JP3621816B2 - Axle weight measurement method for traveling vehicles - Google Patents

Description

【００１８】
静的成分Ｓ_ｉ ≠Ｓ_ｊ ，ｉ≠ｊであるので、差分により、その影響を除くと、次の数式２のようになる。
【００１９】
【数２】

【００２０】
そして、次の数式３を満たすパラメータ｛ａ_１，ａ_２，…， ａ_ｎ ｝を線形最小２乗法により当てはめる。
【００２１】
【数３】

【００２２】
上記当てはめにより得られたパラメータ｛ａ_１，ａ_２，…， ａ_ｎ ｝により、各軸の変動成分ｆ_ｍ （ｋ） に共通な特徴である定係数線形差分方程式は、次の数式４のようになる。
【００２３】
【数４】

【００２４】
さらに、
（１−ａ_１ ・Ｚ^−１−ａ_２ ・Ｚ^−２−…−ａ_ｐ ・Ｚ^−ｎ） ｙ（ｋ） ＝０
となり、この差分方程式の特性方程式は、
１−ａ_１ ・Ｚ^−１−ａ_２ ・Ｚ^−２−…−ａ_ｐ ・Ｚ^−ｎ＝０
となる。
【００２５】
上述の数式４の特性方程式の解をＰ_ｉ 、ｉ ＝１，２，…，ｎとすると、
（Ｚ−Ｐ_１ ）（Ｚ−Ｐ_２ ）…（Ｚ−Ｐ_ｎ ）＝０
となる。これらの解の中で、ｚ平面の上半面に位置する解は、
｛Ｐ_ｉ ｜Ｐ_ｉ ＝ａ_ｉ ＋ｊｂ_ｉ ，ｂ_ｉ ＞０，ｉ∈（１，２，…，ｎ）｝
となり、この解の位相を周波数成分とみなす。即ち、サンプリングの周期（微小サンプリング時間）をＴとすると、｛ω_ｉ ＝∠Ｐ_ｉ ／Ｔ｝が、車体の振動成分の周波数であるとみなす。つまり、この｛ω_ｉ ＝∠Ｐ_ｉ ／Ｔ｝の周波数を、上述の演算手法を用いて周波数推定手段１０にて演算・推定する。
【００２６】
なお、３本の車軸４０…を有する車両２の変動成分の周波数の推定手順を図３と図４のグラフ図にて簡単に説明すると、図３に示すように、車両前端の第１軸の変動軸重値データ列を示すグラフ線１９と、その次の第２軸に対応するグラフ線２０と、後端の第３軸に対応するグラフ線２１の夫々について、前記数式１に示した定係数差分方程式を当てはめて、図４に示すように差分・オフセット分をキャンセルし、周波数を推定する。なお、図４に於て、２２は第１軸の変動成分のグラフ線、２３は第２軸の変動成分のグラフ線、２４は第３軸の変動成分のグラフ線である。
【００２７】
次に、共通位相推定手段１１にて、共通位相を推定する。ここで、走行中の車両２の運動は、前述の変動軸重値データ列に含まれているため、共通の時間軸をとれば各振動成分ω_ｉ の位相ψ_ｉ は各車軸４０で共通のはずである。先ず、変動軸重値データ列の平均値と差分を計算する。各車軸４０毎の振幅の大きさは、各軸重量に比例するので振幅の大きさを合わせる。つまり、差分データ列を変動軸重値データ列の平均値Ｓ′_ｍ で除す。厳密には、Ｓ′_ｍ ≠Ｓ_ｍ であるが、この時点ではＳ_ｍ は未知であるので、Ｓ′_ｍ ≒Ｓ_ｍ と考えて、振幅をほぼ同じ高さ（大きさ）にする。それを数式で表すと次の数式５のようになる。
【００２８】
【数５】

【００２９】
上記数式５のｙ _ｍ ^ｄ （ｋ） の回帰式を次の数式６とする。
【００３０】
【数６】

【００３１】
上記数式６に於て、Ｓ_ｍｉｎ ＝０となるように拘束式を加えることにより回帰式の当てはめ精度の向上を図る。即ち、
Ｖ_ｍｉｎ ＝Ｓ_ｍｉｎ
という条件を付加する。ただし、Ｖ_ｍｉｎ は十分小さい値であり、０としても差し支えない。そして、数式６の正規方程式を構成し、その正規方程式にて｛Ａ_ｉ ，Ｂ_ｉ ｝を導く。そして位相ψ_ｉ は、
ψ_ｉ ＝ ｔａｎ^−１（Ｂ_ｉ ／Ａ_ｉ ）
となる。
【００３２】
以上がピッチングのないバウンシングのみの共通位相推定方法であるが、走行中の車両２にはピッチングが現れる。これを考慮するため、前記数式６における車両検知器４の出力の推定式をバウンシングの位相とピッチングの位相を考えて、数式６の代わりにピッチングを考慮した回帰式にて位相ψ_ｉ を推定する。その回帰式は次の数式７となる。
【００３３】
【数７】

