Method and system for detecting display mileage error of pure electric vehicle instrument

Abstract

The invention provides a pure electric vehicle instrument display mileage error detection method and a pure electric vehicle instrument display mileage error detection system, relates to the technical field of traffic transportation, and can quantify instrument display mileage error results and compare data under different vehicle types and different working conditions to form a unified instrument display mileage evaluation system. The method comprises the following steps: determining a target test mode according to a test environment temperature and a driving range length of a vehicle, and then testing the vehicle; testing data composed of first mileage data and second mileage data displayed by an instrument of the vehicle are collected, the second mileage data comprise mileage data of the rotating hub equipment, and under the condition that the target testing mode is the first shortening method or the second shortening method, the first shortening method and the second shortening method are used for testing the mileage data of the rotating hub equipment; the second mileage data comprises equivalent mileage data determined according to the discharge capacity of the constant-speed section and the average power consumption of the circulation section; and according to the first mileage data and the second mileage data, determining a whole-course error, each segment error and each single-point error, and evaluating the instrument display error of the vehicle by combining a corresponding threshold value.

Description

Technical Field

[0001] The present application relates to the field of transportation technology, and in particular to a method and system for detecting mileage errors displayed on an instrument panel of a pure electric vehicle. Background Art

[0002] Pure electric vehicles are vehicles whose power energy comes from batteries. With the improvement of global environmental awareness and the development of new energy vehicle technology, pure electric vehicles have gradually become one of the mainstreams in the automobile market.

[0003] However, compared to fuel-powered vehicles, pure electric vehicles face more complex technical challenges in predicting range, especially in extreme climates, such as low and high temperatures. A vehicle's battery performance is affected by temperature, and battery consumption from air conditioning and other factors can affect the actual range. Under these conditions, the range displayed on the instrument panel of a pure electric vehicle often differs significantly from the actual drivable distance. This is especially true when the battery charge is low and the displayed range is inflated. This not only causes driver anxiety but also poses safety risks such as sudden stalling and the inability to continue driving due to insufficient battery power, severely hindering the widespread adoption of electric vehicles in cold regions. However, existing evaluation systems for instrument panel range display lack uniformity. Different testing agencies and varying test conditions can lead to incomparable test results, making it difficult to fairly and objectively evaluate the range prediction performance of different vehicle models.

[0004] Therefore, there is an urgent need for an error detection method that can quantify the mileage error results displayed by instruments under different vehicle models and different working conditions and make the data comparable. Summary of the Invention

[0005] The present application provides a method and system for detecting mileage error displayed on the instrument panel of a pure electric vehicle, which can quantify the error results of mileage displayed on the instrument panel under different vehicle models and different operating conditions, make the data comparable, and form a unified evaluation system for mileage displayed on the instrument panel.

[0006] In a first aspect, a method for detecting mileage error displayed on an instrument panel of a pure electric vehicle is provided, comprising: The target test mode is determined according to the test environment temperature and the cruising range length of the vehicle, wherein, in a normal temperature test environment, the test mode adopted by the vehicle with a cruising range of less than or equal to 8 CLTC-P cycle ranges is a continuous working condition method, the continuous working condition method is a test mode in which the test is continuously performed according to the CLTC-P cycle, the test mode adopted by the vehicle with a cruising range of more than 8 CLTC-P cycle ranges is a first shortening method, the first shortening method is a test mode in which the test is performed according to a first test structure, the working condition segments included in the first test structure are in turn a cycle segment, a constant speed segment, a cycle segment and a constant speed segment, the cycle segment of the first test structure includes 2 CLTC-P cycles, in a low temperature test environment or a high temperature test environment, the test mode adopted by the vehicle with a cruising range of less than or equal to 8 CLTC-P cycle ranges is the continuous working condition method, the test mode adopted by the vehicle with a cruising range of more than 8 CLTC-P cycle ranges is a second shortening method, the second shortening method is a test mode in which the test is performed according to a second test structure, the working condition segments included in the second test structure are in turn a cycle segment and a constant speed segment, the cycle segment of the second test structure includes 6 CLTC-P cycles; The vehicle is tested according to the target test mode, and test data of the vehicle is collected, the test data including first mileage data and second mileage data displayed by an instrument, the second mileage data including mileage data of a rotating hub device, wherein, in the case that the target test mode is the first shortening method or the second shortening method, the second mileage data further includes equivalent mileage data determined according to a discharge amount of the constant speed segment and an average power consumption of the cycle segment; The full-range error, the segment errors and the single-point errors are determined according to the first mileage data and the second mileage data, the full-range error is calculated based on sampling points in the full test period, the segment errors are calculated based on sampling points in a high power segment, a medium power segment or a low power segment divided according to the battery state of charge, and the single-point errors are calculated based on a single sampling point; The instrument display error of the vehicle is evaluated according to the full-range error, a full-range error threshold, the segment errors, a segment error threshold, the single-point errors and a single-point error threshold.

