Driving skill training method and system combined with virtual reality

Abstract

The application discloses a driving skill training method and system combined with virtual reality, and belongs to the technical field of driving simulation. The method comprises the following steps: generating a plurality of driving scenes according to a virtual reality driving simulator; training a user to perform driving simulation in the plurality of driving scenes according to a VR wearing device, and obtaining a plurality of sets of driving simulation data sets; constructing a driving operation scoring model; inputting the plurality of sets of driving simulation data sets into the driving operation scoring model, analyzing each set of driving simulation data set according to the driving operation scoring model, and outputting a plurality of driving operation scoring results; and generating a driving skill training report according to the plurality of driving operation scoring results. The application solves the technical problem that the driving skill training in the prior art lacks personalized and data-based evaluation, so that the trainee cannot obtain targeted training and feedback according to the own performance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of driving simulation, in particular to a driving skill training method and system combined with virtual reality. BACKGROUND

[0002] With the rapid development of science and technology, virtual reality (VR) technology has been widely applied in various fields, especially in the field of education and training. Virtual reality has become an important innovative tool due to its immersion and interactivity. In particular, in the field of driving skill training, traditional teaching methods often rely on actual road driving or driving simulators, but these methods have certain limitations, such as high cost, limited training sites, and uncontrollable safety risks. With the maturity of virtual reality technology, driving skill training combined with virtual reality not only can simulate various real driving environments, but also can be repeatedly practiced in a safe and controllable virtual environment, providing a more flexible and efficient learning method for students.

[0003] Currently, traditional driving training relies on real car operation and simple driving simulators, and the learning progress and skill assessment of students usually rely on the subjective judgment of the instructor. Although some driving simulators can provide certain driving scene simulation, there are still the following problems: first, traditional driving simulation training lacks comprehensive and objective evaluation of students' driving operations; second, the training content lacks individualization, and students often cannot effectively train their weak links; third, the training environment and training progress are difficult to adjust in real time, making it difficult to ensure that students train in the best learning state. Therefore, there is an urgent need for a more scientific and efficient driving skill training method that can combine virtual reality technology and data analysis to provide customized training programs for students and improve training effectiveness through real-time feedback and evaluation. SUMMARY

[0004] The present application provides a driving skill training method and system combined with virtual reality, aiming to solve the technical problem that the existing driving skill training lacks individualization and data-based evaluation, so that students cannot obtain targeted training and feedback according to their own performance.

[0005] In view of the above problems, the present application provides a driving skill training method and system combined with virtual reality.

[0006] In a first aspect, a driving skill training method combined with virtual reality is provided. The method comprises connecting a virtual reality driving simulator, generating a plurality of driving scenes based on the virtual reality driving simulator, wherein the virtual reality driving simulator comprises a VR wearable device; training a user to drive in the plurality of driving scenes based on the VR wearable device to obtain a plurality of driving simulation data sets, wherein each driving simulation data set corresponds to one driving scene; constructing a driving operation scoring model, wherein the driving operation scoring model is connected to the virtual reality driving simulator; inputting the plurality of driving simulation data sets into the driving operation scoring model, analyzing each driving simulation data set based on the driving operation scoring model, and outputting a plurality of driving operation scoring results; and generating a driving skill training report based on the plurality of driving operation scoring results.

[0007] In another aspect, a driving skill training system combined with virtual reality is provided. The system comprises a driving scene generation module for connecting a virtual reality driving simulator, generating a plurality of driving scenes based on the virtual reality driving simulator, wherein the virtual reality driving simulator comprises a VR wearable device; a simulation data set acquisition module for training a user to drive in the plurality of driving scenes based on the VR wearable device to obtain a plurality of driving simulation data sets, wherein each driving simulation data set corresponds to one driving scene; a scoring model construction module for constructing a driving operation scoring model, wherein the driving operation scoring model is connected to the virtual reality driving simulator; a scoring result output module for inputting the plurality of driving simulation data sets into the driving operation scoring model, analyzing each driving simulation data set based on the driving operation scoring model, and outputting a plurality of driving operation scoring results; and a training report generation module for generating a driving skill training report based on the plurality of driving operation scoring results.

[0008] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0009] The driving skill training method combined with virtual reality is adopted, wherein a plurality of driving scenes are generated by a virtual reality driving simulator, and driving simulation data sets are obtained based on the performance of students in different scenes. The data sets are analyzed and scored by a driving operation scoring model, which solves the problem of lack of individualization and data-based evaluation in existing driving skill training, so that students cannot obtain targeted training and feedback based on their own performance. This technical solution makes the training process more personalized through accurate scoring and data analysis, and can provide customized training plans and feedback based on the performance of students, thereby improving the effectiveness of driving training and the skill level of students.

[0010] The above description is only a summary of the technical solutions of the present application. In order to enable the technical means of the present application to be more clearly understood, and to be implemented according to the content of the description, and in order to enable the above and other purposes, characteristics and advantages of the present application to be more apparent and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 A flowchart of a driving skill training method combined with virtual reality is provided for the embodiments of the present application;

[0012] Figure 2 A structural schematic diagram of a driving skill training system combined with virtual reality is provided for the embodiments of the present application.

[0013] Reference signs: driving scene generation module 11, simulation data set acquisition module 12, score model construction module 13, score result output module 14, training report generation module 15. DETAILED DESCRIPTION

[0014] The general idea of the technical solutions provided by the present application is as follows:

[0015] The embodiments of the present application provide a driving skill training method and system combined with virtual reality. By combining virtual reality (VR) technology, a virtual reality driving simulator is used to generate multiple driving scenes, and trainees perform driving simulation in these scenes and collect data. By constructing a driving operation scoring model, each set of simulation data is analyzed, and driving score results are output, and personalized driving skill training reports are generated according to these score results. This scheme realizes data-based and personalized evaluation and feedback, and can provide targeted training and improvement suggestions for trainees, improving the efficiency and effectiveness of driving skill training.

