Body monitoring device, body monitoring method, and program

Abstract

To provide a body monitoring device that can determine the physical condition of a subject with greater accuracy using a simple transmitting device. [Solution] The subject's state is classified based on time-series information of signals from multiple transmitters attached to different locations on the subject's body. If the subject's state is classified into a predetermined category, output information corresponding to that category is output to a predetermined output destination.

Description

Technical Field

[0001] The present disclosure relates to a body monitoring device, a body monitoring method, and a program.

Background Art

[0002] A technique for determining whether a care recipient has fallen based on the acceleration transmitted by a wearable module worn by the care recipient is disclosed in Patent Document 1.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the above-described technique, there has been a demand for a technique that can accurately detect the physical state of a target person, such as falling, using a simpler sensor configuration that does not have functions such as an acceleration sensor for detecting acceleration.

[0005] An object of the present disclosure is to provide a body monitoring device, a body monitoring method, and a program that solve the above-described problems.

Means for Solving the Problems

[0006] A body monitoring device according to an aspect of the present disclosure includes a classification unit that classifies the state of a target person based on time-series information of signals from a plurality of transmitting devices respectively attached to different positions of the target person's body, and an output unit that outputs output information corresponding to the classification to a predetermined output destination when the classification of the target person's state is a predetermined classification.

[0007] A body monitoring method according to one aspect of the present disclosure classifies the state of a subject based on time-series information of signals from multiple transmitters attached to multiple different locations on the subject's body, and outputs output information corresponding to the classification to a predetermined output destination when the classification of the subject's state is a predetermined classification.

[0008] In one aspect of the present disclosure, the program causes the computer of a body monitoring device to function as a classification means for classifying the state of a subject based on time-series information of signals from multiple transmitters attached to multiple different locations on the subject's body, and as an output means for outputting output information corresponding to a predetermined classification to a predetermined output destination when the classification of the subject's state is a predetermined classification. [Effects of the Invention]

[0009] According to one embodiment described above, the physical condition of a subject can be determined with greater accuracy using a transmitter with a simple configuration. [Brief explanation of the drawing]

[0010] [Figure 1] This is the first figure illustrating an example of a body monitoring system according to one embodiment of the present disclosure. [Figure 2] This is a hardware configuration diagram of a body monitoring device according to one aspect of the present disclosure. [Figure 3] This is a functional block diagram of a body monitoring device according to one aspect of the present disclosure. [Figure 4] This is the first figure showing the processing flow of a body monitoring device according to one embodiment of the present disclosure. [Figure 5] This figure shows an image of a fall according to one embodiment of the present disclosure. [Figure 6] This is a second figure showing the processing flow of a body monitoring device according to one embodiment of the present disclosure. [Figure 7] The second figure shows an example of a body monitoring system according to one embodiment of the present disclosure. [Figure 8] This is a third figure illustrating an example of a body monitoring system according to one embodiment of the present disclosure. [Figure 9]This figure shows a functional block of another example of the body monitoring device of the present disclosure. [Figure 10] This figure shows the processing flow of another example of the body monitoring device of this disclosure. [Modes for carrying out the invention]

[0011] Each embodiment will be described below with reference to the drawings. Figure 1 is a first figure showing an example of a body monitoring system according to one embodiment of the present disclosure. As shown in Figure 1, the body monitoring system 100 is configured by a communication connection between a body monitoring device 1 and one or more readers 2. Each reader 2 receives signals transmitted by multiple transmitters 3 worn by the person being monitored and transmits those signals to the body monitoring device 1. One example of a transmitter 3 is an RFID tag. The RFID tag may be equipped with a power source such as a battery and actively transmit signals at predetermined intervals. The person being monitored wears the transmitters 3 at multiple different locations on their body, such as their wrists, knees, and neck. The transmitters 3 may repeatedly transmit signals at short intervals, such as tens of milliseconds.

[0012] The body monitoring device 1 acquires information contained in the signal received by the reader 2. Based on this information, the body monitoring device 1 determines whether or not the monitored person has fallen. The body monitoring device 1 may also determine the type of fall. The types of falls can vary, for example, a fall on one's knees, a fall backward, a fall forward, etc. The types of falls may also have classifications such as the severity of the fall.

[0013] The signal transmitted by the transmitting device 3 may include the identifier of the transmitting device (transmitting device ID), the identifier of the monitored person (target person ID), the physical characteristics of the monitored person (height, weight, age, gender), etc. The signal transmitted by the transmitting device 3 may be received by at least one of one or more readers 2 located within a communicable distance and transferred to the body monitoring device 1. When the body monitoring device 1 receives signals from a plurality of readers 2 from one transmitting device 3, the information included in the signal at the earliest reception time in the reader 2 may be specified as the information to be used for subsequent processing. Alternatively, the body monitoring device 1 may specify, based on the identifier of the transmitting device 3 included in the signal, etc., the information of the signal obtained from the reader 2 that has received all the signals of the plurality of transmitting devices 3 worn by a predetermined monitored person as the information to be used for subsequent processing.

