JP2018019611A - Leash type vital measuring device for pet - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a leash type vital measuring device that reduces load on a pet or owner relative to a collar type pet vital sensor.SOLUTION: A leash type vital measuring device for a pet includes: a leash-tension tension detector for detecting tension of a leash 101 between a pet and an owner; a number of steps calculation part for detecting the number of steps by correcting information from an acceleration sensor depending on the tension information; and a display part for presenting the number of steps data to a user.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a vital measurement device that senses vitals such as an activity amount for a pet and displays the result, and a pet insurance service using acquired vital data.

  In recent years, an increasing number of households have pets as members of their families, and the demand for health management for pets as a family is increasing as in humans. Many human vital data acquisition devices for health management have been developed, but there are still few veterinary data acquisition devices for pets. For example, Patent Document 1 discloses a pedometer for pets. This device is a dog pedometer that counts the number of steps of a dog from data obtained from a vibration sensor for a certain period of time attached to a collar of the dog. Patent Document 2 discloses a device that detects not only the number of steps but also dog tremor. Dog tremor is said to represent dog stress, and the number of tremors is one of the indices representing the dog's condition. In this way, several techniques for acquiring various vital data of pets have been developed.

JP 2007-107943 A International Publication No. 2016/006028

  However, it is assumed that these devices are directly attached to pets such as collars, and the burden on the pet side due to wearing of the devices becomes a problem. Such a device is light for humans, but for small dogs it is the same as carrying a heavy weight all the time, so it gets tired with that weight. Also, when trying to lighten the device, it is necessary to reduce the built-in battery, and this shortens the drive time, so there is a problem that the convenience on the owner's side is reduced by removal for charging or replacement.

  The present invention solves the above-described problem, and is a lead-type vital measuring device for pets, which includes a lead tension tension detection unit that detects the tension of a lead between a pet and an owner, and an acceleration sensor according to the tension information. A step number calculating unit that detects the number of steps by correcting the information from and a display unit that presents the step number data to the user.

  According to this configuration, since the pedometer can be attached to the lead, the burden can be shared between the owner and the dog. For this reason, the burden of the weight of the device to a pet can be reduced. Moreover, if it is a lead, since it removes | desorbs from a pet for every walk, since it can serve as both desorption at the time of charge and desorption at the time of a walk, it is also convenient for the owner.

FIG. 1 is a diagram showing an appearance of a lead-type activity meter according to the first embodiment. FIG. 2 is a block diagram of a lead-type activity meter. FIG. 3 is a diagram showing a processing flow up to the calculation of the number of steps of the lead type activity meter. FIG. 4 is a diagram showing a processing flow until calorie calculation of the lead-type activity meter. FIG. 5 is a diagram for explaining the operation of the lead attractive force sensor. FIG. 6 is a diagram showing acceleration sensor data of the activity meter when the lead is pulled. FIG. 7 is a diagram illustrating a processing flow of the fixed model step count calculation processing. FIG. 8 is a diagram showing acceleration sensor data of the activity meter when the lead is loose. FIG. 9 is a diagram illustrating a process flow of the floating model step count calculation process. FIG. 10 is a diagram illustrating a display destination of measurement results. FIG. 11 is a diagram illustrating an example of display in the first embodiment. FIG. 12 is a diagram showing an external appearance (modification 1) of the lead-type activity meter according to the first embodiment. FIG. 13 is a diagram showing an external appearance (modification 2) of the lead-type activity meter according to the first embodiment. FIG. 14 is a diagram illustrating a configuration of a connection portion of the lead-type activity meter (Modification 2). FIG. 15 is a diagram showing the appearance of a pet bed with radio wave radar. FIG. 16 is a block diagram of a pet bed with radio wave radar. FIG. 17 is a diagram showing a processing flow of the pet bed with radio wave radar. FIG. 18 is a diagram illustrating an example of a display of a pet bed with radio wave radar. FIG. 19 is a diagram illustrating an appearance of a pet bed with radio wave radar (modified example). FIG. 20 is a diagram showing a processing flow (modification) of the pet bed with radio wave radar. FIG. 21 is a diagram illustrating an appearance of a pet meal management device. FIG. 22 is a block diagram of a pet meal management device. FIG. 23 is a diagram showing a processing flow of the pet meal management device. FIG. 24 is a diagram illustrating an example of a display of a pet meal management device. FIG. 25 is a diagram illustrating an appearance of a pet meal amount management device (modified example). FIG. 26 is a diagram illustrating a relationship (part 1) between a pet insurance company, a subscriber, and an animal hospital. FIG. 27 is a diagram showing a relationship (part 2) between a pet insurance company, a subscriber, and an animal hospital. FIG. 28 is a diagram illustrating creation of a prediction model based on accumulated data and data analysis. FIG. 29 is a diagram showing a processing flow for calculating premiums for risk subdivided pet insurance. FIG. 30 is a diagram illustrating an example of an insurance premium calculation table for risk subdivided pet insurance. FIG. 31 is a diagram showing a processing flow of a pet insurance with a health plan. FIG. 32 is a diagram illustrating an example of a display of a pet insurance with a health plan. FIG. 33 is a diagram showing a processing flow for fraud detection of an insurance claim application. FIG. 34 is a diagram showing a process flow of individual authentication. FIG. 35 is a diagram showing an example of individual authentication display. FIG. 36 is a diagram showing the relationship among pet insurance companies, subscribers, and animal hospitals in medical advice. FIG. 37 is a diagram showing a processing flow of medical recommendation. FIG. 38 is a diagram showing a display example of vital data presentation to a medical hospital recommended for medical care. FIG. 39 is a diagram showing a processing flow of insurance money processing at the time of medical treatment recommendation. FIG. 40 is a diagram illustrating an example of a display when medical treatment is recommended.

(Knowledge that became the basis of the present invention)
In recent years, the total number of dogs and cats in Japan has exceeded 20 million, more than humans under the age of 15. Under such circumstances, pets are changing their household positions to be treated like family members, and various services and pet products are spreading, and the pet industry as a whole continues to grow. There are two major characteristics of the pet industry: personification and aging. In the former, pets can be dressed like humans or can be relaxed in oxygen capsules. In the latter, the life expectancy of dogs and cats has been extended by the development of pet medicine and the improvement of pet food, etc., and the aging is progressing in the same way as humans. In particular, with regard to aging, various measures are required to make pets live long and healthy.

