007]
[Expression 1]

[0008]
The virtual force command F _hum And torque command N _hum Is finally given to a motor that drives both the left and right wheels 1b of the wheelchair 1. As shown in FIG. 2, the driving force F in the translational direction of the wheelchair 1 is given. _hum , Driving force N in the rotational direction _hum Is generated.
[0009]
Above force command F _hum And torque command N _hum Are given to the translation direction control system 10 and the rotation direction control system 20 shown in FIG.
The translation direction control system 10 includes a control unit 11, compliance control units 12 and 13, a high-pass filter 14, and coordinate conversion units 15 and 16, and a speed signal θ ′ fed back from the wheelchair 1. ^res The reaction force τ generated by the friction with the traveling surface, the collision with the obstacle, etc. (hereinafter collectively referred to as the environment) output by the reaction force estimation means 7 _env , Position signal X representing the current position and speed of wheelchair 1 that is sequentially updated ^cmd , Speed signal X ' ^cmd , And the force command F output from the manipulated variable → force converting means 3 _hum Based on the acceleration command X " ^ref Is output.
On the other hand, the rotation direction control system 20 includes a control unit 21, compliance control units 22 and 23, a high-pass filter 24, and coordinate conversion units 25 and 26, and a speed signal θ ′ fed back from the wheelchair 1. ^res The reaction force τ generated by the external environment output from the reaction force estimation means 7 _env The rotation angle signal φ representing the current rotation angle and rotation speed of the wheelchair 1 that is sequentially updated ^cmd , Rotation speed signal φ ' ^cmd , And the torque command N output from the manipulated variable → force converting means 3 _hum Based on the rotation direction acceleration command φ " ^ref Is output.
In the figure, the differential signals such as velocity and acceleration are shown with one dot or two dots on the top, but in the text, "'" and """are added as described above. Show.
[0010]
Acceleration command X ″ output from the translation direction control system 10 and the rotation direction control system 20 ^ref , Φ ” ^ref Is given to the coordinate conversion means 4, and the angular acceleration command θ "of both wheels of the wheelchair 1 is given. ^ref Furthermore, the angular acceleration command θ " ^ref Is the current command I by the θ ″ → I conversion means 5. ^ref Is converted to
On the other hand, the disturbance estimating means 6 estimates the disturbance applied to the wheelchair 1, and the current signal I corresponding to the amount of disturbance is received from the disturbance estimating means 6. ^cmp Is output. The current command I output from the θ ″ → I conversion means 5 ^ref And the current signal I output by the disturbance estimation means 6 ^cmp Are added to a motor control device (not shown) that drives both wheels of the wheelchair 1 to drive the wheelchair 1.
The reaction force estimating means 7 estimates the external force generated by the external environment and estimates the external force τ. _env Is output. Estimated external force τ output by the reaction force estimation means 7 _env Is a component in the translation direction by the coordinate conversion means 15 and 25 of the translation direction control system 10 and the rotation direction control system 20, respectively. _env N, which is a component in the rotation direction _env The high-frequency components are converted by the high-pass filters 14 and 24 and applied to the compliance control means 13 and 23 of the translation direction control system 10 and the rotation direction control system 20.
[0011]
Compliance control is a control means that corrects the trajectory so that it follows the external force when it receives a force from the outside world. This makes it possible to achieve stable contact with the environment and flexibility in running the wheelchair. It becomes possible to make it.
In the present embodiment, the compliance control is applied to the external force from the environment estimated by the human input by the joystick 2 and the reaction force estimation means 7.
