Human-vehicle driving behavior simulation method, device, system and readable storage medium

Abstract

A kind of human-vehicle driving behavior simulation method, device, system and readable storage medium, the human-vehicle driving behavior simulation method includes: obtaining scene information, the scene information at least includes driving data, environmental data and interactive data;Simulated driving behavior data is generated based on the scene information by pre-set generator, and the simulated driving behavior data is input into pre-set discriminator to generate discrimination result, to iteratively train the pre-set generator and the pre-set discriminator and corresponding simulated driving behavior data;Human-vehicle driving behavior is simulated based on the simulated driving behavior data, and the human-vehicle driving behavior at least includes: the driving interactive behavior of driver, the driving interactive behavior between driver and intelligent automobile and the driving interactive behavior between intelligent automobile.The present application solves the problem that the simulated driving behavior data generated in the prior art is not real and has insufficient social interaction.

Description

Technical Field

[0001] This invention belongs to the field of intelligent vehicle technology, specifically relating to a method, device, system, and readable storage medium for simulating human and vehicle driving behavior. Background Technology

[0002] With the popularization of autonomous driving technology, there are more and more cars with autonomous driving functions in the current transportation network. In the future, higher levels of autonomous driving functions will also become more widespread, and the interaction between human drivers and intelligent cars with autonomous driving functions is becoming increasingly important.

[0003] With the rapid development of autonomous driving technology, the demand for generating simulated driving behavior data is also increasing. Existing driving simulation technologies often fail to fully consider the social factors and interactivity in driving behavior, and cannot simulate how drivers react to the behavior and situations of other road users and autonomous vehicles. This results in simulated driving behavior data often lacking realism and social interactivity, making it impossible to accurately simulate the driver's driving habits and social interactions, as well as the interactions between the driver's car and other vehicles with autonomous driving functions—various situations and behaviors existing in real-world driving environments. This limits the research, testing, and evaluation of autonomous driving technology. For example, currently available simulated driving behavior data cannot simulate long-tail scenarios, making it difficult to assist autonomous driving technology in covering more complete scenarios and hindering the realism and fidelity of the scenarios. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a method, device, system and readable storage medium for simulating human and vehicle driving behavior, which solves at least one of the problems in the background art described above.

[0005] To achieve the above objectives, the present invention adopts the following technical solution.

[0006] The first aspect of this invention provides a method for simulating human and vehicle driving behavior, comprising: Acquire scene information, which includes at least driving data, environmental data, and interaction data; Based on the scene information, a preset generator generates simulated driving behavior data, and the simulated driving behavior data is input into a preset discriminator to generate a discrimination result. The preset generator, the preset discriminator, and the corresponding simulated driving behavior data are then iteratively trained. The simulated driving behavior data is used to simulate human and vehicle driving behavior, which includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles.

[0007] As one embodiment of the present invention, simulating the driving behavior of a human driver based on the simulated driving behavior data includes: Acquire scene information including steering wheel angle, accelerator pedal opening, and brake pedal opening based on actual driving behavior; Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in real driving behavior, the generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the steering wheel angle, accelerator pedal opening and brake pedal opening of the simulated driving behavior. Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in the simulated driving behavior data, the driving behavior of a human driver is simulated when meeting oncoming traffic, overtaking, avoiding obstacles and / or changing lanes.

[0008] As one embodiment of the present invention, simulating driving behavior between a driver and an intelligent vehicle based on the simulated driving behavior data includes: Acquire scene information including real-world driving behavior vehicle sensor data, environmental road rules, and environmental road conditions; Based on vehicle sensor data, environmental road rules, and environmental road conditions from real driving behavior, the generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes vehicle sensor data, environmental road rules, and environmental road conditions from simulated driving behavior. Based on the vehicle sensor data, the environmental road rules, and the environmental road conditions, the steering and shifting behavior of the intelligent vehicle is simulated and predicted. Based on the simulated and predicted steering and shifting behavior of intelligent vehicles, the system simulates the driver's reaction to the intelligent vehicle, the driver's driving behavior, and / or the behavior of the intelligent vehicle after detecting changes in the motion of other vehicles.

[0009] As one embodiment of the present invention, simulating driving behavior between intelligent vehicles based on the simulated driving behavior data includes: Acquire scene information including real-world driving environment conditions, traffic flow, and the driving behavior of other vehicles; Based on the environmental road conditions, traffic flow and other vehicle driving behavior in real driving behavior, the generator generates simulated driving behavior data that meets the preset discrimination threshold of the discriminator. The simulated driving behavior data includes the environmental road conditions, traffic flow and other vehicle driving behavior of the simulated driving behavior. Predict the speed, acceleration, and steering of the intelligent vehicle based on the environmental road conditions, traffic flow, and the driving behavior of other vehicles. Based on the speed, acceleration, and steering of the intelligent vehicle and the driving behavior of other vehicles, the simulation simulates the passing, overtaking, obstacle avoidance, and / or lane changing behaviors between intelligent vehicles.

[0010] As one embodiment of the present invention, after simulating human and vehicle driving behavior based on the simulated driving behavior data, the simulated driving behavior data corresponding to each human and vehicle driving behavior is used as a new input to the preset generator to generate new simulated driving behavior data.

[0011] As one embodiment of the present invention, after simulating human and vehicle driving behavior based on the simulated driving behavior data, the method further includes: Based on a preset evaluation index system, the simulated driving behavior data for human-vehicle driving behavior simulation is evaluated item by item and evaluation results are generated. The evaluation results are used to assist in adjusting the generator, the discriminator and their respective parameters and / or structures.