  With the widespread use of various sensors and networks, health data can be easily acquired and managed by electronic devices such as activity meters. For example, it is possible to acquire a wide range from lifestyle habits such as the number of steps, exercise amount, sleep, and meal amount to vital data such as heart rate, respiratory rate, and weight. Hereinafter, these data are collectively referred to as vital data.

  Although there are still fewer vital sensing devices for pets than humans, for example, Patent Document 1 discloses an activity meter technology for dogs. Similarly, Patent Document 2 discloses a device that detects not only the number of steps but also the tremor of the dog as vital data. Since these devices are directly attached to pets such as collars, even if they are light for humans, they are the same as always carrying heavy loads for dogs with a weight of several kilograms, which is a burden on the dogs. Further, in order to make the device lighter, it is necessary to reduce the built-in battery. As a result, the driving time is shortened, so that there is a problem that the convenience on the owner side is lowered due to the removal for charging or replacement.

  The lead-type vital measuring device for pets of the present invention is a lead-type vital measuring device for pets, and includes a lead tension tension detecting unit that detects the tension of the lead between the pet and the owner, and acceleration according to the tension information. A step number calculating unit that detects the number of steps by correcting information from the sensor, and a display unit that presents the step number data to the user.

  With this configuration, since a pedometer can be attached to the lead, the burden can be shared between the owner and the dog, and the burden on the weight of the device body can be reduced. In addition, since the product form is lead, detachment is performed at the time of a walk, so that the detachment at the time of charging and the detachment at the time of a walk can be performed together, so the convenience of the owner is also high. Charge it when you take it off, that is, when you are not taking a walk.

  The correction in the step count calculation unit may use a model in which the lead step and the threshold are compared, and the step count calculation method is divided when the tension is larger than the threshold and when the tension is smaller than the threshold. .

  According to this, even when a pedometer is attached to a non-fixed object such as a lead, an accurate number of steps can be detected.

  Note that each of the embodiments described below shows a specific example of the present invention. Numerical values, shapes, components, steps, order of steps and the like shown in the following embodiments are merely examples, and are not intended to limit the present invention. In addition, among the constituent elements in the following embodiments, constituent elements that are not described in the independent claims indicating the highest concept are described as optional constituent elements. In all the embodiments, the contents can be combined.

(Embodiment 1)
FIG. 1 shows the appearance of a lead-type activity meter that is a vital measurement device for pets in the present embodiment. The apparatus is mainly composed of an activity meter main body 102 and a battery 103 in the lead 101 portion. This lead 101 is connected to a dog collar 104 in the same manner as usual.

  In this configuration, it is desirable that the activity meter main body 102 be disposed on the lead as close to the dog. This is because if the dog is placed on the owner side, the dog's movement is absorbed in the middle of the lead, and the accuracy of detecting the number of steps may be lost. If it is arranged closer to the dog's collar 104, the movement of the dog can be transmitted directly in the same way as it is worn on the collar. In addition, placing the device as close to the collar as possible has the advantage of reducing the risk of a dog being bitten and broken.

  About the battery 103, it is better to arrange | position to the owner's human side contrary to the active mass meter main body 102. FIG. Since the battery is a heavy device among electronic devices, it directly leads to a burden on the dog. For this reason, the battery 103 is arranged on the human side of the lead so that a person, not a dog, can bear the weight of the battery.

  FIG. 2 shows a functional block diagram of the lead-type activity meter. This block diagram shows the internal configuration of the activity meter main body 102. The battery of the power supply unit 201 is in the battery 103, and a part of the external communication / display unit 209 is displayed on an external smartphone. I do. The lead-type activity meter includes a power supply unit 201 that controls the entire power supply, a lead attractive force sensor 202 that detects lead tension, a triaxial acceleration sensor 203, and a communication unit 204 that exchanges the sensor with the signal processing unit. A detection unit 205 that performs lead tension detection processing, a step calculation unit 206 that calculates the number of steps, a dog load detection unit 207 that detects a load applied to the dog from the lead, an exercise calorie calculation unit 208 that detects exercise calorie of the dog, It also includes an external communication / display unit 209 that communicates results such as the number of steps and calories to an external smartphone, cloud, etc., uploads data, and displays it to the user.

  The contents of each block will be described in order with the processing flow of FIGS. 3 and 4. Two sensors, a lead attractive force sensor 202 and an acceleration sensor 203 are mounted on the lead-type activity meter of the present embodiment. First, reading processing S301 of these sensor data is performed.

  The lead attractive force sensor 202 detects how much tension is applied to the lead between the owner and the dog as a lead tension detection process S302. The lead attractive force sensor 202 is realized by inserting a conductive rubber stretch resistance sensor in a part of the lead. This is a rubber-like resistance. When the rubber is stretched, the resistance is increased. When the rubber is shrunk, the resistance is decreased. The resistance value is converted into a voltage or the like, and used as a sensor value.

  FIG. 5 is a diagram illustrating the operation of the lead attractive force sensor 202. As shown in FIG. 5 (a), when tension is applied to the lead, that is, when the lead is stretched between the dog and the owner, since the rubber inside the lead is stretched, the resistance is kept high. Yes. On the other hand, when the lead is loose between the owner and the dog as in (b) and (c), the resistance is reduced. Since the resistance is rubber resistance, the resistance of the rubber becomes constant as shown in FIG. Instead of the conductive rubber cord sensor, a strain gauge sensor such as a strain gauge sensor may be used, and any method can be used as long as it can detect the tension of the lead.

Next, a threshold value comparison process S303 is performed on the detected tension. A broken line on the graph of FIG. 5 represents a threshold value, and if it is equal to or greater than a certain threshold value, the subsequent processing is branched. In other words, the processing is divided according to whether the lead is tensioned or loose.
If the lead tension is high, it is equivalent to the activity meter being fixed to the collar, so the fixed model step detection process S304 is performed. On the contrary, if the tension is equal to or less than the threshold value, the activity meter is not fixed, so the floating model step count calculation process S305 is performed.

  The fixed model step count detection process S304 will be described with reference to FIGS. First, FIG. 6 shows a waveform of the acceleration sensor when the lead is tensioned. In this case, since an acceleration sensor is fixed to the collar, acceleration is increased or decreased on a certain axis in a cycle corresponding to the number of steps. In the example of FIG. 6, the acceleration is large on the X axis. The acceleration data is converted into the number of steps.

