WO2025057880A1 - Program, map creation device, and learning device - Google Patents

Abstract

Provided is a program for instructing a processor of a map creation device, which communicates with a monitoring device that includes one or more sensors and moves together with a user being monitored, to execute: a process for acquiring an output value of the sensor of the monitoring device and the location of the user of the monitoring device; a process for determining that a prescribed event has occurred to the user of the monitoring device on the basis of information acquired from the monitoring device; a process for creating map information for displaying a map image representing the number or frequency of the events for each place where the prescribed event has occurred; and a process for outputting the map information to another terminal device.

Description

Program, map generation device, and learning device

The present invention relates to a program, a map generating device, and a learning device.
This application claims priority based on Japanese Patent Application No. 2023-150238 filed in Japan on September 15, 2023 and Japanese Patent Application No. 2024-123346 filed in Japan on July 30, 2024, the contents of which are incorporated herein by reference.

In factories and other workplaces, various workplace accidents can occur, such as injuries from falls and heatstroke due to high temperatures. Managers need to monitor and take measures to prevent these workplace accidents from occurring.

Japan Special Publication No. 2022-531490 (A)

However, it was sometimes difficult for managers to constantly monitor the factory and to identify areas in the factory where measures needed to be taken.

The present invention was made in consideration of these circumstances, and one of its objectives is to provide a program, map generation device, and learning device that can monitor and predict the occurrence of workplace accidents in order to prevent them.

　In order to solve the above problems, the program, map generation device, and learning device have the following configurations. Aspect 1 of the present invention is a program for causing a processor of a map generation device, which has one or more sensors and communicates with a monitoring device that moves with a user who is being monitored, to execute the following processes: acquiring the output value of the sensor of the monitoring device and the position of the user of the monitoring device; determining that a specified event has occurred to the user of the monitoring device based on the information acquired from the monitoring device; generating map information for displaying a map image showing the number or frequency of the specified event occurring at each location; and outputting the map information to another terminal device.

In the program of aspect 2 of the present invention, the monitoring device is built into or removably attached to shoes worn by the user.

The program of aspect 3 of the present invention is a program for executing a process in which the specified event is a fall, and the number or frequency of falls at each location of the user is generated as map information.

The program of aspect 4 of the present invention is a program for executing a process in which the predetermined event is the user's slipping, and the number or frequency of each location where the user's slipping is generated as map information.

The program of aspect 5 of the present invention is a program for executing a process in which the specified event is heat stroke, and the number or frequency of occurrence of heat stroke at each location is generated as map information.

The program of aspect 6 of the present invention is a program for executing a process in which the predetermined event is a user carrying a heavy object, and the number or frequency of times that the user has carried a heavy object at each location is generated as map information.

The program of aspect 7 of the present invention is a program for causing a processor of a map generating device to execute the following processes: acquiring a trained model that has been trained to predict the occurrence of a specific event using the past performance of the specific event as input information; generating a prediction map of the occurrence of the specific event using the trained model; and outputting the prediction map.

The program of aspect 8 of the present invention is a program for causing a processor of a map generating device to execute the following processes: acquiring a trained model that has been trained to predict the occurrence and risk of a specified event using the past performance of the specified event as input information; generating a risk level map of the specified event using the trained model; and outputting the risk level map.

The program of aspect 9 of the present invention is a program for causing a processor of the map generating device to execute a process of generating the map information for displaying the movement trajectory of the user, who is the target of monitoring, superimposed on the map image, the movement trajectory being generated based on the position of the user.

The map generating device of aspect 10 of the present invention is a map generating device that has one or more sensors and communicates with a monitoring device that moves with a user to be monitored, and includes an acquisition unit that acquires the output value of the sensor of the monitoring device and the position of the user of the monitoring device, a determination unit that determines that a specified event has occurred to the user of the monitoring device based on the information acquired from the monitoring device, a map generating unit that generates map information for displaying a map image showing the number of times or frequency at which the specified event has occurred, and an output unit that outputs the map information to another terminal device.