[0016] After introducing the basic principles of the present application, the various non-limiting embodiments of the present application will be specifically introduced in combination with the drawings of the specification.

[0017] Embodiment one, as shown in the figure, the embodiments of the present application provide a driving skill training method combined with virtual reality, the method comprises: Figure 1 Step S100: Connect a virtual reality driving simulator, and generate multiple driving scenes according to the virtual reality driving simulator, wherein the virtual reality driving simulator comprises a VR wearable device.

[0018]

[0019] ​Specifically, a virtual reality driving simulator is a system that uses virtual reality (VR) technology to create a virtual driving environment. By simulating real driving scenarios, it provides users with a training experience similar to actual driving. A driving simulator typically includes hardware devices (such as consoles, steering wheels, pedals, seats, etc.) and a software platform (such as a virtual environment generation and interaction system). VR wearables refer to hardware devices that allow users to immerse themselves in a virtual environment, usually including virtual reality headsets (such as Oculus Rift, HTC Vive, Meta Quest, etc.) and accompanying sensors. After wearing this device, users can perceive and interact with the content in the virtual world through vision, hearing, and even touch, achieving a sense of presence. Driving scenarios refer to various virtual driving environments presented in the driving simulator. It includes road conditions, traffic conditions, weather changes, and other factors. For example, it can simulate urban roads, mountain roads, complex intersections, and other driving situations.

[0020] First, the driving simulator connects the software platform with the hardware devices to form a complete training system. The simulator can present various driving environments in real time through connection with a computer system or control unit. The hardware devices include steering wheels, brake pedals, throttle pedals, seats, etc., simulating the real driving operation experience.

[0021] Once the virtual reality system is established, the system software uses 3D modeling and rendering techniques to generate different driving scenarios. These scenarios can include urban roads, mountain highways, expressways, and adverse weather conditions to meet different training needs. For example, in an urban road scenario, factors such as traffic lights, pedestrians, and other vehicles are simulated; in a mountain road scenario, features such as sharp turns and uphill and downhill terrain are simulated.

[0022] VR wearables (such as Oculus Rift, HTC Vive, etc.) connect students with virtual environments through virtual reality headsets. After wearing the headset, students can perceive 360-degree virtual scenes through head movements, and interact with the virtual environment through handheld controllers. For example, when the student turns his head, the sensors in the headset will update the display in real time, presenting the driver's perspective, thereby simulating the real driving experience.

[0023] In the virtual reality system, driving scenarios not only include different geographical environments, but also include different driving difficulties. For example, students can be trained in basic driving skills in ordinary urban road scenarios, or more complex driving environments can be simulated under conditions such as night or rain, increasing the challenge of training.

[0024] By connecting a virtual reality driving simulator and generating multiple driving scenarios, trainees can repeatedly practice in a virtual environment and master the skills to handle different driving situations. Since the system can provide a variety of complex driving environments, trainees can train all-weather without being limited by time and location. This immersive training greatly improves trainees' understanding and response to driving techniques, and reduces the safety risks encountered in actual road training.

[0025] Step S200: The training user drives according to the VR wearable device in the multiple driving scenarios to obtain multiple sets of driving simulation data, wherein each set of driving simulation data corresponds to one driving scenario.

[0026] Specifically, the training user refers to the trainee participating in driving skill training, usually the personnel who conduct driving simulation and skill training through a virtual reality driving simulator. The training user can be a novice driver or a person who already has some driving experience, and the system will provide personalized driving scenarios according to different training needs. The driving simulation data set refers to all the operation data of the trainee in a specific scenario recorded by the system during driving simulation. Each data set usually contains the trainee's behavior during driving, such as acceleration, braking, steering, reaction speed, speed, etc. These data are important basis for evaluating the trainee's driving skills.

[0027] The training user first wears a VR headset and controller and enters the virtual reality system through the VR wearable device. After wearing the device, the user's field of vision is completely immersed in the virtual driving scene, as if he is in the actual driving environment. For example, the user can see streets, traffic lights, pedestrians, etc. in the urban road scene, as if he is in actual driving.

[0028] The system customizes different driving scenarios for each trainee. For example, the first scenario is a sunny urban road, simulating the trainee's performance in a busy urban area; the second scenario is a rainy highway, requiring the trainee to respond to slippery roads, rain and fog, etc. The third scenario is a night mountain road, simulating the difficulty of night driving. The trainee selects the scenario for simulation driving according to his training plan.

[0029] In each driving scenario, the system records the trainee's driving behavior in real time, including but not limited to acceleration, braking, steering, speed change, reaction speed, attention concentration, etc. For example, in the urban road scenario, the system records whether the trainee can brake in time when encountering a red light, whether he obeys traffic rules, whether he has frequent sudden braking or acceleration, etc. In the rainy highway scenario, the system records the trainee's speed control, whether he maintains a safe following distance, and his driving stability on slippery roads, etc.

[0030] Each set of driving simulation data corresponds to a specific driving scenario. For example, in an urban road scenario, the system records all driving behavior data to form a data set; in a rain highway scenario, another data set records the student's performance in this scenario. In this way, the system can generate independent data sets for each scenario, facilitating subsequent analysis and scoring.

[0031] Through this step, the system can capture the student's performance in different driving environments in real time and comprehensively, providing data support for subsequent skill assessment. The simulation data set in each scenario not only helps to analyze the student's driving ability under specific conditions, but also provides a basis for the development of individualized training programs. For example, if the student performs poorly in the rain scenario, the system can arrange more rain driving training according to the data feedback. Through the accumulation of multiple scenarios and multiple data sets, the student can gradually improve their response ability in various driving environments, ultimately achieving comprehensive driving skill improvement.

