JP7010806B2 - Vibration displacement estimation program, device and method using filter phase lead - Google Patents

Description

The present invention relates to a technique for obtaining vibration displacement from measurement data related to vibration, and particularly to a technique for deriving vibration displacement using a digital filter.

At present, it is quite common for users who have boarded mobile objects such as automobiles, trains, and ships to enjoy various services using terminals such as smartphones and tablet computers.

At this time, in many cases, the user obtains desired information by looking at the image displayed on the terminal screen, but the image is shaken by the vibration accompanying the movement of the moving body, and it is often very difficult to see. .. In addition, it is also a problem that the eyes get tired and the symptoms of motion sickness appear by continuing to see the shaking image in this way.

As a measure for improving the visibility of an image displayed in such a shaking environment, for example, in Patent Document 1, a touch panel 107 for displaying an image and an acceleration sensor 303 for detecting the movement of a device are detected. The touch panel 107 includes a CPU 301 that calculates the movement amount of the image to be displayed on the touch panel 107 based on the movement of the device, and the touch panel 107 is characterized in that the display position of the image is moved based on the movement amount L calculated from the CPU 301. The mobile phone 100 is disclosed.

Here, the CPU 301 detects the acceleration of the mobile phone 100 in motion from the output value of the acceleration sensor 303, and calculates the movement amount L of the mobile phone 100 from the detected acceleration. This movement amount L is obtained, for example, by the CPU 301 integrating the acceleration twice in time. Next, the CPU 301 moves the image 601 displayed on the screen of the touch panel 107 before the movement of the mobile phone on the screen based on the calculated movement amount L.

JP-A-2015-141700

As described above, conventionally, the mobile terminal (mobile phone 100 in Patent Document 1) cannot directly measure the vibration displacement (movement amount L), and therefore, for example, with respect to the acceleration data measured by using an acceleration sensor. The vibration displacement is derived by applying the second-order integral (two integrals) filter processing.

Next, the mobile terminal moves the display position of the image in the screen based on the derived vibration displacement, and adjusts the image display so as not to be affected by the vibration.

However, in the conventional technique as described above, it takes time for the vibration displacement derivation process, the display image adjustment process, and the like, and the delay time (delay time) which is the delay from the acceleration measurement time point at the image display time point where the display position is adjusted ( The existence of latency) becomes a big problem. That is, in order to reflect the vibration received by the mobile terminal on the image display in real time, it is necessary to predict the vibration displacement in the future for this delay time.

Here, various filters are usually used to derive the vibration displacement, but in general, the frequency characteristics of the phase in the output signal differ greatly for each filter, and as a result, the vibration generated in a wide frequency band is predicted. Due to the difference in the phase of the displacement to be performed, it is very difficult to determine the predicted amount of vibration displacement suitable for adjusting the image position.

Therefore, it is an object of the present invention to provide a vibration displacement estimation program, an apparatus and a method capable of estimating a vibration displacement without depending on a displacement prediction in a future time.

According to the present invention, a computer mounted on a device capable of estimating vibration displacement from a measured amount related to vibration acceleration or velocity and suppressing the influence of the vibration on the processing result output of the processing performed by itself based on the vibration displacement. Is a vibration displacement estimation program that makes
An integral approximation filter means that applies an integral approximation filter process equivalent to an integral whose phase advances more than the integral filter process to the measured quantity.
Compensation amount determining means for calculating the phase advance amount in the measured quantity subjected to the integral approximation filter processing, and
The computer is made to function as a vibration displacement determining means for determining the vibration displacement based on the measured quantity subjected to the integral approximation filter processing at the previous time point corresponding to the phase lead .
The integral approximation filter means is the phase lead calculated by the compensation amount determining means, and the value of the phase lead calculated for the frequency related to the vibration affecting the processing result output is the processing of the present apparatus. The parameters in the integral approximation filter processing are adjusted so that the value corresponds to the delay time until the processing result output in the above or the value is within a predetermined range from this equivalent value, and the integral approximation filter processing is performed.
A vibration displacement estimation program is provided.

Further, in the vibration displacement estimation program according to the present invention, the compensation amount determining means also calculates the gain fluctuation amount in the measured quantity subjected to the integral approximation filter processing.
It is also preferable that the vibration displacement determining means determines the vibration displacement by correcting the gain fluctuation amount with respect to the measured quantity subjected to the integral approximation filter processing at the previous time point corresponding to the phase advance amount.

Further, as an embodiment of the vibration displacement estimation program according to the present invention, the vibration displacement estimation program is a low frequency removal filter that removes low frequency components with respect to the measured amount before and / or after the integral approximation filter processing. Further functioning the computer as a low frequency removal filter means to apply processing,
It is also preferable that the compensation amount determining means calculates the phase lead amount in the measured quantity subjected to the low frequency elimination filter processing and the integral approximation filtering processing.

Further, in the vibration displacement estimation program according to the present invention, the above device is a device that estimates the vibration displacement from the measured quantity related to the vibration acceleration.
It is also preferable that the integral approximation filter means performs an integral approximation filter process corresponding to the second-order integral in which the phase advances to the measured quantity related to the acceleration.

Further, as another embodiment of the vibration displacement estimation program according to the present invention, the vibration displacement estimation program further functions the computer as a frequency determining means for determining an instantaneous or quasi-instantaneous frequency in the measured quantity.
It is also preferable that the compensation amount determining means calculates the phase lead amount at the instantaneous or quasi-instantaneous frequency in the measured quantity subjected to the integral approximation filtering process.

Further, as a further embodiment of the vibration displacement estimation program according to the present invention, the vibration displacement determining means is a vibration displacement with respect to the measured amount subjected to the integral approximation filter processing at the previous time point corresponding to the phase advance portion. It is also preferable to determine the vibration displacement based on the amount multiplied by a predetermined function value indicating a monotonous increase with respect to the frequency in order to reduce the vibration displacement in the low frequency region where estimation is unnecessary to zero or a minute amount.

Further, as a specific application example of the above-mentioned other embodiment, the above-mentioned device is based on the determined vibration displacement, and the display position of the image in the image display as the processing result output by the image processing of this device. It is also preferable to correct.

