JP3728895B2 - Exercise intensity detector - Google Patents

Description

【００５３】
数１においてｂは、マザー・ウエーブレットψ（ｘ）をトランスレート（平行移動）する際に用いるパラメータであり、一方、ａはスケール（伸縮）する際のパラメータである。したがって、数１においてウエーブレットψ（（ｘ−ｂ）／ａ）は、マザー・ウエーブレットψ（ｘ）をｂだけ平行移動し、ａだけ伸縮したものである。この場合、スケールパラメータａに対応してマザー・ウエーブレットψ（ｘ）の幅は伸長されるので、１／ａは周波数に対応するものとなる。
【００５４】
ここで、ウエーブレット変換部２０の詳細な構成について説明する。図１１はウエーブレット変換部２０の詳細な構成を示すブロック図である。このウエーブレット変換部２０は、上記した数１の演算処理を行う構成であって、クロックＣＫが供給され、クロック周期で演算処理が行われるようになっている。図示するようにウエーブレット変換部２０は、マザー・ウエーブレットψ（ｘ）を記憶する基底関数記憶部Ｗ１、スケールパラメータａを変換するスケール変換部Ｗ２、バッファメモリＷ３、トランスレートを行う平行移動部Ｗ４および乗算部Ｗ５から構成される。なお、基底関数記憶部Ｗ１に記憶するマザー・ウエーブレットψ（ｘ）としては、ガボールウエーブレットの他、メキシカンハット、Ｈａａｒウエーブレット、Ｍｅｙｅｒウエーブレット、Ｓｈａｎｎｏｎウエーブレット等が適用できる。
【００５５】
まず、基底関数記憶部Ｗ１からマザー・ウエーブレットψ（ｘ）が読み出されると、スケール変換部Ｗ２はスケールパラメータａの変換を行う。ここで、スケールパラメータａは周期に対応するものであるから、ａが大きくなると、マザー・ウエーブレットψ（ｘ）は時間軸上で伸長される。この場合、基底関数記憶部Ｗ１に記憶されるマザー・ウエーブレットψ（ｘ）のデータ量は一定であるので、ａが大きくなると単位時間当たりのデータ量が減少してしまう。スケール変換部Ｗ２は、これを補うように補間処理を行うとともに、ａが小さくなると間引き処理を行って、関数ψ（ｘ／ａ）を生成する。このデータはバッファメモリＷ３に一旦格納される。
【００５６】
次に、平行移動部Ｗ４はバッファメモリＷ３からトランスレートパラメータｂに応じたタイミングで関数ψ（ｘ／ａ）を読み出すことにより、関数ψ（ｘ／ａ）の平行移動を行い関数ψ（ｘ−ｂ／ａ）を生成する。
【００５７】
次に、乗算部Ｗ５には、図示せぬＡ／Ｄ変換器を介して体動除去脈波波形ＭＨ’をＡ／Ｄ変換して得た体動除去脈波データＭＨ’が供給される。乗算部Ｗ４は、変数１／ａ^1/2、関数ψ（ｘ−ｂ／ａ）および体動除去脈波データＭＨ’を乗算してウエーブレット変換を行い、体動除去脈波解析データＭＫＤを生成する。この例において、体動除去脈波解析データＭＫＤは、０Ｈｚ〜０．５Ｈｚ、０．５Ｈｚ〜１．０Ｈｚ、１．０Ｈｚ〜１．５Ｈｚ、１．５Ｈｚ〜２．０Ｈｚ、２．０Ｈｚ〜２．５Ｈｚ、２．５Ｈｚ〜３．０Ｈｚ、３．０Ｈｚ〜３．５Ｈｚ、３．５Ｈｚ〜４．０Ｈｚといった周波数領域に分割されて出力される。
【００５８】
１−２：呼吸成分生成部
次に、呼吸成分生成部２１は、体動除去脈波解析データＭＫＤを各周波数領域で比較し、最大のエネルギー成分を有する領域を特定し、それ以上の周波数成分を除去して、呼吸波形解析データＶＫＤを生成する。最大エネルギー成分を有する周波数領域以上を除去したのは、最大エネルギー成分を有する周波数領域に脈波成分の基本波周波数成分が存在するからである。
【００５９】
いまここで、体動除去脈波解析データＭＫＤが図１２に示すものであるとすれば、各期間ｔ１〜ｔ８において、図１３に斜線で示す領域が最大エネルギー成分として特定される。この場合、斜線で示した領域以上の高域周波数領域は、「０」に置換されて、図１４に示すデータが、呼吸波形解析データＶＫＤとして生成される。
【００６０】
１−３：逆ウエーブレット変換部
次に、逆ウエーブレット変換部２２は、ウエーブレット変換部２０と相補的な関係にあり、以下に示す数２を演算し、呼吸波形データＶＤを生成し、これにＡ／Ｄ変換を施して呼吸波形ＶＨを出力する。
【数２】

【００６１】
このようにして、呼吸成分抽出部１４は、体動除去脈波データＭＨ’に基づいて呼吸波形ＶＨを抽出する。次に、評価部１５は、ゼロクロスコンパレータ２３、デューティ比検出部２４および運動強度テーブル２５から構成される。
【００６２】
１−４：ゼロクロスコンパレータ
まず、ゼロクロスコンパレータ２３は、例えば、図１５に示すようにコンデンサＣとオペアンプＯＰから構成される。コンデンサＣの値は、呼吸波形ＶＨが十分通過するように設定される。オペアンプＯＰは、呼吸波形ＶＨをゼロレベルと比較して矩形波Ｓを生成するが、コンデンサＣを介して呼吸波形ＶＨがオペアンプＯＰに供給されるので、矩形波Ｓは、呼吸波形ＶＨの平均値レベルを閾値として波形整形を行っている。
【００６３】
ところで、上述したように、人の呼吸は運動強度Ｘが大きくなるにつれ、吸入時間と排出時間の差がなくなる。このため、運動強度Ｘが大きくなるにつれ、矩形波Ｓのデューティ比は５０％に近ずく。
【００６４】
１−５：デューティ比検出部
次に、デューティ比検出部２４の回路図を図１６に、また、そのタイミングチャートを図１７に示す。クロック信号ＣＫ（図１７（ａ）参照）は、各々ゲート２４１，２４２の一方の入力に供給される。また、ゲート２４１の他方の入力には矩形波Ｓ（図１７（ｂ）参照）が供給され、ゲート２４２の他方の入力にはインバータ２４０によって反転された矩形波Ｓが供給される。ここで、各ゲート２４１，２４２によって、クロック信号ＣＫは制限されるから、ゲート２４１の出力信号は図１７（ｃ）に示すように矩形波Ｓがハイレベルの期間のみクロック信号ＣＫを通過させる。一方、ゲート２４２の出力信号は図１７（ｄ）に示すように矩形波Ｓがローレベルの期間にのみクロック信号ＣＫを通過させる。
【００６５】
各ゲート２４１，２４２の出力信号はカウンタ２４３，２４４に各々供給されるから、カウンタ２４３のカウント値Ｃ１は矩形波Ｓのハイレベル期間を、カウンタ２４４のカウント値Ｃ２は矩形波Ｓのローレベル期間を示すものとなる。
【００６６】
そして、除算器２４５は、Ｃ１／Ｃ２を算出し、これをデューティ比として出力する。なお、図１７に示す時刻Ｔにおいて除算演算が実行され、またカウンタ２４３，２４４はその直後にリセットされるようになっている。
【００６７】
この例における除算結果ＤＲは、ハイレベル期間Ｃ１／ローレベル期間Ｃ２であるから、運動強度Ｘが大きくなるにつれ、演算結果ＤＲは「１」に近づく。しかしながら、運動強度Ｘがある限界を越えて増大すると、呼吸波形が大幅に乱れるため、そのような運動強度Ｘにあっては、演算結果ＤＲの変化が激しい。また逆に、通常の運度強度Ｘの領域では、呼吸波形のデューティ比が急に変化するようなことはない。以下に説明する構成は、演算結果ＤＲの連続性、換言すれば、デューティ比の連続性を検出することによって、運動強度Ｘの限界値Ｘmaxを特定するものである。
【００６８】
演算結果ＤＲは、メモリ２４６に供給され記憶される。なお、記憶内容は、次の演算結果ＤＲが出力される毎に更新されるようになっている。減算器２４７が現在の演算結果ＤＲから直前の演算結果ＤＲ’を減算すると、比較器２４８は減算結果ΔＤＲが予め定められた範囲内にあるか否かを判定する。具体的には、以下に示す式を満たすか否かを判定する。
＋Ｋ＞ΔＤＲ＞−Ｋ
ここで＋Ｋ，−Ｋは、運動強度Ｘが限界値Ｘmaxを越えて、呼吸波形のデューティ比の連続性がなくなるか否かが判定できるように定められる。
【００６９】
上記した式を満たす場合には、通常の運動強度であると判定され比較器２４８の出力信号はハイレベルになる。一方、上記した式を満たさない場合には、運動強度Ｘが限界値Ｘmaxを越えていると判定され比較器２４８の出力信号はローレベルになる。
【００７０】
合成器２４９は、比較器２４８の出力信号がハイレベルであれば、演算結果ＤＲを出力し、一方、比較器２４８の出力信号がローレベルであれば、演算結果ＤＲが取り得ない値、例えば、「０」を出力する。
【００７１】
１−６：運動強度テーブル
次に、運動強度テーブル２５（図９参照）は、ＲＯＭ等で構成されており、そこには、演算結果ＤＲに対応付けられて運動強度Ｘが格納されている。したがって、演算結果ＤＲを参照して運動強度テーブル２５にアクセスすれば、運動強度Ｘを求めることができる。また、演算結果ＤＲが取り得ない値が入力された場合、例えば、「０」の場合には、限界値Ｘmaxが出力されるようになっている。
これにより、脈波波形から呼吸波形を抽出して、そのデューティ比から運動強度Ｘを求めることができる。
【００７２】
２：第２実施形態の動作
次に、図面を参照して第２実施形態の動作を説明する。図１８は第２実施形態に係わる運動強度検出装置１のフローチャートである。図において、ステップＳ１〜ステップＳ５までの処理は、図９に示す第１実施形態の動作と同様であり、脈波波形から体動波形が除去されて、体動除去脈波波形ＭＨ’が生成される。
【００７３】
この後、ウエーブレット変換部２０によって、体動除去脈波データＭＨ’にウエーブレット変換処理が施され、体動除去脈波解析データＭＫＤが生成される。この体動除去脈波解析データＭＫＤには、脈波成分と呼吸成分が含まれるが、脈波成分は呼吸成分よりも高い周波数領域に存在し、かつ、脈波成分のエネルギーは呼吸成分のエネルギーと比較して大きい。このため、呼吸成分生成部２１は、体動除去脈波解析データＭＫＤの最大エネルギー周波数領域以上を「０」に置換して、呼吸波形データＶＫＤを生成する（ステップＳ２１）。
【００７４】
次に、逆ウエーブレット変換部２２が、呼吸波形データＶＫＤに逆ウエーブレット変換を施して呼吸波形ＶＨを生成すると、ゼロクロスコンパレータ２３は呼吸波形ＶＨをその平均値レベルで比較して矩形波Ｓを生成する。この後、デューティ比検出部２４は、矩形波Ｓのデューティ比を検出する（ステップＳ２３）。
【００７５】
次に、運動強度テーブル２５が、デューティ比検出部２４の出力データを参照して運動強度Ｘを求めると（ステップＳ２４）、表示部１６は運動強度Ｘを表示する（ステップＳ２５）。これにより、運動強度Ｘが被験者に告知される。
【００７６】
このように、第２実施形態にあっては、体動検出部１１および波形処理部１２によって脈波波形ＭＨに重畳している体動成分ＭＨｔを生成し、これを除去しているので、呼吸成分抽出部１４は運動中であってもウエーブレット変換を用いて呼吸波形を抽出することが可能である。また、呼吸波形のデューティ比に基づいて運動強度Ｘを算出したので、被験者の負担にならず、かつ、簡易に運動強度Ｘを知ることができる。
【００７７】
Ｅ．第３実施形態
１．第３実施形態の構成
本発明の第３実施形態に係わる運動強度検出装置１の構成を図面を参照しつつ説明する。第３実施形態に係わる運動強度検出装置１の外観構成は第１実施形態と同様である。また、第３実施形態に係わる運動強度検出装置１の電気的構成については、体動除去をＦＦＴ処理した後に行う点を除いて、図５に示す第１実施形態に係わる運動強度検出装置１と同様である。
【００７８】
図１９は第３実施形態に係わる運動強度検出装置１の構成を示すブロック図である。図において、３０，３１は第１のＦＦＴ処理部、第２のＦＦＴ処理部であって、ＣＰＵ等によって構成される。第１のＦＦＴ処理部３０は、脈波波形ＭＨにＦＦＴ処理を施して、脈波解析データＭＦＤを生成する。また、第２のＦＦＴ処理部３１は、体動波形ＴＨにＦＦＴ処理を施して、体動解析データＴＦＤを生成する。
【００７９】
次に、体動除去部１３は、脈波解析データＭＦＤの各スペクトラム周波数成分の内、体動解析データＴＦＤの各スペクトラム周波数に対応するスペクトラム周波数成分成分を除去して、体動除去脈波解析データＭＫＤを生成する。この体動除去脈波解析データＭＫＤにおいて、低域の周波数領域における最大ピーク周波数が呼吸成分の基本波周波数Ｆv1であり、高域の周波数領域における最大ピーク周波数が脈波の基本周波数Ｆm1となる。
【００８０】
次に、呼吸成分抽出部１４、評価部１５および表示部１６は、第１実施形態と同様であるので、ここでは説明を省略する。
【００８１】
２．第３実施形態の動作
次に、第３実施形態に係わる運動強度検出装置１の動作を図面を参照しつつ説明する。図２０は第３実施形態に係わる運動強度検出装置１の動作を示すフローチャートである。
まず、装置本体が運動強度測定モードに設定されると（ステップＳ１）、脈波検出部１０は脈波波形ＭＨを検出する。この後、第１のＦＦＴ処理部３０は脈波波形ＭＨにＦＦＴ処理を行い、脈波解析データＭＦＤを生成する（ステップＳ３２）。一方、体動検出部１１が、被験者の体動を示す体動波形ＴＨを検出すると、第２のＦＦＴ処理部３１は体動波形ＴＨにＦＦＴ処理を行い、体動解析データＴＦＤを生成する。
【００８２】
次に、体動除去部１３は、脈波解析データＭＦＤから体動成分を除去して体動除去脈波解析データＭＫＤを生成する。図２１は、脈波解析データＭＦＤ、体動解析データＴＦＤおよび体動除去脈波解析データＭＫＤの関係について、その一例を示す図である。この図を用いて、体動除去の動作を説明する。まず、図２１（ａ）は脈波解析データＭＦＤの内容を示しており、図２１（ｂ）は体動解析データＴＦＤの内容を示している。この例にあっては、体動除去部１３は、体動解析データＴＦＤに基づいて、図２１（ｂ）に示す各スペクトル周波数Ｆt1〜Ｆt6を特定する。この後、体動除去部１３は、脈波解析データＭＦＤの各スペクトラム周波数成分の内、スペクトル周波数Ｆt1〜Ｆt6に対応するスペクトラム周波数成分を除去して、図２１（ｃ）に示す体動除去脈波解析データＭＫＤを生成する。