[0014] The body monitoring device 1 may have the function of the reader 2. In this case, the body monitoring device 1 directly receives the signal transmitted by the transmitting device 3.

[0015] Figure 2 is a hardware configuration diagram of a body monitoring device according to an aspect of the present disclosure. As shown in Figure 2, the body monitoring device 1 is a computer equipped with hardware such as a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, a storage device 104, and a communication module 105. The body monitoring device 1 may be a PC (Personal Computer), a cloud server, or the like. When the body monitoring device 1 has the function of the reader 2, it includes an RFID antenna 106.

[0016] Figure 3 is a functional block diagram of a body monitoring device according to an aspect of the present disclosure. [[ID=##]] As shown in Figure 3, when the CPU 101 of the body monitoring device 1 executes a body monitoring program, it exhibits functions such as a control unit 11, a classification unit 12, a region specifying unit 13, an output unit 14, a storage unit 15, a learning unit 16, and a learning model storage unit 17.

[0017] The control unit 11 controls each functional unit. The classification unit 12 classifies the state of the subject based on the time-series information of the signals of the plurality of transmitting devices 3 respectively attached to different positions of the body of the monitored subject. The area specifying unit 13 specifies the type of the spatial area corresponding to the position of the monitored subject based on the time-series information of the signals of each of the plurality of transmitting devices. The output unit 14 outputs output information corresponding to the classification to a predetermined output destination when the classification of the state of the subject is a predetermined classification. The storage unit 15 records learning data for generating a learning model used by the classification unit 12 to classify the state of the subject in the storage unit.

[0018] The learning unit 16 generates a learning model used by the classification unit 12 to classify the state of the subject. The learning model storage unit 17 stores the learning model. The learning unit 16 and the learning model storage unit 17 may not be provided as functions of the body monitoring device 1 and may be functions exhibited by other learning devices.

[0019] Then, by the above-described functions, the body monitoring device 1 classifies the state of the subject based on the time-series information of the signals of each of the transmitting devices 3 respectively attached to different positions of the body of the subject, and when the classification is a predetermined classification, outputs output information corresponding to the classification to a predetermined output destination. The time-series information of the signals of each of the transmitting devices 3 may include transmission information such as transmission time, transmission device ID, monitored subject ID, information indicating physical characteristics (height, weight, age, gender), and reception intensity on the reception side of the signal.

[0020] The body monitoring device 1 may input the time-series information of the signals of each of the plurality of transmitting devices in a learning model obtained by machine learning the relationship between the time-series information of the signals of each of the plurality of transmitting devices over a measurement time period and correct information indicating the classification of at least one of the states of the monitored subject being fallen or non-fallen, and obtain the state of the subject being either fallen or non-fallen to classify the state of the monitored subject.

[0021] Furthermore, the body monitoring device 1 may be configured to input the time-series information of the signals from each of the multiple transmitters 3 over a measurement period, physical characteristic information, and correct information indicating the classification of the subject's state as either fallen or not fallen, into a learning model obtained by machine learning the relationship between these two states, and input the time-series information of the signals from each of the multiple transmitters 3 and the physical characteristic information of the person being monitored, thereby acquiring the state of the person being monitored as either fallen or not, and classifying the state of the person being monitored.

[0022] Furthermore, the body monitoring device 1 may input the time-series information of the signals from each of the multiple transmitters 3 into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters 3 over the measurement time period and the correct information indicating the classification of the monitored person's state into one of multiple different types of falls or non-falls, thereby acquiring the monitored person's state into one of multiple different types of falls or non-falls and classifying the monitored person's state.

[0023] Furthermore, the body monitoring device 1 may use machine learning to obtain a learning model that shows the relationship between the time-series information of the signals from each of the multiple transmitters 3 over the measurement time period, physical characteristic information, and correct information indicating the classification of the monitored person's state into one of multiple different types of falls or non-falls, by inputting the time-series information of the signals from each of the multiple transmitters 3 and the physical characteristic information into the learning model, thereby acquiring the monitored person's state as one of multiple different types of falls or non-falls and classifying the person's state.

[0024] Figure 4 is a first diagram showing the processing flow of a body monitoring device according to one embodiment of the present disclosure. The learning unit 16 of the body monitoring device 1 pre-generates a learning model that the classification unit 12 uses to classify the subject's condition.

[0025] Specifically, the learning unit 16 generates a learning model by machine learning the relationship between the time-series information of signals from multiple transmitters 3 worn by a particular person, which is recorded in the storage device 104, and the correct information indicating the classification of the person's state as either fallen or not fallen, using time-series information of signals from a large number of people in either fallen or not fallen states (step S101). Using this learning model, the body monitoring device 1 can take the time-series information of signals from each of the multiple transmitters 3 as input, acquire the state of the person in either fallen or not fallen, and classify the state of the person being monitored.