In the map generating device of aspect 11 of the present invention, the monitoring device is built into or removably attached to shoes worn by the user.

The learning device of aspect 12 of the present invention is for executing an acquisition unit that acquires past records of specific events that have occurred to a user of the monitoring device and a model generation unit that generates a trained model that has been trained to be able to predict the occurrence of the specific events.

According to the present invention, by outputting a map containing information about work-related accidents, it is possible to implement efficient monitoring and preventive measures against work-related accidents.

FIG. 1 is an overall view of a map output system 1. FIG. 11 shows an example of a map image displayed by a second terminal device 400. 13 is a diagram showing an example of a work accident occurrence prediction map image displayed by the second terminal device 400. FIG. 13 is a diagram showing an example of a work accident risk level map displayed by the second terminal device 400. FIG. 13 is a diagram showing an example of a work accident risk level map including traffic line data displayed by the second terminal device 400. FIG. FIG. 3 is a diagram showing an example of a map image including flow line data in the map image of FIG. 2 . FIG. 13 is a diagram showing an example of processing performed by a model generating unit 620.

Below, we will explain embodiments of the program, learning device, and map generation method of the present invention with reference to the drawings.

FIG. 1 is an overall view of a map output system 1. The map output system 1 includes, for example, a monitoring device 100 and a map generating device 200. The monitoring device 100 is connected to the map generating device 200 via a first terminal device 300, a second terminal device 400, and a network NW. The network NW includes, for example, the Internet, a LAN (Local Area Network), a wireless base station, a provider device, and the like. The monitoring device 100 and the first terminal device 300 are connected by short-range wireless communication, for example, Bluetooth (registered trademark). Multiple beacon transmitters 500 are installed, for example, in a factory where the user works.

The beacon transmitter 500 wirelessly transmits a signal (e.g., an RFID signal) once every few seconds within a radius of several meters to several tens of meters. The beacon transmitter 500 is installed, for example, at regular intervals in the user's work area. The beacon transmitter 500 includes, for example, a transmitter ID. It is assumed that the monitoring device 100 knows in advance the installation location of the beacon transmitter 500 corresponding to the transmitter ID.

The components of each device will be described below. The monitoring device 100 includes, for example, a communication unit 110, pressure sensors 120L and 120R, acceleration sensors 130L and 130R, a temperature sensor 140, a heartbeat sensor 150, a position detection unit 160, a report unit 170, and a beacon receiver 180. The monitoring device 100 is a device that is built into or detachably attached to a shoe, and includes components for the left foot and the right foot. The communication unit 110, the temperature sensor 140, the heartbeat sensor 150, the position detection unit 160, the report unit 170, and the beacon receiver 180 are provided on either the left foot or the right foot, or on both. Hereinafter, when there is no need to distinguish between the sensor for the left foot and the sensor for the right foot, the symbols L and R indicating whether the sensor is for the left foot or the right foot may be omitted in the description. It is also preferable that there are multiple pressure sensors 120L and 120R for the left and right feet.
Although not particularly limited, the number of sensors included in the monitoring device 100 may be 1 or more, 2 or more, or 5 or more. Furthermore, although not particularly limited, the number of sensors included in the monitoring device 100 may be 100 or less, 10 or less, or 5 or less.
In addition, the temperature sensor 140, heart rate sensor 150, position detection unit 160, reporting unit 170, and beacon receiver 180 may be used as part of the monitoring device 100 in a mobile terminal, a mobile phone, or a smart watch.

