JP7092624B2 - Behavior identification device, behavior identification method and program - Google Patents

Description

The present invention relates to a behavior identification device, a behavior identification method and a program.

The milking period of dairy cows is after mating (artificial insemination) of dairy cows to give birth to calves (postpartum), and this period is called the lactation period. It is known that the lactation period is about 305 days. In addition, it is common to stop milking about 60 days before the next expected delivery date in preparation for the next delivery. This 60-day period is called the dry milk period. Dairy cows repeat calving, lactation, and dryness every year.

Dairy farmers who raise dairy cows need to mate (artificial insemination) dairy cows every year during the dry season because they continuously milk from dairy cows. When mating dairy cows (artificial insemination), it is necessary to discover the estrus of the dairy cows. To detect estrus in dairy cows, it is necessary to check the external signs of dairy cows and to confirm the behaviors that are often seen during estrus. In addition, the breeding record using a ledger is used to manage the estrus period.

Abe Ryo, "Agricultural Science Seminar Basics of Livestock Breeding", New Edition, Agricultural Fisheries Village Cultural Association, April 2008, p.109, p.122-124.

Here, for example, in a large-scale dairy farm where the number of cows raised is several thousand or more, it is a heavy burden on the dairy farmer to confirm the external signs and behavior of each dairy cow. .. On the other hand, for example, if it is possible to easily confirm the behavior that is often seen during estrus by identifying the behavior of the dairy cow that is being raised, the burden on the dairy farmer can be reduced.

In addition, identifying the behavior of dairy cows leads not only to the discovery of estrus but also to the detection of abnormal behavior due to illness, injury, etc., and can contribute to the health management of dairy cows.

One embodiment of the present invention has been made in view of the above points, and an object thereof is to specify the behavior of livestock.

In order to solve the above problems, one embodiment of the present invention is a behavior specifying device for specifying the behavior of a livestock, and the acceleration data measured by an acceleration sensor mounted on the livestock and the pressure pressure mounted on the livestock. A storage means for storing pressure data measured by a sensor, an index value calculation means for calculating a predetermined index value from one or more acceleration data during a predetermined time, the index value, and the pre-created said. When the first specific means for specifying the behavior of the livestock and the behavior specified by the first specific means are predetermined behaviors based on the specific model for specifying the behavior of the livestock, the predetermined behavior is specified. It is characterized by having a second specifying means for identifying the behavior of the livestock based on one or more of the pressure data and one or more of the acceleration data over time.

The behavior of livestock can be identified.

It is a figure which shows an example of the whole structure of the behavior specifying system which concerns on 1st Embodiment. It is a figure for demonstrating an example of the behavior identification of a cow. It is a figure which shows an example of the measurement data stored in the measurement data storage part. It is a figure which shows an example of a behavior specific model. It is a figure which shows an example of the functional structure of the action specifying processing unit which concerns on 1st Embodiment. It is a flowchart which shows an example of the action specifying process by the motion intensity analysis which concerns on 1st Embodiment. It is a flowchart which shows an example of the action specifying process by the corrected atmospheric pressure analysis which concerns on 1st Embodiment. It is a flowchart which shows an example of the calculation process of the corrected atmospheric pressure difference minimum value which concerns on 1st Embodiment. It is a figure for demonstrating the relationship between a tag orientation and a twist. It is a figure which shows an example of the functional structure of the action specifying processing part which concerns on the 2nd Embodiment. It is a flowchart which shows an example of the calculation process of the corrected atmospheric pressure difference minimum value which concerns on 2nd Embodiment. It is a flowchart which shows an example of the calculation process for a tag which concerns on 2nd Embodiment.

Hereinafter, embodiments of the present invention (hereinafter, also referred to as “embodiments”) will be described with reference to the drawings. In each of the following embodiments, the behavior identification system 1 for specifying the behavior of cattle will be described as an example of livestock. Livestock are not limited to cattle.

[First Embodiment]
First, the overall configuration of the behavior specifying system 1 according to the present embodiment will be described with reference to FIG. FIG. 1 is a diagram showing an example of the overall configuration of the behavior specifying system 1 according to the first embodiment.

As shown in FIG. 1, the behavior identification system 1 according to the present embodiment measures a behavior identification device 10 for identifying a cow's behavior, one or more tags 20 attached to the cow, and a reference atmospheric pressure. A reference barometric pressure sensor 30 is included. The tag 20 is preferably attached by being fixed to the neck portion of the cow with a belt or the like.

The tag 20 is a device attached to a cow. One tag 20 is attached to one cow. The tag 20 includes an acceleration sensor that measures the acceleration of the cow wearing the tag 20 (acceleration of three axes of x-axis, y-axis, and z-axis). Here, the coordinate system composed of the x-axis, the y-axis, and the z-axis is fixed to the tag 20, and the tag 20 is, for example, in a state where the cow stands up and the neck is not bent (straight). The cow is mounted so that the right direction is the positive direction of the x-axis, the traveling direction is the positive direction of the y-axis, and the gravity direction is the positive direction of the z-axis with respect to the traveling direction of the cow.

The tag 20 transmits measurement data including the acceleration sensor value measured by the acceleration sensor to the action specifying device 10 at predetermined time intervals (for example, every 2 seconds). The measurement data transmitted to the action specifying device 10 is stored (stored) in the measurement data storage unit 200 described later.

The reference atmospheric pressure sensor 30 is installed at a predetermined position in the barn (for example, on the ground in the barn) and measures the reference atmospheric pressure. The reference atmospheric pressure sensor 30 transmits reference atmospheric pressure data including a reference atmospheric pressure sensor value indicating the measured atmospheric pressure to the action specifying device 10 at predetermined time intervals (for example, every 2 seconds). The reference atmospheric pressure data transmitted to the action specifying device 10 is stored (stored) in the reference atmospheric pressure data storage unit 400, which will be described later.

The behavior identification device 10 is one or more computers that identify the behavior of a cow. The behavior specifying device 10 includes a behavior specifying processing unit 100, a measurement data storage unit 200, a behavior specifying model 300, a reference barometric pressure data storage unit 400, and a behavior data storage unit 500.

The behavior specifying processing unit 100 identifies the behavior of the cow by motion intensity analysis based on the measurement data stored in the measurement data storage unit 200 and the behavior specifying model 300. Further, when a predetermined action is specified by the action intensity analysis, the action specifying processing unit 100 has the measurement data stored in the measurement data storage unit 200 and the reference pressure data stored in the reference pressure data storage unit 400. And, based on the behavior data stored in the behavior data storage unit 500, more accurate behavior of the cow is specified by the corrected pressure pressure analysis. The action specifying processing unit 100 is realized by a process of causing a CPU (Central Processing Unit) or the like to execute one or more programs installed in the action specifying device 10.

The measurement data storage unit 200 stores the measurement data received from the tag 20. The measurement data storage unit 200 can be realized by using, for example, an auxiliary storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). The measurement data storage unit 200 stores a plurality of measurement data at predetermined time intervals (for example, every 2 seconds).

The behavior identification model 300 is a model for identifying behavior from the acceleration sensor values included in the measurement data. The behavior identification model 300 is created in advance by a machine learning method such as SVM (Support Vector Machine) according to, for example, the type of livestock and the type of behavior to be specified. The behavior specific model 300 is stored in, for example, an auxiliary storage device such as an HDD or SSD.

The reference atmospheric pressure data storage unit 400 stores the reference atmospheric pressure data received from the reference atmospheric pressure sensor 30. The reference atmospheric pressure data storage unit 400 can be realized by using, for example, an auxiliary storage device such as an HDD or SSD. The reference atmospheric pressure data storage unit 400 stores a plurality of reference atmospheric pressure data at predetermined time intervals (for example, every 2 seconds).