[0026] Alternatively, the learning unit 16 generates a learning model by machine learning the relationship between the time-series information of signals from each of the multiple transmitters 3 worn by a person being monitored, the physical characteristics information of the person being monitored, and the correct information indicating the classification of the person's state as either fallen or not fallen, which is recorded in the storage device 104. Using this learning model, the body monitoring device 1 can take the time-series information of signals from each of the multiple transmitters 3 and the physical characteristics information of the person being monitored as input, acquire the state of the person being monitored as either fallen or not, and classify the state of the person being monitored.

[0027] Alternatively, the learning unit 16 generates a learning model by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters 3 and the correct information indicating the state of the monitored person, which is classified into one of several different types of falls or non-falls. Using this learning model, the body monitoring device 1 can take the time-series information of the signals from each of the multiple transmitters 3 as input, acquire the state of the monitored person, which is either one of several different types of falls or non-falls, and classify the state of the monitored person.

[0028] Alternatively, the learning unit 16 generates a learning model by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters 3, the physical characteristic information, and the correct information indicating the state of the monitored person, which is either one of several different types of falls or a state of not falling. Using this learning model, the body monitoring device 1 can take the time-series information of the signals from each of the multiple transmitters 3 and the physical characteristic information as input, acquire the state of the monitored person, which is either one of several different types of falls or a state of not falling, and classify the state of the person.

[0029] The learning unit 16 stores the information necessary to construct the generated learning model (for example, weight coefficients that make up the neural network) as learning model data in the learning model storage unit 17 (step S102).

[0030] In generating the learning model described above, the learning unit 16 acquires learning data (data contained in the signal transmitted by the RFID tag, which is the transmitting device 3) that is pre-recorded in a storage unit such as the memory device 104, and is received via the reader 2 when the monitored person falls or does not fall. This learning data is recorded in the storage unit and linked to the respective labels for falling and not falling, and the learning unit 16 reads the learning data along with these labels. In the learning process step, the learning unit 16 learns the read learning data and generates a learning model by machine learning that can correctly classify whether the monitored person has fallen or not based on the data contained in the signal transmitted by the monitored person's transmitting device 3. Information contained in the signal acquired when actually determining whether the monitored person has fallen or not may also be saved as learning data and used to update the learning model.

[0031] The learning unit 16 may also evaluate the learning model. In the evaluation process, the learning unit 16 inputs evaluation data (information contained in the signal transmitted by the transmitter 3 when the vehicle falls over or does not fall over) into the generated learning model, and obtains the classification result indicating whether the vehicle falls over or does not fall over, which is output by the learning model. If the accuracy rate of the classification result is above a predetermined threshold, the learning unit 16 terminates the learning process. If the accuracy rate is below the predetermined threshold, the learning unit 16 may update the evaluation data and retrain the vehicle.

[0032] The signal transmitted by the transmitter 3 can include information about the physical characteristics of the person being monitored, such as height, weight, age, and gender, as parameters. By including parameters containing information about physical characteristics as features in the training data during learning, the tendency to fall can also be learned.

[0033] Figure 5 shows an image of a fall according to one embodiment of the present disclosure. The example shown in Figure 5 illustrates a case where the monitored person has transmitters 3 attached to three locations: near the neck, on the wrist of one arm, and on the knee of one leg. The monitored person in Figure 5(a) is a relatively tall, healthy individual. The monitored person in Figure 5(b) is an elderly person. As shown in Figure 5(a), if the monitored person nearly falls for any reason and puts one knee on the ground, the transmitter 3 attached near the neck will begin to fall from a relatively high position, and when the person puts one knee on the ground, the height of the transmitter 3 will move to a relatively low height, which is a considerable distance from the position where the monitored person would be standing. Also, as shown in Figure 5(b), if the person who falls is older, the movement of the transmitter 3 in three-dimensional space may initially be gradual when not falling. Alternatively, in older monitored persons, the position of the transmitter 3 is often at ground (floor) height when falling. Furthermore, the movement of the transmitter 3 during a fall also changes depending on height and age. By linking physical characteristics with the movement of the transmitter 3 using machine learning, it is possible to improve the accuracy of classifying falls and non-falls based on physical characteristics.

[0034] Figure 6 is a second diagram showing the processing flow of a body monitoring device according to one embodiment of the present disclosure. Next, we will explain the fall detection process performed by the body monitoring device 1. Each transmitter 3 attached to the person being monitored transmits a signal at a predetermined interval, such as every 10 milliseconds (step S201). The signal may include the transmission time, transmitter ID, person being monitored ID, and information indicating physical characteristics (height, weight, age, gender). This information included in the signal will be called the transmitted information. The transmitted information does not have to include the person being monitored ID. In this case, the body monitoring device 1 may store in advance the transmitter ID of the transmitter 3 attached to the person being monitored and the person being monitored ID that identifies the person being monitored, and the body monitoring device 1 may identify the person being monitored that is linked to the transmitter ID included in the received signal and include it in the transmitted information.