[0007] In a feasible design, in the case that the target test mode is the first shortening method or the second shortening method, the step of collecting the equivalent mileage data includes: After excluding parts corresponding to the acceleration process and the soaking process in the constant speed segment, the constant speed segment is divided into a plurality of micro constant speed segments according to a preset time interval; The discharge amount corresponding to each micro constant speed segment is determined; The average power consumption of the corresponding cycle segment is determined; The equivalent mileage data of the micro constant speed segment is determined according to the following formula: ;​ in, Indicates micro constant speed segment Equivalent mileage data, Indicates micro constant speed segment The discharge amount, i represents the number of the micro constant speed segment, Indicates the average power consumption of the cycle segment, where, when the test method is the first shortening method, It represents the average power consumption of the 3rd CLTC-P cycle and the 4th CLTC-P cycle, when the test method is the second shortening method, It represents the average value of the power consumption of the 5th CLTC-P cycle and the 6th CLTC-P cycle.

[0008] In a feasible design, the vehicle test data is collected, including: Obtain the video of the instrument display screen recorded by the video recording device during the test; Extract the mileage data displayed by each instrument from the video at a frequency of once per second; The mileage data displayed on each instrument is sampled according to a first sampling rule to obtain first mileage data. The first mileage data includes mileage data corresponding to each sampling point. The first sampling rule is used to stipulate that the mileage data corresponding to the sampling point of the cycle segment is the absolute value of the difference between the mileage data displayed on the instrument at the current sampling point and the mileage data displayed on the instrument at the previous sampling point. The sampling point of the cycle segment is the time after the end of the current cycle segment and before the beginning of the next cycle segment. It is also stipulated that after the constant speed segment is divided into multiple micro-constant speed segments according to a preset time interval, the time of the endpoint of each micro-constant speed segment is used as the sampling point, and the absolute value of the difference between the mileage data displayed on the instrument at the current sampling point and the mileage data displayed on the instrument at the previous sampling point is used as the mileage data corresponding to the current sampling point.

[0009] In a feasible design, the mileage data displayed by each instrument is extracted from the video at a frequency of once per second, including: The video is collected at a frequency of 5 frames per second to form an image data sequence; The pre-trained digital recognition model independently identifies the mileage number area in each frame of the image data sequence, and outputs five sets of results per second containing specific numbers and corresponding confidence levels. Based on the highest confidence screening mechanism, the number with the highest confidence is selected from the five groups of results corresponding to each second as the mileage data displayed on the instrument corresponding to that second.

[0010] In a feasible design, the process of identifying the mileage number area in each frame image includes the following steps: If the current frame image is the initial frame in the image data sequence, the coordinates of the four vertices of the mileage number area in the initial frame are determined by the target detection algorithm, and the offsets of the four vertices of the mileage number area in the initial frame relative to the preset image boundary are recorded as four reference parameters; If the current frame image is not the initial frame in the image data sequence, the four vertex coordinates of the mileage digital area of ​​the current frame are determined by calculating the screen translation amount based on the phase correlation tracking technology, and the similarity between the offset of each vertex coordinate of the current frame relative to the preset screen boundary and the corresponding reference parameter of the initial frame is calculated. If all similarities are higher than the similarity threshold, the mileage digital area of ​​the next frame is continued to be identified until the mileage digital area of ​​each frame image in the image data sequence is identified. If there is a similarity lower than the similarity threshold, the four vertex coordinates of the mileage digital area of ​​the current frame are re-determined based on the phase correlation tracking technology, and the similarity between the offset of each vertex coordinate of the current frame relative to the preset screen boundary and the corresponding reference parameter of the initial frame is calculated again until all similarities of the current frame are higher than the similarity threshold.

[0011] In a feasible design, the vehicle test data is collected, including: The cumulative mileage of the vehicle is recorded in real time by the rotating hub device, and a cumulative value is generated every second to obtain a cumulative value sequence; The accumulated value sequence is sampled according to the second sampling rule to obtain the mileage data of the hub device. The second sampling rule is used to stipulate that the mileage data corresponding to the sampling point of the cycle segment is the difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point. The sampling point of the cycle segment is the moment after the end of the current cycle segment and before the beginning of the next cycle segment, and after the constant speed segment is divided into multiple micro-constant speed segments according to the preset time interval, the moment of the end point of each micro-constant speed segment is used as the sampling point, and the difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point is used as the mileage data corresponding to the current sampling point.

[0012] In a feasible design, the full-distance error includes the full-distance mean absolute error and the full-distance root mean square error. The full-distance error is determined based on the first mileage data and the second mileage data, including: Obtaining the instrument display mileage data of each valid sampling point except the zero value point and the initial point in the first mileage data; Obtaining measured mileage data for each valid sampling point in the second mileage data, wherein when the target test method is the continuous operating method, the measured mileage data is mileage data of the hub device; when the target test method is the first shortened method or the second shortened method, the measured mileage data is equivalent mileage data; The overall mean absolute error is determined according to the following formula: ; in, represents the overall mean absolute error, Indicates the number of valid sampling points, Indicates the number of valid sampling points in the full test cycle, Indicates valid sampling points The instrument displays mileage data. Indicates valid sampling points Measured mileage data; The full-process root mean square error is determined according to the following formula: ; in, It represents the root mean square error of the whole process.