According to the present invention, a vibration displacement estimation device capable of estimating the vibration displacement from the measured amount related to the acceleration or velocity of the vibration and suppressing the influence of the vibration on the processing result output of the processing performed by itself based on the vibration displacement. And,
An integral approximation filter means that applies an integral approximation filter process equivalent to an integral whose phase advances more than the integral filter process to the measured quantity.
Compensation amount determining means for calculating the phase advance amount in the measured quantity subjected to the integral approximation filter processing, and
It has a vibration displacement determining means for determining a vibration displacement based on a measured quantity subjected to the integral approximation filter processing at a previous time point corresponding to the phase advance portion.
The integral approximation filter means is the phase lead calculated by the compensation amount determining means, and the value of the phase lead calculated for the frequency related to the vibration affecting the processing result output is the processing of the present apparatus. The parameters in the integral approximation filter processing are adjusted so that the value corresponds to the delay time until the processing result output in the above or the value is within a predetermined range from this equivalent value, and the integral approximation filter processing is performed.
A vibration displacement estimation device characterized by this is provided.

According to the present invention, it is a vibration displacement estimation program that functions a computer mounted on a device capable of estimating vibration displacement from a measured quantity related to vibration acceleration or velocity.
An integral approximation filter means that applies an integral approximation filter process equivalent to an integral whose phase advances more than the integral filter process to the measured quantity.
Frequency determining means for determining the instantaneous or quasi-instantaneous frequency in the measured quantity,
Compensation amount determining means for calculating the phase lead at the instantaneous or quasi-instantaneous frequency of the measured quantity subjected to the integral approximation filter processing, and
The vibration displacement at the time of estimating the vibration displacement is determined based on the measured amount subjected to the integral approximation filter processing at the time point corresponding to the calculated phase advance amount when viewed from the time of estimation of the vibration displacement. Vibration displacement determination means
A vibration displacement estimation program is provided to make the computer work.
In the vibration displacement estimation program according to the present invention, the compensation amount determining means also calculates the gain fluctuation amount at the instantaneous or quasi-instantaneous frequency in the measured quantity subjected to the integral approximation filter processing.
The vibration displacement determining means corrects the gain fluctuation amount for the measured amount subjected to the integral approximation filter processing at the time point corresponding to the phase advance amount when viewed from the vibration displacement estimation time point. Therefore, it is also preferable to determine the vibration displacement at the time of estimating the vibration displacement .

According to the present invention, the vibration displacement can be estimated without relying on the displacement prediction in the future time.

It is a functional block diagram which shows the functional structure in one Embodiment of the vibration displacement estimation apparatus by this invention. It is a graph for demonstrating the adjustment of the filtering parameter in the pre-filter unit, the integral approximation filter unit and the post-filter unit which concerns on this invention. It is a schematic diagram which shows the specific example which corrected the display image position in a smartphone with the estimated vibration displacement which concerns on this invention. It is a flowchart which shows one Embodiment of the vibration displacement estimation method by this invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[Vibration displacement estimation device]
FIG. 1 is a functional block diagram showing a functional configuration according to an embodiment of the vibration displacement estimation device according to the present invention.

According to the embodiment shown in FIG. 1, the smartphone 1 which is the vibration displacement estimation device according to the present invention is an image display device having an image display unit (touch panel display 103) capable of displaying an image. Further, the user holding the smartphone 1 can view and view the image (including the moving image) displayed on the touch panel display 103 of the smartphone 1 inside the car 2 as a moving body or in the state of being in the car 2. I'm watching.

Here, the automobile 2 is running and gives vibrations of various frequencies to the smartphone 1, and as a result, the user has to keep looking at the shaken display image as it is, which is very difficult to see. In particular, vibration at a high frequency of several hertz (Hz) or more causes the displayed image to blur or have an afterimage every moment, and the visibility is significantly reduced.

On the other hand, the smartphone 1 of the present embodiment estimates the vibration displacement of its own vibration based on the acceleration measurement data momentarily acquired by the equipped acceleration sensor 101 or the like, and is further based on the estimated vibration displacement. Therefore, the display position of the image displayed on the touch panel display 103 is changed to improve the visibility of the displayed image.

Here, the smartphone 1 realizes more appropriate correction / adjustment of the image display position by estimating the vibration variation in consideration of the latency (delay time) of the processing in the device. That is, it is a major feature that the vibration displacement estimation process is performed in more real time.

By the way, the vibration displacement estimation device (and image display device) in the present embodiment is not limited to a smartphone, for example, a handheld terminal such as a tablet type / notebook computer, a mobile terminal, or a wristwatch type. It may be a wearable terminal such as a terminal or an HMD (Head Mounted Display). It is also possible to use a personal computer (PC) brought into the moving body. Further, it may be a monitor-dedicated device such as a seat back monitor device installed in the moving body, or a car navigation device.

Similarly, in FIG. 1, the smartphone 1 is specifically characterized as a vibration displacement estimation device.
(A) An integral approximation filter unit 112 that performs "integral approximation filter processing" equivalent to an integral whose phase advances more than the integral filter processing on a measured quantity related to vibration acceleration.
(B) Compensation amount determination unit 115 for calculating "phase lead" in a measured quantity subjected to "integral approximation filter processing", and
(C) It has a vibration displacement determining unit 116 that determines a vibration displacement based on a measured quantity subjected to "integral approximation filtering processing" at a previous time point corresponding to a "phase advance component".

Here, the "integral approximation filter processing" of the above configuration (A) is a digital filter processing equivalent to an integral of a type in which the phase of the output signal advances more than the simple integral filter processing. The specific form will be explained in detail later. When processing a measure related to vibration acceleration as in the above configuration (A), the "integral approximation filter processing" is a second-order integral whose phase advances more than a simple second-order integral (two-time integral) filter process. It will be a considerable filtering process.

On the other hand, when the smartphone 1 can acquire the measured quantity related to the speed in vibration, the "integral approximation filter processing" for the measured quantity related to this speed is more than the simple first-order integral (one-time integral) filter processing. It can be a filter process equivalent to the first-order integral in which the phase advances. For example, when a measure related to the vibration angular velocity of rotational vibration with the Z axis as the rotation axis in the display screen coordinate system described later is acquired using the gyro sensor 102, the “integral approximation filter processing” of the above configuration (A). Is a process equivalent to the first-order integration, and it is also possible to estimate the vibration angle.

As described above, the smartphone 1 can determine the vibration displacement based on the measured quantity subjected to the "integral approximation filter processing" at the previous time point corresponding to the calculated "phase advance". Specifically, the filter processing output is selected by going back by the amount that the filter processing output for the measured quantity advances on the time axis (related to the phase), and it is based on the selected (more real-time) filtering output. This makes it possible to estimate the vibration displacement in real time.

That is, according to the smartphone 1, even if there is a latency in the device processing, the vibration displacement of the smartphone 1 is performed by using only the output signals from the acceleration sensor 101 and the gyro sensor 102 without depending on the displacement prediction in the future time. It can be estimated.

