JP3584669B2 - Drive control device for electric vehicles - Google Patents

Description

切換制御部５００にて採用されている制御切換信号生成論理を、表１に示す。この表に示すように、切換制御部５００は、右後輪１０ＲＲの角加速度ｄωＲ／ｄｔ及び左後輪１０ＲＬの角加速度ｄωＬ／ｄｔのいずれかでも所定のしきい値ＴＨ以上になったら、少なくとも左右いずれかの駆動輪にスリップ又はその傾向が生じていると見なし、ＴＲＣ／ＡＢＳ相当制御部２００にて補正されたトルク指令ＴＲ及びＴＬがモータ制御部３６Ｒ及び３６Ｌに出力されるよう、制御切換信号を発生させる。角加速度ｄωＲ／ｄｔ及びｄωＬ／ｄｔがいずれもしきい値ＴＨを下回っているときには、切換制御部５００は、走行安定制御部３００にて生成されたトルク指令ＴＲ及びＴＬがモータ制御部３６Ｒ及び３６Ｌに出力されるよう、制御切換信号を発生させる。本実施形態では、従って、在来エンジン車におけるＴＲＣやＡＢＳに相当する機能を、車輪速ＶＲＲ及びＶＲＬの検出からモータ制御部３６Ｒ及び３６Ｌへのトルク指令出力に至る電気情報処理にて実現できるため、油圧を制御する機構等は不要である。
【００３０】
上述の制御切換論理を実現するには、例えば、角加速度ｄωＲ／ｄｔ及びｄωＬ／ｄｔを検出及び判定する手段、並びにこの判定の結果に基づき且つ表１の論理に従い制御切換信号を発生させる手段があればよい。これらの手段を備えた切換制御部５００の一例を、図８に示す。この図では、車輪加速度演算部５０６は車輪速演算部１２２にて検出された右後輪１０ＲＲの車輪速ＶＲＲを微分することにより右後輪１０ＲＲの加速度ｄＶＲＲ／ｄｔを求め、角加速度演算部５１０は加速度ｄＶＲＲ／ｄｔを車輪半径Ｒにて除すことにより右後輪１０ＲＲの角加速度ｄωＲ／ｄｔを求め、スリップ判定部５１４は角加速度ｄωＲ／ｄｔをしきい値ＴＨと比較している。同様に、車輪加速度演算部５０８は車輪速演算部１２４にて検出された車輪速ＶＲＬを微分することにより左後輪１０ＲＬの加速度ｄＶＲＬ／ｄｔを求め、角加速度演算部５１２は加速度ｄＶＲＬ／ｄｔを車輪半径Ｒにて除すことにより左後輪１０ＲＬの角加速度ｄωＬ／ｄｔを求め、スリップ判定部５１６は角加速度ｄωＬ／ｄｔをしきい値ＴＨと比較している。制御動作選択部５１８は、スリップ判定部５１４及び５１６における判定結果に基づきかつ表１の論理に従い、制御切換信号を生成する。このような構成により、前述の論理に従う制御切換が実現される。
【００３１】
（５）ＴＲＣ／ＡＢＳ相当制御部の機能
【表２】