[0012] A second aspect of the present invention provides a human-vehicle driving behavior simulation device, comprising: The acquisition unit is at least used to acquire scene information, which includes at least driving data, environmental data, and interaction data; the training unit is at least used to generate simulated driving behavior data based on the scene information using a preset generator, input the simulated driving behavior data into a preset discriminator to generate a discrimination result, and iteratively train the preset generator, the preset discriminator, and the corresponding simulated driving behavior data. The simulation unit is used at least to simulate human and vehicle driving behavior based on the simulated driving behavior data, wherein the human and vehicle driving behavior includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles.

[0013] A third aspect of the present invention provides a human-vehicle driving behavior simulation system, comprising: The data acquisition module is used to acquire scene information, which includes at least driving data, environmental data, and interaction data. The GAN module is at least used to generate simulated driving behavior data based on the scene information via a preset generator, input the simulated driving behavior data into a preset discriminator to generate a discrimination result, and iteratively train the preset generator and the preset discriminator. The simulation module is used at least to simulate human and vehicle driving behavior based on the simulated driving behavior data, wherein the human and vehicle driving behavior includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles.

[0014] In one embodiment of the present invention, the simulation module includes at least a first simulation submodule, the first simulation submodule being used for at least: Acquire scene information including steering wheel angle, accelerator pedal opening, and brake pedal opening based on actual driving behavior; Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the steering wheel angle, accelerator pedal opening and brake pedal opening of the simulated driving behavior. Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in the simulated driving behavior data, the driving behavior of a human driver, such as meeting oncoming traffic, overtaking, avoiding obstacles and changing lanes, is simulated.

[0015] In one embodiment of the present invention, the simulation module includes at least a second simulation submodule, the second simulation submodule being used for at least: Acquire scene information including real-world driving behavior vehicle sensor data, environmental road rules, and environmental road conditions; Based on vehicle sensor data, environmental road rules, and environmental road conditions from real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes vehicle sensor data, environmental road rules, and environmental road conditions from simulated driving behavior. Based on the vehicle sensor data, the environmental road rules, and the environmental road conditions, the steering and shifting behavior of the intelligent vehicle is simulated and predicted. Based on the simulated and predicted steering and shifting behavior of intelligent vehicles, the system simulates the reaction behavior of human drivers to intelligent vehicles, the driving behavior of human drivers, and the behavior of intelligent vehicles after detecting changes in the motion of other vehicles.

[0016] As one embodiment of the present invention, the simulation module includes at least a third simulation submodule, the third simulation submodule being used for at least: Acquire scene information including real-world driving environment conditions, traffic flow, and the driving behavior of other vehicles; Based on the environmental road conditions, traffic flow and other vehicle driving behavior in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the environmental road conditions, traffic flow and other vehicle driving behavior of the simulated driving behavior. Predict the speed, acceleration, and steering of the intelligent vehicle based on the environmental road conditions, traffic flow, and the driving behavior of other vehicles. Based on the speed, acceleration, and steering of the intelligent vehicle and the driving behavior of other vehicles, the simulation shows the passing, overtaking, obstacle avoidance, and lane changing behaviors between intelligent vehicles.

[0017] A fourth aspect of the present invention provides an electronic device, comprising: At least one processor; and at least one memory communicatively connected to the processor, wherein the memory stores program instructions executable by the processor, and the processor invokes the program instructions to perform the steps of the human-vehicle driving behavior simulation method as described in the first aspect of the present invention.

[0018] The fifth aspect of the present invention provides a readable storage medium storing a computer program, which is executed by a processor as the steps of the human-vehicle driving behavior simulation method described in the first aspect of the present invention.

[0019] In summary, compared with the prior art, the present invention has at least one of the following beneficial technical effects: 1. This invention, through adversarial training of a preset generator and a preset discriminator, can generate more realistic and socially interactive simulated human driver and / or intelligent vehicle driving behavior data; 2. This invention uses GAN to simulate the driving habits and social interaction behaviors of human drivers (human-to-human), the interaction behaviors between a human driver's car and an intelligent car with autonomous driving function activated (human-to-car), and the interaction behaviors between intelligent cars with autonomous driving function activated (car-to-car). This improves the realism and social interactivity of simulated driving behavior data and provides strong support for the research, development, testing and evaluation of autonomous driving technology. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating a method for simulating human and vehicle driving behavior in a specific embodiment of the present invention.

[0022] Figure 2 This is a block diagram of a human-vehicle driving behavior simulation device in a specific embodiment of the present invention.

[0023] Figure 3 This is a block diagram of a human-vehicle driving behavior simulation system in a specific embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Furthermore, it should be understood that the specific embodiments described herein are only for illustration and explanation of this application and are not intended to limit this application.

[0026] It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments of this application. Furthermore, the descriptions of each embodiment in the following embodiments have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0027] Generative Adversarial Networks (GANs) consist of two basic networks: a generator (G, also known as a generator network) and a discriminator (D, also known as a discriminator network). The generator produces new data, often based on a set of noise or random numbers, while the discriminator determines whether the generated data or the real data is true. The generator is unlabeled, making it an unsupervised network; the discriminator is labeled, making it a supervised network, with labels indicating "false" or "true" (0 and 1). As a deep learning model, GANs have the ability to generate and discriminate samples, and can be used to simulate the driving behavior of different drivers.

[0028] The intelligent vehicle in this invention is specifically a vehicle that has activated its autonomous driving function.

[0029] The human-vehicle driving behavior in this invention includes, but is not limited to, the driving behavior of a person (taking a human driver as an example) in different scenarios, the driving behavior between a driver and an intelligent vehicle, and the driving behavior between intelligent vehicles. Driving behavior is a summary term and can be understood as driving habits, interactive behavior, collaborative behavior, or communication behavior, etc. Its name can be changed according to the specific implementation scenario without changing its essence as driving behavior included in this invention, which will not be elaborated here.

