Automatic driving function scene data testing method and system and medium
By collecting and analyzing data of autonomous driving vehicles in real time, a self-regressive moving average vehicle performance prediction model and POT hyperparameter optimization model are constructed, which solves the problem of insufficient timeliness and accuracy of test results in traditional testing methods, realizes automated testing, and improves the Efficiency and accuracy meet different user needs.
Patent Information
- Application Number
- CN202411188219.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-06
AI Technical Summary
Traditional autonomous vehicle testing methods rely on a lot of manpower and time, making it difficult to fully cover all possible scenarios, and the timeliness and accuracy of the test results are insufficient.
Using a scenario data testing method for autonomous driving functions, by collecting on-board sensors, controllers and status data in real time, a first-order single-sequence autoregressive moving average vehicle performance prediction model and POT hyperparameter optimization model are constructed to perform performance prediction and optimization , and establish a vehicle functional scenario test evaluation function to realize automated testing.
Accurate testing of vehicle performance is achieved, timeliness and accuracy of test results are ensured, manual intervention is reduced, testing efficiency and accuracy is improved, and a comprehensive understanding of the vehicle's autonomous driving performance is met, and the testing needs of different users are met.
Smart Images

Figure CN119089201A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of autonomous driving technology, and in particular to an autonomous driving function scenario data testing method, system and medium. Background Art
[0002] With the rapid development of autonomous driving technology, the demand for testing autonomous vehicles in various complex scenarios is increasing. Traditional testing methods often rely on a lot of manpower and time for recording, analyzing test data, and outputting test results, and it is difficult to fully cover all possible scenarios. Summary of the invention
[0003] In view of the above problems, the present invention provides an autonomous driving function scenario data testing method, system and medium, which can not only collect and analyze data in real time to ensure the timeliness and accuracy of the test results, but also the entire testing process does not require human intervention, thereby improving the testing efficiency and accuracy.
[0004] In order to achieve the above-mentioned purpose and other related purposes, the technical solution provided by the present invention is as follows:
[0005] A method for testing scene data of an autonomous driving function, the method comprising:
[0006] M1. The autonomous driving vehicle is driving in a preset test scenario, collecting data information of various on-board sensors, data information of the vehicle controller, and data information of the vehicle status in real time, and performing normalization processing to obtain normalized data information of various on-board sensors, vehicle controller, and vehicle status;
[0007] M2. Based on the normalized data information of the vehicle sensors, vehicle controller and vehicle status, a vehicle performance prediction model based on the autoregressive moving average of the first-order single integral sequence is constructed to predict the performance of the vehicle and obtain the predicted vehicle performance test data information;
[0008] M3. Based on the predicted vehicle performance test data information, construct a POT hyperparameter optimization model for vehicle performance, optimize the vehicle performance test data, and obtain optimized vehicle performance test data information;
[0009] M4. Based on the optimized vehicle performance test data information, establish a vehicle function scenario test evaluation function G, evaluate the vehicle performance, and obtain the vehicle performance evaluation data information.
[0010] Furthermore, in step M2, the construction of the autoregressive moving average vehicle performance prediction model based on the first-order single integral sequence to predict the performance of the vehicle includes:
[0011] M21. Based on the normalized data information of the vehicle sensors, vehicle controller and vehicle status, a first-order single integral sequence function f of vehicle-related data is established.
[0012]
[0013] Among them, x1 is the processed data information of each on-board sensor, x2 is the data information of the vehicle controller, x3 is the data information of the vehicle status, α1, α2 and α3 are the serialization constant parameters of the vehicle-related data, and the vehicle-related data is serialized to obtain the first-order single integral sequence data information of the vehicle-related data;
[0014] M22. Input the first-order single integral sequence data information of the vehicle-related data into the autoregressive moving average vehicle performance prediction model for training and learning, and determine the vehicle performance prediction function F,
[0015]
[0016] Among them, y is the first-order single integral sequence data information of vehicle-related data, g is the mean function of the first-order single integral sequence of vehicle-related data, λ1, λ2 and λ3 are the prediction factors of vehicle performance, and the trained autoregressive moving average vehicle performance prediction model is obtained;
[0017] M23. Based on the trained autoregressive moving average vehicle performance prediction model, the first-order single integral sequence data information of the vehicle-related data is input to predict the performance of the vehicle, and obtain the predicted vehicle performance test data information.
[0018] Furthermore, the mean function g of the first-order single integral sequence of the vehicle-related data is,
[0019] Among them, y is the first-order single integral sequence data information of vehicle-related data.