【００３４】
この数式７から正規方程式（記載省略）を構成し、その正規方程式にて｛Ａ_ｉ ，Ｂ_ｉ ｝と、次の数式８を導く。そしてピッチングを考慮した共通位相ψ_ｉ を推定する。即ち、この共通位相ψ_ｉ を、上述の計算方法に基づいて共通位相推定手段１１にて演算・推定する。
【００３５】
【数８】

【００３６】
なお、３本の車軸４０…を有する車両２の変動成分の共通位相推定手順を図４と図５のグラフ図にて簡単に説明すると、図４に示すように変動軸重値データ列の差分・オフセット分をキャンセルし、次に、図５に示すように、静的成分（変動軸重値データ列の平均値）にて規制化して振幅の高さをほぼ同一にすることにより、共通位相（各軸に共通する位相）を推定する。なお、図５に於て、２５は第１軸に対応するグラフ線、２６は第２軸に対応するグラフ線、２７は第３軸に対応するグラフ線である。
【００３７】
次に、振幅・静止軸重値推定手段１２にて、各軸の変動成分の振幅を推定する。ここまでに、周波数と共通位相が推定されており、（ω_ｉ ，ψ_ｉ ），ｉ ＝１，２，…，ｐは既知であるから、次の数式９に示した式（１５）について線形最小２乗法で当てはめを行う。
【００３８】
【数９】

【００３９】
具体的には上記数式９より正規方程式（記載省略）を構成し、その正規方程式にて振幅と各軸重の静的成分を推定する。つまり、周波数推定手段１０から送られる周波数と共通位相のデータを振幅・静止軸重値推定手段１２にて、数式９について線形最小２乗法を適用する。
【００４０】
なお、３本の車軸４０…を有する車両２の変動成分の共通位相推定手順を図６のグラフ図にて簡単に説明すると、破線にて示すグラフ線２８は周波数推定手段１０にて推定した周波数ωと共通位相推定手段１１にて推定した共通位相ψを有するサインカーブであり、その共通位相ψに合わせて、第１軸（グラフ線１９）、第２軸（グラフ線２０）と、第３軸（グラフ線２１）の夫々について振幅と軸重の静的成分を車軸毎に推定する。
【００４１】
その後、後処理手段１３にて、静止荷重値を算出する。つまり、静止軸重値Ｓ_ｍ が得られているので、静止荷重値Ｗは、次の数式１０のようになる。
【００４２】
【数１０】

【００４３】
また、後処理手段１３は、軸重値または荷重値を後段の装置に出力する。
【００４４】
上述のように、この走行車両の軸重測定方法によれば、走行中の車両２の静止時の重量（静止荷重値）を高精度に推定することができる。従って、高速道路の料金所等における車重の超過の判定を精度よく行うことができる。また、軸重を検出するための載荷板やロードセル等の部品（軸重検出器３）として、従来から使用されていたものをそのまま利用できるという利点がある。
【００４５】
なお、図１〜図６にて説明した軸重測定方法では、軸重値を精度良く測定することができない場合がある。
【００４６】
これに対応するため、一旦、数式９で表される各車軸４０の変動成分の振幅を推定した後、過大な振幅について振幅制限を行って、その後、各車軸４０の変動成分の振幅を推定して、車両２の静止軸重値を推定する。
【００４７】
例えば、第３軸の振幅が過大であった場合、最大振幅をとるべき第２軸の振幅を基準として、静的成分の代用として、平均値法で得られた平均値Ｓ′_ｍ を使用して、制限式
Ａ_３ ＝Ａ_２ ・Ｓ′_３ ／Ｓ′_２
を正規方程式に追加して振幅の制限を行った後、再度振幅の推定を行って、静止軸重値を推定する。
【００４８】
また、ピッチングが顕著なときは、数式９の合成の前段階の形である次の数式１１において、
Ａ_３ ＝Ａ_２ ・Ｓ′_３ ／Ｓ′_２
Ｂ_３ ＝Ｂ_２ ・Ｓ′_２ ／Ｓ′_３
を正規方程式に追加し、同様の再推定を行えばよい。ただし、数式１１におけるＡ_ｉ はバウンシングの振幅，Ｂ_ｉ はピッチングの振幅である。
【００４９】
【数１１】