[0032] Step S300: Build a driving operation scoring model, wherein the driving operation scoring model is connected to the virtual reality driving simulator.

[0033] Specifically, the driving operation scoring model is an algorithm model used to evaluate the performance of drivers during simulated driving. The model analyzes various operational behaviors of students in driving simulation (such as acceleration, braking, steering, etc.) and gives a comprehensive score to reflect the student's driving skill level and operational safety. The scoring model is usually based on specific driving behavior characteristics, such as reaction time, driving fluency, safety, etc.

[0034] When building the driving operation scoring model, the scoring criteria and evaluation indicators of the model need to be defined first. For example, the model scores based on the following behaviors of the student during driving: brake characteristics: whether to brake in time and smoothly when needed, and whether the braking force is appropriate; steering characteristics: whether to maintain smooth steering and avoid sharp turns; speed control: whether to adjust speed reasonably according to road conditions, avoiding too fast or too slow; reaction characteristics: when encountering unexpected situations (such as traffic signal changes, obstacles, etc.), whether the student's reaction speed is timely and effective.

[0035] To achieve these scoring criteria, the model can use machine learning or rule engine methods to automatically calculate a comprehensive score based on the student's operation data. For example, the model can analyze the student's behavior data in the simulation scenario, identify the student's driving habits, and evaluate the quality of driving operation based on these habits.

[0036] Specifically, to build the driving operation scoring model, first, define the scoring criteria and evaluation indicators to clearly identify the driving behavior characteristics that need to be assessed. Each characteristic should be quantified according to actual driving requirements to form a calculable parameter. Second, collect the operation data of the trainees in the virtual reality driving simulator, including acceleration, braking, steering, speed, reaction time, etc. Then, select a suitable algorithm to build the scoring model. Rule-based models can be used to judge the pros and cons of trainee operations based on pre-set thresholds; machine learning algorithms such as decision trees, support vector machines (SVM), etc. can also be used to train the model based on historical data, allowing it to automatically identify and score trainee driving behavior. Next, design a data transmission mechanism to transmit real-time data from the simulator to the scoring model for processing. Finally, verify and optimize the scoring model to ensure its accuracy and stability. By continuously adjusting model parameters, the rationality and individuality of the scoring results are improved, and ultimately a comprehensive assessment of trainee driving skills is achieved.

[0037] The driving operation scoring model needs to be connected with the virtual reality driving simulator for data transmission. This is usually achieved through an API interface or data transmission protocol, where the simulator sends the trainee's operation data (such as acceleration, braking, steering angle, speed, etc.) to the scoring model in real time. The scoring model receives these data and analyzes them according to pre-set rules and algorithms, ultimately outputting the trainee's operation score in each driving scenario.

[0038] By building the driving operation scoring model and connecting it with the virtual reality driving simulator, the system can provide objective and quantitative driving evaluation, helping trainees understand their strengths and weaknesses in driving, and adjusting their training plans based on the scoring results.

[0039] Step S400: input the multiple sets of driving simulation data sets into the driving operation scoring model, analyze each set of driving simulation data set according to the driving operation scoring model, and output multiple driving operation scoring results.

[0040] Specifically, the driving operation scoring results are the scores output by the model after analyzing each set of simulation data set, which reflect the trainee's driving skill level in a specific scenario. For example, the scoring results will show how well the trainee performs in braking, steering, reaction speed, etc.

[0041] During training, each driving behavior of the trainee (such as braking, acceleration, steering, etc.) is recorded to form an independent data set. The system will input these data sets into the scoring model according to the scene.

[0042] After receiving the data, the driving operation scoring model analyzes it based on pre-set rules or machine learning algorithms. The model first extracts features from the data, identifying important driving behavior characteristics such as braking characteristics, steering angles, speed changes, etc. Then, it scores each feature according to the established evaluation criteria, and finally, it outputs a comprehensive operation score.

[0043] Each data set corresponds to a specific driving scenario, and the scoring model evaluates each data set to output a driving operation score result. For example, when simulating urban road driving, the model outputs the student's scores in braking, acceleration, reaction speed, etc.; in the rain on the highway scenario, the model evaluates the student's ability to respond to wet and slippery roads and the level of speed control. Each score reflects the student's skill level in a specific driving scenario.

[0044] Through this step, the system can comprehensively analyze the student's driving behavior and automatically and accurately assess the student's driving ability in different scenarios. The score of each driving scenario can reflect the student's performance in a specific driving situation, such as whether the brake is timely, whether the speed is appropriate, whether the reaction is rapid, etc. This not only provides precise skill feedback to the student, but also helps the coach to develop a personalized training plan, focusing on the student's weaknesses for training and improvement, thereby improving the student's overall driving ability.

[0045] Step S500: According to the plurality of driving operation scoring results, a driving skill training report is generated.

[0046] Specifically, the driving skill training report is a comprehensive document that shows the student's performance evaluation results in various aspects during the training process. The report not only includes the student's scores in each scenario, but also includes the student's overall performance trend, improvement direction, and future training suggestions. The report is usually used to help the student and the coach understand the student's progress and develop a reasonable plan for subsequent training.

[0047] First, the system collects the student's scoring results in various driving scenarios from previous training sessions. These scoring results include multiple dimensions of scoring (such as driving stability, reaction speed, operation fluency, etc.). These results are digital scores obtained by analyzing the student's performance in different scenarios through the driving operation scoring model.

[0048] The system will summarize and analyze the collected scoring results. Here, the evaluation indicators (such as driving stability, reaction time, safety, etc.) will be summarized into specific performance in each scenario, or even summarized through certain data processing methods (such as weighted average, maximum value, etc.).

