Behavioral analysis device and behavioral analysis method
The behavior analysis device addresses inefficiencies in manufacturing by generating models of human behavior patterns, detecting irregularities, and evaluating proficiency, thereby supporting process improvements through automated monitoring and feedback.
Patent Information
- Application Number
- JP2020028092
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2020-02-21
- Publication Date
- 2024-11-13
- Estimated Expiration
- 2040-02-21
AI Technical Summary
Existing methods struggle to efficiently monitor and improve human behavior in manufacturing sites where workers exhibit varying efficiency due to different work methods, leading to inefficiencies and potential errors, requiring skilled supervision for process improvement.
A behavior analysis device that acquires position information, generates models of periodic human behavior patterns, detects irregularities, and evaluates proficiency levels by comparing individual behavior with expert models, facilitating automated identification of improvement areas.
Enables automated detection of irregular behaviors and proficiency assessment, supporting process improvements by highlighting areas needing enhancement and providing actionable insights for managers.
Smart Images

Figure 0007581629000001 
Figure 0007581629000002 
Figure 0007581629000003
Abstract
Description
[Technical field]
[0001] The present invention relates to a technique for analyzing human behavior. [Background technology]
[0002] Improving production efficiency is an important issue in manufacturing sites. In sites where many manual tasks are performed and each worker is required to perform complex tasks, such as cell production, each worker has a different way of proceeding with the task and the way of moving, which tends to cause variation in the work efficiency of each worker. Therefore, it is desirable to find waste in each worker's movements and to find areas where delays and mistakes are likely to occur, and to improve the process and work content. However, in the past, the only way to improve the process and work content was for supervisors and skilled workers to monitor the movements of each worker and find wasteful movements, delays, and tasks that are likely to cause mistakes, and the reality was that it took a lot of time and advanced skills to improve the process and work content.
[0003] Patent Document 1 proposes a method of supporting workers by providing a system that makes it easy to search for the know-how and examples of experts. However, it is difficult to improve work efficiency or processes by simply providing such a system. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] JP 2007-148938 A Summary of the Invention [Problem to be solved by the invention]
[0005] The present invention has been made in consideration of the above-mentioned circumstances, and an object of the present invention is to provide a technique for assisting in monitoring human behavior and finding areas requiring improvement. [Means for solving the problem]
[0006] The present disclosure includes a behavior analysis device characterized by having an acquisition unit that acquires location information of a person performing a predetermined behavior within a predetermined area, a data storage unit that accumulates time series data of the location information acquired by the acquisition unit, and a data analysis unit that generates a model representing an average change in the location information over one period by analyzing a periodic pattern of changes in the location information from the time series data.
[0007] This configuration makes it possible to automatically generate a model that represents the trends of human behavior (changes in location information). This model can be used for various purposes, such as comparing differences in behavior between individuals, understanding the characteristics of each individual's behavior (good points, bad points, wasteful movements, etc.), and finding areas that need improvement in a specific area or specific behavior.
[0008] The behavior analysis device may further include an evaluation unit that detects non-stationary behavior by comparing the acquired position information for one cycle with the model. This configuration is useful for monitoring human behavior because it can automatically detect when a person performs non-stationary behavior. For example, the evaluation unit may determine that the behavior is non-stationary when the difference in position information at a corresponding time point between the position information for one cycle and the model and / or the difference in the length of one cycle exceeds a predetermined threshold. This makes it possible to detect non-stationary behavior with simple processing. The behavior analysis device may further include an output unit that issues a notification when non-stationary behavior is detected.
[0009] The data storage unit may store time-series data of one or more people, and the data analysis unit may analyze a periodic pattern of changes in location information for each person to generate a model for each person. This makes it possible to grasp different tendencies for each person. The evaluation unit may compare the location information for one cycle with the model for each person. This makes it possible to accurately detect non-stationary behavior for each person.
