JP4936147B2 - Work monitoring system - Google Patents

Description

  The present invention relates to a system for monitoring work conditions.

For example, a task that can be a catastrophe if a mistake is made, such as in air traffic control operations, requires that an accurate judgment be made instantaneously. Such work requires a moderate tension state, but it is difficult to maintain that state for a long time. Therefore, the controller changes work regularly and takes breaks.
In order to prevent each controller from getting tired and making mistakes, the timing of the break is set for safety. However, there are individual differences in the ease of fatigue, etc., and depending on the person, it may be possible to forcibly take a break and the concentration will be interrupted, making it impossible to work accurately.

Therefore, it is conceivable to monitor the mental and physical state of the worker during actual work, and instruct a break or work stop based on this.
Thus, in order to monitor the working state of the worker in this way, a system that utilizes the brain activity level during the working is conventionally known.
In particular, it is known that the worker's cerebral neocortex controls the work, and its activity is closely related to workability, so in conventional systems, the activity of the worker's cerebral neocortex during work is known. The degree is measured, and a break or alternation is instructed based on the activity index.
This system is premised on the idea that human fatigue appears in the activity index of the cerebral neocortex and that mistakes are likely to occur when the activity index of the cerebral neocortex decreases due to fatigue or relaxation. When the value of the activity index measured during the work falls below the set value, a warning is issued to alert or stop the work. However, the set value is determined uniformly by ignoring individual differences.
Japanese Patent No. 3686465 Seiichi Shiomi, “Brain function model considered from speech analysis”, Kansei Engineering Research Papers, Japan, Japan Society for Kansei Engineering, February 2004, Vol. 4, No. 1, p. 3-12

As described above, the conventional system has the following problems because individual differences are ignored in determining the setting value of the brain activity.
For example, when detecting the brain activity level during work and comparing the detected value with a set value set uniformly, a warning may occur depending on the person even if the brain activity level exceeds the set value. On the other hand, even if the brain activity was lower than the set value, there was almost no mistake. In other words, there are individual differences in the relationship between work mistakes and brain activity, so if the above set value is uniformly determined, the person to be monitored will accurately monitor whether the work can be performed correctly. There was a problem that I could not.

  An object of the present invention is to provide a monitoring system capable of accurately monitoring whether or not an appropriate state in which an operation can be correctly performed is performed in consideration of individual characteristics.

According to a first aspect of the present invention, there is provided a storage unit storing an activity index appropriate range of the cerebral neocortex for each monitoring target person who satisfies the required work accuracy for a specific monitoring target work, and the monitoring target. A voice data input unit for inputting voice data of a monitoring subject during work, an output unit, and a processing unit, the processing unit based on the voice data input from the voice data input unit Calculate the cortex activity index, determine whether the calculated activity index of the cerebral neocortex has deviated from the appropriate range of activity index, and output a warning signal to the output section when it deviated from the appropriate range of activity index It is characterized in that it is output.
The activity index of the cerebral neocortex calculated from the voice data is a brain activity index calculated based on the SiCECA algorithm described in Non-Patent Document 1.

A second invention is the first invention assumes, each monitored task to the active index appropriate range stored in the storage unit in association is running a test task set for each of the monitored work Work accuracy per test work unit length calculated by comparing the work data input according to the number of correct answers and processing time based on pre-stored evaluation criteria, and voice data collected during this test work Cerebral neocortex activity index appropriate range that satisfies the required work accuracy of the monitored work specified based on the correspondence table that correlates with the activity index of the cerebral neocortex calculated based on It has the characteristics.
The test work based on the monitoring target work is a work that requires the ability required for the monitoring target work. For example, the monitoring target work itself, a simulated work with the same work type, or a work type. This includes tasks that require the same ability even if they are different.

