Work support device, work support method, and work support program

Abstract

This allows us to evaluate the degree to which each action performed by a worker during their work is dangerous for that worker. [Solution] The work support device (100) comprises a processor (110) and a memory (120). The processor (110) works in cooperation with the memory (120) to acquire a work plan (121) and worker information (122). Based on the work plan (121) and worker information (122), it performs an action evaluation of the target worker's planned work actions and uses the action evaluation to evaluate the degree of risk to the target worker's planned work actions.

Description

Technical Field

[0001] The present disclosure relates to a work support system, a work support method, and a work support program.

Background Art

[0002] Patent Document 1 describes a danger warning device including a monitoring unit that acquires information about current workers, a risk management unit that determines conditions under which accidents of the same type are likely to occur based on past accidents and determines whether the acquired information meets the determined conditions, and a warning unit that warns workers of the occurrence of accidents of the same type when the acquired information meets the determined conditions. At this time, in the display of management information, management information in a plurality of stores within a peripheral area at a predetermined distance from the location of a specific store is displayed so as to be comparable between stores. Here, the risk management unit creates a probability distribution of the occurrence of the accident type for each individual condition (location, time, biometric information, ···).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, accidents and dangerous situations depend not only on location and time but also on the actions of the work scheduled for a specific target worker. That is, the probability that a worker will encounter an accident or danger varies greatly depending on what actions each worker takes during work.

[0005] An object of the present disclosure is to provide a technology capable of evaluating the degree of danger of each work for a target worker among the actions during the work performed by the worker.

Means for Solving the Problems

[0006] A work support device according to one aspect of the present disclosure comprises a processor and a memory, wherein the processor cooperates with the memory to acquire a work plan and worker information, performs an action evaluation of the target worker's scheduled work actions from the work plan and the worker information, and uses the action evaluation to evaluate the degree of risk of the target worker's scheduled work actions.

[0007] A work support method according to one aspect of the present disclosure involves instructing a computer to acquire a work plan and worker information, to perform an action evaluation of the target worker's planned work actions based on the work plan and worker information, and to use the action evaluation to evaluate the degree of risk of the target worker's planned work actions.

[0008] Furthermore, a work support device according to one aspect of the present disclosure comprises a processor and memory, wherein the processor cooperates with the memory to acquire real-time work video and worker information of the target worker included in the work video, divides the work video into units of work actions, performs an action evaluation of the target worker's actions based on the work video and the worker information, and uses the action evaluation to evaluate the degree of risk to the target worker's actions.

[0009] A work support program according to one aspect of this disclosure instructs a computer to acquire a work plan and worker information, to perform an action evaluation of the target worker's planned work actions based on the work plan and worker information, and to use the action evaluation to evaluate the degree of risk of the target worker's planned work actions.

[0010] These comprehensive or specific embodiments may be implemented as systems, devices, methods, integrated circuits, computer programs, or recording media, or as any combination of systems, devices, methods, integrated circuits, computer programs, and recording media. [Effects of the Invention]

[0011] According to this disclosure, it is possible to evaluate how dangerous each action performed by a worker during their work is to that worker. [Brief explanation of the drawing]

