Evaluation method of attention allocation mechanism for pedestrians crossing the street based on virtual reality simulation

Abstract

The present invention provides a method for evaluating the attention allocation mechanism of pedestrians crossing the street based on virtual reality simulation, including: step S1, obtaining eye tracking data during the simulated pedestrian crossing process and the trajectory coordinate data of the pedestrians and each moving vehicle; step S2, converting each moving vehicle into an axis-aligned bounding box, and simulating and generating the gaze ray of the pedestrian according to the eye tracking data, and obtaining the maximum vertex value and the minimum vertex value of the axis-aligned bounding box according to the trajectory coordinate data; step S3, based on the gaze ray, obtaining the maximum parameter and the minimum parameter of the gaze ray entering the axis-aligned bounding box according to the maximum vertex value and the minimum vertex value; step S4, matching the maximum parameter and the minimum parameter with the intersection condition to obtain the gaze duration result, so as to characterize the attention allocation of pedestrians when crossing the street. The beneficial effect is that the present invention can solve the defects of the traditional method that the experimental environment is uncontrollable, the data acquisition is not comprehensive, there are safety issues, and the accuracy is low.

Description

Technical Field

[0001] The invention relates to the technical field of modern traffic safety research, and in particular to a method for evaluating a pedestrian crossing attention allocation mechanism based on virtual reality simulation. Background Art

[0002] In the field of modern traffic safety research, virtual reality (VR) technology has gradually become an effective tool. Through VR technology, researchers can build a virtual environment that highly simulates reality, allowing participants to conduct experiments under safe and controllable conditions. Especially in the study of pedestrian crossing behavior, VR technology can simulate various complex traffic scenarios and obtain the behavioral data of participants, thereby analyzing their behavioral patterns and decision-making processes.

[0003] Eye tracking technology has a wide range of applications in the field of human-computer interaction. By capturing and recording the movement of human eyes, researchers can understand how individuals allocate their attention in specific situations. In traffic behavior research, eye tracking technology can reveal how pedestrians look at different target objects in the traffic environment while crossing the street.

[0004] Traditional methods for studying the attention allocation mechanism of pedestrians crossing the street usually rely on field observations and questionnaires. These methods have certain defects. First, the experimental environment is uncontrollable. In a real traffic environment, the experimental conditions are complex and changeable, and it is difficult to control variables, which may affect the accuracy and reliability of the experimental results; second, data acquisition is not comprehensive. Traditional observation and questionnaire survey methods are difficult to comprehensively and accurately obtain various behavioral data of pedestrians, resulting in certain limitations in the research results; third, there are safety hazards. Field experiments may have safety risks, especially when studying pedestrian crossing behavior, participants may face actual traffic dangers; fourth, the repeatability is poor. Due to the inconsistency of experimental conditions, it is difficult for traditional methods to repeat experiments under the same conditions, affecting the repeatability and verifiability of the research. Summary of the invention

[0005] The technical problem to be solved by the present invention is to solve the defects of the traditional method, such as uncontrollable experimental environment, incomplete data acquisition, safety issues and low accuracy. In order to overcome the defects of the above-mentioned prior art (or related technology), the present invention provides a method for evaluating the attention allocation mechanism of pedestrians crossing the street based on virtual reality simulation.

[0006] The present invention provides a method for evaluating the attention allocation mechanism of pedestrians crossing the street based on virtual reality simulation. A virtual simulation environment based on a three-dimensional coordinate system is constructed in advance according to a real pedestrian crossing scene and a plurality of moving vehicles in the pedestrian crossing scene, and a participant is made to wear a VR eye tracking device to simulate pedestrians crossing the street in the virtual simulation environment. The method for evaluating the attention allocation mechanism of pedestrians crossing the street includes the following steps:

[0007] Step S1, obtaining the eye tracking data collected by the VR eye tracking device during the process of simulating pedestrians crossing the street and the trajectory coordinate data of the pedestrians and each of the moving vehicles simulated in the virtual simulation environment;