[0013] In a feasible design, the segmentation error includes the segmentation mean absolute error and the segmentation root mean square error. Determining the segmentation error of the high-battery segment based on the first mileage data and the second mileage data includes the following steps: The segmented mean absolute error of the high power segment is determined according to the following formula: ; in, Represents the segmented mean absolute error of the high power segment, Indicates the number of the first valid sampling point in the high power segment. Indicates the number of the last valid sampling point in the high power segment; The segmented RMS error of the high power segment is determined according to the following formula; ; in, Indicates the segmented RMS error of the high battery segment.

[0014] In a feasible design, determining each single-point error based on the first mileage data and the second mileage data includes the following steps: The absolute value of the difference between the instrument displayed mileage of each valid sampling point in the first mileage data and the measured mileage data of the corresponding valid sampling point in the second mileage data is determined as the single point error of the corresponding valid sampling point.

[0015] In a second aspect, a system for detecting mileage errors displayed on an instrument panel of a pure electric vehicle is provided, comprising: The test mode determination module is configured to determine a target test mode according to a test environment temperature and a cruising range length of the vehicle. When the test environment temperature is a normal temperature, the test mode adopted by the vehicle with a cruising range ≤8 CLTC-P cycle ranges is a continuous duty mode, the continuous duty mode is a test mode in which tests are continuously performed according to the CLTC-P cycle, and the test mode adopted by the vehicle with a cruising range >8 CLTC-P cycle ranges is a first shortening mode, the first shortening mode is a test mode in which tests are performed according to a first test structure, the first test structure includes, in sequence, a cycle section, a constant speed section, a cycle section, and a constant speed section, and the cycle section of the first test structure includes 2 CLTC-P cycles. When the test environment temperature is a low temperature or a high temperature, the test mode adopted by the vehicle with a cruising range ≤8 CLTC-P cycle ranges is the continuous duty mode, and the test mode adopted by the vehicle with a cruising range >8 CLTC-P cycle ranges is a second shortening mode, the second shortening mode is a test mode in which tests are performed according to a second test structure, the second test structure includes, in sequence, a cycle section and a constant speed section, and the cycle section of the second test structure includes 6 CLTC-P cycles. The test module is configured to test the vehicle according to the target test mode and collect test data of the vehicle, the test data including first mileage data and second mileage data displayed by the instrument, and the second mileage data including mileage data of a rotary hub device. When the target test mode is the first shortening mode or the second shortening mode, the second mileage data further includes equivalent mileage data determined according to a discharge amount of the constant speed section and an average power consumption of the cycle section. The error calculation module is configured to determine a total error, a section error, and a single-point error according to the first mileage data and the second mileage data. The total error is calculated based on sampling points in a whole test period. The section error is calculated based on sampling points in a high-power section, a medium-power section, or a low-power section divided according to a battery state of charge. The single-point error is calculated based on a single sampling point. The evaluation module is configured to evaluate an instrument display error of the vehicle according to the total error, a total error threshold, the section error, a section error threshold, the single-point error, and a single-point error threshold.

[0016] The embodiment of the application first accurately determines the target test method according to the test environment temperature and the cruising range length of the vehicle, constructs a comprehensive and unified test framework, and can find an appropriate test method for vehicles with long or short cruising ranges in normal temperature, low temperature or high temperature environments. This innovative initiative provides a standardized operation process and data acquisition basis for subsequent error detection, ensures the comparability of test data under different vehicle models and different working conditions from the source, and builds a solid framework for solving technical problems. Specifically, for vehicles with a cruising range of ≤8 CLTC-P cycle ranges in a normal temperature test environment, the continuous working condition method is adopted. This test method is highly consistent with the actual driving conditions of the vehicle, and can fully reflect the cruising performance of the vehicle in common driving scenarios such as urban roads. Because the test time of short-range vehicles is relatively short, the continuous working condition method can completely capture the energy consumption and range display characteristics of the vehicle at different driving stages, ensuring the comprehensiveness and accuracy of error detection. For vehicles with a cruising range of >8 CLTC-P cycle ranges, the first shortening method is adopted, and the test structure includes cycle segments, constant speed segments, cycle segments and constant speed segments in turn. This design takes into account the long cruising range of the vehicle, and simulates the high-speed driving scenario of the vehicle through the setting of the constant speed segment. Such a test structure can effectively shorten the test time while ensuring the representativeness of the test results, balancing the test efficiency and accuracy of long-range vehicles. In low temperature or high temperature test environments, vehicles with a cruising range of ≤8 CLTC-P cycle ranges also use the continuous working condition method. This is because, under extreme temperature conditions, the energy consumption and cruising performance of the vehicle will be significantly affected, and the continuous working condition method can fully expose the performance of the vehicle under these special working conditions. For vehicles with a cruising range of >8 CLTC-P cycle ranges, the second shortening method is adopted, and the test structure is a cycle segment followed by a constant speed segment. This design takes into account the test efficiency of long-range vehicles under extreme temperatures, and simulates the energy consumption of the vehicle during high-speed driving or stable driving through the setting of the constant speed segment, ensuring the reliability of the test results.