According to the first invention, it is possible to compare the activity index appropriate range of the cerebral neocortex corresponding to the specific work with the activity index of the cerebral neocortex during the actual work. It is possible to accurately monitor whether or not the work of the monitoring subject is in an appropriate state during the work.
In other words, it was found that the correlation between the activity index of the cerebral neocortex and the work accuracy differs depending on the type of work depending on the worker, and therefore the present invention relates to the cerebral neocortex for each monitoring target for each work. The appropriate range of activity index was stored, and the appropriate range of activity index was compared with the activity index of the cerebral neocortex during work. Accordingly, it is possible to accurately determine whether the work is appropriate for each work and for each person to be monitored, and output a warning signal at an appropriate timing.
Since warning signals can be output at an appropriate timing in this way, accidents and operational errors can be reduced.

In addition, since the activity index of the cerebral neocortex during actual work is calculated based on the voice data of the monitoring subject, for example, the activity index of the cerebral neocortex is calculated from the amount of oxygen and blood flow in the brain. Compared with the case of calculating, the apparatus for detecting the activity index of the cerebral neocortex can be simplified. The blood flow and oxygen consumption of the cerebral neocortex are measured by PET (positron
emission tomography) and topographic devices, but compared to such devices, the audio data input unit of the present invention can be overwhelmingly miniaturized, and it is also necessary to attach a biological information detection device to the monitoring subject. Absent.

  According to the second invention, since the appropriate activity index range of the cerebral neocortex is set from the test work based on the actual work, an accurate appropriate activity range can be set for each work. As a result, the situation of the person being monitored can be monitored more accurately.

1 to 7 show an embodiment of the present invention.
As shown in FIG. 1, the work monitoring system of this embodiment includes a processing unit 1, an audio data input unit 2, an activity index appropriate range storage unit 3, and an output unit 4 connected to the processing unit 1. I have.
The voice data input unit 2 has a function of capturing a voice uttered during the work of the monitoring subject and inputting it to the processing unit 1, and includes a microphone and an A / D conversion means.

The activity index appropriate range storage unit 3 is a storage unit that stores the activity index appropriate range of the cerebral neocortex of a monitoring target person for a specific work, and is a work and monitoring target person as shown in FIG. The worker and the appropriate degree of activity index range are stored in association with each other. The “activity index appropriate range” refers to a range of brain activity that satisfies the required work accuracy required for the specific work. The required work accuracy is set in advance depending on the type of work. For example, in the case of work that does not allow a slight failure, the required accuracy is set high.
The method for determining the appropriate range of activity index of the cerebral neocortex will be described in detail later, but in short, when the activity index of the cerebral neocortex is within the appropriate range of activity index, the work This means that the work accuracy satisfies the required accuracy. And such an activity index appropriate range is stored in the activity index appropriate range storage unit 3 for each work and for each worker who is a monitoring target.

  In addition, the processing unit 1 takes in the voice data of the monitoring subject who is performing the work through the voice data input unit 2 during the work, and calculates the activity index of the cerebral neocortex from the voice data. It has a function. The processing unit 1 compares the calculated brain activity with the appropriate activity index range corresponding to the work, and warns when the calculated activity index of the cerebral neocortex of the monitoring subject deviates from the range. A function of outputting a signal to the output unit 4 is provided. The output unit 4 to which the warning signal is input issues a warning to the outside or turns on a warning lamp.

Next, how to specify the appropriate range of activity index of the cerebral neocortex will be described.
Figure 3 shows the proper range calculation device for calculating the activity index proper range of neocortex for a particular task. The appropriate range calculation device of this embodiment is a device for executing a test work with the monitoring target person of the work monitoring system of this embodiment as an operator, and calculating an appropriateness index appropriate range for this test work. In the form, it is a separate device from the work monitoring system shown in FIG. However, this activity index appropriate range calculation device can also be incorporated into the work monitoring system of the present invention.

The appropriate range calculation apparatus includes a calculation unit 11 that calculates an activity index appropriate range of the cerebral neocortex for the monitoring target work of the monitoring target person of the work monitoring system, and the calculation unit 11 becomes a monitoring target person. The voice data input unit 12 for inputting the worker's voice data is connected to the work data input unit 13 for inputting the work data as the execution result of the test work.
The voice data input unit 12 is for capturing voice data that is the basis of the activity index of the cerebral neocortex of the worker, and specifically comprises a microphone and A / D conversion means. In addition, an operation unit 14 such as a mouse or a keyboard is connected to the work data input unit 13 for an operator to input work data for a test work.

