Safety notification training method and system based on VR eye movement tracking

Abstract

The invention provides a safety notification training method and system based on VR eye movement tracking, and belongs to the field of virtual reality technology and educational training. The method comprises the following steps: acquiring personal information of a student, establishing a binding relationship between the student information and a serial number of a student end VR device, and uploading the binding relationship to a server; the student side VR device obtains personal information from the server according to the device serial number and loads a corresponding training scene and a configuration table; acquiring student eye tracking data in real time; whether the student watches the safety notification content or not is judged based on eye tracking data in cooperation with a ray detection technology; recording the learning progress of the student in real time, and uploading data to a server; the teacher side is used for synchronously obtaining student learning data, and a compliance training report is automatically generated based on eye movement data, learning records and electronic signatures. According to the method, the 3D model, the dynamic text and voice explanation are combined, the learning interest and concentration of students are improved, and attention to key safety information is ensured by monitoring the sight focus of the students in real time.

Description

Technical Field

[0001] The present invention belongs to the field of virtual reality technology and education and training, and in particular relates to a safety notification training method and system based on VR eye tracking. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] With the rapid development of virtual reality (VR) technology, its application in education and training is becoming increasingly widespread, especially in safety training. Traditional safety training methods primarily rely on paper-based materials, on-site demonstrations, or video tutorials, but these methods are gradually revealing their limitations in practical applications. When it comes to evaluating training effectiveness, traditional methods struggle to accurately quantify trainees' mastery of safety knowledge and objectively reflect their attention and understanding of key information. Training processes are generally characterized by a lack of interactivity, and the one-way knowledge transfer model can easily distract trainees, reducing their engagement and impacting training effectiveness.

[0004] In terms of scene reproduction capabilities, traditional training methods have limited ability to simulate complex industrial environments or high-risk work scenarios, making it difficult to provide a realistic and immersive experience. This leaves trainees lacking sufficient response capabilities when faced with emergencies in actual operations. Regarding data management, traditional training recording, statistics, and analysis rely primarily on manual operations, which is not only inefficient but also prone to data errors and omissions, making it difficult to provide timely and accurate data support for corporate safety management decisions. Furthermore, when it comes to compliance verification, traditional training methods struggle to achieve full-process recording and electronic archiving of the training process, failing to meet the stringent requirements of modern enterprise safety management for traceable and auditable training records.

[0005] In recent years, some VR training systems have begun to introduce eye-tracking technology to improve training effectiveness, but they still face many technical challenges in practical applications. The existing system's eye movement data collection accuracy is insufficient, making it difficult to accurately judge the trainee's gaze state and level of cognition of key safety information. The lack of a standardized mechanism for system configuration management makes it difficult to update training content and cannot quickly respond to changes in corporate training needs. There are delays in data synchronization between the teacher and student ends, which affects the real-time monitoring and management efficiency of the training process. The report generation function is generally simple and cannot automatically integrate multi-dimensional training data and generate complete reports that meet compliance requirements. A lot of manual intervention is still required. These technical shortcomings limit the in-depth application and promotion of VR eye tracking technology in the field of safety training. Summary of the Invention

[0006] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a safety notification training method and system based on VR eye tracking.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: A first aspect of the present invention provides a safety notification training method based on VR eye tracking; A safety information training method based on VR eye tracking, including: S1, obtain the personal information entered by the student, establish a binding relationship between the student information and the serial number of the student's VR device, and upload the binding relationship information to the server; S2: The student's VR device obtains personal information from the server based on the device serial number and loads the corresponding training scenario and configuration table; S3: Real-time collection of student eye tracking data. Based on this data and combined with radiographic detection technology, it is determined whether the student is paying attention to the safety information. If the student's gaze deviates from the target content or the duration of eye closure exceeds a preset threshold, the learning process is considered interrupted. S4, records students’ learning progress in real time and uploads the data to the server; S5 uses the teacher’s end to synchronously obtain students’ learning data and automatically generates compliance training reports based on eye movement data, learning records and electronic signatures.