[0030] like Figure 1 As shown, the first aspect of the present invention provides a method for simulating human and vehicle driving behavior, which mainly includes the following steps.

[0031] Step S110: Obtain scene information, which includes at least driving data, environmental data, and interaction data.

[0032] Specifically, the scene information includes real driving data, environmental road data, and interaction data with other road users collected through high-resolution cameras, radar, GPS, and vehicle sensors. This includes data such as human driver behavior data, vehicle status data, weather data, lighting data, and data from other intelligent vehicles. The data involved in this application is not limited to these; it can also originate from surveillance video data of actual driving scenarios and various data from V2X, such as traffic flow. Generally, the scene information acquired by this invention is not limited to a single data set or group of data, nor is the data name limited. Any scene information that can be used to simulate human and vehicle driving behavior in this invention, and which can be obtained in the prior art, falls within the scope of protection of this invention. Examples include steering wheel angle, accelerator pedal opening, brake pedal opening, vehicle sensor data, environmental road rules, environmental road conditions, traffic flow, lighting intensity, and the driving behavior of other vehicles.

[0033] Step S120: Based on the scene information, simulated driving behavior data is generated by a preset generator, and the simulated driving behavior data is input into a preset discriminator to generate a discrimination result, so as to iteratively train the preset generator, the preset discriminator and the corresponding simulated driving behavior data.

[0034] Specifically, step S120 is used to define the model architecture. Based on the characteristics of GAN, corresponding generators and discriminators are pre-configured to obtain preset generators and preset discriminators. A Deep Convolutional Neural Network (DCNN) is configured as the generator. The input to the preset generator is random noise and corresponding scene information, and the output of the preset generator is simulated driving behavior data representing simulated driving behavior. Specifically, the simulated driving behavior includes four types of driving behaviors: meeting oncoming traffic, overtaking, obstacle avoidance, and lane changing. The function of the preset generator is to generate simulated driving behavior data to learn and extract information from scene information that reflects the driver's reaction, the behavior of other road users, and situational features, such as weather, traffic signs, and road type information, specifically rain and fog, intersections, highways, and urban roads, etc.

[0035] Furthermore, a multilayer perceptron (MLP) is configured as a preset discriminator, taking simulated driving behavior data as input and outputting the discrimination result. The preset discriminator's role is to generate discrimination results to assess the realism of the simulated driving behavior.

[0036] Furthermore, the present invention improves the realism of the simulated driving behavior generated by the preset generator and improves the accuracy of the preset discriminator in judging the realism of the simulated driving behavior by means of the following methods.

[0037] Initialize model parameters: Use a random number generator to initialize the parameters of the preset generator and the preset discriminator (the preset generator is hereinafter referred to as the generator, and the preset discriminator is hereinafter referred to as the discriminator).

[0038] Training generator: Using random noise and scene information as input, a generator network structure generates simulated driving behavior data, such as driver behavior data that simulates human driver driving habits or interactions. Essentially, the data from intelligent vehicles and real human drivers is of the same data type, encompassing four interaction behaviors: meeting oncoming traffic, overtaking, obstacle avoidance, and lane changing. The difference lies in whether the driving entity is a computer or a human. This distinction is made because the driving logic of computers differs from that of humans, stemming from their internal algorithms and driving styles. Furthermore, the random noise, represented by random numbers, drives the generator to produce diverse data samples. Different random numbers input into the generator produce different results, allowing for continuous generation of varied outcomes. The simulated driving behavior data generated by the generator is then input into a discriminator to obtain its output, or discrimination result. A loss function is calculated based on the discriminator's output, measuring the realism and quality of the generated simulated driving behavior data. Specifically, the generator's loss function can be a cross-entropy loss function. The generator's parameters are updated using the gradient descent optimization algorithm to minimize the loss function. After each update of the generator's parameters, a new batch of simulated driving behavior data is generated and input into the discriminator.

[0039] Training the discriminator: Using scene information from human drivers and intelligent vehicles as input, the discrimination result is obtained through the discriminator's network structure. Similarly, the loss function is calculated based on the discriminator's discrimination result. In order to handle outliers while ensuring robust loss measurement, the discriminator's loss function is called Huber Loss. Huber Loss is a loss function that combines the characteristics of mean squared error and mean absolute error. When the error is small, it is close to the mean squared error, and when the error is large, it is close to the mean absolute error. A suitable discrimination threshold can be obtained based on this. An optimization algorithm is used to update the discriminator's parameters to minimize the loss function. After each parameter update, a new batch of scene information and simulated driving behavior data generated by the generator are re-evaluated. Specifically, because the discriminator updates its parameters based on the evaluation results each time—for example, if real driving behavior data is input but the discriminator interprets it as fake—the discriminator parameters need to be optimized.

[0040] Iterative optimization: During training, the steps of training the generator and discriminator need to be iterated repeatedly, and the relevant parameters of the generator and discriminator are continuously updated. In each iteration, the generator and discriminator compete and learn from each other, enabling the generator to gradually generate more realistic simulated data, while the discriminator can more accurately judge the authenticity of the data. This iterative optimization process aims to improve the training effect of GAN and the realism of simulated human and vehicle driving behavior.

[0041] Step S130: Simulate human-vehicle driving behavior based on the simulated driving behavior data. The human-vehicle driving behavior includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles.

