Road marking optimization design method based on eye movement

Abstract

The invention relates to an eye movement-based road marking optimization design method, which comprises the following steps of: acquiring eye movement data of a driver in a driving process by using an eye tracker, analyzing data such as a fixation point, fixation time and frequency, glancing amplitude, glancing duration, pupil area, pupil area change rate and the like, calculating indexes such as eye movement discrete degree and the like, and evaluating the visual load of the driver. The method comprises the following steps: identifying a marking line with high attention of a driver, judging the marking line as an important marking line and a problem marking line, analyzing the problems of improper marking line design, unobvious position or confusion with other road features and the like, following design specifications and principles of traffic marking lines, and adjusting key parameters of the marking line, such as angle, spacing, color and line width. And based on the adjusted road marking, collecting eye movement data by using an eye tracker by using a virtual environment simulation driving technology, calculating the visual load of the driver on the adjusted marking, and evaluating the optimization effect. According to the road marking optimization design method based on the eye movement data, marking lines with high attention degree of a driver are recognized by analyzing the watching characteristics of the driver on the road marking lines, the cause of problem marking lines is further analyzed, optimization and virtual driving verification are conducted on the problem marking lines, and therefore the road marking optimization design method based on the eye movement data is given.

Description

Technical Field

[0001] The present invention relates to the field of engineering technology, and particularly to a method for optimizing road markings based on eye movement data, providing an analysis basis for optimizing road traffic facilities. Background Art

[0002] In the ever-changing progress of the transportation field, the importance of traffic safety has become increasingly prominent, prompting the scientific research community and the industrial community to continuously deepen the exploration of improving road safety. Road traffic markings are an important part of traffic safety facilities. Optimizing these facilities can effectively reduce the incidence of traffic accidents. With the continuous growth and improvement of highways at all levels across the country, traffic accidents have also become a by-product of traffic development. Due to the high driving speed on high-grade highways, traffic accidents usually result in terrible consequences. There are many ordinary highways in the country, and the operating speeds vary greatly on different sections, which easily causes accidents such as rear-end collisions. With the rapid development of technology, as a cutting-edge innovation, eye-catching technology is gradually being applied to the research and practice of driving safety. At the same time, advanced machine learning technology is combined to identify and analyze traffic targets in the road environment ahead. By deeply integrating the identified traffic targets with the driver's eye gaze data, the fixation index of the driver on the road markings is detected. This provides a scientific basis for evaluating the effectiveness of road traffic markings, ensuring that they are effectively perceived by drivers, thereby further optimizing the traffic environment design and reducing traffic accidents to a certain extent.

[0003] Currently, there are few methods for optimizing road traffic markings, especially in the application of assisted driving based on eye movement data. Most existing methods use machine learning technology for marking recognition and damage repair, usually relying on a large amount of labeled data, resulting in limited recognition accuracy and efficiency. Although eye-catching technology has gradually been applied in the research of driving safety, existing methods rarely deeply integrate the driver's eye movement data with road marking recognition, failing to fully utilize eye movement data to evaluate the fixation characteristics and visual load of drivers on road markings, and there is a scarcity in the application of optimizing the design of road markings.

[0004] Therefore, the present invention proposes an optimized design method for road markings based on eye movement. By calculating indicators such as the degree of eye movement dispersion, the visual load of the driver is evaluated, and the road markings that need to be optimized are identified, thereby improving road safety and meeting the needs of the assisted driving system for perceiving road markings in the traffic environment. By studying the fixation characteristics of drivers on road markings and calculating the visual load, the optimization adjustment of road traffic markings is carried out accordingly. Summary of the Invention

[0005] The present invention proposes an optimized design method for road markings based on eye movement, aiming to collect eye movement data of drivers during driving, study the fixation characteristics of drivers on road markings, provide an analysis basis for the design and optimization of road markings, optimize the design of road markings, and verify the optimization effect through virtual driving means.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] An optimized design method for road markings based on eye movement, comprising:

[0008] Collect eye movement data of drivers during driving, study the fixation characteristics of drivers on road markings, evaluate the visual load of drivers on road markings, classify the road markings that receive much attention into important markings and problem markings, conduct problem analysis and appropriate adjustment on the problem markings, based on the adjusted road markings, collect the eye movement data of drivers again in a virtual driving environment, evaluate the visual load of drivers on road markings, and verify the effectiveness of the optimized design of road markings.

[0009] 1. An optimized design method for road markings based on eye movement, characterized in that it comprises:

[0010] Collect eye movement data of drivers during driving, study the fixation characteristics of drivers on road markings, evaluate the visual load of drivers on road markings, classify the road markings that receive much attention into important markings and problem markings, conduct problem analysis and appropriate adjustment on the problem markings, based on the adjusted road markings, collect the eye movement data of drivers again in a virtual driving environment, evaluate the visual load of drivers on road markings, and verify the effectiveness of the optimized design of road markings.

