Vehicle ADAS road test data processing method

Abstract

The invention discloses a whole vehicle ADAS road test data processing method, which comprises the following steps: based on ADAS test content, carrying out test single item classification on road test, and respectively carrying out data statistics on the test single items to obtain classified first elements; classifying the test scenes, respectively performing data statistics on the test scenes to obtain classified second elements, acquiring time characteristics of the first elements and the second elements in the test period, and associating the test scenes and the test single items based on the same time characteristics in the first elements and the second elements for data processing to obtain statistical items; according to the invention, the data of the ADAS road test of the whole vehicle is efficiently processed, and the indexes and curves for evaluating the ADAS function of the vehicle are obtained; a data processing framework is constructed to be used for dynamically adjusting and changing test scenes of different projects, so that the accuracy and efficiency of data processing are improved; according to the method, the data is associated with the test scene and the test content by relying on the test log establishment time and the test content mapping, so that the universality and the flexibility of data processing logic are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle road testing, and in particular to a method for processing whole vehicle ADAS road testing data. Background Art

[0002] Advanced Driver Assistance System (ADAS) is an important part of the wave of automotive intelligent technology. It is a technology for the transition from traditional cars to fully autonomous cars. ADAS provides drivers with a series of safety and convenience functions by integrating multiple sensors (such as radar, cameras, ultrasonic sensors, etc.) and control systems. These functions are designed to prevent traffic accidents, improve driving comfort, and reduce the burden on drivers by assisting drivers in perception, judgment, and operation.

[0003] With the improvement of intelligence level, the safety of ADAS system is also facing new challenges. This requires automobile manufacturers and suppliers to consider not only the risk of system functional failure, but also the risk of failure caused by system function, performance limitations or driver misoperation when designing and implementing ADAS system to ensure the safety and reliability of the system. ADAS road testing is a key link to ensure that ADAS system can achieve the expected performance, reliability and safety in the actual traffic environment. This test is usually carried out in a real vehicle environment to verify the performance of the system under various road and traffic conditions.

[0004] In real road tests, testers will simulate various driving scenarios of users, including normal driving, emergency situations, driver errors, etc., to evaluate the response time and accuracy of the system in these situations. At the same time, the human-computer interaction interface of the system will be evaluated to ensure that the driver can easily understand the warnings and instructions of the system and interact with it smoothly. In addition, real road tests will also include the performance of the system under extreme conditions, such as high traffic density, bad weather or complex road conditions, as well as the safety and reliability when the sensor is interfered or fails. During the real road test, a large amount of data will be collected, including the vehicle's dynamics data, the system's perception, decision-making, execution data and environmental data, which will be analyzed to evaluate the performance of the system and identify any potential problems.

[0005] The current vehicle ADAS road test lacks data processing methods and systems, has low efficiency in processing large amounts of data, lacks automated processing capabilities, has scattered results that are not adaptable, and is time-consuming. There is also a lack of corresponding calculation methods for some indicator values ​​used for ADAS functions, and a lack of data support. The above problems need to be solved urgently. Summary of the invention

[0006] The object of the present invention is to provide a method for processing vehicle ADAS road test data to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a method for processing vehicle ADAS road test data, comprising:

[0008] Based on the ADAS test content, the road test is classified into test items, and data statistics are performed on the test items to obtain the first element of classification, wherein the test items are used to extract the single functions that need to be counted and generate statistical items;

[0009] The test scenarios are classified, and data statistics are performed on each test scenario to obtain the second element of the classification. During the test cycle, the time characteristics of the first element and the second element are obtained, and based on the same time characteristics in the first element and the second element, the test scenarios and test items are associated to perform data processing to obtain statistical items.

[0010] Furthermore, based on the ADAS test content, the road test is classified into test items, and the first element of the classification is obtained by performing data statistics on the test items respectively, which also includes:

[0011] Get data signals, set the algorithm strategy for the test items, and obtain statistics for the single function based on the data signals and the algorithm strategy for the test items.

