JP2008188092A - Data processing method, data processing apparatus, and data processing program - Google Patents

Abstract

The present invention relates to a data processing method, a data processing apparatus, and a program for improving data processing speed and data processing accuracy with a simple configuration when analyzing a walking state of a quadruped animal.
Vibration data detection steps A10 and A20 for detecting vibrations generated by walking movement of a quadruped animal as time-series vibration data, and amplitude for converting the vibration data to extract the magnitude of amplitude energy Energy conversion step A30, basic vibration data generation step A50 for generating time-series basic vibration data from the converted data, and basic data points that characterize fluctuations in the vibration data are extracted based on the basic vibration data. Basic data point extraction step A60, extraction range setting step A70 for setting a predetermined time range based on the detection time of the basic data point as the extraction range of the vibration data, and vibration data included in the extraction range And a feature data point extraction step A80 for extracting feature data points from the inside.
[Selection] Figure 3

Description

  The present invention relates to a data processing method, a data processing device, and a data processing program for diagnosing a walking state by analyzing a walking rhythm during a walking movement of a quadruped animal.

2. Description of the Related Art Conventionally, a technique for diagnosing a subject's health condition by extracting biological rhythm information of a human body such as a walking rhythm or a blinking rhythm and analyzing the periodic variation thereof is known.
For example, Patent Document 1 discloses a configuration in which information on vibration (acceleration fluctuation) of a limb is analyzed as a walking rhythm when a person walks, and a decrease in physiological function or aging is determined. Here, fractal analysis, frequency (FFT) analysis, self-function analysis, chaos analysis and the like are listed as analysis methods. By observing the fluctuations of periodic vibration data using these data analysis techniques, it is possible to extract linear or nonlinear structures inherent in walking rhythms, which makes it difficult to diagnose diseases and physiology of the cranial nervous system that are difficult to diagnose. It is possible to objectively diagnose the decline in function.
JP 2000-166877 A

However, as a result of investigations by the present inventors, it has been found that when a quadruped animal is a diagnosis target, it is difficult to accurately extract a walking rhythm and diagnosis is difficult.
That is, in the walking movement of a quadruped animal, different vibrations are generated depending on the movement of each of the four limbs. Therefore, if you focus on the movement of one arbitrary foot and try to extract only the rhythm of the movement of that foot, the rhythm resulting from the movement of the other three feet will be mixed in and the walking interval will be accurately extracted I could not.

In particular, the phase difference of the walking rhythm of each limb during quadrupedal walking is approximately half of the phase difference of the walking rhythm of each limb during biped walking. It was difficult to easily extract the limb walking rhythm.
The present invention has been devised in view of such a problem, and a data processing method and data processing capable of improving data processing speed and data processing accuracy with a simple configuration when analyzing the walking state of a quadruped animal. An object is to provide an apparatus and a program.

  In order to achieve the above object, the data processing method of the present invention according to claim 1 includes a vibration data detection step for detecting vibrations generated as a result of walking movement of a quadruped animal as time-series vibration data, and the vibration data. An amplitude energy conversion step for performing data conversion to extract the magnitude of amplitude energy in the vibration data detected in the detection step, and for grasping the characteristics of the vibration for the data obtained in the amplitude energy conversion step The basic vibration data generating step for generating the time-series basic vibration data by performing the signal processing as the pre-processing, and the fluctuation of the vibration data based on the basic vibration data generated in the basic vibration data generating step A basic data point extracting step for extracting basic data points characterizing the data, and the basic data points extracted in the basic data point extracting step An extraction range setting step for setting a predetermined time range based on the detection time as the extraction range of the vibration data, and among the vibration data detected in the vibration data detection step, the extraction range setting step sets the range. And a feature data point extracting step of extracting the vibration data characterizing the vibration as a feature data point from the vibration data included in the extraction range.

Here, as a quadruped animal, specifically, humans grasp and manage the behavior of domestic animals, pets, laboratory animals such as cattle, pigs, sheep, horses, dogs, cats, hamsters, rats, mice, etc. And animals to do. Of these, cows, horses, and dogs are preferable.
Examples of the data conversion process for extracting the magnitude of the amplitude energy include an absolute value process and a square process.

  Note that the preprocessing refers to general irreversible (with irreversible changes) arithmetic processing for making it easy to find fluctuations in the vibration data, for example, unnecessary processing included in the vibration data. It includes filter processing for removing information (noise), Hilbert transform processing, Fourier transform processing, signal processing using an averaging method, wavelet analysis processing, fractal analysis processing, and the like. In addition, arbitrary signal addition and subtraction, proportional processing, integration processing, differentiation processing, and the like are included.

  According to a second aspect of the present invention, there is provided a data processing method according to the present invention, in addition to the configuration of the first aspect, in the amplitude energy conversion step, the vibration data obtained in the vibration data detection step is an absolute value of the absolute value. It is characterized in that it is converted into value vibration data, and the absolute value vibration data is further subjected to envelope processing to convert the absolute value vibration data into envelope vibration data.

According to a third aspect of the present invention, there is provided a data processing method according to the present invention, wherein, in addition to the configuration of the first or second aspect, a linear analysis is performed on the data converted in the amplitude energy conversion step in the basic vibration data generation step. The basic vibration data is generated by performing the signal processing using a technique.
According to a fourth aspect of the present invention, there is provided a data processing method according to the present invention, wherein, in addition to the configuration according to any one of the first to third aspects, the fundamental vibration data generation step converts the amplitude energy conversion step. A predetermined frequency component set in advance is filtered from the data.

  According to a fifth aspect of the present invention, there is provided a data processing method according to the present invention, in addition to the configuration according to any one of the first to fourth aspects, the basic vibration data generating step converts the amplitude energy conversion means. Wave data obtained by smoothing envelope vibration data is generated as the basic vibration data, and in the basic data point extraction step, a point on the basic vibration data correlated with the wavelength of the basic vibration data is extracted as the basic data point. It is characterized by doing.

  “Smoothing the envelope vibration data” means converting the waveform of the envelope vibration data into a smooth wave (waveform) by signal processing. In addition, “the point on the fundamental vibration data that correlates with the wavelength of the fundamental vibration data” refers to the peak of the wave (the point at which the amplitude is maximum, the maximum point or the minimum point in one wavelength) or zero point. Thus, it means a point that appears periodically.

According to a sixth aspect of the present invention, in addition to the configuration of the fifth aspect, in the extraction range setting step, the time near the detection time of the basic data point is set as the extraction range. It is a feature.
According to a seventh aspect of the present invention, in addition to the configuration of the sixth aspect, in the feature data point extraction step, the detection time of the vibration peak in the vibration data included in the extraction range is set as the feature data. It is characterized by extracting as points.