Further, the arithmetic unit 11 includes a test scenario storage unit 15 that stores a scenario of a test work to be executed by the apparatus, an evaluation standard storage unit 16 that stores an evaluation standard for evaluating the result of the test work, and a test A required work accuracy storage unit 17 that stores work accuracy required for work, a data storage unit 18 that stores data input from the voice data input unit 12 and the work data input unit 13, and a data output unit 19 Connected.
The data output unit 19 has a function of outputting an activity index appropriate range, which is a processing result of the calculation unit 11.
Here, the display 10 is connected to the data output unit 19 so that the appropriate activity index range of the cerebral neocortex calculated by the calculation unit 11 is displayed. The displayed activity index appropriate range is manually input to the activity index appropriate range storage unit 3 of the work monitoring system. The function of the unit 11 may be incorporated, and the appropriate activity index range calculated by the processing unit 1 may be calculated and stored in the activity index appropriate range storage unit 3.

In the appropriate range calculation device, the display 10 connected to the data output unit 19 displays the appropriate range of the activity index calculated by the calculation unit 11 as described above. The work is displayed.
FIG. 4 shows, for example, a test operation for a police officer in charge of certifying a violator in a traffic speed violation control, in which the vehicle information displayed in the video is input to the worker. . The vehicle information to be input by the worker is the lane in which the worker is traveling, whether the vehicle is a large vehicle, a normal vehicle, a light vehicle, or the like, color, number, or the like.

  In addition to the moving image window 10a, the display 10 includes a lane number input button 10b including “1”, “2”, and “3” for specifying a lane, and “large”, “normal”, and “light”. A car model input button 10c, a color input button 10d composed of “dark”, “medium”, and “light”, a number input button 10e composed of numbers “0” to “9”, and the like are displayed.

Then, after the operator clicks these input buttons with the mouse, the operator clicks the enter button 10f to input information on the car passing through the moving image area 10a. Here, when the lane number input button 10b is clicked, the moving image is temporarily stopped to make the vehicle stationary so that it can be easily seen. Further, when the mold input button 10c, the color input button 10d, and the number input button 10e are clicked and the enter button 10f is clicked, the moving image is restarted.
In the figure, reference numeral 10g is a number display column for displaying the input number, and reference numeral 10h is a clear button for correcting the input data.

The work data of the present invention is automobile information input by the worker.
A program for displaying the screen for the test work and stopping or playing back the moving image according to the work data input by the worker is stored in the test scenario storage unit 15 as a test work scenario. Store in advance.
Further, in the evaluation criterion storage unit 16, the above-mentioned information regarding all the vehicles appearing in the moving image is stored as correct data in association with, for example, the elapsed time from the start of moving image reproduction or the progress of the moving image. . The processing unit 1 compares the correct answer data with the work data input by the worker, and calculates work accuracy described later.

The test scenario stored in the test scenario storage unit 15 is also set with a timing for outputting an utterance invitation signal for prompting the worker who is performing the test work to speak. The computing unit 11 outputs the utterance invitation signal at the output timing of the utterance invitation signal set in this test scenario. The output of the utterance invitation signal is, for example, displaying words to be read by the worker on the display 10 or outputting a voice signal for prompting utterance from, for example, a speaker.
Then, the calculation unit 11 outputs the above-mentioned generation attraction signal and simultaneously turns on the function of the voice data input unit 12 so that the voice of the operator can be captured.