[0012] Further, the test scenarios are classified, data statistics are performed on the test scenarios respectively to obtain the second element of the classification, the time characteristics of the first element and the second element are obtained within the test cycle, and the test scenarios and test items are associated with each other based on the same time characteristics in the first element and the second element to perform data processing to obtain statistical items, and further includes:

[0013] The acquired data signal format is converted into a processable format, the sampling frequency is unified, and the name of the data signal is unified.

[0014] Furthermore, in the test cycle, the time characteristics of the first element and the second element are obtained, and based on the same time characteristics in the first element and the second element, the test scenario and the test item are associated to perform data processing to obtain statistical items, which also includes:

[0015] Obtaining a log, wherein the log includes a time characteristic of a first element and a second element;

[0016] Associating the data signal after conversion into a processable format with the log;

[0017] Format the log and calibrate the test cycle;

[0018] Set time tags, classify and label the classification of test items and test scenarios in each test cycle, and associate the data signal with the first element and the second element through the time tags.

[0019] Further, obtaining data signals, setting an algorithm strategy for a single test item, and obtaining statistics of a single function according to the data signals and the algorithm strategy for the single test item also include:

[0020] The acquired data signals include SACC / ACC function status signal, SACC / ACC function takeover request signal, ACC deceleration request signal, and target object status signal;

[0021] Statistics of single functions obtained based on data signals and algorithm strategies of single tests include mileage in different speed ranges and function usage rates in different speed ranges per 100 kilometers, statistics on takeover requests in each speed range, statistics on ACC stop-start times, statistics on ACC exits due to functional abnormalities, statistics on ACC exits due to driver braking for safety reasons, statistics on SACC exits due to functional abnormalities, distribution statistics on the longitudinal distance between the vehicle and the target object in front when the vehicle starts to decelerate at different speeds, distribution statistics on acceleration response time after the vehicle recognizes that the vehicle in front has left at different speeds, distribution statistics on average lateral acceleration when passing through a curve at different speeds, and distribution statistics on average yaw acceleration when passing through a curve at different speeds.

[0022] Furthermore, the mileage in different speed ranges and the function usage rate statistics in different speed ranges per 100 kilometers also include:

[0023] According to the SACC / ACC function status signal, all activated data segments are intercepted;

[0024] These data segments are divided into segments according to the speed range, and the trapezoidal integration method is used for each segment to obtain the mileage in different speed ranges. The calculation formula is:

[0025] ∫abf(x)dx≈2n(ba)[f(x0)+2f(x1)+2f(x2)+...+2f(x n-1 )+f(x n )]——Formula (1);

[0026] Where a and b are the lower and upper limits of the integral interval, respectively; n is the number of subintervals into which the interval [a, b] is divided; x n For each subinterval endpoint, x0 = ax0 = a and x n =bx n =b; f(x) is the integrand;

[0027] At the same time, the function usage rate per 100 kilometers is obtained; the usage time of different speed ranges can be calculated according to the sampling frequency, and the calculation formula is:

[0028]

[0029] Where n is the total number of data points and f is the frequency of each data point.

[0030] Furthermore, the longitudinal distance distribution statistics between the vehicle and the target object ahead when the vehicle starts to decelerate at different vehicle speeds also include:

[0031] intercepting all data segments of ACC controlling vehicle deceleration according to the ACC deceleration request signal;

[0032] The longitudinal distance signal value between the vehicle and the target object in front at the starting moment of each data segment is the longitudinal distance between the vehicle and the target object in front when the vehicle starts to decelerate. The heat map is drawn with speed and longitudinal distance as coordinates.

[0033] Furthermore, the distribution statistics of acceleration response time after the vehicle recognizes the front vehicle leaving at different vehicle speeds also include:

[0034] According to the target state signal, all data segments without the target are intercepted;

[0035] The starting time is when the front vehicle leaves. From this time to when the vehicle starts to accelerate is the acceleration response time. The heat map is drawn with speed and acceleration response time as coordinates.

