Intelligent network connection automobile test method based on vehicle-road cloud integration

Abstract

The invention discloses an intelligent networked automobile test method based on vehicle-road cloud integration. The method comprises the following steps: S1, obtaining vehicle end data, road end data and cloud end data; s2, transmitting the obtained vehicle end data, road end data and cloud data to a vehicle and road cloud data center, and performing data optimization processing through the vehicle and road cloud data center to obtain a multi-source data optimization processing result; s3, generating real road extremely cold environment scene data according to the multi-source data optimization processing result; s4, testing and evaluating the vehicle according to the real road extremely cold environment scene data; the optimization processing effect of the vehicle-road cloud integrated data is improved, the operation performance of the intelligent connected vehicle in the extremely cold environment scene is tested and evaluated, and then the operation reliability of the intelligent connected vehicle in the extremely cold environment is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile testing, and in particular to an intelligent networked automobile testing method based on vehicle-road-cloud integration. BACKGROUND

[0002] In a new round of scientific and technological revolution and industrial change, the automobile has not been a simple means of transportation, but has gradually transformed into a comprehensive scientific and technological product integrating intelligent interaction, automatic control, external communication and artificial intelligence. After the automobile industry is deeply integrated with the Internet of Things, communication and other fields, people realize real-time interaction between smart vehicles and intelligent roads, and on the basis of dynamic traffic information collection and fusion in all time and space, carry out vehicle active control and road cooperative management, fully realize the effective cooperation of vehicle-road-cloud, and ultimately achieve the purpose of improving traffic efficiency and ensuring traffic safety.

[0003] However, if the effective cooperation of vehicle-road-cloud is to be realized, the design and testing of the cooperation scheme need to be carried out, especially the intelligent networked automobile testing in extremely cold scenes. In the existing testing method, although the software-in-the-loop testing can solve the problem of testing efficiency, the authenticity of the testing scene is insufficient, especially the road testing, which not only cannot meet the demand of safe driving, but also the road scene cannot be reproduced, and the scene repeatability testing of the intelligent driving electric control system cannot be carried out, so the running performance testing demand of the intelligent networked automobile in the extremely cold environment cannot be met. SUMMARY

[0004] The embodiment of the present application provides an intelligent networked automobile testing method based on vehicle-road-cloud integration, which solves the above technical problems.

[0005] Therefore, the present application provides an intelligent networked automobile testing method based on vehicle-road-cloud integration, which comprises the following steps: S1, acquiring vehicle-end data, road-end data and cloud-end data; S2, transmitting the acquired vehicle-end data, road-end data and cloud-end data to a vehicle-road-cloud data center, and performing data optimization processing through the vehicle-road-cloud data center to obtain a multi-source data optimization processing result; S3, generating real road extremely cold environment scene data according to the multi-source data optimization processing result; S4, testing and evaluating the vehicle according to the real road extremely cold environment scene data.

[0006] Optionally, in step S2, the transmitting of the acquired vehicle-end data, road-end data and cloud-end data to the vehicle-road-cloud data center comprises: transmitting the vehicle-end data, road-end data and cloud-end data to the vehicle-road-cloud data center in real time through 5G communication technology.

[0007] Optionally, in step S2, the vehicle-road-cloud data center performs data processing, including: performing standardization processing and integrated analysis on the vehicle-end data, road-end data and cloud-end data to form vehicle-road-cloud integrated data; Based on the LSTM network, the vehicle-road-cloud integrated data is divided into a training set and a test set, and after parameter optimization, a new evaluation optimization model is constructed; Based on the new evaluation optimization model, the membership information of the vehicle-road-cloud multi-source data in the vehicle-road-cloud integrated data to the matched regional hierarchical framework information is analyzed, and the influence degree information of the regional hierarchy is calculated according to the multi-target data type information obtained by the vehicle-road-cloud multi-source data; According to the membership information and influence degree information, the current different vehicle-road-cloud data is fuzzy evaluated, and the evaluation score of each vehicle-road-cloud data is obtained, and the regional hierarchical framework of the current different vehicle-road-cloud data is determined, the comprehensive integral of the regional hierarchical framework formed by the current different vehicle-road-cloud data is calculated according to the evaluation score of the current different vehicle-road-cloud data and the matched regional hierarchical framework information, and the integral of each vehicle-road-cloud data is updated through the new evaluation optimization model according to the comprehensive integral of the regional hierarchical framework and the evaluation score of each vehicle-road-cloud data, and each vehicle-road-cloud data is matched and configured in the regional hierarchical framework.