[0008] As a further technical solution, the eye tracking data includes position information and rotation information of the eyeball relative to the camera.

[0009] As a further technical solution, the eye movement data is obtained by calling the PXR_MotionTracking.GetEyeTrackingData function of the VR device, and ray detection is performed using Physics.Raycast to determine whether the trainee's gaze point matches the safety notification content.

[0010] As a further technical solution, the configuration table of the training scene is implemented based on Unity's ScriptableObject, including the description, ID, scene name, 3D model and learning text of the safety notification point. The teacher can remotely update the configuration table and synchronize it to the student side.

[0011] As a further technical solution, students’ learning progress is recorded in real time and the data is uploaded to the server, including: When learning is completed, the current informed knowledge is marked as learned locally and stored in the informed learning data; All learning data will be serialized into Json and uploaded to the server, saved in the database, and stored in the corresponding folder according to the student's personal information.

[0012] As a further technical solution, the server stores student login information, device status and learning records. The learning records include the ID of the learning notification point and the learning completion status. The data is synchronized in real time between the teacher and student ends through the HTTP protocol.

[0013] As a further technical solution, the generation of the training report includes the following sub-steps: Obtain students' electronic signatures and learning records from the server; Compare the learned and unlearned safety notification points according to the configuration table; Fill the comparison results into the preset report template and export them as PDF or Word format files.

[0014] A second aspect of the present invention provides a safety notification training system based on VR eye tracking.

[0015] A safety information training system based on VR eye tracking, including: The student information acquisition module is configured to: obtain the personal information input by the student, establish a binding relationship between the student information and the device serial number, and upload the binding relationship information to the server; The training scenario generation module is configured as follows: the student-side VR device obtains personal information from the server according to the device serial number and loads the corresponding training scenario and configuration table; The content notification module is configured to: start the VR training program and collect the trainee's eye tracking data in real time through the VR device's eye tracking API; use the eye tracking data in conjunction with radiographic detection technology to determine whether the trainee is paying attention to the safety notification content; if it is detected that the trainee's gaze deviates from the target content or the duration of eye closure exceeds a preset threshold, the learning is determined to be interrupted; The learning progress recording module is configured to: record the student's learning progress in real time and upload the data to the server; The training report generation module is configured to: use the teacher's end to synchronously obtain student learning data, and automatically generate a compliance training report based on eye movement data, learning records and electronic signatures.

[0016] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps of a safety information training method based on VR eye tracking as described in the first aspect of the present invention.

[0017] The fourth aspect of the present invention provides an electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps of the safety notification training method based on VR eye tracking as described in the first aspect of the present invention are implemented.

[0018] One or more of the above technical solutions have the following beneficial effects: (1) This invention uses VR technology to create a 1:1 replica of an industrial environment, combined with 3D models, dynamic text, and voice explanations, allowing trainees to be immersed in the environment and improve their learning interest and concentration. At the same time, eye tracking technology can monitor the trainees' gaze in real time to ensure that they are truly paying attention to key safety information rather than simply "going through the motions." The system also supports interactive learning, such as simulated operations and risk point marking, making training more practical and helping trainees to more firmly grasp safety knowledge.

[0019] (2) The present invention achieves full automation of the training process through digital management. The teacher can remotely configure the training content and monitor the students' progress in real time, eliminating the need for manual statistics of learning status. The system automatically records students' learning data (such as gaze duration and completion rate) and generates standardized reports, significantly reducing the management burden. In addition, VR equipment is reusable, avoiding the repeated investment in materials and venues in traditional training, significantly reducing corporate training costs in the long run.

[0020] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0022] Figure 1 This is a flow chart of the method of the first embodiment.

[0023] Figure 2 It is a box plot of the training effects of different training methods in the first embodiment.

[0024] Figure 3 This is a system structure diagram of the second embodiment. DETAILED DESCRIPTION

[0025] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0026] It should be noted that the terms used herein are for describing particular embodiments only and are not intended to limit the exemplary embodiments according to the present invention.