ＴＲＣ／ＡＢＳ相当制御部２００は、角加速度ｄωＲ／ｄｔ及びｄωＬ／ｄｔに基づき右後輪１０ＲＲに係るフィードバックトルクΔＴＲ及び左後輪１０ＲＬに係るフィードバックトルクΔＴＬを表２の論理に従い求め、仮確定されたトルク指令ＴＲからフィードバックトルクΔＴＲを、また仮確定されたトルク指令ＴＬからフィードバックトルクΔＴＬを減ずる。表２の左半分に現れているしきい値ＴＨω１〜ＴＨω３や、右半分に現れているゲインＧ１及びＧ２並びに補正項Ｓ１及びＳ２は、モータ１２Ｒ及び１２Ｌの状態に応じ且つ特性に従い、ＴＲＣ／ＡＢＳ相当制御部２００が設定する。しきい値ＴＨω１〜ＴＨω３、ゲインＧ１及びＧ２並びに補正項Ｓ１及びＳ２をモータ１２Ｒ及び１２Ｌの状態及び特性に従属させることにより、本実施形態においては、モータ１２Ｒ及び１２Ｌの出力トルクが広範囲に亘って変化する例えば市街地走行向けの車両であっても、好適なタイミングでＴＲＣ／ＡＢＳ相当制御を起動できまたＴＲＣ／ＡＢＳ相当制御に適するトルクを発生させることができる。即ち、特開平２―２９９４０２号公報や特開平３―２７７０１号公報に記載のフォークリフト等に比べ、スリップを好適に抑制乃至解消可能な電気自動車が得られる。なお、ＴＲＣ／ＡＢＳ相当制御部２００の機能、作用及び効果に関しては、本願と同じ出願人に係る特開平８―１８２１１９号公報や、特願平９−８６９３号をも参照されたい。
【００３２】
図９に、ＴＲＣ／ＡＢＳ相当制御部２００の一例構成を示す。図中、右の角加速度判定部２０２は、角加速度ｄωＲ／ｄｔをしきい値ＴＨω１〜ＴＨω３（但しＴＨω１＞ＴＨω２＞０＞ＴＨω３）と比較することにより、角加速度ｄωＲ／ｄｔが、ＴＨω１超、ＴＨω１以下ＴＨω２超、ＴＨω２以下ＴＨω３超、及びＴＨω３以下のうち、どの領域に属しているかを判定する。フィードバックトルク演算部２０４は角加速度ｄωＲ／ｄｔがＴＨω１超の領域に属しているときに、フィードバックトルク演算部２０６は角加速度ｄωＲ／ｄｔがＴＨω１以下ＴＨω２超の領域に属しているときに、フィードバックトルク演算部２０８は角加速度ｄωＲ／ｄｔがＴＨω２以下ＴＨω３超の領域に属しているときに、フィードバックトルク演算部２１０は角加速度ｄωＲ／ｄｔがＴＨω３以下の領域に属しているときに、それぞれ、右後輪１０ＲＲに係るフィードバックトルクΔＴＲを演算する。フィードバックトルク演算部２０４〜２１０にて使用される演算式は、それぞれ、表２の右半分の上から１番目、２番目、３番目及び４番目の式である（符号同順）。加算器２１２は、仮確定されたトルク指令ＴＲからフィードバックトルクΔＴＲを減ずることにより、補正されたトルク指令ＴＲを生成し、これをモータ制御部３６Ｒに出力する。
【００３３】
同様に、図中、左の角加速度判定部２１４は、角加速度ｄωＬ／ｄｔをしきい値ＴＨω１〜ＴＨω３と比較することにより、角加速度ｄωＬ／ｄｔが、ＴＨω１超、ＴＨω１以下ＴＨω２超、ＴＨω２以下ＴＨω３超、及びＴＨω３以下のうち、どの領域に属しているかを判定する。フィードバックトルク演算部２１６は角加速度ｄωＬ／ｄｔがＴＨω１超の領域に属しているときに、フィードバックトルク演算部２１８は角加速度ｄωＬ／ｄｔがＴＨω１以下ＴＨω２超の領域に属しているときに、フィードバックトルク演算部２２０は角加速度ｄωＬ／ｄｔがＴＨω２以下ＴＨω３超の領域に属しているときに、フィードバックトルク演算部２２２は角加速度ｄωＬ／ｄｔがＴＨω３以下の領域に属しているときに、それぞれ、左後輪１０ＲＬに係るフィードバックトルクΔＴＬを演算する。フィードバックトルク演算部２１６〜２２２にて使用される演算式は、それぞれ、表２の右半分の上から５番目、６番目、７番目及び８番目の式である（符号同順）。加算器２２４は、仮確定されたトルク指令ＴＬからフィードバックトルクΔＴＬを減ずることにより、補正されたトルク指令ＴＬを生成し、これをモータ制御部３６Ｌに出力する。
【００３４】
しきい値ＴＨω１〜ＴＨω３、ゲインＧ１及びＧ２並びに補正項Ｓ１及びＳ２を決定しているのは、回転数判定部２２６及び２２８、係数決定部２３０及び２３２並びにしきい値演算部２３４及び２３６である。また、その際には係数決定用マップ２３８が使用される。これらのうち回転数判定部２２６は、回転数演算部１３０にて演算されたモータ１２Ｒの回転数ＮＲをベース回転数ＮＢと比較する。ここでいうベース回転数ＮＢは、図６、図７及び図１０に示されるように、モータ１２Ｒ（及び１２Ｌ）のトルクが定トルク線により制約される定トルク領域と、モータ１２Ｒ（及び１２Ｌ）のトルクが定パワー線により制約される定パワー領域との境を示す回転数である。係数決定部２３０は、アクセルがオンしているかオフしているかに関する情報をオン／オフ判定部１０２から入力すると共に、回転数ＮＲがベース回転数ＮＢを上回っているか否かに関する情報を回転数判定部２２６から入力し、これらの入力に基づき、現在のモータ１２Ｒの動作点が図１０中の領域１〜４のうちどの領域に属しているかを判定する。領域１は定トルク力行領域、２は定パワー力行領域、３は定トルク回生領域、４は定パワー回生領域である。同様に、回転数判定部２２８は、回転数演算部１３２にて演算されたモータ１２Ｌの回転数ＮＬをベース回転数ＮＢと比較する。係数決定部２３２は、アクセルがオンしているかオフしているかに関する情報をオン／オフ判定部１０２から入力すると共に、回転数ＮＬがベース回転数ＮＢを上回っているか否かに関する情報を回転数判定部２２８から入力し、これらの入力に基づき、現在のモータ１２Ｌの動作点が領域１〜４のうちどの領域に属しているかを判定する。
【００３５】
係数決定部２３０は、モータ１２Ｒの動作点が現在所属している領域にて用いるべき係数値を、係数決定用マップ２３８から獲得する。係数決定部２３２は、モータ１２Ｌの動作点が現在所属している領域にて用いるべき係数値を、係数決定用マップ２３８から獲得する。係数決定用マップ２３８は、各種の係数をモータ動作点の所属領域に対応づけて記憶している。より具体的には、係数決定用マップ２３８上では、図１０に示されるように、しきい値ＴＨω１〜ＴＨω３を決定するための係数ａ１〜ａ３、ｂ１〜ｂ３、ｃ１〜ｃ３及びｄ１〜ｄ３並びにフィードバックゲインＧ１及びＧ２を決定するための係数Ａ１、Ａ２、Ｂ１、Ｂ２、Ｃ１、Ｃ２、Ｄ１及びＤ２が、各領域毎に一般に異なる値に設定される。また、補正項Ｓ１及びＳ２は力行か回生かに応じ一般に異なる値に設定される。右のしきい値演算部２３４は、決定された係数を用いて
【数２】
（アクセルオン時）
ＴＨω１＝ａ１×ｅｘｐ（ｂ１×ＶＡ＋ｃ１×ＶＲＲ＋ｄ１×ＶＡ×ＶＲＲ）
ＴＨω２＝ａ２×ｅｘｐ（ｂ２×ＶＡ＋ｃ２×ＶＲＲ＋ｄ２×ＶＡ×ＶＲＲ）
ＴＨω３＝ａ３×ｅｘｐ（ｂ３×ＶＡ＋ｃ３×ＶＲＲ＋ｄ３×ＶＡ×ＶＲＲ）
Ｇ１ ＝Ａ１×ｅｘｐ（Ｂ１×ＶＡ＋Ｃ１×ＶＲＲ＋Ｄ１×ＶＡ×ＶＲＲ）
Ｇ２ ＝Ａ２×ｅｘｐ（Ｂ２×ＶＡ＋Ｃ２×ＶＲＲ＋Ｄ２×ＶＡ×ＶＲＲ）
（アクセルオフ時）
ＴＨω１＝ａ１×ｅｘｐ（ｂ１×ＦＢ＋ｃ１×ＶＲＲ＋ｄ１×ＦＢ×ＶＲＲ）
ＴＨω２＝ａ２×ｅｘｐ（ｂ２×ＦＢ＋ｃ２×ＶＲＲ＋ｄ２×ＦＢ×ＶＲＲ）
ＴＨω３＝ａ３×ｅｘｐ（ｂ３×ＦＢ＋ｃ３×ＶＲＲ＋ｄ３×ＦＢ×ＶＲＲ）
Ｇ１ ＝Ａ１×ｅｘｐ（Ｂ１×ＦＢ＋Ｃ１×ＶＲＲ＋Ｄ１×ＦＢ×ＶＲＲ）
Ｇ２ ＝Ａ２×ｅｘｐ（Ｂ２×ＦＢ＋Ｃ２×ＶＲＲ＋Ｄ２×ＦＢ×ＶＲＲ）
の演算を実行することにより、しきい値ＴＨω１〜ＴＨω３及びフィードバックゲインＧ１及びＧ２を決定する。同様に、左のしきい値演算部２３６は、次の式中のＶＲＲに代えてＶＲＬを用いた式にて、しきい値ＴＨω１〜ＴＨω３及びフィードバックゲインＧ１及びＧ２を決定する。尚、ＶＲＲは車輪速演算部１２２から、ＶＲＬは車輪速演算部１２４から、ＶＡはアクセル開度演算部１０４から、ＦＢはブレーキ力演算部１０６から入力する。
【００３６】
このように、モータ１２Ｒ及び１２Ｌの特性に従い決定したしきい値、フィードバックゲイン及び補正項を用いてフィードバックトルクΔＴＲ及びΔＴＬを決定しているため、モータ１２Ｒ及び１２Ｌの特性を広い領域に亘って使用する用途でも、モータ１２Ｒ及び１２Ｌの状態や車両操縦者の要求に応じた適切なフィードバックトルクΔＴＲ及びΔＴＬを発生させることができる。また、このフィードバックトルクΔＴＲ及びΔＴＬを用いて補正したトルク指令ＴＲ及びＴＬを、角加速度ｄωＲ／ｄｔ及びｄωＬ／ｄｔのうち少なくとも一方がしきい値ＴＨ以上であるときにモータ制御部３６Ｒ及び３６Ｌに出力するようにしているため、スリップ又はその傾向を好適に防止又は抑制できる。なお、本実施形態では左右のモータ１２Ｒ及び１２Ｌが同一特性であることを仮定しているため、左右で係数決定用マップ２３８を共用しているが、同一特性でない場合等には、別々のマップとしてもよい。
【００３７】
なお、本実施形態では、角加速度ｄωＲ／ｄｔ及びｄωＬ／ｄｔをしきい値判定しているが、スリップ率＝｜（駆動輪の車輪速−車体速）／（駆動輪の車輪速）｜をしきい値判定してもよい。但し、ＴＲＣ／ＡＢＳ相当制御への切換を実行する実施形態においては、ＴＲＣ／ＡＢＳ相当制御において駆動輪の角加速度を使用するから、演算負担軽減等の点では駆動輪の角加速度の方が好ましい。なお、車輪速は、左右の従動輪の車輪速を検出しその平均値を求めることにより得ることができる。
【００３８】
（６）走行安定制御部及び学習管理部の機能
図１１に、走行安定制御部３００及び学習管理部６００の機能構成を示す。走行安定制御部３００は、車両走行中に車体に走行不安定性が発生しないようあるいは走行不安定性が軽減されるよう、力行トルク演算部１１２又は回生トルク演算部１１８にて求められたトルク指令Ｔに基づき、モータ制御部３６Ｒ及び３６Ｌに供給すべきトルク指令ＴＲ及びＴＬを生成する機能を有している。
【００３９】
かかる機能を実現すべく、走行安定制御部３００は、上述のようにトルク指令Ｔを入力する一方で、ヨーレイトセンサ５０、横加速度センサ５４及び舵角センサ４８の出力や、車体速演算部１３４にて求められた車体速ＶＳを入力する。走行安定制御部３００は、ヨーレイトセンサ５０の出力信号をヨーレイトγｔを示すデータに変換するヨーレイト演算部３０２、横加速度センサ５４の出力信号を横加速度Ｇｙを示すデータに変換する横加速度演算部３０４、及び舵角センサ４８の出力信号を舵角δｔを示すデータに変換する舵角演算部３０６を有している。更に、横加速度Ｇｙ、ヨーレイトγｔ及び車体速ＶＳに基づき次の式
【数３】
ｄβ／ｄｔ＝Ｇｙ／ＶＳ−γｔ
に基づきすべり角速度ｄβ／ｄｔを求めるすべり角速度演算部３０８や、すべり角速度演算部３０８にて得られたすべり角速度ｄβ／ｄｔを時間で積分することによりすべり角度βを求めるすべり角度演算部３１０が、設けられている。このようにして得られる各種の情報、すなわち車体の運動状態を示すヨーレイトγｔ及びすべり角度β、車両操縦者からの旋回要求を示す舵角γｔ、並びに車両操縦者からの加減速要求を示すトルク指令Ｔは、走行安定制御部３００内に設けられている制御用ニューラルネットワーク３１２及び学習管理部６００内に設けられている更新用ニューラルネットワーク６０２にそれぞれ入力される。これら、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、いずれも、車両走行開始前に予め、γｔ、β、δｔ及びＴに応じたトルク指令ＴＲ及びＴＬの生成に関し学習しており、車両走行中は、入力信号であるγｔ、β、δｔ及びＴに基づき出力信号たるＴＲ及びＴＬを生成する。
【００４０】
制御用ニューラルネットワーク３１２にて生成されたトルク指令ＴＲ及びＴＬは、原則としてそのままモータ制御部３６Ｒ及び３６Ｌに供給される。ただし、制御用ニューラルネットワーク３１２にて生成されるトルク指令ＴＲ及びＴＬがモータ１２Ｒ及び１２Ｌの出力可能範囲外となることを防ぐべく、制御用ニューラルネットワーク３１２の後段にはトルク制限演算部３１４が設けられており、走行安定制御部３００からモータ制御部３６Ｒ及び３６Ｌへのトルク指令供給はこのトルク制限演算部３１４による制限を受ける。トルク制限演算部３１４は、例えば、図１２に示す内容を有するマップを用いて実現することができる。なお、走行安定制御部３００から出力されるトルク指令ＴＲ及びＴＬに代え、前述の如く、ＴＲＣ／ＡＢＳ相当制御部２００の出力が用いられることがあることは、いうまでもない。
【００４１】
制御用ニューラルネットワーク３１２にて生成されたトルク指令ＴＲ及びＴＬは、このようにしてモータ制御部３６Ｒ及び３６Ｌに出力される一方で、学習管理部６００内に設けられている学習制御部６０４にも入力される。学習制御部６０４は、後述のように、トルク指令ＴＲ及びＴＬや、各ニューラルネットワークへの入力信号γｔ、β、δｔ及びＴに基づき、教師信号たるトルク指令ＴＲＳ及びＴＬＳを生成し、更新用ニューラルネットワーク６０２に与える。更新用ニューラルネットワーク６０２は、この教師信号により、車両走行中に、入力信号γｔ、β、δｔ及びＴに応じたトルク指令ＴＲ及びＴＬの生成に関し学習する。更新用ニューラルネットワーク６０２は、更に、この学習の結果得られるシナプス結合係数Ｗｊｉ及びＶｋｊ並びにニューロンしきい値θｊ及びλｋを制御用ニューラルネットワーク３１２に送信し、制御用ニューラルネットワーク３１２におけるシナプス結合係数及びニューロンしきい値をこれをもって更新する。
【００４２】
図１３に、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２の機能構成を概念的に示す。この図に示すように、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、それぞれ、入力層３１２ａ又は６０２ａ、中間層３１２ｂ又は６０２ｂ及び出力層３１２ｃ又は６０２ｃを有している。なお、また、この図では、ｉ番目（ｉ＝１〜４）の入力層ニューロン３１２ｄ又は６０２ｄからｊ番面（ｊ＝１〜５）の中間層ニューロン３１２ｅ又は６０２ｅに至るシナプス結合３１２ｇ又は６０２ｇの係数がＷｊｉとあらわされており、また、ｊ番目の中間層ニューロン３１２ｅ又は６０２ｅからｋ番目（ｋ＝１又は２）の出力層ニューロン３１２ｆ又は６０２ｆに至るシナプス結合３１２ｈ又は６０２ｈの係数がＶｋｊとあらわされている。
【００４３】
本実施形態では、入力信号が４種類であり出力信号が２種類であるため、入力層３１２ａ又は６０２ａに設けられているニューロン３１２ｄ又は６０２ｄの個数は４個、出力層３１２ｃ又は６０２ｃに設けられているニューロン３１２ｆ又は６０２ｆの個数は２個である。ただし、本発明を実施するに際しては、本実施形態で使用している４種類の入力信号のうちいずれかを省略したり、あるいは他の種類の入力信号を付加することも可能である。また、本実施形態では、中間層３１２ｂ又は６０２ｂの個数が一層でありかつ中間層３１２ｂ又は６０２ｂ内のニューロン３１２ｅ又は６０２ｅの個数が５個であるが、本発明を実施するに際し、中間層の層数や中間層ニューロンの個数に格別の限定を施す必要はない。更に、この図では、自己の出力を入力のひとつとするニューロンは用いられていないが、本発明を実施するに際してはそのようなニューロンを用いても構わない。
【００４４】
図１４に、車両走行開始前に制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２が図示しない装置の管理の下に車両走行開始前（出荷前等）に実行する学習の手順を示す。この図に示すように、学習に際しては、まずシナプス結合係数Ｗｊｉ及びＶｋｊや、中間層ニューロン３１２ｅ又は６０２ｅのしきい値θｊや、出力層ニューロン３１２ｆ又は６０２ｆのしきい値ηｋが、初期的に乱数設定される（７００）。ただし、この初期化に用いる乱数は、比較的小さな範囲内で発生させる。
【００４５】
初期化を行った後に、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、図示しない外部の装置から学習パターンを入力する（７０２）。ここでいう学習パターンは、入力信号γｔ、β、δｔ及びＴの組み合わせと、この入力に応じて出力層３１２ｃ又は６０２ｃから得られるべきトルク指令ＴＲ及びＴＬとを対応付けたものである。制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、ステップ７００にて初期化されたシナプス結合係数及びニューロンしきい値並びにステップ７０２にて入力した学習パターンを構成する入力信号γｔ、β、δｔ及びＴに基づき、次の式
【数４】
Ｕｊ＝ΣＷｊｉ・Ｉｉ＋θｉ
Ｈｊ＝ｆ（Ｕｊ）
の式に従い中間層ニューロン３１２ｅ及び６０２ｅの出力Ｈｊを計算する（７０４）。ただし、上の式中、Σはｉについての総和をあらわしており、ｆはシグモイド関数
【数５】
ｆ（ｘ）＝１／｛１＋ｅｘｐ（−ｘ）｝
である。制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、続いて、次の式
【数６】
Ｓｋ＝ΣＶｋｊ・Ｈｊ＋ηｋ
Ｏｋ＝ｆ（Ｓｋ）
の式に従い、出力層ニューロン３１２ｆ又は６０２ｆの出力Ｏｋを計算する（７０６）。ただし、上の式中、Σはｊについての総和である。
【００４６】
このようにして出力層ニューロン３１２ｆ又は６０２ｆの出力Ｏｋを計算した後、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、ステップ７０２にて入力した学習パターンを構成するトルク指令ＴＲ及びＴＬを教師信号Ｔｋとして、次の式
【数７】
λｋ＝（Ｏｋ−Ｔｋ）・Ｏｋ・（１−Ｏｋ）
に従い、出力層ニューロン３１２ｆ又は６０２ｆの出力誤差λｋを計算する（７０８）。制御用ニューラルネットワーク３１２又は更新用ニューラルネットワーク６０２は、続いて、次の式
【数８】
σｊ＝Σλｋ・Ｖｋｊ・Ｈｊ・（１−Ｈｊ）
に従い、中間層ニューロン３１２ｅ又は６０２ｅの出力誤差σｊを計算する（７１０）。ただし、上の式中、Σはｋについての総和である。
【００４７】
制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、このようにして得られた出力誤差λｋ及びσｊを用いて、シナプス結合係数Ｗｊｉ及びＶｋｊ並びにニューロンしきい値θｊ及びηｋを更新する。すなわち、次の式
【数９】
Ｖｋｊ＝Ｖｋｊ＋ε・λｋ・Ｈｊ
ηｋ＝ηｋ＋φ・λｋ
に従いシナプス結合係数Ｖｋｊ及びニューロンしきい値λｋを更新し（７１２）、また、次の式
【数１０】
Ｗｊｉ＝Ｗｊｉ＋ε・σｊ・Ｉｉ
θｊ＝θｊ＋φ・σｊ
に従いシナプス結合係数Ｗｊｉ及びニューロンしきい値θｊを更新する（７１４）。ただし、ε及びφは定数である。
【００４８】
このようにしてシナプス結合係数Ｗｊｉ及びＶｋｊ並びにニューロンしきい値θｊ及びηｋを更新した後、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、ステップ７０２にて入力した学習パターンとは異なる学習パターンを選択し（７１６）、再度、ステップ７０２以降の動作を実行する。このようにして学習パターンを更新しながらステップ７０２〜７１４の動作が繰り返し実行された結果、予め準備されている全ての学習パターンに関する学習が終了するに至ると（７１８）、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、一通りの学習が終了したことを記憶し、改めて、全ての学習パターンに関する学習を実行する（７２０）。すべての学習パターンについての学習の繰り返し回数が所定の制限回数に至ると（７２２）、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２は、車両走行開始前に実行すべき学習の手順を終了する。
【００４９】
なお、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２双方に関し上述した学習手順を実行するようにしてもよいが、制御用ニューラルネットワーク３１２及び更新用ニューラルネットワーク６０２のうちいずれかのみが学習を実行し、ステップ７００〜７２２に至る学習が終了した段階で残りの１個のニューラルネットワークにその結果を転送するようにしてもよい（７２４）。このようにすれば、車両走行開始時におけるシナプス結合係数Ｗｋｊ及びＶｋｊ並びにニューロンしきい値θｊ及びηｋに関し、車両走行開始時点（出荷時点）での両ニューラルネットワーク間の相違をなくすことができる。
【００５０】
また図１４中のステップ７００〜７２２による学習の手法は、誤差逆伝搬法（バックプロパゲーション法）と呼ばれている。すなわち、ニューラルネットワークに対しある組み合わせにかかる入力信号を与えておいて得られた出力信号を、予め想定しておいた出力信号すなわち教師信号と比較し、教師信号に対する出力信号の差
【数１１】
Ｅ＝Σ（Ｔｋ−Ｏｋ）^２／２
を最小とするようシナプス結合係数やニューロンしきい値を更新していく手法である。なお、上の式中Σはｋについての総和をあらわしている。教師信号Ｔｋに対する出力信号Ｏｋの誤差Ｅをシナプス結合係数Ｖｋｊに関して最小化するには、
【数１２】