[0008] Step S2, converting each of the moving vehicles into an axis-aligned bounding box, simulating and generating a gaze ray of a pedestrian according to the eye tracking data, and obtaining a maximum vertex value and a minimum vertex value of the axis-aligned bounding box in a three-dimensional coordinate system according to the trajectory coordinate data;

[0009] Step S3, based on the gaze ray, obtaining a maximum parameter and a minimum parameter of the gaze ray entering the axis-aligned bounding box according to the maximum value and the minimum value of the vertex;

[0010] Step S4, matching the maximum parameter and the minimum parameter with a pre-configured intersection condition to obtain a corresponding gaze duration result to characterize the attention allocation of pedestrians when crossing the street.

[0011] Compared with the prior art, the pedestrian crossing attention allocation mechanism evaluation method based on virtual reality simulation has the following advantages:

[0012] The present invention takes into account that the traditional method for measuring pedestrian crossing indicators relies on field measurement or video viewing, and there are problems such as uncontrollable experimental environment, incomplete data acquisition and greater danger. Therefore, a virtual simulation environment based on a three-dimensional coordinate system is constructed according to actual pedestrian crossing scenes and multiple moving vehicles in pedestrian crossing scenes. Under the premise of ensuring safety, the controllability and immersion of the pedestrian simulation crossing process are improved, and the data obtained by using a VR eye tracking device is more comprehensive; at the same time, the present invention takes into account that the method of simply judging whether a pedestrian is looking at a vehicle by the coordinates of pedestrians and moving vehicles in a virtual simulation scene has low accuracy. Therefore, by defining the spatial range of the moving vehicle as an axis-aligned bounding box, the intersection of the pedestrian's gaze ray and the axis-aligned bounding box is calculated to obtain the gaze duration result, thereby improving the accuracy, and solving the defects of uncontrollable experimental environment, incomplete data acquisition, safety issues and low accuracy in the traditional method.

[0013] In a possible implementation, the step S1 includes:

[0014] Step S11, defining the gaze source coordinates of the pedestrian in the virtual simulation environment, obtaining the position coordinates where the gaze ray of the pedestrian falls as the gaze point coordinates through the VR eye tracking device, and using the gaze source coordinates and the gaze point coordinates as the eye tracking data;

[0015] Step S12, collecting the x-direction coordinate value, y-direction coordinate value, z-direction coordinate value of each of the traveling vehicles in the three-dimensional coordinate system and the length, width, and height of each of the traveling vehicles as the trajectory coordinate data.

[0016] In a possible implementation, in step S2, the gaze ray is generated by the following calculation formula:

[0017] ;

[0018] R(p) represents the coordinate vector of any point on the gaze ray;

[0019] o represents the gaze source coordinates, ;

[0020] d represents the direction vector between the normalized gaze source coordinates and the gaze point coordinates, ;

[0021] p represents a parameter value from the gaze source coordinates to the intersection of the gaze ray and the bounding box on the corresponding axis.

[0022] In a possible implementation, in step S2, the maximum value of the vertex and the minimum value of the vertex are obtained by the following calculation formula:

[0023] ;

[0024] ;

[0025] in,

[0026] represents the maximum value of the vertex;

[0027] represents the minimum value of the vertex;

[0028] The x-direction coordinate value of the moving vehicle corresponding to the axis-aligned bounding box;

[0029] The y-direction coordinate value of the moving vehicle corresponding to the axis-aligned bounding box;

[0030] The z-direction coordinate value of the moving vehicle corresponding to the axis-aligned bounding box;

[0031] L Indicates the length of the moving vehicle corresponding to the axis-aligned bounding box;

[0032] WIndicates the width of the moving vehicle corresponding to the axis-aligned bounding box;

[0033] H Indicates the height of the moving vehicle corresponding to the axis-aligned bounding box.