[0017] Secondly, when testing the vehicle according to the target test method and collecting the test data of the vehicle, if the target test method is the first shortening method or the second shortening method, the second range data collected not only includes the range data of the hub device, but also innovatively introduces equivalent range data determined according to the discharge amount of the constant speed segment and the average energy consumption of the cycle segment. This improvement effectively solves the problem of large deviation of the total range of the hub caused by high-speed discharge in the shortening method test, making the actual range data more accurate and objective in reflecting the actual driving conditions of the vehicle. The introduction of equivalent range data not only enriches the dimension of test data, but also enhances the scientificity and comparability of data, further improves the accuracy of vehicle performance evaluation under different working conditions, provides more reliable data support for subsequent error calculation, and significantly improves the accuracy and reliability of error detection.

[0018] Finally, by comprehensively analyzing and calculating the collected first mileage data and second mileage data, the overall error, the error of each segment and the error of each single point are determined. The overall error reflects the overall deviation trend, the error of each segment is accurately positioned to the error characteristics of different power segments, and the error of each single point is refined to the display accuracy of each sampling time. This multi-dimensional error quantification method makes the error result more detailed and comparable. At the same time, by comparing with the preset error threshold, the accurate evaluation of the display error of the vehicle instrument is realized, which provides a strong basis for the accuracy improvement of the instrument display mileage of the pure electric vehicle and the performance evaluation of different vehicle models. This comprehensive and detailed error evaluation system can not only accurately identify the mileage display error of different vehicle models under various working conditions, but also provide accurate data support for subsequent vehicle improvement and optimization, so that the displayed range error of different vehicle models and different working conditions can be detected, the results can be quantified, and the data can be compared, and the evaluation system of the instrument display mileage error is unified. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0020] Figure 1 is a schematic flow chart of a pure electric vehicle instrument display mileage error detection method provided by an exemplary embodiment of the present application; Figure 2 is a schematic flow chart of another pure electric vehicle instrument display mileage error detection method provided by an exemplary embodiment of the present application; Figure 3 is a schematic diagram of a pure electric vehicle instrument display mileage error detection system provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0022] In order to solve the problem of lack of uniformity in the existing evaluation system of instrument display mileage, the application provides a pure electric vehicle instrument display mileage error detection method and device. The method establishes a corresponding test matrix according to multiple temperature scenes and multiple working condition structures, and establishes a hierarchical error evaluation system of "total error + segmented error + single point error", so that the instrument display range error under different vehicle models and different working conditions can be detected, the results can be quantified, and the data can be compared, thereby achieving the purpose of unifying the evaluation system of instrument display mileage.

[0023] Figure 1 An example of a pure electric vehicle instrument display mileage error detection method provided by an example embodiment of the application is shown in the flowchart as shown in Figure 1 S110, determining a target test mode according to the test environment temperature and the cruising range length of the vehicle.

[0024] Among them, under the normal temperature test environment temperature, the test mode adopted by the vehicle with a cruising range ≤8 China Light-duty Vehicle Test Cycle-Passenger Car (CLTC-P) cycle range is continuous working condition method, the continuous working condition method is a test mode for continuously testing according to CLTC-P cycle, the test mode adopted by the vehicle with a cruising range >8 CLTC-P cycle range is first shortening method, the first shortening method is a test mode for testing according to the first test structure, the working condition segments included in the first test structure (also referred to as the first test matrix) are in turn cycle segment (Dynamic Segment, DS), constant segment (Constant Segment, CSS), cycle segment, constant segment, under low temperature test environment temperature or high temperature test environment temperature, the test mode adopted by the vehicle with a cruising range ≤8 CLTC-P cycle range is continuous working condition method, the test mode adopted by the vehicle with a cruising range >8 CLTC-P cycle range is second shortening method, the second shortening method is a test mode for testing according to the second test structure, the working condition segments included in the second test structure (also referred to as the second test matrix) are in turn cycle segment and constant segment.

[0025] For example, the first cycle segment and the second cycle segment of the first test structure each include two CLTC-P cycles. The cycle segment of the second test structure includes 6 CLTC-P cycles.

[0026] For example, the working condition speed of the constant speed segment of the first test structure and the working condition speed of the constant speed segment of the second test structure are each set to a constant value of 100 (km) / h.

[0027] For example, the normal temperature test environment temperature is 23℃. ​

[0028] Exemplarily, the low-temperature test environment temperature is -7℃.

[0029] Exemplarily, the high-temperature test environment temperature is 30℃.

[0030] It can be seen that, in the continuous mode, the vehicle starts from full power, and keeps running according to the vehicle speed specified in the CLTC-P. If the vehicle cannot meet the specified vehicle speed requirement for 4 seconds continuously, the test is ended. In the first shortened method or the second shortened method, the vehicle starts from full power, and runs according to the specified test structure, the vehicle speed specified in the CLTC-P, and the vehicle speed specified in the constant speed section. If the vehicle cannot meet the specified vehicle speed requirement for 4 seconds continuously, the test is ended.