[0036] Furthermore, the average lateral acceleration distribution statistics when passing through a curve at different vehicle speeds also include:

[0037] First, all the data segments where SACC is activated are intercepted according to the SACC status signal to determine whether the vehicle is turning:

[0038] Perform Gaussian filtering on the steering wheel angle signal to remove noise;

[0039] Set the threshold and set the fluctuation data points near 0 to 0;

[0040] Take out all the data segments of the steering wheel angle signal that are not 0, and set the threshold values ​​of the angle and duration;

[0041] The data that meets the turning angle threshold size and duration are taken out, that is, the data when passing the curve;

[0042] Calculate the average lateral acceleration in this segment of data, and use the lateral acceleration and speed of this segment of data as coordinates to draw a heat map of the average lateral acceleration distribution when passing through a curve at different vehicle speeds. The calculation formula is:

[0043]

[0044] Where: SmoothedValue is the smoothed value; w(i) is the Gaussian weight function, which represents the weight of the original data point valuei when calculating the average value; valuei is the i-th data point in the original data sequence.

[0045] Furthermore, the average yaw acceleration distribution statistics when passing through a curve at different vehicle speeds also include:

[0046] When passing through a curve, all data segments where SACC is activated are intercepted according to the SACC status signal;

[0047] The average yaw acceleration is calculated based on the yaw angular velocity of the data segment, and the formula is:

[0048] See formula (1-4), and draw the thermal map with angular velocity and velocity as coordinates.

[0049]

[0050] Where: θ i (t) represents the yaw angle at time t, in degrees; n is the number of discrete points; Δt i is the time interval of the ith discrete point.

[0051] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a whole vehicle ADAS road test data processing method to efficiently process the whole vehicle ADAS road test data to obtain indicators and curves for evaluating the vehicle ADAS function; on the one hand, the whole vehicle ADAS road test method and scene characteristics are sorted out to obtain data classification and statistical characteristics to construct a data processing framework to adapt to the dynamically adjusted and changing test scenes of different projects, thereby obtaining formatted output data indicators and icons; the data volume of the whole vehicle ADAS road test is huge, and the entire test process involves multiple cities or regions, multiple test road scenes and multiple test contents. Tests of different road scenes and test contents are often interspersed and dynamically adjusted. It is necessary to classify the data when processing the data, but the previous reliance on manual data subpackaging is inefficient. For this problem, we rely on the mapping of test log creation time and test content to associate data with test cities, test scenarios, and test content. When processing a single data file, we can directly classify it based on the label, shorten the data processing link, and reduce manual processing procedures. On the other hand, we convert the data into parquet format for storage, unify the frequency of each data signal to 10Hz, and fill in the default value. We divide the speed range and calculate the mileage based on the vehicle speed signal. We calculate the cumulative calculation of the mileage / frequency of function use, exit, and takeover frequency based on each function status signal, calculate the speed and acceleration distribution of the vehicle when the function is activated, and calculate the distance distribution, acceleration and deceleration response time distribution, and target type distribution based on the target signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 The present invention is a flowchart of a method for processing vehicle ADAS road test data in an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] In the description of the present invention, unless otherwise specified, "plurality" means two or more than two; the orientations or positional relationships indicated by the terms "upper", "lower", "left", "right", "inner", "outer", "front end", "rear end", "head", "tail", etc. are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0055] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0056] Please refer to the attached drawings of the specification, the present invention provides a technical solution: a method for processing vehicle ADAS road test data, comprising the following steps:

[0057] S102, based on the ADAS test content, the road test is classified into test items, and data statistics are performed on the test items to obtain the first element of the classification, wherein the test items are used to extract the single functions that need to be counted and generate statistical items;