[0042] This invention uses GANs to simulate the driving habits and social interaction behaviors of human drivers (human-to-human), the interaction behaviors between a human driver's car and an intelligent car with autonomous driving function activated (human-to-car), and the interaction behaviors between intelligent cars with autonomous driving function activated (car-to-car). This improves the realism and social interactivity of simulated driving behavior data and provides strong support for the research, development, testing and evaluation of autonomous driving technology.

[0043] In one embodiment of the present invention, the driving behavior of a driver is simulated based on the simulated driving behavior data, including: acquiring scene information including the steering wheel angle, accelerator pedal opening and brake pedal opening of the actual driving behavior; Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the steering wheel angle, accelerator pedal opening and brake pedal opening of the simulated driving behavior. The simulated driving behavior data, including steering wheel angle, accelerator pedal opening, and brake pedal opening, simulates the driver's actions such as meeting oncoming traffic, overtaking, avoiding obstacles, and changing lanes. The speed, acceleration, and steering of the intelligent vehicle are specifically reflected by the steering wheel angle, accelerator pedal opening, and brake pedal opening.

[0044] Specifically, the embodiment introduces a Generative Adversarial Network (GAN) to simulate the behavior and reactions of human drivers, utilizing the generator and discriminator within the GAN for adversarial training. During training, the generator is responsible for generating fake data similar to real driving behavior data, i.e., simulated driving behavior data, to reflect simulated driving behavior and reactions. The discriminator is responsible for determining whether this data is real or generated fake data. The generator continuously optimizes the generated fake data, making it difficult for the discriminator to distinguish between real and fake data, while the discriminator continuously updates its parameters to improve its ability to distinguish between real and fake data. Both real and simulated driving behaviors refer to four types of driver interaction behaviors: meeting oncoming traffic, overtaking, obstacle avoidance, and lane changing. The specific measurement parameters are steering wheel angle, accelerator pedal opening, and brake pedal opening. In terms of specific measurement parameters, human drivers achieve driving through three parameters: steering wheel angle, accelerator pedal opening, and brake pedal parameters. For example, to simulate a human driver's overtaking behavior, it is necessary to increase the accelerator pedal opening and steer to overtake the vehicle in front. These interaction behaviors are all driver operations, while the driver's actual operations when driving a vehicle are only acceleration, deceleration, and steering.

[0045] In this way, the driving habits and social interaction behaviors of human drivers can be learned, and simulated driving behavior data with realism and social interaction can be generated. This data is used for the research, development, testing and evaluation of autonomous driving technology, thereby improving the performance and safety of autonomous driving systems.

[0046] Furthermore, during the training process, this invention uses a loss function to optimize the parameters of the driver's driving behavior so that it can better simulate the driver's social driving habits. Regularization is introduced to improve the simulation of the driver's driving behavior and ensure the consistency of different data.

[0047] In one embodiment of the present invention, simulating driving behavior between a driver and an intelligent vehicle based on the simulated driving behavior data includes: Acquire scene information including real-world driving behavior vehicle sensor data, environmental road rules, and environmental road conditions; Based on vehicle sensor data, environmental road rules, and environmental road conditions from real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes vehicle sensor data, environmental road rules, and environmental road conditions from simulated driving behavior. Based on the vehicle sensor data, the environmental road rules, and the environmental road conditions, the steering and speed change behavior of the intelligent vehicle is simulated and predicted. The speed change behavior specifically refers to acceleration or deceleration behavior. Specifically, the sensor data includes steering wheel angle, accelerator pedal opening, and brake pedal opening. The environmental road rules include traffic rules, such as not running red lights and yielding to pedestrians. The environmental road conditions specifically refer to highways or urban roads and the driving conditions of other vehicles. Based on the simulated and predicted steering and shifting behavior of intelligent vehicles, the system simulates the reaction behavior of human drivers to intelligent vehicles, the driving behavior of drivers, and the behavior of intelligent vehicles after detecting changes in the motion of other vehicles.

[0048] This invention can generate more realistic and socially interactive driving behavior data between human drivers and intelligent vehicles.

[0049] In one embodiment of the present invention, the simulation based on the driving behavior between the simulated intelligent vehicles includes: Acquire scene information including real-world driving environment conditions, traffic flow, and the driving behavior of other vehicles; Based on the environmental road conditions, traffic flow and other vehicle driving behavior in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the environmental road conditions, traffic flow and other vehicle driving behavior of the simulated driving behavior. The speed, acceleration, and steering of the intelligent vehicle are predicted based on the environmental road conditions, traffic flow, and the driving behavior of other vehicles. Specifically, traffic flow is used to count the number of vehicles within a preset geographical area; an excessive number of vehicles will interfere with the vehicle's driving. The driving behavior of other vehicles indicates whether it will interfere with the vehicle's behavior. The vehicle's speed, acceleration, steering, and the driving behavior of other vehicles are used to simulate passing, overtaking, obstacle avoidance, and lane changing behaviors between intelligent vehicles.

[0050] In this invention, by predicting the possible behaviors of other intelligent vehicles and then inputting this information into a preset generator, more realistic and socially interactive simulated driving behavior data can be generated.

[0051] In one embodiment of the present invention, after simulating human and vehicle driving behavior based on the simulated driving behavior data, the simulated driving behavior data corresponding to each human and vehicle driving behavior is used as a new input to the preset generator to generate new simulated driving behavior data.

[0052] Specifically, regardless of which of the above-mentioned interactive behaviors, the main model structure used is the same. The main difference lies in the specific content of the required scene information. Some data comes from people, some from people and cars, and some from cars.

[0053] In this way, the present invention solves the problem of insufficient realism and social interactivity in the generation of simulated driving behavior data in the prior art.

