JP5179686B2 - Driving behavior risk calculation device - Google Patents

Description

【００４８】
ここで、ａ，ｂ，ｃ，ｄ，ｅは、それぞれのパラメータに対応する重み付け係数である。各重み付け係数は、視行動の分類と時間に応じて例えば次のように異なっている。例えば、重み付け係数ａは、注視時間が長いときは小さく、注視時間が短いときは大きくなる。また、重み付け係数ｂは、瞬きしている時間が長いときは大きくなり、瞬きしている時間が短いときは小さくなる。重み付け係数ｃは、サッケード眼球運動をしている時間が長いときは大きくなり、サッケード眼球運動をしている時間が短いときは小さくなる。重み付け係数ｄは、追従眼球運動をしている時間が長いときは小さくなり、追従眼球運動をしている時間が短いときは大きくなる。重み付け係数ｅは、輻輳運動をしている時間が長いときは大きくなり、追従眼球運動をしている時間が短いときは小さくなる。
【００４９】
そして、運転行動危険度演算部３３は、式（１）に従って運転手の注意状態を考慮した危険度ＤＲを求め、さらに車両状態を考慮して運転行動危険度ＤＲ１を演算する。例えば、車速が速くなるに従って危険度ＤＲを高くしたり、操舵角が大きくなるに従って危険度ＤＲを高くしたり、ブレーキ油圧が高くなるに従って危険度ＤＲを低くすることによって、運転行動危険度ＤＲ１を求める。
【００５０】
以上のように、第１の実施の形態に係る運転行動危険度演算装置１は、ドライバの視行動を分類して注意状態を推定し、ドライバの注意状態と車速等の車両状態とに基づいて運転行動危険度ＤＲ１を演算することによって、ドライバの注意状態とドライバの実際の運転行動とを総合的に考慮した危険度の判定を行うことができる。
【００５１】
［第２の実施の形態］
つぎに、本発明の第２の実施の形態について説明する。なお、第１の実施の形態と同一の部位については同一の符号を付し、重複する説明については省略する。本実施の形態では、マイクロコンピュータ３０は、第１の実施の形態と異なる演算手法を用いて運転行動危険度ＤＲ２を演算する。
【００５２】
本実施の形態に係るマイクロコンピュータ３０は、機能的には図４に示すように構成されている。すなわち、マイクロコンピュータ３０は、ドライバ用撮像画像に基づいてドライバの視行動を検出する視行動検出部３１と、ドライバの視行動に基づいて瞬きの時間等を検出する視線フィルタ３２と、車外撮像画像から対象物を検出する対象物検出部３４と、を備えている。
【００５３】
さらに、マイクロコンピュータ３０は、ＧＰＳ信号に基づいて車両の現在位置を検出する位置検出部３５と、対象物毎の環境危険度ＫＲを演算する環境危険度演算部３６と、環境危険度ＫＲを演算するためのパラメータテーブルや、演算された環境危険度ＫＲを記憶する環境危険度記憶部３７と、運転行動危険度ＤＲ２を演算する運転行動危険度演算部３８と、を備えている。
【００５４】
（パラメータテーブル）
ここで、環境危険度記憶部３７には、対象物パラメータテーブル、向きパラメータテーブル、速度パラメータテーブル、大きさパラメータテーブル、位置パラメータテーブルが記憶されている。
【００５５】
対象物パラメータテーブルは、図５に示すように、対象物の種類（カテゴリ）と、対象物の各種類の細分類と、細分類毎の危険度パラメータとで構成されている。なお、危険度パラメータとは、事故誘因の可能性に関連した値をいい、例えば、事故の統計、保険料算定データ、運転規範データ等を考慮して求められた値である。
【００５６】
対象物の種類としては、例えば、歩行者、２輪車、自動車がある。歩行者の細分類としては、高齢者、子供、中学生−成人がある。２輪車の細分類としては、自転車、バイクがある。自動車の細分類としては、普通車、大型車がある。高齢者、子供、中学生−成人、自転車、バイク、普通車、大型車には、それぞれ危険度パラメータｔ１、ｔ２、・・・、ｔ７が付されている。
【００５７】
一般に、歩行者、２輪車、自動車の順に危険度が低くなっていき、詳しくは、高齢者、子供、中学生−成人、自転車、バイク、普通車、大型車の順に危険度が低くなると考えられる。すなわち、高齢者の歩行者が最も危険度が高く、大型車が最も危険度が低くなっている。そこで、危険度パラメータは、ｔ１＞ｔ２＞・・・＞ｔ７になっている。
【００５８】
向きパラメータテーブルは、図６に示すように、対象物の向きと、向き毎の危険度パラメータとで構成されている。対象物の向きとしては、車両の進行方向に対して直交、車両の進行方向と同方向、車両の進行方向と対向がある。直交、同方向、対向には、それぞれ危険度パラメータｍ１、ｍ２、ｍ３が付されている。また、直交、同方向、対向の順に危険度が低くなっている。そこで、危険度パラメータは、ｍ１＞ｍ２＞ｍ３になっている。
【００５９】
速度パラメータテーブルは、図７に示すように、対象物の速度と、速度毎の危険度パラメータとで構成されている。対象物の速度としては、高速、歩行速度（例えば、時速４ｋｍ前後）、停止がある。高速、歩行速度、停止には、それぞれ危険度パラメータｓ１、ｓ２、ｓ３が付されている。また、高速、歩行速度、停止の順に危険度が低くなっている。そこで、危険度パラメータは、ｓ１＞ｓ２＞ｓ３になっている。
【００６０】
大きさパラメータテーブルは、図８に示すように、対象物の大きさと、大きさ毎の危険度パラメータとで構成されている。対象物の大きさは特に限定されるものではなく、対象物の大きさ毎に危険度パラメータｄ１、ｄ２、ｄ３・・・が付されている。なお、危険度が高い「大きさ」には大きな値の危険度パラメータが付され、危険度が低い「大きさ」には小さな値の危険度パラメータが付されている。
【００６１】
また、位置パラメータテーブルは、図９に示すように、対象物の位置と、位置毎の危険度パラメータとで構成されている。対象物の位置は特に限定されるものではなく、対象物の位置毎に危険度パラメータｐ１、ｐ２、ｐ３・・・が付されている。なお、危険度が高い「位置」には大きな値の危険度パラメータが付され、危険度が低い「位置」には小さな値の危険度パラメータが付されている。
【００６２】
（環境危険度ＫＲの演算）
そして、マイクロコンピュータ３０は、車両周囲の環境の対象物の種類毎に環境危険度ＫＲを以下のようにして演算する。
【００６３】
対象物検出部３４は、車外用ＣＣＤカメラ１３で得られた車外撮像画像にパターンマッチング等の所定の画像処理を施して、車外撮像画像から対象物を抽出して検出する。また、対象物検出部３４は、レーダ装置１４から出射され受光されるまでの赤外線レーザの往復時間等に基づいて対象物までの相対的な距離や速度を認識する。そして、対象物検出部３４は、これらの認識結果に基づいて、対象物の種類毎に、車両の進行方向に対する向き、速度、大きさ、位置（自車に対する相対的な位置）を検出し、検出結果を環境危険度演算部３６に供給する。
【００６４】
なお、ＩＤ信号を発信する発信器を歩行者に保持させたり、このような発信器を２輪車や自動車に取り付けてもよい。この場合、対象物検出部３４は、各発信器から発信されたＩＤ信号に基づいて、歩行者、２輪車、自動車を検出することができる。
【００６５】
位置検出部３５は、ＧＰＳ信号受信回路１１で受信されたＧＰＳ信号に基づいて自車の現在位置を検出し、自車の現在位置を環境危険度演算部３６に供給する。
【００６６】
環境危険度演算部３６は、対象物の種類毎に危険度を演算する。具体的には、環境危険度演算部３６は、環境危険度記憶部３７に記憶された図５から図９に示した各テーブルを参照し、さらに対象物検出部３４や位置検出部３５で検出された情報を用いて、対象物、向き、速度、大きさ、位置にそれぞれ対応する危険度パラメータを選択する。そして、選択した危険度パラメータを用いて、対象物の種類に応じた環境危険度ＫＲを示す関数Ｆを演算する。なお、危険度パラメータが例えばａ₁，ａ₂，ａ₃，・・・とあるときは、環境危険度ＫＲを示す関数はＦ（ａ₁，ａ₂，ａ₃，・・・）として表すことができる。本実施の形態では、環境危険度ＫＲを示す関数として例えば式（２）のように表すことができる。
【００６７】
【数２】

【００６８】
式（２）において、ｔは対象物の種類の危険度パラメータ、ｍは向きの危険度パラメータ、ｓは速度の危険度パラメータ、ｄは大きさの危険度パラメータ、ｐは位置の危険度パラメータである。Ａ、Ｂ、Ｃ、Ｄ、Ｅは、それぞれ所定の重み付け係数である。
【００６９】
環境危険度演算部３６は、式（２）に従って対象物毎に環境危険度ＫＲを演算し、対象物毎の環境危険度ＫＲを運転行動危険度演算部３８に供給する。環境危険度演算部３６は、上述したような環境危険度ＫＲの演算処理を行う毎に、環境危険度ＫＲの履歴データを環境危険度記憶部３７に記憶する。
【００７０】
図１０は、対象物の種類毎に演算された環境危険度ＫＲの履歴データを示す図である。履歴データは、環境危険度ＫＲが演算された日時、環境危険度ＫＲ、演算時に用いた危険度パラメータ（対象物、向き、速度、大きさ、位置）で構成されている。
【００７１】
これにより、環境危険度演算部３６は、検出された車両周囲の環境における対象物と、レーダ装置１４を用いて検出された対象物の向きや速度等の属性と、危険度パラメータと、に基づいて歩行者や自動車等の移動可能な対象物の種類毎に環境危険度ＫＲを演算することができる。このとき、対象物の種類の危険度パラメータだけでなく、対象物の向き、速度、大きさ、位置それぞれの危険度パラメータを用いているので、対象物の総合的な環境危険度ＫＲを求めることができる。
【００７２】
（運転行動危険度ＤＲの演算）
視行動検出部３１は、ドライバ用ＣＣＤカメラ１８で撮像されたドライバ用撮像画像に基づいて、ドライバの視行動として例えば、視線の向きや眼球運動を検出する。
【００７３】
視線フィルタ３２は、視行動検出部３１で検出されたドライバの視行動を注視、瞬き、サッケード眼球運動、追従眼球運動、輻輳運動の５つに分類する。また、視線フィルタ３２は、視行動として注視を検出したときは、注視時間も検出する。そして、視線フィルタ３２は、ドライバの視行動から求められた結果を運転行動危険度演算部３８に供給する。
【００７４】
運転行動危険度演算部３８は、環境危険度演算部３６で演算された対象物毎の環境危険度ＫＲ_nに対して、視線フィルタ３２の検出結果に基づいて重み付け係数Ｅ_nを求め、更に自車の車速ｖを用いて運転行動危険度ＤＲ２を演算する。以下、運転行動危険度ＤＲ２の演算について説明する。
【００７５】
図１１は、運転行動危険度演算部３８の演算内容を模式的に示した図である。運転行動危険度演算部３８は、最初に、環境危険度演算部３６で演算された環境危険度ＫＲ_nに対応する重み付け係数Ｅ_nを生成し、各環境危険度ＫＲ_nに重み付けを行う。
【００７６】
重み付け係数Ｅ_nは、視線フィルタ３２の出力によって決定されるパラメータであり、具体的には、対象物にドライバの視線があるときは値が小さくなり、対象物にドライバの視線がないときは値が大きくなる。また、重み付け係数Ｅ_nは、対象物を注視している時間が長いときは値が小さくなり、対象物を注視している時間が短いときは値が大きくなる。
【００７７】
対象物にドライバの視線がある場合においては、さらに、注視、瞬き、サッケード眼球運動、追従眼球運動、輻輳運動の分類に応じて重み付け係数Ｅ_nの値を変えてもよい。例えば、注視や追従眼球運動の場合には重み付け係数Ｅ_nの値を小さくし、瞬き、サッケード眼球運動及び輻輳運動の場合には重み付け係数Ｅ_nの値を大きくすることもできる。
【００７８】
運転行動危険度演算部３８は、環境危険度演算部３６で演算された各対象物の環境危険度ＫＲ_nと当該環境危険度ＫＲ_nに対応する重み付け係数Ｅ_nとの積の総和を演算して、式（３）に示す中間危険度ＭＲを求める。
【００７９】
【数３】

【００８０】
次に、運転行動危険度演算部３８は、各車輪速センサ１７で検出された車輪速信号に基づいて自車の車速ｖを演算し、この自車の車速ｖを中間危険度ＭＲに乗ずる。そして、ｖ・ＭＲに対して心理量変換を行うことで、運転行動危険度ＤＲ２を求める。ここで、心理量変換を示す関数をｇとすると、運転行動危険度ＤＲ２は次の式（４）で表される。
【００８１】
【数４】

【００８２】
なお、運転行動危険度演算部３８は、関数ｇを用いることなく運転行動危険度ＤＲ２を求めることもできる。例えば、予めｖ・ＭＲを心理変換したときの運転行動危険度ＤＲ２の関係を示すマップをメモリに記憶しておき、このマップを参照して運転行動危険度ＤＲ２を求めてもよい。
【００８３】
（最終画像生成処理）
運転行動危険度演算部３８は、運転行動危険度ＤＲ２の演算処理を行うだけでなく、さらに、車外撮像画像、車速、操舵角等に基づいて最終画像の生成処理を行う。ここでは、図１２に示す車外撮像画像を基礎として最終画像を生成することについて説明する。
【００８４】
最初に、運転行動危険度演算部３８は、視線フィルタ３２の検出結果に基づいて、車外撮像画像に対して、視線を向けていない場所にマスクをかける処理を行う。図１３は、車外撮像画像にかけられたマスクの状態を説明する図である。
【００８５】
図１３（Ａ）は、ドライバが前方を注視しているときのマスクを示す図である。ドライバは、前方を注視しているときは、注視している場所の両端はあまり見ていない。そこで、運転行動危険度演算部３８は、撮像画像の両端が見えないようにマスク処理を行う。
【００８６】
同図（Ｂ）は、ドライバが右寄り前方を注視しているときのマスクを示す図である。ドライバは、右寄り前方を注視しているときは、注視している場所以外（正面及び左側）はあまり見ていない。そこで、運転行動危険度演算部３８は、撮像画像の正面及び左側が見えないようにマスク処理を行う。なお、運転行動危険度演算部３８は、ドライバが左寄り前方を注視しているときは、これとを逆の処理、すなわち撮像画像の正面及び右側が見えないようにマスク処理を行えばよい。
【００８７】