[0027] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0028] Example 1 This embodiment discloses a safety notification training method based on VR eye tracking; like Figure 1 As shown in FIG, a safety information training method based on VR eye tracking includes: S1, obtain the personal information entered by the student, establish a binding relationship between the student information and the serial number of the student's VR device, and upload the binding relationship information to the server; S2: The student's VR device obtains personal information from the server based on the device serial number and loads the corresponding training scenario and configuration table; S3: Real-time collection of student eye tracking data. Based on this data and combined with radiographic detection technology, it is determined whether the student is paying attention to the safety information. If the student's gaze deviates from the target content or the duration of eye closure exceeds a preset threshold, the learning process is considered interrupted. S4, records students’ learning progress in real time and uploads the data to the server; S5 uses the teacher’s end to synchronously obtain students’ learning data and automatically generates compliance training reports based on eye movement data, learning records and electronic signatures.

[0029] Specifically, it also includes the following: In step S1, the personal information entered by the student when logging into the device is obtained. This personal information includes the student's name, ID, and handwritten signature data. The student information is then bound to the device's serial number. The device's text code, which includes the upload server's domain name, the device's serial number, and the student's personal information, is then uploaded to the server. The server then creates folders with corresponding names based on the device's serial number to store the student's learning information.

[0030] In step S2, the student's VR device is equipped with a built-in VR program developed based on Pico, primarily used for training students on hidden danger information. After the student logs in, the VR device synchronizes their login information with the server based on the device code. The student then selects a different learning scenario, each of which includes a 1:1 replica of the actual factory site and a corresponding configuration table.

[0031] If each safety notification scenario is loaded using the conventional data loading process, the initial package required for loading is large, resulting in longer loading times, inability to dynamically update resources later, and very limited editing flexibility. Furthermore, scripts for executing notification points and the UI displaying information must be manually added, resulting in increased performance overhead and potentially causing loading lags. In this embodiment, the configuration table is developed based on UnityScriptableObject and defines a SafetyInfo class with the following structure: "desription" represents an overview of the notification point, "index" represents the notification point ID, "group" represents the scene name to which the notification point belongs, "name" represents the notification point name, "reason" represents the learning content of the notification point, and "model" represents the 3D model of the notification point. "index" is an Int parameter, "model" is a Transform parameter, and all other parameters are String types. After all content is configured, it is saved in the program as a persistent resource. Upon opening the student's VR device, the system automatically generates all notification points based on the entered configuration table and places them in the scene. The positions are determined by the 3D model. Specifically, for each learning scenario, the system first clones the 3D model selected from the configuration table "model" directly into the scene, maintaining the same position as the original model. It then automatically adds the controller raycasting script, collision detector, model highlighting script, notification point information storage script, and UI guideline script to the cloned model, automatically configuring all script parameters to complete the safety notification function. After the scripts are added, the system automatically generates the notification point UI, synchronizing the content in the configuration table "reason" with the UI text to display the specific information of the notification point. The generated UI is added to the corresponding notification point model for easy script invocation. Each notification point is generated during the loading phase through the above process. Finally, it is added to the scene one by one in ID order based on the contents of the configuration table "index". This avoids the problem of loading all models at once and allows the configuration table contents to be modified at any time to dynamically update resources. In this embodiment, the use of ScriptableObject to develop the configuration table is primarily due to Unity's extensive support for ScriptableObject compared to traditional JSON or XML, faster loading speed, more flexible configuration, and the ability to better meet enterprise notification modification requirements.