[0012] [Figure 1] A diagram showing an example of the configuration of the work support system according to Embodiment 1. [Figure 2A] Block diagram showing an example of the configuration of the operation creation unit according to Embodiment 1. [Figure 2B] This figure shows an example of a planned work operation table used by the operation creation unit according to Embodiment 1. [Figure 3A] Block diagram showing an example of the configuration of the operation evaluation unit according to Embodiment 1. [Figure 3B] This figure shows an example of a worker information table used by the operation evaluation unit according to Embodiment 1. [Figure 3C] This figure shows an example of a proficiency information table used by the operation evaluation unit according to Embodiment 1. [Figure 4A] This figure shows an example of updating proficiency information when a new worker is present according to Embodiment 1. [Figure 4B] This figure shows an example of a worker information table when a new worker exists according to Embodiment 1. [Figure 4C] This figure shows an example of a proficiency information table when there are new workers according to Embodiment 1. [Figure 5] A diagram showing an example of the configuration of the risk assessment unit according to Embodiment 1. [Figure 6A] This figure shows an example of reinforcement learning for the generation coefficient calculation model according to Embodiment 1. [Figure 6B] A graph showing an example of the probability density function used by the generation coefficient calculation model according to Embodiment 1. [Figure 6C] This figure shows an example of an accident probability table used by the accident coefficient calculation model according to Embodiment 1. [Figure 7A] A diagram showing an example of reinforcement learning for a damage calculation model according to Embodiment 1. [Figure 7B]Figure showing an example of the damage data table used by the damage calculation model according to Embodiment 1 [Figure 8A] Block diagram showing an example of the configuration of the risk simulation video generation unit according to Embodiment 1 [Figure 8B] Figure showing an example of the sample video table used by the risk simulation video generation unit according to Embodiment 1 [Figure 8C] Figure showing an example of the worker information (for video creation) table used by the risk simulation video generation unit according to Embodiment 1 [Figure 9] Flowchart showing an example of the work support method according to Embodiment 1 [Figure 10] Figure showing an example of the configuration of the work support system according to Embodiment 2 ]> [Figure 11A] Block diagram showing an example of the configuration of the motion creation unit according to Embodiment 2 [Figure 11B] Figure for explaining the reinforcement learning of motion division in the motion creation unit according to Embodiment 2 [Figure 12A] Block diagram showing an example of the configuration of the motion creation unit when using the motion definition table according to Embodiment 2 [Figure 12B] Figure showing an example of the motion definition table used by the motion creation unit according to Embodiment 2 [Figure 13A] Figure showing an example of the configuration when the continuous motion prediction unit according to Embodiment 2 learns using past work motion data [Figure 13B] Figure showing an example of the configuration when the evaluation target motion of the continuous motion prediction unit according to Embodiment 2 is determined [Figure 14] Flowchart showing an example of the work support method according to Embodiment 2 [Figure 15] Flowchart showing an example of the work support method when performing continuous motion prediction according to Embodiment 2

Mode for Carrying Out the Invention

[0013] Embodiments of the present disclosure will be described in detail below, with appropriate reference to the drawings. However, descriptions that are unnecessarily detailed may be omitted. For example, detailed descriptions of already well-known matters and redundant descriptions of substantially identical configurations may be omitted. This is to avoid the following description becoming unnecessarily verbose and to facilitate understanding for those skilled in the art. The accompanying drawings and the following description are provided to enable those skilled in the art to fully understand the present disclosure and are not intended to limit the subject matter of the claims. The functions of one configuration shown in this embodiment may be realized by two or more physical configurations, or the functions of two or more configurations may be realized by, for example, one physical configuration.

[0014] (Embodiment 1) First, the work support system 1 will be described. Figure 1 is a diagram showing an example of the configuration of the work support system according to Embodiment 1. Figure 2A is a block diagram showing an example of the configuration of the motion creation unit according to Embodiment 1. Figure 2B is a diagram showing an example of a planned work motion table used by the motion creation unit according to Embodiment 1. Figure 3A is a block diagram showing an example of the configuration of the motion evaluation unit according to Embodiment 1. Figure 3B is a diagram showing an example of a worker information table used by the motion evaluation unit according to Embodiment 1. Figure 3C is a diagram showing an example of a proficiency information table used by the motion evaluation unit according to Embodiment 1. Figure 4A is a diagram showing an example of updating proficiency information when a new worker is present according to Embodiment 1. Figure 4B is a diagram showing an example of a worker information table when a new worker is present according to Embodiment 1. Figure 4C is a diagram showing an example of a proficiency information table when a new worker is present according to Embodiment 1. Figure 5 is a diagram showing an example of the configuration of the risk assessment unit according to Embodiment 1. Figure 6A is a diagram showing an example of reinforcement learning of the occurrence coefficient calculation model according to Embodiment 1. Figure 6B is a graph showing an example of a probability density function used by the occurrence coefficient calculation model according to Embodiment 1. Figure 6C shows an example of an accident occurrence probability table used by the occurrence coefficient calculation model according to Embodiment 1. Figure 7A shows an example of reinforcement learning for the damage calculation model according to Embodiment 1. Figure 7B shows an example of a damage data table used by the damage calculation model according to Embodiment 1. Figure 8A is a block diagram showing an example of the configuration of the hazard simulation video generation unit according to Embodiment 1. Figure 8B shows an example of a sample video table used by the hazard simulation video generation unit according to Embodiment 1. Figure 8C shows an example of a worker information (for video creation) table used by the hazard simulation video generation unit according to Embodiment 1.