Further, the data processing method of the present invention described in claim 8 is based on the feature data point detected in the feature data point extraction step, in addition to the configuration described in any one of claims 1 to 7. The method further includes a nonlinear analysis calculation processing step for extracting a nonlinear structure from the vibration data.
The data processing apparatus of the present invention according to claim 9 includes vibration data detection means for detecting vibrations generated by walking movement of a quadruped animal as time-series vibration data, and the vibrations detected by the vibration data detection means. Amplitude energy conversion means for performing data conversion to extract the magnitude of the amplitude energy in the data, and signal processing as preprocessing for grasping the characteristics of the vibration for the data obtained by the amplitude energy conversion means And basic vibration data generating means for generating time-series basic vibration data, and basic data points that characterize fluctuations of the vibration data are extracted based on the basic vibration data generated by the basic vibration data generating means. Basic data point extraction means, and extraction of the vibration data within a predetermined time range based on the detection time of the basic data points extracted by the basic data point extraction means An extraction range setting means for setting as an enclosure, and characterizing the vibration from vibration data included in the extraction range set by the extraction range setting means among the vibration data detected by the vibration data detection means A feature data point extracting means for extracting vibration data as feature data points is provided.

  According to a tenth aspect of the present invention, there is provided the data processing device according to the tenth aspect, wherein the amplitude energy converting means has an absolute value comprising the absolute value of the vibration data obtained by the vibration data detecting means. The absolute value vibration data is converted into vibration data, envelope processing is further performed on the absolute value vibration data, and the absolute value vibration data is converted into envelope vibration data.

The data processing device of the present invention according to claim 11 is the configuration according to claim 9 or 10, wherein the fundamental vibration data generating means is linear with respect to the envelope vibration data converted by the amplitude energy converting means. It is characterized by comprising linear signal processing means for performing the signal processing using an analysis method and generating the fundamental vibration data.
A data processing device according to a twelfth aspect of the present invention is the data processing device according to any one of the ninth to eleventh aspects, wherein the fundamental vibration data generating means is converted by the amplitude energy converting means. It has a filter processing means for filtering a predetermined frequency component set in advance from vibration data.

  A data processing apparatus according to a thirteenth aspect of the present invention is the data processing apparatus according to any one of the ninth to twelfth aspects, wherein the fundamental vibration data generating means is converted by the amplitude energy converting means. The basic data point extracting means extracts a point on the basic vibration data correlated with the wavelength of the basic vibration data as the basic data point. It is a feature.

The data processing apparatus of the present invention according to claim 14 is characterized in that, in the configuration according to claim 13, the extraction range setting means sets a time near the detection time of the basic data point as the extraction range. It is said.
The data processing device of the present invention according to claim 15 is the data processing device according to claim 14, wherein the feature data point extraction means sets the detection time of the vibration peak in the vibration data included in the extraction range to the feature data point. It is characterized by extracting as.

A data processing device according to a sixteenth aspect of the present invention is the configuration according to any one of the ninth to fifteenth aspects, wherein the vibration is based on the feature data point detected by the feature data point extracting means. The present invention is further characterized by further comprising a nonlinear analysis calculation processing means for extracting a nonlinear structure in the data.
A data processing program of the present invention according to claim 17 is a data processing for causing a computer to function as an amplitude energy converting means, a basic vibration data generating means, a basic data point extracting means, an extraction range setting means, and a characteristic data point extracting means. A program for converting the amplitude energy into the amplitude energy conversion means for extracting the magnitude of the amplitude energy in the vibration caused by the walking motion of the quadruped animal, detected as the vibration data, and generating the basic vibration data. Means performs signal processing as pre-processing for grasping the characteristics of the vibration on the data converted by the amplitude energy conversion means to generate time-series basic vibration data, and to extract the basic data points Means for characterizing fluctuations in the vibration data based on the basic vibration data generated by the basic vibration data generating means. And the extraction range setting means sets a predetermined time range based on the detection time of the basic data points extracted by the basic data point extraction means as the vibration data extraction range, Vibration data that characterizes the vibration from vibration data included in the extraction range set by the extraction range setting means among the vibration data detected by the vibration data detection means. Are extracted as feature data points.

According to the data processing method, data processing apparatus, and data processing program of the present invention (claims 1, 9, and 17), when generating the basic vibration data, first, the magnitude of the amplitude energy of the vibration generated with the walking motion is extracted. By doing so, it is possible to easily grasp the walking rhythm of only the limb of interest, and to improve the calculation accuracy of subsequent periodic analysis processing.
In addition, in setting the extraction range of vibration data, basic vibration data subjected to signal processing as preprocessing is used, and in the extraction of more specific feature data points, it is extracted from vibration data included in the extraction range. It is possible to extract vibration data that characterizes the vibration caused by the walking movement of a paw animal very accurately.

According to the data processing method and data processing apparatus of the present invention (claims 2 and 10), the amplitude energy fluctuation is obtained by taking the absolute value of the vibration data and extracting the envelope prior to the signal processing as the preprocessing. The basic data points that characterize fluctuations in vibration data can be easily extracted.
According to the data processing method and data processing apparatus of the present invention (claims 3 and 11), by using the linear analysis method as the preprocessing, the processing can be easily completed in a short time with a simple configuration.

According to the data processing method and data processing apparatus (claims 4 and 12) of the present invention, signal processing can be easily performed by filtering vibration data. It is also possible to quickly select information sufficient to extract basic data points.
According to the data processing method and data processing apparatus of the present invention (claims 5 and 13), the point on the fundamental vibration data correlated with the wavelength of the fundamental vibration data generated by smoothing the vibration data is detected. Therefore, basic data points can be easily detected. Further, by referring to the basic data points, it is possible to easily grasp the change in the wavelength of the basic vibration data, that is, the change in the vibration period.

  According to the data processing method and the data processing apparatus of the present invention (claims 6 and 14), since the time near the detection time of the basic data point on the basic vibration data is set as the extraction range, the basic data point of the basic vibration data is set. The vibration data included in the range away from can be excluded from the feature data point extraction targets. That is, it is possible to easily extract information on a portion having a strong correlation with vibration data that characterizes walking motion.