が小さくなるようＶｋｊを更新していけばよい。前述の数９は、数１２の右辺にあらわれるＨｊの係数をλｋとおいたものである。同様に、シナプス結合係数Ｗｊｉに関し誤差Ｅを最小化するには、
【数１３】

を最小化すればよい。上の式中、Σはｋに関する総和をあらわしている。前述の数８は、数１３の右辺にあらわれるＩｉの係数をσｊとおいたものである。
【００５１】
図１５〜図１７に、制御用ニューラルネットワーク３１２、更新用ニューラルネットワーク６０２及び学習制御部６０４の、車両走行中における動作の流れを示す。これらの手順は、対応する部材によって所定周期毎に繰り返し実行されるものである。また、本実施形態の特徴の一つは、次に述べるように、車両走行開始前に行った学習の結果に基づき制御用ニューラルネットワーク３１２が車両走行中にトルク指令ＴＲ及びＴＬを実行する一方で、更新用ニューラルネットワーク６０２が、学習制御部６０４の管理の下に実際の入力信号に基づく学習を行い、制御用ニューラルネットワーク３１２の組織例えばシナプス結合係数やニューロンしきい値を、学習により得られた更新用ニューラルネットワーク６０２の組織例えばシナプス結合係数やニューロンしきい値によって逐次更新していく点にある。
【００５２】
まず、図１５に示されるように、制御用ニューラルネットワーク３１２は、ヨーレイトγｔ、すべり角度β、舵角δｔ及びトルク指令Ｔを入力し（８００）、ステップ７０４及び７０６と同様の原理にて中間層ニューロン３１２ｅの出力Ｈｊや出力層ニューロン３１２ｆの出力Ｏｋを順に計算する（８０２，８０４）。制御用ニューラルネットワーク３１２は、計算により得られた出力Ｏｋすなわちトルク指令ＴＲ及びＴＬを、トルク制限演算部３１４を介しモータ制御部３６Ｒ及び３６Ｌに供給する一方で、学習制御部６０４に供給する。学習制御部６０４は、図１７に示されるように、制御用ニューラルネットワーク３１２にて生成されたトルク指令ＴＲ及びＴＬのほか、ヨーレイトγｔ、すべり角度β、舵角δｔ及びトルク指令Ｔを入力し（１０００）、トルク指令ＴＲにかかる教師信号ＴＲＳ及びトルク指令ＴＬに係る教師信号ＴＬＳを次の式
【数１４】
ＴＲＳ＝ＦＲ（γｔ，β、δｔ，Ｔ）
ＴＬＳ＝ＦＬ（γｔ，β、δｔ，Ｔ）
を発生させ（１００４）、この教師信号ＴＲＳ及びＴＬＳを更新用ニューラルネットワーク６０２に供給する（１００６）。
【００５３】
更新用ニューラルネットワーク６０２は、ヨーレイトγｔ、すべり角度β、舵角δｔ及びトルク指令Ｔに加え、この教師信号ＴＲＳ及びＴＬＳを入力する（図１６：９００）。更新用ニューラルネットワーク６０２は、ステップ７０４〜７１０と同様の手順によって、中間層ニューロン６０２ｅの出力Ｈｊ、出力層ニューロン６０２ｆの出力Ｏｋ、出力層ニューロン６０２ｆの出力誤差λｋ、中間層ニューロン６０２ｅの出力誤差σｊを、順に計算する（９０２〜９０８）。更新用ニューラルネットワーク６０２は、ステップ７１２及び７１４と同様の手順によって、シナプス結合係数Ｗｊｉ及びＶｋｊ並びにニューロンしきい値θｊ及びλｋを更新する（９１０，９１２）。更新用ニューラルネットワーク６０２は、更新後のシナプス結合係数Ｗｊｉ及びＶｋｊ並びにニューロンしきい値θｊ及びλｋ、制御用ニューラルネットワーク３１２に転送する（９１４）。制御用ニューラルネットワーク３１２は、シナプス結合係数Ｗｊｉ及びＶｋｊ並びにニューロンしきい値θｊ及びλｋが更新用ニューラルネットワーク６０２から転送されてきたとき（８０６）、転送されてきたこれらの情報に基づき、自らのシナプス結合係数Ｗｊｉ及びＶｋｊ並びにニューロンしきい値θｊ及びλｋを更新する（８０８）。
【００５４】
従って、本実施形態では、入力信号たるヨーレイトγｔ、すべり角度β、舵角δｔ及びトルク指令Ｔに基づき、制御用ニューラルネットワーク３１２を用いてトルク指令ＴＲ及びＴＬが生成され、車輪角加速度が大きくない限り（すなわちスリップ又はその傾向が生じていないとみなせる限り）このトルク指令ＴＲ及びＴＬを用いてモータ１２Ｒ及び１２Ｌの制御が行われる。また、制御用ニューラルネットワーク３１２の組織は、先に図１４に示した手順によって、車両走行開始前に形成されている。従って、図１４に示す手順を実行するに際し、実車の走行により得られたデータや数値シミュレーション（模擬演算）により得られたデータを学習パターンに設定しておけば、車体の基礎運動方程式のみでは十分に制御モデルに取り込むことができない要因にも対処することができ、従って従来に比べ制御の信頼性や精度が向上する。また、車両走行中に、制御用ニューラルネットワーク３１２にて生成されたトルク指令ＴＲ及びＴＬと実走行環境下での入力信号γｔ、φ、δｔ及びＴとに基づき学習制御部６０４が教師信号ＴＲＳ及びＴＬＳを発生させ、更新用ニューラルネットワーク６０２がこの教師信号ＴＲＳ及びＴＬＳに基づき行った学習の結果を制御用ニューラルネットワーク３１２に転送しその組織を更新するようにしたため、車両走行中に実際に発生した加減速要求や車体運動状態に応じて制御用ニューラルネットワーク３１２の組織を更新することができ、より信頼性の高い制御が実現されることになる。更に、制御用ニューラルネットワーク３１２によるトルク指令生成処理と更新用ニューラルネットワーク６０２による学習とが同時並列的に実行されるため、処理の効率化による制御の迅速化という面も生ずる。
【００５５】
さらに、本実施形態において各ニューラルネットワークへの入力信号として用いられている量は、いずれも、車体の基礎運動方程式にあらわれている量又はその量から導出できる量である。まず、車輪に左右差がなく、コーナリングフォースにも左右差がないと考えると、車体の基礎運動方程式は、
【数１５】

となる。変数名とその意味に関しては、図１８を参照されたい。他方、各車輪の回転の運動方程式は、車輪に左右差がなく、また指令したトルクがそのまま正確に実出力になると仮定すると、
【数１６】