The behavior data storage unit 500 stores behavior data indicating the behavior specified by the behavior identification processing unit 100. The action data storage unit 500 can be realized by using, for example, an auxiliary storage device such as an HDD or SSD. The behavior data storage unit 500 stores behavior data indicating the behavior specified by the behavior identification processing unit 100 at predetermined time intervals (for example, every 10 minutes).

The configuration of the behavior specifying system 1 shown in FIG. 1 is an example, and may be another configuration. For example, the behavior identification device 10 may be composed of a plurality of computers. Further, for example, a device (cloud server or the like) connected to the action specifying device 10 via a network may have all or a part of the functions of the action specifying processing unit 100. Further, for example, instead of the tag 20, an acceleration sensor and a barometric pressure sensor may be attached to the cow separately.

<Identification of cow behavior>
Here, the behavior of the cow specified by the behavior specifying system 1 according to the present embodiment and the outline of the specifying method will be described with reference to FIG. 2. FIG. 2 is a diagram for explaining an example of specifying the behavior of a cow.

First, it is assumed that there are four behaviors of the cow specified by the behavior specifying system 1 according to the present embodiment, in ascending order of movement, "standing", "lying", "ruminant", and "activity". Therefore, in the present embodiment, it is specified which of the four behaviors of "standing", "lying", "ruminant" and "activity" is performed by the cow. The magnitudes of the "standing" and "lying" movements may be reversed. That is, they may be "lying", "standing", "ruminant", and "activity" in ascending order of movement.

"Standing up" is an action that indicates the state in which a cow is standing. "Lying" is an action that indicates the state in which a cow is lying down. In addition, both "standing" and "lying" are states in which the cow is not moving (in other words, the state in which the cow is standing and standing still or the state in which the cow is lying and standing still).

"Ruminant" is an action that indicates a state in which a cow is ruminating (the action of returning food once swallowed to the mouth and chewing it again). "Activity" is an action that indicates a state in which a cow is performing a predetermined action. For example, a state in which a cow is walking, a state in which a cow is eating food, or a state in which a cow is drinking water. It is an action that indicates such things.

Which of the behaviors of "standing", "lying", "ruminant" and "activity" is performed by the cow is specified by the motion intensity analysis. In the motion intensity analysis, whether the behavior of the cow is "standing", "lying", "ruminant" or "activity" based on the acceleration sensor value included in the measurement data and the behavior identification model 300. Identify. More specifically, in the motion intensity analysis, the maximum value of the fluctuation width of the L2 norm of the accelerometer value during the predetermined time (for example, 10 minutes) to the predetermined index value (accelerometer value during the predetermined time). And the standard deviation of the L2 norm of the accelerometer value). The maximum value of the swing width of the L2 norm is the difference between the average value of the L2 norm of the acceleration sensor value during the predetermined time and the value of the L2 norm of each acceleration sensor value during the predetermined time. Is the value that maximizes the absolute value of.

Then, in the behavior specific model 300 classified into the four areas of "standing", "lying", "ruminant", and "activity", the point indicating the calculated index value is included in any of these four areas. Identify the behavior.

In addition, when the motion intensity analysis identifies "standing", "lying" or "ruminant", the corrected barometric pressure analysis identifies more accurate behavior. This is because the motion intensity analysis may not be able to accurately identify the posture of the cow during the behavior identified by using the behavior identification model 300.

In the corrected barometric pressure analysis, the first correction is performed using the acceleration sensor value and the barometric pressure sensor value included in each measurement data for a predetermined time (for example, 20 minutes) and the reference barometric pressure sensor value included in each reference barometric pressure data. After calculating the atmospheric pressure difference minimum value and the second corrected atmospheric pressure difference minimum value, the behavior of the cow is specified based on the corrected atmospheric pressure difference minimum value and a predetermined threshold value. The corrected barometric pressure difference minimum value is the minimum value (for example, the minimum value in 10 minutes) of the corrected barometric pressure difference value obtained by correcting the difference between the reference barometric pressure sensor value and the barometric pressure sensor value with the acceleration sensor value.

More specifically, for example, when the action 10 minutes ago (that is, the previous action) is "lying", the first corrected barometric pressure difference minimum value between 10 minutes ago and the present is It is determined whether or not the value is reduced by a predetermined threshold value or more from the second minimum corrected atmospheric pressure difference between 20 minutes before and 10 minutes before. Then, when the first corrected atmospheric pressure difference minimum value decreases by a predetermined threshold value or more from the second corrected atmospheric pressure difference minimum value, it is specified as "standing". On the other hand, when the first corrected atmospheric pressure difference minimum value does not decrease by a predetermined threshold value or more from the second corrected atmospheric pressure difference minimum value, it is specified that the behavior is the same as the previous one (that is, "lying").

Also, for example, when the action 10 minutes ago (that is, the previous action) is other than "lying" (that is, when it is "standing" or "ruminant"), the first corrected barometric pressure difference is minimized. It is determined whether or not the value has increased by a predetermined threshold value or more from the second minimum corrected atmospheric pressure difference. Then, when the first corrected atmospheric pressure difference minimum value increases by a predetermined threshold value or more from the second corrected atmospheric pressure difference minimum value, it is specified as “lying”. On the other hand, when the first corrected atmospheric pressure difference minimum value does not increase by a predetermined threshold value or more from the second corrected atmospheric pressure difference minimum value, the behavior is the same as the previous one (that is, "standing" or "ruminant"). To specify.

As described above, the behavior specifying system 1 according to the present embodiment has four behaviors of "standing", "lying", "ruminant" and "activity" from the acceleration sensor value received from the tag 20 attached to the cow. To identify. In addition, the behavior identification system 1 according to the present embodiment has an acceleration sensor value, a barometric pressure sensor value, and a reference barometric pressure sensor when the behavior of the cow is identified as "standing", "lying down", or "rebelling". Identify more accurate behavior from the value.

Thereby, it is possible to identify whether the behavior of the cow is "standing", "lying", "ruminant" or "activity" by the motion intensity analysis. In addition, when the motion intensity analysis identifies "standing", "lying" or "ruminant", the corrected barometric pressure analysis shows that the cow's behavior is "standing", "lying" or "previous behavior". Which one can be specified with higher accuracy.

Then, by displaying the behavior of these identified cows on, for example, a display device (display) of the behavior specifying device 10, it becomes possible to easily confirm the behavior of the cow. For this reason, for example, it is possible to detect behaviors that are often seen during estrus of cattle and abnormal behaviors due to illness or injury at an early stage, and it is possible to easily manage the estrus period and health management of cattle. become.

<Measurement data stored in the measurement data storage unit 200>
Next, the measurement data stored in the measurement data storage unit 200 according to the present embodiment will be described with reference to FIG. FIG. 3 is a diagram showing an example of measurement data stored in the measurement data storage unit 200. When the action specifying device 10 receives the measurement data from the tag 20, the action specifying processing unit 100 may store (store) the received measurement data in the measurement data storage unit 200.

As shown in FIG. 3, the measurement data storage unit 200 stores one or more measurement data for each tag ID that identifies the tag. Since one tag 20 is attached to one cow, the tag ID may be information for identifying the cow (cattle individual identification information, etc.).

Each measurement data includes a date and time, an acceleration sensor value, and a barometric pressure sensor value. The date and time is, for example, the date and time when the tag 20 transmits the measurement data. The date and time may be the date and time when the action specifying device 10 receives the measurement data.