[0010] The behavior analysis device may further include an evaluation unit that determines the proficiency of each of the plurality of people with respect to the predetermined behavior by relatively evaluating a plurality of models generated from the time-series data of the plurality of people. With this configuration, the proficiency of each person can be automatically and easily grasped. The evaluation unit may further select a model of an expert from the plurality of models based on the determined proficiency, and compare the model of the evaluation subject with the model of the expert. Such a comparison makes it possible to evaluate the good or bad of the behavior of the evaluation subject, and to detect areas in the behavior of the evaluation subject that need improvement. For example, the behavior analysis device may further include an output unit that outputs information indicating the difference between the model of the evaluation subject and the model of the expert. For example, by providing such information to a manager or supervisor, it is possible to support the discovery of areas that need improvement and efforts to improve processes.
[0011] The evaluation unit may determine the proficiency of the plurality of people by relatively evaluating the time length of each of the plurality of models. This allows the proficiency to be determined by simple processing. For example, when the predetermined behavior is a task consisting of one or more steps, the evaluation unit may determine the proficiency of each step of the plurality of people by determining the working time for each step from each of the plurality of models and classifying each step based on the length of the working time. This allows the evaluation of each step and the identification of areas requiring improvement.
[0012] The acquisition unit may acquire position information of a person based on information acquired from a sensor that senses a person present in the predetermined area. The sensor may be, for example, an image sensor, a human presence sensor, or a sensor that detects a person's position in combination with a device carried by the person. The acquisition unit may identify a person performing the predetermined behavior by face recognition, or may acquire identification information for identifying a person performing the predetermined behavior from an external source.
[0013] The present invention may be understood as a behavior analysis device having at least some of the above means, or as a behavior evaluation device, a behavior monitoring device, an anomaly detection device, a proficiency evaluation device, a process improvement support device, or the like. The present invention may also be understood as a behavior analysis method, a behavior evaluation method, a behavior monitoring method, an anomaly detection method, a proficiency evaluation method, a process improvement support method, or the like, including at least some of the above processing. The present invention may also be understood as a program for causing a processor to execute each step of such a method, or a computer-readable recording medium non-temporarily storing the program. Note that the above means and processing can be combined with each other as much as possible to constitute the present invention. Effect of the Invention
[0014] According to the present invention, it is possible to provide a technique for assisting in monitoring human behavior and finding areas requiring improvement. [Brief description of the drawings]
[0015] [Figure 1] FIG. 1 is a diagram showing an example of analysis of the behavior of workers on a production line. [Diagram 2] FIG. 2 is a block diagram showing an example of the configuration of a surveillance system including a behavior analysis device. [Diagram 3] FIG. 3 is a flowchart of data analysis and model generation by the behavior analysis device. [Figure 4] FIG. 4 is a schematic diagram of the data analysis and model generation process. [Diagram 5] FIG. 5 is a flowchart of anomaly detection by the behavior analysis device. [Figure 6] FIG. 6 is a schematic diagram of the flow of anomaly detection. [Figure 7] FIG. 7 is a flowchart of proficiency determination by the behavior analysis device. [Figure 8] FIG. 8 is a schematic diagram of the flow of skill level determination. [Figure 9] FIG. 9 is a flowchart of model comparison by the behavior analysis device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0016] <Application Examples> Referring to FIG. 1, an example in which the present invention is applied to the analysis of worker behavior on a production line will be described as one application example of the present invention.
[0017] In a production line such as a cell production system, workers move within a designated work area and perform multiple work processes in order. At this time, since workers perform their work according to a specified procedure, a certain periodicity and regularity may appear in the behavior of the workers. For example, the left side of Figure 1 is a graph plotting the temporal change in the location information of the workers (the horizontal axis shows time and the vertical axis shows the location coordinates), and it can be seen that a periodic pattern (dashed line) appears in the change in the location information. In this method, by analyzing the periodic pattern of the change in the location information from the time series data, a model is generated that represents the average change in the location information of each worker in one period. This model represents the statistical tendency of the behavior of the worker, that is, the stationary behavior.
[0018] Such models can be used, for example, to visualize the tendencies of each worker, to distinguish and detect abnormal behavior (non-regular behavior), to evaluate the level of proficiency in a task, and to compare the behavior of workers. The right side of Figure 1 shows an example of how the model can be used. For example, when comparing the models of an expert and a newcomer A, the shape of the pattern is roughly the same, but the pattern of the newcomer A is more horizontal (i.e., has a longer period), so it can be seen that the newcomer A has a lower overall level of proficiency in the task. Also, when comparing the models of an expert and a newcomer B, it can be seen that the newcomer B has a lower level of proficiency in the process corresponding to position X. Also, by comparing the behavior of the workers that is captured sequentially with the model, it is possible to easily distinguish and detect abnormal behavior of the workers. If it is confirmed that abnormal behavior occurs frequently at the same position (process), there may be a problem in that process. For example, by comparing it with the history of the machine used in that process, it can be useful for discovering abnormalities or failures in the machine, and for improving the work procedures in that process.
