JP5155733B2 - Signal processing method and signal processing apparatus - Google Patents
Signal processing method and signal processing apparatus Download PDFInfo
- Publication number
- JP5155733B2 JP5155733B2 JP2008129214A JP2008129214A JP5155733B2 JP 5155733 B2 JP5155733 B2 JP 5155733B2 JP 2008129214 A JP2008129214 A JP 2008129214A JP 2008129214 A JP2008129214 A JP 2008129214A JP 5155733 B2 JP5155733 B2 JP 5155733B2
- Authority
- JP
- Japan
- Prior art keywords
- signal
- frequency
- low
- phase
- processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Indication And Recording Devices For Special Purposes And Tariff Metering Devices (AREA)
Description
本発明は、センサの出力信号からそのセンサに負荷された物理量に対応した直流成分を速やかに且つ精度よく得るための技術に関する。 The present invention relates to a technique for quickly and accurately obtaining a DC component corresponding to a physical quantity loaded on a sensor from an output signal of the sensor.
各種の物理量を検出するためのセンサには、その物理量の変化に対して過渡的な応答を示すものが多い。 Many sensors for detecting various physical quantities show a transient response to changes in the physical quantities.
例えば、ロードセル等のように物品の質量を検出するためのセンサは、物品の荷重を受けて変形し、その変形量に応じた電圧の信号を出力するが、センサに対する物品の荷重が急激に行なわれた場合、このセンサの系の固有振動モードが励起されてセンサに伝達されるため、その出力信号は非線形振動をする。 For example, a sensor for detecting the mass of an article, such as a load cell, is deformed by receiving the load of the article and outputs a voltage signal corresponding to the amount of deformation, but the load of the article on the sensor is abruptly performed. In this case, since the natural vibration mode of this sensor system is excited and transmitted to the sensor, the output signal exhibits non-linear vibration.
この出力信号の非線形振動は時間が経過するにしたがって減衰して、最終的には物品の質量Ｍに対応した一定の値（直流値）に収束するが、ライン等で物品の質量検査を連続的に行う場合、この振動が完全に収束するまで待っていたのでは効率的な検査がおこなえない。 This non-linear vibration of the output signal is attenuated as time passes, and finally converges to a constant value (DC value) corresponding to the mass M of the article. In this case, an efficient inspection cannot be performed by waiting until this vibration is completely converged.
そこで、一般的には低域通過フィルタにより雑音成分を除去することが行われているが、数１０Ｈｚ以下の雑音成分を除去するためのフィルタの時定数はかなり大きく、収束予想値を高速に得ることは困難であった。 Therefore, noise components are generally removed by a low-pass filter. However, the time constant of the filter for removing noise components of several tens of Hz or less is quite large, and an expected convergence value is obtained at high speed. It was difficult.
この問題を解決するための一つの技術として、本願出願人は、センサの出力信号に含まれる交流信号成分を抽出してその位相を反転させて原信号と加算することにより、センサの出力信号の雑音成分を除去する技術を開示している（特許文献１）。 As one technique for solving this problem, the applicant of the present application extracts an alternating current signal component included in the sensor output signal, inverts the phase thereof, and adds the original signal to the sensor signal. A technique for removing a noise component is disclosed (Patent Document 1).
しかし、上記特許文献１の技術を用いても、床振動等で生じる数１０Ｈｚ以下の雑音成分を精度よく且つ速やかに除去することが困難であった。
However, even if the technique of
特に、位相反転用として一般的に用いられるヒルベルト変換器の周波数特性は基本的にハイパス型であり、極低域までその周波数特性を延ばそうとすれば、そのタップ数が非常に大きくなり、その処理による大きな遅延が生じてしまう。 In particular, the frequency characteristics of a Hilbert transformer that is generally used for phase inversion is basically a high-pass type. If you try to extend the frequency characteristics to an extremely low frequency range, the number of taps will become very large, Cause a large delay.
この問題を解決する方法として、低域の雑音成分な対しては、その利得低下分を見込んでデジタルフィルタによる利得補償を行う方法も採用できるが、その場合においても、処理遅延を短くするためには狭帯域なフィルタ処理をせざるを得ず、低域雑音成分の周波数が変動してしまうと、雑音の除去効果が著しく低下してしまう。 As a method of solving this problem, a method of compensating for gain by a digital filter in anticipation of the gain reduction can be adopted for low-frequency noise components, but even in that case, in order to shorten the processing delay However, if the frequency of the low-frequency noise component fluctuates, the noise removal effect will be significantly reduced.
本発明は、この点を改善し、センサの出力信号に含まれる低周波雑音成分の周波数変動があった場合でも、速やかに且つ精度よく除去して、センサに負荷された物理量に対応する直流成分を得ることができる信号処理方法および装置を提供することを目的としている。 The present invention improves this point, and even when there is a frequency variation of a low frequency noise component included in the output signal of the sensor, the DC component corresponding to the physical quantity loaded on the sensor can be quickly and accurately removed. It is an object of the present invention to provide a signal processing method and apparatus capable of obtaining the above.
前記目的を達成するために、本発明の請求項１の信号処理方法は、
センサの出力信号に対するＡ／Ｄ変換処理で得られた時系列の信号を入力信号として受け、該入力信号に含まれる所定周波数（Ｆａ）より高い高域雑音成分を除去し、該高域雑音成分が除去された第１処理信号（Ｙ（ｋ））から前記所定周波数以下の低域雑音成分を除去して、前記入力信号の直流成分を求める信号処理方法であって、
前記低域雑音成分を除去する処理は、
移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインで前記第１処理信号を増幅する段階（Ｓ５）と、
前記増幅された信号に対して、ヒルベルト変換処理を２段行い、前記増幅された信号に対して１８０度の位相差を有する第１移相信号（Ｎｉ）と、該第１移相信号に対して９０度の位相差を有する第２移相信号（Ｎｑ）とを生成する段階（Ｓ６）と、
前記第１移相信号と第２移相信号とに基づいて、前記増幅された信号に含まれる前記除去対象の低域雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出する段階（Ｓ７）と、
前記算出された瞬時振幅と、前記第１処理信号に含まれる前記除去対象の低域雑音成分の瞬時振幅との誤差を検出する段階（Ｓ８）と、
前記算出した瞬時振幅、瞬時位相、瞬時周波数および前記瞬時振幅の誤差に基づいて前記第１処理信号に含まれる前記除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した第１雑音信号を生成する段階（Ｓ９）と、
前記第１処理信号と前記第１雑音信号とを加算して、前記第１処理信号から前記除去対象の低域雑音成分を除去した第２処理信号（Ｚ（ｋ））を求める段階（Ｓ１０）とを含むことを特徴とする。
In order to achieve the object, a signal processing method according to
A time-series signal obtained by A / D conversion processing for the output signal of the sensor is received as an input signal, a high-frequency noise component higher than a predetermined frequency (Fa) included in the input signal is removed, and the high-frequency noise component A signal processing method for removing a low-frequency noise component having a frequency equal to or lower than the predetermined frequency from the first processed signal (Y (k)) from which a DC component of the input signal is obtained,
The process of removing the low-frequency noise component is:
Amplifying the first processing signal with a fixed gain set in advance to compensate for a gain drop of a low-frequency noise component to be removed that is predicted to be generated by the phase shift processing (S5);
The Hilbert transform process is performed on the amplified signal in two stages, and the first phase-shifted signal (Ni) having a phase difference of 180 degrees with respect to the amplified signal and the first phase-shifted signal are processed. Generating a second phase shift signal (Nq) having a phase difference of 90 degrees (S6),
Calculating an instantaneous amplitude, an instantaneous phase, and an instantaneous frequency of the low-frequency noise component to be removed included in the amplified signal based on the first phase-shifted signal and the second phase-shifted signal (S7); ,
Detecting an error between the calculated instantaneous amplitude and the instantaneous amplitude of the low-frequency noise component to be removed included in the first processing signal (S8);
Based on the calculated instantaneous amplitude, instantaneous phase, instantaneous frequency, and error in the instantaneous amplitude, the first low-frequency noise component included in the first processing signal has the same amplitude, the same frequency, and the phase is inverted. Generating a noise signal (S9);
Adding the first processed signal and the first noise signal to obtain a second processed signal (Z (k)) obtained by removing the low-frequency noise component to be removed from the first processed signal (S10); It is characterized by including.