[0015] The work support system 1 includes a work support device 100, a plan generation device 200, a forecasting system 300, multiple worker terminals 400, and multiple store / head office terminals 500. The work support device 100, the plan generation device 200, the forecasting system 300, the worker terminals 400, and the store / head office terminals 500 can send and receive information and data to and from each other via wired cables or a communication network. Examples of wired cables include HDMI® cables and USB cables. Examples of communication networks include wired LAN, wireless LAN, LTE, 4G, 5G, the internet, and VPN (Virtual Private Network).

[0016] Work support system 1 is used, for example, in logistics companies and retail companies, in companies that have multiple locations such as warehouses, logistics centers, supermarkets, and stores. When dangers or accidents are predicted for a created work plan, it notifies workers and organizational managers via video or alerts through worker terminals 400 and store / headquarters terminals 500. In work support system 1, the work support device 100 outputs video or alerts when dangers or accidents are predicted for each worker's work plan generated by the plan generation device 200 using information such as workload prediction data generated by the prediction system 300.

[0017] The work support device 100, the plan generation device 200, the prediction system 300, the worker terminal 400, and the store / head office terminal 500 are electronic devices such as personal computers, servers, smartphones, and tablets, respectively. "Workers" include full-time employees, contract employees, part-time workers, and temporary workers, regardless of their employment contract.

[0018] The work support device 100 comprises a processor 110, a memory 120, and a display unit 130. The processor 110 comprises an action creation unit 111, an action evaluation unit 112, a risk assessment unit 113, and a risk simulation video generation unit 114. The memory 120 stores work plan information 121, worker information 122, and sample videos 123.

[0019] The processor 110 implements various functions by executing programs stored in the memory 120. The processor 110 calculates various correlation values ​​and outputs corresponding information, although details will be described later. Examples of the processor 110 include a CPU (Central Processing Unit), MPU (Micro Processing Unit), controller, LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), PLD (Programmable Logic Device), or FPGA (Field-Programmable Gate Array). The plan generation device 200, forecasting system 300, worker terminal 400, and store / head office terminal 500 should also be equipped with similar processors.

[0020] The motion creation unit 111 will be explained with reference to Figure 2. Based on the work plan information 121, the motion creation unit 111 creates (determines) the motions to be evaluated for risk assessment. By inputting the planned work motion TBL (Table) into the motion determination model, the motions to be evaluated (walking, lifting, etc.) are determined. For example, if there is a task of carrying luggage, the motions can be divided into actions such as lifting the luggage, walking, and putting the luggage down. In other words, the motions to be evaluated are determined by retrieving the corresponding motion information from the planned work motion TBL in Figure 2B for the target worker shown in the work plan information 121. The planned work motion TBL lists the contents of the actions (walking, lifting, etc.) associated with each work ID (carrying things, etc.).

[0021] Next, the motion evaluation unit 112 will be explained with reference to Figures 3A, 3B, 3C, 4A, 4B, and 4C. The motion evaluation unit 112 performs a motion evaluation of the target worker's planned work movements based on the work plan information 121 and worker information 122 acquired by the processor. That is, the motion evaluation unit 112 calculates ability values ​​such as proficiency, experience, qualifications, and physical abilities for each worker when performing each work movement. When information on the movement to be evaluated and worker information 122 are input to the motion evaluation unit 112, ability values ​​such as the proficiency of the movement to be evaluated for the target worker are output. In other words, it is preferable to perform motion evaluation based on the target worker's proficiency for the planned work movements. As shown in Figure 3B, worker information 122 lists each worker's qualification information, etc. The proficiency information TBL, which is proficiency information 125 in Figure 3C, includes the skill level for each movement of each worker, and is created through a proficiency evaluation model by referring to past motion information, etc., and is updated periodically. In other words, the motion evaluation unit 112 refers to the proficiency information TBL for the target work action performed by the target worker and outputs a motion evaluation, such as proficiency, via the motion evaluation judgment model. Actions that are clearly of low risk should be excluded from the processes after motion evaluation.

