Risk management systems and risk management methods

Abstract

The invention proposes a system and method for risk management. In this system, the upper layer is connected to multiple lower layers. The risk management system manages predicted risks, including first-sensor data collected by the first lower layer capable of collecting first-sensor data, and second-sensor data collection by the second lower layer capable of collecting second-sensor data different from the first-sensor data. It performs arithmetic processing to generate first-normalized data based on the first-sensor data, second-normalized data based on the second-sensor data, and the normalization processing unit generates integrated normalized data by performing arithmetic processing to combine the first-normalized and second-normalized data into a single dataset. The predictive processing unit generates predicted risk data based on the integrated normalized data.

Description

Risk management system and risk management method

[0001] The present invention relates to a risk management system and a risk management method.

[0002] In recent years, the importance of employee labor management has increased in corporate activities. In particular, there is a need to manage risks that may impair employee health, such as working environment and overtime hours, so that employees can work safely and comfortably. Patent Document 1 discloses an example of a health monitoring system.

[0003] JP 2009-61221 A

[0004] In the case of the system described in Patent Document 1, in conventional crisis monitoring systems, the risk level is set based on information acquired by a specific device. Therefore, when different types of information are acquired by multiple different types of devices, it is not possible to combine the information to determine the risk level.

[0005] In view of the above problems, one object of the present invention is to perform risk management easily and accurately using information acquired by a plurality of types of devices.

[0006] According to one embodiment of the present invention, there is provided a risk management system that is an upper-layer system connected to multiple lower-layer systems. The risk management system includes an acquisition unit, a normalization processing unit, and a prediction processing unit. The acquisition unit acquires first sensing data acquired by a first lower-layer system capable of acquiring a first type of data and second sensing data acquired by a second lower-layer system capable of acquiring a second type of data different from the first sensing data. The normalization processing unit generates first normalized data based on the first sensing data and performs calculations to generate second normalized data based on the second sensing data, and also performs calculations to combine the first normalized data and the second normalized data into a single integrated data to generate integrated normalized data. The prediction processing unit generates predicted risk data based on the integrated normalized data.

[0007] The risk management system may include a learning processing unit. The learning processing unit may generate an integrated normalized machine learning model by machine learning on second integrated normalized data generated by a normalization processing unit using first sensing data and second sensing data acquired in advance. The prediction processing unit may apply the integrated normalized data to the integrated normalized machine learning model to generate the risk data.

[0008] In the above-mentioned risk management system, the normalization processing unit may set a first normalization reference point corresponding to the first sensing data and a second normalization reference point corresponding to the second sensing data based on predetermined conditions, apply the first sensing data and the first normalization reference point to a first normalization method for generating the first normalized data to generate the first normalized data, and apply the second sensing data and the second normalization reference point to a second normalization method for generating the second normalized data to generate the second normalized data.

[0009] In the above risk management system, the first normalization method may be set based on first identification information associated with the first sensing data, and the second normalization method may be set based on second identification information associated with the second sensing data.

[0010] In the above-mentioned risk management system, the normalization processing unit may assign a weight value to at least one of the first normalized data, the second normalized data, and the integrated normalized data based on predetermined conditions, and the prediction processing unit may adjust the risk data based on the weight value.

[0011] In the above risk management system, the acquisition unit may acquire at least one of user information and environmental information, and the predetermined condition may be set based on at least one of the user information and the environmental information.

[0012] The risk management system may further include a warning information sending unit that sends warning information to a user terminal in accordance with the risk data.

[0013] According to one embodiment of the present invention, a risk management method is provided, which includes: a risk management system, which is an upper layer system connected to a plurality of lower layer systems, acquiring first sensing data acquired by a first lower layer system capable of acquiring a first type of data and second sensing data acquired by a second lower layer system capable of acquiring a second type of data different from the first sensing data; generating first normalized data based on the first sensing data; generating second normalized data based on the second sensing data; performing calculations to combine the first normalized data and the second normalized data into a single integrated data to generate integrated normalized data; and generating risk data based on the integrated normalized data.

[0014] In the above risk management method, the method may further include generating an integrated normalized machine learning model by machine learning targeting second integrated normalized data generated by performing the normalization process using the first sensing data and second sensing data previously acquired by the acquisition unit, and generating the predicted risk data may generate the risk data by applying the integrated normalized data to the integrated normalized machine learning model.

[0015] In the above risk management method, generating the first normalized data and the second normalized data may include setting a first normalization reference point corresponding to the first sensing data and a second normalization reference point corresponding to the second sensing data based on a predetermined condition, applying the first sensing data and the first normalization reference point to a first normalization method for generating the first normalized data to generate the first normalized data, and applying the second sensing data and the second normalization reference point to a second normalization method for generating the second normalized data to generate the second normalized data.

[0016] In the above risk management method, the first normalization method may be set based on first identification information associated with the first sensing data, and the second normalization method may be set based on second identification information associated with the second sensing data.

[0017] In the above risk management method, generating the risk data may include assigning a weight value to at least one of the first normalized data, the second normalized data, and the integrated normalized data based on a predetermined condition, and adjusting the generated risk data based on the weight value.

[0018] In the above risk management method, when acquiring the first sensing data and the second sensing data, at least one of user information and environmental information may be acquired, and the specified condition may be set based on at least one of the user information and the environmental information.