[0058] It can be understood that the present embodiment first deconstructs the ADAS road test, including the ADAS test content and the corresponding algorithm strategy, test process, test location, and statistical items; ADAS is subdivided into multiple test items according to the test content, which is the first element of data statistics classification, and the single function that needs to be counted is extracted; ADAS road test content generally includes separate test items for each function, and the test items are based on the test mileage as the standard. The statistical items are random, so they are test items with data processing requirements, which can be divided into SSACC / SACC function driving tests, AEB special tests, parking special tests and other special tests, and different test items have different processing logics; the multi-dimensional classification of ADAS road test content and algorithm strategies, by subdividing the ADAS road test into multiple test items, combined with the classification of test locations or scenes, creates a multi-dimensional data statistical classification system, which can more accurately manage and analyze test data.

[0059] Specifically, based on the ADAS test content, the road test is classified into test items, and the first element of the classification is obtained by performing data statistics on the test items respectively, which also includes:

[0060] Get data signals, set the algorithm strategy for the test items, and obtain statistics for the single function based on the data signals and the algorithm strategy for the test items.

[0061] Among them, the original data signal generally contains several signals, each signal is a discrete point of a specific sampling frequency. Data processing is to select a specific signal within a specific time period for statistical calculation, so it is necessary to clarify the two elements of the time period and the required signal.

[0062] S104. Classify the test scenarios, perform data statistics on the test scenarios respectively to obtain the second element of the classification, obtain the time characteristics of the first element and the second element within the test cycle, and associate the test scenarios and test items based on the same time characteristics in the first element and the second element to perform data processing to obtain statistical items.

[0063] It is understandable that the test scenario of ADAS road test includes the test location and the test scenario road conditions. For example, the ADAS road test locations include Chongqing, Guangzhou, Changchun, Hulunbuir and cities along the way, and the test scenario road conditions include urban areas, highways, expressways, etc. During the test process, within a certain period of time, the test location, test road conditions and test items are associated through this time period, so the statistical results will include the test location, test road conditions, and the content of the test items; based on the time characteristic analysis of the test process: analyze the process management and operation methods of each region and each function test in the test cycle to obtain the time characteristics of each classification element. These time characteristics serve as logical conditions for building a data processing framework, which improves the accuracy and efficiency of data processing.

[0064] Specifically, the test scenarios are classified, data statistics are performed on the test scenarios respectively to obtain the second element of the classification, the time characteristics of the first element and the second element are obtained within the test cycle, and the test scenarios and test items are associated with each other based on the same time characteristics in the first element and the second element to perform data processing to obtain statistical items, and also includes:

[0065] The acquired data signal format is converted into a processable format, the sampling frequency is unified, and the name of the data signal is unified.

[0066] Among them, in this embodiment, it is necessary to unify the data format, including data format conversion, signal sampling frequency unification, and signal name unification, in order to adapt the data processing logic to different vehicle model items; the statistical items are generally the target data statistical results of different ADAS functions required by the test standards, and the signals required for calculation are given at the same time, based on which the specific data processing logic of each statistical item can be analyzed; the test process refers to the process management and operation methods of analyzing the tests of various regions and functions in the test cycle, that is, the algorithm strategy, to obtain the time characteristics of each classification element as the logical condition for building the data processing framework; the statistical items are the target result outputs, and the statistical items of each function and the data signals and calculation methods required for each statistical item are determined, which is the basic logic for statistical calculation of the classified data; in this embodiment, it is preferred to convert the data into parquet format for storage, unify the frequencies of all data signals to 10Hz, and fill in the default values.

[0067] Specifically, in the test cycle, the time characteristics of the first element and the second element are obtained, and based on the same time characteristics in the first element and the second element, the test scenario and the test item are associated to perform data processing to obtain statistical items, and further includes:

[0068] Obtaining a log, wherein the log includes a time characteristic of a first element and a second element;

[0069] Associating the data signal after conversion into a processable format with the log;

[0070] Format the log and calibrate the test cycle;

[0071] Set time tags, classify and label the classification of test items and test scenarios in each test cycle, and associate the data signal with the first element and the second element through the time tags.