According to the data processing method and data processing apparatus (claims 7 and 15) of the present invention, it is possible to accurately extract the characteristic data points characterizing the walking motion included in the vibration data.
According to the data processing method and data processing apparatus of the present invention (claims 8 and 16), the nonlinear structure in the vibration data can be extracted based on the accurately extracted feature data points, and the data is highly reliable. Analysis can be performed.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
1 to 5 show a data processing apparatus according to an embodiment of the present invention. FIG. 1 is a block diagram showing the overall configuration of the data processing apparatus, and FIG. 2 explains processing contents in the data processing apparatus. (A) is a time-series graph of vibration data related to data processing in the amplitude energy conversion unit, basic vibration data generation unit, and basic data point extraction unit, and (b) is data processing in the feature data point extraction unit. FIG. 3 is a flowchart showing the contents of control in the data processing apparatus, FIG. 4 is a schematic diagram showing a configuration example of the data processing apparatus using a computer, and FIG. 5 is a diagram of the data processing apparatus. It is a time series graph which shows the analysis process of the vibration data as a comparative example with respect to the processing content.

[1. Constitution]
[1-1. overall structure]
In the present embodiment, a data processing apparatus that analyzes a walking rhythm in cow walking motion will be described as a specific example. That is, this data processing device is a device that detects acceleration of vibration caused by the walking movement of the cow as vibration data, performs data processing on the data, diagnoses the walking state of the cow, and outputs it.

  As shown in FIG. 1, the data processing apparatus includes a vibration data detection device (vibration data detection means) 1, a first data processing section 9, and a second data processing section (nonlinear analysis calculation processing means) 10. The The first data processing unit 9 performs arithmetic processing for facilitating grasping the characteristics of the signal detected by the vibration data detection device 1, while the second data processing unit 10 performs substantial data processing. As an analysis of walking rhythm.

  The first data processing unit 9 and the second data processing unit 10 schematically show functional units that are arithmetically processed inside the computer, and each function is configured as an individual program. The first data processing unit 9 in this embodiment functions as an amplitude energy conversion unit, a basic vibration data generation unit, a basic data point extraction unit, an extraction range setting unit, and a feature data point extraction unit. The second data processing unit 10 functions as a nonlinear analysis calculation processing means.

An example of the configuration of the data processing apparatus using a computer is shown in FIG. The computer 12 includes the vibration data detection device 1, a storage device (ROM, RAM, etc.) 13, a central processing unit (CPU) 14, a monitor 15 as an output interface, a keyboard 16 and a mouse 17 as input interfaces. It is configured. Here, the first data processing unit 9 and the second data processing unit 10 according to the data processing apparatus are stored as programs in the storage device 13.
Hereinafter, the signal processing contents in the data processing apparatus will be conceptually described with reference to the block diagram of FIG.

[1-2. Vibration data detector]
The vibration data detection device 1 is a sensor that detects vibrations that occur with the walking movement of a quadruped animal. This vibration includes not only information directly detected by the sensor, but also information obtained by processing the information by calculation or the like and obtaining a corresponding parameter value as an estimated value.

In this embodiment, an acceleration sensor for detecting the magnitude of the acceleration generated with the walking movement of the cow is applied as the vibration data detection apparatus 1 and is mounted on the limb of the target cow.
As this acceleration sensor, a uniaxial to triaxial sensor may be arbitrarily used according to the type of animal to be measured and the purpose of analysis. It is preferable to use a triaxial acceleration sensor for detecting acceleration acting on the sensor. In this specific example, a three-axis acceleration sensor is used, and the detected information of the acceleration in the vertical direction and the detected time information are detected as vibration data S as shown in FIG. To be input.
As other specific examples of sensors that detect vibration, a speed sensor, a displacement sensor (position sensor), and the like are also conceivable.

[1-3. First data processing unit]
The first data processing unit 9 is a functional unit that performs the processing specified in claim 1 on the vibration data S. As shown in FIG. 1, the vibration data storage unit 11, the amplitude energy conversion unit (amplitude energy conversion means) ) 6, basic vibration data generation unit (basic vibration data generation unit) 2, basic data point extraction unit (basic data point extraction unit) 3, extraction range setting unit (extraction range setting unit) 4 and feature data point extraction unit (features) Data point extraction means) 5.

  The processing applied here is to convert the data by applying mathematical and electrical processing to the vibration data S in order to make it easy to grasp the characteristics (in other words, signal processing). . The processing can be classified into analog signal processing and digital signal processing depending on the type of signal to be processed. The first data processing unit 9 performs processing included in the category of digital signal processing.

[1-3-1. Vibration data storage unit]
The vibration data storage unit 11 is a functional unit that stores the vibration data S input from the vibration data detection device 1. As shown in FIG. 1, the vibration data S stored here is divided into two systems from the vibration data storage unit 11 and is input to each of the amplitude energy conversion unit 6 and the feature data point extraction unit 5. ing. That is, raw information that has not been processed at all is input to each of the amplitude energy conversion unit 6 and the feature data point extraction unit 5.

The vibration data S input to the vibration data storage unit 11 may be a series of time-series data detected during a predetermined time in the vibration data detection device 1 or the vibration data detection device 1. It may be individual measurement data detected at any time.
In the former case, the vibration data S as time series data is input to the amplitude energy conversion unit 6 and the feature data point extraction unit 5 as it is. In the latter case, each vibration data S is stored in the vibration data storage unit 11 until the total number of vibration data S from the vibration data detection device 1 is equal to or greater than a predetermined number set in advance. The vibration data S as a whole is input to the amplitude energy conversion unit 6 and the feature data point extraction unit 5 at the stage where the values are sufficiently arranged. In the present embodiment, the latter case will be described.

Note that the time-series vibration data S indicates a complicated variation including a periodic peak, as shown in FIG. This is because the vibration data detection apparatus 1 simultaneously detects not only the vibration associated with the foot movement of the limb on which the vibration data detection apparatus 1 is worn, but also the vibration associated with the foot movement of another limb body. That is, the vibrations of the limbs that are independent of each other interfere with each other, and the walking step of the limb with the vibration data detection device 1 (the time interval from when the limb touches the ground again to the ground) and the vibration data S The correspondence is difficult to grasp.
Therefore, in the present invention, such a correspondence can be more easily grasped by a configuration including an amplitude energy conversion unit 6 described below, which performs an amplitude energy conversion process on the vibration data S.

[1-3-2. Amplitude energy conversion unit]
The amplitude energy conversion unit 6 converts time-series vibration data S input from the vibration data storage unit 11 into absolute value vibration data S _A composed of absolute values thereof, and also converts the converted absolute value vibration data S _A. Is subjected to envelope processing.
First, here, the absolute value vibration data S _A is calculated according to the following equation.
S _A = | S | (Formula 1)

That is, here, the magnitude of the amplitude energy in the time-series vibration data S is extracted. Subsequently, for each data of the time-series absolute value vibration data S _A , an envelope process for linking the local maximum points (data points having a larger value than the adjacent data points before and after) is performed, and the envelope vibration data S _E is obtained. Is converted. That is, here, the envelope vibration data S _E is obtained as a polygonal line that envelops the fine amplitude of the absolute value vibration data S _A.

