JP2024050173A - Work monitoring device, work monitoring method, and program - Google Patents

Abstract

[Problem] To provide a work monitoring device, work monitoring method, and program capable of suppressing the occurrence of work-related accidents. [Solution] A work monitoring device comprising an analysis information acquisition unit that acquires data detected in association with the movements of a worker as analysis information indicating information for analyzing the work state of the worker, an analysis unit that analyzes the work state of the worker based on the acquired analysis information, and a sign detection unit that detects signs of the occurrence of a work-related accident in the work of the worker based on the analysis results of the work state of the worker. [Selected Figure] Figure 2

Description

The present invention relates to a work monitoring device, a work monitoring method, and a program.

Traditionally, in production lines in factories and the like, not only work is done by machines, but many tasks are also done by people (workers). For example, when multiple workers perform the same task, the amount of time each worker spends on the task varies depending on their level of proficiency, with workers with lower levels of proficiency spending more time on the task. For this reason, production lines where workers with lower levels of proficiency work are performing have lower production efficiency than production lines where workers with higher levels of proficiency work. As a result, various technologies have been proposed to suppress the decline in production efficiency of production lines where workers with lower levels of proficiency work are performing.

For example, the following Patent Document 1 discloses a technology for detecting movements that should be improved during a given task, based on information about the movements obtained from a worker while working. The information about the movements is, for example, information that indicates changes in skeletal position and changes in acceleration over time. This technology allows workers with low levels of proficiency in a task to prevent a decline in the production efficiency of the production line by improving the detected movements.

JP 2019-40421 A

While it is important to prevent a decline in work efficiency at manufacturing sites, it is also important to ensure the safety of workers and prevent the occurrence of industrial accidents. However, the technology in Patent Document 1 above cannot detect whether or not a worker is working safely from the detected movements, and does not take into consideration the prevention of industrial accidents.

In view of the above problems, the object of the present invention is to provide a work monitoring device, a work monitoring method, and a program that can reduce the occurrence of work-related accidents.

In order to solve the above-mentioned problems, a work monitoring device according to one aspect of the present invention includes an analysis information acquisition unit that acquires data detected in association with the movements of a worker as analysis information indicating information for analyzing the work state of the worker, an analysis unit that analyzes the work state of the worker based on the acquired analysis information, and a sign detection unit that detects signs of an industrial accident occurring during the work of the worker based on the analysis results of the work state of the worker.

A work monitoring method according to one aspect of the present invention includes an analysis information acquisition process in which an analysis information acquisition unit acquires data detected in association with the movements of a worker as analysis information indicating information for analyzing the work state of the worker, an analysis process in which an analysis unit analyzes the work state of the worker based on the acquired analysis information, and a sign detection process in which a sign detection unit detects signs of an industrial accident occurring during the work of the worker based on the analysis result of the work state of the worker.

A program according to one aspect of the present invention causes a computer to function as: analytical information acquisition means for acquiring data detected in association with the movements of a worker as analytical information indicating information for analyzing the working state of the worker; analysis means for analyzing the working state of the worker based on the acquired analytical information; and symptom detection means for detecting symptoms of an industrial accident occurring during the work of the worker based on the analysis results of the working state of the worker.

The present invention can reduce the occurrence of work-related accidents.

1 is a block diagram showing an example of a schematic configuration of a work monitoring system according to an embodiment of the present invention. FIG. 2 is a block diagram showing an example of the functional configuration of the wearable device according to the present embodiment. FIG. 4 is a sequence diagram showing an example of a processing flow in the work monitoring system according to the present embodiment. 10 is a flowchart showing an example of a process flow in an analysis process according to the present embodiment.

The following describes in detail an embodiment of the present invention with reference to the drawings.

1. Overview of the work monitoring system
The schematic configuration of the work monitoring system according to this embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing an example of the schematic configuration of the work monitoring system according to this embodiment.

