Method and device for comprehensive evaluation of confidence of shaft-coupled whole vehicle-in-the-loop simulation system

Abstract

The application provides a kind of axle coupling type whole vehicle in loop simulation system confidence comprehensive evaluation method and device, the present application is around axle coupling type whole vehicle in loop simulation system, it is proposed to characterize the certainty numerical evaluation index, probability evaluation index and semantic description evaluation index of its confidence level, realize the detailed description of the multi-dimensional characteristics of simulation system, the weight of the above three types of indexes and interval representation is modeled and mapped conversion using bivalent fuzzy set, the confidence evaluation grade corresponding to the confidence level is obtained using EKM algorithm, to comprehensively evaluate the confidence level of axle coupling type whole vehicle in loop simulation system.The method proposed in the application effectively solves the core limitation of single evaluation dimension in the prior art, and significantly reduces the uncertainty disturbance in the evaluation process, realizes the scientific and comprehensive evaluation of the confidence level of axle coupling type whole vehicle in loop simulation system.

Description

Technical Field

[0001] This invention relates to the field of intelligent connected vehicle testing technology, and in particular to a method and apparatus for comprehensive evaluation of the confidence level of an axis-coupled whole-vehicle-in-the-loop simulation system. Background Technology

[0002] As the automotive industry deeply integrates electrification, intelligence, and connectivity, the functionality and system complexity of intelligent connected vehicles are increasing exponentially. To ensure their safety, reliability, compliance, and user experience in real-world road operations, scientific, systematic, and comprehensive testing and evaluation of intelligent connected vehicles has become crucial. Developing efficient, reliable, and comprehensive testing technologies has become a key link in promoting the maturity of intelligent connected vehicle technology and its industrial implementation.

[0003] Current testing and evaluation methods for intelligent connected vehicles mainly fall into three categories: closed-track testing, open road testing, and simulation testing. Traditional closed-track or open road testing methods face inherent bottlenecks such as long testing cycles, high costs, difficulty in reproducing extreme and dangerous scenarios, and limited test coverage. Simulation testing technology, with its outstanding advantages of high efficiency, high safety, reproducibility, and low cost, has become a core means to solve the bottlenecks of real-vehicle testing. By constructing a high-precision virtual simulation environment, it is possible to simulate a vast number of conventional, edge, and even extremely dangerous traffic scenarios in the laboratory, enabling rapid verification and iterative optimization of vehicle perception, decision-making, and control system functions, greatly accelerating the R&D process and improving the breadth and depth of testing.

[0004] Specifically, simulation testing can be categorized into vehicle-in-the-loop (VIL) simulation, hardware-in-the-loop (HIL) simulation, software-in-the-loop (Software-in-the-loop) simulation, and driver-in-the-loop (HIL) simulation, based on factors such as the level of the test object and the degree of hardware integration. VIL simulation systems can be further divided into site-based VIL simulation systems and axle-coupled VIL simulation systems, depending on the load platform structure. Axle-coupled VIL simulation systems are typically conducted on real roads or dedicated test tracks, applying loads and excitations to the vehicle using actual road surface and environmental conditions. Their system structure is relatively open and dependent on the real physical environment. In contrast, axle-coupled VIL simulation systems, as an advanced testing method, place the actual vehicle's drive / brake wheels directly on a dynamometer that dynamically simulates road loads and operating conditions. This allows the vehicle's powertrain, braking system, and related control systems to operate on their actual mechanical interfaces, introducing a higher degree of real vehicle dynamics characteristics into the entire perception, decision-making, and control chain. Compared to pure simulation or VIL simulation that only includes partial hardware, this method more realistically reflects the vehicle's comprehensive response under dynamically changing road and driving conditions.

[0005] However, the effectiveness of simulation testing is highly dependent on the accuracy and confidence level of its model. If the simulation environment deviates significantly from the real world, the test results will be difficult to use directly to guide design and verification, and may even introduce the risk of misjudgment. Therefore, constructing a scientific and accurate confidence evaluation method is a prerequisite for the practical application and standardization of shaft-coupled vehicle-in-the-loop simulation testing technology, and is also the key to ensuring its value in intelligent connected vehicle testing.

[0006] Currently, there is a lack of systematic methods and standards for the quantitative evaluation of the performance of shaft-coupled vehicle-in-the-loop simulation systems and the reliability of their test results. On the one hand, some evaluation methods rely on expert experience and subjective judgment, which is affected by uncertainties such as human cognitive bias and experience differences, resulting in poor repeatability of evaluation results. On the other hand, most quantitative evaluation methods are based on analysis of only a single type of data, making it difficult to comprehensively evaluate the confidence level of the vehicle-in-the-loop simulation system, resulting in shortcomings such as one-sided evaluation dimensions and inaccurate evaluation results.

[0007] Overall, for the confidence evaluation requirements of shaft-coupled vehicle-in-the-loop simulation systems, how to construct a comprehensive confidence evaluation method based on multimodal data fusion to overcome the technical shortcomings of existing confidence evaluation methods, such as single evaluation dimensions and significant uncertainty interference, and ultimately achieve a dual improvement in the accuracy and comprehensiveness of confidence evaluation, has become a key technical problem that urgently needs to be solved in this field. Summary of the Invention

[0008] In view of the above problems, the present invention proposes a method and apparatus for comprehensive evaluation of the confidence of a shaft-coupled vehicle-in-the-loop simulation system to overcome or at least partially solve the above problems.