FIG. 2A shows the vibration data S and the envelope vibration data S _E by a thin solid line and a thick solid line, respectively. The envelope vibration data S _E obtained here is input to the basic vibration data generation unit 2.

[1-3-3. Basic vibration data generator]
The basic vibration data generation unit 2 is a functional unit that performs processing for extracting necessary signal components from the vibration data S detected by the vibration data detection device 1. In this embodiment, the digital filter processing for detecting the peak of acceleration change during walking, that is, the feature point detected in the same cycle as the cycle of the walking step is performed, and obtained by the amplitude energy conversion unit 6. The envelope vibration data S _E is subjected to three types of filter processing: a low-pass filter, a phase filter, and a high-pass filter.

The specific content of the processing performed by the basic vibration data generation unit 2 varies depending on the type and content of data to be processed or the purpose of analysis, but generally It is preferable to perform linear analysis methods such as filter processing and Fourier transform processing alone or in combination.
The low-pass filter is a filter that reduces (or removes) a high-frequency vibration component from an input signal including various frequency components. This low-pass filter suppresses noise components with higher frequencies than the frequency of acceleration fluctuations accompanying the cow's walking movement (that is, the reciprocal of the time interval from when one foot touches the ground to the ground again). ing.

In addition, the vibration of the fine time interval seen in the vibration data S shown with the thin continuous line in Fig.2 (a) is a high frequency noise component. In addition, the envelope vibration data S _E indicated by a thick solid line in FIG. 2A also includes fine vibration although the time interval is longer than that of the vibration data S. In the low pass filter, such a high frequency component of the envelope vibration data S _E is suppressed.

The phase filter is a filter that delays the phase change of the input signal. Here, the phase of the input signal is delayed by a quarter period of the acceleration fluctuation.
On the other hand, the high-pass filter is a filter that reduces (or removes) a vibration component having a frequency lower than the frequency of the acceleration fluctuation accompanying the walking movement of the cow. The high-pass filter suppresses the drift component of the signal detected by the vibration data detection device 1. By applying these filter processes, the time series data of the acceleration fluctuation accompanying the walking movement of the cow is clarified from the envelope vibration data S _E.

For example, when the above-described filter processing is performed on the envelope vibration data S _E indicated by a thick solid line in FIG. 2A, a signal indicated by a broken line is extracted. That is, from the raw information detected by the vibration data detection device 1, the feature of the walking step of one limb during the walking movement of the cow is extracted as a wave having a period corresponding to the walking step. To a time-series signal thus taken out, hereinafter referred to as fundamental vibration data S _B.
Since this basic vibration data S _B is a signal extracted through the phase filter, the time when the magnitude of this signal becomes 0 (that is, the position of the zero point) is the peak of the original envelope vibration data S _E. This corresponds to the time when the position (maximum value, minimum value) is detected.

[1-3-4. Basic data point extraction unit]
The basic data point extraction unit 3 extracts information characterizing the acceleration fluctuation based on the basic vibration data S _B generated by the basic vibration data generation unit 2. Here, as information characterizing the acceleration fluctuation, the peak position of the input vibration data S estimated from the basic vibration data S _B obtained by the filter processing is extracted. That is, the basic data point extraction unit 3 pays attention to the 0 points of the basic vibration data S _B whose signal value changes from negative to positive before and after that, and the envelope vibration data S _E corresponding to the 0 point. To extract. More specifically, so as to extract the envelope vibration data S _E time and same time of zero point during FIGS. 2 (a) as the basic data points D _B.

As described above, one cycle of the fundamental vibration data S _B corresponds (correlation) to one limb of the step during cattle walking. Therefore, periodic variation of the step would appear as a periodic variation of the fundamental vibration data S _B. Therefore, in the present embodiment, using the zero point of the fundamental vibration data S _B as a cue to know the period of the fundamental vibration data S _B (which varies the signal value at the before and after from negative to positive), the basic data points It is doing the extraction of D _B.

In this, as shown in FIG. 2 (a), between the actual peak location and the basic data points D _B of the vibration data S it can be seen that a slight shift occurs. This is due to information lost by the filter processing in the basic vibration data generation unit 2. That is, as a result of missing the characteristic information contained in the vibration data S by performing filtering occurs, it is the actual peak position and the position of the zero point of the fundamental vibration data S _B is got different .

  In general, the more irreversible arithmetic processing operations are performed as preprocessing for an input signal, the more the original information contained in the input signal is lost regardless of the processing content. That is, even if it is a processing operation for making it easy to find the fluctuation of the input signal, some information is included in the information obtained as a result. Of course, it is conceivable to pursue preprocessing with as small an error as possible within the limit that makes it easy to grasp fluctuations in the input signal, but the accuracy of the calculation obtained as a result is limited.

Further, as a typical technique for reducing such errors, there is a technique of matching the peak position and the basic data points D _B of the vibration data S by adjusting the filtering content. For example, by finely adjusting the delay amount of the phase change in the phase filter, it is conceivable to uniformly reduce the error between the peak position and the basic data points D _B of the vibration data S.
However, such a fine adjustment is a somewhat uncertain technique that must be relied on by a skilled engineer, and there is a risk that the data processing accuracy may become unstable. In addition, since such fine adjustment needs to be performed according to each target data group (for example, each individual from which a signal is detected), when handling a plurality (large amount) of data groups, Processing takes time and adjustment work is complicated. Moreover, because it can not adjust unless Verify the base data point D _B obtained by performing an actual pretreatment, in such a method is not sufficient data processing speed is obtained.

Therefore, in the present invention, simply as a fundamental vibration data S not only to extract the basic data points D _B based on _B, with an extraction range setting unit 4 and the feature data point extraction section 5 will be described below configuration.

[1-3-5. Extraction range setting section]
Extraction range setting unit 4 is configured to set the near time time that the basic data points D _B was detected as the extraction range A. Time is near, it means that the time before and after the time at which the basic data points D _B are detected. Here, as shown in FIG. 2 (b), the extraction range A is the basic data points D _B is detected before the time the predetermined time t ₁ and the basic data points D _B of a predetermined time after the time at which the detected t ₂ It consists of and. The extraction range A set here is input to the feature data point extraction unit 5 to be described later.