  Specifically, the step count detection process as shown in FIG. 7 is performed. First, peak axis extraction processing S701 for selecting an axis having a large change among the three-axis sensors is performed. This process simply selects an axis that varies greatly and periodically. Next, smoothing processing S702 is performed on the extracted axis. Depending on the accuracy of the three-axis acceleration sensor, even small movements are detected as acceleration, and therefore, all smoothing filters remove noise other than the acceleration of the action level such as the number of steps. Thereafter, a peak detection process S703 is performed on the waveform subjected to the smoothing process. In this process, when the window is cut for a certain period of time, the lower limit value and the upper limit value of the acceleration are extracted, and the opening width of the lower limit peak and the upper limit peak is detected. Finally, a walking determination process S704 is performed to determine whether or not there is an acceleration change due to walking with respect to the change on the width or time axis. In this process, it is determined whether or not the acceleration window corresponds to the number of steps by looking at the size of the lower and upper limit widths and the interval between adjacent waveforms. For example, when running, the acceleration tends to increase, so the lower and upper limits widen, but the distance between adjacent windows must be close. In addition, when walking slowly, the distance between the windows becomes long, but the lower and upper limits are narrower because they are walking. Thus, the determination which modeled the acceleration which generate | occur | produces by walking is performed in walk determination process S704.

  Next, the floating model step count calculation process S305 will be described with reference to FIGS. FIG. 8 shows the waveform of the acceleration sensor when no tension is applied to the lead. In this case, since the acceleration sensor is not fixed, a large acceleration increase / decrease occurs in each axis. In the example of FIG. 8, it can be seen that accelerations are greatly increased in the X axis, the Y axis, and the Z axis. For this waveform, it is necessary to detect the number of steps, and processing different from the case of attaching an activity meter to a normal collar is required.

  A specific floating model step count calculation process S305 when the lead is loose, that is, when the activity meter is not fixed, will be described with reference to FIG. First, a superimposition process S901 of acceleration components appearing on the XYZ axes is performed. In this example, the waveform is superimposed by simply taking the logical product (OR) of the three axes. Note that other methods may be used as the method of superposition.

  Next, a smoothing process S902 is performed on the superimposed waveforms. Since the activity meter is not fixed, the amount of noise changes compared to the case where it is fixed, but here the same smoothing process is performed. Next, the peak of the minimum value and the maximum value of adjacent waveforms is detected in a window having a certain time width. Since the activity meter is not fixed, it tends to be smaller than when the peak-to-peak is fixed. For this reason, if it is determined whether it is running or walking from the upper and lower limits, an erroneous determination is made. Here, this width correction is performed as non-fixed correction processing S904. Finally, in the walking determination process S905, model determination using the width and the relationship with adjacent windows is performed in the same manner as described above. The above processing is performed by the step count calculation unit 206.

  Finally, the detected number of steps is subjected to display processing S306 in the external communication / display unit 209. The display process S306 is a process for notifying the user (= owner) of information such as the total number of detected steps. As shown in FIG. 10, it may be displayed on the activity meter main body 102 itself, or may be displayed on a smartphone 1001 or a tablet (not shown) using communication such as Bluetooth (registered trademark). Further, data may be uploaded from the communication network of the smartphone 1001 or the activity meter main body 102 to the cloud server 1002.

  Next, we will explain how to measure calories consumed by dogs. The usual calorie is calculated from the number of steps and exercise intensity, but in the case of walking a dog, there is the influence of human pulling the dog with the lead, so in order to accurately measure the calorie, the influence from the owner, that is, the owner It is necessary to take into account. For this purpose, the load obtained by the tension from the lead is also input to the motion calorie calculation.

  After performing the sensor data reading S301 and the lead tension detection process S302 as shown in FIG. 4, the dog load detection unit 207 performs the load detection process S403. The load received by the dog is detected from the resistance change as shown in FIG. 5A, and the load obtained by integrating the load amount with time becomes the load amount to the dog, and this calculation is performed in the load detection process S403. Finally, the calorie calculation unit 208 calculates the calorie by adding the steps and the load from the lead. If the dog type is determined, the calorie consumption per unit step is determined, so that the calorie from the number of steps can be obtained by multiplying the detected number of steps. For the load from the lead, a similar table is prepared for each dog type, and calorie is detected by performing multiplication.

  Finally, display processing S405 is performed. Since the display is the same as the display process S306 in the case of the number of steps, the description is omitted.

  An example of display on the smartphone 1001 will be described with reference to FIG. The device of this embodiment is a vital device specialized for walking a dog. Therefore, as shown in the figure, the history of the walk is displayed as a whole, such as the date and time, distance, and health activity level viewed from the obtained vital data. In addition, since the owner's smartphone is used, the owner's step count and calorie consumption are also displayed along with the dog step count, calorie consumption, pulling distance, and messiness. The pulling distance indicates the distance that the dog walks by pulling the reed with respect to the total distance, and the naughtiness indicates the degree of discipline. Specifically, the degree of tampering is calculated by comparing the pulling distance with respect to the entire travel distance, the value calculated from the pulling strength, and a predetermined table.

  The number of steps may be visualized by a bar graph or a pie graph. Without giving a value for the specific number of steps taken in the walk, it may be expressed as more or shorter than the national average, or the result may be displayed as a symbolic mark such as a smile or a crying face. In addition, the number of steps taken today may be displayed as a ranking in the whole country, compared to the value collected on the cloud side. In addition, the display may display not only for the owner but also the number of steps of the whole family counted in the cloud, or only for the dog.

(Modification of activity meter)
As a form of implementation of this embodiment, the lead-type activity meter may be a battery-integrated activity meter 1201 in which the battery and the activity meter body are housed in the same individual as shown in FIG. When the weight of the battery is light or when the dog is large, the burden on the dog is reduced. Therefore, it is conceivable that the battery is mounted as an integrated type in consideration of ease of manufacture and price. Further, in the case of the battery-integrated activity meter 1201, the device becomes heavy, so it may be mounted not at a position close to the dog as shown in FIG. Further, the activity meter may be implemented by merging with other functions such as a rewinding function of the lead.