同図（Ｃ）は、ドライバがまばたきをしているときのマスクを示す図である。ドライバは、まばたきをしているときは、前方全部をあまり見ていない。そこで、運転行動危険度演算部３８は、撮像画像の全体が見えないようにマスク処理を行う。なお、運転行動危険度演算部３８は、まばたきの頻度に応じてマスク処理の濃度の調整を行ってもよい。例えば、単位時間当たりのまばたきの回数が所定値より少ないときはマスクの濃度を薄くしたり、単位時間当たりのまばたきの回数が所定値以上のときはマスクの濃度を濃くしてもよい。
【００８８】
つぎに、運転行動危険度演算部３８は、車速に基づいて１秒後及び２秒後の自車の到達距離を演算する。さらに、運転行動危険度演算部３８は、操舵角センサ１６で検出された操舵角に基づいて１秒後の自車の進行方向を演算する。図１４は、１秒後の到達距離（実線）、２秒後の到達距離（点線）及び１秒後の自車の進行方向（矢印）を示す図である。
【００８９】
そして、運転行動危険度演算部３８は、車外撮像画像に対してドライバの視線に応じてマスク処理を行い、さらに、１秒後の到達距離、２秒後の到達距離及び１秒後の自車の進行方向を重ねて、最終画像を生成する。つまり、図１２の車外撮像画像、図１３（Ａ）から（Ｃ）のいずれか１つの画像、図１４の画像を重ね合わせて最終画像を生成する。
【００９０】
図１５は、車外撮像画像に上述した画像処理を施すことによって生成された最終画像を示す図である。図１５（Ａ）はドライバが前方を注視しているとき、同図（Ｂ）はドライバが右寄り前方を注視しているとき、同図（Ｃ）はドライバがまばたきをしているときに生成された最終画像を示す図である。
【００９１】
また、図１６は、マイクロコンピュータ３０によって演算された運転行動危険度（０〜＋１）の結果を示す図である。横軸はフレーム番号を示し、このフレーム番号は車外用ＣＣＤカメラ１３及びドライバ用ＣＣＤカメラ１８が撮像する１フレーム毎の撮像画像を示している。また、３０フレームは１秒に相当する。したがって、フレーム番号の３０，６０，９０・・・は、１秒，２秒，３秒・・・に対応している。
【００９２】
図１６において、Ａ点では、４秒１０フレーム、ＤＲ２＝０．５４である。Ｂ点では、１９秒２６フレーム、ＤＲ２＝０．０５である。Ｃ点では、２２秒２２フレーム、ＤＲ２＝１．００である。Ｄ点では、２４秒１１フレーム、ＤＲ２＝０．６０である。
【００９３】
図１７は、Ａ点における車外撮像画像（Ａ）と、Ａ点におけるドライバ用撮像画像（Ｂ）を示す図である。このとき、車速は４．７［ｋｍ／ｈ］、ドライバは左寄り前方を注視している。ドライバは左側にある自動車周辺を注視しているが、右側に歩行者がいるので、運転行動危険度ＤＲ２は少し高い値（０．５４）になった。
【００９４】
図１８は、Ｂ点における車外撮像画像（Ａ）と、Ｂ点におけるドライバ用撮像画像（Ｂ）を示す図である。このとき、車速は０．０［ｋｍ／ｈ］、ドライバは左目視している。ドライバは左を目視しているために前方の歩行者を確認していないが、自車は完全に停止しているので、運転行動危険度ＤＲ２は低い値（０．０５）になった。
【００９５】
図１９は、Ｃ点における車外撮像画像（Ａ）と、Ｃ点におけるドライバ用撮像画像（Ｂ）を示す図である。このとき、車速は１．０［ｋｍ／ｈ］、ドライバは左目視している。ドライバは左を目視して前方の歩行者を確認してなく、かつ、自車は動いているので、運転行動危険度ＤＲ２は最高値（１．００）になった。
【００９６】
図２０は、Ｄ点における車外撮像画像（Ａ）と、Ｄ点におけるドライバ用撮像画像（Ｂ）を示す図である。このとき、車速は４．１［ｋｍ／ｈ］、ドライバは右寄り前方を注視している。ドライバは右寄り前方を注視して右側から自動車等が来ないかを確認しているが、前方の２人の歩行者を確認していないので、運転行動危険度ＤＲ２（０．６０）は少し高くなった。
【００９７】
（最終処理）
マイクロコンピュータ３０は、上述したように運転行動危険度ＤＲ２を演算し、更に最終画像を生成すると、ＬＣＤ１９に最終画像及び運転行動危険度ＤＲ２を表示させる。これにより、ドライバは、ＬＣＤ１９に表示される最終画像を確認することによって、自分が今どの場所を注意しているかを認識することができる。また、ドライバは、運転行動危険度ＤＲ２を確認することによって、現在の客観的な危険度を認識することができ、危険度が高かった場合には事故が起きないように注意して運転を行うことができる。
【００９８】
なお、マイクロコンピュータ３０は、運転行動危険度ＤＲ２の値が所定の閾値より大きい場合には事故の起こる可能性があると判定し、ドライバに注意を喚起すべく、スピーカ２０から警告音を発するようにしてもよい。
【００９９】
以上のように、第２の実施の形態に係る運転行動危険度演算装置１は、車両周囲の環境における対象物の危険度ＫＲを演算するだけでなく、その対象物をどの程度ドライバが注意しているかを考慮して運転行動危険度ＤＲ２を演算することによって、車両周囲環境及びドライバの動作を総合的に判断して、より客観的な危険度を演算することができる。そして、運転行動危険度ＤＲ２を表示することで、ドライバに注意を促すことができる。
【０１００】
また、運転行動危険度演算装置１は、視線に応じて生成された最終画像を表示することによって、ドライバが注意している場所を視覚により認識させることができる。この結果、運転行動危険度演算装置１は、車両の周囲環境に起因する事故やドライバの不注意等に起因する事故の両方を未然に防止することができる。
【０１０１】
［第３の実施の形態］
つぎに、本発明の第３の実施の形態について説明する。なお、上述した実施の形態と同一の部位については同一の符号を付し、重複する説明については省略する。本実施の形態では、マイクロコンピュータ３０は、第１及び第２の実施の形態と異なる手法を用いて運転行動危険度ＤＲ３を演算する。
【０１０２】
マイクロコンピュータ３０は、機能的には図２１に示すように構成されており、ドライバ用撮像画像に基づいてドライバの視行動を検出する視行動検出部３１と、ドライバの瞬きの時間等を検出する視線フィルタ３２と、車外撮像画像から対象物を検出する対象物検出部３４と、ドライバの視線移動等に基づいて運転行動危険度ＤＲ３を演算する運転行動危険度演算部３９とを備えている。
【０１０３】
図２２は、上空から自車を見たときのドライバの横方向の視線の向きを分類した図である。自車の進行方向を含む領域を領域Ａとし、時計回り方向に領域Ｂ、領域Ｃ、・・・、領域Ｈとした。なお、横方向の視線の向きは、このように８つの領域に分類する代わりに、角度を用いて数値化してもよい。
【０１０４】
図２３は、車外用ＣＣＤカメラ１３により撮像された車外撮像画像の模式的な図である。車外撮像画像は、横方向において、中央部分を領域Ａ、領域Ａの左側領域を領域Ｈ、領域Ａの右側領域を領域Ｂに分類している。領域Ａ、領域Ｂ、領域Ｈは、図２２に示したものとそれぞれ対応しており、自車の進行方向を基準方向としたときの向きに対応している。
【０１０５】
また、車外撮像画像は、縦方向において上から順に領域イ、領域ロ、領域ハ、領域ニに分類されている。領域イから領域ニは、それぞれ視線の高低の向きを示している。なお、視線の高低の向きは、角度を用いて数値化してもよい。
【０１０６】
図２４は、車外撮像画像とドライバの視線の向きを示す図である。Ｘ₁，Ｘ₂，・・・，Ｘ₅は、それぞれ対象物検出部３４によって検出された歩行者や車両等の障害物である。ここでは、自車が図２４に示すような障害物が存在する道路を走行し、ドライバが同図の矢印に従って視線を移動した場合を例に挙げて説明する。
【０１０７】
運転行動危険度演算部３９は、対象物検出部３４からの出力に基づいて道路上の各障害物Ｘ₁〜Ｘ₅を認識すると共に、視線フィルタ３２からの出力に基づいて図２４に示す矢印に沿って移動するドライバの視行動を検出する。なお、ドライバの視行動としては、注視と眼球追従運動を使用するのが好ましい。また、ドライバが所定時間以上注視している場合のみを用いると、さらに高精度に運転行動危険度ＤＲ３を求めることができる。
【０１０８】
運転行動危険度演算部３９は、最初の視線が領域Ａにある場合では、ドライバは領域Ａに存在する２個の障害物Ｘ₂及びＸ₃を見ていると判定する。これ以前では、ドライバは障害物を見ていなかったので、運転行動危険度演算部３９は、ドライバが記憶している障害物の個数（以下「記憶個数」という。）を０とする。
【０１０９】
つぎに、運転行動危険度演算部３９は、視線が領域Ｈに移動したことを検出すると、ドライバは領域Ｈに存在する１個の障害物Ｘ₁を見ていると判定する。ドライバはこの直前に２個の障害物Ｘ₂及びＸ₃を見ていたので、運転行動危険度演算部３９は、記憶個数を２に設定する。
【０１１０】
運転行動危険度演算部３９は、視線が再び領域Ａに移動したことを検出すると、ドライバは領域Ａに存在する障害物Ｘ₂及びＸ₃を見ていると判定する。ドライバはこの直前に１個の障害物Ｘ₁を見ていたので、運転行動危険度演算部３９は、記憶個数を１に設定する。
【０１１１】
そして、運転行動危険度演算部３９は、視線が領域Ｂに移動したことを検出すると、ドライバは領域Ｂに存在する障害物Ｘ₄及びＸ₅を見ていると判定する。ドライバは領域Ｂに視線を移動させる前までに領域Ｈにおける障害物Ｘ₁と領域Ａにおける障害物Ｘ₂及びＸ₃を見ていたので、運転行動危険度演算部３９は、記憶個数を３に設定する。
【０１１２】
図２５は、ドライバの視線移動順序、ドライバが現在見ている障害物の個数及びそのときの記憶個数の関係を示す図である。記憶個数は、ドライバが現在見ている障害物以外に把握している障害物の個数を示している。したがって、記憶個数が多くなるに従ってドライバは注意して自車を運転しなければならないため、記憶個数はドライバにかかる負担も示している。
【０１１３】
記憶個数は、現在時点から所定時間前までにドライバが見た障害物の個数が好ましい。したがって、運転行動危険度演算部３９は、時間の経過と共に、所定時間前に見た障害物の個数を順次消去すると共に、新たにドライバが見たと認識された障害物の個数を順次追加することによって、記憶個数を常に更新するようになっている。
【０１１４】
また、運転行動危険度演算部３９は、障害物の再確認に要した時間を考慮して記憶個数を設定することもできる。このとき、運転行動危険度演算部３９は、例えば障害物の再確認に要した時間が所定時間以上であった場合は、当該障害物の個数を記憶している時間を長くすればよい。
【０１１５】
このように運転行動危険度演算部３９は、ドライバの視線が移動する毎にドライバが記憶している障害物の個数を更新すると、図２６に示すように、記憶個数と閾値ｔｈとを比較する。運転行動危険度演算部３９は、記憶個数と閾値ｔｈとを比較した結果、記憶個数が閾値ｔｈ以上であるときは運転行動危険度ＤＲ３が高いと判定し、記憶個数が閾値ｔｈ未満であるときは運転行動危険度ＤＲ３が低いと判定する。
【０１１６】
なお、本実施の形態では閾値ｔｈを６個に設定しているが、これに限定されるものではない。閾値ｔｈは、ドライバの年齢などの個人差に応じて変えるのが好ましい。例えば、事故の起こしやすいドライバの場合は閾値ｔｈを５個に設定したり、長い間事故を起こしたことがないドライバの場合は閾値ｔｈを７個に設定してもよい。
【０１１７】
マイクロコンピュータ３０は、このような処理を経て、例えば記憶個数が閾値ｔｈ以上になって運転行動危険度ＤＲ３が高いと判定したときは、速度を下げるようにドライバに促す画面をＬＣＤ１９に表示したり、速度を下げることを促す警報をスピーカ２０から出力する。
【０１１８】
以上のように、第３の実施の形態に係る運転行動危険度演算装置は、注視領域における障害物の数を認識し、ドライバが障害物を記憶している記憶個数を推測し、人間の記憶特性を考慮してドライバの負荷が大きくなりすぎないかを判定することができる。そして、推測された記憶個数が閾値ｔｈ以上になると、運転行動危険度ＤＲ３が高くなっていると判定して、ドライバに車速の低下を促して、事故の可能性を大きく低減することができる。
【０１１９】
なお、本発明は、このような実施の形態に限定されるものではない。例えば、運転行動危険度演算部３９は、視線フィルタ３２の出力結果を用いてドライバの視線の移動を検出したが、例えば操舵角と車速とに基づいてドライバの視線の動きを推定してもよい。これにより、視線検出のための画像処理の負担を省略することができるので、運転行動危険度ＤＲ３の演算速度を速くすることができる。
【０１２０】
［第４の実施の形態］
つぎに、本発明の第４の実施の形態について説明する。なお、上述した実施の形態と同一の部位や回路に対しては同一の符号を付し、その詳細な説明は省略する。第４の実施の形態では、図２７に示す構成の運転行動危険度演算装置１Ａを例に挙げて説明する。
【０１２１】
本実施の形態に係る運転行動危険度演算装置１Ａは、データ計算装置４０との間でデータ通信を行って、運転行動危険度ＤＲ４を演算するものである。運転行動危険度演算装置１Ａは、具体的には、運転行動危険度演算装置１の構成回路に加えて、更に、データ演算装置４０との間でデータの送受信を行う通信回路２１を備えている。
【０１２２】
通信回路２１は、データバス５を介して供給されたデータに所定の変調処理を行って、通信アンテナ２１ａを介してデータ演算装置４０に送信する。また、通信回路２１は、データ演算装置４０からのデータを、通信アンテナ２１ａを介して受信し、所定の復調処理を行った後、復調処理済みのデータをマイクロコンピュータ３０に供給する。一方、データ演算装置４０は、アンテナ４０ａを介して、運転行動危険度演算装置１Ａとの間でデータの送受信を行うことができる。
【０１２３】
（パラメータテーブルの更新）
運転行動危険度演算装置１Ａは、データ演算装置４０から送信される情報に基づいて、常に最新のパラメータテーブルを保持することができる。
【０１２４】
ここで、データ演算装置４０は、車両毎に、図５から図９と同様に構成されたパラメータテーブルを記憶している。そして、データ演算装置４０は、パラメータテーブルが更新処理されると、更新されたパラメータテーブルのデータを、所定の車両の運転行動危険度演算装置１Ａに送信する。