[0022] In the case of new employees or other workers for whom sufficient past information is not available, it is not possible to generate proficiency information 125 or input motion data because there is no past work data for that individual. In other words, during periods when there is insufficient data to quantify the worker's proficiency and other abilities, the proficiency information 125 of a similar worker is used in the reference person determination model based on the worker information 122. The reference person determination model, for example, inputs the worker information 122 (new worker information) of a new worker into a classification tree model, refers to the proficiency information 125 of a worker with a high similarity to the existing worker information TBL, and creates (copies or uses) the new worker's proficiency information. The worker information TBL with the new worker information added is shown in Figure 4B, and the updated proficiency information TBL using it is shown in Figure 4C. The proficiency information TBL should be updated in the proficiency evaluation model by learning the worker information and corresponding motion data using supervised classification tree learning or similar methods.

[0023] Next, the risk assessment unit 113 will be explained. The risk assessment unit 113 uses motion evaluation to assess the risk of the planned actions of the target worker. That is, for each work action, the risk is assessed from the perspective of the probability of occurrence and the extent of damage. The risk is assessed by calculating the occurrence coefficient for each type of risk and the amount of damage that would result if the risk occurs. For example, the risk r can be calculated as follows. r = oc*d r: risk factor, risk coefficient (evaluation result) oc: occurrence coef, occurrence coefficient (the probability that the accident will occur during the relevant work) d:damage, damage (yen) (amount of damage in the event of the accident)

[0024] The occurrence coefficient is calculated using a regression-based model, which combines data from the accident probability TBL (number of accidents) in Figure 6C, created from past cases, with proficiency information 125. The probability density function in Figure 6B is recommended.

[0025] Furthermore, damages are calculated based on the amount of loss incurred at the time of the accident. Damages are calculated by adding up the value of the damaged goods, medical expenses for injuries, and labor costs during the recovery period. This total amount is then used as damage data and calculated using a damage calculation model that employs reinforcement learning.

[0026] If the risk coefficient (degree of danger) exceeds a threshold, alerts or videos can be output via the display unit 130, worker terminal 400, or store / head office terminal 500 to inform workers, supervisors, and other relevant parties of the potential danger or accident. It is also advisable to evaluate the degree of danger based on the calculation results of the probability of an accident occurring or the amount of damage.

[0027] Next, the hazard simulation video generation unit 114 will be described. If the hazard coefficient (hazard level) evaluated by the hazard assessment unit 113 is above a threshold, the hazard simulation video generation unit 114 creates an instruction document to be input into the video generation model based on the hazard content of the hazard assessment result. That is, the processor 110 further acquires a sample video, and if the hazard level evaluation is above a threshold, it inputs at least a portion of the worker information and the sample video into the video generation model to generate and output a hazard simulation video. The instruction document may be implemented by creating a standard text in advance. The processor 110 creates and acquires an instruction document to be input into the video generation model corresponding to the hazard level evaluation, and inputs the instruction document into the video generation model. Then, it acquires an image or video (showing the face and body shape) of the target worker from the worker information 122, combines it with the instruction document, and generates a hazard simulation video by superimposing the face onto the sample video 123, etc. That is, the hazard simulation video may be generated by combining the sample video with the face information 1226 of the target worker included in the worker information 122. The facial information 1226 may include photos or videos. Sample videos 123 may be accident videos or dangerous videos, and are listed by task or situation in a sample video TBL as shown in Figure 8B. Images or videos of the target workers are listed and managed in a worker information (for video creation) TBL as shown in Figure 8C. The generated hazard simulation video may be displayed on one of the following: the display unit 130, the worker terminal 400, or the store / head office terminal 500. The generated video may be an accident simulation video. By seeing a dangerous video with their own face superimposed in advance, workers can realistically perceive the level of danger and be more careful about danger and accidents.

[0028] The work plan information 121 contained in memory 120 is, for example, the work plan for each worker for the target date. It is preferable that it also includes the details of the work assigned to each worker for each time slot.

[0029] Worker information 122 may include any of the following: length of service information 1221, age information 1222, qualification information 1223, or past accident count information 1224. Worker information 122 may also include physical ability information indicating the worker's physical capabilities. Length of service information 1221 includes the employee's length of service. Age information 1222 includes the age of each worker. Qualification information 1223 includes work-related qualifications held by the worker. Past accident count information 1224 includes the number of accidents each worker has caused in the past. It may also include proficiency information 1225 and facial information 1226. Length of service information 1221, age information 1222, qualification information 1223, and past accident count information 1224 may be composed of proficiency information 1225 alone or in combination, or may include other information. Facial information 1226 may include photographs or videos.