となる。変数名とその意味に関しては、図１９を参照されたい。更に、コーナリングフォースは、
【数１７】

と表される。従って、数１６及び数１７を数１５に代入して整理すると、
【数１８】

という連立微分方程式が得られる。式中、ｓはラプラス演算子である。
【００５６】
本実施形態において入力信号として用いている量のうち、ヨーレイトγｔ、すべり角度β及び舵角δｔは、数１８に示されている連立微分方程式にあらわれている。また、トルク指令Ｔは、車体速ＶＳ及びモーメントＭの形で、同式にあらわれている。従って、γｔ、φ、δｔ及びＴを各ニューラルネットワークへの入力信号として用いている本実施形態によれば、車体の基礎運動方程式によって表現されている加減速要求や操舵要求と車体の運動状態との関係を走行安定制御に好適に反映させることができ、従来に比べ正確な制御を実現することができる。更に、制御用ニューラルネットワーク３１２を用いたトルク指令ＴＲ及びＴＬの生成に関しては、車体が旋回しているか否かといった状態の如何による適用制限は存在しない。従って、車両が直進しているときであっても、スリップ等によるすべり角度βやヨーレートγｔの変化を抑制することが可能である。
【００５７】
また、前述したように、４種類述べた入力信号のうちいずれかを省略することも可能であるし、また他の量に置き換えることも可能である。その際には、比較的感度の低いヨーレイトγｔの他に、それよりも感度が高い量を入力信号として用いるのが好ましい。すなわち、本実施形態のごとく、比較的感度の低いヨーレイトγｔの他に比較的感度の高いすべり角度βを用いるようにすれば、例えば特開平５−９１６０７号公報にて開示されているヨーレイトの目標制御に比べ、正確で信頼性の高い制御を実現することができる。更に、学習制御部６０４にて教師信号ＴＲＳ及びＴＬＳを生成する際に、制御用ニューラルネットワーク３１２や更新用ニューラルネットワーク６０２への入力信号としても用いられているヨーレイトγｔ、すべり角度β、舵角δｔ及びトルク指令Ｔを用いることにより、比較的信頼性の高い教師信号ＴＲＳ及びＴＬＳを生成することができる。例えば、本願出願人が特願平９−６８５７１号にて提案したフィードバック制御論理に従い教師信号ＴＲＳ及びＴＬＳを生成するようにすれば、走行中の学習に際しても車体の基礎運動方程式によるモデルを反映させることができるため、制御用ニューラルネットワーク３１２の出力に基づき更新用ニューラルネットワーク６０２が学習しその結果に基づき制御用ニューラルネットワーク３１２の組織を更新するという学習／組織更新のループが不安定になることはない。
【００５８】
【発明の効果】
以上説明したように、本発明の好ましい実施形態においては、車体運動をモデル化した車体の基礎運動方程式ではなく、実車試験又は模擬演算結果に基づき制御用ニューラルネットワークが予め学んでいる内容に基づきモータに対する指令を生成すると共に、加減速要求、車体運動状態、制御用ニューラルネットワーク出力等に基づき車両走行中に学習する更新用ニューラルネットワークを設け車両走行中も制御用ニューラルネットワークの組織更新を行えるようにしたため、基礎運動方程式に基づく車体運動のモデル化に内在する問題点を解消でき、また、実使用環境下における舵角の過大過小、操舵後に受ける外乱、制御系の遅れその他様々な制御誤差要因に対処できる。その結果、制御の信頼性、精度、応答性等が向上する。
【図面の簡単な説明】
【図１】本発明を実施するのに適する電気自動車のシステム構成を示すブロック図である。
【図２】ホイルインモータの構造の一例を示す断面図であり、図中右上の円内は一部拡大図である。
【図３】図１のシステムにおける制動力配分を示す図である。
【図４】車両制御部の機能構成を示すブロック図である。
【図５】トルク指令仮確定部の機能構成を示すブロック図である。
【図６】力行トルクマップの内容を示す図である。
【図７】回生トルクマップの内容を示す図である。
【図８】切換制御部の機能構成を示すブロック図である。
【図９】ＴＲＣ／ＡＢＳ相当制御部の機能構成を示すブロック図である。
【図１０】係数決定用マップの内容を示す図である。
【図１１】走行安定制御部及び学習管理部の機能構成を示すブロック図である。
【図１２】トルク制限演算部の機能を示すマップである。
【図１３】制御用及び更新用ニューラルネットワークの機能構成を概念的に示す図である。
【図１４】車両走行開始前に制御用及び更新用ニューラルネットワークが実行する学習の手順を示すフローチャートである。
【図１５】車両走行中に制御用ニューラルネットワークが実行する指令生成及び組織更新の手順を示すフローチャートである。
【図１６】車両走行中に更新用ニューラルネットワークが実行する学習及びその結果の転送の手順を示すフローチャートである。
【図１７】車両走行中に学習制御部が実行する教師信号生成手順を示すフローチャートである。
【図１８】車体の運動を既述するための所定数及び変数を定義する平面的概念図である。
【図１９】車輪の回転運動を既述するための所定数及び変数を定義する概念図である。
【符号の説明】
１０ＲＲ 右後輪、１０ＲＬ 左後輪、１２Ｒ，１２Ｌ モータ、３４Ｒ，３４Ｌ インバータ、３６Ｒ，３６Ｌ モータ制御部、３８ 車両制御部、４０ＲＲ，４０ＲＬ，４０ＦＲ，４０ＦＬ 車輪速センサ、４２ アクセルセンサ、４４ ブレーキセンサ、４６ シフトポジションスイッチ、４８ 舵角センサ、５０ ヨーレイトセンサ、５４ 横加速度センサ、１００ トルク指令仮確定部、２００ ＴＲＣ／ＡＢＳ相当制御部、３００ 走行安定制御部、３０２ ヨーレイト演算部、３０４ 横加速度演算部、３０６ 舵角演算部、３０８ すべり角速度演算部、３１０ すべり角度演算部、３１２ 制御用ニューラルネットワーク、３１４ トルク制限演算部、４００ 制御切換部、５００ 切換制御部、６００ 学習管理部、６０２ 更新用ニューラルネットワーク、６０４ 学習制御部、ＴＲ，ＴＬ，Ｔ トルク指令、φ 車体すべり角度、γｔ ヨーレイト、δｔ 舵角、Ｗｉｊ，Ｖｋｊ シナプス結合係数、θｊ，λｋ ニューロンしきい値。[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a drive control device that is mounted on an electric vehicle provided with a motor corresponding to each drive wheel and controls each motor so that the running stability of the electric vehicle is enhanced.
[0002]
[Prior art]
For conventional engine vehicles, so-called four-wheel steering (4WS) has been developed in order to achieve and improve the performance of smoothly executing acceleration / deceleration, turning, lane change, and the like, that is, running stability. However, in the current 4WS, a mechanism related to the steering and an electronic control system (or a hydraulic control system in some cases) for controlling the steering are somewhat large-scale and complicated, so that a large mounting space is required. However, there are problems such as that the speed is slightly slow and a failure is likely to occur. On the other hand, running stability is also required for an electric vehicle that is a vehicle propelled by a vehicle-mounted motor. It is conceivable to mount a 4WS for a conventional engine vehicle on an electric vehicle, but in this case, the above-described problems cannot be avoided. Also, especially in pure electric vehicles that only have a motor as the propulsion source of the vehicle, it is common to install a much larger battery than that of a conventional engine vehicle, so the mounting space problem is It becomes even more noticeable.
[0003]
Under such circumstances, various types of running stability control means or methods that can be mounted on an electric vehicle have been studied. For example, in JP-A-1-298903, JP-A-1-298904, JP-A-1-298905 and JP-A-5-176418, a plurality of motors are arranged so that the left and right drive wheels can be driven independently of each other. With respect to an electric vehicle of the type, that is, an electric vehicle with left and right driving wheels independently driven, a method of performing motor torque control or speed control in accordance with a difference in traveling locus between left and right driving wheels is disclosed. In the left and right drive wheel independent drive type electric vehicle, different torques can be generated in each of the left and right drive wheels. Therefore, an effect similar to 4WS in a conventional gasoline vehicle can be achieved by using a large-scale and complicated mechanism such as 4WS. Can be achieved without a control system. As a typical example of the type capable of performing the control described in each of the above publications, there is an electric vehicle of a type in which a vehicle driving motor is embedded in or integrated with each of left and right drive wheels, that is, a wheel-in motor type electric vehicle. . A wheel-in motor type electric vehicle has a motor for each drive wheel, so there is no need for a distribution mechanism such as a differential gear, so low transmission loss and low energy consumption can be used. There are advantages such that the motor can be made smaller than that of a normal so-called one-motor type electric vehicle because it only needs to be able to be driven, and since the motor is built in the drive wheels, the integration is high and the cabin space is widened.
[0004]
[Problems to be solved by the invention]
However, the difference between the running trajectories of the left and right driving wheels alone cannot sufficiently cope with the rotational movement of the vehicle body around the center of gravity, the movement of the vehicle body in the lateral direction, the slip of the driving wheels, and the like, and cannot achieve sufficient running stability. For example, when steering is performed for turning, various amounts of yaw rate, lateral acceleration, slip angle, and the like are generated in the vehicle body. These are quantities whose values are determined according to the vehicle speed, the wheel speed, and the like in addition to the cornering force, the lateral force, the road surface friction coefficient, and the like, and the values cannot be uniquely estimated from the travel locus difference.
[0005]
In order to solve such deficiencies, the present applicant has already proposed a technology for controlling the output (torque) of each motor so that the response of the yaw rate or the slip angle to the steering angle matches the target response (Japanese Patent Application No. Hei 10-26139). No. 9-8693). By using such a control method, that is, the target yaw rate adaptation control and the slip angle adaptation control (for example, the slip angle zero control), it becomes possible to appropriately cope with the rotational motion, the lateral motion, and the like, thereby improving the running stability of the vehicle. Further, the steering angle is an amount that is immediately generated with the steering by the vehicle operator, and since the control based on this steering angle is performed in the previous proposal, the effect of reducing the frequency of pedal operation by the vehicle operator also occurs. In other words, even if the vehicle operator steers while driving with the accelerator pedal, and as a result the traveling becomes unstable, before the vehicle operator senses it and starts to depress the brake pedal, Since the target yaw rate adaptation control and the target slip angle adaptation control are activated and the running instability is eliminated or reduced, the operation of the brake pedal by the vehicle operator becomes unnecessary, and the running with the accelerator pedal depressed becomes possible.
[0006]
The inventor has been studying to further improve the drive control device according to the prior proposal. One of the points that need to be improved as a result of this study is that if control is performed based solely on the steering angle, the steering angle will be too large or too small, disturbances after steering (road surface, cross wind, etc.), delays in the control system, etc. As a result, unexpected disturbance (control error) occurs in the vehicle body motion. Another point that needs to be improved is to model the body motion in the form of the basic body motion equation and perform target yaw rate adaptation control or target slip angle adaptation control according to the control logic derived based on this model. The control error due to the difference between the vehicle and this model cannot be eliminated, and various fluctuation factors (e.g., changes in vehicle characteristics with time, changes in loaded weight, etc.) in an actual use environment cannot be dealt with.
[0007]
A first object of the present invention is to provide a neural network for control for maintaining and improving the running stability of a vehicle, that is, running stability control in a left-right driving wheel independent driving type electric vehicle represented by a wheel-in motor type electric vehicle. By introducing it, it is possible to deal with various control error factors, including control error factors that cannot be expressed only by the basic body motion equation of the vehicle body and therefore can not be reflected in the control logic, so that the reliability and accuracy of control Is to improve. A second object of the present invention is to further improve the reliability, accuracy, responsiveness, and the like of running stability control by performing parallel and distributed processing of running stability control using a neural network and updating the organization of the neural network. It is in. A third object of the present invention is to update the organization of a neural network for running stability control based on the results learned from acceleration / deceleration requests actually generated during running and the vehicle body motion state, thereby improving control reliability. And to further improve the accuracy.
[0008]
[Means for Solving the Problems]
A preferred embodiment of the present invention is a drive control for controlling the running of an electric vehicle by individually giving commands relating to the outputs to a plurality of motors for individually driving the driving wheels of the electric vehicle. An embodiment according to the device, wherein the generation of the command according to the acceleration / deceleration request and the vehicle body motion state, the neural network for control and the neural network for updating learned before starting the vehicle traveling based on the actual vehicle test or the simulation result. And a running learning management means for providing a teacher signal to the updating neural network during running of the vehicle based on an input signal indicating an acceleration / deceleration request and a vehicle body motion state and an output of the controlling neural network. The control neural network generates and outputs a command to be individually given to each of the plurality of motors according to the input signal while the vehicle is running. The neural network for updating, based on the input signal indicating the acceleration / deceleration request and the vehicle body movement state and the output of the control neural network, i.e., the instruction to the motor, even during the running of the vehicle, receives the command corresponding to the acceleration / deceleration request and the vehicle body movement state. Learn about generating. The updating neural network updates the organization of the controlling neural network based on the result of learning. This tissue update transmits the synapse coupling coefficient and the neuron threshold value in the updating neural network to the control neural network, and thereby updates / replaces the synaptic coupling coefficient and the neuron threshold value in the control neural network at least partially. For example, it can be executed even while the vehicle is running.
[0009]
As described above, in the present embodiment, a command for the motor is generated based on the content learned in advance by the control neural network based on the result of the actual vehicle test or the simulation operation instead of the basic vehicle body motion equation modeling the vehicle body motion. Therefore, no control error occurs due to the difference that would occur between the model and the actual vehicle motion when the vehicle motion is modeled. In addition, since the control neural network can be renewed even while the vehicle is running, it is possible to quickly and accurately cope with various factors such as an excessively small or large steering angle, a disturbance received after steering, a control system delay, and the like in an actual use environment. . Therefore, control reliability, accuracy, responsiveness, and the like are improved.
[0010]
The present invention can be expressed not only as a "drive control device" but also as a "drive control method", "neutral network use method", "electric vehicle", and the like. Such changes to the wording can be easily made by those who refer to the disclosure of the present application.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.
[0012]
(1) System configuration
FIG. 1 shows a system configuration of an electric vehicle suitable for carrying out the present invention. This electric vehicle is a rear-wheel drive, and the right rear wheel 10RR and the left rear wheel 10RL are drive wheels, and the right front wheel 10FR and the left front wheel 10FL are driven wheels. However, the present invention can be applied to a front wheel drive vehicle and a four wheel drive vehicle.
[0013]
The electric vehicle in FIG. 1 is a wheel-in motor type. That is, the right vehicle running motor 12R in the figure is incorporated in the right rear wheel 10RR, and the left vehicle running motor 12L is incorporated in the left rear wheel 10RL. FIG. 2 shows an example of a method of assembling using the right rear wheel 10RR as an example. In this figure, the rotor 18 is fixed inside the wheel 16 so as to be able to rotate integrally with the tire 14, while the stator 24 is fixed to the rear axle 22 via the motor shaft 20, and the stator 24 is connected to the rotor 18 via bearings or the like. Has been joined. Further, while a rotor magnet (permanent magnet) 26 is fixed to the inner wall surface of the rotor 18 so as to face the stator 24 with a small gap, a stator winding 28 is wound around the stator 24, and a current is applied to the stator winding 28. Is connected to the stator winding 28 via the inside of the motor shaft 20. In such a structure, by supplying an alternating current to the stator winding 28 via the cable 30, the rotor 18 rotates and generates propulsion of the vehicle. In practicing the present invention, other types of structures may be used, or other types of structures in which motors are provided for the left and right driving wheels may be used.
[0014]
The battery 32 shown in FIG. 1 is a drive power supply source for the motors 12R and 12L, and the discharge output thereof is supplied to the motor 12R via the inverter 34R and to the motor 12L via the inverter 34L. . Inverters 34R and 34L are a type of power converter. That is, the inverter 34R converts the discharge output (DC) of the battery 32 into a power format (three-phase AC in this figure) suitable for the motor 12R under the control of the motor control unit 36R, and the inverter 34L controls the motor control unit 36L. Is converted into a power format suitable for the motor 12L. The motor control unit 36R controls the inverter 34R in accordance with the torque command TR from the vehicle control unit 38 to output a torque corresponding to the torque command TR from the motor 12R. Similarly, the motor control unit 36L controls the inverter 34L according to the torque command TL from the vehicle control unit 38 to output a torque corresponding to the torque command TL from the motor 12L. The motor control units 36R and 36L also have a function of insulating and separating the corresponding one of the inverters 34R and 34L from the vehicle control unit 38. In addition, the control of the inverters 34R and 34L by the motor control units 36R and 36L is based on the phase current detection values of the motors 12R and 12L obtained from a current sensor (not shown) or the motors 12R and 12L obtained from the rotor angular position and the like. This is performed based on the estimated value of each phase current. The present invention is applicable not only to a pure electric vehicle but also to a so-called hybrid vehicle.
[0015]
The vehicle control unit 38 is a control member having functions of controlling output torques of the motors 12R and 12L, monitoring and controlling the state of each component mounted on the vehicle, notifying a vehicle occupant of the vehicle state, and other functions. This can be realized mainly by software modification of the control unit (ECU). Outputs of sensors provided in various parts of the vehicle are input to the vehicle control unit 38, and the vehicle control unit 38 uses the outputs of the sensors for motor output control and vehicle state monitoring.