  In the present embodiment, in addition to the above, as shown in FIG. 13, the activity meter main body 102 may be mounted on the collar and only the battery 103 may be mounted on the lead 101 side. The active mass meter is lightweight because it has only a few electronic circuits and sensors such as acceleration. In addition, since the collar type can be securely fixed to the dog, there is an advantage that the amount of activity of the dog can be detected with little error. On the other hand, since it is necessary to charge the battery once every several days or weeks, bringing the lead 101 to the lead makes it possible to share the detachment of the lead and the detachment for charging. Since the effort related to detachment is reduced, the convenience of the user can be ensured as in the case of the lead-type activity meter. If it is the above structure, the fall of the convenience accompanying detection accuracy and charge can be prevented simultaneously.

  In the case of this mounting form, since it is necessary to secure an electrical contact between the collar 104 and the lead 101, an example of the configuration is shown in FIG. As shown in FIG. 14A, the power cable 1401 may be connected to the collar-side power connector 1402 separately from the mechanical connection to the collar 104 by a ring. According to this configuration, the power supply side does not need to ensure mechanical strength, and therefore design is easy. Further, as shown in FIG. 14B, the connection to the collar and the power supply are made using the ring 1403 whose inner side is a conductor for connecting the power supply and the connector 1404 in which the ring fits in the same width inside the collar. A connector that can also be used for connection may be used. Further, as shown in FIG. 14C, the same ring 1403 may be attached to the collar side, and the power source may be connected by connecting the rings. In this case, since the ring may come off for a moment, it is necessary to make sure that the ring is in close contact with a conductive rubber material.

  In addition, although the said Example is described supposing the lead of a dog, it does not restrict | limit to a dog actually. Other than dogs, cats, rabbits, horses, raccoon dogs, wild boars, and sheep can be used and can be applied to all animals.

(Pet bed with built-in radio radar)
Several devices are described for obtaining vital data in daily life without stressing pets. FIG. 15 shows a pet bed 1501 with a vital measurement function using a radio wave radar. A radio wave radar in the microwave band is mounted on the bed, and a transmission antenna 1502 for the radar is mounted on the center of the bed, and four receiving antennas 1503A to 1503D are mounted on the outer periphery of the bed.

  The radio wave radar is a method for detecting the movement of an object by applying a radio wave to a moving object and detecting a Doppler shift of the reflected wave. For this reason, it is possible to detect respiration from the movement of the lungs of the pet and to detect the body movement from the movement of the heart beat and the movement of the whole body. However, it is difficult to detect these three at the same time because the magnitudes of the movements are different. In particular, the movement of the pet's heart is much smaller than that of humans and is difficult to detect even with radar. In order to increase detection accuracy, it is necessary to drive the radar at a position as close as possible to the pet's heart so as not to be buried in other noises. However, a pet bed is not necessarily guaranteed to sleep in the same orientation as a human being, so a pet-specific measure different from that of a human is required. In the present invention, as shown in FIG. 15, the detection accuracy is improved by placing a weight sensor 1504 in the center and arranging several receiving antennas 1503A to 1503D around. A specific processing flow and the like will be described later.

  Pet beds are difficult to connect to a power outlet using a cable. If you have a power cable, you will not be able to place your bed where you want your pet to be because the storage space is limited. In addition, it is desirable to keep the cables from being visible to the dog as much as possible because there is a risk of electric power because the dog bites the electric cable. For this reason, in the pet bed 1501 with radio wave radar of the present invention, the entire system is driven by a battery.

  However, it is necessary to devise measures to reduce power consumption because there is a battery with a large capacity if radio waves are constantly transmitted by battery drive. For this reason, in this embodiment, the weight sensor 1504 is arranged near the center of the bed, and the radio wave radar is driven at regular intervals only when the dog is riding to measure the vitality of the ped. By doing in this way, since it becomes electric power which can be driven enough even with a battery, the user's vitality can be measured without being charged frequently, without being limited in place, and without fear of electric shock.

  A specific configuration diagram and processing flow will be described below with reference to FIGS. 16 and 17, respectively. The pet bed 1501 with the radio wave radar first performs the weight detection process S1701. In this process, the weight information obtained from the weight sensor 1504 is converted into weight information. The weight detection may use a pressure sensor or a strain gauge sensor. In the weight detection process S1701, a process of converting physical quantities of these sensors into information on actual weight is performed.

  Next, the weight detection unit 1605 performs threshold value comparison processing S1702 for the detected weight. The purpose here is to determine whether a dog or cat is on the basis of the weight. The set threshold value may be set by the user or may be set to about half of the average dog weight. Also, it may be set in advance to a weight that can react even with a puppy of about 200 g.

  When the weight is below the threshold value, the radio wave radar is not driven and the standby state is continued. When the weight exceeds the threshold value, that is, when the dog is placed on the bed, the radio wave radar driving process is advanced. First, the weight detection unit 1605 performs the distribution detection process S1703. This distribution detection processing unit detects which side of the bed the dog is sleeping on. For example, when the weight sensor detects a relatively large pressure (weight), there is a possibility that the body is on the weight sensor. Further, since the head can be determined from the distribution of the center of gravity, the direction of the head is detected. In FIG. 15, the description will be made assuming that the head is facing leftward.

  Next, the antenna switching unit 1603 performs antenna switching processing S1704 of the receiving antenna. Since it was detected in the previous detection that the dog's head is facing the left side, the receiving antenna close to the heart is 1503D. By switching the reception to the receiving antenna 1503D, the reflected wave of the Doppler can be captured at a position closest to the pet's heart, so that high-quality data with less noise can be acquired.

  Next, radar drive processing S1705 is performed to drive the radar on both the transmission side and the reception side. The transmission wave generator 1602 generates a microwave band (2.4 GHz) sine wave or the like, and irradiates the pet with radio waves from the transmission antenna 1502. The transmission data is designed to spread evenly around. A reflected wave that hits an object (here, a pet) is received from the receiving antenna 1503D, and vital detection processing S1706 is performed by the Doppler processing unit 1604.

  In the vital detection processing S1706, vital sensing is performed by obtaining a Doppler shift from the difference between the transmission waveform and the reception waveform, and separating and extracting the frequency component corresponding to the vital. In the case of a dog's heartbeat, since it is about 60 to 120 times per second, a component of 60 to 120 Hz appears strongly in the obtained result.