【００５０】
また、予期せぬ車両の挙動により、振幅制限を行っても、静止軸重値が信頼できないことがあるため、静止軸重値の推定にあたって、変動軸重値データ列の平均値Ｓ′_ｍ をそのまま代用するのも安全策として好ましい。
【００５１】
具体的には、振幅の推定と共に得られる静止軸重値について
（静止軸重値−平均値）／平均値≧Ｒ
となった場合に、推定静止軸重値として平均値を採用すればよい。ここで、Ｒは定数であり、例えば、Ｒ＝０．２５とするのが好ましいが、それ以外の数値としても良い場合がある。
【００５２】
次に、本発明に係る走行車両の軸重測定方法の他の実施の形態について説明する。この場合、図７のフローチャート図に示すように、先ず、変動軸重値データ列を作成し、その後、変動成分の周波数を推定し、次に、その周波数を演算して各車軸４０毎に変動成分の位相と振幅及び静止軸重値を推定する。
【００５３】
この軸重測定方法に於て、変動軸重値データ列の作成と変動成分の周波数の推定は、図１〜図４にて説明した方法と同様の方法にて行う。しかして、各車軸４０毎の変動成分の位相と振幅の推定方法を説明すると、既に、変動成分がｐ個の周波数成分をもつことが分かっていると仮定し、軸重計の出力の推定値ｙ′_ｍ （ｋ） をｐ個の正弦波で次の数式１２として表現する。
【００５４】
【数１２】

【００５５】
これを最小２乗法を用いて、次の数式１３の値となるように、パラメータ｛Ａ_ｉ ，Ｂ_ｉ ，Ｓ_ｍ ｝，ｉ＝１，２，…，ｐを推定する。Ｓ_ｍ は第ｍ軸の静的成分（静止軸重値）である。即ち、当てはめにより各軸毎に変動成分の位相と振幅及び静止軸重値を同時に推定する。
【００５６】
【数１３】

