Pre-shift briefing AI intelligent device

Abstract

The application relates to the technical field of safety management, and specifically discloses an AI intelligent device for pre-shift briefing. The system comprises: a behavior collection module which extracts a face visual line, a body posture, a group interaction feature and a weighted mapping into an attention distraction index by collecting a personnel video stream; an overlapping analysis module which identifies pre-shift distracted personnel based on the attention distraction index, and judges a trigger overlapping analysis signal in combination with environmental interference; a position optimization module which constructs a position adjustment model through a random forest algorithm, adjusts the position of the distracted personnel and checks the feasibility; a pre-shift evaluation module which calculates a distraction optimization coefficient before and after position adjustment, and judges information transmission effectiveness; and a pre-shift optimization module which optimizes a period, a time length and a content sequence of a key period based on personnel content mastery and a concentrated distraction period. The device is favorable for improving pre-shift briefing information transmission effectiveness, and extracting pre-shift safety and operation guidance for engineering construction.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of safety management, and particularly relates to a pre-shift briefing AI intelligent device. BACKGROUND

[0002] In the field of highway, municipal, building construction and other engineering construction, the pre-shift briefing means that the management personnel need to face the workers and explain the specific work tasks (such as the range of steel binding, the concrete pouring process) of the day, the potential risk points (such as high-altitude falling protection, mechanical collision avoidance, temporary power specification) in the field, and the safety protection requirements and emergency disposal process.

[0003] However, the traditional pre-shift briefing mode cannot guarantee effective information transmission: on the one hand, workers often have problems of inattention during the briefing process, such as frequently deviating from the briefing display screen, whispering with others, and walking around casually, but the management personnel can only make subjective judgments and cannot quantify the degree of distraction of the workers;

[0004] On the other hand, the core information of the briefing (such as the protection details of high-risk operations and the key nodes of the task) is concentrated in a specific period, and if the workers are inattentive during this period, it is easy to cause incomplete information reception and fuzzy memory; ultimately, these problems will directly affect the subsequent work link, that is, the workers may have irregular operations (such as not wearing protective equipment as required) and ignore risk hidden dangers due to not fully grasping the briefing precautions, thereby increasing the probability of work safety risks and quality problems.

[0005] Therefore, the present application provides a pre-shift briefing AI intelligent device. SUMMARY

[0006] The present application aims to provide a pre-shift briefing AI intelligent device to solve the above background problems.

[0007] The purpose of the present application can be achieved by the following technical solutions:

[0008] A pre-shift briefing AI intelligent device: comprising the following modules:

[0009] The behavior acquisition module is used for acquiring the behavior characteristics of the personnel during the pre-shift briefing, performing attention mapping processing on the behavior characteristics, and obtaining an attention distraction index;

[0010] The overlapping analysis module is used for performing attention distraction analysis on the personnel based on the attention distraction index, identifying the pre-shift distraction personnel, extracting the distraction period of the attention of the pre-shift distraction personnel, and the key period of the pre-shift briefing, performing time sequence overlapping analysis on the distraction period and the key period, obtaining a coincidence determination coefficient, and judging whether an overlapping analysis signal is triggered or not;

[0011] Position optimization module: if the overlap analysis signal is triggered, the pre-shift briefing position of the pre-shift distraction personnel is adjusted, and the attention distraction index of all personnel after position adjustment is obtained;

[0012] Pre-shift evaluation module: based on the adjusted attention distraction index, the attention of the pre-shift briefing personnel is optimized and evaluated to obtain a distraction optimization coefficient; based on the distraction optimization coefficient, the optimization effectiveness of the key period is analyzed to determine whether the information transmission effectiveness of the pre-shift briefing meets the standard;

[0013] Pre-shift optimization module: if it does not meet the standard, the period division characteristics of the pre-shift distraction personnel on the key period are obtained; a period division model is constructed, the period division characteristics are input into the period division model, and the key period is adjusted based on the pre-shift briefing adjustment content output by the model.

[0014] As a further scheme of the application: the manner of attention mapping processing of the behavior characteristics is:

[0015] Among them, the behavior characteristics include: facial line of sight characteristics, body posture characteristics and group interaction characteristics;

[0016] The facial line of sight characteristics, body posture characteristics and group interaction characteristics are weighted and summed to obtain the attention distraction index.

[0017] As a further scheme of the application: the manner of obtaining the facial line of sight characteristics is:

[0018] The effective visual range of each personnel in the pre-shift briefing is obtained; if the visual angle deviates from the effective visual range, the time length of the visual angle deviating from the effective visual range in the monitoring period is obtained to obtain the deviation time length;

[0019] The proportion of the deviation time length in the monitoring period time length is calculated to obtain the deviation time ratio, and the deviation time ratio is taken as the facial line of sight characteristics.

[0020] As a further scheme of the application: the manner of performing the time sequence overlap analysis is:

[0021] The distraction period of the attention of the pre-shift distraction personnel is extracted; if the distraction period and the key period coincide and overlap in the time dimension, the overlapping period is marked as an overlap period, and the number of overlap periods is obtained as the number of overlap periods;

[0022] The average distraction index of the individual in the overlap period of the pre-shift distraction personnel is calculated;

[0023] The number of key periods is obtained, the proportion of the number of overlap periods in the number of key periods is calculated to obtain the period overlap ratio;

[0024] The average distraction index and the period overlap ratio are multiplied to obtain the coincidence determination coefficient.

[0025] As a further scheme of the present application: the determination method of the pre-shift scattered personnel is:

[0026] Obtain the characteristic variables of all personnel before the shift briefing, and perform clustering processing on the characteristic variables of all personnel through a clustering algorithm;

[0027] Extract the personnel in the heavy scattered group in the clustering result and the corresponding attention scattering index, and if the attention scattering index of the personnel in the heavy scattered group shows an upward trend in N monitoring periods, mark the personnel as pre-shift scattered personnel.