  Finally, display processing S1707 for displaying the result on a smartphone or the like is performed. FIG. 18 shows an example of display. Since the radio wave radar can take respiration, heartbeat, and body movement, these values are displayed on the user's smartphone. Further, since sleep can be automatically detected from these data, it is possible to detect sleep time and the like. Specifically, if the body motion decreases and the heart rate falls below normal, it is determined as sleeping. In addition to sleep, information that can be known from body and internal organs movements such as breathing, heartbeat, and body movement may be described on the display unit.

  In the above example, the reflected wave from the heart and stomach area is efficiently detected by changing the position of the receiving antenna using one transmitting antenna and one receiving antenna. As shown in FIG. The transmission / reception antennas may be three antennas and may be 3 × 3 MIMO (multiple-input and multiple-output). By doing so, it is possible to efficiently receive radio waves from a certain direction of the pet's heart by beamforming the radio waves instead of the antenna position, so there is no need to change and receive the positions. In addition, any method that can obtain radio wave directivity may be used without using MIMO.

  In the above example, the pet orientation is detected from the weight distribution, but the detection process may not be performed. For example, data from all receiving antennas may be acquired sequentially and the data from the antenna with the best signal quality may be adopted, or the beam forming may be used to use the data from the direction with the best signal quality. Also good.

  In the above description, the position of the receiving antenna is switched based on the result of the weight sensor. However, the transmission frequency may be switched. For example, when a small dog rides, it may be switched to the millimeter wave range such as the 60 GHz band instead of the 2.4 GHz band. The antennas need not all have the same shape. For example, the shape of the antenna near the center and the antenna near the corner may be changed.

  Further, the antenna switching function may be omitted as shown in FIG. When the detection accuracy of the radio wave radar is improved, the vital data of the pet can be acquired regardless of the antenna position. Therefore, the configuration may be limited to only the weight detection function and the vital detection function.

(Pet food management sensor)
A mat device 2101 for managing the amount of pet meal is shown in FIG. A pet food dish 2102 is placed on the device for use. The mat device is capable of detecting the amount and distribution of pressure, and has a basic function that can detect how much and what amount is on the mat. As a result, when a heavy object is on the range of a pet dish, the object is recognized as a meal. When a distribution about the size of the pet's foot is detected, it is determined as a pet. In addition, since the way of walking, weight, and the like differ depending on the pet and it appears in the distribution of the way of walking, the individual authentication of the pet is performed from such information (handling the way of walking like a fingerprint). In addition, when the distribution of pets occurs around the plate 2102 and the weight near the plate further decreases, the amount of food for each individual pet is counted by counting as the amount of pet food eaten by the pet. Can be managed.

  A specific configuration and processing flow will be described with reference to FIGS. The mat device 2101 has a configuration in which pressure sensors 2201 are spread on a two-dimensional plane. First, data is sent from the pressure sensor 2201 to the pressure distribution calculation unit 2203 through the communication unit 2202. In this part, a weight detection process S2301 is performed.

  Specifically, a process of calculating the weight from the area of the pressure distribution and the total pressure applied to the area is performed. For example, in the case of a resistance change type pressure sensor, the magnitude of pressure is obtained from data converted into voltage values, and the area is obtained from the number of these sensors.

  Next, the threshold value determination unit 2204 performs a comparison process S2302 whether or not the threshold value A is greater than or equal to threshold value A. The threshold value A is smaller than the threshold value B, which will be described later. If the pressure is from the weight of several hundred grams in the pressure distribution in the range of the dog's foot, Is. That is, if it is more than this threshold, recognition processing S2303 is applied as a dog. An individual recognition unit 2205 performs individual recognition of a dog. Individual recognition of the dog is performed using the relationship between the magnitude of the pressure, the stride where the pressure occurs, and the time. For example, in a small dog, a small pressure is generated over a short distance. For a large dog, a large pressure distribution is generated with a large stride, so that a certain degree of individual identification is possible. Of course, if the pressure sensor becomes finer and the detected pressure becomes more accurate, the pressure sensor can be placed to identify the individual by only the pressure distribution on the sole of the foot when walking.

  Next, the threshold value B and comparison processing S2304 are similarly performed. This process is a threshold process for detecting the position of the plate 2102 on which the pet food is placed. If there is pressure distribution in the size of a pet dish, the area is detected as the area where the dish is placed. Detailed recognition is performed by the weight detection unit 2206 as a pet food amount detection process S2305. Here, the amount of pet food reduced by constantly monitoring the weight can be detected.

  Since the individual authentication and the existence of the pet near the plate can be confirmed together, the reduced weight is automatically recorded as the amount eaten by the pet in this pet food amount detection process S2305.

  Finally, these sensing results are displayed on the screen of the smartphone 1001 in communication / display processing S2306. When data is not displayed, the detected result is uploaded to the cloud server 1002. FIG. 24 shows an example of a screen displayed on the smartphone 1001. The name of the registered pet, the amount of meal eaten by the pet, and the meal time can be acquired as data.

  In addition, as shown in FIG. 25 (a), some food may be attached to the plate 2102 to sense and acquire pet food data. For example, you may attach the vibration sensor 2501, the weight sensor 2502, and the food component sensor 2503 to the surface which puts the pet food of a plate. Based on information from the vibration sensor 2501, it is possible to detect an action of a pet eating. It's unclear how much the pet ate, but you can only record when you ate it. For example, when the amount given per day is a fixed amount, the amount of meal can be detected only by timing.

  Further, the weight sensor 2502 can detect the amount of pet food placed or the amount reduced by using only a plate. If the food component sensor 2503 is attached together with the weight sensor 2502, the type of pet food can also be determined. For example, even a food component sensor that detects the amount of salt can be identified if the salt content is different in a predetermined pet food.

  Moreover, you may attach the near-infrared spectroscopy apparatus 2504 to a plate. With this spectroscopic device, several components such as moisture and fat content of pet food on the plate 2102 can be detected by spectroscopic measurement. In addition, calorie can be detected by combining the detection results of these components, so that not only the amount of meal but also the calorie intake can be detected at the same time. Moreover, since this also becomes a near-infrared camera, it becomes possible to identify an individual by dog face recognition or the like.

(Application of vital data to pet insurance services)
Vital data such as the number of steps, heart rate, breathing, sleep, and meal are valuable data for managing the health of the pet. An example of applying this data to various pet insurance services will be described below.