[0034] In a possible implementation, in step S3, the minimum parameter and the maximum parameter are obtained by the following calculation formula:

[0035] ;

[0036] ;

[0037] ;

[0038] ;

[0039] Indicates the parameter value when the ray in each axis (x, y, z) direction intersects with the minimum face of the bounding box;

[0040] Represents the parameter value of the minimum value of the vertex in each axis (x, y, z) direction;

[0041] Represents the parameter value of the gaze source coordinates in each axis (x, y, z) direction;

[0042] Indicates the parameter value when the ray in each axis (x, y, z) direction intersects with the maximum face of the bounding box;

[0043] Represents the parameter value of the maximum value of the vertex in each axis (x, y, z) direction;

[0044] The parameter range representing the earliest time that the ray enters the bounding box in each direction;

[0045] The parameter range representing the latest time at which the ray leaves the bounding box in each direction;

[0046] The farthest parameter value representing the ray leaving the bounding box is used as the maximum parameter;

[0047] The closest parameter value representing the ray entering the bounding box is used as the minimum parameter;

[0048] The parameter value representing the direction vector between the normalized gaze source coordinates and the gaze point coordinates in each axis (x, y, z) direction.

[0049] In a possible implementation manner, before executing step S4, the method further includes:

[0050] Count the error values ​​obtained in step S3, and replace all the error values ​​with or .

[0051] In a possible implementation manner, the preconfigured intersection condition includes two matching items, and step S4 includes:

[0052] Step S41, determining whether the maximum parameter and the minimum parameter satisfy both the matching items:

[0053] If yes, outputting a matching result representing the intersection of the gaze ray and the axis-aligned bounding box;

[0054] If not, outputting a matching result indicating that the gaze ray does not intersect the axis-aligned bounding box;

[0055] Step S42, counting the number of the matching results representing the intersection of the gaze ray and the axis-aligned bounding box, and multiplying the number by a pre-configured motion capture frequency of the VR eye tracking device to calculate an effective gaze duration of the pedestrian on the moving vehicle;

[0056] Step S43, counting the total gaze time of pedestrians in the process of simulating pedestrians crossing the street, and taking the ratio of the effective gaze time to the total gaze time as the gaze time result.

[0057] In a possible implementation, in step S41, the expressions of the two matching items are as follows:

[0058] ;

[0059] The farthest parameter value representing the ray leaving the bounding box is used as the maximum parameter;

[0060] The closest parameter value that the ray enters the bounding box is represented as the minimum parameter. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is a flow chart of the steps of the present invention;

[0062] Figure 2 is a specific flow chart of step S1 of the present invention;

[0063] Figure 3 A schematic diagram of the origin of the coordinate axes of a pedestrian of the present invention;

[0064] Figure 4 A schematic diagram of the origin of the coordinate axes of a moving vehicle of the present invention;

[0065] Figure 5 is a specific flow chart of step S4 of the present invention;

[0066] Figure 6 A schematic diagram of the intersection of a pedestrian's gaze ray and an axis-aligned bounding box according to the present invention;

[0067] Figure 7 This is a histogram showing the comparison of gaze points for single-point distance discrimination and axis-aligned bounding box discrimination of the present invention. DETAILED DESCRIPTION

[0068] First, those skilled in the art should understand that these implementations are only used to explain the technical principles of the embodiments of the present invention, and are not intended to limit the protection scope of the embodiments of the present invention. Those skilled in the art can make adjustments to them as needed to adapt to specific application scenarios.

[0069] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0070] See also Figure 1 The embodiment of the present invention discloses a method for evaluating the attention allocation mechanism of pedestrians crossing the street based on virtual reality simulation. By constructing a virtual simulation environment, a real pedestrian crossing scene and a moving vehicle are built. The eye tracking data of the participants in the simulated pedestrian crossing process, especially the gaze point coordinates of the gaze ray, are recorded by having the participants wear VR eye tracking equipment. At the same time, the physical coordinates of the pedestrians in the virtual simulation environment can also be obtained. By calculating whether the gaze ray of the pedestrian in the virtual simulation environment intersects with the spatial range of the moving vehicle, it is determined whether the participant has looked at the moving vehicle. Finally, the gaze duration results of the participants are summarized to analyze and evaluate the attention allocation mechanism of pedestrians crossing the street.