【０１２５】
運転行動危険度演算装置１Ａは、パラメータテーブルをデータ演算装置４０から送信されたパラメータテーブルに更新し、常に最新の危険度パラメータを用いて環境危険度ＫＲを正確に演算することができる。この結果、運転行動危険度演算装置１Ａは、最新の環境状況に応じた運転行動危険度ＤＲ４を演算することができる。
【０１２６】
（データ演算装置４０による環境危険度ＫＲの演算）
データ演算装置４０は、運転行動危険度演算装置１Ａに代わって、対象物の種類に応じた環境危険度ＫＲを演算することもできる。このために、運転行動危険度演算装置１Ａ及びデータ演算装置４０は、以下の処理を実行する。
【０１２７】
運転行動危険度演算装置１Ａのマイクロコンピュータ３０は、第２の実施の形態と同様に、車両周囲の環境にある対象物を認識し、認識した対象物の種類毎に、車両の進行方向に対する向き、速度、大きさ、位置を認識する。そして、マイクロコンピュータ３０は、これらのデータを、通信回路２１を介してデータ演算装置４０に送信する。
【０１２８】
データ演算装置４０は、運転行動危険度演算装置１Ａから送信されたデータを受信し、受信したデータに基づいて対象物の種類に応じた環境危険度ＫＲを演算する。そして、データ演算装置４０は、演算された対象物の種類に応じた環境危険度ＫＲを示すデータを運転行動危険度演算装置１Ａに送信する。このとき、データ演算装置４０は、環境危険度ＫＲを演算する毎に、図１０と同様にして、危履歴データを記憶するのが好ましい。
【０１２９】
運転行動危険度演算装置１Ａのマイクロコンピュータ３０は、通信回路２１を介して対象物の種類に応じた環境危険度ＫＲを受信すると、環境危険度ＫＲを用いて第２の実施の形態と同様にして運転行動危険度ＤＲ４を演算する。
【０１３０】
これにより、運転行動危険度演算装置１Ａは、データ量の多いパラメータテーブルを記憶する必要もなく、また、環境危険度ＫＲの演算処理を行う必要がないので、その演算処理のための負担を少なくすることができる。
【０１３１】
なお、本発明は、上述した実施の形態に限定されるものではなく、特許請求の範囲に記載された範囲内で、種々の設計上の変更を行うことができる。
【０１３２】
例えば上述した実施の形態では、対象物、向き、速度、大きさ、位置のそれぞれに危険度パラメータを付していたが、その他、地域毎に又は時間毎に危険度パラメータを付すこともできる。
【０１３３】
また、上述した各危険度パラメータは、本発明の一例であり、これに限定されるものではない。したがって、パラメータテーブルとしては、図５から図９に示したものに限定されず、例えば、自車の車両位置、前方信号の色、天候、環境の明るさに応じてパラメータテーブルを設けてもよいのは勿論である。
【０１３４】
【発明の効果】
本発明に係る運転行動危険度演算装置は、視行動検出手段で検出された視行動と、車両状態計測手段で計測された車両状態と、に基づいて運転行動危険度を演算して出力することによって、運転手が確認している場所や方向と運転手が実際に運転行動している車両状態とを考慮して、運転手の運転行動による危険度を正確に求めることができる。
【０１３５】
本発明に係る運転行動危険度演算装置は、環境危険度演算手段で演算された対象物の種類に応じた危険度と、視行動検出手段で検出された運転者の視行動と、に基づいて運転行動危険度を演算して出力することにより、車両周囲の環境と運転者の確認状態を総合的に考慮した運転行動危険度を求めることができる。
【０１３６】
本発明に係る運転行動危険度演算装置は、環境認識手段で認識された対象物と、視行動検出手段で検出された運転者の視行動と、に基づいて運転者が記憶する対象物の個数を推定し、推定された対象物の個数に基づいて運転行動危険度をして出力して演算することにより、多くの対象物を把握しながら運転するときの運転負担を示す運転行動危険度を演算することができる。
【図面の簡単な説明】
【図１】本発明の第１の実施の形態に係る運転行動危険度演算装置の構成を示すブロック図である。
【図２】運転行動危険度演算装置に備えられたマイクロコンピュータの機能的な構成を示すブロック図である。
【図３】マイクロコンピュータに記憶された視行動テーブルの構成を示す図である。
【図４】本発明の第２の実施の形態に係る運転行動危険度演算装置に備えられたマイクロコンピュータの機能的な構成を示すブロック図である。
【図５】環境危険度演算装置に備えられたマイクロコンピュータに記憶された対象物パラメータテーブルの構成を示す図である。
【図６】マイクロコンピュータに記憶された向きパラメータテーブルの構成を示す図である。
【図７】マイクロコンピュータに記憶された速度パラメータテーブルの構成を示す図である。
【図８】マイクロコンピュータに記憶された大きさパラメータテーブルの構成を示す図である。
【図９】マイクロコンピュータに記憶された位置パラメータテーブルの構成を示す図である。
【図１０】危険度の履歴データの構成を示す図である。
【図１１】運転行動危険度演算部の演算内容を模式的に示した図である。
【図１２】最終画像の基礎となる車外撮像画像の一例を示す図である。
【図１３】車外撮像画像にかけられたマスクの状態を説明する図である。
【図１４】１秒後の到達距離（実線）、２秒後の到達距離（点線）及び１秒後の自車の進行方向（矢印）を示す図である。
【図１５】車外撮像画像に画像処理を施すことによって生成された最終画像を示す図である。
【図１６】マイクロコンピュータによって演算された運転行動危険度の結果を示す図である。
【図１７】Ａ点における車外撮像画像（Ａ）と、Ａ点におけるドライバ用撮像画像（Ｂ）を示す図である。
【図１８】Ｂ点における車外撮像画像（Ａ）と、Ｂ点におけるドライバ用撮像画像（Ｂ）を示す図である。
【図１９】Ｃ点における車外撮像画像（Ａ）と、Ｃ点におけるドライバ用撮像画像（Ｂ）を示す図である。
【図２０】Ｄ点における車外撮像画像（Ａ）と、Ｄ点におけるドライバ用撮像画像（Ｂ）を示す図である。
【図２１】本発明の第３の実施の形態におけるマイクロコンピュータの機能的な構成を示すブロック図である。
【図２２】上空から自車を見たときのドライバの横方向の視線の向きを分類した図である。
【図２３】車外用ＣＣＤカメラにより撮像された車外撮像画像の模式的な図である
【図２４】車外撮像画像とドライバの視線の向きを示す図である。
【図２５】ドライバの視線移動順序、ドライバが現在見ている障害物の個数及びそのときの記憶個数の関係を示す図である。
【図２６】ドライバが記憶している障害物の個数である記憶個数と閾値ｔｈとを比較する状態を説明する図である。
【図２７】本発明の第４の実施の形態に係る運転行動危険度演算装置の構成を示すブロック図である。
【符号の説明】
１，１Ａ 運転行動危険度演算装置
１３ 車外用ＣＣＤカメラ
１５ マスタシリンダ油圧センサ
１６ 操舵角センサ
１７ 車輪速センサ
１８ ドライバ用ＣＣＤカメラ
１９ ＬＣＤ
２０ スピーカ
３０ マイクロコンピュータ[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a driving behavior risk calculation device, and more particularly to a driving behavior risk calculation device that calculates the risk of an accident or the like based on a driver's line of sight or movement.
[0002]
[Prior art and problems to be solved by the invention]
Japanese Laid-Open Patent Publication No. 6-215300 proposes a driving status warning device (hereinafter referred to as “Prior Art 1”) that issues a warning in accordance with the driver's intention and the driving status of the vehicle. Prior art 1 focuses on detecting the road environment such as the position of an obstacle, and further estimates the driver's driving intention based on the steering wheel operation and the vehicle speed. And the prior art 1 predicts the area | region where a dangerous substance may exist from the detected road environment and the driver's estimated intention, and confirms whether there is an obstacle in the predicted area | region. Alarms are given as necessary.
[0003]
However, since the prior art 1 handles and detects pedestrians, vehicles, and other obstacles in the same manner and issues a warning to the driver, the warning may be issued even when the driver thinks that it is not dangerous. That is, there is a problem that the driver feels bothered when the case where the prior art 1 issues an alarm and the case where the driver determines that the driver is subjectively dangerous.
[0004]
Moreover, the prior art 1 is intended for an automobile-only road or a main road, and is not suitable for a general road used by both vehicles and pedestrians. Therefore, there is a problem that it is not possible to cope with a sudden jump of a vehicle or a person on a general road.
[0005]
In Japanese Patent Laid-Open No. 7-167668, a travel information providing device (hereinafter referred to as “Prior Art 2”) that reliably notifies appropriate information regarding a situation that the driver is not aware of while reducing the driver's trouble as much as possible. Proposed. Prior art 2 determines whether the driver is paying sufficient attention to the obstacle, but in actual driving scenes, it is common to change the confirmation time for each gaze target, and the obstacle is uniquely detected. There is little effect to prevent accidents.
[0006]
Japanese Laid-Open Patent Publication No. 2000-262897 proposes a safe driving evaluation apparatus (hereinafter referred to as “Prior Art 3”) that prompts a driver to recognize the evaluation result of driving details and promotes safe driving. Although the prior art 3 detects the direction of the driver's line of sight and uses it for driving evaluation, it can detect only a side-view driving. Therefore, the prior art 3 determines that the driver is “safe” if the driver is not looking aside, and does not issue an alarm.
[0007]
Moreover, since the prior art 3 is not suitable for a general road used by both vehicles and pedestrians, there is a problem that it is impossible to deal with a sudden jump of a vehicle or a person on the general road.