[0049] The key part of generating the report is the comprehensive performance analysis of the multiple scenario score results. For example, the system calculates the average score of the trainee in all scenarios and finds out the strengths and weaknesses. According to the performance of the trainee, the system automatically generates training recommendations based on data analysis results. Finally, the system generates a driving skill training report in a clear and intuitive form. The report is usually presented in the form of charts, words and tables, which is easy for the trainee and the coach to understand. The charts can include the trainee's score curve in each dimension, comparison chart of each scenario, etc.

[0050] Through this process, the trainee can clearly understand his performance and progress in the training process, and find out his strengths and weaknesses in time, and then develop a targeted improvement plan. At the same time, the coach or training institution can also adjust the training content according to the content of the report and provide more personalized training programs. The report is not only a feedback tool for the trainee, but also an important means to improve the overall training quality, which helps the trainee continuously optimize his driving skills.

[0051] Further, according to the driving operation scoring model, each set of driving simulation data is analyzed, and the method comprises: the driving operation scoring model extracts features from each set of driving simulation data to obtain low-dimensional features, the low-dimensional features including brake features, speed features, steering features and reaction features; high-dimensional feature convolution is performed according to the low-dimensional features to output high-dimensional features, the high-dimensional features including driving stability, driving fluency, reaction rate and driving safety; the driving operation scoring model analyzes the high-dimensional features to obtain the driving operation scoring results corresponding to each set of driving simulation data.

[0052] Specifically, low-dimensional features are basic, concise, and descriptive features extracted from raw driving data, often single indicators that directly reflect the trainee's driving operations. For example, braking features include braking force and timing; speed features refer to speed variation; steering features describe steering angle and steering smoothness; and reaction features involve the trainee's response speed and decision-making in unexpected situations. These low-dimensional features are the basis for high-dimensional features, providing preliminary evidence for subsequent analysis. High-dimensional features are more complex and abstract features obtained by combining, calculating, or convolving low-dimensional features. They can comprehensively reflect the trainee's overall performance during driving. For example, driving stability can be evaluated by combining braking smoothness, speed control stability, and other low-dimensional features; driving fluency involves the coherence and fluency of operations during driving. Feature convolution is the process of combining and calculating low-dimensional features through mathematical methods (such as convolutional neural networks or other algorithms) to obtain high-dimensional features. Convolution operations are used in this step to extract deeper driving behavior patterns, making the analysis results from low-dimensional features more comprehensive and accurate. Driving stability refers to whether the trainee can smoothly control the vehicle during driving, avoiding sudden braking or sharp turns. Driving fluency measures the coherence and fluency of operations during driving, especially the transition between steering and acceleration, braking. Reaction rate refers to the speed and accuracy of the trainee's response to unexpected situations. Driving safety refers to whether the trainee follows safe driving rules, such as observing speed limits, maintaining a reasonable distance from the vehicle in front, and responding to traffic signals in a timely manner.

[0053] During the driving simulation process, the system collects various operation data from the trainee in real time. First, these data are decomposed into a series of low-dimensional features, i.e., simple operation indicators such as braking features, speed features, steering features, and reaction features. For example, braking features include braking response time delay and force; steering features include steering angle and steering smoothness; and reaction features describe the trainee's response speed when facing traffic light changes or emergency situations. All these low-dimensional features will provide basic data for the calculation of high-dimensional features.

[0054] Next, the scoring model performs high-dimensional feature convolution on the extracted low-dimensional features. Through convolution operations, the system combines different low-dimensional features to generate comprehensive, high-level features. For example, driving stability can be derived by combining braking smoothness and speed control stability; and driving fluency is calculated from the coordination between acceleration and braking, steering smoothness, and other low-dimensional features. This convolution process enables the model to analyze the trainee's overall driving performance by combining multiple data, rather than just single operation indicators.

[0055] After the convolution process, the model will generate a series of high-dimensional features, such as driving stability, driving fluency, reaction rate, and driving safety, etc. These high-dimensional features can more comprehensively and accurately reflect the driving ability of the student. For example, if the student's reaction is stable when braking, and the vehicle speed control is stable, the model will output a high score of driving stability; if the student's steering is smooth and reaction is quick when changing lanes, the fluency score will be higher.

[0056] Finally, the scoring model analyzes each set of driving simulation data based on these high-dimensional features and outputs the corresponding driving operation score results. Each set of score results reflects the overall performance of the student in that driving scenario. For example, in the city driving simulation, if the student brakes in time and steers smoothly, the score result is higher; if the student's reaction is slow and speed control is improper in the highway scenario, the score result is lower.

[0057] Through this feature extraction and convolution analysis process, the scoring model can more comprehensively and accurately evaluate the driving ability of the student. Compared with relying solely on low-dimensional features, combining high-dimensional features, the model can provide more detailed feedback to help students understand their specific performance in the driving process.

[0058] Further, after outputting multiple driving operation score results, the method further comprises: setting multiple driving score levels for each driving scenario; matching the driving score levels according to the multiple driving operation score results in each driving scenario to obtain the matching driving score level in each driving scenario; and integrating the multiple matching driving score levels in the multiple driving scenarios to generate a comprehensive driving score level.

[0059] Specifically, the driving score level is a classification evaluation of the student's operation performance in a specific driving scenario. Usually, the student's driving level is divided into different levels according to the score results. The score level can be numerical (such as 1-100 points) or graded (such as excellent, good, medium, and poor). Different score levels represent the skill level of the student in that driving scenario. Score level matching is to compare the student's driving operation score results with the preset score standard to determine the student's level in a specific scenario. The score level matching of each scenario will determine the student's specific score in that scenario. The comprehensive driving score level is a comprehensive evaluation result obtained by weighting or averaging the score results in multiple driving scenarios, reflecting the overall performance of the student in all simulation scenarios. This comprehensive score can provide a comprehensive skill evaluation for the student, facilitating the student to understand their driving ability in different situations.