[0011] 2. The optimized design method for road markings based on eye movement according to claim 1, characterized by comprising the following steps:

[0012] Step S1: Use an eye tracker to collect eye movement data, including fixation points, fixation time and frequency, saccade amplitude, saccade duration, pupil area, and pupil area change rate.

[0013] Step S2: Calculate indexes such as eye movement dispersion degree to evaluate the visual load of drivers.

[0014] Step S3: Analyze the eye movement data indexes, identify the road markings with high driver attention, and determine important markings and problem markings.

[0015] Step S4: Analyze that the problem markings are improperly designed, not obvious in position, or confused with other road features, etc.

[0016] Step S5: Follow the design specifications and principles of traffic signs, and adjust the key parameters of the road markings, such as angle, spacing, color, and line width.

[0017] Step S6: Use an eye tracker to test the driver's reaction, and utilize virtual driving to calculate the driver's visual load during the adjustment of the road markings to verify the optimization effect.

[0018] 3. The method for optimizing the design of road markings based on eye movement according to claim 2, wherein the method for distinguishing important road markings and problematic road markings in step S3 is as follows:

[0019] Important road markings refer to those directly related to the driving task, which provide important cognition that drivers must pay attention to during driving and occupy a necessary position in the driver's vision. Important road markings are usually within the driver's central visual field and should have obvious visual characteristics so that drivers can quickly identify and obtain important information. Since the driver's cognition is limited, the design of important road markings should convey necessary driving information to the driver without increasing additional visual load. Problematic road markings refer to those related to the driving task, but due to certain reasons (such as improper design, inconspicuous position, confusion with other road features, etc.), the information they convey is deviated, resulting in drivers paying excessive attention to them. Problematic road markings will cause drivers to pay excessive attention to them, increase visual load, and affect driving safety.

[0020] The design purpose of important road markings is for quick recognition. The fixation time is usually not too long, maintained within a reasonable range, and the fixation frequency should be moderate, not leading to excessive visual search. At the same time, the change rate of pupil area remains within the normal range, indicating that the driver's attention to the road markings is necessary rather than unusual. Problematic road markings will cause an increase in the driver's fixation time because drivers need to spend more time understanding these road markings, and will also cause an increase in the fixation frequency, reflecting the driver's excessive attention or confusion. At the same time, it will cause an increase in the change rate of pupil area, indicating that the driver's attention to these road markings is unusual.

[0021] Therefore, if a road marking has a long fixation time and a high fixation frequency, but the change rate of pupil area is normal and it is directly related to the driving task, then it is determined that the road marking is an important road marking. If a road marking has a long fixation time and a high fixation frequency, and the change rate of pupil area is large, indicating that the driver pays excessive attention or is distracted by it, then it is determined that the road marking is a problematic road marking.

[0022] 4. The method for optimizing the design of road markings based on eye movement according to claim 2, wherein the determination method for analyzing improper design, inconspicuous position, or confusion with other road features of problematic road markings in step S4 is as follows:

[0023] The fixation heatmap is used to display the concentrated area of the driver's fixation. By mapping the fixation point data onto a two-dimensional plane, the depth of the displayed color represents the fixation density. Poorly designed road markings can lead to abnormal fixation distributions, which can be visually shown through the fixation heatmap. If the fixation density in a certain road marking area is abnormally high, it indicates that the road marking is poorly designed, causing the driver to over-focus. On the contrary, if the fixation density in a certain road marking area is abnormally low, it indicates that the position of the road marking is not obvious and fails to effectively guide the driver's visual attention. Poorly designed road markings can lead to abnormal saccades. If a certain road marking causes the driver's saccade amplitude to be abnormally large or the saccade duration to be abnormally long, it indicates that the road marking is poorly designed, resulting in a decrease in the driver's visual search efficiency. Poorly designed road markings can lead to an increase in cognitive load, causing an increase in the pupil change rate.

[0024] The confusion matrix can analyze the accuracy of the driver's recognition of road markings. Poorly designed road markings can lead to an increase in recognition errors and reduce the accuracy of road marking recognition. The gaze heatmap can display the concentrated area of the driver's gaze. Road markings with an inconspicuous position will result in sparse fixation points shown in the gaze heatmap.

[0025] Using image processing technology, visual features such as the features of road markings and the surrounding environment are extracted. These features cover color, shape, texture, contrast, etc. The purpose of feature extraction is to identify the fundamental features in the image, such as edges, corners, textures, etc. By calculating the confusion index between the road markings and other features, the degree of confusion of the road markings is evaluated. A high confusion index indicates that the road markings are confused with other road features.