[0008]
The present invention has been proposed in order to solve the above-described problem, and it is possible to estimate a driver's recognition state with respect to an obstacle and accurately calculate the risk level when driving the vehicle. An object is to provide an arithmetic device.
[0009]
[Means for Solving the Problems]
The invention described in claim 1 is a visual behavior detecting means for detecting a visual behavior of a driver, a vehicle state measuring means for measuring a vehicle state, a visual behavior detected by the visual behavior detecting means, and the vehicle state measurement. Driving behavior risk calculation means for calculating the driving behavior risk based on the vehicle state measured by the means, and driving behavior risk output for outputting the driving behavior risk calculated by the driving behavior risk calculation means Means.
[0015]
The invention according to claim 1 recognizes an object in the environment around the vehicle, and determines the direction, speed, and relative position of the object with respect to the traveling direction of the vehicle. Place Detecting environment recognition means and object recognized by the environment recognition means Every time the object Risk parameters related to the possibility of accident triggers determined in advance according to the type of the object, and the direction, speed, and position of the object detected by the environment recognition means In place The environmental risk level calculating means for calculating the risk level determined from the risk level parameter determined in advance, the visual behavior detecting means for detecting the driver's visual behavior, and the target calculated by the environmental risk level calculating means. object Every environment Risk level and driver's visual behavior detected by the visual behavior detection means The weighting factor according to the visual behavior of the driver with respect to the object determined for each object based on Driving risk calculation means for calculating the driving action risk level based on the driving action risk degree output means for outputting the driving action risk degree calculated by the driving action risk degree calculation means. .
[0017]
The environmental risk level calculation means includes a risk parameter related to the possibility of an accident trigger predetermined according to the type of the object, and the direction, speed, and position of the object detected by the environment recognition means. Set From the risk parameters determined in advance according to each, environment Calculate the risk. The visual behavior detection means detects the direction of the driver's line of sight and eye movement based on, for example, an image of the driver.
[0018]
The driving behavior risk calculation means is an object calculated by the environmental risk calculation means. Every environment By using the degree of risk and the driver's visual behavior detected by the visual behavior detector, the driving behavior risk is calculated in consideration of the environment around the vehicle and the driver's confirmation state.
[0019]
The driving action risk level output means is not particularly limited as long as it can call attention to a person, and it is only necessary to stimulate the human senses according to the risk level. In particular, it is preferable to output images and sounds. This can alert the driver and prevent traffic accidents.
[0020]
The invention according to claim 2 is the visual behavior of the driver detected by the visual behavior detection means in the invention of claim 1. A weighting factor according to the visual behavior of the driver with respect to the object determined for each object based on In the environmental risk calculation means Calculation Of each object environment It is characterized by calculating the driving behavior risk level by weighting the risk level and adding the environmental risk level of each weighted object.
[0021]
According to a second aspect of the present invention, the driving behavior risk calculating means is the environmental risk calculating means according to the driver's visual behavior detected by the visual behavior detecting means. Calculation Of each object environment Weight the risk. For example, when the line of sight of the object is present, the weight of the risk of the object is reduced, and when the line of sight of the object is not, the weight of the risk of the object is increased. Further, the weight may be changed according to the gaze time of the object. For example, the weight may be decreased when the gaze time is long, and may be increased when the gaze time is short.
[0022]
Then, by adding the environmental risk level of each weighted object, it is possible to calculate not only the level of the environmental risk level for the target object but also the driving action risk level considering the driver's attention. For example, even if there is an object with a high environmental risk, if the driver is gazing at the object, the weighting is reduced, so the driving action risk is low. On the contrary, even if there is an object whose environmental risk is not so high, if the driver does not see the object, the weight is increased, so that the driving action risk becomes high. Therefore, it is possible to calculate a driving behavior risk level that comprehensively considers the environmental risk level of each object and the driver's attention state.
[0023]
Claim 3 The described invention is claimed. 2 In the described invention, further comprising vehicle speed detecting means for detecting a vehicle speed, the driving behavior risk calculating means calculates an intermediate risk by adding an environmental risk of each of the weighted objects, and is calculated. The driving behavior risk level is calculated based on the intermediate risk level and the vehicle speed detected by the vehicle speed detecting means.
