JP2004348517A - Human error diagnostic system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a human error diagnostic system for diagnosing a human error occurrence probability. <P>SOLUTION: The human error diagnostic system diagnoses occurrence probability of human error of the type caused by interrupting a correct schema normally when a note of correct schema escapes by a Stroop task (S12, S13) by double tasks combined with 4-digit numerical value recitation task (S11, S14, and S15) and the Stroop task (S12, S13). <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a human error diagnosis system for diagnosing the possibility of occurrence of a human error that is a cause of an accident caused by a human factor, and particularly to the possibility of occurrence of a specific human error classified based on a cognitive occurrence mechanism. The present invention relates to a human error diagnosis system for diagnosing.
[0002]
[Prior art]
An accident occurs because there is a wrong event that is different from the flow of events that should be performed. The occurrence of this erroneous event is often caused by human error, that is, human error. However, human errors often have various causes intricately involved, and it is extremely difficult to analyze the causes and plan countermeasures.
[0003]
Conventionally, a safety management support system disclosed in Patent Literature 1 below is known as a system for analyzing such human errors and supporting the planning of measures to prevent recurrence.
[0004]
This safety management support system includes an accident database that classifies and stores an accident that has occurred according to the time, place, and magnitude of responsibility of the accident. In addition, for each accident, data indicating a party error (human error), direct factor data indicating a direct cause of the accident (for example, “ignore signal”), and background factor data indicating a background factor of the accident ( For example, a human error database for storing “drinking driving”, “ignorance”, and the like is provided. It also has a countermeasure database that stores countermeasures taken for past accidents.
[0005]
Then, the safety management support system having such a configuration uses the information in the database to analyze the accident by performing tallying such as “analysis of the previous year” and “analysis of the factor ratio,” It is configured to draft.
[0006]
[Patent Document 1]
JP-A-10-105567
[0007]
[Problems to be solved by the invention]
However, in order to effectively prevent human error, it is considered effective to diagnose the possibility of human error for each individual and take measures in consideration of the result of this diagnosis. No diagnostic device has been reported.
[0008]
The present invention has been made in view of such a problem, and an object of the present invention is to provide a human error diagnosis system for diagnosing the possibility of a human error in order to effectively prevent the occurrence of a human error. .
[0009]
[Means for Solving the Problems]
In order to solve the above problems, a human error diagnosis system according to the present invention performs various operations in a human error diagnosis system that causes a person to be diagnosed to execute a predetermined problem and diagnoses the possibility of occurrence of a human error. An arithmetic unit, an input device for the subject to input a response to the task, storage means for holding information for executing the task and information on the response of the subject to the task, and A display device for presenting to a person to be diagnosed, wherein the storage device includes, as the predetermined task, an interference-free stimulus in a situation where there is no interruption due to a habitual schema, and the habitual schema. Holds information on the stimulus with interference, which is a situation in which a human error is likely to occur due to the interruption of the The interfering stimulus and the non-interfering stimulus are each presented to the display device in a random order a predetermined number of times, and from the response information to the interfering stimulus and the response information to the non-interfering stimulus held in the storage device. It is characterized by diagnosing the possibility of occurrence of a human error of a type that is interrupted by a habitualized schema.
[0010]
In addition, the human error diagnosis method according to the present invention, by using a human error diagnosis system including an arithmetic unit, a storage device, an input device, and a display device, allows a person to be diagnosed to execute a Stroop task to reduce the possibility of a human error. A human error diagnosis method for diagnosing, wherein the arithmetic unit is configured to cause the display device to have an interference-based Stroop stimulus and an interference-free Stroop stimulus based on information on an interference-based Stroop stimulus and an interference-free Stroop stimulus stored in the storage device. A presenting step of presenting the stimuli a plurality of times in a random order, a storing step of storing, in the storage device, response information input by the subject to the Stroop stimulus via the input device, Based on the reaction information, the type of human A diagnostic step of diagnosing the likelihood of over, characterized by comprising a.
[0011]
Further, the human error diagnosis program according to the present invention allows a computer having an arithmetic device, a storage device, an input device, and a display device to execute a Stroop task for a person to be diagnosed to diagnose the possibility of occurrence of a human error. Is a human error diagnostic program for executing, based on the information on the Stroop stimulus with interference and the Stroop stimulus without interference stored in the storage device, the Stroop stimulus with interference on the display device, Presenting the interference-free Stroop stimulus a plurality of times in a random order, holding the reaction information input by the subject to the Stroop stimulus via the input device in the storage device, and the arithmetic device Is a type that a habitual schema interrupts based on the reaction information Characterized in that to execute the steps of diagnosing the likelihood of human error, to the computer.
[0012]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0013]
(1st Embodiment)
First, a first embodiment of the present invention will be described. FIG. 1 is a diagram illustrating a schematic configuration of a human error diagnosis system 1 according to the first embodiment. As shown in FIG. 1, a human error diagnosis system 1 according to the present embodiment includes a personal computer main body (hereinafter, referred to as a “PC main body”) 10, a display 20 as an output display device, a keyboard 30 as an input device, and The dedicated key 40 is used to select an option when the subject performs a Stroop task to be described later. The dedicated key 40 has three keys for selecting one from three options.
[0014]
The PC body 10 also includes an arithmetic unit 11 for performing various operations, a storage device 12 for storing information necessary for diagnosing human errors, which will be described later, and information input by the subject through the keyboard 30, and a time for measuring time. Timer 13 is included. The timer 13 can measure time in units of 1 ms.
[0015]
The human error diagnosis system 1 having the above configuration is a system that allows a person to be diagnosed to execute a Stroop task, thereby diagnosing a possibility of causing a specific error type based on a model of a cognitive occurrence mechanism. In this way, by diagnosing the degree of the possibility of causing a specific error type for each individual, it is possible to carry out appropriate personnel assignment on site, and to the individuals who are likely to cause the error type. In such a case, it is possible to call attention so as not to cause the error, and to carry out effective prevention training for the error type.