[0028] As a further scheme of the present application: the method for obtaining the characteristic variables is:

[0029] Extract the number of periods in which the scattered periods appear continuously in N monitoring periods, and calculate the ratio of the number of periods in which the scattered periods appear continuously to the number N of monitoring periods to obtain the scattering frequency;

[0030] Calculate the mean of the attention scattering index of each individual in the number of periods in which the scattered periods appear continuously to obtain the individual average scattering index;

[0031] Obtain the mean of the individual average scattering index of all personnel at the time of the shift briefing to obtain the group average scattering index; calculate the deviation proportion of the individual average scattering index and the group average scattering index to obtain the scattering intensity deviation;

[0032] Take the scattering intensity deviation, the individual average scattering index, and the scattering frequency as the characteristic variables of each personnel.

[0033] As a further scheme of the present application: the method for optimizing and evaluating the attention of the personnel before the shift briefing is:

[0034] Obtain and calculate the change amount of the individual average scattering index of the single pre-shift scattered personnel before and after the position adjustment to obtain the scattering adjustment amount;

[0035] Calculate the ratio of the scattering adjustment amount to the individual average scattering index before the adjustment to obtain the scattering optimization coefficient of the overall period;

[0036] Obtain the scattering adjustment amount of the key period after the position adjustment, and calculate the ratio of the individual average scattering index before the adjustment to obtain the scattering optimization coefficient of the key period;

[0037] If the scattering optimization coefficient of the overall period and the scattering optimization coefficient of the key period are both positive, and the individual average scattering index in the key period is lower than the preset scattering boundary value, it indicates that the attention regulation of the single pre-shift scattered personnel is effective.

[0038] As a further scheme of the present application: the method for performing the optimization effectiveness analysis is:

[0039] Based on single effective personnel data, the key period is split to obtain a high-optimal period and a medium-optimal period;

[0040] Combined with the average dispersion index of the individual before adjustment, the high-optimal period and the medium-optimal period are obtained by establishing a period-by-period optimization evaluation process to obtain a high-optimal period overall optimization coefficient and a medium-optimal period overall optimization coefficient;

[0041] The associated operation violation rate in the high-optimal period and the overlap duration of the environmental event interference and the key period are obtained;

[0042] After excluding the overlap duration of the environmental event interference and the key period, the partial correlation interference check is performed based on the high-optimal period overall optimization coefficient and the associated operation violation rate in the high-optimal period to obtain a partial correlation coefficient;

[0043] The standard effectiveness criterion is constructed, and if the high-optimal period overall optimization coefficient and the medium-optimal period overall optimization coefficient and the partial correlation coefficient meet the standard effectiveness criterion, the information transmission effectiveness of the pre-shift briefing meets the standard.

[0044] As a further scheme of the present application, the partial correlation interference check is performed in the following manner:

[0045] For the high-optimal period overall optimization coefficient X, the high-optimal period associated operation violation rate Y, and the overlap duration Z of the environmental interference and the key period, the Pearson correlation coefficients of X and Y, X and Z, and Y and Z are calculated when the partial correlation coefficient is calculated after controlling Z;

[0046] The Pearson correlation coefficients of X and Y, X and Z, and Y and Z are substituted into the partial correlation coefficient formula to obtain the partial correlation coefficient.

[0047] As a further scheme of the present application, the key period is adjusted in the following manner:

[0048] The AI intelligent device is used to obtain the degree of mastery of the pre-shift dispersion personnel, and the group average dispersion index of the key period is obtained to construct a group dispersion sequence; the group dispersion sequence is split and time-dimensionally backtracked to obtain a concentrated dispersion period;

[0049] The attribution analysis is performed in combination with the degree of mastery, the concentrated dispersion period, and the high-optimal period and the medium-optimal period to obtain an attribution label;

[0050] The key period, the degree of mastery, the concentrated dispersion period, and the attribution label of the pre-shift briefing content of the key period are used as period division features;

[0051] M period division features are obtained, a period division dataset containing the M period division features is constructed, a gradient regression model algorithm based on multi-feature input is used to construct a period division model, and the period and duration of the key period of the pre-shift briefing content and the content order are output;

[0052] Adjust the next period of pre-shift briefing based on the time period and length of the output and the content order.

[0053] Advantages of the present application:

[0054] (1) The panoramic camera and the directional tracking camera are used to collect the pre-shift briefing video stream, the YOLOv8 algorithm is used to separate the personnel independent image area, the face visual line, the body posture and the group interaction features are extracted and weighted mapped into the attention distraction index, which is beneficial to convert the behavior features of the personnel in the pre-shift briefing process into quantitative attention state data, provides basic data support for subsequent personnel attention analysis and optimization, and realizes the monitoring and quantization of the personnel attention state.

[0055] (2) The pre-shift briefing personnel are identified based on the attention distraction index, the distraction period and the key period are extracted and time sequence overlap analysis is performed, and the interference misjudgment principle is constructed combined with the environmental events to judge whether the overlap analysis signal is triggered, which is beneficial to locate the key personnel and time correlation of the attention distraction in the key period, and provides targeted judgment basis for whether the personnel position needs to be adjusted.

[0056] (3) After the overlap analysis signal is triggered, the personnel basic position data and the distraction feature data are integrated to construct a data set, a position mobilization model is constructed by using the random forest algorithm to output the initial optimal position, and the final optimal position is determined through multidimensional feasibility verification, which is beneficial to improve the attention environment of the distracted personnel in the pre-shift briefing; the distraction optimization coefficient is calculated based on the attention distraction index after the position adjustment, the evaluation equation set is constructed combined with the risk weight of the high-optimal period and the medium-optimal period, and the environmental influence is excluded through the partial correlation interference verification, the information transmission effectiveness of the pre-shift briefing is judged in multiple dimensions, the briefing effect can be comprehensively evaluated from the individual optimization effect, the risk dimension and the environmental interference, and the information transmission quality of the pre-shift briefing is provided.