[0020] Furthermore, the constraints of the vehicle voltage ratio prediction factors λ1, λ2 and λ3 are as follows:
[0021]
[0022] Furthermore, the serialization constant parameters α1, α2 and α3 of the vehicle-related data are,
[0023]
[0024] Among them, x1 is the data information of each on-board sensor after processing, x2 is the data information of the vehicle controller, and x3 is the data information of the vehicle status.
[0025] Furthermore, the method further comprises:
[0026] M6. Based on the performance evaluation data information of the vehicle, a preset threshold is set. If the performance evaluation of the vehicle is greater than the preset threshold, the vehicle's autonomous driving performance meets the requirements. If the performance evaluation of the vehicle is less than the preset threshold, the vehicle's autonomous driving performance does not meet the requirements.
[0027] Furthermore, in step M3, the POT hyperparameter optimization model of vehicle performance is constructed to optimize the vehicle performance test data, including:
[0028] M31. Based on the predicted vehicle performance test data information, construct a vehicle performance test data set and divide it into a test data set and a training data set;
[0029] M32. Input the training data set into the POT hyperparameter optimization model of vehicle performance for training and learning, and determine the vehicle performance test data optimization function W,
[0030]
[0031] Among them, z i is the i-th training data set, n is the sample size, ω i is the POT hyperparameter optimization factor of the i-th training data set, and the POT hyperparameter optimization model of the trained vehicle performance is obtained;
[0032] M33. Based on the trained POT hyperparameter optimization model of vehicle performance, input the test set, optimize and adjust the model parameters, and obtain the optimized POT hyperparameter optimization model of vehicle performance;
[0033] M34. Based on the POT hyperparameter optimization model of the optimized vehicle performance, the predicted vehicle performance test data information is input, the vehicle performance test data is optimized, and the optimized vehicle performance test data information is obtained.
[0034] Furthermore, the vehicle function scenario test evaluation function G is,
[0035]
[0036] Among them, r is the performance test data information of the optimized vehicle, and θ1 and θ2 are the evaluation change factors of the vehicle function scenario test.
[0037] In order to achieve the above-mentioned object and other related objects, the present invention further provides a system for implementing any one of the above-mentioned automatic driving function scenario data testing methods, the system comprising:
[0038] Data acquisition module, used to collect data information of various on-board sensors, data information of vehicle controller and data information of vehicle status in real time;
[0039] The data analysis module is used to construct an autoregressive moving average vehicle performance prediction model based on a first-order single integral sequence, predict the vehicle performance, obtain the predicted vehicle performance test data information, and construct a POT hyperparameter optimization model for vehicle performance to optimize the vehicle performance test data;
[0040] Test scenario management module, responsible for managing and configuring test scenarios;
[0041] The result output module is used to establish a vehicle functional scenario test evaluation function G, evaluate the performance of the vehicle, and obtain the vehicle performance evaluation data information.
[0042] In order to achieve the above-mentioned purpose and other related purposes, the present invention also provides a computer-readable storage medium, which stores a computer program programmed or configured to execute any one of the autonomous driving function scenario data testing methods.
[0043] The present invention has the following positive effects:
[0044] 1. The present invention predicts the performance of a vehicle by constructing an autoregressive moving average vehicle performance prediction model based on a first-order single integral sequence, and optimizes the performance test data of the vehicle by combining the construction of a POT hyperparameter optimization model of vehicle performance. It can not only accurately test the performance of the vehicle, but also collect and analyze data in real time to ensure the timeliness and accuracy of the test results.
[0045] 2. The present invention evaluates the performance of the vehicle by establishing a vehicle functional scenario test evaluation function G, which can not only help users fully understand the vehicle's autonomous driving performance, but also meet the testing needs of different users. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 It is a schematic diagram of the method flow of the present invention;
[0047] Figure 2 A schematic diagram of a process for constructing an autoregressive moving average vehicle performance prediction model based on a first-order single integral sequence according to the present invention;
[0048] Figure 3 A schematic diagram of a process for constructing a POT hyperparameter optimization model for vehicle performance according to the present invention;
[0049] Figure 4 It is a schematic diagram of the system framework of the present invention. DETAILED DESCRIPTION
[0050] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.