The predetermined times t ₁ and t ₂ are times that can be arbitrarily set. For example, it may be determined as a ratio with respect to the wavelength λ of the basic vibration data S _B or may be a preset value. In the present embodiment, each predetermined time t ₁ and t ₂ is set to be larger than the time error estimated to be caused by the above-described filtering process. The predetermined times t ₁ and t ₂ may be relatively short when the amount of information lost by the filtering process is relatively small, while the predetermined time t ₁ and t ₂ may be relatively short when performing the filtering process with a relatively large amount of information lost. Times t ₁ and t ₂ may be set relatively long.

[1-3-6. Feature data point extraction unit]
The feature data point extraction unit 5 extracts a signal characterizing the vibration, that is, an actual peak position as the feature data point D _C from the vibration data S included in the extraction range A set by the extraction range setting unit 4. To do. The calculation process here is obtained by detecting the maximum value of the envelope vibration data S _E within the extraction range A as shown in FIG. As a result, the time when the limb with the vibration data detection device 1 attached to the ground and the vertical acceleration at that time are accurately extracted as the feature data point D _C. In FIG. 2B, the position of the feature data point D _{C on} the graph is indicated by a symbol +.

Since the envelope vibration data S _E is obtained by concatenating the maximum points of the absolute value vibration data S _A , the envelope vibration data S _E is substantially determined from the vibration data S included in the extraction range A by the above calculation. Thus, the actual peak position is extracted.
In this way, the extraction range setting unit 4 and the feature data point extraction unit 5 return to the original envelope vibration data S _E again based on the basic data point D _B , and the feature data point D _C from the envelope vibration data S _E. Functions to extract. That is, the base data point D _B is set but contains errors, the extraction range A in the extraction range setting unit 4, it is assumed that information indicating the original features in the vicinity exists. Further, the feature data point extraction unit 5 reduces the calculation labor and calculation time by using only the envelope vibration data S _E included in the extraction range A as a calculation target, and the feature data point D from the envelope vibration data S _E. The accuracy is ensured by extracting _C.

[1-4. Second data processing unit]
The second data processing unit 10 is a functional unit for performing substantial data processing on the data processed in the first data processing unit 9, and as shown in FIG. 1, the analysis unit 7 and the determination unit 8. It is configured with. In the second data processing unit 10, the characteristics of the peak interval time of the acceleration data of the limb mounted with the vibration data detection device 1 are analyzed. As an analysis method in this embodiment, a non-linear analysis method for observing fluctuations in peak interval time is used.

  Here, “fluctuation” refers to a slight waveform shift (spatial and temporal change or movement that is partially irregular) that is observed when a certain wave changes every moment. . For example, not only time-series data that detects body movements associated with walking as acceleration fluctuations, but also vital signs such as respiratory rate, heart rate, and brain waves are used as time-series data, the peak interval and period of those waves are It is known to exhibit complex fluctuations that are not constant. On the other hand, a number of methods have been proposed for analyzing the structure that seems to dominate the behavior among such complex changes that appear irregular. The second data processing unit 10 analyzes the nonlinear structure existing behind the fluctuation by observing the degree of fluctuation of the peak interval time using such a method.

Specific analysis methods include known analysis methods such as spectrum analysis (FFT analysis), fractal analysis (multifractal analysis, detrend fluctuation analysis, etc.), chaos analysis, and wavelet analysis. Detrend fluctuation analysis, which is one of the analysis methods, is used.
The detrend fluctuation analysis method is a statistical analysis method in which the complexity of the wave to be analyzed is evaluated by a value called a scaling index. In the present embodiment, the acceleration data of either the left or right foot is aligned in the data alignment unit 6, the scaling index of the acceleration data is calculated in the analysis unit 7, and the evaluation is performed in the determination unit 8. ing.

[1-4-1. Analysis Department]
Analyzing unit 7 calculates the scaling exponent characteristic data points D _C input from the feature data point extraction unit 5 of the first data processing unit 9. Specifically, divided into n intervals based feature data points D _C on the detection time, the respective feature data points D _C is detected time intervals in each section and (walking interval) and its Trend The least square error (variance) F is calculated, and the slope α of the logarithmic plot of each of the division number n and the variance F is calculated as a scaling index. Note that the trend means a trend of data transition in each section. For example, the data in each section is approximated to a straight line.

In this way, if by changing the division number n of sections, also would be varied observation scale characteristic data points D _C group, dispersion F is calculated also changes. On the other hand, if a linear relationship is recognized in each logarithmic plot of the division number n and the variance F, it means that a scale in self-similarity exists between the division number n and the variance F. That is, here, the magnitude of the self-similarity of the degree of variation in the actual walking interval time when the section to be observed is changed is calculated as the scaling index α.

[1-4-2. Judgment unit]
The determination unit 8 determines the walking state based on the scaling index α calculated by the analysis unit 7. In general, it is known that the correlation between the division number n and the variance F can be determined by the value of the scaling index α. For example, when 0.5 <α <1, long-range correlation is recognized, and when α = 1, 1 / f fluctuation correlation is recognized. If there is a correlation but no fractal property is observed, α> 1.

  Note that “1 / f fluctuation” means fluctuation among the aforementioned fluctuations such that the magnitude (power spectrum) of the fluctuation component is 1 / f with respect to the frequency f. For example, if a spectrum analysis is performed on a wave that exists in nature, such as a buzzing sound of a stream, a wind pressure of a breeze, a shape of a wood grain, or a chirping sound of a bird, a 1 / f fluctuation whose power spectrum is inversely proportional to the frequency f can be observed. In recent years, it has been found that fluctuations are also observed in biological signals (vital signs) emitted from humans and animals, and in particular, 1 / f fluctuation is used as an index for judging the health state of an observation target. Many usefulness has been reported.

  Based on these characteristics, the determination unit 8 determines that the better the walking state is, the closer the scaling index α is to α = 1. The determination result here is output to the monitor 15 described above.

[2. flowchart]
The control contents in the data processing apparatus will be described with reference to the flowchart shown in FIG.
In step A10, acceleration detection information and detection time information thereof are detected as vibration data S by the acceleration sensor as the vibration data detection device 1. The vibration data S detected here is input to the vibration data storage unit 11 of the first data processing unit 9 and stored therein.

  In subsequent step A20, it is determined whether or not the total number of vibration data S stored in the vibration data storage unit 11 is equal to or larger than a predetermined number set in advance. That is, in this step, it is determined whether or not the number of data to be signal processed is sufficient. If the total number of vibration data S is greater than or equal to the predetermined number, the process proceeds to step A30. If the total number of vibration data S is less than the predetermined number, the process returns to step A10. Thus, steps A10 to A20 are repeatedly executed until the number of data is sufficient.