【００５７】
この測定方法によれば、走行車両のピッチングやバウンシングの関係を考慮する必要が無く、演算を簡素化できる。
【００５８】
なお、本発明によれば、車軸の本数が３本の場合以外にも、適用可能であり、２本，４本，５本は勿論のこと、６本以上の整数本であっても適用できる。
【００５９】
【発明の効果】
本発明は上述の構成により、次のような著大な効果を奏する。
【００６０】
請求項１記載の走行車両の軸重測定方法によれば、走行中の車両２の静止時の重量（静止荷重値）を高精度に推定することができる。従って、高速道路の料金所等に於て、車重の超過等の判定を精度よく行うことができる。また、軸重を検出するための載荷板やロードセル等の部品として、従来から使用されていたものをそのまま利用できるという利点がある。また、ピッチングやバウンシングの関係を考慮した静止荷重値を推定できる。
【００６１】
請求項２記載の走行車両の軸重測定方法によれば、走行中の車両２の静止荷重値を一層高精度に推定することができる。従って、車重の超過等の判定を一層精度よく行うことができる。
【００６２】
請求項３記載の走行車両の軸重測定方法によれば、走行中の車両２の静止時の重量（静止荷重値）を高精度に推定することができる。従って、高速道路の料金所等に於て、車重の超過等の判定を精度よく行うことができる。また、軸重を検出するための載荷板やロードセル等の部品として、従来から使用されていたものをそのまま利用できるという利点がある。また、ピッチングやバウンシングの関係を考慮する必要が無く、演算を簡素化できる。
【図面の簡単な説明】
【図１】軸重測定装置とそれに接続される機器を説明するブロック図である。
【図２】本発明の走行車両の軸重測定方法の実施の一形態のフローチャート図である。
【図３】３軸の車両の変動軸重値データ列を説明するグラフ図である。
【図４】周波数の推定方法を説明するグラフ図である。
【図５】共通位相の推定方法を説明するグラフ図である。
【図６】振幅の推定方法を説明するグラフ図である。
【図７】本発明の走行車両の軸重測定方法の他の実施の形態のフローチャート図である。
【符号の説明】
２ 車両
４０ 車軸[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method for measuring axle load of a traveling vehicle.
[0002]
[Prior art]
Conventionally, in a vehicle axle load measuring device used on an expressway or the like, each axle load of a running vehicle has been sequentially measured. Since the traveling vehicle vibrates up and down, measuring each axle weight means measuring a value obtained by adding a dynamic fluctuation amount to the axle weight value at rest. The conventional axle load measuring device has no means for estimating the axle load value at rest in correspondence with the dynamic fluctuation amount.
[0003]
[Problems to be solved by the invention]
By the way, the length of the loading plate of the axle load measuring device is about 760 mm in the traveling direction of the vehicle, and the distance between the load transducers (load cells) is mostly about 660 to 670 mm. Assuming that the speed at which the vehicle passes through the toll gate is about 20 km / h and the ground contact width of the tire is 250 mm, the time for which the tire is completely on the loading plate (measurement time) is 0.09. It will be about seconds. Further, it has been found from experience that the frequency of ground pressure fluctuation (frequency of fluctuation component) given to the road surface of the vehicle affecting measurement is around 3 Hz. Therefore, if the fluctuation waveform is a sine wave, only about one third of one cycle can be measured. In the conventional axle load measuring device, since the axle load value is estimated from the measurement result of only one axle weight, it is impossible to estimate the stationary axle load value.
[0004]
Therefore, an object of the present invention is to provide a method for measuring the axle load of a traveling vehicle, which can solve the above-described problems and can accurately estimate the static load value of the traveling vehicle.
[0005]
[Means for Solving the Problems]
In order to achieve the above-mentioned object, the axle load measuring method for a traveling vehicle according to the present invention measures a variable axle weight value of a running vehicle many times at a predetermined minute sampling time by a single axle weight detector. A variable axle weight value data string for each axle, and the fluctuation component of each axle is expressed by a constant coefficient difference equation having common parameters on the assumption that it is a continuous vibration. Applying the above constant coefficient difference equation to estimate the frequency of the fluctuation component of the vehicle while traveling, then estimating the common phase that is the phase of the fluctuation component of the vehicle body based on the frequency, The common phase is used to estimate the amplitude of the fluctuation component of each axle and the target stationary axle weight value.
[0006]
In addition, a variable axle weight value of a running vehicle is measured many times at a predetermined minute sampling time by a single axle weight detector to obtain a variable axle weight value data string for each axle, and the fluctuation component of each axle is assuming a continuous vibration represented by a constant coefficient difference equation having a common parameter for the variation axle load value data string frequency fluctuation component of the traveling vehicle by applying the constant coefficient difference equation Next, the common phase, which is the phase of the fluctuation component of the vehicle body, is estimated based on the frequency, and the amplitude of the fluctuation component of each axle is estimated by calculating the common phase, and then estimated. As a result, the amplitude is limited for the excessive amplitude, and then the amplitude of the fluctuation component of each axle is estimated to estimate the stationary axle weight value of the vehicle.
[0007]
In addition, a variable axle weight value of a running vehicle is measured many times at a predetermined minute sampling time by a single axle weight detector to obtain a variable axle weight value data string for each axle, and the fluctuation component of each axle is assuming a continuous vibration represented by a constant coefficient difference equation having a common parameter for the variation axle load value data string frequency fluctuation component of the traveling vehicle by applying the constant coefficient difference equation Then, using the frequency, the phase and amplitude of the fluctuation component and the target stationary axle weight value are estimated for each axle.
[0008]
DETAILED DESCRIPTION OF THE INVENTION
The present invention will be described in detail below based on the illustrated embodiment.
[0009]
In FIG. 1, reference numeral 1 denotes an axle load measuring device. The axle load measuring device 1 includes an axle load detector 3 embedded in a road surface 30, a vehicle detector 4 for discriminating one vehicle 2 at a time, It comprises an arithmetic processing unit 5 that performs arithmetic processing on signals from the axle load detector 3 and the vehicle detector 4, a printer 14, and a warning indicator 15. The axle load detector 3 includes a load plate (not shown) that detects the load (axle weight) received from the wheels 2a and the wheels 2a of the vehicle 2 for each axle 40.