  FIG. 26 is a diagram showing a relationship between a pet insurance subscriber, a pet 2601, a pet insurance company 2602, and an animal hospital 2603 when the vital data is used for pet insurance. The pet insurance subscriber and the pet 2601 use the vital sensor (other than the one described in the present embodiment) to acquire the vital data of the pet. When a pet becomes ill, he / she receives medical care at an animal hospital 2603 and pays medical expenses to the animal hospital at that time. Thereafter, the pet insurance subscriber applies for insurance to the pet insurance company 2602 and receives the insurance money. The pet insurance company 2602 collects not only vital data but also data on pet illnesses and treatment costs incurred according to insurance applications from subscribers.

  In this example, the pet insurance subscriber first rebuilds all the medical expenses, but as shown in FIG. 27, he / she pays 30% share at the animal hospital like a human being, There is also a pattern of applying for medical expenses. Even in this case, there is no change in the fact that three data are collected by the insurance company.

  As shown in FIG. 28, these three data indicate all the relationships from vital data to illness and medical expenses data, and thus become valuable assets that can realize new services utilizing them. Specifically, machine learning (deep learning) using vital data such as meals of past data as input explanatory variables, disease data of past data a little later than that, and treatment cost data as objective variables And regression), a model for predicting medical expenses and disease risk can be generated from vital data. With this predictive model, you can determine what kind of dog you are eating is at risk of getting sick in the future and how much you are at risk of spending treatment. Various services can be provided to insurance companies and insurance subscribers on vital data acquired by this prediction model. Specific service examples will be described below.

  For example, by predicting a medical cost risk that occurs within one year from the acquired vital data, it is possible to realize a risk-divided pet medical insurance that can set an optimal insurance premium. FIG. 29 shows a process flow diagram for calculating the insurance premium for this insurance.

  First, processing S2901 for extracting vital data for a certain period is performed. In this example, the average value, the minimum value per day, and the maximum value are calculated by looking at data for one month for a certain period. Next, processing S2902 for predicting future medical expenses for the next year using the medical cost prediction model created from the accumulated data is performed. As shown in FIG. 28, the medical cost prediction model is obtained by machine learning using past data as an average variable, daily minimum and maximum values as explanatory variables, and treatment cost (= medical cost) data as an objective variable. Generate.

  Note that it is not necessary to use an average value, minimum value, maximum value, or the like, and all data may be input as explanatory variables, or only a predetermined data range may be input.

  Next, processing S2903 for setting the health rank from the predicted annual medical expenses is performed. For example, the rank is determined by a table as shown in FIG. In this example, for example, if the predicted annual medical cost is between 5,001 and 40,000 yen for a 5-year-old Shiba Inu, the health AA rank is determined. These health ranks may be created for each dog type as in the example, or the dog type data may be incorporated as an explanatory variable in the prediction model and may be determined including the dog type on the model.

  Next, processing S2904 for determining the premium for the next month or the next year according to the rank is performed. Set by referring to the insurance product table that sets the insurance premiums that will be insured from the predicted medical expenses and subscriber distribution for the following year.

  Finally, screen processing S2905 for presenting to the user as an insurance premium is performed. The screen processing may be presented to the user with a smartphone or a personal computer, or may be printed on paper and sent, and any means for presenting to the user is not limited.

  The above is to optimize the insurance premiums for the following month or year based on the data after enrollment (for example, discount), but there is also insurance that discounts insurance premiums on the assumption that it will be implemented with the commitment of health measures at the time of enrollment Is possible. For example, as shown in FIG. 31, a health plan is presented from an insurance company at the time of enrollment, and monitoring is performed with a vital sensor to determine whether the health plan is being implemented. This is a mechanism for penalizing premiums if they are not implemented.

  This will be described with reference to the processing flow of FIG. First, processing S3101 for detecting a disease risk of a dog type is performed. This is also modeled by data analysis shown in FIG. A predictive model is generated from machine learning with the explanatory variable as the breed and age of the past data accumulated and the objective variable as the disease data or treatment cost data. This makes it possible to predict the disease risk by dog type and age.

  Next, the process proceeds to processing S3102 for extracting a countermeasure for reducing a disease risk. Similar to the previous data analysis, a predictive model is generated from machine learning using the breed and age as well as vital data as explanatory variables and the objective variable as disease data or treatment cost data. Using this model, it becomes possible to predict how the medical expenses will change when the vitality data is further changed by the insurance breed's breed and age. In other words, it becomes possible to detect what kind of vital data will be kept and how much the medical cost will be reduced by the breed and age.

  For example, in the case of a 5-year-old Shiba Inu, the average medical expenses for the next year is 30,000 yen, but if you walk an average of 10,000 steps a day, it will be 10,000 yen. it can. In other words, this is the same as a dog walking about 10,000 steps can get a medical insurance premium. In addition, it is possible to predict how much medical expenses will be reduced due to changes in vital data such as meals and exercise, such as a medical expenses of 3,000 yen for dogs who are taking supplement A.

  In the above, the average value of the vital data at that time is input as an explanatory variable. However, if there is sufficient data, it is possible to generate a prediction model using the vital data change amount as an explanatory variable instead of the vital data.

  Next, processing S3103 for generating and presenting to the user a health plan for the target dog extracted from the predicted medical expenses is performed. This processing is performed by selecting a prediction model that is highly effective and easy for the user to implement from the previous prediction model. For example, the user presents and selects several plans, such as 10,000 steps / day, or supplement A once a week. Alternatively, as shown in FIG. 32A, a hurdle for implementation may be investigated by a preliminary questionnaire from a user on the system side, and a health improvement plan may be automatically generated from the result data and proposed. In this case, an item that can sense the improvement effect is desirable. For example, since the number of steps and meals can be sensed, these items are displayed with priority.

  After the user subscribes to the insurance with this health plan, the process proceeds to a process of confirming the execution with the vital sensor data. First, data acquisition processing S3104 is performed from vital data possessed by the insurance subscriber. The acquired vital data is compared with the health plan presented in step S3103 (S3105). If the plan is executed as planned, the premium continuation process S3106 is performed as shown. If the plan is not as planned, the premium is increased. The application S3107 is executed. These processes not only change the process of insurance premium claim amount, but also present the results to the user as shown in FIG. 32 (b) to improve motivation. In this example, only the result is displayed, but it is better if it can be displayed from a warning such as “There is a risk that the insurance premium may increase”. On the other hand, it is also effective to praise the user and increase motivation like "You are following a health plan and are very good".