[0030] Sample video 123 shows scenes of accidents and dangerous situations.

[0031] The display unit 130 may include a liquid crystal display device, an organic EL device, or other display device. The plan generation device 200, the forecasting system 300, the worker terminal 400, and the store / head office terminal 500 may also be equipped with similar display units.

[0032] The plan generation device 200 generates a work plan using the workload prediction data from the prediction system 300 and / or information such as work performance.

[0033] The forecasting system 300 generates information such as workload forecasting data. The worker terminal 400 is a terminal used by each worker among the employees of each store. The store / headquarters terminal 500 is a terminal installed in each store and at headquarters. For example, if a company in the logistics industry uses this work support system 1, at least one of the work support device 100, the plan generation device 200, the forecasting system 300, and the store / headquarters terminal 500 may be installed at headquarters or some of the branches or warehouses, and the worker terminal 400 is used by the target workers among the employees of each store or warehouse.

[0034] Next, the flow of the planning analysis method will be described. Figure 9 is a flowchart showing an example of a work support method according to Embodiment 1. The order of the steps in the flow may be changed unless there are constraints. The following flow may be executed by the computer running the program.

[0035] First, the processor 110 obtains work plan information 121 and worker information 122 from memory 120 or an external source (ST1). Then, based on the work plan information 121, the processor 110 creates (determines) the actions to be evaluated for risk assessment (ST2). By inputting the planned work action TBL into the action determination model, the actions to be evaluated (walking, lifting, etc.) are determined. Then, the processor 110 performs an action evaluation for the planned work actions of the target worker (ST3). The action evaluation may include an assessment of proficiency. Then, the processor 110 evaluates the risk level for the planned work actions of the target worker (ST4). If the risk level is above the threshold (ST5:YES), the processor 110 generates a risk simulation video (ST6). If the risk level is below the threshold (ST5:NO), the processor 110 returns to step ST1.

[0036] (Embodiment 2) In Embodiment 2, the work support device 100 also includes a processor 110 and a memory 120, and the processor 110 works in cooperation with the memory 120 to perform the following processing: The processor 110 acquires real-time work video captured by the camera 600 and worker information 122 of the target worker included in the work video, divides the work video into units of work movements using the motion analysis unit 115, performs a motion evaluation of the target worker's movements based on the work video and worker information, and uses the motion evaluation to evaluate the degree of risk of the target worker's movements. The processor 110 also predicts the continuation of the target's movements, evaluates the degree of risk of the continuation, and outputs an alert if it exceeds a threshold.

[0037] Next, Embodiment 2 will be described with reference to Figures 10 to 15. In Embodiment 2, real-time video is analyzed to evaluate the degree of risk of the operation. Note that elements common to Embodiment 1 will not be explained. Figure 10 is a diagram showing an example of the configuration of the work support system according to Embodiment 2. Figure 11A is a block diagram showing an example of the configuration of the motion creation unit according to Embodiment 2. Figure 11B is a diagram illustrating reinforcement learning for motion division in the motion creation unit according to Embodiment 2. Figure 12A is a block diagram showing an example of the configuration of the motion creation unit when using the motion definition table according to Embodiment 2. Figure 12B is a diagram showing an example of the motion definition table used by the motion creation unit according to Embodiment 2. Figure 13A is a diagram showing an example of the configuration when the continuous motion prediction unit according to Embodiment 2 learns from past work operation data. Figure 13B is a diagram showing an example of the configuration when the operation to be evaluated by the continuous motion prediction unit according to Embodiment 2 has been determined.

[0038] As shown in Figure 10, in Embodiment 2, the work support system 1 is further equipped with a camera 600, and the processor 110 is newly equipped with a motion analysis unit 115 and a continuous motion prediction unit 116.

[0039] Camera 600 films the workers performing their tasks. In Embodiment 2, the motion creation unit 111 creates (determines) the actions to be evaluated from the motion data of the video captured by camera 600. The motion analysis unit 115 extracts motion data from the captured video. Since the video is filmed continuously, the continuous motion data is divided into individual motion datasets using the motion division model in Figure 11A. Each action is converted into the corresponding work action name using the motion determination model. Motion analysis can be performed by dividing existing video into motion units. In Embodiment 2, the motion creation unit 111 may also create (determine) the actions to be evaluated using the work plan information 121.