The acceleration sensor value is a value of acceleration measured by the acceleration sensor included in the tag 20. The acceleration sensor value includes an x component indicating an acceleration component in the x-axis direction, a y component indicating an acceleration component in the y-axis direction, and a z component indicating an acceleration component in the z-axis direction. For example, the measurement data of the date and time “t ₁ ” includes the x component “x ₁ ” of the acceleration sensor value, the y component “y ₁ ”, and the z component “z ₁ ”. Similarly, for example, the measurement data of the date and time “t ₂ ” includes the x component “x ₂ ” of the acceleration sensor value, the y component “y ₂ ”, and the z component “z ₂ ”. In the following, the x component of the acceleration sensor value will be referred to as the "x component acceleration sensor value", the y component of the acceleration sensor value will be referred to as the "y component acceleration sensor value", and the z component of the acceleration sensor value will be referred to as the "z component acceleration sensor value". show.

The barometric pressure sensor value is a barometric pressure value measured by the barometric pressure sensor included in the tag 20. The barometric pressure sensor value "p ₁ " is included in the measurement data of the date and time "t ₁ ". Similarly, for example, the measurement data at the date and time “t ₂ ” includes the barometric pressure sensor value “p ₂ ”.

For example, when each tag 20 transmits measurement data to the action specifying device 10 every 2 seconds, if Δt = ti _{+ 1} − ti ( _i is an integer of 1 or more), then Δt = 2 [seconds]. be.

As described above, the measurement data stored in the measurement data storage unit 200 includes the date and time, the acceleration sensor value, and the barometric pressure sensor value for each tag ID.

<Behavior specific model 300>
Next, the behavior identification model 300 for identifying the behavior from the acceleration sensor values included in the measurement data will be described with reference to FIG. FIG. 4 is a diagram showing an example of the behavior specific model 300.

As shown in FIG. 4, in the behavior specific model 300, the maximum value of the fluctuation width of the L2 norm of the acceleration sensor value during a predetermined time (for example, 10 minutes) is the horizontal axis, and the standard deviation of the L2 norm is the vertical axis. When this is done, it is represented as a relationship graph showing the relationship between the maximum value and the standard deviation value of these fluctuation widths and the behavior.

For example, if the region D1 includes the maximum swing width of the L2 norm of the accelerometer value in a certain 10 minutes and the standard deviation, the behavior of the cow in the 10 minutes is specified as "standing up". Further, for example, when the maximum value and the standard deviation of the swing width of the L2 norm of the acceleration sensor value in a certain 10 minutes are included in the region D2, the behavior of the cow in the 10 minutes is specified as “lying”.

Similarly, for example, if the region D3 includes the maximum value and the standard deviation of the swing width of the L2 norm of the accelerometer value in a certain 10 minutes, the behavior of the cow in the 10 minutes is specified as "ruminant". In the region D3 specified as "ruminant", further, when the ruminant operation is performed while the cow is standing (this case is referred to as "standing ruminant"), the ruminant is performed while the cow is lying down. It may be divided into the case of performing an operation (this case is referred to as "lying rumination").

Similarly, for example, when the maximum value and the standard deviation of the swing width of the L2 norm of the acceleration sensor value in a certain 10 minutes are included in the region D4, the behavior of the cow in the 10 minutes is specified as "activity". To.

In the example shown in FIG. 4, the cases where the regions D1 to D4 of the behavior specific model 300 do not overlap each other are shown, but the portion where two or more regions of the regions D1 to D4 overlap each other is shown. May exist.

In this way, the behavior specific model 300 is a model in which a region representing the relationship between the maximum value and standard deviation of the L2 norm of the acceleration sensor value during a predetermined time (for example, 10 minutes) and the behavior of the cow is defined. be. Such a behavior specific model 300 is created in advance by a machine learning method such as SVM. The SVM is an example, and may be created by various machine learning methods such as a neural network. Further, for example, even when the cow performs the same operation, the specific acceleration sensor value differs depending on the type of the acceleration sensor and the like. However, the positional relationship between the regions D1 to D4 indicating each action is substantially constant. Hereinafter, the maximum value of the swing width of the L2 norm is also referred to as “the maximum value of the L2 norm”.

<Functional configuration of behavior identification processing unit 100>
Next, the functional configuration of the behavior specifying processing unit 100 according to the present embodiment will be described with reference to FIG. FIG. 5 is a diagram showing an example of the functional configuration of the behavior specifying processing unit 100 according to the first embodiment.

As shown in FIG. 5, the action specifying processing unit 100 according to the present embodiment includes an acquisition unit 101, a preprocessing unit 102, an index value calculation unit 103, a correction difference minimum value calculation unit 104, and an action specification unit. 105 and are included.

The acquisition unit 101 acquires measurement data from the measurement data storage unit 200. At this time, the acquisition unit 101 acquires measurement data for a predetermined time (for example, 10 minutes) from the measurement data storage unit 200 for each tag ID, for example.

Further, the acquisition unit 101 acquires reference atmospheric pressure data from the reference atmospheric pressure data storage unit 400. At this time, the acquisition unit 101 acquires the reference atmospheric pressure data during the predetermined time.

Further, the acquisition unit 101 acquires the latest behavior data (that is, the behavior data indicating the behavior specified immediately before) among the behavior data stored in the behavior data storage unit 500.

The pre-processing unit 102 performs pre-processing on the measurement data acquired by the acquisition unit 101. The pre-processing includes, for example, defect complementation (resampling) processing of measurement data, noise removal processing, calculation processing of representative values for each predetermined interval (for example, average value for each predetermined interval), and the like.

The index value calculation unit 103 calculates the index value from the acceleration sensor value included in the measurement data after the preprocessing by the preprocessing unit 102. For example, the index value calculation unit 103 calculates the maximum value and standard deviation of the L2 norm of the acceleration sensor value included in the measurement data after the preprocessing as the index value.

The correction difference minimum value calculation unit 104 calculates the correction pressure difference minimum value from the barometric pressure sensor value and the acceleration sensor value included in the measurement data and the reference barometric pressure sensor value included in the reference barometric pressure data. For example, the correction difference minimum value calculation unit 104 is based on the barometric pressure sensor value and the acceleration sensor value included in the measurement data in 20 minutes and the reference barometric pressure data included in the reference barometric pressure data in the 20 minutes. The first corrected atmospheric pressure difference minimum value in the latter half 10 minutes and the second corrected atmospheric pressure difference minimum value in the first half 10 minutes are calculated. The first half and the second half are the first half and the second half when the dates and times are arranged in ascending order (oldest order).

The corrected atmospheric pressure difference minimum value is the minimum value of the corrected atmospheric pressure difference value during a predetermined time (for example, 10 minutes). The corrected barometric pressure difference value is the average value of the barometric pressure sensor values included in each measurement data at the predetermined time (for example, 10 minutes) at 1-minute intervals, and each of the reference barometric pressure data included in the 10-minute period. The barometric pressure difference, which is the difference between the reference barometric pressure sensor value and the average value at 1-minute intervals, is the average value of the acceleration sensor values of a predetermined component (for example, y component) included in each measurement data at 1-minute intervals. It is corrected by.

The action identification unit 105 has four actions (“standing”, “lying”, “rebellion”, and “activity” from the maximum value and standard deviation of the L2 norm calculated by the index value calculation unit 103 and the behavior identification model 300. ) Is specified (behavior is specified by motion intensity analysis).

Then, the action specifying unit 105 stores the action data indicating the specified action in the action data storage unit 500. Further, the action specifying unit 105 may display information indicating the specified action on a display device such as a display, or another device connected to the action specifying device 10 (for example, a PC, a smartphone, or a tablet). It may be output to a terminal, etc.).