The monitoring device 100 may be anything that moves with the user who is the subject of monitoring. For example, the monitoring device 100 may be attached to the user's body or clothing, or attached like a wristwatch strap. In this case, the burden of attaching and carrying the device on the user who is the subject of monitoring is reduced. The monitoring device 100 may be built into an accessory, or may be attached in a detachable manner. Attached also includes being stored. Accessories include, for example, safety shoes, helmets, belts, work clothes, and other items worn by workers. The user will not forget to wear such accessories, so forgetting to carry the monitoring device 100 is suppressed. In addition, the monitoring device 100 may be an insole (insole) of a shoe, a wristwatch, a finger ring, a necklace, a bracelet, a name tag, etc. The shoes into which the insole is inserted are the shoes worn by the user when working in a facility such as a factory. In this way, if the insole, which is the monitoring device 100, is inserted into the shoes that the user normally wears, the user is more likely to forget to carry the monitoring device 100, which is more preferable.

The communication unit 110 communicates with the map generating device 200, for example, via the first terminal device 300 and the network NW. Alternatively, the communication unit 110 may communicate with the map generating device 200 by connecting to the network NW without going through the first terminal device 300.

The beacon receiver 180 receives the signal emitted by the beacon transmitter 500 and obtains the transmitter ID.

The pressure sensor 120 is a device that uses a pressure-sensitive element to measure the pressure of the user's foot, converts it into an electrical signal, and outputs it.

The acceleration sensor 130 is, for example, a three-axis acceleration sensor, and is a device that converts the acceleration of each of the three axes into an electrical signal and outputs it. The acceleration of each of the three axes detected by the acceleration sensor includes a weight acceleration component, which makes it possible to detect the inclination and vibration of the user's foot. The acceleration sensor 130 may be, for example, a gyro sensor.

The temperature sensor 140 measures the temperature of the user's feet and/or the outside air temperature, converts it into an electrical signal, and outputs it.

The heart rate sensor 150, for example, shines light on the user's leg, detects the light reflected by blood vessels, and measures the pulse rate from the change in the amount of light over time. The heart rate sensor 150 is a device that converts the measured pulse rate into an electrical signal and outputs it. Note that the heart rate sensor 150 may also be a microwave sensor.

The position detection unit 160 detects the position of the monitoring device 100 (i.e., the user's position) based on, for example, the transmitter ID acquired by the beacon receiver 180. The position detection unit 160 can detect the installation location of the beacon transmitter 500 from the transmitter ID, so it may detect the installation location of the detected beacon transmitter 500 as the user's position, or, if multiple transmitter IDs are acquired, it may detect the user's position based on the principle of triangulation taking into account the reception strength. The position detection unit 160 may also detect the user's position using information from a GPS receiver (not shown) of the first terminal device 300.

The reporting unit 170 outputs the output values of the various sensors, together with, for example, the position detected by the position detection unit 160, to the map generating device 200 via the communication unit 110 and the first terminal device 300. At this time, the reporting unit 170 adds the identification information of the monitoring device 100 and outputs the output values and the position to the map generating device 200. The output information is acquired and managed by the map generating device 200. For example, the map generating device 200 manages information in which the identification information of the monitoring device 100, the output values, and the position are associated with each other as management information.

The map generating device 200 includes, for example, a communication unit 210, an acquisition unit 220, a determination unit 230, a map generating unit 240, and an output unit 250. The components other than the communication unit 210 are realized by, for example, a hardware processor such as a CPU (Central Processing Unit) executing a program (software). Some or all of these components may be realized by hardware (including circuitry) such as an LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), or GPU (Graphics Processing Unit), or may be realized by collaboration between software and hardware. The program may be stored in advance in a storage device (a storage device with a non-transient storage medium) such as a hard disk drive (HDD) or flash memory, or may be stored in a removable storage medium (a non-transient storage medium) such as a DVD or CD-ROM, and installed by inserting the storage medium into a drive device.

The communication unit 210 is a communication interface for connecting to the network NW. The communication unit 210 communicates with the monitoring device 100 via the first terminal device 300.

The acquisition unit 220 acquires, for example, the output values of the pressure sensor 120, acceleration sensor 130, temperature sensor 140, and heart rate sensor 150, as well as the user's current location, from the monitoring device 100.