[0051] Embodiment 1: As shown in 1, a method for testing scene data of an autonomous driving function comprises:
[0052] M1. The autonomous driving vehicle is driving in a preset test scenario, collecting data information of various on-board sensors, data information of the vehicle controller, and data information of the vehicle status in real time, and performing normalization processing to obtain normalized data information of various on-board sensors, vehicle controller, and vehicle status;
[0053] M2. Based on the normalized data information of the vehicle sensors, vehicle controller and vehicle status, a vehicle performance prediction model based on the autoregressive moving average of the first-order single integral sequence is constructed to predict the performance of the vehicle and obtain the predicted vehicle performance test data information;
[0054] M3. Based on the predicted vehicle performance test data information, construct a POT hyperparameter optimization model for vehicle performance, optimize the vehicle performance test data, and obtain optimized vehicle performance test data information;
[0055] M4. Based on the optimized vehicle performance test data information, establish a vehicle function scenario test evaluation function G, evaluate the vehicle performance, and obtain the vehicle performance evaluation data information.
[0056] In this embodiment, if Figure 2 As shown, in step M2, the construction of an autoregressive moving average vehicle performance prediction model based on a first-order single integral sequence and the prediction of vehicle performance include:
[0057] M21. Based on the normalized data information of the vehicle sensors, vehicle controller and vehicle status, a first-order single integral sequence function f of the vehicle-related data is established.
[0058]
[0059] Among them, x1 is the processed data information of each on-board sensor, x2 is the data information of the vehicle controller, x3 is the data information of the vehicle status, α1, α2 and α3 are the serialization constant parameters of the vehicle-related data, and the vehicle-related data is serialized to obtain the first-order single integral sequence data information of the vehicle-related data;
[0060] M22. Input the first-order single integral sequence data information of the vehicle-related data into the autoregressive moving average vehicle performance prediction model for training and learning, and determine the vehicle performance prediction function F,
[0061]
[0062] Among them, y is the first-order single integral sequence data information of vehicle-related data, g is the mean function of the first-order single integral sequence of vehicle-related data, λ1, λ2 and λ3 are the prediction factors of vehicle performance, and the trained autoregressive moving average vehicle performance prediction model is obtained;
[0063] M23. Based on the trained autoregressive moving average vehicle performance prediction model, the first-order single integral sequence data information of the vehicle-related data is input to predict the performance of the vehicle, and obtain the predicted vehicle performance test data information.
[0064] In this embodiment, the mean function g of the first-order single integral sequence of the vehicle-related data is,
[0065]
[0066] Among them, y is the first-order single integral sequence data information of vehicle-related data.
[0067] In this embodiment, the constraints of the vehicle voltage ratio prediction factors λ1, λ2 and λ3 are:
[0068]
[0069] In this embodiment, the serialization constant parameters α1, α2 and α3 of the vehicle-related data are,
[0070]
[0071] Among them, x1 is the data information of each on-board sensor after processing, x2 is the data information of the vehicle controller, and x3 is the data information of the vehicle status.
[0072] In this embodiment, the method further includes:
[0073] M6. Based on the performance evaluation data information of the vehicle, a preset threshold is set. If the performance evaluation of the vehicle is greater than the preset threshold, the vehicle's autonomous driving performance meets the requirements. If the performance evaluation of the vehicle is less than the preset threshold, the vehicle's autonomous driving performance does not meet the requirements.
[0074] Example 2: Based on the method for testing automatic driving function scenario data in Example 1, the present invention is further illustrated and described below.
[0075] As shown in 1, a method for testing scene data of an autonomous driving function includes:
[0076] M1. The autonomous driving vehicle is driving in a preset test scenario, collecting data information of various on-board sensors, data information of the vehicle controller, and data information of the vehicle status in real time, and performing normalization processing to obtain normalized data information of various on-board sensors, vehicle controller, and vehicle status;
[0077] M2. Based on the normalized data information of the vehicle sensors, vehicle controller and vehicle status, a vehicle performance prediction model based on the autoregressive moving average of the first-order single integral sequence is constructed to predict the performance of the vehicle and obtain the predicted vehicle performance test data information;
[0078] M3. Based on the predicted vehicle performance test data information, construct a POT hyperparameter optimization model for vehicle performance, optimize the vehicle performance test data, and obtain optimized vehicle performance test data information;
[0079] M4. Based on the optimized vehicle performance test data information, establish a vehicle function scenario test evaluation function G, evaluate the vehicle performance, and obtain the vehicle performance evaluation data information.