[0009] According to one aspect of the present invention, a comprehensive confidence evaluation method for a shaft-coupled vehicle-in-the-loop simulation system is provided, comprising:

[0010] Based on the virtual environment generation module, sensor data injection module, data interaction module, and power load module in the simulation system, multimodal evaluation index data is determined to characterize the confidence features of the shaft-coupled vehicle-in-the-loop simulation system. The multimodal evaluation index data includes deterministic numerical evaluation indexes, probabilistic evaluation indexes, and semantic description evaluation indexes. Deterministic numerical evaluation indexes refer to evaluation indexes that are obtained based on direct measurement or quantitative calculation, with clear and reproducible numerical results. Probabilistic evaluation indexes refer to evaluation indexes that are based on statistical analysis methods and describe the performance fluctuation law of the simulation system and the probability of failure or deviation. Semantic description evaluation indexes refer to evaluation indexes constructed through subjective evaluation methods such as semantic description or hierarchical assignment for subjective perceptions or complex characteristics in the simulation system that are difficult to quantify.

[0011] Set the weight data corresponding to the multimodal evaluation index data, and convert the multimodal evaluation index data and weight data into a type-two fuzzy set representation;

[0012] The confidence level of the axle-coupled vehicle-in-the-loop simulation system is generated based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data. The confidence level corresponding to the confidence level is obtained by using the EKM algorithm, so as to comprehensively evaluate the confidence level of the axle-coupled vehicle-in-the-loop simulation system.

[0013] Optionally, the deterministic numerical evaluation index refers to an evaluation index that is obtained based on direct measurement or quantitative calculation, and whose numerical results are clear and reproducible.

[0014] The deterministic numerical evaluation index includes one or more of the following: steady-state torque control error, anti-drag resistance simulation error, vehicle longitudinal attitude simulation error, vehicle lateral attitude simulation error, wheel speed synchronization error, acceleration consistency error, virtual camera target detection position error, virtual lidar point cloud ranging error, virtual millimeter-wave radar speed measurement error, and scene slope resistance simulation error.

[0015] Optionally, the probability evaluation index refers to an evaluation index based on statistical analysis methods that describes the performance fluctuation law of the simulation system and the probability of failure or deviation.

[0016] The probability evaluation indicators include one or more of the following: camera video signal injection failure probability, lidar signal injection failure probability, millimeter-wave radar signal injection failure probability, CAN bus error probability, and Ethernet packet loss probability.

[0017] Optionally, the semantic description evaluation index refers to an evaluation index constructed through subjective evaluation methods such as semantic description and hierarchical assignment for subjective perceptions or complex characteristics in the simulation system that are difficult to quantify.

[0018] The semantic description evaluation indicators include one or more of the following: the realism of lighting and weather effects, the naturalness of scene interaction behavior, the perceptible degree of system latency, and the naturalness of vehicle posture changes.

[0019] Optionally, the weight data corresponding to the multimodal evaluation index data includes:

[0020] The deterministic numerical evaluation index, probabilistic evaluation index, and semantic description evaluation index in the multimodal evaluation index data are assigned corresponding weights in the form of interval values.

[0021] Optionally, the confidence level of the shaft-coupled vehicle-in-the-loop simulation system is generated based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data. The confidence level corresponding to the confidence level is obtained by using the EKM algorithm, including:

[0022] Based on the type II fuzzy set representation corresponding to the multimodal evaluation index data and weight data, the evaluation index and weight are weighted and combined to obtain the confidence level of the shaft-coupled vehicle in-loop simulation system.

[0023] The type II fuzzy set representation of the confidence level is obtained using the EKM algorithm;

[0024] The type-2 fuzzy set representation of the confidence level is converted into the final confidence level rating.

[0025] Optionally, converting the type-2 fuzzy set representation of the confidence level into the final confidence level rating includes:

[0026] Define multi-level confidence level labels and convert the confidence level labels into type-2 fuzzy set representations;

[0027] Calculate the similarity between the confidence level and the confidence level label, and take the label with the highest similarity as the final evaluation of the confidence level of the shaft-coupled vehicle-in-the-loop simulation system.

[0028] According to another aspect of the present invention, a comprehensive evaluation device for the confidence level of an axle-coupled vehicle-in-the-loop simulation system, which applies the confidence level comprehensive evaluation method for axle-coupled vehicle-in-the-loop simulation system described in any one of the above claims, is also provided, the device comprising:

[0029] The index generation module, based on the virtual environment generation module, sensor data injection module, data interaction module, and power load module in the simulation system, determines multimodal evaluation index data to characterize the confidence features of the axle-coupled vehicle-in-the-loop simulation system. The multimodal evaluation index data includes deterministic numerical evaluation indexes, probabilistic evaluation indexes, and semantic description evaluation indexes. Deterministic numerical evaluation indexes refer to evaluation indexes that are obtained based on direct measurement or quantitative calculation, with clear and reproducible numerical results. Probabilistic evaluation indexes refer to evaluation indexes that are based on statistical analysis methods and describe the performance fluctuation law and the probability of failure or deviation of the simulation system. Semantic description evaluation indexes refer to evaluation indexes constructed through subjective evaluation methods such as semantic description or hierarchical assignment for subjective perceptions or complex characteristics in the simulation system that are difficult to quantify.