1 is a system for monitoring work performed by a worker. The work monitoring system 1, for example, monitors whether the work performed by the worker is performed normally, and monitors whether there are any signs of a work-related accident (hereinafter also referred to as an "occupational accident") occurring during the work performed by the worker.
The work monitoring system 1 analyzes the work state of the worker when performing these monitoring operations. The work state of the worker includes, for example, a motion state in which the worker is performing motion for the work and a biological state of the worker performing the work. When analyzing the work state of the worker, the work monitoring system 1 acquires analysis information indicating information for analyzing the work state of the worker. To analyze the motion state, the work monitoring system 1 acquires, for example, acceleration information accompanying the motion of the worker as analysis information. Furthermore, to analyze the biological state, the work monitoring system 1 acquires, for example, biological information accompanying the motion of the worker as analysis information.

If the analysis of the worker's work status shows that the worker is not performing the work normally (i.e., there is an abnormality in the work status), the work monitoring system 1 notifies the worker that there is an abnormality in the work status (hereinafter, also referred to as an "abnormality notification"). The abnormality notification may include instructions for the worker to improve the work status. This allows the worker to check the notification and improve his or her own work status, thereby preventing the occurrence of an industrial accident.
Furthermore, when a sign of an industrial accident is detected during work by a worker, the work monitoring system 1 notifies the work supervisor of the detection of a sign of an industrial accident (hereinafter also referred to as an "injury occurrence sign notification"). This allows the supervisor to check the notification and improve the work condition of the worker, thereby preventing the occurrence of an industrial accident. Even if an industrial accident has already occurred at the time of receiving the notification, by checking the notification and rushing to the site, it is possible to respond without delay from the occurrence of the industrial accident, and to minimize the damage.

Below, the work monitoring system 1 will be described using as an example an example of monitoring the work that workers repeatedly perform to manufacture products on a production line at a manufacturing site such as a factory. Note that application examples of the work monitoring system 1 are not limited to this example. For example, as long as a worker is performing work, there are no particular limitations on the industry, location, work content, etc. to which the work monitoring system 1 is applied. Examples of industries include the manufacturing industry and the construction industry.

An example of an industrial accident that may occur in a manufacturing site such as a factory is being caught or entangled in a machine. For example, when working next to a line where products are flowing, if a worker notices dirt on the products flowing in front of the rollers, the worker needs to stop the line and clean it. However, since the dirt is very fine, the worker decides that he can clean it without stopping the line, and follows the products as they flow by with his upper body. This cleaning action is different from the actions taken in normal work. At this time, the worker is unable to remove the dirt due to working in a different position than usual, and his finger gets caught in the rollers by mistake. The worker is able to pull out his finger, but the intense pain causes him to crouch down on the spot, holding his finger. This action of crouching down and not being able to move can also be said to be different from the actions taken in normal work.
For this reason, movements that differ from normal work movements, such as the cleaning movements and crouching movements mentioned above, can be said to be movements that can lead to industrial accidents.

As shown in FIG. 1, the work monitoring system 1 includes a wearable device 10, a supervisor terminal 20, and a management DB (Data Base) 30.

(1) Wearable Device 10
The wearable device 10 is a terminal worn by a worker and is an example of a work monitoring device. Note that the manner in which the wearable device 10 is worn is not particularly limited as long as it is capable of acquiring necessary analysis information, and may be any manner such as a glove type, a wristband type, a watch type, a glasses type, or a head-mounted type.

The wearable device 10 is communicatively connected to the supervisor terminal 20 and the management DB 30 via the network NW. In communication with the supervisor terminal 20, the wearable device 10 transmits a notice of a possible industrial accident. In communication with the management DB 30, the wearable device 10 transmits work information. Work information is information acquired regarding the work being performed by a worker. Specifically, the work information is, for example, information indicating the results of an analysis of the work status or information indicating the content of past work.

(2) Supervisor terminal 20
The supervisor terminal 20 is a terminal used by a supervisor to manage workers and their work, etc. The supervisor terminal 20 is, for example, a personal computer (PC), a smartphone, a tablet terminal, or the like.

The supervisor terminal 20 is communicatively connected to the wearable device 10 and the management DB 30 via the network NW. In communication with the wearable device 10, the supervisor terminal 20 receives a notification of a warning sign of an industrial accident. In communication with the management DB 30, the supervisor terminal 20 transmits a request to refer to work information and receives the requested work information.

