An automatic driving environment complexity evaluation method and equipment

Abstract

This invention provides a method and equipment for evaluating the environmental complexity of autonomous driving. The method senses the driving environment elements of an autonomous vehicle, analyzes the composition and interaction mechanisms of static and dynamic driving environment elements, and determines the influencing factors of traffic environment complexity. Based on the complexity influencing factors and the spatiotemporal characteristics of driving environment elements, it constructs equivalent mass, driving strategy contribution, and static traffic driving environment complexity coefficients, and introduces them into a universal gravitation model to establish a spatiotemporal quantified gravity model to calculate the instantaneous longitudinal traffic environment complexity. The instantaneous complexity corresponding to multiple dynamic elements is weighted and summed to obtain the overall instantaneous complexity, and then the cumulative longitudinal traffic environment complexity is obtained through time integration, thus completing the environmental complexity evaluation. This invention can comprehensively and accurately quantify traffic environment complexity, effectively improving the environmental adaptability and driving safety of autonomous vehicles in complex scenarios.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving environment complexity evaluation technology, specifically an autonomous driving environment complexity evaluation method and equipment. Background Technology

[0002] Currently, scholars both domestically and internationally have conducted extensive research on the complexity assessment of road traffic environments. Some studies extract traffic environment factors through simulation software or low-cost cameras to classify and evaluate road environments. Other studies divide environmental complexity into multiple aspects such as road environment, lane obstacles, and road surface conditions, constructing assessment index systems for specific scenarios. In urban road scenarios, some methods attempt to divide the traffic environment into static and dynamic environments and establish complexity calculation models based on information entropy. However, most existing methods focus on the analysis of single dimensions or local elements, failing to systematically integrate the composition and interaction mechanisms of static and dynamic driving environment elements, resulting in an incomplete characterization of traffic environment complexity. Furthermore, although some studies introduce parameters such as relative speed and relative distance, borrowing the basic form of gravity models to express complexity, they lack clear calculation methods for determining key parameters, making it difficult to achieve dynamic, real-time, and quantifiable complexity assessment.

[0003] Meanwhile, existing evaluation methods often focus on driver or vehicle status, neglecting to assess the adaptability of autonomous vehicles in complex traffic environments. This results in evaluation results that fail to effectively correspond to the complexity of traffic scenarios. Particularly in autonomous driving systems, how to extract and integrate multi-dimensional information from people, vehicles, roads, and the environment in real time to reflect the impact of dynamic changes in the traffic environment on vehicle perception and decision-making remains a weak point in current research. The lack of a systematic understanding of the interaction between the static environmental base and dynamic participants makes it difficult for existing methods to accurately reflect the true complexity of traffic environments, limiting the improvement of the environmental adaptability and driving safety of autonomous vehicles in complex scenarios. Summary of the Invention

[0004] To address the technical problems mentioned in the background section, this invention proposes a method and equipment for evaluating the environmental complexity of autonomous driving.

[0005] Therefore, the technical solution adopted by the present invention is as follows: A method for evaluating the environmental complexity of autonomous driving includes: Step 1: Perceive the driving environment elements of the autonomous vehicle, analyze the composition and interaction mechanism of the static and dynamic driving environment elements, and determine the factors affecting the complexity of the traffic environment. Step 2: Construct complexity parameters based on the complexity influencing factors and the spatiotemporal characteristics of driving environment elements. These complexity parameters include equivalent mass, driving strategy contribution, and static traffic driving environment complexity coefficient. Replace the physical mass in the universal gravitation model with the equivalent mass, and introduce the driving strategy contribution and static traffic driving environment complexity coefficient into the universal gravitation model to obtain a spatiotemporal quantized gravity model. The calculation result of the spatiotemporal quantized gravity model is the instantaneous longitudinal traffic environment complexity of the autonomous vehicle. When multiple dynamic driving environment elements exist in the same traffic scenario, perform a weighted summation of the instantaneous longitudinal traffic environment complexity corresponding to each dynamic driving environment element to obtain the overall instantaneous longitudinal traffic environment complexity of the autonomous vehicle. Step 3: Perform time integration on the overall instantaneous longitudinal traffic environment complexity to obtain the cumulative longitudinal traffic environment complexity, thereby completing the evaluation of the autonomous driving environment complexity.