[0031] S120, testing the vehicle according to the target test mode, and collecting test data of the vehicle.

[0032] The test data includes first mileage data and second mileage data displayed by the instrument, and the second mileage data includes mileage data of the hub device. In the case where the target test mode is the first shortened method or the second shortened method, the second mileage data further includes equivalent mileage data determined according to the discharge amount of the constant speed section and the average power consumption of the cycle section. The average power consumption is the average discharge amount consumed by the vehicle per kilometer in the cycle section, the constant speed section, or the micro-constant speed section, and the unit is watt-hour per kilometer.

[0033] In a feasible design, the first mileage data of the vehicle is collected by the following method, including: acquiring a video of the instrument display screen recorded by the video recording device during the test; extracting the instrument display mileage data once per second; sampling the instrument display mileage data according to a first sampling rule to obtain the first mileage data, the first mileage data including mileage data corresponding to each sampling point, the first sampling rule being used to specify that the mileage data corresponding to a sampling point in the cycle section is the absolute value of the difference between the instrument display mileage data of the current sampling point and the instrument display mileage data of the last sampling point, and the sampling point in the constant speed section is the time point after the end of the current cycle section and before the start of the next cycle section, and after the constant speed section is divided into a plurality of micro-constant speed sections according to a preset time interval, the instrument display mileage data of the current sampling point and the instrument display mileage data of the last sampling point are taken as the absolute value of the difference as the mileage data corresponding to the current sampling point.

[0034] For example, when testing the vehicle by using the first shortening method, the starting time of the first CLTC-P cycle in the first cycle section is the 1st second, the ending time of the first CLTC-P cycle is the 200th second, and the starting time of the second CLTC-P cycle is the 202nd second. Then, the sampling point corresponding to the first CLTC-P cycle is the 201st second, and the mileage data corresponding to the sampling point is the absolute value (10 km) of the difference (-10 km) between the instrument display mileage data (590 km) corresponding to the 201st second and the instrument display mileage data (600 km) corresponding to the 0th second (i.e., the time before the start of the current cycle section). Similarly, the ending time of the second CLTC-P cycle is the 401st second, the corresponding sampling point is the 402nd second, and the mileage data corresponding to the sampling point is the absolute value (12 km) of the difference (-12 km) between the instrument display mileage data (578 km) corresponding to the 402nd second and the instrument display mileage data (590 km) corresponding to the 201st second.

[0035] It can be seen that the sampling point of the present application can be understood as a sampling time.

[0036] For example, the video recording device is equipped with a high-definition camera with a resolution of 1920x1080 pixels and a frame rate of 60 frames per second.

[0037] For example, the video recording device is equipped with an anti-shake bracket, a heat dissipation module, and a data processing module.

[0038] The anti-shake bracket is made of high-strength shock-absorbing material, and the length of the bracket can be adjusted. The heat dissipation module is used to ensure the stability of the device during long-term operation. The data processing module is used to analyze the video recorded by the video recording device to extract the instrument display mileage data at a frequency of once per second. The video recording device, the anti-shake bracket, the heat dissipation module, and the data processing module constitute an instrument display data acquisition device, which belongs to the test module.

[0039] The above examples can achieve accurate monitoring and analysis of vehicle performance by collecting test data of the vehicle. First, the video of the instrument display screen recorded by the video recording device during the test is obtained. Second, the mileage data displayed by each instrument is extracted from the video at a frequency of once per second, and continuous and detailed vehicle driving mileage information can be obtained. Then, according to the sampling rule, the extracted instrument display mileage data is sampled, and the mileage data corresponding to each sampling point constitutes the first mileage data, which provides a reliable basis for subsequent vehicle performance analysis and evaluation.

[0040] In a feasible design, the video is extracted at a frequency of once per second to extract the instrument display mileage data as follows: The video is collected at a frequency of 5 frames per second to form an image data sequence. For the mileage number region in each image in the image data sequence, an independent recognition is performed through the pre-trained number recognition model, and 5 groups of results containing specific numbers and corresponding confidence levels per second are output. Based on the highest confidence screening mechanism, the highest confidence number in the 5 groups of results corresponding to each second is selected as the instrument display mileage data corresponding to this second.

[0041] Exemplarily, when the number with the highest confidence appears more than a preset number threshold in the 5 images corresponding thereto, the output is confirmed, which can effectively filter transient recognition errors. The number threshold is set according to actual needs, which is not limited in the present application.

[0042] The design samples the video at a frequency of 5 frames per second to form an image data sequence, which provides a sufficient dense data basis for subsequent number recognition, and guarantees the continuity and integrity of the data. For the mileage number region in each image, an independent recognition is performed through the pre-trained number recognition model and 5 groups of results are output, which introduces a redundancy design, effectively increases the accuracy and reliability of the recognition, and even in the case of errors in part of the results, correct data can still be obtained through multiple results. Further, based on the highest confidence screening mechanism, the highest confidence number in the result corresponding to each sampling point is selected as the final instrument display mileage data, which directly filters out low-quality and unreliable recognition results, thereby ensuring that the extracted mileage data has high accuracy and reliability, and effectively improving the overall quality and efficiency of extracting instrument display mileage data from the video.