[0010]
The arithmetic processing unit 5 includes an AD converter 6 that receives a signal from the axle load detector 3, and an axle that receives the signal from the AD converter 6 and the signal from the vehicle detector 4 to detect the number of axles of the vehicle 2. The number detection means 7, the inter-axis time detection means 8 for detecting the time between the axes, the number-of-axes detection means 7, the inter-axis time detection means 8, and the signals from the AD converter 6 are received for each axle 40. The load signal pre-processing means 9 which computes the converted load signal to obtain the variable axle weight value data string for each axle 40, and the variable axle weight value data string calculated by the load signal pre-processing means 9 is calculated. A frequency estimation means 10 for estimating the frequency of the axle load fluctuation, a common phase estimation means 11 for calculating the estimated frequency and the fluctuation axle weight value data string sent from the frequency estimation means 10 to estimate the common phase of the vehicle body, Estimated common phase and estimated frequency sent from the common phase estimation means 11 The variable / shaft weight value data string is calculated by using the number and the amplitude / static shaft weight value estimating means 12 for estimating the amplitude and the static shaft weight value of the shaft weight fluctuation and the amplitude / static shaft weight value estimating means 12 are sent. Post-processing means 13 for calculating a static load value from the static axle weight value (by a calculation method such as a sum of static axle weight values) and determining whether or not the value exceeds an allowable weight value.
[0011]
Further, the post-processing means 13 of the arithmetic processing unit 5 has a printer 14 for printing a shaft weight value, a vehicle weight value, etc., a warning indicator 15 for displaying a warning when the static load value exceeds the allowable weight value, A photographer 16 for taking a photograph of the vehicle 2 exceeding the allowable weight value is electrically connected. Alternatively, the transmission processing means 17 may be electrically connected to the post-processing means 13 and the photographer 16, and the central monitoring means 18 may be connected to the transmission processing means 17 via a connection line such as a communication line. By doing so, centralized management becomes possible.
[0012]
By the way, the length of the loading plate of the axle load measuring device is about 760 mm in the traveling direction of the vehicle 2, and the distance between the load converters (load cells) is about 660 to 670 mm in most cases. The speed at which the vehicle 2 passes through the toll gate is limited to about 20 km / h, and assuming that the ground contact width of the tire is 250 mm, the time (measurement time) that the tire is completely on the loading plate is 0. It will be about 09 seconds. Further, it has been found from experience that the frequency of the ground pressure fluctuation (frequency of the fluctuation component) applied to the road surface 30 of the vehicle 2 affecting the measurement is around 3 Hz. Therefore, if the fluctuation waveform is a sine wave, only about one third of one cycle can be measured.
[0013]
The axial weight measuring method for a traveling vehicle according to the present invention is a method for accurately estimating a static axial weight value (a stationary weight value) from such a variable axial weight data string having a fluctuation component that is less than one cycle. In other words, considering that the amount of change in the axle load value of each axle is affected by the vibration of the rigid body, the change in axle load measured in a short time on a single loading plate is captured for one vehicle. The stationary shaft weight value is estimated by estimating the waveform (frequency, phase, amplitude) of the contact pressure applied to the road surface for one vehicle being measured.
[0014]
Thus, the axle load measuring method of this traveling vehicle will be described with reference to the block diagram of FIG. 1 and the flowchart of FIG. First, the vehicle detector 4 detects the vehicle 2 and the axle load detector 3 starts measuring the axle load value of the running vehicle 2. At this time, measurement is performed many times for each predetermined minute sampling time. Then, the load signal preprocessing means 9 extracts only the data when the wheel 2a is completely placed on the loading plate of the axle weight detector 3, and the variable axle weight consisting of a large number of discretized axle weight value data. Value data string. Also, count the number of axles. If necessary, signal noise is removed using filtering / smoothing.
[0015]
Next, the frequency estimation means 10 estimates the frequency of the fluctuation component. Explaining this in detail, each data of the variable axial weight value data string from the load signal preprocessing means 9 is
y _m (k) m = 1, 2,..., N k = 0, 1, 2.
And N is the number of axes, m is the axis number, and k is the number of samples per axis. y _m (k) can be expressed as the sum of the fluctuation component f _m (k) and the static axis weight value S _m (static component) as follows.
_{y m (k) = f m} (k) + S m
[0016]
Here, assuming that f _m (k) is a continuous vibration, f _m (k) is expressed by a constant coefficient difference equation, and its vibration component is estimated. The constant coefficient nth-order difference equation of the m-th axis weight detector output is as shown in the following equation 1.
[0017]
[Expression 1]