[0006] Furthermore, the traffic environment complexity influencing factors include static traffic environment complexity influencing factors and dynamic traffic environment complexity influencing factors; The static traffic environment complexity influencing factor is a static indicator that reflects the inherent characteristics of the static environment and affects the complexity of the traffic environment. The static indicator includes positive indicators and negative indicators. The dynamic traffic environment complexity influencing factor is a dynamic indicator that reflects the motion characteristics of dynamic driving environment elements and their interaction with autonomous vehicles.

[0007] Furthermore, the formula for calculating the equivalent mass is as follows: in, and Self-driving cars and dynamic driving environment elements Equivalent quality, and Self-driving cars and dynamic driving environment elements Category and Self-driving cars and dynamic driving environment elements driving speed, The first undetermined coefficient, For autonomous vehicles Direction of motion and autonomous vehicles and dynamic driving environment elements The angle between the relative distances, For dynamic driving environment elements Direction of motion and autonomous vehicles and dynamic driving environment elements The angle between the relative distances.

[0008] Furthermore, the formula for calculating the contribution of the driving strategy is as follows: in, For autonomous vehicles For dynamic driving environment elements Contribution of driving strategy and These are the second and third undetermined coefficients, respectively. It is a natural exponential function.

[0009] Furthermore, the calculation process for the static traffic driving environment complexity coefficient is as follows: Calculate the first Attribute values ​​of a static feature With optimal attribute value correlation : in, Attribute values ​​for all static features With optimal attribute value The minimum absolute difference, Attribute values ​​for all static features With optimal attribute value The maximum absolute difference For resolution coefficients, The number of static features; The complexity coefficient of the static traffic driving environment is calculated based on the correlation degree, and the calculation formula is as follows: in, This represents the complexity coefficient of the static traffic driving environment.

[0010] Furthermore, the spatiotemporal quantization gravity model is as follows: in, For the instantaneous longitudinal traffic environment complexity, For autonomous vehicles and dynamic driving environment elements The relative distance, This is the fourth undetermined coefficient.

[0011] Furthermore, the instantaneous longitudinal traffic environment complexity The calculation formula is as follows: in, , where q is the weight coefficient of the dynamic driving environment element in the driving environment of the autonomous vehicle; The cumulative longitudinal traffic environment complexity The calculation formula is as follows: Where ∫dt represents the integration operation over time. It is a time variable.

[0012] An autonomous driving environment complexity evaluation device, comprising: The sensing module is used to sense the driving environment elements; The sending module is used to send the driving environment elements to the receiving module; A receiving module is used to receive data sent by the sending module; The storage module is used to store the first undetermined coefficient, the second undetermined coefficient, the third undetermined coefficient, the fourth undetermined coefficient, the weighting coefficient, and the driving environment elements; The calculation module is used to construct the equivalent mass, driving strategy contribution, and static traffic environment complexity coefficients, and to calculate the instantaneous longitudinal traffic environment complexity, the overall instantaneous longitudinal traffic environment complexity, and the cumulative longitudinal traffic environment complexity.

[0013] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the autonomous driving environment complexity evaluation method.

[0014] Compared with the prior art, the advantages of the present invention are as follows: 1. This invention constructs a spatiotemporal quantification model of longitudinal traffic environment complexity that covers all elements of the driving environment through an improved universal gravitation model, thereby achieving a unified representation of the dynamic and spatiotemporal dimensions of traffic environment complexity.

[0015] 2. This invention integrates all elements of driving environment information, including people, vehicles, roads, and the environment, and introduces key parameters such as category, distance, and speed. It comprehensively considers the combined effects of static and dynamic traffic environments, clarifies the factors influencing the complexity of the traffic environment, and forms a more comprehensive method for measuring complexity.