(3) Management DB 30
The management DB 30 is a DB server that manages work information. The management DB 30 is configured, for example, by one or more server devices (for example, cloud servers).
The management DB 30 is communicatively connected to the wearable device 10 and the supervisor terminal 20 via the network NW. In communication with the wearable device 10, the management DB 30 receives work information. In communication with the supervisor terminal 20, the management DB 30 receives a reference request for work information and receives the requested work information.

2. Functional configuration of wearable device
The schematic configuration of the work monitoring system 1 according to the present embodiment has been described above. Next, the functional configuration of the wearable device 10 according to the present embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram showing an example of the functional configuration of the wearable device 10 according to the present embodiment.

As shown in FIG. 2, the wearable device 10 includes a communication unit 110, an input unit 120, an acceleration information acquisition unit 130, a biometric information acquisition unit 140, a memory unit 150, a control unit 160, and an output unit 170.

(1) Communication unit 110
The communication unit 110 has a function of transmitting and receiving various information. For example, the communication unit 110 communicates with the supervisor terminal 20 and the management DB 30 via the network NW to transmit and receive various information.

(2) Input unit 120
The input unit 120 has a function of receiving input from an operator. The input unit 120 is configured with an input device that the wearable device 10 has as hardware, such as a button, a touch panel, a microphone, and the like.

(3) Acceleration information acquisition unit 130
The acceleration information acquisition unit 130 has a function of acquiring acceleration information. The acceleration information acquisition unit 130 is realized by an acceleration sensor that is provided as hardware in the wearable device 10. The acceleration information acquisition unit 130 detects acceleration caused by a worker wearing the wearable device 10 performing work using the acceleration sensor, and acquires acceleration information indicating the acceleration.

(4) Biometric information acquisition unit 140
The bioinformation acquisition unit 140 has a function of acquiring bioinformation. The bioinformation acquisition unit 140 is realized by a biosensor that the wearable device 10 has as hardware. For example, the biosensor is a heartbeat sensor or an electroencephalogram sensor. In the case of a heartbeat sensor, the bioinformation acquisition unit 140 detects the heart rate of a worker wearing the wearable device 10 while the worker is working using the heartbeat sensor, and acquires bioinformation indicating the heart rate. In the case of an electroencephalogram sensor, the bioinformation acquisition unit 140 detects brain waves of a worker wearing the wearable device 10 while the worker is working using the electroencephalogram sensor, and acquires bioinformation indicating the brain waves.

(5) Storage unit 150
The storage unit 150 has a function of storing various information. The storage unit 150 is configured by a storage medium provided as hardware in the wearable device 10, such as a hard disk drive (HDD), a solid state drive (SSD), a flash memory, an electrically erasable programmable read only memory (EEPROM), a random access read/write memory (RAM), a read only memory (ROM), or any combination of these storage media.
As shown in FIG. 2, the storage unit 150 stores a motion state analysis model 151 , a biological state analysis model 152 , and a sign detection model 153 .

(5-1) Operational state analysis model 151
The motion state analysis model 151 is a model for analyzing a motion state. For example, the motion state analysis model 151 is a model (trained model) that has learned the relationship between a specific motion in a task and acceleration information detected when a worker performs the specific motion. For this reason, the motion state analysis model 151 receives the acceleration information acquired by the analysis information acquisition unit 161 as an input and outputs the degree to which the worker is performing the specific motion. The degree to which the worker is performing the specific motion is, for example, the degree of deviation between the specific motion in a task and the specific motion actually performed by the worker.

The motion state analysis model 151 is prepared for each worker and for each task performed by the worker. That is, a dedicated motion state analysis model 151 is prepared for each task performed by each worker. Therefore, in order to generate the motion state analysis model 151, each worker performs the motions for each task while wearing the wearable device 10 in advance, and acquires acceleration information corresponding to the motions for each task as learning data.