[0019] <Embodiment> (Monitoring system configuration) A behavior analysis device according to an embodiment of the present invention will be described with reference to Fig. 2. Fig. 2 is a block diagram showing an example of the configuration of a surveillance system including a behavior analysis device.
[0020] The monitoring system 1 is a system for monitoring the working status of workers on a factory production line, and includes a behavior analysis device 10 and a sensor 11 as main components. The specific configuration of the behavior analysis device 10 and the sensor 11 is not important. For example, the behavior analysis device 10 and the sensor 11 may be connected to each other by wire or wirelessly so as to be able to communicate with each other, or may be configured as an integrated unit (i.e., the behavior analysis device 10 and the sensor 11 are built into one housing). In the former configuration, the number of behavior analysis devices 10 and the number of sensors 11 are not limited to 1:1, and may be 1:N, N:1, or N:N (N is an integer of 2 or more). In the latter configuration, the control system of the sensor 11 and the function of the behavior analysis device 10 may be implemented in the same processor.
[0021] (Sensor) The sensor 11 is a device for sensing the position of a worker present on the production line. There is no limit to the type of sensor 11 as long as it can sense the position of the worker. For example, it may be an image sensor installed to capture an image of the worker's activity area, or a human presence sensor that detects the position of the worker within the activity area. Examples of the human presence sensor include an infrared sensor and a radio wave sensor. Alternatively, a mechanism may be used in which the position of the worker is detected by combining the sensor 11 with a device carried by the worker (such as a tag, a smartphone, a BLE device, or a transmitter). The sensing results of the sensor 11 are sequentially taken into the behavior analysis device 10.
[0022] In this embodiment, an image sensor is used as the sensor 11. This is because the image sensor has advantages such as being able to monitor a wide area with one sensor, being able to simultaneously obtain position information of multiple workers, and being able to measure position information with high accuracy. The image sensor has a wide-field camera (for example, a fisheye camera or an omnidirectional camera) and an image processing unit that processes images captured by the camera. The image processing unit may have, for example, a function to detect a human face or human body from an image, a function to track the detected face or human body, and a function to identify an individual by face recognition or human body recognition. The image processing unit is composed of, for example, a processor and a memory, and the above-mentioned functions are realized by the processor reading and executing a program stored in the memory. Note that all or a part of the above-mentioned functions may be realized by a processor such as an ASIC or an FPGA.
[0023] (behavior analysis device) The behavior analysis device 10 is a device that performs behavior analysis of a worker working on a production line using a sensing result captured from a sensor 11. The behavior analysis device 10 of this embodiment has, as main components, an acquisition unit 101, a data accumulation unit 102, a data analysis unit 103, an evaluation unit 104, and an output unit 105. The behavior analysis device 10 can be configured, for example, by a general-purpose computer equipped with a processor (CPU), a ROM, a RAM, a storage (HDD, SSD, etc.), an input device (keyboard, pointing device, etc.), a display device, etc. In this case, the components 101 to 105 shown in FIG. 2 are realized by the processor reading and executing a program stored in the ROM or storage. Note that all or a part of the components 101 to 105 may be realized by a processor such as an ASIC or FPGA. In addition, the functions of the behavior analysis device 10 may be realized by using cloud computing or distributed computing.