[0040] The continuous action prediction unit 116 predicts the future work actions of the worker currently being filmed. In other words, when the camera 600 monitors the work in real time, the continuous action prediction unit 116 predicts accidents and hazards from work actions that may occur in the near future if the current work actions continue, in order to prevent danger, and assesses the hazards. Multiple predicted continuous work actions are transmitted to the hazard assessment unit 113.

[0041] As shown in Figure 11B, the motion splitting model automatically divides continuous motion data into individual work actions. At this time, the presence of separate training motion data for each action allows for reinforcement learning using supervised classification. Training is performed using pre-prepared motion data for each action, enabling the model to identify each data point individually. The motion splitting can also be enhanced by targeting randomly combined motion data. The output of the motion splitting model is to divide a single continuous motion dataset into multiple action motion data parts, which are then output sequentially.

[0042] Alternatively, the system may be trained using pre-prepared motion data categorized by action, enabling it to identify each action individually. For training, the system may be trained on randomly combined motion data, allowing it to simultaneously differentiate between actions and determine which action is being performed.

[0043] Alternatively, as shown in Figure 12A, a motion decision model can be used that performs supervised classification learning, using motion data for each action and a motion definition table as training data. Existing motion data for each action is associated with the corresponding motion definition. As shown in the motion definition TBL in Figure 12B, the "walking" motion dataset is associated with the motion definition TBL "A0001", classification learning is performed on the motion data part, and motion definitions for the corresponding classifications are added.

[0044] As shown in Figure 13A, the continuous action prediction unit 116 acquires past work action data from the action determination model of the action creation unit 111 shown in Figures 11A and 11B, inputs this data into the continuous action prediction model to determine the predicted continuous work action, which is the next task, and inputs the predicted action into the risk assessment unit 113. Alternatively, as shown in Figure 13B, the action to be evaluated output by the action creation unit 111 may be input into the continuous action prediction model to determine the predicted continuous work action, which is the next task, and input the predicted action into the risk assessment unit 113.

[0045] Camera 600 can be a standard surveillance camera or similar.

[0046] Next, the flow of the planning analysis method will be described. Figure 14 is a flowchart showing an example of a work support method according to Embodiment 2. Figure 15 is a flowchart showing an example of a work support method when predicting continuous operation according to Embodiment 2. Note that the order of the flows may be changed unless there are constraints. The following flows may be executed by the computer running the program.

[0047] First, the processor 110 acquires real-time work video footage and worker information 122 captured by the camera 600 from memory 120 or an external source (ST11). Then, the processor 110 divides the real-time work video footage into units of work movements by performing motion analysis on the video (ST12). Then, based on the results of the motion analysis, the processor 110 creates (determines) the movements to be evaluated for risk assessment (ST13). Next, the processor 110 performs a motion evaluation of the planned work movements of the target worker (ST14). The motion evaluation may include an assessment of proficiency, etc. Then, the processor 110 evaluates the risk level of the planned work movements of the target worker (ST15). If the risk level is above a threshold (ST16: YES), the processor 110 issues an alert (ST17). An alert includes displaying on at least one of the display unit 130, the worker terminal 400, and the store / head office terminal 500 that the level of danger is above a threshold or that the likelihood of danger or accidents is high. Also, if the level of danger is below the threshold (ST16: NO), the processor 110 returns to step ST11.

[0048] Furthermore, the flow for predicting continuous actions will be explained with reference to Figure 15. First, the processor 110 acquires real-time work video and worker information 122 captured by the camera 600 from memory 120 or from an external source (ST21). Then, the processor 110 divides the work video into units of work actions by performing motion analysis on the real-time work video (ST22). Then, the processor 110 creates (determines) the actions to be evaluated for risk assessment based on the results of the motion analysis (ST23). Then, the processor 110 performs a motion evaluation of the planned actions of the target worker (ST24). The motion evaluation may include an evaluation of proficiency, etc. Then, the processor 110 predicts continuous actions from the real-time work video (ST25). Then, the processor 110 evaluates the risk level of the predicted continuous actions of the target worker (ST26). If the risk level is above a threshold (ST27: YES), the processor 110 issues an alert (ST28). An alert includes displaying on at least one of the display unit 130, worker terminal 400, and store / head office terminal 500 that the level of danger is above a threshold or that the likelihood of danger or accidents is high. Furthermore, if the level of danger is below the threshold (ST27: NO), the processor 110 returns to step ST21. This allows for the output of alerts after predicting continued operation, thus enabling more proactive prevention of dangers and accidents.