As will be described later, the compliance control means 13 and 23 and the compliance control means 12 and 22 correspond to a buffer control system composed of a second-order delay element schematically represented by a mass, a spring, and a damper. Estimated external force τ as a signal _env F _env Ingredient, N _env Convert components to slowly changing output signals.
Therefore, for example, when a wheelchair collides with an obstacle, even if an external force that changes rapidly occurs, the signals fed back to the control means 11 and 21 of the translation direction control system 10 and the rotation direction control system 20 change slowly. It is possible to make the control system have a soft response.
Similarly, by inputting the operation force of the joystick 2 to the compliance control means 22 and 23, the output of the joystick 2 can be converted into a slowly changing signal. Even if the joystick 2 is operated suddenly, the control system Can give a soft response.
[0012]
Further, the estimated external force τ output by the reaction force estimating means 7 _env F _env The component is input to the low-pass filter 17 and the low frequency component is extracted and input to the adaptive control means 18.
As will be described later, the adaptive control means 18 is F _env The frictional force of the running surface is obtained from the low frequency component of the motor, and according to the frictional force, the force command F in the operation amount → force converting means 3 _hum Conversion coefficient [K in the above equation (1) _F ] (Hereinafter also referred to as gain).
For example, when the friction coefficient of the running surface increases, the gain is increased to increase the driving force applied to the wheelchair even if the joystick operation amount is the same. Thereby, a desired operation characteristic can be obtained even when, for example, the condition of the passenger or the running surface changes.
[0013]
4 and 5 are detailed block diagrams of the translation direction control system 10 and the rotation direction control system 20. This figure shows the translation direction control system of FIG. 1 and the rotation direction control system separately in two figures, FIG. 4 shows the configuration of the translation direction control system 10, and FIG. 5 shows the rotation direction control system. The configuration of the system 20 is shown, and the same components as those in FIG.
4 and 5 show the joystick 2, the operation amount → force conversion means 3, the coordinate conversion means 4, the θ ″ → I conversion means 5, the wheelchair 1, the disturbance estimation means 6, and the reaction force estimation means 7, respectively. However, these are common parts as shown in FIG.
4 and 5, the wheel rotation angle θ of the wheelchair 1 is shown. ^res , Rotation speed θ ' ^res In FIG. 1, the wheelchair 1 and the block 1 c are collectively referred to as a wheelchair 1, and the output of the wheelchair 1 is the rotational speed θ ′. ^res It is said.
[0014]
Hereafter, each part of FIG. 4, FIG. 5 is demonstrated.
(1) Disturbance estimation means 6
In this embodiment, the disturbance estimating means 6 for estimating the disturbance acting on the wheelchair is used. The disturbance estimation means 6 estimates the sum total of disturbances acting on the wheelchair including parameter fluctuations, and cancels the disturbances by feeding back the estimated disturbances. Thereby, a robust control system can be configured.
The disturbance estimation means 6 is, for example, Toshiyuki Murakami, Ryo Nakamura, Minokata, and Hiroshi Onishi: “Force sensorless compliance control of multi-degree-of-freedom robot based on reaction force estimation observer”, Journal of the Robotics Society of Japan, Vol. 11, no. 5, pp. 765-768, (1993) (hereinafter referred to as paper 1), Naoki Nishikawa, Yasutaka Fujimoto, Toshiyuki Murakami, and Hiroshi Onishi: "Variable compliance control of a three-dimensional two-legged walking robot considering environmental changes" Vol. 119-D, no. 12, pp. 1507-1513 (1999) (hereinafter referred to as paper 2), and those described in paper 1 and paper 2 can also be used in this embodiment.
FIG. 6 shows a block diagram of the disturbance estimation means 6. As shown in the figure, the disturbance estimation means 6 is used to determine the rotational speed θ ′ of the wheel of the wheelchair. ^res And the input current I _a , Estimate the disturbance acting on the wheelchair 1, _cmp Is output.
The estimated disturbance is expressed by the following equation (3).
[0015]
[Expression 2]