(5-2) Biological state analysis model 152
The biological condition analysis model 152 is a model for analyzing a biological condition. The biological condition analysis model 152 is a model (trained model) that has learned the relationship between a specific motion in a task and biological information detected when a worker performs the specific motion. For this reason, the biological condition analysis model 152 receives the biological information acquired by the analysis information acquisition unit 161 as an input and outputs the degree to which the worker's biological condition is normal. The degree to which the worker's biological condition is normal is, for example, the degree of deviation between normal biological information when a specific motion in a task is performed and biological information when the worker is actually performing the specific motion.

The biological condition analysis model 152 is prepared for each worker and for each task performed by the worker. That is, a dedicated biological condition analysis model 152 is prepared for each task performed by each worker. Therefore, in order to generate the biological condition analysis model 152, each worker performs the movements for each task while wearing the wearable device 10 in advance, and obtains biological information corresponding to the movements for each task as learning data.

(5-3) Predictor detection model 153
The sign detection model 153 is a model that detects signs of the occurrence of an industrial accident. The sign detection model 153 is a model (trained model) that has learned the relationship between the motion state or biological state and the probability of the occurrence of an industrial accident. The sign detection model 153 receives as input information output from the motion state analysis model 151 or biological state analysis model 152 (the analysis unit 162 described later), and outputs the probability of the occurrence of an industrial accident (hereinafter also referred to as the "occupational accident occurrence probability").
Here, an example of the probability of occurrence of an industrial accident that is output according to a combination of the motion state and biological information input to the sign detection model 153 will be described.

For example, assume that the degree of deviation indicated by the motion state input to the sign detection model 153 is large, and the degree of deviation indicated by the biometric information is also large. In this case, it is estimated from the motion state that the worker is performing an unusual motion, and it is estimated from the biometric information that the worker's heart rate and brain waves are disturbed. An unusual motion may be performed, for example, when responding to an unexpected situation, when avoiding an accident, or when an accident has already occurred. Disturbances in heart rate and brain waves may occur when a worker is working in a hurry, or when an accident has already occurred and the worker is injured. For this reason, when a worker is performing an unusual motion and it is estimated that the worker's heart rate and brain waves are disturbed, it is estimated that the worker is in a state where an industrial accident is highly likely to occur. Therefore, the sign detection model 153 outputs, for example, a high probability of an industrial accident occurring.

Let us also assume that the degree of deviation indicated by the motion state input to the sign detection model 153 is large, while the degree of deviation indicated by the biometric information is small. In this case, it is estimated from the motion state that the worker is behaving differently than usual, and from the biometric information that the worker's heart rate and brain waves are stable. From this, it is estimated that the worker is behaving differently than usual but is responding calmly, and therefore it is estimated that the worker is in a state where an industrial accident is likely to occur. Therefore, the sign detection model 153 outputs, for example, a slightly higher probability of an industrial accident occurring.

Let us also assume that the degree of deviation indicated by the motion state input to the sign detection model 153 is small, and the degree of deviation indicated by the biometric information is large. In this case, it is estimated from the motion state that the worker is performing the same motions as usual, and from the biometric information that the worker's heart rate and brain waves are disturbed. From this, it is estimated that the worker is performing the same motions as usual in a hurry, and therefore the worker is estimated to be in a state where a work-related accident is somewhat likely to occur. Therefore, the sign detection model 153 outputs, for example, a slightly lower probability of a work-related accident occurring.

Also, assume that the degree of deviation indicated by the motion state input to the sign detection model 153 is small, and the degree of deviation indicated by the biometric information is also small. In this case, it is estimated from the motion state that the worker is performing the same motions as usual, and from the biometric information that the worker's heart rate and brain waves are stable. From this, it is estimated that the worker is calmly performing the same motions as usual, and therefore that the worker is in a state where an industrial accident is unlikely to occur. Therefore, the sign detection model 153 outputs, for example, a low probability of an industrial accident occurring.

The trained models in the motion state analysis model 151, biological state analysis model 152, and sign detection model 153 may be models trained using a deep learning model. The trained models in each model may also be models obtained by learning using a non-deep learning type algorithm that can predict non-linear time series data. In this case, the learning time can be shortened, and an AI model can be created using the latest information at each system response. This enables highly accurate predictions in real time.