[0016]
In describing the diagnosis processing by the human error diagnosis system 1, first, an error type classified based on the above-described model of the cognitive occurrence mechanism will be described in detail.
[0017]
First, FIG. 2 is a diagram illustrating a range of a human error targeted by the human error diagnosis system 1. As shown in the figure, the “human error” according to the present embodiment excludes “violation” in which the user does not consciously do what to do in a scene where he knows what to do. Also, in the present embodiment, “human error” is defined as “good definition error”, which is an error that occurs when a task to be performed is known, and “bad”, which is an error that occurs when the task is unknown. Definition error ". The target of this human error separation system is this “good definition error”.
[0018]
FIG. 3 is a diagram showing a model for explaining the occurrence of the good definition error. FIG. 3A is a diagram illustrating “when the act is correctly performed”, and FIG. 3B is a diagram illustrating “when the act is incorrectly performed”. As shown in FIG. 7A, if attention is paid to the “correct schema”, the “correct schema” is stimulated, and the “correct action” is performed. However, as shown in FIG. 3B, if the attention does not turn to the “correct schema”, that is, the association strength between “stimulus” and “correct schema” is small, or “stimulus” and “wrong schema” If the strength of the association between "" and the output of "correct schema" conflicts with the output of "wrong schema" and the output of "wrong schema" wins, the "wrong schema" will be stimulated and Act is performed. Here, the "schema" is a structure of knowledge or a framework of cognition, and refers to a mutual relationship (causal relationship or logical relationship) of knowledge held by a large amount of humans.
[0019]
Next, FIG. 4 is a diagram showing a human error classification system based on the model of the cognitive occurrence mechanism according to the present embodiment. FIG. 5 is a diagram illustrating nine error types classified according to the classification system.
[0020]
As shown in FIG. 4, in the classification system according to the present embodiment, first, in step 1 (hereinafter, “step” is referred to as “S”), “attention” (D), “attention deviating (D)”, and “sustained attention decaying (V)”. Assuming that, as the correct course, attention shifts from the first schema (previous schema) to the second schema (correct schema), the “delay of attention shift” means that the shift from the first schema to the second schema This means that the shift to attention has been made, but the transition has been delayed. Further, "attention deviating" means that attention to the second schema has been deviated to another (for example, an incorrect schema or the like). Further, “persistent attention decay” means that attention to the second schema has attenuated.
[0021]
Next, in S2, classification is performed based on the presence or absence of an incorrect schema. In other words, the classification is made based on whether or not there is a schema in which attention has been paid and an erroneous action has been performed instead of the correct schema (second schema). If the erroneous schema is “none”, it means that the erroneous act has not been executed by the erroneous schema, and the correct act by the correct schema has simply not been executed.
[0022]
Also, in the case of the above-mentioned "attention switching delay", the conversion from the first schema to the second schema (correct schema) is performed with a delay, that is, attention is directed to the first schema. Therefore, the first schema always exists as an erroneous schema, and is “Yes”. The cases of "attention deviating" and "persistent attention decaying" are classified as "present" or "absent (NN)" depending on the presence or absence of an incorrect schema. In the case of "none", the subsequent classification is not performed, and in the case of "distraction of attention (D)"-"none (NN)", the error type of "(5) DNN", "attenuation of persistent attention" If “(V)” − “none (NN)”, the error type is classified as “(9) VNN”.
[0023]
Next, in S3, the association between the erroneous schema and the stimulus that have been paid attention instead of the correct schema (second schema) is classified according to the enhanced factor. The reason is that "immediate activity (P)" means that the attention was immediately given to the wrong schema, and "high experience frequency (") means that attention is frequently given to the wrong schema. F)). In the case of the "attention switching delay", the attention is immediately focused on the first schema, which is an erroneous schema, and therefore all become "immediately active (P)".
[0024]
Next, in S4, classification is performed according to the type of the erroneous schema. In S3, if the schema is classified as "last active", the erroneous schema is the schema to which attention has been immediately focused, and therefore all are classified as "last active schema (P)". However, in the case of “attention switching delay” − “presence” − “previous activation”, the incorrect schema is the first schema to which attention was paid before, and is not called the immediately preceding activation schema. , "The previous schema (A)". In the case of "attention switching delay (S)"-"present"-"previous activity (P)"-"prior schema (A)", the error type of "(1) SPA" and "attention" In the case of “deviation (D)” − “present” − “previous activity (P)” − “previous activity schema (P)”, an error type of “(4) DPP” and “attenuation of persistent attention (V ))-"Yes"-"Previous Activity (P)"-"Previous Activity Schema (P)", it is classified as an error type of "(8) VPP".
[0025]
Further, in S3, when the classification is classified as “high experience frequency (F)”, the “normal schema rate high schema (G)”, which is normally a correct schema with a high probability, or an efficient schema It is classified into two types of a certain “efficient erroneous schema (B)”.
[0026]
Therefore, in the case of "attention deviating (D)"-"present"-"high experience frequency (F)"-"positive schema rate high schema (G)", the error type of "(2) DFG", " In the case of "attention deviating (D)"-"present"-"high experience frequency (F)"-"efficient erroneous schema (B)", the error type of "(3) DFB" In the case of "decay (V)"-"Yes"-"high experience frequency (F)"-"positive schema rate high schema (G)", the error type of "(6) VFG", "decay of continuous attention" In the case of “(V)” − “Yes” − “High experience frequency (F)” − “Efficient erroneous schema (B)”, the error type is classified as “(7) VFB”.
[0027]
Subsequently, each error type will be described in further detail with reference to FIG.