[0072] Among them, log formatting is to format and record the test log content based on the ADAS road test method and scenario deconstruction. The entire test cycle is divided into days, and then the test content in each unit time period is classified and labeled, and the data entity is associated with each classification element through the time label; ADAS road test process analysis is to obtain the time characteristics of the test items, test locations, and scenarios in the entire test cycle, and obtain the time period elements for data calculation. First, the analysis is carried out with a single day as the time unit. A single day can be divided into multiple shifts, mainly single shift, double shift, triple shift and temporary shift. The arrangement of test items is uncertain, that is, the plan is dynamically adjusted, the schedule and shift arrangement of each project are interspersed, and there are different test contents in a single day, so the time period of the test project needs to be specific to the time period within a single day. The schedule of the test site is generally continuous, that is, the test of a single city is completed before the next city; sorting out the above content can obtain the two major elements of data processing, the time period of statistical data and the data signal; the time period needs to be divided, and the granularity of the division is the time period within a single day, and the basis for the division is the test project and location. Based on this, the data processing framework is built: first, the log record content is formatted, including the elements of the sample vehicle number, date, shift, test location, and test content. The time period of each sample vehicle can be divided into categories, and the time periods of the same category can be merged. Then, the data files are processed one by one, the data format is converted, and it is parsed into a format that can be processed programmatically, and the signal sampling frequency and signal name are unified. The data timestamp is obtained, and the data is associated with the time log, that is, the test content and location to which it belongs are divided according to the time interval where the timestamp is located, and then the corresponding calculation processing logic is selected. Finally, the data of each category is aggregated to form the final result; log formatting and data association technology: based on the ADAS road test method and scenario deconstruction, the test log content is formatted and recorded, and the data entity is associated with each classification element through the time tag. This technology helps to ensure the consistency and traceability of the data. Therefore, the construction of the data processing framework includes steps such as raw data format conversion, log formatting processing, data and log association, and data statistical algorithm, forming a complete data processing system. The framework can adapt to different vehicle projects and improve the versatility and flexibility of data processing logic.

[0073] Specifically, the data signal is obtained, the algorithm strategy of the test item is set, and the statistics of the single function are obtained according to the data signal and the algorithm strategy of the test item, which also includes:

[0074] The acquired data signals include SACC / ACC function status signal, SACC / ACC function takeover request signal, ACC deceleration request signal, and target object status signal;

[0075] Statistics of single functions obtained based on data signals and algorithm strategies of single tests include mileage in different speed ranges and function usage rates in different speed ranges per 100 kilometers, statistics on takeover requests in each speed range, statistics on ACC stop-start times, statistics on ACC exits due to functional abnormalities, statistics on ACC exits due to driver braking for safety reasons, statistics on SACC exits due to functional abnormalities, distribution statistics on the longitudinal distance between the vehicle and the target object in front when the vehicle starts to decelerate at different speeds, distribution statistics on acceleration response time after the vehicle recognizes that the vehicle in front has left at different speeds, distribution statistics on average lateral acceleration when passing through a curve at different speeds, and distribution statistics on average yaw acceleration when passing through a curve at different speeds.

[0076] Among them, the statistical items are clarified and formulated with data processing logic, and the statistical items of each function as well as the data signals and calculation strategies required for each statistical item are clarified, providing the basic logic of statistical calculation for the classified data; the speed intervals are divided and the mileage is calculated based on the vehicle speed signal, and the mileage / frequency of function use, exit, and takeover frequency are accumulated according to the functional status signals. The speed and acceleration distribution of the vehicle when the function is activated is calculated, and the distance distribution to the target object, the acceleration and deceleration response time distribution, and the target object type distribution are calculated according to the target object signal.