[0024]
Claim 3 In the described invention, the driving behavior risk calculation means comprehensively considers the environmental risk of each object and the driver's attention state by adding the weighted environmental risks of each object. The calculated intermediate risk is calculated. Then, the driving behavior risk level is calculated based on the calculated intermediate risk level and the vehicle speed detected by the vehicle speed detecting means. Considering the vehicle speed, it is difficult to avoid obstacles when the vehicle speed is high, and there is room to avoid obstacles when the vehicle speed is slow. Because sex changes. Thereby, not only the environmental risk level of each object and the driver's attention state but also the driving state of the vehicle can be taken into consideration, and the driving behavior risk level can be calculated from a more comprehensive viewpoint.
[0029]
Claim 4 The invention described in claim 1 to claim 1 3 The vehicle traveling direction measured by the vehicle state measuring means is further provided with vehicle state measuring means for measuring the traveling direction of the vehicle. And the visual behavior detected by the visual behavior detection means, the driver calculates at least one of the time and frequency of watching the traveling direction of the vehicle, and calculates at least one of the calculated time and frequency. The driving behavior risk is calculated based on one.
Claim 5 The invention described in claim 1 to claim 1 3 The visual behavior detection means for classifying the visual behavior detected by the visual behavior detection means to estimate the driver's attention state, wherein the driving behavior risk calculation The means further calculates a driving behavior risk level using the attention state estimated by the visual attention state detection means.
[0030]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.
[0031]
[First Embodiment]
FIG. 1 is a block diagram showing a configuration of a driving behavior risk calculation device 1 according to the first embodiment of the present invention.
[0032]
The driving behavior risk calculation device 1 includes a GPS signal receiving circuit 11 that receives a GPS (Global Positioning System) signal, a DVD drive 12 that reads map information recorded on a DVD disc, and obstacles on and around the road. And a radar device 14 for recognizing a front obstacle by an infrared radar.
[0033]
The GPS signal receiving circuit 11 receives a GPS signal having time and GPS satellite position information via the GPS antenna 11 a and supplies the GPS signal to the microcomputer 30 via the data bus 5. The DVD drive 12 reads the map information from the DVD disk based on the position information where the vehicle is currently traveling, and supplies it to the microcomputer 30 via the data bus 5.
[0034]
The on-vehicle CCD camera 13 is installed so that an object in front of the vehicle can be imaged. Then, the outside CCD camera 13 supplies the captured outside image to the microcomputer 30 via the data bus 5. In order to recognize a front obstacle, the radar device 14 sharply squeezes a pulsed infrared radar with respect to the obstacle and transmits it in a two-dimensional direction, and receives the infrared radar reflected by the obstacle. The radar device 14 is not limited to transmitting and receiving infrared radars, but may be one that transmits and receives radio wave radars.
[0035]
Further, the driving behavior risk calculation device 1 includes a master cylinder hydraulic sensor 15 that detects a brake hydraulic pressure in a master cylinder (not shown), a steering angle sensor 16 that detects a steering angle, and a wheel that detects a rotational angular velocity of the vehicle. A speed sensor 17, a driver CCD camera 18 that captures the face of the driver, an LCD (Liquid Crystal Display) 19 that displays the environmental risk level, a speaker 20 that outputs the environmental risk level by voice, and a risk level And a microcomputer 30 for performing calculations and other controls.
[0036]
The master cylinder oil pressure sensor 15 detects the brake oil pressure in the master cylinder according to the depression force of the brake pedal of the driver, and supplies it to the microcomputer 30 via the data bus 5. The steering angle sensor 16 generates a steering angle signal corresponding to the rotation angle of a steering shaft (not shown) and supplies the steering angle signal to the microcomputer 30 via the data bus 5.
[0037]
Although the wheel speed sensor 17 is provided on each wheel of the vehicle, only one is shown in FIG. Each wheel speed sensor 17 generates a wheel speed signal according to the rotation of a rotor made of a magnetic material having an outer peripheral portion formed in a tooth shape, and supplies the wheel speed signal to the microcomputer 30 via the data bus 5.
[0038]
The driver CCD camera 18 is set in the vehicle so that the driver's face can be imaged from the right front or the left front when the driver is facing the front. Although only one driver CCD camera 18 may be provided, two driver CCD cameras 18 may be provided so that the driver's face can be imaged from the right front and the left front so that the driver's face can be easily recognized. The driver CCD camera 18 and the outside CCD camera 13 can capture images synchronously, and generate a captured image of 30 frames per second.
[0039]
The microcomputer 30 includes a CPU (Central Processing Unit) (not shown), a RAM (Random Access Memory) that functions as a data work area and temporarily stores necessary calculation results, and a program that executes predetermined processing. And a ROM (Read Only Memory) in which a plurality of tables are stored.
[0040]
(Functional configuration of microcomputer 30)
The microcomputer 30 is functionally configured as shown in FIG. 2, and performs a visual behavior detection unit 31 that detects the visual behavior of the driver based on the driver captured image, and classifies the visual behavior of the driver. A line-of-sight filter 32 and a driving behavior risk calculating unit 33 that calculates the driving behavior risk DR1 are provided.
[0041]
The visual behavior detector 31 detects, for example, the direction of the line of sight and eye movement as the visual behavior of the driver based on the driver captured image captured by the driver CCD camera 18. The visual behavior detection process is not particularly limited. For example, in the document “5th Robotics Symposia Proceedings, p220-p225, 2000”, “Development of Face / Gaze Measurement System and Application to Motion Recognition ( Yoshio Matsumoto, Alexander Zenlinsky, Tsukasa Ogasawara) ”can be used.
[0042]
The line-of-sight filter 32 classifies the driver's visual behavior detected by the visual behavior detector 31 into five types: gaze, blink, saccade eye movement, following eye movement, and convergence movement. The line-of-sight filter 32 also detects the gaze time when gaze is detected as the visual behavior. The line-of-sight filter 32 supplies the result obtained from the driver's visual behavior to the driving behavior risk calculating unit 33.
[0043]
The driving behavior risk calculation unit 33 determines the traveling direction of the host vehicle based on the steering angle detected by the steering angle sensor 16. Then, using the line-of-sight behavior classified by the line-of-sight filter 32, the degree of driving behavior risk in consideration of how long and how often the traveling direction of the vehicle is viewed from the present to a predetermined time ago. Calculate DR1. For example, when the traveling direction is viewed for a predetermined reference time or more, or the traveling direction is viewed for a predetermined reference number or more, the driving behavior risk DR1 is low. On the other hand, when the traveling direction has not been viewed for a predetermined reference time or more, or the traveling direction has not been viewed for a predetermined reference number of times or more, the driving behavior risk DR1 becomes high.
[0044]
Furthermore, the driving action risk degree calculation unit 33 can increase the driving action risk degree DR1 when the vehicle speed is equal to or higher than a predetermined reference speed or the brake hydraulic pressure is equal to or higher than a predetermined reference pressure. Conversely, the driving action risk degree calculation unit 33 can reduce the driving action risk degree DR1 when the vehicle speed is less than a predetermined reference speed or the brake hydraulic pressure is less than a predetermined reference pressure.
[0045]
In addition, the driving behavior risk calculating unit 33 estimates the driver's attention state using the line-of-sight behavior classified by the line-of-sight filter 32, and calculates the driving behavior risk DR1 in consideration of the estimated attention state. You can also. Here, the driving behavior risk calculating unit 33 stores a visual behavior table for specifying parameters according to the classification of visual behavior. As shown in FIG. 3, the visual behavior table gives parameters P1, P2, P3, P4, and P5 to gaze, blink, saccade eye movement, following eye movement, and convergence movement.
[0046]
Then, the driving behavior risk level calculation unit 33 calculates the following formula (1) to obtain the risk level DR.
[0047]
[Expression 1]

[0048]
Here, a, b, c, d, and e are weighting coefficients corresponding to the respective parameters. Each weighting coefficient differs, for example, as follows according to the classification of visual behavior and time. For example, the weighting coefficient a is small when the gaze time is long and large when the gaze time is short. Further, the weighting coefficient b increases when the blinking time is long, and decreases when the blinking time is short. The weighting coefficient c increases when the saccade eye movement time is long, and decreases when the saccade eye movement time is short. The weighting coefficient d decreases when the tracking eye movement is long, and increases when the tracking eye movement is short. The weighting coefficient e increases when the time during which the converging movement is performed is long, and decreases when the time during which the following eye movement is performed is short.
[0049]
Then, the driving behavior risk calculating unit 33 obtains a risk DR considering the driver's attention state according to the equation (1), and further calculates the driving behavior risk DR1 considering the vehicle state. For example, by increasing the risk DR as the vehicle speed increases, increasing the risk DR as the steering angle increases, or decreasing the risk DR as the brake hydraulic pressure increases, the driving action risk DR1 is increased. Ask.
[0050]
As described above, the driving behavior risk calculation device 1 according to the first embodiment estimates the attention state by classifying the driver's visual behavior, and based on the driver's attention state and the vehicle state such as the vehicle speed. By calculating the driving behavior risk DR1, it is possible to determine the risk considering the driver's attention state and the driver's actual driving behavior comprehensively.
[0051]
[Second Embodiment]
Next, a second embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected about the site | part same as 1st Embodiment, and the overlapping description is abbreviate | omitted. In the present embodiment, the microcomputer 30 calculates the driving behavior risk DR2 using a calculation method different from that of the first embodiment.
[0052]
The microcomputer 30 according to the present embodiment is functionally configured as shown in FIG. That is, the microcomputer 30 includes a visual behavior detection unit 31 that detects the visual behavior of the driver based on the captured image for the driver, a visual line filter 32 that detects blinking time based on the visual behavior of the driver, and a captured image outside the vehicle. And an object detection unit 34 for detecting the object from.
[0053]
Further, the microcomputer 30 calculates the environmental risk degree KR, a position detection part 35 that detects the current position of the vehicle based on the GPS signal, an environmental risk degree calculation part 36 that calculates the environmental risk degree KR for each object. And an environmental risk storage unit 37 that stores the calculated environmental risk KR, and a driving behavior risk calculation unit 38 that calculates the driving behavior risk DR2.
[0054]
(Parameter table)
Here, the environmental risk degree storage unit 37 stores an object parameter table, an orientation parameter table, a speed parameter table, a size parameter table, and a position parameter table.
[0055]
As shown in FIG. 5, the object parameter table is composed of object types (categories), sub-classification of each type of object, and risk parameters for each sub-classification. The risk parameter refers to a value related to the possibility of accident incentives, for example, a value obtained in consideration of accident statistics, insurance premium calculation data, driving norm data, and the like.