[0032] In step S3, when the trainee starts to learn specific notifications, the eye tracking data is obtained by calling the Pico hardware API interface to determine the trainee's attention to the learning content. In Pico, Pico provides an interface for the development of the eye tracking part, allowing developers to call the eye tracking sensor on Pico for eye tracking when needed, and return the eye data stream captured by the Pico sensor. Specifically, the eye data PXRPose is obtained through the PXR_MotionTracking.GetEyeTrackingData function. The eyeDatas array in PXRPose stores the left eye data and right eye data captured by the device. The data contains the position information and rotation information of the eyeball relative to the camera. Through these two pieces of information, Physics.Raycast is used in Unity to perform ray detection to determine whether the current notification text is being read, to ensure that the trainee is indeed reading the notification that needs to be learned. When the student starts reading the notification text, the audio of the text will also be played. The current student's eye opening and closing status can be obtained through Pico's PXR_MotionTracking.GetEyeOpenness. During the audio playback, if the student's eyes are detected to be away from the text or the eyes are closed for more than 3 seconds, it means that the current notification learning is interrupted and the current notification is considered not learned.

[0033] In step S4, when the current learning is complete, the current learning knowledge is marked as learned and stored locally in the learning data. The learning data is an array that stores the indexes of all learned learning points. For details, refer to the data storage ScriptabeObject structure. When the student completes the learning and exits the scene, the system serializes all the learning data into JSON and uploads it to the server, saving it in the database and storing it in the corresponding folder according to the name ID entered by the student when logging in.

[0034] Furthermore, in step S5, the teacher's client stores all device information under the current domain name and can access the database to obtain personnel information. Furthermore, the teacher can query and generate learning reports for the student. In this embodiment, the teacher and student clients utilize a bidirectional data synchronization mechanism based on HTTP to obtain the various statuses of devices and students within the current network from the server. The teacher also manages database read and write operations, isolating the student and teacher clients from directly accessing and modifying data.

[0035] Specifically, once the teacher's app starts running, it reads student and device data from the database at the configured query interval. A pre-created table, Pico List, is created in the database. This table contains the device serial number (Pico Equipment), device status (Pico Status), login user name (Player Name), login user ID (Player ID), the learning scenario (Subject Enum), and uploaded data (Report Data). Each time a student logs in, the student constructs the URL: https: / / {IP} / gamemanager.php?login&mac={MAC} and sends an HTTP request to the server. https (a secure version of HTTP) represents the protocol, {IP} represents the host IP address, / gamemanager.php represents the requested resource path, and login&mac={MAC} represents the parameter, which in this case is the VR device serial number. The server processes the request and returns a JSON-formatted 'result' field, which can be either 'success' or 'failure'. The student parses the JSON result to determine whether the login succeeded or failed. During the logout phase, the student sends a request with the parameters logout=pico&mac={MAC}. The server queries the database for a Pico Equipment with the same serial number 'mac', sets the device's Pico Status to Idle, and clears the login user's name (Player Name), login user ID (Player ID), and the Subject Enum for the scene being learned. The student also periodically queries the device's status based on the configured time parameters, sending a request with the parameters query=pico&mac={MAC}. The server responds with 'result', 'Player ID', and 'Player Name'. 'result' indicates the device's login status, while 'Player ID' and 'Player Name' represent the user ID and name associated with the device's serial number, as determined by the server's query of the 'mac' field in the database. The student then parses the 'result' field to monitor the device's status in real time. When switching scenes, the student sends a request with the parameters currentscene&mac={}&subject={}. The server updates the learned scene's name in the Subject Enum with the contents of the 'subject' field in the database based on the 'mac' field. When uploading learning data, the student will send a request with the parameters uploadreport&mac={}&data={}. The server will query the serial number 'mac' field in the database and upload the learning content 'data' to the Report Data in the database.