[0026] 5. The method for optimizing the design of road markings based on eye movement according to claim 2, wherein the method for adjusting the key parameters of the road markings in step S5 is as follows:

[0027] Following the traffic road marking design specifications and principles, adjust the key parameters of the road markings, such as angle, spacing, color, and line width. Through on-site measurement or relevant road marking design documents, collect the parameters of the existing road markings, such as angle, spacing, color, and line width. Evaluate whether the existing road marking parameters comply with the traffic road marking design specifications and principles, including evaluating the visibility, recognition accuracy, and visual interference of the road markings. According to the traffic road marking design specifications, determine the optimal angle, spacing, color, and line width of the road markings, calculate the adjusted angle, adjusted spacing, color contrast, and adjusted line width, and perform the optimization design of the road markings based on this.

[0028] 6. The method for optimizing the design of road markings based on eye movement according to claim 2, wherein the method for using virtual driving to calculate the driver's visual load on the adjusted road markings to verify the optimization effect in step S6 is as follows:

[0029] Based on the adjusted road markings, use an eye tracker to test the driver's reaction according to virtual environment simulation driving technology, calculate the visual load of the driver on the adjusted markings, and evaluate the optimization effect. Use 3D modeling software to create a 3D model of the road environment. The model covers road geometry, markings, traffic signs, vehicles, and other relevant elements. During the modeling process, ensure the accuracy and adjustability of the road markings for subsequent optimization experiments. Apply real textures and materials to the 3D model to enhance the environmental realism. Set the marking textures according to the actual marking materials and colors to ensure similar visual characteristics in the virtual environment as in reality. Simulate real-world lighting conditions, including natural light and artificial light sources. Explore the lighting changes at different times of the day to evaluate the visibility of the markings under different lighting conditions. Design interactive elements in the virtual environment, such as adjustable marking parameters (angle, spacing, color, line width), for optimization experiments.

[0030] Select the hardware and software platforms of the driving simulator to assist in high-precision driving behavior simulation. Integrate the eye tracker with the simulation driving system to collect the driver's eye movement data in real time. The eye tracker has a high sampling rate and accuracy to ensure the accuracy and reliability of the eye movement data. Ensure the data synchronization of the simulation driving system and the eye tracker for accurate timestamping and event marking. Explore the simulator latency and the calibration accuracy of the eye tracker to ensure data consistency and comparability. Design the simulation driving experiment scenarios and tasks to evaluate the driver's reaction to the adjusted markings. The experiment scenarios cover the marking designs of traffic conditions under different road conditions to comprehensively evaluate the optimization effect of the markings.

[0031] In the simulation driving experiment, collect the driver's eye movement data in real time, including fixation points, fixation time and frequency, saccade amplitude, saccade duration, pupil area, and pupil area change rate. Use the collected fixation point data to generate a fixation point heat map through the Gaussian kernel smoothing algorithm to display the areas where the driver's fixation is concentrated and reveal the driver's attention to different marking areas. Identify abnormal visual search patterns by calculating the saccade amplitude and saccade duration. Calculate the visual load of the driver on the adjusted markings to evaluate the actual optimization effect of the markings.

[0032] Compare the experimental data before and after optimization, and conduct statistical analysis on the experimental data to determine the significance and actual effect of the marking optimization. Based on the statistical analysis of the experimental results, put forward claims for marking optimization to guide the design and optimization of actual road markings.

[0033] The beneficial effects of the present invention are:

[0034] By analyzing the eye movement data, it helps to analyze the driver's attention to road markings during actual driving, and based on this, optimize the design of road markings. By comparing the eye movement characteristics of drivers on unadjusted and adjusted road markings in a virtual driving environment, it is verified that the optimized design method of road markings is more targeted and effective, thus effectively improving road driving safety. Description of the Drawings

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0036] Figure 1 It is a flowchart of the implementation of the optimized design method of the road marking of the present invention. Detailed Description of the Invention

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0038] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0039] As Figure 1 shown, the present invention discloses an optimized design method of road markings based on eye movement, including:

[0040] Collect the eye movement data of the driver during driving, study the fixation characteristics of the driver on the road markings, evaluate the visual load of the driver on the road markings, classify the highly concerned road markings into important markings and problem markings, analyze the problems of the problem markings and make appropriate adjustments. Based on the adjusted road markings, collect the eye movement data of the driver again in a virtual driving environment, evaluate the visual load of the driver on the road markings, and verify the effectiveness of the optimized design of the road markings.