[0057] (4) When the information transmission is not up to standard, the personnel content mastery is obtained through rapid testing, the concentrated distraction period is positioned and attributed combined with the sliding window algorithm, and the XGBoost gradient regression model is constructed based on multiple features to output the period, length and content order optimization scheme of the key period, which is beneficial to provide data-driven direction for the subsequent adjustment of the key period of the pre-shift briefing, and improve the transmission effect of the key content in the pre-shift briefing. BRIEF DESCRIPTION OF DRAWINGS

[0058] The present application will be further described below with reference to the accompanying drawings.

[0059] Figure 1 is a module diagram of the pre-shift briefing AI intelligent device of the present application;

[0060] Figure 2Flow chart of whether to trigger an overlap analysis signal in the present application. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0062] Embodiment 1

[0063] Please refer to Figure 1 The present application is a pre-shift briefing AI intelligent device, which comprises the following modules:

[0064] The behavior collection module is used to collect the behavior characteristics of personnel during the pre-shift briefing, and the behavior characteristics are subjected to attention mapping processing to obtain an attention distraction index.

[0065] The process of collecting the behavior characteristics of personnel during the pre-shift briefing and subjecting the behavior characteristics to attention mapping processing to obtain an attention distraction index is as follows:

[0066] Preferably, the panoramic camera and the directional tracking camera of the AI intelligent device based on the pre-shift briefing are used to collect the video stream during the pre-shift briefing.

[0067] It should be noted that the pre-shift briefing is a preposition safety and operation guidance communication link between the construction team management personnel and the team operation personnel (workers) in the highway, municipal, building construction and other engineering construction fields before daily operation, and the communication content includes: specific operation tasks (such as process content, construction range), potential risk points on site (such as high-altitude falling, mechanical collision, electrical hazards, etc.), targeted safety protection requirements (such as protective equipment wearing standards) and emergency disposal procedures, as well as synchronous operation technical standards and matters needing attention.

[0068] The independent image area of each personnel during the pre-shift briefing is separated from the video stream based on the video stream through the YOLOv8 algorithm.

[0069] As understood by those skilled in the art, the YOLOv8 algorithm processes the pre-shift briefing video stream frame by frame by loading a pre-trained personnel target detection model, identifies the personnel target in each frame of image in real time and frames the boundary box, and then segments the independent image area of each personnel from the video stream image according to the boundary box.

[0070] The face line-of-sight features, body posture features and group interaction features of the personnel in the monitoring period are captured from the independent image area.

[0071] The face line-of-sight feature, the limb posture feature and the group interaction feature are taken as the behavior features of the personnel;

[0072] The face line-of-sight feature is collected in the following manner:

[0073] The size of the display screen of the pre-shift briefing and the standing distance of the personnel from the display screen are obtained, and the effective visual range of each personnel is calculated based on an optical line-of-sight algorithm;

[0074] The angle of view of the personnel to the display screen in the pre-shift briefing is obtained, and if the angle of view deviates from the effective visual range, the time length of the deviation of the angle of view from the effective visual range in the monitoring period is obtained to obtain the deviation time length;

[0075] As understood by those skilled in the art, the effective visual range of each personnel is calculated in the following manner: according to the actual width and height dimensions of the display screen of the pre-shift briefing (such as 1.5 m x 0.8 m) and the horizontal standing distance of the personnel from the left and right edges of the display screen captured by the directional tracking camera in real time, the horizontal field of view angles of the personnel looking at the left and right edges of the display screen and the vertical field of view angles of the personnel looking at the upper and lower edges of the display screen are calculated based on the geometric relationship of the field of view angle = 2 x arctan (object height / 2 x object distance) in the optical line-of-sight algorithm, and the area enclosed by the two direction field of view angles is the effective visual range in which the personnel can clearly receive the content of the display screen;

[0076] The angle of view of the personnel to the display screen in the pre-shift briefing is obtained in the following manner: the center points of the eyes of the personnel are located by the camera, and the center point coordinates of the display screen content display area are combined to calculate the included angle between the line connecting the center points of the eyes and the personnel to the center point of the display screen content and the horizontal and vertical boundaries of the effective visual range, and the included angle is the actual angle of view of the personnel to the display screen, so as to determine whether the angle of view falls within the effective visual range;

[0077] The proportion of the deviation time length to the time length of the monitoring period is calculated to obtain the deviation time ratio, and the deviation time ratio is taken as the face line-of-sight feature;

[0078] The frequency of the occurrence of invalid actions of the personnel in the monitoring period is obtained as the limb posture feature;

[0079] The cumulative time length of the turning of the head of the personnel to other personnel and the cumulative time length of the leaving of the personnel from the preset area in the monitoring period are obtained;

[0080] The two cumulative time lengths are summed up, and the sum is processed by ratio with the time length of the monitoring period to obtain the distraction time ratio, and the distraction time ratio is taken as the group interaction feature;

[0081] The face line-of-sight feature, the limb posture feature and the group interaction feature are weighted and summed up to obtain the attention distraction index, so as to realize the attention mapping processing of the behavior features;

[0082] It should be noted that the determination method of the weight is: 500 times of pre-shift briefing complete records are accumulated, each record needs to contain: specific numerical value of 3 behavior characteristics, and corresponding attention distraction degree score, (the attention distraction degree score is combined with the distraction behavior record observed by artificial observation and the subsequent 1 hour operation violation times, and the comprehensive assignment is 0-1) ; Then the Pearson correlation coefficient algorithm is used to calculate the correlation strength of each behavior characteristic and the distraction degree score, and the correlation coefficient r1 of the facial line of sight feature and the distraction degree is about 0.65, the correlation coefficient r2 of the body posture feature is about 0.45, and the correlation coefficient r3 of the group interaction feature is about 0.35;

[0083] Then the three correlation coefficients are normalized, and the facial line of sight feature weight = 0.65 / (0.65+0.45+0.35) = 0.5, the body posture feature weight = 0.45 / 1.45 approximately 0.3, and the group interaction feature weight = 0.35 / 1.45 approximately 0.2 are calculated.