[0030] The conversion module is used to set the weight data corresponding to the multimodal evaluation index data and convert the multimodal evaluation index data and weight data into a type-two fuzzy set representation.

[0031] The model evaluation module is used to generate the confidence level of the axle-coupled vehicle-in-the-loop simulation system based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data. The EKM algorithm is used to obtain the confidence level corresponding to the confidence level, so as to comprehensively evaluate the confidence level of the axle-coupled vehicle-in-the-loop simulation system using the confidence level.

[0032] The present invention also provides a computer-readable storage medium for storing program code for executing the confidence comprehensive evaluation method for the shaft-coupled vehicle-in-the-loop simulation system described in any of the preceding claims.

[0033] The present invention also provides a computing device, the computing device including a processor and a memory: the memory is used to store program code and transmit the program code to the processor; the processor is used to execute the confidence comprehensive evaluation method of the shaft-coupled whole vehicle in-loop simulation system described above according to the instructions in the program code.

[0034] To address the shortcomings of existing confidence evaluation methods for axle-coupled vehicle-in-the-loop simulation systems, such as limited dimensionality and significant uncertainty interference, this invention provides a confidence evaluation method and apparatus for axle-coupled vehicle-in-the-loop simulation systems based on multimodal data fusion. This method constructs multimodal confidence evaluation indicators encompassing deterministic numerical, probabilistic, and semantic descriptions, as well as indicator weights represented by intervals. By transforming multimodal data, a standardized representation under a unified scale is obtained, effectively eliminating uncertainties in the evaluation process and achieving a scientific and comprehensive evaluation of the confidence level of axle-coupled vehicle-in-the-loop simulation systems.

[0035] Compared to conventional shaft-coupled vehicle-in-the-loop simulation confidence evaluation methods, the method proposed in this invention is more scientific and comprehensive, specifically in the following aspects:

[0036] (1) The confidence evaluation method of the shaft-coupled whole vehicle in-loop simulation system based on multimodal data fusion proposed in this invention solves the problems of single evaluation dimension and significant subjective interference in the existing confidence evaluation methods, and realizes scientific and comprehensive confidence evaluation.

[0037] (2) This invention constructs a multimodal confidence evaluation index covering deterministic numerical values, probabilistic and semantic descriptions, as well as index weights represented by intervals. By transforming multimodal data, a standardized representation under a unified scale is obtained, and the uncertainty in the evaluation process is effectively eliminated, thereby significantly improving the accuracy and reliability of the confidence evaluation results.

[0038] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below.

[0039] The above and other objects, advantages and features of the present invention will become more apparent to those skilled in the art from the following detailed description of specific embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0040] Figure 1 A schematic diagram of the confidence comprehensive evaluation method for the shaft-coupled vehicle-in-the-loop simulation system according to an embodiment of the present invention is shown. Detailed Implementation

[0041] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0042] This invention provides a comprehensive confidence evaluation method for a shaft-coupled vehicle-in-the-loop simulation system, applicable to such systems. The shaft-coupled vehicle-in-the-loop simulation system includes a virtual environment generation module, a sensor data injection module, a data interaction module, and a power load module. The virtual environment generation module constructs high-precision simulation scenarios of roads, climate, and traffic, generating dynamic operating condition commands to provide a virtual scenario foundation for the entire system's simulation operation. The sensor data injection module simulates signals from various vehicle sensors such as cameras, radar, and GPS, accurately injecting them into the actual vehicle ECU to support its perception and decision-making logic verification. The data interaction module enables real-time transmission and interaction of multi-source data between modules, unifying the time reference and ensuring closed-loop data flow between the virtual environment, hardware devices, and actual vehicle components. The power load module simulates load characteristics such as road resistance, gradient, and adhesion coefficient using a dynamometer, providing precise load feedback to the actual vehicle chassis to achieve dynamic matching between the load and the virtual operating conditions.

[0043] like Figure 1 As shown, the confidence comprehensive evaluation method of the shaft-coupled vehicle-in-the-loop simulation system of this invention includes the following steps S1 to S3.

[0044] S1: Based on the virtual environment generation module, sensor data injection module, data interaction module and power load module in the simulation system, determine the multimodal evaluation index data used to characterize the confidence features of the axle-coupled vehicle in-loop simulation system. The multimodal evaluation index data includes deterministic numerical evaluation index, probabilistic evaluation index and semantic description evaluation index.

[0045] S2: Set the weight data corresponding to the multimodal evaluation index data, and convert the multimodal evaluation index data and weight data into a type II fuzzy set representation.