[0028]
As shown in the figure, first, (1) SPA is an error of the type "switching is delayed when it is necessary to rapidly switch from a certain task or decision to another task or decision". For example, regarding an operation of an aircraft, a human error that a switching of an operation pattern is delayed in a situation where an urgent operation pattern needs to be changed due to a sudden change in the state of the aircraft corresponds to (1) SPA.
[0029]
(2) The DFG is an error of the type “when correct attention to a correct operation or judgment is distracted by another, normally the correct operation or judgment interrupts”. For example, in a case where train operation is delayed, the braking force is weakened (unusual) due to rainfall or a large number of passengers, so the brake must be applied at a different timing than usual. Human error, such as that attention was deviated due to the delay of the train and the brakes were applied at normal timing and could not be stopped completely, corresponds to (2) DFG.
[0030]
(3) The DFB is an error of the type "an efficient erroneous work or judgment is interrupted when attention is paid to a correct work or judgment". For example, in a case where a train operation delay has occurred, a human being who has to confirm that the departure signal is green at the time of train departure, but has lost attention due to the train delay and did not check the signal The error corresponds to (3) DFB. Such errors occur because the starting signal is mostly blue and it is more efficient not to confirm.
[0031]
(4) The DPP is an error of the type "when the attention to the correct work or judgment is deviated to another, the immediately preceding work or judgment is interrupted". For example, in a case where the operation of a train driven by a driver who has driven an express on the previous route has been delayed, the driver's attention has been deviated to the train delay, and the stop pattern during the express operation has been interrupted. A human error such as passing a semi-stop station corresponds to (4) DPP.
[0032]
(5) DNN is an error of the type "when correct attention is paid to correct work or judgment, correct work or judgment is not performed". For example, in the case where a train is stuck between stations due to a failure and a rescue train for towing is requested, even if the failure has been repaired, wait at the site until the rescue train arrives or notify the command post that the repair has been made A human error, such as a need to move the train because the user is distracted by the delay of the train due to a failure and forgets the task of waiting at the site or contacting the command center, and moving the train, falls under (5) DNN.
[0033]
(6) VFG is an error of the type "when the continuous attention from the correct work or judgment attenuates, the correct work or judgment normally interrupts". For example, if a slow section is provided due to track construction, the driver must drive while searching for a slow signal and its warning sign, that is, paying continuous attention to the warning sign. (6) VFG caused a human error, such as the driver was distracted and the vehicle was driven according to the normal driving pattern without paying attention to the warning sign. Applicable.
[0034]
(7) VFB is an error of the type "Efficient work or judgment interrupts when continuous attention from correct work or judgment attenuates". For example, in the case of a train garage entry operation, in the case where the exchange signal signal has failed, the driver puts the garage in the back while checking the worker's exchange signal by looking out the window, and from the worker You have to pay attention to the change signal continuously until the stop signal is issued, but since you know the position of the stop roughly, the continuous attention is distracted and the efficient A human error, such as performing a complicated operation and delaying the application of a brake when a stop signal is sent from a worker, corresponds to (7) VFB.
[0035]
(8) VPP is an error of the type “when the last attention is attenuated from the correct work or judgment, the immediately preceding work or judgment is interrupted”. For example, a train driver must continue to pay attention to the next stop, but the driver loses constant attention, and the pattern of the express stop on the previous route comes to mind. Therefore, a human error such as passing through a stop station corresponds to (8) VPP.
[0036]
(9) VNN is an error of the type "correct work or judgment is not performed when continuous attention from correct work or judgment is attenuated". For example, when a rescue train is requested due to a train failure between stations, the operator of the rescue train must drive while continuously paying attention to the failed train, but the attention becomes distracted and the failed train Human error, such as forgetting to do so and delaying the discovery of a faulty train and applying the brakes, could be considered.
[0037]
The error types classified based on the cognitive occurrence mechanism have been described above. However, the human error diagnosis system 1 according to the first embodiment has the above-mentioned (2) DFG type errors (hereinafter, referred to as “DFG errors”). ) Is a system for diagnosing the likelihood of occurrence. Hereinafter, the diagnosis process will be described in detail.
[0038]
As a task for diagnosing the possibility of a DFG error occurrence according to the present embodiment, a “Stroop task due to a double task” is used. The Stroop problem is a problem that utilizes the Stroop effect, which is a phenomenon in which when a color name is written in a character having a color different from the color, the name of the character color is delayed. In the present embodiment, a numerical recitation task is added to this Stroop task to form a dual task. The Stroop effect is described in detail in Hiroyuki Shimada, "Stroop Effect-An Approach from Cognitive Psychology" (Haifukan, 1994/11).
[0039]
The dual task Stroop task consists of three components: a four digit number, a Stroop stimulus, and a Stroop option. The 4-digit numerical value is a 4-digit numerical value in which numerical values from 0 to 9 are arranged at random, and is presented to the subject by voice from a speaker of the PC main body. The arithmetic unit 11 generates a random 4-digit numerical value.
[0040]
In addition, there are two types of Stroop stimulus: a stimulus with interference and a stimulus without interference, and one of them is presented at the center of the display in a random order. The interference-stimulated stimulus is a word (for example, a hiragana word) indicating a color name colored with a color different from the color name is displayed on a display and presented to the subject. In the present embodiment, the word "red" colored red, the word "red" colored red, the word "red" colored blue, the word "blue" colored blue, and the yellow colored Either the word "Aka" or the word "Ao" colored yellow is presented. The non-interference stimulus is presented with one of the symbol “++” colored red, the symbol “++” colored blue, and the symbol “++” colored yellow. In addition, the Stroop option is one in which three words, "red", "blue", and "blue", which are colored in black, are arranged in random order and presented at the bottom of the display.