[0024] The acquisition unit 101 has a function of acquiring data of the sensing result from the sensor 11. The sensing result includes, for example, position information of the detected worker and time information indicating the detection time. The sensing result may also include information other than the position information and time information, such as the ID of the worker (identification information indicating who the worker is) and the number of the production line. The position information is, for example, a coordinate value indicating the position where the worker is present. The coordinate system of the position information may be a sensor coordinate system or a global coordinate system. A two-dimensional coordinate system indicating the position in a plane may be used, or a one-dimensional coordinate system may be used when the movement of the worker can be regarded as a simple reciprocating movement. In this embodiment, a configuration is adopted in which the sensor 11 (image sensor) determines the position information of the worker, but raw data (image data in the case of an image sensor) may be taken in from the sensor 11, and the acquisition unit 101 may analyze the raw data (detect a face or human body in the case of image data) to recognize the position information of the worker. The acquisition unit 101 may also identify the worker from the image data by a technology such as face recognition. Alternatively, identification information of the worker may be externally input to the acquisition unit 101 separately from the sensing result. For example, identification information read from an ID card carried by the worker may be input to the acquisition unit 101 together with time information, and linked to position information based on the time information. Alternatively, a user (operator of the behavior analysis device 10) may manually input the ID of the worker.
[0025] The data accumulation unit 102 has a function of accumulating the time-series data of the sensing results acquired by the acquisition unit 101 in a non-volatile storage. Fig. 2 shows a schematic example of the accumulated time-series data. In the example of Fig. 2, the production line number, location information, and time information indicating the detection date and time are accumulated for each worker ID.
[0026] The data analysis unit 103 has a function of generating a model that represents an average change in location information in one period by analyzing a periodic pattern of changes in location information from the time-series data. At this time, the data analysis unit 103 analyzes the time-series data for each worker and generates a model for each worker. In other words, it can be said that the data analysis unit 103 generates a model that represents the statistical tendency (regular behavior) of the behavior of each worker.
[0027] The evaluation unit 104 has a function of detecting unsteady behavior by comparing one period of position information newly acquired by the acquisition unit 101 with a model generated by the data analysis unit 103. The evaluation unit 104 also has a function of determining the proficiency of each worker by relatively evaluating a plurality of models generated from the time-series data of a plurality of workers, and a function of comparing models (behavioral tendencies) between an expert and an unexpert.
[0028] The output unit 105 has a function of outputting information obtained by the data analysis unit 103 and the evaluation unit 104. The output unit 105 may output information to a display device included in the behavior analysis device 10, or may transmit information to an external device. For example, the output unit 105 may transmit a notification message to a terminal of an administrator or supervisor, may transmit a warning signal or a control signal to another device, or may emit sound, light, vibration, or the like.
[0029] (Model Generation) An example of the data analysis and model generation process will be described with reference to Fig. 3 and Fig. 4. Fig. 3 is a flowchart of the data analysis and model generation by the behavior analysis device 10, and Fig. 4 is a schematic diagram of the flow of the data analysis and model generation. The process of Fig. 3 is preferably executed in a state where a statistically sufficient amount of time-series data has been accumulated in the data accumulation unit 102.
[0030] In step S300, the data analysis unit 103 determines the person to be analyzed. For example, when data on a plurality of workers is stored in the data storage unit 102, the data analysis unit 103 may select the person to be analyzed in order of the worker ID. Alternatively, the user may specify the person to be analyzed. The subsequent processes of steps S301 to S304 are executed for each person to be analyzed.
[0031] In step S301, the data analysis unit 103 reads the time-series data 40 of the position information of the person to be analyzed from the data storage unit .
[0032] In step S302, the data analysis unit 103 divides the time series data of the location information into periods and extracts a large number of samples 41 of "changes in the location information for one period." The start and end points of one period may be determined by a peak (maximum or minimum) in the time series data of the location information, or may be determined based on the value of the location information. For example, if a task is made up of n work steps, from step 1 to step n, and a worker repeatedly performs the tasks of step 1 to step n, the end of one period (and the start of the next period) may be determined when the value of the worker's location information enters the range corresponding to the work location of step 1.
[0033] In step S303, the data analysis unit 103 generates a model 42 that represents an average change in the position information in one period, using the multiple samples 41 extracted in step S302. For example, the data analysis unit 103 performs curve fitting using the least squares method or the like on data in which the position information of multiple samples is plotted with the time phase aligned, Alternatively, the data analysis unit 103 may calculate an average value of position information at the same phase (time point) among a plurality of samples, and generate a model using a sequence of average value points (or a curve fitted to the sequence of points). Also, a model may be generated using a method other than the method described here. Note that all samples extracted in step S302 may be used for model generation, or model generation may be performed using only samples with approximately the same period length.