[0035] When the reader 2 receives a signal, it transmits the reception time, reception intensity, and transmission information obtained from the signal to the body monitoring device 1 (step S202). Alternatively, if the body monitoring device 1 has the functions of the reader 2, it may directly receive the signal transmitted by the transmitter 3, detect the reception time and reception intensity of the signal, and obtain the transmission information contained in the signal. The control unit 11 outputs the set of reception time, reception intensity, and transmission information to the storage unit 15. The storage unit 15 records the reception time, reception intensity, and transmission information in a transmission information table (step S203).

[0036] Furthermore, if signals transmitted simultaneously from a single transmitter 3 are received by multiple readers 2, the storage unit 15 may calculate the spatial coordinates of the transmitter 3 based on the received signal strength at each reader 2 and record the value of these spatial coordinates in association with the reception time, received signal strength, and transmission information. Alternatively, if signals transmitted simultaneously from a single transmitter 3 are received by multiple readers 2, the storage unit 15 may combine the transmission information of these multiple signals and record it in a transmission information table in association with the reception time, received signal strength at each reader 2, the reader ID of the receiving reader 2, spatial coordinates, and transmission information. Signals with the same transmitter ID and received by each reader 2 within a predetermined time difference range can be presumed to be signals transmitted from a single transmitter 3. The storage unit 15 pre-stores the coordinates of the reader 2 relative to a certain origin in three-dimensional space and calculates the spatial coordinates of the transmitter 3 based on these coordinates and the received signal strength. Known techniques may be used for calculating these spatial coordinates. The storage unit 15 does not need to calculate the spatial coordinates of the transmitter 3.

[0037] Each of the transmitters 3 worn by one or more designated monitored individuals transmits a signal, thereby accumulating information such as reception time, reception strength, transmission information, and spatial coordinates in the transmission information table.

[0038] The control unit 11 of the body monitoring device 1 instructs the classification unit 12 to start processing at predetermined intervals or based on the detection of a trigger for the start of the determination process (step S301). The classification unit 12 monitors the status of one monitored person at predetermined intervals, for example, every 10 seconds. Specifically, the classification unit 12 identifies the ID of a monitored person from the monitored person table. Based on the reception time included in the time from the first time, which indicates the start of the interval to identify a predetermined time interval, to the second time, which indicates the end of the interval after the first time, the classification unit 12 identifies and reads from the transmission information table the transmission information including the ID of the identified monitored person, the reception strength at each reader 2 associated with it, the reader ID of the receiving reader 2, spatial coordinates, etc. (step S302). In this way, the classification unit 12 can obtain the transmission information of signals transmitted from multiple transmitters 3 worn by a certain monitored person, the reception strength at each reader 2 of the signal containing that transmission information, the reader ID of the receiving reader 2, and spatial coordinates within a predetermined interval (for example, 10 seconds). This acquired information will be referred to as judgment information. The judgment information may not include one or more of the following: the received signal strength at each reader 2 containing the transmitted information, the reader ID of the receiving reader 2, or the spatial coordinates. The transmitted information included in the judgment information may not include one or more of the information indicating the physical characteristics (height, weight, age, gender) mentioned above. The transmitted information may not include physical characteristics at all.

[0039] The classification unit 12 inputs judgment information into a neural network using a learned model (step S303). As a result, the classification unit 12 obtains a judgment result from the neural network indicating whether the person has fallen or not. If the learned model is not a model generated using information indicating physical characteristics, the transmitted information included in the judgment information that the classification unit 12 inputs into the neural network does not need to include information indicating physical characteristics.

[0040] If the learning model is generated as a model that outputs a fall type, the classification unit 12 obtains a determination result indicating whether a fall occurred or not, and also indicating the fall type (step S304). The fall type may indicate, for example, a type indicating that the person fell forward, a type indicating that the person fell backward, a type indicating that the person fell and the right knee touched the ground, a type indicating that the person fell and the left knee touched the ground, the degree of force with which the person fell, etc.

[0041] When the classification unit 12 obtains a determination result, it outputs the determination result to the output unit 14. The output unit 12 outputs the determination result to a predetermined output destination connected via communication (step S305). The predetermined output destination may be a predetermined terminal device, or it may be a warning sound generator installed in the room or indoors where the monitored person is located. The output unit 14 may output the determination result to the predetermined output destination only if the determination result indicates a fall. In this case, the determination result may include information indicating a fall and information indicating the type of fall. The output destination device may output presentation information in a manner corresponding to the type of fall. The presentation information may be screen information displayed on a display or a warning sound emitted from a speaker. The output destination of the determination result may be an output device installed in a predetermined facility (such as a train station or city hall), or it may be a computer managed by a company that provides monitoring services, such as a security company.