[0054] It should be noted that, because the discriminator has a certain optimization algorithm, different types of simulated driving behavior data—such as data between human drivers, between human drivers and intelligent vehicles, and between intelligent vehicles—will generate increasingly different models after continuous training. The parameters trained on these three models will also differ. Similarly, due to the differences in driving behaviors such as meeting oncoming traffic, overtaking, obstacle avoidance, and lane changing, the input and output of data between human drivers, between human drivers and intelligent vehicles, and between intelligent vehicles will also produce four different types. Combining the three driving interaction behaviors and the four driving behaviors, this application can summarize 4*3=12 typical human-vehicle driving behavior models.

[0055] In one embodiment of the present invention, after simulating human and vehicle driving behavior based on the simulated driving behavior data, the method further includes: Based on a preset evaluation index system, the simulated driving behavior data for human-vehicle driving behavior simulation is evaluated item by item and evaluation results are generated. The evaluation results are used to assist in adjusting the generator, the discriminator and their respective parameters and / or structures.

[0056] Specifically, a pre-defined automated evaluation mechanism can assess the authenticity and validity of simulated driving behavior data. This mechanism compares the statistical characteristics and behavioral patterns of simulated driving behavior data with those of real driving behavior data. This ensures the quality of the simulated data, thereby improving the accuracy and reliability of autonomous driving technology research, testing, and evaluation.

[0057] Specifically, the preset evaluation index system shall at least satisfy data consistency, logical rationality, scenario adaptability and / or prediction accuracy.

[0058] Specifically, data consistency assessment includes evaluating whether simulated driving behavior data is consistent with real driving behavior data in terms of statistical characteristics and distribution, which is achieved by comparing the distribution, mean, standard deviation and other indicators of simulated driving behavior data and real driving behavior data.

[0059] Specifically, the behavioral logic assessment includes: evaluating whether the simulated driver behavior and intelligent vehicle behavior conform to actual driving logic and traffic rules, and specifically assessing whether the driver's decisions and reactions in behaviors such as meeting oncoming traffic, overtaking, avoiding obstacles, and changing lanes are reasonable.

[0060] Specifically, scenario adaptability assessment includes evaluating the adaptability and generalization ability of simulated driving behavior data under different traffic scenarios. This can be achieved by testing simulated data in three traffic scenarios (urban roads, highways, and intersections).

[0061] Specifically, the accuracy assessment includes: for simulated driving behavior data involving behavior prediction, evaluating the accuracy of the prediction model is crucial, which is achieved by comparing the consistency of the prediction results with the actual driver behavior or intelligent vehicle behavior.

[0062] To run an automated evaluation mechanism, an evaluation script or tool can be pre-set. This script or tool can automatically load simulated driving behavior data and perform item-by-item evaluation according to a pre-set evaluation index system. The evaluation results are presented in three forms: numerical, graphical, and report, so that R&D personnel can intuitively understand the authenticity and validity of the simulated driving behavior data.

[0063] In addition, to improve the accuracy and effectiveness of the evaluation, expert review and user feedback can be introduced to further improve the evaluation process. Statistical analysis methods can also be used to evaluate the authenticity and effectiveness of the simulated data, using three indicators such as mean squared error, mean absolute error, and correlation coefficient to measure the consistency between the simulated data and the real data.

[0064] If the evaluation results indicate that the simulated data has a large bias, the GAN will initiate a reverse adjustment strategy to adjust the training and optimization process. The adjustment strategy includes adjusting the loss function, optimizing the structure of the generator and discriminator, and adjusting the training parameters. After adjustment, the training and evaluation will be carried out again until the preset evaluation criteria are met.

[0065] This invention, through adversarial training of a pre-set generator and a pre-set discriminator, can generate more realistic and socially interactive simulated human driver and intelligent vehicle driving behavior data. It also effectively simulates long-tail scenarios, helping autonomous driving technology cover a wider range of diverse, low-probability, or sudden scenarios. Compared to traditional methods of generating simulated driving behavior data, this invention offers greater realism and fidelity. Furthermore, this invention can be applied to multiple fields such as autonomous vehicle development, driving behavior analysis, traffic planning, and safety assessment. More realistic simulated driving behavior helps improve the safety and reliability of autonomous driving systems. Simultaneously, this invention can also be used to study driver behavior patterns and traffic flow characteristics, providing strong support for traffic planning and optimization.

[0066] like Figure 2 As shown, a second aspect of the present invention provides a human-vehicle driving behavior simulation device, comprising: The acquisition unit 210 is used to acquire scene information, which includes at least driving data, environmental data, and interaction data. Training unit 220 generates simulated driving behavior data based on the scene information via a preset generator, and inputs the simulated driving behavior data into a preset discriminator to generate a discrimination result, so as to iteratively train the preset generator, the preset discriminator and the corresponding simulated driving behavior data; The simulation unit 230 is used at least to simulate human and vehicle driving behavior based on the simulated driving behavior data, wherein the human and vehicle driving behavior includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles.

[0067] like Figure 3 As shown, a third aspect of the present invention provides a human-vehicle driving behavior simulation system, comprising: The data acquisition module 310 is used to acquire scene information, which includes at least driving data, environmental data, and interaction data. GAN module 320 is at least used to generate simulated driving behavior data based on the scene information via a preset generator, input the simulated driving behavior data into a preset discriminator to generate a discrimination result, and iteratively train the preset generator and the preset discriminator; The simulation module 330 is used at least to simulate human and vehicle driving behavior based on the simulated driving behavior data, wherein the human and vehicle driving behavior includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles.