[0041] Eye movement data, including fixation points, fixation duration and frequency, saccade amplitude, saccade duration, pupil area, and pupil area change rate, are collected using an eye tracker. Visualize the eye movement data (such as fixation point heatmaps, saccade path maps) to check for abnormal patterns in the data, and at the same time use statistical methods (such as box plots) to identify outliers in the data. Ensure that the data set does not contain invalid or incorrect data points. Directly delete those data points identified as invalid, or if necessary, correct the invalid data points through interpolation or other data repair techniques. Identify the sources of noise in the data, which may come from measurement errors of the eye tracker, environmental interference, or atypical behavior of the subjects. Use frequency domain analysis methods such as Fourier transform to identify high-frequency noise components in the data. Use appropriate filters to remove the noise in the data while retaining useful information. Ensure that the eye movement data is temporally aligned with other experimental data (such as video recordings, vehicle sensor data). Achieve synchronization by correcting the timestamps of the eye movement data and other data. Ensure that the data points between different data sources are temporally aligned for joint analysis. Use interpolation methods and resampling methods to perform temporal alignment between different data sources. Use data analysis software to effectively preprocess the eye movement data to ensure the accuracy and reliability of the data.

[0042] Fixation generally refers to the state when the driver's line of sight stays within the central visual acuity range of 2 - 5° for a minimum time exceeding 80 - 100 ms. The length of fixation time represents the duration that the driver focuses their line of sight on a certain area. Generally speaking, fixation time reflects the driver's attention level to that area, and can indirectly reflect the energy used by the driver to obtain information from that area and the difficulty of processing the information.

[0043] According to existing research findings, when the fixation duration is less than 200 ms, it can be considered that the measured driver can complete the fixation on the target in a short time, and it is not difficult to obtain relevant information during driving. When there are more fixable targets, the driver will increase their mental effort to obtain more target information, resulting in a shorter fixation duration for a single target and an increase in the proportion of fixation points with a fixation duration greater than 200 ms.

[0044] In addition, by observing the number of fixation points of the driver in different areas, the degree of interest of the driver in that area can be reflected to a certain extent. Currently, there are three types of driver fixation point division methods adopted at home and abroad, namely the mechanical division method, the fixation point statistics method, and the dynamic clustering method, etc. Here, the fixation point statistics method is adopted, and the fixation point data is collected through an eye tracker, and the fixation point distribution of the driver during driving is statistically analyzed frame by frame. After statistics, the fixation points of the driver at the plane intersection are divided into eight areas, named Area A, Area B, Area C, Area D, Area E, Area F, Area G, and Area H, which are the far road, the far left, the far right, the left side of the road, the right side of the road, the cab, the left side of the vehicle, and the right side of the vehicle, and the objects in each area are analyzed in detail.

[0045] According to the characteristics that the fixation point distribution of the driver during driving mainly concentrates on the right side of the road and the far road, the traffic signs at the intersection are set at the right side position of the road in the upstream functional area. The content of the traffic signs should be as concise as possible, and the number of traffic signs should be as small as possible, so as to facilitate the driver to recognize the information on the traffic signs faster.

[0046] When the vehicle enters a complex intersection, due to the different levels of the traffic flow rate at the intersection, the fixation characteristics of the driver will also change. When the traffic flow rate at the intersection increases, the visual characteristics of the driver change from stable to unstable. The design of the intersection considers the driver factors, finds out the maximum traffic capacity tending to the driver's driving safety, and improves the implementation effect.

[0047] Generally speaking, the saccade amplitude and the saccade duration can respectively characterize the saccade activities of the driver from the two levels of space and time. The saccade amplitude is defined as the amplitude swept by the driver's eyes from the end of the last fixation to the start of the next fixation, reflecting the depth of fixation. Usually, the saccade range is 1 - 40°, the saccade speed can reach 400 - 600° / s, and the saccade duration is 30 - 120 ms. During driving, the saccade behavior represents the transfer of the driver's visual point between the concerned targets, mainly for searching, discovering, supplementing information, targets or abnormal situations related to driving.

[0048] When the road conditions are clear, sufficient information can be obtained with just one fixation, and a larger amplitude can be swept when the line of sight transfers to the next fixation point; if the road conditions are more complex, the amount of information obtained by one fixation is less, then the next saccade amplitude will be smaller to obtain sufficient information near that fixation point. When the driver recognizes the intersection, it is necessary to increase the saccade amplitude and the saccade duration to obtain the road information as soon as possible. The saccade duration of the driver during the recognition process mainly concentrates in the time period of [50 ms - 550 ms], and the saccade amplitude of the driver mainly concentrates in [3.5° - 7.3°]. Therefore, the signs and markings at the intersection should be set at appropriate positions within the driver's saccade range.