ところで、体動波形ＴＨは、例えば、腕の振りの加速度そのものとして検出されるが、血流は血管や組織の影響を受けるので、脈波解析データＭＦＤの体動成分と体動解析データＴＦＤは一致しない。具体的には、図２１（ｂ）と図２１（ａ）に示すように、スペクトル周波数Ｆt1〜Ｆt6に対応する各スペクトラム周波数成分が、脈波解析データＭＦＤと体動解析データＴＦＤで異なる。このため、この例では、脈波解析データＭＦＤから体動解析データＴＦＤを減算するのではなく、スペクトル周波数Ｆt1〜Ｆt6に対応するスペクトラム周波数成分を除去するようにしている。これにより、体動成分を十分に除去したを生成することができる。
【００８３】
次に、呼吸成分抽出部１３は、体動除去脈波解析データＭＫＤに基づいて、その各スペクトル周波数成分のうち、最大ピーク周波数を特定する（ステップＳ３５）。この場合、脈波成分の基本周波数Ｆm1が特定される。この後、第１実施形態で図９を用いて説明したステップＳ７〜ステップＳ１１の処理が行われ、表示部１６に運動強度Ｘが表示される。
【００８４】
このように、第３実施形態にあっては、脈波波形ＭＨおよび体動波形ＴＨにＦＦＴ処理を各々施して、体動成分を除去しているので、第１実施形態で説明した波形処理部１２を省略することができる。これにより、呼吸成分抽出部１４は運動中であっても呼吸成分を抽出することが可能である。また、呼吸成分の歪率Ｋに基づいて運動強度Ｘを評価部１５で算出したので、被験者の負担にならず、かつ、簡易に運動強度Ｘを知ることができる。
【００８５】
Ｆ．第４実施形態
上述した第１〜第３実施形態にあっては、体動を除去するため、体動検出部１０を用いて体動波形を検出し、これに基づいて脈波波形ＭＨの周波数成分から体動成分を除去するようにしたが、第４実施形態は体動検出部１０を用いることなく体動成分を除去するものである。
【００８６】
１．第４実施形態の全体構成
本発明の第４実施形態に係わる運動強度検出装置１の構成を図面を参照しつつ説明する。第４実施形態に係わる運動強度検出装置１の外観構成は第１実施形態と同様である。また、第４実施形態に係わる運動強度検出装置１の電気的構成を図２２に示す。なお、図１９に示す構成と同一の構成には同一の符号を付す。
【００８７】
図２２において、図１９に示す第３実施形態の運動強度検出装置１と構成が相違するのは、体動検出部１１および第２のＦＦＴ処理部３１が設けられていない点、体動除去部１３の替わりに脈波成分除去部１４が設けられている点、呼吸成分抽出部１３の内部構成を変更した呼吸成分抽出部１３’が設けられている点である。以下、相違点について説明する。
【００８８】
２．脈波成分除去部
次に、脈波成分除去部１３’は、ローパスフィルタで構成され、脈波解析データＭＦＤから脈波成分を除去して、脈波成分除去解析データＭＤ’を生成する。ここで、ローパスフィルタのカットオフ周波数は、脈波成分の基本波周波数よりも若干低く選ばれる。その理由は、体動成分の基本波周波数および呼吸成分の基本波周波数は、脈波成分の基本波周波数より低いからである。具体的には、安静時において測定された脈波成分の基本波周波数より若干低くカットオフ周波数を設定する。
例えば、脈波解析データＭＦＤとローパスフィルタのカットオフ周波数ｆｃが図２３に示す関係にあるとすれば、脈波成分除去解析データＭＤ’は図２４に示すものとなる。
【００８９】
３．呼吸成分抽出部
次に、呼吸成分抽出部１３’は脈波成分除去解析データＭＤ’から呼吸成分を抽出するものであって、ＣＰＵ等によって構成される。図２５は呼吸成分抽出部１３’の詳細な機能構成を示すブロック図である。
【００９０】
図において、スペクトル抽出部４０は脈波成分除去解析データＭＤ’の各スペクトル周波数から、２つのスペクトル周波数を１組として抽出し、低い方のスペクトル周波数を基本周波数テーブル４１に出力するとともに、高い方のスペクトル周波数を差分検出部４２に出力する。
【００９１】
例えば、脈波成分除去解析データＭＤ’が図２４に示すものであれば、各スペクトル周波数ｆ1〜ｆ14のなから、任意のスペクトル周波数を１組として抽出する。この場合、抽出されるスペクトル周波数の組としては、₁₄Ｃ₂組ある。また、スペクトル周波数の組が、ｆ1とｆ3であれば、ｆ1が基本周波数テーブル４１に、ｆ3が差分検出部４２に出力される。
【００９２】
次に、基本周波数テーブル４１は、ＲＯＭ等によって構成され、そこには呼吸成分の基本波周波数ｆm1に対応付けて体動成分の基本波周波数Ｆt1が予め格納されてある。この基本周波数テーブル４１の内容は、実測値から構成される。
本発明者らは、基本周波数テーブル４１のデータを設定するにあたり、被験者に対し走行速度を段階的に変化させ、走行ピッチと呼吸数の関係を実測した。図２６は、その実験結果を示したものである。走行ピッチとは、単位時間当たりの歩数である。この例にあっては、脈波検出部１０（脈波検出用センサユニット１３０）は、図３に示すように指の根本に装着されるので、これによって検出される脈波波形ＭＨ中に存在する体動成分は、腕の振りに左右される。腕の振りと走行ピッチの関係は、力強く振るか滑らかに振るかにによって異なるが、２ピッチに対して１回の腕の振りであるのが通常である。また、１回の腕の振りの周期が体動波形の１周期に相当する。したがって、走行ピッチ（回／分）をＰ、呼吸数（回／分）をＶとすれば、体動成分の基本周波数Ｆt1、呼吸成分の基本周波数Ｆv1は、走行ピッチＰと呼吸数Ｖを用いて以下の式で与えられる。
Ｆt1＝Ｐ／（６０・２）、 Ｆv1＝Ｖ／６０
上記した式を用いて、図２６に示すグラフを変換すると、体動成分の基本周波数Ｆt1と呼吸成分の基本周波数Ｆv1の関係が得られる。これを図２７に示す。基本周波数テーブル４１の内容は、例えば、図２７に示すものとなる。
【００９３】
次に、差分検出部４２は、スペクトル抽出部４０から出力される他方のスペクトル周波数と基本周波数テーブル４１から出力される周波数の差分を検出する。仮に、スペクトル抽出部４０によって抽出されたスペクトル周波数の組が、体動成分の基本周波数Ｆt1と呼吸成分の基本周波数Ｆv1であるならば、基本周波数テーブル４１にはＦv1が供給されＦt1が出力されるから、差分検出部４２の出力は「０」となる。一方、スペクトル抽出部４０によって抽出されたスペクトル周波数の組がＦv1とＦ（ただしＦv1＜Ｆ）であるならば、差分検出部４２の出力は、「｜Ｆ−Ｆt1｜」となる。したがって、差分検出部４２の出力が最も小さくなるスペクトル周波数の組が、Ｆt1，Ｆv1となる。
【００９４】
次に、比較部４３は、スペクトル抽出部４０から出力されるスペクトル周波数の組毎に、差分検出部４２の出力を比較して、その値が最も小さくなる組を特定し、当該組を構成するスペクトル周波数の低い方を出力する。この場合、特定される組は、Ｆt1，Ｆv1であり、また、Ｆt1＞Ｆv1の関係があるから、比較部４３からは呼吸成分の基本波周波数Ｆv1が出力される。
【００９５】
次に、高調波周波数生成部４４は、呼吸成分の基本波周波数Ｆv1を整数倍して、Ｆv2、Ｆv3、Ｆv4…を生成し、これらに対応する各レベルＬ１、Ｌ２、Ｌ３、Ｌ４…を呼吸成分として出力する。
このようにして、生成された呼吸成分は、第１実施形態で説明した評価部１５に供給され、歪率Ｋに基づいて運動強度Ｘが生成され、これが表示部１６に表示される。
【００９６】
このように、本実施形態によれば、体動成分の基本周波数Ｆt1と呼吸成分の基本周波数Ｆv1の関係に着目し、体動成分と呼吸成分を呼吸成分抽出部１４’で分離するようにしたので、体動検出部１１、第２のＦＦＴ処理部３１等を用いなくとも呼吸成分に基づいて運動強度Ｘを求めることができる。これにより、小型化、軽量化を図ることができ、被験者により使い勝手のよい運動強度検出装置１を提供することができる。
【００９７】
Ｇ．変形例
本発明は、上述した実施形態に限定されるものではなく、以下に述べる各種の変形が可能である。
（１）運動強度Ｘが大きくなると、骨格筋の酸素消費量が増加するため、呼吸数と脈拍数が増加する。ここで、呼吸数と脈拍数には一定の関係がある。図２８は、走行時における呼吸数と脈拍数の一例を示す図である。第１〜第４実施形態における呼吸成分抽出部１４において、脈拍数との関連でフィルタリング処理を行ってもよい。
【００９８】
具体的には、呼吸数と脈拍数の関係を予め格納したテーブルを設けておき、まず、このテーブルを用いて脈拍数（６０／ｆm1）から推定される呼吸数（６０／Ｆv1）を求める。そして、推定される呼吸成分の基本波周波数を中心周波数とするバンドパスファイルタを用いて、呼吸成分の基本周波数Ｆｖ１を抽出する。なお、この場合のフィルタリング処理はデジタル的に行えばよい。
これにより、より正確に呼吸成分を抽出することが可能となる。
【００９９】
（２）上述した第１実施形態において、第３次高調波成分Ｆv3は呼吸波形の特徴を最も良く表していると考えられるので、基本波周波数Ｆv1の第３次高調波成分Ｆv3に着目して、運動強度Ｘを求めるようにしてもよい。この場合、呼吸成分抽出部１４は、基本波周波数Ｆv1とその第３次高調波Ｆv3を抽出する。評価部１５は、これらに対応するレベルＬ１，Ｌ３からＬ３／Ｌ1を算出し、Ｌ３／Ｌ1と運動強度Ｘとの関係を予め格納した運動強度テーブルを参照して運動強度Ｘを求める。これにより、歪率Ｋを算出しなくともよいから、演算処理を簡易にすることができ、この結果、処理を高速にでき、また、ＣＰＵの負担を軽減できる。
【０１００】
（３）上述した第２実施形態にあっては、ＦＦＴを用いて周波数解析を行ったが、本発明はこれに限定されるのではなく、周波数解析を行うのであれば、いかなる手法を用いてもよく、例えば、ウエーブレット変換を用いることができる。ウエーブレット変換では、短い時間で周波数解析を行うことができるが、時間を短くすると周波数解析が粗くなる。そこで、解析時間の一単位（時間分解能）をある程度時間を長くとることによって、周波数領域を細かくすることができる。
【０１０１】
（４）上述した第２実施形態において、ウエーブレット変換部２０は、基底関数展開を行いウエーブレット変換を実現したが、本発明はこれに限定されるものではなく、ウエーブレット変換をフィルタバンクによって実現してもよい。フィルタバンクの構成例を図２９に示す。図において、フィルタバンクは３段で構成されており、その基本単位は、高域フィルタ１Ａおよびデシメーションフィルタ１Ｃと、低域フィルタ１Ｂおよびデシメーションフィルタ１Ｃである。高域フィルタ１Ａと低域フィルタ１Ｂは、所定の周波数帯域を分割して、高域周波数成分と低域周波数成分を各々出力するようになっている。この例にあっては脈波データＭＤの周波数帯域として０Ｈｚ〜４Ｈｚを想定しているので、一段目の高域フィルタ１Ａの通過帯域は２Ｈｚ〜４Ｈｚに設定され、一方、一段目の低域フィルタ１Ｂの通過帯域は０Ｈｚ〜２Ｈｚに設定される。また、デシメーションフィルタ１Ｃは、１サンプルおきにデータを間引く。
こうして生成されたデータが次段に供給されると、周波数帯域の分割とデータの間引きが繰り返され、最終的には、０Ｈｚ〜４Ｈｚの周波数帯域を８分割したデータＭ１〜Ｍ８が得られる。
【０１０２】
また、高域フィルタ１Ａと低域フィルタ１Ｂとは、その内部に遅延素子(Ｄフリップフロップ)を含むトランスバーサルフィルタで構成すればよい。ところで、人の脈拍数は４０〜２００の範囲にあり、脈波波形ＭＨの基本波周波数は、生体の状態に応じて刻々と変動する。この場合、基本波周波数に同期して、分割する帯域を可変することができれば、動的な生体の状態に追従した情報を得ることができる。そこで、トランスバーサルフィルタに供給するクロックを脈波波形ＭＨとさせることによって、分割する帯域を適応的に可変してもよい。
【０１０３】
（５）上述した第２実施形態において、ウエーブレット変換部２０を図２９に示すフィルタバンクで構成した場合には、逆ウエーブレット変換部２２を図３０に示す逆フィルタバンクで構成してもよい。図において、逆フィルタバンクは３段で構成されており、その基本単位は、高域フィルタ２Ａおよび補間フィルタ２Ｃと、低域フィルタ１Ｂおよび補間フィルタ２Ｃと、加算器２Ｄである。高域フィルタ２Ａと低域フィルタ２Ｂは、所定の周波数帯域を分割して、高域周波数成分と低域周波数成分を各々出力するようになっている。また、補間フィルタ２Ｃは、２サンプル毎に１サンプルを内挿補間する。
【０１０４】
ここで、波形を再現するためには、図２９に示すフィルタバンクと図３０に示すフィルタバンクに完全再構成フィルタバンクを用いる必要がある。この場合、高域フィルタ１Ａ，２Ａおよび低域フィルタ１Ｂ，２Ｂの特性は、以下の関係があることが必要である。
Ｈ0(-Z)Ｆ0(Z)＋Ｈ1(-Z)Ｆ1(Z)＝０
Ｈ0(Z)Ｆ0(Z)＋Ｈ1(-Z)Ｆ1(Z)＝２Z^-L
【０１０５】
また、高域フィルタ２Ａと低域フィルタ２Ｂとは、その内部に遅延素子(Ｄフリップフロップ)を含むトランスバーサルフィルタで構成すればよい。
【０１０６】