The determination unit 230 determines whether the user has fallen, slipped, suffered from heat stroke, or carried a heavy object based on the information acquired by the acquisition unit 220.

(1) The determination unit 230 determines whether the user has fallen based on, for example, the output values of the pressure sensor 120 and the acceleration sensor 130. The determination unit 230 may determine whether the user has fallen based only on the output value of the pressure sensor 120. For example, the determination unit 230 determines that the user has fallen when the output values of the pressure sensors 120 of both feet indicate a value less than a threshold value. The determination unit 230 may also determine that the user has fallen when, for example, the output values of the pressure sensors 120 of both feet indicate a value less than a threshold value and the angle of the acceleration sensor 130 changes suddenly (for example, the resultant acceleration exceeds the threshold value within one second). When the determination unit 230 determines that the user has fallen, it outputs the number of falls and the location of the falls. The determination unit 230 acquires information on the location where the fall occurred from the acquisition unit 220. The number of falls and the location of the falls are combined to form the fall information.

(2) The determination unit 230 may determine whether or not the user has slipped, for example, based on the output value of the acceleration sensor 130. The determination unit 230 determines that the user has slipped if the angle of the acceleration sensor 130 changes suddenly (for example, the resultant acceleration exceeds a threshold within one second). If the determination unit 230 determines that the user has slipped, it outputs the number of times the user has slipped and the location of the slip. As for the location of the slip, the determination unit 230 acquires the location of the user's slip from the acquisition unit 220. The number of times the user has slipped and the location of the slip are combined to form semi-fall information.

(3) For example, when the output value of the temperature sensor 140 is equal to or greater than a threshold value, the determination unit 230 may determine that the user has heat stroke or is likely to have heat stroke. For example, the determination unit 230 may use the value of an outside temperature gauge (not shown) installed in a factory and the output value of the temperature sensor 140 to determine whether the user has heat stroke or is likely to have heat stroke. When the determination unit 230 determines that the user has heat stroke or is likely to have heat stroke, it acquires the location where heat stroke has been determined from the acquisition unit 220. The acquired information that the user has heat stroke and the location where the heat stroke occurred are regarded as heat stroke information.

(4) The determination unit 230 may determine whether or not the user has held a heavy object based on, for example, the output value of the pressure sensor 120. The determination unit 230 determines that a heavy object has been held when the output values of the pressure sensors 120 of both feet exceed a threshold value for a certain period of time. When the determination unit 230 determines that the user has held a heavy object, it acquires from the acquisition unit 220 the location where the heavy object was held. The acquired information that a heavy object has been held and the location where the heavy object was held are regarded as weight information.

The above-mentioned specified events are workplace accidents or events that could become workplace accidents. Each time a specified event occurs, the determination unit 230 obtains the location where the specified event occurred from the acquisition unit 220, and links each occurrence of the specified event to the location.

The map generating unit 240 receives various types of information from the determining unit 230 and generates map information. The map generating unit 240 may generate map information for each type of information, or may generate one piece of map information for all the information.

The output unit 250 outputs the map information generated by the map generation unit 240 to the second terminal device 400 and displays it as a map image. The map information may be, for example, an image in JPEG (Joint Photographic Experts Group) format, or may be information such as parameters for displaying the map image on the second terminal device 400 (such as the number of occurrences of an event associated with a coordinate).

The memory unit 260 is configured with a storage medium, such as a HDD, flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), RAM (Random Access Read/Write Memory), ROM (Read Only Memory), or any combination of these storage media. The memory unit 260 stores the trained model 632, which will be described later.

The first terminal device 300 and the second terminal device 400 are devices that can be connected to the network NW, such as, for example, a smartphone or a tablet terminal. The first terminal device 300 is owned by the user, and the second terminal device 400 is owned by the user's administrator. For example, a monitoring app (application software) runs on the second terminal device 400, and displays the output values of various sensors and the map output from the map generating device 200 as images.