[0046] S3: Generate the confidence level of the shaft-coupled vehicle-in-the-loop simulation system based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data. Use the EKM algorithm to obtain the confidence level corresponding to the confidence level, so as to comprehensively evaluate the confidence level of the shaft-coupled vehicle-in-the-loop simulation system using the confidence level.

[0047] The method proposed in this invention effectively overcomes the core limitation of the single evaluation dimension in the prior art, and also significantly reduces the uncertainty disturbance in the evaluation process, realizing a scientific and comprehensive evaluation of the confidence level of the shaft-coupled vehicle-in-the-loop simulation system.

[0048] To achieve a comprehensive and accurate evaluation of the confidence level of axle-coupled vehicle-in-the-loop simulation system, and to avoid the limitations of a single type of indicator, such as one-sided evaluation dimensions and distorted results, this invention constructs a three-category differentiated evaluation indicator framework based on the core influencing factors and evaluation logic of confidence level. This framework enables multi-dimensional characterization and quantification of the confidence level characteristics of axle-coupled vehicle-in-the-loop simulation system. The definitions and core characterization dimensions of each indicator are as follows:

[0049] I. Deterministic Numerical Evaluation Indicators

[0050] Deterministic numerical evaluation indicators refer to evaluation indicators that are obtained through direct measurement or quantitative calculation, with clearly defined and reproducible numerical results. Optionally, the deterministic numerical evaluation indicators in this implementation include one or more of the following: steady-state torque control error, anti-drag resistance simulation error, vehicle longitudinal attitude simulation error, vehicle lateral attitude simulation error, wheel speed synchronization error, acceleration consistency error, virtual camera target detection position error, virtual lidar point cloud ranging error, virtual millimeter-wave radar speed measurement error, and scene slope resistance simulation error. Specifically, it can be expressed as:

[0051]

[0052] In the formula, Indicates a deterministic numerical evaluation index. This represents the steady-state torque control error, used to describe the percentage error between the coupler's commanded torque and the actual output torque under steady-state conditions. This represents the simulation error of anti-drag resistance, used to describe the percentage error between the torque of vehicle driving resistance (including air resistance, rolling resistance, etc.) simulated by the coupler under coasting conditions and the theoretical resistance torque calculated based on the vehicle model and standard formulas. This represents the longitudinal attitude simulation error of the vehicle body, used to describe the percentage error between the vehicle pitch angle calculated by the simulation model and the measured value from the high-precision inertial measurement unit (IMU). This represents the lateral attitude simulation error of the vehicle body, used to describe the percentage error between the vehicle roll angle calculated by the simulation model and the measured value by the IMU. This represents the wheel speed synchronization error, used to describe the percentage error between the non-driving wheel speeds output by the simulation model and the actual vehicle wheel speeds. This represents the acceleration consistency error, used to describe the percentage error between the longitudinal acceleration calculated by the vehicle dynamics model and the acceleration value measured by the IMU. This represents the virtual camera target detection position error, used to describe the percentage error between the position of the virtual camera perceived on the two-dimensional image plane and the theoretical position projected onto the image plane based on the ground truth. This represents the ranging error of a virtual lidar point cloud, used to describe the percentage error between the average distance calculated from the virtual lidar point cloud data and the true distance. This represents the velocity measurement error of the virtual millimeter-wave radar, used to describe the percentage error between the target velocity output by the virtual millimeter-wave radar and the theoretical true value. This represents the simulation error of slope resistance in a scenario. It describes the percentage error between the equivalent slope value calculated from the output torque of the shaft-coupled dynamometer keeping the actual vehicle stationary in a virtual scenario with a specific slope and the theoretical slope value set in the scenario.

[0053] II. Probability Evaluation Indicators

[0054] Probabilistic evaluation metrics refer to evaluation metrics based on statistical analysis methods that describe the performance fluctuation patterns and the probability of failure or deviation in a simulation system. Probabilistic evaluation metrics include one or more of the following: camera video signal injection failure probability, lidar signal injection failure probability, millimeter-wave radar signal injection failure probability, CAN bus error probability, and Ethernet packet loss probability. Specifically, they can be expressed as:

[0055]

[0056] In the formula, This represents a probability evaluation index. This indicates the probability of camera video signal injection failure. This indicates the probability of LiDAR signal injection failure. This indicates the probability of millimeter-wave radar signal injection failure. Indicates the CAN bus error probability. This indicates the probability of Ethernet packet loss.

[0057] III. Semantic Description Evaluation Indicators

[0058] Semantic description evaluation metrics refer to evaluation metrics constructed through subjective evaluation methods such as semantic description and hierarchical assignment for subjective perceptions or complex characteristics in simulation systems that are difficult to quantify. Semantic description evaluation metrics include one or more of the following: realism of lighting and weather effects, naturalness of scene interaction behavior, perceptible system latency, and naturalness of vehicle posture changes. Specifically, they can be expressed as:

[0059]

[0060] In the formula, This represents a semantic description evaluation metric. This indicates the realism of the lighting and weather effects. Indicates the naturalness of scene interaction behavior. Indicates the perceptible level of system latency. This indicates the naturalness of changes in vehicle body posture.