The action specifying unit 105 may specify the above four actions from the maximum value and standard deviation of the L2 norm and the processing result of a program or the like that generates data equivalent to the action specifying model 300.

Further, when the "standing", "lying" or "rebellion" is specified above, the action specifying unit 105 further specifies the action indicated by the action data acquired by the acquisition unit 101 (that is, the action specified immediately before). More accurate behavior is specified from the first corrected atmospheric pressure difference minimum value and the second corrected atmospheric pressure difference minimum value calculated by the corrected differential minimum value calculation unit 104 according to the corrected behavior (corrected pressure analysis). Identify the action).

That is, when the previous action is "lying" and the first corrected atmospheric pressure difference minimum value decreases by a predetermined threshold value or more from the second corrected atmospheric pressure difference minimum value, the action specifying unit 105 "stands up". To specify. On the other hand, in the action specifying unit 105, when the previous action is "lying" and the first corrected pressure difference minimum value is not reduced by a predetermined threshold value or more from the second corrected pressure difference minimum value. Identify the same as the previous action.

Further, in the action specifying unit 105, when the previous action is other than "lying" and the first corrected atmospheric pressure difference minimum value increases by a predetermined threshold value or more from the second corrected atmospheric pressure difference minimum value, " Specify as "lying". On the other hand, in the action specifying unit 105, the previous action is other than "lying", and the first corrected pressure difference minimum value does not increase by a predetermined threshold value or more from the second corrected pressure difference minimum value. If so, specify that it is the same as the previous action.

The predetermined threshold value is, for example, a value preset by the user. Such a threshold is determined, for example, by an empirical rule regarding the behavior identification of cattle.

<Behavior identification process>
Next, the behavior specifying process according to the present embodiment will be described.

≪Behavior identification processing by motion intensity analysis≫
First, the behavior identification process by motion intensity analysis will be described with reference to FIG. FIG. 6 is a flowchart showing an example of the action specifying process by the motion intensity analysis according to the first embodiment. The action specifying process shown in FIG. 6 is repeatedly executed, for example, every 10 minutes. However, the action specifying process shown in FIG. 6 may be executed, for example, at a preset date and time, or may be executed at predetermined predetermined time intervals.

First, the acquisition unit 101 acquires measurement data for the past predetermined time (for example, the past 10 minutes) from the measurement data storage unit 200 for each tag ID (step S11). In this way, the acquisition unit 101 acquires measurement data in a predetermined time unit (for example, in units of 10 minutes) from the measurement data storage unit 200 for each cow (tag ID). It should be noted that such a predetermined time is not limited to 10 minutes, and can be set to any time by, for example, the user of the action specifying device 10.

Hereinafter, the description will be continued assuming that the measurement data s ₁ , s ₂ , ..., S _M for the past 10 minutes of the tag ID "Tag 1" have been acquired by the acquisition unit 101.

Next, the preprocessing unit 102 performs preprocessing on the measurement data s ₁ , s ₂ , ..., S _M acquired by the acquisition unit 101 (step S12). That is, the preprocessing unit 102 performs defect complementation (resampling) processing, noise removal processing, and the like for measurement data. The defect completion process is a process for complementing a defect when the data cannot be acquired in order to improve the accuracy of the measurement data. Further, the noise removal process is a process of removing data indicating a momentary motion of a cow (for example, a motion of momentarily shaking the body, a motion of momentarily causing the body to be greatly shaken, etc.).

As a result, the measurement data after the pretreatment are renumbered, and the measurement data s ₁ , s ₂ , ..., S _N are obtained. Here, when resampling is performed at 2-second intervals, N = 300.

Next, the index value calculation unit 103 determines the index value (maximum value of L2 norm and the maximum value of the L2 norm) from the acceleration sensor values included in the measurement data s ₁ , s ₂ , ..., S _N after the preprocessing by the preprocessing unit 102. The standard deviation) is calculated (step S13). That is, when the L2 norms of the accelerometer values included in the measurement data s ₁ , s ₂ , ..., S _N after the preprocessing are a ₁ , a ₂ , ... a _N , respectively, the index value calculation unit. 103 calculates the standard deviation σ and the maximum value m of these a ₁ , a ₂ , ... a _N. Here, the maximum value m is the maximum value of the difference between the average value of the L2 norms a ₁ , a ₂ , ... a _N and each L2 norm a _i (i = 1, ..., N). That is.

Next, the action identification unit 105 has four actions (“standing”, “lying”, “ruminant” and “reverse” from the standard deviation σ and the maximum value m calculated by the index value calculation unit 103 and the behavior identification model 300. "Activity") is specified (step S14). That is, the action specifying unit 105 identifies in which of the areas D1 to D4 on the behavior specifying model 300 the standard deviation σ and the maximum value m calculated by the index value calculating unit 103 are included. Identify the behavior.

Next, the action specifying unit 105 determines whether or not the action specified in step S14 is "standing", "lying", or "ruminant" (step S15).

When it is determined in step S15 that the specified action is "standing", "lying" or "ruminant", the action specifying processing unit 100 identifies a more accurate action by the corrected atmospheric pressure analysis (step S16). .. The details of the process of this step (behavior identification process by corrected atmospheric pressure analysis) will be described later.

On the other hand, when it is determined in step S15 that the specified action is not "standing", "lying" or "rebellion" (that is, when the specified action is "activity"), or following step S16, The behavior specifying unit 105 stores the behavior data indicating the specified behavior in the behavior data storage unit 500 (step S17). At this time, the behavior specifying unit 105 creates, for example, behavior data associated with the specified behavior and the current date and time, and then stores the created behavior data in the behavior data storage unit 500.

That is, when it is determined in step S15 that the specified action is not "standing", "lying" or "lying", the action specifying unit 105 acts on the action data in which the action "activity" is associated with the current date and time. It is stored in the data storage unit 500. Similarly, following step S16, the action specifying unit 105 stores the action data associated with the action specified in step S16 with the current date and time in the action data storage unit 500.

≪Behavior identification processing by corrected barometric pressure analysis≫
Next, the details of the process of step S16 (behavior specification process by corrected atmospheric pressure analysis) will be described with reference to FIG. 7. FIG. 7 is a flowchart showing an example of the action specifying process by the corrected atmospheric pressure analysis according to the first embodiment.

Here, as described above, "ruminant" is divided into "standing ruminant" and "lying ruminant". Therefore, in FIG. 7, "standing" includes "standing ruminant", and "lying" includes "lying ruminant".

First, the correction difference minimum value calculation unit 104 calculates the first correction pressure difference minimum value and the second correction pressure difference minimum value (step S21). Here, the first corrected atmospheric pressure difference minimum value is, for example, the corrected atmospheric pressure difference minimum value between the past 20 minutes ago and the past 10 minutes ago, and the second corrected atmospheric pressure difference minimum value is, for example, the past 10 minutes. It is the minimum corrected atmospheric pressure difference between minutes before (that is, between the past 10 minutes before and the present time). The details of the process of this step (calculation process of the minimum corrected atmospheric pressure difference) will be described later.

Next, the acquisition unit 101 acquires the latest behavior data (that is, the behavior data indicating the behavior specified immediately before) among the behavior data stored in the behavior data storage unit 500 (step S22). ..

Next, the action specifying unit 105 determines whether or not the action indicated by the action data acquired in step S22 is “lying” (step S23). That is, the action specifying unit 105 determines whether or not the previously specified action is “lying”.