In step A30, the amplitude energy conversion unit 6, the vibration data S of chronologically input from the vibration data storage unit 11 is converted into an absolute value vibration data S _A. That is, in this step, the magnitude of the amplitude energy inherent in the vibration data S is extracted.
In subsequent step A40, the absolute value vibration data S _A converted in step A30 is further converted into envelope vibration data S _E. That is, here, as indicated by a thick solid line in FIG. 2A, time-series data enveloping the amplitude of the absolute value vibration data S _A is obtained.

In step A50, the basic vibration data generation unit 2 performs three types of filter processing on the time series data of the envelope vibration data S _E, a low-pass filter, a phase filter, and a high-pass filter. By these filter processing, vibration components of high frequency and low frequency are reduced centering on acceleration fluctuation accompanying the walking movement of the cow, and the feature of walking time interval is extracted as a wave having a corresponding period. This wave is the basic vibration data S _B indicated by a broken line in FIG. Note that the time at which the magnitude of the basic vibration data S _B becomes 0 by the phase filter processing corresponds to the time at the peak position of the time-series vibration data S.

In subsequent step A60, the basic data point extraction unit 3 extracts the basic data point D _B based on the basic vibration data S _B. Here, the envelope vibration data S _E corresponding to the zero point of the basic vibration data S _B whose signal value changes from negative to positive before and after is extracted as the basic data point D _B. As shown in FIG. 2 (a), between the actual vibration data peak positions and the basic data points D _B of S, it can be seen that some errors occurs.

In further successive step A70, the extraction range setting unit 4, near the time of the time when the basic data points D _B is detected is set as the extraction range A. Extraction range A is across the time at which the basic data points D _B is detected, it is a previous predetermined time t ₁ and thereafter the width of the predetermined time t ₂ and time which is the sum of. Since each of the predetermined times t ₁ and t ₂ is set to be larger than the time error estimated to be caused by the filter processing, the peak of the actual vibration data S is located in the extraction range A. That is, as shown in FIG. 2 (b), the extraction range A is provided with a can absorb certain errors occurring between the actual peak location and the basic data points D _B of the vibration data S width .

In step A80, the feature data point extraction unit 5, the maximum value of the vibration data S contained in the extraction ranges A are extracted as feature data points D _C, the process proceeds to step A90. As shown in FIG. 2B, the feature data point D _C is the maximum value of the envelope vibration data S _E , that is, the maximum value of the actual vibration data S, and is an accurate signal that characterizes the walking state.
Then, in the subsequent step A90, these feature data points D _C are analyzed in the second data processing unit 10, and this flow ends.

A description of the specific analysis flow and determination in the second data processing unit 10 and the output flow thereof will be omitted, but the scaling index α of the fluctuation of the walking interval time is calculated for the time series feature data point D _C , It is determined whether or not the walking state is good with reference to the state where α = 1. Further, such a determination result is output to the monitor 15. Incidentally, when the analysis is completed once the feature data points D _C, the total number of vibration data S according to the determination in step A20 is reset.

[3. effect]
[3-1. Effects of two signal processing processes]
As described above, according to the data processing device of the present embodiment, the vibration data S input from the vibration data detection device 1 is processed in each of the two signal processing steps. One is the amplitude energy conversion unit 6, a signal processing for setting the extraction range A in the fundamental vibration data generation unit 2 and the basic data extracting unit 3, the other is the feature data points D _C in the feature data point extraction unit 5 It is signal processing for extraction.

In the former signal processing process, pre-processing is performed on the vibration data S, but the extraction range A set as a result has a width that can tolerate the error, and therefore the influence of the error due to the pre-processing. Can be offset. Further, since there have been extraction process directly feature data points D _C without being subjected to the pretreatment in the latter signal process, it can be taken out accurate feature data points D _C.

In other words, in general, signal processing as preprocessing makes it easy to grasp the characteristics of vibration data, but characteristic information “missing” occurs. However, according to this data processing apparatus, with reference to the result of the process, to extract the feature data points D _C from the original vibration data without "missing", it is possible to improve the reliability of data processing.
Moreover, when extracting the feature data points D _C in the feature data point extraction unit 5, the extraction range vibration data S contained in A only (envelope vibration data S _E) is referred to, the vibration data S other than extraction region A is excluded Therefore, it is possible to easily extract only the information of the part that is considered to have a strong correlation with the vibration data S characterizing the cow's walking movement, and to secure a sufficient data processing speed.

Further, the feature data point extraction unit 5 is adapted to extract as the characteristic data points D _C the maximum value of the vibration data S contained in the extraction ranges A. With this configuration, it is possible to accurately extract feature data points even from vibration data S during walking of a quadruped animal that is noisy and feature extraction is difficult, and the walking rhythm can be accurately determined. I can grasp it.

[3-2. Effect of amplitude energy conversion]
Further, by taking the absolute value of the vibration data S in the amplitude energy conversion unit 6, it is possible to accurately grasp the magnitude of the amplitude energy of the vibration generated with the walking motion.
For example, when collecting vibration data focusing on the walking of any one limb during walking of a quadruped animal, the vibration data is partially amplified under the influence of vibration caused by the movement of other limbs. Alternatively, it is easy to cancel, but the information on the amplitude energy inherent in the information on acceleration is not easily influenced by vibrations of other limbs. That is, by extracting the magnitude of the amplitude energy, it is possible to easily grasp the walking rhythm of only the limb of interest, and it is possible to improve the calculation accuracy of the subsequent cycle analysis processing.

In addition, by extracting the envelope of the absolute value vibration data S _{A in} the amplitude energy conversion unit 6, it is possible to grasp the outline of the fluctuation of the amplitude energy, and to easily extract the basic data points characterizing the fluctuation of the vibration data S. can do. Note that the entire information amount can be compressed (reduced) by the envelope process without losing information on the peak portion of the vibration data S.

Here, as a comparative example with the data processing apparatus of the present invention, the amplitude energy conversion in the amplitude energy conversion unit 6 (that is, the conversion of the vibration data S into the absolute value vibration data S _A and the envelope vibration data S _E ) is not performed. In this case, the basic vibration data S _B ′ and the basic data point D _B ′ are shown graphically in FIG. This is a diagram showing a calculation result when time-series vibration data S output from the vibration data storage unit 11 is directly input to the basic vibration data generation unit 2 instead of the envelope vibration data S _E.

As shown in FIG. 5, since the amplitude energy conversion processing of the vibration data S is not performed on the basic vibration data S _B ′, the fluctuation period thereof is the same as that of the basic vibration data S _B shown in FIG. Compared to about half. In other words, this indicates that the vibration of the limb body on which the vibration data detecting device 1 is worn is mixed with the vibration of the limb body other than the limb body, so that an accurate walking rhythm cannot be grasped. Further, the waveform itself of the basic vibration data S _B ′ has an unstable shape, and a portion where the zero point cannot be extracted is also seen.