[0016]
For example, a wheel speed sensor 40RR (for example, a resolver) provided on a wheel (16 in FIG. 2) of the right rear wheel 10RR outputs a signal indicating the wheel speed VRR of the right rear wheel 10RR (for example, a pulse signal for each minute angular position displacement). ), And the wheel speed sensor 40RL provided on the wheel of the left rear wheel 10RL generates a signal indicating the wheel speed VRL of the left rear wheel 10RL. Similarly, the wheel speed sensor 40FR provided on the wheel of the right front wheel 10FR outputs a signal indicating the wheel speed VFR of the right front wheel 10FR, and the wheel speed sensor 40FL provided on the wheel of the left front wheel 10FL outputs the wheel speed of the left front wheel. A signal indicating VFL is generated. The accelerator sensor 42 receives a signal indicating an amount of depression of an accelerator pedal (not shown), that is, an accelerator opening VA, and the brake sensor 44 transmits a signal indicating the amount of depression of a brake pedal 56, that is, a braking force FB, to a shift position switch. Reference numeral 46 generates a signal indicating a closing range of a shift lever (not shown) (and a shift lever position within the range in an engine brake range or the like), that is, a shift position. Further, the steering angle sensor 48 outputs a signal indicating the steering angle δt that changes according to the operation of a steering wheel (not shown), the yaw rate sensor 50 outputs a signal indicating the yaw rate γt acting on the vehicle body, and the lateral acceleration sensor 54 outputs the signal indicating the yaw rate γt. Are generated, respectively, indicating the lateral acceleration Gy acting on. The outputs of these sensors are converted into data in a format that can be processed by the vehicle control unit 38 when input to the vehicle control unit 38. The vehicle control unit 38 uses the converted data to determine the torque commands TR and TL, switch control methods, and the like.
[0017]
In FIG. 1, a braking system that hydraulically brakes the front wheels and regeneratively brakes the rear wheels is used. That is, when the brake pedal 56 is depressed, the hydraulic pressure generated in the master cylinder 58 is transmitted to the left and right wheel cylinders 60R and 60L in response to the depression, and acts on the left and right brake wheels 62R and 62L to apply braking torque to the left and right front wheels. Is applied, the vehicle control unit 38 generates torque commands TR and TL relating to regeneration in accordance with the braking force (for example, the hydraulic pressure of the master cylinder 58) FB detected using the brake sensor 44. Therefore, as shown in FIG. 3, the distribution of the braking force in the vehicle of FIG. 1 is a distribution in which both the oil pressure regeneration and the braking force FB (“pedal input” on the horizontal axis) increase. As described above, since the hydraulic system and the regenerative system are separated from each other after the brake sensor 44, even if either the hydraulic pressure or the regeneration fails, the vehicle can be evacuated on the other side. In addition, since the hydraulic system is not provided with mechanisms such as valves and pumps or an electric system for driving and controlling the same, the system configuration is simplified compared to a system that shuts off the hydraulic pressure during regenerative operation, for example. Become. One of the reasons why it is not necessary to provide a hydraulic system with a mechanism such as a valve or a pump or an electric system for driving and controlling the same is as described later, by using the control of the output torque of the motors 12R and 12L. This is a characteristic configuration of the present embodiment that performs stability control.
[0018]
(2) Overall function of vehicle control unit
FIG. 4 shows a functional configuration of the vehicle control unit 38 in the present embodiment. Although the present invention can be implemented by hardware logic or software, the following description uses a block diagram for clarifying a functional configuration. The vehicle control unit 38 includes a torque command provisional determination unit 100, a TRC / ABS equivalent control unit 200, a traveling stability control unit 300, a control switching unit 400, a switching control unit 500, and a learning management unit 600. The torque command provisional determination unit 100 captures the output of the accelerator sensor 42, the brake sensor 44, the shift position switch 46, and the like as information indicating a request from the vehicle operator, and at the same time, controls the rotation speeds NR of the motors 12R and 12L to be controlled. The NL-related information is taken from the wheel speed sensors 40RR and 40RL, and the torque required to generate an output in response to a request such as acceleration or deceleration from the vehicle operator under the current motor rotation speed is calculated based on the information. The left and right driving wheels 10RR and 10RL are obtained, and the obtained torques are used as torque commands TR and TL. However, at this stage, the torque commands TR and TL are not finally determined, but are only "tentatively determined". The torque command provisional determination unit 100 further generates a torque command T (described later) determined based on the vehicle body speed VS.
[0019]
The TRC / ABS equivalent control unit 200 executes an operation having an effect equivalent to TRC / ABS in a conventional engine vehicle by correcting the temporarily determined torque command TR and TL under specific conditions. The traveling stability control unit 300 generates torque commands TR and TL based on the torque command T from the torque command provisional determination unit 100 and the output of each sensor, and generates a moment M around the center of gravity of the vehicle body by the torque command. Prevent unstable running during steering. The learning management unit 600 updates the organization of the neutral network in the running stability control unit 300 during running. Both the TRC / ABS equivalent control unit 200 and the traveling stability control unit 300 utilize the characteristic of the left and right wheel independent drive type electric vehicle that the torque commands can be individually given to the left and right motors 12R and 12L, respectively. This operation is performed. Therefore, in the present embodiment, it is possible to provide a function that contributes to the running stability of the vehicle, including a function corresponding to TRC / ABS in a conventional engine vehicle, without a hydraulic control system or a complicated electronic control system.
[0020]
The control switching unit 400 and the switching control unit 500 selectively switch the control logic according to the vehicle state or the driving state. The control switching section 400 is provided after the TRC / ABS equivalent control section 200 and the traveling stability control section 300, and the switching control section 500 supplies a control signal to the control switching section 400. That is, whether to output the torque command TR and TL provisionally determined by the torque command provisional determination unit 100 and corrected by the TRC / ABS equivalent control unit 200, or whether the torque command TR and TL generated by the traveling stability control unit 300 Whether to output TL is determined by the control switching unit 400 based on a control signal from the switching control unit 500.
[0021]
(3) Function of torque command provisional confirmation section
As shown in FIG. 5, the torque command provisional determination unit 100 includes an on / off determination unit 102 that determines whether or not the accelerator pedal is depressed based on the output of the accelerator sensor 42. When the on / off determining unit 102 determines that the accelerator is on, that is, the accelerator pedal is being depressed, the accelerator opening calculating unit 104 calculates the accelerator opening VA based on the output of the accelerator sensor 42, and the accelerator is off, that is, the accelerator is off. When it is determined that the pedal has not been depressed, the braking force calculation unit 106 calculates the braking force (stepping force) FB based on the output of the brake sensor 44. At this time, the shift position set by the shift position switch 46 is referred to by the accelerator opening calculator 104 and the brake force calculator 106, and the accelerator opening VA and the brake force FB are calculated and determined according to the shift position at that time. Is done. The calculated accelerator opening VA is supplied to powering torque calculators 108, 110 and 112, and the braking force FB is supplied to regenerative torque calculators 114, 116 and 118.
[0022]
On the other hand, the wheel speed calculator 122 converts a signal, for example, a pulse signal, from the wheel speed sensor 40RR into data indicating the wheel speed VRR, and the wheel speed calculator 124 indicates a signal from the wheel speed sensor 40RL as the wheel speed VRL. The wheel speed calculation unit 126 converts the signal from the wheel speed sensor 40FR into data indicating the wheel speed VFR, and the wheel speed calculation unit 128 converts the signal from the wheel speed sensor 40FL into the wheel speed VFL. Convert to data. The rotation speed calculation unit 130 arranged downstream of the wheel speed calculation unit 122 converts the wheel speed VRR into the rotation speed NR by multiplying by 60 / (2πR), and is arranged downstream of the wheel speed calculation unit 124. The rotation speed calculation unit 132 converts the wheel speed VRL to the rotation speed NL by multiplying by 60 / (2πR). The vehicle speed calculation unit 134 disposed after the wheel speed calculation units 126 and 128 is a vehicle speed calculation unit.
(Equation 1)
VS = (VFR + VFL) / 2
Is calculated, and the rotation speed calculation unit 136 arranged downstream of the vehicle speed calculation unit 134 converts the vehicle speed VS into the rotation speed N by multiplying by 60 / (2πR). Here, the number of rotations is expressed in units of rpm, and the wheel speed is expressed in units of m / sec. R is a wheel radius.
[0023]
Of the rotation speed information obtained in this manner, the rotation speed NR of the right motor 12R is provided to the powering torque calculation unit 108 and the regenerative torque calculation unit 114, and the rotation speed NL of the left motor 12L is provided to the powering torque calculation unit 110 and The regenerative torque calculator 116 and the average rotational speed N corresponding to the vehicle speed VS are input to the powering torque calculator 112 and the regenerative torque calculator 118, respectively.
[0024]
The right powering torque calculation unit 108 refers to the powering torque map 138 using the rotation speed NR of the right motor 12R as a key, and obtains the maximum powering torque that can be output from the right motor 12R at the rotation speed NR at that time. The torque command TR for the right motor 12R is determined by dividing the maximum powering torque by the accelerator opening VA. As shown in FIG. 6, the powering torque map 138 used here is a means for holding the rotation speed vs. maximum torque characteristic in a powering region (a region where the rotation speed> 0 and the torque> 0). By referring to the rotational speed NR as a key as described above, the maximum powering torque (point on the curve of VA = 100% in FIG. 6) at that rotational speed NR is obtained. Further, assuming that the accelerator opening VA at that time is x%, the power running torque to be output, that is, the accelerator opening VA and the rotation at that time are obtained by proportional processing of multiplying the obtained maximum power running torque by x / 100. A torque command TR (point on the curve of VA = x% in FIG. 6) corresponding to the number NR can be obtained. The powering torque calculation unit 110 on the left side also obtains a powering torque to be output, that is, a torque command TL corresponding to the accelerator opening VA and the rotational speed NL at that time in the same procedure. The powering torque calculation unit 112 also generates information in which the torque command T, that is, the vehicle speed VS, corresponding to the accelerator opening VA and the rotation speed N at that time is converted into a torque command in the same procedure.
[0025]
The right regenerative torque calculation unit 114 refers to the regenerative torque map 140 using the rotation speed NR of the right motor 12R as a key, and obtains the maximum regenerative torque that can be output from the right motor 12R at the rotation speed NR at that time. By dividing this maximum regenerative torque by the braking force FB, the torque command TR for the right motor 12R is determined. As shown in FIG. 7, the regenerative torque map 140 used here is means for holding the rotational speed vs. maximum torque characteristic in a regenerative region (region where rotational speed> 0 and torque <0). Thus, by referring to the rotational speed NR as a key, the maximum regenerative torque at that rotational speed NR (the point on the curve of FB = 100% in FIG. 7) is obtained. Further, assuming that the braking force FB at that time is x%, the regenerative torque to be output, that is, the braking force FB and the rotation speed NR at that time, are proportionally processed by multiplying the obtained maximum regenerative torque by x / 100. (The point on the curve of FB = x% in FIG. 7) that conforms to the above. The regenerative torque calculation unit 116 on the left side also obtains a regenerative torque to be output, that is, a torque command TL corresponding to the braking force FB and the rotational speed NL at that time in the same procedure. The regenerative torque calculation unit 118 also generates information obtained by converting the torque command T, that is, the vehicle speed VS, into a torque command based on the braking force FB and the rotation speed at that time.
[0026]
The powering / regeneration switching unit 120 switches between commanding the powering torque and the commanding the regenerative torque according to the output of the on / off determining unit 102. That is, whether the torque commands TR, TL, and T determined by the powering torque calculation units 108, 110, and 112 are supplied to the subsequent stage, or the torque commands TR, TL determined by the regenerative torque calculation units 114, 116, and 118. And T are supplied to the subsequent stage according to the on / off state of the accelerator. Hereinafter, for convenience of description, the torque commands TR and TL determined by the torque command temporary determination unit 100 will be referred to as “temporarily determined” torque commands TR and TL. T is not a command to the motor itself, but is referred to as a torque command for convenience.
[0027]
In FIG. 5, the left and right motoring torque map 138 and the regenerative torque map 140 are shared, but this is an example in the case where the left and right motors 12R and 12L have the same characteristics. Use individual maps with. 6 and 7, only the maximum power running torque or the maximum regenerative torque is mapped, but a number of torque curves may be mapped using the accelerator opening VA or the braking force FB as a parameter. . If such a mapping is performed, the torque command can be obtained by referring to the map based on the pair of the rotation speed and the accelerator opening or the braking force, so that the powering or the regenerative torque calculation units 108 to 118 There is no need to perform pro-rata processing. On the other hand, mapping only the number of rotations versus the maximum torque characteristic can save storage space for holding the map. Furthermore, it is simpler to execute the calculation of the torque commands TR and TL in a time-division manner by a single arithmetic unit, instead of simultaneously and concurrently using the left and right individual functions and arithmetic members. In order to adopt such a configuration, for example, a timing clock that alternately provides the execution timing of the calculation for the left driving wheel and the execution timing of the calculation for the right driving wheel is generated, and the calculation synchronized with the timing clock is performed. do it. The same applies to the calculation of various coefficients and correction amounts described later.
[0028]
(4) Function of switching control unit
Specifically, the control switching unit 400 shown in FIG. 4 outputs the torque commands TR and TL generated by the traveling stability control unit 300 to the motor control units 36R and 36L, or the torque command provisional determination unit. It switches whether to output the torque commands TR and TL tentatively determined at 100 and corrected by the TRC / ABS equivalent control unit 200. That is, the temporarily determined torque commands TR and TL are corrected by the TRC / ABS equivalent control unit 200 only when the wheel angular acceleration satisfies a predetermined condition, and output to the motor control units 36R and 36L. This switching of the control logic is executed according to a control switching signal supplied from the switching control unit 500.
[0029]
[Table 1]