（６）また、上述した各実施形態においては、表示部１６を告知手段の一例として説明したが、装置から人間に対して告知をするための手段としては以下説明するようなものが挙げられる。これら手段は五感を基準に分類するのが適当かと考えられる。なお、これらの手段は、単独で使用するのみならず複数の手段を組み合わせても良いことは勿論である。そして、以下説明するように、例えば視覚以外に訴える手段を用いれば、視覚障害者であっても告知内容を理解することができ、同様に、聴覚以外に訴える手段を用いれば聴覚障害者に対して告知を行うことができ、障害を持つ使用者にも優しい装置を構成できる。
【０１０７】
まず、聴覚に訴える告知手段としては、脈象の分析・診断結果などを知らせるための目的、あるいは警告の目的でなされるものなどがある。例えば、ブザーの他、圧電素子、スピーカが該当する。また、特殊な例として、告知の対象となる人間に携帯用無線呼出受信機を持たせ、告知を行う場合にはこの携帯用無線呼出受信機を装置側から呼び出すようにすることが考えられる。また、これらの機器を用いて告知を行うにあたっては、単に告知するだけではなく、何らかの情報を一緒に伝達したい場合も多々ある。そうした場合、伝えたい情報の内容に応じて、以下に示す音量等の情報のレベルを変えれば良い。例えば、音高、音量、音色、音声、音楽の種類（曲目など）である。
【０１０８】
次に、視覚に訴える告知手段が用いられるのは、装置から各種メッセージ，測定結果を知らせる目的であったり、警告をするためであったりする。そのための手段として以下のような機器が考えられる。例えば、ディスプレイ装置、ＣＲＴ（陰極線管表示装置），ＬＣＤ（液晶表示ディスプレ）、プリンタ、Ｘ−Ｙプロッタ、ランプなどがある。なお、特殊な表示装置として眼鏡型のプロジェクターがある。また、告知にあたっては以下に示すようなバリエーションが考えられる。例えば、数値の告知におけるデジタル表示，アナログ表示の別、グラフによる表示、表示色の濃淡、数値そのまま或いは数値をグレード付けして告知する場合の棒グラフ表示、円グラフ、フェイスチャート等である。
【０１０９】
次に、触覚に訴える告知手段は、警告の目的で使用されることがあると考えられる。そのための手段として以下のようなものがある。まず、腕時計等の携帯機器の裏面から突出する形状記憶合金を設け、この形状記憶合金に通電するようにする電気的刺激がある。また、腕時計等の携帯機器の裏から突起物（例えばあまり尖っていない針など）を出し入れ可能な構造としてこの突起物によって刺激を与える機械的刺激がある。
【０１１０】
次に、嗅覚に訴える告知手段は、装置に香料等の吐出機構を設けるようにして、告知する内容と香りとを対応させておき、告知内容に応じた香料を吐出するように構成しても良い。ちなみに、香料等の吐出機構には、マイクロポンプなどが最適である。
【０１１１】
（７）上述した各実施形態においては、脈波検出手段ｆ１の一例として脈波検出用センサユニット１３０を取りあげ説明したが、本発明はこれに限定されるものではなく、脈動を検出できるものであれば、どのようなものであってもよい。
【０１１２】
例えば、脈波検出用センサユニット１３０は反射光を利用したものであったが、透過光を利用したものであってもよい。ところで、波長領域が７００ｎｍ以下の光は、指の組織を透過しにくい傾向がある。このため、透過光を利用する場合は、発光部から波長が６００ｎｍ〜１０００ｎｍの光を照射し、照射光を組織→血管→組織の順に透過させ、この透過光の光量変化を検出する。透過光は血液中のヘモグロビンの吸収を受けるので、透過光の光量変化を検出することによって、脈波波形を検出することができる。
【０１１３】
この場合、発光部には、ＩｎＧａＡｓ系（インジウム−ガリウム−砒素）やＧａＡｓ系（ガリウム−砒素）のレーザー発光ダイオードが好適である。ところで、波長が６００ｎｍ〜１０００ｎｍの外光は組織を透過し易いので、受光部に外光が入射すると脈波信号のＳ／Ｎが劣化してしまう。そこで、発光部から偏光したレーザー光を照射し、透過光を偏光フィルタを介して受光部で受光するようにしてもよい。これにより、外光の影響を受けることなく、脈波信号を良好なＳ／Ｎ比で検出することができる。
【０１１４】
この場合には、図３１（ａ）に示すように、発光部２００を締着具１４５の締め付け側に設け、時計本体側には受光部２０１を設けている。この場合、発光部２００から照射された光は、血管１４３を透過した後、橈骨２０２と尺骨２０３の間を通って、受光部２０１に達する。なお、透過光を用いる場合には、照射光は組織を透過する必要があるため、組織の吸収を考慮すると、その波長は６００ｎｍ〜１０００ｎｍであることが望ましい。
【０１１５】
また、同図（ｂ）は検出部位を耳朶とする例である。把持部材２０４と把持部材２０５は、バネ２０７で付勢され、軸２０６を中心に回動できるようになっている。また、把持部材２０４と把持部材２０５には、発光部２００と受光部２０１が設けられている。この脈波検出部を用いる場合には、耳朶を把持部材２０４と把持部材２０５で把持して脈波を検出する。なお、反射光を用いる場合には、同図（ｃ）に示すように指尖部から脈波波形ＭＨを検出するようにしてもよい。
【０１１６】
次に、光電式脈波センサを眼鏡と組み合わせた使用態様を説明する。なお、この眼鏡の形態では、使用者に対する告知手段としての表示装置も一緒に組み込まれた構造になっている。したがって、脈波検出部として以外に表示装置としての機能についても併せて説明する。
図３２は、脈波検出部が接続された装置を眼鏡に取り付けた様子を表わす斜視図である。図のように、装置本体は本体７５ａと本体７５ｂに分かれ、それぞれ別々に眼鏡の蔓７６に取り付けられており、これら本体が蔓７６内部に埋め込まれたリード線を介して互いに電気的に接続されている。
【０１１７】
本体７５ａは表示制御回路を内蔵しており、この本体７５ａのレンズ７７側の側面には全面に液晶パネル７８が取り付けられ、また、該側面の一端には鏡７９が所定の角度で固定されている。さらに本体７５ａには、光源（図示略）を含む液晶パネル７８の駆動回路と、表示データを作成するための回路が組み込まれている。この光源から発射された光は、液晶パネル７８を介して鏡７９で反射されて、眼鏡のレンズ７７に投射される。また、本体７５ｂには、装置の主要部が組み込まれており、その上面には各種のボタンが設けられている。なお、これらボタン８０，８１の機能は装置毎に異なる。また。光電式脈波センサを構成するＬＥＤ３２およびフォトトランジスタ３３（図４を参照）はパッド８２，８３に内蔵されると共に、パッド８２，８３を耳朶へ固定するようになっている。これらのパッド８２，８３は、本体７５ｂから引き出されたリード線８４，８４によって電気的に接続されている。
【０１１８】
次に、圧力センサによって脈波波形ＭＨを検出する例を説明する。図３３（ａ）は圧力センサを用いた脈波診断装置の外観構成を示す斜視図である。この図に示すように、脈波診断装置１には、一対のバンド１４４，１４４が設けられており、その一方の締着具１４５の締め付け側には、圧力センサ１３０’の弾性ゴム１３１が突出して設けられている。締着具１４５を備えるバンド１４４は、圧力センサ１３０による検出信号を供給するべくＦＰＣ（Flexible Printed Circuit）基板を軟性プラスチックで被覆した構造（詳細は図示省略）となっている。
【０１１９】
また、使用時においては、図３３（ｂ）に示すように、締着具１４５に設けられた弾性ゴム１３１が橈骨動脈１４３の近傍に位置するべく、腕時計１４６が被験者の左腕１４７に巻回される。このため、脈波を恒常的に検出することが可能となる。なお、この巻回については通常の腕時計の使用状態と何等変わることがない。こうして弾性ゴム１３１が、被験者の橈骨動脈１４３近傍に押圧されると、該動脈の血流変動（すなわち脈波）が弾性ゴム１３１を介して圧力センサ１３０’に伝達され、圧力センサ１３０’はこれを血圧として検知する。
【０１２０】
【発明の効果】
以上、説明したように本発明の発明特定事項によれば、脈波波形から呼吸成分を抽出し、これを周波数解析することによって、運動強度を算出するようにしたので、簡易な構成で運動強度を算出することができる。
【図面の簡単な説明】
【図１】 運動強度と呼吸波形の関係を示す図である。
【図２】 本実施形態に係わる運動強度検出装置の機能ブロック図である。
【図３】 第１実施形態に係わる運動強度検出装置の外観構成を示す斜視図である。
【図４】 同実施形態に係わる脈波検出用センサユニットの回路図である。
【図５】 同実施形態に係わる運動強度検出装置の電気的な構成を示すブロック図である。
【図６】 体動除去脈波データＭＨ’にＦＦＴ処理を施した結果を示す図である。
【図７】 図６における低域周波数領域ＬＦの拡大図である。
【図８】 同実施形態の表示部の一態様を示すフェイスチャートである。
【図９】 同実施形態に係わる運動強度検出装置の動作を示すフローチャートである。
【図１０】 第２実施形態に係わる呼吸成分抽出部および評価部の内部構成を示すブロック図である。
【図１１】 同実施形態に係わるウエーブレット変換部の構成を示すブロック図である。
【図１２】 同実施形態に係わる体動除去脈波解析データＭＫＤの一例を示す図である。
【図１３】 図１２に示した体動除去脈波解析データＭＫＤの最大エネルギー領域を示す図である。
【図１４】 同実施形態に係わる呼吸波形解析データＶＫＤの一例を示す図である。
【図１５】 同実施形態に係わるゼロクロスコンパレータの回路図である。
【図１６】 同実施形態に係わるデューティ比検出部の回路図である。
【図１７】 同実施形態に係わるデューティ比検出部のタイミングチャートである。
【図１８】 同実施形態に係わる運動強度検出装置の動作を示すフローチャートである。
【図１９】 第３実施形態に係わる運動強度検出装置の電気的な構成を示すブロック図である。
【図２０】 同実施形態に係わる運動強度検出装置の動作を示すフローチャートである。
【図２１】 同実施形態に係わる脈波解析データＭＦＤ、体動解析データＴＦＤおよび体動除去脈波解析データＭＫＤの関係について、その一例を示す図である。
【図２２】 第４実施形態に係わる運動強度検出装置の電気的な構成を示すブロック図である。
【図２３】 同実施形態に係わる脈波解析データＭＦＤとカットオフ周波数ｆｃの関係の一例を示す図である。
【図２４】 同実施形態に係わる脈波成分除去解析データＭＤ’の一例を示す図である。
【図２５】 同実施形態に係わる呼吸成分抽出部１３’の詳細な構成を示すブロック図である。
【図２６】 同実施形態において走行ピッチと呼吸数の関係を実測した結果を示す図である。
【図２７】 同実施形態において体動成分の基本周波数Ｆt1と呼吸成分の基本周波数Ｆv1の関係を示すグラフである。
【図２８】 呼吸数と脈拍数の関係を示す図である。
【図２９】 フィルタバンクの構成例を示すブロック図である。
【図３０】 逆フィルタバンクの構成例を示すブロック図である。
【図３１】 変形例に係わる光電式脈波センサの例を示す図である。
【図３２】 変形例において光電式脈波センサを眼鏡に応用した例を示す図である。
【図３３】 変形例において圧力センサを用いた脈波診断装置の外観構成を示す斜視図である。
【符号の説明】
ＭＨ 脈波波形
１０ 脈波検出部（脈波検出手段）
１３０ 脈波検出用センサユニット（脈波検出手段）
１１ 体動検出部（体動検出手段）
ＴＨ 記体動波形
ＭＨ’体動除去脈波波形，体動除去脈波データ
１３ 体動除去部（体動除去手段）
１４ 呼吸成分抽出手部（呼吸成分抽出手段）
１５ 評価部（運度強度生成手段）
ＭＫＤ 体動除去脈波解析データ
２０ ウエーブレット変換部
ＶＫＤ 呼吸波形解析データ
２１ 呼吸成分生成部（呼吸波形生成部）
２２ 逆ウエーブレット変換部（呼吸波形生成部）
ＶＨ 呼吸波形
３０ 第１のＦＦＴ処理部（第１の周波数解析部、周波数解析部）
３１ 第２のＦＦＴ処理部（第２の周波数解析部）
１３’脈波成分除去部、
４１ 基本周波数テーブル
４２ 差分検出部（周波数特定部）
４３ 比較部（周波数特定部）
４４ 高調波周波数生成部（周波数特定部、抽出部）
１６ 表示部（告知手段）[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an exercise intensity detection apparatus suitable for detecting exercise intensity from a pulse waveform.
[0002]
[Prior art]
As a method for measuring exercise intensity such as running or weight training, a method for measuring the lactic acid concentration in blood is known. In this measurement method, focusing on the fact that lactic acid is a fatigue substance, it is determined that the exercise intensity is high when the lactic acid concentration is high.