(6) Control unit 160
The control unit 160 has a function of controlling the overall operation of the wearable device 10. The control unit 160 is realized, for example, by causing a central processing unit (CPU) or a graphics processing unit (GPU), which is provided as hardware in the wearable device 10, to execute a program.
As shown in FIG. 2 , the control unit 160 includes an analysis information acquisition unit 161 , an analysis unit 162 , a sign detection unit 163 , and an output control unit 164 .

(6-1) Analysis information acquisition unit 161
The analytical information acquisition unit 161 has a function of acquiring analytical information. The analytical information acquisition unit 161 acquires data detected in association with the motion of the worker as analytical information. For example, the analytical information acquisition unit 161 acquires acceleration information detected and acquired by the acceleration information acquisition unit 130 in association with the motion of the worker as analytical information. In addition, the analytical information acquisition unit 161 acquires bioinformation detected and acquired by the bioinformation acquisition unit 140 in association with the motion of the worker as analytical information.

(6-2) Analysis Unit 162
The analysis unit 162 has a function of analyzing the working state of the worker based on the analysis information acquired by the analysis information acquisition unit 161. For example, the analysis unit 162 analyzes the motion state of the worker based on the acceleration information, using the motion state analysis model 151. More specifically, the analysis unit 162 inputs the acceleration information acquired by the analysis information acquisition unit 161 to the motion state analysis model 151 stored in the storage unit 150, and obtains the degree to which the worker is performing a specific motion as an output (analysis result).

The analysis unit 162 also analyzes the worker's biological condition based on the biological information and using the biological condition analysis model 152. More specifically, the analysis unit 162 inputs the biological information acquired by the analysis information acquisition unit 161 to the biological condition analysis model 152 stored in the storage unit 150, and obtains the degree to which the worker's biological condition is normal as an output (analysis result).

In addition, when the analysis of the movement state indicates that the worker is performing the movement normally, the analysis unit 162 may analyze the speed at which the worker performs the work by calculating the speed of the movement from the acceleration information. When the analysis of the speed of performance indicates that the speed of performance is slow, it may be estimated that a delay will occur in the progress of the work.

(6-3) Premonition detection unit 163
The sign detection unit 163 has a function of detecting signs of an industrial accident occurring during the work of the worker, based on the analysis result of the working state of the worker by the analysis unit 162. For example, the sign detection unit 163 detects a sign using the sign detection model 153, based on at least one of the motion state and the biological state. More specifically, the sign detection unit 163 inputs the analysis result output from the analysis unit 162 to the sign detection model 153 stored in the storage unit 150, and obtains the probability of an industrial accident occurring as an output.

In addition, the precursor detection unit 163 may not detect a precursor when it is estimated that the worker is performing a specific task based on the motion state indicated by the analysis results, and may detect a precursor when it is estimated that the worker is not performing a specific task.
In addition, the precursor detection unit 163 may not detect precursors when the biological condition of the worker during the work is estimated to be normal based on the biological condition indicated by the analysis results, and may detect precursors when the biological condition of the worker during the work is estimated to be abnormal.

(6-4) Output control unit 164
The output control unit 164 has a function of controlling the output of various information. For example, when an abnormality occurs in the work status of a worker, the output control unit 164 causes the output unit 170 to output an abnormality notification. In this way, the output control unit 164 can inform the worker that an abnormality occurs in the work status.
Furthermore, when a sign of an industrial accident occurring during the work of a worker is detected, the output control unit 164 transmits an industrial accident occurrence sign notification to the supervisor terminal 20 and causes it to be output. In this way, the output control unit 164 can inform the supervisor that there is a sign of an industrial accident occurring during the work of the worker.

In addition, when the output control unit 164 estimates that a delay will occur in the progress of the work based on the results of monitoring the speed at which the worker performs the work by the analysis unit 162, the output control unit 164 may cause the output unit 170 to output an instruction to improve the speed at which the work is performed.