[0019] In the risk management method, warning information may be transmitted to a user terminal in accordance with the risk data.

[0020] By using one embodiment of the present invention, risk management can be easily and accurately performed using information acquired by multiple types of devices.

[0021] 1 is a diagram showing the hardware configuration of a risk management collaboration system according to one embodiment of the present invention. FIG. 2 is a functional block diagram composed of each component of a risk management collaboration system according to one embodiment of the present invention. FIG. 3 is a schematic diagram showing the flow of data from a lower-layer system to an upper-layer system. FIG. 4 is an example of a dataset including first sensing data and first normalized data. FIG. 5 is an example of a dataset including second sensing data and second normalized data. FIG. 6 is an example of a dataset including third sensing data and third normalized data. FIG. 7 is an example of a dataset including integrated normalized data. FIG. 8 is an example of a dataset including integrated normalized data and risk data. FIG. 9 is a flow diagram of a risk management method according to one embodiment of the present invention. FIG. 10 is a flow diagram of a risk management method according to one embodiment of the present invention. FIG. 11 is a flow diagram of a risk management method according to one embodiment of the present invention. FIG. 12 is a flow diagram of a risk management method according to one embodiment of the present invention. FIG. 13 is a flow diagram of a risk management method according to one embodiment of the present invention. FIG. 14 is a flow diagram of a risk management method according to one embodiment of the present invention. FIG. 15 is a data table used in normalization processing. FIG. 16 is a flow diagram of a risk management method according to one embodiment of the present invention.

[0022] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the present invention can be implemented in many different forms, and should not be construed as being limited to the description of the embodiments exemplified below. The drawings may be diagrammatic in order to clarify the explanation, but they are merely examples and do not limit the interpretation of the present invention.

[0023] Furthermore, the letters "first" and "second" attached to each element are convenient symbols used to distinguish between elements and have no further meaning unless otherwise specified. In the drawings referred to in this embodiment, identical parts or parts having similar functions are designated by the same or similar symbols (symbols consisting of the numeral xxx with A, B, 1, 2, etc. attached thereto), and repeated explanations thereof may be omitted. Furthermore, some components may be omitted from the drawings. In addition, if a person having ordinary knowledge in the field to which the present invention pertains would recognize such components, no special explanations will be given.

[0024] First Embodiment A risk management collaboration system and a risk management method according to a first embodiment of the present invention will be described in detail with reference to the drawings.

[0025] (1-1. Hardware Configuration of the Risk Management Collaboration System) Figure 1 shows the hardware configuration of the risk management collaboration system 1. In a portion of the configuration of the risk management collaboration system 1, risk management processing is executed to perform more accurate risk management. As shown in Figure 1, the risk management collaboration system 1 includes a risk management system 10 (also referred to as an upper-layer system), a lower-layer system 20 (also referred to as a sensing device), and a user terminal 30. The risk management system 10 functions as an application server that acquires sensing data input from multiple lower-layer systems 20, stores it in a database for risk management, normalizes and integrates each acquired sensing data, and generates risk data. The risk management collaboration system 1 is composed of multiple hierarchical systems, such as lower-layer systems and upper-layer systems.

[0026] The lower layer system 20 has the function of acquiring various types of sensing data. The lower layer system 20 includes multiple lower layer systems. In this example, a first lower layer system 20-1, a second lower layer system 20-2, and a third lower layer system 20-3 are provided. Specifically, the first lower layer system 20-1 has the function of acquiring heat index data. The second lower layer system 20-2 has the function of acquiring cumulative workload data. The third lower layer system 20-3 has the function of acquiring cumulative stay time data. Note that when it is not necessary to explain each lower layer system separately, they will be explained as the lower layer system 20.

[0027] The user terminal 30 is a terminal that inputs information related to sensing data to the risk management system 10 and the lower layer system 20 and transmits and receives various information. In Figure 1, one user terminal 30 is provided, but multiple user terminals may be provided. The configuration of each hardware in this embodiment is shown below.

[0028] The risk management system 10 has a control unit 11, a memory unit 12, a communication unit 13, and a display unit 14. The control unit 11, the memory unit 12, the communication unit 13, and the display unit 14 are connected via a bus. In this embodiment, a risk management server is used for the risk management system 10. Note that the risk management system 10 may use an information processing device such as a PC (Personal Computer), or may be composed of multiple information processing devices.

[0029] The control unit 11 is a computer, and uses a CPU (Central Processing Unit), ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array), or other arithmetic processing circuit to control processing based on instructions defined in software (program) for executing risk management processing. In addition, a user interface for executing the risk management program may be provided on the display unit 14 in response to an instruction from the control unit 11.

[0030] The storage unit 12 may use semiconductor memory such as an SSD (Solid State Drive), as well as magnetic recording media (magnetic tape, magnetic disk, etc.), optical recording media, magneto-optical recording media, and other storage devices. The storage unit 12 stores the risk management program and also functions as a risk management database 12a that stores sensing data, normalized data, etc. used in the risk management program. The storage unit 12 may be provided on a server different from the risk management system 10 as appropriate and function as a database.

[0031] The communication unit 13 connects to the network 40 under the control of the control unit 11 and transmits and receives information to and from external devices. The communication unit 13 includes, for example, a modem or a network interface card (NIC). Communication between the communication unit 13 and the lower layer system 20 and the user terminal 30 is performed via the Internet (specifically, SSL (Secure Sockets Layer) / TLS (Transport Layer Security) or VPN (Virtual Private Network)) or an intranet.