[0034] In step S304, the data analysis unit 103 registers the model 42 generated in step S303 in the data accumulation unit 102 together with the worker ID of the person to be analyzed and information on the period of the time-series data used in the data analysis. The data registered here will hereinafter be referred to as "individual smoothed data" meaning data that defines the steady (level) behavior of each worker. Note that the data accumulation unit 102 may be configured to register multiple individual smoothed data with different time-series data periods for the same worker. This is because a worker's proficiency in a task increases over time, and the way he or she behaves may change.
[0035] In step S305, the data analysis unit 103 checks whether the processing for all the workers has been completed, and if not, the process returns to step S301 to analyze the data for the next worker.
[0036] (Detection of abnormalities) An example of anomaly detection processing will be described with reference to Figs. 5 and 6. Fig. 5 is a flowchart of anomaly detection by the behavior analysis device 10, and Fig. 6 is a schematic diagram of the flow of anomaly detection. The processing of Fig. 5 can be executed once personal leveled data is generated and registered. Note that the processing of Fig. 5 may be executed online (i.e., on data successively acquired from the sensor 11 while the production line is in operation) or offline (i.e., for the purpose of later evaluation of data accumulated in the data accumulation unit 102, etc.). Here, the online processing will be described.
[0037] In step S500, the evaluation unit 104 determines an evaluation target. For example, when personal leveling data of a plurality of workers is registered in the data storage unit 102, the evaluation unit 104 may select the evaluation target in order of the worker ID. Alternatively, the user may specify the evaluation target. The subsequent processes of steps S501 to S504 are executed for each evaluation target.
[0038] In step S501, the evaluation unit 104 reads data 60 of position information for the most recent one period of the person to be evaluated from the data storage unit 102. The start point and end point of one period may be determined in the same manner as described in step S302. The evaluation unit 104 may fit a curve to the sequence of points of the position information for one period.
[0039] In step S502, the evaluation unit 104 reads the individual smoothed data 61 of the evaluation subject from the data storage unit 102, and compares the position information 60 for one period acquired in step S501 with the individual smoothed data 61. For example, the evaluation unit 104 may calculate the difference 62 in the value of the position information at the corresponding time (phase) between the position information 60 for one period and the individual smoothed data 61, or the difference 63 in the length of one period. If these differences 62 and 63 exceed a predetermined threshold (YES in step S503), the output unit 105 notifies that an abnormality (unsteady behavior of the evaluation subject) has been detected (step S504). At this time, it is preferable to notify not only the presence or absence of an abnormality, but also information indicating the work line or work process where the abnormality occurred. The work process where the abnormality occurred can be estimated, for example, from the position 64 where the deviation between the position information 60 for one period and the individual smoothed data 61 is the largest.
[0040] In step S505, the evaluation unit 104 checks whether the processing of all the workers has been completed. If the evaluation is not completed, the process returns to step S501 and the evaluation of the next worker is performed.
[0041] (Determination of proficiency) An example of the proficiency determination process will be described with reference to Fig. 7 and Fig. 8. Fig. 7 is a flowchart of proficiency determination by the behavior analysis device 10, and Fig. 8 is a schematic diagram of the flow of proficiency determination. The process in Fig. 7 can be executed when personal leveled data of multiple workers is generated and registered. In this embodiment, the personal leveled data of multiple workers is relatively evaluated to determine the proficiency of each of the multiple workers for the task.
[0042] In step S700, the evaluation unit 104 acquires personal leveling data of multiple workers from the data storage unit 102. Then, in step S701, the evaluation unit 104 calculates the work time for each process from the personal leveling data. Since the correspondence between the value of the location information and the work location of each process is known in advance, it is possible to determine the start and end points of each process based on the value of the location information, as shown in Fig. 8.
[0043] In step S702, the evaluation unit 104 calculates the average and variance of the work time for each process from the data of multiple workers. Then, in step S703, the evaluation unit 104 sets a threshold value for classifying workers according to their proficiency for each process. For example, as shown in FIG. 8, three classes may be classified: those whose work time is shorter than "average - 1σ" (σ is standard deviation) are "experts," those whose work time is in the range of "average - 1σ" to "average + 1σ" are "average," and those whose work time is longer than "average + 1σ" are "novices." Note that the method of setting the threshold value and the number of classes are not limited to this and may be designed as appropriate.