Thus, when the amplitude energy conversion is not performed, it is difficult to extract an accurate zero point corresponding to the peak position of the vibration data S, and accordingly, the accuracy of the basic data point D _B ′ is also lowered. As a result, it can be seen that it is difficult to extract accurate feature data points D _C.

[3-3. Other effects]
The filter processing as preprocessing in the fundamental vibration data generation unit 2 is general signal processing such as a low-pass filter, a phase filter, and a high-pass filter, and is easy to implement and can be processed in a short time. . On the other hand, the calculation processing in the feature data point extraction unit 5 does not require a complicated calculation and can quickly obtain a result. In particular, since the processing content in the first data processing unit 9 of the present embodiment is configured by signal processing using a linear analysis technique, for example, the configuration is simple compared to signal processing using a nonlinear analysis technique. There is an advantage.

Further, the data processing apparatus is merely a configuration including a fundamental vibration data based on S _B not only extracts the basic data points D _B, the extraction range setting unit 4 and the feature data point extraction section 5. That is, as a conventional method for reducing the error generated by the filter processing, there is a method of finely adjusting the delay amount of the phase change in the phase filter. According to the present invention, accurate data is extracted. There thus after there is no need for such fine-tuning, is actually a check of the basic data points D _B obtained by performing a pre-processing is not required. Therefore, it is possible to completely automate the entire signal processing process from the detection of the vibration data S and the input of the vibration data S to the first data processing unit 9 to the output of the result in the second data processing unit 10, The signal processing effort can be significantly reduced.

Furthermore, the data processing apparatus of this embodiment, after the data processing in the first data processing unit 9, a nonlinear analysis method for observing the fluctuations of the peak interval time of the characteristic data points D _C is used. As described above, although could be rather difficult to grasp the non-linear structure once to remove noise if pretreatment, feature data points D _C inputted to the second data processing unit 10 in the present data processing apparatus, the noise Since the information is extracted from raw information that has not been removed, characteristic information is not lost and the nonlinear structure can be accurately grasped. That is, it is possible to perform nonlinear analysis by extracting accurate feature information without discarding information that looks like noise at first glance.
Thus, according to the present data processing device, the data processing speed and the data processing accuracy can be improved with a simple configuration.

[4. Others]
As mentioned above, although embodiment of this invention was described, this invention is not limited to the said embodiment, A various deformation | transformation is possible in the range which does not deviate from the meaning of this invention.
For example, in the above-described embodiment, the data processing functions in the first data processing unit 9 and the second data processing unit 10 are configured as programs, but means for realizing these functions is not limited thereto. For example, each of the first data processing unit 9 and the second data processing unit 10 may be configured as a one-chip microcomputer incorporating a ROM, a RAM, a CPU, or the like, or formed as an electronic circuit such as a digital circuit or an analog circuit. May be.

  As described above, the data processing apparatus of the present invention can automate the signal processing process from the detection of the vibration data S to the output of the result. When the data processing device according to the present invention is configured, a small display device having the same function as the monitor 15 according to the present invention and a microsensor having the same function as the vibration data detection device 1 are mounted, so It is also possible to manufacture a small-sized processing apparatus.

In the above-described embodiment, the amplitude energy conversion unit 6 converts the vibration data S into the absolute value vibration data S _A and performs envelope processing to extract the magnitude of the amplitude energy. . However, instead of such a configuration, information that correlates with the amplitude energy may be extracted by squaring the vibration data S. Alternatively, a configuration in which envelope processing is performed on the square-transformed data may be considered. As described above, any data conversion method can be used as long as a parameter that correlates with the amplitude energy of vibration is extracted.

In the above-described embodiment, an acceleration sensor for detecting an acceleration signal is applied as the signal detection unit 1. However, as a signal to be calculated by the data processing apparatus, it relates to the state of various objects. Various parameters are possible.
In the above-described embodiment, the vibration data S detected by the vibration data detection device 1 is directly input to the first data processing unit 9, but the vibration data detection device 1 and the first data processing unit 9 may be separated. For example, the time series data of the vibration data S detected by the vibration data detection device 1 is stored in some storage medium, and the time series data is sent to the first data processing unit 9 when an arithmetic process is required. It is possible to input. In this case, the time series data may be input to the vibration data storage unit 11 or may be input to the amplitude energy conversion unit 6 and the feature data point extraction unit 5 without using the vibration data storage unit 11. Good.

  In the above-described embodiment, the basic vibration data generation unit 2 performs three types of filter processing: a low-pass filter, a phase filter, and a high-pass filter. However, a band-pass filter or a notch filter may be used in combination. Note that the preprocessing in the basic vibration data generation unit 2 refers to general irreversible (with irreversible changes) arithmetic processing for making it easy to find parameter variations. In other words, the specific processing content need not be the filter processing as long as it is an arithmetic processing for making it easy to find parameter variations. For example, it is conceivable to use Hilbert transform processing, envelope processing, Fourier transform processing, signal processing using an averaging method, wavelet analysis processing, fractal analysis processing, and the like. Also included are arbitrary signal addition and subtraction, proportional processing, integration processing, differentiation processing, and the like.

When the first data processing unit 9 and the second data processing unit 10 are configured using electronic circuits such as a digital circuit and an analog circuit, an analog circuit is used instead of the digital filter described in the above embodiment. Apply a filter. That is, the data processing performed in the first data processing unit 9 may be analog signal processing.
In the above-described embodiment, the basic data extracting unit 3, the vibration data S corresponding to the zero point of the fundamental vibration data S _B is adapted to be extracted, such extraction object, the present data processing It can set suitably according to the calculation object in an apparatus. Position and width of the extraction range A in the extraction range setting unit 4, and is the same for the position for taking out the characteristic data points D _C in the feature data point extraction section 5.

Moreover, in the above-mentioned embodiment, although the method of a detrend fluctuation | variation analysis is used in the analysis part 7, an analysis method is not limited to this. In consideration of the general effect of the preprocessing on the analysis result, when the analysis method in the analysis unit 7 is a nonlinear analysis method, the analysis result is more accurate than the case of the linear analysis method. Can be expected.
In the above-described embodiment, the walking rhythm during the walking movement of the cow is an analysis target. However, it is possible to analyze the walking rhythm of an arbitrary quadruped animal such as a horse, sheep, or pig.