[0041]
Next, the configuration content of the Stroop task by the dual task will be described. The task according to the present embodiment includes 60 trials, and the above-described Stroop stimulus with and without interference is presented 60 times in a random order.
[0042]
FIG. 6 is a diagram showing a breakdown of 60 trials. As shown in the figure, 30 stimuli with interference and 30 stimuli without interference are tried. In addition, the stimuli with interference include the word “blue” colored red, the word “blue” colored red, the word “red” colored blue, the word “blue” colored blue, and the color yellow Each of the word “Aka” and the word “Ao” colored yellow is tried five times. In the non-interference stimulation, the symbol “++” colored in red, the symbol “++” colored in blue, and the symbol “++” colored in yellow are tried 10 times each.
[0043]
Next, the procedure of each trial will be described with reference to FIG. FIG. 7 is a flowchart showing the procedure of the n-th trial. As shown in the drawing, when the n-th trial is started (S10), first, in S11, a random 4-digit numerical value is presented to the subject from the speaker of the PC main body 10. This four-digit numerical value is temporarily stored in the storage device 12.
[0044]
One second after S11 ends, the Stroop stimulus and the Stroop option are presented on the display 20 (S12). FIG. 8 is a diagram illustrating an example of a screen for presenting a Stroop stimulus and a Stroop option. The center of the figure is the Stroop stimulus, and in this example, the word “Ao” 100 colored in red, which is the Stroop stimulus with interference, is presented. The lower part is a Stroop option, in which the words "Aka" 101, "Blue" 102, and "Ao" 103, which are colored in black, are presented side by side in this order. In this example, since the Stroop stimulus is a character colored in red, selecting the option “Aka” 101 gives a correct answer. At the same time as the start of S12, the timer 13 starts.
[0045]
Next, when the person to be diagnosed responds (answers the color of the character) (S13), the timer 13 is stopped, and information on the reaction of the person to be diagnosed is recorded in the storage device 12. The response of the subject is made by pressing the dedicated key 40 corresponding to the option representing the color of the Stroop stimulus. FIG. 9 is a diagram showing information on a reaction recorded in the storage device 12. In S13, for each trial, the type of Stroop stimulus (0: no interference, 1: interference), the correctness of the Stroop reaction indicating whether the subject has selected the correct option (0: wrong, 1: Correct), the Stroop reaction time (ms) measured by the timer 13 is recorded. As shown in the column of trial 5 in FIG. 9, when 2 seconds have elapsed without a response from the subject, the correctness of the Stroop reaction for that trial is not recorded, and the Stroop reaction time is assumed to be 2000 ms. Are recorded in the storage device 12.
[0046]
Subsequently, in S14, the numerical value input screen is presented on the display 20, and the timer 13 is started again. The person to be diagnosed inputs the four-digit numerical value presented in S11 via the keyboard 30 by relying on his / her memory (S15). Here, the arithmetic unit 11 determines whether or not the four-digit numerical value temporarily recorded in the storage device 12 in S11 matches the four-digit numerical value input by the subject in S15. The correctness is further recorded in the storage device 12 (see FIG. 9).
[0047]
If three seconds have elapsed after the numerical value input screen was presented in S14 without any input by the subject, in S15, the input numerical value for the trial is automatically determined to be "wrong." Are recorded in the storage device 12. When the correctness of the input numerical value is recorded in the storage device 12 in this way, the timer 13 is also stopped, and the n-th trial ends (S16). Then, when the n-th trial is completed, the (n + 1) -th trial is started immediately, and the above-described procedure is repeated until all the 60 trials are completed. Note that the above-described procedure is executed by the arithmetic device 11 controlling the storage device 12, the timer 13, the display 20, and the like, and the information input by the subject from the keyboard 30 and the dedicated key 40 is stored in the storage device 12. Is also executed under the control of the arithmetic unit 11.
[0048]
Next, a process in which the human error diagnosis system 1 according to the present embodiment diagnoses the possibility of the occurrence of a DFG error from the information (FIG. 9) obtained by the Stroop task will be described in detail. In this process, the arithmetic unit 11 calculates the values of the Stroop error response rate and the Stroop interference amount from the information recorded in the storage device 12, and diagnoses the possibility of occurrence based on these values. .
[0049]
A method for calculating the Stroop false reaction rate will be described. The calculation of the Stroop error response rate targets only trials related to the Stroop stimulus with interference. First, the number of trials in which the correctness of the input numerical value is “false” (the trial in which the numerical input correctness is “0” in FIG. 9) is subtracted from 30 trials related to the Stroop stimulus with interference, and the “corrected interference trial number” is calculated. Ask. Next, among the 30 trials related to the Stroop stimulus with interference, the trial of the correctness of the input numerical value is “correct” and the trial of the Stroop reaction is “false” (in FIG. 9, the trial in which the true / false of the Stroop reaction is “0”). ) Is added to the number of trials in which the correctness of the input numerical value is “correct” and the Stroop reaction time is 2000, thereby obtaining the “number of incorrect responses with corrected interference”. The Stroop erroneous reaction rate is obtained by dividing the number of erroneous reactions with corrected interference by the number of trials with corrected interference, and is obtained by the following equation 1.
(Stroop false reaction rate) = (number of false reactions with corrected interference) / (number of trials with corrected interference) (1)
[0050]
Next, a method of calculating the Stroop interference amount will be described. First, out of the 60 trials, trials in which the numerical input is an erroneous response (in FIG. 9, numerical input correctness is “0”), trials in which the Stroop reaction is an erroneous response (in FIG. 9, "0"), and runs with a Stroop reaction time of 2000 ms are excluded. Then, regarding the Stroop reaction time of the remaining trials, the “average reaction time” and “standard deviation” are obtained, and the Stroop reaction time is equal to or less than “(average reaction time) − (2 × standard deviation)” and “( Trials greater than or equal to (mean reaction time) + (2 × standard deviation) are further excluded.