[0049] The pupil is a disc-shaped small hole in the center of the iris of the human eye. Its most important function is to control the amount of light entering the human eye. It can change the focal length like a camera lens. As the light intensity changes, the iris stretches or compresses to change the pupil area, enabling objects to be clearly imaged on the retina. Affected by the external environment, the diameter of the pupil usually varies within the range of 1.5 mm - 8.5 mm. When it is extremely narrowed or expanded, the diameter can also be less than 1.0 mm or greater than 9.0 mm.

[0050] In addition to being affected by light factors, the size of the pupil of a functionally normal human eye is also affected by cognitive load. When other factors remain unchanged, when the cognitive load is low, the pupil area decreases correspondingly, and when the cognitive load is high, the pupil area also increases correspondingly. According to the principles of ergonomics, it is considered that when the pupil area is smaller, the type of sign setting is more reasonable. On the contrary, when the pupil area is larger, the cognitive load of the driver is greater, and the sign setting is more unreasonable. Since the horizontal and vertical pupil diameters are not exactly the same, the pupil is regarded as an ellipse, and the elliptical area formula is used to calculate the pupil area. The formula is as follows:

[0051]

[0052] Where X is the horizontal pupil diameter and Y is the vertical pupil diameter.

[0053] At the same time, in order to reduce the error caused by the differences between different drivers, the pupil area change rate is used as an observation index here:

[0054]

[0055] In the formula, is the pupil change rate, S driving is the pupil area during driving, and S station is the pupil area in the resting state.

[0056] There are a large number of vehicles at road plane intersections and the intersection conditions are relatively complex. During the process of passing through the intersection, the pupil area and the pupil area change rate of the driver will gradually increase. The visual load of the driver increases when passing through the intersection, and the driver's tension level increases. Speed limit signs and speeding capture devices can be set at the intersection so that the vehicles entering the intersection range can maintain a relatively safe speed range, reducing the driver's sense of tension to a certain extent.

[0057] Calculate metrics such as eye movement dispersion to evaluate the driver's visual load. Eye movement dispersion is an important metric used to measure the distribution of fixation points of a driver during driving. It reflects the driver's visual attention concentration and visual load. A higher dispersion usually means that the driver frequently switches attention between multiple targets, resulting in increased cognitive load and thus affecting driving safety. By calculating the eye movement dispersion, the visual load status of the driver in a specific environment can be identified, and then the road markings and traffic signs can be optimized to improve road safety.

[0058] Calculate the average position of the fixation points:

[0059]

[0060]

[0061] where N is the total number of fixation points, and x i and y i are the coordinates of each fixation point.

[0062] Calculate the standard deviation of the fixation points:

[0063]

[0064]

[0065] Calculate the distance between fixation points:

[0066]

[0067] Calculate the average distance:

[0068]

[0069] Calculate the standard deviation of the dynamic distribution:

[0070]

[0071] Calculate the eye movement dispersion η:

[0072]

[0073] where G represents the static distribution of fixation points, and G_dynamic represents the dynamic distribution of fixation points.

[0074] Analyze the eye movement data metrics, identify the road markings with high driver attention, and determine important road markings and problem road markings.

[0075] Important markings refer to the markings that are directly related to the driving task and provide important cognition that drivers must pay attention to during driving, occupying a necessary position in the driver's vision. Important markings are usually located within the driver's central vision and should have obvious visual characteristics so that drivers can quickly identify and obtain important information. Since the driver's cognition is limited, the design of important markings should convey necessary driving information to the driver without increasing additional visual load. Problematic markings refer to the markings that are related to the driving task but, due to certain reasons (such as improper design, inconspicuous position, confusion with other road features, etc.), result in deviations in the information they convey, causing drivers to pay excessive attention. Problematic markings will cause drivers to pay excessive attention to them, increasing visual load and affecting driving safety.

[0076] The design purpose of important markings is for quick recognition. The fixation time is usually not too long, maintained within a reasonable range, and the fixation frequency should be moderate, not causing excessive visual search. At the same time, the pupil area change rate remains within the normal range, indicating that the driver's attention to the markings is necessary rather than unusual. Problematic markings will cause an increase in the driver's fixation time because the driver needs to spend more time understanding these markings, and will also cause an increase in the fixation frequency, reflecting the driver's excessive attention or confusion. At the same time, it will cause an increase in the pupil area change rate, indicating that the driver's attention to these markings is unusual.

[0077] Therefore, if a marking has a long fixation time and a high fixation frequency, but the pupil area change rate is normal and it is directly related to the driving task, then this marking is determined to be an important marking. If a marking has a long fixation time and a high fixation frequency, and the pupil area change rate is large, indicating that the driver pays excessive attention or is distracted by it, then this marking is determined to be a problematic marking.