Table 1 shows the control switching signal generation logic employed in the switching control section 500. As shown in this table, the switching control unit 500 determines that at least one of the angular acceleration dωR / dt of the right rear wheel 10RR and the angular acceleration dωL / dt of the left rear wheel 10RL becomes equal to or greater than the predetermined threshold value TH. Control switching is performed such that the torque command TR and TL corrected by the TRC / ABS equivalent control unit 200 are output to the motor control units 36R and 36L, assuming that a slip or a tendency is generated in one of the left and right drive wheels. Generate a signal. When the angular accelerations dωR / dt and dωL / dt are both less than the threshold value TH, the switching control unit 500 sends the torque commands TR and TL generated by the traveling stability control unit 300 to the motor control units 36R and 36L. A control switching signal is generated so as to be output. In the present embodiment, therefore, a function corresponding to TRC or ABS in a conventional engine vehicle can be realized by electrical information processing from detection of wheel speeds VRR and VRL to output of torque commands to motor control units 36R and 36L. No mechanism for controlling the oil pressure is required.
[0030]
To implement the above control switching logic, for example, means for detecting and determining the angular accelerations dωR / dt and dωL / dt, and means for generating a control switching signal based on the result of this determination and according to the logic in Table 1 are provided. I just need. FIG. 8 shows an example of the switching control section 500 provided with these means. In this figure, the wheel acceleration calculator 506 obtains the acceleration dVRR / dt of the right rear wheel 10RR by differentiating the wheel speed VRR of the right rear wheel 10RR detected by the wheel speed calculator 122, and calculates the angular acceleration calculator 510. Calculates the angular acceleration dωR / dt of the right rear wheel 10RR by dividing the acceleration dVRR / dt by the wheel radius R, and the slip determination unit 514 compares the angular acceleration dωR / dt with the threshold TH. Similarly, the wheel acceleration calculator 508 determines the acceleration dVRL / dt of the left rear wheel 10RL by differentiating the wheel speed VRL detected by the wheel speed calculator 124, and the angular acceleration calculator 512 calculates the acceleration dVRL / dt. By dividing by the wheel radius R, the angular acceleration dωL / dt of the left rear wheel 10RL is obtained, and the slip determination unit 516 compares the angular acceleration dωL / dt with the threshold TH. The control operation selecting unit 518 generates a control switching signal based on the determination results of the slip determining units 514 and 516 and according to the logic shown in Table 1. With such a configuration, control switching according to the above-described logic is realized.
[0031]
(5) Function of TRC / ABS equivalent control unit
[Table 2]