By the way, the moving body on which the smartphone 1 (the user who owns the smartphone) gets in is not limited to the automobile 2. For example, large vehicles such as buses and trucks, railroad vehicles including linear cars, monorails, ships, elevators, aircraft including airships, mega floats, submersibles, spacecraft, space stations, etc., inside or inside them. If the smartphone 1 can be used on board, various things are applicable.

[Device function configuration, vibration displacement estimation program]
Hereinafter, the functional configuration of the smartphone 1 in the present embodiment will be described in detail with reference to the functional block diagram also shown in FIG. According to the functional block diagram, the smartphone 1 has an acceleration sensor 101, a gyro sensor 102, a touch panel display 103, an analysis result storage unit 104, a communication interface unit 105, and a processor memory. Incidentally, FIG. 1 shows only the functional configuration related to the vibration displacement estimation function in the smartphone 1.

Here, this processor memory stores one embodiment of the vibration displacement estimation program according to the present invention, and also has a computer function, and vibration displacement estimation is performed by executing this vibration displacement estimation program. Carry out the process. Of course, the vibration displacement estimation device according to the present invention is not limited to the smartphone, and various forms of the device including the computer equipped with the vibration displacement estimation program according to the present invention are the vibration displacement estimation device. Can be adopted as.

Further, the processor memory includes a pre-filter unit 111 including the parameter adjustment unit 111p, an integral approximation filter unit 112 including the parameter adjustment unit 112p, a post-filter unit 113 including the parameter adjustment unit 113p, and a quasi-instantaneous frequency determination unit 114. It also has a compensation amount determination unit 115, a vibration displacement determination unit 116 including a sigmoid function adjustment unit 116s, an input / output control unit 117 including an image display control unit 117g, a sensor integration unit 121, and an application unit 122.

It should be noted that these functional components can be regarded as the functions of the vibration displacement estimation program stored in the processor memory. Further, the process flow shown by connecting the functional components of the smartphone 1 in FIG. 1 with arrows is also understood as an embodiment of the vibration displacement estimation method according to the present invention.

Similarly, in the functional block diagram of FIG. 1, the acceleration sensor 101 is a measuring means that measures the acceleration received by the smartphone 1 itself and outputs a measured amount related to the acceleration, and can be, for example, a known 3-axis acceleration sensor. .. Further, the gyro sensor 102 is a measuring means for measuring an angular velocity (or angular acceleration) as a change in the orientation of the device of the smartphone 1 and outputting a measured amount related to the angular velocity, and is, for example, a known 3-axis gyroscope. May be good.

Here, in the present embodiment, as the display screen coordinate system of the touch panel display 103, an XYZ orthogonal coordinate system having a horizontal X axis, a vertical Y axis, and a depth Z axis on the display screen is set. .. On the other hand, it is also preferable to set the measurement axes of the acceleration sensor 101 and the gyro sensor 102 as the three axes of these XYZ so that the vibration displacement in the X-axis and Y-axis directions can be finally calculated.

Further, the sensor integration unit 121 inputs the measured amount related to the acceleration and the measured amount related to the angular velocity from the acceleration sensor 101 and the gyro sensor 102, respectively, and uses a known sensor fusion process to input the measured amount related to the angular velocity and the Z-axis. Based on the measured quantity related to the acceleration in the direction, the measured quantity related to the acceleration in the X-axis direction and the measured quantity related to the acceleration in the Y-axis direction are corrected, and the measured quantity related to the acceleration in the X and Y-axis directions with higher accuracy is corrected. Is output to the pre-filter unit 111. At that time, the accuracy of the measured quantity related to the angular velocity may be improved by the same known sensor fusion treatment.

Here, the measured quantity output from the sensor integration unit 121 has an ultra-low frequency different from the vibration to be corrected for the display position, such as gravitational acceleration, acceleration during acceleration / deceleration of the automobile 2, and centrifugal acceleration during curve progress. Acceleration component is mixed.

Therefore, the pre-filter unit 111 performs a low-frequency removal filter process for removing the ultra-low frequency component as a pre-process for the integral approximation filter process described later with respect to the measured quantity input from the sensor integration unit 121. Specifically, in the present embodiment, these ultra-low frequency components are removed by DC cut filter processing. This DC cut filter processing can be the processing shown in the following equation.
(1) Recurrence formula: y [i] ＝ x [i] －x [i-1] ＋ r ₁ ＊ y [i-1]
(2) Transfer function: h (z) = y (z) / x (z) = (1-z- ¹ ) / (1-r ₁ * z ^-1 )
y (z) ＝ x (z) －z ^－1 ＊ x (z) ＋ r ₁ ＊ z ^－1 ＊ y (z)

Here, in the above equation (1), x [i] is the i-th filter input signal sample, and x [i-1] is the filter input signal sample at the time point one before the i-th. y [i] is the i-th filter output signal sample, and y [i-1] is the filter output signal sample at the time point one before the i-th filter. In addition, r ₁ is a filtering parameter that takes a value from 0 to 1. Further, in the above equation (2), x (z) and y (z) are Z-transform results of the input signal sample x [i] and the output signal sample y [i], respectively (x (z) = Z ( x [i]), y (z) = Z (y [i])). Furthermore, "*" in the following formulas including the above formula represents multiplication.

This DC cut filter processing is a kind of high-pass filter processing, and is a processing for advancing the phase of the output signal. The parameter adjusting unit 111p of the pre-filter unit 111 can adjust the value of the above parameter r ₁ to control the phase advance in the output signal of the DC cut filter processing. Further, it is also preferable that the pre-filter unit 111 stores, for example, the output signal (measured quantity processed by the DC cut filter) for the past 1 second as a time-series signal in order to determine the quasi-instantaneous frequency described later.

Similarly, in the functional block diagram of FIG. 1, the integral approximation filter unit 112 has a phase higher than that of the integral filter process with respect to the measured quantity (output signal of the prefilter unit 111) that has been DC-cut filtered and input from the prefilter unit 111. Perform "integral approximation filter processing" equivalent to the forward integral. In the present embodiment, since the output signal of the pre-filter unit 111 is a measured quantity related to acceleration, the "integral approximation filter processing" is a second-order integral (two integrals) in which the phase advances in order to convert to a displacement equivalent amount. ) It is a considerable filtering process.