[0056]
Examples of the types of objects include pedestrians, two-wheeled vehicles, and automobiles. The subcategories of pedestrians include elderly people, children, junior high school students and adults. Bicycles and motorcycles are classified into two-wheeled vehicles. There are two types of automobiles: ordinary cars and large cars. Elderly persons, children, junior high school students-adults, bicycles, motorcycles, ordinary cars, and large cars are assigned risk parameters t1, t2,.
[0057]
In general, the risk level decreases in the order of pedestrians, motorcycles, and automobiles. Specifically, it is considered that the risk levels decrease in the order of elderly people, children, junior high school students-adults, bicycles, motorcycles, ordinary cars, and large cars. . That is, elderly pedestrians have the highest degree of danger and large vehicles have the lowest degree of danger. Therefore, the risk parameter is t1>t2>...> T7.
[0058]
As shown in FIG. 6, the orientation parameter table is composed of the orientation of the object and the risk parameter for each orientation. As the direction of the object, there are orthogonal to the traveling direction of the vehicle, the same direction as the traveling direction of the vehicle, and facing the traveling direction of the vehicle. Risk parameters m1, m2, and m3 are assigned to the orthogonal, the same direction, and the opposite, respectively. Also, the degree of risk decreases in the order of orthogonal, same direction, and facing. Therefore, the risk parameter is m1>m2> m3.
[0059]
As shown in FIG. 7, the speed parameter table is composed of the speed of the object and the risk parameter for each speed. The speed of the object includes high speed, walking speed (for example, around 4 km / h) and stop. Risk parameters s1, s2, and s3 are assigned to high speed, walking speed, and stop, respectively. Moreover, the danger level is low in the order of high speed, walking speed, and stop. Therefore, the risk parameter is s1>s2> s3.
[0060]
As shown in FIG. 8, the size parameter table is composed of the size of the object and the risk parameter for each size. The size of the object is not particularly limited, and risk parameters d1, d2, d3,... Are assigned for each size of the object. It should be noted that a “magnitude” with a high risk level is assigned a large risk parameter, and a “magnitude” with a low risk level is assigned a small risk parameter.
[0061]
Further, as shown in FIG. 9, the position parameter table is composed of the position of the object and the risk parameter for each position. The position of the object is not particularly limited, and risk parameters p1, p2, p3,... Are assigned for each position of the object. It should be noted that a “position” with a high degree of risk is assigned a risk parameter with a large value, and a “position” with a low degree of risk is assigned a risk parameter with a small value.
[0062]
(Calculation of environmental risk KR)
The microcomputer 30 calculates the environmental risk KR for each type of environmental object around the vehicle as follows.
[0063]
The object detection unit 34 performs predetermined image processing such as pattern matching on the outside-captured image obtained by the outside CCD camera 13, and extracts and detects the target from the outside-captured image. Further, the object detection unit 34 recognizes the relative distance and speed to the object based on the round trip time of the infrared laser until it is emitted from the radar device 14 and received. And the object detection part 34 detects the direction with respect to the advancing direction of a vehicle, speed, a magnitude | size, and a position (relative position with respect to the own vehicle) for every kind of object based on these recognition results, The detection result is supplied to the environmental risk calculation unit 36.
[0064]
In addition, you may make a pedestrian hold | maintain the transmitter which transmits ID signal, or such a transmitter may be attached to a two-wheeled vehicle or a motor vehicle. In this case, the object detection unit 34 can detect pedestrians, two-wheeled vehicles, and automobiles based on the ID signals transmitted from the transmitters.
[0065]
The position detection unit 35 detects the current position of the host vehicle based on the GPS signal received by the GPS signal receiving circuit 11 and supplies the current position of the host vehicle to the environmental risk degree calculation unit 36.
[0066]
The environmental risk level calculation unit 36 calculates the risk level for each type of object. Specifically, the environmental risk level calculation unit 36 refers to the tables shown in FIGS. 5 to 9 stored in the environmental risk level storage unit 37, and is further detected by the object detection unit 34 and the position detection unit 35. Using the obtained information, risk parameters corresponding to the object, orientation, speed, size, and position are selected. Then, using the selected risk parameter, a function F indicating the environmental risk KR corresponding to the type of the object is calculated. The risk parameter is, for example, a ₁ , A ₂ , A _Three ,..., The function indicating the environmental risk KR is F (a ₁ , A ₂ , A _Three ,... In the present embodiment, the function indicating the environmental risk KR can be expressed as, for example, Expression (2).
[0067]
[Expression 2]

[0068]
In Equation (2), t is a risk parameter for the type of object, m is a risk parameter for direction, s is a risk parameter for speed, d is a risk parameter for magnitude, and p is a risk parameter for position. is there. A, B, C, D, and E are predetermined weighting coefficients.
[0069]
The environmental risk level calculation unit 36 calculates the environmental risk level KR for each target according to the equation (2), and supplies the environmental risk level KR for each target to the driving behavior risk level calculation unit 38. The environmental risk level calculation unit 36 stores history data of the environmental risk level KR in the environmental risk level storage unit 37 each time the environmental risk level KR is calculated.
[0070]
FIG. 10 is a diagram showing history data of the environmental risk KR calculated for each type of object. The history data includes the date and time when the environmental risk KR is calculated, the environmental risk KR, and the risk parameters (object, orientation, speed, size, and position) used during the calculation.
[0071]
Thereby, the environmental risk degree calculation unit 36 is based on the detected object in the environment around the vehicle, attributes such as the direction and speed of the object detected using the radar device 14, and the risk parameter. Thus, the environmental risk KR can be calculated for each type of movable object such as a pedestrian or a car. At this time, since not only the risk parameter of the object type but also the risk parameters of the direction, speed, size, and position of the object are used, the overall environmental risk KR of the object is obtained. Can do.
[0072]
(Calculation of driving behavior risk DR)
The visual behavior detector 31 detects, for example, the direction of the line of sight and eye movement as the visual behavior of the driver based on the driver captured image captured by the driver CCD camera 18.
[0073]
The line-of-sight filter 32 classifies the driver's visual behavior detected by the visual behavior detector 31 into five types: gaze, blink, saccade eye movement, following eye movement, and convergence movement. The line-of-sight filter 32 also detects the gaze time when gaze is detected as the visual behavior. The line-of-sight filter 32 supplies the result obtained from the driver's visual behavior to the driving behavior risk calculating unit 38.
[0074]
The driving behavior risk calculation unit 38 calculates the environmental risk KR for each object calculated by the environmental risk calculation unit 36. _n On the other hand, based on the detection result of the line-of-sight filter 32, the weighting coefficient E _n Further, the driving behavior risk DR2 is calculated using the vehicle speed v of the host vehicle. Hereinafter, calculation of the driving behavior risk DR2 will be described.
[0075]
FIG. 11 is a diagram schematically showing the calculation contents of the driving behavior risk calculation unit 38. The driving behavior risk calculation unit 38 first calculates the environmental risk KR calculated by the environmental risk calculation unit 36. _n The weighting factor E corresponding to _n Generate each environmental risk KR _n Is weighted.
[0076]
Weighting factor E _n Is a parameter determined by the output of the line-of-sight filter 32. Specifically, the value is small when the object has the driver's line of sight, and the value is large when the object does not have the driver's line of sight. Also, the weighting coefficient E _n The value decreases when the time of gazing at the object is long, and increases when the time of gazing at the object is short.
[0077]
When the object has a driver's line of sight, the weighting coefficient E is further determined according to the classification of gaze, blink, saccade eye movement, following eye movement, and convergence movement. _n The value of may be changed. For example, in the case of gaze and tracking eye movement, the weighting coefficient E _n In the case of blinking, saccade eye movement and vergence movement, the weighting coefficient E is reduced. _n The value of can also be increased.
[0078]
The driving behavior risk calculation unit 38 is the environmental risk KR of each object calculated by the environmental risk calculation unit 36. _n And the environmental risk KR _n The weighting factor E corresponding to _n Is calculated to obtain an intermediate risk MR shown in Equation (3).
[0079]
[Equation 3]

[0080]
Next, the driving behavior risk calculation unit 38 calculates the vehicle speed v of the host vehicle based on the wheel speed signal detected by each wheel speed sensor 17, and multiplies the vehicle speed v of the host vehicle by the intermediate risk MR. Then, the driving behavior risk DR2 is obtained by performing psychological conversion on v · MR. Here, when the function indicating the psychological amount conversion is g, the driving behavior risk DR2 is expressed by the following equation (4).
[0081]
[Expression 4]

[0082]
In addition, the driving action risk degree calculation unit 38 can also obtain the driving action risk degree DR2 without using the function g. For example, a map indicating the relationship of the driving behavior risk DR2 when v · MR is psychologically converted in advance may be stored in the memory, and the driving behavior risk DR2 may be obtained with reference to this map.
[0083]
(Final image generation process)
The driving behavior risk calculation unit 38 not only performs the calculation processing of the driving behavior risk DR2, but also performs final image generation processing based on the vehicle outside captured image, the vehicle speed, the steering angle, and the like. Here, generation of a final image based on the outside-vehicle captured image shown in FIG. 12 will be described.
[0084]
First, based on the detection result of the line-of-sight filter 32, the driving action risk degree calculation unit 38 performs a process of masking a place where the line of sight is not directed on the outside-vehicle captured image. FIG. 13 is a diagram for explaining the state of the mask applied to the outside-vehicle captured image.
[0085]
FIG. 13A is a diagram illustrating a mask when the driver is gazing forward. When the driver is gazing at the front, the driver does not look at both ends of the gazing area. Therefore, the driving behavior risk calculation unit 38 performs a mask process so that both ends of the captured image cannot be seen.
[0086]
FIG. 5B is a diagram showing the mask when the driver is gazing at the front right side. When the driver is gazing at the front right side, he / she does not see much other than the gazing place (front and left sides). Therefore, the driving behavior risk calculation unit 38 performs a mask process so that the front and left sides of the captured image cannot be seen. Note that when the driver is gazing at the left front, the driving behavior risk calculation unit 38 may perform the reverse process, that is, perform the mask process so that the front and right sides of the captured image cannot be seen.
[0087]
FIG. 6C is a diagram showing a mask when the driver is blinking. When the driver is blinking, the driver does not see much of the front. Therefore, the driving behavior risk calculation unit 38 performs a mask process so that the entire captured image cannot be seen. Note that the driving behavior risk degree calculation unit 38 may adjust the mask processing density according to the blinking frequency. For example, the mask density may be decreased when the number of blinks per unit time is less than a predetermined value, or the mask density may be increased when the number of blinks per unit time is greater than or equal to a predetermined value.
[0088]
Next, the driving behavior risk calculation unit 38 calculates the reach distance of the host vehicle after 1 second and 2 seconds after based on the vehicle speed. Further, the driving behavior risk calculation unit 38 calculates the traveling direction of the host vehicle after 1 second based on the steering angle detected by the steering angle sensor 16. FIG. 14 is a diagram illustrating an arrival distance after 1 second (solid line), an arrival distance after 2 seconds (dotted line), and a traveling direction (arrow) of the host vehicle after 1 second.