[0061] As mentioned in step S2 above, weight data corresponding to the multimodal evaluation index data is set, and the multimodal evaluation index data and weight data are converted into a type-two fuzzy set representation. Setting the weight data corresponding to the multimodal evaluation index data includes assigning corresponding weights to the deterministic numerical evaluation index, probabilistic evaluation index, and semantic description evaluation index in the multimodal evaluation index data using interval values.

[0062] In the confidence assessment process, two main limitations are encountered. Firstly, the evaluation indicators themselves generally possess uncertainty. For deterministic numerical and probabilistic evaluation indicators, measurement errors, system noise, environmental disturbances, and model simplification errors during testing lead to inherent fluctuations and reliability issues in measurement and statistical results. For semantic descriptive evaluation indicators, subjective uncertainty arises due to individual ambiguity in semantic understanding, differences in evaluators' experience backgrounds, and inconsistencies in subjective scoring scales, resulting in differences in the evaluators' cognitive abilities. Secondly, evaluation data from different modalities differ fundamentally in their dimensions and representational dimensions. Deterministic indicators are mostly physical scalars, probabilistic indicators exhibit statistical distributions, while subjective evaluations are discrete semantic levels. This heterogeneity makes it difficult to standardize, quantify, and directly compare multi-source data within a unified reference framework, constituting a core technical bottleneck for the effective fusion of cross-modal data.

[0063] Similar to evaluation indicators, weight allocation also exhibits significant subjective uncertainty (stemming from expert cognitive ambiguity and experience differences). Therefore, this embodiment of the invention uses interval values ​​to represent evaluation weights to accurately quantify this uncertainty. On the other hand, the forms of expression for evaluation indicators and indicator weights are highly diverse, encompassing four heterogeneous forms: deterministic numerical values, probability distributions, semantic descriptions, and interval values. The mathematical properties and operational rules of these different forms differ fundamentally, leading to limitations in the applicability of traditional direct calculation methods. In practical applications, the interval values ​​corresponding to the weights can be set according to the specific circumstances.

[0064] Meanwhile, in order to overcome the above-mentioned technical limitations, the embodiments of the present invention convert the evaluation index containing deterministic numerical values, probabilities and semantic descriptions, as well as the evaluation weights represented by intervals, into a type II fuzzy set representation, specifically including the following steps S2-1 to S2-4.

[0065] S2-1: Model the deterministic numerical evaluation index as a type II fuzzy set.

[0066] Assuming all deterministic numerical evaluation indicators It is any value .Will Standardize to the range of 0-100, i.e. Due to the existence of uncertainty, the following settings are made: The standard deviation is , The standard deviation is Therefore, Model as a type II fuzzy set , j Indicates the index number. Specifically, it is expressed as:

[0067]

[0068] In the formula, FOU The uncertain region of a fuzzy set is defined by all ordered pairs that satisfy a specific condition. x,u )composition, x This indicates the basic values ​​of the evaluation indicators, and their range is from... , and To determine, u express x The first-level membership degree of a fuzzy set, whose value range is determined by the lower membership function. membership function Confirmed, respectively represented as:

[0069]

[0070] In the formula, The mean is The standard deviation is Gaussian function, , .

[0071] S2-2: Model the probability evaluation index as a type II fuzzy set.

[0072] Assuming all probability evaluation metrics It is any value .Will Standardize to the range of 0-100, that is, satisfy... . The standard deviation is expressed as Therefore, Model as a type II fuzzy set , j Indicates the index number. Corresponding FOU It can be represented as:

[0073]

[0074] In the formula, c This represents the coverage factor, set to 2, and the subordinate membership function. membership function They are represented as follows:

[0075]

[0076] In the formula, , , This represents the center offset, with a value of 5. , , Take 0.2.

[0077] S2-3: Model the semantic description evaluation index as a type II fuzzy set.

[0078] A 7-level subjective semantic evaluation table is defined, comprising {Extremely Trustworthy (EC), Very Trustworthy (VC), Moderately Trustworthy (MC), Weakly Trustworthy (WC), Less Trustworthy (LC), Moderately Untrustworthy (HC), and Untrustworthy (NC)}. Each level of subjective semantic evaluation is mapped to a numerical representation from 0 to 100, resulting in the subjective evaluation. Assuming all semantic description evaluation metrics It is any value ,have For the indicators at both ends (extremely reliable EC, unreliable NC), a piecewise Gaussian representation is used. Model as a type II fuzzy set , j Indicates the index number. The corresponding FOU can be represented as:

[0079]

[0080] That is, for an extremely reliable EC, its membership function is... membership function They are represented as follows:

[0081]

[0082] For an untrusted NC, its membership function membership function They are represented as follows:

[0083]

[0084] For the five intermediate evaluation levels {Very Trustworthy (VC), Moderately Trustworthy (MC), Weakly Trustworthy (WC), Less Trustworthy (LC), and Moderately Untrustworthy (HC)}, a one-sided Gaussian representation is used, with its membership function... membership function They are represented as follows:

[0085]

[0086] In the formula, Indicates its standard deviation, , and These represent the boundary adjustment parameters, l and e Generally, 3 to 5 are used. h Generally, 2 to 4 are used.