[0071] Continue to see Figure 1 The specific steps of the pedestrian crossing attention allocation mechanism evaluation method in the embodiment of the present invention are:

[0072] Step S1, obtaining eye tracking data collected by a VR eye tracking device during a simulated pedestrian crossing process and trajectory coordinate data of pedestrians and vehicles simulated in a virtual simulation environment;

[0073] Step S2, converting each moving vehicle into an axis-aligned bounding box, simulating and generating a gaze ray of a pedestrian according to the eye tracking data, and obtaining a maximum vertex value and a minimum vertex value of the axis-aligned bounding box in a three-dimensional coordinate system according to the trajectory coordinate data;

[0074] Step S3, based on the gaze ray, obtaining the maximum parameter and the minimum parameter of the gaze ray entering the axis-aligned bounding box according to the maximum value and the minimum value of the vertex;

[0075] Step S4, matching the maximum parameter and the minimum parameter with the pre-configured intersection condition to obtain the corresponding gaze duration result, so as to characterize the attention allocation of pedestrians when crossing the street.

[0076] Continue to see Figure 1 The virtual simulation environment in the present invention is implemented based on the virtual environment development engine. During the experiment, the VR eye tracking device is connected to the virtual simulation platform for subsequent operations. The virtual simulation environment in the experiment is a two-way four-lane scene without traffic lights. The initial position of the pedestrian is on one side of the lane. Facing the oncoming vehicles, the pedestrian needs to cross the street at a time he thinks is appropriate. The entire experimental process built on the virtual environment development engine includes the following four steps:

[0077] Step 1: Create the starting and ending points for pedestrians to cross the street;

[0078] Step 2: Create a trigger point (task node) for the vehicle to move.

[0079] Step 3: Draw the driving route and driving status of the moving vehicle;

[0080] Step 4: Connect each task module;

[0081] During the experiment, participants wore VR eye tracking devices and used joysticks to control the virtual characters in the virtual simulation scene to walk freely and interact with moving vehicles in the virtual simulation scene. During this process, the VR eye tracking device could track the participants' eye movements in real time, thereby collecting eye tracking data. At the same time, the virtual environment development engine would also record the position coordinate data of the virtual characters and moving vehicles in the scene.

[0082] See also Figure 2 , step S1 comprises:

[0083] Step S11, defining the gaze source coordinates of the pedestrian in the virtual simulation environment, obtaining the position coordinates where the gaze ray of the pedestrian falls as the gaze point coordinates through the VR eye tracking device, and using the gaze source coordinates and the gaze point coordinates as eye tracking data;

[0084] Step S12, collecting the x-direction coordinate value, y-direction coordinate value, z-direction coordinate value of each traveling vehicle in the three-dimensional coordinate system and the length, width and height of each traveling vehicle as trajectory coordinate data.

[0085] Continue to see Figure 2 , the sizes of pedestrians and vehicles in the virtual environment development engine are defined as follows:

[0086] The height of a pedestrian is defined as 170 cm. Figure 3 The origin of the coordinate axis shown in the figure represents the coordinates of the pedestrian in the scene (Track), the physical coordinates of the pedestrian itself are the gaze source coordinates of the pedestrian's gaze ray (GazeOrigin), and the position coordinates of the pedestrian's gaze ray obtained by the VR eye tracker are the gaze point coordinates (EyeHitSite), where: Figure 3 There is no limitation on the directions of the x-axis, y-axis and z-axis, which can be adjusted according to the actual situation.

[0087] The vehicle is defined as 568cm in length, 198cm in width, and 150cm in height. Figure 4 The origin of the coordinate axis shown represents the coordinates of the moving vehicle in the scene (VehicleTrack), where Figure 4 There is no limitation on the directions of the x-axis, y-axis and z-axis, which can be adjusted according to the actual situation.