Hereinafter, a procedure for calculating the appropriate range of the activity index of the worker by executing the test work using the screen shown in FIG. 4 in the appropriate range calculation apparatus of this embodiment will be described.
When the start signal is input, the calculation unit 11 causes the display 10 to display a test work screen including the moving image window 10 a based on the scenario stored in the test scenario storage unit 15.
The operator inputs necessary information by operating the operation unit 4 such as a mouse. However, before starting the test operation, the operator information is also input. The worker information is information for identifying a worker who performs a test work, such as name and gender. However, the worker attributes may be stored in advance in this system, and only the worker ID may be input during the test work.

When the test work is started and the worker inputs work data, the work data is input to the calculation unit 11 via the work data input unit 13. The calculation unit 11 stores the input work data in the data storage unit 18 in association with the progress of the scenario, for example, the playback time of the moving image, the number of frames of the moving image data, and the like.
Further, the calculation unit 11 displays a reading screen on the display 10 as the utterance invitation signal and takes in the audio data from the audio data input unit 12 when the above-described scenario is advanced and the above-described utterance invitation timing comes. The calculation unit 11 also stores the acquired voice data in the data storage unit 18 in association with the degree of progress of the scenario.

When all the test work is completed according to the above scenario, the calculation unit 11 executes the following processing on the work data and the voice data stored in the data storage unit 18.
First, the calculation unit 11 compares the collected work data with the correct answer data stored in the evaluation criterion storage unit 16 to calculate the work accuracy per unit length of the test work according to the progress of the scenario. This unit length can be set for each test operation such as time and the number of frames.
The work accuracy is calculated by comparing the work data input during the unit length with the correct answer data to be input during the unit length. It is calculated from the excess or deficiency of the data, the data input timing, etc., and is a value indicating how accurately a specific test operation has been performed. The criterion for determining the correct answer and the calculation method of the work accuracy are stored in advance in the evaluation reference storage unit 16 for each test work.

  In the case of a test work in which the moving vehicle shown in FIG. 4 is observed and the information is input, the evaluation criteria for the number of vehicles corresponding to the input work data and the evaluation of the correctness of the input information content Standards, evaluation standards for input timing, etc. are defined respectively. Note that the above-mentioned evaluation of the input timing is input when the same correct data is input, when the work data is input immediately after the car appears in the moving image window 10a, and immediately before disappearing from the moving image window 10a. Since the reaction rate differs depending on the case, the evaluation is changed for each reaction rate.

Then, for each item, the work data is evaluated according to the evaluation criteria, and the result is integrated to calculate the work accuracy.
Further, the computing unit 11 calculates an activity index of the cerebral neocortex from the collected voice data according to the SiCECA algorithm described in Non-Patent Document 1.

Then, the calculation unit 11 associates the calculated activity index of the cerebral neocortex with the work accuracy per unit length corresponding to the input of the voice data that is the basis of the activity index. The correspondence between these data is as follows.
As shown in FIG. 5, when the test work is continuously performed and the test work data is also continuously input, the calculation unit 11 divides the progress time of the test work into unit times Δt1, Δt2,. Thus, the work accuracy is calculated for each unit time. On the other hand, when the audio data V1 is input within Δt1 and another audio data V2 is input within Δt2, the activity index of the cerebral neocortex calculated based on the audio data V1 is set to the work accuracy of the unit time Δt1. Correlating, the activity index of the cerebral neocortex calculated based on the voice data V2 is associated with the work accuracy of the unit time Δt2.

  However, depending on the test work scenario, voice data may be input outside the unit time in which the work data is input. For example, when an utterance attraction signal is output every time a work is divided and voice data is input each time, the work accuracy per unit length immediately before or after the voice data is input is set to Correlate with activity index of cerebral neocortex based on data. Note that whether to select immediately before or after the input of audio data is set in advance.

The calculation unit 11 creates a correspondence table between the work accuracy calculated as described above and the activity index of the cerebral neocortex. The values of this correspondence table can be represented by the graph of FIG. 6 with the horizontal axis representing the activity index S of the cerebral neocortex and the vertical axis representing the work accuracy A. A solid line graph G1 shown in FIG. 6 is data related to the test work I of a specific worker X. The graph G1 is a set of plots of a plurality of points P1, P2, P3,... Pm representing work accuracy and activity index of the cerebral neocortex corresponding to the unit length of each scenario progression. The correlation between the activity index S of the cerebral neocortex and the work accuracy A is shown. Further, the graph G2 is data of another worker Y.
Thus, even when the same test work I is performed, the correspondence graph between the work accuracy and the activity index of the cerebral neocortex differs depending on the worker.