2 is a diagram showing an example of a map image displayed by the second terminal device 400. Based on various information determined by the determination unit 230, FIG. 2 shows the locations where a specific event occurred with a black circle, and further shows the frequency that indicates how many times what specific event occurred during an observation period (e.g., one month). Although FIG. 2 shows the number of times per month, for example, all past times may be displayed, or the number of times per year may be displayed. These numbers, frequencies, or details of the specific event may be displayed when an operation is performed on the map image at the location where the specific event occurred or in its vicinity. In other words, the "map image showing the number or frequency of each location where the specific event occurred" may be displayed after the above operation is performed. This information may be displayed on a separate screen, or may be displayed superimposed on the map image.

The map generating unit 240 may generate not only the map information that is the basis for the map image shown in FIG. 2, but also a work-related accident prediction map. FIG. 3 shows an example of a work-related accident prediction map image displayed by the second terminal device 400. The work-related accident prediction map is information in map format that predicts what type of work-related accidents will occur. The work-related accident prediction map image shows, for example, locations where work-related accidents are predicted to occur with black circles, and indicates what type of work-related accident will occur with a level such as high, medium, or low based on the output values of various sensors and the past number of occurrences of specified events. The level may be expressed as a numerical value, for example.

The map generation unit 240 may, for example, determine a threshold value for each predetermined event, and if the output values of various sensors or the past occurrences of a predetermined event exceed a first threshold, indicate the level as "high", if the output values are between the first and second thresholds (first threshold > second threshold), indicate the level as "medium", and if the output values are below the threshold, indicate the level as "low". Information indicating these levels may be displayed when an operation is performed on the map image at or near the location where the predetermined event occurred. This information may be displayed on a separate screen, or may be superimposed on the map image.

The map generating unit 240 may generate a work accident risk level map in addition to generating the work accident occurrence prediction map shown in FIG. 3. FIG. 4 is a diagram showing an example of a work accident risk level map displayed by the second terminal device 400. The work accident risk level map shows the predicted risk of work accidents in map form. The risk of a work accident is, for example, a numerical value obtained by multiplying the type of work accident predicted to occur by the degree of impact. The higher the numerical value of the "work accident risk", the higher the risk. The degree of impact is a numerical value that is set in advance, taking into consideration the impact of the work accident, such as the aftereffects on the user, and the impact on the business. In FIG. 4, for example, a factory is divided into sections of 3 m x 3 m, and mesh processing is performed so that the risk of work accidents can be displayed for each section. The risk of work accidents is indicated by the shade of color, with the darker the color, the higher the risk, and the lighter the color, the lower the risk.

The map generating unit 240 may identify a place where a physically demanding task is performed as a place where a work-related accident is likely to occur, or may determine that the risk of a work-related accident occurring is high. The map generating unit 240 may also obtain places where the physical load is high by inputting information on the physical load into the trained model 632. The process of generating the trained model 632 will be described later. A place where the output value of the pressure sensor 120 is higher than the threshold value for a certain period of time only for one foot, or a place where the output value of the pressure sensor 120 is a similar value periodically, may also be identified as a place where a work-related accident is likely to occur, or may determine that the risk of a work-related accident occurring is high. The information indicating the above-mentioned risk level may be displayed in response to a user's operation. For example, when a user performs an operation on a specified area, information indicating the risk level of the area or the vicinity of the area in response to the operation may be displayed. This information may be displayed on a separate screen, or may be displayed superimposed on the map image.

For the purposes of processing by the map generation unit 240, the determination unit 230 determines that the user is performing work that places a high physical strain on the user if, for example, the output value of the heart rate sensor 150 exceeds a threshold value for a certain period of time. The map generation unit 240 may consider work that places a high physical strain on the user as work that is likely to cause work-related accidents, and add this to the materials used to generate a work-related accident prediction map or a work-related accident risk level map.