[0044] In step S704, the evaluation unit 104 determines the person to be evaluated. For example, the evaluation unit 104 may select the person to be evaluated in order of the worker ID. Alternatively, the user may specify the person to be evaluated. In step S705, the evaluation unit 104 calculates the working time of each process from the individual leveling data of the person to be evaluated (if the calculation has already been performed in step S701, the calculation result may be reused), and determines the proficiency for each process by comparing with the threshold set in step S703. In this embodiment, for example, the proficiency for each process is set as 2 (expert), 1 (average), and 0 (rookie). Then, in step S706, the evaluation unit 104 calculates the overall proficiency by adding up the proficiency for each process. For example, when there are three processes, the minimum value of the overall proficiency score is 0 (when the proficiency for each process is all 0 (rookie)), and the maximum value is 6 (when the proficiency for each process is all 2 (expert)).
[0045] In step S707, the evaluation unit 104 registers the data on the proficiency for each process and the overall proficiency calculated in steps S705 and S706 in the data accumulation unit 102 together with the worker ID of the person to be evaluated and information on the period of the personal leveling data used in the evaluation. The data registered here will be referred to as "proficiency data" hereinafter. Note that the data accumulation unit 102 may be configured to register multiple pieces of proficiency data for the same worker, each with a different period of personal leveling data. This is because proficiency can improve over time.
[0046] In step S708, the evaluation unit 104 checks whether the processing for all the workers has been completed, and if not, the process returns to step S704 to calculate the skill level of the next worker.
[0047] (Compared to an expert) An example of the process of model comparison with an expert will be described with reference to Fig. 9. Fig. 9 is a flowchart of model comparison by the behavior analysis device 10. The process of Fig. 9 can be executed when personal leveled data and proficiency data are generated and registered.
[0048] In step S900, the evaluation unit 104 determines an evaluation target. For example, the user may specify the evaluation target, or the evaluation unit 104 may automatically select the evaluation target. In step S901, the evaluation unit 104 reads the individual leveled data of the evaluation target from the data accumulation unit 102.
[0049] In step S902, the evaluation unit 104 selects one expert based on the proficiency data, and reads the individual leveled data of the expert from the data storage unit 102. For example, the worker with the highest overall proficiency may be selected as the expert, or if attention is focused on only a specific process, the worker with both high proficiency in that process and high overall proficiency may be selected as the expert.
[0050] In step S903, the evaluation unit 104 compares the personal leveled data of the evaluation target with the personal leveled data of the expert. Then, in step S904, the output unit 105 outputs information indicating the difference between the personal leveled data of the evaluation target and the personal leveled data of the expert. For example, as shown on the right side of FIG. 1, the personal leveled data of the evaluation target and the personal leveled data of the expert may be displayed overlapping on the time axis. In addition, information indicating the difference in work time for each process, the difference in total work time, the process with the largest difference in work time, etc. may be output.
[0051] (Advantages of this embodiment) According to the device of the present embodiment described above, it is possible to automatically generate individual leveling data, which is a model that represents the tendency of human behavior (changes in location information). This individual leveling data can be used for various purposes, such as for comparing differences in behavior between individuals, for grasping characteristics of each individual's behavior (good points, bad points, wasteful movements, etc.), and for discovering areas requiring improvement in a specific area or specific behavior. In addition, since it is possible to automatically detect when a person performs non-regular behavior, it is useful for monitoring human behavior. In addition, it is possible to automatically and easily grasp the proficiency of each person. Furthermore, by comparing with an expert, it is possible to evaluate the good and bad of the behavior of the person being evaluated, and to detect areas requiring improvement in the behavior of the person being evaluated. For example, by providing such information to a manager or supervisor, it is possible to support the discovery of areas requiring improvement and efforts to improve processes.
[0052] <Other> The above embodiment merely exemplifies the configuration of the present invention. The present invention is not limited to the above specific form, and various modifications are possible within the scope of the technical idea. For example, the above embodiment is directed to analyzing and evaluating the behavior of workers on a production line, but the present invention can be used for other objects and applications. Other possible objects are cases where a person is scheduled to perform a specific action within a specific area. For example, the present invention can be applied to the application of analyzing the behavior of passersby at ticket gates at stations or gates at facilities.