[0051]
Subsequently, among the remaining trials, the average value of the Stroop reaction time of the trial relating to the Stroop stimulus without interference is obtained, and is defined as “average reaction time without interference”. In addition, among the remaining trials, the average value of the Stroop reaction times of the trials related to the Stroop stimulus with the interference is obtained, and is defined as “average reaction time with the interference”. The Stroop interference amount is obtained by subtracting the average reaction time without interference from the average reaction time with interference, and is obtained by the following equation 2.
(Stroop interference amount) = (Average reaction time with interference) − (Average reaction time without interference) (2)
[0052]
Next, a criterion for diagnosing the possibility of occurrence of a DFG error from the thus determined diagnostic index of the Stroop error response rate and the Stroop interference amount will be described. FIG. 10 is a diagram showing this diagnostic criterion. First, with regard to the Stroop error response rate, a person whose “Stroop error response rate ≧ 0.2” is diagnosed as a person who has a high possibility of generating a DFG error. Regarding the Stroop interference amount, a person with “Stroop interference amount ≧ 80” is diagnosed as a person who is likely to generate a DFG error, and a person with “50 ≦ Stroop interference amount <80” indicates a DFG error. Is diagnosed as a person having a medium probability of occurrence, and a person with “Stroop interference amount <50” is diagnosed as a person having a low possibility of generating a DFG error. Note that these thresholds used for diagnosis can be appropriately changed according to conditions and the like.
[0053]
As for the final diagnosis result, the diagnosis result based on the Stroop erroneous reaction rate is compared with the diagnosis result based on the Stroop interference amount, and the worse diagnosis result is used as the final diagnosis result. This is to make a diagnosis on the safe side. For example, when the diagnosis result based on the Stroop error response rate is “high possibility” and the diagnosis result based on the Stroop interference amount is “small possibility”, the final diagnosis result is “high possibility of DFG error”. .
[0054]
Here, the reason why the possibility of occurrence of a DFG error can be diagnosed by the Stroop task by the dual task will be described in detail. As described with reference to FIGS. 4 and 5, a DFG error is a situation in which, in an unusual situation, work or judgment must be performed based on the schema to be executed this time. It is an error that when deviated, the work and judgment based on the inadvertently habitual schema are interrupted. If it is the same situation as usual, it is better to perform the work and judgment based on the habitual schema that is always done, but this time it is a different situation, so based on the schema to be executed this time Work and judgment must be done. Of course, the unusual situation is not completely different, but is similar enough to inadvertently retrieve the habitual schema that you do all the time.
[0055]
In such a situation, the habitual schema that is regularly used is easier to retrieve than the schema to be executed this time. This is because human information processing is habitual based on the frequency of experience, and the habitual schema is easily unintentionally extracted without paying attention. In this way, even in unusual situations, usually habitual schema will be extracted, but this makes it impossible to perform correct work and judgment in unusual situations Become. In fact, people can be more flexible, even in unusual situations, and usually do the right thing and make decisions by extracting the schema to be done this time.
[0056]
The caution function enables such a flexible response as described in FIG. FIG. 11 is a diagram illustrating a model for explaining the occurrence of a DFG error. FIG. 11A shows a model of “when an action is performed correctly” in a different situation, and FIG. 11B shows a model of “when an incorrect action is performed” in a different situation. ing.
[0057]
As shown in FIG. 11 (a), if attention is directed to the correct schema to be executed this time, a habitual erroneous schema is not stimulated, and a correct action based on the correct schema is executed. Will be. On the other hand, when attention is deprived of other factors such as an unexpected situation of the external environment and internal factors such as troubles of the next job, as shown in FIG. Attention deviates from As a result, as shown in the figure, a habitual erroneous schema is stimulated, and an erroneous action based on this is executed.
[0058]
Further, in the Stroop task according to the dual task according to the present embodiment, the phenomenon such as the interruption of the habitual schema in the DFG error is presented by a character having a color different from the color that the color name word means. In such a case, in the Stroop task of answering the color of the character, the phenomenon that the meaning of the color name word (customized schema) is interrupted corresponds. In addition, a numerical recitation task added to the Stroop task corresponds to other matters that are distracted. The DFG error that occurs in the Stroop task with interference is to answer the character color name of the word in a situation where the character color name of the word must be answered instead of the meaning of the color name word (unusual situation). The color name word based on the habitual schema (wrong schema) that inadvertently executes the answer of the meaning of the color name word where the answer (correct action) must be made based on the schema (correct schema) It is thought that this is caused by answering (incorrect action) the meaning of.
[0059]
Thus, a Stroop erroneous reaction with interference that occurs in a Stroop task due to a dual task can be considered to be a DFG error itself. Therefore, the Stroop erroneous reaction rate (Equation 1) according to the first embodiment is an index that directly indicates the possibility of a DFG error.
[0060]
However, false reactions in the Stroop task generally occur very rarely. For this reason, it is difficult to accurately diagnose the possibility of a potential human error using only the Stroop error response rate. Therefore, in the present embodiment, the diagnosis is performed in combination with the amount of Stroop interference (Equation 2) in which a similar mechanism is assumed. The Stroop interference amount is the difference between the average reaction time with interference and the average reaction time without interference in the Stroop task, but the possibility of causing a DFG error depends on the answer of the color of the color name word (ie, correct action). This is because it is considered to appear as such a time delay. That is, the amount of Stroop interference according to the first embodiment is an index that indirectly indicates the possibility of occurrence of a DFG error. In general, a human response to a non-interference Stroop stimulus is about 500 ms, and a Stroop interference amount is about several tens ms.
[0061]
Here, the functions of the human error diagnosis system 1 according to the present embodiment are realized by installing a program for realizing these functions in the PC main body 10.