[0060] First, specific grading standards are set for each driving scenario. For example, the grading levels for urban road driving scenarios can be set as: excellent, good, medium, pass, and fail. This setting divides the performance of students into different levels according to the operation score results calculated by the scoring model. Similar grading levels can also be set for other scenarios, such as highways and mountain roads.

[0061] The scoring results of students in each driving scenario are compared with the set grading levels to determine the grading level of students in each scenario. For example, in the urban driving scenario, if the student's score is 85, according to the grading level standard, his level is "good"; if in the highway scenario, the score is 75, it is "medium". The grading level of each scenario corresponds to the performance of the student in that specific scenario.

[0062] After the student completes the simulation training in multiple scenarios, the system will integrate the grading levels in each scenario. Usually, this integration process can use weighted average to integrate the grading levels of all scenarios. For example, the grading levels in different scenarios such as urban driving, mountain driving and highway driving can be assigned weights according to their importance. For example, highway driving is considered more challenging, so it is given a higher weight. Finally, the system outputs a comprehensive driving grading level reflecting the student's overall performance in all scenarios. For example, if the student performs well in urban driving and mountain driving, but performs medium in highway driving, the comprehensive grading level is "good" or "medium", and the specific level is calculated according to the weight of each scenario.

[0063] This process enables the scoring system to comprehensively evaluate the driving ability of students according to the performance of different driving scenarios. By converting the scoring results of each scenario into specific levels and calculating the weighted average according to the importance, a comprehensive and objective skill evaluation is finally generated. Students not only can see their specific performance in each scenario, but also can get a comprehensive and objective skill evaluation. This comprehensive score can help students understand their strengths and weaknesses comprehensively, so as to improve and improve targetedly. At the same time, coaches or training institutions can develop more targeted training plans according to the comprehensive score of students to further improve the driving skills of students.

[0064] Further, after outputting the plurality of driving operation score results, the method further comprises: feeding back the plurality of driving operation score results to the virtual reality driving simulator, generating a plurality of driving training scenes, and the training user trains based on the plurality of driving training scenes; the virtual reality driving simulator compares the plurality of driving operation score results with a pre-stored driving operation score threshold value, screens N driving scenes less than the driving operation score threshold value, wherein N is a positive integer greater than or equal to 1; and identifies the N driving scenes as a plurality of driving training scenes to be trained by the current training user.

[0065] Specifically, the driving operation score threshold value is a pre-set score or standard, representing the minimum or expected driving performance of the trainee in a certain scene. The score threshold value can be set according to the training purpose, scene difficulty or trainee level. For example, the score threshold value of a certain scene requires the trainee's driving stability score to exceed 80 points to pass the training of the scene. The N driving scenes refer to the set of driving simulation scenes screened according to the score threshold value, wherein N is a positive integer greater than or equal to 1, indicating the number of scenes in which the trainee performs poorly and needs further training. The screened scenes will be used as the training target for the trainee.

[0066] The trainee's performance in the plurality of driving simulation scenes has been obtained through the scoring model, and the driving operation score results contain evaluation data of various driving skill indicators. These score results will be fed back to the virtual reality driving simulator as the basis for decision-making for the next training content. The virtual reality driving simulator will generate a series of new driving training scenes according to the score results to strengthen the training of the trainee's weaknesses.

[0067] The virtual reality driving simulator compares the driving operation score results in each scene with the pre-stored driving operation score threshold value. Each driving scene has a set score standard (i.e. threshold value), and when the trainee's score is lower than the threshold value, the simulator considers that the scene needs more training. For example, if the score threshold value of a certain scene is set to 70 points, and the trainee's score is 65 points, then the scene will be marked as a scene that needs to be retrained.

[0068] According to the comparison result, the virtual reality driving simulator will screen a plurality of scenes with scores lower than the score threshold value, and the number is N, wherein N is a positive integer (greater than or equal to 1). These screened scenes are the scenes that the trainee needs to focus on training at present. The trainee will conduct more in-depth training in these scenes to improve his weaknesses.

[0069] The selected driving scenarios will be identified as "training-ready" scenarios, where the trainee will undergo intensive training. For example, the trainee is required to practice in more complex urban road or highway scenarios to improve their braking response or steering control precision. The trainee's training system will adapt to these training-ready scenarios, ensuring that the training content is highly targeted.

[0070] This step realizes a personalized training program that can adjust training content in real-time based on the trainee's performance in different scenarios. Through the feedback system, the trainee can focus on scenarios where they perform poorly and target their improvement.

[0071] Further, the method for generating a plurality of driving scenarios based on the virtual reality driving simulator includes: constructing a driving scenario database; inputting the identity information and training project information of the training user; performing identity verification on the identity information of the training user, and if the identity verification is passed, performing project scenario screening in the driving scenario database according to the training project information, and outputting a screening scenario set; and generating a plurality of driving scenarios corresponding to the training user according to the screening scenario set.

[0072] Specifically, the driving scenario database is a database system that stores various driving simulation scenario data. It contains a variety of driving situations and environmental settings, such as urban roads, highways, mountain roads, and other common or special driving scenarios. Each scenario contains specific information related to it, such as scenario description, environmental conditions, traffic conditions, road structure, etc. The identity information of the training user refers to the personal identity information of the trainee in the training system, including name, age, driver's license information, trainee number, etc. These information are used to uniquely identify the trainee and track and record during the training process. The training project information refers to the training content, course arrangement, goals, etc. related to the training user. For example, the trainee is participating in "basic driving skill training" or "advanced driving skill training". Different training projects will involve different driving scenarios and simulation training requirements. Project scenario screening is the process of selecting a set of driving scenarios related to the project from the driving scenario database according to the training project information. The screening scenario set refers to a group of driving simulation scenarios selected from the database according to the training project requirements. These scenarios are customized to meet specific training goals, and the trainee will train in these scenarios.