[0043] In a feasible design, to solve the dynamic offset problem of the number region in the video, the present application adopts a precise positioning and tracking strategy to realize the recognition of the mileage number region in each image: If the current image is the initial frame in the image data sequence, the four vertex coordinates of the mileage number region in the initial frame are determined through a target detection algorithm, and the offsets of the four vertex coordinates of the mileage number region in the initial frame relative to the preset picture boundary are recorded as four reference parameters. If the current frame image is not the initial frame in the image data sequence, the four vertex coordinates of the mileage digital area of ​​the current frame are determined by calculating the screen translation amount based on the phase correlation tracking technology, and the similarity between the offset of each vertex coordinate of the current frame relative to the preset screen boundary and the corresponding reference parameter of the initial frame is calculated. If all similarities are higher than the similarity threshold, the mileage digital area of ​​the next frame is continued to be identified until the mileage digital area of ​​each frame image in the image data sequence is identified. If there is a similarity lower than the similarity threshold, the four vertex coordinates of the mileage digital area of ​​the current frame are re-determined based on the phase correlation tracking technology, and the similarity between the offset of each vertex coordinate of the current frame relative to the preset screen boundary and the corresponding reference parameter of the initial frame is calculated again until all similarities of the current frame are higher than the similarity threshold.

[0044] The similarity threshold is set according to actual needs and is not limited in this application.

[0045] By analyzing the phase spectrum information of adjacent frames, the above example maintains high tracking accuracy in scenes with varying lighting or slight blur, ensuring that the digital area remains within the recognition frame. Simultaneously, the system continuously calculates the similarity between the current frame and the initial reference frame, triggering a relocalization mechanism when the similarity falls below a threshold to prevent tracking failures caused by severe jitter.

[0046] For example, before recognizing the mileage digits in each frame, the image is preprocessed using adaptive histogram equalization to enhance the contrast between the digits and the background. Illumination robustness features are embedded in the digit recognition model to reduce the impact of light and dark fluctuations on the recognition results and enhance adaptability to changing lighting conditions. A multi-frame differencing method is also used to remove anomalous pixels in reflective areas of the image, and morphological operations are combined to restore digit edges.

[0047] In a feasible design, the mileage data of the hub equipment is collected by the following methods, including: The cumulative mileage of the vehicle is recorded in real time by the rotating hub device, and a cumulative value is generated every second to obtain a cumulative value sequence; The accumulated value sequence is sampled according to the second sampling rule to obtain the mileage data of the hub device. The second sampling rule is used to stipulate that the mileage data corresponding to the sampling point of the cycle segment is the difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point. The sampling point of the cycle segment is the moment after the end of the current cycle segment and before the beginning of the next cycle segment, and after the constant speed segment is divided into multiple micro-constant speed segments according to the preset time interval, the moment of the end point of each micro-constant speed segment is used as the sampling point, and the difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point is used as the mileage data corresponding to the current sampling point.

[0048] The above example records the cumulative mileage of the vehicle in real time and generates a cumulative value every second to obtain a continuous cumulative value sequence, thereby realizing the collection of mileage data of the hub device. According to specific sampling rules, the cumulative value sequence is sampled to obtain the mileage data of the hub device. The sampling rule defines the sampling points of the cycle segment, where the mileage data of each sampling point is the difference between the hub cumulative value of the current sampling point and the hub cumulative value of the previous sampling point. In addition, for the constant speed segment, the rule further stipulates that the segment is divided into multiple micro-constant speed segments at preset time intervals, and the moments of the endpoints of these micro-constant speed segments are used as sampling points, and the difference between the hub cumulative value of the current sampling point and the hub cumulative value of the previous sampling point is used as the corresponding mileage data. This technical solution ensures the accurate collection of mileage data, while optimizing the data processing process and improving data processing efficiency and accuracy.

[0049] Through the synergistic effect of the above technologies, the present application solution can stably and accurately extract dashboard mileage data in complex environments, providing reliable support for subsequent data analysis and application.

[0050] In a feasible design, when the target test method is the first shortening method or the second shortening method, the step of collecting equivalent mileage data includes: After excluding the parts corresponding to the acceleration process and the vehicle immersion process in the constant speed section, the constant speed section is divided into multiple micro constant speed sections according to the preset time intervals; Determine the discharge amount corresponding to each micro-constant speed segment; Determine the average power consumption of the corresponding cycle segment; Determine the micro constant speed section according to the following formula (1): Equivalent mileage data : ;Formula (1); in, Indicates micro constant speed segment Equivalent mileage data (in kilometers), Indicates micro constant speed segment The discharge capacity (watt-hours), Indicates the number of the micro constant speed segment, Indicates the average power consumption of the cycle segment, where, when the test method is the first shortening method, It represents the average power consumption of the 3rd CLTC-P cycle and the 4th CLTC-P cycle, when the test method is the second shortening method, It represents the average value of the power consumption of the 5th CLTC-P cycle and the 6th CLTC-P cycle.