Specifically, in the integral approximation filter processing of the present embodiment, the filter processing as shown in the following equation can be performed twice.
(3) Recurrence formula: y [i] ＝ r ₂ ＊ y [i-1] ＋ x [i]
(4) Transfer function: h (z) = y (z) / x (z) = 1 / (1-r ₂ * z ^-1 )
y (z) ＝ x (z) ＋ r ₂ ＊ z ^－1 ＊ y (z)

Here, r ₂ in the above equations (3) and (4) is a filtering parameter that takes a value from 0 to 1. In this integral approximation filter processing, by introducing this parameter r ₂ , the phase in the output signal can be advanced more than the simple integral filter processing (y [i] ＝ y [i-1] ＋ x [i])). It is possible. Here, the parameter adjusting unit 112p of the integral approximation filter unit 112 can adjust the value of this parameter r ₂ to control the phase advance in the output signal of the present integral approximation filter processing.

Furthermore, by introducing this parameter r ₂ , it is possible to maintain the filtering process in a stable direction without diverging. In fact, if a simple integral filter process (y [i] = y [i-1] + x [i]) is simply performed twice to perform second-order integration, the displacement as an output signal may diverge. However, according to the main integral approximation filter processing equivalent to the second-order integral, it is possible to reduce such divergence by adjusting r ₂ .

Further, the "integral approximation filter processing" according to the present invention is not limited to the filter processing described above, and can be various as long as it is a filter process equivalent to a (second-order) integral whose phase advances more than a simple integral filter process. Can be adopted as "integral approximation filtering".

For example, a process of performing the filter process twice as shown in the following equation may be adopted.
(5) Recurrence formula: y [i] ＝ r ₂ ＊ y [i-1] ＋ (x [i] －p ＊ x [i-1])
(6) Transfer function: h (z) = y (z) / x (z) = (1-p * z ^-1 ) / (1-r ₂ * z ^-1 )
y (z) ＝ (x (z) －p ＊ z ^－1 ＊ (z)) ＋ r ₂ ＊ z ^－1 ＊ y (z)

Here, both r ₂ and p in the above equations (5) and (6) are filtering parameters that take a value from 0 to 1. By the way, the filter processing represented by the above equations (5) and (6) is a high-pass filter element (-p * x [i-1]) with respect to the filter processing represented by the above equations (3) and (4). (If P = 0, it matches the filter processing of the above equations (3) and (4)).

Further, when the integral approximation filter processing in which the above equations (5) and (6) are performed twice is adopted, the parameter adjusting unit 112p adjusts the values of these parameters r ₂ and p, and the integral approximation filter processing is performed. It controls the phase advance in the output signal.

Here, in the output signal (measured quantity after the second-order integral processing corresponding to the displacement) output from the integral approximation filter unit 112, the DC component may be accumulated and remain by the two integral equivalent processes. Therefore, next, the post-filter unit 113 performs a low-frequency elimination filter process for removing the ultra-low frequency component on the measured quantity (output signal of the integral approximation filter unit 112) that has been subjected to the integral approximation filter processing.

Specifically, in the present embodiment, the post filter unit 113 removes DC components by DC cut filter processing. This DC cut filter process can be the process shown in the following equation, which has the same form as the above equations (1) and (2).
(7) Recurrence formula: y [i] ＝ x [i] －x [i-1] ＋ r ₃ ＊ y [i-1]
(8) Transfer function: h (z) = y (z) / x (z) = (1-z- ¹ ) / (1 _- r3 * z ^-1 )
y (z) ＝ x (z) －z ^－1 ＊ x (z) ＋ r ₃ ＊ z ^－1 ＊ y (z)

In the above equations (7) and (8), r ₃ is also a filtering parameter that takes a value from 0 to 1 like r ₁ . Further, the parameter adjusting unit 113p of the post-filter unit 113 can also adjust the value of the parameter r ₃ to control the phase advance in the output signal of the DC cut filter processing.

Incidentally, in the present embodiment, the post-filter unit 113 always stores the output signal after the post-filter processing for the past N (N is a natural number of sufficient magnitude) sample. As will be described later, the vibration displacement determination unit 116 selects one of these N samples in consideration of the compensation phase component, and determines the estimated vibration displacement based on the selected output signal sample. It is.

The quasi-instantaneous frequency determination unit 114 determines the "instantaneous" or "quasi-instantaneous" frequency in the measured quantity for which the vibration displacement is estimated. Of these, the "instantaneous" frequency is calculated by a known method using complex signal processing, but in the present embodiment, in order to suppress the processing delay as much as possible, the quasi-instantaneous frequency determining unit 114 calculates the "quasi-instantaneous" frequency. .. Incidentally, after this, the compensation amount determination unit 115 calculates the phase lead portion and the gain fluctuation portion of the amplitude of the signal (which has been subjected to the post-filter processing) at the quasi-instantaneous frequency determined here.

Specifically, the quasi-instantaneous frequency determination unit 114 uses a positive / negative threshold value for zero cross determination (on the time axis) in the time-series signal (measurement amount processed by the DC cut filter) output from the pre-filter unit 111. Two past zero-cross points are detected, the number of samples corresponding to the half cycle of the signal is calculated from these points, and the quasi-instantaneous frequency is determined using the sampling frequency. Here, the output signal of the pre-filter unit 111 is more likely to zero-cross than the output signal of the integral approximation filter unit 112 and the output signal of the post-filter unit 113 in the subsequent stage, resulting in a stable result (quasi-instantaneous frequency). Is easy to obtain, so it is more suitable for calculating the quasi-instantaneous frequency.

Here, the quasi-instantaneous frequency determined in this way tends to change significantly in a short time when the target time-series signal is a signal related to complicated vibration. Therefore, it is also preferable to perform a low-pass filter processing on the time-series signal before the quasi-instantaneous frequency determination processing so that a quasi-instantaneous frequency that is stable to some extent can be obtained.

Further, when a bias occurs in the frequency estimation error derived from the input acceleration waveform, it is also preferable to set a coefficient for removing the bias in the filtering process and adjust the coefficient value. Further, unlike the case described above, in the case of an implementation environment in which a time-series signal related to a simple vibration in which vibrations of a plurality of frequencies are not mixed can be obtained, the acceleration waveform in the measured quantity input to the prefilter unit 111 is quasi. It is also preferable to determine the instantaneous frequency. Of course, the frequency may be estimated from each of the post-integration correction waveform and the acceleration waveform, and a predetermined average thereof may be used as the quasi-instantaneous frequency.

Further, as a more preferable method, it is also possible to estimate the quasi-instantaneous frequency by using a method having high robustness (resistant to waveform collapse) such as frequency estimation by the autocorrelation method.