[0018]
Since the static components S _i ≠ S _j and i ≠ j, if the influence is excluded due to the difference, the following Expression 2 is obtained.
[0019]
[Expression 2]

[0020]
Then, parameters {a ₁ , a ₂ ,..., _An } satisfying the following mathematical formula 3 are applied by the linear least square method.
[0021]
[Equation 3]

[0022]
Based on the parameters {a ₁ , a ₂ ,..., _An } obtained by the above fitting, a constant coefficient linear difference equation that is a common feature to the fluctuation component f _m (k) of each axis is expressed as become.
[0023]
[Expression 4]

[0024]
further,
(1-a ₁ · Z ⁻¹ −a ₂ · Z ⁻² −... −a _p · Z ⁻ⁿ ) y (k) = 0
The characteristic equation of this difference equation is
1-a ₁ · Z ⁻¹ −a ₂ · Z ⁻² −... −a _p · Z ⁻ⁿ = 0
It becomes.
[0025]
If the solution of the characteristic equation of Equation 4 above is P _i , i = 1, 2,..., N,
(ZP ₁ ) (ZP ₂ )... (ZP _n ) = 0
It becomes. Among these solutions, the solution located on the upper half of the z plane is
{P _i | P _i = a _i + jb _i , b _i > 0, iε (1, 2,..., N)}
Thus, the phase of this solution is regarded as a frequency component. That is, if the sampling period (minute sampling time) is T, {ω _i = ∠P _i / T} is regarded as the frequency of the vibration component of the vehicle body. That is, the frequency estimation means 10 calculates and estimates the frequency of {ω _i = 演算 P _i / T} using the above-described calculation method.
[0026]
The procedure for estimating the frequency of the fluctuation component of the vehicle 2 having the three axles 40 will be briefly described with reference to the graphs of FIGS. 3 and 4. As shown in FIG. The graph line 19 indicating the variable axis weight value data string, the graph line 20 corresponding to the second axis next thereto, and the graph line 21 corresponding to the third axis at the rear end are each represented by the above-described formula 1. A coefficient difference equation is applied to cancel the difference / offset and estimate the frequency as shown in FIG. In FIG. 4, 22 is a graph line of the fluctuation component of the first axis, 23 is a graph line of the fluctuation component of the second axis, and 24 is a graph line of the fluctuation component of the third axis.
[0027]
Next, the common phase estimation means 11 estimates the common phase. Here, motion of the vehicle 2 during traveling, because it is in a fluctuation axle load value data string above, taking a common time axis of each vibration component omega _i phase [psi _i is common to each axle 40 It should be. First, the average value and the difference of the variable axis weight value data string are calculated. Since the amplitude of each axle 40 is proportional to the weight of each axle, the amplitude is matched. That is, the difference data string is divided by the average value S ′ _m of the variable axis weight value data string. Strictly speaking, S ′ _m ≠ S _m , but since S _m is unknown at this point, it is considered that S ′ _m ≈S _m and the amplitudes are set to substantially the same height (size). This can be expressed by the following mathematical formula 5.
[0028]
[Equation 5]

[0029]
The regression equation of y _m ^d (k) in the above equation 5 is defined as the following equation 6.
[0030]
[Formula 6]

[0031]
In the above equation 6, a constraint equation is added so that S _min = 0, thereby improving the accuracy of fitting the regression equation. That is,
V _min = S _min
Add the condition. However, V _min is a sufficiently small value and may be 0. Then, a normal equation of Equation 6 is constructed, and {A _i , B _i } is derived from the normal equation. And the phase ψ _i is
ψ _i = tan ⁻¹ (B _i / A _i )
It becomes.
[0032]
The above is the common phase estimation method only for bouncing without pitching, but pitching appears in the traveling vehicle 2. In order to take this into consideration, considering the bouncing phase and the pitching phase as the estimation formula for the output of the vehicle detector 4 in the equation 6, the phase ψ _i is estimated by a regression equation that considers pitching instead of the equation 6. . The regression equation is the following Equation 7.
[0033]
[Expression 7]

[0034]
A normal equation (not shown) is constructed from this equation 7, and {A _i , B _i } and the following equation 8 are derived from the normal equation. Then, the common phase ψ _i considering the pitching is estimated. That is, the common phase ψ _i is calculated and estimated by the common phase estimation means 11 based on the above calculation method.
[0035]
[Equation 8]

[0036]
The common phase estimation procedure of the fluctuation component of the vehicle 2 having the three axles 40 will be briefly described with reference to the graphs of FIGS. 4 and 5. As shown in FIG. -By canceling the offset and then regulating with the static component (average value of the variable axis weight value data string) and making the amplitude heights approximately the same as shown in FIG. (Phase common to each axis) is estimated. In FIG. 5, 25 is a graph line corresponding to the first axis, 26 is a graph line corresponding to the second axis, and 27 is a graph line corresponding to the third axis.
[0037]
Next, the amplitude / stationary axis weight value estimation means 12 estimates the amplitude of the fluctuation component of each axis. Since the frequency and the common phase have been estimated so far, and (ω _i , ψ _i ), i = 1, 2,..., P are known, linearity is obtained with respect to equation (15) shown in the following equation 9. Fit using the least squares method.
[0038]
[Equation 9]