[0053] <Appendix 1> (1) an acquisition unit (101) for acquiring location information of a person performing a predetermined action within a predetermined area; a data storage unit (102) that stores time-series data of the location information acquired by the acquisition unit (101); a data analysis unit (103) that generates a model representing an average change in the location information in one period by analyzing a periodic pattern of the change in the location information from the time series data; A behavior analysis device (10) comprising: [Explanation of symbols]
[0054] 1: Surveillance system 10: Behavior analysis device 11: Sensor
Claims
1. An acquisition unit that acquires location information of a person who repeatedly performs a task consisting of multiple steps within a predetermined area; a data storage unit that stores time-series data of the location information acquired by the acquisition unit; a data analysis unit that performs a process of dividing the time series data into periods, extracting a plurality of samples that represent changes in position information for one period from the start to the end of the task, and generating, for each person, a model that represents an average change in position information during one period from the start to the end of the task, using the plurality of samples; an evaluation unit that determines a proficiency level of each of the plurality of people for the task by relatively evaluating a plurality of models generated from the time-series data of the plurality of people, The behavioral analysis device is characterized in that the evaluation unit determines the proficiency of each process of each of the multiple people by calculating the work time for each process from each of the multiple models and classifying each process based on the length of the work time.
2. The evaluation unit further detects non-steady behavior of an evaluation subject selected from the plurality of people by comparing one period of position information acquired from the evaluation subject with a model of the evaluation subject.
2. The behavior analysis device according to claim 1 .
3. The evaluation unit determines that the behavior is unsteady when a difference in the position information at a corresponding time point between the position information for one period and the model and / or a difference in the length of one period exceeds a predetermined threshold.
3. The behavior analysis device according to claim 2.
4. The device further includes an output unit that notifies the user when an abnormal behavior is detected.
4. The behavior analysis device according to claim 3.
5. The evaluation unit further selects a model of an expert from among the plurality of models based on the determined proficiency level, and compares the model of the expert with the model of the person to be evaluated selected from the plurality of people.
2. The behavior analysis device according to claim 1 .
6. The method further includes an output unit that outputs information indicating a difference between the model of the person to be evaluated and the model of the expert.
6. The behavior analysis device according to claim 5.
7. The evaluation unit determines the proficiency levels of the plurality of people by relatively evaluating the time lengths of the plurality of models.
7. The behavior analysis device according to claim 1, 5, or 6.
8. The acquisition unit acquires position information of people based on information acquired from a sensor that senses people present in the predetermined area.
8. The behavior analysis device according to claim 1, wherein the behavior analysis device is a device for analyzing a behavior of a user.
9. The sensor is an image sensor, a human presence sensor, or a sensor that detects the position of a person in combination with a device carried by the person.
9. The behavior analysis device according to claim 8.
10. The acquisition unit identifies the person performing the task by face recognition.
10. The behavior analysis device according to claim 1, wherein the behavior analysis device is a device for analyzing a behavior of a user.
11. 10. The behavior analysis device according to claim 1, wherein the acquisition unit acquires identification information for identifying the person performing the task from an external source.
12. A step in which a processor acquires location information of a person who repeatedly performs a task consisting of multiple steps within a predetermined area; a processor divides the accumulated time series data of the location information into periods, extracts a plurality of samples representing changes in the location information for one period from the start to the end of the task, and performs a process of generating, for each person, a model representing an average change in the location information for one period from the start to the end of the task using the plurality of samples; A step of determining a proficiency level of each of the plurality of people for the task by relatively evaluating a plurality of models generated from the time-series data of the plurality of people, A behavioral analysis method characterized in that in the step of determining the level of proficiency, a processor determines the working time for each process from each of the multiple models, and classifies each process based on the length of working time, thereby determining the level of proficiency for each process for each of the multiple people.
13. A program for causing a processor to execute each step of the behavior analysis method according to claim 12.
Citation Information
Patent Citations
Equipment inspection support system, method and program
JP2007148938A
Operation analysis device and operation analysis method
JP2011248664A
Computer program, operation analysis support method, and operation analysis support device
JP2017151520A
Work support system, management server, mobile terminal, work support method, and program
JP2017228030A
Production management system and production management method
JP2018067170A