[0016]
Where g _dis Is the gain of the disturbance estimation means 6, and the disturbance τ estimated by the disturbance estimation means 6 _dis Is the current signal I _cmp The current command I ^ref And added to the wheelchair 1.
[0017]
(2) Reaction force estimation means 7
In this embodiment, reaction force estimation means 7 is used to detect an external force generated by the environment. The reaction force estimation means 7 is described in the paper 1 and paper 2 described above, and the one described in the paper 1 and paper 2 can also be used in this embodiment. FIG. 7 shows a block diagram of the reaction force estimating means 7. The reaction force estimation means 7 _init Is the same as the disturbance estimation means 6 except that the wheel speed of the wheelchair is θ ′. ^res And the input current I _a , Estimate the external force acting on the wheelchair 1, τ _env Is output.
Further, the estimated external force is expressed by the above equation (4).
Where g _reac Is the gain of the reaction force estimation means 7 and τ _init Is the difference from the nominal value of the inertia matrix. External force τ estimated by the reaction force estimation means 7 _env F is a component in the translation direction by the coordinate conversion means 15, 25. _env , N which is a component in the rotation direction _env Is converted to
The conversion is performed by the following equation (5).
[0018]
[Equation 3]

[0019]
F converted by the coordinate conversion means 15 and 25 _env Ingredient, N _env The components are input to the high pass filters 14 and 24, and only the high frequency components are extracted and input to the compliance control means 13 and 23. Accordingly, the compliance control means 14 and 24 respond to only the high frequency component of the reaction force acting on the wheelchair 1 and are not affected by the frictional force on the running surface, the weight of the occupant, and the like.
[0020]
(3) Compliance control means 12, 13, 22, 23
As described above, in this embodiment, compliance control is applied to human input from the joystick 2 and external force from the environment estimated by the reaction force estimation means 7.
First, a virtual impedance characteristic matrix Mc for human input _hum , Dc _hum , Kc _hum Is set, the virtual impedance model is given by the following equation (6).
[0021]
[Expression 4]

[0022]
Where Ac _hum Is the amplification gain of human input. When equation (6) is shown in a block diagram, it becomes a second-order lag element as shown in FIG.
Similarly, when a virtual impedance model is set for an external force from the environment, equation (7) is obtained.
[0023]
[Equation 5]

[0024]
Where Ac _env Is a feedback gain of external force. If the equation (7) is represented by a block, it becomes a second-order lag element as shown in FIG.
The compliance control means 12 and 13 shown in FIG. _hum , X'c _hum , X ″ c _hum , And position, velocity, acceleration feedback signal Xc _env , X'c _env , X ″ c _env 8 and FIG. 9 are modified so that the position, speed, acceleration command value, position, speed, and acceleration feedback signal are controlled. Output to the means 11 to realize the compliance operation.
In the above, F in the translation direction _hum , F _env Although the components have been described, the compliance control means 22 and 23 are the same as described above, and the above-described equations (6) and (7), F in FIGS. _hum , F _env N _hum , N _env And X may be replaced with φ.
[0025]
(4) Control means 11, 21
The control means 11 has a position signal X indicating the current position and speed. ^cmd , Speed signal X ' ^cmd And the position, speed, and acceleration command value Xc output from the compliance control means 12 _hum , X'c _hum , X ″ c _hum , And the position, velocity, acceleration feedback signal Xc output from the compliance control means 13 _env , X'c _env , X ″ c _env And the angular velocity θ ′ of the wheel of the wheelchair 1 ^res Is converted by the coordinate conversion means 16 to the velocity feedback signal X ′ ^res Is entered.
As shown in FIG. 4, the control means 11 ^cmd X ′ output from the coordinate conversion means 16 ^res Position feedback signal X integrated ^res Position signal Xc output from the compliance control means 12, 13 based on the deviation from _hum , Xc _env Is multiplied by the coefficient Kp to obtain the position deviation, and the speed feedback signal X ′ ^res Speed command X ′ corresponding to the current position ^cmd The speed signal X′c output from the compliance control means 12, 13 from the deviation from _hum , X'c _env Is multiplied by the coefficient Kv to obtain the speed deviation, and the acceleration signal X ″ c output from the compliance control means 12 and 13 is further calculated from the sum of the position deviation and the speed deviation. _hum , X ″ c _env By reducing the acceleration command X "in the translation direction. ^ref Ask for.
That is, the control means 11 calculates the translation direction acceleration command X ″ according to the following equation (8). ^ref Is output. Thereby, the compliance operation can be realized.
[0026]
[Formula 6]