Here, if the person who is training during the exercise can know the exercise intensity himself, it is convenient to perform health management and scientific training. However, in order to measure the lactic acid concentration, it is necessary to collect blood and examine the concentration, so it is impossible to perform measurement while continuing exercise.
Therefore, the present inventors have repeatedly studied on an index representing exercise intensity, and found that there is a close relationship between the respiratory waveform and the exercise intensity.
[0003]
[Problems to be solved by the invention]
Here, how to measure the respiratory waveform is an important problem. First, as a method of measuring the respiratory waveform of a subject who is resting, such as a sick person, wrap a band around the chest or abdomen to measure the stretched state, put a thermocouple in the nostril, count fluctuations in resistance value, etc. This method is general. However, it is extremely troublesome and inconvenient for a subject who performs daily health care or a subject who performs training to wear such a thing.
[0004]
By the way, when the frequency component of the fluctuation of the RR cycle of the electrocardiogram of the subject in a resting state is analyzed, a component corresponding to the respiratory rate exists. Since the pulse wave is synchronized with the electrocardiogram, the fluctuation frequency component of the pulse wave period (or pulse wave amplitude) includes a component corresponding to the respiratory waveform.
Therefore, an apparatus for measuring a respiratory waveform based on an electrocardiogram or a pulse wave by extracting such a component is known. For example, Japanese Patent Laid-Open No. 62-22627 discloses a technique for measuring a series of pulse intervals, measuring a change period of these pulse intervals, and calculating a respiration rate by the reciprocal of the change period.
[0005]
In Japanese Utility Model Laid-Open No. 4-51912, the first respiratory rate is detected based on the fluctuation period of the RR interval of the electrocardiogram waveform or the fluctuation of the envelope of the peak value of the pulse waveform, and the reciprocation of the abdominal surface of the subject A technique is disclosed in which a second respiration rate is detected by detecting exercise, and the lower one of the first and second respiration rates is recorded and displayed.
[0006]
In addition, in the actual utility model 4-136207, the respiratory rate is estimated based on the period of fluctuation of the amplitude of the pulse wave waveform, and the average value of the pulse wave waveform (swell of the low frequency component) is calculated. A technique for reducing the influence of swell and noise by using data when the inclination is small is disclosed.
[0007]
Japanese Patent Laid-Open No. 6-142082 discloses a technique for multiplying the subject's pulse rate and the maximum blood pressure value obtained sequentially, and calculating the respiration rate based on the pulsation cycle of the multiplied value. In addition, Japanese Utility Model Publication No. 6-22325 discloses a technique for determining the respiration rate of a living body based on a change period of a curve connecting peak values of pulse waves.
[0008]
However, if the subject is exercising, the electromyogram waveform is superimposed on the electrocardiogram waveform, and the body motion component is superimposed on the pulse wave. Since these components may have a higher level than the component corresponding to the respiratory waveform, there is a problem that the respiratory waveform is erroneously calculated.
[0009]
The present invention has been made in view of the above-described circumstances, and an object thereof is to provide an exercise intensity detection device that extracts a respiratory component from a pulse wave waveform and easily detects exercise intensity based on the extracted respiratory component.
[0010]
[Means for Solving the Problems]
In order to solve the above-mentioned problems, in the invention according to claim 1, a pulse wave detecting means for detecting a pulse wave waveform from a detection part of a living body, and a body for detecting a body movement waveform indicating the body movement of the living body. Body motion removal for generating a body motion component in the pulse wave waveform based on the body motion waveform and removing the body motion component from the pulse wave waveform to generate a body motion removal pulse wave waveform Means, a respiratory component extracting means for extracting a respiratory component based on the body motion removal pulse wave waveform, and a strength intensity generating means for calculating an exercise intensity based on the respiratory component extracted by the respiratory component extracting means It is characterized by comprising.
[0011]
In the invention according to claim 2, the respiratory component extraction unit includes a wavelet conversion unit that performs wavelet conversion on the body motion removal pulse wave waveform to generate body motion removal pulse wave analysis data. The respiratory waveform analysis data is generated by removing the frequency component corresponding to the pulse wave component from the body motion removal pulse wave analysis data, the inverse waveform is applied to the respiratory waveform analysis data, and the respiratory waveform is used as the respiratory component. A respiration waveform generation unit for generating the respiration waveform.
[0012]
Further, in the invention according to claim 3, the mobility intensity generating means is based on a ratio of frequency components obtained by performing frequency analysis on the respiratory components extracted by the respiratory component extracting means, The exercise intensity is calculated.
[0013]
In the invention according to claim 4, the mobility intensity generating means calculates a distortion rate from each frequency component obtained by performing frequency analysis on the respiratory component extracted by the respiratory component extracting means. The exercise intensity is calculated based on the distortion rate.
[0014]
Further, in the invention according to claim 5, the mobility intensity generating means and the fundamental wave frequency component obtained by performing frequency analysis on the respiratory component extracted by the respiratory component extracting means and the third order A ratio of harmonic components is calculated, and the exercise intensity is calculated based on the ratio.
[0015]
In the invention according to claim 6, the respiratory component extracting means extracts a respiratory waveform as the respiratory component, and the mobility intensity generating means is the respiratory component extracted by the respiratory component extracting means. A duty ratio of the waveform is detected, and the exercise intensity is generated based on the duty ratio.
[0016]
In the invention according to claim 7, the body motion removal means includes a first frequency analysis unit that analyzes a frequency spectrum of the pulse wave waveform, and a first frequency analysis unit that analyzes the frequency spectrum of the body motion waveform. 2 and the frequency spectrum analyzed by the second frequency analysis unit are removed from the frequency spectrum analyzed by the first frequency analysis unit to remove body motion. A body motion removal unit that generates a body motion removal spectrum, wherein the respiratory component extraction unit extracts a frequency spectrum corresponding to a fundamental wave component of the respiratory component from the body motion removal spectrum, and the exercise intensity generation unit Is based on the level of the frequency spectrum corresponding to the fundamental component of the respiratory component and the level of the frequency spectrum corresponding to its harmonic component. And calculates the exercise intensity.
[0017]
In the invention according to claim 8, the breathing component extraction means specifies a band determined according to a pulse rate from the body motion removal spectrum, and from the frequency spectrum within the band. A frequency spectrum corresponding to the fundamental wave component of the respiratory component is extracted.
[0018]
Further, in the invention according to claim 9, the mobility intensity generating means is based on a spectrum level corresponding to a fundamental wave component of the respiratory component and a spectrum level corresponding to a harmonic component thereof. A distortion rate of a respiratory waveform is calculated, and the exercise intensity is calculated based on the distortion rate.
[0019]
In the invention according to claim 10, the mobility intensity generating means includes a spectrum level corresponding to the fundamental wave component of the respiratory component and a spectrum level corresponding to the third harmonic component. And the exercise intensity is calculated based on the ratio.
[0020]
In the invention according to claim 11, a pulse wave detection unit that detects a pulse wave waveform from a detection part of a living body, a respiratory component extraction unit that extracts a respiratory component from the pulse wave waveform, and the respiratory component It is characterized by comprising a mobility intensity generating means for calculating exercise intensity based on the respiratory component extracted by the extracting means.
[0021]
In the invention according to claim 12, the respiratory component extraction means performs a frequency analysis on the pulse wave waveform to generate a pulse wave analysis data, and a pulse from the pulse wave analysis data. A pulse wave component removing unit that removes a wave component, a fundamental frequency table that associates and stores a relationship between a body motion fundamental wave frequency and a respiratory fundamental wave frequency in advance, and the fundamental wave frequency table. Based on the frequency specifying unit for specifying the respiratory fundamental wave frequency and the body movement fundamental wave frequency from the inside, and calculating the respective harmonic frequencies based on the respiratory basic frequency specified by the frequency specifying unit, and extracting respiratory components And an extraction unit.
[0022]
Further, in the invention according to claim 13, the mobility intensity generating means is based on a spectrum level corresponding to the fundamental wave component of the respiratory component and a spectrum level corresponding to the harmonic component thereof. A distortion rate of a respiratory waveform is calculated, and the exercise intensity is calculated based on the distortion rate.
[0023]
Further, in the invention according to claim 14, the mobility intensity generating means includes a spectrum level corresponding to a fundamental wave component of the respiratory component and a spectrum level corresponding to a third harmonic component thereof. And the exercise intensity is calculated based on the ratio.
[0024]
The invention according to claim 15 is characterized by further comprising a notifying means for notifying the exercise intensity generated by the mobility intensity generating means.
[0025]
DETAILED DESCRIPTION OF THE INVENTION
A. principle
As the exercise intensity increases, the oxygen consumption of the skeletal muscle increases, so the respiration rate increases. Here, the relationship between the exercise intensity and the respiratory waveform is shown in FIG. In the figure, the + direction on the vertical axis indicates inhalation and the − direction indicates discharge. FIG. 4A shows a breathing waveform at rest (exercise intensity X1), FIG. 2B shows a breathing waveform at exercise intensity X2, FIG. 3C shows a breathing waveform at exercise intensity X3, and FIG. ) Shows a respiratory waveform at the exercise intensity X4, and the exercise intensities X1 to X4 have the following relationship.
X1 <X2 <X3 <X4
[0026]
From this figure, when resting, the inhalation time is longer than the excretion time, but as the exercise intensity increases, the difference between the inhalation time and the excretion time decreases, the respiratory waveform gradually approaches a sine wave, and the exercise intensity further increases. It can be seen that the respiratory waveform is greatly disturbed.
[0027]
Here, the fact that the respiratory waveform approaches a sine wave means that the harmonic component with respect to the fundamental component is reduced. In particular, unless the exercise intensity exceeds a certain limit, the respiratory waveform changes from a sawtooth wave to a sine wave, so that the third harmonic component decreases as the exercise intensity increases.
Therefore, an index of exercise intensity can be obtained by frequency analysis of the respiratory waveform.
[0028]
The present applicant pays attention to this point and extracts the respiratory component from the pulse wave waveform and analyzes the frequency to detect the exercise intensity.
[0029]
B. Functional configuration
Next, functions of the exercise intensity detection device according to the present embodiment will be described. FIG. 2 is a functional block diagram of the exercise intensity detection device according to the present embodiment. In the figure, f1 is a pulse wave detection means for detecting a pulse wave waveform from a detection part of a living body. For example, an optical pulse wave sensor or a pressure sensor is applicable. Further, f2 is a body motion detecting means for detecting a body motion waveform indicating the body motion of the living body. For example, an acceleration sensor is applicable.
[0030]
Further, f3 is a body motion removing unit, which generates a body motion component in the pulse waveform based on the body motion waveform, and removes the body motion component from the pulse waveform to generate a body motion removed pulse waveform. . Specifically, by applying appropriate waveform processing to the body motion waveform and subtracting this from the pulse waveform, or analyzing the frequency spectrum of the pulse waveform and the frequency spectrum of the body motion waveform, respectively. A body motion removal waveform can be generated by removing a waveform having the same frequency as the frequency spectrum of the waveform from the frequency spectrum of the pulse wave waveform.
[0031]
Further, f4 is a respiratory component extracting means for extracting a respiratory component based on the body motion removal pulse wave waveform. The respiratory component extraction means f4 includes, for example, a wavelet conversion unit that performs wavelet conversion on a body motion removal pulse wave waveform to generate body motion removal pulse wave analysis data, and a pulse wave component from the body motion removal pulse wave analysis data. A respiratory waveform generation unit that generates respiratory waveform analysis data by removing the frequency component corresponding to the above, and applies a reverse wavelet to the respiratory waveform analysis data to generate a respiratory waveform as the respiratory component. Further, f5 is a mobility intensity generating means for calculating the exercise intensity based on the respiratory component extracted by the respiratory component extracting means. In this case, the exercise intensity can be calculated based on the ratio of frequency components obtained by performing frequency analysis on the extracted respiratory components.
[0032]
C. First embodiment
1. Configuration of the first embodiment
The configuration of the exercise intensity detection device 1 according to the first embodiment of the present invention will be described with reference to the drawings.
1-1: External configuration of the first embodiment
FIG. 3 is a perspective view showing an external configuration of the exercise intensity detection device 1 according to the first embodiment. In FIG. 3, the exercise intensity detection device 1 of this example includes a device main body 110 having a wristwatch structure, a cable 120 connected to the device main body 110, and a pulse wave detection sensor provided on the distal end side of the cable 120. The unit 130 is generally configured. A connector piece 80 is formed on the distal end side of the cable 120, and this connector piece 80 is detachable from the connector portion 70 formed on the 6 o'clock side of the apparatus main body 10. The apparatus main body 10 is provided with a wristband 60 which is wound around the arm from 12 o'clock in the wristwatch and fixed in the 6 o'clock direction. With the wristband 60, the apparatus main body 110 is detachable from the arm. The pulse wave detection sensor unit 130 is attached to the base of the index finger while being shielded by the sensor fixing band 140. In this way, when the pulse wave detection sensor unit 130 is attached to the base of the finger, the cable 120 can be shortened, so the cable 120 does not get in the way during running. Further, when the distribution of the body temperature from the palm to the fingertip is measured, the temperature of the fingertip is remarkably reduced when it is cold, but the temperature of the base of the finger is not relatively lowered. Therefore, if the pulse wave detection sensor unit 130 is attached to the base of the finger, the pulse rate and the like can be accurately measured even when running outdoors on a cold day.
[0033]
The device main body 110 also includes a resin watch case 200 (main body case). On the surface side of the watch case 200, in addition to the current time and date, the pitch and pulse rate during running and walking are also shown. A liquid crystal display device 210 with an EL backlight that displays exercise intensity, respiratory rate, and the like is configured. In addition to the segment display area, the liquid crystal display device 210 includes a dot display area. In the dot display area, various types of information can be displayed graphically.
[0034]
Further, inside the watch case 200, a change in exercise intensity, a pulse rate or a respiration rate is obtained based on the pulse wave waveform MH measured by the pulse wave detection sensor unit 130 and displayed on the liquid crystal display device 210. For this purpose, a control unit including a microcomputer that performs various controls and data processing is configured. The control unit is also configured with a timer circuit so that the normal time, lap time, split time, and the like can be displayed on the liquid crystal display device 210. In addition, button switches 111 to 115 for performing external operations such as time adjustment and display mode switching are configured on the outer periphery of the watch case 200.