[0068] This system is used in multiple fields, including the development of autonomous vehicles, driving behavior analysis, traffic planning, and safety assessment. By simulating driving behavior more realistically, it helps improve the safety and reliability of autonomous driving systems. Simultaneously, this system can also be used to study driver behavior patterns and traffic flow characteristics, providing strong support for traffic planning and optimization. Furthermore, this system features real-time performance, scalability, modular design, security and privacy protection, and ease of integration and deployment.

[0069] Please continue reading. Figure 3 In one embodiment of the present invention, the simulation module 330 includes at least a first simulation submodule 3301, which is at least used for: Acquire scene information including steering wheel angle, accelerator pedal opening, and brake pedal opening based on actual driving behavior; Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the steering wheel angle, accelerator pedal opening and brake pedal opening of the simulated driving behavior. Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in the simulated driving behavior data, the driving behavior of a human driver, such as meeting oncoming traffic, overtaking, avoiding obstacles and changing lanes, is simulated.

[0070] Specifically, the first simulation submodule 3301, also known as the human driving habit module, can simulate how human drivers react to the behavior and situations of other road users. The human driving habit module simulates the behavior and reactions of human drivers by introducing GAN and uses the preset generator and preset discriminator in GAN for adversarial training.

[0071] During training, a pre-generated generator is responsible for generating fake data similar to real driving data, i.e., simulated driving behaviors and reactions. A pre-generated discriminator is responsible for determining whether this data is real or generated fake data. The pre-generated generator continuously optimizes the generated fake data, making it difficult for the pre-generated discriminator to distinguish between real and fake data. The pre-generated discriminator, in turn, continuously updates its parameters to improve its ability to differentiate between real and fake data. The simulated driving behaviors and reactions refer to four types of interactive behaviors: meeting oncoming traffic, overtaking, obstacle avoidance, and lane changing. Specific parameters for measurement include steering wheel angle, accelerator pedal opening, and brake pedal opening.

[0072] In this way, the human driving habit module can learn the driving habits and social interaction behaviors of human drivers, and generate simulated driving behavior data with realism and social interaction. This data is used for the research, testing, and evaluation of autonomous driving technology, improving the performance and safety of autonomous driving systems.

[0073] During training, a loss function is used to optimize the parameters of the human driving habits module, enabling it to better simulate the social driving habits of drivers. Furthermore, this invention introduces regularization to enhance the simulation of social habits and ensure consistency across different data sets.

[0074] Please continue reading. Figure 3 In one embodiment of the present invention, the simulation module 330 includes at least a second simulation submodule 3302, the second simulation submodule 3302 being used for at least: Acquire scene information including real-world driving behavior vehicle sensor data, environmental road rules, and environmental road conditions; Based on vehicle sensor data, environmental road rules, and environmental road conditions from real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes vehicle sensor data, environmental road rules, and environmental road conditions from simulated driving behavior. Based on the vehicle sensor data, the environmental road rules, and the environmental road conditions, the steering and shifting behavior of the intelligent vehicle is simulated and predicted. Based on the simulated and predicted steering and shifting behavior of intelligent vehicles, the system simulates the driver's reaction to the intelligent vehicle, the driver's driving behavior, and the behavior of the intelligent vehicle after detecting changes in the motion of other vehicles.

[0075] Specifically, the second simulation submodule 3302, also known as the interaction module between human drivers and intelligent vehicles, is similar to the human driving habits module. This module considers how human drivers cooperate with other intelligent vehicles that have activated autonomous driving functions. Through this module, the interaction behavior between human drivers and intelligent vehicles can be simulated.

[0076] It should be noted that simulating the interaction between human drivers and other intelligent vehicles that have activated autonomous driving functions is a complex but crucial part. This module can simulate scenarios in multi-vehicle environments, where human drivers interact with other vehicles that are also equipped with autonomous driving systems. The simulation and prediction models for the driving behavior of human drivers and intelligent vehicles mainly involve two levels of simulation and prediction.

[0077] Level 1: Simulation and prediction of intelligent vehicle behavior, which involves simulating and predicting the behavior of intelligent vehicles, including steering, acceleration, and deceleration, based on sensor data, traffic rules, and road conditions.

[0078] Level Two: Simulation and Prediction of Human Driver Reactions: Based on the simulation and prediction of the steering and shifting behavior of intelligent vehicles, predict how human drivers will react to these behaviors.

[0079] In one application scenario of the invention, the interaction steps between the driver and other intelligent vehicles that have activated autonomous driving functions are as follows.

[0080] Environment initialization: First, set up a simulated traffic environment, including roads, traffic signals, and 10 cars. Of the 10 cars, 5 are ordinary cars driven by human drivers, and 5 are intelligent cars with autonomous driving functions activated.

[0081] Behavioral prediction: Then, for each intelligent vehicle that activates its autonomous driving function, a behavior prediction model is used to predict its possible behavior. At the same time, for human drivers, their possible reactions also need to be predicted.

[0082] Interactive simulation: Next, based on these predictions, the interaction between human drivers and other intelligent vehicles is simulated. This includes the human driver's reaction to the behavior of the intelligent vehicle, adjusting their own driving strategy, and the corresponding measures taken by the intelligent vehicle that has activated the autonomous driving function when it detects changes in the motion of other vehicles.

[0083] Result evaluation: Finally, assess the impact of this interaction on overall traffic flow and how it affects the performance of autonomous driving systems.

[0084] Here, the module receives real driving behavior data as input and uses a simulation model to predict the driver's possible behavior. This information is then input into the generator to generate more realistic and socially interactive simulated driving behavior data.