[0008] Optionally, the construction of the new evaluation optimization model includes: performing multi-layer fusion analysis of the vehicle-road-cloud multi-target type data in the vehicle-road-cloud integrated data under the regional hierarchical framework, retraining the evaluation optimization model by adding a new level LSTM model and a target function to obtain a new evaluation optimization model, wherein the expression of the target function is:

[0009] In the formula, f(·) is a system function; x is the training set sample data of the evaluation optimization model; y is the test set sample data of the evaluation optimization model; θ1 is a parameter set of a shallow LSTM model of the evaluation optimization model; θ2 is a parameter set in an LSTM model network of a new level of the evaluation optimization model; is a loss function; R is a real number field, which represents finding a set of parameter sets θ1 of a shallow LSTM model and parameter sets θ2 in an LSTM model network of a new level in the real number field R, so that the loss function is minimized.

[0010] Optionally, the multi-layer fusion analysis under the regional hierarchical framework includes: updating each layer of vehicle-road-cloud data under the regional hierarchical framework through the new evaluation optimization model, weighting and fusing the multi-layer vehicle-road-cloud data fusion probability scores, analyzing each vehicle-road-cloud data in each layer of vehicle-road-cloud data according to the fusion probability scores, and determining the result data for output.

[0011] Optionally, the vehicle-road cloud integration data includes data standards of analysis domain, vehicle domain, traffic domain, environment domain, infrastructure domain, user behavior domain, communication domain, security domain, maintenance domain, policy and regulation domain, map and navigation domain, and event and log domain, forming an ODS database with data integration and cleaning.

[0012] Optionally, in step S4, the vehicle is tested and evaluated according to the real road extreme cold environment scene data, including loading the real road extreme cold environment scene data after modeling into a laboratory vehicle-mounted computing unit through an HIL cabinet to test scene perception, planning, decision-making, and control algorithm, and transmitting the test results to a vehicle simulation motion bench through the HIL cabinet for response and function evaluation.

[0013] From the above technical solutions, the embodiments of the present application have the following advantages: The intelligent connected vehicle testing method based on vehicle-road cloud integration provided by the present application has the following advantages compared with the prior art: the vehicle-end data, road-end data, and cloud-end data are obtained; the obtained vehicle-end data, road-end data, and cloud-end data are transmitted to a vehicle-road cloud data center, and data optimization processing is performed through the vehicle-road cloud data center to obtain multi-source data optimization processing results; real road extreme cold environment scene data is generated according to the multi-source data optimization processing results; the vehicle is tested and evaluated according to the real road extreme cold environment scene data; the running performance of the intelligent connected vehicle in the extreme cold environment scene is tested and evaluated, thereby improving the reliability of the intelligent connected vehicle in the extreme cold environment.

[0014] The vehicle-end data, road-end data, and cloud-end data are standardized and integrated for analysis to form vehicle-road cloud integration data; the vehicle-road cloud integration data is divided into a training set and a test set based on an LSTM network, and parameter optimization is performed to construct a new evaluation optimization model, thereby improving the optimization processing effect of the vehicle-road cloud integration data. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly express and illustrate the technical solutions of the embodiments of the present application, the drawings needed for the description of the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0016] Figure 1 A step diagram of the intelligent connected vehicle testing method based on vehicle-road cloud integration provided in the embodiments of the present application. DETAILED DESCRIPTION

[0017] In order for those skilled in the art to better understand the scheme of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0018] For the sake of understanding, please refer to Figure 1 An embodiment of the intelligent networked vehicle testing method based on vehicle-road-cloud integration provided by the present application comprises the following steps: S1, acquiring vehicle-end data, road-end data and cloud-end data; Specifically, the vehicle-end data is acquired by installing sensors such as laser radar, camera and millimeter wave radar on the vehicle to collect parameter data such as steering wheel angle, brake pedal stroke, throttle opening, vehicle speed and driving trajectory in real time during the driving process of the vehicle; The road-end data is acquired by deploying roadside sensing devices such as high-definition cameras and millimeter wave radars on both sides of the road to monitor information data such as traffic flow, vehicle driving state and road conditions in real time; The cloud-end data is acquired by the cloud through a communication tool to obtain position information data of all road traffic participants.