[0089]
Then, the driving behavior risk calculation unit 38 performs mask processing on the captured image outside the vehicle according to the line of sight of the driver, and further reaches the arrival distance after 1 second, the arrival distance after 2 seconds, and the host vehicle after 1 second. The final image is generated by superimposing the traveling directions. That is, the final image is generated by superimposing the outside-vehicle captured image of FIG. 12, any one of the images of FIGS. 13A to 13C, and the image of FIG.
[0090]
FIG. 15 is a diagram illustrating a final image generated by performing the above-described image processing on a captured image outside the vehicle. 15A is generated when the driver is gazing forward, FIG. 15B is generated when the driver is gazing forward, and FIG. 15C is generated when the driver is blinking. FIG.
[0091]
FIG. 16 is a diagram illustrating the result of the driving behavior risk (0 to +1) calculated by the microcomputer 30. The horizontal axis indicates a frame number, and this frame number indicates a captured image for each frame captured by the outside CCD camera 13 and the driver CCD camera 18. 30 frames corresponds to 1 second. Therefore, the frame numbers 30, 60, 90... Correspond to 1 second, 2 seconds, 3 seconds.
[0092]
In FIG. 16, at point A, 4 frames 10 seconds, DR2 = 0.54. At point B, 19 seconds 26 frames, DR2 = 0.05. At point C, 22 seconds 22 frames, DR2 = 1.00. At point D, 24 seconds and 11 frames, DR2 = 0.60.
[0093]
FIG. 17 is a diagram showing a captured image (A) outside the vehicle at point A and a captured image for driver (B) at point A. At this time, the vehicle speed is 4.7 [km / h], and the driver is gazing at the left front. The driver is gazing around the car on the left side, but because there are pedestrians on the right side, the driving behavior risk DR2 is a little higher (0.54).
[0094]
FIG. 18 is a diagram illustrating an outside-vehicle captured image (A) at point B and a driver-captured image (B) at point B. At this time, the vehicle speed is 0.0 [km / h], and the driver visually observes the left. The driver has not confirmed the pedestrian ahead because he / she looks at the left, but since the vehicle is completely stopped, the driving action risk DR2 has become a low value (0.05).
[0095]
FIG. 19 is a diagram illustrating a captured image (A) outside the vehicle at point C and a captured image for driver (B) at point C. At this time, the vehicle speed is 1.0 [km / h], and the driver visually observes the left. Since the driver did not confirm the pedestrian ahead by visually observing the left and the vehicle was moving, the driving behavior risk DR2 reached the highest value (1.00).
[0096]
FIG. 20 is a diagram showing a captured image (A) outside the vehicle at point D and a captured image for driver (B) at point D. At this time, the vehicle speed is 4.1 [km / h], and the driver is gazing at the right front. The driver looks at the right front and confirms whether a car or the like is coming from the right side, but since the two pedestrians in front are not confirmed, the driving behavior risk DR2 (0.60) is a little high became.
[0097]
(Final processing)
When the microcomputer 30 calculates the driving behavior risk DR2 as described above and further generates a final image, the microcomputer 19 causes the LCD 19 to display the final image and the driving behavior risk DR2. Thereby, the driver can recognize which place he / she is paying attention to by confirming the final image displayed on the LCD 19. In addition, the driver can recognize the current objective risk level by checking the driving action risk level DR2, and drive with care so that an accident does not occur if the risk level is high. be able to.
[0098]
The microcomputer 30 determines that there is a possibility of an accident when the value of the driving behavior risk DR2 is greater than a predetermined threshold value, and emits a warning sound from the speaker 20 to alert the driver. It may be.
[0099]
As described above, the driving behavior risk calculation device 1 according to the second embodiment not only calculates the risk KR of an object in the environment around the vehicle, but also how much the driver is careful about the object. By calculating the driving action risk level DR2 in consideration of whether the vehicle is driving, it is possible to comprehensively determine the surrounding environment of the vehicle and the operation of the driver, and to calculate a more objective risk level. Then, the driver's attention can be urged by displaying the driving behavior risk DR2.
[0100]
Further, the driving behavior risk calculation device 1 can visually recognize a place where the driver is careful by displaying the final image generated according to the line of sight. As a result, the driving behavior risk calculation device 1 can prevent both accidents caused by the surrounding environment of the vehicle and accidents caused by driver's carelessness.
[0101]
[Third Embodiment]
Next, a third embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected about the site | part same as embodiment mentioned above, and the overlapping description is abbreviate | omitted. In the present embodiment, the microcomputer 30 calculates the driving behavior risk DR3 by using a method different from the first and second embodiments.
[0102]
The microcomputer 30 is functionally configured as shown in FIG. 21, and detects a visual behavior detection unit 31 that detects the visual behavior of the driver based on the driver captured image, a blinking time of the driver, and the like. A line-of-sight filter 32, an object detection unit 34 that detects an object from a captured image outside the vehicle, and a driving action risk degree calculation unit 39 that calculates a driving action risk degree DR 3 based on movement of the driver's line of sight and the like are provided.
[0103]
FIG. 22 is a diagram in which the direction of the line of sight of the driver when viewing the vehicle from above is classified. A region including the traveling direction of the host vehicle is defined as a region A, and a region B, a region C,. Note that the direction of the line of sight in the horizontal direction may be quantified using an angle instead of being classified into eight regions in this way.
[0104]
FIG. 23 is a schematic diagram of an externally captured image captured by the external CCD camera 13. In the laterally picked-up image, in the lateral direction, the central portion is classified as region A, the left region of region A is classified as region H, and the right region of region A is classified as region B. Region A, region B, and region H correspond to those shown in FIG. 22 and correspond to directions when the traveling direction of the host vehicle is the reference direction.
[0105]
In addition, the captured images outside the vehicle are classified in the vertical direction from top to bottom, region a, region b, region c, and region d. Regions i to d indicate the direction of the line of sight. The direction of the line of sight may be quantified using an angle.
[0106]
FIG. 24 is a diagram illustrating a captured image outside the vehicle and the direction of the driver's line of sight. X ₁ , X ₂ , ..., X _Five Are obstacles such as pedestrians and vehicles detected by the object detection unit 34, respectively. Here, a case where the own vehicle travels on a road with an obstacle as shown in FIG. 24 and the driver moves his / her line of sight according to the arrow in FIG. 24 will be described as an example.
[0107]
Based on the output from the object detection unit 34, the driving behavior risk calculation unit 39 calculates each obstacle X on the road. ₁ ~ X _Five , And the visual behavior of the driver moving along the arrow shown in FIG. 24 is detected based on the output from the visual line filter 32. Note that it is preferable to use gaze and eye following movement as the visual behavior of the driver. Further, when only the case where the driver is gazing for a predetermined time or more is used, the driving behavior risk DR3 can be obtained with higher accuracy.
[0108]
When the first line of sight is in the area A, the driving behavior risk calculation unit 39 determines that the driver has two obstacles X present in the area A. ₂ And X _Three Determine that you are watching. Prior to this, since the driver did not see the obstacle, the driving behavior risk calculation unit 39 sets the number of obstacles stored in the driver (hereinafter referred to as “stored number”) to zero.
[0109]
Next, when the driving behavior risk calculating unit 39 detects that the line of sight has moved to the region H, the driver detects one obstacle X existing in the region H. ₁ Determine that you are watching. The driver has two obstacles X just before this ₂ And X _Three Therefore, the driving behavior risk calculation unit 39 sets the number of memories to 2.
[0110]
When the driving behavior risk calculation unit 39 detects that the line of sight has again moved to the region A, the driver detects the obstacle X present in the region A. ₂ And X _Three Determine that you are watching. The driver has one obstacle X just before this ₁ Therefore, the driving behavior risk calculation unit 39 sets the stored number to 1.
[0111]
When the driving behavior risk calculating unit 39 detects that the line of sight has moved to the region B, the driver detects the obstacle X present in the region B. _Four And X _Five Determine that you are watching. Before the driver moves his line of sight to the area B, the obstacle X in the area H ₁ And obstacle X in area A ₂ And X _Three Therefore, the driving behavior risk calculation unit 39 sets the number of memories to 3.
[0112]
FIG. 25 is a diagram illustrating the relationship between the driver's line-of-sight movement order, the number of obstacles the driver is currently viewing, and the number of stored objects at that time. The stored number indicates the number of obstacles grasped by the driver other than the obstacle currently being viewed. Therefore, since the driver has to drive the vehicle with caution as the number of stored items increases, the stored number indicates the burden on the driver.
[0113]
The number of stored objects is preferably the number of obstacles seen by the driver from the present time to a predetermined time before. Therefore, the driving behavior risk calculation unit 39 sequentially deletes the number of obstacles seen a predetermined time ago as time passes, and sequentially adds the number of obstacles recognized as seen by the driver. Therefore, the stored number is constantly updated.
[0114]
In addition, the driving behavior risk calculation unit 39 can set the number of memories in consideration of the time required for reconfirmation of the obstacle. At this time, for example, when the time required for reconfirmation of the obstacle is equal to or longer than a predetermined time, the driving behavior risk calculation unit 39 may increase the time for storing the number of obstacles.
[0115]
In this way, when the driver's line of sight moves, the driving behavior risk level calculation unit 39 updates the number of obstacles stored by the driver, and compares the stored number with the threshold th as shown in FIG. . As a result of comparing the stored number with the threshold th, the driving behavior risk calculating unit 39 determines that the driving action risk DR3 is high when the stored number is equal to or greater than the threshold th, and the stored number is less than the threshold th. Determines that the driving behavior risk DR3 is low.
[0116]
In the present embodiment, the threshold th is set to six, but is not limited to this. The threshold th is preferably changed according to individual differences such as the age of the driver. For example, the threshold th may be set to 5 for a driver who is likely to cause an accident, or the threshold th may be set to 7 for a driver who has not caused an accident for a long time.
[0117]
The microcomputer 30 displays a screen for prompting the driver to reduce the speed on the LCD 19 when, for example, the number of memories exceeds the threshold th and it is determined that the driving action risk DR3 is high through such processing. Then, an alarm for urging the speed down is output from the speaker 20.
[0118]
As described above, the driving behavior risk calculation device according to the third embodiment recognizes the number of obstacles in the gaze area, estimates the number of memories in which the driver stores the obstacles, and determines the human memory characteristics. It is possible to determine whether the load on the driver becomes too large in consideration. When the estimated stored number is equal to or greater than the threshold th, it is determined that the driving action risk DR3 is high, and the driver can be prompted to decrease the vehicle speed, thereby greatly reducing the possibility of an accident.
[0119]
The present invention is not limited to such an embodiment. For example, the driving behavior risk calculation unit 39 detects the movement of the driver's line of sight using the output result of the line-of-sight filter 32, but may estimate the movement of the driver's line of sight based on, for example, the steering angle and the vehicle speed. . As a result, the burden of image processing for line-of-sight detection can be omitted, so that the calculation speed of the driving behavior risk DR3 can be increased.