The TRC / ABS equivalent control unit 200 obtains the feedback torque ΔTR for the right rear wheel 10RR and the feedback torque ΔTL for the left rear wheel 10RL based on the angular accelerations dωR / dt and dωL / dt according to the logic of Table 2, and is temporarily determined. Then, the feedback torque ΔTR is subtracted from the obtained torque command TR, and the feedback torque ΔTL is subtracted from the temporarily determined torque command TL. The threshold values THω1 to THω3 appearing in the left half of Table 2 and the gains G1 and G2 and the correction terms S1 and S2 appearing in the right half depend on the TRC / ABS according to the state of the motors 12R and 12L and according to the characteristics. The equivalent control unit 200 sets. By making the threshold values THω1 to THω3, the gains G1 and G2, and the correction terms S1 and S2 dependent on the states and characteristics of the motors 12R and 12L, in the present embodiment, the output torques of the motors 12R and 12L are wide. Even in a vehicle that changes, for example, for traveling in an urban area, the TRC / ABS equivalent control can be started at a suitable timing, and a torque suitable for the TRC / ABS equivalent control can be generated. That is, an electric vehicle can be obtained in which the slip can be suitably suppressed or eliminated as compared with a forklift described in JP-A-2-299402 and JP-A-3-27701. For the function, operation and effect of the TRC / ABS equivalent control unit 200, refer to Japanese Patent Application Laid-Open No. 8-182119 and Japanese Patent Application No. 9-8693, which are assigned to the same applicant as the present application.
[0032]
FIG. 9 shows an example configuration of the TRC / ABS equivalent control section 200. In the figure, the right angular acceleration determination unit 202 compares the angular acceleration dωR / dt with threshold values THω1 to THω3 (though THω1>THω2>0> THω3), so that the angular acceleration dωR / dt exceeds THω1, It is determined which region belongs to THω1 or less and THω2 or more, THω2 or less and THω3 or less, and THω3 or less. The feedback torque calculation unit 204 calculates the feedback torque when the angular acceleration dωR / dt belongs to the region where THω1 or more and THω2 or less and the angular acceleration dωR / dt belongs to the region where THω2 or less and THω2. Arithmetic unit 208 determines that the angular acceleration dωR / dt belongs to the region of THω2 or less and exceeds THω3, and feedback torque arithmetic unit 210 determines that the angular acceleration dωR / dt belongs to the region of THω3 or less. The feedback torque ΔTR related to the wheel 10RR is calculated. The arithmetic expressions used in the feedback torque arithmetic units 204 to 210 are the first, second, third, and fourth expressions from the upper right of Table 2 (signs in the same order). The adder 212 generates a corrected torque command TR by subtracting the feedback torque ΔTR from the temporarily determined torque command TR, and outputs the corrected torque command TR to the motor control unit 36R.
[0033]
Similarly, in the figure, the left angular acceleration determination unit 214 compares the angular acceleration dωL / dt with the threshold values THω1 to THω3, so that the angular acceleration dωL / dt is higher than THω1, higher than THω1 and higher than THω2, and lower than THω2. It is determined which of the areas exceeds THω3 and is equal to or less than THω3. The feedback torque calculator 216 determines when the angular acceleration dωL / dt belongs to a region exceeding THω1 and the feedback torque calculator 218 determines when the angular acceleration dωL / dt belongs to a region equal to or less than THω1 and exceeds THω2. Arithmetic operation unit 220 determines that the angular acceleration dωL / dt belongs to a region of THω2 or less and exceeds THω3, and feedback torque arithmetic operation unit 222 determines that the angular acceleration dωL / dt belongs to a region of THω3 or less. The feedback torque ΔTL for the wheel 10RL is calculated. The arithmetic expressions used in the feedback torque arithmetic units 216 to 222 are the fifth, sixth, seventh, and eighth expressions from the top of the right half of Table 2 (in the same sign order). The adder 224 generates a corrected torque command TL by subtracting the feedback torque ΔTL from the temporarily determined torque command TL, and outputs the corrected torque command TL to the motor control unit 36L.
[0034]
The threshold values THω1 to THω3, the gains G1 and G2, and the correction terms S1 and S2 are determined by the rotation speed determination units 226 and 228, the coefficient determination units 230 and 232, and the threshold calculation units 234 and 236. . In this case, the coefficient determination map 238 is used. Among these, the rotation speed determination unit 226 compares the rotation speed NR of the motor 12R calculated by the rotation speed calculation unit 130 with the base rotation speed NB. As shown in FIGS. 6, 7, and 10, the base rotation speed NB is a constant torque region where the torque of the motor 12R (and 12L) is restricted by a constant torque line, and the motor 12R (and 12L). Is a rotation speed indicating a boundary with a constant power region in which the constant torque is restricted by the constant power line. The coefficient determination unit 230 receives information on whether the accelerator is on or off from the on / off determination unit 102, and determines the information on whether the rotation speed NR exceeds the base rotation speed NB. Based on these inputs from the unit 226, it is determined which of the areas 1 to 4 in FIG. 10 the current operating point of the motor 12R belongs to. Region 1 is a constant torque power running region, 2 is a constant power power running region, 3 is a constant torque regeneration region, and 4 is a constant power regeneration region. Similarly, the rotation speed determination unit 228 compares the rotation speed NL of the motor 12L calculated by the rotation speed calculation unit 132 with the base rotation speed NB. The coefficient determination unit 232 inputs information on whether the accelerator is on or off from the on / off determination unit 102, and determines the information on whether the rotation speed NL exceeds the base rotation speed NB based on the rotation speed. Input is made from the unit 228, and based on these inputs, it is determined which of the regions 1 to 4 the current operating point of the motor 12L belongs to.
[0035]
The coefficient determination unit 230 obtains a coefficient value to be used in the area to which the operating point of the motor 12R currently belongs from the coefficient determination map 238. The coefficient determination unit 232 obtains a coefficient value to be used in the area to which the operating point of the motor 12L currently belongs from the coefficient determination map 238. The coefficient determination map 238 stores various coefficients in association with the area to which the motor operating point belongs. More specifically, on the coefficient determination map 238, as shown in FIG. 10, the coefficients a1 to a3, b1 to b3, c1 to c3, and d1 to d3 for determining the threshold values THω1 to THω3 Coefficients A1, A2, B1, B2, C1, C2, D1, and D2 for determining feedback gains G1 and G2 are generally set to different values for each region. Further, the correction terms S1 and S2 are generally set to different values depending on whether the vehicle is running or regenerated. The threshold calculator 234 on the right uses the determined coefficient
(Equation 2)
(When the accelerator is on)
THω1 = a1 × exp (b1 × VA + c1 × VRR + d1 × VA × VRR)
THω2 = a2 × exp (b2 × VA + c2 × VRR + d2 × VA × VRR)
THω3 = a3 × exp (b3 × VA + c3 × VRR + d3 × VA × VRR)
G1 = A1 × exp (B1 × VA + C1 × VRR + D1 × VA × VRR)
G2 = A2 × exp (B2 × VA + C2 × VRR + D2 × VA × VRR)
(When accelerator is off)
THω1 = a1 × exp (b1 × FB + c1 × VRR + d1 × FB × VRR)
THω2 = a2 × exp (b2 × FB + c2 × VRR + d2 × FB × VRR)
THω3 = a3 × exp (b3 × FB + c3 × VRR + d3 × FB × VRR)
G1 = A1 × exp (B1 × FB + C1 × VRR + D1 × FB × VRR)
G2 = A2 × exp (B2 × FB + C2 × VRR + D2 × FB × VRR)
, The threshold values THω1 to THω3 and the feedback gains G1 and G2 are determined. Similarly, the left threshold value calculation unit 236 determines the threshold values THω1 to THω3 and the feedback gains G1 and G2 by using a formula using VRL instead of VRR in the following formula. It should be noted that VRR is input from the wheel speed calculator 122, VRL is input from the wheel speed calculator 124, VA is input from the accelerator opening calculator 104, and FB is input from the brake force calculator 106.
[0036]
As described above, since the feedback torques ΔTR and ΔTL are determined using the threshold value, the feedback gain, and the correction term determined according to the characteristics of the motors 12R and 12L, the characteristics of the motors 12R and 12L are used over a wide area. In such a case, it is possible to generate appropriate feedback torques ΔTR and ΔTL according to the states of the motors 12R and 12L and the requirements of the vehicle operator. The torque commands TR and TL corrected using the feedback torques ΔTR and ΔTL are transmitted to the motor control units 36R and 36L when at least one of the angular accelerations dωR / dt and dωL / dt is equal to or greater than the threshold value TH. Since the output is performed, slip or the tendency can be suitably prevented or suppressed. In this embodiment, since the left and right motors 12R and 12L are assumed to have the same characteristics, the left and right motors share the coefficient determination map 238. It may be.
[0037]
In the present embodiment, the angular accelerations dωR / dt and dωL / dt are determined as threshold values. However, the slip ratio = | (wheel speed of driving wheel−body speed) / (wheel speed of driving wheel) | A threshold may be determined. However, in the embodiment in which the switching to the TRC / ABS equivalent control is executed, the angular acceleration of the driving wheel is used in the TRC / ABS equivalent control. Therefore, the angular acceleration of the driving wheel is more preferable from the viewpoint of reducing the calculation load. . The wheel speed can be obtained by detecting the wheel speeds of the left and right driven wheels and calculating the average value.
[0038]
(6) Functions of the running stability control unit and the learning management unit
FIG. 11 shows a functional configuration of the traveling stability control unit 300 and the learning management unit 600. The running stability control unit 300 transmits the torque command T obtained by the powering torque calculating unit 112 or the regenerative torque calculating unit 118 so that running instability does not occur in the vehicle body during running of the vehicle or running instability is reduced. It has a function of generating torque commands TR and TL to be supplied to the motor control units 36R and 36L based on these.
[0039]
In order to realize such a function, the driving stability control unit 300 inputs the torque command T as described above, while outputting the yaw rate sensor 50, the lateral acceleration sensor 54, the steering angle sensor 48, and the vehicle speed calculating unit 134. The vehicle speed VS obtained is input. The running stability controller 300 converts the output signal of the yaw rate sensor 50 into data indicating the yaw rate γt, the yaw rate calculator 304 converts the output signal of the lateral acceleration sensor 54 into data indicating the lateral acceleration Gy, And a steering angle calculation unit 306 that converts an output signal of the steering angle sensor 48 into data indicating a steering angle δt. Further, based on the lateral acceleration Gy, the yaw rate γt, and the vehicle speed VS, the following equation is obtained.
(Equation 3)
dβ / dt = Gy / VS−γt
The slip angular velocity calculating section 308 for calculating the slip angular velocity dβ / dt based on the equation (1), and the slip angle calculating section 310 for calculating the slip angle β by integrating the slip angular velocity dβ / dt obtained by the slip angular velocity calculating section 308 with time, Is provided. Various information obtained in this manner, that is, a yaw rate γt and a slip angle β indicating the motion state of the vehicle body, a steering angle γt indicating a turning request from the vehicle operator, and a torque command indicating an acceleration / deceleration request from the vehicle operator T is input to the control neural network 312 provided in the running stability control unit 300 and the updating neural network 602 provided in the learning management unit 600, respectively. Both of the control neural network 312 and the update neural network 602 have learned beforehand about the generation of the torque commands TR and TL according to γt, β, δt and T before the vehicle starts running. In the middle, the output signals TR and TL are generated based on the input signals γt, β, δt, and T.
[0040]
The torque commands TR and TL generated by the control neural network 312 are, in principle, supplied to the motor control units 36R and 36L as they are. However, in order to prevent the torque commands TR and TL generated by the control neural network 312 from being outside the output possible range of the motors 12R and 12L, a torque limit calculation unit 314 is provided at a subsequent stage of the control neural network 312. The torque command supply from the traveling stability control unit 300 to the motor control units 36R and 36L is restricted by the torque restriction calculation unit 314. The torque limit calculation unit 314 can be realized using, for example, a map having the contents shown in FIG. Needless to say, the output of the TRC / ABS equivalent control unit 200 may be used instead of the torque commands TR and TL output from the traveling stability control unit 300 as described above.
[0041]
While the torque commands TR and TL generated by the control neural network 312 are output to the motor control units 36R and 36L in this way, the learning control unit 604 provided in the learning management unit 600 also outputs the torque commands TR and TL. Is entered. The learning control unit 604 generates the torque commands TRS and TLS as teacher signals based on the torque commands TR and TL and the input signals γt, β, δt and T to each neural network, as described later, and updates the neural network for updating. Give to network 602. The updating neural network 602 learns the generation of the torque commands TR and TL according to the input signals γt, β, δt, and T while the vehicle is running, based on the teacher signal. The updating neural network 602 further transmits the synaptic coupling coefficients Wji and Vkj and the neuron thresholds θj and λk obtained as a result of the learning to the controlling neural network 312, and the synaptic coupling coefficients and the neurons in the controlling neural network 312. Update the threshold with this.
[0042]
FIG. 13 conceptually shows a functional configuration of the control neural network 312 and the update neural network 602. As shown in this figure, the control neural network 312 and the update neural network 602 each have an input layer 312a or 602a, an intermediate layer 312b or 602b, and an output layer 312c or 602c. In this figure, the synapse connection 312g or 602g from the i-th (i = 1 to 4) input layer neuron 312d or 602d to the j-th plane (j = 1 to 5) middle layer neuron 312e or 602e is shown. The coefficient is represented by Wji, and the coefficient of the synaptic connection 312h or 602h from the jth intermediate layer neuron 312e or 602e to the kth (k = 1 or 2) output layer neuron 312f or 602f is represented by Vkj. Have been.
[0043]
In the present embodiment, since there are four types of input signals and two types of output signals, the number of neurons 312d or 602d provided in the input layer 312a or 602a is four, and the number of neurons 312d or 602d is provided in the output layer 312c or 602c. The number of neurons 312f or 602f is two. However, when implementing the present invention, it is also possible to omit any of the four types of input signals used in the present embodiment or to add other types of input signals. Further, in the present embodiment, the number of the intermediate layers 312b or 602b is one and the number of the neurons 312e or 602e in the intermediate layer 312b or 602b is five. There is no particular need to limit the number or number of hidden neurons. Further, in this figure, a neuron whose own output is one of the inputs is not used, but such a neuron may be used in practicing the present invention.
[0044]
FIG. 14 shows a learning procedure executed by the control neural network 312 and the update neural network 602 before the vehicle starts running (before shipping, etc.) under the management of a device (not shown). As shown in this figure, at the time of learning, first, the synaptic coupling coefficients Wji and Vkj, the threshold value θj of the intermediate layer neuron 312e or 602e, and the threshold value ηk of the output layer neuron 312f or 602f are initially set to random numbers. It is set (700). However, the random numbers used for this initialization are generated within a relatively small range.
[0045]
After the initialization, the control neural network 312 and the update neural network 602 input a learning pattern from an external device (not shown) (702). Here, the learning pattern is obtained by associating a combination of the input signals γt, β, δt, and T with the torque commands TR and TL to be obtained from the output layer 312c or 602c according to the input. The control neural network 312 and the updating neural network 602 include the synapse coupling coefficient and the neuron threshold value initialized in step 700 and the input signals γt, β, δt, and T constituting the learning pattern input in step 702. Based on the following formula
(Equation 4)
Uj = ΣWji · Ii + θi
Hj = f (Uj)
The output Hj of the hidden neurons 312e and 602e is calculated according to the following equation (704). Where Σ represents the sum of i, and f is the sigmoid function
(Equation 5)
f (x) = 1 / {1 + exp (-x)}
It is. The controlling neural network 312 and the updating neural network 602 are then represented by the following equations:
(Equation 6)
Sk = ΣVkj · Hj + ηk
Ok = f (Sk)
The output Ok of the output layer neuron 312f or 602f is calculated according to the equation (706). Here, in the above equation, Σ is the sum of j.
[0046]
After calculating the output Ok of the output layer neuron 312f or 602f in this manner, the control neural network 312 and the update neural network 602 transmit the torque commands TR and TL constituting the learning pattern input in step 702 to the teacher signal. As Tk, the following equation
(Equation 7)
λk = (Ok−Tk) · Ok · (1-Ok)
, The output error λk of the output layer neuron 312f or 602f is calculated (708). The control neural network 312 or the update neural network 602 then proceeds to the following equation:
(Equation 8)
σj = Σλk · Vkj · Hj · (1-Hj)
, The output error σj of the hidden neuron 312e or 602e is calculated (710). Here, in the above equation, Σ is the sum of k.
[0047]
The control neural network 312 and the updating neural network 602 update the synaptic coupling coefficients Wji and Vkj and the neuron thresholds θj and ηk using the output errors λk and σj obtained in this manner. That is, the following equation
(Equation 9)
Vkj = Vkj + ε · λk · Hj
ηk = ηk + φ · λk
Update the synaptic coupling coefficient Vkj and the neuron threshold value λk according to (712).
(Equation 10)
Wji = Wji + ε · σj · Ii
θj = θj + φ · σj
Update the synaptic coupling coefficient Wji and the neuron threshold value θj according to (714). Here, ε and φ are constants.
[0048]
After updating the synaptic coupling coefficients Wji and Vkj and the neuron thresholds θj and ηk in this manner, the control neural network 312 and the update neural network 602 generate a learning pattern different from the learning pattern input in step 702. The selection is made (716), and the operation after step 702 is executed again. As a result of repeating the operations of steps 702 to 714 while updating the learning pattern in this way, when the learning for all the prepared learning patterns is completed (718), the control neural network 312 and The updating neural network 602 stores the fact that one type of learning has been completed, and executes learning for all learning patterns again (720). When the number of repetitions of learning for all learning patterns reaches a predetermined limit number (722), the control neural network 312 and the updating neural network 602 end the learning procedure to be executed before the vehicle starts running.
[0049]
The learning procedure described above may be executed for both the control neural network 312 and the update neural network 602, but only one of the control neural network 312 and the update neural network 602 executes the learning. When the learning up to steps 700 to 722 is completed, the result may be transferred to the remaining one neural network (724). By doing so, it is possible to eliminate the difference between the two neural networks at the start of vehicle running (at the time of shipment) with respect to the synaptic coupling coefficients Wkj and Vkj and the neuron thresholds θj and ηk at the start of running of the vehicle.
[0050]
The learning method in steps 700 to 722 in FIG. 14 is called an error back propagation method (back propagation method). That is, an output signal obtained by giving an input signal according to a certain combination to the neural network is compared with a previously assumed output signal, that is, a teacher signal, and the difference between the output signal and the teacher signal is compared.
(Equation 11)
E = Σ (Tk−Ok) ² / 2
In this method, the synaptic coupling coefficient and the neuron threshold are updated so as to minimize. In the above equation, Σ represents the sum of k. To minimize the error E of the output signal Ok with respect to the teacher signal Tk with respect to the synaptic coupling coefficient Vkj,
(Equation 12)