[0084] It can be understood that the role of obtaining the attention distraction index is:

[0085] Role one, providing basis for pre-shift distraction personnel identification. The attention distraction index is used to set the demarcation value to mark the distraction period, and then calculate the distraction frequency, individual average distraction index and distraction intensity deviation, and form the personnel attention characteristic variable. These variables are input into the clustering algorithm, and the normal, mild and severe distraction groups can be divided, and then the rising trend of the index of the personnel in the severe distraction group is monitored to locate the pre-shift distraction personnel that need to be focused on.

[0086] Role two, providing data basis for distraction and key period time series overlap analysis. When the attention distraction index in the monitoring period is greater than or equal to the preset demarcation value, the period is marked as a distraction period. The number of overlapping periods and the period overlap ratio of the distraction period and the key period are calculated, the coincidence determination coefficient is calculated combined with the average distraction index in the overlapping period, and the interference misjudgment principle constructed by the environmental event is combined to judge whether the overlap analysis signal required by the position adjustment is triggered.

[0087] The overlap analysis module: based on the attention distraction index, the attention distraction of personnel is analyzed, the pre-shift distraction personnel is identified, the distraction period of the pre-shift distraction personnel attention is extracted, and the key period of the pre-shift briefing is extracted. The time series overlap analysis of the distraction period and the key period is performed, the coincidence determination coefficient is obtained, and it is judged whether the overlap analysis signal is triggered;

[0088] Among them, the way of identifying the pre-shift distraction personnel based on the attention distraction index is:

[0089] Obtain the attention dissipation index of a single person over N monitoring periods, set a threshold value for the attention dissipation index, and mark the monitoring period as a dissipation period if the attention dissipation index within a monitoring period is higher than or equal to the preset threshold value.

[0090] Extract the number of consecutive periods of dispersion within N monitoring periods, calculate the ratio of the number of consecutive periods to the number of monitoring periods N, and obtain the dispersion frequency.

[0091] The average distractibility index of an individual is obtained by calculating the mean of the distractibility index within the number of consecutive cycles of distractibility.

[0092] The average individual disorganization index of all personnel during the pre-shift briefing is used to obtain the group average disorganization index.

[0093] The deviation percentage between the individual average disorganization index and the group average disorganization index is calculated to obtain the disorganization intensity deviation.

[0094] It should be noted that the deviation ratio is calculated as follows: (Individual average dispersion index - Group average dispersion index) / Group average dispersion index;

[0095] The dispersion intensity deviation, the individual's average dispersion index, and the dispersion frequency are used as characteristic variables for each person.

[0096] The characteristic variables of all personnel in the pre-shift briefing were obtained, and the characteristic variables of all personnel were clustered using the K-means clustering algorithm to obtain the normal disorganization group, the mild disorganization group, and the severe disorganization group.

[0097] Obtain the information of the severely inattentive group and their corresponding inattention index. If the inattention index of the severely inattentive group shows an upward trend within N monitoring periods, then mark the person as an inattentive person before shift.

[0098] Those skilled in the art will understand that the K-means clustering algorithm is used to cluster the feature variables of all individuals as follows: The feature variables of all individuals are organized into a data matrix; then, the number of clusters is set to 3 (corresponding to normal, mild, and severe disorganized groups). K-means first randomly selects 3 initial cluster centers; by iteratively calculating the distance of each individual to the 3 centers, individuals are assigned to the closest group, and the center position of each group is updated. This process is repeated until the centers no longer change; the final 3 groups will naturally distinguish normal, mild, and severe disorganized individuals based on differences in feature variables, providing a grouping basis for subsequent screening of severe groups and judging the upward trend of the index.

[0099] The determination manner that the attention distraction index presents an upward trend in the N monitoring periods is that: the attention distraction indexes of the personnel in the severe distraction group in the N monitoring periods are arranged in time sequence, linear fitting is performed with the monitoring period sequence as the horizontal axis and the distraction index as the vertical axis, if the slope of the regression straight line obtained by fitting is positive, and N / 2 fitting coordinate points are distributed above the straight line, it is determined that an upward trend is presented;

[0100] The manner of extracting the key period of the pre-shift briefing is that:

[0101] In some embodiments, the person skilled in the art extracts the time period related to the core information (such as safety specifications and task critical nodes) as the key period of the pre-shift briefing according to the content process of the pre-shift briefing;

[0102] It should be noted that the time division dimensions of the key period of the pre-shift briefing and the monitoring period are the same;

[0103] The process of performing time sequence overlap analysis on the distraction period and the key period, obtaining a coincidence determination coefficient and determining whether an overlap analysis signal is triggered is that:

[0104] If the distraction period and the key period overlap in the time dimension, the overlapping period is marked as an overlap period, and the number of overlap periods is obtained as the number of overlap periods;

[0105] The average distraction index of the individual in the overlap period is calculated;

[0106] The number of key periods is obtained, the proportion of the number of overlap periods to the number of key periods is calculated, and a period overlap ratio is obtained;

[0107] The average distraction index and the period overlap ratio are multiplied to obtain a coincidence determination coefficient;

[0108] It can be understood that the physical meaning of the coincidence determination coefficient is that the correlation and influence degree of the attention distraction state of the pre-shift distraction personnel and the key period (core information transmission period) of the pre-shift briefing in the time dimension is quantified. The actual degree of influence of the core information transmission caused by the distraction of the personnel in the key period is reflected by the product of the individual average distraction index in the overlap period (reflecting the severity of the attention distraction of the personnel in the overlap period) and the period overlap ratio (the proportion of the number of overlap periods to the number of key periods, reflecting the time overlap range of the distraction and the key period), which provides a quantitative basis for determining whether an overlap analysis signal is triggered;

[0109] The current environmental event is collected, and an interference misjudgment principle is constructed;

[0110] The manner of collecting the current environmental event and constructing the interference misjudgment principle is that:

[0111] If a sudden environmental disturbance (such as equipment failure or external noise) occurs during the pre-shift briefing, it is marked as an environmental event. At the same time, the overlap time between the environmental event and the key cycle is obtained. If the overlap time is ≥30% of the key cycle, the interference misjudgment principle is met; otherwise, it is not met.