[0042] If the classification unit 12 determines that the result indicates a fall, it may output the determination information to the area identification unit 13. In this case, the area identification unit 13 may detect the location where the monitored person fell based on the spatial coordinates of the transmitter 3 included in the determination information. For example, when the body monitoring device 1 monitors a monitored person who is indoors, it stores identification information of areas such as stairs, living room, toilet, and bathroom, along with their spatial coordinates, in advance. The area identification unit 13 identifies the identification information of an area close to the spatial coordinates of the transmitter 3 included in the determination information and outputs the identification information (such as the name) of that area to the output unit 14 (step S306). The output unit 14 may further output the identification information of the area where the monitored person fell to another output destination.

[0043] If the storage unit 15 determines that the result indicates a fall, it associates the identifier indicating a fall with the determination information used to output the determination result and records it in the storage unit. If the storage unit 15 determines that the result indicates no fall, it associates the identifier indicating no fall with the determination information used to output the determination result and records it in the storage unit (step S307). The learning unit 16 may learn the determination information and the identifier indicating a fall or no fall (correct data) using machine learning and update the learning model (step S308).

[0044] (Other Embodiment 1) Figure 7 is a second figure showing an example of a body monitoring system according to one embodiment of the present disclosure. In other embodiments, multiple body monitoring devices 1 may communicate with each other to constitute a body monitoring system 100. In this case, each body monitoring device 1 transmits to other body monitoring devices 1 pairs of information, such as a judgment result (fall or not fall) and judgment information, which it has stored in its memory based on the signals it has received. Each body monitoring device 1 may use supervised learning to machine learn the relationship between the judgment result and judgment information stored in its memory based on the signals it has received, and the relationship between the judgment result and judgment information received from other devices, in order to generate the above-mentioned learning model.

[0045] (Another Embodiment 2) The physical monitoring device 1 may communicate with the electronic medical record management device 5. The electronic medical record management device 5 stores data on the physical characteristics of the person being monitored and is installed in hospitals, welfare facilities, etc. The physical monitoring device 1 may send request information for physical characteristics, including the person being monitored's ID, to the electronic medical record management device 5. As a result, the electronic medical record management device 5 may read the physical characteristic information associated with the person being monitored's ID from its database and send it to the physical monitoring device 1, allowing the physical monitoring device 1 to acquire the person's physical characteristic information. The physical monitoring device 1 may generate a learning model using the physical characteristic information acquired from the electronic medical record management device 5, or it may input the physical characteristic information acquired from the electronic medical record management device 5 into the learning model to obtain a judgment result for determining the state of the person being monitored. This eliminates the need for administrators, etc., to input the physical characteristic information of the person being monitored.

[0046] (Other Embodiment 3) Figure 8 is a third figure showing an example of a body monitoring system according to one embodiment of the present disclosure. In the above-described embodiment 1, the body monitoring device 1 acquires information contained in the signal transmitted by the transmitter 3 via the reader 2. However, another device may acquire the information contained in the signal transmitted by the transmitter 3 via the reader 2, and the body monitoring device 1 may acquire that information. For example, the electronic medical record management device 5 may acquire the reception time, reception strength, and transmission information based on the signal transmitted by the transmitter 3 from the reader 2, and use that information to perform the processing shown in embodiment 1.

[0047] (Another Embodiment 4) The output unit 14 of the body monitoring device 1 may, when the determination result indicates a fall, transmit a predetermined control signal to an output destination installed in the area identified by the area identification unit 13. For example, if the area identification unit 13 identifies an escalator, the output unit 14 may output a control signal indicating a stop to the escalator's emergency device, thereby causing the escalator to be stopped immediately. This prevents secondary accidents such as falls of the monitored person.

[0048] (Other Embodiments 5) The process described above calculates the spatial coordinates of the monitored person's falls, and these spatial coordinates can be used to identify conditions that make falls more likely. This allows for the identification of the causes of falls in the environment of facilities where monitored persons reside, and can be used to inform future facility design. For example, by analyzing statistical data on the number of falls, it is possible to identify characteristics such as floor material, step locations, and stair heights in areas with a high number of falls, thereby investigating conditions that are likely to cause falls. Another example is the ability to determine the optimal handrail height from data on the number of falls by height.

[0049] Specifically, the physical monitoring device 1 stores the identification information of each of several predetermined areas, linked to the number of times a fall was determined in the judgment results. The physical monitoring device 1 calculates the fall rate for each of the multiple areas in the total number of times a fall was determined, and records it linked to the identification information of each area. The physical monitoring device 1 also stores the identification information of each area, linked to the physical characteristic information included in the judgment information used to calculate the state in which a fall was determined. This allows administrators to make improvements to each area of ​​the facility, etc., based on the fall rate and physical characteristic information corresponding to the predetermined area.

[0050] The body monitoring system 100 can be operated on any scale by installing multiple leaders 2 to match the number of people being monitored and the size of the facility in which it is operated.