[0035]
Next, the pulse wave detection sensor unit 130 includes an LED 32, a phototransistor 33, and the like as shown in FIG. When the switch SW is turned on and a power supply voltage is applied, light is emitted from the LED 32, reflected by a blood vessel or tissue, received by the phototransistor 33, and a pulse wave signal M is detected. Here, the emission wavelength of the LED is selected in the vicinity of the absorption wavelength peak of hemoglobin in the blood. For this reason, a light reception level changes according to a blood flow rate. Therefore, the pulse wave waveform can be detected by detecting the light reception level.
As the LED 32, an InGaN-based (indium-gallium-nitrogen-based) blue LED is suitable. The emission spectrum of the blue LED has an emission peak at 450 nm, for example, and the emission wavelength region is in the range from 350 nm to 600 nm. In this case, a GaAsP-based (gallium-arsenic-phosphorus-based) phototransistor may be used as the phototransistor 33 in correspondence with the LED having such light emission characteristics. In the photoreception wavelength region of the phototransistor 33, for example, the main sensitivity region is in the range from 300 nm to 600 nm, and there is also a sensitivity region at 300 nm or less. When such a blue LED and the phototransistor 33 are combined, a pulse wave is detected in the wavelength region from 300 nm to 600 nm, which is the overlapping region. In this case, there are the following advantages.
[0036]
First, of the light included in the external light, light having a wavelength region of 700 nm or less tends to be difficult to transmit through the finger tissue, so the external light is applied to the finger portion that is not covered with the sensor fixing band. However, only light in a wavelength region that does not reach the phototransistor 33 via the finger tissue and does not affect detection reaches the phototransistor 33. On the other hand, since light in a wavelength region lower than 300 nm is almost absorbed by the skin surface, even if the light receiving wavelength region is 700 nm or less, the substantial light receiving wavelength region is 300 nm to 700 nm. Therefore, the influence of external light can be suppressed without covering the finger with a large scale. Moreover, hemoglobin in blood has a large extinction coefficient for light having a wavelength of 300 nm to 700 nm, and is several times to about 100 times or more larger than that for light having a wavelength of 880 nm. Therefore, as shown in this example, when light of a wavelength region (300 nm to 700 nm) having a large light absorption characteristic is used as detection light in accordance with the light absorption characteristic of hemoglobin, the detection value changes with sensitivity according to the blood volume change. Therefore, the S / N ratio of the pulse wave waveform MH based on the blood volume change can be increased.
[0037]
1-2: Electrical configuration of the first embodiment
Next, the electrical configuration of the first embodiment is shown in FIG. In the figure, reference numeral 10 denotes a pulse wave detection unit, which corresponds to the above-described pulse wave detection sensor unit 130. The pulse wave detection unit 10 detects a pulse wave waveform MH indicating the magnitude of pulsation. Reference numeral 11 denotes a body motion detection unit, which is composed of, for example, an acceleration sensor and is provided inside the watch case 200. As a result, a body movement waveform TH indicating body movement caused by arm swinging during running is detected.
[0038]
Reference numeral 12 denotes a waveform processing unit that applies a certain waveform process to the body motion waveform TH, and reference numeral 13 denotes a body motion removal unit. The reason why the waveform processing is performed is that the body motion removal unit 13 accurately removes the body motion component. Here, if the body motion component in the pulse wave waveform MH is represented by MHt and the true pulse wave component (body motion removal pulse waveform) is represented by MH ′, MH = MHt + MH ′. For example, the body motion waveform TH is detected as the acceleration of the swing of the arm itself, but since the blood flow is affected by blood vessels and tissues, the body motion component MHt is a dull body motion waveform TH. For this reason, the waveform processing unit 12 is configured by an appropriate low-pass filter. The low-pass filter format and coefficients are calculated from values obtained through experiments. Thereby, the body motion component MHt can be obtained from the body motion waveform TH. The body motion removal unit 13 generates a body motion removal pulse wave waveform MH ′ by subtracting the body motion component MHt from the pulse wave waveform MH.
[0039]
Next, reference numeral 14 denotes a respiratory component extraction unit, which includes a CPU (Central Processing Unit), an A / D converter, and the like. In this example, the body motion removal pulse wave waveform MH ′ is converted from an analog signal to a digital signal by the A / D converter and then taken into the CPU as body motion removal pulse wave data MH ′. ing. The respiratory component extraction unit 14 performs FFT processing on the body motion removal pulse wave data MH ′ to perform frequency analysis.
[0040]
FIG. 6 is a diagram schematically illustrating a result of performing the FFT process on the body motion removal pulse wave data MH ′. In this figure, the maximum peak frequency in the low frequency region LF is the fundamental frequency Fv1 of the respiratory component, and the maximum peak frequency in the high frequency region HF is the fundamental frequency component Fm1 of the pulse wave. FIG. 7 is an enlarged view of the low frequency region LF in FIG. From this figure, it can be seen that the respiratory component is composed of the fundamental frequency Fv1, harmonics Fv2, Fv3, Fv4,. In this example, the respiratory component extraction unit 14 first performs FFT processing on the body motion removal pulse wave waveform MH ′ to identify the maximum peak frequency. Since the fundamental wave component of the pulse wave is maximum, Fm1 is specified as the maximum peak frequency. Thereafter, the maximum peak frequency is specified in a frequency region lower than Fm1. Since the frequency component lower than the pulse wave component corresponds to the respiratory component, the fundamental frequency Fv1 of the respiratory component is specified here. Thereafter, the respiratory component extraction unit 14 detects the level L1 of Fv1 and the levels L2, L3, L4... Of the harmonic frequencies Fv2, Fv3, Fv4. In this example, the harmonic frequency is limited to less than Fm1. This is because a pulse wave component exists when Fm1 is exceeded, and therefore, if Fm1 is an integral multiple of Fv1, the respiratory component cannot be separated.
[0041]
Next, reference numeral 15 denotes an evaluation unit, which includes a CPU and a ROM. The CPU calculates the distortion rate K of the respiratory waveform based on L1, L2, L3, L4... Detected by the respiratory component extraction unit 14. Specifically, the distortion rate K is calculated according to the following calculation formula.
K = (L2²+ L3²+ L4²…)^1/2/ L1
By the way, as described above, at rest, the inhalation time is longer than the excretion time, but as the exercise intensity X increases, the difference between the inhalation time and the excretion time decreases, and the respiratory waveform gradually approaches a sine wave, Furthermore, when the exercise intensity X increases, the respiratory waveform is greatly disturbed. That is, unless the exercise intensity X exceeds a certain limit, the higher the exercise intensity X, the lower the ratio of the harmonic component to the fundamental wave component. When the exercise intensity X exceeds a certain limit, the ratio of the harmonic component to the fundamental component suddenly increases. This indicates that there is a fixed relationship between the distortion rate K of the respiratory waveform and the mobility intensity X. The distortion K decreases as the exercise intensity X increases, and when the exercise intensity X exceeds a certain limit, the distortion K increases. Means a surge. Therefore, if the relationship between the distortion rate K and the exercise intensity X is obtained in advance, the exercise intensity X can be obtained from the distortion rate K.
[0042]
Next, the exercise intensity X is stored in the ROM in association with the distortion rate K. Therefore, if the ROM is accessed using the distortion rate K as an address, the exercise intensity X can be calculated. In this sense, the ROM functions as an exercise intensity table. It should be noted that the luck strength X may be graded such as 5 levels or 3 levels. In this case, the exercise intensity X stored in the ROM may be expressed by a predetermined number of steps.
[0043]
Next, reference numeral 16 denotes a display unit, which corresponds to the liquid crystal display device 210 described above. The display unit 16 may display the exercise intensity X as a numerical value as it is, or may display it as a bar graph using a dot display area. Further, when the evaluation unit 16 performs grading on the exercise intensity X, characters and symbols corresponding to the stage may be displayed. For example, if the exercise intensity corresponding to walking is X1, the exercise intensity corresponding to jogging is X2, the exercise intensity of short-distance running is X3, and the exercise intensity that is too high and impairs health is X4, Display a message such as “Light exercise”, “Medium exercise” for X2, “Strong exercise” for X3, and “Dangerous” for X4. Further, the exercise intensity X may be displayed in association with a face chart as shown in FIG.
[0044]
2. Operation of the first embodiment
Next, the operation of the first embodiment will be described with reference to the drawings. FIG. 9 is a flowchart showing the operation of the first embodiment. Here, a case where the subject who has stopped starts running and gradually increases its speed will be described as an example. First, when the subject operates the button switches 111 to 115 (see FIG. 3) to set the exercise intensity measurement mode (step S1), the pulse wave detector MH detects the pulse wave waveform MH (step S2).
[0045]
Next, when the body motion detection unit 11 detects a body motion waveform TH indicating the body motion of the subject (step S3), the waveform processing unit 12 performs waveform processing on the body motion waveform TH (step S4). Since the waveform processing is processing for converting the body motion waveform TH into the body motion component MHt in the pulse waveform MH as described above, the body motion removal is performed by subtracting the body motion component MHt from the pulse waveform MH. The unit 13 generates a body motion removal pulse wave waveform MH ′ (step S5).
[0046]
Next, the respiratory component extraction unit 13 performs frequency analysis by performing FFT on the body motion removal pulse wave waveform MH ′. And based on this analysis result, the maximum peak frequency is specified among each frequency component of the body motion removal pulse wave waveform MH '(step S6). In this case, the fundamental frequency Fm1 of the pulse wave component is specified. Thereafter, the respiratory component extraction unit 13 detects the fundamental frequency Fv1 of the respiratory component by specifying the maximum peak frequency less than Fm1 (step S7). Next, the respiratory component extraction unit 13 calculates a respiratory frequency component. Specifically, the harmonic frequencies Fv2, Fv3, Fv4,... Are detected by multiplying the fundamental frequency Fv1 by an integer, and the levels L1 corresponding to the fundamental frequency Fv1, and the harmonic frequencies Fv2, Fv3, Fv4,. , L2, L3, L4.
[0047]
Next, the evaluation unit 15 calculates the distortion rate K of the respiratory frequency component based on the respiratory frequency components L1, L2, L3, L4... (Step S9). As described above, since the relationship between the distortion rate K and the exercise intensity X is stored in the ROM in advance, the exercise intensity X is obtained by accessing the ROM based on the distortion rate K (step S10). Thereafter, the exercise intensity X is displayed on the display unit 16, whereby the exercise intensity X is notified to the subject.
[0048]
As described above, in the first embodiment, the body motion detection unit 11 and the waveform processing unit 12 generate the body motion component MHt superimposed on the pulse wave waveform MH and remove it, so that the breathing The component extraction unit 14 can extract a respiratory component even during exercise. Moreover, since the exercise intensity X is calculated based on the distortion factor K of the respiratory component, the exercise intensity X can be easily known without burdening the subject.
[0049]
D. Second embodiment
1. Configuration of the second embodiment
The configuration of the exercise intensity detection device 1 according to the second embodiment of the present invention will be described with reference to the drawings. The appearance configuration of the exercise intensity detection device 1 according to the second embodiment is the same as that of the first embodiment. The electrical configuration of the exercise intensity detection device 1 according to the second embodiment is the exercise intensity detection according to the first embodiment shown in FIG. 5 except for the internal components of the respiratory component extraction unit 14 and the evaluation unit 15. It is the same as the device 1. Hereinafter, the respiratory component component extraction unit 14 and the evaluation unit 15 will be described.
[0050]
FIG. 10 is a block diagram showing an internal configuration of the respiratory component extraction unit 14 and the evaluation unit 15 according to the second embodiment. First, the respiratory component extraction unit 14 includes a wavelet conversion unit 20, a respiratory component generation unit 21, and an inverse wavelet conversion unit 22.
[0051]
1-1: Wavelet converter
The wavelet transform unit 20 performs well-known wavelet transform on the body motion removal pulse waveform MH ′ output from the body motion removal unit 13 to generate body motion removal pulse wave analysis data MKD.
[0052]
In general, in a time-frequency analysis in which a signal is simultaneously captured from both time and frequency, a wavelet is a unit for cutting out a signal portion. The wavelet transform represents the size of each part of the signal cut out in this unit. In order to define the wavelet transform, a function ψ (x) localized in terms of time and frequency is introduced as a mother wavelet as a basis function. Here, the wavelet transformation of the function f (x) by the mother wavelet ψ (x) is defined as follows.
[Expression 1]

[0053]
In Equation 1, b is a parameter used when the mother wavelet ψ (x) is translated (translated), while a is a parameter used when scaling (stretching). Accordingly, in Equation 1, the wavelet ψ ((x−b) / a) is obtained by translating the mother wavelet ψ (x) by b and expanding and contracting by a. In this case, since the width of the mother wavelet ψ (x) is expanded corresponding to the scale parameter a, 1 / a corresponds to the frequency.
[0054]
Here, a detailed configuration of the wavelet conversion unit 20 will be described. FIG. 11 is a block diagram showing a detailed configuration of the wavelet conversion unit 20. The wavelet conversion unit 20 is configured to perform the arithmetic processing of Equation 1 described above, is supplied with a clock CK, and is configured to perform arithmetic processing at a clock cycle. As shown in the figure, the wavelet transform unit 20 includes a basis function storage unit W1 that stores a mother wavelet ψ (x), a scale transform unit W2 that transforms a scale parameter a, a buffer memory W3, and a translation unit that performs translation. W4 and a multiplication unit W5. As the mother wavelet ψ (x) stored in the basis function storage unit W1, in addition to a Gabor wavelet, a Mexican hat, a Haar wavelet, a Meyer wavelet, a Shannon wavelet, or the like can be applied.
[0055]
First, when the mother wavelet ψ (x) is read from the basis function storage unit W1, the scale conversion unit W2 converts the scale parameter a. Here, since the scale parameter a corresponds to the period, the mother wavelet ψ (x) is expanded on the time axis when a increases. In this case, since the amount of data of the mother wavelet ψ (x) stored in the basis function storage unit W1 is constant, the amount of data per unit time decreases as a increases. The scale conversion unit W2 performs an interpolation process so as to compensate for this, and performs a thinning process when a becomes smaller to generate a function ψ (x / a). This data is temporarily stored in the buffer memory W3.