[0112] like Figure 2 As shown, the overlap determination coefficient is compared with the preset overlap determination threshold. If the overlap determination coefficient is higher than or equal to the preset overlap determination threshold, and the environmental event does not meet the interference misjudgment principle, then the overlap analysis signal is triggered.

[0113] If the overlap determination coefficient is lower than the preset overlap determination threshold, or if the environmental event meets the interference misjudgment principle, the change of the overlap determination coefficient will be continuously monitored.

[0114] It should be noted that the overlap judgment threshold is obtained as follows: The overlap judgment threshold needs to be quantitatively determined by combining historical data and actual risk scenarios: First, collect complete data from at least 500 pre-shift briefings (including the overlap judgment coefficient for each time, the violation rate of the subsequent hour of operation, and hazard records). Group the overlap judgment coefficients according to numerical ranges (e.g., 0.1-0.2, 0.2-0.3, etc.) and calculate the average violation rate for each group. Then, through ROC curve analysis, find the critical point that "makes the overlap judgment coefficient of the high violation rate group (e.g., violation rate ≥ 5%) recognized and the misjudgment rate of the low violation rate group (e.g., violation rate < 1%) ≤ 5%". This critical point is the initial overlap judgment threshold.

[0115] Example 2

[0116] like Figure 1 As shown, this invention is an AI-powered intelligent device for pre-shift briefings, and also includes the following modules:

[0117] Location optimization module: If the overlap analysis signal is triggered, the location of the pre-shift briefing for the personnel who are distracted before the shift is adjusted, and the attention distraction index of all personnel after the location adjustment is obtained.

[0118] The method for adjusting the location of the pre-shift briefing for employees who are not attentive before their shift is as follows:

[0119] Preferably, the area where people stand is located by using a directional tracking camera to obtain basic position data and dispersion feature data for each person, and the two sets of data are combined to obtain a dispersion data set;

[0120] It should be noted that the basic location data includes the personnel's location coordinates, as well as the distance between the personnel and the display screen, the horizontal angle, and the vertical angle (for example, personnel coordinates (3,2) correspond to a distance of 4 meters from the display screen and a horizontal angle of 12°), and whether there are any obstructions (such as pillars or equipment) around the location (marked with 0-1, where 1 indicates obstruction).

[0121] The distraction feature data includes: attention distraction index and face line of sight deviation time ratio under the position, limb invalid action frequency;

[0122] Integrate the distraction data set of all personnel, construct a position adjustment model through a random forest algorithm, input the distraction data set into the position adjustment model, and output the initial preferred position of each pre-shift distraction personnel;

[0123] The skilled person in the art can understand that the position adjustment model is constructed by a random forest algorithm in the following manner: taking the basic position data (coordinates, distance, angle, and shielding mark) as input features, and taking "whether the attention distraction index corresponding to the position is lower than the preset optimization threshold" (such as 0.5, 1 for yes and 0 for no) as the target label; the distraction data set is divided into a training set and a validation set in the ratio of 8:2, a forest composed of multiple decision trees is constructed using the training set (each tree is trained by randomly selecting part of the samples and features), and the prediction result of "whether the position is preferred" is output through the majority voting mechanism; the model accuracy is evaluated using the validation set (such as requiring ≥85%), and the number of trees (such as 50-200), maximum depth, and other parameters are adjusted to optimize the model, so that the model can output the preferred position that can reduce the distraction index based on the input position features;

[0124] The preferred position is subjected to feasibility verification, and if the feasibility verification is satisfied, the initial preferred position is taken as the preferred position;

[0125] For example, the preferred position is subjected to feasibility verification in the following manner: verification 1: whether the preferred position is within the "pre-shift briefing preset effective area" (such as a 10m×8m rectangular area, coordinates outside the area are excluded);

[0126] Verification 2: whether the distance between the preferred position and other personnel is ≥1.2 meters (complying with the safety distance to avoid adjusting personnel to be crowded and increasing interaction distraction);

[0127] Verification 3: if the preferred position is shielded (such as a model not recognizing temporarily placed tools), a sub-optimal position is automatically selected from the "alternative position library" (pre-labeled positions without shielding and good vision) to ensure that the adjustment is executable;

[0128] Based on the voice prompt of the pre-shift briefing and the help and guidance of the staff, the pre-shift distraction personnel are changed from the current position to the preferred position through the pre-shift briefing;

[0129] The behavior features of the personnel during the pre-shift briefing after the position change are obtained, and the attention mapping process is performed again to obtain the attention distraction index.

[0130] Pre-shift evaluation module: based on the adjusted attention distraction index, the attention of the pre-shift disclosure personnel is optimized and evaluated to obtain a distraction optimization coefficient, and based on the distraction optimization coefficient, the optimization effectiveness of the key period is analyzed to determine whether the information transmission effectiveness of the pre-shift disclosure meets the standard;

[0131] The way to evaluate the distraction optimization coefficient of the pre-shift disclosure personnel is:

[0132] The change amount of the individual average distraction index of the single pre-shift distraction personnel before and after position adjustment is obtained and calculated to obtain a distraction adjustment amount;

[0133] The ratio of the distraction adjustment amount to the individual average distraction index before adjustment is calculated to obtain the distraction optimization coefficient of the overall period;

[0134] The distraction adjustment amount of the key period after position adjustment is obtained, and the ratio of the individual average distraction index before adjustment is calculated to obtain the distraction optimization coefficient of the key period;

[0135] If the distraction optimization coefficient of the overall period and the distraction optimization coefficient of the key period are both positive, and the individual average distraction index in the key period is lower than the preset distraction boundary value, it indicates that the attention regulation of the single pre-shift distraction personnel is effective;

[0136] The way to evaluate the distraction optimization coefficient of the pre-shift disclosure personnel is:

[0137] Based on the single effective personnel data, the key period is split to obtain a high-optimization period and a medium-optimization period;