また、本発明の請求項２の信号処理方法は、請求項１記載の信号処理方法において、
前記低域雑音成分を除去する処理は、
前記固定ゲインで処理信号を増幅する処理から、処理信号と生成した雑音信号とを加算して低域雑音成分を除去するまでの一連の処理（Ｓ５〜Ｓ１０）を、前記第２処理信号に含まれる各雑音成分に対して繰り返し行う（Ｓ１１、１２）ことを特徴とする。
The signal processing method according to
The process of removing the low-frequency noise component is:
The second process signal includes a series of processes (S5 to S10) from the process of amplifying the process signal with the fixed gain to adding the process signal and the generated noise signal to remove the low-frequency noise component. It is characterized by repeatedly performing each noise component (S11, 12).
また、本発明の請求項３の信号処理装置は、
センサ（１１）の出力信号に対するＡ／Ｄ変換処理を行うＡ／Ｄ変換器（２１）と、
前記Ａ／Ｄ変換器の出力信号を受け、所定周波数（Ｆａ）より高い高域雑音成分を除去し、該高域雑音成分を除去して得られた第１処理信号（Ｙ（ｋ））を出力する高域雑音除去部（２２）と、
前記第１処理信号を受け、前記所定周波数以下の低域雑音成分を除去する低域雑音除去部（３０）とを有する信号処理装置であって、
前記低域雑音除去部が、
移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインで前記第１処理信号を増幅するゲイン補償手段（３１）と、
前記ゲイン補償手段によって増幅された信号に対して、ヒルベルト変換処理を２段行い、前記増幅された信号に対して１８０度の位相差を有する第１移相信号（Ｎｉ）と、該第１移相信号に対して９０度の位相差を有する第２移相信号（Ｎｑ）とを生成する移相手段（３２）と、
前記移相手段によって得られた第１移相信号と第２移相信号とに基づいて、前記ゲイン補償手段によって増幅された信号に含まれる前記除去対象の低域雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出する雑音信号情報算出手段（３３）と、
前記雑音信号情報算出手段によって算出された瞬時振幅と、前記第１処理信号に含まれる前記除去対象の低域雑音成分の瞬時振幅との誤差を検出する振幅誤差検出手段（３４）と、
前記雑音信号情報算出手段によって算出された瞬時振幅、瞬時位相、瞬時周波数および振幅誤差検出手段によって検出された前記瞬時振幅の誤差に基づいて、前記第１処理信号に含まれる前記除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した第１雑音信号を生成する雑音信号生成手段（３５）と、
前記第１処理信号と前記第１雑音信号とを加算して、前記第１処理信号から前記除去対象の低域雑音成分を除去した第２処理信号（Ｚ（ｋ））を求める加算手段（３６）とを備えていることを特徴とする。
The signal processing device according to
An A / D converter (21) that performs A / D conversion processing on the output signal of the sensor (11);
The output signal of the A / D converter is received, a high frequency noise component higher than a predetermined frequency (Fa) is removed, and a first processing signal (Y (k)) obtained by removing the high frequency noise component is obtained. An output high-frequency noise removal unit (22);
A signal processing device having a low-frequency noise removing unit (30) that receives the first processing signal and removes a low-frequency noise component of the predetermined frequency or less;
The low-frequency noise removing unit is
Gain compensation means (31) for amplifying the first processing signal with a fixed gain set in advance in order to compensate for a gain drop of a low-frequency noise component to be removed that is predicted to be generated by the phase shift processing;
The signal amplified by the gain compensation means is subjected to two stages of Hilbert transform processing, and the first phase-shifted signal (Ni) having a phase difference of 180 degrees with respect to the amplified signal and the first phase-shifted signal. Phase shifting means (32) for generating a second phase shifting signal (Nq) having a phase difference of 90 degrees with respect to the phase signal;
Based on the first phase shift signal and the second phase shift signal obtained by the phase shift means, the instantaneous amplitude and instantaneous phase of the low frequency noise component to be removed included in the signal amplified by the gain compensation means And noise signal information calculating means (33) for calculating an instantaneous frequency;
Amplitude error detection means (34) for detecting an error between the instantaneous amplitude calculated by the noise signal information calculation means and the instantaneous amplitude of the low-frequency noise component to be removed included in the first processing signal;
Based on the instantaneous amplitude, instantaneous phase, instantaneous frequency, and error of the instantaneous amplitude detected by the amplitude error detecting means calculated by the noise signal information calculating means, the low frequency range to be removed included in the first processing signal A noise signal generating means (35) for generating a first noise signal having the same amplitude and the same frequency as the noise component and having a phase inverted;
Adding means (36) for adding the first processed signal and the first noise signal to obtain a second processed signal (Z (k)) obtained by removing the low-frequency noise component to be removed from the first processed signal. ).
また、本発明の請求項４の信号処理装置は、請求項３記載の信号処理装置において、
前記低域雑音除去部は、
前記ゲイン補償手段、前記移相手段、前記雑音信号情報算出手段、前記振幅誤差検出手段、前記雑音信号生成手段および前記加算手段を一組の雑音除去ブロックとし、該雑音除去ブロックを除去対象の低域雑音成分の数分直列接続して構成され、各雑音除去ブロックでそれぞれ除去対象の低域雑音成分を除去することを特徴とする。
The signal processing device according to
The low-frequency noise removing unit is
The gain compensation unit, the phase shift unit, the noise signal information calculation unit, the amplitude error detection unit, the noise signal generation unit, and the addition unit constitute a set of noise removal blocks, and the noise removal block is a low target to be removed. It is configured by connecting in series the number of band noise components, and each noise removal block removes low band noise components to be removed.
このように本発明では、センサの出力信号に対するＡ／Ｄ変換処理で得られた時系列の信号に対して高域雑音除去処理を行い、その高域雑音除去処理後の第１処理信号に対し、移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインで増幅し、ヒルベルト変換処理を２段行い、増幅された信号に対して１８０度の位相差を有する第１移相信号と、その第１移相信号に対して９０度の位相差を有する第２移相信号とを生成し、それらの信号に基づいて、前記増幅された信号に含まれる前記除去対象の低域雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出し、その算出された瞬時振幅と、前記第１処理信号に含まれる前記除去対象の低域雑音成分の瞬時振幅との誤差を検出し、算出した瞬時振幅、瞬時位相、瞬時周波数および前記瞬時振幅の誤差に基づいて前記第１処理信号に含まれる前記除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した第１雑音信号を生成して、第１雑音信号とを加算することにより、第１処理信号から除去対象の低域雑音成分を除去している。 As described above, in the present invention, the high-frequency noise removal processing is performed on the time-series signal obtained by the A / D conversion processing on the output signal of the sensor, and the first processed signal after the high-frequency noise removal processing is performed. , Amplifying with a fixed gain set in advance to compensate for the gain reduction of the low-frequency noise component to be removed, which is predicted to be generated by the phase-shifting process, performing the Hilbert transform process in two stages, and for the amplified signal A first phase-shifted signal having a phase difference of 180 degrees and a second phase-shifted signal having a phase difference of 90 degrees with respect to the first phase-shifted signal are generated, and the amplified signal is amplified based on these signals. The instantaneous amplitude, instantaneous phase and instantaneous frequency of the low-frequency noise component to be removed included in the received signal are calculated, and the calculated instantaneous amplitude and the low-frequency noise component to be removed included in the first processing signal are calculated. Detects the error from the instantaneous amplitude of A first noise signal having the same amplitude and the same frequency as the low-frequency noise component to be removed included in the first processing signal and having the phase inverted based on the instantaneous amplitude, instantaneous phase, instantaneous frequency, and error of the instantaneous amplitude. Is added to the first noise signal to remove the low-frequency noise component to be removed from the first processed signal.
このため、除去対象の低域雑音成分の周波数変動による振幅誤差が無くなり、低域雑音成分を速やかに且つ精度よく除去でき、直流成分を検出することができる。 For this reason, the amplitude error due to the frequency fluctuation of the low-frequency noise component to be removed is eliminated, the low-frequency noise component can be removed quickly and accurately, and the DC component can be detected.
また、上記処理を繰り返し行うことで、複数の低域雑音成分の除去が可能となる。 In addition, by repeatedly performing the above processing, a plurality of low frequency noise components can be removed.
以下、図面に基づいて本発明の実施の形態を説明する。
先ず始めに、本発明の信号処理方法の一実施形態を図１のフローチャートに基づいて説明する。
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, an embodiment of the signal processing method of the present invention will be described based on the flowchart of FIG.