[0078] Analyze the improper design, inconspicuous position, or confusion with other road features of problematic markings. The fixation point heat map is used to display the area where the driver's fixations are concentrated. By mapping the fixation point data onto a two-dimensional plane, the fixation point density is represented by the shade of color. Markings with improper design will result in abnormal fixation point distribution, and this abnormality can be visually displayed through the fixation point heat map.

[0079]

[0080] H(x,y) represents the value of the heat map at the position (x,y).

[0081] w i is the weight of fixation point i, usually related to the fixation time or the number of fixations.

[0082] (x i ,y i ) are the coordinates of fixation point i.

[0083] σ is the standard deviation of the Gaussian kernel, which controls the smoothness of the heat map.

[0084] If the fixation density in a certain marking area is abnormally high, it indicates that the marking design is improper, leading to excessive attention from the driver. On the contrary, if the fixation density in a certain marking area is abnormally low, it indicates that the marking position is not obvious enough to effectively guide the driver's visual attention.

[0085] The saccade amplitude is the angle through which the eye turns from the current fixation point to the next fixation point. For example, the saccade amplitude from the first fixation point to the second fixation point is denoted as ρ, which can be calculated through the horizontal and vertical visual angles. According to the minimum angle theorem, the calculation formula is as follows:

[0086] cosρ = cos|θ2 - θ1|cos|λ2 - λ1|

[0087] In the formula: λ1 and θ1 are the vertical and horizontal visual angles of the first fixation point respectively; λ2 and θ2 are the vertical and horizontal visual angles of the second fixation point respectively.

[0088] The time taken from the start to the end of an eye saccade movement is the saccade time, which reflects the time spent by the driver during the visual search process. The average saccade time is the ratio of the total saccade time to the total number of saccades during the driver's driving on the road section. The formula is as follows:

[0089]

[0090] In the formula: T s is the average saccade time (ms); T t,s is the total saccade time (ms); n is the total number of saccades.

[0091] Improperly designed markings can lead to abnormal saccades. If a certain marking causes the driver's saccade amplitude to be abnormally large or the saccade duration to be abnormally long, it indicates that the marking design is improper, resulting in a decrease in the driver's visual search efficiency.

[0092] The assessment of cognitive load can be achieved by measuring the change rate of pupil area. Improperly designed markings can lead to an increase in cognitive load, causing an increase in the pupil change rate.

[0093] The confusion matrix can analyze the driver's recognition accuracy of markings. Improperly designed markings can lead to an increase in recognition errors and reduce the marking recognition accuracy.

[0094]

[0095] Among them, C ij represents the value in the i-th row and j-th column of the confusion matrix, N ij is the number of times of misidentifying marking j as marking i, N jis the total number of times the marking line j is recognized.

[0096] The gaze heat map can show the area where the driver's line of sight is concentrated. Marking lines with an unclear position will result in a sparse display of fixation points in the gaze heat map.

[0097] Using image processing technology, visual features such as marking lines and surrounding environmental features are extracted. These features cover color, shape, texture, contrast, etc. The purpose of feature extraction is to identify the fundamental features in the image, such as edges, corners, textures, etc. Commonly used feature extraction methods include SIFT, SURF, and ORB, etc.

[0098] By calculating the confusion index between the marking line and other features, the confusion degree of the marking line is evaluated. The calculation of the confusion index can be achieved through the following formula:

[0099]

[0100] Among them, the confusion index represents the degree of overlap between the marking line area and other feature areas. A high confusion index indicates that the marking line is confused with other road features.

[0101] Following the traffic marking design specifications and principles, key parameters of the marking line are adjusted, such as angle, spacing, color, and line width. Through on-site measurement or relevant road marking design documents, parameters such as the angle, spacing, color, and line width of the existing marking line are collected. Evaluate whether the existing marking line parameters meet the traffic sign design specifications and principles, including evaluating the visibility, recognition accuracy, and visual interference of the marking line. According to the traffic sign design specifications, determine the optimal angle, spacing, color, and line width of the marking line, calculate the adjusted angle, adjusted spacing, color contrast, and adjusted line width, and carry out the optimized design of the marking line based on this.

[0102] Based on the adjusted road marking line, using the eye tracker according to the virtual environment simulation driving technology to test the driver's reaction, calculate the visual load of the driver on the adjusted marking line, and evaluate the optimization effect. Use 3D modeling software to create a 3D model of the road environment. The model covers road geometry, marking lines, traffic signs, vehicles, and other relevant elements. During the modeling process, it is necessary to ensure the accuracy and adjustability of the road marking lines for subsequent optimization experiments. Apply real textures and materials to the 3D model to improve the environmental realism. The marking line texture is set according to the actual marking line material and color to ensure that it has similar visual characteristics in the virtual environment as in reality. Simulate real-world lighting conditions, covering natural light and artificial light sources. Explore the lighting changes at different times to evaluate the visibility of the marking line under different lighting conditions. Design interactive elements in the virtual environment, such as adjustable marking line parameters (angle, spacing, color, line width), for optimization experiments.