  As described above, predicting the future illness risk or medical expenses of the insured person from the attribute information and vital data such as meals, exercise, etc., and the health plan and insurance premium generated based on the prediction information are insured If there is a step of presenting to the subscriber or the insured person, and comprising the step of confirming whether or not the insurance plan is implemented by a vital sensor after joining the insurance plan, the implementation of the plan can be confirmed to the insured person A feature of the present invention is an insurance premium management system that continues the first insurance premium and changes the amount of insurance claimed to the second insurance premium if implementation cannot be confirmed. According to this, since the insurance subscriber can apply the activity for the future health improvement to the discount of the current insurance premium, the motivation for the user's health improvement is improved.

  The system may further include a step of displaying a warning in advance when the possibility that the implementation meets the standard is low after the step of confirming the implementation of the health plan of the insured person. According to this, the user can take not only the result but also a proactive measure, and there is an effect that the possibility of withdrawal from the insurance due to the premium premium is reduced.

  Vital data may also be used to detect fraudulent insurance applications. Unlike humans, pets are difficult to authenticate, and if you have several pets of the same type, there is a risk that you will be enrolled in insurance and apply for insurance with another similar pet. As a countermeasure against this impersonation, detection is performed using a vital sensor and a prediction model.

  As a concept of processing, even if a wearable vital sensor is attached to an insured pet only before and after an illness, it can be detected as spoofing because of sudden fluctuation of vitals. In addition, since time-series vital data by disease has been accumulated in the past, it is possible to create a vital change model at the time of disease from these data and detect whether the wearable vital sensor has been replaced by pattern matching. How it works.

  Processing will be described with reference to FIG. First, vital data around a medical treatment date received at an animal hospital is acquired for a disease subject to insurance application (S3301). At this time, the acquisition period may be changed depending on the type of illness. For example, vital data is acquired over a long range of one week to several months in the case of a sudden disease such as a cold for one week before and after, and in the case of a disease that progresses gradually, such as cancer.

  Next, processing S3302 which extracts the prediction vital model of the disease for which insurance is applied is performed. This is also generated from the accumulated data shown in FIG. This is a pet of the same type rather than a prediction model, and before and after vital data at the time of having the disease is extracted, and feature points common to those data are extracted. For example, in the case of a cold, the amount of activity and exercise intensity gradually decline, and the characteristics of vitals appear according to the disease.

  Similarity determination processing S3303 between this model and the vital data of the pet of the insurance application is performed. If they are similar, there is a high possibility that the pet from which the vital data has been acquired has the applied disease, and insurance payment processing S3304 is performed.

  If the similarity cannot be obtained, or if the vitals suddenly fluctuate, there is a possibility that a replacement has occurred, so the process branches to the confirmation step. Thereafter, it may be confirmed mechanically, but in this embodiment, confirmation request S3305 is made to a human operator. If there is no problem in the confirmation S3306, the insurance money is paid (S3304). In addition, if there is fraud by confirmation, insurance money is not paid (S3307).

  Requesting confirmation from the operator seems inefficient at first glance, but confirmation through humans is a major deterrent to fraud when considering the effect of suppressing fraud of insurance subscribers. For example, it is known that if a person makes a random check and makes a call, the number of persons who commit fraud will be greatly reduced simply by telling the subscriber at the time of contract.

  Further, the wearable vital sensor may be used as an individual identification ID device in the same manner as the microchip other than when applying for insurance. In this case, it is necessary to detect whether it is attached to the original individual when the removal occurs. The process for this is demonstrated using FIG.

  Here, a collar-type wearable vital sensor is assumed. First, it is detected S3401 whether or not the wearable device has been detached. For the attachment / detachment, for example, a switch that electrically detects that the collar has been removed may be mounted, or may be detected by acquiring acquired vital data.

  Next, processing S3402 which displays on the smart phone 1001 of the insurance subscriber as a history that it has been detached is performed. An example of the screen is shown in FIG. Thus, the effect of fraud suppression can be expected simply by showing the user to the desorption history.

  When the wearable vital sensor is attached again, vital data for a certain period is acquired S3403. By comparing this vital data with its own vital data acquired in the past in S3404, it is identified whether or not they are the same individual. Similarity determination processing S3405 is performed, and if it is the same individual, authenticated processing S3406 is performed as the same individual. If a different vital is issued, the same individual re-authentication process S3407 is performed. The re-authentication process is confirmed by telephone or by sending a photo. Any other method may be used as long as it can be confirmed that they are the same individual.

  In addition, a disease sign may be detected from the information of the vital sensor, and medical advice may be provided to the insured person. When a medical recommendation is made, a pet insurance subscriber is notified in cooperation with an animal hospital as shown in FIG.

  The pet insurance subscriber and the pet 2601 pay vital data and insurance premiums to the pet insurance company 2602. The pet insurance company 2602 pays the insurance money to the pet insurance subscriber in accordance with the insurance application. This is the basic operation of a pet insurance company, but the healthier the pet, the better it is, and the insurance company can reduce the amount of insurance paid. The above can be realized if early disease risk can be detected, medical advice is provided, and the subscriber can go to the animal hospital 2603. A mechanism for this is realized using a vital sensor.

  An outline of the process will be described first. A prediction model is generated from the vital data received from the pet insurance subscriber and the pet 2601 and the accumulated data of all the insurance subscribers in the past, and the disease risk of the subject is detected. When there is a strong suspicion of a disease risk, a pet insurance company 2602 directly makes a medical recommendation to pet insurance subscribers and pets 2601. In this case, notification is made by a smartphone application or telephone. If there is a strong suspicion, but the judgment cannot be made, the vital data is sent to the animal hospital 2603 and the animal hospital asks for confirmation. If the animal hospital determines that the suspicion of illness is strong, the animal hospital recommends medical care to pet insurance subscribers. This medical recommendation makes it easier for an insurance company to increase the amount of medical expenses that a pet insurance subscriber and the medical care of a pet 2601 have than the normal insurance burden ratio.

  System processing will be described with reference to FIG. First, feature extraction processing S3702 is performed from vital data for a certain period. The certain period is in the range of several days to several weeks from the viewpoint of processing amount, but it is not necessary to be limited to this range. As the feature quantity, some vital data feature quantities such as time fluctuations and frequency components are extracted from average values and peaks.