[0077] Specifically, the mileage in different speed ranges and the function usage rate statistics in different speed ranges per 100 kilometers also include:

[0078] According to the SACC / ACC function status signal, all activated data segments are intercepted;

[0079] These data segments are divided into segments according to the speed range, and the trapezoidal integration method is used for each segment to obtain the mileage in different speed ranges. The calculation formula is:

[0080] ∫abf(x)dx≈2n(ba)[f(x0)+2f(x1)+2f(x2)+...+2f(x n-1 )+f(x n )]——Formula (1);

[0081] Where a and b are the lower and upper limits of the integral interval, respectively; n is the number of subintervals into which the interval [a, b] is divided; x nFor each subinterval endpoint, x0 = ax0 = a and x n =bx n =b; f(x) is the integrand;

[0082] At the same time, the function usage rate per 100 kilometers is obtained; the usage time of different speed ranges can be calculated according to the sampling frequency, and the calculation formula is:

[0083]

[0084] Where n is the total number of data points and f is the frequency of each data point.

[0085] Specifically, the longitudinal distance distribution statistics between the vehicle and the target object in front when the vehicle starts to decelerate at different vehicle speeds also include:

[0086] intercepting all data segments of ACC controlling vehicle deceleration according to the ACC deceleration request signal;

[0087] The longitudinal distance signal value between the vehicle and the target object in front at the starting moment of each data segment is the longitudinal distance between the vehicle and the target object in front when the vehicle starts to decelerate. The heat map is drawn with speed and longitudinal distance as coordinates.

[0088] Specifically, the distribution statistics of the acceleration response time after the vehicle recognizes the front vehicle leaving at different vehicle speeds also include:

[0089] According to the target state signal, all data segments without the target are intercepted;

[0090] The starting time is when the front vehicle leaves. From this time to when the vehicle starts to accelerate is the acceleration response time. The heat map is drawn with speed and acceleration response time as coordinates.

[0091] Specifically, the average lateral acceleration distribution statistics when passing through a curve at different vehicle speeds also include:

[0092] First, all the data segments where SACC is activated are intercepted according to the SACC status signal to determine whether the vehicle is turning:

[0093] Perform Gaussian filtering on the steering wheel angle signal to remove noise;

[0094] Set the threshold and set the fluctuation data points near 0 to 0;

[0095] Take out all the data segments of the steering wheel angle signal that are not 0, and set the threshold values ​​of the angle and duration;

[0096] The data that meets the turning angle threshold size and duration are taken out, that is, the data when passing the curve;

[0097] Calculate the average lateral acceleration in this segment of data, and use the lateral acceleration and speed of this segment of data as coordinates to draw a heat map of the average lateral acceleration distribution when passing through a curve at different vehicle speeds. The calculation formula is:

[0098]

[0099] Where: SmoothedValue is the smoothed value; w(i) is the Gaussian weight function, which represents the weight of the original data point valuei when calculating the average value; valuei is the i-th data point in the original data sequence.

[0100] Specifically, the average yaw acceleration distribution statistics when passing a curve at different vehicle speeds also include:

[0101] When passing through a curve, all data segments where SACC is activated are intercepted according to the SACC status signal;

[0102] The average yaw acceleration is calculated based on the yaw angular velocity of the data segment, and the formula is:

[0103] See formula (1-4), and draw the thermal map with angular velocity and velocity as coordinates.

[0104]

[0105] Where: θ i (t) represents the yaw angle at time t, in degrees; n is the number of discrete points; Δt i is the time interval of the ith discrete point.

[0106] Preferably, according to the SACC / ACC function request takeover signal, a data segment of the request status is intercepted, each segment of data is a request takeover, and the number of request takeovers in each speed interval is counted according to the speed when each takeover request signal is sent.

[0107] Preferably, the number of ACC follow-stop-to-start statistics: first, all ACC function activated data segments are intercepted according to the ACC function status signal, and all speed segments with a speed of 0 km / h are intercepted in the data segment, that is, ACC controls the vehicle to be stationary, and it is determined whether the end time of each segment is at the last time point of the ACC activation state, and whether the start time of each segment is at the first time point of the ACC activation state. If neither is located, it means that ACC controls the vehicle to follow and start once. The distance between the vehicle and the front vehicle at the end time of each data segment is the distance between the vehicle and the front vehicle when the vehicle starts following the front vehicle.