[0103] Select a driving simulator hardware and software platform to assist in high-precision driving behavior simulation. Integrate an eye tracker with the driving simulation system to collect drivers' eye movement data in real time. The eye tracker features a high sampling rate and precision to ensure the accuracy and reliability of the eye movement data. Ensure the data synchronization between the driving simulation system and the eye tracker for precise timestamping and event marking. Explore the simulator latency and the calibration accuracy of the eye tracker to ensure data consistency and comparability. Design driving simulation experiment scenarios and tasks to evaluate drivers' responses to the adjusted road markings. The experiment scenarios cover the road marking designs under different road conditions to comprehensively evaluate the optimization effect of the road markings.

[0104] In the driving simulation experiment, collect drivers' eye movement data in real time, including fixation points, fixation duration and frequency, saccade amplitude, saccade duration, pupil area, and pupil area change rate. Using the collected fixation point data, generate a fixation point heatmap through the Gaussian kernel smoothing algorithm to display the areas where drivers' fixations are concentrated and reveal the degree of attention of drivers to different road marking areas. Identify abnormal visual search patterns by calculating the saccade amplitude and saccade duration. Calculate the visual load of drivers on the adjusted road markings to evaluate the actual effect of the road marking optimization.

[0105] Compare the experimental data before and after optimization, and conduct statistical analysis on the experimental data to determine the significance and actual effect of the road marking optimization. Based on the statistical analysis of the experimental results, put forward claims for road marking optimization to guide the design and optimization of actual road markings.

[0106] The embodiments described above are only descriptions of the preferred embodiments of the present invention and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A road marking optimization design method based on eye movement, characterized in that: include: Collect the driver's eye movement data during driving, study the driver's gaze characteristics on road markings, evaluate the driver's visual load on road markings, divide the road markings that have attracted much attention into important markings and problematic markings, analyze the problematic markings and make appropriate adjustments, and based on the adjusted road markings, collect the driver's eye movement data again in a virtual driving environment, evaluate the driver's visual load on road markings, and verify the effectiveness of the road marking optimization design.

2. The method for optimizing road markings based on eye movement according to claim 1, characterized in that: The following steps are involved: Step S1: using an eye tracker to collect eye movement data, including fixation point, fixation time and frequency, saccade amplitude, saccade duration, pupil area and pupil area change rate. Step S2: Calculate indicators such as eye movement discreteness to evaluate the driver's visual load. Step S3: Analyze the eye movement data indicators, identify the road markings that the driver pays more attention to, and determine the important road markings and problematic road markings. Step S4: Analyze the problem of improper marking design, unclear location or confusion with other road features. Step S5: Follow the design specifications and principles of traffic signs and adjust key parameters of the markings, such as angle, spacing, color, and line width. Step S6: Use an eye tracker to test the driver's reaction, use virtual driving to calculate the driver's visual load when adjusting the marking line, and verify the optimization effect.

3. The road marking optimization design method based on eye movement according to claim 2 is characterized in that: The method for distinguishing important markings and problematic markings in step S3 is: Important markings are those that are directly related to the driving task and provide important cognition that the driver must pay attention to during driving. They occupy a necessary position in the driver's vision. Important markings are usually located in the driver's central field of vision and must have obvious visual features so that the driver can quickly identify and obtain important information. Since the driver's cognition is limited, the design of important markings should convey necessary driving information to the driver without adding additional visual load. Problematic markings are those that are related to the driving task, but due to some reasons (such as improper design, unclear location, confusion with other road features, etc.), the information they convey deviates, causing the driver to pay too much attention to them. Problematic markings can cause the driver to pay too much attention to them, increase visual load, and affect driving safety. Important markings are designed to be quickly identified, and the fixation time is usually not too long, maintained within a reasonable range, and the fixation frequency should be moderate, not causing excessive visual search, and the pupil area change rate remains within the normal range, indicating that the driver's attention to the markings is necessary and not unusual. Problematic markings will cause the driver to increase the fixation time on them because the driver needs to spend more time to understand these markings, and will cause an increase in the fixation frequency, reflecting the driver's excessive attention or confusion. At the same time, it will cause an increase in the pupil area change rate, indicating that the driver's attention to these markings is unusual. Therefore, if the fixation time and frequency of a certain marking are long, but the pupil area change rate is normal and directly related to the driving task, the marking is judged to be an important marking. If the fixation time and frequency of a certain marking are long, and the pupil area change rate is large, indicating that the driver is over-focusing or distracted, the marking is judged to be a problem marking.