[0032] The lower layer system 20 has a control unit 21, a storage unit 22, a communication unit 23, and a display unit 24. The control unit 21, the storage unit 22, the communication unit 23, and the display unit 24 are connected via a bus. In this embodiment, the lower layer system 20 may be configured as a single information processing device or may be configured as multiple information processing devices.

[0033] The control unit 21 is a computer that uses a CPU, ASIC, FPGA, or other arithmetic processing circuit to control processing based on instructions specified in a program for acquiring sensing data and a program related to the risk management program. In addition, the display unit 24 of the lower layer system 20 may be provided with a user interface for executing the above-mentioned programs in response to instructions from the control unit 21.

[0034] The storage unit 22 may be a semiconductor memory such as an SSD, or may be a magnetic recording medium, an optical recording medium, a magneto-optical recording medium, or any other storage element. The storage unit 22 stores programs related to the risk management program, and also functions as a program for acquiring sensing data and a database for storing sensing data. The storage unit 22 may be provided in a server different from the lower layer system 20 as appropriate, and may function as a database.

[0035] The communication unit 23 has a transceiver and communicates the risk management system 10, the user terminal 30, and other related information via the network 40. The communication unit 23 includes, for example, a modem or a NIC. The communication between the communication unit 23 and the risk management system 10 and the user terminal 30 uses the Internet (specifically, SSL / TLS or VPN) or an intranet.

[0036] The user terminal 30 has a display unit 31, a control unit 32, a memory unit 33, an operation unit 34, and a communication unit 35. The display unit 31, the control unit 32, the memory unit 33, the operation unit 34, and the communication unit 35 are connected via a bus. The user terminal 30 may be a smartphone, a mobile phone (feature phone), a tablet terminal, a notebook PC, a desktop PC, an IoT (Internet of Things) device (for example, a device equipped with a power source, a control function, a communication function, and an information storage function), or the like, and any device that can communicate with the risk management system 10 via a network is applicable.

[0037] The display unit 31 is a display device such as a liquid crystal display or an organic EL (Electro Luminescence) display, and the display content is controlled by a signal input from the control unit 32 .

[0038] The control unit 32 is a type of computer and includes a CPU, ASIC, FPGA, or other arithmetic processing circuit. The control unit 32 executes a program stored in a storage unit 33 such as a memory based on operations on the display unit 31 and the operation unit 34. The control unit 32 also transmits information for instructing the execution of processing related to the risk management program stored in the storage unit 12 of the risk management system 10.

[0039] The operation unit 34 includes a controller, a button, or a switch. When the user uses the operation unit 34 to move up, down, left, or right, press, rotate, or perform other actions, information based on the actions is input to the control unit 32. Note that if the display unit 31 is a display device (touch panel) having a touch sensor, the display unit 31 and the operation unit 34 may be located in the same place.

[0040] The communication unit 35 has the function of transmitting and receiving data to and from the risk management system 10. For example, a transceiver via a LAN is used for the communication unit 35. Note that the communication unit 35 is not limited to a transceiver via a LAN, and may be provided with a transceiver for mobile terminal communication (e.g., LTE communication), or a transceiver for short-range wireless communication. The user terminal 30 is connected to the risk management system 10 via a network 40.

[0041] (1-2. Configuration of the risk management control unit 100) Fig. 2 is a functional block diagram made up of the components of the risk management collaboration system 1. Fig. 3 is a schematic diagram showing the flow of data to the risk management system 10 (upper layer system).

[0042] The risk management system 10 has a risk management control unit 100 that controls a program (risk management program) that realizes the risk management function. The risk management control unit 100 includes an acquisition unit 110, a normalization processing unit 120, a learning processing unit 130, a prediction processing unit 140, an analysis unit 150, and a warning information transmission unit 160.

[0043] The acquisition unit 110 has a function of acquiring various types of sensing data and other data (including normalized data) necessary for generating risk data in the risk management collaboration system 1. For example, the acquisition unit 110 may have a function of acquiring first sensing data acquired or generated in the first lower layer system 20-1, second sensing data acquired or generated in the second lower layer system 20-2, and third sensing data acquired or generated in the third lower layer system 20-3.

[0044] The normalization processing unit 120 has a function of performing calculation processing to normalize the acquired sensing data. In this embodiment, "normalization" refers to setting a reference value for the actual data (sensing data) acquired by the lower layer system 20 and generating relative data based on the reference point. Normalization makes it possible to use the same judgment index for data with different data formats or units. This allows each sensing data to be judged in the same way. Details of the normalization process will be described later.

[0045] 4 is an example of a first data set 510 of first sensed data and data associated with the first sensed data. In this example, the first data set 510 includes heat index data 510a and first normalized data 510b acquired in a first lower-layer system.

[0046] 5 shows an example of the second sensing data and the second data set 520 of data associated with the second sensing data. In this example, the second data set 520 includes cumulative workload data 520a and second normalized data 520b acquired in the second lower layer system.

[0047] 6 is an example of the third sensing data and the third data set 530 of data associated with the third sensing data. In this example, the third data set 530 includes cumulative dwell time data 530a and third normalized data 530b acquired in the third lower layer system.