[0085] Please continue reading. Figure 3 In one embodiment of the present invention, the simulation module 330 includes at least a third simulation submodule 3303, which is at least used for: Acquire scene information including real-world driving environment conditions, traffic flow, and the driving behavior of other vehicles; Based on the environmental road conditions, traffic flow and other vehicle driving behavior in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the environmental road conditions, traffic flow and other vehicle driving behavior of the simulated driving behavior. Predict the speed, acceleration, and steering of the intelligent vehicle based on the environmental road conditions, traffic flow, and the driving behavior of other vehicles. Based on the speed, acceleration, and steering of the intelligent vehicle and the driving behavior of other vehicles, the simulation shows the passing, overtaking, obstacle avoidance, and lane changing behaviors between intelligent vehicles.

[0086] Specifically, the third simulation submodule 3303, also known as the intelligent vehicle interaction module, is similar to the first and second simulation submodules mentioned above. It simulates the interactive behaviors between intelligent vehicles. This module considers communication protocols and standards, behavior prediction and decision coordination, and safety and obstacle avoidance. The intelligent vehicle interaction module simulates the communication and collaborative behaviors between intelligent vehicles, similar to social interactions between human drivers. This module simulates how intelligent vehicles interact with other vehicles; like the human driving habits module, the interactive behaviors include meeting oncoming traffic, overtaking, obstacle avoidance, and lane changing.

[0087] The behavior prediction model is based on deep learning. This model predicts vehicle behavior, including speed, acceleration, and steering, by analyzing road conditions ahead, traffic flow, and the driving behaviors of other vehicles. Deep learning-based prediction models can learn from large amounts of data and utilize the pattern recognition capabilities of neural networks to accurately predict vehicle behavior. This approach exists in existing technologies and will not be elaborated upon further.

[0088] Specifically, when the interaction module between intelligent vehicles is running, it first needs to input vehicle data from real-world traffic scenarios into the behavior prediction model. Then, the model predicts the future behavior of the vehicles based on this data. Next, the interaction module generates corresponding interactive behaviors based on the prediction results and information from other vehicles. These behaviors include passing, overtaking, obstacle avoidance, and lane changing. Finally, these simulated interactive behaviors are output to a simulation environment for use in the research, testing, and evaluation of autonomous driving technology.

[0089] Here, this module receives data from other smart cars and other relevant contextual data as input, and predicts the possible behaviors of other smart cars through a predictive model. This information is then input into a preset generator to generate more realistic and socially interactive simulated driving behavior data.

[0090] Please continue reading. Figure 3 In one embodiment of the present invention, the human-vehicle driving behavior simulation system further includes an evaluation module 340. The evaluation module evaluates the simulated driving behavior data for human-vehicle driving behavior simulation item by item based on at least a preset evaluation index system and generates evaluation results. The evaluation results are used to assist in adjusting the generator, the discriminator and their respective parameters and / or structures.

[0091] Specifically, the evaluation module can be used to compare the statistical characteristics and behavioral patterns of simulated driving behavior with real driving behavior, and automatically evaluate the authenticity and validity of simulated driving behavior data.

[0092] The present invention proposes a human driving behavior simulation system based on generative adversarial networks, which considers not only the interaction between simulated drivers and other road users by utilizing human driving habits, but also the interaction between human drivers and intelligent vehicles, as well as the interaction between intelligent vehicles.

[0093] like Figure 4 As shown, a fourth aspect of the present invention provides an electronic device, comprising: At least one processor; and at least one memory communicatively connected to the processor, wherein the memory stores program instructions executable by the processor, and the processor invokes the program instructions to perform the steps of the method as described in any of the above embodiments.

[0094] The fifth aspect of the present invention discloses a readable storage medium storing a computer program, which is executed by a processor as described in any of the above embodiments.

[0095] Computer-readable storage media can include: any entity or device capable of carrying computer programs, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc. A computer program includes computer program code. Computer program code can be in the form of source code, object code, executable files, or certain intermediate forms, etc. Computer-readable storage media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc.

[0096] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a system including a processing module or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0098] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for simulating human-vehicle driving behavior, characterized in that, include: Acquire scene information, which includes at least driving data, environmental data, and interaction data; Pre-configure the corresponding generator and discriminator to obtain the preset generator and preset discriminator. Based on the scene information, generate simulated driving behavior data through the preset generator, and input the simulated driving behavior data into the preset discriminator to generate a discrimination result, so as to iteratively train the preset generator, the preset discriminator and the corresponding simulated driving behavior data; Based on the simulated driving behavior data, the human-vehicle driving behavior is simulated, which includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles. The simulation of human driver behavior based on the simulated driving behavior data includes: Acquire scene information including steering wheel angle, accelerator pedal opening, and brake pedal opening in real driving behavior; Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the steering wheel angle, accelerator pedal opening and brake pedal opening of the simulated driving behavior. Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in the simulated driving behavior data, the driving behavior of a human driver is simulated when meeting oncoming traffic, overtaking, avoiding obstacles and / or changing lanes.

2. The method for simulating human-vehicle driving behavior according to claim 1, characterized in that, Simulating driving behavior between the driver and the intelligent vehicle based on the simulated driving behavior data includes: Acquire scene information including real-world driving behavior vehicle sensor data, environmental road rules, and environmental road conditions; Based on vehicle sensor data, environmental road rules, and environmental road conditions from real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes vehicle sensor data, environmental road rules, and environmental road conditions from simulated driving behavior. Based on the vehicle sensor data, the environmental road rules, and the environmental road conditions, the steering and shifting behavior of the intelligent vehicle is simulated and predicted. Based on the simulated and predicted steering and shifting behavior of intelligent vehicles, the system simulates the driver's reaction to the intelligent vehicle, the driver's driving behavior, and / or the behavior of the intelligent vehicle after detecting changes in the motion of other vehicles.