[0087] S2-4: Determine the weights corresponding to each type of evaluation index and model them as a type II fuzzy set.

[0088] For the three categories of evaluation indicators—deterministic numerical values, probabilistic indicators, and semantic description evaluation indicators (a total of 19 indicators: 10 deterministic numerical indicators, 5 probabilistic indicators, and 4 semantic description evaluation indicators)—the weight of each evaluation indicator was determined through expert subjective judgment. Since the type-two fuzzy set method does not require the sum of weights to be 1, the weights in this embodiment are... With interval To express, This represents the lower limit of the weight of this indicator (the most conservative estimate). This represents the upper limit of the weight of this indicator (the most optimistic estimate), and satisfies... .and, and They represent and standard deviationj Indicates the weight index, when When, the weight of the corresponding deterministic numerical index, when When, the weight of the corresponding probability index, when When, the weight of the corresponding semantic description evaluation index.

[0089] Will Model as a type II fuzzy set Specifically, it is expressed as:

[0090]

[0091] Its membership function membership function They are represented as follows:

[0092]

[0093] Furthermore, in step S3 above, the confidence level of the shaft-coupled vehicle-in-the-loop simulation system is generated based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data. The confidence level evaluation corresponding to the confidence level is obtained by using the EKM algorithm, including the following steps S3-1 to S3-3.

[0094] S3-1: Based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data, the evaluation indexes and weights are weighted and combined to obtain the confidence level of the shaft-coupled vehicle-in-the-loop simulation system. It can be represented as:

[0095]

[0096] S3-2: Use the EKM algorithm to obtain the type-2 fuzzy set representation corresponding to the confidence level.

[0097] Since the evaluation indicators and weights are mapped to type II fuzzy sets, direct weighted calculation is not possible. Therefore, this embodiment of the invention uses the Enhanced Karnik-Mendel (EKM) algorithm to calculate the confidence level. The upper membership function (UMF) and lower membership function (LMF) are obtained. Type II fuzzy set representation: .

[0098] Taking UMF calculation as an example, the membership interval [0,1] is divided into... indivual The cut-off set, that is:

[0099]

[0100] In the formula, Indicates the first f indivual Cutoff set. Assume the evaluation metrics are in... The UMF interval at that location is represented as Weight in The UMF interval at that location is represented as For each The left endpoint of the UMF interval after the evaluation index and weight are fused is solved using the EKM iterative formula. Right endpoint :

[0101]

[0102] In the formula, They represent the first j Each weight in The left and right endpoints of the UMF interval at that location. They respectively represent the first j The evaluation index corresponding to each weight is The left and right endpoints of the UMF interval at that location. , These represent the iteration boundaries of the left and right endpoints of the UMF, respectively. All points in... place , Perform a connection to obtain the confidence level. The UMF. Similarly, the confidence level is obtained by following the same steps. LMF.

[0103] S3-3: Confidence level The type-II fuzzy set representation is converted into the final confidence level. Specifically, firstly, multi-level confidence level labels are defined, and the confidence level labels are converted into type-II fuzzy set representations; then, the similarity between the confidence level and the confidence level label is calculated, and the label with the highest similarity is taken as the final confidence level of the shaft-coupled vehicle-in-the-loop simulation system.

[0104] Similar to steps S2-3 above, first, 7 levels of confidence labels are predefined and converted into type-2 fuzzy set representations, resulting in... , i Indicates the confidence level. Then calculate the confidence level. With confidence labels Similarity:

[0105]

[0106] In the formula, Indicates confidence level With the i Confidence labels similarity, G Represents the main variable x The total number of discrete sampling points within [0, 100] , They represent In the sampling points The upper membership value and lower membership value at the location, , They represent In the sampling points The upper and lower membership values ​​at each point are calculated. Finally, the label with the highest similarity is taken as the final evaluation of the confidence level of the shaft-coupled vehicle-in-the-loop simulation system.

[0107] In a specific embodiment of this invention, measured data for various evaluation indicators were collected based on a shaft-coupled vehicle-in-the-loop simulation test platform for a certain vehicle model. Considering that deterministic and probabilistic indicators are inverse indicators (lower values ​​are better), they were normalized using preset performance thresholds, mapping them to positive confidence scores within the range of 0-100. Higher scores indicate better performance and higher confidence. For semantic description-based evaluation indicators, confidence levels were preset based on expert experience, as shown in Table 1. Evaluations were then conducted according to Table 1 to obtain specific confidence scores.

[0108] Table 1. Relationship between the credibility level and confidence score of semantic description indicators

[0109]

[0110] The statistical results of the confidence scores for each evaluation indicator are shown in Table 2. Based on the analysis of historical data, the standard deviation of the confidence scores for each indicator was further determined; for the deterministic indicators, the standard deviation of their standard deviations was also determined. Regarding the weighting, an interval format was used, with the left and right endpoints of the interval representing the lower and upper limits of the weights, respectively, and the standard deviations corresponding to the upper and lower limits were determined for each.