図１に示しているように、この実施形態の信号処理方法は、質量、圧力等の物理量Ｍを受けたセンサの出力信号ｘ（ｔ）を、その周波数成分の上限の２倍より十分高い周波数でオーバサンプリングして、量子化雑音が極めて少ないデジタルの原信号Ｘ（ｋ）に変換する（Ｓ１）。 As shown in FIG. 1, in the signal processing method of this embodiment, the output signal x (t) of a sensor that has received a physical quantity M such as mass or pressure has a frequency sufficiently higher than twice the upper limit of its frequency component. Oversampling is performed to convert the digital original signal X (k) with very little quantization noise (S1).
そして、この原信号Ｘ（ｋ）に含まれる所定周波数Ｆａ（例えば数１０Ｈｚ）以上の高域雑音成分を除去する（Ｓ２）。この所定周波数Ｆａは、後述する各移相手段の周波数振幅特性の平坦部と低域側の傾斜部の境界近傍に設定されているものとする（図３参照）。 Then, a high frequency noise component having a frequency equal to or higher than a predetermined frequency Fa (for example, several 10 Hz) included in the original signal X (k) is removed (S2). This predetermined frequency Fa is assumed to be set in the vicinity of the boundary between the flat portion of the frequency amplitude characteristic of each phase shift means to be described later and the inclined portion on the low frequency side (see FIG. 3).
高域雑音成分の除去処理については後述するが、例えば前記した特許文献１の処理が使用可能であり、この処理により、例えば、図２の（ａ）のような信号ｘ（ｔ）に対して、図２の（ｂ）のように低域の雑音成分のみが重畳した信号Ｙ（ｋ）を出力する。
Although the high-frequency noise component removal process will be described later, for example, the process of
高域雑音除去処理で所定周波数Ｆａ以上の雑音成分が短時間に除去された第１処理信号Ｙ（ｋ）は、低域雑音除去処理を受けるが、その低域雑音成分の存在を先に調べておくことで、以後の処理に無駄がなくなる。 The first processed signal Y (k) from which the noise component of the predetermined frequency Fa or higher is removed in a short time by the high-frequency noise removal processing is subjected to the low-frequency noise removal processing, but the existence of the low-frequency noise component is examined first. By doing so, there is no waste in subsequent processing.
先ず始めに、スペクトラム解析処理（Ｓ３）を行い、周波数Ｆａ以下の所定レベル以上の雑音成分の有無を調べる（Ｓ４）。 First, spectrum analysis processing (S3) is performed, and the presence / absence of a noise component of a predetermined level equal to or lower than the frequency Fa is checked (S4).
ここで、図３に示しているように、周波数Ｆａ以下の帯域に測定誤差上、無視できないレベルの雑音成分Ｎ１、Ｎ２があり、Ｎ１のレベルが大きい場合を仮定する。この場合、第１処理信号Ｙ（ｋ）に対するレベルの大きいＮ１に対する雑音除去処理が優先的に行われる（レベルが同等の場合には周波数が高い方を優先する）。 Here, as shown in FIG. 3, it is assumed that there are noise components N1 and N2 of a level that cannot be ignored due to measurement errors in the band below the frequency Fa, and the level of N1 is large. In this case, the noise removal processing for N1 having a large level with respect to the first processing signal Y (k) is preferentially performed (when the level is equal, the higher frequency is prioritized).
この低域雑音除去処理は、基本的に元の信号に含まれる交流の雑音成分をヒルベルト変換処理により１８０゜移相して加算することで雑音成分を相殺するものである。 This low-frequency noise removal processing basically cancels the noise component by shifting the AC noise component included in the original signal by 180 ° by the Hilbert transform processing and adding it.
ただし、前記したようにヒルベルト変換器はハイパス型で、図３で示したように周波数Ｆａより下の傾斜部で周波数とともに利得が低下し、図２の（ｃ）のように、１８０°移相処理した後の信号の低域成分の振幅が大きく減衰してしまう。この利得が低下する周波数は、ヒルベルト変換処理のタップ数を多くすることでより低い方へずらすことができるが、そのために大きな遅延時間が生じてしまう。 However, as described above, the Hilbert transformer is a high-pass type, and as shown in FIG. 3, the gain decreases with the frequency at the slope portion below the frequency Fa, and the phase shift by 180 ° as shown in FIG. The amplitude of the low frequency component of the processed signal is greatly attenuated. The frequency at which the gain decreases can be shifted to a lower level by increasing the number of taps in the Hilbert transform process, but this causes a large delay time.
そこで、この発明では、移相処理により生じると予測される除去対象の低域雑音成分（この場合Ｎ１）の利得低下を補償するために予め設定された固定ゲインＧａで第１処理信号Ｙ（ｋ）を増幅する（Ｓ５）。なお、このゲインＧａは雑音Ｎ１に対する周波数解析結果とヒルベルト変換処理の周波数特性により既知となる。 Therefore, in the present invention, the first processing signal Y (k) is set with a fixed gain Ga set in advance to compensate for the gain reduction of the low-frequency noise component to be removed (N1 in this case) that is predicted to be generated by the phase shift processing. ) Is amplified (S5). The gain Ga is known from the frequency analysis result for the noise N1 and the frequency characteristics of the Hilbert transform process.
そして、この増幅した処理信号Ｇａ・Ｙ（ｋ）に対してヒルベルト変換処理を２段行い、信号Ｇａ・Ｙ（ｋ）に対して１８０°移相した第１移相信号Ｎｉとそれに対して９０度移相した第２移相信号Ｎｑを生成する（Ｓ６）。 The amplified processed signal Ga · Y (k) is subjected to two stages of Hilbert transform processing, and the first phase-shifted signal Ni phase-shifted 180 ° with respect to the signal Ga · Y (k) and 90 ° thereto. A second phase-shifted signal Nq that has been phase-shifted is generated (S6).
次に、第１移相信号Ｎｉ、第２移相信号Ｎｑから雑音信号の瞬時振幅、瞬時位相、瞬時周波数を算出する（Ｓ７）。 Next, the instantaneous amplitude, instantaneous phase, and instantaneous frequency of the noise signal are calculated from the first phase shift signal Ni and the second phase shift signal Nq (S7).
即ち、ある時刻ｍＴ（ｍ＝ｋ）における移相信号Ｎｉ（ｍＴ）、第２の移相信号Ｎｑ（ｍＴ）に対して、瞬時位相φ（ｍＴ）、瞬時振幅α（ｍＴ）、瞬時周波数ｆ（ｍＴ）は、以下のように求められる。 That is, for a phase shift signal Ni (mT) and a second phase shift signal Nq (mT) at a certain time mT (m = k), an instantaneous phase φ (mT), an instantaneous amplitude α (mT), and an instantaneous frequency f (MT) is obtained as follows.
φ（ｍＴ）＝ｔａｎ−１［Ｎｑ（ｍＴ）／Ｎｉ（ｍＴ）］ ……（１）
α（ｍＴ）＝｛［Ｎｑ（ｍＴ）］２＋［Ｎｉ（ｍＴ）］２｝１／２ ……（２）
ｆ（ｍＴ）＝１／｛２π［φ（ｍＴ）−φ（ｍＴ−Ｔ）］｝／Ｔ ……（３）
φ (mT) = tan −1 [Nq (mT) / Ni (mT)] (1)
α (mT) = {[Nq (mT)] 2 + [Ni (mT)] 2 } 1/2 (2)
f (mT) = 1 / {2π [φ (mT) −φ (mT−T)]} / T (3)
一方、前記した除去対象の雑音成分の周波数は、環境の変化などにより変動する可能性があり、前記した低域特性から僅かな周波数変動により瞬時振幅が大きく変動してしまう。 On the other hand, the frequency of the noise component to be removed may fluctuate due to a change in environment or the like, and the instantaneous amplitude greatly fluctuates due to a slight frequency fluctuation due to the low frequency characteristics described above.
そこで、算出された瞬時振幅α（ｍＴ）と、第１処理信号Ｙ（ｋ）に含まれる除去対象の低域雑音成分の瞬時振幅との誤差ｅ（比あるいは差のいずれでもよい）を検出する（Ｓ８）。 Therefore, an error e (which may be either a ratio or a difference) between the calculated instantaneous amplitude α (mT) and the instantaneous amplitude of the low-frequency noise component to be removed included in the first processing signal Y (k) is detected. (S8).
この誤差ｅは、例えば第１処理信号Ｙ（ｋ）の微分処理成分（交流成分）と、第１移相信号Ｎｉの差分成分と微分処理成分との比（あるいは差）を求める。 For this error e, for example, the ratio (or difference) between the differential processing component (AC component) of the first processing signal Y (k) and the differential component and differential processing component of the first phase shift signal Ni is obtained.