Similarly, in the functional block diagram of FIG. 1, the compensation amount determination unit 115 is the phase lead portion at the determined quasi-instantaneous frequency in the output signal (measurement amount after post-filter processing) of the post-filter unit 113. Compensated phase (number of compensated phase samples) ",
(B) The "compensation gain", which is the gain fluctuation of the amplitude at the determined quasi-instantaneous frequency, is calculated. Here, the signal input to the compensation amount determination unit 115 is a time-series signal that has undergone pre-filter processing, integral approximation filter processing (corresponding to second-order integration), and post-filter processing in the present embodiment.

Specifically, the compensation amount determination unit 115 calculates the phase lead of the filter output signal with respect to the filter input signal at the quasi-instantaneous frequency for each of the above three filter processes performed, based on the transfer function thereof. After adding the phase lead of, it is converted into the number of samples to calculate the "compensation phase sample number". In addition, the gain, which is the ratio of the input / output signals of each filter, is calculated, and these gains are multiplied to calculate the “compensation gain”.

First, the calculation of the compensation phase sample number p_pre of the DC cut filter performed by the pre-filter unit 111 is shown. By converting the recurrence formula (y [i] ＝ x [i] －x [i-1] ＋ r ₁ * y [i-1]) in the time region shown in the above equation (1) into the Z region. , The transfer function (h (z) = (1-z- ¹ ) / (1-r ₁ * z- ¹ )) shown in the above equation (2) is calculated, and this transfer function is used as the frequency ( When converted to the ω) region, the following equation (9) h (ω) = (1-exp (-jωT)) / (1-r ₁ * exp (-jωT)) (j is an imaginary unit)
Is obtained. Here, T is the reciprocal of the sampling frequency F _S (Hz) (T = 1 / F _S ).

Next, by using the above equation (9), the compensation phase sample number p_pre in the pre-filter processing can be calculated by the following equation.
(10) p_pre ＝ F _S ＊ delay (Unit is the number of samples)
delay = (2πf) ^-1 * henkaku (unit is seconds)
henkaku ＝ Arg ((1-exp (-j * 2πfT)) / (1-r ₁ * exp (-j * 2πfT))) (Unit is rad)
Here, f is the quasi-instantaneous frequency (Hz), and Arg (CN) is a function that returns the argument of the complex number CN.

Further, the compensation gain g_pre (ratio) in the pre-filter processing can be calculated by the following equation.
(11) g_pre = 1 / Mod ((1-exp (-j * 2πfT)) / (1-r ₁ * exp (-j * 2πfT)))
Here, Mod (CN) is a function that returns the absolute value (radius) of the complex number CN.

Incidentally, the compensation phase sample number p_post and the compensation gain g_post in the post filter processing using the DC cut filter are also calculated by the equations in which r ₁ is replaced with r ₃ in the above equations (10) and (11), respectively. ..

Next, the calculation of the compensation phase sample number p_mid of the integral approximation filter performed by the integral approximation filter unit 112 is shown. By converting the recurrence formula (y [i] = r ₂ * y [i-1] + x [i]) in the time region shown in the above equation (3) to the Z region, the above equation (4) can be obtained. The transfer function shown (h (z) = 1 / (1-r ₂ * z ^-1 )) is calculated. When this transfer function is converted into the frequency (ω) region, the following equation (12) h (ω) = 1 / (1-r ₂ * exp (-jωT)) (j is an imaginary unit, T = 1 / F _S )
Is obtained.

Next, by using the above equation (12), the compensation phase sample number p_mid in the integral approximation filter processing equivalent to the second-order integral can be calculated by the following equation.
(13) p_mid = 2 * F _S * delay (Unit is the number of samples, Count 2 is processed twice)
delay = (2πf) ^-1 * henkaku (unit is seconds)
henkaku = Arg (1 / (1-r ₂ * exp (-j * 2πfT))) (Unit is rad, f is quasi-instantaneous frequency)

The compensation gain g_mid (ratio) in the pre-filter processing can be calculated by the following equation.
(14) g_mid = 1 / (Mod (1 / (1-r ₂ * exp (-j * 2πfT)))) ² (The square of the denominator is processed twice)

By the way, when the above equations (5) and (6) are used as the modification mode of the integral approximation filter processing, the compensation phase sample number p_mid and the compensation gain g_mid (ratio) in the integral approximation filter processing equivalent to the second-order integral are respectively. It can be calculated by the following equations (15) and (16).
(15) p_mid = 2 * F _S * delay (Unit is the number of samples, Count 2 is processed twice)
delay = (2πf) ^-1 * henkaku (unit is seconds)
henkaku = Arg ((1-p * exp (-j * 2πfT)) / (1-r ₂ * exp (-j * 2πfT))) (Unit is rad, f is quasi-instantaneous frequency)
(16) g_mid = 1 / (Mod ((1-r ₂ * exp (-j * 2πfT)) / (1-r ₂ * exp (-j * 2πfT)))) ² (The square of the denominator is processed twice. according to)

From the number of compensated phase samples p_pre, p_mid and p_post calculated as described above, the total number of compensated phase samples p_all related to all filter processing is calculated by the following equation (17) p_all ＝ p_pre ＋ p_mid ＋ p_post.
Calculated by. From the compensation gains g_pre, g_mid, and g_post, the total compensation gain g_all related to all filtering is calculated by the following equation (18) g_all = g_pre * g_mid * g_post.
Calculated by.

Similarly, in the functional block diagram of FIG. 1, the vibration displacement determination unit 116 is the output signal (integral approximation filter processing) of the post filter unit 113 at the previous time point (previous time point) corresponding to the total number of compensated phase samples (phase advance) p_all. And the vibration displacement is determined based on the measured amount that has been post-filtered.

Specifically, the vibration displacement determination unit 116 stores the output time-series signal of the post filter unit 113, and selects the sample position of one of the stored past N samples in consideration of the compensation phase. The vibration displacement is determined by correcting the total compensation gain (gain fluctuation) g_all for the output signal sample selected in. This can be expressed using a mathematical formula as follows. Hereinafter, the measured amount related to the acceleration in the X-axis direction of the display screen is dealt with, and the vibration displacement in the X-axis direction is calculated. Naturally, the same processing is performed in the Y-axis direction as well.