[0019] S2, transmitting the acquired vehicle-end data, road-end data and cloud-end data to a vehicle-road-cloud data center and performing data optimization processing through the vehicle-road-cloud data center to obtain a multi-source data optimization processing result; Specifically, the vehicle-end data, road-end data and cloud-end data are transmitted to the vehicle-road-cloud data center in real time through 5G communication technology, that is, the collected data is transmitted to the vehicle-road-cloud data center through 5G communication technology by using the vehicle-mounted communication module; the road-end data is transmitted to the vehicle-road-cloud data center by the roadside unit (RSU).

[0020] Further, the vehicle-road cloud data center performs data processing, including: standardizing and integrating the vehicle-end data, road-end data, and cloud-end data to form vehicle-road cloud integrated data; the vehicle-road cloud integrated data includes data standards of analysis domain, vehicle domain, traffic domain, environment domain, infrastructure domain, user behavior domain, communication domain, safety domain, maintenance domain, policy and regulation domain, map and navigation domain, and event and log domain, forming an ODS database with data integration and cleaning; the analysis domain stores data after preliminary analysis, providing support for subsequent data analysis and evaluation optimization; the vehicle domain stores vehicle-related data, including vehicle identification, location, speed, and driving state; the traffic domain stores traffic-related data, including traffic flow, traffic pattern, and traffic accident record; the environment domain records environmental factor data, including weather, temperature, humidity, and illumination; the infrastructure domain records road infrastructure status and data, including traffic signal, camera, and sensor; the user behavior domain collects user behavior data, including travel habit and preference; the communication domain records communication data between vehicles, infrastructure, pedestrians, and network; the safety domain stores traffic safety-related data, including accident report and safety warning; the maintenance domain stores vehicle maintenance and diagnosis data; the policy and regulation domain stores traffic regulation and road use policy data; the map and navigation domain includes map data, navigation information, and real-time traffic condition data; the event and log domain records system operation and event log data. This further achieves more intelligent and efficient traffic management services, improves road safety, optimizes traffic flow, reduces congestion, and provides personalized travel services.

[0021] Further, based on the LSTM network, the vehicle-road cloud integrated data is divided into a training set and a test set, and after parameter optimization, a new evaluation optimization model is constructed; based on the new evaluation optimization model, the membership information of the vehicle-road cloud multi-source data in the matched regional hierarchical framework information is analyzed, and the influence degree information of the regional hierarchy is calculated according to the multi-target data type information obtained by the vehicle-road cloud multi-source data; according to the membership information and the influence degree information, the current different vehicle-road cloud data is fuzzy evaluated and the evaluation score of each vehicle-road cloud data is obtained, and the regional hierarchical framework of the current different vehicle-road cloud data is determined, the comprehensive integral of the regional hierarchical framework formed by the current different vehicle-road cloud data is calculated according to the evaluation score of the current different vehicle-road cloud data and the matched regional hierarchical framework information, and the integral of each vehicle-road cloud data is updated through the new evaluation optimization model according to the comprehensive integral of the regional hierarchical framework and the evaluation score of each vehicle-road cloud data, and the integral of each vehicle-road cloud data is matched and configured in the regional hierarchical framework. Regional hierarchical framework: This model may contain a multi-level regional framework, which is used to distinguish and process data of different geographical or logical regions. This hierarchical method helps to more accurately manage and optimize traffic flow, and more effectively allocate resources. Information analysis and matching integral: The model can analyze the relevant information of each regional hierarchy and calculate a matching integral. Represents the consistency, completeness or prediction accuracy of the data, used to evaluate the data quality of different regions. Precise matching and configuration: According to the matching integral, the model can perform precise matching and configuration in the corresponding framework hierarchy. It can be applied to adjust the timing of traffic signal lights, optimize vehicle path planning, or adjust traffic flow control strategies. Achieve the effect of optimizing the processing of vehicle-road cloud integrated data Further, the construction of the new evaluation optimization model includes: performing multi-layer fusion analysis of the vehicle-road cloud multi-target type data in the vehicle-road cloud integrated data under the regional hierarchical framework, retraining the evaluation optimization model by adding a new level LSTM model and a target function to obtain a new evaluation optimization model, wherein the expression of the target function is:

[0022] In the formula, f(·) is a system function; x is the training set sample data of the evaluation optimization model; y is the test set sample data of the evaluation optimization model; θ1 is the parameter set of the shallow LSTM model of the evaluation optimization model; θ2 is the parameter set in the LSTM model network of the new level of the evaluation optimization model; is a loss function; R is a real number field, which means finding a set of parameters θ1 of a shallow LSTM model and a set of parameters θ2 of a new level of LSTM model network in the real number field R, so that the loss function is minimized. The multi-layer fusion analysis under the regional level framework is carried out, including: updating each layer of car-road cloud data under the regional level framework by the new evaluation optimization model, then weighting and fusing the multi-layer car-road cloud data fusion probability score, analyzing each car-road cloud data in each layer of car-road cloud data according to the fusion probability score, and determining the result data for output. The multi-layer training method of LSTM is adopted. This means that the model is not only a single-layer LSTM network, but also contains multiple LSTM layers, each of which can learn different abstract features of data, and the data processing effect and robustness: by increasing the multi-layer training, the model can be more accurate when processing car-road cloud integrated data, and the robustness of the model to abnormal values, noise and uncertainty is improved, which is crucial to ensure the stability and reliability of the transportation system. The effect of improving the optimization processing of car-road cloud integrated data is realized.

[0023] S3, generating real road extreme cold environment scene data according to the multi-source data optimization processing result; Specifically, according to the multi-source data optimization processing result, a digital test field and an environment simulation model are developed to simulate the road environment, vehicle dynamics characteristics and the like under extreme cold weather, the road spectrum model is combined with the vehicle dynamics model for joint simulation to generate realistic extreme cold scene simulation data.

[0024] S4, testing and evaluating the vehicle according to the real road extreme cold environment scene data; Specifically, the extreme cold scene simulation data after modeling is loaded into the laboratory vehicle-mounted computing unit through the HIL cabinet for testing of scene perception, planning, decision-making and control algorithm, and the test results are transmitted to the vehicle dynamics model or combined with the vehicle simulation motion bench for response and function evaluation; the running performance test and evaluation of the intelligent connected vehicle in the extreme cold environment scene are realized, and the reliability of the intelligent connected vehicle in the extreme cold environment is improved.

[0025] It needs to be supplemented that an environment simulation cabin is built in the laboratory, and the generated extreme cold scene data is injected into the perception system of the intelligent connected vehicle through perception injection technology for in-loop testing. Then, through dark box testing and the like, the perception, decision-making and control capabilities of the vehicle in the extreme cold scene are verified. In the cold region test site, different road types (such as icy road surface and snow-covered road surface) are selected for actual road testing of the intelligent connected vehicle. In the testing process, the running data of the vehicle is collected for correction and optimization of the extreme cold scene data.

[0026] In the real road environment of the extremely cold area, intelligent networked vehicles are deployed for actual operation test, system reliability, safety and adaptability of the vehicle in the real extremely cold scene are verified, and actual operation data are collected to provide reliable data basis for further optimization of the extremely cold scene test.

[0027] The terms "first", "second", "third", "fourth" and the like in the description of this application and the above-mentioned drawings, if any, are used to distinguish similar objects, and do not necessarily have to describe a particular order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented, for example, in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0028] The above-described embodiments are merely used to illustrate the technical solutions of the present application, but not limit it; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A test method for intelligent connected vehicles based on vehicle-road-cloud integration, characterized in that: The following steps are involved: S1. Obtain vehicle-side data, road-side data, and cloud-side data; S2. Transmitting the acquired vehicle-side data, road-side data, and cloud-side data to a vehicle-road cloud data center, and performing data optimization processing by the vehicle-road cloud data center to obtain a multi-source data optimization processing result; S3. Generating real road extreme cold environment scene data based on the multi-source data optimization processing results; S4. Then test and evaluate the vehicle based on the real road extreme cold environment scenario data.