そして、前記算出した瞬時振幅α（ｍＴ）、瞬時位相φ（ｍＴ）、瞬時周波数ｆ（ｍＴ）および前記瞬時振幅の誤差ｅに基づいて、瞬時振幅ｅ・α（ｍＴ）、瞬時周波数ｆ（ｍＴ）で、瞬時位相φ（ｍＴ）に処理の遅延などを考慮した初期位相を含めた雑音成分信号、即ち、第１処理信号Ｙ（ｋ）に含まれる除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した第１雑音信号ＮＬ１（ｋ）を生成する（Ｓ９）。 Then, based on the calculated instantaneous amplitude α (mT), instantaneous phase φ (mT), instantaneous frequency f (mT), and error e of the instantaneous amplitude, the instantaneous amplitude e · α (mT), the instantaneous frequency f (mT) ), The noise component signal including the initial phase in consideration of processing delay or the like in the instantaneous phase φ (mT), that is, the same amplitude as the low-frequency noise component to be removed included in the first processing signal Y (k), A first noise signal N L1 (k) having the same frequency and inverted phase is generated (S9).
このようにして得られた第１雑音信号ＮＬ１（ｋ）と第１処理信号Ｙ（ｋ）とを加算して、第１処理信号Ｙ（ｋ）から除去対象の低域雑音成分Ｎ１を除去した第２処理信号（Ｚ（ｋ））を求める（Ｓ１０）。 The first noise signal N L1 (k) thus obtained and the first processed signal Y (k) are added to remove the low-frequency noise component N1 to be removed from the first processed signal Y (k). The second processed signal (Z (k)) is obtained (S10).
つまり、図２の（ｂ）の第１処理信号Ｙ（ｋ）に対して、それに含まれる除去対象のレベル最大の雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出し、瞬時振幅については周波数変動による誤差補正を行うことで、図２の（ｃ）のように第１処理信号Ｙ（ｋ）に含まれる除去対象の雑音成分を１８０゜移相した第１雑音信号ＮＬ１（ｋ）を生成し、第１処理信号Ｙ（ｋ）と加算することで、図２の（ｄ）のように除去対象のレベル最大の雑音成分Ｎ１が除去された第２処理信号Ｚ（ｋ）を得ている。 That is, for the first processing signal Y (k) in FIG. 2B, the instantaneous amplitude, instantaneous phase, and instantaneous frequency of the noise component with the maximum level to be removed included therein are calculated. By performing error correction due to fluctuation, the first noise signal N L1 (k) obtained by shifting the noise component to be removed included in the first processed signal Y (k) by 180 ° as shown in FIG. By generating and adding the first processed signal Y (k), the second processed signal Z (k) from which the noise component N1 having the maximum level to be removed is removed as shown in FIG. 2D is obtained. Yes.
この第２処理信号Ｚ（ｋ）に無視できないレベルの雑音がなければ（Ｓ１１）、第２処理信号Ｚ（ｋ）が最終の処理結果として出力されることになるが、図３に示したように数Ｈｚの低い周波数の床振動等によりさらに低域の雑音成分Ｎ２が存在している場合、図２の（ｄ）のように第２処理信号Ｚ（ｋ）にその雑音成分が重畳している。 If the second processing signal Z (k) does not have a level of noise that cannot be ignored (S11), the second processing signal Z (k) is output as the final processing result, as shown in FIG. If a low frequency noise component N2 exists due to a floor vibration of a low frequency of several Hz, the noise component is superimposed on the second processed signal Z (k) as shown in FIG. Yes.
その場合、第２の処理信号Ｚ（ｋ）に対して前記処理Ｓ５〜Ｓ１０と同様の雑音除去処理がなされる（Ｓ１２）。 In that case, a noise removal process similar to the processes S5 to S10 is performed on the second processed signal Z (k) (S12).
即ち、第２処理信号Ｚ（ｋ）に対して、移相処理により生じると予測される除去対象の低域雑音成分Ｎ２の利得低下を補償するために予め設定された固定ゲインＧｂで増幅し、ヒルベルト変換処理を２段行い、増幅された信号Ｇｂ・Ｚ（ｋ）に対して１８０度の位相差を有する第１移相信号Ｎｉと、その第１移相信号に対して９０度の位相差を有する第２移相信号Ｎｑを生成し、それらの信号に基づいて、信号信号Ｇｂ・Ｚ（ｋ）に含まれる除去対象の低域雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出し、その算出された瞬時振幅と、第２処理信号Ｚ（ｋ）に含まれる除去対象の低域雑音成分Ｎ２の瞬時振幅との誤差ｅを検出し、算出した瞬時振幅、瞬時位相、瞬時周波数および瞬時振幅の誤差に基づいて、第２処理信号Ｚ（ｋ）に含まれる除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した第２雑音信号ＮＬ２（ｋ）を図２の（ｅ）のように生成して、第２処理信号Ｚ（ｋ）と加算することにより、第２処理信号Ｚ（ｋ）から低域雑音成分Ｎ２を除去し、図２の（ｆ）のように、低域雑音が除去された信号Ｗ（ｋ）を求め、これを最終の処理結果として出力する。 That is, the second processed signal Z (k) is amplified by a preset fixed gain Gb in order to compensate for the gain reduction of the low-frequency noise component N2 to be removed, which is predicted to be generated by the phase shift processing, A Hilbert transform process is performed in two stages, a first phase shift signal Ni having a phase difference of 180 degrees with respect to the amplified signal Gb · Z (k), and a phase difference of 90 degrees with respect to the first phase shift signal A second phase-shifted signal Nq having, and based on those signals, calculate the instantaneous amplitude, instantaneous phase and instantaneous frequency of the low-frequency noise component to be removed included in the signal signal Gb · Z (k), An error e between the calculated instantaneous amplitude and the instantaneous amplitude of the low-frequency noise component N2 to be removed included in the second processed signal Z (k) is detected, and the calculated instantaneous amplitude, instantaneous phase, instantaneous frequency, and instantaneous Based on the amplitude error, the second processed signal Z (k) Low-frequency noise component and the amplitude of the removal target contained, and a second noise signal N L2 phase is the inverse of the (k) at the same frequency generated as in FIG. 2 (e), the second processed signal Z ( k) is added to remove the low-frequency noise component N2 from the second processed signal Z (k) to obtain a signal W (k) from which the low-frequency noise has been removed as shown in FIG. 2 (f). This is output as the final processing result.
この最終の処理結果として出力された第３処理信号Ｗ（ｋ）から、センサの出力信号ｘ（ｔ）の直流成分Ｖｍ、即ち、センサに負荷された物理量を正確に予測できる。 From the third processing signal W (k) output as the final processing result, the DC component Vm of the sensor output signal x (t), that is, the physical quantity loaded on the sensor can be accurately predicted.
上記例では、無視できないレベルの２つの低域雑音成分を除去するために、２段の低域雑音除去処理を行っていたが、さらに多くの無視できないレベルの低域雑音がある場合には処理Ｓ１１に戻り、再び同様の処理が繰り返される。 In the above example, in order to remove two low-frequency noise components having a level that cannot be ignored, two-stage low-frequency noise removal processing is performed. Returning to S11, the same processing is repeated again.
このような一連の処理で低域の雑音成分は除去されるが、前記したように、低域の雑音成分に周波数の変動があると、生成した雑音信号の振幅誤差が大きく変動することになるが、この低域雑音の除去処理は、各低域雑音成分について連続的に行われ、雑音成分の各瞬時値と振幅誤差の検出およびそれによる雑音信号の振幅補正を行うので、低域の雑音成分に周波数の変動があっても、それによる誤差の増加が抑制され、常に精度の高い直流値の検出を高速に行うことができる。 Although a low frequency noise component is removed by such a series of processes, as described above, if there is a frequency variation in the low frequency noise component, the amplitude error of the generated noise signal greatly varies. However, this low-frequency noise removal processing is continuously performed for each low-frequency noise component, and each instantaneous value of the noise component and the amplitude error are detected and the amplitude of the noise signal is corrected accordingly. Even if the component has a frequency variation, an increase in error due to the variation is suppressed, and a highly accurate DC value can always be detected at high speed.
なお、上記のように複数段の雑音除去処理を行う場合には、処理の前段にデシメーション処理を行い、それに続くヒルベルト変換器の特性を共通化することもできる。 Note that when performing multi-stage noise removal processing as described above, decimation processing can be performed before the processing, and the characteristics of the subsequent Hilbert transformer can be made common.