[0049] (Summary of this disclosure) Based on the above description of embodiments, the following technologies are disclosed.

[0050] <Technology 1> A work support device (100) according to one embodiment comprises a processor (110) and a memory (120). The processor (110) cooperates with the memory (120) to acquire a work plan (121) and worker information (122). Based on the work plan (121) and the worker information (122), it performs an action evaluation of the planned work actions of the target worker and uses the action evaluation to evaluate the degree of risk of the planned work actions of the target worker. This allows us to evaluate how dangerous each action performed by a worker is for that particular worker.

[0051] <Technology 2> In the work support device (100) described in Technology 1, the processor (110) further acquires a sample video, and if the risk level is above a predetermined threshold, it inputs at least a portion of the worker information (122) and the sample video into a video generation model to generate and output a simulated risk video. This allows for an assessment of the degree of danger each task poses to a worker during their work. If the danger level exceeds a threshold, a simulated danger video can be generated, encouraging workers to be more careful about hazards and accidents.

[0052] <Technology 3> In the work support device (100) described in Technology 2, the processor (110) acquires an instruction document to be input to the video generation model corresponding to the risk assessment, and inputs the instruction document to the video generation model. This allows for the evaluation of the degree of danger each task poses to a worker during their work, and by effectively generating hazard simulation videos when the danger level exceeds a threshold, it can encourage workers to be more careful about hazards and accidents.

[0053] <Technology 4> In the work support device (100) described in any one of the technologies 1 to 3, the processor (110) performs the operation evaluation based on the worker's proficiency with the planned operations. This allows us to evaluate how dangerous each task is for a worker based on their level of proficiency in that particular task.

[0054] <Technology 5> In the work support device (100) described in any of the technologies 1 to 4, the processor (110) performs the risk assessment based on the calculation result of the probability of an accident occurring or the amount of damage. This allows for a more detailed assessment of the degree to which each action performed by a worker during their work is dangerous for that worker.

[0055] <Technology 6> In the work support device (100) described in any of the technologies 1 to 5, the worker information (122) includes at least one of the following: years of service (1221), age (1222), qualifications (1223), and number of past accidents (1224). This allows us to evaluate the degree to which each task is dangerous for a worker based on their detailed proficiency in that particular task.

[0056] <Technology 7> In the work support device (100) described in any of technologies 1 to 6, the hazard simulation video is generated by combining the sample video 123 with the face information 1226 of the target worker included in the worker information 122. This allows for an assessment of the degree of danger each task is to the worker, and if the danger level exceeds a threshold, it generates a simulated danger video that raises a greater sense of urgency, thereby encouraging workers to be more careful about dangers and accidents.

[0057] <Technology 8> A work support device (100) according to one embodiment comprises a processor (110) and a memory (120), wherein the processor (110) cooperates with the memory (120) to acquire real-time work video and worker information (122) of the target worker included in the work video, divide the work video into units of work movements, perform an action evaluation of the target worker's actions based on the work video and the worker information (122), and use the action evaluation to evaluate the degree of risk to the target worker's actions. This allows for the evaluation of the degree to which each action performed by a worker during their work is dangerous for that worker, using real-time video footage of the work in progress.

[0058] <Technology 9> In the work support device described in Technical 8, the processor (110) predicts the continuation of the operation of the target, evaluates the risk level of the continuation of the operation, and outputs an alert if the risk level is above a predetermined threshold. This allows for the evaluation of the degree of danger each action performed by a worker during work using real-time video footage, and enables alerts to be issued if the level of danger exceeds a threshold.

[0059] <Technology 10> In the work support device (100) described in Technical 8 or 9, the processor (110) divides the work video into units of work movements by motion analysis. This allows for the evaluation of the degree of danger each task poses to a worker by performing motion analysis, even when using real-time work video.