[0027]
Position signal X indicating the current position and speed ^cmd , Speed signal X ' ^cmd Are sequentially updated each time the wheelchair moves according to the operation command of the joystick 2 and the position and speed of the wheelchair are updated.
In this embodiment, the position signal X ^cmd , Speed signal X ' ^cmd The current position and speed are given as follows. When the person does not operate the joystick 2, the wheelchair 1 is stopped. ^cmd , Speed signal X ' ^cmd By giving a signal that changes along a predetermined trajectory as follows, the wheelchair can be moved along the trajectory, and the wheelchair can be displaced from the trajectory according to the operation of the joystick 2.
In the above, F in the translation direction _hum , F _env Although the components have been described, the control means 21 also performs the rotation direction acceleration command φ ″ in the same manner as described above. ^ref Is output. In the case of the control means 21, X in the above equation (8) may be replaced with φ.
[0028]
In the present embodiment, as described above, since compliance control is applied to human input from the joystick 2 and external force from the environment estimated by the reaction force estimation means 7, position control, rotation Position / rotation direction response characteristics with respect to direction commands, position / rotation direction response characteristics with respect to human force, and position / rotation direction response characteristics with respect to external forces from the environment can be designed independently.
Therefore, it is possible to achieve desired compliance characteristics while maintaining position / rotation direction control performance.
It is also possible to design the compliance characteristics independently for the environmental modes in the translational and rotational directions. For example, when one wheel of a wheelchair has tripped over a pebble, the translation direction can be softened and the rotation direction can be controlled tightly so that it can get over the stone.
Therefore, by changing the compliance characteristic, it is possible to change the running characteristic of the wheelchair so that it can be easily used by a person according to the purpose.
[0029]
(5) Coordinate conversion means 16, 26, coordinate conversion means 4, and θ ″ → I conversion means 5
The wheelchair 1 is represented by a model shown in FIG. 10, and parameters of the wheelchair 1 are defined as follows.
x ₀ : X coordinate of the reference point in the world coordinate system
y ₀ : Y coordinate of the reference point in the world coordinate system
φ: Wheelchair direction angle
R: Radius of drive wheel
W: Tread
θr ': Right wheel rotation angle
θl ′: Left wheel rotation angle
Midpoint Po (x of drive wheel 1b of wheelchair 1 above ₀ , Y ₀ ) Is a control reference point, the position and posture of the wheelchair 1 are given by the following equation (9).
The kinematics of the wheelchair 1 is given by the following equation (10), and the following equation (12) is obtained by performing time differentiation on the equation (11).
Thus, the acceleration command X ″ in the world coordinate system shown in the following equation (12) ^ref , Acceleration command φ ” ^ref From the angular acceleration command θ ” ^ref Can be requested.
[0030]
[Expression 7]

[0031]
The coordinate conversion means 16, 26 is the rotational acceleration θ ′ of the wheelchair 1. ^res By the above equation (10), the speed feedback signal X ′ ^res , Φ ^'res Convert to
The coordinate conversion means 4 is an acceleration command X ″ output from the control means 11, 21. ^ref , Φ " ^ref Rotation acceleration command θ ”for both wheels of the wheelchair according to the above equation (12) ^ref Convert to
[0032]
Further, the position of the center of gravity of the wheelchair is P shown in FIG. ₀ , The kinetic energy function K is defined by equation (13).
Thus, by solving the Lagrange equation, the dynamic relational expression is defined as the following equations (14) and (15). Here, Mθ is an equivalent inertia matrix.
[0033]
[Equation 8]

[0034]
The θ ″ → I conversion means 5 calculates the rotational acceleration command θ ″ according to the above equation (14). ^ref Is converted into a torque command, and the torque command is converted into a current command I for the motor that drives both wheels of the wheelchair. ^ref Convert to
[0035]
(6) Adaptive control means 18
As described above, the adaptive control means 18 performs F _env The frictional force of the running surface is obtained from the low frequency component of the motor, and according to the frictional force, the force command F in the operation amount → force converting means 3 _hum Conversion coefficient [K in the above equation (1) _F 〕change.
As shown in FIG. 4, the external force F applied to the wheelchair 1 from the environment estimated by the reaction force estimation means 7. _env Is input to the low-pass filter 17.
F _env Among them, the low-frequency component output by the low-pass filter 17 is a steady disturbance, and can be regarded as a component due to friction between the wheelchair 1 and the floor.
This friction force is used as the wheelchair speed X ' ^res Dividing by the environmental friction coefficient μ _env Can be identified.
Environmental friction coefficient μ _env Includes all of the passenger's weight and friction on the running surface. It is therefore possible to adapt to these variations.
And this friction coefficient μ _env The desired environmental friction coefficient μ ₀ Divided by the operation amount → the coefficient K in the force conversion means 3 _F Ask for.
Thus, the coefficient K _F By making the variable, desired operation characteristics can be obtained even when the conditions of the passenger and the running surface change. That is, the assist ratio can be made variable according to the passenger and the situation of the running surface, and a person can obtain the same operation amount even when the situation of the running surface changes by only applying a certain force.
[0036]
As described above, in the present embodiment, the reaction force from the environment is estimated by the reaction force estimation means, and the estimated reaction force is detected by the high-pass filters 14 and 24 and the low-pass filter 17 as a high-frequency component and a low-frequency component. Since the high-frequency component is a component due to collision with an obstacle, compliance control is performed to reduce the impact force, and the low-frequency component is the friction force between the wheel and the running surface, so the friction force Accordingly, the coefficient K in the operation amount → force converting means 3 as described above. _F It is changing.
For this reason, soft control can be performed for a sudden reaction force caused by a collision with an obstacle, etc., and an appropriate assist ratio should be set for a change in the friction coefficient of the running surface. Can do.
That is, even when the road surface becomes a jari road or the weight of the wheelchair increases, a human can obtain the same speed by applying the same force. Therefore, a significant improvement in operability can be expected. In addition, when an operation is accidentally collided with an obstacle, the impact force is alleviated, so that safety can be improved.
[0037]
In addition, although the case where a joystick was used was demonstrated in the said Example, as shown in FIG. 11, the two torque sensors 31a and 31b are attached to the wheelchair 1, and the output of torque sensor 31a, 31b is made into the force of a translation direction. You may comprise so that it may convert into the force of a rotation direction.
If comprised in this way, the wheelchair 1 can be moved to arbitrary directions according to the force which a person applies to the said torque sensors 31a and 31b similarly to the case where the said joystick is used.
Further, in an electric cart, a golf cart or the like that is advanced while a handle is pushed by a person, the torque sensors 31a and 31b are attached to the handle portion, and the electric cart or the golf cart or the like is driven by the control device shown in this embodiment. If comprised in this way, it will become possible to move an electric trolley | bogie, a golf cart, etc. to arbitrary directions with comparatively small force.
When using the torque sensors 31a and 31b, the manipulated variable → force converting means 3 is calculated by the following formula (16). ^H , Torque N in rotation direction ^H Convert to
Even when the torque sensors 31a and 31b are used, when the value of the coefficient in the equation (16) is changed by the output of the adaptive control means 18, the situation of the occupant and the running surface changes as in the case of using the joystick. In addition, desired operation characteristics can be obtained.
[0038]
[Equation 9]