[0016] 3. By adopting an improved universal gravitation model, this invention simplifies the calculation process, enhances the model's ease of operation and adaptability to various scenarios, and is effectively applicable to a variety of real-world traffic scenarios, thus possessing greater practicality and operability. Attached Figure Description

[0017] 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.

[0018] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the basic framework for calculating the environmental complexity of the spatiotemporal quantization gravity model of the present invention; Figure 3 This is a schematic diagram illustrating the complex longitudinal driving environment for the spatiotemporal quantization gravity model of the present invention. Figure 4 This is a schematic diagram of the device structure for calculating the environmental complexity of the spatiotemporal quantization gravity model of the present invention. Detailed Implementation

[0019] To achieve the above objectives, the present invention provides a method for evaluating the complexity of autonomous driving environments. Please refer to the following technical solution. Figures 1 to 4 ,include: Step 1: Perceive the driving environment elements of the autonomous vehicle, analyze the composition and interaction mechanism of the static and dynamic driving environment elements, and determine the factors affecting the complexity of the traffic environment. Driving environment elements include static driving environment elements and dynamic driving environment elements; Static driving environment elements provide the basic environmental framework for autonomous vehicle operation and are fundamental elements that determine the basic complexity of the traffic environment. Their attributes do not change significantly in a short period of time. These elements include road type, number of lanes, road marking clarity, road marking type, road surface type, types of fixed road obstacles, distribution of road obstacles, degree of road damage, road coverage, density of road obstacle distribution, and the degree of light and shadow and image blur. Road types include urban roads, highways, and rural roads, etc. Road marking types include solid lines, dashed lines, and stop lines, etc. Road surface types include asphalt and cement, etc. Types of fixed road obstacles include construction barriers, roadblocks, and landslide debris, etc. The methods for obtaining these elements are as follows: Basic attribute data such as road type and number of lanes are retrieved directly from the road basic information database; The status attribute values ​​of road marking clarity, road obstacle distribution density, road damage level, road coverage, and light and shadow blurring are determined by expert calibration or scoring. Among them, the spatial distribution location of road obstacles can be supplemented by the relative distance data between them and autonomous vehicles directly detected by millimeter-wave vehicle radar, thus improving the spatial feature information of static elements.

[0020] Dynamic driving environment elements are road participation elements with motion characteristics in the traffic environment. They are core elements that determine the dynamic changes in the complexity of the traffic environment. Their attributes change in real time with time and space, including the category and quantity of dynamic driving environment elements, the driving speed and direction of movement of each participant such as motor vehicles, non-motor vehicles, and pedestrians, the relative distance, relative speed, and angle of movement between each participant and the autonomous vehicle, and the changes in the motion state of each participant (lane changing, cutting in, turning back, etc.). The methods of obtaining these elements are as follows: The number of dynamic driving environment elements, their relative distance to the vehicle, relative speed, and the angle of motion direction are directly detected and obtained by millimeter-wave vehicle radar. At the same time, the radar can capture the real-time motion trajectory of the dynamic elements, indirectly deduce their motion state changes based on trajectory changes, and combine the angle of motion direction to help deduce the motion direction of the dynamic elements. The speed of the autonomous vehicle itself and the surrounding dynamic elements is directly collected by speed sensors; The categories of dynamic driving environment elements are determined by the Delphi method to complete attribute determination and specific category value determination, which can distinguish motor vehicles, non-motor vehicles and pedestrians, as well as specific vehicle models and specific types of non-motor vehicles (electric vehicles or bicycles). Millimeter-wave vehicle radar works synchronously with speed sensors, and can indirectly deduce the relative acceleration between dynamic elements and autonomous vehicles, as well as the spatiotemporal characteristics of dynamic elements such as speed and position changes at different time points, based on the parameters directly acquired by each. After sensing the driving environment elements, the driving environment elements are uniformly transmitted to the preset data receiving module through the preset data sending module; The data receiving module classifies and stores all driving environment elements, providing a traceable and complete data set for subsequent analysis of the composition of static and dynamic driving environment elements.