[0048] 7 shows an example of an integrated normalized data dataset 540. In this example, the integrated normalized data dataset 540 is formed by sequentially concatenating and integrating first normalized data 510b, second normalized data 520b, and third normalized data 530b.

[0049] 8 shows an example of a dataset 550 for constructing a machine learning model. The dataset 550 includes risk data 560 serving as training data, along with concatenated first normalized data 510b, second normalized data 520b, and third normalized data 530b.

[0050] The learning processing unit 130 has the function of generating an integrated normalized machine learning model by machine learning targeting the second integrated normalized data generated by the normalization processing unit 120 using the first sensing data and second sensing data acquired in advance.

[0051] The prediction processing unit 140 has a function of generating predicted risk data based on the integrated normalized data. The prediction processing unit 140 also has a function of performing arithmetic processing to combine the normalized data generated by the normalization processing into a single piece of data (integrated normalized data), and can generate risk data based on the integrated normalized data.

[0052] The analysis unit 150 has a function of comparing and analyzing the generated risk data with risk data set under predetermined conditions, and determining whether or not there is a risk.

[0053] The warning information transmission unit 160 has a function of generating and transmitting warning information for notifying the user terminal 30 of the existence of a risk based on the risk data. The warning information may be display data such as character data or image data to be displayed on the display unit 31, or may be music data.

[0054] The lower layer system 20 executes processing prescribed in a program related to the risk management processing program stored in the memory unit 22. The lower layer system 20 has a receiving unit 210 and a transmitting unit 220 as functional units. The receiving unit 210 has a function of receiving various data from the risk management system 10 and the user terminal 30. The transmitting unit 220 has a function of transmitting various risk management information including sensing data to the risk management system 10.

[0055] The user terminal 30 executes processing prescribed in a program related to the risk management processing program stored in the memory unit 33. The user terminal 30 has a receiving unit 310 and a transmitting unit 320 as functional units. The receiving unit 310 has a function of receiving warning information based on risk data from the risk management system 10. The transmitting unit 320 has a function of transmitting part of the information related to the risk management program (e.g., user information) to the risk management system 10 and the lower-layer system 20.

[0056] (1-3. Risk management control processing) Next, we will explain the risk management control processing based on instructions from the risk management program in the risk management control unit 100. Figure 9 is a flow diagram of the risk management control processing. As shown in Figure 9, the risk management control processing includes a first risk management control processing S100, a second risk management control processing S200, and a third risk management control processing S300.

[0057] The first risk management control process S100 includes processes for generating and acquiring first sensing data (heat index data), second sensing data (cumulative workload data), and third sensing data (cumulative stay time data). The second risk management control process S200 includes processes for setting a reference point for sensing data, generating normalized data from the sensing data, and generating risk data. The third risk management control process S300 includes a comparison and analysis process and a warning information generation process. Each risk management control process will be explained separately.

[0058] (1-3-1. First Risk Management Control Process) Figure 10 is a flow diagram of the first risk management control process S100. As shown in Figure 10, first, a user of a user terminal 30 inputs user data and environmental data into the user terminal 30 (S101). Figure 11 shows a data set 600 of the input user data and environmental data. In this example, a user ID 610 and a user name 620 are input as user data, and a location ID 630 and a site name 640 are input as environmental data in the data set 600. The user terminal 30 transmits the input user data and environmental data to the risk management system 10 (S103), and the risk management system 10 receives and acquires the transmitted user data (S105).

[0059] The risk management system 10 generates and transmits sensing data generation instruction information to the lower layer system 20 (S107, S109). Upon receiving the generation instruction information (S111), the lower layer system 20 generates or acquires sensing data based on the user information (S111). At this time, the sensing data may be pre-generated data stored in the memory unit 22 (database) of the lower layer system 20, or may be newly generated or acquired based on the sensing data generation instruction information. In addition, the lower layer system 20 may acquire new sensing data every time a certain period of time has passed.

[0060] In this embodiment, heat index (WBGT: Wet Bulb Globe Temperature) data is acquired as first sensing data from the first lower layer system 20-1. The heat index is an index relating to the transfer of heat (heat balance) between the human body and the outside air, and is generated based on factors that have a large impact on the heat balance of the human body. The factors used include (1) humidity, (2) the surrounding thermal environment such as solar radiation and radiation, and (3) air temperature. The heat index data may be used in combination with geographical information. Note that a weather forecast DB device is used as the first lower layer system 20-1 in this embodiment.

[0061] The second lower layer system 20-2 generates and acquires cumulative workload data as the second sensing data. In this case, a weather forecast DB device is used as the second lower layer system 20-2. The cumulative workload data is the amount of time (overtime hours) exceeding the basic work hours per month. A work performance and health sensor DB device may also be used as the second lower layer system 20-2.

[0062] The third lower layer system 20-3 generates cumulative stay time data as the third sensing data. The cumulative stay time data per month is calculated by subtracting the arrival time from the departure time on each work day. A work performance and health sensor DB device may be used as the third lower layer system 20-3.

[0063] 3, the lower layer system 20 transmits the acquired sensing data to the risk management system 10 (S113). Upon receiving the sensing data (S115), the risk management system 10 stores the received sensing data in the risk management database 12a provided in the storage unit 12 (S115). This completes the first risk management control process S100.