For example, if the output value of the pressure sensor 120 is higher than the threshold value for only one foot for a certain period of time, the determination unit 230 determines that a work-related accident is likely to occur. Work-related accidents that are likely to occur include, for example, falls and injuries such as back pain. The map generation unit 240 may add information about work-related accidents that are likely to occur to the materials used to generate the work-related accident prediction map and the work-related accident risk level map.

For example, if the output value of the pressure sensor 120 periodically indicates a similar value, the determination unit 230 determines that repetitive work is occurring. The map generation unit 240 may determine that repetitive work is likely to cause injuries such as back pain, and add information about the locations where repetitive work occurs to the materials used to generate the work accident prediction map and the work accident risk level map.

The map generating unit 240 may, for example, generate a map by overlaying movement line data on the generated map. The movement line data is obtained by acquiring the user's movement route from the position detecting unit 160 and connecting points with lines. The movement line data is an example of "the movement trajectory of the user generated based on the position of the user who is the target of monitoring." The map generating unit 240 may generate a map that displays the movement line data alone.

FIG. 5 shows an example of an occupational accident risk level map including movement line data displayed by the second terminal device 400. The dotted lines indicate the movement lines of users. The map generating unit 240 may generate a map including movement line data of multiple users. The map generating unit 240 may also generate a map including multiple movement line data of a target user (e.g., movement line data such as work on different dates and times), or may generate a map including movement line data of different users.

The map generation unit 240 may generate map information that superimposes the traffic line data on the map of FIG. 2 or FIG. 3 described above. Depending on the user's operation, a map may be displayed that displays any or all of the following information: traffic line data, the number of occurrences of a specific event, the frequency of a specific event, and details of the specific event. This allows the administrator to easily check the information he or she wants to obtain.

For example, by displaying a map with movement line data overlaid on the map generated as described above, it may be possible to identify the cause of a specific event. For example, by checking the movement line data of a person who has fallen just before the fall, it may be possible to identify the cause of the fall. It is also useful for determining whether the cause of the specific event is the person who caused the specific event or the environment. For example, by analyzing the behavior of each specific event, it may be possible to determine that one person causes the specific events in a similar pattern, even if they appear to be many at first glance. The results of this clarification can be used to plan preventive measures for work-related accidents. Furthermore, by generating a map that displays movement line data alone after having the function of displaying the specific event, it becomes easy to analyze using the movement line data, even if the movement line data is complex and difficult to see when it is displayed overlaid with the specific event. In addition, multiple movement line data can be displayed on one map. In this case, it is possible to easily determine whether the cause of a specific event is the person or the environment by analyzing multiple movement line data.

Incidentally, the movement line data may be included in the occupational accident risk level map as shown in FIG. 5, or may be included in the map image of FIG. 2 or the occupational accident prediction map of FIG. 3. FIG. 6 is a diagram showing an example of a map image including the movement line data of FIG. 2. By including the movement line data in the map image, it is possible to easily compare the occurrence of a specified event with the movement line data to identify the cause of the occurrence of the specified event or to consider plans to prevent the occurrence of the specified event.

The map generating device 200 may display information indicating the movement line data of the specified monitoring device 100 (user) on the map. For example, the map generating device 200 may refer to the management information to identify the movement line data associated with the identification information of the monitoring device 100 used by the specified user, and display a map including the identified movement line data. The identification information of the management information is associated with information indicating whether a specific event has occurred, details of the specific event, and other information. More specifically, the identification information of the management information is associated with the location history when the specific event occurred. For example, when the location where the specific event occurred is operated on the map of FIG. 2 or FIG. 6, the map generating device 200 displays the movement line data where the specific event occurred on the map. In this way, the movement line data is displayed, or the movement line data is displayed in association with the specific event, making it easier to investigate the cause of a work accident or to analyze whether a specific event occurred due to human factors.