When it is determined in step S23 that the previous action is "lying", the action specifying unit 105 determines that the first corrected atmospheric pressure difference minimum value is a predetermined threshold value TH ₁ from the second corrected atmospheric pressure difference minimum value. It is determined whether or not the decrease is (> 0) or more (step S24). In other words, the action specifying unit 105 determines whether or not the amount of change in the first corrected pressure difference minimum value with respect to the second corrected pressure difference minimum value is equal to or less than a predetermined threshold value TH ₁ ′ (<0). .. The threshold value TH ₁ ′ is obtained by reversing the positive and negative values of the threshold value TH ₁ .

In step S24, when it is determined that the first corrected atmospheric pressure difference minimum value is decreased by a predetermined threshold value TH ₁ or more from the second corrected atmospheric pressure difference minimum value, the action specifying unit 105 is in the corresponding 10 minutes (that is, the past). The behavior of the cow during (from 10 minutes before to the current time) is specified as "standing" (step S25). In this case, the amount of change in the first corrected atmospheric pressure difference minimum value with respect to the second corrected atmospheric pressure difference minimum value is equal to or less than the predetermined threshold value TH ₁ ′, and the altitude of the tag 20 attached to the cow is equal to or higher than the predetermined threshold value TH ₁ . ,It is rising. Therefore, it is considered that the cow has changed from the "lying" state to the "standing" state.

On the other hand, when it is not determined in step S24 that the first corrected atmospheric pressure difference minimum value is reduced by a predetermined threshold value TH ₁ or more from the second corrected pressure pressure difference minimum value, the action specifying unit 105 takes the 10 minutes. Identify the behavior of the cow as the same as the previous behavior (ie, "lying") (step S26). In this case, the amount of change in the first corrected atmospheric pressure difference minimum value with respect to the second corrected atmospheric pressure difference minimum value is equal to or less than the predetermined threshold value TH ₁ ′, and the altitude of the tag 20 attached to the cow is equal to or higher than the predetermined threshold value TH ₁ . Has not risen. Therefore, the cow is considered to remain in the "lying" state.

In step S23, when it is determined that the previous action is other than "lying" (that is, "standing"), the action specifying unit 105 has the first corrected pressure difference minimum value as the second corrected pressure difference. It is determined whether or not the predetermined threshold value TH ₂ (<0) or more is increased from the minimum value (step S27). In other words, the action specifying unit 105 determines whether or not the amount of change in the first corrected pressure difference minimum value with respect to the second corrected pressure difference minimum value is equal to or greater than a predetermined threshold value TH ₂ '(> 0). .. The threshold value TH ₂ ′ is obtained by reversing the positive and negative values of the threshold value TH ₂ .

In step S27, when it is determined that the first corrected atmospheric pressure difference minimum value has increased by a predetermined threshold value TH ₂ or more from the second corrected atmospheric pressure difference minimum value, the behavior specifying unit 105 determines the behavior of the cow in the 10 minutes. It is specified as "lying" (step S28). In this case, the amount of change in the first corrected atmospheric pressure difference minimum value with respect to the second atmospheric pressure difference minimum value is the predetermined threshold value TH ₂ ′ or more, and the altitude of the tag 20 attached to the cow is the predetermined threshold value TH ₂ ′ or more. , Is descending. Therefore, it is considered that the cow has changed from the "standing" state to the "lying" state.

On the other hand, if it is not determined in step S27 that the first corrected atmospheric pressure difference minimum value has increased by a predetermined threshold value TH ₂ or more from the second corrected pressure pressure difference minimum value, the action specifying unit 105 takes the 10 minutes. The behavior of the cow is identified as the same as the previous behavior (that is, "standing up") (step S29). In this case, the amount of change in the first corrected atmospheric pressure difference minimum value with respect to the second atmospheric pressure difference minimum value is the predetermined threshold value TH ₂ ′ or more, and the altitude of the tag 20 attached to the cow is the predetermined threshold value TH ₂ ′ or more. Is not descending. Therefore, the cow is considered to remain in the "standing" state.

<< Calculation processing of the minimum corrected atmospheric pressure difference >>
Next, the details of the process of step S21 (calculation process of the minimum corrected atmospheric pressure difference value) will be described with reference to FIG. FIG. 8 is a flowchart showing an example of the calculation process of the minimum corrected atmospheric pressure difference according to the first embodiment.

First, the acquisition unit 101 acquires the pressure sensor value and the y component acceleration sensor value included in the measurement data of the corresponding tag ID (that is, the tag ID “Tag1”) for the past 20 minutes from the measurement data storage unit 200, and at the same time. The reference pressure sensor value included in the reference pressure data for the past 20 minutes is acquired from the reference pressure data storage unit 400 (step S31).

The acquisition unit 101 does not have to acquire the barometric pressure sensor value and the y component acceleration sensor value up to the past 10 minutes from the measurement data storage unit 200. In this case, the barometric pressure sensor value and the y-component acceleration sensor value included in the measurement data acquired in step S11 of FIG. 6 may be used as the barometric pressure sensor value and the y-component acceleration sensor value up to the past 10 minutes.

Further, acquiring the measurement data for the past 20 minutes in step S31 is an example, and the present invention is not limited to this. For example, when the measurement data up to the past ΔT is acquired in step S11 of FIG. 6 with a predetermined time width as ΔT, in the above step S31, the barometric pressure sensor value, the y component acceleration sensor value and the y component acceleration sensor value up to the past 2ΔT are obtained. It suffices if the reference barometric pressure sensor value is acquired.

Further, the y component acceleration sensor value is acquired in step S31 above so that the positive direction of the y-axis is the traveling direction of the cow when the cow is standing and the neck is not bent. This is because is installed. When the tag 20 is attached so that the traveling direction is in the positive direction of the other axis (for example, x-axis or z-axis), if the acceleration sensor value of the component of the other axis is acquired. good.

In the following, in step S31 above, the barometric pressure sensor values p ₁ , ..., P _M'and the y component acceleration sensor values y ₁ , ..., y _M'for the past 20 minutes and the reference barometric pressure for the past 20 minutes. The description continues assuming that the sensor values q ₁ , ..., Q _L are acquired by the acquisition unit 101.

Next, the pretreatment unit 102 has the barometric pressure sensor values p ₁ , ..., P _M'acquired by the acquisition unit 101, the y component acceleration sensor values y ₁ , ..., y _M' , and the reference pressure. Preprocessing is performed on the sensor values q ₁ , ..., Q _L (step S32). That is, the pretreatment unit 102 performs the following two pretreatments (pretreatment 1) and (pretreatment 2). However, it is possible to perform only (pretreatment 1) without performing (pretreatment 2).

(Pretreatment 1) Atmospheric pressure sensor values p ₁ , ..., p _M' , y component acceleration sensor values y ₁ , ..., y _M' , and reference barometric pressure sensor values q ₁ , ..., q _L. And resample. As a result, the barometric pressure sensor value, the y component acceleration sensor value, and the reference barometric pressure sensor value after resampling are renumbered, and the barometric pressure sensor values p ₁ , ..., P _N'and the y component acceleration sensor value y ₁ , ..., Y _N'and the reference barometric pressure sensor values q ₁ , ..., Q _N'are obtained. Here, when resampling is performed at 2-second intervals, N'= 600.

(Pretreatment 2) At regular intervals (for example, 1 minute intervals), the average value of the barometric pressure sensor values p ₁ , ..., P _N'and the y component acceleration sensor values y ₁ , ..., y _N ' And the mean value of the reference barometric pressure sensor values q ₁ , ..., Q _N'are calculated. As a result, the average value μ _p1 , ..., μ _pR of the barometric pressure sensor values at each predetermined interval and the average value μ _y1 , ..., μ _yR of the y component acceleration sensor values at each predetermined interval are obtained. , The average value μ _q1 , ···, μ _qR of the reference barometric pressure sensor values at the predetermined intervals can be obtained. Here, when the average value is calculated at 1-minute intervals, R = 20.