[0108] Preferably, the number of ACC exits caused by functional abnormalities is counted: first, all data segments of ACC function activation are intercepted according to the ACC function status signal, and the end time of each segment is the ACC function exit time. It is determined whether the brake pedal signal is triggered, the takeover request signal is triggered, and the ACC function switch signal is triggered at this moment. If none of them are triggered, it means that the ACC function has exited abnormally once.

[0109] Preferably, the number of times ACC is exited due to driver braking for safety reasons is counted: first, all data segments activated by the ACC function are intercepted according to the ACC function status signal, and the end time of each segment is the ACC function exit time. At this moment, if the brake pedal signal is triggered, the request to take over signal is not triggered, and the ACC function switch signal is not triggered, if the conditions are met, the ACC function is abnormally exited once due to driver braking for safety reasons.

[0110] Preferably, the number of SACC exits due to functional abnormalities is counted: first, all data segments of SACC function activation are intercepted according to the SACC function status signal, and the end time of each segment is the SACC function exit time. It is determined whether the brake pedal signal is triggered, whether the takeover request signal is triggered, whether the SACC function switch signal is triggered, and whether the steering wheel hand torque exceeds the function exit threshold. If none of the above is true, it means that the SACC function has exited abnormally once.

[0111] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A vehicle ADAS road test data processing method, characterized in that: include: Based on the ADAS test content, the road test is classified into test items, and data statistics are performed on the test items to obtain the first element of classification, wherein the test items are used to extract the single functions that need to be counted and generate statistical items; The test scenarios are classified, and data statistics are performed on each test scenario to obtain the second factor of classification. During the test cycle, the time characteristics of the first factor and the second factor are obtained, and based on the same time characteristics in the first factor and the second factor, the test scenarios and test items are associated to perform data processing to obtain statistical items.

2. A vehicle ADAS road test data processing method according to claim 1, characterized in that: Based on the ADAS test content, the road test is classified into test items, and the data statistics of the test items are respectively obtained to obtain the first element of classification, which also includes: Get data signals, set the algorithm strategy for the test items, and obtain statistics for the single function based on the data signals and the algorithm strategy for the test items.

3. A vehicle ADAS road test data processing method according to claim 2, characterized in that: The test scenarios are classified, and data statistics are performed on the test scenarios to obtain the second element of the classification. During the test cycle, the time characteristics of the first element and the second element are obtained, and the test scenarios and test items are associated with each other based on the same time characteristics in the first element and the second element to perform data processing to obtain statistical items, and also includes: The acquired data signal format is converted into a processable format, the sampling frequency is unified, and the name of the data signal is unified.

4. A vehicle ADAS road test data processing method according to claim 3, characterized in that: In the test cycle, the time characteristics of the first element and the second element are obtained, and based on the same time characteristics in the first element and the second element, the test scenario and the test item are associated to perform data processing to obtain statistical items, and further includes: Obtaining a log, wherein the log includes a time characteristic of a first element and a second element; Associating the data signal after conversion into a processable format with the log; Format the log and calibrate the test cycle; Set time tags, classify and label the classification of test items and test scenarios in each test cycle, and associate the data signal with the first element and the second element through the time tags.

5. A vehicle ADAS road test data processing method according to claim 2, characterized in that: Get data signals, set the algorithm strategy for the test items, and obtain the statistics of the single function according to the data signals and the algorithm strategy for the test items, including: The acquired data signals include SACC / ACC function status signal, SACC / ACC function takeover request signal, ACC deceleration request signal, and target object status signal; Statistics of single functions obtained based on data signals and algorithm strategies of single tests include mileage in different speed ranges and function usage rates in different speed ranges per 100 kilometers, statistics on takeover requests in each speed range, statistics on ACC stop-start times, statistics on ACC exits due to functional abnormalities, statistics on ACC exits due to driver braking for safety reasons, statistics on SACC exits due to functional abnormalities, distribution statistics on the longitudinal distance between the vehicle and the target object in front when the vehicle starts to decelerate at different speeds, distribution statistics on acceleration response time after the vehicle recognizes that the vehicle in front has left at different speeds, distribution statistics on average lateral acceleration when passing through a curve at different speeds, and distribution statistics on average yaw acceleration when passing through a curve at different speeds.