[0080] In this embodiment, if Figure 3 As shown, in step M3, the POT hyperparameter optimization model of vehicle performance is constructed to optimize the vehicle performance test data, including:
[0081] M31. Based on the predicted vehicle performance test data information, construct a vehicle performance test data set and divide it into a test data set and a training data set;
[0082] M32. Input the training data set into the POT hyperparameter optimization model of vehicle performance for training and learning, and determine the vehicle performance test data optimization function W,
[0083]
[0084] Among them, z i is the i-th training data set, n is the sample size, ωi is the POT hyperparameter optimization factor of the i-th training data set, and the POT hyperparameter optimization model of the trained vehicle performance is obtained;
[0085] M33. Based on the trained POT hyperparameter optimization model of vehicle performance, input the test set, optimize and adjust the model parameters, and obtain the optimized POT hyperparameter optimization model of vehicle performance;
[0086] M34. Based on the POT hyperparameter optimization model of the optimized vehicle performance, the predicted vehicle performance test data information is input, the vehicle performance test data is optimized, and the optimized vehicle performance test data information is obtained.
[0087] In this embodiment, the vehicle function scenario test evaluation function G is:
[0088]
[0089] Among them, r is the performance test data information of the optimized vehicle, and θ1 and θ2 are the evaluation change factors of the vehicle function scenario test.
[0090] In this embodiment, if Figure 4 As shown, the present invention also provides a system for implementing any of the above-mentioned automatic driving function scenario data testing methods, the system comprising: a data acquisition module, for real-time acquisition of data information of various on-board sensors, data information of the vehicle controller, and data information of the vehicle status;
[0091] The data analysis module is used to construct an autoregressive moving average vehicle performance prediction model based on a first-order single integral sequence, predict the vehicle performance, obtain the predicted vehicle performance test data information, and construct a POT hyperparameter optimization model for vehicle performance to optimize the vehicle performance test data;
[0092] Test scenario management module, responsible for managing and configuring test scenarios;
[0093] The result output module is used to establish a vehicle functional scenario test evaluation function G, evaluate the performance of the vehicle, and obtain the vehicle performance evaluation data information.
[0094] Among them, data collection for following-constant speed following-stop scenarios: in the autonomous driving following scenario, all sensors, controllers and chassis data on the autonomous driving vehicle are collected.
[0095] Data analysis: The collected data is analyzed in real time according to the preset vehicle-following driving test scenario, including key indicators such as starting response time, acceleration and acceleration change rate, following vehicle speed, target vehicle speed, following vehicle speed fluctuation change rate, deceleration and deceleration change rate, and following-stop distance, which can comprehensively evaluate the vehicle's autonomous driving performance in this scenario.
[0096] Test scenario management: Users can customize test scenarios through this software system interface, such as setting different scenario test conditions, traffic flow, and performance indicators.
[0097] Result output: This system software generates detailed test reports and visual charts based on test data analysis and scenario performance indicators, including key performance indicators such as response time and path tracking accuracy, and gives evaluation suggestions. In the real-time test process, if the safety or functional scenario indicators cannot be achieved, a red font pop-up window will be displayed to remind you.
[0098] In this embodiment, the present invention provides a computer-readable storage medium, on which is stored a computer program programmed or configured to execute any one of the autonomous driving function scenario data testing methods described.
[0099] Any reference to memory, storage, database or other media used in the embodiments provided in the present application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM) and memory bus dynamic RAM (RDRAM).
[0100] In summary, the present invention can not only collect and analyze data in real time, ensuring the timeliness and accuracy of the test results, but also the entire test process does not require human intervention, thereby improving the test efficiency and accuracy.
[0101] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.
Claims
1. A method for testing scene data of an autonomous driving function, characterized in that: The method comprises: M1. The autonomous driving vehicle is driving in a preset test scenario, collecting data information of various on-board sensors, data information of the vehicle controller, and data information of the vehicle status in real time, and performing normalization processing to obtain normalized data information of various on-board sensors, vehicle controller, and vehicle status; M2. Based on the normalized data information of the vehicle sensors, vehicle controller and vehicle status, a vehicle performance prediction model based on the autoregressive moving average of the first-order single integral sequence is constructed to predict the performance of the vehicle and obtain the predicted vehicle performance test data information; M3. Based on the predicted vehicle performance test data information, construct a POT hyperparameter optimization model for vehicle performance, optimize the vehicle performance test data, and obtain optimized vehicle performance test data information; M4. Based on the optimized vehicle performance test data information, establish a vehicle function scenario test evaluation function G, evaluate the vehicle performance, and obtain the vehicle performance evaluation data information.