[0064] (1-3-2. Second Risk Management Control Processing) Figure 12 is a flow diagram of the second risk management control processing S200. The second risk management control processing S200 is executed following the first risk management control processing S100. First, the risk management system 10 generates normalized data based on the acquired sensing data (S201). The generation of normalized data will be described in detail below.

[0065] 13 is a flow diagram of the normalized data generation process S201. First, a normalization reference value (S2011), a normalized maximum value (S2013), and a normalized minimum value (S2015) associated with each piece of sensing data are set. The normalization reference value, the normalized maximum value, and the normalized minimum value can be set arbitrarily.

[0066] In the case of the heat index data, which is the first sensing data, in this example, the heat index data "15°C" is set as the normalization reference value "0". Similarly, the heat index data "30°C" is set as the normalization maximum value "1". Similarly, the heat index data "0°C" is set as the normalization minimum value "-1".

[0067] Once the above settings have been made, the sensing data is then normalized based on a predetermined formula (normalization function) (S2017). In this example, the normalization is performed based on the following formula 1 (linear function). If identification information is set for each sensing data, the risk management system 10 may set a normalization method based on that information and perform calculation processing. In Equation 1, "Normalvalue" indicates normalized data. "Value" indicates sensing data. "Max" indicates the maximum value of sensing data. "Min" indicates the minimum value of sensing data. "t" indicates the maximum normalized value. "s" indicates the minimum normalized value. In this example, in Equation 1, the sensing data and the normalized data have a linear relationship.

[0068] When the sensing data and the normalization reference value are applied to Equation 1 and calculation processing is performed, first normalized data is generated uniquely corresponding to each heat index data as shown in FIG.

[0069] The second sensing data is also subjected to normalization processing in the same manner as the first sensing data. In the case of the cumulative workload data, which is the second sensing data, the cumulative workload data of "70 hours" is set as the normalization reference value of "0". Similarly, the cumulative workload data of "80 hours" is set as the maximum normalization value of "1". Similarly, the cumulative workload data of "60 hours" is set as the minimum normalization value of "-1".

[0070] Once the above settings are made, normalization processing is performed based on Equation 1. At this time, second normalized data is generated uniquely corresponding to each piece of cumulative workload data, as shown in FIG.

[0071] The third sensing data is also subjected to normalization processing in the same manner as the first sensing data. In the case of the cumulative stay time data, which is the third sensing data, the monthly cumulative stay time data of "120 hours" is set as the normalization reference value of "0". Similarly, the monthly cumulative stay time data of "130 hours" is set as the maximum normalization value of "1". Similarly, the cumulative stay time data of "110 hours" is set as the minimum normalization value of "-1".

[0072] Once the above settings are made, normalization processing is performed based on Equation 1. At this time, third normalized data is generated uniquely corresponding to each piece of cumulative stay time data, as shown in FIG.

[0073] Next, the generated normalized data is concatenated and integrated (S203). In this embodiment, the first normalized data, the second normalized data, and the third normalized data are concatenated and integrated in this order, as shown in Figs. 3 and 7 .

[0074] Next, the risk management system 10 applies the integrated normalized data to a machine learning model using the concatenated normalized data (S205) to generate predicted risk data (S207). In this example, the risk management system 10 generates risk data as output data. In this example, the risk data is set with "0" as the minimum value and "100" as the maximum value. Note that symbols such as C1, C2, ..., Cn may be used for the risk data.

[0075] The machine learning model used in this embodiment will be described in detail below. FIG. 14 is a flow diagram showing the machine learning model generation method S205. As shown in FIG. 14, the normalization processing unit 120 of the risk management system 10 generates normalized data using sensing data previously acquired in the risk management database 12a of the storage unit 12 (S2051). The previously acquired sensing data may be, for example, one week's worth of data. Next, the normalization processing unit 120 generates normalized data (also referred to as second integrated normalized data) by concatenating and integrating the normalized data using the normalized data (S2053). Next, the learning processing unit 130 generates risk data that serves as training data in accordance with the second integrated normalized data (S2055) and stores the risk data in the database. FIG. 8 is an example of a dataset 550 for constructing a machine learning model. The dataset 550 includes risk data 560 that serves as training data, along with concatenated first normalized data 510b, second normalized data 520b, and third normalized data 530b. The training data may be set in advance by information input by a user. The learning processing unit 130 performs machine learning using an artificial intelligence (AI) algorithm on the integrated normalized data (and the training data) (S2057), and forms a machine learning model (S2059).

[0076] FIG. 15A shows an example of a dataset of sensing data used as prediction data and normalized data corresponding to each sensing data. FIG. 15B shows an example of a dataset of concatenated and integrated integrated normalized data and risk data generated from the integrated normalized data. In this example, predicted heat index data of "29°C," predicted cumulative workload data of "75 hours," and predicted cumulative stay time data of "128 hours" are acquired as sensing data. At this time, first normalized data of "0.93" is generated from the predicted heat index data of "29°C." Similarly, second normalized data of "0.50" is generated from the predicted cumulative workload data of "75 hours." Furthermore, third normalized data of "0.80" is generated from the predicted cumulative stay time data of "128 hours." Next, the first normalized data, second normalized data, and third normalized data are concatenated and integrated to generate integrated normalized data (0.93, 0.50, 0.80), which is input as machine learning input data. As a result, the risk data output is 90. This completes the second risk management control process S200.