As described above in detail, according to the human error diagnosis system according to the present embodiment, it is possible to accurately diagnose the possibility of occurrence of a DFG error classified based on a model of a cognitive occurrence mechanism. In addition, based on the diagnosis result, it is possible to alert a person who is likely to have a DFG error to a DFG error and to carry out a DFG error prevention training, thereby effectively preventing the occurrence of a human error. can do. In addition, it is possible to effectively prevent the occurrence of human error by allocating human resources in consideration of the possibility of occurrence of DFG errors.
[0062]
(Second embodiment)
Next, a second embodiment of the present invention will be described. Since the human error diagnosis system according to the second embodiment has substantially the same configuration as the human error diagnosis system according to the first embodiment, description of the same configuration will be omitted, and only different configurations will be described. This will be described in detail.
[0063]
The human error diagnosis system 1 according to the second embodiment is a system for diagnosing the possibility of occurrence of the above-mentioned (6) VFG type error (hereinafter, referred to as “VFG error”). Is very different. For this reason, in the present embodiment, a “Stroop problem due to a monitoring problem” is used instead of the “Stroop problem due to a double problem” in the first embodiment as a problem for diagnosing the possibility of occurrence of a VFG error. Diagnosis.
[0064]
The Stroop task by the monitoring task is composed of three components: filler stimulus, Stroop stimulus, and Stroop option. The Stroop stimulus and the Stroop option are the same as those in the first embodiment. The filler stimulus is one in which a black star “★★” is presented in the center of the gray display screen.
[0065]
Next, the configuration content of the Stroop task by the monitoring task will be described. The Stroop task according to the present embodiment also includes 60 trials as in the first embodiment, and the breakdown is the same as the breakdown shown in FIG. However, the procedure in each trial is different from the procedure of the first embodiment.
[0066]
FIG. 12 is a flowchart showing the procedure of the n-th trial according to the present embodiment. As shown in the drawing, when the n-th trial is started (S20), first, in S21, a black star “★★” as the filler stimulus 1 is displayed on the gray display screen. Presented in the center for 1 second. Subsequently, one second after S21, the filler stimulus 2 is similarly presented for one second (S22). This presentation of the filler stimulus is similarly repeated m times (S24). The number m is randomly set for each trial so that the minimum number of trials is 20 times, the maximum number of trials is 40 times, and the total number of trials is 1800 times (30 times on average).
[0067]
One second after the m-th filler stimulus is presented in S24, the Stroop stimulus and the Stroop option are presented on the display 20 (S25). This S25 is the same as S12 of the first embodiment. Subsequently, in S26, when the person to be diagnosed responds and the processing corresponding thereto is executed by the arithmetic unit 11, the n-th trial ends after one second (S27). In this S26, similarly to S13 of the first embodiment, if 2 seconds have elapsed without any response from the subject, the correctness of the Stroop reaction is not recorded for the trial, and the Stroop reaction time is set to 2000 ms. Recorded in the storage device 12.
[0068]
FIG. 13 is a diagram illustrating information recorded in a storage device according to the second embodiment. As shown in the figure, in the second embodiment, since there is no step corresponding to the numerical value input (S15) of the first embodiment, the correctness of the numerical value input is determined from the items of FIG. 9 according to the first embodiment. Excluded, the type of Stroop stimulus, correctness of Stroop response and Stroop response time are recorded.
[0069]
Also in the second embodiment, when the n-th trial is completed, the (n + 1) -th trial is started immediately, and the procedure shown in FIG. 12 is repeated until the 60 trials are completed. Note that the above-described procedure is executed by the arithmetic device 11 controlling the storage device 12, the timer 13, the display 20, and the like, and the information input by the subject from the keyboard 30 and the dedicated key 40 is stored in the storage device 12. Is also executed under the control of the arithmetic unit 11.
[0070]
Next, a process in which the human error diagnosis system 1 according to the present embodiment diagnoses the possibility of occurrence of a VFG error from the information (FIG. 13) obtained by the Stroop task will be described in detail. In this process, the arithmetic unit 11 calculates the values of the Stroop error response rate and the Stroop interference amount from the information recorded in the storage device 12, and diagnoses the possibility of occurrence based on these values. .
[0071]
First, the method of calculating the Stroop error response rate is the same as that in the first embodiment, and is calculated by the following equation 1.
(Stroop false reaction rate) = (number of false reactions with corrected interference) / (number of trials with corrected interference) (1)
[0072]
Next, a method for calculating the amount of Stroop interference will be described. In the second embodiment, the calculation method is slightly different from that of the first embodiment because there is no right or wrong input numerical value. First, out of the 60 trials, trials in which the Stroop reaction is an erroneous reaction (in FIG. 13, the Stroop reaction correct / incorrect is “0”) and trials in which the Stroop reaction time is 2000 ms are excluded. The subsequent processing is the same as that of the first embodiment. For the Stroop reaction time of the remaining trials, the “average reaction time” and “standard deviation” are obtained, and the Stroop reaction time is calculated as “(average reaction time) − ( Trials less than or equal to “2 × standard deviation)” and trials with “(average reaction time) + (2 × standard deviation)” or more are further excluded.
[0073]
Subsequently, among the remaining trials, the average value of the Stroop reaction time of the trial relating to the Stroop stimulus without interference is obtained, and is defined as “average reaction time without interference”. In addition, among the remaining trials, the average value of the Stroop reaction times of the trials related to the Stroop stimulus with the interference is obtained, and is defined as “average reaction time with the interference”. The Stroop interference amount is obtained by subtracting the average reaction time without interference from the average reaction time with interference, and is obtained by the following equation 2.