Hereinafter, assuming that the average value is calculated at 1-minute intervals, the average value of the barometric pressure sensor value in 1 minute is the "average barometric pressure sensor value", and the average value of the y component acceleration sensor value in 1 minute is the "average y". The "component acceleration sensor value" and the average value of the reference atmospheric pressure sensor values in one minute are also referred to as "average reference atmospheric pressure sensor values".

Next, the correction difference minimum value calculation unit 104 calculates the barometric pressure difference value from the average barometric pressure sensor values μ _p1 , ..., μ _pR and the average reference barometric pressure sensor values μ _q1 , ..., μ _qR ( Step S33). Assuming that the atmospheric pressure difference value is _ui (i = 1, ..., R), the atmospheric pressure difference value _ui is calculated by, for example, _ui = μ _pi − μ _qi . However, it may be calculated as _ui = μ _qi − μ _pi .

Next, the correction difference minimum value calculation unit 104 calculates the correction pressure difference value from the pressure difference values u ₁ , ..., U _R and the average y component acceleration sensor values μ _{y 1} , ..., _μ y R. (Step S34). The corrected atmospheric pressure difference value is set to vi ( _i = 1, ..., R), and the corrected atmospheric pressure difference value _{vi is calculated by vi = u i + μ} y _i × C. Here, C is a preset constant (parameter). In this way, the corrected barometric pressure difference value v _i is calculated by correcting the _barometric pressure difference value u _i by the average y component acceleration sensor value μ y i.

Next, the correction difference minimum value calculation unit 104 calculates the first correction difference minimum value and the second correction difference minimum value from the correction pressure difference value vi ( _i = 1, ..., R) ( Step S35). That is, the correction difference minimum value calculation unit 104 uses the correction pressure difference value vi ( _{i =} 1, ..., R) included in the time width from the present to the past 10 minutes ago. The minimum value of is the first corrected atmospheric pressure difference minimum value, and the minimum value of the corrected atmospheric pressure difference value _vi included in the time width from the past 10 minutes ago to the past 20 minutes ago is the second corrected atmospheric pressure difference minimum value. ..

The correction difference minimum value calculation unit 104 may use, for example, an average value or a maximum value instead of the above-mentioned minimum value. That is, for example, the average value of the corrected atmospheric pressure difference value _vi included in the time width from the present to the past 10 minutes ago is the "first corrected atmospheric pressure difference average value", and the time from the past 10 minutes ago to the past 20 minutes ago. The average value of the corrected atmospheric pressure difference value _vi included in the width may be used as the “second corrected atmospheric pressure difference average value” in place of the first corrected atmospheric pressure difference minimum value and the second corrected atmospheric pressure difference minimum value. .. Similarly, for example, the maximum value of the corrected atmospheric pressure difference value _vi included in the time width from the present to the past 10 minutes ago is set to the "first corrected atmospheric pressure difference maximum value", from the past 10 minutes ago to the past 20 minutes ago. Even if the maximum value of the corrected atmospheric pressure difference value _vi included in the time width is used as the "second corrected atmospheric pressure difference maximum value" in place of the first corrected atmospheric pressure difference minimum value and the second corrected atmospheric pressure difference minimum value. good.

<Summary of the first embodiment>
As described above, the behavior specifying system 1 according to the present embodiment can specify the behavior of the cow from the measurement data received from the tag 20 worn by the cow. At this time, in the behavior identification system 1 according to the present embodiment, by using the index value calculated from the acceleration sensor value included in the measurement data and the behavior identification model 300 created in advance by the machine learning method, the cow Behavior can be identified (behavior identification by motion intensity analysis). In particular, by using the maximum value of the L2 norm and the standard deviation as index values, it is possible to identify "ruminant", which is a behavior peculiar to cattle, with high accuracy. In other words, it becomes possible to distinguish "ruminant", which is a behavior peculiar to cows, from behaviors such as "standing" and "lying" with high accuracy.

Further, in the behavior identification system 1 according to the present embodiment, when the behavior of the cow is specified as "standing", "lying" or "rebellion" by the behavior identification by the motion intensity analysis, a predetermined component (for example, y component) is used. ) Acceleration sensor value, pressure sensor value, and reference pressure sensor value, more accurate behavior can be specified (behavior identification by pressure pressure analysis). This makes it possible to identify "standing" (including "standing ruminant") or "lying" (including "lying ruminant") with high accuracy. In particular, when the barometric pressure difference value is corrected by the acceleration sensor value of the y component, for example, even when the neck is lowered while standing up, it is erroneously specified as "lying". It will be possible to identify "standing" with high accuracy.

Then, by displaying the behavior of the cow identified in this way on, for example, a display device (display or the like) of the behavior identification device 10 and providing it to a dairy farmer or the like, confirmation of the behavior often seen at the time of estrus of the cow can be confirmed. In addition to being able to easily perform it, it will also be possible to detect abnormal behavior due to illness or injury at an early stage, so that it will be possible to easily manage the estrus period and health management of cattle. Become.

[Second embodiment]
Next, the second embodiment will be described. In the first embodiment, with the cow standing upright and the neck not bent, the right direction with respect to the traveling direction of the cow is the positive direction of the x-axis, and the traveling direction is the positive direction of the y-axis. , The case where the tag 20 is attached so that the direction of gravity is the positive direction of the z-axis has been described. However, for example, the tag 20 may be attached upside down, or the tag 20 may be turned upside down due to a twist of the belt (hereinafter, also simply referred to as “twist”). In this case, since the sign of the y-component acceleration sensor value is inverted, the atmospheric pressure difference value may not be appropriately corrected by the y-component acceleration sensor value in the corrected atmospheric pressure analysis.

Therefore, in the second embodiment, a case where the behavior is specified by the corrected atmospheric pressure analysis will be described in consideration of the tag orientation, the twist, and the like indicating the front and back of the tag 20.

In the second embodiment, the differences from the first embodiment will be mainly described, and the description of the same components as those in the first embodiment will be omitted.

<Relationship between tag orientation and twist>
Here, the relationship between the tag orientation of the tag 20 and the twist of the belt for attaching the tag 20 to the cow will be described with reference to FIG. FIG. 9 is a diagram for explaining the relationship between the tag orientation and the twist. After that, with the cow standing and the neck not bent, the right direction is the positive direction of the x-axis, the traveling direction is the positive direction of the y-axis, and gravity with respect to the traveling direction of the cow. The tag direction of the tag 20 attached so that the direction is the positive direction of the z-axis is set to "+1" (positive direction). On the other hand, the tag orientation of the tag 20 mounted upside down with respect to the tag 20 whose tag orientation is "+1" is set to "-1" (negative direction).

FIG. 9A shows a case where the tag orientation is “+1” and no twist occurs. In this case, the right direction with respect to the traveling direction of the cow is the positive direction of the x-axis, the traveling direction is the positive direction of the y-axis, and the gravity direction is the positive direction of the z-axis.

FIG. 9B shows a case where the tag orientation is “-1” and no twist occurs. In this case, the left direction with respect to the traveling direction of the cow is the positive direction of the x-axis, the opposite direction to the traveling direction is the positive direction of the y-axis, and the gravity direction is the positive direction of the z-axis.

FIG. 9C shows a case where the tag orientation is “+1” and twisting occurs. In this case, the right direction with respect to the traveling direction of the cow is the positive direction of the x-axis, the opposite direction to the traveling direction is the positive direction of the y-axis, and the opposite direction to the gravity direction is the positive direction of the z-axis. Will be.