First, the vibration displacement determination unit 116 is subjected to the output time-series signal of the post filter unit 113.
(19) x_post [i ₀ ], x_post [i ₀ -1], x_post [i ₀ -2], ..., x_post [i ₀ -N]
Is saved. In the above equation (19), the time point (current time) at which the vibration displacement is estimated is set to i ₀ . The vibration displacement determination unit 116 is a signal sample past x_post [i ₀ -p_all] by the number of compensating phase samples p_all calculated for the quasi-instantaneous frequency f [i ₀ ] at this time i ₀ from these signal samples. ] Is selected as the output signal sample at time point i ₀ . Here, if p_all is not an integer, the value obtained by interpolating from the values of the signal samples before and after the p_all sample position may be used as the output signal sample, or the integer value I (rounded off p_all) obtained. You can also select x_post [i ₀ -I (p_all)] corresponding to p_all) as the output signal sample.

The vibration displacement determination unit 116 then multiplies the selected x_post [i ₀ -p_all] by the compensation gain g_all, which is also calculated for the quasi-instantaneous frequency f [i ₀ ], and the following equation (20) X_out [i ₀ ] = G_all * x_post [i ₀ -p_all]
Determines the oscillating displacement X_out [i ₀ ] estimated at time point i ₀ .

In this way, according to the present embodiment, the vibration displacement _{X_out} [ i ₀ ] can be estimated. In addition, this vibration displacement estimation process can handle vibrations in a wider frequency band than before.

Incidentally, it is also preferable that the various compensation amounts determined by the compensation amount determination unit 115 and the vibration displacement determined (estimated) by the vibration displacement determination unit 116 are temporarily stored in the analysis result storage unit 104 and appropriately read out. Further, these various compensation amounts and estimated vibration displacements are transmitted from the input / output control unit 117 to an external information processing device, for example, a seat back monitor device installed in the automobile 2 via the communication interface unit 105, and the device concerned. It may be used in.

Here, a preferred embodiment in which the vibration displacement is estimated and the image display position on the screen of the touch panel display 103 is corrected (display screen shake correction) will be described.

In the present embodiment, the sigmoid function adjusting unit 116s of the vibration displacement determining unit 116 has a problem of estimation accuracy in the low frequency region where the need for correction is low with respect to the vibration displacement X_out [i ₀ ] estimated at the time point i ₀ . It is also preferable to multiply a predetermined function (sigmoid function in this embodiment) value indicating a monotonous increase with respect to the frequency in order to deal with the above. In this case, the vibration displacement determining unit 116 determines the vibration displacement based on the multiplied amount.

In fact, in the case of low frequency vibration, the image is not difficult to see (because the eyes can easily follow the movement of the image), so the correction process of the image display position is almost unnecessary. On the other hand, the quasi-instantaneous frequency determined by the above method is often unstable in the low frequency region, and further, the phase and gain of the output signal of the above-mentioned DC cut filter processing and integral approximation filter processing are determined. It tends to show large fluctuations in the low frequency region.

Therefore, by performing the sigmoid function value multiplication process as described above, it is possible to adjust the vibration displacement in the low frequency region, where vibration displacement estimation is generally unnecessary in image display position correction, to zero or a minute amount. It is. The sigmoid function is a well-known function that is also used in a signal transduction model of nerve cells, and is a monotonically increasing function of a frequency having a value between 0 and 1.

Here, specifically, the correction frequency band setting process described above will be described using a mathematical formula. The estimated vibration displacement X _C _out [i ₀ ] after correction is given by the following equation (21) X _C _out [i 0]. ₀ ] = X_out [i ₀ ] * (1 / (1 + exp (-g_gg * (f [i ₀ ] -g_gc))))
= G_all * x_post [i ₀ -p_all] * (1 / (1 + exp (-g_gg * (f [i ₀ ] -g_gc))))
It is represented by. Here, g_gg is the gradient of the sigmoid function and is set to 4, for example. Further, g_gc is a frequency at which the sigmoid function takes a median value (0.5), and is set to, for example, 2.5 (Hz). Furthermore, f [i ₀ ] is the quasi-instantaneous frequency at time point i ₀ .

It should be noted that for vibrations having a frequency higher than the maximum frequency f_max of the target vibrations, the phase cannot be matched between the estimated displacement and the actual vibration displacement due to the latency of the system. Therefore, under the condition that vibration with a frequency higher than the design value f_max occurs, the monotonic subtraction function obtained by subtracting the sigmoid function value from the value 1 (the calculated output result changes from 1 to 0 with respect to the input of the quasi-instantaneous frequency). Function) can be adopted. In this case, in the sigmoid function, g_gc is set to f_max, or a frequency slightly higher than f_max is set, and g_gg is set to a steep one. This makes it possible to set the vibration displacement not to be estimated for vibrations with frequencies higher than the maximum frequency f_max.

Next, a specific example of correcting the display image position on the smartphone 1 with the corrected estimated vibration displacement will be described.

Here, after the predetermined application program is executed in the application unit 122 of the smartphone 1, the image related to the execution is displayed on the screen of the touch panel display 103 via the image display control unit 117g of the input / output control unit 117. Consider a situation in which a latency of approximately 50 milliseconds (msec) occurs until this is done. Further, it is assumed that the execution cycle F _S of the application program is 5 msec (that is, the frequency is 200 Hz). In such a case, the image display is delayed by about 10 (= 50/5) samples for the output signal sample.

Therefore, the pre-filter unit 111, the integral approximation filter unit 112, and the post-filter unit 113
-Number of compensating phase samples for the maximum frequency f_max in vibration received by the smartphone 1 (accelerometer 101) p_all (f_max)
Is 10 samples corresponding to the above latency (50 msec) (or, for example, between 10 plus or minus 1 sample), parameters r ₁ , parameter r ₂ (and p), and parameter r, respectively. It is also preferable to adjust ₃ .

Incidentally, the above-mentioned maximum frequency f_max is also set in advance to the maximum frequency that can be supported in consideration of the execution content and execution status of the application program.

FIG. 2 is a graph for explaining adjustment of filtering parameters in the pre-filter unit 111, the integral approximation filter unit 112, and the post-filter unit 113. Here, this graph is a graph showing the frequency dependence of the movement change amount (unit is the number of samples) in each filter processing. The maximum frequency f_max is set to 10Hz.

According to FIG. 2, in all of the pre-filter processing, the integral approximation filtering processing, and the post-filter processing, the movement change amount (number of samples) is a very large value in the ultra-low frequency region, and the frequency is high. As it becomes smaller, it decreases rapidly, and each shows a tendency to gradually approach a predetermined value.

Here, in this example, by adjusting the parameters (r ₁ , r ₂ (and p) and r ₃ ) in each filtering process, the maximum frequency f_max of the movement change (number of samples) as a result of these filtering processes. The total value at (10Hz) matches the processing latency (50msec, 10 samples) on the smartphone 1. This makes it possible to estimate the vibration displacement used for correcting the display image position on the smartphone 1 in a form more in line with the current situation, considering the latency of the processing on the smartphone 1.