[0021] The interaction mechanism between static and dynamic driving environment elements is as follows: Static driving environment elements directly determine the range, patterns, and behavioral choices of dynamic driving environment elements. For example, a static environment of a single lane or narrow road will limit the speed of motor vehicles and lane-changing behavior; a static intersection with a high density of obstacles will increase the complexity of the trajectory of dynamic elements; while a multi-lane highway with clear markings will reduce the complexity of the motion constraints of dynamic elements. In the same static driving environment, the real-time motion characteristics of dynamic driving environment elements can change the perceived complexity of the overall traffic environment. For example, in a static urban road environment of moderate complexity, if millimeter-wave vehicle radar detects disordered movements such as high-speed lane changes and oncoming traffic, and the speed sensor collects that the speed is close to the road speed limit, the perceived complexity of the overall environment will be significantly improved.

[0022] The complexity of static driving environment elements (obtained by a comprehensive assessment of static driving environment elements calibrated or scored by experts) is positively correlated with the amplification effect of the movement and changes of dynamic driving environment elements. For example, in a highly complex static environment at an urban intersection, the disordered movement of a small number of dynamic driving environment elements can significantly increase the overall traffic environment complexity; while in a low-complexity static environment on a straight highway, the regular movement of dynamic driving environment elements has a relatively small impact on the overall traffic environment complexity.

[0023] The factors influencing traffic environment complexity include static traffic environment complexity factors and dynamic traffic environment complexity factors; The static traffic environment complexity influencing factor is a static indicator that can reflect the inherent characteristics of the static environment and has a significant impact on the complexity of the traffic environment. It includes two types of indicators: positive indicators and negative indicators. Positive indicators: Indicators with higher attribute values ​​indicate higher traffic environment complexity, including road obstacle distribution density, light and shadow blurring degree, and image blurring degree; Negative indicators: Indicators with higher attribute values ​​indicate lower traffic environment complexity, including road type, number of lanes, clarity of road markings, and degree of road damage; Both positive and negative indicators were dimensionless, and the attribute values ​​were normalized to... The interval, the formula is as follows: No. The dimensionless formula for the positive indicator is as follows: in, The dimensionless processing of the first The attribute values ​​of a positive indicator, For the first The attribute values ​​of a positive indicator, For the first The minimum attribute value of a positive indicator is obtained by statistically analyzing the full sample of the original data for that indicator. For the first The maximum attribute value of a positive indicator is obtained by statistical analysis of the full sample of the original data for that indicator. No. The dimensionless formula for a negative indicator: in, The dimensionless processing of the first The attribute values ​​of a negative indicator. For the first The attribute values ​​of a negative indicator. For the first The minimum attribute value of a negative indicator is obtained by statistical analysis of the entire sample of the original data for that indicator. For the first The maximum attribute value of a negative indicator is obtained by statistical analysis of the full sample of the original data for that indicator. The dynamic traffic environment complexity influencing factor is a dynamic indicator that reflects the motion characteristics of dynamic driving environment elements and their interaction with autonomous vehicles. It includes the type and quantity of dynamic driving environment elements, the relative distance, relative speed, and angle of motion between dynamic elements and autonomous vehicles, as well as the driving strategy response of autonomous vehicles to changes in dynamic elements. The driving strategy response is indirectly derived based on the real-time parameters of dynamic elements detected by millimeter-wave vehicle radar and speed sensors, combined with the strategy logic of the autonomous driving system.