FIG. 9D shows a case where the tag orientation is “-1” and twisting occurs. In this case, the left direction with respect to the traveling direction of the cow is the positive direction of the x-axis, the traveling direction is the positive direction of the y-axis, and the opposite direction to the gravity direction is the positive direction of the z-axis.

<Functional configuration of behavior identification processing unit 100>
Next, the functional configuration of the behavior specifying processing unit 100 according to the present embodiment will be described with reference to FIG. FIG. 10 is a diagram showing an example of the functional configuration of the behavior specifying processing unit 100 according to the second embodiment.

As shown in FIG. 10, the behavior specifying processing unit 100 according to the present embodiment further includes a tag orientation calculation unit 106.

The tag orientation calculation unit 106 calculates whether the tag orientation of the tag 20 attached to the cow is "+1" or "-1".

Further, the correction difference minimum value calculation unit 104 according to the present embodiment calculates the correction pressure difference minimum value in consideration of the tag orientation calculated by the tag orientation calculation unit 106 and the presence / absence of twisting of the tag 20. ..

<Behavior identification process>
Next, the behavior specifying process according to the present embodiment will be described. In the behavior specifying process according to the present embodiment, the calculation process of the corrected atmospheric pressure difference minimum value is different from that of the first embodiment. Therefore, the calculation process of the corrected atmospheric pressure difference minimum value will be described below.

<< Calculation processing of the minimum corrected atmospheric pressure difference >>
The calculation process of the minimum corrected atmospheric pressure difference value (details of step S21 in FIG. 7) according to the present embodiment will be described with reference to FIG. FIG. 11 is a flowchart showing an example of the calculation process of the corrected atmospheric pressure difference minimum value according to the second embodiment.

First, the acquisition unit 101 measures the pressure sensor value, the y component acceleration sensor value, and the z component acceleration sensor value included in the measurement data of the corresponding tag ID (that is, the tag ID “Tag1”) for the past 20 minutes. In addition to acquiring from 200, the reference pressure sensor value included in the reference pressure data for the past 20 minutes is acquired from the reference pressure data storage unit 400 (step S41).

The acquisition unit 101 does not have to acquire the barometric pressure sensor value, the y component acceleration sensor value, and the z component acceleration sensor value up to the past 10 minutes from the measurement data storage unit 200. In this case, as the pressure sensor value, the y component acceleration sensor value, and the z component acceleration sensor value up to the past 10 minutes, the pressure sensor value, the y component acceleration sensor value, and the y component acceleration sensor value included in the measurement data acquired in step S11 of FIG. The z-component acceleration sensor value may be used.

Further, acquiring the measurement data for the past 20 minutes in step S41 is an example, and the present invention is not limited to this. For example, when the measurement data up to the past ΔT is acquired in step S11 of FIG. 6 with a predetermined time width as ΔT, in the above step S41, the pressure sensor value and the y component acceleration sensor value up to the past 2ΔT are obtained. It suffices if the z component acceleration sensor value and the reference pressure sensor value are acquired.

In the following, in step S41 above, the barometric pressure sensor values p ₁ , ..., P _M ', y component acceleration sensor values y ₁ , ..., y _M'and z component acceleration sensor values z ₁ , for the past 20 minutes. ..., Z _M'and the reference barometric pressure sensor values q ₁ , ..., Q _L for the past 20 minutes will be described as being acquired by the acquisition unit 101.

Next, the preprocessing unit 102 has the barometric pressure sensor values p ₁ , ..., P _M'acquired by the acquisition unit 101, the y component acceleration sensor values y ₁ , ..., y _M' , and the z component. Preprocessing is performed on the acceleration sensor values z ₁ , ..., Z _M'and the reference barometric pressure sensor values q ₁ , ..., Q _L (step S42). That is, the pretreatment unit 102 performs the following two pretreatments (pretreatment 1) and (pretreatment 2). However, it is possible to perform only (pretreatment 1) without performing (pretreatment 2).

(Pretreatment 1) Barometric pressure sensor value p ₁ , ..., p _M' , y component acceleration sensor value y ₁ , ..., y _M' , z component acceleration sensor value z ₁ , ..., z _{Resample M'and} the reference barometric pressure sensor values q ₁ , ..., Q _L. As a result, the pressure sensor value after resampling, the y component acceleration sensor value, the z component acceleration sensor value, and the reference pressure sensor value are renumbered, and the pressure sensor values p ₁ , ..., P _N'and the y component. Acceleration sensor values y ₁ , ..., y _N' , z component acceleration sensor values z ₁ , ..., z _N'and reference pressure sensor values q ₁ , ..., q _N'are obtained. Here, when resampling is performed at 2-second intervals, N'= 600.

(Pretreatment 2) At intervals (for example, 1 minute interval), the average value of the barometric pressure sensor values p ₁ , ..., P _N'and the y component acceleration sensor values y ₁ , ..., y _N ' , The average value of the z component acceleration sensor values z ₁ , ..., Z _N ', and the average value of the reference pressure sensor values q ₁ , ..., Q _N'are calculated. As a result, the average value of the pressure sensor values at each predetermined interval μ _p1 , ..., μ _pR and the average value of the y component acceleration sensor values at each predetermined interval μ _y1 , ..., Μ _yR . , The mean value μ _z1 , ···, μ _zR of the y component acceleration sensor value at each predetermined interval, and the mean value μ _q1 , ···, μ _qR of the reference pressure sensor value at each predetermined interval. can get. Here, when the average value is calculated at 1-minute intervals, R = 20.

Hereinafter, assuming that the average value is calculated at 1-minute intervals, the average value of the pressure sensor values in 1 minute is the "average pressure sensor value", and the average value of the y component acceleration sensor values in 1 minute is the "average y". "Component acceleration sensor value", the average value of the z component acceleration sensor value in 1 minute is also referred to as "average z component acceleration sensor value", and the average value of the reference pressure sensor value in 1 minute is also referred to as "average reference pressure sensor value".

Next, the correction difference minimum value calculation unit 104 calculates the barometric pressure difference value from the average barometric pressure sensor values μ _p1 , ..., μ _pR and the average reference barometric pressure sensor values μ _q1 , ..., μ _qR ( Step S43). Assuming that the atmospheric pressure difference value is _ui (i = 1, ..., R), the atmospheric pressure difference value _ui is calculated by, for example, _ui = μ _pi − μ _qi . However, it may be calculated as _ui = μ _qi − μ _pi .

Next, the tag orientation calculation unit 106 calculates the tag orientation (“+1” or “-1”) of the tag 20 (that is, the tag 20 of the tag ID “Tag1”) (step S44). The details of the process of this step (calculation process for tag orientation) will be described later.

Next, the correction difference minimum value calculation unit 104 has a barometric pressure difference value u ₁ , ..., U _R , an average y component acceleration sensor value μ _{y 1} , ..., μ _y R, and an average z component acceleration sensor value μ. The corrected atmospheric pressure difference value is calculated from _z1 , ..., μ _zR and the tag orientation calculated in step S44 above (step S45). The corrected atmospheric pressure difference value is vi ( _i = 1, ..., R), the tag orientation is e, and the corrected atmospheric pressure difference value _vi is _vi = _ui + e × sign (μ _zi ) × μ _yi × C. Is calculated by. Here, C is a preset constant (parameter), and sign is a sign function. However, sign (0) is assumed to return a positive number.