Note that the horizontal and vertical axes do not include time elements related to the progress of the test work. Therefore, the graph of FIG. 6 does not show a change in work accuracy over time. There is no correlation with temporal changes.
However, when the activity index of the cerebral neocortex is high, it can be called normal tension, and when the low normal state is relaxed, if the monitored work continues for a long time, it becomes relaxed due to fatigue, and the cerebral neocortex Cortex activity index tends to be low.
In addition, another experiment confirms that, for the same worker, the work accuracy is almost the same if the activity index of the cerebral neocortex is the same regardless of the work time during the same work. . Therefore, the graph characteristic shown in FIG. 6 showing the correlation between the activity index of the cerebral neocortex and the work accuracy is considered to represent the basic ability of the worker X or the worker Y who is an individual in the work. .

On the other hand, even for the same worker, it has been confirmed that the correlation between the activity index of the cerebral neocortex and the work accuracy differs depending on the type of work. FIG. 7 shows a correspondence table between the activity index of the cerebral neocortex and the work accuracy when the worker X executes the test work I and the test work II. This is indicated by a chain line graph G3.
The test operation I and the test operation II correspond to the operations I and II that are actually monitored, respectively. As described above, the test work I and II and the monitoring target work I and II do not have to be exactly the same, but the test work requires the same ability as the case where the monitoring target work is executed, and is being monitored. What is necessary is just to reproduce the correspondence between the brain activity level and the work accuracy. Accordingly, the appropriate activity index range calculated based on the test operations I and II can be applied to the monitoring target tasks I and II as they are.

When the correspondence tables shown in the graphs of FIGS. 6 and 7 are created as described above, the calculation unit 11 specifies the required work accuracy corresponding to the test work from the required work accuracy storage unit 17 and satisfies this requirement. Identify the range of activity index of the cerebral neocortex. Then, the calculation unit 11 outputs the range of the activity index of the identified cerebral neocortex to the data output unit 19 as the activity index appropriate range.
For example, when the required work accuracy is A1, the appropriate activity index range that satisfies the graph G1 in FIG. 6 is Sx1. In other words, the worker X satisfies the required work accuracy of the test work I when the activity index of the cerebral neocortex is within the activity index appropriate range Sx1.

Similarly, the appropriate activity index range Sy1 for the work I of the worker Y and the appropriate activity index range Sx2 for the work II of the worker X can be specified. As shown in FIG. 2, the activity index appropriate range specified in this way is stored in the activity index appropriate range storage unit 3 for each work and for each worker.
FIG. 7 shows the appropriate range of the activity index for the work I and the work II specified with the required work accuracy being constant, but actually, the required work accuracy is different if the work is different. However, the required work accuracy is set for each work.

As described above, when the appropriate activity index appropriate range is calculated for each work and for each worker, and the value is stored in the activity index appropriate range storage unit 3, the operation monitoring system can be operated at any time.
The procedure for monitoring actual work by this system will be described below.
Before operating the work monitoring system of this embodiment and starting the actual work, the operator X does not show a work ID for identifying the work, a worker ID for identifying the worker, and the like. Enter from. Based on the input work ID and worker ID, the processing unit 1 specifies the activity index appropriate range Sx1 for the work I of the worker X from the activity index appropriate range storage unit 3, and this activity index The appropriate range Sx1 is stored and the process proceeds to the following process.