2. The method for testing scene data of an autonomous driving function according to claim 1, characterized in that: In step M2, the construction of an autoregressive moving average vehicle performance prediction model based on a first-order single integral sequence to predict vehicle performance includes: M21. Based on the normalized data information of the vehicle sensors, vehicle controller and vehicle status, a first-order single integral sequence function f of vehicle-related data is established. Among them, x1 is the processed data information of each on-board sensor, x2 is the data information of the vehicle controller, x3 is the data information of the vehicle status, α1, α2 and α3 are the serialization constant parameters of the vehicle-related data, and the vehicle-related data is serialized to obtain the first-order single integral sequence data information of the vehicle-related data; M22. Input the first-order single integral sequence data information of the vehicle-related data into the autoregressive moving average vehicle performance prediction model for training and learning, and determine the vehicle performance prediction function F, Among them, y is the first-order single integral sequence data information of vehicle-related data, g is the mean function of the first-order single integral sequence of vehicle-related data, λ1, λ2 and λ3 are the prediction factors of vehicle performance, and the trained autoregressive moving average vehicle performance prediction model is obtained; M23. Based on the trained autoregressive moving average vehicle performance prediction model, the first-order single integral sequence data information of the vehicle-related data is input to predict the performance of the vehicle, and obtain the predicted vehicle performance test data information.
3. The method for testing scene data of an autonomous driving function according to claim 2, characterized in that: The mean function g of the first-order single integral sequence of the vehicle-related data is, Among them, y is the first-order single integral sequence data information of vehicle-related data.
4. The method for testing scene data of an autonomous driving function according to claim 2, characterized in that: The constraints of the vehicle voltage ratio prediction factors λ1, λ2 and λ3 are:
5. The method for testing scene data of an autonomous driving function according to claim 2, characterized in that: The serialization constant parameters α1, α2 and α3 of the vehicle-related data are, Among them, x1 is the data information of each on-board sensor after processing, x2 is the data information of the vehicle controller, and x3 is the data information of the vehicle status.
6. The method for testing scene data of an autonomous driving function according to claim 1, characterized in that: The method further comprises: M6. Based on the performance evaluation data information of the vehicle, a preset threshold is set. If the performance evaluation of the vehicle is greater than the preset threshold, the vehicle's autonomous driving performance meets the requirements; if the performance evaluation of the vehicle is less than the preset threshold, the vehicle's autonomous driving performance does not meet the requirements.
7. The method for testing scene data of an automatic driving function according to claim 1, characterized in that: In step M3, the POT hyperparameter optimization model of vehicle performance is constructed to optimize the vehicle performance test data, including: M31. Based on the predicted vehicle performance test data information, construct a vehicle performance test data set and divide it into a test data set and a training data set; M32. Input the training data set into the POT hyperparameter optimization model of vehicle performance for training and learning, and determine the vehicle performance test data optimization function W, Among them, z i is the i-th training data set, n is the sample size, ω i is the POT hyperparameter optimization factor of the i-th training data set, and the POT hyperparameter optimization model of the trained vehicle performance is obtained; M33. Based on the trained POT hyperparameter optimization model of vehicle performance, input the test set, optimize and adjust the model parameters, and obtain the optimized POT hyperparameter optimization model of vehicle performance; M34. Based on the POT hyperparameter optimization model of the optimized vehicle performance, the predicted vehicle performance test data information is input, the vehicle performance test data is optimized, and the optimized vehicle performance test data information is obtained.
8. The method for testing scene data of an autonomous driving function according to claim 1, characterized in that: The vehicle function scenario test evaluation function G is: Among them, r is the performance test data information of the optimized vehicle, and θ1 and θ2 are the evaluation change factors of the vehicle function scenario test.
9. A system for implementing the method for testing scene data of an autonomous driving function according to any one of claims 1 to 8, characterized in that: The system comprises: Data acquisition module, used to collect data information of various on-board sensors, data information of vehicle controllers and data information of vehicle status in real time; The data analysis module is used to construct an autoregressive moving average vehicle performance prediction model based on a first-order single integral sequence, predict the vehicle performance, obtain the predicted vehicle performance test data information, and construct a POT hyperparameter optimization model for vehicle performance to optimize the vehicle performance test data; Test scenario management module, responsible for managing and configuring test scenarios; The result output module is used to establish a vehicle functional scenario test evaluation function G, evaluate the performance of the vehicle, and obtain the vehicle performance evaluation data information.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program that is programmed or configured to execute the automatic driving function scenario data testing method as described in any one of claims 1 to 8.