[0111] Table 2. Confidence scores and weights of evaluation indicators

[0112]

[0113] Substituting the data from Table 2 into the calculation, we can obtain the type-II fuzzy set representation of each evaluation index and its corresponding weight. Based on this, the EKM algorithm is used to perform a weighted fusion calculation on the indexes and weights, and the resulting fusion is still presented in the form of a type-II fuzzy set.

[0114] In this embodiment of the invention, the confidence levels divided in Table 1 are selected as evaluation labels. The specific operation is as follows: For each confidence level, the median of its score range is taken as the confidence score of that level, and the corresponding standard deviation is determined by combining industry knowledge and data statistics results; then, referring to the type II fuzzy set conversion method of deterministic numerical indicators, these confidence labels are converted into type II fuzzy sets.

[0115] The weighted fusion result of the indicators and weights was compared with the type II fuzzy set of each confidence label to calculate the similarity. The results are shown in Table 3. Finally, the confidence label with the highest similarity was selected as the confidence evaluation conclusion of the axle-coupled vehicle-in-the-loop simulation system. The evaluation level corresponding to this example is medium confidence.

[0116] Table 3 Similarity Results

[0117]

[0118] Based on the same inventive concept, this embodiment of the invention also provides a comprehensive confidence evaluation device for axle-coupled vehicle-in-the-loop simulation system, used to execute the comprehensive confidence evaluation method for axle-coupled vehicle-in-the-loop simulation system described in the above embodiment. The device includes an index generation module, a conversion module, and a model evaluation module.

[0119] The index generation module, based on the virtual environment generation module, sensor data injection module, data interaction module, and power load module in the simulation system, determines multimodal evaluation index data to characterize the confidence features of the axle-coupled vehicle-in-the-loop simulation system. The multimodal evaluation index data includes deterministic numerical evaluation indexes, probabilistic evaluation indexes, and semantic description evaluation indexes. Deterministic numerical evaluation indexes refer to evaluation indexes that are obtained based on direct measurement or quantitative calculation, with clear and reproducible numerical results. Probabilistic evaluation indexes refer to evaluation indexes that are based on statistical analysis methods and describe the performance fluctuation law and the probability of failure or deviation of the simulation system. Semantic description evaluation indexes refer to evaluation indexes constructed through subjective evaluation methods such as semantic description or hierarchical assignment for subjective perceptions or complex characteristics in the simulation system that are difficult to quantify.

[0120] The conversion module is used to set the weight data corresponding to the multimodal evaluation index data and convert the multimodal evaluation index data and weight data into a type-two fuzzy set representation.

[0121] The model evaluation module is used to generate the confidence level of the axle-coupled vehicle-in-the-loop simulation system based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data. The EKM algorithm is then used to obtain the confidence level corresponding to the confidence level, so as to comprehensively evaluate the confidence level of the axle-coupled vehicle-in-the-loop simulation system using the confidence level evaluation. The specific functions of each module in this embodiment can be found in the above method embodiment, and will not be repeated here.

[0122] An optional embodiment of the present invention also provides a computer-readable storage medium for storing program code for executing the confidence comprehensive evaluation method for the shaft-coupled vehicle-in-the-loop simulation system described in the above embodiments.

[0123] An optional embodiment of the present invention also provides a computing device, the computing device including a processor and a memory: the memory is used to store program code and transmit the program code to the processor; the processor is used to execute the confidence comprehensive evaluation method of the shaft-coupled whole vehicle in-the-loop simulation system described in the above embodiment according to the instructions in the program code.

[0124] Those skilled in the art will clearly understand that the specific working process of the systems, devices, modules and units described above can be referred to the corresponding process in the foregoing method embodiments. For the sake of brevity, it will not be repeated here.

[0125] Furthermore, the functional units in the various embodiments of the present invention can be physically independent of each other, or two or more functional units can be integrated together, or all functional units can be integrated into one processing unit. The integrated functional units described above can be implemented in hardware, or in software or firmware.

[0126] Those skilled in the art will understand that if the integrated functional unit is implemented in software and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or all or part of it, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computing device (e.g., a personal computer, server, or network device) to execute all or part of the steps of the methods described in the embodiments of the present invention when running the instructions. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0127] Alternatively, all or part of the steps of the foregoing method embodiments can be implemented by hardware (such as a computing device, personal computer, server, or network device) related to program instructions. The program instructions can be stored in a computer-readable storage medium. When the program instructions are executed by the processor of the computing device, the computing device executes all or part of the steps of the methods described in the various embodiments of the present invention.

[0128] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that within the spirit and principles of the present invention, modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein; and these modifications or substitutions do not cause the corresponding technical solutions to depart from the protection scope of the present invention.