[0088] Continue to see Figure 2 After the simulation of pedestrian crossing is completed, the trajectory coordinates of pedestrians and vehicles can be exported in csv format through the virtual environment development engine. The coordinate data in these files correspond to each other according to the synchronized time. The pedestrian trajectory file contains the time node, the x-coordinate, y-coordinate, and z-coordinate of the pedestrian at the current moment, as well as the x-coordinate, y-coordinate, and z-coordinate of the position where the gaze ray reaches at the current moment; the vehicle trajectory file contains the time node, the ID of the current vehicle, and the x-coordinate, y-coordinate, and z-coordinate of the vehicle at the current moment; using the time node, the pedestrian trajectory file is merged with the trajectory file of the vehicle at the same time, and finally a file containing the time node, the vehicle ID, the vehicle position, and the pedestrian and gaze point positions is obtained, which is used as the basis for subsequent numerical calculations.

[0089] Continue to see Figure 1Based on the existing data, simply calculating and comparing the distance between the gaze ray and the position of the moving vehicle cannot accurately determine whether the pedestrian is actually looking at the moving vehicle. The present invention uses the size data of the moving vehicle, defines the moving vehicle as an axis-aligned bounding box (AABB), and calculates the intersection of the gaze ray and the AABB model to more accurately determine the gaze of the pedestrian. The gaze ray involved in this process can be expressed as:

[0090] ;

[0091] Among them, R( p ) is the coordinate vector of any point on the gaze ray, o is the gaze source coordinate, i.e. the current coordinate value of the pedestrian, ; d is the normalized direction vector of the look-at ray, ; p The parameter value representing the intersection of the gaze ray from the gaze source coordinates with the bounding box on the corresponding axis.

[0092] Continue to see Figure 1 , the spatial range of the moving vehicle is defined as an axis-aligned bounding box (AABB), whose minimum vertex value (a vertex of the box) and maximum vertex value (the diagonal vertex of the box) are:

[0093] ;

[0094] ;

[0095] in, and are the maximum and minimum vertex values ​​of the axis-aligned bounding box, respectively. , and are the center coordinates of the moving vehicle, and L, W, and H are the length, width, and height of the moving vehicle.

[0096] Continue to see Figure 1 , the formula for calculating the intersection parameters of the pedestrian's gaze ray with the axis-aligned bounding box boundary on each coordinate axis is as follows:

[0097] ;

[0098] in, Represents the distance vector from the origin of the look ray to the minimum value of the AABB model vertex. Represents the distance vector from the origin of the gaze ray to the maximum value of the AABB model vertex. The above two formulas need to be calculated on the x, y, and z axes respectively;

[0099] In order to handle the possible division by zero in the calculation process, all error (NAN) values ​​in the above calculation process are replaced by or The specific replacement method is expressed as as well as ;

[0100] Then the parameter range of the gaze ray entering the AABB model and the parameter range of the gaze ray leaving the AABB model are calculated, which can be expressed as , so the closest parameter that the gaze ray enters the AABB model is the largest in all axis directions. , expressed as , the farthest parameter of the ray leaving the AABB model, that is, the smallest in all axis directions , expressed as .

[0101] See also Figure 5 , the pre-configured intersection condition includes two matching items, and step S4 includes:

[0102] Step S41, determine whether the maximum parameter and the minimum parameter satisfy two matching items at the same time:

[0103] If yes, output the matching result representing the intersection of the gaze ray and the axis-aligned bounding box;

[0104] If not, output the matching result indicating that the gaze ray does not intersect the axis-aligned bounding box;

[0105] Step S42, counting the number of matching results representing the intersection of the gaze ray and the axis-aligned bounding box, and multiplying the number by the motion capture frequency of the pre-configured VR eye tracking device to calculate the effective gaze time of the pedestrian on the moving vehicle;

[0106] Step S43, counting the total gaze time of pedestrians in the process of simulating pedestrians crossing the street, and taking the ratio of the effective gaze time to the total gaze time as the gaze time result.

[0107] Continue to see Figure 5 In step S41, the intersection condition is determined. Specifically, if the following conditions are met at the same time, it is considered that the gaze ray intersects with the spatial range of the vehicle, otherwise it does not intersect:

[0108] ;

[0109] .