4. The road marking optimization design method based on eye movement according to claim 2 is characterized in that: The method for determining whether the problem marking is improperly designed, unclearly positioned, or confused with other road features in step S4 is as follows: The gaze heat map is used to display the driver's concentrated gaze area. By mapping the gaze data onto a two-dimensional plane, the displayed color depth indicates the gaze density. Improperly designed markings can lead to abnormal gaze distribution, which can be intuitively displayed through the gaze heat map. If the gaze density in a certain marking area is abnormally high, this indicates that the marking is improperly designed, causing the driver to pay too much attention. On the contrary, if the gaze density in a certain marking area is abnormally low, this indicates that the position of the marking is not obvious and fails to effectively guide the driver's visual attention. Improperly designed markings can lead to abnormal scans. If a certain marking causes the driver to scan with an abnormally large amplitude or an abnormally long duration, this indicates that the marking is improperly designed, resulting in a decrease in the driver's visual search efficiency. Improperly designed markings can increase cognitive load and increase the pupil change rate. The confusion matrix can analyze the accuracy of the driver's recognition of road markings. Improperly designed road markings will increase recognition errors and reduce the accuracy of road marking recognition. The sight heat map can show the driver's concentrated sight area. Road markings with unclear positions will cause the sight heat map to show sparse gaze points. Image processing technology is used to extract visual features such as markings and surrounding environment features. These features include color, shape, texture, contrast, etc. The purpose of feature extraction is to identify fundamental features in the image, such as edges, corners, textures, etc. The degree of confusion of the markings is evaluated by calculating the confusion index between the markings and other features. A high confusion index indicates that the markings are confused with other road features.

5. The method for optimizing road markings based on eye movement according to claim 2, characterized in that: The method for adjusting the key parameters of the marking line in step S5 is: Follow the traffic marking design specifications and principles to adjust the key parameters of the markings, such as angle, spacing, color, and line width. Collect the existing marking angle, spacing, color, line width and other parameters through on-site measurement or relevant road marking design documents. Evaluate whether the existing marking parameters meet the traffic marking design specifications and principles, including the visibility, recognition accuracy, and visual interference of the markings. According to the traffic marking design specifications, determine the optimal angle, spacing, color, and line width of the markings, calculate the adjustment angle, spacing, color contrast, and line width, and optimize the marking design based on this.

6. The road marking optimization design method based on eye movement according to claim 2 is characterized in that: The method of using virtual driving to calculate the visual load of the driver on the adjusted marking line to verify the optimization effect in step S6 is: Based on the adjusted road markings, an eye tracker is used to test the driver's reaction according to the virtual environment simulation driving technology, calculate the driver's visual load on the adjusted markings, and evaluate the optimization effect. Use 3D modeling software to create a 3D model of the road environment. The model covers road geometry, markings, traffic signs, vehicles and other related elements. The accuracy and adjustability of the road markings must be ensured during the modeling process to facilitate subsequent optimization experiments. Apply real textures and materials to the 3D model to improve the realism of the environment. The marking texture is set according to the actual marking material and color to ensure that it has similar visual characteristics in the virtual environment as in reality. Simulate real-world lighting conditions, including natural light and artificial light sources. Explore the changes in lighting at different time periods to evaluate the visibility of the markings under different lighting conditions. Design interactive elements in the virtual environment, such as adjustable marking parameters (angle, spacing, color, line width), to facilitate optimization experiments. Select driving simulator hardware and software platforms to assist in high-precision driving behavior simulation. Integrate eye trackers with the driving simulation system to collect driver eye movement data in real time. Eye trackers have high sampling rates and precision to ensure accuracy and reliability of eye movement data. Ensure synchronization of driving simulation system and eye tracker data, and perform accurate timestamps and event markings. Explore simulator latency and eye tracker calibration accuracy to ensure data consistency and comparability. Design driving simulation experimental scenarios and tasks to evaluate driver reactions to adjusted road markings. Experimental scenarios cover road marking designs for traffic conditions under different road conditions to fully evaluate the road marking optimization effect. In the simulated driving experiment, the driver's eye movement data is collected in real time, covering the gaze point, gaze time and frequency, glance amplitude, glance duration, pupil area and pupil area change rate. Using the collected gaze point data, a gaze point heat map is generated through the Gaussian kernel smoothing algorithm to show the driver's focus area and reveal the driver's attention to different marking areas. By calculating the glance amplitude and glance duration, abnormal visual search patterns are identified. The driver's visual load when adjusting the markings is calculated to evaluate the actual effect of marking optimization. Compare the experimental data before and after optimization, and conduct statistical analysis on the experimental data to determine the significance and actual effect of road marking optimization. Conduct statistical analysis based on the experimental results, propose road marking optimization proposals, and guide the actual road marking design and optimization.

Images