[0051] In this embodiment, the constant speed segment under the first shortening method or the second shortening method is first eliminated from the acceleration and vehicle soaking sections, and then subdivided into a number of micro-constant speed segments according to a fixed duration. Subsequently, the discharge amount of each micro-constant speed segment is converted into an equivalent mileage based on the average power consumption of the corresponding cycle segment. In this way, the mileage under a high power consumption condition (constant speed segment) can be converted into an equivalent mileage under a low power consumption condition (cycle segment) according to the power consumption ratio, thereby eliminating the significant deviation caused by the high-speed discharge due to the shortening method resulting in the total hub mileage reaching only 70% to 80% of the rated energy of the battery. The final measured mileage data and the instrument display mileage are consistent and comparable under a unified energy consumption benchmark, thereby improving the accuracy and repeatability of instrument display error detection for pure electric vehicles.

[0052] S130 , determining the overall error, each segment error, and each single point error based on the first mileage data and the second mileage data.

[0053] The full-process error is calculated based on the sampling points within the entire test cycle. The segmented error is calculated based on the sampling points of the high-charge segment (charge > 80%), medium-charge segment (20% ≤ charge ≤ 80%), or low-charge segment (charge < 20%) divided by the battery state of charge range. The single-point error is calculated based on a single sampling point.

[0054] For example, this application divides all sampling points into five categories based on the battery state of charge (SOC) from high to low, where the zero-value point and the initial point are not included in the error calculation. The zero-value point is the sampling point where the instrument displays 0 or "——" mileage, or the instrument displays no mileage. The initial point is the sampling point corresponding to the vehicle's displayed mileage before the test begins. The sampling points in the high-battery range, low-battery range, and medium-battery range below do not include the zero-value point and the initial point.

[0055] The high-battery period sampling points refer to the sampling points corresponding to the first two CLTC-P cycles during the test. For example, during a continuous duty cycle test, the high-battery period sampling points include the sampling points corresponding to the first CLTC-P cycle and the sampling points corresponding to the second CLTC-P cycle.

[0056] During the first shortened test, the low-battery segment includes the sampling points corresponding to the last two CLTC-P cycles and all sampling points corresponding to the constant-speed segment after the last CLTC-P cycle. During the continuous operating mode and second shortened tests, the low-battery segment includes all sampling points where the battery's state of charge (SOC) is below 20%.

[0057] The sampling points of the medium power segment include all sampling points that are temporally located between the last sampling point of the high power segment and the first sampling point of the low power segment.

[0058] In the above example, by dividing the sampling points into high, low, and medium battery ranges based on the battery's state of charge, error calculations can be performed separately for each range. This allows for a comprehensive assessment of the display accuracy of a pure electric vehicle's instrument panel at different battery states, providing users with more reliable mileage information.

[0059] In a feasible design, the full-distance error includes the full-distance mean absolute error and the full-distance root mean square error. The full-distance error is determined based on the first mileage data and the second mileage data in the following manner: Obtaining the instrument display mileage data of each valid sampling point except the zero value point and the initial point in the first mileage data; Obtaining measured mileage data for each valid sampling point in the second mileage data, wherein when the target test method is the continuous operating method, the measured mileage data is mileage data of the hub device; when the target test method is the first shortened method or the second shortened method, the measured mileage data is equivalent mileage data; The average absolute error of the whole process is determined according to the following formula (2): ;Formula (2); in, represents the overall mean absolute error, Indicates the number of valid sampling points, Indicates the number of valid sampling points in the full test cycle, Indicates valid sampling points The instrument shows the mileage data (in kilometers). Indicates valid sampling points Measured mileage data (in kilometers); The root mean square error of the whole process is determined according to the following formula (3): ;Formula (3); in, It represents the root mean square error of the whole process.

[0060] The embodiment innovatively introduces a double-index collaborative evaluation system of the whole average absolute error and the whole root mean square error by accurately focusing on the fine calculation process of the whole error. In the data preprocessing link, the effective sampling point instrument display mileage data is accurately screened out by eliminating the zero value point and the initial point in the first mileage data, effectively avoiding the interference of invalid data on error calculation. In the determination of the measured mileage data, the conditional judgment formula data selection is carried out according to the target test mode, to ensure that the data accurately meets the whole error calculation requirements. In the whole average absolute error calculation, the absolute value mean of the difference between the effective sampling point instrument display mileage and the measured mileage can effectively describe the average deviation degree of the instrument display mileage and the measured mileage in the whole test period. The whole root mean square error is calculated by the square root of the square error mean, which further amplifies the influence of large error points, making the error evaluation more sensitive. This double-index system quantifies the error from different dimensions, not only significantly improves the accuracy of the whole error evaluation, but also enhances the stability and reliability of the evaluation results, providing a solid data support for subsequent error judgment.