[0110] See also Figure 6, the first intersection point on the right side of the figure indicates the position of the pedestrian, the second intersection point indicates the intersection point where the gaze ray enters the AABB model, the third intersection point indicates the intersection point where the gaze ray leaves the AABB model, and the dotted line segment indicates the part of the gaze ray inside the AABB model;

[0111] After the calculation is completed, the gaze ray is intersected with the spatial range of the moving vehicle, that is, the coordinate ID of the moving vehicle at the current moment is returned as 1, and the coordinate ID of the vehicle not being gazed is returned as 0. Then the number of coordinate IDs with a return value of 1 is summarized and output. This value is the number of times the pedestrian gazes at the moving vehicle during the entire model pedestrian crossing process. The motion capture frequency of the VR eye tracking device for eye movement is 25ms / time. The multiplication of the number of gazes and the motion capture frequency is the effective gaze time of the pedestrian on the moving vehicle. The ratio of the effective gaze time to the total gaze time of the pedestrian recorded in the entire model pedestrian crossing process can reflect the attention distribution of the pedestrian when crossing the street.

[0112] See Table 1 below:

[0113] Table 1 Error rate table of single point distance judgment and axis-aligned bounding box judgment

[0114] ;

[0115] In order to verify the accuracy of the method of the present invention, the AOI (area of ​​interest) eye movement data that can be directly measured is used for verification. The gaze situation of the AOI is calculated by using the present method and the traditional single-point distance judgment method. The calculation results are compared and analyzed with the standard AOI data to obtain the error rate table of single-point distance judgment and axis-aligned bounding box judgment as shown in Table 1 above. The results show that the average absolute error of the axis-aligned bounding box method for calculating and judging the gaze situation of pedestrians on moving vehicles is reduced by 71.45% compared with the single-point distance judgment method. At the same time, Figure 7 As shown, in the obtained gaze point coordinates, the accuracy is improved by 105 gaze point units on average; it can be seen that the method described in the present invention more accurately simulates the actual shape and size of the moving vehicle by using the axis-aligned bounding box (AABB), and can better reflect the actual space occupancy of the moving vehicle compared to the single-point distance judgment method, overcomes the defect that the single-point distance judgment easily ignores the spatial extension and shape characteristics of the moving vehicle and thus causes misjudgment, and is more in line with the actual situation in judging the gaze of pedestrians on the vehicle, thereby improving the accuracy of the judgment.

[0116] In the description of the present invention, the description with reference to the terms "one embodiment", "some embodiments", "in the present embodiment", "specific example", or "some examples" etc. means that the specific features, mechanisms, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, mechanisms, materials or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0117] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for evaluating pedestrian attention allocation mechanism when crossing a street based on virtual reality simulation, characterized in that: A virtual simulation environment based on a three-dimensional coordinate system is constructed in advance according to a real pedestrian crossing scene and multiple vehicles in the pedestrian crossing scene, and the participants are made to wear VR eye tracking devices to simulate pedestrian crossing in the virtual simulation environment. The pedestrian crossing attention allocation mechanism evaluation method includes the following steps: Step S1, obtaining the eye tracking data collected by the VR eye tracking device during the process of simulating pedestrians crossing the street and the trajectory coordinate data of the pedestrians and each of the moving vehicles simulated in the virtual simulation environment; Step S2, converting each of the moving vehicles into an axis-aligned bounding box, simulating and generating a gaze ray of a pedestrian according to the eye tracking data, and obtaining a maximum vertex value and a minimum vertex value of the axis-aligned bounding box in a three-dimensional coordinate system according to the trajectory coordinate data; Step S3, based on the gaze ray, obtaining a maximum parameter and a minimum parameter of the gaze ray entering the axis-aligned bounding box according to the maximum value and the minimum value of the vertex; Step S4, matching the maximum parameter and the minimum parameter with a pre-configured intersection condition to obtain a corresponding gaze duration result to characterize the attention allocation of pedestrians when crossing the street; The preconfigured intersection condition includes two matching items, and step S4 includes: Step S41, determining whether the maximum parameter and the minimum parameter satisfy both the two matching items: If yes, outputting a matching result representing the intersection of the gaze ray and the axis-aligned bounding box; If not, outputting a matching result indicating that the gaze ray does not intersect the axis-aligned bounding box; Step S42, counting the number of the matching results representing the intersection of the gaze ray and the axis-aligned bounding box, and multiplying the number by a pre-configured motion capture frequency of the VR eye tracking device to calculate an effective gaze duration of the pedestrian on the moving vehicle; Step S43, counting the total gaze time of pedestrians in the process of simulating pedestrians crossing the street, and taking the ratio of the effective gaze time to the total gaze time as the gaze time result; In step S41, the expressions of the two matching items are as follows: ； The maximum parameter represents the farthest parameter value of the ray from the bounding box of the axis-aligned bounding box; The closest parameter value representing the ray entering the bounding box of the axis-aligned bounding box is used as the minimum parameter.