[0039]
Specifically, a normal equation (not shown) is constructed from Equation 9 above, and the amplitude and the static component of each axial load are estimated using the normal equation. That is, the frequency and common phase data sent from the frequency estimation means 10 are applied to the equation 9 by the linear least square method using the amplitude / static axis weight value estimation means 12.
[0040]
The common phase estimation procedure of the fluctuation component of the vehicle 2 having the three axles 40... Is simply described with reference to the graph of FIG. 6. The graph line 28 shown by the broken line indicates the frequency estimated by the frequency estimation means 10. ω and a sine curve having a common phase ψ estimated by the common phase estimation means 11, and in accordance with the common phase ψ, a first axis (graph line 19), a second axis (graph line 20), and a third For each axis (graph line 21), the static components of amplitude and axle load are estimated for each axle.
[0041]
Thereafter, the post-processing means 13 calculates a static load value. That is, since the static shaft weight value _Sm is obtained, the static load value W is expressed by the following Expression 10.
[0042]
[Expression 10]

[0043]
Further, the post-processing means 13 outputs the axle load value or the load value to the subsequent apparatus.
[0044]
As described above, according to the axle load measuring method of the traveling vehicle, the stationary weight (static load value) of the traveling vehicle 2 can be estimated with high accuracy. Accordingly, it is possible to accurately determine whether the vehicle weight is exceeded at a tollgate on an expressway. Further, there is an advantage that conventionally used components (axial load detector 3) such as a loading plate and a load cell for detecting the axial load can be used as they are.
[0045]
1 to 6 may not be able to measure the axial load value with high accuracy.
[0046]
To cope with this, once the amplitude of the fluctuation component of each axle 40 represented by Expression 9 is estimated, the amplitude of the excessive amplitude is limited, and then the amplitude of the fluctuation component of each axle 40 is estimated. Thus, the stationary axle weight value of the vehicle 2 is estimated.
[0047]
For example, when the amplitude of the third axis is excessive, the average value S ′ _m obtained by the average value method is used as a substitute for the static component with reference to the amplitude of the second axis that should take the maximum amplitude. Thus, the limiting expression A ₃ = A ₂ · S ′ ₃ / S ′ ₂
Is added to the normal equation to limit the amplitude, and then the amplitude is estimated again to estimate the static axis weight value.
[0048]
Further, when pitching is remarkable, in the following formula 11 which is the form of the previous stage of the synthesis of formula 9,
A ₃ = A ₂ · S ′ ₃ / S ′ ₂
B ₃ = B ₂ · S ′ ₂ / S ′ ₃
Can be added to the normal equation and re-estimated in the same way. In Equation 11, A _i is the bouncing amplitude, and B _i is the pitching amplitude.
[0049]
[Expression 11]

[0050]
In addition, even if the amplitude is limited due to unexpected vehicle behavior, the stationary axle weight value may not be reliable. Therefore, when estimating the stationary axle weight value, the average value S ′ _m of the variable axle weight value data string is calculated. It is also preferable as a safety measure to substitute as it is.
[0051]
Specifically, the static axle weight value obtained together with the amplitude estimation (static axle weight value−average value) / average value ≧ R
In this case, an average value may be adopted as the estimated stationary shaft weight value. Here, R is a constant. For example, R = 0.25 is preferable, but other numerical values may be used.
[0052]
Next, another embodiment of the axle load measuring method for a traveling vehicle according to the present invention will be described. In this case, as shown in the flowchart of FIG. 7, first, a variable axis weight value data string is created, then the frequency of the variable component is estimated, and then the frequency is calculated to vary for each axle 40. Estimate the phase and amplitude of components and the static axis weight value.
[0053]
In this axial load measuring method, the generation of the variable axial load value data string and the estimation of the frequency of the variable component are performed by the same method as described with reference to FIGS. Thus, the estimation method of the phase and amplitude of the fluctuation component for each axle 40 will be explained. It is assumed that the fluctuation component already has p frequency components, and the estimated value of the output of the axle load meter y ′ _m (k) is expressed by the following Equation 12 with p sine waves.
[0054]
[Expression 12]

[0055]
Using the least square method, parameters {A _i , B _i , S _m }, i = 1, 2,..., P are estimated so as to be the values of the following formula 13. S _m is a static component (static axis weight value) of the m-th axis. That is, the phase and amplitude of the fluctuation component and the stationary axis weight value are simultaneously estimated for each axis by fitting.
[0056]
[Formula 13]