[0077] (1-3-3. Third Risk Management Control Processing) FIG. 16 is a flow diagram of the third risk management control processing S300. The third risk management control processing S300 is executed following the second risk management control processing S200. First, the risk management system analyzes the generated risk data (S301). At this time, the risk management system 10 determines whether the risk data exceeds a set value (S303). The set value is set arbitrarily. In this example, the risk management system 10 determines whether the risk data exceeds 80.

[0078] When the risk data does not exceed the set value (S303; No), the risk management system 10 determines that there is no risk (S305). On the other hand, when the risk data exceeds the set value (S303; Yes), the risk management system 10 determines that there is a risk (S307). At this time, the risk management system 10 generates warning information (S309). The risk management system 10 also selects a warning destination to which the warning information is to be sent. The warning information includes information notifying the risk of future work as well as the risk data. The risk management system 10 sends the warning information to the user terminal 30 (S311), and the user terminal 30 receives the warning information (S313). In this example, the risk management system 10 sends the warning information to the user terminal 30 of the user at site A. By the user terminal 30 receiving the warning information, the user can be aware of the risks of current or future work.

[0079] By using this embodiment, information obtained from various lower-level systems is converted into normalized data, allowing various types of data to be evaluated using the same criteria. Furthermore, by linking and integrating the normalized data, it can be judged as a single piece of data, making it easier to make comprehensive risk judgments. Therefore, by using this embodiment, risk management can be performed easily and accurately using information obtained from multiple types of devices.

[0080] Furthermore, the risk management system of this embodiment can learn risk information using multiple types of data obtained by the lower-layer system, and can further make future predictions based on that risk information.

[0081] Second Embodiment In this embodiment, a risk management control process different from that in the first embodiment will be described. Specifically, an example in which normalized data is generated using a nonlinear function will be described.

[0082] 17 is a flow diagram of the normalization process S201A in this embodiment. First, a normalization reference value (S2011A) and a normalization maximum value (S2013A) associated with each sensing data are set. In this example, the normalization reference value and the normalization minimum value may be the same. The normalization reference value and the normalization maximum value can be set arbitrarily.

[0083] Next, normalized data is generated using a normalization function (S2015A). In this embodiment, a nonlinear function is used as the normalization function. Specifically, the nonlinear function may be a quadratic function as shown in Equation 2, or an exponential function as shown in Equation 3. "Normalvalue" indicates normalized data. "Value" indicates sensing data. In Equation 2 and Equation 3, a, b, and c may be set arbitrarily.

[0084] As described above, normalized data can be generated using a nonlinear function for the acquired sensing data. In other words, by using this embodiment, appropriate normalized data can be generated even when the sensing data and the normalized data have a nonlinear relationship.

[0085] Third Embodiment In this embodiment, a risk management control process different from that in the first embodiment will be described. Specifically, an example in which normalized data is generated based on a data table will be described.

[0086] Figure 18 is a flow diagram of the normalized data generation process S201B. Figure 19 is a data table 700 showing the relationship between heat index data and normalized data. The data table 700 includes heat index data 710, warning level data 720, and normalized data 730. In this example, warning level data is set independent of the heat index data. As shown in Figure 19, the heat index data is set in the range of 0°C to 39°C.

[0087] 18, first, a normalization reference value is set (S2011B). In this example, when the heat index data is "0°C", the warning level data and normalized data are "0". In this embodiment, the normalization reference value and the normalized minimum value may be set to be the same.

[0088] Next, a normalized maximum value is set (S2013B). In this example, when the heat index data is "39°C", the maximum value of the warning level data is "5". In this case, the normalized maximum value is "1". The normalized data is set as one-fifth of the warning level data.

[0089] Normalized data for a heat index range of 1°C to 38°C is generated as follows: When the heat index is 24°C or less, the warning level data is set to 0 because there is little impact on the human body. Therefore, the corresponding normalized data is also set to 0.

[0090] When the heat index is in the range of 25°C or higher and 29°C or lower, the warning level data increases in increments of 0.1 from 0.1 to 0.5. At this time, the normalized data is set in increments of 0.02 from 0.02 to 0.1. When the heat index is in the range of 30°C or higher and 38°C or lower, the warning level data is set to increase in magnitude due to the strong impact on the human body. Accordingly, the normalized data is also set to increase in magnitude.

[0091] In this embodiment, normalized data can be easily generated for input heat index data by referring to a data table. As a result, even if the sensing data and the normalized data are not in a linear relationship, appropriate normalized data can be generated by referring to the data table.

[0092] In this embodiment, normalized data is generated via intermediate values ​​called warning level data. By using the warning level data, risk information already acquired by the user can also be used as an index for risk assessment. Therefore, the meaning of the numerical values ​​of the normalized data can be clarified.

[0093] Fourth Embodiment In this embodiment, a risk management control process different from that in the first embodiment will be described. Specifically, an example of weighting normalized data will be described. Note that portions that overlap with the first embodiment of the present invention will be omitted as appropriate.