[0036] Once the teacher's app starts running, it reads student and device data from the database at the configured query interval, constructing the URL: https: / / {IP} / gamemanager.php?query=pc&mac={MAC}. The server processes the request and returns JSON-formatted data based on the corresponding device serial number in the database. The teacher's UI visualizes the returned data. Each time a student logs in to a device, the teacher retrieves the player name and player ID from the Pico List database and displays them after the name and ID on the UI. Device status is categorized as "in use" or "idle." The returned Pico Status determines whether the device is logged in. If logged in, the corresponding UI highlight appears. The teacher displays the retrieved scene name (Subject Enum) in the scene name position on the UI, providing real-time visibility into the student's learning status and enabling teacher monitoring. During the logout phase, the teacher sends a request with the parameters "download report&mac={MAC}." The server queries the database for a Pico Equipment with the matching serial number "mac," sets the device status to "idle," and clears the player ID, player name, and subject. When the learning is completed, the teacher sends a request with the parameters download autoreport=pc&mac = {MAC} to the server. The server will process the request and return the ReportData corresponding to the device serial number 'mac' from the database. Finally, the teacher parses the learning data ReportData and stores the learning data in the report template in a serialized manner. The report generation progress on the teacher UI interface will be automatically loaded until the report generation is completed and downloaded to the report path in the settings for storage.

[0037] Furthermore, during the report generation process, the name, ID, and learning scenario name obtained from the database are combined and written to the corresponding location. By comparing with the Scriptable Object notification list stored on the teacher's side (the teacher's list and the list in the program are always consistent), after the comparison is completed, the content that has been learned and the content that has not been learned are determined. The learned content will be listed one by one, and the unlearned content will only display the ID number of the notification point and fill in the report. The complete learning report content will be generated and exported to PDF and Word format files and stored in the report path on the teacher's side. The system automatically generates a security notification archive signature report to record the focus of personnel and confirmation status.

[0038] In addition, in order to verify the effectiveness of the method of the present invention, the training effects of three training methods, PPT, video and VR of this system, were compared and analyzed. The box-line diagram of the training effect is shown in the figure below. Figure 2 As shown. The training effect is the difference in the number of hidden dangers identified before and after training. The horizontal line in the middle of the box in the box plot represents the median; the whiskers extending from the box represent the distribution range of the data, but do not include outliers; the upper and lower boundaries of the box reflect the degree of dispersion of the data; points outside the whiskers in the box plot are usually regarded as outliers. Figure 2 It can be seen that under the VR training method, the training effect is at a relatively high level. VR training has a good effect in improving the relevant abilities of trainees and can enable most trainees to achieve relatively high training results.

[0039] Furthermore, to further clarify the effectiveness of the three training methods, we used the Kirkpatrick Assessment Model's learning layer assessment as an example to analyze the differences in the actual effectiveness of the different training models. The results are shown in Table 1. The data includes mean, standard deviation, Shapiro-Wilk test results, significance level, and effect size.

[0040] Table 1 Comparison of training effectiveness

[0041] The results show that VR training has the best overall effect in improving the difference in the number of hidden dangers identified, followed by the PPT group, and the video group is the worst.

[0042] Example 2 This embodiment discloses a safety notification training system based on VR eye tracking; like Figure 2 As shown, a safety information training system based on VR eye tracking includes: The student information acquisition module is configured to: obtain the personal information input by the student, establish a binding relationship between the student information and the device serial number, and upload the binding relationship information to the server; The training scenario generation module is configured as follows: the student-side VR device obtains personal information from the server according to the device serial number and loads the corresponding training scenario and configuration table; The content notification module is configured to: start the VR training program and collect the trainee's eye tracking data in real time through the VR device's eye tracking API; use the eye tracking data in conjunction with radiographic detection technology to determine whether the trainee is paying attention to the safety notification content; if it is detected that the trainee's gaze deviates from the target content or the duration of eye closure exceeds a preset threshold, the learning is determined to be interrupted; The learning progress recording module is configured to: record the student's learning progress in real time and upload the data to the server; The training report generation module is configured to: use the teacher's end to synchronously obtain student learning data, and automatically generate a compliance training report based on eye movement data, learning records and electronic signatures. Example 3 The purpose of this embodiment is to provide a computer-readable storage medium.

[0043] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a safety notification training method based on VR eye tracking as described in Example 1.

[0044] Example 4 The purpose of this embodiment is to provide an electronic device.

[0045] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps of the safety notification training method based on VR eye tracking as described in Example 1 are implemented.