[0073] First, a driving scenario database is constructed, and different driving scenarios and environmental conditions are stored in it. This database contains a rich variety of scenario types, such as urban roads, highways, mountain roads, etc., to adapt to different training needs and situations. Specifically, the structure of the database is first defined, including scenario types, environmental conditions, traffic conditions, road features, etc. Then, the specific data of different driving scenarios is collected and sorted, such as the specific description of each scenario, traffic signs, weather conditions, speed limits, etc. of urban roads, highways, mountain roads, etc. Then, these data are entered into the database, which can use a relational database (such as MySQL) or a non-relational database (such as MongoDB). In addition, scenarios need to be associated with other data (such as user preferences, driving skill levels, etc.), to ensure that the system can filter and generate appropriate training scenarios according to different needs.

[0074] Then, the identity information and training project information of the training users are entered into the system, such as the student's name, driving license level, training course content, etc. These information not only helps the system to distinguish different students, but also helps to select appropriate scenarios according to the training goals of the students. For example, for beginners, the system will prefer to choose simple urban driving scenarios, while for students with some driving experience, more challenging scenarios such as mountain roads or complex traffic environments will be selected.

[0075] Next, the system will verify the identity of the student's identity information to ensure that each student can access the correct personalized training content. If the identity verification is passed (for example, the student successfully logs into the system and verifies the identity), the system will filter the project scenarios according to the student's training project information (such as course arrangement or skill requirements). For example, if the student participates in "basic driving skill training", the system will filter out simple driving scenarios related to this project, such as flat urban roads or low-speed loop scenarios. If the student's project is "advanced driving skill training", the system will select more complex scenarios, such as night driving, driving in the rain, or mountain driving, etc.

[0076] Finally, according to the filtered scenario set, the system generates a series of driving scenarios that meet the needs of the training user. Each scenario is customized according to the student's training project, ensuring the relevance and effectiveness of the training. For example, if the student needs to practice highway driving, the system will generate a set of simulation training scenarios related to highways, including different traffic flow, weather conditions, traffic signs, etc.

[0077] Through this process, the system can provide personalized and targeted driving training for each student. Students can get the most suitable training scenarios according to their needs and training goals, which not only improves the efficiency of training, but also ensures that students can improve their driving skills in the shortest time.

[0078] Further, after outputting the filtered scenario set, the method further includes: wherein each driving scenario in the driving scenario database carries identification information, which is an indicator measuring the operation difficulty of the driving scenario; obtaining an identification information set corresponding to the filtered scenario set; performing operation difficulty distribution uniformity identification on the identification information set, and optimizing the plurality of driving scenarios with a preset operation difficulty distribution uniformity as a target, and outputting the plurality of optimized driving scenarios.

[0079] Specifically, the identification information is in the driving scenario database, and each driving scenario has an identification information indicating the operation difficulty of the scenario. This is usually a numerical value or a level, which measures the driving difficulty of the scenario. The operation difficulty can be set according to various factors, such as traffic density, road complexity, weather conditions, etc. For example, the driving difficulty of a highway is low, while the difficulty of a mountain road or a complex urban intersection is high. The operation difficulty distribution uniformity identification refers to identifying the operation difficulty distribution of the plurality of filtered driving scenarios and evaluating whether it is uniform. If the difficulty distribution of a scenario set is concentrated, the training effect is not ideal, and the improvement of the student's ability is limited. The plurality of optimized driving scenarios is obtained by adjusting and re-filtering the operation difficulty distribution, so that the student can experience more balanced challenges. For example, if there are too many low-difficulty scenarios in a training project, the optimization process will introduce more high-difficulty scenarios to ensure that the student's skills are balanced.

[0080] In the previous steps, the system has filtered a plurality of scenarios from the driving scenario database according to the student's needs. These scenarios all have an operation difficulty identification information, i.e., a preset difficulty value for each scenario. The system extracts the identification information set from the filtered scenarios to obtain the difficulty values of all scenarios. For example, for the filtered scenarios including a highway, an urban road, and a mountain road, the system extracts the difficulty values of these scenarios, which are 3, 5, and 8, respectively.

[0081] The system will perform operation difficulty distribution uniformity identification according to the extracted difficulty value set. The goal of this identification process is to check whether the difficulty distribution of the current filtered scenarios is uniform. For example, if most of the scenarios have difficulty values concentrated in a low range (such as 2 to 4), and there are not enough high-difficulty scenarios, then the difficulty distribution is not uniform and needs to be adjusted. The system uses statistical algorithms (such as uniformity calculation, variance analysis, etc.) to judge the uniformity of the current difficulty distribution.

[0082] According to the result of operation difficulty distribution uniformity recognition, the system will optimize the current screening scene. The purpose of optimization is to make the difficulty of the scene set more balanced. For example, if the system finds that the current training scene is too simple, the optimization process will add some more challenging scenes or adjust the scene order to gradually improve the skills of the trainee and finally achieve a balanced difficulty training experience.

[0083] After the optimization process is completed, the system generates a new optimized scene set, and the scenes in this set have a more reasonable difficulty distribution. For example, the system extracts several groups of medium difficulty scenes from low difficulty urban road and highway scenes, and adds some higher challenge mountain road, complex intersection or special weather condition scenes to ensure that the trainee gradually improves from simple to complex and from easy to difficult.

[0084] Through this process, the trainee can receive more balanced and gradually improved training content, avoiding situations where the training difficulty is too high or too low. Through difficulty distribution optimization, the system can provide more appropriate training challenges for the trainee, helping them gradually master higher difficulty driving skills in the process of continuous improvement and improving training effectiveness.

[0085] Further, the training user drives in the plurality of driving scenes according to the VR wearing device, and the method further comprises: acquiring a cumulative simulation time length of the training user in the plurality of driving scenes; if the cumulative simulation time length is greater than a preset cumulative simulation time length, a first reminder information is generated, and the virtual reality driving simulator reminds the training user to pause training according to the first reminder information.