[0051] In the example described above, the state of the monitored person was classified based on the learning model's determination result whether they were in a fallen or not. However, in other examples, other states of the monitored person may be classified based on the learning model's determination result. For example, the body monitoring device 1 may determine whether the person is in a sleep state or a non-sleep state, or whether they are walking or not, based on determination information such as signals transmitted from the transmitter 3, received intensity, spatial coordinates, and physical characteristic information.

[0052] Furthermore, in the process described above, the body monitoring device 1 uses the spatial coordinates of the transmitter 3 as information for determination. However, it is also possible to generate a learning model using the velocity and acceleration included in the signal from the transmitter 3, and the direction of movement that can be calculated based on those velocities and accelerations, or to obtain a determination result of the body state using the learning model.

[0053] Figure 9 shows a functional block diagram of another example of a body monitoring device. Figure 10 shows the processing flow for another example of a body monitoring device. The body monitoring device 1 includes at least a classification means 91 and an output means 92. The classification means classifies the state of the person being monitored based on time-series information of signals from multiple transmitters 3, each attached to multiple different locations on the person being monitored (step S301). The output means outputs output information corresponding to a predetermined classification to a predetermined output destination when the classification of the monitored person's status is a predetermined classification (step S302).

[0054] In related technologies, fall detection using RFID tags involves attaching an antenna close to the floor to detect when the RFID tag drops below a certain height. In this case, it was difficult to distinguish between a person sitting or lying down intentionally and a fall. Furthermore, other fall detection methods also presented difficulties in managing a large number of monitored individuals over a wide area.

[0055] On the other hand, the processing of the body monitoring device 1 in this disclosure focuses on the absolute and relative positions of multiple RFID tags and detects the height, speed, direction of movement, and movement specific to falls, thereby distinguishing them from other actions and enabling accurate fall detection. Furthermore, by recording information on the physical characteristics of the person being monitored on the RFID tags, it is possible to detect falls in a way that is tailored to the individual's situation, characterized by height, weight, age, gender, etc.

[0056] The body monitoring device 1 of this disclosure has been described above. According to the above process, the physical condition of the subject can be determined with greater accuracy using a transmitter with a simple configuration.

[0057] While the present disclosure has been described above with reference to embodiments, the present disclosure is not limited to the embodiments described above. Various modifications to the structure and details of the present disclosure may be made that can be understood by those skilled in the art within the scope of the present disclosure.

[0058] Some or all of the above embodiments may also be described as follows, but are not limited to the following:

[0059] (Note 1) A classification means for classifying the state of the subject based on time-series information of signals from multiple transmitters attached to multiple different locations on the subject's body, An output means that outputs output information corresponding to a predetermined classification to a predetermined output destination when the classification of the subject's condition is a predetermined classification, A body monitoring device equipped with the following features.

[0060] (Note 2) The classification means inputs the time-series information of the signals from each of the multiple transmitters into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state as either fallen or not fallen, thereby acquiring the subject's state as either fallen or not fallen and classifying the subject's state. The body monitoring device described in Appendix 1.

[0061] (Note 3) The classification means classifies the state of the subject based on the subject's physical characteristics information, based on the information contained in the signal transmitted from the transmitter. The body monitoring device described in Appendix 1.

[0062] (Note 4) The classification means inputs the time-series information of the signals from each of the multiple transmitters and the physical characteristic information into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the state of the subject, which is classified as either fallen or not fallen, in order to acquire the state of the subject, which is either fallen or not, and classify the state of the subject. The body monitoring device described in Appendix 3.

[0063] (Note 5) The classification means inputs the time-series information of the signals from each of the multiple transmitters into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state into one of multiple different types of falls or non-falls, thereby acquiring the subject's state as one of the multiple different types of falls or non-falls and classifying the subject's state. The body monitoring device described in Appendix 1.

[0064] (Note 6) The classification means inputs the time-series information of the signals from each of the multiple transmitters and the physical characteristic information into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the state of the subject, which is classified into one of multiple different types of falls or non-falls, in order to acquire the state of the subject, which is one of multiple different types of falls or non-falls, and classify the state of the subject. The body monitoring device described in Appendix 3.

[0065] (Note 7) The system includes a region identification means that identifies the type of spatial region corresponding to the location of the subject based on the time-series information of the signals from each of the plurality of transmitting devices, The output means outputs the type of the spatial region to a predetermined output destination when the classification of the subject's state is a predetermined classification. A body monitoring device as described in any one of the appendices 2 through 6.

[0066] (Note 8) A survey means for investigating areas with a high frequency of falls based on the state of falls and the number of falls classified for each spatial area, A body monitoring device as described in Appendix 7, comprising the following features.

[0067] (Note 9) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices and the correct information indicating the classification of the state of the subject, which is either fallen or not fallen. A body monitoring device as described in Appendix 2, comprising the following features.