[0094] FIG. 20 is a flow diagram of the normalized data generation process S201C. In this embodiment, after generating the normalized data (S2017), a weight value may be assigned to the normalized data (S2019). In this example, a weight value is assigned to the first normalized data generated from the heat index data. When the weight value is assigned, the risk data may be adjusted based on the weighted first normalized data. This allows the first normalized data to be more important than the second normalized data generated from the cumulative workload data and the third normalized data generated from the cumulative stay time data. As a result, risk management in the risk management system can be improved.

[0095] In this embodiment, weights are assigned after generating the normalized data, but the present invention is not limited to this. For example, weighting may be performed after concatenation and integration (concatenated and integrated normalized data).

[0096] Fifth Embodiment In this embodiment, a risk management control process different from that of the first embodiment will be described. Specifically, an example in which a setting value is changed when a user terminal receives warning information in a third risk management control process will be described. Note that portions that overlap with the first embodiment of the present invention will be omitted as appropriate.

[0097] FIG. 21 is a flow diagram of the third risk management control process S300D. When the user terminal 30 receives warning information (S313), if the user determines that there is no risk in the current situation, the user can input response data into the user terminal 30. For example, when warning information is displayed on the display unit 31 of the user terminal 30, a question such as "Is this warning information appropriate for you?" may be displayed along with push buttons for answering "Yes" and "No." When the user presses the response button, the response data is input (S315). At this time, the user terminal 30 determines whether the setting value for determining whether to generate warning information needs to be modified. When the user presses "No," it is determined that there is no need to modify the setting value (S317; No), and the third control process S300D ends.

[0098] When the user presses "Yes", the user terminal 30 determines that the setting value needs to be corrected (S317; Yes), and generates and transmits setting value correction instruction information (S319, S321).

[0099] When the risk management system 10 receives the set value correction instruction information (S323), it performs the set value correction process (S325). This completes the third risk management control process S300D. Note that after the set value is corrected, the risk data analysis process may be performed again.

[0100] In this embodiment, even if a risk is determined to exist and warning information is sent, feedback can be given in accordance with the actual risk, thereby enabling more accurate risk management.

[0101] (Modifications) Within the scope of the concept of the present invention, a person skilled in the art may conceive of various modifications and alterations, and it is understood that these modifications and alterations also fall within the scope of the present invention. For example, to the above-described embodiments, a person skilled in the art may appropriately add, delete, or modify components, or add, omit, or change conditions of processing, and these modifications are also included within the scope of the present invention as long as they maintain the gist of the present invention.

[0102] In the first embodiment of the present invention, an example is shown in which the first lower layer system 20-1, the second lower layer system 20-2, and the third lower layer system 20-3 are provided as the lower layer systems 20, but the present invention is not limited to this. Multiple pieces of sensing data may be generated from one lower layer system.

[0103] Furthermore, although the first embodiment of the present invention illustrates an example in which three pieces of sensing data are used, the present invention is not limited to this. For example, two pieces of sensing data may be used, or four or more pieces of sensing data may be used. In other words, in the present invention, it is sufficient that at least two or more pieces of sensing data (first sensing data and second sensing data) are used.

[0104] Furthermore, in the first embodiment of the present invention, when setting the normalization reference value, the normalization reference value may be set based on user information such as the user's age, gender, years of service, and environmental information (or geographical information). Furthermore, when normalizing data using a linear function, a predetermined coefficient or constant may be set as appropriate. This allows the normalization data to be adjusted for each user. Therefore, risk management tailored to each user can be performed, such as by setting the normalization data to be high for sensing data that is determined to pose a high risk for each individual.

[0105] Furthermore, in the first embodiment of the present invention, an example was shown in which warning information was generated when risk data satisfied a predetermined condition, but the present invention is not limited to this. For example, warning information may be generated when normalized data generated from each piece of sensing data is out of the range from the normalized minimum value "-1" to the normalized maximum value "1" and is therefore different from normal. This allows warning information to be generated more quickly, enabling more accurate risk management.

[0106] In the first embodiment of the present invention, as shown in Figures 4 to 7, numerical data is used for the first normalized data, second normalized data, and third normalized data. However, the present invention is not limited to this. For example, alphabets or other characters other than numerical data may be used. Specifically, when the first normalized data is "A," the second normalized data is "A," and the third normalized data is "B," the integrated normalized data becomes "AAB." In this case, risk data may be generated according to the number of "As." When there are many "As," the risk data may be determined to be "high," and warning information may be sent.

[0107] In the first embodiment of the present invention, the first normalized data, the second normalized data, and the third normalized data are concatenated and integrated in this order, but the present invention is not limited to this. For example, the concatenation order may be set according to the priority of the normalized data.

[0108] Furthermore, in the first embodiment of the present invention, heat index data, cumulative workload data, and cumulative stay time data are used as sensing data, but the present invention is not limited to this. For example, worker body temperature data, hydration amount data, pulse data, heart rate data, breathing data, body tilt data, time data, age data, and female data may also be used. By acquiring various types of sensing data from more lower-level systems, the risk management system 10 can expand machine learning models and generate more accurate risk data. Therefore, highly accurate risk management can be performed.

[0109] In the first embodiment of the present invention, an example is shown in which the integrated normalized data is generated by concatenating and integrating the first normalized data, the second normalized data, and the third normalized data, but the present invention is not limited to this. For example, the integrated normalized data may be generated by performing a predetermined arithmetic process, such as addition or multiplication, on the first normalized data, the second normalized data, and the third normalized data.