[0056]
Next, the translation unit W4 reads the function ψ (x / a) from the buffer memory W3 at a timing according to the translation parameter b, thereby translating the function ψ (x / a) and performing the function ψ (x−a). b / a) is generated.
[0057]
Next, body motion removal pulse wave data MH ′ obtained by A / D conversion of the body motion removal pulse wave waveform MH ′ via an A / D converter (not shown) is supplied to the multiplication unit W5. The multiplication unit W4 is a variable 1 / a^1/2, Function ψ (x−b / a) and body motion removal pulse wave data MH ′ are subjected to wavelet transform to generate body motion removal pulse wave analysis data MKD. In this example, the body motion removal pulse wave analysis data MKD is 0 Hz to 0.5 Hz, 0.5 Hz to 1.0 Hz, 1.0 Hz to 1.5 Hz, 1.5 Hz to 2.0 Hz, 2.0 Hz to 2. Hz. The output is divided into frequency regions such as 5 Hz, 2.5 Hz to 3.0 Hz, 3.0 Hz to 3.5 Hz, and 3.5 Hz to 4.0 Hz.
[0058]
1-2: Respiratory component generator
Next, the respiratory component generation unit 21 compares the body motion removal pulse wave analysis data MKD in each frequency region, identifies a region having the maximum energy component, removes a frequency component higher than that, and performs respiratory waveform analysis. Data VKD is generated. The reason why the frequency region having the maximum energy component is removed is that the fundamental frequency component of the pulse wave component exists in the frequency region having the maximum energy component.
[0059]
Now, assuming that the body motion removal pulse wave analysis data MKD is as shown in FIG. 12, in each period t1 to t8, a region indicated by hatching in FIG. 13 is specified as the maximum energy component. In this case, the high frequency region above the region indicated by the oblique lines is replaced with “0”, and the data shown in FIG. 14 is generated as the respiratory waveform analysis data VKD.
[0060]
1-3: Inverse wavelet conversion unit
Next, the inverse wavelet conversion unit 22 has a complementary relationship with the wavelet conversion unit 20 and calculates the following equation 2 to generate respiratory waveform data VD, which is subjected to A / D conversion. A respiratory waveform VH is output.
[Expression 2]

[0061]
In this way, the respiratory component extraction unit 14 extracts the respiratory waveform VH based on the body motion removal pulse wave data MH ′. Next, the evaluation unit 15 includes a zero cross comparator 23, a duty ratio detection unit 24, and an exercise intensity table 25.
[0062]
1-4: Zero cross comparator
First, the zero-cross comparator 23 includes a capacitor C and an operational amplifier OP, for example, as shown in FIG. The value of the capacitor C is set so that the respiratory waveform VH passes sufficiently. The operational amplifier OP compares the respiration waveform VH with the zero level to generate a rectangular wave S. Since the respiration waveform VH is supplied to the operational amplifier OP via the capacitor C, the rectangular wave S is the average value of the respiration waveform VH. Waveform shaping is performed using the level as a threshold.
[0063]
By the way, as described above, as the exercise intensity X increases in human breathing, the difference between the inhalation time and the discharge time disappears. For this reason, as the exercise intensity X increases, the duty ratio of the rectangular wave S approaches 50%.
[0064]
1-5: Duty ratio detection unit
Next, a circuit diagram of the duty ratio detection unit 24 is shown in FIG. 16, and a timing chart thereof is shown in FIG. The clock signal CK (see FIG. 17A) is supplied to one input of each of the gates 241 and 242. Further, a rectangular wave S (see FIG. 17B) is supplied to the other input of the gate 241, and a rectangular wave S inverted by the inverter 240 is supplied to the other input of the gate 242. Here, since the clock signal CK is limited by the gates 241 and 242, the output signal of the gate 241 allows the clock signal CK to pass only during a period in which the rectangular wave S is at a high level as shown in FIG. On the other hand, as shown in FIG. 17D, the output signal of the gate 242 allows the clock signal CK to pass only during a period in which the rectangular wave S is at a low level.
[0065]
Since the output signals of the gates 241 and 242 are supplied to the counters 243 and 244, respectively, the count value C1 of the counter 243 is the high level period of the rectangular wave S, and the count value C2 of the counter 244 is the low level period of the rectangular wave S. Will be shown.
[0066]
Then, the divider 245 calculates C1 / C2 and outputs this as a duty ratio. Note that a division operation is executed at time T shown in FIG. 17, and the counters 243 and 244 are reset immediately thereafter.
[0067]
Since the division result DR in this example is the high level period C1 / the low level period C2, as the exercise intensity X increases, the calculation result DR approaches “1”. However, if the exercise intensity X increases beyond a certain limit, the respiration waveform is greatly disturbed. Therefore, in such exercise intensity X, the calculation result DR changes greatly. On the other hand, in the region of the normal mobility intensity X, the duty ratio of the respiratory waveform does not change suddenly. The configuration described below specifies the limit value Xmax of the exercise intensity X by detecting the continuity of the calculation result DR, in other words, the continuity of the duty ratio.
[0068]
The calculation result DR is supplied to and stored in the memory 246. The stored contents are updated every time the next calculation result DR is output. When the subtractor 247 subtracts the immediately previous calculation result DR ′ from the current calculation result DR, the comparator 248 determines whether or not the subtraction result ΔDR is within a predetermined range. Specifically, it is determined whether or not the following expression is satisfied.
+ K> ΔDR> −K
Here, + K and -K are determined so that it can be determined whether or not the exercise intensity X exceeds the limit value Xmax and the continuity of the duty ratio of the respiratory waveform is lost.
[0069]
If the above equation is satisfied, it is determined that the exercise intensity is normal, and the output signal of the comparator 248 becomes high level. On the other hand, when the above equation is not satisfied, it is determined that the exercise intensity X exceeds the limit value Xmax, and the output signal of the comparator 248 becomes low level.
[0070]
The synthesizer 249 outputs the operation result DR if the output signal of the comparator 248 is high level, while the value that the operation result DR cannot take if the output signal of the comparator 248 is low level, for example, “0” is output.
[0071]
1-6: Exercise intensity table
Next, the exercise intensity table 25 (see FIG. 9) is composed of a ROM or the like, in which exercise intensity X is stored in association with the calculation result DR. Therefore, if the exercise intensity table 25 is accessed with reference to the calculation result DR, the exercise intensity X can be obtained. In addition, when a value that cannot be taken as the operation result DR is input, for example, in the case of “0”, the limit value Xmax is output.
Thereby, the respiratory waveform is extracted from the pulse waveform, and the exercise intensity X can be obtained from the duty ratio.
[0072]
2: Operation of the second embodiment
Next, the operation of the second embodiment will be described with reference to the drawings. FIG. 18 is a flowchart of the exercise intensity detection device 1 according to the second embodiment. In the figure, the processing from step S1 to step S5 is the same as the operation of the first embodiment shown in FIG. 9, and the body motion waveform is removed from the pulse waveform to generate the body motion removed pulse waveform MH ′. Is done.
[0073]
Thereafter, the wavelet conversion unit 20 performs wavelet conversion processing on the body motion removal pulse wave data MH ′ to generate body motion removal pulse wave analysis data MKD. The body motion removal pulse wave analysis data MKD includes a pulse wave component and a respiratory component. The pulse wave component exists in a higher frequency region than the respiratory component, and the energy of the pulse wave component is the energy of the respiratory component. Big compared to. Therefore, the respiratory component generation unit 21 generates respiratory waveform data VKD by replacing “0” or more with the maximum energy frequency region of the body motion removal pulse wave analysis data MKD (step S21).
[0074]
Next, when the inverse wavelet transform unit 22 performs the inverse wavelet transform on the respiration waveform data VKD to generate the respiration waveform VH, the zero-cross comparator 23 compares the respiration waveform VH with the average value level to generate the rectangular wave S. Generate. Thereafter, the duty ratio detection unit 24 detects the duty ratio of the rectangular wave S (step S23).
[0075]
Next, when the exercise intensity table 25 obtains the exercise intensity X with reference to the output data of the duty ratio detection unit 24 (step S24), the display unit 16 displays the exercise intensity X (step S25). Thereby, the exercise intensity X is notified to the subject.
[0076]
As described above, in the second embodiment, the body motion detection unit 11 and the waveform processing unit 12 generate the body motion component MHt superimposed on the pulse wave waveform MH and remove it, so that the breathing The component extraction unit 14 can extract a respiratory waveform using wavelet transform even during exercise. Further, since the exercise intensity X is calculated based on the duty ratio of the respiratory waveform, the exercise intensity X can be easily known without burdening the subject.
[0077]
E. Third embodiment
1. Configuration of the third embodiment
The configuration of the exercise intensity detection device 1 according to the third embodiment of the present invention will be described with reference to the drawings. The appearance configuration of the exercise intensity detection device 1 according to the third embodiment is the same as that of the first embodiment. The electrical configuration of the exercise intensity detection device 1 according to the third embodiment is the same as that of the exercise intensity detection device 1 according to the first embodiment shown in FIG. 5 except that the body motion removal is performed after the FFT process. It is the same.
[0078]
FIG. 19 is a block diagram showing a configuration of the exercise intensity detection device 1 according to the third embodiment. In the figure, reference numerals 30 and 31 denote a first FFT processing unit and a second FFT processing unit, which are constituted by a CPU or the like. The first FFT processing unit 30 performs FFT processing on the pulse wave waveform MH to generate pulse wave analysis data MFD. In addition, the second FFT processing unit 31 performs FFT processing on the body motion waveform TH to generate body motion analysis data TFD.
[0079]
Next, the body motion removing unit 13 removes the spectrum frequency component corresponding to each spectrum frequency of the body motion analysis data TFD from among the spectrum frequency components of the pulse wave analysis data MFD, so that the body motion removed pulse wave analysis is performed. Data MKD is generated. In this body motion removal pulse wave analysis data MKD, the maximum peak frequency in the low frequency range is the fundamental frequency Fv1 of the respiratory component, and the maximum peak frequency in the high frequency range is the fundamental frequency Fm1 of the pulse wave.
[0080]
Next, since the respiratory component extraction unit 14, the evaluation unit 15, and the display unit 16 are the same as those in the first embodiment, description thereof is omitted here.
[0081]
2. Operation of the third embodiment
Next, the operation of the exercise intensity detection device 1 according to the third embodiment will be described with reference to the drawings. FIG. 20 is a flowchart showing the operation of the exercise intensity detection device 1 according to the third embodiment.
First, when the apparatus main body is set to the exercise intensity measurement mode (step S1), the pulse wave detector 10 detects the pulse wave waveform MH. Thereafter, the first FFT processing unit 30 performs FFT processing on the pulse wave waveform MH to generate pulse wave analysis data MFD (step S32). On the other hand, when the body motion detection unit 11 detects the body motion waveform TH indicating the body motion of the subject, the second FFT processing unit 31 performs an FFT process on the body motion waveform TH and generates body motion analysis data TFD.
[0082]
Next, the body motion removal unit 13 removes the body motion component from the pulse wave analysis data MFD to generate body motion removal pulse wave analysis data MKD. FIG. 21 is a diagram illustrating an example of the relationship among the pulse wave analysis data MFD, the body motion analysis data TFD, and the body motion removal pulse wave analysis data MKD. The operation for removing body movement will be described with reference to FIG. 21A shows the contents of the pulse wave analysis data MFD, and FIG. 21B shows the contents of the body motion analysis data TFD. In this example, the body motion removal unit 13 specifies each of the spectral frequencies Ft1 to Ft6 shown in FIG. 21 (b) based on the body motion analysis data TFD. Thereafter, the body motion removal unit 13 removes the spectrum frequency components corresponding to the spectrum frequencies Ft1 to Ft6 from the spectrum frequency components of the pulse wave analysis data MFD, and the body motion removal pulse shown in FIG. Wave analysis data MKD is generated.
By the way, the body motion waveform TH is detected as, for example, the acceleration of the swing of the arm itself, but since the blood flow is affected by blood vessels and tissues, the body motion component of the pulse wave analysis data MFD and the body motion analysis data TFD are It does not match. Specifically, as shown in FIGS. 21B and 21A, each spectrum frequency component corresponding to the spectrum frequencies Ft1 to Ft6 is different between the pulse wave analysis data MFD and the body motion analysis data TFD. Therefore, in this example, the body frequency analysis data TFD is not subtracted from the pulse wave analysis data MFD, but the spectrum frequency components corresponding to the spectrum frequencies Ft1 to Ft6 are removed. Thereby, it is possible to generate a body motion component sufficiently removed.
[0083]
Next, the respiratory component extraction unit 13 specifies the maximum peak frequency among the spectral frequency components based on the body motion removal pulse wave analysis data MKD (step S35). In this case, the fundamental frequency Fm1 of the pulse wave component is specified. Thereafter, the processing in steps S7 to S11 described in the first embodiment with reference to FIG. 9 is performed, and the exercise intensity X is displayed on the display unit 16.
[0084]
As described above, in the third embodiment, the pulse wave waveform MH and the body motion waveform TH are each subjected to the FFT processing to remove the body motion component, and thus the waveform processing unit described in the first embodiment. 12 can be omitted. Thereby, the respiratory component extraction unit 14 can extract a respiratory component even during exercise. Further, since the exercise intensity X is calculated by the evaluation unit 15 based on the distortion factor K of the respiratory component, the exercise intensity X can be easily known without burdening the subject.
[0085]
F. Fourth embodiment
In the first to third embodiments described above, in order to remove body movement, the body movement waveform is detected using the body movement detection unit 10, and based on this, the body movement is detected from the frequency component of the pulse wave waveform MH. Although the component is removed, the fourth embodiment removes the body motion component without using the body motion detection unit 10.
[0086]
1. Overall configuration of the fourth embodiment
The configuration of the exercise intensity detection device 1 according to the fourth embodiment of the present invention will be described with reference to the drawings. The appearance configuration of the exercise intensity detection device 1 according to the fourth embodiment is the same as that of the first embodiment. FIG. 22 shows an electrical configuration of the exercise intensity detection device 1 according to the fourth embodiment. In addition, the same code | symbol is attached | subjected to the structure same as the structure shown in FIG.