(7) Output unit 170
The output unit 170 has a function of outputting various information. The output unit 170 is configured by an output device included in the wearable device 10 as hardware, for example, a display device such as a display device or a touch screen (touch panel), or an audio output device such as a speaker.

For example, when an abnormality occurs in the work status of a worker, the output unit 170 outputs an abnormality notification input from the output control unit 164. The output unit 170 may display the abnormality notification on a display device, or may output the abnormality notification as audio from an audio output device.

<3. Processing flow>
The functional configuration of the wearable device 10 according to the present embodiment has been described above. Next, a process flow according to the present embodiment will be described with reference to Fig. 3 and Fig. 4. Fig. 3 is a sequence diagram showing an example of a process flow in the work monitoring system 1 according to the present embodiment. Fig. 4 is a flowchart showing an example of a process flow in an analysis process according to the present embodiment.

As shown in FIG. 3, first, a worker puts on the wearable device 10 and starts work (step S101).
Next, the analytical information acquisition unit 161 of the wearable device 10 acquires, as analytical information, the acceleration information acquired by the acceleration information acquisition unit 130 of the wearable device 10 and the biological information acquired by the biological information acquisition unit 140 (step S102).
Next, the analysis unit 162 of the wearable device 10 performs an analysis process based on the analysis information acquired by the analysis information acquisition unit 161 (step S103).

Here, the flow of the analysis process in step S103 will be described with reference to FIG.
As shown in FIG. 4, first, the analysis unit 162 analyzes the motion state of the worker by using the acceleration information acquired as the analysis information and the motion state analysis model 151 (step S201).
Next, the analysis unit 162 determines whether the worker is performing normal operations based on the analysis result of the operation state (step S202). If the worker is performing normal operations (step S202/YES), the process proceeds to step S203. On the other hand, if the worker is not performing normal operations (step S202/NO), the process proceeds to step S204.

When the process proceeds to step S203, the analysis unit 162 analyzes the speed at which the worker performs the task (step S203). After the analysis, the process proceeds to step S203.
When the process proceeds to step S204, the analysis unit 162 analyzes the biological condition of the operator by using the biological information acquired as the analysis information and the biological condition analysis model 152 (step S204).
Next, the sign detection unit 163 of the wearable device 10 detects signs of an industrial accident occurring during the worker's work using the motion state and biological state analyzed by the analysis unit 162 and the sign detection model 153 (step S205).

3, if there is an abnormality in the working state of the worker (step S104/YES), the process proceeds to step S105. On the other hand, if there is no abnormality in the working state of the worker (step S104/NO), the process repeats from step S102.
When the process proceeds to step S105, the output control unit 164 of the wearable device 10 causes the output unit 170 to output an abnormality notification (step S105).
The worker checks the abnormality notification output by the output unit 170 and takes action in response to the abnormality notification (step S106).

Next, if a sign of an industrial accident occurring is detected (step S107/YES), the process proceeds to step S108. On the other hand, if a sign of an industrial accident occurring is not detected (step S107/NO), the process proceeds to step S110.

When the process proceeds to step S108, the output control unit 164 transmits an industrial accident sign notification to the supervisor terminal 20 and causes it to be output (step S108).
The supervisor checks the industrial accident warning notice sent to the supervisor terminal 20 and takes action (step S109).

When the process proceeds to step S110, the output control unit 164 transmits the work information to the management DB 30 (step S110). The management DB 30 stores the work information received from the wearable device 10 (step S111).
The supervisor terminal 20 transmits a request to the management DB 30 to refer to the work information, making the work information available for reference (step S112). The supervisor checks the work information from the supervisor terminal 20 (step S113).

As described above, the wearable device 10 (work monitoring device) according to this embodiment includes an analysis information acquisition unit 161 that acquires data detected in association with the movements of the worker as analysis information indicating information for analyzing the work state of the worker, an analysis unit 162 that analyzes the work state of the worker based on the acquired analysis information, and a sign detection unit 163 that detects signs of an industrial accident occurring during the work of the worker based on the analysis results of the work state of the worker.