[0024] Step 2: Construct complexity parameters based on the complexity influencing factors and the spatiotemporal characteristics of the driving environment elements. These complexity parameters include equivalent mass, driving strategy contribution, and static traffic driving environment complexity coefficient. Calculate the instantaneous longitudinal traffic environment complexity of the autonomous vehicle based on a preset spatiotemporal quantized gravity model and the complexity parameters. When multiple dynamic driving environment elements exist in the same traffic scenario, perform a weighted summation of the instantaneous longitudinal traffic environment complexity corresponding to each dynamic driving environment element to obtain the overall instantaneous longitudinal traffic environment complexity of the autonomous vehicle. The spatiotemporal quantized gravity model is obtained by replacing the mass in the universal gravitation model with the equivalent mass and incorporating the driving strategy contribution and static traffic driving environment complexity coefficient into the universal gravitation model. Equivalent mass refers to the type and speed of the autonomous vehicle and the dynamic driving environment elements, as shown in the following formula: in, and Self-driving cars and dynamic driving environment elements Equivalent quality, and Self-driving cars and dynamic driving environment elements The category and attribute values ​​are determined using the Delphi method. and Self-driving cars and dynamic driving environment elements The driving speed of autonomous vehicles is regulated. The direction of motion is the positive direction. The first undetermined coefficient, For autonomous vehicles Direction of motion and autonomous vehicles and dynamic driving environment elements The angle between the relative distances, For dynamic driving environment elements Direction of motion and autonomous vehicles and dynamic driving environment elements The angle between the relative distances is positive in the clockwise direction; The first set of undetermined coefficients is based on fitting and calibration of real-vehicle test and simulation data from mainstream driving scenarios for autonomous vehicles. It also matches the detection accuracy of millimeter-wave vehicle radar and speed sensors, controlling the weight of the speed component on the equivalent mass to avoid distortion in the equivalent mass calculation due to excessive amplification or reduction of the speed term. Furthermore, the coefficient values ​​must ensure that the equivalent mass is positive, the model calculation is stable, and it adapts to the complex quantification needs of typical traffic scenarios such as urban roads and highways. The value range is 0.0005. Up to 0.0020 ,For example: The first undetermined coefficient for the mainstream urban road scenario is set at 0.001. ; The first undetermined coefficient for the highway scenario is taken as: ; The first undetermined coefficient for the rural road scene is taken as: ; If the millimeter-wave vehicle radar and speed sensor detect that a dynamic driving environment element is stationary, the equivalent mass of that element is directly taken as the attribute value corresponding to its category.

[0025] The formula for calculating the contribution of driving strategy is as follows: in, For autonomous vehicles For dynamic driving environment elements Contribution of driving strategy and These are the second and third undetermined coefficients, respectively. It is a natural exponential function; The second undetermined coefficient is calibrated based on the relative speed change characteristics of autonomous vehicles and dynamic driving environment elements under different scenarios. Its function is to adjust the nonlinear influence of the relative speed term on the contribution of driving strategy, adapt to the dynamic interaction rules of each scenario, and ensure that the contribution changes with relative speed and motion angle in line with the real traffic scenario. The value range is 0.2 to 0.8. For example, the second undetermined coefficient is 0.5 for urban roads, 0.6 for highways, and 0.4 for rural roads. The third undetermined coefficient serves to calibrate the baseline amplitude of the driving strategy's contribution. By fitting the baseline data of the driving strategy response across all scenarios, the output range of the exp function is matched to ensure accuracy under different motion states. The overall value is within a reasonable dimensionless range, providing a stable benchmark for subsequent speed term adjustments. The value range is 0.5 to 1.1. For example, the third undetermined coefficient for urban roads is 0.8, for highways it is 0.9, and for rural roads it is 0.7.

[0026] The complexity coefficient of the static traffic driving environment is determined by the grey relational analysis method, and the calculation process is as follows: Calculate the first Attribute values ​​of a static feature With optimal attribute value correlation : in, Attribute values ​​for all static features With optimal attribute value The minimum absolute difference, Attribute values ​​for all static features With optimal attribute value The maximum absolute difference is the resolution coefficient, a standard parameter for grey relational analysis, with a value range of [0,1], typically taken as 0.5. The number of static features; The complexity coefficient of the static traffic driving environment is calculated based on the correlation degree, and the calculation formula is as follows: in, This represents the complexity coefficient of the static traffic driving environment.