[0087]
In FIG. 22, the configuration differs from the exercise intensity detection device 1 of the third embodiment shown in FIG. 19 in that the body motion detection unit 11 and the second FFT processing unit 31 are not provided, and the body motion removal unit 13 is that a pulse wave component removing unit 14 is provided instead of 13 and a respiratory component extracting unit 13 ′ in which the internal configuration of the respiratory component extracting unit 13 is changed is provided. Hereinafter, differences will be described.
[0088]
2. Pulse wave component removal unit
Next, the pulse wave component removal unit 13 ′ is configured by a low-pass filter, removes the pulse wave component from the pulse wave analysis data MFD, and generates pulse wave component removal analysis data MD ′. Here, the cutoff frequency of the low-pass filter is selected slightly lower than the fundamental frequency of the pulse wave component. This is because the fundamental wave frequency of the body motion component and the fundamental wave frequency of the respiratory component are lower than the fundamental wave frequency of the pulse wave component. Specifically, the cutoff frequency is set slightly lower than the fundamental frequency of the pulse wave component measured at rest.
For example, if the pulse wave analysis data MFD and the cut-off frequency fc of the low pass filter are in the relationship shown in FIG. 23, the pulse wave component removal analysis data MD ′ is as shown in FIG.
[0089]
3. Respiratory component extraction unit
Next, the respiratory component extraction unit 13 'extracts a respiratory component from the pulse wave component removal analysis data MD', and is configured by a CPU or the like. FIG. 25 is a block diagram showing a detailed functional configuration of the respiratory component extraction unit 13 '.
[0090]
In the figure, the spectrum extraction unit 40 extracts two spectrum frequencies as one set from each spectrum frequency of the pulse wave component removal analysis data MD ′, outputs the lower spectrum frequency to the basic frequency table 41, and outputs the higher one. Are output to the difference detector 42.
[0091]
For example, if the pulse wave component removal analysis data MD 'is as shown in FIG. 24, an arbitrary spectral frequency is extracted as one set from each of the spectral frequencies f1 to f14. In this case, as a set of extracted spectral frequencies,₁₄C₂There are pairs. If the set of spectral frequencies is f1 and f3, f1 is output to the basic frequency table 41 and f3 is output to the difference detector 42.
[0092]
Next, the fundamental frequency table 41 is configured by a ROM or the like, in which the fundamental wave frequency Ft1 of the body motion component is stored in advance in association with the fundamental wave frequency fm1 of the respiratory component. The contents of the basic frequency table 41 are composed of actually measured values.
In setting the data of the basic frequency table 41, the present inventors actually measured the relationship between the running pitch and the respiration rate by changing the running speed stepwise for the subject. FIG. 26 shows the experimental results. The traveling pitch is the number of steps per unit time. In this example, since the pulse wave detection unit 10 (pulse wave detection sensor unit 130) is attached to the base of the finger as shown in FIG. 3, it is present in the pulse wave waveform MH detected thereby. The body motion component to be influenced depends on the swing of the arm. The relationship between the arm swing and the running pitch varies depending on whether the arm swings vigorously or smoothly, but it is usually one arm swing for every two pitches. In addition, one arm swing period corresponds to one period of the body movement waveform. Therefore, if the running pitch (times / minute) is P and the respiratory rate (times / minute) is V, the fundamental frequency Ft1 of the body motion component and the fundamental frequency Fv1 of the respiratory component use the running pitch P and the respiratory rate V. Is given by
Ft1 = P / (60.2), Fv1 = V / 60
When the graph shown in FIG. 26 is converted using the above equation, the relationship between the fundamental frequency Ft1 of the body motion component and the fundamental frequency Fv1 of the respiratory component is obtained. This is shown in FIG. The contents of the basic frequency table 41 are, for example, as shown in FIG.
[0093]
Next, the difference detection unit 42 detects the difference between the other spectrum frequency output from the spectrum extraction unit 40 and the frequency output from the fundamental frequency table 41. If the set of spectrum frequencies extracted by the spectrum extraction unit 40 is the fundamental frequency Ft1 of the body motion component and the fundamental frequency Fv1 of the respiratory component, Fv1 is supplied to the fundamental frequency table 41 and Ft1 is output. Therefore, the output of the difference detection unit 42 is “0”. On the other hand, if the set of spectrum frequencies extracted by the spectrum extraction unit 40 is Fv1 and F (where Fv1 <F), the output of the difference detection unit 42 is “| F−Ft1 |”. Therefore, the set of spectral frequencies that minimizes the output of the difference detector 42 is Ft1 and Fv1.
[0094]
Next, the comparison unit 43 compares the output of the difference detection unit 42 for each set of spectrum frequencies output from the spectrum extraction unit 40, identifies the set having the smallest value, and configures the set. The one with the lower spectral frequency is output. In this case, the specified group is Ft1, Fv1, and since there is a relationship of Ft1> Fv1, the comparison unit 43 outputs the fundamental wave frequency Fv1 of the respiratory component.
[0095]
Next, the harmonic frequency generation unit 44 generates Fv2, Fv3, Fv4... By multiplying the fundamental wave frequency Fv1 of the respiratory component by an integer, and breathes the levels L1, L2, L3, L4. Output as a component.
The respiratory component thus generated is supplied to the evaluation unit 15 described in the first embodiment, and an exercise intensity X is generated based on the distortion rate K, which is displayed on the display unit 16.
[0096]
Thus, according to this embodiment, focusing on the relationship between the fundamental frequency Ft1 of the body motion component and the fundamental frequency Fv1 of the respiratory component, the body motion component and the respiratory component are separated by the respiratory component extraction unit 14 ′. Therefore, the exercise intensity X can be obtained based on the respiratory component without using the body motion detection unit 11, the second FFT processing unit 31, and the like. Thereby, size reduction and weight reduction can be achieved, and the exercise intensity detection device 1 that is more convenient for the subject can be provided.
[0097]
G. Modified example
The present invention is not limited to the above-described embodiments, and various modifications described below are possible.
(1) As the exercise intensity X increases, the oxygen consumption of the skeletal muscle increases, so the respiratory rate and pulse rate increase. Here, there is a certain relationship between the respiratory rate and the pulse rate. FIG. 28 is a diagram showing an example of respiratory rate and pulse rate during running. In the respiratory component extraction unit 14 in the first to fourth embodiments, a filtering process may be performed in relation to the pulse rate.
[0098]
Specifically, a table storing the relationship between the respiratory rate and the pulse rate in advance is provided, and first, the respiratory rate (60 / Fv1) estimated from the pulse rate (60 / fm1) is obtained using this table. Then, the fundamental frequency Fv1 of the respiratory component is extracted using a bandpass filter having the estimated fundamental frequency of the respiratory component as the center frequency. Note that the filtering process in this case may be performed digitally.
Thereby, it becomes possible to extract a respiratory component more correctly.
[0099]
(2) In the first embodiment described above, the third harmonic component Fv3 is considered to best represent the characteristics of the respiratory waveform, so pay attention to the third harmonic component Fv3 of the fundamental frequency Fv1. The exercise intensity X may be obtained. In this case, the respiratory component extraction unit 14 extracts the fundamental frequency Fv1 and its third harmonic Fv3. The evaluation unit 15 calculates L3 / L1 from the levels L1 and L3 corresponding to these, and obtains the exercise intensity X with reference to the exercise intensity table in which the relationship between L3 / L1 and the exercise intensity X is stored in advance. Thereby, since it is not necessary to calculate the distortion factor K, the arithmetic processing can be simplified. As a result, the processing can be performed at a high speed and the burden on the CPU can be reduced.
[0100]
(3) In the second embodiment described above, frequency analysis is performed using FFT. However, the present invention is not limited to this, and any method can be used as long as frequency analysis is performed. For example, wavelet transformation can be used. In wavelet transform, frequency analysis can be performed in a short time, but if the time is shortened, the frequency analysis becomes coarse. Therefore, the frequency region can be made finer by taking a unit of analysis time (time resolution) as long as possible.
[0101]
(4) In the second embodiment described above, the wavelet transform unit 20 performs the basis function expansion to realize the wavelet transform. However, the present invention is not limited to this, and the wavelet transform is performed by a filter bank. It may be realized. A configuration example of the filter bank is shown in FIG. In the figure, the filter bank is composed of three stages, and its basic units are a high-pass filter 1A and a decimation filter 1C, and a low-pass filter 1B and a decimation filter 1C. The high-pass filter 1A and the low-pass filter 1B divide a predetermined frequency band and output a high-frequency component and a low-frequency component, respectively. In this example, since 0 Hz to 4 Hz is assumed as the frequency band of the pulse wave data MD, the pass band of the first-stage high-pass filter 1A is set to 2 Hz to 4 Hz, while the first-stage low-pass filter The pass band of 1B is set to 0 Hz to 2 Hz. The decimation filter 1C thins out data every other sample.
When the data thus generated is supplied to the next stage, the frequency band division and the data thinning are repeated, and finally, data M1 to M8 obtained by dividing the frequency band of 0 Hz to 4 Hz into 8 are obtained.
[0102]
Further, the high-pass filter 1A and the low-pass filter 1B may be configured by a transversal filter including a delay element (D flip-flop) therein. By the way, a person's pulse rate is in the range of 40 to 200, and the fundamental frequency of the pulse wave waveform MH fluctuates every moment according to the state of the living body. In this case, if the band to be divided can be varied in synchronization with the fundamental frequency, information that follows the dynamic state of the living body can be obtained. Therefore, the band to be divided may be adaptively varied by setting the clock supplied to the transversal filter to the pulse wave waveform MH.
[0103]
(5) In the second embodiment described above, when the wavelet transform unit 20 is configured by the filter bank shown in FIG. 29, the inverse wavelet transform unit 22 may be configured by the inverse filter bank shown in FIG. . In the figure, the inverse filter bank is composed of three stages, and the basic unit is a high-pass filter 2A, an interpolation filter 2C, a low-pass filter 1B, an interpolation filter 2C, and an adder 2D. The high-pass filter 2A and the low-pass filter 2B divide a predetermined frequency band and output a high-frequency component and a low-frequency component, respectively. The interpolation filter 2C interpolates one sample every two samples.
[0104]
Here, in order to reproduce the waveform, it is necessary to use a completely reconstructed filter bank for the filter bank shown in FIG. 29 and the filter bank shown in FIG. In this case, the characteristics of the high-pass filters 1A and 2A and the low-pass filters 1B and 2B need to have the following relationship.
H0 (-Z) F0 (Z) + H1 (-Z) F1 (Z) = 0
H0 (Z) F0 (Z) + H1 (-Z) F1 (Z) = 2Z^-L
[0105]
Further, the high-pass filter 2A and the low-pass filter 2B may be configured by a transversal filter including a delay element (D flip-flop) therein.
[0106]
(6) Moreover, in each embodiment mentioned above, although the display part 16 was demonstrated as an example of a notification means, what is demonstrated below as a means for notifying a person from an apparatus is mentioned. It is considered appropriate to classify these means based on the five senses. Of course, these means are not only used alone, but a plurality of means may be combined. And, as explained below, for example, if a means that appeals to something other than sight is used, the contents of the notification can be understood even by a visually impaired person. Similarly, if a means that appeals to something other than hearing is used, This makes it possible to make a device that is friendly to users with disabilities.
[0107]
First, there are auditory notification means for the purpose of notifying the analysis / diagnosis result of the phenotype or for the purpose of warning. For example, a buzzer, a piezoelectric element, and a speaker are applicable. As a special example, it is conceivable that a person to be notified has a portable wireless call receiver, and when making a notification, this portable wireless call receiver is called from the apparatus side. In addition, when making a notification using these devices, there are many cases where it is desired not only to notify but also to transmit some information together. In such a case, the level of information such as volume shown below may be changed according to the content of the information to be transmitted. For example, the pitch, volume, tone color, voice, and music type (song number, etc.).
[0108]
Next, the visual notification means is used for the purpose of notifying various messages and measurement results from the apparatus or for giving a warning. The following devices can be considered as means for that purpose. For example, there are a display device, a CRT (cathode ray tube display device), an LCD (liquid crystal display), a printer, an XY plotter, a lamp, and the like. Note that there is a glasses-type projector as a special display device. In addition, the following variations can be considered for notification. For example, there are digital display and numerical display in the notification of numerical values, display by graph, shading of display color, bar graph display when displaying numerical values as they are or after numerical values are graded, a pie chart, a face chart, and the like.
[0109]
Next, it is considered that a tactile notification means may be used for warning purposes. The following means are available for that purpose. First, there is an electrical stimulus for providing a shape memory alloy that protrudes from the back surface of a portable device such as a wristwatch and energizing the shape memory alloy. Further, there is a mechanical stimulus that gives a stimulus by the protrusion as a structure in which a protrusion (for example, a needle that is not very sharp) can be taken in and out from the back of a portable device such as a wristwatch.
[0110]
Next, the notification means appealing to the sense of smell may be configured such that the device is provided with a discharge mechanism such as a fragrance, the content to be notified is associated with the scent, and the fragrance according to the notification content is discharged. good. Incidentally, a micropump or the like is optimal for a discharge mechanism for fragrance or the like.
[0111]
(7) In each of the above-described embodiments, the pulse wave detection sensor unit 130 has been described as an example of the pulse wave detection means f1, but the present invention is not limited to this and can detect pulsation. Anything is acceptable.
[0112]
For example, the pulse wave detection sensor unit 130 uses reflected light, but may use transmitted light. By the way, light having a wavelength region of 700 nm or less tends to hardly pass through the finger tissue. For this reason, when using transmitted light, light having a wavelength of 600 nm to 1000 nm is irradiated from the light emitting unit, and the irradiated light is transmitted in the order of tissue → blood vessel → tissue, and a change in the amount of the transmitted light is detected. Since the transmitted light is absorbed by hemoglobin in the blood, the pulse wave waveform can be detected by detecting the change in the amount of the transmitted light.