[0039]
In order to verify the effect of the present invention, an experiment was performed in which a human operated the wheelchair 1 of the present example and pushed an object on the way.
The gain K _F 12 to 14 show the experimental results when the value is constant. In this experiment, the position of the wheelchair (Fig. 12), virtual force command value (Fig. 13), and environment through a low-pass filter when the joystick is fully pushed down and applied to an object on the way (external force increased) The horizontal force in FIGS. 12 to 14 is time, and the arrow indicates the time when it hits an object.
As shown in FIGS. 12 to 14, the gain K _F Even when is constant, the compliance control is performed on the external force from the environment, so that the impact force at the time of contact with the object can be relaxed and stable contact with the object can be achieved.
However, the gain K _F Therefore, it can be seen that the speed of the wheelchair becomes slow and stops due to the increase of the frictional force with the floor by pushing the object.
[0040]
The adaptive control means 18 is provided to provide the gain K _F FIGS. 15 to 17 show the experimental results when V is made variable according to the environment.
As in FIGS. 12 to 14, the position of the wheelchair (FIG. 15), virtual force command value (FIG. 16), low-pass filter when the joystick is fully pushed down and applied to an object on the way (external force increased) The external force from the environment (Fig. 17) was obtained, the horizontal axis is time, and the arrow indicates the point when it hits an object.
When the adaptive control means 18 is provided, as shown in FIGS. 15 to 17, even when the environmental friction force increases, the gain K _F It can be seen that the desired operating characteristics can be realized without being affected by the frictional force of the object.
Further, a large static friction works when the wheelchair starts to move, but in this example, it is possible to start smoothly.
Through the above experiment, the friction coefficient of the environment is identified, and the gain K according to the value is determined. _F It was confirmed that the desired operating characteristics can be realized by adapting to changes in the environment by making the variable.
[0041]
【The invention's effect】
As described above, the following effects can be obtained in the present invention.
(1) An external force applied to the traveling vehicle by a reaction force estimation means or the like by converting the output of a joystick for operating the traveling vehicle or a torque sensor into a virtual translation direction force command and a rotation direction force command. Because compliance control is performed for human input from the joystick and torque sensor and external force from the environment, the external force from the environment changes suddenly when it collides with an obstacle. The generated impact can be mitigated and soft control can be performed against external forces from the environment.
(2) In particular, since compliance control is performed for human input and external force from the above environment, position responsiveness to human force and position responsiveness to external force can be designed independently. It is possible to ensure desired compliance characteristics against external force while maintaining the position control performance.
(3) The frictional force between the traveling vehicle and the traveling surface is obtained from the low frequency component of the external force applied to the traveling vehicle, and the gain of the joystick or torque sensor is changed by a value corresponding to the frictional force. Thus, even if the variation in the weight of the traveling vehicle or the state of the traveling surface changes, desired operation characteristics can be obtained.
[Brief description of the drawings]
FIG. 1 is a diagram showing an overall configuration of a control device according to an embodiment of the present invention.
FIG. 2 is a diagram showing a conceptual configuration of an electric wheelchair to which the present embodiment is applied.
FIG. 3 is a diagram modeling a joystick.
FIG. 4 is a detailed block diagram of a translation direction control system.
FIG. 5 is a detailed block diagram of a rotation direction control system.
FIG. 6 is a block diagram of disturbance estimation means.
FIG. 7 is a block diagram of reaction force estimation means.
FIG. 8 is a block diagram of conversion from human input to a position command.
FIG. 9 is a block diagram of conversion from an external force to a position command.
FIG. 10 is a diagram modeling a wheelchair.
FIG. 11 is a diagram illustrating input by a torque sensor.
FIG. 12 is a diagram showing a change in position of a wheelchair when the gain is made constant.
FIG. 13 is a diagram showing a change in a virtual force command value when the gain is made constant.
FIG. 14 is a diagram showing fluctuations in external force through a low-pass filter when the gain is constant.
FIG. 15 is a diagram showing a change in position of a wheelchair when a variable gain is set.
FIG. 16 is a diagram showing a change in a virtual force command value when a variable gain is used.
FIG. 17 is a diagram showing fluctuations in external force through a low-pass filter when a variable gain is used.
[Explanation of symbols]
1 Electric wheelchair
2 Joystick
3 Operation signal → Force conversion means
4 Coordinate conversion means
5 θ ”→ I conversion means
6 Disturbance estimation means
7 Reaction force estimation means
10 Translation direction control system
11 Control means
12, 13 Compliance control means
14 High-pass filter
15,16 Coordinate conversion means
17 Low-pass filter
18 Adaptive control means
20 Rotation direction control system
21 Control means
22, 23 Compliance control means
24 High-pass filter
25, 26 Coordinate conversion means
31a, 31b Torque sensor