6. A vehicle ADAS road test data processing method according to claim 5, characterized in that: Mileage in different speed ranges and function usage statistics in different speed ranges per 100 kilometers, including: According to the SACC / ACC function status signal, all activated data segments are intercepted; These data segments are divided into segments according to the speed range, and the trapezoidal integration method is used for each segment to obtain the mileage in different speed ranges. The calculation formula is: ∫abf(x)dx≈2n(ba)[f(x0)+2f(x1)+2f(x2)+…+2f(x n-1 )+f(x n )]——in (1): Where a and b are the lower and upper limits of the integral interval, respectively; n is the number of subintervals into which the interval [a, b] is divided; x n For each subinterval endpoint, x0 = ax0 = a and x n =bx n =b; f(x) is the integrand; At the same time, the function usage rate per 100 kilometers is obtained; the usage time of different speed ranges can be calculated according to the sampling frequency, and the calculation formula is: Where n is the total number of data points and f is the frequency of each data point.

7. A vehicle ADAS road test data processing method according to claim 5, characterized in that: The longitudinal distance distribution statistics between the vehicle and the target object in front when the vehicle starts to decelerate at different speeds also include: intercepting all data segments of ACC controlling vehicle deceleration according to the ACC deceleration request signal; The longitudinal distance signal value between the vehicle and the target object in front at the starting moment of each data segment is the longitudinal distance between the vehicle and the target object in front when the vehicle starts to decelerate. The heat map is drawn with speed and longitudinal distance as coordinates.

8. A vehicle ADAS road test data processing method according to claim 5, characterized in that: The distribution statistics of the acceleration response time after the vehicle recognizes the front vehicle leaving at different vehicle speeds also include: According to the target state signal, all data segments without the target are intercepted; The starting time is when the front vehicle leaves. From this time to when the vehicle starts to accelerate is the acceleration response time. The heat map is drawn with speed and acceleration response time as coordinates.

9. A vehicle ADAS road test data processing method according to claim 5, characterized in that: Statistics of average lateral acceleration distribution when passing a curve at different vehicle speeds, including: First, all the data segments where SACC is activated are intercepted according to the SACC status signal to determine whether the vehicle is turning: Perform Gaussian filtering on the steering wheel angle signal to remove noise; Set the threshold and set the fluctuation data points near 0 to 0; Take out all the data segments of the steering wheel angle signal that are not 0, and set the threshold values ​​of the angle and duration; The data that meets the turning angle threshold size and duration are taken out, that is, the data when passing the curve; Calculate the average lateral acceleration in this segment of data, and use the lateral acceleration and speed of this segment of data as coordinates to draw a heat map of the average lateral acceleration distribution when passing through a curve at different vehicle speeds. The calculation formula is: Where: SmoothedValue is the smoothed value; w(i) is the Gaussian weight function, which represents the weight of the original data point valuei when calculating the average value; valuei is the i-th data point in the original data sequence.

10. A vehicle ADAS road test data processing method according to claim 5, characterized in that: Statistics of average yaw acceleration distribution when passing a curve at different vehicle speeds, including: When passing through a curve, all data segments where SACC is activated are intercepted according to the SACC status signal; The average yaw acceleration is calculated based on the yaw angular velocity of the data segment, and the formula is: See formula (1-4), and draw the thermal map with angular velocity and velocity as coordinates. Where: θ i (t) represents the yaw angle at time t, in degrees; n is the number of discrete points; Δt i is the time interval of the ith discrete point.

Images