[0113]
In this case, an InGaAs-based (indium-gallium-arsenic) or GaAs-based (gallium-arsenic) laser light-emitting diode is suitable for the light-emitting portion. By the way, since external light having a wavelength of 600 nm to 1000 nm easily passes through the tissue, the S / N of the pulse wave signal is deteriorated when the external light is incident on the light receiving unit. Therefore, it is also possible to irradiate polarized laser light from the light emitting unit and receive the transmitted light at the light receiving unit via a polarizing filter. As a result, the pulse wave signal can be detected with a good S / N ratio without being affected by external light.
[0114]
In this case, as shown in FIG. 31A, the light emitting unit 200 is provided on the fastening side of the fastener 145, and the light receiving unit 201 is provided on the watch body side. In this case, the light emitted from the light emitting unit 200 passes through the blood vessel 143 and then passes between the ribs 202 and the ulna 203 to reach the light receiving unit 201. In addition, when using transmitted light, since irradiation light needs to permeate | transmit a structure | tissue, when the absorption of a structure | tissue is considered, it is desirable that the wavelength is 600 nm-1000 nm.
[0115]
FIG. 5B shows an example in which the detection site is the earlobe. The gripping member 204 and the gripping member 205 are biased by a spring 207 so that the gripping member 204 and the gripping member 205 can rotate around a shaft 206. Further, the gripping member 204 and the gripping member 205 are provided with a light emitting unit 200 and a light receiving unit 201. When this pulse wave detection unit is used, a pulse wave is detected by holding the earlobe with the holding member 204 and the holding member 205. When using reflected light, the pulse wave waveform MH may be detected from the fingertip as shown in FIG.
[0116]
Next, a usage mode in which a photoelectric pulse wave sensor is combined with glasses will be described. In the form of glasses, a display device as a notification means for the user is also incorporated. Therefore, the function as a display device in addition to the pulse wave detection unit will be described together.
FIG. 32 is a perspective view illustrating a state in which the apparatus to which the pulse wave detection unit is connected is attached to the glasses. As shown in the figure, the apparatus main body is divided into a main body 75a and a main body 75b, which are separately attached to the spectacle vine 76, and these main bodies are electrically connected to each other via lead wires embedded in the vine 76. ing.
[0117]
The main body 75a has a built-in display control circuit. A liquid crystal panel 78 is attached to the entire surface of the main body 75a on the lens 77 side, and a mirror 79 is fixed to one end of the side surface at a predetermined angle. Yes. Further, a driving circuit for a liquid crystal panel 78 including a light source (not shown) and a circuit for creating display data are incorporated in the main body 75a. The light emitted from this light source is reflected by the mirror 79 through the liquid crystal panel 78 and projected onto the lens 77 of the glasses. Further, the main part of the apparatus is incorporated in the main body 75b, and various buttons are provided on the upper surface thereof. The functions of these buttons 80 and 81 are different for each apparatus. Also. The LED 32 and the phototransistor 33 (see FIG. 4) constituting the photoelectric pulse wave sensor are built in the pads 82 and 83, and the pads 82 and 83 are fixed to the earlobe. These pads 82 and 83 are electrically connected by lead wires 84 and 84 drawn from the main body 75b.
[0118]
Next, an example in which the pulse waveform MH is detected by the pressure sensor will be described. FIG. 33A is a perspective view showing an external configuration of a pulse wave diagnostic apparatus using a pressure sensor. As shown in this figure, the pulse wave diagnostic device 1 is provided with a pair of bands 144 and 144, and the elastic rubber 131 of the pressure sensor 130 ′ projects from the tightening side of one of the fasteners 145. Is provided. The band 144 including the fastener 145 has a structure (details not shown) in which a flexible printed circuit (FPC) substrate is covered with a soft plastic so as to supply a detection signal from the pressure sensor 130.
[0119]
In use, as shown in FIG. 33 (b), the wristwatch 146 is wound around the left arm 147 of the subject so that the elastic rubber 131 provided on the fastener 145 is positioned in the vicinity of the radial artery 143. The For this reason, it becomes possible to detect a pulse wave constantly. This winding does not change from the normal usage state of a wristwatch. When the elastic rubber 131 is pressed in the vicinity of the radial artery 143 of the subject in this way, the blood flow fluctuation (that is, the pulse wave) of the artery is transmitted to the pressure sensor 130 ′ via the elastic rubber 131, and the pressure sensor 130 ′ Is detected as blood pressure.
[0120]
【The invention's effect】
As described above, according to the invention specific matter of the present invention, the exercise intensity is calculated by extracting the respiratory component from the pulse wave waveform and performing frequency analysis on the extracted respiratory component. Can be calculated.
[Brief description of the drawings]
FIG. 1 is a diagram showing the relationship between exercise intensity and respiratory waveform.
FIG. 2 is a functional block diagram of the exercise intensity detection device according to the present embodiment.
FIG. 3 is a perspective view showing an external configuration of the exercise intensity detection device according to the first embodiment.
FIG. 4 is a circuit diagram of a pulse wave detection sensor unit according to the embodiment.
FIG. 5 is a block diagram showing an electrical configuration of the exercise intensity detection device according to the embodiment.
FIG. 6 is a diagram showing a result of performing FFT processing on body motion removal pulse wave data MH ′.
7 is an enlarged view of a low frequency region LF in FIG. 6. FIG.
FIG. 8 is a face chart showing one aspect of the display unit of the embodiment.
FIG. 9 is a flowchart showing the operation of the exercise intensity detection device according to the embodiment.
FIG. 10 is a block diagram showing an internal configuration of a respiratory component extraction unit and an evaluation unit according to the second embodiment.
FIG. 11 is a block diagram showing a configuration of a wavelet conversion unit according to the embodiment.
FIG. 12 is a diagram showing an example of body motion removal pulse wave analysis data MKD according to the embodiment.
13 is a diagram showing a maximum energy region of body motion removal pulse wave analysis data MKD shown in FIG.
FIG. 14 is a diagram showing an example of respiratory waveform analysis data VKD according to the embodiment.
FIG. 15 is a circuit diagram of a zero cross comparator according to the same embodiment;
FIG. 16 is a circuit diagram of a duty ratio detection unit according to the embodiment;
FIG. 17 is a timing chart of the duty ratio detection unit according to the embodiment;
FIG. 18 is a flowchart showing the operation of the exercise intensity detection device according to the embodiment.
FIG. 19 is a block diagram showing an electrical configuration of the exercise intensity detection device according to the third embodiment.
FIG. 20 is a flowchart showing the operation of the exercise intensity detection device according to the embodiment.
FIG. 21 is a diagram showing an example of the relationship among pulse wave analysis data MFD, body motion analysis data TFD, and body motion removal pulse wave analysis data MKD according to the embodiment;
FIG. 22 is a block diagram showing an electrical configuration of the exercise intensity detection device according to the fourth embodiment.
FIG. 23 is a diagram showing an example of a relationship between pulse wave analysis data MFD and a cut-off frequency fc according to the same embodiment.
FIG. 24 is a diagram showing an example of pulse wave component removal analysis data MD ′ according to the embodiment.
FIG. 25 is a block diagram showing a detailed configuration of a respiratory component extraction unit 13 'according to the embodiment.
FIG. 26 is a diagram showing a result of actual measurement of a relationship between a running pitch and a respiration rate in the same embodiment.
FIG. 27 is a graph showing a relationship between a fundamental frequency Ft1 of a body motion component and a fundamental frequency Fv1 of a respiratory component in the same embodiment.
FIG. 28 is a diagram showing the relationship between respiratory rate and pulse rate.
FIG. 29 is a block diagram illustrating a configuration example of a filter bank.
FIG. 30 is a block diagram illustrating a configuration example of an inverse filter bank.
FIG. 31 is a diagram illustrating an example of a photoelectric pulse wave sensor according to a modification.
FIG. 32 is a diagram showing an example in which the photoelectric pulse wave sensor is applied to eyeglasses in a modified example.
FIG. 33 is a perspective view showing an external configuration of a pulse wave diagnostic apparatus using a pressure sensor in a modified example.
[Explanation of symbols]
MH pulse waveform
10 Pulse wave detector (pulse wave detection means)
130 Pulse wave detection sensor unit (pulse wave detection means)
11 Body motion detection unit (body motion detection means)
TH cursive waveform
MH 'body motion elimination pulse wave waveform, body motion elimination pulse wave data
13 body motion removal unit (body motion removal means)
14 Respiratory component extraction hand (respiratory component extraction means)
15 Evaluation part (means for generating strength of handling)
MKD body motion removal pulse wave analysis data
20 Wavelet converter
VKD respiratory waveform analysis data
21 Respiratory component generator (respiratory waveform generator)
22 Inverse wavelet converter (respiration waveform generator)
VH respiratory waveform
30 First FFT processing unit (first frequency analysis unit, frequency analysis unit)
31 2nd FFT process part (2nd frequency analysis part)
13 'pulse wave component removal unit,
41 Basic frequency table
42 Difference detection unit (frequency identification unit)
43 Comparison part (frequency identification part)
44 Harmonic frequency generation unit (frequency identification unit, extraction unit)
16 Display (notification means)

Claims (15)

Pulse wave detection means for detecting a pulse wave waveform from a detection site of a living body;
Body motion detecting means for detecting a body motion waveform indicating the body motion of the living body;
Body motion removing means for generating a body motion component in the pulse wave waveform based on the body motion waveform, and removing the body motion component from the pulse wave waveform to generate a body motion removal pulse wave waveform;
A respiratory component extracting means for extracting a respiratory component based on the body motion removal pulse wave waveform;
An exercise intensity detection apparatus comprising: a exercise intensity generation unit that calculates exercise intensity based on the respiratory component extracted by the respiratory component extraction unit. The respiratory component extraction means includes
A wavelet conversion unit that generates wave motion removal pulse wave analysis data by performing wavelet conversion on the body motion removal pulse wave waveform;
A respiratory waveform analysis data is generated by removing a frequency component corresponding to a pulse wave component from the body motion removal pulse wave analysis data, and a reverse waveform is applied to the respiratory waveform analysis data to generate a respiratory waveform as the respiratory component. The exercise intensity detection device according to claim 1, further comprising: a respiratory waveform generation unit that performs the operation. 2. The exercise intensity generation unit calculates the exercise intensity based on a ratio of frequency components obtained by performing frequency analysis on the respiratory component extracted by the respiratory component extraction unit. Or the exercise intensity detection apparatus of 2. The mobility intensity generation unit calculates a distortion rate from each frequency component obtained by performing frequency analysis on the respiratory component extracted by the respiratory component extraction unit, and calculates the exercise intensity based on the distortion rate. The exercise intensity detection device according to claim 3. The mobility intensity generating means calculates a ratio between a fundamental frequency component and a third harmonic component obtained by performing frequency analysis on the respiratory component extracted by the respiratory component extracting means, and based on the ratio The exercise intensity detection device according to claim 3, wherein the exercise intensity is calculated. The respiratory component extraction means extracts a respiratory waveform as the respiratory component,
3. The exercise intensity generation unit detects a duty ratio of the respiration waveform extracted by the respiration component extraction unit, and generates the exercise intensity based on the duty ratio. 4. The exercise intensity detection device described. The body movement removing means includes
A first frequency analysis unit for analyzing a frequency spectrum of the pulse wave waveform;
A second frequency analysis unit for analyzing a frequency spectrum of the body movement waveform;
The frequency spectrum analyzed by the second frequency analysis unit is removed from the frequency spectrum analyzed by the first frequency analysis unit to generate a body motion removal spectrum from which body motion has been removed. A body movement removal unit
The respiratory component extraction means extracts a frequency spectrum corresponding to the fundamental component of the respiratory component from the body movement removal spectrum,
The exercise intensity generating unit calculates the exercise intensity based on a frequency spectrum level corresponding to a fundamental wave component of the respiratory component and a frequency spectrum level corresponding to a harmonic component thereof. 2. The exercise intensity detection device according to 1. The breathing component extracting means identifies a band determined according to the pulse rate from the body movement removal spectrum, and extracts a frequency spectrum corresponding to a fundamental wave component of the breathing component from the frequency spectrum within the band. The exercise intensity detection device according to claim 7. The mobility intensity generating means calculates a distortion rate of a respiratory waveform based on a spectrum level corresponding to a fundamental wave component of the respiratory component and a spectrum level corresponding to a harmonic component thereof, and based on the distortion rate The exercise intensity detection device according to claim 7 or 8, wherein the exercise intensity is calculated. The mobility intensity generating means obtains a ratio between a spectrum level corresponding to the fundamental wave component of the respiratory component and a spectrum level corresponding to the third harmonic component, and calculates the exercise intensity based on the ratio. The exercise intensity detection device according to claim 7 or 8, wherein the exercise intensity detection device is calculated. Pulse wave detection means for detecting a pulse wave waveform from a detection site of a living body;
A respiratory component extraction means for extracting a respiratory component from the pulse wave waveform;
An exercise intensity detection apparatus comprising: a exercise intensity generation unit that calculates exercise intensity based on the respiratory component extracted by the respiratory component extraction unit. The respiratory component extraction means includes
A frequency analysis unit for generating a pulse wave analysis data by performing frequency analysis on the pulse wave waveform;
A pulse wave component removing unit for removing a pulse wave component from the pulse wave analysis data;
A fundamental frequency table in which the relationship between the body motion fundamental wave frequency and the respiratory fundamental wave frequency is stored in association with each other;
With reference to the fundamental frequency table, a frequency identifying unit that identifies a respiratory fundamental wave frequency and a body motion fundamental wave frequency from the analysis data;
An exercise intensity detection device comprising: an extraction unit that calculates each harmonic frequency based on the respiration fundamental frequency identified by the frequency identification unit and extracts a respiration component. The mobility intensity generating means calculates a distortion rate of a respiratory waveform based on a spectrum level corresponding to a fundamental wave component of the respiratory component and a spectrum level corresponding to a harmonic component thereof, and based on the distortion rate The exercise intensity detection device according to claim 11, wherein the exercise intensity is calculated. The mobility intensity generating means obtains a ratio between a spectrum level corresponding to the fundamental wave component of the respiratory component and a spectrum level corresponding to the third harmonic component, and calculates the exercise intensity based on the ratio. The exercise intensity detection device according to claim 11 or 12, wherein the exercise intensity detection device is calculated. The exercise intensity detection device according to claim 1, further comprising a notification unit that notifies the exercise intensity generated by the mobility intensity generation unit.

Images