When the actual work I is started and the worker X utters voice during execution, voice data is input from the voice data input unit 2 to the processing unit 1. The voice data captured by the voice data input unit 2 may be any voice generated by the worker X. For example, in the case of a police officer in charge who recognizes a violator by controlling traffic speed violation, Try to speak the middle lane, large cars, regular cars, light cars, etc., colors, numbers, etc.
Also, in the case of work that requires a safety check with a voice, the sound for checking the safety can be captured. If the work does not require utterance, the processing unit 1 is set so that a signal for prompting utterance is periodically sent from the output unit 4.

  When the voice data is input from the voice data input unit 2, the processing unit 1 processes the voice data based on the SiCECA algorithm, and calculates the activity index of the cerebral neocortex. However, unlike the calculation unit 11, the processing unit 1 of this work monitoring system calculates an activity index of the cerebral neocortex each time audio data is input from the audio data input unit 2.

The processing unit 1 compares the calculated activity index of the cerebral neocortex with the previously specified activity index appropriate range Sx1, and determines whether or not the activity index appropriate range Sx1 has deviated. When the calculated activity index of the cerebral neocortex is within the activity index appropriate range Sx1, the processing is continued as it is. When the calculated activity index is outside the activity index appropriate range Sx1, a warning signal is displayed. Is output to the output unit 4 so that an alarm or the like is issued.
The state where the alarm is issued is a state where the activity index of the cerebral neocortex at the time of the worker X deviates from the activity index appropriate range Sx1, and the work accuracy cannot satisfy the required work accuracy. It can be said. When such a state occurs, the warning signal is output to alert the worker or interrupt the work, thereby preventing work mistakes and the like.

In particular, since the relative relationship between the activity index of the cerebral neocortex and the work accuracy is found to be different for each work and for each worker, the appropriate range of the activity index for each work and each worker is specified and stored. As a result, the timing at which the required work accuracy of the work to be monitored cannot be satisfied can be determined more accurately for each worker, and a warning signal can be output at the correct timing.
Further, in this work monitoring system, the worker who is the monitoring target can calculate the brain activity level of the worker simply by emitting a sound at a predetermined timing during actual work. Compared with the case where the activity index is detected, the apparatus can be miniaturized. Furthermore, since the operator does not need to attach a special device to the body, workability is not deteriorated.

It is a work monitoring system block diagram of an embodiment. It is the table which showed the example of the data which the activity index appropriate range memory | storage part has memorize | stored. It is a block diagram of an activity index appropriate range calculation apparatus. It is the figure which showed the test work screen for calculating an activity index appropriate range. It is a figure for demonstrating audio | voice data input timing. It is the graph which showed the correspondence table of the activity index of the cerebral neocortex and the work precision of the workers X and Y who performed the work I. It is the graph which showed the correspondence table of the activity index of the cerebral neocortex when the worker X performed the work I and the work II, and work precision.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Processing part 2 Voice data input part 3 Activity index appropriate range storage part 4 Output part Sx1, Sx2 Activity index appropriate range Sy1 (worker X) Activity index appropriate range (worker Y)

Claims (2)

A storage unit that stores the appropriate activity index appropriate range of the cerebral neocortex for each monitoring target person who satisfies the required work accuracy for the specific monitoring target work , and monitoring during the monitoring target work during the work A voice data input unit that inputs voice data of the subject, an output unit, and a processing unit are provided, and the processing unit calculates an activity index of the cerebral neocortex based on the voice data input from the voice data input unit. A work monitoring system that calculates and determines whether or not the calculated activity index of the cerebral neocortex has deviated from the appropriate range of activity index, and outputs a warning signal to the output unit when it deviates from the appropriate range of activity index. The appropriate range of the activity index stored in the storage unit in association with each monitoring target work is based on the work data input in accordance with the execution of the test work set for each monitoring target work as a pre-stored evaluation criterion. The work accuracy per unit length of the test work calculated by comparing the number of correct answers and processing time based on this, and the activity index of the cerebral neocortex calculated based on the voice data collected during this test work The work monitoring system according to claim 1, which is an appropriate range of an activity index of the cerebral neocortex that satisfies the required work accuracy of the work to be monitored, which is specified based on the associated correspondence table.

Images