このように、実施形態の信号処理方法では、センサ１の出力信号に対するＡ／Ｄ変換処理で得られた時系列の信号に対して高域雑音除去処理を行い、その結果得られた第１処理信号に対して、移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインで増幅し、ヒルベルト変換処理を２段行い、増幅された信号に対して１８０度の位相差を有する第１移相信号と、その第１移相信号に対して９０度の位相差を有する第２移相信号とを生成し、それらの信号に基づいて、前記増幅された信号に含まれる前記除去対象の低域雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出し、その算出された瞬時振幅と、前記第１処理信号に含まれる前記除去対象の低域雑音成分の瞬時振幅との誤差を検出し、算出した瞬時振幅、瞬時位相、瞬時周波数および前記瞬時振幅の誤差に基づいて前記第１処理信号に含まれる前記除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した雑音信号を生成して、第１処理信号と加算することにより、第１処理信号に含まれる低域雑音成分を除去している。
As described above, in the signal processing method of the embodiment, the high-frequency noise removal processing is performed on the time-series signal obtained by the A / D conversion processing on the output signal of the
このため、除去対象の低域雑音成分の周波数変動による振幅誤差が無くなり、低域雑音成分を速やかに且つ精度よく除去でき、直流成分を検出することができる。 For this reason, the amplitude error due to the frequency fluctuation of the low-frequency noise component to be removed is eliminated, the low-frequency noise component can be removed quickly and accurately, and the DC component can be detected.
なお、高域雑音除去の処理は任意であり、簡単にはローパスフィルタのみで高域雑音成分を除去してもよく、また、前記した特許文献１の処理で極めて高速な雑音除去処理を行うこともできる。
Note that the high-frequency noise removal process is arbitrary, and a high-frequency noise component may be simply removed using only a low-pass filter. In addition, the extremely high-speed noise removal process may be performed using the process described in
図４は、上記実施形態の信号処理方法を用いた実施形態の信号処理装置２０の構成を示している。
FIG. 4 shows the configuration of the
この信号処理装置２０は、ロードセル等のセンサ１１の出力信号ｘ（ｔ）を、Ａ／Ｄ変換器２１により例えば数ｋＨｚでオーバサンプリングして、量子化雑音が少ないデジタルの原信号Ｘ（ｋ）を得て、高域雑音除去部２２に入力している。
The
高域雑音除去部２２は、移相手段２３、遅延手段２４および加算手段２５により構成されている。
The high-frequency noise removal unit 22 includes a
原信号Ｘ（ｋ）を受ける移相手段２３は、９０度の移相を行う２つのヒルベルト変換器２３ａ、２３ｂを直列に接続したものであり、そのタップ数は、周波数Ｆａ以上の周波数成分に対して平坦な周波数特性が得られる最小数に設定されている。 The phase shift means 23 that receives the original signal X (k) is a unit in which two Hilbert transformers 23a and 23b that perform a phase shift of 90 degrees are connected in series, and the number of taps is a frequency component equal to or higher than the frequency Fa. On the other hand, the minimum number is set so that a flat frequency characteristic can be obtained.
なお、上記のようにヒルベルト変換器を用いた移相手段２３の場合、原理的に直流分は除去されているので、原信号に重畳している交流成分に対して１８０度移相した雑音信号ＮＨ（ｋ）を生成できる。ただし、低域の雑音成分については元の成分に対して振幅が低下している。 In the case of the phase shift means 23 using the Hilbert converter as described above, since the DC component is removed in principle, the noise signal shifted by 180 degrees with respect to the AC component superimposed on the original signal. N H (k) can be generated. However, the amplitude of the low frequency noise component is lower than the original component.
また、遅延手段２４は、信号Ｘ（ｋ）に対し、移相手段２３の移相処理に必要な処理時間分の遅延を与え、加算器２５に出力する。
The delay unit 24 gives a delay corresponding to the processing time necessary for the phase shift processing of the
加算手段２５には、原信号Ｘ（ｋ）と、その原信号に含まれていた周波数Ｆａ以上の高域雑音成分を１８０度移相した雑音信号ＮＨ（ｋ）とが入力されることになり、その加算処理により高域雑音成分が相殺除去された第１処理信号Ｙ（ｋ）が前記図２の（ｂ）のように出力されることになる。 The addition means 25 is supplied with the original signal X (k) and the noise signal N H (k) obtained by shifting the high frequency noise component of the frequency Fa or higher included in the original signal by 180 degrees. Thus, the first processed signal Y (k) from which the high-frequency noise component has been canceled out by the addition process is output as shown in FIG.
このようにして得られた第１処理信号Ｙ（ｋ）は、スペクトラム解析手段２６、低域雑音除去部３０に入力される。 The first processing signal Y (k) obtained in this way is input to the spectrum analyzing means 26 and the low-frequency noise removing unit 30.
スペクトラム解析手段２６は、第１処理信号Ｙ（ｋ）に対するスペクトラム解析を行い、周波数Ｆａ以下の雑音成分を求める。
The
雑音分布判定手段２７は、スペクトラム解析手段２６で無視できないレベルの雑音成分があるか否かを判定する。この判定結果によって後述する信号切換用のスイッチ２８が切り替わる。
The noise distribution determination unit 27 determines whether there is a noise component at a level that cannot be ignored by the
第１処理信号Ｙ（ｋ）は、低域雑音除去部３０のゲイン補償手段３１および遅延手段３８に入力される。 The first processed signal Y (k) is input to the gain compensation unit 31 and the delay unit 38 of the low frequency noise removing unit 30.
ゲイン補償手段３１は、後述の移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインＧａで第１処理信号Ｙ（ｋ）を増幅して、移相手段３２に入力する。 The gain compensation means 31 amplifies the first processing signal Y (k) with a fixed gain Ga set in advance in order to compensate for the gain drop of the low-frequency noise component to be removed that is predicted to be caused by the phase shift processing described later. And input to the phase shift means 32.
移相手段３２は、例えば図５のように、９０度の移相処理を行う直列接続された２つのヒルベルト変換器３２ａ、３２ｂとヒルベルト変換処理の遅延時間と等しい遅延を与える遅延手段３２ｃとにより構成されている。なお、ヒルベルト変換器３２ａ、３２ｂのタップ数は、周波数Ｆａより低い周波数成分に対しては、前記図３で示したように周波数が低くなる程振幅が減衰する特性となり、その帯域の信号の振幅を下げてしまうが、その振幅低下は前記したようにその雑音成分の周波数解析で得られた周波数に対応する固定のゲインを用いたゲイン補償手段３１によって補償されている。 For example, as shown in FIG. 5, the phase shift means 32 includes two Hilbert transformers 32a and 32b connected in series for performing a phase shift process of 90 degrees and a delay means 32c for giving a delay equal to the delay time of the Hilbert transform process. It is configured. It should be noted that the number of taps of the Hilbert transformers 32a and 32b is such that the frequency component lower than the frequency Fa has a characteristic that the amplitude decreases as the frequency decreases, as shown in FIG. However, the amplitude reduction is compensated by the gain compensation means 31 using a fixed gain corresponding to the frequency obtained by the frequency analysis of the noise component as described above.
移相手段３２は、増幅された信号Ｙ（ｋ）′＝Ｇａ・Ｙ（ｋ）に対して１８０度の位相差を有する第１移相信号Ｎｉと、その第１移相信号に対して９０度の位相差を有する第２移相信号Ｎｑとを生成して、雑音信号情報算出手段３３へ出力する。 The phase shift means 32 includes a first phase shift signal Ni having a phase difference of 180 degrees with respect to the amplified signal Y (k) ′ = Ga · Y (k), and 90 for the first phase shift signal. The second phase shift signal Nq having a phase difference of degree is generated and output to the noise signal information calculation means 33.
雑音信号情報算出手段３３は、第１移相信号Ｎｉと第２移相信号Ｎｑとに基づいて、信号Ｙ（ｋ）′に含まれる除去対象の低域雑音成分Ｎ１の瞬時振幅α、瞬時位相φおよび瞬時周波数ｆを、前記した式（１）〜（３）にしたがって算出し、その結果を雑音信号生成手段３５に出力する。 Based on the first phase shift signal Ni and the second phase shift signal Nq, the noise signal information calculation means 33 calculates the instantaneous amplitude α and instantaneous phase of the low-frequency noise component N1 to be removed included in the signal Y (k) ′. φ and instantaneous frequency f are calculated according to the above-described equations (1) to (3), and the results are output to the noise signal generating means 35.