With this configuration, it is possible to detect signs of an industrial accident occurring during the work of a worker from the working condition of the worker. This allows the worker or supervisor to be notified that a sign of an industrial accident occurring has been detected, making it possible to improve the working condition before an industrial accident occurs.
Therefore, the wearable device 10 according to this embodiment makes it possible to prevent the occurrence of work-related accidents.

4. Modifications
The above describes the embodiment. Next, modified examples of the above embodiment will be described. Each modified example described below may be applied to the embodiment alone or in combination. Each modified example may be applied in place of the configuration described in the embodiment, or may be applied in addition to the configuration described in the embodiment.

In the above-described embodiment, an example has been described in which the functions of the work monitoring device are realized by the wearable device 10, but the present invention is not limited to such an example. For example, some or all of the functions of the work monitoring device may be realized by one or more server devices (e.g., cloud servers). Also, some or all of the functions of the work monitoring device may be realized by a quantum computer.

In the above-described embodiment, the function of the acceleration information acquisition unit 130 is realized by an acceleration sensor, and an example is described in which acceleration information detected by the acceleration sensor is acquired, but the invention is not limited to such an example. For example, the acceleration information acquisition unit 130 may acquire an image (moving image or multiple still images) showing the worker working, captured by a camera installed near the worker, and calculate the acceleration based on the image, thereby acquiring acceleration information.

In addition, in the above-described embodiment, an example has been described in which the motion state analysis model 151 and the biological condition analysis model 152 are prepared for each worker and each task, but the present invention is not limited to such an example. For example, a single motion state analysis model 151 and a single biological condition analysis model 152 may be prepared for multiple workers. In other words, multiple workers may be able to use the same model.

In the above-described embodiment, an example has been described in which the sign detection unit 163 uses the sign detection model 153 that has learned the relationship between the motion state or biological state and the probability of an industrial accident occurring in the sign detection, but the example is not limited to this. For example, the sign detection unit 163 may perform sign detection using a model that has learned the relationship between motions that are likely to cause industrial accidents and the probability of an industrial accident occurring. In this case, the sign detection unit 163 may detect a sign when it is estimated that the worker is performing a motion that is likely to cause an industrial accident in the course of work based on the motion state indicated by the analysis result, and may not detect a sign when it is estimated that the worker is not performing a motion that is likely to cause an industrial accident in the course of work.

In addition, in the above-mentioned embodiment, an example has been described in which the analysis unit 162 uses a trained model in analyzing the motion state, analyzing the biological state, and detecting signs, but the present invention is not limited to such an example. For example, the analysis unit 162 may analyze the motion state and the biological state based on a comparison of the acquired acceleration information and biological information with a preset threshold value. The analysis unit 162 may also perform sign detection based on a comparison of the results of the analysis of the motion state and the analysis of the biological state with a preset threshold value.

In addition, in the above-described embodiment, information that is the reason for judging the working status of the worker may be calculated using a technology or method classified as Explainable AI, and may be displayed on the wearable device 10, the supervisor terminal 20, etc.

The above describes the modified examples of the embodiment.
In addition, some or all of the functions of the work monitoring system 1 and the wearable device 10 (work monitoring device) in the above-mentioned embodiment may be realized by a computer. In that case, a program for realizing the functions may be recorded in a computer-readable recording medium, and the program recorded in the recording medium may be read into a computer system and executed to realize the functions. In addition, the term "computer system" as used herein includes hardware such as an OS and peripheral devices. In addition, the term "computer-readable recording medium" refers to portable media such as a flexible disk, an optical magnetic disk, a ROM, and a CD-ROM, and a storage device such as a hard disk built into a computer system. Furthermore, the term "computer-readable recording medium" may include a medium that dynamically holds a program for a short period of time, such as a communication line when a program is transmitted via a network such as the Internet or a communication line such as a telephone line, and a medium that holds a program for a certain period of time, such as a volatile memory inside a computer system that is a server or client in that case. Furthermore, the above program may be for realizing a part of the above-mentioned functions, or may be capable of realizing the above-mentioned functions in combination with a program already recorded in a computer system, or may be realized by using a programmable logic device such as an FPGA (Field Programmable Gate Array).