Claims (5)

A method for controlling a traveling vehicle that feeds back at least a position signal of the traveling vehicle and drives the driving means to move the traveling vehicle based on a deviation from a command value,
Convert the operation amount into force command,
The force command is input to the first secondary delay element to obtain a command signal,
Further, the external force applied to the traveling vehicle is obtained,
The external force is input to the second secondary delay element to obtain a feedback signal of the external force,
Based on the feedback signal from the traveling vehicle, the command signal output by the first second-order lag element, and the feedback signal of the external force output by the second second-order lag element, a driving signal for the traveling vehicle is obtained. A method for controlling a traveling vehicle, wherein the traveling of the traveling vehicle is controlled by the driving means. The external force applied to the traveling vehicle is input to a low-pass filter that cuts high-frequency components, and based on the output of the low-pass filter, the frictional force between the traveling vehicle and the traveling surface is obtained,
2. The traveling vehicle control method according to claim 1, wherein a conversion coefficient into the force signal is changed in accordance with the frictional force. 3. The method for controlling a traveling vehicle according to claim 1, wherein an external force applied to the traveling vehicle is input to the second second-order lag element through a high-pass filter that cuts a low frequency component. A traveling vehicle control device that feeds back at least a position signal of a traveling vehicle, drives a driving unit based on a deviation from a command value, and moves the traveling vehicle,
Means for converting the manipulated variable into a force command;
First compliance control means for inputting the force command to a first secondary delay element and obtaining a command signal;
Means for obtaining an external force applied to the traveling vehicle;
A high-pass filter that cuts low frequency components of the external force,
A second compliance control means for inputting the output of the high-pass filter into a second second-order lag element and obtaining a feedback signal of an external force;
Control means for controlling the driving means based on a feedback signal from the traveling vehicle, a command signal output by the first second-order lag element, and a feedback signal of external force output by the second second-order lag element; A traveling vehicle control device comprising: Means for inputting an external force applied to the traveling vehicle to a low-pass filter that cuts a high-frequency component, and obtaining a frictional force between the traveling vehicle and the traveling surface based on an output of the low-pass filter;
5. The traveling vehicle control device according to claim 4, further comprising means for changing a conversion coefficient into the force signal by a value corresponding to the frictional force.

Images