[0086] Specifically, the cumulative simulation time length refers to the total time of the training user in the virtual reality driving simulator. It is a statistical measure of the cumulative operation time of the user in multiple driving scenes, used to measure the training time of the trainee. The preset cumulative simulation time length is a maximum training time limit set by the system, which is used to prevent the user from performing long simulation training and avoid affecting the training effect due to fatigue. This time length can be set according to the training needs, physical condition of the trainee, and the best training time recommended by scientific research. The first reminder information refers to the prompt information generated by the system when the cumulative simulation time length of the trainee exceeds the preset time. The purpose of this reminder information is to remind the trainee to rest or pause training, avoid excessive fatigue, and ensure training effectiveness and physical health.

[0087] After the trainee wears the VR wearing device and enters the virtual driving simulation environment, the system will track and record the simulation driving time of the trainee in real time. The system can accurately calculate the training time of the trainee in each driving scene and add up these times to form the cumulative simulation time length of the trainee.

[0088] The system sets a preset cumulative simulation duration according to the training needs and health recommendations of the trainee. For example, assuming the preset cumulative simulation duration is 2 hours, when the total duration of the trainee's training in the virtual reality system exceeds 2 hours, the system triggers a reminder mechanism. The preset duration can also be flexibly adjusted according to different scenarios and the training status of the trainee to avoid physical fatigue or decreased attention caused by excessive training.

[0089] Once the system detects that the cumulative simulation duration of the trainee exceeds the preset value, the system automatically generates a first reminder message prompting the trainee to rest or pause training. The system can deliver this information to the trainee through the interface of the virtual reality driving simulator or the voice reminder function. After the trainee sees the prompt information, the system can also pause the training process or forcibly terminate the current driving scenario as needed. This mechanism ensures that the trainee does not face unnecessary health risks due to excessive training.

[0090] Through this step, the system can effectively manage the training duration of the trainee, avoid fatigue accumulation caused by long training, and thus improve the efficiency and effectiveness of training. Controlling the training duration not only helps to improve the trainee's attention concentration, but also ensures their physical health and reduces potential risks caused by fatigue driving.

[0091] In summary, the driving skill training method combined with virtual reality provided by the embodiments of the present application has the following technical effects:

[0092] 1. By combining a virtual reality driving simulator with a driving operation scoring model, an immersive training environment is provided, and precise data-based evaluation of trainee performance is achieved. This solution objectively reflects the trainee's operation ability in different driving scenarios, providing a reliable basis for personalized training, thereby improving training effectiveness and trainee skill level.

[0093] 2. By building a driving scenario database and selecting appropriate training scenarios based on the trainee's identity information, personalized customization of training content is achieved. This solution makes training more in line with the needs of trainees, improves training efficiency, and ensures that trainees train in scenarios suitable for their abilities, thereby better improving driving skills.

[0094] 3. By monitoring the cumulative simulation duration of the trainee and providing reminders, it avoids fatigue caused by long training, and improves the scientificity and safety of training. This mechanism effectively manages the training duration of the trainee, ensuring that the trainee can maintain the best state for driving training, thereby optimizing training effectiveness.

[0095] Embodiment two, based on the same inventive concept as the driving skill training method combined with virtual reality in the preceding embodiments, as Figure 2 shown, the embodiments of the present application provide a driving skill training system combined with virtual reality, which comprises:

[0096] The driving scene generation module 11 is configured to connect a virtual reality driving simulator, and generate a plurality of driving scenes according to the virtual reality driving simulator, wherein the virtual reality driving simulator comprises a VR wearable device; the simulation data set acquisition module 12 is configured to train a user to drive in the plurality of driving scenes according to the VR wearable device, and acquire a plurality of groups of driving simulation data sets, wherein each group of driving simulation data sets corresponds to one driving scene; the scoring model construction module 13 is configured to construct a driving operation scoring model, wherein the driving operation scoring model is connected with the virtual reality driving simulator; the scoring result output module 14 is configured to input the plurality of groups of driving simulation data sets into the driving operation scoring model, analyze each group of driving simulation data sets according to the driving operation scoring model, and output a plurality of driving operation scoring results; and the training report generation module 15 is configured to generate a driving skill training report according to the plurality of driving operation scoring results.

[0097] Further, the scoring result output module 14 is further configured to perform the following steps: the driving operation scoring model extracts features from each group of driving simulation data sets, and acquires low-dimensional features, wherein the low-dimensional features comprise brake features, speed features, steering features and reaction features; high-dimensional features are convolved according to the low-dimensional features, and high-dimensional features are output, wherein the high-dimensional features comprise driving stability, driving fluency, reaction rate and driving safety; and the driving operation scoring model analyzes the high-dimensional features, and acquires a driving operation scoring result corresponding to each group of driving simulation data sets.

[0098] Further, the scoring result output module 14 is further configured to perform the following steps: a plurality of driving scoring levels are set for each driving scene; the plurality of driving operation scoring results are matched with the driving scoring levels in each driving scene, and a matched driving scoring level in each driving scene is obtained; and a plurality of matched driving scoring levels in the plurality of driving scenes are integrated, and a comprehensive driving scoring level is generated.

[0099] Further, the scoring result output module 14 is further configured to perform the following steps: the plurality of driving operation scoring results are fed back to the virtual reality driving simulator, a plurality of driving training scenes are generated, and the training user is trained based on the plurality of driving training scenes; the virtual reality driving simulator compares the plurality of driving operation scoring results with a pre-stored driving operation scoring threshold, and screens N driving scenes less than the driving operation scoring threshold, wherein N is a positive integer greater than or equal to 1; and the N driving scenes are identified as a plurality of driving training scenes to be trained by the current training user.