[0138] Combined with the individual average distraction index before adjustment, an optimization evaluation of the split period (high-optimization period and medium-optimization period) is established to obtain a high-optimization period overall optimization coefficient and a medium-optimization period overall optimization coefficient;

[0139] Preferably, the way to establish the optimization evaluation of the split period is:

[0140] The individual average distraction index of the attention regulation effective personnel before position adjustment (reflecting the distraction base level before adjustment) and the distraction optimization coefficient in the high-optimization period (reflecting the distraction improvement effect of the high-optimization period after position adjustment) are obtained;

[0141] For each attention regulation effective personnel, the product of the individual average distraction index before adjustment and the high-optimization period distraction optimization coefficient is calculated, and the product results of all attention regulation effective personnel are accumulated to obtain an optimization total contribution value of the high-optimization period (reflecting the optimization effect sum of all effective personnel in the high-optimization period);

[0142] Add the individual average distraction index of all valid personnel before adjustment to obtain the optimization benchmark total value of the high-optimization period (reflecting the distraction benchmark sum before adjustment of all valid personnel);

[0143] Perform ratio processing on the optimization total contribution value and the optimization benchmark total value to obtain the overall optimization coefficient of the high-optimization period (the higher the value, the more significant the overall attention improvement effect in the high-optimization period, which meets the disclosure requirement of safety first);

[0144] Obtain the individual average distraction index of the attention regulation valid personnel before position adjustment and the distraction optimization coefficient in the medium-optimization period;

[0145] For each attention regulation valid personnel, calculate the product of the individual average distraction index before adjustment and the distraction optimization coefficient in the medium-optimization period, and add the product results of all attention regulation valid personnel to obtain the optimization total contribution value of the medium-optimization period;

[0146] Add the individual average distraction index of all valid personnel before adjustment to obtain the optimization benchmark total value of the medium-optimization period;

[0147] Calculate the ratio of the optimization total contribution value and the optimization benchmark total value to obtain the overall optimization coefficient of the medium-optimization period;

[0148] Obtain the associated operation violation rate in the high-optimization period and the overlap duration of environmental event interference and the key period;

[0149] It should be noted that the associated operation violation rate is obtained by comparing the number of violation events in the construction period corresponding to the operation content involved in the high-optimization period with the total number of operations in that period through intelligent monitoring or safety patrol records on site;

[0150] After excluding the overlap duration of environmental event interference and the key period, based on the overall optimization coefficient of the high-optimization period, the associated operation violation rate in the high-optimization period is subjected to partial correlation interference verification to obtain the partial correlation coefficient;

[0151] The partial correlation interference verification method is as follows: in the pre-shift disclosure scenario, for the overall optimization coefficient of the high-optimization period (X), the associated operation violation rate of the high-optimization period (Y), and the overlap duration of environmental interference and the key period (Z), when calculating the partial correlation coefficient after controlling Z, the Pearson correlation coefficients of X and Y, X and Z, and Y and Z are calculated. 、 、 );

[0152] Based on the Pearson correlation coefficients of X and Y, X and Z, and Y and Z, the partial correlation coefficient formula is substituted , by subtracting the common influence term of Z on X and Y and standardizing the processing, the net correlation coefficient of X and Y after eliminating environmental interference, i.e., the partial correlation coefficient, is obtained.

[0153] The standard effectiveness criterion is constructed, and if the high-optimization period overall optimization coefficient and the medium-optimization period overall optimization coefficient and the partial correlation coefficient meet the standard effectiveness criterion, the information transmission effectiveness of the pre-shift briefing meets the standard;

[0154] If not, it is determined that the information transmission effectiveness of the pre-shift briefing does not meet the standard;

[0155] For example, the way to construct the standard effectiveness criterion is: based on the accident data of the pre-shift briefing of the project in the past three years and 500+ effective control records, the core threshold is determined through ROC curve analysis, that is, the high-optimization period (safety type) overall weighted optimization coefficient is greater than or equal to 22% (the threshold corresponds to the identification critical point that the safety violation rate is less than or equal to 3% after optimization of the high-optimization period), and the medium-optimization period (task type) overall weighted optimization coefficient is greater than or equal to 18% (matching the risk weight of 0.4, corresponding to the critical point that the task type violation rate is less than or equal to 5%); then through the statistical significance verification of 300+ groups of "optimization coefficient-violation rate-environmental interference" data, it is determined that the partial correlation coefficient is less than or equal to -0.5 after controlling the environmental interference (at this time, the negative correlation between the optimization of the high-optimization period and the decrease of the violation rate is significant at the p<0.05 level, and the casual correlation is excluded); at the same time, combined with the analysis of historical standard cases, the average decrease of the group overall operation violation rate is greater than or equal to 30%, so it is included in the criterion; finally, the high-optimization is greater than or equal to 22%, the medium-optimization is greater than or equal to 18%, the partial correlation coefficient is less than or equal to -0.5, and all three standard criteria need to be met, that is, the information transmission effectiveness of the pre-shift briefing meets the standard.

[0156] The pre-shift optimization module: if it does not meet the standard, it is used to obtain the cycle division characteristics of the key period of the pre-shift scattered personnel; the cycle division model is constructed, the cycle division characteristics are input into the cycle division model, and the pre-shift briefing adjustment content is output based on the model to adjust the key period;

[0157] The way to obtain the cycle division characteristics of the key period of the pre-shift scattered personnel is:

[0158] Preferably, if the information transmission effectiveness of the pre-shift briefing does not meet the standard, the key period content quick test is pushed to all pre-shift scattered personnel through the touch screen and voice interaction of the AI intelligent device, and the quick test score is obtained as the mastery degree;

[0159] At the same time, the group average dispersion index of the key period is obtained, and the group dispersion sequence is constructed;

[0160] The group dispersion sequence is split and time-dimensionally backtracked by using the sliding window algorithm, and the concentrated dispersion period is obtained;