[0027] The spacetime quantization gravity model is as follows: in, For the instantaneous longitudinal traffic environment complexity, For autonomous vehicles and dynamic driving environment elements The relative distance, The fourth undetermined coefficient; The calibration of the fourth undetermined coefficient is designed with reference to the physical meaning of the gravitational constant and is based on real vehicle test data and traffic simulation fitting results for autonomous driving in all scenarios. At the same time, it adapts to the requirement of unified formula dimensions to ensure that the instantaneous longitudinal traffic environment complexity calculation result is a reasonable dimensionless value without excessive amplification or reduction. It is also fine-tuned according to the interaction characteristics of static and dynamic elements in different road scenarios and matched with the detection accuracy of millimeter-wave vehicle radar and speed sensors. Finally, the complexity value is made to fit the complexity gradient of real traffic scenarios, with a value range of 0.7 to 1.3. For example, the value of the fourth undetermined coefficient for urban roads is 1.0, the value of the fourth undetermined coefficient for highways is 0.9, and the value of the fourth undetermined coefficient for rural roads is 1.1.

[0028] Instantaneous longitudinal traffic environment complexity The calculation formula is as follows: in, Let q be the weight coefficient of the dynamic driving environment element q in the autonomous vehicle driving environment. For example, the weight of a pedestrian in urban roads is set to 0.4, and the calibration rule is as follows: The values ​​are assigned based on a comprehensive assessment of the element type threat level (e.g., pedestrian > non-motorized vehicle > motorized vehicle), relative distance (closer, higher weight), motion interaction state (moving towards each other / crossing each other > moving in the same direction), and collision time (shorter TTC, higher weight). All dynamic elements The sum must be 1; The weight distribution is adjusted according to the characteristics of the road scene. For example, urban roads focus on pedestrians, while highways focus on motor vehicles. Through real-vehicle testing and simulation fitting, we ensure that the weight allocation matches the subjective perception of human drivers regarding the complexity of the environment.

[0029] Step 3: Perform time integration on the overall instantaneous longitudinal traffic environment complexity to obtain the cumulative longitudinal traffic environment complexity, thereby completing the evaluation of the autonomous driving environment complexity.

[0030] Cumulative longitudinal traffic environment complexity The calculation formula is as follows: Where ∫dt represents the integration operation over time. It is a time variable, specifically the real-time travel time.

[0031] An autonomous driving environment complexity evaluation device, comprising: The sensing module is used to sense the driving environment elements; The sending module is used to send the driving environment elements to the receiving module; A receiving module is used to receive data sent by the sending module; The storage module is used to store the first undetermined coefficient, the second undetermined coefficient, the third undetermined coefficient, the fourth undetermined coefficient, the weighting coefficient, and the driving environment elements; The calculation module is used to construct the equivalent mass, driving strategy contribution, and static traffic environment complexity coefficients, and to calculate the instantaneous longitudinal traffic environment complexity, the overall instantaneous longitudinal traffic environment complexity, and the cumulative longitudinal traffic environment complexity.

[0032] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the autonomous driving environment complexity evaluation method.

[0033] This invention proposes a method and equipment for evaluating the complexity of autonomous driving environments. By constructing an improved gravitational model, it incorporates static and dynamic driving environment elements and their interactions into a unified framework, and introduces key parameters such as equivalent mass and driving strategy contribution. This solves the problem that existing technologies lack comprehensive and dynamic quantification capabilities for traffic environment complexity, thereby improving the environmental adaptability and driving safety of autonomous vehicles in complex scenarios.

[0034] In summary, this invention improves the universal gravitation model by deeply integrating static and dynamic driving environment elements and introducing key parameters such as equivalent mass and driving strategy contribution. It constructs a spatiotemporal quantitative model that can comprehensively characterize the complexity of the traffic environment, achieving continuous evaluation from instantaneous to cumulative, and improving the environmental adaptability and driving safety of the autonomous driving system in complex scenarios.