一方、振幅誤差検出手段３４は、雑音信号情報算出手段３３によって算出された瞬時振幅と、第１処理信号に含まれる除去対象の低域雑音成分の瞬時振幅との誤差ｅを検出するし、その結果を雑音信号生成手段３５に出力する。
On the other hand, the amplitude
雑音信号生成手段３５は、雑音信号情報算出手段３３によって算出された瞬時振幅、瞬時位相、瞬時周波数および振幅誤差検出手段３４によって検出された瞬時振幅の誤差ｅに基づいて、第１処理信号に含まれる除去対象の低域雑音成分Ｎ１と同振幅、同周波数で且つ位相が反転した第１雑音信号ＮＬ１（ｋ）を生成する。
The noise signal generator 35 is included in the first processing signal based on the instantaneous amplitude, instantaneous phase, instantaneous frequency calculated by the noise signal information calculator 33 and the instantaneous amplitude error e detected by the
加算手段３６は、第１処理信号Ｙ（ｋ）と第１雑音信号ＮＬ１（ｋ）を加算して、第１処理信号Ｙ（ｋ）からレベル最大の低域雑音成分Ｎ１を除去した第２処理信号（Ｚ（ｋ））を求める。なお、実際には第１雑音信号ＮＬ１（ｋ）を生成するために必要な処理遅延分を遅延手段３８によって第１処理信号Ｙ（ｋ）に与えてから加算手段３６へ入力する。 The adding means 36 adds the first processed signal Y (k) and the first noise signal N L1 (k), and removes the low-frequency noise component N1 having the maximum level from the first processed signal Y (k). A processing signal (Z (k)) is obtained. In practice, the processing delay necessary for generating the first noise signal N L1 (k) is given to the first processing signal Y (k) by the delay means 38 and then input to the adding means 36.
このようにして得られた第２処理信号Ｚ（ｋ）に前記したようにさらに雑音Ｎ２が含まれて場合、第２処理信号Ｚ（ｋ）は、２段目目の低域雑音処理を行うためのゲイン補償手段４１に入力され、前記同様の低域雑音除去処理を受ける。 As described above, when the noise N2 is further included in the second processed signal Z (k) obtained in this way, the second processed signal Z (k) performs the second-stage low-frequency noise processing. And is subjected to the same low-frequency noise removal processing as described above.
即ち、ゲイン補償手段４１は、移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインＧｂで第２処理信号Ｚ（ｋ）を増幅して、移相手段４２に入力する。 That is, the gain compensation means 41 amplifies the second processing signal Z (k) with a fixed gain Gb set in advance to compensate for the gain reduction of the low-frequency noise component to be removed that is predicted to occur due to the phase shift processing. And input to the phase shift means 42.
移相手段４２は、前記した移相手段３２と同様に構成され、増幅された信号Ｚ（ｋ）′＝Ｇｂ・Ｚ（ｋ）に対して１８０度の位相差を有する第１移相信号Ｎｉと、その第１移相信号に対して９０度の位相差を有する第２移相信号Ｎｑとを生成して、雑音信号情報算出手段４３へ出力する。 The phase shift means 42 is configured in the same manner as the phase shift means 32 described above, and has a first phase shift signal Ni having a phase difference of 180 degrees with respect to the amplified signal Z (k) ′ = Gb · Z (k). And a second phase shift signal Nq having a phase difference of 90 degrees with respect to the first phase shift signal, and output to the noise signal information calculation means 43.
雑音信号情報算出手段４３は、第１移相信号Ｎｉと第２移相信号Ｎｑとに基づいて、信号Ｚ（ｋ）′に含まれる除去対象の低域雑音成分Ｎ２の瞬時振幅α、瞬時位相φおよび瞬時周波数ｆを、前記した式（１）〜（３）にしたがって算出し、その結果を雑音信号生成手段４５に出力する。 The noise signal information calculation means 43 is based on the first phase shift signal Ni and the second phase shift signal Nq, and the instantaneous amplitude α and instantaneous phase of the low-frequency noise component N2 to be removed included in the signal Z (k) ′. φ and instantaneous frequency f are calculated according to the above-described equations (1) to (3), and the results are output to the noise signal generating means 45.
一方、振幅誤差検出手段４４は、雑音信号情報算出手段４３によって算出された瞬時振幅と、第２処理信号Ｚ（ｋ）に含まれる除去対象の低域雑音成分の瞬時振幅との誤差ｅを検出するし、その結果を、雑音信号生成手段４５に出力する。 On the other hand, the amplitude error detector 44 detects an error e between the instantaneous amplitude calculated by the noise signal information calculator 43 and the instantaneous amplitude of the low-frequency noise component to be removed included in the second processed signal Z (k). Then, the result is output to the noise signal generation means 45.
雑音信号生成手段４５は、雑音信号情報算出手段４３によって算出された瞬時振幅、瞬時位相、瞬時周波数および振幅誤差検出手段４４によって検出された瞬時振幅の誤差ｅに基づいて、第２処理信号に含まれる除去対象の低域雑音成分Ｎ２と同振幅、同周波数で且つ位相が反転した第２雑音信号ＮＬ２（ｋ）を生成する。 The noise signal generation unit 45 is included in the second processing signal based on the instantaneous amplitude, the instantaneous phase, the instantaneous frequency, and the instantaneous amplitude error e detected by the amplitude error detection unit 44 calculated by the noise signal information calculation unit 43. The second noise signal N L2 (k) having the same amplitude, the same frequency and the inverted phase as the low-frequency noise component N2 to be removed is generated.
加算手段４６は、第２処理信号Ｚ（ｋ）と第２雑音信号ＮＬ２（ｋ）を加算して、第２処理信号Ｚ（ｋ）からレベル最大の低域雑音成分Ｎ２を除去した第３処理信号（Ｗ（ｋ））を求める。なお、実際には第２雑音信号ＮＬ２（ｋ）を生成するために必要な処理遅延分を遅延器４８によって第２処理信号Ｚ（ｋ）に与えてから加算手段４６へ入力する。
The adding means 46 adds the second processed signal Z (k) and the second noise signal N L2 (k), and removes the low-frequency noise component N2 having the maximum level from the second processed signal Z (k). A processing signal (W (k)) is obtained. In practice, the processing delay necessary for generating the second noise signal N L2 (k) is given to the second processing signal Z (k) by the
この第３処理信号Ｗ（ｋ）は、スイッチ２８を介して最終の処理結果として出力されることになる。 The third processing signal W (k) is output as the final processing result via the switch 28.
以上、レベルの大きい２つの雑音が存在する場合について説明したが、雑音分布判定手段２７によって一つの雑音信号が存在すると判定された場合には、第２処理信号Ｚ（ｋ）をスイッチ２８で最終の処理結果として選択出力すればよい。 As described above, the case where two noises having a large level are present has been described. However, when it is determined by the noise distribution determination unit 27 that one noise signal exists, the second processed signal Z (k) is finalized by the switch 28. It is sufficient to select and output as a result of the processing.
また、３つ以上の低域雑音成分に対応することも可能であり、その場合には、前記したゲイン補償手段、移相手段、雑音信号情報算出手段、振幅誤差検出手段、雑音信号生成手段、および加算手段を一組の雑音除去ブロックとし、その雑音除去ブロックを除去対象の低域雑音成分の数分直列接続して、各雑音除去ブロックでそれぞれの低域雑音成分を除去すればよい。現実的な低域雑音除去部としては、上記雑音除去ブロックを３〜４段構成にしておき、測定に有害な３〜４までの低域雑音成分を除去すればよい。
It is also possible to correspond to three or more low-frequency noise components, in which case the above-described gain compensation means, phase shift means, noise signal information calculation means, amplitude error detection means, noise signal generation means, The addition means may be a set of noise removal blocks, and the noise removal blocks may be connected in series by the number of low-frequency noise components to be removed, and the respective low-frequency noise components may be removed by each noise removal block. As a realistic low-frequency noise removal unit, the above-described noise removal block may be configured in 3 to 4 stages, and low-
次に、上記移相処理を用いた場合の実施形態のシミュレーションを図６〜図８に示す。ここで図６〜図８の縦軸の単位はグラム、横軸の単位は秒であり、シミュレーション条件は、以下の通りである。 Next, simulations of the embodiment in the case of using the phase shift process are shown in FIGS. Here, the unit of the vertical axis in FIGS. 6 to 8 is gram, the unit of the horizontal axis is second, and the simulation conditions are as follows.
自由度バネマスダンパ系に０．３秒の台形波（重量２０００グラム）を入力（この振幅を１００パーセントとする）
固有振動数 約４５Ｈｚ（振幅２０パーセント）
第２の周波数範囲の低域雑音 ４．５Ｈｚ（振幅５パーセント）
第１の周波数範囲の低域雑音 ２２Ｈｚ（ともに振幅１パーセント）
高域雑音 ５１Ｈｚ、６１Ｈｚ、６３Ｈｚ、７１Ｈｚ（ともに振幅３パーセント）
サンプリングレートＴ＝５ｍｓ
A trapezoidal wave (weight: 2000 grams) of 0.3 seconds is input to the spring mass damper system of freedom (this amplitude is assumed to be 100 percent).