The above describes in detail the embodiments of the present invention with reference to the drawings, but the specific configuration is not limited to the above, and various design changes can be made without departing from the spirit of the present invention.

1...work monitoring system, 10...wearable device, 20...supervisor terminal, 30...management DB, 110...communication unit, 120...input unit, 130...acceleration information acquisition unit, 140...biometric information acquisition unit, 150...storage unit, 151...motion state analysis model, 152...biometric state analysis model, 153...sign detection model, 160...control unit, 161...analysis information acquisition unit, 162...analysis unit, 163...sign detection unit, 164...output control unit, 170...output unit, NW...network

Claims (13)

an analysis information acquisition unit that acquires data detected in association with a motion of a worker as analysis information indicating information for analyzing a working state of the worker;
an analysis unit that analyzes a work state of the worker based on the acquired analysis information;
a sign detection unit that detects a sign of an industrial accident occurring during the work of the worker based on the analysis result of the work state of the worker;
A work monitoring device comprising: The work state of the worker includes a motion state in which the worker is performing a motion for the work and a biological state of the worker performing the work,
The sign detection unit detects the sign based on at least one of the motion state and the biological state.
The work monitoring device according to claim 1 . The sign detection unit uses a sign detection model that has learned a relationship between the motion state or the biological state and the probability of occurrence of the industrial accident, and outputs the probability of occurrence of the industrial accident using an analysis result output from the analysis unit as an input.
The work monitoring device according to claim 2 . the sign detection unit does not detect the sign when it is estimated that the worker is performing a specific motion in the work based on the motion state indicated by the analysis result, and detects the sign when it is estimated that the worker is not performing the specific motion in the work.
The work monitoring device according to claim 2 . The sign detection unit detects the sign when it is estimated that the worker is performing an action that is likely to cause the industrial accident during the work based on the motion state indicated by the analysis result, and does not detect the sign when it is estimated that the worker is not performing an action that is likely to cause the industrial accident during the work.
The work monitoring device according to claim 2 . the sign detection unit does not detect the sign when the biological condition of the worker during the work is estimated to be normal based on the biological condition indicated by the analysis result, and detects the sign when the biological condition of the worker during the work is estimated to be abnormal.
The work monitoring device according to claim 2 . The analysis information acquisition unit acquires acceleration information associated with a motion of the worker as the analysis information,
The analysis unit analyzes the motion state based on the acceleration information.
The work monitoring device according to claim 2 . The analysis unit uses a motion state analysis model that has learned a relationship between a specific motion in the work and acceleration information detected when the worker performs the specific motion, and outputs a degree to which the worker is performing the specific motion, using the acceleration information acquired by the analysis information acquisition unit as an input.
The work monitoring device according to claim 7. the analysis information acquisition unit acquires biological information associated with the motion of the worker as the analysis information;
The analysis unit analyzes the biological condition based on the biological information.
The work monitoring device according to claim 2 . The analysis unit uses a biological condition analysis model that has learned a relationship between a specific movement in the work and biological information detected when the worker performs the specific movement, and outputs a degree to which the biological condition of the worker is normal, using the biological information acquired by the analysis information acquisition unit as an input.
The work monitoring device according to claim 9. an output control unit that outputs a notification to a terminal used by a supervisor of the work when a sign of an industrial accident occurring during the work of the worker is detected;
The work monitoring device according to claim 1 , further comprising: an analysis information acquisition step in which an analysis information acquisition unit acquires data detected in association with a motion of the worker as analysis information indicating information for analyzing a working state of the worker;
an analysis step of analyzing a working state of the worker based on the acquired analysis information by an analysis unit;
a sign detection step in which a sign detection unit detects a sign of an industrial accident occurring in the work of the worker based on an analysis result of the work state of the worker;
A work monitoring method comprising: Computer,
an analysis information acquisition means for acquiring data detected in association with the motion of the worker as analysis information indicating information for analyzing the working state of the worker;
an analysis means for analyzing a working state of the worker based on the acquired analysis information;
a sign detection means for detecting a sign of an industrial accident occurring during the work of the worker based on the analysis result of the work state of the worker;
A program to function as a

Images