[0035] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for evaluating the environmental complexity of autonomous driving, characterized in that, include: Step 1: Perceive the driving environment elements of the autonomous vehicle, analyze the composition and interaction mechanism of the static and dynamic driving environment elements, and determine the factors affecting the complexity of the traffic environment. Step 2: Construct complexity parameters based on the complexity influencing factors and the spatiotemporal characteristics of driving environment elements. These complexity parameters include equivalent mass, driving strategy contribution, and static traffic driving environment complexity coefficient. Replace the physical mass in the universal gravitation model with the equivalent mass, and introduce the driving strategy contribution and static traffic driving environment complexity coefficient into the universal gravitation model to obtain a spatiotemporal quantized gravity model. The calculation result of the spatiotemporal quantized gravity model is the instantaneous longitudinal traffic environment complexity of the autonomous vehicle. When multiple dynamic driving environment elements exist in the same traffic scenario, perform a weighted summation of the instantaneous longitudinal traffic environment complexity corresponding to each dynamic driving environment element to obtain the overall instantaneous longitudinal traffic environment complexity of the autonomous vehicle. Step 3: Perform time integration on the overall instantaneous longitudinal traffic environment complexity to obtain the cumulative longitudinal traffic environment complexity, thereby completing the evaluation of the autonomous driving environment complexity.

2. The method according to claim 1, characterized in that, The factors influencing the complexity of the traffic environment include static traffic environment complexity factors and dynamic traffic environment complexity factors. The static traffic environment complexity influencing factor is a static indicator that reflects the inherent characteristics of the static environment and affects the complexity of the traffic environment. The static indicator includes positive indicators and negative indicators. The dynamic traffic environment complexity influencing factor is a dynamic indicator that reflects the motion characteristics of dynamic driving environment elements and their interaction with autonomous vehicles.

3. The method according to claim 2, characterized in that, The formula for calculating the equivalent mass is as follows: in, and Self-driving cars and dynamic driving environment elements Equivalent quality, and Self-driving cars and dynamic driving environment elements Category and Self-driving cars and dynamic driving environment elements driving speed, The first undetermined coefficient, For autonomous vehicles Direction of motion and autonomous vehicles and dynamic driving environment elements The angle between the relative distances, For dynamic driving environment elements Direction of motion and autonomous vehicles and dynamic driving environment elements The angle between the relative distances.

4. The method according to claim 3, characterized in that, The formula for calculating the contribution of the driving strategy is as follows: in, For autonomous vehicles For dynamic driving environment elements Contribution of driving strategy and These are the second and third undetermined coefficients, respectively. It is a natural exponential function.

5. The method according to claim 4, characterized in that, The calculation process for the static traffic driving environment complexity coefficient is as follows: Calculate the first Attribute values ​​of a static feature With optimal attribute value correlation : in, Attribute values ​​for all static features With optimal attribute value The minimum absolute difference, Attribute values ​​for all static features With optimal attribute value The maximum absolute difference For resolution coefficients, The number of static features; The complexity coefficient of the static traffic driving environment is calculated based on the correlation degree, and the calculation formula is as follows: in, This represents the complexity coefficient of the static traffic driving environment.

6. The method according to claim 5, characterized in that, The spatiotemporal quantization gravity model is as follows: in, For the instantaneous longitudinal traffic environment complexity, For autonomous vehicles and dynamic driving environment elements The relative distance, This is the fourth undetermined coefficient.

7. The method according to claim 6, characterized in that, The instantaneous longitudinal traffic environment complexity The calculation formula is as follows: in, , where q is the weight coefficient of the dynamic driving environment element in the driving environment of the autonomous vehicle; The cumulative longitudinal traffic environment complexity The calculation formula is as follows: Where ∫dt represents the integration operation over time. It is a time variable.

8. An autonomous driving environment complexity evaluation device, used to execute the method described in any one of claims 1-7, characterized in that, include: The sensing module is used to sense the driving environment elements; The sending module is used to send the driving environment elements to the receiving module; A receiving module is used to receive data sent by the sending module; The storage module is used to store the first undetermined coefficient, the second undetermined coefficient, the third undetermined coefficient, the fourth undetermined coefficient, the weighting coefficient, and the driving environment elements; The calculation module is used to construct the equivalent mass, driving strategy contribution, and static traffic environment complexity coefficients, and to calculate the instantaneous longitudinal traffic environment complexity, the overall instantaneous longitudinal traffic environment complexity, and the cumulative longitudinal traffic environment complexity.

9. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the autonomous driving environment complexity evaluation method according to any one of claims 1 to 7.

Images