[0100] Further, the driving scene generation module 11 is further configured to perform the following steps: constructing a driving scene database; inputting the identity information and training project information of the training user; verifying the identity information of the training user, and if the verification is passed, performing project scene screening in the driving scene database according to the training project information, and outputting a set of screened scenes; and generating a plurality of driving scenes corresponding to the training user according to the set of screened scenes.

[0101] Further, the driving scene generation module 11 is further configured to perform the following steps: wherein each driving scene in the driving scene database carries identification information, which is an index for measuring the operation difficulty of the driving scene; obtaining a set of identification information corresponding to the set of screened scenes; performing operation difficulty distribution uniformity identification on the set of identification information, and optimizing the plurality of driving scenes with a preset operation difficulty distribution uniformity as a target, and outputting the plurality of optimized driving scenes.

[0102] Further, the simulation data set acquisition module 12 is further configured to perform the following steps: obtaining the cumulative simulation time length of the training user in the plurality of driving scenes; and if the cumulative simulation time length is greater than a preset cumulative simulation time length, generating a first reminder information, and the virtual reality driving simulator reminds the training user to pause the training according to the first reminder information.

[0103] Any step of the above method can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor to implement any method in the embodiments of the present application. No redundant limitation is made here.

[0104] Further, the above-mentioned first or second order relationship also represents a specific concept, and / or refers to the selection of multiple elements individually or collectively. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A driving skills training method incorporating virtual reality, characterized in that: The method includes: Connect to a virtual reality driving simulator and generate multiple driving scenarios based on the virtual reality driving simulator, wherein the virtual reality driving simulator includes a VR wearable device; The driving scenarios include, but are not limited to, urban roads, mountain roads, highways, and adverse weather conditions; The training users conduct driving simulations in multiple driving scenarios using the VR wearable device to obtain multiple sets of driving simulation datasets, where each set of driving simulation datasets corresponds to a driving scenario; Construct a driving operation scoring model, wherein the driving operation scoring model is connected to the virtual reality driving simulator; The multiple sets of driving simulation datasets are input into the driving operation scoring model. The driving operation scoring model is used to analyze each set of driving simulation datasets and output multiple driving operation scoring results. Based on the multiple driving operation scores, a driving skills training report is generated; The method for analyzing each driving simulation dataset based on the driving operation scoring model includes: The driving operation scoring model extracts features from each driving simulation dataset to obtain low-dimensional features, including braking features, speed features, steering features, and reaction features. The high-dimensional features are convolved based on the low-dimensional features to output high-dimensional features, which include driving stability, driving smoothness, reaction rate and driving safety. The driving operation scoring model analyzes the high-dimensional features to obtain the driving operation scoring results corresponding to each driving simulation dataset. After outputting multiple driving operation score results, the method further includes: Multiple driving score levels can be set for each driving scenario; Based on the multiple driving operation score results, a score level is matched for each driving scenario to obtain the matched driving score level for each driving scenario; By integrating multiple matching driving score levels under the aforementioned driving scenarios, a comprehensive driving score level is generated.

2. The method as described in claim 1, characterized in that, After outputting multiple driving operation score results, the method also includes: The multiple driving operation scoring results are fed back to the virtual reality driving simulator to generate multiple driving training scenarios, and the training users train based on the multiple driving training scenarios. The virtual reality driving simulator compares the multiple driving operation score results with a pre-stored driving operation score threshold, and filters out N driving scenarios that are less than the driving operation score threshold, where N is a positive integer greater than or equal to 1. The N driving scenarios are identified as multiple driving training scenarios that the current training user needs to be trained on.

3. The method as described in claim 1, characterized in that, The method for generating multiple driving scenarios based on the virtual reality driving simulator includes: Build a driving scenario database; Enter the identity information of the training users and the training program information; The identity information of the training user is verified. If the identity verification is successful, the project scenario is filtered in the driving scenario database according to the training project information, and the filtered scenario set is output. Based on the selected scenario set, multiple driving scenarios corresponding to the training user are generated.

4. The method as described in claim 3, characterized in that, After outputting the set of filtered scenarios, the method also includes: Each driving scenario in the driving scenario database carries identification information, which is an indicator for measuring the difficulty of operating the driving scenario. Obtain the set of identifier information corresponding to the set of filtered scenarios; The set of identification information is used to identify the uniformity of the operation difficulty distribution. The multiple driving scenarios are optimized with the preset uniformity of the operation difficulty distribution as the target, and the optimized multiple driving scenarios are output.

5. The method as described in claim 1, characterized in that, The method for training users to perform driving simulations in multiple driving scenarios using the VR wearable device further includes: Obtain the cumulative simulation time of the training users in the multiple driving scenarios; If the cumulative simulation time exceeds the preset cumulative simulation time, a first reminder message is generated, and the virtual reality driving simulator reminds the training user to pause the training based on the first reminder message.

6. A driving skills training system incorporating virtual reality, characterized in that: The system is used to perform the driving skills training method incorporating virtual reality as described in any one of claims 1 to 5, the system comprising: A driving scenario generation module is used to connect to a virtual reality driving simulator and generate multiple driving scenarios based on the virtual reality driving simulator, wherein the virtual reality driving simulator includes a VR wearable device; The simulation dataset acquisition module is used to train users to perform driving simulations in multiple driving scenarios using the VR wearable device, and to acquire multiple sets of driving simulation datasets, wherein each set of driving simulation datasets corresponds to a driving scenario; A scoring model construction module is used to construct a driving operation scoring model, wherein the driving operation scoring model is connected to the virtual reality driving simulator; The scoring result output module is used to input the multiple sets of driving simulation datasets into the driving operation scoring model, analyze each set of driving simulation datasets according to the driving operation scoring model, and output multiple driving operation scoring results. The training report generation module is used to generate a driving skills training report based on the multiple driving operation scoring results.

Images