[0161] For example, the way of splitting and identifying the time dimension of the backtracking is as follows: if the total length of the pre-shift is 30 minutes, the group average distraction index is collected in units of 1 minute to form a group distraction sequence (such as 0.4 in the first minute, 0.35 in the second minute, …, 0.5 in the 30th minute), and the high priority key period is from the 10th minute to the 20th minute; the window size of the sliding window algorithm is set to 3 minutes (to ensure that the minimum effective period of continuous distraction is covered), and the step size is 1 minute (to avoid missing the period), and the sliding starts from the first minute: the average distraction index of the first window (1-3 minutes) is 0.38, the average distraction index of the second window (2-4 minutes) is 0.36, …, when the sliding window reaches the 14th-16th minute, the average index is 0.72 (≥ the preset high distraction threshold 0.7), the average index of the 15th-17th minute window is 0.75, and the average index of the 16th-18th minute window is 0.73, all of which meet the threshold; the three continuous and overlapping windows are combined to obtain the concentrated distraction period of the 14th-18th minute, and the period is completely within the high priority key period (10-20 minutes), which provides a basis for subsequent attribution analysis;

[0162] The attribution analysis is performed in combination with the mastery degree, the concentrated distraction period, and the high priority period and the medium priority period to obtain an attribution label;

[0163] The way of performing the attribution analysis is as follows: first, a judgment criterion is set: the mastery degree is taken as a boundary at 80% (≥ 80% is good mastery, and < 80% is insufficient mastery, which is set by a person skilled in the art according to experience), the concentrated distraction period is taken as a judgment basis whether it is completely / partially within the corresponding period, and the high priority period (safety type) has a higher priority than the medium priority period (task type);

[0164] If the mastery degree of the high priority period is < 70%, the concentrated distraction period (such as the 14th-18th minute) is completely within the high priority period (the 10th-20th minute), the mastery degree of the medium priority period is ≥ 80%, and the concentrated distraction period does not overlap with the medium priority period, the attribution label is “the high priority period overlaps with the attention trough, resulting in ineffective delivery of safety type content”;

[0165] If the mastery degree of the high priority period is < 70%, the concentrated distraction period does not overlap with the high priority period, but the content understanding difficulty of the high priority period is marked as level 5 (the highest), the attribution label is “the content difficulty of the high priority period is too high, and the personnel cannot understand it, resulting in insufficient mastery”;

[0166] If the mastery degree of the medium priority period is < 60%, the concentrated distraction period completely overlaps with the medium priority period (the 21st-28th minute), the mastery degree of the high priority period is ≥ 85%, and there is no concentrated distraction, the attribution label is “the medium priority period is arranged in the attention trough, and the task type content delivery is insufficient”, and the corresponding attribution label is output by combination and matching of multi-dimensional data;

[0167] The key cycle, the mastery degree, the concentration and distraction period and the attribution label of the pre-shift briefing content of the key cycle are taken as the cycle division features;

[0168] M cycle division features are acquired, and a cycle division dataset containing the M cycle division features is constructed;

[0169] Preferably, M = 200;

[0170] A gradient regression model algorithm based on multi-feature input is used to construct a cycle division model;

[0171] As understood by those skilled in the art, when constructing the cycle division model, the features in the cycle division dataset are first converted into numerical values that can be processed by the model;

[0172] That is, the attribution label is converted into a numerical vector through one-hot encoding, the concentration and distraction period is converted into a length numerical value according to the start minute-end minute, the mastery degree is kept in the form of a percentage decimal, and the original key cycle is split into multi-dimensional features according to the period start value, length and risk weight to form a complete input feature matrix as the dataset;

[0173] The dataset is divided into a training set and a validation set according to an 8:2 ratio, an XGBoost gradient regression model is selected, and the period start value, length and content order priority of the optimized key cycle are taken as three independent regression targets; during training, the hyperparameters are optimized through a grid search method (wherein the number of trees is set to 100-200, the learning rate is 0.05-0.1, and the maximum depth is 3-5 layers), the mean square error (MSE) is taken as the loss function, L1 regularization is added to avoid overfitting, and the model prediction accuracy is monitored in real time during the training process until the error of the model on the validation set is stable and there is no overfitting trend, thereby forming a cycle division model that can output the period, length and content order;

[0174] The cycle division dataset is input into the cycle division model, and the period and length of the key cycle of the pre-shift briefing content and the content order are output;

[0175] The next cycle of pre-shift briefing is adjusted based on the output period and length and content order.

[0176] The above describes one embodiment of the present application in detail, but the content described is only a preferred embodiment of the present application and cannot be considered as limiting the scope of the implementation of the present application. Any equivalent changes and improvements made within the scope of the present application should still be included in the scope of the present application.