2. The method for evaluating pedestrian crossing attention allocation mechanism based on virtual reality simulation according to claim 1 is characterized in that: The step S1 comprises: Step S11, defining the gaze source coordinates of the pedestrian in the virtual simulation environment, obtaining the position coordinates where the gaze ray of the pedestrian falls as the gaze point coordinates through the VR eye tracking device, and using the gaze source coordinates and the gaze point coordinates as the eye tracking data; Step S12, collecting the x-direction coordinate value, y-direction coordinate value, z-direction coordinate value of each of the traveling vehicles in the three-dimensional coordinate system and the length, width, and height of each of the traveling vehicles as the trajectory coordinate data.

3. The method for evaluating pedestrian crossing attention allocation mechanism based on virtual reality simulation according to claim 2 is characterized in that: In step S2, the gaze ray is generated by the following calculation formula: ； R(p) represents the coordinate vector of any point on the gaze ray; o represents the gaze source coordinates, ; d represents the direction vector between the normalized gaze source coordinates and the gaze point coordinates, ; p represents a parameter value from the gaze source coordinates to the intersection of the gaze ray and the bounding box on the corresponding axis.

4. The method for evaluating pedestrian crossing attention allocation mechanism based on virtual reality simulation according to claim 2 is characterized in that: In step S2, the maximum value and the minimum value of the vertex are obtained by the following calculation formula: ； ； in, represents the maximum value of the vertex; represents the minimum value of the vertex; The x-direction coordinate value of the moving vehicle corresponding to the axis-aligned bounding box; The y-direction coordinate value of the moving vehicle corresponding to the axis-aligned bounding box; The z-direction coordinate value of the moving vehicle corresponding to the axis-aligned bounding box; L represents the length of the moving vehicle corresponding to the axis-aligned bounding box; W Indicates the width of the moving vehicle corresponding to the axis-aligned bounding box; H Indicates the height of the moving vehicle corresponding to the axis-aligned bounding box.

5. The method for evaluating pedestrian crossing attention allocation mechanism based on virtual reality simulation according to claim 2 is characterized in that: In step S3, the minimum parameter and the maximum parameter are obtained by the following calculation formula: ； ； ； ； Indicates the parameter value when the ray in each axis (x, y, z) direction intersects with the minimum face of the bounding box; Represents the parameter value of the minimum value of the vertex in each axis (x, y, z) direction; Represents the parameter value of the gaze source coordinates in each axis (x, y, z) direction; Indicates the parameter value when the ray in each axis (x, y, z) direction intersects with the maximum face of the bounding box; Represents the parameter value of the maximum value of the vertex in each axis (x, y, z) direction; The parameter range representing the earliest time that the ray enters the bounding box in each direction; The parameter range representing the latest time at which the ray leaves the bounding box in each direction; The farthest parameter value representing the ray leaving the bounding box is used as the maximum parameter; The closest parameter value representing the ray entering the bounding box is used as the minimum parameter; The parameter value representing the direction vector between the normalized gaze source coordinates and the gaze point coordinates in each axis (x, y, z) direction.

6. The method for evaluating pedestrian crossing attention allocation mechanism based on virtual reality simulation according to claim 1 is characterized in that: Before executing step S4, the method further includes: Count the error values ​​obtained in step S3, and replace all the error values ​​with or .

Images