[0110] In the first embodiment of the present invention, an example is shown in which training data is generated using one week's worth of sensing data acquired in advance to generate a machine learning model, but the present invention is not limited to this. For example, the sensing data acquired in advance may be one year's worth.

[0111] Furthermore, new training data may be generated every predetermined period (e.g., one week). This updates the machine learning model, enabling more accurate risk data to be generated. This improves the accuracy of risk management.

[0112] 1... Risk management collaboration system, 10... Risk management system (upper layer system), 11... Control unit, 12... Memory unit, 13... Communication unit, 14... Display unit, 20... Lower layer system, 21... Control unit, 22... Memory unit, 23... Communication unit, 24... Display unit, 30... User terminal, 31... Display unit, 32... Control unit, 33... Memory unit, 34... Operation unit, 35... Communication unit, 40... Network, 100... Risk management control unit, 110... Acquisition unit, 120... Normalization processing unit, 130... Learning processing unit, 140... Prediction processing unit, 150... Analysis unit, 160... Warning information transmission unit, 210... Receiving unit, 220... Transmitting unit, 310... Receiving unit, 320... Transmitting unit

Claims

1. A risk management system which is an upper layer system connected to a plurality of lower layer systems and manages predicted risks, the risk management system comprising: An acquisition unit that acquires first sensing data acquired by a first lower layer system capable of acquiring first type of data, and second sensing data acquired by a second lower layer system capable of acquiring a second type of data different from the first sensing data; A normalization processing unit that performs arithmetic processing for generating first normalized data based on the first sensing data, performs arithmetic processing for generating second normalized data based on the second sensing data, and generates integrated normalized data by performing arithmetic processing for making the first normalized data and the second normalized data into one integrated data; A prediction processing unit that generates risk data predicted based on the integrated normalized data; wherein the acquisition unit acquires user information; The arithmetic processing for generating the first normalized data when generating the predicted risk data is defined by a first normalization arithmetic expression using a first normalization reference value and a first normalization method; The arithmetic processing for generating the second normalized data when generating the predicted risk data is defined by a second normalization arithmetic expression using a second normalization reference value and a second normalization method; The first normalization arithmetic expression is set based on the user information; The second normalization arithmetic expression is set based on the user information. A risk management system.

2. The risk management system according to claim 1, further comprising a learning processing unit that generates an integrated normalized machine learning model by machine learning on the second integrated normalized data generated by the normalization processing unit using the first sensing data and the second sensing data acquired in advance by the acquisition unit; The prediction processing unit generates the risk data by applying the integrated normalized data to the integrated normalized machine learning model.

3. The risk management system according to claim 1 or 2, wherein each of the first normalization reference value and the second normalization reference value is set based on the user information.

4. The first normalization method is set based on first identification information associated with the first sensing data, and the second normalization method is set based on second identification information associated with the second sensing data. The risk management system according to claim 3.

5. The normalization processing unit assigns a weight value to at least one of the first normalized data, the second normalized data, and the integrated normalized data based on a predetermined condition, and the prediction processing unit adjusts risk data based on the weight value. The risk management system according to any one of claims 1 to 4.

6. The predetermined condition is set based on the user information. The risk management system according to claim 5.

7. The risk management system according to any one of claims 1 to 6, further including a warning information transmission unit that transmits warning information to a user terminal according to the risk data.

8. An upper-layer system connected to a plurality of lower-layer systems, the risk management system for managing predicted risks, obtains first sensing data obtained by a first lower-layer system capable of obtaining first-type data, and second sensing data obtained by a second lower-layer system capable of obtaining second-type data different from the first sensing data, generates first normalized data based on the first sensing data, generates second normalized data based on the second sensing data, and performs normalization processing for generating integrated normalized data by performing arithmetic processing for making the first normalized data and the second normalized data into one aggregated data, generates risk data predicted based on the integrated normalized data, includes obtaining user information when obtaining the first sensing data and the second sensing data, the arithmetic processing for generating the first normalized data when generating the predicted risk data is defined by a first normalization arithmetic expression using a first normalization reference value and a first normalization method, the arithmetic processing for generating the second normalized data when generating the predicted risk data is defined by a second normalization arithmetic expression using a second normalization reference value and a second normalization method, the first normalization arithmetic expression is set based on the user information, the second normalization arithmetic expression is set based on the user information, risk management method.

9. Further includes generating an integrated normalized machine learning model by machine learning for the second integrated normalized data generated by performing the normalization processing using the previously obtained first sensing data and the second sensing data, generating the predicted risk data includes generating the risk data by applying the integrated normalized data to the integrated normalized machine learning model, the risk management method according to claim 8.

10. Each of the first normalization reference value and the second normalization reference value is set based on the user information, the risk management method according to claim 8 or 9.

11. The first normalization method is set based on first identification information associated with the first sensing data, and the second normalization method is set based on second identification information associated with the second sensing data. The risk management method according to claim 10.

12. Generating the risk data includes assigning a weight value to at least one of the first normalized data, the second normalized data, and the integrated normalized data based on a predetermined condition, and adjusting the generated risk data based on the weight value. The risk management method according to any one of claims 8 to 11.

13. The predetermined condition is set based on the user information. The risk management method according to claim 12.

14. Transmitting warning information to a user terminal according to the risk data. The risk management method according to any one of claims 8 to 13.

Images