Claims

1. An AI-powered intelligent device for pre-shift briefings, characterized in that: Includes the following modules: Behavior data collection module: used to collect the behavioral characteristics of personnel during pre-shift briefings, perform attention mapping processing on the behavioral characteristics, and obtain the attention dissipation index; Overlap Analysis Module: Based on the attention dissipation index, perform attention dissipation analysis on personnel, identify personnel who are inattentive before shift, extract the attention dissipation cycle of personnel who are inattentive before shift, as well as the key cycle of pre-shift briefing, perform time-series overlap analysis on the dissipation cycle and key cycle, obtain the overlap judgment coefficient, and determine whether the overlap analysis signal is triggered. Location optimization module: If the overlap analysis signal is triggered, the location of the pre-shift briefing for the personnel who are distracted before the shift is adjusted, and the attention distraction index of all personnel after the location adjustment is obtained. Pre-shift assessment module: Based on the adjusted attention dissipation index, the attention of the personnel conducting the pre-shift briefing is optimized and assessed to obtain the dissipation optimization coefficient. Based on the dissipation optimization coefficient, the effectiveness of optimization in key periods is analyzed to determine whether the effectiveness of information transmission during the pre-shift briefing meets the standards. Pre-shift optimization module: If the target is not met, it is used to obtain the periodic division characteristics of key periods for employees who are not focused before the shift. Construct a cycle segmentation model, input cycle segmentation features into the cycle segmentation model, and adjust key cycles based on the pre-shift briefing adjustment content output by the model; The method for performing the aforementioned time series overlap analysis is as follows: Extract the period of distraction of the attention of the personnel who are distracted before the shift. If the period of distraction and the period of focus overlap in the time dimension, mark the overlapping period as the overlapping period and obtain the number of overlapping periods as the number of overlapping periods. Calculate the average disengagement index of individuals with pre-shift disengagement during overlapping periods; Obtain the number of key cycles, calculate the ratio of overlapping cycles to the total number of key cycles, and obtain the cycle overlap ratio. The overlap determination coefficient is obtained by multiplying the average dispersion index with the period overlap ratio. The method for optimizing the assessment of the attention of personnel during pre-shift briefings is as follows: Obtain and calculate the change in the average disorganization index of a single pre-shift disorganized employee before and after the position adjustment, and obtain the disorganization adjustment amount; The ratio of the dispersion adjustment amount to the individual average dispersion index before adjustment is calculated to obtain the dispersion optimization coefficient of the overall cycle. Obtain the dispersion adjustment amount of the key period after the position adjustment, and calculate the ratio of the individual average dispersion index before the adjustment to obtain the dispersion optimization coefficient of the key period. If the overall cycle's dispersion optimization coefficient and the key cycle's dispersion optimization coefficient are both positive, and the individual average dispersion index within the key cycle is lower than the preset dispersion threshold, then it indicates that the attention regulation of a single pre-shift distracted person is effective. The method for performing the optimization effectiveness analysis is as follows: Based on individual valid personnel data, key cycles are broken down into high-optimal cycles and medium-optimal cycles; By combining the individual average dispersion index before adjustment, a periodic optimization evaluation process is established to obtain the overall optimization coefficient for high-optimal periods and the overall optimization coefficient for medium-optimal periods. Obtain the violation rate of related operations within the high-optimization period, as well as the overlap duration between environmental incident interference and key periods; After excluding the overlap between environmental events and key cycles, the partial correlation coefficient is obtained by checking the partial correlation interference based on the overall optimization coefficient of the high-optimal cycle and the violation rate of related operations within the high-optimal cycle. Construct a criterion for the effectiveness of meeting the standards. If the overall optimization coefficient and the partial correlation coefficient of the high-optimal cycle and the medium-optimal cycle meet the criterion for the effectiveness of meeting the standards, then the information transmission effectiveness of the pre-shift briefing is met. The method for adjusting key cycles is as follows: By using AI-powered smart devices to obtain the pre-shift awareness of disorganized personnel and the average disorganization index of key periods, a disorganization sequence is constructed. The disorganization sequence is then broken down and backtracked over time to identify concentrated disorganization periods. Attribution analysis was conducted by combining the degree of mastery, periods of concentrated and dispersed activity, and high-optimal and medium-optimal cycles to obtain attribution labels; The key periods, mastery levels, concentrated and scattered periods, and attribution labels of the pre-shift briefing content for key periods are used as the characteristics for period division. Obtain M periodicity features and construct a periodicity dataset containing the M periodicity features; based on a gradient regression model algorithm with multiple feature inputs, construct a periodicity model and output the time period, duration, and content order of the key periods in the pre-shift briefing content; The pre-shift briefing for the next cycle will be adjusted based on the output time period, duration, and content order.

2. The AI-powered intelligent device for pre-shift briefing as described in claim 1, characterized in that: The attention mapping process for the aforementioned behavioral features is as follows: Among them, behavioral characteristics include: facial gaze characteristics, body posture characteristics, and group interaction characteristics; The inattention index is obtained by weighting and summing facial gaze features, body posture features, and group interaction features.

3. The AI-powered intelligent device for pre-shift briefing according to claim 2, characterized in that: The method for obtaining the facial gaze features is as follows: Obtain the effective field of view for each person during the pre-shift briefing. If the viewing angle deviates from the effective field of view, obtain the duration of the deviation within the monitoring period. The deviation time is calculated as a proportion of the monitoring cycle time to obtain the deviation time ratio, which is then used as a facial gaze feature.

4. The AI-powered intelligent device for pre-shift briefing according to claim 1, characterized in that: The method for identifying employees who are disorganized before their shift is as follows: Obtain the characteristic variables of all personnel during the pre-shift briefing, and then use a clustering algorithm to cluster the characteristic variables of all personnel. Extract the individuals in the severely inattentive group from the clustering results, along with their corresponding inattention index. If the inattention index of the severely inattentive group shows an upward trend over N monitoring periods, then the individuals are marked as inattentive before their shift.

5. The AI-powered intelligent device for pre-shift briefing according to claim 4, characterized in that: The method for obtaining the feature variables is as follows: Extract the number of consecutive periods of dispersion within N monitoring periods, and calculate the ratio of the number of consecutive periods to the number of monitoring periods N to obtain the dispersion frequency. The average distractibility index of an individual is obtained by calculating the mean of the distractibility index within the number of consecutive cycles of distractibility. The average individual disorganization index of all personnel during the pre-shift briefing is used to obtain the group average disorganization index. The deviation percentage between the individual average disorganization index and the group average disorganization index is calculated to obtain the disorganization intensity deviation. The dispersion intensity deviation, the individual's average dispersion index, and the dispersion frequency are used as characteristic variables for each person.

6. The AI-powered intelligent device for pre-shift briefing according to claim 5, characterized in that: The method for performing the partial correlation interference verification is as follows: When calculating the partial correlation coefficients after controlling Z for the overall optimization coefficient X of the high-optimal cycle, the violation rate Y of the associated operations of the high-optimal cycle, and the overlap time Z of environmental interference and key cycles, the Pearson correlation coefficients of X and Y, X and Z, and Y and Z are calculated. The partial correlation coefficient is obtained by substituting the Pearson correlation coefficients of X and Y, X and Z, and Y and Z into the formula for the partial correlation coefficient.

Images