3. The method for simulating human-vehicle driving behavior according to claim 1, characterized in that, Simulating driving behavior among intelligent vehicles based on the simulated driving behavior data includes: Acquire scene information including real-world driving environment conditions, traffic flow, and the driving behavior of other vehicles; Based on the environmental road conditions, traffic flow and other vehicle driving behavior in real driving behavior, the preset generator generates simulated driving behavior data that meets the preset discrimination threshold of the discriminator. The simulated driving behavior data includes the environmental road conditions, traffic flow and other vehicle driving behavior of the simulated driving behavior. Predict the speed, acceleration, and steering of the intelligent vehicle based on the environmental road conditions, traffic flow, and the driving behavior of other vehicles. Based on the speed, acceleration, and steering of the intelligent vehicle and the driving behavior of other vehicles, the simulation simulates the passing, overtaking, obstacle avoidance, and / or lane changing behaviors between intelligent vehicles.

4. The method for simulating human-vehicle driving behavior according to any one of claims 1-3, characterized in that, After simulating human and vehicle driving behavior based on the simulated driving behavior data, the simulated driving behavior data corresponding to each human and vehicle driving behavior is used as a new input to the preset generator to generate new simulated driving behavior data.

5. The method for simulating human-vehicle driving behavior according to any one of claims 1-3, characterized in that, After simulating human and vehicle driving behavior based on the simulated driving behavior data, the process also includes: Based on a preset evaluation index system, the simulated driving behavior data for human-vehicle driving behavior simulation is evaluated item by item and evaluation results are generated. The evaluation results are used to assist in adjusting the preset generator, preset discriminator and their corresponding parameters and / or structures.

6. A vehicle driving behavior simulation device based on the vehicle driving behavior simulation method of claim 1, characterized in that, include: The acquisition unit is used to acquire scene information, which includes at least driving data, environmental data, and interaction data. The training unit is at least used to generate simulated driving behavior data based on the scenario information via a preset generator; The simulation unit is used at least to simulate human and vehicle driving behavior based on the simulated driving behavior data, wherein the human and vehicle driving behavior includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between the intelligent vehicles.

7. A human-vehicle driving behavior simulation system, characterized in that, include: The data acquisition module is used to acquire scene information, which includes at least driving data, environmental data, and interaction data. The GAN module is at least used to pre-configure the corresponding generator and discriminator to obtain the preset generator and preset discriminator, generate simulated driving behavior data based on the scene information through the preset generator, and input the simulated driving behavior data into the preset discriminator to generate a discrimination result, so as to iteratively train the preset generator and the preset discriminator; The simulation module is at least used to simulate human-vehicle driving behavior based on the simulated driving behavior data, wherein the human-vehicle driving behavior includes at least: the driver's driving interaction behavior, the driving interaction behavior between the driver and the intelligent vehicle, and the driving interaction behavior between intelligent vehicles; wherein the simulation module includes at least a first simulation submodule, the first simulation submodule being at least used for: Acquire scene information including steering wheel angle, accelerator pedal opening, and brake pedal opening in real driving behavior; Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in real driving behavior, the preset generator generates simulated driving behavior data that meets the discrimination threshold of the preset discriminator. The simulated driving behavior data includes the steering wheel angle, accelerator pedal opening and brake pedal opening of the simulated driving behavior. Based on the steering wheel angle, accelerator pedal opening and brake pedal opening in the simulated driving behavior data, the driving behavior of a human driver is simulated when meeting oncoming traffic, overtaking, avoiding obstacles and / or changing lanes.

8. The human-vehicle driving behavior simulation system according to claim 7, characterized in that, The simulation module includes at least a second simulation submodule, which is at least used for: Acquire scene information including real-world driving behavior vehicle sensor data, environmental road rules, and environmental road conditions; Based on vehicle sensor data, environmental road rules, and environmental road conditions from real driving behavior, the preset generator generates simulated driving behavior data that meets the preset discrimination threshold. The simulated driving behavior data includes vehicle sensor data, environmental road rules, and environmental road conditions from simulated driving behavior. Based on the vehicle sensor data, the environmental road rules, and the environmental road conditions, the steering and shifting behavior of the intelligent vehicle is simulated and predicted. Based on the simulated and predicted steering and shifting behavior of intelligent vehicles, the system simulates the reaction behavior of human drivers to intelligent vehicles, the driving behavior of human drivers, and / or the behavior of intelligent vehicles after detecting changes in the motion of other vehicles.

9. The human-vehicle driving behavior simulation system according to claim 7, characterized in that, The simulation module includes at least a third simulation submodule, which is used for at least: Acquire real-world driving behavior data, including road conditions in the actual driving environment, traffic flow, and the driving behavior of other vehicles. Based on the environmental road conditions, traffic flow and other vehicle driving behavior in real driving behavior, the preset generator generates simulated driving behavior data that meets the preset discrimination threshold of the discriminator. The simulated driving behavior data includes the environmental road conditions, traffic flow and other vehicle driving behavior of the simulated driving behavior. Predict the speed, acceleration, and steering of the intelligent vehicle based on the environmental road conditions, traffic flow, and the driving behavior of other vehicles. Based on the speed, acceleration, and steering of the intelligent vehicle and the driving behavior of other vehicles, the simulation simulates the passing, overtaking, obstacle avoidance, and / or lane changing behaviors between intelligent vehicles.

10. An electronic device, characterized in that, include: At least one processor; And at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor invokes the program instructions to perform the steps of the human-vehicle driving behavior simulation method as described in any one of claims 1-5.

11. A readable storage medium storing a computer program, characterized in that, The computer program is executed by a processor using the steps of the human-vehicle driving behavior simulation method as described in any one of claims 1-5.

Images