Vkj may be updated so that is smaller. In the above equation 9, the coefficient of Hj appearing on the right side of equation 12 is set as λk. Similarly, to minimize the error E with respect to the synaptic coupling coefficient Wji,
(Equation 13)

Should be minimized. In the above equation, Σ represents the sum regarding k. Equation 8 above is obtained by setting the coefficient of Ii appearing on the right side of Equation 13 as σj.
[0051]
FIGS. 15 to 17 show the flow of operations of the control neural network 312, the update neural network 602, and the learning control unit 604 while the vehicle is running. These procedures are repeatedly executed by the corresponding members at predetermined intervals. One of the features of the present embodiment is that the control neural network 312 executes the torque commands TR and TL while the vehicle is running, based on the result of learning performed before the vehicle starts running, as described below. , The update neural network 602 performs learning based on the actual input signal under the control of the learning control unit 604, and obtains the tissue of the control neural network 312, for example, the synaptic coupling coefficient and the neuron threshold by learning. The point is that the update is performed sequentially according to the tissue of the updating neural network 602, for example, the synaptic coupling coefficient or the neuron threshold value.
[0052]
First, as shown in FIG. 15, the control neural network 312 receives the yaw rate γt, the slip angle β, the steering angle δt, and the torque command T (800), and uses the same principle as in steps 704 and 706 to input the intermediate layer. The output Hj of the neuron 312e and the output Ok of the output layer neuron 312f are sequentially calculated (802, 804). The control neural network 312 supplies the output Ok obtained by the calculation, that is, the torque commands TR and TL to the motor control units 36R and 36L via the torque limit calculation unit 314, while supplying the output Ok to the learning control unit 604. As shown in FIG. 17, the learning control unit 604 inputs the yaw rate γt, the slip angle β, the steering angle δt, and the torque command T, in addition to the torque commands TR and TL generated by the control neural network 312 ( 1000), the teacher signal TRS related to the torque command TR and the teacher signal TLS related to the torque command TL are expressed by the following equations.
[Equation 14]
TRS = FR (γt, β, δt, T)
TLS = FL (γt, β, δt, T)
Is generated (1004), and the teacher signals TRS and TLS are supplied to the updating neural network 602 (1006).
[0053]
The updating neural network 602 receives the teacher signals TRS and TLS in addition to the yaw rate γt, the slip angle β, the steering angle δt, and the torque command T (FIG. 16: 900). The neural network for update 602 outputs the output Hj of the intermediate layer neuron 602e, the output Ok of the output layer neuron 602f, the output error λk of the output layer neuron 602f, and the output error σj of the intermediate layer neuron 602e by the same procedure as in steps 704 to 710. Are calculated in order (902-908). The updating neural network 602 updates the synaptic coupling coefficients Wji and Vkj and the neuron threshold values θj and λk in the same procedure as in steps 712 and 714 (910, 912). The updating neural network 602 transfers the updated synaptic coupling coefficients Wji and Vkj, the neuron thresholds θj and λk, and the control neural network 312 (914). When the synaptic coupling coefficients Wji and Vkj and the neuron thresholds θj and λk are transferred from the updating neural network 602 (806), the control neural network 312 determines its own synapse based on the transferred information. The coupling coefficients Wji and Vkj and the neuron threshold values θj and λk are updated (808).
[0054]
Therefore, in the present embodiment, the torque commands TR and TL are generated using the control neural network 312 based on the yaw rate γt, the slip angle β, the steering angle δt, and the torque command T, which are input signals, and the wheel angular acceleration is not large. The motors 12R and 12L are controlled using the torque commands TR and TL as long as it is possible (that is, as long as slip or the tendency is not considered to occur). In addition, the organization of the control neural network 312 is formed before the vehicle starts running by the procedure shown in FIG. Therefore, when the procedure shown in FIG. 14 is executed, if the data obtained by running the actual vehicle or the data obtained by the numerical simulation (simulation operation) is set in the learning pattern, only the basic motion equation of the vehicle body is sufficient. In addition, it is possible to cope with a factor that cannot be taken into the control model, and therefore, the reliability and accuracy of the control are improved as compared with the related art. Further, while the vehicle is traveling, the learning control unit 604 uses the teacher signals TRS and TLS is generated, and the update neural network 602 transfers the result of learning performed based on the teacher signal TRS and the TLS to the control neural network 312 to update its organization. The organization of the control neural network 312 can be updated according to the acceleration / deceleration request and the vehicle body motion state, and more reliable control is realized. Further, since the torque command generation processing by the control neural network 312 and the learning by the update neural network 602 are performed simultaneously and in parallel, the speed of the control is increased due to the efficiency of the processing.
[0055]
Further, in the present embodiment, the quantities used as input signals to the respective neural networks are all the quantities appearing in the basic motion equation of the vehicle body or the quantities derivable from the quantities. First, considering that there is no left-right difference in the wheels and no difference in left-right cornering force, the basic motion equation of the vehicle body is
(Equation 15)

It becomes. See FIG. 18 for the variable names and their meanings. On the other hand, the equation of motion of the rotation of each wheel is as follows, assuming that there is no left-right difference between the wheels and that the commanded torque is exactly the actual output as it is
(Equation 16)

It becomes. See FIG. 19 for variable names and their meanings. In addition, the cornering force
[Equation 17]

It is expressed as Therefore, when rearranging equations 16 and 17 into equation 15,
(Equation 18)

Is obtained. Where s is the Laplace operator.
[0056]
Among the quantities used as input signals in the present embodiment, the yaw rate γt, the slip angle β, and the steering angle δt appear in the simultaneous differential equations shown in Expression 18. Further, the torque command T is expressed in the same formula in the form of the vehicle speed VS and the moment M. Therefore, according to the present embodiment in which γt, φ, δt, and T are used as input signals to each neural network, the acceleration / deceleration request and the steering request represented by the basic motion equation of the vehicle body and the motion state of the vehicle body Can be suitably reflected in the running stability control, and more accurate control can be realized as compared with the related art. Further, regarding the generation of the torque commands TR and TL using the control neural network 312, there is no application restriction depending on the state such as whether or not the vehicle body is turning. Therefore, even when the vehicle is traveling straight, it is possible to suppress changes in the slip angle β and the yaw rate γt due to a slip or the like.
[0057]
In addition, as described above, any one of the four types of input signals can be omitted, or can be replaced with another amount. In this case, it is preferable to use, as an input signal, an amount having a higher sensitivity than the yaw rate γt having a relatively low sensitivity. That is, as in the present embodiment, if the slip angle β with relatively high sensitivity is used in addition to the yaw rate γt with relatively low sensitivity, for example, the target yaw rate disclosed in Japanese Patent Application Laid-Open No. 5-91607 is used. Compared to control, accurate and highly reliable control can be realized. Further, when the learning control unit 604 generates the teacher signals TRS and TLS, the yaw rate γt, the slip angle β, and the steering angle δt, which are also used as input signals to the control neural network 312 and the update neural network 602, are used. By using the torque command T and the torque command T, teacher signals TRS and TLS with relatively high reliability can be generated. For example, if the teacher signals TRS and TLS are generated in accordance with the feedback control logic proposed by the present applicant in Japanese Patent Application No. 9-68571, the model based on the basic body motion equation of the vehicle body is reflected even during learning during traveling. Therefore, the learning / tissue update loop in which the update neural network 602 learns based on the output of the control neural network 312 and updates the organization of the control neural network 312 based on the result becomes unstable. Absent.
[0058]
【The invention's effect】
As described above, in the preferred embodiment of the present invention, the motor is not based on the basic motion equation of the vehicle body that models the vehicle body motion, but on the basis of what the control neural network has learned in advance based on the results of actual vehicle tests or simulation calculations. And an update neural network that learns while the vehicle is running based on acceleration / deceleration requests, vehicle body motion status, output of the control neural network, etc. so that the organization of the control neural network can be updated even while the vehicle is running. As a result, the problems inherent in the modeling of the vehicle body motion based on the basic motion equation can be solved.In addition, in the actual use environment, the steering angle is too large or too small, disturbance received after steering, control system delay, and other various control error factors. I can deal with it. As a result, control reliability, accuracy, responsiveness, and the like are improved.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a system configuration of an electric vehicle suitable for carrying out the present invention.
FIG. 2 is a cross-sectional view showing an example of the structure of a wheel-in motor, and a circle in the upper right in the figure is a partially enlarged view.
FIG. 3 is a diagram showing a braking force distribution in the system of FIG. 1;
FIG. 4 is a block diagram illustrating a functional configuration of a vehicle control unit.
FIG. 5 is a block diagram illustrating a functional configuration of a torque command provisional determination unit.
FIG. 6 is a diagram showing the contents of a powering torque map.
FIG. 7 is a diagram showing the contents of a regenerative torque map.
FIG. 8 is a block diagram illustrating a functional configuration of a switching control unit.
FIG. 9 is a block diagram illustrating a functional configuration of a TRC / ABS equivalent control unit.
FIG. 10 is a diagram showing the contents of a coefficient determination map.
FIG. 11 is a block diagram illustrating a functional configuration of a traveling stability control unit and a learning management unit.
FIG. 12 is a map showing functions of a torque limit calculation unit.
FIG. 13 is a diagram conceptually showing a functional configuration of a control and update neural network.
FIG. 14 is a flowchart showing a learning procedure executed by the control and update neural networks before the vehicle starts running.
FIG. 15 is a flowchart showing a procedure of command generation and tissue update executed by the control neural network during vehicle running.
FIG. 16 is a flowchart showing a procedure of learning executed by the updating neural network during traveling of the vehicle and a procedure of transferring a result of the learning.
FIG. 17 is a flowchart illustrating a teacher signal generation procedure executed by the learning control unit while the vehicle is traveling.
FIG. 18 is a conceptual plan view that defines a predetermined number and variables for describing the motion of the vehicle body.
FIG. 19 is a conceptual diagram defining a predetermined number and variables for describing the rotational motion of the wheel.
[Explanation of symbols]
10RR right rear wheel, 10RL left rear wheel, 12R, 12L motor, 34R, 34L inverter, 36R, 36L motor control unit, 38 vehicle control unit, 40RR, 40RL, 40FR, 40FL wheel speed sensor, 42 accelerator sensor, 44 brake sensor , 46 shift position switch, 48 steering angle sensor, 50 yaw rate sensor, 54 lateral acceleration sensor, 100 torque command provisional determination section, 200 TRC / ABS equivalent control section, 300 running stability control section, 302 yaw rate calculation section, 304 lateral acceleration calculation Section, 306 steering angle calculation section, 308 slip angular velocity calculation section, 310 slip angle calculation section, 312 control neural network, 314 torque limit calculation section, 400 control switching section, 500 switching control section, 600 learning management section, 602 updating neural network, 04 learning control unit, TR, TL, T torque command, phi vehicle slip angle, [gamma] t yaw rate, .DELTA.t steering angle, Wij, Vkj synaptic coupling coefficient, .theta.j, .lamda.k neuron threshold.

Claims (1)

A drive control device that controls the running of the electric vehicle by individually giving a command regarding the output to each of a plurality of motors for individually driving the driving wheels of the electric vehicle. Based on the result of the test or the simulation operation, learning has been completed regarding the acceleration / deceleration request and the generation of the command corresponding to the vehicle body motion state. The neural network for control and the generation of the above-mentioned command according to the acceleration / deceleration request and the vehicle body motion state based on the actual vehicle test or the simulation calculation result before starting the vehicle traveling have been learned. The above learning is continued based on the output of the network, and based on the result, the updating neural network for updating the organization of the control neural network is updated. Neural network and drive control apparatus characterized by comprising: a travel in learning management means for providing a training signal to the update neural network during the vehicle traveling based on the output of the input signal and the control neural network.

Images