In this way, the corrected barometric pressure difference value v _i is calculated by correcting the _barometric pressure difference value u _i with the average y component acceleration sensor value μ y i considering the tag orientation and the sign of the average z component acceleration sensor value μ _z i. Will be done. As a result, the atmospheric pressure difference value can be appropriately corrected by the y component acceleration sensor value even when the front and rear are different depending on the mounting method of the tag 20 or when the belt is twisted. Here, the influence of the front and back of the tag 20 being different depending on the tag orientation e is corrected, and the influence of the twisting caused by sign (μ _zi ) is corrected.

Next, the correction difference minimum value calculation unit 104 calculates the first correction difference minimum value and the second correction difference minimum value from the correction pressure difference value vi ( _i = 1, ..., R) ( Step S46). That is, the correction difference minimum value calculation unit 104 uses the correction pressure difference value vi ( _{i =} 1, ..., R) included in the time width from the present to the past 10 minutes ago. The minimum value of is the first corrected atmospheric pressure difference minimum value, and the minimum value of the corrected atmospheric pressure difference value _vi included in the time width from the past 10 minutes ago to the past 20 minutes ago is the second corrected atmospheric pressure difference minimum value. ..

As in the first embodiment, the correction difference minimum value calculation unit 104 may use, for example, an average value or a maximum value instead of the above-mentioned minimum value.

≪Calculation process for tag orientation≫
Next, the details of the process (calculation process for tag orientation) in step S44 will be described with reference to FIG. FIG. 12 is a flowchart showing an example of the tag orientation calculation process according to the second embodiment.

First, the acquisition unit 101 acquires the y-component acceleration sensor value and the z-component acceleration sensor value of the corresponding tag ID (that is, the tag ID “Tag1”) for the past 24 hours from the measurement data storage unit 200 (step S51). .. The past 24 hours is an example, and the y-component acceleration sensor value and the z-component acceleration sensor value during an arbitrary time may be acquired.

Next, the tag orientation calculation unit 106 calculates the following index values every 10 minutes (step S51).

-Yz value = average y component acceleration sensor value x sign (average z component acceleration sensor value)
-Y component standard deviation Here, the average y component acceleration sensor value is the average value of the y component acceleration sensor values in 10 minutes. Similarly, the average z-component acceleration sensor value is the average value of the z-component acceleration sensor values in 10 minutes. The z-component standard deviation is the standard deviation of the y-component acceleration sensor value in 10 minutes.

As a result, the yz value and the y component standard deviation corresponding to each other are obtained every 10 minutes. It should be noted that the calculation of the index value every 10 minutes is an example, and the index value may be calculated every arbitrary time width.

Next, the tag orientation calculation unit 106 calculates the median y _a of the yz values whose corresponding y component standard deviation exceeds the predetermined threshold value _SA among the yz values calculated in step S52 above (step). S53). Here, the threshold value _SA can be set to an arbitrary value, and for example, _SA = 150 or the like may be set.

Next, the tag orientation calculation unit 106 calculates the median y _b of the yz values whose corresponding y component standard deviation is below the predetermined threshold value SB among the yz values calculated in step S52 above (step ₎ . S54). Here, the threshold value _SB can be set to an arbitrary value, and for example, _SB = 40 or the like may be set.

Finally, the tag orientation calculation unit 106 calculates the tag orientation e by e = sign ( _{ya − y b )} (step S55).

The above step S54 may not be executed. In this case, in step S555 described above, the tag orientation may be calculated by e ₌ sign (ya).

<Summary of the second embodiment>
As described above, in the behavior identification system 1 according to the present embodiment, even when the tag orientation of the tag 20 is different or a twist occurs, the cattle can be analyzed by the corrected atmospheric pressure analysis as in the first embodiment. It becomes possible to identify "standing" or "lying" with high accuracy.

The present invention is not limited to the above-described embodiment disclosed specifically, and various modifications and modifications can be made without departing from the scope of claims.

1 Behavior identification system 10 Behavior identification device 20 Tags 30 Reference pressure sensor 100 Behavior identification processing unit 101 Acquisition unit 102 Preprocessing unit 103 Index value calculation unit 104 Correction difference minimum value calculation unit 105 Behavior identification unit 200 Measurement data storage unit 300 Behavior identification Model 400 Reference pressure data storage unit 500 Behavior data storage unit

Claims (7)

It is a behavior identification device that identifies the behavior of livestock.
A storage means for storing the acceleration data measured by the acceleration sensor mounted on the livestock and the barometric pressure data measured by the barometric pressure sensor mounted on the livestock.
An index value calculating means for calculating a predetermined index value from one or more of the acceleration data during a predetermined time, and
A first specific means for specifying the behavior of the livestock based on the index value and a specific model for specifying the behavior of the livestock prepared in advance.
When the behavior specified by the first specific means is a predetermined behavior, the behavior of the livestock is specified based on one or more of the barometric pressure data and one or more of the acceleration data during a predetermined time. The second specific means and
Have,
The acceleration data includes acceleration values of a plurality of components.
The second specific means is
The behavior of the livestock using the correction value obtained by correcting each of the barometric pressure values included in the one or more barometric pressure data with each of the acceleration values of a predetermined component among the acceleration values included in the one or more acceleration data. A behavior identification device characterized by identifying . The second specific means is
Of the one or more correction values, a first minimum value indicating the minimum value of the correction value in the first period and a second minimum value indicating the minimum value of the correction value in the second period are calculated. Then, by comparing the amount of change between the first minimum value and the second minimum value with a predetermined threshold value, the behavior of the livestock is specified according to the comparison result. The behavior specifying device according to claim 1 . It has a direction calculating means for calculating a value indicating the direction in which the accelerometer is attached to the livestock.
The second specific means is
Each of the barometric pressure values included in one or more of the barometric pressure data, each of the acceleration values of the predetermined first component among the acceleration values included in the one or more of the acceleration data, and the acceleration of the predetermined second component. The behavior specifying device according to claim 1 or 2 , wherein the behavior of the livestock is specified by using the correction value corrected by the code of the value and the value indicating the direction. The second specific means is
Each of the average values of the barometric pressure values contained in one or more of the barometric pressure data in the predetermined third period is the acceleration value of the predetermined first component of the acceleration values included in the one or more acceleration data. The livestock using each of the average values in the third period, the sign of the average value of the acceleration value of the predetermined second component in the third period, and the correction value corrected by the value indicating the direction. The behavior specifying device according to claim 3 , wherein the behavior is specified. The acceleration value of the first component is an acceleration value of a component indicating the traveling direction of the livestock or a component indicating the direction opposite to the traveling direction.
The behavior specifying device according to claim 3 or 4 , wherein the acceleration value of the second component is an acceleration value of a component indicating the direction of gravity or a component indicating the direction opposite to the direction of gravity. .. A behavioral identification device that identifies the behavior of livestock,
A storage procedure for storing the acceleration data measured by the acceleration sensor mounted on the livestock and the barometric pressure data measured by the barometric pressure sensor mounted on the livestock in a storage means.
An index value calculation procedure for calculating a predetermined index value from one or more of the acceleration data during a predetermined time, and
A first specific procedure for specifying the behavior of the livestock based on the index value and a specific model for specifying the behavior of the livestock prepared in advance, and
When the behavior specified by the first specific procedure is a predetermined behavior, the behavior of the livestock is specified based on one or more of the barometric pressure data and one or more of the acceleration data during a predetermined time. The second specific procedure and
And run
The acceleration data includes acceleration values of a plurality of components.
The second specific procedure is
The behavior of the livestock using the correction value obtained by correcting each of the barometric pressure values included in the one or more barometric pressure data with each of the acceleration values of a predetermined component among the acceleration values contained in the one or more acceleration data. A behavioral identification method characterized by identifying . A program for making a computer function as each means in the behavior specifying device according to any one of claims 1 to 5 .

Images