[0068] (Note 10) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices, the physical characteristic information, and the correct information indicating the classification of the subject's state as either fallen or not fallen. A body monitoring device as described in Appendix 4, comprising the following features.

[0069] (Note 11) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices and the correct information indicating the classification of the state of the subject into one of several different types of falls or non-falls, A body monitoring device as described in Appendix 2, comprising the following features.

[0070] (Note 12) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices, the physical characteristic information, and the correct information indicating the classification of the subject's state into one of several different types of falls or non-falls. A body monitoring device as described in Appendix 4, comprising the following features.

[0071] (Note 13) The subject's condition is classified based on the time-series information of signals from multiple transmitters attached to different locations on the subject's body. If the classification of the subject's condition falls under a predetermined category, output information corresponding to that classification is output to a predetermined output destination. Body monitoring methods.

[0072] (Note 14) The relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state as either fallen or not fallen is used to obtain a learning model, into which the time-series information of the signals from each of the multiple transmitters is input to acquire the subject's state as either fallen or not fallen and classify the subject's state. The method of physical monitoring described in Appendix 13.

[0073] (Note 15) Based on the information contained in the signal transmitted from the transmitter, the subject's condition is classified based on the subject's physical characteristics. The method of physical monitoring described in Appendix 13.

[0074] (Note 16) The relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the correct information indicating the classification of the subject's state as either fallen or not fallen is used to obtain a learning model, into which the time-series information of the signals from each of the multiple transmitters and the physical characteristic information are input to acquire the subject's state as either fallen or not fallen and classify the subject's state. The method of physical monitoring described in Appendix 15.

[0075] (Note 17) The relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state into one of several different types of falls or non-falls is obtained by machine learning, and the time-series information of the signals from each of the multiple transmitters is input into the learning model, which then acquires the subject's state as one of the several different types of falls or non-falls and classifies the subject's state. The method of physical monitoring described in Appendix 13.

[0076] (Note 18) The relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the correct information indicating the classification of the subject's state into one of several different types of falls or non-falls is obtained by machine learning, and the time-series information of the signals from each of the multiple transmitters and the physical characteristic information are input into the learning model, which then acquires the subject's state into one of several different types of falls or non-falls and classifies the subject's state. The method of physical monitoring described in Appendix 15.

[0077] (Note 19) Based on the time-series information of the signals from each of the multiple transmitting devices, the type of spatial region corresponding to the location of the subject is identified. If the classification of the subject's condition is a predetermined classification, the type of the spatial region is output to a predetermined output destination. A method of physical monitoring described in any one of the appendices 14 to 18.

[0078] (Note 20) Based on the state of falls and the number of falls classified for each spatial region, the regions with a high frequency of falls will be investigated. A method of body monitoring as described in Appendix 19, comprising the following:

[0079] (Note 21) The learning model is generated by machine learning the relationship between the time-series information of the signals from each of the multiple transmitting devices and the correct information indicating the classification of the subject's state as either fallen or not fallen. A method of body monitoring as described in Appendix 14, comprising the following:

[0080] (Note 22) The learning model is generated by machine learning the relationship between the time-series information of the signals from each of the multiple transmitting devices, the physical characteristic information, and the correct information indicating the classification of the subject's state as either fallen or not fallen. A method of body monitoring as described in Appendix 16, comprising:

[0081] (Note 23) The learning model is generated by machine learning the relationship between the time-series information of the signals from each of the multiple transmitting devices and the correct information indicating the classification of the subject's state into one of several different types of falls or non-falls. A method of body monitoring as described in Appendix 14, comprising the following:

[0082] (Note 24) The learning model is generated by machine learning the relationship between the time-series information of the signals from each of the multiple transmitting devices, the physical characteristic information, and the correct information indicating the classification of the subject's state into one of several different types of falls or non-falls. A method of body monitoring as described in Appendix 16, comprising:

[0083] (Note 25) The computer of the body monitoring device, A classification means for classifying the state of a subject based on time-series information of signals from multiple transmitters attached to multiple different locations on the subject's body. An output means that outputs output information corresponding to a predetermined classification to a predetermined output destination when the classification of the subject's condition is a predetermined classification. A program that makes it function as such.

[0084] (Note 26) The classification means inputs the time-series information of the signals from each of the multiple transmitters into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state as either fallen or not fallen, thereby acquiring the subject's state as either fallen or not fallen and classifying the subject's state. The program described in Appendix 25.

[0085] (Note 27) The classification means classifies the state of the subject based on the subject's physical characteristics information, based on the information contained in the signal transmitted from the transmitter. The program described in Appendix 25.

[0086] (Note 28) The classification means inputs the time-series information of the signals from each of the multiple transmitters and the physical characteristic information into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the state of the subject, which is classified as either fallen or not fallen, in order to acquire the state of the subject, which is either fallen or not, and classify the state of the subject. The program described in Appendix 27.