1 is a block diagram showing an overall configuration of a data processing apparatus according to an embodiment of the present invention. It is a graph for demonstrating the data processing content in this data processor, (a) is a time series graph of the vibration data which concerns on the data processing in an amplitude energy conversion part, a basic vibration data generation part, and a basic data extraction part, (b) ) Is a time-series graph of vibration data related to data processing in the feature data point extraction unit. It is a flowchart which shows the control content in this data processor. It is a schematic diagram which shows the structural example of this data processing apparatus using a computer. It is a time series graph which shows the analysis process of the vibration data as a comparative example.

Explanation of symbols

1 Vibration data detection device (vibration data detection means)
2 Basic vibration data generation unit (basic vibration data generation means, linear signal processing means, filter processing means)
3 Basic data point extraction unit (Basic data point extraction means)
4 Extraction range setting section (extraction range setting means)
5. Feature data point extraction unit (feature data point extraction means)
6 Amplitude energy conversion unit (amplitude energy conversion means)
7 Analysis Unit 8 Determination Unit 9 First Data Processing Unit 10 Second Data Processing Unit (Nonlinear Analysis Calculation Processing Unit)
11 Vibration Data Storage Unit 12 Computer 13 Storage Device 14 Central Processing Unit (CPU)
15 Monitor

Claims (17)

A vibration data detection step for detecting vibrations caused by the walking movement of a quadruped animal as time-series vibration data;
An amplitude energy conversion step for performing data conversion to extract the magnitude of amplitude energy in the vibration data detected in the vibration data detection step;
A basic vibration data generation step for generating time-series basic vibration data by performing signal processing as preprocessing for grasping the characteristics of the vibration with respect to the data obtained in the amplitude energy conversion step;
A basic data point extraction step for extracting basic data points characterizing fluctuations in the vibration data based on the basic vibration data generated in the basic vibration data generation step;
An extraction range setting step for setting a predetermined time range based on the detection time of the basic data point extracted in the basic data point extraction step as the extraction range of the vibration data;
A feature of extracting vibration data characterizing the vibration as a feature data point from vibration data included in the extraction range set in the extraction range setting step among the vibration data detected in the vibration data detection step. A data processing method comprising a data point extraction step. In the amplitude energy conversion step, the vibration data obtained in the vibration data detection step is converted into absolute value vibration data composed of the absolute value, and the absolute value vibration data is further subjected to envelope processing to obtain the absolute value. Convert vibration data to envelope vibration data
The data processing method according to claim 1, wherein: The fundamental vibration data generation step generates the basic vibration data by performing the signal processing on the data converted in the amplitude energy conversion step using a linear analysis method. 3. The data processing method according to 1 or 2. The predetermined frequency component set in advance is filtered from the data converted in the amplitude energy conversion step in the basic vibration data generation step. Data processing method. In the basic vibration data generation step, wave data obtained by smoothing the data converted by the amplitude energy conversion means is generated as the basic vibration data.
5. The fundamental data point extracting step, wherein a point on the fundamental vibration data that correlates with a wavelength of the fundamental vibration data is extracted as the fundamental data point. Data processing method. 6. The data processing method according to claim 5, wherein in the extraction range setting step, a time near the detection time of the basic data point is set as the extraction range. 7. The data processing method according to claim 6, wherein, in the feature data point extraction step, a vibration peak detection time in vibration data included in the extraction range is extracted as the feature data point. 8. The method according to claim 1, further comprising a nonlinear analysis calculation processing step for extracting a nonlinear structure in the vibration data based on the feature data point detected in the feature data point extraction step. The data processing method according to item 1. Vibration data detecting means for detecting vibrations caused by a quadruped walking movement as time-series vibration data;
Amplitude energy conversion means for performing data conversion to extract the magnitude of amplitude energy in the vibration data detected by the vibration data detection means;
Basic vibration data generating means for generating time-series basic vibration data by performing signal processing as preprocessing for grasping the characteristics of the vibration on the data obtained by the amplitude energy conversion means;
Basic data point extracting means for extracting basic data points characterizing fluctuations in the vibration data based on the basic vibration data generated by the basic vibration data generating means;
Extraction range setting means for setting a predetermined time range based on the detection time of the basic data points extracted by the basic data point extraction means as the extraction range of the vibration data;
A feature of extracting vibration data characterizing the vibration as a feature data point from vibration data included in the extraction range set by the extraction range setting unit among the vibration data detected by the vibration data detection unit. A data processing apparatus comprising a data point extracting means. The amplitude energy conversion means converts the vibration data obtained by the vibration data detection means into absolute value vibration data consisting of the absolute value, and further performs envelope processing on the absolute value vibration data to obtain the absolute value. The data processing apparatus according to claim 9, wherein vibration data is converted into envelope vibration data. The fundamental vibration data generating means has linear signal processing means for generating the fundamental vibration data by performing the signal processing on the envelope vibration data converted by the amplitude energy converting means using a linear analysis method. The data processing apparatus according to claim 9 or 10, characterized by comprising: The fundamental vibration data generating means includes filter processing means for filtering a predetermined frequency component set in advance from the envelope vibration data converted by the amplitude energy converting means. The data processing device according to any one of 9 to 11. The basic vibration data generation means generates wave data obtained by smoothing the data converted by the amplitude energy conversion means as the basic vibration data,
The basic data point extracting means extracts a point on the basic vibration data correlated with the wavelength of the basic vibration data as the basic data point. Data processing equipment. 14. The data processing apparatus according to claim 13, wherein the extraction range setting means sets a time near the detection time of the basic data point as the extraction range. 15. The data processing apparatus according to claim 14, wherein the feature data point extracting unit extracts a vibration peak detection time in vibration data included in the extraction range as the feature data point. The nonlinear analysis calculation processing means for extracting a nonlinear structure in the vibration data based on the feature data points detected by the feature data point extraction means, further comprising: The data processing apparatus according to item 1. A data processing program for causing a computer to function as amplitude energy conversion means, basic vibration data generation means, basic data point extraction means, extraction range setting means, and feature data point extraction means,
The amplitude energy converting means performs data conversion to extract the magnitude of the amplitude energy in the vibration caused by the walking movement of the quadruped animal detected as vibration data,
The basic vibration data generation means generates time-series basic vibration data by performing signal processing as preprocessing for grasping the characteristics of the vibration on the data converted by the amplitude energy conversion means,
The basic data point extracting means extracts basic data points characterizing fluctuations in the vibration data based on the basic vibration data generated by the basic vibration data generating means,
The extraction range setting means sets a predetermined time range based on the detection time of the basic data points extracted by the basic data point extraction means as the vibration data extraction range,
Vibration data that characterizes the vibration from vibration data included in the extraction range set by the extraction range setting means among the vibration data detected by the vibration data detection means. A data processing program characterized by extracting as feature data points.

Images