[0120]
[Fourth Embodiment]
Next, a fourth embodiment of the present invention will be described. In addition, the same code | symbol is attached | subjected to the site | part and circuit same as embodiment mentioned above, and the detailed description is abbreviate | omitted. In the fourth embodiment, a driving behavior risk calculating device 1A having the configuration shown in FIG. 27 will be described as an example.
[0121]
The driving behavior risk calculation device 1A according to the present embodiment performs data communication with the data calculation device 40 to calculate the driving behavior risk DR4. Specifically, the driving behavior risk calculation device 1A includes a communication circuit 21 that transmits and receives data to and from the data calculation device 40 in addition to the constituent circuits of the driving behavior risk calculation device 1. .
[0122]
The communication circuit 21 performs predetermined modulation processing on the data supplied via the data bus 5 and transmits the data to the data arithmetic device 40 via the communication antenna 21a. Further, the communication circuit 21 receives data from the data arithmetic device 40 via the communication antenna 21 a, performs a predetermined demodulation process, and then supplies the demodulated data to the microcomputer 30. On the other hand, the data calculation device 40 can transmit and receive data to and from the driving behavior risk calculation device 1A via the antenna 40a.
[0123]
(Update parameter table)
The driving behavior risk calculating device 1A can always hold the latest parameter table based on the information transmitted from the data calculating device 40.
[0124]
Here, the data calculation device 40 stores a parameter table configured in the same manner as in FIGS. 5 to 9 for each vehicle. Then, when the parameter table is updated, the data calculation device 40 transmits the updated parameter table data to the driving behavior risk calculation device 1A of the predetermined vehicle.
[0125]
The driving behavior risk calculation device 1A can update the parameter table to the parameter table transmitted from the data calculation device 40, and always accurately calculate the environmental risk KR using the latest risk parameter. As a result, the driving behavior risk calculation device 1A can calculate the driving behavior risk DR4 according to the latest environmental situation.
[0126]
(Calculation of environmental risk KR by the data calculation device 40)
The data calculation device 40 can calculate the environmental risk KR corresponding to the type of the object instead of the driving behavior risk calculation device 1A. For this purpose, the driving behavior risk calculating device 1A and the data calculating device 40 execute the following processing.
[0127]
Similarly to the second embodiment, the microcomputer 30 of the driving behavior risk calculation device 1A recognizes an object in the environment around the vehicle, and determines the direction of the vehicle in the traveling direction for each type of the recognized object. Recognize speed, size and position. The microcomputer 30 transmits these data to the data calculation device 40 via the communication circuit 21.
[0128]
The data calculation device 40 receives the data transmitted from the driving behavior risk calculation device 1A, and calculates the environmental risk KR according to the type of the object based on the received data. Then, the data calculation device 40 transmits data indicating the environmental risk KR corresponding to the calculated type of the object to the driving behavior risk calculation device 1A. At this time, it is preferable that the data calculation device 40 stores danger history data in the same manner as in FIG. 10 every time the environmental risk KR is calculated.
[0129]
When the microcomputer 30 of the driving behavior risk calculation device 1A receives the environmental risk KR corresponding to the type of the object via the communication circuit 21, the microcomputer 30 uses the environmental risk KR as in the second embodiment. To calculate the driving behavior risk DR4.
[0130]
As a result, the driving behavior risk calculation device 1A does not need to store a parameter table having a large amount of data, and does not need to perform the calculation processing of the environmental risk KR, thereby reducing the burden for the calculation processing. can do.
[0131]
The present invention is not limited to the above-described embodiments, and various design changes can be made within the scope described in the claims.
[0132]
For example, in the above-described embodiment, the risk parameter is attached to each of the object, the direction, the speed, the size, and the position. However, the risk parameter may be attached for each region or for each time.
[0133]
Moreover, each risk parameter mentioned above is an example of this invention, and is not limited to this. Therefore, the parameter table is not limited to those shown in FIGS. 5 to 9, and for example, a parameter table may be provided according to the vehicle position of the host vehicle, the color of the front signal, the weather, and the brightness of the environment. Of course.
[0134]
【Effect of the invention】
The driving behavior risk calculating device according to the present invention calculates and outputs the driving behavior risk based on the visual behavior detected by the visual behavior detecting means and the vehicle state measured by the vehicle state measuring means. Thus, it is possible to accurately determine the degree of danger due to the driving behavior of the driver in consideration of the location and direction confirmed by the driver and the state of the vehicle in which the driver is actually driving.
[0135]
The driving behavior risk calculation device according to the present invention is based on the risk according to the type of the object calculated by the environmental risk calculation means and the driver's visual behavior detected by the visual behavior detection means. By calculating and outputting the driving behavior risk level, it is possible to obtain a driving behavior risk level that comprehensively considers the environment around the vehicle and the confirmation state of the driver.
[0136]
The driving behavior risk calculation device according to the present invention includes the number of objects stored by the driver based on the object recognized by the environment recognition means and the driver's visual behavior detected by the visual behavior detection means. The driving behavior risk that indicates the driving burden when driving while grasping many objects is calculated by calculating and outputting the driving behavior risk based on the estimated number of objects. It can be calculated.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a driving behavior risk degree computing device according to a first embodiment of the present invention.
FIG. 2 is a block diagram showing a functional configuration of a microcomputer provided in the driving behavior risk degree computing device.
FIG. 3 is a diagram showing a configuration of a visual behavior table stored in a microcomputer.
FIG. 4 is a block diagram showing a functional configuration of a microcomputer provided in a driving behavior risk calculation device according to a second embodiment of the present invention.
FIG. 5 is a diagram showing a configuration of an object parameter table stored in a microcomputer provided in the environmental risk degree computing device.
FIG. 6 is a diagram showing a configuration of an orientation parameter table stored in a microcomputer.
FIG. 7 is a diagram showing a configuration of a speed parameter table stored in a microcomputer.
FIG. 8 is a diagram showing a configuration of a size parameter table stored in a microcomputer.
FIG. 9 is a diagram showing a configuration of a position parameter table stored in a microcomputer.
FIG. 10 is a diagram illustrating a configuration of risk history data.
FIG. 11 is a diagram schematically showing calculation contents of a driving action risk degree calculation unit.
FIG. 12 is a diagram illustrating an example of a captured image outside the vehicle that is a basis of a final image.
FIG. 13 is a diagram illustrating a state of a mask applied to an outside-vehicle captured image.
FIG. 14 is a diagram showing an arrival distance after 1 second (solid line), an arrival distance after 2 seconds (dotted line), and a traveling direction (arrow) of the host vehicle after 1 second.
FIG. 15 is a diagram illustrating a final image generated by performing image processing on a captured image outside the vehicle.
FIG. 16 is a diagram illustrating a result of driving behavior risk calculated by a microcomputer.
FIG. 17 is a diagram illustrating a captured image (A) outside the vehicle at point A and a captured image for driver (B) at point A;
FIG. 18 is a diagram illustrating an outside-vehicle captured image (A) at point B and a driver-captured image (B) at point B.
FIG. 19 is a diagram illustrating a captured image (A) outside the vehicle at point C and a captured image for driver (B) at point C.
FIG. 20 is a diagram showing an outside-vehicle captured image (A) at point D and a driver-captured image (B) at point D.
FIG. 21 is a block diagram showing a functional configuration of a microcomputer according to a third embodiment of the present invention.
FIG. 22 is a diagram in which the direction of the line of sight of the driver when viewing the vehicle from above is classified.
FIG. 23 is a schematic diagram of an outside-captured image captured by a CCD camera for outside the vehicle.
FIG. 24 is a diagram showing the captured image outside the vehicle and the direction of the driver's line of sight.
FIG. 25 is a diagram illustrating a relationship between a driver's line-of-sight movement order, the number of obstacles that the driver is currently viewing, and the number of stored objects at that time;
FIG. 26 is a diagram illustrating a state in which a stored number that is the number of obstacles stored by a driver is compared with a threshold value th.
FIG. 27 is a block diagram showing a configuration of a driving behavior risk calculation device according to a fourth embodiment of the present invention.
[Explanation of symbols]
1,1A Driving behavior risk calculation device
13 External CCD camera
15 Master cylinder oil pressure sensor
16 Steering angle sensor
17 Wheel speed sensor
18 CCD camera for driver
19 LCD
20 Speaker
30 Microcomputer

Claims (5)

Recognizes the environmental objects around the vehicle, and environment recognition means for detecting a relative position for said object with respect to the traveling direction with respect to the direction, speed, and the vehicle of the vehicle,
For each object recognized by the environment recognizing means, a risk parameter related to the possibility of an accident trigger determined in advance according to the type of the object, and the object detected by the environment recognizing means orientation, velocity, and environmental risk calculating means for calculating an environmental risk determined from the risk parameters predetermined in accordance with the respective position,
Visual behavior detecting means for detecting the visual behavior of the driver;
The driving for the target object determined for each target object based on the environmental risk level for each target object calculated by the environmental risk level calculating means and the driver's visual behavior detected by the visual action detecting means. Driving behavior risk calculating means for calculating the driving behavior risk based on a weighting factor according to the visual behavior of the person ,
Driving behavior risk output means for outputting the driving behavior risk calculated by the driving behavior risk calculation means;
A driving behavior risk computing device. The driving behavior risk calculating means is a weighting factor according to the driver's visual behavior with respect to the target determined for each target based on the driver's visual behavior detected by the visual behavior detecting means. , it performs weighting to environmental risk of each object that is calculated by the environmental risk calculation means, and calculates the driving behavior risk by adding the environmental hazards of each object weighted The driving behavior risk calculation device according to claim 1. It further comprises vehicle speed detection means for detecting the vehicle speed,
The driving behavior risk calculating means calculates an intermediate risk by adding an environmental risk of each weighted object, and calculates the calculated intermediate risk and the vehicle speed detected by the vehicle speed detecting means. The driving behavior risk calculating device according to claim 2, wherein the driving behavior risk is calculated based on the driving behavior risk. Vehicle condition measuring means for measuring the traveling direction of the vehicle,
The driving behavior risk degree calculating means is further configured to allow the driver to see the traveling direction of the vehicle based on the traveling direction of the vehicle measured by the vehicle state measuring means and the visual behavior detected by the visual behavior detecting means. 4. The driving behavior risk level is calculated based on at least one of the calculated time and frequency, and at least one of the calculated time and frequency is calculated. The driving behavior risk calculation device according to claim 1. Visual attention state detection means for classifying the visual behavior detected by the visual behavior detection means and estimating the driver's attention state;
The driving behavior risk level calculating means further calculates a driving behavior risk level using the caution state estimated by the visual caution state detection means. The driving behavior risk calculation device according to the item.

Images