The trained model 632 is generated, for example, by a learning device 600 that is separate from the map generating device 200. The learning device 600 includes, for example, an acquisition unit 610, a model generating unit 620, and a storage unit 630. The components other than the storage unit 630 are realized, for example, by a processor such as a CPU or GPU executing a program (software). Some or all of these functional units may be realized by hardware such as an LSI, ASIC, or FPGA, or may be realized by a combination of software and hardware. The program may be stored in advance in a storage device such as an HDD or flash memory, or may be stored in a removable storage medium such as a DVD or CD-ROM, and installed in the storage device by inserting the storage medium into a drive device.

The storage unit 630 is configured with a storage medium, such as a HDD, a flash memory, an EEPROM, a RAM, a ROM, or any combination of these storage media. The storage unit 630 stores information such as a trained model 632, a training dataset 634, and model setting information 636.

[Learning stage/Phase 1]
The functions of each part of the learning device 600 will be described below, divided into a learning stage, an inference stage, and a re-learning stage. In the learning stage, the acquisition unit 610 acquires input information from the monitoring device 100, which is the original data of the learning data set 634. The original data is, for example, fall information, half-fall information, heat stroke information, and weight information that are the source of the learning data 634A, and past records of a specific event in the factory that are the source of the teacher data 634B. The learning data set 634 is obtained by associating the learning data 634A with the teacher data 634B. The teacher data 634B may be past records manually input, or may be past records in other factories acquired from outside.

The acquisition unit 610 includes a preprocessing unit 612. The preprocessing unit 612 performs preprocessing on the original data to generate a learning dataset 634. The preprocessing unit 612 performs, for example, normalization processing.

The model generation unit 620 generates the trained model 632 based on the training data 634A and the teacher data 634B. As described above, the training data 634A is obtained by performing preprocessing on various information acquired from the monitoring device 100. The teacher data 634B is obtained by performing preprocessing on the past performance of a specific event in the factory. The number of elements in the training data 634A matches the number of input nodes specified by the model setting information 636. The model setting information 636 is information that specifies the number of input nodes, the number of output nodes, the connection mode of intermediate nodes, etc. of the machine learning model that is the basis of the trained model 632.

FIG. 7 is a diagram showing an example of processing performed by the model generation unit 620. The model generation unit 620 learns the parameters of the machine learning model using a technique such as backpropagation so that the output of the machine learning model when learning data 634A is used as input data approaches teacher data 634B. For example, the machine learning model at the point when the above processing has been executed a specified number of times is finalized as the trained model 632.

[Inference stage/phase 2]
The learning device 600 transmits, for example, the trained model 632 to the map generation device 200. The transmitted trained model 632 is stored in the storage unit 260 of the map generation device 200. The map generation unit 240 generates a work accident occurrence prediction map and a work accident risk level map using the trained model 632. The trained model 632 may be installed in the map generation device 200 by an external medium such as a storage medium.

[Relearning Phase/Phase 3]
The acquisition unit 610 acquires the fall information, partial fall information, heat stroke information, and weight information from the determination unit 230. The acquisition unit 610 periodically acquires new various information. The learning device 600 acquires the new various information, incorporates it as teacher data 634B so as to obtain more accurate data, and repeats the processing of phase 1 to perform re-learning.

The learning device 600 repeats phases 2 and 3 as feedback to improve the accuracy of the trained model 632 constructed in phase 1.

The functions of the learning device 600 may be included in the map generating device 200, for example.

The map output system 1 described above outputs a map including information about a specific event, which visualizes the frequency of occurrence of the specific event and the location where the specific event occurred, allowing for efficient monitoring and preventive measures against work-related accidents. In addition, the map output system 1 outputs the location where the specific event occurred in map format, allowing the manager to understand where the specific event is occurring. The map output system 1 generates a work-related accident prediction map and a work-related accident risk level map using the trained model 632 generated by the learning device 600, allowing for measures and prevention against future work-related accidents. The map output system 1 outputs a map including movement line data, allowing the movement line of a skilled user to be visualized, and work efficiency can be improved by referring to the movement line of a skilled user. In addition, the map output system 1 outputs a map including movement line data, allowing the manager to understand entry and exit of prohibited areas.