2. The intelligent connected vehicle testing method based on vehicle-road-cloud integration according to claim 1, characterized in that: In step S2, the acquired vehicle-side data, road-side data and cloud-side data are transmitted to the vehicle-road cloud data center, including: transmitting the vehicle-side data, road-side data and cloud-side data to the vehicle-road cloud data center in real time through 5G communication technology.

3. The intelligent connected vehicle testing method based on vehicle-road-cloud integration according to claim 1, characterized in that: In step S2, the vehicle-road-cloud data center performs data processing, including: standardizing and integrating the vehicle-side data, road-side data, and cloud-side data to form vehicle-road-cloud integrated data; Then, based on the LSTM network, the vehicle-road-cloud integrated data is divided into a training set and a test set, and after parameter optimization, a new evaluation optimization model is constructed; Based on the new evaluation optimization model, the membership information of the multi-source vehicle-road-cloud data in the vehicle-road-cloud integrated data to the matched regional hierarchical framework information is analyzed, and the influence information of the regional hierarchical level is calculated based on the multi-target data type information obtained from the multi-source vehicle-road-cloud data; Based on the membership information and influence information, a fuzzy evaluation is performed on the current different vehicle-road cloud data and a corresponding evaluation score for each vehicle-road cloud data is obtained, and the regional hierarchical framework of the current different vehicle-road cloud data is determined. Based on the evaluation scores of the current different vehicle-road cloud data and the matched regional hierarchical framework information, the comprehensive integral of the regional hierarchical framework formed by the current different vehicle-road cloud data is calculated. Based on the comprehensive integral of the regional hierarchical framework and the evaluation score of each vehicle-road cloud data, the integral of each vehicle-road cloud data is updated through the new evaluation optimization model, and then the integral of each vehicle-road cloud data is matched and configured with the regional hierarchical framework.

4. The intelligent connected vehicle testing method based on vehicle-road-cloud integration according to claim 3 is characterized in that: The constructing of the new evaluation optimization model includes: performing a multi-layer fusion analysis on the vehicle-road-cloud multi-objective type data in the vehicle-road-cloud integrated data under the regional hierarchical framework, and retraining the evaluation optimization model by adding a new layer of LSTM model and objective function to obtain a new evaluation optimization model, wherein the expression of the objective function is: Where f(·) is the system function; x is the training set sample data of the evaluation optimization model; y is the test set sample data of the evaluation optimization model; θ1 is the parameter set of the shallow LSTM model of the evaluation optimization model; θ2 is the parameter set of the LSTM model network of the new layer of the evaluation optimization model; is the loss function; R is the real number field, which means finding a set of parameter sets θ1 of the shallow LSTM model and a set of parameter sets θ2 in the new level LSTM model network in the real number field R to minimize the loss function.

5. The intelligent connected vehicle testing method based on vehicle-road-cloud integration according to claim 4 is characterized in that: The multi-layer fusion analysis under the regional hierarchical framework includes: updating each layer of vehicle-road-cloud data under the regional hierarchical framework through the new evaluation optimization model, and then weightedly fusing the fusion probability scores of the multi-layer vehicle-road-cloud data. According to the fusion probability scores, each vehicle-road-cloud data in each layer of vehicle-road-cloud data is analyzed and outputted after being determined as result data.

6. The intelligent connected vehicle testing method based on vehicle-road-cloud integration according to claim 5, characterized in that: The vehicle-road-cloud integrated data includes: data standards for analysis domain, vehicle domain, traffic domain, environment domain, infrastructure domain, user behavior domain, communication domain, security domain, maintenance domain, policy and regulation domain, map and navigation domain, and event and log domain, forming an ODS database with data integration and cleaning capabilities.

7. The intelligent connected vehicle testing method based on vehicle-road-cloud integration according to claim 1, characterized in that: In step S4, the vehicle is tested and evaluated based on the real road extreme cold environment scene data, including: loading the real road extreme cold environment scene data after modeling into the laboratory vehicle-mounted computing unit through the HIL cabinet to test the scene perception, planning, decision-making and control algorithms, and responding and functionally evaluating the test results on the vehicle simulation motion test bench through the HIL cabinet.

Images