Natural frequency about 45Hz (20% amplitude)
Low frequency noise in the second frequency range 4.5 Hz (5% amplitude)
Low frequency noise in the first frequency range 22 Hz (both 1% amplitude)
High frequency noise 51Hz, 61Hz, 63Hz, 71Hz (both amplitudes are 3%)
Sampling rate T = 5ms
図６は、センサの出力Ｘ（ｋ）とその信号に対して高域雑音除去処理を行った結果Ｙ（ｋ）を示している。 FIG. 6 shows the output Y (k) of the sensor output X (k) and the result of performing high-frequency noise removal processing on the signal.
ここで、高域雑音除去処理の下限周波数Ｆａを３０Ｈｚとし、それ以上で周波数平坦なタップ数５０のヒルベルト変換処理を用いている。 Here, the lower limit frequency Fa of the high-frequency noise removal processing is set to 30 Hz, and the Hilbert transform processing with 50 taps that is flat in frequency is used.
この信号Ｙ（ｋ）には、低周波の４．５Ｈｚ、２２Ｈｚの雑音が残っているが、２２Ｈｚの雑音成分に対して、前記した最初の低域雑音除去処理を行って得られたのが図７の第２処理信号Ｚ（ｋ）である。この第２処理信号Ｚ（ｋ）には、第１処理信号Ｙ（ｋ）の２２Ｈｚの雑音成分が除去されているが、４．５Ｈｚの雑音が残留していることが判る。 In this signal Y (k), low frequency 4.5 Hz and 22 Hz noise remains, but the result obtained by performing the first low-frequency noise removal processing on the 22 Hz noise component is obtained. This is the second processing signal Z (k) in FIG. It can be seen that the 22 Hz noise component of the first processed signal Y (k) is removed from the second processed signal Z (k), but 4.5 Hz noise remains.
この信号Ｚ（ｋ）に対して、４．５Ｈｚの雑音成分を除去対象として上記同様の処理を行って得られたのが図８の第３処理信号Ｗ（ｋ）であり、４．５Ｈｚの雑音が抑圧され、２０００グラムで平坦な時間領域が得られており、この時間帯を重量検出タイミングとすることで、高精度な測定が行えることがわかる。 The third processing signal W (k) in FIG. 8 is obtained by performing the same processing as described above on this signal Z (k) with a noise component of 4.5 Hz as an object of removal. Noise is suppressed and a flat time region of 2000 grams is obtained, and it can be seen that highly accurate measurement can be performed by using this time zone as the weight detection timing.
なお、上記の条件で、低域の雑音成分に対して±１０％の周波数変動を与えた場合でも、前記した振幅誤差の補正効果により、１／１０００以内の測定精度を確保できることを確認している。また、上記高域を含めて３段の雑音除去処理に要する時間は、約２７５ｍｓｅｃであり、数Ｈｚの低域雑音成分をきわめて高速に除去できることがわかる。 It should be noted that even when a frequency variation of ± 10% is given to the low frequency noise component under the above conditions, it is confirmed that the measurement accuracy within 1/1000 can be ensured by the effect of correcting the amplitude error described above. Yes. The time required for the three-stage noise removal processing including the high frequency is about 275 msec, and it can be seen that a low frequency noise component of several Hz can be removed very quickly.
１１……センサ、２０……信号処理装置、２１……Ａ／Ｄ変換器、２２……高域雑音除去部、２３……移相手段、２４……遅延手段、２５……加算手段、２６……スペクトラム解析手段、２７……雑音分布判定手段、３０……低域雑音除去部、３１、４１……ゲイン補償手段、３２、４２……移相手段、３２ａ、３２ｂ、４２ａ、４２ｂ……ヒルベルト変換器、３３、４２……雑音信号情報算出手段、３４、４４……振幅誤差検出手段、３５、４５……雑音信号生成手段、３６、４６……加算手段、３８、４８……遅延手段 DESCRIPTION OF SYMBOLS 11 ... Sensor, 20 ... Signal processing apparatus, 21 ... A / D converter, 22 ... High frequency noise removal part, 23 ... Phase shift means, 24 ... Delay means, 25 ... Adder means, 26 …… Spectrum analysis means, 27 …… Noise distribution determination means, 30 ...... Low-frequency noise removing unit, 31, 41 …… Gain compensation means, 32, 42 …… Phase shift means, 32 a, 32 b, 42 a, 42 b Hilbert transformer, 33, 42... Noise signal information calculation means, 34, 44... Amplitude error detection means, 35, 45... Noise signal generation means, 36, 46.
Claims (4)
前記低域雑音成分を除去する処理は、
移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインで前記第１処理信号を増幅する段階（Ｓ５）と、
前記増幅された信号に対して、ヒルベルト変換処理を２段行い、前記増幅された信号に対して１８０度の位相差を有する第１移相信号（Ｎｉ）と、該第１移相信号に対して９０度の位相差を有する第２移相信号（Ｎｑ）とを生成する段階（Ｓ６）と、
前記第１移相信号と第２移相信号とに基づいて、前記増幅された信号に含まれる前記除去対象の低域雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出する段階（Ｓ７）と、
前記算出された瞬時振幅と、前記第１処理信号に含まれる前記除去対象の低域雑音成分の瞬時振幅との誤差を検出する段階（Ｓ８）と、
前記算出した瞬時振幅、瞬時位相、瞬時周波数および前記瞬時振幅の誤差に基づいて前記第１処理信号に含まれる前記除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した第１雑音信号を生成する段階（Ｓ９）と、
前記第１処理信号と前記第１雑音信号とを加算して、前記第１処理信号から前記除去対象の低域雑音成分を除去した第２処理信号（Ｚ（ｋ））を求める段階（Ｓ１０）とを含むことを特徴とする信号処理方法。 A time-series signal obtained by A / D conversion processing for the output signal of the sensor is received as an input signal, a high-frequency noise component higher than a predetermined frequency (Fa) included in the input signal is removed, and the high-frequency noise component A signal processing method for removing a low-frequency noise component having a frequency equal to or lower than the predetermined frequency from the first processed signal (Y (k)) from which a DC component of the input signal is obtained,
The process of removing the low-frequency noise component is:
Amplifying the first processing signal with a fixed gain set in advance to compensate for a gain drop of a low-frequency noise component to be removed that is predicted to be generated by the phase shift processing (S5);
The Hilbert transform process is performed on the amplified signal in two stages, and the first phase-shifted signal (Ni) having a phase difference of 180 degrees with respect to the amplified signal and the first phase-shifted signal are processed. Generating a second phase shift signal (Nq) having a phase difference of 90 degrees (S6),
Calculating an instantaneous amplitude, an instantaneous phase, and an instantaneous frequency of the low-frequency noise component to be removed included in the amplified signal based on the first phase-shifted signal and the second phase-shifted signal (S7); ,
Detecting an error between the calculated instantaneous amplitude and the instantaneous amplitude of the low-frequency noise component to be removed included in the first processing signal (S8);
Based on the calculated instantaneous amplitude, instantaneous phase, instantaneous frequency, and error in the instantaneous amplitude, the first low-frequency noise component included in the first processing signal has the same amplitude, the same frequency, and the phase is inverted. Generating a noise signal (S9);
Adding the first processed signal and the first noise signal to obtain a second processed signal (Z (k)) obtained by removing the low-frequency noise component to be removed from the first processed signal (S10); And a signal processing method.
前記固定ゲインで処理信号を増幅する処理から、処理信号と生成した雑音信号とを加算して低域雑音成分を除去するまでの一連の処理（Ｓ５〜Ｓ１０）を、前記第２処理信号に含まれる各雑音成分に対して繰り返し行う（Ｓ１１、１２）ことを特徴とする請求項１記載の信号処理方法。 The process of removing the low-frequency noise component is:
The second process signal includes a series of processes (S5 to S10) from the process of amplifying the process signal with the fixed gain to adding the process signal and the generated noise signal to remove the low-frequency noise component. The signal processing method according to claim 1, wherein the signal processing is repeatedly performed on each noise component (S 11, 12).