　Although the above describes the form for carrying out the present invention using an embodiment, the present invention is in no way limited to such an embodiment, and various modifications and substitutions can be made without departing from the spirit of the present invention.

The present invention makes it possible to provide a program, a learning device, and a map generation method that can monitor and predict the occurrence of work-related accidents to prevent them.

REFERENCE SIGNS LIST 100 Monitoring device 110 Communication unit 120 Pressure sensor 130 Acceleration sensor 140 Temperature sensor 150 Heart rate sensor 160 Position detection unit 170 Report unit 180 Beacon receiver 200 Map generation device 210 Communication unit 220 Acquisition unit 230 Determination unit 240 Map generation unit 250 Output unit 260 Storage unit 300 First terminal device 400 Second terminal device 500 Beacon transmitter 600 Learning device 610 Acquisition unit 612 Preprocessing unit 620 Model generation unit 630 Storage unit 632 Learned model 634 Learning dataset 636 Model setting information

Claims (12)

A processor of a map generating device that has one or more sensors and communicates with a monitoring device that moves with a user who is a target of monitoring,
A process of acquiring an output value of a sensor of the monitoring device and a position of a user of the monitoring device;
A process of determining that a predetermined event has occurred to a user of the monitoring device based on information acquired from the monitoring device;
A process of generating map information for displaying a map image showing the number or frequency of occurrence of the predetermined event for each location;
outputting the map information to another terminal device;
A program for executing. The monitoring device is built into or detachably attached to shoes worn by the user.
The program according to claim 1. The predetermined event is a fall,
A process of generating the number or frequency of falls at each location of the user as the map information;
3. The program according to claim 1 or 2 for executing the above. The predetermined event is that the user has slipped,
A process of generating the number or frequency of the user's slipping at each location as the map information;
3. The program according to claim 1 or 2 for executing the above. The predetermined event is heat stroke,
A process of generating the number or frequency of occurrence of heat stroke for each location as the map information;
3. The program according to claim 1 or 2 for executing the above. The predetermined event is the user carrying a heavy object,
generating the map information based on the number or frequency of times that the user has carried a heavy object for each location;
3. The program according to claim 1 or 2 for executing the above. A processor of the map generating device,
A process of acquiring a trained model that has been trained to predict the occurrence of the specified event using past records of the specified event as input information;
A process of generating an occurrence prediction map of the predetermined event using the trained model;
A process of outputting the occurrence prediction map;
3. The program according to claim 1 or 2 for executing the above-mentioned. A processor of the map generating device,
A process of acquiring a trained model that has been trained to predict the occurrence and risk of the specified event using past records of the specified event as input information;
A process of generating a risk level map of the predetermined event using the trained model;
A process of outputting the risk level map;
3. The program according to claim 1 or 2 for executing the above-mentioned. A processor of the map generating device,
A process of generating map information for superimposing and displaying a movement trajectory of the user, which is generated based on the position of the user who is the target of monitoring, on the map image;
3. The program according to claim 1 or 2 for executing the above. A map generating device that has one or more sensors and communicates with a monitoring device that moves with a user to be monitored,
an acquisition unit that acquires an output value of a sensor of the monitoring device and a position of a user of the monitoring device;
a determination unit that determines that a predetermined event has occurred to a user of the monitoring device based on information acquired from the monitoring device;
a map generating unit that generates map information for displaying a map image showing the number or frequency of occurrence of the predetermined event;
an output unit that outputs the map information to another terminal device;
A map generating device comprising: The monitoring device is built into or detachably attached to shoes worn by the user.
The map generating device of claim 10. The learning device is
An acquisition unit that acquires past records of predetermined events that have occurred to a user of the monitoring device;
A model generation unit that generates a trained model that is trained to predict the occurrence of the predetermined event;
Equipped with learning equipment.

Images