Claims

1. A comprehensive evaluation method for the confidence level of an axle-coupled vehicle-in-the-loop simulation system, characterized in that, include: Based on the virtual environment generation module, sensor data injection module, data interaction module, and power load module in the simulation system, multimodal evaluation index data is determined to characterize the confidence features of the shaft-coupled vehicle-in-the-loop simulation system. The multimodal evaluation index data includes deterministic numerical evaluation indexes, probabilistic evaluation indexes, and semantic description evaluation indexes. The deterministic numerical evaluation indexes refer to evaluation indexes that are obtained based on direct measurement or quantitative calculation, with clear and reproducible numerical results. The probabilistic evaluation indexes refer to evaluation indexes that are based on statistical analysis methods and describe the performance fluctuation law and the probability of failure or deviation of the simulation system. The semantic description evaluation indexes refer to evaluation indexes constructed through subjective evaluation methods such as semantic description or hierarchical assignment for subjective perception or complex characteristics in the simulation system that are difficult to quantify. Set the weight data corresponding to the multimodal evaluation index data, and convert the multimodal evaluation index data and weight data into a type-two fuzzy set representation; The confidence level of the axle-coupled vehicle-in-the-loop simulation system is generated based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data. The confidence level corresponding to the confidence level is obtained by using the EKM algorithm, so as to comprehensively evaluate the confidence level of the axle-coupled vehicle-in-the-loop simulation system.

2. The method according to claim 1, characterized in that, The deterministic numerical evaluation index includes one or more of the following: steady-state torque control error, anti-drag resistance simulation error, vehicle longitudinal attitude simulation error, vehicle lateral attitude simulation error, wheel speed synchronization error, acceleration consistency error, virtual camera target detection position error, virtual lidar point cloud ranging error, virtual millimeter-wave radar speed measurement error, and scene slope resistance simulation error.

3. The method according to claim 1, characterized in that, The probability evaluation indicators include one or more of the following: camera video signal injection failure probability, lidar signal injection failure probability, millimeter-wave radar signal injection failure probability, CAN bus error probability, and Ethernet packet loss probability.

4. The method according to claim 1, characterized in that, The semantic description evaluation indicators include one or more of the following: the realism of lighting and weather effects, the naturalness of scene interaction behavior, the perceptible degree of system latency, and the naturalness of vehicle posture changes.

5. The method according to any one of claims 1 to 4, characterized in that, The weight data corresponding to the multimodal evaluation index data includes: The deterministic numerical evaluation index, probabilistic evaluation index, and semantic description evaluation index in the multimodal evaluation index data are assigned corresponding weights in the form of interval values.

6. The method according to any one of claims 1 to 4, characterized in that, The confidence level of the shaft-coupled vehicle-in-the-loop simulation system is generated based on the type-II fuzzy set representation corresponding to the multimodal evaluation index data and weight data. The confidence level corresponding to the confidence level is obtained by using the EKM algorithm, including: Based on the type II fuzzy set representation corresponding to the multimodal evaluation index data and weight data, the evaluation index and weight are weighted and combined to obtain the confidence level of the shaft-coupled vehicle in-loop simulation system. The type II fuzzy set representation of the confidence level is obtained using the EKM algorithm; The type-2 fuzzy set representation of the confidence level is converted into the final confidence level rating.

7. The method according to claim 6, characterized in that, The conversion of the type-2 fuzzy set representation of confidence into the final confidence rating level includes: Define multi-level confidence level labels and convert the confidence level labels into type-2 fuzzy set representations; Calculate the similarity between the confidence level and the confidence level label, and take the label with the highest similarity as the final evaluation of the confidence level of the shaft-coupled vehicle-in-the-loop simulation system.

8. A comprehensive evaluation device for the confidence level of a shaft-coupled vehicle-in-the-loop simulation system applied to the comprehensive evaluation method for confidence level of a shaft-coupled vehicle-in-the-loop simulation system according to any one of claims 1 to 7, characterized in that, include: The index generation module, based on the virtual environment generation module, sensor data injection module, data interaction module, and power load module in the simulation system, determines multimodal evaluation index data to characterize the confidence features of the axle-coupled vehicle-in-the-loop simulation system. The multimodal evaluation index data includes deterministic numerical evaluation indexes, probabilistic evaluation indexes, and semantic description evaluation indexes. Deterministic numerical evaluation indexes refer to evaluation indexes that are obtained based on direct measurement or quantitative calculation, with clear and reproducible numerical results. Probabilistic evaluation indexes refer to evaluation indexes that are based on statistical analysis methods and describe the performance fluctuation law and the probability of failure or deviation of the simulation system. Semantic description evaluation indexes refer to evaluation indexes constructed through subjective evaluation methods such as semantic description or hierarchical assignment for subjective perceptions or complex characteristics in the simulation system that are difficult to quantify. The conversion module is used to set the weight data corresponding to the multimodal evaluation index data and convert the multimodal evaluation index data and weight data into a type-two fuzzy set representation. The model evaluation module is used to generate the confidence level of the axle-coupled vehicle-in-the-loop simulation system based on the type-two fuzzy set representation corresponding to the multimodal evaluation index data and weight data, and to obtain the confidence level corresponding to the confidence level using the EKM algorithm, so as to comprehensively evaluate the confidence level of the axle-coupled vehicle-in-the-loop simulation system using the confidence level.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program code for executing the confidence comprehensive evaluation method of the shaft-coupled whole vehicle in-loop simulation system according to any one of claims 1 to 7.

10. A computing device, characterized in that, The computing device includes a processor and memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the confidence comprehensive evaluation method of the shaft-coupled whole vehicle-in-the-loop simulation system according to any one of the instructions in the program code.

Images