(Stroop interference amount) = (Average reaction time with interference) − (Average reaction time without interference) (2)
[0074]
Next, a criterion for diagnosing the possibility of a VFG error from the diagnostic index of the Stroop erroneous reaction rate and the Stroop interference amount thus obtained will be described. The diagnostic criterion in the present embodiment is the same as the diagnostic criterion according to the first embodiment shown in FIG. Therefore, regarding the Stroop false reaction rate, a person with “Stroop false reaction rate ≧ 0.2” is diagnosed as a person who has a high possibility of generating a VFG error. Regarding the Stroop interference amount, a person with “Stroop interference amount ≧ 80” is diagnosed as having a high possibility of generating a VFG error, and a person with “50 ≦ Stroop interference amount <80” indicates a VFG error. Is diagnosed as a person having a medium probability of occurrence, and a person with “Stroop interference amount <50” is diagnosed as a person having a low possibility of generating a VFG error.
[0075]
Also, regarding the final diagnosis result, the diagnosis result based on the Stroop erroneous reaction rate and the diagnosis result based on the Stroop interference amount are compared, and the worse diagnosis result is used as the final diagnosis result. For example, when the diagnosis result based on the Stroop error response rate is “high possibility” and the diagnosis result based on the Stroop interference amount is “small possibility”, the final diagnosis result is “VFG error possibility”. .
[0076]
Next, the reason why the possibility of occurrence of the VFG error can be diagnosed by the Stroop task by the monitoring task will be described in detail. As described with reference to FIGS. 4 and 5, a VFG error is a situation that is unusual and requires attention for a certain period of time. This is an error in which the work or the judgment based on the schema that has been inadvertently habitually interrupted when the place where the judgment or judgment had to be made was blurred.
[0077]
That is, the mechanism of generating the VFG error is almost the same as the mechanism of generating the DFG error in the first embodiment. However, in the case of the DFG error, the attention to the correct schema is attenuated because the attention is deviated to other matters. In contrast, VFG errors differ in that persistent attention to the correct schema is attenuated due to blurring.
[0078]
FIG. 14 is a diagram illustrating a model for explaining the occurrence of a VFG error. FIG. 14A shows a model of “when an action is performed correctly” in a different situation, and FIG. 14B shows a model of “when a wrong action is performed” in a different situation. ing.
[0079]
As shown in FIG. 14 (a), if attention is directed to the correct schema to be executed this time, a habitual erroneous schema is not stimulated, and a correct action based on the correct schema is executed. Will be. On the other hand, if it is blurred, attention to the correct schema to be executed this time is attenuated, as shown in FIG. As a result, as shown in the figure, a habitual erroneous schema is stimulated, and an erroneous action based on this is executed.
[0080]
In the Stroop task according to the monitoring task according to the present embodiment, the phenomenon (customized schema) of the color name word interrupts the phenomenon such as the interruption of the habitual schema in the VFG error in the Stroop task. Is supported. In addition, the phenomenon in which attention is attenuated due to blurring is realized by a monitoring task of continuously waiting for a Stroop stimulus in a situation where an irregular number of filler stimuli is presented.
[0081]
As described above, similarly to the first embodiment, the Stroop erroneous reaction with interference that occurs in the Stroop task due to the monitoring task is considered to be the VFG error itself. That is, the Stroop error response rate (Equation 1) according to the second embodiment is an index that directly indicates the possibility of occurrence of a VFG error.
[0082]
Regarding the amount of Stroop interference (Equation 2), as in the first embodiment, the possibility of causing a VFG error is considered to appear as a delay in the time required to answer the color of the character of the color name word. Therefore, the Stroop interference amount according to the second embodiment is an index that indirectly indicates the possibility of a VFG error.
[0083]
Here, the functions of the human error diagnosis system 1 according to the present embodiment are realized by installing a program for realizing these functions in the PC main body 10.
[0084]
As described above in detail, according to the human error diagnosis system according to the present embodiment, it is possible to accurately diagnose the possibility of occurrence of a VFG error classified based on a model of a cognitive occurrence mechanism. In addition, based on the diagnosis result, it is possible to alert an individual who is likely to have a VFG error to a VFG error and to carry out a VFG error prevention training, thereby effectively preventing the occurrence of a human error. can do. In addition, it is possible to effectively prevent the occurrence of a human error by allocating personnel in consideration of the possibility of a VFG error.
[0085]
As described above, the present invention has been described in detail based on the first and second embodiments. However, the embodiments of the present invention are not limited to these, and various modifications may be made without departing from the gist of the present invention. Is possible.
[0086]
For example, the number of trials and the like in the first and second embodiments can be changed as appropriate, and the more the number of trials, the more accurate diagnosis can be made.
[0087]
Further, in the first and second embodiments, the possibility of occurrence of the DFG error and the VFG error classified based on the error classification system shown in FIGS. 4 and 5 is diagnosed. The target error types are not limited to these two error types. That is, according to the Stroop task according to the first and second embodiments, it is possible to diagnose the possibility of occurrence of an error type such as a habitual schema which is always performed in a different situation than usual. It is possible to diagnose not only the possibility of occurrence of an error type classified based on the classification system of FIGS. 4 and 5 but also the possibility of occurrence of a human error such as a customary schema interrupt. Then, by adding a “double task” or a “monitoring task” to the Stroop task in accordance with the error type, it is possible to diagnose the possibility of occurrence in accordance with the error type.
[0088]
【The invention's effect】
According to the present invention, it is possible to provide a human error diagnostic system capable of accurately diagnosing the possibility of a human error in order to effectively prevent the occurrence of a human error.
[Brief description of the drawings]
FIG. 1 is a diagram showing a schematic configuration of a human error diagnosis system according to a first embodiment of the present invention.
FIG. 2 is a diagram illustrating a range of a human error targeted by the human error diagnosis system according to the first embodiment of the present invention.
FIG. 3 is a diagram illustrating a model for explaining occurrence of a good definition error according to the first embodiment of the present invention.