[0060] <Technology 11> A work support method according to one embodiment acquires a work plan (121) and worker information (122), performs an action evaluation of the target worker's planned work actions from the work plan (121) and the worker information (122), and uses the action evaluation to evaluate the degree of risk of the target worker's planned work actions. This allows us to evaluate how dangerous each action performed by a worker is for that particular worker.

[0061] <Technology 12> A work support program according to one embodiment acquires a work plan (121) and worker information (122), instructs the computer to perform an action evaluation of the target worker's planned work actions based on the work plan (121) and the worker information (122), and uses the action evaluation to evaluate the degree of risk of the target worker's planned work actions. This allows us to evaluate how dangerous each action performed by a worker is for that particular worker.

[0062] While embodiments have been described above with reference to the attached drawings, this disclosure is not limited to such examples. It is clear to those skilled in the art that various modifications, alterations, substitutions, additions, deletions, and equivalents can be conceived within the scope of the claims, and these are also understood to fall within the technical scope of this disclosure. Furthermore, the components of the embodiments described above can be combined in any way without departing from the spirit of the invention. [Industrial applicability]

[0063] The technology disclosed herein can evaluate the degree to which each action performed by an employee during work is dangerous to the employee in question. [Explanation of symbols]

[0064] 1. Work support system 100 Work support devices 110 processors 111 Action Creation Unit 112 Operation Evaluation Unit 113 Risk Assessment Department 114 Danger Simulation Video Generation Unit 115 Motion Analysis Unit 116 Continuous Operation Prediction Unit 120 memory 121 Work Plan Information 122 Worker Information 1221 Years of Service Information 1222 Age Information 1223 Credential Information 1224 Accident Count Information 123 Sample Videos 130 Display section 200 Plan Generation Device 300 Prediction Systems 400 worker terminals 500 stores / head office terminals 600 Camera

Claims

1. Equipped with a processor and memory, The aforementioned processor cooperates with the memory, Obtain the work plan and worker information. Based on the work plan and the worker information, an action evaluation is performed on the planned actions of the worker in question. The degree of risk to the planned actions of the worker in question is evaluated using the aforementioned action evaluation. Work support device.

2. The aforementioned processor further acquires a sample video, If the aforementioned risk level is above a predetermined threshold, at least a portion of the worker information and the sample video are input to the video generation model to generate and output a simulated risk video. The work support device according to claim 1.

3. The aforementioned processor, Obtain an instruction document to be input into the video generation model corresponding to the risk assessment, The instruction document is input into the video generation model. The work support device according to claim 2.

4. The aforementioned processor, The performance evaluation is performed based on the worker's proficiency with the planned actions. The work support device according to claim 1 or 2.

5. The aforementioned processor, The risk level is assessed based on the probability of an accident occurring or the calculation of the amount of damage. The work support device according to claim 1 or 2.

6. The aforementioned worker information includes at least one of the following: years of service, age, qualifications, and number of past accidents. The work support device according to claim 1 or 2.

7. The aforementioned hazard simulation video is generated by combining the sample video with the facial information of the target worker included in the worker information. The work support device according to claim 2.

8. Equipped with a processor and memory, The aforementioned processor cooperates with the memory, The real-time work video and the worker information of the target worker included in the said work video are acquired. The aforementioned work video is divided into units of work actions, Based on the aforementioned work video and the aforementioned worker information, an action evaluation is performed on the target worker's actions. The degree of risk to the target worker's actions is evaluated using the aforementioned performance evaluation. Work support device.

9. The aforementioned processor, Predicting the continuation of the operation of the aforementioned target, The degree of risk of the aforementioned continued operation is evaluated, If the aforementioned risk level is above a predetermined threshold, an alert is issued. The work support device according to claim 8.

10. The aforementioned processor, The aforementioned work video is divided into units of movement for the work by performing motion analysis. The work support device according to claim 8 or 9.

11. Obtain the work plan and worker information. Based on the work plan and the worker information, an action evaluation is performed on the planned actions of the worker in question. The degree of risk to the planned actions of the worker in question is evaluated using the aforementioned action evaluation. Work support method.

12. Obtain the work plan and worker information. Based on the work plan and the worker information, an action evaluation is performed on the planned actions of the worker in question. The computer is instructed to use the aforementioned motion evaluation to assess the degree of risk to the planned actions of the worker in question. Work support program.

Images