  Next, the disease risk prediction model is compared with this vital data feature amount S3701. The prediction model is generated from the accumulated data shown in FIG. Since each vital data corresponding to each disease is accumulated, a common feature amount is detected from the accumulated vital data for each disease. For detection of the feature amount, a method such as convolution or deep learning may be used, or a method such as mathematical Fourier analysis may be used.

  The feature amount of the disease risk prediction model and the vital data is compared S3703. If the similarity is greater than or equal to the threshold value 1, the procedure proceeds to the medical recommendation step, and if the similarity is lower, the procedure returns to monitoring.

  When the similarity is very high, that is, when the similarity equal to or greater than the threshold value 2 is determined S3704, medical insurance is directly recommended to the user from the insurance company (S3705). If the similarity is relatively small but there is a strong suspicion of disease risk, the animal hospital is notified of the details of the vital data S3706. At this time, as shown in FIG. 38, some latest vital data and normal vital data are sent side by side. Further, the candidate for the disease is sent to the smartphone 1001 on the animal hospital side together with the accuracy and presented.

  On the side of the animal hospital, the received vital data is looked up and medical advice is given to the user at the judgment of the hospital side (S3707). At this time, if in doubt, the veterinarian will call the subscriber from the veterinary hospital and ask the veterinarian whether he / she is recommended by conducting a medical inquiry while looking at the data.

  In addition, with respect to the medical expenses that have been consulted by the medical recommendation, the applied insurance money varies as shown in FIG. With respect to the medical expenses that have been received after receiving the recommendation, the insurance premium is increased (S3902). If the patient is not encouraged to visit the hospital, the insurance premium payment S3903 is made as usual. As shown in FIG. 40, notifying the user of this fact has an effect of increasing the consultation rate by the medical recommendation. Also, since there is a slight decrease in careless visits when there is no medical recommendation, a reduction in insurance payments can be expected.

  In addition, the vital device for pets may be other than those described above. For example, a thermometer, weight, blood pressure, or the like may be used. Also, environmental data other than vital data may be used. For example, information such as room temperature and humidity may be used as one of vital data.

  Although the above processing is presented as a service for pets, it may be applied to humans. In this case, the pet insurance company is replaced with a medical insurance company, the subscriber is not a pet but a human, the animal hospital is replaced with a human hospital, and vital devices are replaced with human vital devices.

  According to the present disclosure, it is possible to provide a lead-type vital measurement device for pets that can reduce the burden of the weight of the device on the pet, which is useful in the technical field of managing pet health.

DESCRIPTION OF SYMBOLS 101 Lead 102 Activity meter main body 103 Battery 104 Collar 201 Power supply part 202 Lead attractive force sensor 203 Acceleration sensor 204 Communication part 205 Lead tension detection part 206 Step count calculation part 207 Dog load detection part 208 Exercise calorie calculation part 209 External communication / display part S301 Sensor data read processing S302 Read tension detection processing S303 Detection tension threshold comparison processing S304 Fixed model step count calculation processing S305 Floating model step count calculation processing S306 Display processing (step count)
S403 Load detection process S404 Exercise calorie calculation process S405 Display process (calorie)
S701 Peak axis extraction processing S702 Smoothing processing S703 Peak detection processing S704 Walking determination processing S901 Overlay processing S902 Smoothing processing S903 Peak detection processing S904 Non-fixed correction processing S905 Walking determination processing 1001 Smartphone 1002 Cloud server 1201 Activity meter Body type)
1401 Power connector (male)
1402 Power connector (female)
1403 Power connector ring 1404 Power connector ring receiver 1501 Pet bed 1502 Transmission antenna 1503 Reception antenna 1504 Weight sensor 1601 Communication unit 1602 Transmission wave generation unit 1603 Antenna switching unit 1604 Doppler processing unit 1605 Weight detection unit 1606 Control unit 1607 Power unit S1701 Weight Detection processing S1702 Weight determination processing S1703 Distribution detection processing S1704 Antenna switching processing S1705 Radar drive processing
S1706 Vital detection processing S1707 Display processing 2101 Mat device 2102 Pet tableware 2201 Pressure sensor 2202 Communication unit 2203 Pressure distribution calculation unit 2204 Threshold judgment unit 2205 Individual recognition unit 2206 Weight detection unit 2207 Control unit 2208 Power supply unit S2301 Weight detection processing S2302 First threshold determination processing S2303 Individual identification processing S2304 Determination processing S2305 Pet food amount detection processing S2306 Communication / display processing 2501 Vibration sensor 2502 Weight sensor 2503 Food component sensor 2504 Near-red spectroscopic device 2601 Pet insurance subscriber and pet 2602 Pet Insurance company 2603 Animal hospital S2901 Vital data extraction processing for a certain period S2902 Future medical cost prediction processing using medical cost prediction model S2903 Health rank Setting processing S2904 Next year (next month) insurance premium determination processing S2905 Screen display processing S3101 Dog-type disease risk detection processing S3102 Disease risk reduction countermeasure extraction processing S3103 Health plan user presentation processing S3104 for target dog Vital sensor data Acquisition processing S3105 Plan execution confirmation determination processing S3106 Insurance premium continuation processing S3107 Insurance premium premium processing S3301 Vital data acquisition processing for medical diseases S3302 Disease prediction vital model extraction processing for medical diseases S3303 Similarity determination processing S3304 Insurance payment processing S3305 Confirm with operator Request processing
S3306 Confirmation determination input processing from the operator (determination is performed by the operator)
S3307 Insurance payment cancellation process S3401 Desorption determination process S3402 Desorption history screen display process S3403 Vital data acquisition process for a certain period S3404 Vital data comparison process with a certain period S3405 Similarity determination process S3406 Authenticated issuance process S3407 Re-authentication Processing S3701 Vital data feature extraction processing for a certain period S3702 Vital comparison processing with disease risk prediction model S3703 Similarity first determination processing S3704 Similarity second determination processing S3705 Medical treatment recommendation notification processing S3706 Data transfer processing to animal hospital S3707 medical treatment recommendation (implemented at the hospital side)
S3901 Medical examination recommendation process S3902 Application medical expenses payment insurance premium processing S3903 Normal insurance payment processing

Claims (1)

A lead type vital measuring device for pets,
A lead tension tension detector that detects the tension of the lead between the pet and the owner;
A step number calculating unit for detecting the number of steps by correcting information from the acceleration sensor according to the tension information;
And a display unit for presenting the step count data to a user.

Images