[0057]
According to this measurement method, it is not necessary to consider the relationship between pitching and bouncing of the traveling vehicle, and the calculation can be simplified.
[0058]
In addition, according to this invention, it is applicable besides the case where the number of axles is three, and it can be applied not only to 2, 4, and 5, but also to an integer of 6 or more. .
[0059]
【The invention's effect】
The present invention has the following remarkable effects by the above-described configuration.
[0060]
According to the axle load measuring method of the traveling vehicle according to the first aspect, the stationary weight (static load value) of the traveling vehicle 2 can be estimated with high accuracy. Accordingly, it is possible to accurately determine whether the vehicle weight has been exceeded or the like at a tollgate on an expressway. Moreover, there is an advantage that what has been used conventionally can be used as it is as a component such as a loading plate or a load cell for detecting the axial weight. In addition, a static load value can be estimated in consideration of the relationship between pitching and bouncing.
[0061]
According to the axle load measuring method of the traveling vehicle according to the second aspect, the stationary load value of the traveling vehicle 2 can be estimated with higher accuracy. Therefore, it is possible to determine the vehicle weight excess or the like more accurately.
[0062]
According to the axle load measuring method of the traveling vehicle according to claim 3, the stationary weight (static load value) of the traveling vehicle 2 can be estimated with high accuracy. Accordingly, it is possible to accurately determine whether the vehicle weight has been exceeded or the like at a tollgate on an expressway. In addition, there is an advantage that components conventionally used as components such as a loading plate and a load cell for detecting the axial weight can be used as they are. Further, it is not necessary to consider the relationship between pitching and bouncing, and the calculation can be simplified.
[Brief description of the drawings]
FIG. 1 is a block diagram for explaining a shaft weight measuring device and devices connected thereto.
FIG. 2 is a flowchart of one embodiment of the method for measuring axle load of a traveling vehicle according to the present invention.
FIG. 3 is a graph illustrating a variable axial weight value data string of a three-axis vehicle.
FIG. 4 is a graph illustrating a frequency estimation method.
FIG. 5 is a graph illustrating a common phase estimation method.
FIG. 6 is a graph illustrating an amplitude estimation method.
FIG. 7 is a flowchart of another embodiment of the method for measuring axle load of a traveling vehicle according to the present invention.
[Explanation of symbols]
2 Vehicle 40 axle

Claims (3)

The variable axle weight value of the traveling vehicle (2) is measured a number of times at a predetermined minute sampling time by a single axle weight detector (3) to form a variable axle weight value data string for each axle (40), the fluctuation component of each axle (40), expressed at a constant coefficient difference equation having a common parameter assumed to be sustained oscillations for the fluctuation axle load value data string by applying the constant coefficient difference equation travel The frequency of the fluctuation component of the vehicle (2) in the vehicle is estimated, and then the common phase that is the phase of the fluctuation component of the vehicle body is estimated based on the frequency. A method for measuring the axle load of a traveling vehicle, comprising estimating an amplitude of a fluctuation component of the axle (40) and a stationary axle weight value. The variable axle weight value of the traveling vehicle (2) is measured a number of times at a predetermined minute sampling time by a single axle weight detector (3) to form a variable axle weight value data string for each axle (40), the fluctuation component of each axle (40), expressed at a constant coefficient difference equation having a common parameter assumed to be sustained oscillations for the fluctuation axle load value data string by applying the constant coefficient difference equation travel The frequency of the fluctuation component of the vehicle (2) in the vehicle is estimated, then the common phase that is the phase of the fluctuation component of the vehicle body is estimated based on the frequency, and the common phase is calculated to calculate each axle. The amplitude of the fluctuation component of (40) is estimated, and then the amplitude is limited with respect to the excessive amplitude. Then, the amplitude of the fluctuation component of each axle (40) is estimated, and the stationary axle weight of the vehicle (2) is estimated. A method for measuring the axle load of a traveling vehicle, wherein the value is estimated. The variable axle weight value of the traveling vehicle (2) is measured a number of times at a predetermined minute sampling time by a single axle weight detector (3) to form a variable axle weight value data string for each axle (40), the fluctuation component of each axle (40), expressed at a constant coefficient difference equation having a common parameter assumed to be sustained oscillations for the fluctuation axle load value data string by applying the constant coefficient difference equation travel The frequency of the fluctuation component of the vehicle (2) in the vehicle is estimated, and then the phase and amplitude of the fluctuation component and the stationary axle weight value are estimated for each axle (40) using the frequency. A method for measuring the axle load of a traveling vehicle.

Images