[0087] (Note 29) The classification means inputs the time-series information of the signals from each of the multiple transmitters into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state into one of multiple different types of falls or non-falls, thereby acquiring the subject's state as one of the multiple different types of falls or non-falls and classifying the subject's state. The program described in Appendix 25.

[0088] (Note 30) The classification means inputs the time-series information of the signals from each of the multiple transmitters and the physical characteristic information into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the state of the subject, which is classified into one of multiple different types of falls or non-falls, in order to acquire the state of the subject, which is one of multiple different types of falls or non-falls, and classify the state of the subject. The program described in Appendix 27.

[0089] (Note 31) The system includes a region identification means that identifies the type of spatial region corresponding to the location of the subject based on the time-series information of the signals from each of the plurality of transmitting devices, The output means outputs the type of the spatial region to a predetermined output destination when the classification of the subject's state is a predetermined classification. The program described in any one of the appendices 26 through 30.

[0090] (Note 32) A survey means for investigating areas with a high frequency of falls based on the state of falls and the number of falls classified for each spatial area, The program described in Appendix 31, which includes the following:

[0091] (Note 33) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices and the correct information indicating the classification of the state of the subject, which is either fallen or not fallen. The program described in Appendix 26, which includes the following:

[0092] (Note 34) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices, the physical characteristic information, and the correct information indicating the classification of the subject's state as either fallen or not fallen. The program described in Appendix 28, which includes the following:

[0093] (Note 35) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices and the correct information indicating the classification of the state of the subject into one of several different types of falls or non-falls, The program described in Appendix 26, which includes the following:

[0094] (Note 36) A learning means that generates the learning model by machine learning the relationship between the time-series information of the signals from each of the plurality of transmitting devices, the physical characteristic information, and the correct information indicating the classification of the subject's state into one of several different types of falls or non-falls. The program described in Appendix 28, which includes the following: [Explanation of Symbols]

[0095] 1...Body monitoring device 2. Leader 3. Transmitter 11. Control Unit 12...Classification section 13...Area identification part 14.. Output section 15...Storage Department 16. Learning Department 17. Learning Model Memory Unit

Claims

1. A classification means for classifying the state of the subject based on time-series information of signals from multiple transmitters attached to multiple different locations on the subject's body, An output means that outputs output information corresponding to a predetermined classification to a predetermined output destination when the classification of the subject's condition is a predetermined classification, A body monitoring device equipped with the following features.

2. The classification means inputs the time-series information of the signals from each of the multiple transmitters into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state as either fallen or not fallen, thereby acquiring the subject's state as either fallen or not fallen and classifying the subject's state. The body monitoring device according to claim 1.

3. The classification means identifies the physical characteristics information of the subject based on the information contained in the signal transmitted from the transmitting device, and classifies the state of the subject based on the physical characteristics information. The body monitoring device according to claim 1.

4. The classification means inputs the time-series information of the signals from each of the multiple transmitters and the physical characteristic information into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the state of the subject, which is classified as either fallen or not fallen, in order to acquire the state of the subject, which is either fallen or not, and classify the state of the subject. The body monitoring device according to claim 3.

5. The classification means inputs the time-series information of the signals from each of the multiple transmitters into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters and the correct information indicating the classification of the subject's state into one of multiple different types of falls or non-falls, thereby acquiring the subject's state as one of the multiple different types of falls or non-falls and classifying the subject's state. The body monitoring device according to claim 1.

6. The classification means inputs the time-series information of the signals from each of the multiple transmitters and the physical characteristic information into a learning model obtained by machine learning the relationship between the time-series information of the signals from each of the multiple transmitters, the physical characteristic information, and the state of the subject, which is classified into one of multiple different types of falls or non-falls, in order to acquire the state of the subject, which is one of multiple different types of falls or non-falls, and classify the state of the subject. The body monitoring device according to claim 3.

7. The system includes a region identification means that identifies the type of spatial region corresponding to the location of the subject based on the time-series information of the signals from each of the plurality of transmitting devices, The output means outputs the type of the spatial region to a predetermined output destination when the classification of the subject's state is a predetermined classification. The body monitoring device according to claim 1.

8. A survey means for investigating areas with a high frequency of falls based on the state of falls and the number of falls classified for each spatial area, The body monitoring device according to claim 7, comprising:

9. The subject's condition is classified based on the time-series information of signals from multiple transmitters attached to different locations on the subject's body. If the classification of the subject's condition falls under a predetermined category, output information corresponding to that classification is output to a predetermined output destination. Physical monitoring methods.

10. The computer of the body monitoring device, A classification means for classifying the state of a subject based on time-series information of signals from multiple transmitters attached to multiple different locations on the subject's body. An output means that outputs output information corresponding to a predetermined classification to a predetermined output destination when the classification of the subject's condition is a predetermined classification. A program that makes it function as such.

Images