[0046] The steps involved in the apparatuses of Examples 2, 3, and 4 above correspond to those of Method Example 1. For detailed implementations, please refer to the relevant description of Example 1. The term "computer-readable storage medium" should be understood to mean a single medium or multiple media containing one or more instruction sets; it should also be understood to include any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and causing the processor to perform any method of the present invention.

[0047] Those skilled in the art will appreciate that the modules or steps of the present invention described above can be implemented using a general-purpose computer device. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. Alternatively, they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0048] Although the above describes the specific embodiments of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without any creative work are still within the scope of protection of the present invention.

Claims

1. A safety notification training method based on VR eye tracking, characterized in that: include: S1, obtain the personal information entered by the student, establish a binding relationship between the student information and the serial number of the student's VR device, and upload the binding relationship information to the server; S2: The student's VR device obtains personal information from the server based on the device serial number and loads the corresponding training scenario and configuration table; S3, collects students’ eye tracking data in real time; Based on eye tracking data and radiographic detection technology, it is determined whether the trainee is paying attention to the safety information content. If the trainee's gaze deviates from the target content or the duration of eye closure exceeds the preset threshold, it is considered as a learning interruption. S4, records students’ learning progress in real time and uploads the data to the server; S5 uses the teacher’s end to synchronously obtain students’ learning data and automatically generates compliance training reports based on eye movement data, learning records and electronic signatures.

2. A safety notification training method based on VR eye tracking according to claim 1, characterized in that: The eye tracking data includes position information and rotation information of the eyeball relative to the camera.

3. The safety notification training method based on VR eye tracking according to claim 1, characterized in that: The eye movement data is obtained by calling the PXR_MotionTracking.GetEyeTrackingData function of the VR device, and ray detection is performed using Physics.Raycast to determine whether the trainee's gaze point matches the safety notification content.

4. The safety notification training method based on VR eye tracking according to claim 1, characterized in that: The configuration table of the training scenario is implemented based on Unity's Scriptable Object, including the description, ID, scene name, 3D model and learning text of the safety notification point. The teacher can remotely update the configuration table and synchronize it to the student side.

5. The safety notification training method based on VR eye tracking according to claim 1, characterized in that: Record students' learning progress in real time and upload data to the server, including: When learning is completed, the current informed knowledge is marked as learned locally and stored in the informed learning data; All learning data will be serialized into Json and uploaded to the server, saved in the database, and stored in the corresponding folder according to the student's personal information.

6. The safety notification training method based on VR eye tracking according to claim 1, characterized in that: The server stores student login information, device status and learning records. The learning records include the ID of the learning notification point and the learning completion status. The data is synchronized in real time between the teacher and the student via the HTTP protocol.

7. The safety notification training method based on VR eye tracking according to claim 1, characterized in that: The generation of the training report includes the following sub-steps: Obtain students' electronic signatures and learning records from the server; Compare the learned and unlearned safety notification points according to the configuration table; Fill the comparison results into the preset report template and export them as PDF or Word format files.

8. A safety information training system based on VR eye tracking, characterized by: include: The student information acquisition module is configured to: obtain the personal information input by the student, establish a binding relationship between the student information and the device serial number, and upload the binding relationship information to the server; The training scenario generation module is configured as follows: the student-side VR device obtains personal information from the server according to the device serial number and loads the corresponding training scenario and configuration table; The content notification module is configured to: start the VR training program and collect the trainee’s eye tracking data in real time through the VR device’s eye tracking API; Based on eye tracking data and radiographic detection technology, it is determined whether the trainee is paying attention to the safety information content. If the trainee's gaze deviates from the target content or the duration of eye closure exceeds the preset threshold, it is considered as a learning interruption. The learning progress recording module is configured to: record the student's learning progress in real time and upload the data to the server; The training report generation module is configured to: use the teacher's end to synchronously obtain student learning data, and automatically generate a compliance training report based on eye movement data, learning records and electronic signatures.

9. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps of the safety notification training method based on VR eye tracking are implemented.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the safety notification training method based on VR eye tracking are implemented as described in any one of claims 1 to 7.

Images