Incidentally, it has been explained that the output signal after the post-filter processing is always stored only for the past N samples in the post-filter unit 113, but this N value is the total of the phase changes of the three filter processings in FIG. It can be determined from the graph of values.

Here, for example, assuming that the frequency for which the vibration displacement is estimated is 2 to 10 Hz, the total value of the phase changes in FIG. 2 shows a maximum value of 88 (number of samples) at 2 Hz in this frequency range. Therefore, it is possible to determine that the output signal after the post-filter processing should be stored, for example, for the past 90 samples (N = 90).

In adjusting each parameter, it is also preferable to adjust the parameters so as not to increase the error more than a predetermined value, considering that the error of the estimated value of vibration displacement itself may increase depending on the value of these parameters. .. Further, it is also preferable to adjust the parameters so as to suppress the error as much as possible in consideration of the fact that the error tends to increase even when the vibration of the displacement estimation target is a vibration greatly deviating from the sine wave.

Further, when the vibration of the vibration displacement estimation target is shown as a continuous waveform instead of an impulse shape, there is almost no problem even if the vibration is corrected and controlled by the backing. Moreover, when the positive and negative of the waveform of the estimated vibration displacement is reversed, the vibration displacement whose phase is advanced by 180 ° is obtained. Therefore, by performing the positive / negative inversion processing of the vibration displacement waveform in this way, it is possible to supplement the latency by half a cycle of the maximum frequency f_max in the corresponding vibration.

FIG. 3 is a schematic diagram showing a specific example in which the display image position on the smartphone 1 is corrected based on the estimated vibration displacement described above.

According to FIG. 3, the smartphone 1 is subjected to vibration and is displaced by D in the X-axis plus direction at the time point t ₀ (time point i ₀ at the sample position). On the other hand,
(A) The quasi-instantaneous frequency determination unit 114 determines the quasi-instantaneous frequency f [i ₀ ] for the output time-series signal (measured quantity related to the acceleration of the vibration) of the pre-filter unit 111.
(B) The compensation amount determination unit 115 calculates the total number of compensation phase samples p_all and the total compensation gain g_all at the quasi-instantaneous frequency f [i ₀ ].
Here, the total number of compensated phase samples p_all is a value corresponding to the latency of processing in the smartphone 1. Then
(C) The vibration displacement determination unit 116 uses the calculated p_all and g_all to obtain the vibration displacement X _C _out [i ₀ ] (the above equation (21)) for the quasi-instantaneous frequency f [i ₀ ] at the time point t ₀ . Calculate and determine.

Here, for the vibration displacement X _C _out [i 0] in (c) above, the vibration displacement _{x _post [i 0 -p_all} ] calculated at the past time point by the advance of the phase (time) in the vibration displacement calculation process is used. It is determined by the above, and it is almost the same as the actual vibration displacement D.

The image display control unit 117g displays the image displayed on the touch panel display 103 in the direction opposite to the direction of the vibration displacement X _C _out [i ₀ ] (X-axis plus direction) (X-axis minus direction). X _C _out [i ₀ ] is displaced by the absolute value and displayed.

By performing the display image position correction process as described above not only in the X-axis direction but also in the Y-axis direction at each sample time, the user can use the smartphone 1 even though the smartphone 1 is shaking. It is possible to enjoy an image display that is almost unaffected by shaking.

[Vibration displacement estimation method]
FIG. 4 is a flowchart schematically showing an embodiment of the vibration displacement estimation method according to the present invention. Incidentally, this embodiment is a vibration displacement estimation method for vibration in the X-axis direction or the Y-axis direction, but as another embodiment, it is also possible to estimate the displacement by Z-axis rotation. In this case, unlike the flowchart of FIG. 4, the rotational displacement is calculated by performing the first-order integral filter processing on the “measured quantity related to the angular velocity” from the gyro sensor.

(S101) Acquire the output signals (measured quantities) of the acceleration sensor 101 and the gyro sensor 102.
(S102) A known sensor fusion process is performed on the acquired output signal to generate an accelerated measurement amount with noise removed and improved accuracy.
(S103) Pre-filtering is performed on the generated acceleration measure to remove ultra-low frequency acceleration components that cause noise such as gravitational acceleration and centrifugal acceleration. Further, the acceleration measurement amount after the pre-filter processing is stored as a time-series signal for later use for quasi-instantaneous frequency determination.

(S104) An integral approximation filter process corresponding to the second-order integral is performed on the pre-filtered acceleration measure, and an "estimated amount related to displacement" is generated.
(S105) A post-filter process is performed on the generated "estimated amount related to displacement" to remove the DC component accumulated / remaining by the integral equivalent process, and the DC component is accumulated as a time-series signal.

(S106) Using the time-series signal of the pre-filtered acceleration measure, the quasi-instantaneous frequency in the vibration to be the displacement estimation target is determined.
(S107) The number of compensated phase samples (phase lead) and compensation gain at the quasi-instantaneous frequency are calculated, and the time-series signal of the "estimator related to displacement" subjected to post-filter processing is used based on the number of compensated phase samples. One signal sample is selected to determine the vibration displacement taking into account the compensation gain.

(S108) The determined vibration displacement is subjected to a sigmoid function multiplication process to determine the adjusted vibration displacement used for correcting the image display position.
(S109, S110) The touch panel display 103 displays an image whose display position has been corrected based on the determined adjusted vibration displacement.

As described in detail above, according to the present invention, it is possible to determine the vibration displacement based on the measured quantity subjected to the "integral approximation filter processing" at the previous time point corresponding to the calculated "phase lead". can. Specifically, the filter processing output is selected by going back by the amount that the filter processing output for the measured quantity advances on the time axis (related to the phase), and it is based on the selected (more real-time) filtering output. This makes it possible to estimate the vibration displacement in real time.

That is, according to the present invention, even when there is latency in the device processing, the vibration displacement can be estimated without depending on the displacement prediction in the future time.

By the way, in the future, we will navigate not only in transportation such as automobiles and trains, but also in various mobile objects such as bicycles and motorcycles, and in various environments such as sea, underwater (seabed), and even in the atmosphere and outer space. It is expected that the chances of viewing and viewing the image displayed on the device will further increase in the moving body. It is considered that the present invention greatly contributes to the realization of an easy-to-see image display with improved visibility even in such a situation.