前記Ａ／Ｄ変換器の出力信号を受け、所定周波数（Ｆａ）より高い高域雑音成分を除去し、該高域雑音成分を除去して得られた第１処理信号（Ｙ（ｋ））を出力する高域雑音除去部（２２）と、
前記第１処理信号を受け、前記所定周波数以下の低域雑音成分を除去する低域雑音除去部（３０）とを有する信号処理装置であって、
前記低域雑音除去部が、
移相処理により生じると予測される除去対象の低域雑音成分の利得低下を補償するために予め設定された固定ゲインで前記第１処理信号を増幅するゲイン補償手段（３１）と、
前記ゲイン補償手段によって増幅された信号に対して、ヒルベルト変換処理を２段行い、前記増幅された信号に対して１８０度の位相差を有する第１移相信号（Ｎｉ）と、該第１移相信号に対して９０度の位相差を有する第２移相信号（Ｎｑ）とを生成する移相手段（３２）と、
前記移相手段によって得られた第１移相信号と第２移相信号とに基づいて、前記ゲイン補償手段によって増幅された信号に含まれる前記除去対象の低域雑音成分の瞬時振幅、瞬時位相および瞬時周波数を算出する雑音信号情報算出手段（３３）と、
前記雑音信号情報算出手段によって算出された瞬時振幅と、前記第１処理信号に含まれる前記除去対象の低域雑音成分の瞬時振幅との誤差を検出する振幅誤差検出手段（３４）と、
前記雑音信号情報算出手段によって算出された瞬時振幅、瞬時位相、瞬時周波数および振幅誤差検出手段によって検出された前記瞬時振幅の誤差に基づいて、前記第１処理信号に含まれる前記除去対象の低域雑音成分と同振幅、同周波数で且つ位相が反転した第１雑音信号を生成する雑音信号生成手段（３５）と、
前記第１処理信号と前記第１雑音信号とを加算して、前記第１処理信号から前記除去対象の低域雑音成分を除去した第２処理信号（Ｚ（ｋ））を求める加算手段（３６）とを備えていることを特徴とする信号処理装置。 An A / D converter (21) that performs A / D conversion processing on the output signal of the sensor (11);
The output signal of the A / D converter is received, a high frequency noise component higher than a predetermined frequency (Fa) is removed, and a first processing signal (Y (k)) obtained by removing the high frequency noise component is obtained. An output high-frequency noise removal unit (22);
A signal processing device having a low-frequency noise removing unit (30) that receives the first processing signal and removes a low-frequency noise component of the predetermined frequency or less;
The low-frequency noise removing unit is
Gain compensation means (31) for amplifying the first processing signal with a fixed gain set in advance in order to compensate for a gain drop of a low-frequency noise component to be removed that is predicted to be generated by the phase shift processing;
The signal amplified by the gain compensation means is subjected to two stages of Hilbert transform processing, and the first phase-shifted signal (Ni) having a phase difference of 180 degrees with respect to the amplified signal and the first phase-shifted signal. Phase shifting means (32) for generating a second phase shifting signal (Nq) having a phase difference of 90 degrees with respect to the phase signal;
Based on the first phase shift signal and the second phase shift signal obtained by the phase shift means, the instantaneous amplitude and instantaneous phase of the low frequency noise component to be removed included in the signal amplified by the gain compensation means And noise signal information calculating means (33) for calculating an instantaneous frequency;
Amplitude error detection means (34) for detecting an error between the instantaneous amplitude calculated by the noise signal information calculation means and the instantaneous amplitude of the low-frequency noise component to be removed included in the first processing signal;
Based on the instantaneous amplitude, instantaneous phase, instantaneous frequency, and error of the instantaneous amplitude detected by the amplitude error detecting means calculated by the noise signal information calculating means, the low frequency range to be removed included in the first processing signal A noise signal generating means (35) for generating a first noise signal having the same amplitude and the same frequency as the noise component and having a phase inverted;
Adding means (36) for adding the first processed signal and the first noise signal to obtain a second processed signal (Z (k)) obtained by removing the low-frequency noise component to be removed from the first processed signal. And a signal processing device.
前記ゲイン補償手段、前記移相手段、前記雑音信号情報算出手段、前記振幅誤差検出手段、前記雑音信号生成手段および前記加算手段を一組の雑音除去ブロックとし、該雑音除去ブロックを除去対象の低域雑音成分の数分直列接続して構成され、各雑音除去ブロックでそれぞれ除去対象の低域雑音成分を除去することを特徴とする請求項３記載の信号処理装置。 The low-frequency noise removing unit is
The gain compensation unit, the phase shift unit, the noise signal information calculation unit, the amplitude error detection unit, the noise signal generation unit, and the addition unit constitute a set of noise removal blocks, and the noise removal block is a low target to be removed. 4. The signal processing apparatus according to claim 3, wherein the signal processing apparatus is configured to be connected in series by the number of band noise components, and each of the noise removal blocks removes a low band noise component to be removed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2008129214A JP5155733B2 (en) | 2008-05-16 | 2008-05-16 | Signal processing method and signal processing apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2008129214A JP5155733B2 (en) | 2008-05-16 | 2008-05-16 | Signal processing method and signal processing apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JP2009276268A JP2009276268A (en) | 2009-11-26 |
| JP5155733B2 true JP5155733B2 (en) | 2013-03-06 |
Family
ID=41441829
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP2008129214A Expired - Fee Related JP5155733B2 (en) | 2008-05-16 | 2008-05-16 | Signal processing method and signal processing apparatus |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP5155733B2 (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5524277B2 (en) * | 2012-05-29 | 2014-06-18 | アンリツ株式会社 | Signal processing apparatus and signal processing method |
| CN110045748B (en) * | 2019-04-03 | 2022-05-03 | 深圳高速工程检测有限公司 | Aircraft control method, aircraft control device, computer equipment and storage medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4488496B2 (en) * | 2004-03-24 | 2010-06-23 | アンリツ株式会社 | Signal processing method and signal processing apparatus |
| JP4324594B2 (en) * | 2006-02-03 | 2009-09-02 | アンリツ株式会社 | Digital modulation signal demodulator |
-
2008
- 2008-05-16 JP JP2008129214A patent/JP5155733B2/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| JP2009276268A (en) | 2009-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP5155733B2 (en) | Signal processing method and signal processing apparatus | |
| US11335356B2 (en) | Digital audio processing device, digital audio processing method, and digital audio processing program | |
| JP4729553B2 (en) | Signal processing method and signal processing apparatus | |
| JP4819742B2 (en) | Signal processing method and signal processing apparatus | |
| JP4488496B2 (en) | Signal processing method and signal processing apparatus | |
| US20200073007A1 (en) | Metal detection apparatus | |
| CN106258000B (en) | Acceleration detector and active noise controller | |
| JP4788656B2 (en) | Power test system | |
| JP6515733B2 (en) | Digital protection relay | |
| WO2016002217A1 (en) | Current detection device and current detection method | |
| JP4638981B2 (en) | Signal processing device | |
| JP2019117999A (en) | Noise removal filter device and noise removal method | |
| JP4643350B2 (en) | Strain sensor signal processing device | |
| JP6265870B2 (en) | Digital protection relay device | |
| JP4814136B2 (en) | Signal processing method and signal processing apparatus | |
| US20180085067A1 (en) | Signal source identifying device for biological information, and signal source identifying method for biological information | |
| JP2019020756A (en) | Sensor signal processor | |
| US20190145818A1 (en) | Processing apparatus, processing system, physical quantity measurement apparatus, and measurement method | |
| CN111044758B (en) | Acceleration sensor output value correction method and acceleration sensor | |
| JP2016034173A (en) | Electric quantity meter | |
| JP6352117B2 (en) | Filter device and filtering method | |
| JP6373765B2 (en) | Filter device and filtering method | |
| US10914794B2 (en) | Measuring apparatus, method, and storage medium | |
| US20150168189A1 (en) | Method for determination of the time of flight of the signals in the signal paths of a coriolis flow meter | |
| JP2008076300A (en) | Measuring unit and material testing machine |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A621 | Written request for application examination |
Free format text: JAPANESE INTERMEDIATE CODE: A621 Effective date: 20110209 |
|
| TRDD | Decision of grant or rejection written | ||
| A01 | Written decision to grant a patent or to grant a registration (utility model) |
Free format text: JAPANESE INTERMEDIATE CODE: A01 Effective date: 20121127 |
|
| A61 | First payment of annual fees (during grant procedure) |
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20121207 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20151214 Year of fee payment: 3 |
|
| R150 | Certificate of patent or registration of utility model |
Free format text: JAPANESE INTERMEDIATE CODE: R150 |
|
| R250 | Receipt of annual fees |
Free format text: JAPANESE INTERMEDIATE CODE: R250 |
|
| LAPS | Cancellation because of no payment of annual fees |