FIG. 4 is a diagram illustrating a human error classification system based on a model of a cognitive generation mechanism according to the first embodiment of the present invention.
FIG. 5 is a diagram illustrating details of an error type according to the first embodiment of the present invention.
FIG. 6 is a diagram illustrating a configuration of a Stroop task according to a dual task according to the first embodiment of the present invention.
FIG. 7 is a flowchart illustrating a trial procedure according to the first embodiment of the present invention.
FIG. 8 is a diagram illustrating an example of a screen for presenting a Stroop stimulus and a Stroop option according to the first embodiment of the present invention.
FIG. 9 is a diagram showing information recorded in a storage device according to the first embodiment of the present invention.
FIG. 10 is a diagram showing diagnostic criteria according to the first embodiment of the present invention.
FIG. 11 is a diagram illustrating a model for explaining occurrence of a DFG error according to the first embodiment of the present invention.
FIG. 12 is a flowchart illustrating a trial procedure according to the second embodiment of the present invention.
FIG. 13 is a diagram showing information recorded in a storage device according to a second embodiment of the present invention.
FIG. 14 is a diagram showing a model for explaining occurrence of a VFG error according to the embodiment of the present invention.
[Explanation of symbols]
1 Human error diagnosis system
10 PC body
11 arithmetic unit
12 Storage device
13 Timer
20 Display
30 keyboard
40 Dedicated key

Claims (9)

In a human error diagnosis system that causes a subject to execute a predetermined task and diagnoses the possibility of a human error,
An arithmetic device for performing various calculations; an input device for the subject to input a response to the task; and storage means for storing information for executing the task and information on the response of the subject to the task. And a display device for presenting the subject to the subject,
The storage device may include, as the predetermined task, an interference-free stimulus in a situation where there is no interruption due to a habitual schema and an interference where a human error is likely to occur due to the interruption of the habitual schema. Holds information about the stimulus,
The arithmetic unit presents the interference stimulus and the non-interference stimulus to the display device in a random order each of a predetermined number of times, and the response information to the interference stimulus held in the storage device and the interference-free stimulus. A human error diagnosis system for diagnosing the possibility of occurrence of a human error of a type interrupted by a habitual schema based on information on a reaction to a stimulus. In a human error diagnosis system that allows a subject to execute a Stroop task to diagnose the possibility of a human error,
An arithmetic device that performs various arithmetic processes; an input device for the subject to input a response to the Stroop task; information for executing the Stroop task; and information about the response of the subject to the Stroop task. A storage device for holding, and a display device for presenting the Stroop task to the subject,
The storage device, for the Stroop task, holds information on Stroop stimulation with interference and Stroop stimulation without interference,
The computing device presents the interference-based Stroop stimulus and the interference-free Stroop stimulus to the display device in a random order each of a predetermined number of times, and response information to the interference-based Stroop stimulus held in the storage device. A human error diagnosing system for diagnosing the possibility of occurrence of a human error of a type interrupted by a habitual schema, based on reaction information to the Stroop stimulus without interference. The arithmetic unit presents a color name word colored in a color different from the color name on the display device as the interference-based Stroop stimulus, and the character or the color painted in a certain color as the interference-free Stroop stimulus. Configured to present a symbol on the display device;
Further, the arithmetic unit records, in the storage device, information as to whether or not the name of the color coloring the color name word is correctly answered as the response information to the interference-induced Stroop stimulus, 3. The human error diagnosis system according to claim 2, wherein information as to whether or not the name of the color coloring the character or the symbol is correctly answered is recorded in the storage device as response information to the information. The human error according to claim 2, wherein the arithmetic device calculates a difference between a time required for a correct response to the Stroop stimulus with interference and a time required for a response to the Stroop stimulus without interference. Diagnostic system. The storage device retains information on yet another task, and the arithmetic device presents the another task to the subject together with presenting the Stroop stimulus, and responds to each of the Stroop stimuli. The human error diagnosis system according to any one of claims 2 to 4, wherein the information includes correct / wrong information on the another task, and diagnoses a possibility of occurrence of a human error by a Stroop task due to a dual task. . The human error diagnosis system according to claim 5, wherein the arithmetic unit presents, as the another task, a recitation task for distracting the subject's attention from the Stroop task. An arithmetic device, a storage device, a human error diagnostic system including an input device and a display device, a human error diagnostic method for performing a Stroop task for a person to be diagnosed and diagnosing the possibility of occurrence of a human error,
The computing device presents the display device with the interference Stroop stimulus and the interference-free Stroop stimulus a plurality of times in random order on the display device based on the information on the Stroop stimulus with interference and the Stroop stimulus without interference stored in the storage device. Presenting step,
A storage step of storing, in the storage device, response information input by the subject to the Stroop stimulus via the input device,
A diagnosis step of diagnosing the possibility of occurrence of a human error of a type interrupted by a habitualized schema based on the reaction information. The presenting step, when presenting each of the Stroop stimuli, is a step of presenting together a recitation task to divert the attention of the subject from the Stroop task,
The human error diagnosis method according to claim 7, wherein the storing step is a step of storing correct / incorrect information for the recitation task. A human error diagnosis program for causing a computer including an arithmetic device, a storage device, an input device, and a display device to execute a Stroop task for a subject to perform a task of diagnosing a possibility of occurrence of a human error. ,
The computing device presents the display device with the interference Stroop stimulus and the interference-free Stroop stimulus a plurality of times in random order on the display device based on the information on the Stroop stimulus with interference and the Stroop stimulus without interference stored in the storage device. Steps to
Holding the response information input by the subject to the Stroop stimulus via the input device in the storage device,
Diagnosing the possibility of occurrence of a human error of a type interrupted by a habitual schema based on the reaction information, based on the reaction information, causing the computer to execute the computer.

Images