Further, the application destination of the vibration displacement estimation process according to the present invention is not limited to the vibration (sway) correction in the image display of the image display device used in the moving body. For example, it is applied to vibration correction for captured images in a portable camera such as a wearable camera, vibration suppression processing generated from a drive system in a moving body, and vibration suppression control in a powered machine such as an electric motor. It is possible.

In the various embodiments of the present invention described above, various changes, modifications and omissions within the scope of the technical idea and viewpoint of the present invention can be easily carried out by those skilled in the art. The above description is merely an example and does not intend to limit the present invention. The present invention is limited only to the scope of claims and their equivalents.

1 Smartphone (vibration displacement estimation device)
101 Acceleration sensor 102 Gyro sensor 103 Touch panel display 104 Analysis result storage unit 105 Communication interface unit 111 Pre-filter unit 111p, 112p, 113p Parameter adjustment unit 112 Integral approximation filter unit 113 Post filter unit 114 Semi-instantaneous frequency determination unit 115 Compensation amount determination Part 116 Vibration displacement determination part 116s Sigmoid function adjustment part 117 Input / output control part 117g Image display control part 121 Sensor integration part 122 Application part 2 Automobile (moving body)

Claims (10)

Vibration displacement is estimated from the measured amount related to the acceleration or velocity of vibration, and the vibration displacement estimation that makes the computer installed in the device capable of suppressing the influence of the vibration on the processing result output of the processing performed by itself based on the vibration displacement function. It ’s a program,
An integral approximation filter means that applies an integral approximation filter process equivalent to an integral whose phase advances more than the integral filter process to the measured quantity.
Compensation amount determining means for calculating the phase advance amount in the measured quantity subjected to the integral approximation filter processing, and
The computer is made to function as a vibration displacement determining means for determining the vibration displacement based on the measured quantity subjected to the integral approximation filter processing at the previous time point corresponding to the phase lead .
The integral approximation filter means is a phase lead amount calculated by the compensation amount determining means, and the value of the phase lead amount calculated for the frequency related to the vibration affecting the processing result output is the value of the device. The parameters in the integral approximation filter processing are adjusted so that the value corresponds to the delay time until the processing result output in the processing or the value is within a predetermined range from the corresponding value, and the integral approximation filter processing is performed. implement
This is a vibration displacement estimation program. The compensation amount determining means also calculates the gain fluctuation amount in the measured quantity subjected to the integral approximation filter processing.
The vibration displacement determining means is characterized in that the vibration displacement is determined by correcting the gain fluctuation amount with respect to the measured amount subjected to the integral approximation filter processing at the previous time point corresponding to the phase advance amount. The vibration displacement estimation program according to claim 1. The vibration displacement estimation program further functions the computer as a low-frequency elimination filter means for performing a low-frequency elimination filter process for removing low-frequency components on the measured quantity before and / or after the integral approximation filter process.
The vibration displacement estimation program according to claim 1 or 2, wherein the compensation amount determining means calculates a phase lead amount in a measured quantity subjected to the low frequency elimination filter process and the integral approximation filter process. The device is a device that estimates vibration displacement from a measured quantity related to vibration acceleration.
The vibration displacement estimation according to any one of claims 1 to 3, wherein the integral approximation filter means performs an integral approximation filter process corresponding to a second-order integral in which a phase advances to a measured quantity related to the acceleration. program. The vibration displacement estimation program further functions the computer as a frequency determining means for determining an instantaneous or quasi-instantaneous frequency in the measured quantity.
The method according to any one of claims 1 to 4, wherein the compensation amount determining means calculates a phase advance amount at the instantaneous or quasi-instantaneous frequency in the measured quantity subjected to the integral approximation filtering process. Vibration displacement estimation program. The vibration displacement determining means reduces the vibration displacement in the low frequency region, which does not require vibration displacement estimation, to zero or a minute amount with respect to the measured amount subjected to the integral approximation filter processing at the previous time point corresponding to the phase advance. The vibration displacement estimation program according to any one of claims 1 to 5, wherein the vibration displacement is determined based on a quantity obtained by multiplying a predetermined function value indicating a monotonous increase with respect to a frequency. The device according to any one of claims 1 to 6 , wherein the device corrects the display position of the image in the image display as the processing result output by the image processing of the device based on the determined vibration displacement. The described vibration displacement estimation program. A vibration displacement estimation program that operates a computer mounted on a device that can estimate vibration displacement from measured quantities related to vibration acceleration or velocity.
An integral approximation filter means that applies an integral approximation filter process equivalent to an integral whose phase advances more than the integral filter process to the measured quantity.
Frequency determining means for determining the instantaneous or quasi-instantaneous frequency in the measured quantity,
Compensation amount determining means for calculating the phase lead at the instantaneous or quasi-instantaneous frequency of the measured quantity subjected to the integral approximation filter processing, and
The vibration displacement at the time of estimation of the vibration displacement is determined based on the measured amount subjected to the integral approximation filter processing at the time point corresponding to the calculated phase advance amount when viewed from the time of estimation of the vibration displacement. A vibration displacement estimation program characterized by operating a computer as a means for determining vibration displacement. The compensation amount determining means also calculates the gain fluctuation amount at the instantaneous or quasi-instantaneous frequency in the measured quantity subjected to the integral approximation filter processing.
The vibration displacement determining means corrects the gain fluctuation amount for the measured amount subjected to the integral approximation filter processing at the time point corresponding to the phase advance amount when viewed from the vibration displacement estimation time point. The vibration displacement estimation program according to claim 8 , wherein the vibration displacement at the time of estimating the vibration displacement is determined by the operation. It is a vibration displacement estimation device that can estimate the vibration displacement from the measured amount related to the acceleration or velocity of the vibration and suppress the influence of the vibration on the processing result output of the processing that is performed by itself based on the vibration displacement.
An integral approximation filter means that applies an integral approximation filter process equivalent to an integral whose phase advances more than the integral filter process to the measured quantity.
Compensation amount determining means for calculating the phase advance amount in the measured quantity subjected to the integral approximation filter processing, and
It has a vibration displacement determining means for determining a vibration displacement based on a measured quantity subjected to the integral approximation filter processing at a previous time point corresponding to the phase advance portion.
The integral approximation filter means is a phase lead amount calculated by the compensation amount determining means, and the value of the phase lead amount calculated for the frequency related to the vibration affecting the processing result output is the value of the device. The parameters in the integral approximation filter processing are adjusted so that the value corresponds to the delay time until the processing result output in the processing or the value is within a predetermined range from the corresponding value, and the integral approximation filter processing is performed. implement
A vibration displacement estimator characterized by this .

Images