[0061] In a feasible design, the segment error includes a segment average absolute error and a segment root mean square error, and determining the segment error of the high power segment according to the first mileage data and the second mileage data includes the following steps: The segment average absolute error of the high power segment is determined according to the following formula (4): ; formula (4); wherein, denotes the segment average absolute error of the high power segment, denotes the number of the first effective sampling point of the high power segment, denotes the number of the last effective sampling point of the high power segment; The segment root mean square error of the high power segment is determined according to the following formula (5): ; formula (5); wherein, denotes the segment root mean square error of the high power segment.

[0062] The embodiment introduces a double-index evaluation system of segmented average absolute error and segmented root mean square error by segment error refinement calculation specially for high electricity section, and realizes the accurate quantification of the instrument display mileage error in high electricity section. The design not only determines the range of the start and end sampling points of the high electricity section, but also calculates the average absolute error and the root mean square error based on the difference between the instrument display mileage and the measured mileage of each sampling point in the range. The segmented average absolute error can accurately reflect the average deviation degree between the instrument display and the measured mileage in the high electricity section, and provide basic data for overall error evaluation. The segmented root mean square error is more focused on highlighting the influence of larger errors, which is helpful to find significant error points in the high electricity section in time. This targeted segmented error evaluation method makes the error analysis of the high electricity section more accurate and comprehensive, and facilitates in-depth analysis of the accuracy of the instrument display mileage in different electricity sections, and provides a strong basis for subsequent error optimization and vehicle performance improvement.

[0063] In a feasible design, the segmented error includes a segmented average absolute error and a segmented root mean square error, and determining the segmented error of the middle electricity section according to the first mileage data and the second mileage data includes the following steps: The segmented average absolute error of the middle electricity section is determined according to the following formula (6): ; formula (6); wherein, denotes the segmented average absolute error of the middle electricity section, denotes the number of the first effective sampling point of the middle electricity section, denotes the number of the last effective sampling point of the middle electricity section; The segmented root mean square error of the middle electricity section is determined according to the following formula (7): ; formula (7); wherein, denotes the segmented root mean square error of the middle electricity section.

[0064] The embodiment introduces a double-index evaluation system of segmented average absolute error and segmented root mean square error by segment error refinement calculation specially for the medium power section, and realizes the precise quantification of the instrument display mileage error in the medium power section. The design not only clearly defines the range of the start and end sampling points of the medium power section, but also calculates the average absolute error and the root mean square error based on the difference between the instrument display mileage and the measured mileage of each sampling point in the range. The segmented average absolute error can accurately reflect the average deviation degree of the instrument display and the measured mileage in the medium power section, and provide basic data for overall error evaluation. The segmented root mean square error is more focused on highlighting the influence of larger errors, which is helpful to timely find the significant error points in the medium power section. This targeted segmented error evaluation method makes the error analysis of the medium power section more accurate and comprehensive, and facilitates in-depth analysis of the accuracy of the instrument display mileage under different power sections, and provides a strong basis for subsequent error optimization and vehicle performance improvement.

[0065] In a feasible design, the segmented error includes a segmented average absolute error and a segmented root mean square error, and determining the segmented error of the low power section according to the first mileage data and the second mileage data includes the following steps: The segmented average absolute error of the low power section is determined according to the following formula (8): ; formula (8); wherein, denotes the segmented average absolute error of the low power section, denotes the number of the first effective sampling point of the low power section, denotes the number of the last effective sampling point of the low power section; The segmented root mean square error of the low power section is determined according to the following formula (9): ; formula (9); wherein, denotes the segmented root mean square error of the low power section.

[0066] The embodiment introduces a double-index evaluation system of segmented average absolute error and segmented root mean square error through segmented error refinement calculation specifically for the low power segment, realizing the accurate quantification of the instrument display mileage error in the low power segment. The design not only clearly defines the start and end sampling point range of the low power segment, but also calculates the average absolute error and the root mean square error based on the difference between the instrument display mileage and the measured mileage of each sampling point in the range. The segmented average absolute error can accurately reflect the average deviation degree of the instrument display and the measured mileage in the low power segment, providing basic data for overall error evaluation; while the segmented root mean square error focuses more on highlighting the impact of larger errors, which helps to timely find significant error points in the low power segment. This targeted segmented error evaluation method makes the error analysis of the low power segment more accurate and comprehensive, facilitating in-depth analysis of the accuracy of instrument display mileage under different power intervals, and providing a strong basis for subsequent error optimization and vehicle performance improvement.

[0067] In a feasible design, determining each single-point error according to the first mileage data and the second mileage data includes the following steps: The absolute value of the difference between the instrument display mileage of each valid sampling point in the first mileage data and the measured mileage data of the corresponding valid sampling point in the second mileage data is determined as the single-point error of the corresponding valid sampling point.

[0068] The embodiment realizes accurate evaluation of the accuracy of mileage display at each sampling point by directly calculating the absolute value of the difference between the instrument display mileage and the measured mileage of each valid sampling point to obtain the single-point error. Presenting the error in the form of absolute value can intuitively reflect the specific deviation degree of the instrument display and the actual driving mileage of the sampling point, and the error quantification is detailed to each sampling time, making the error analysis more precise. This facilitates quick positioning of abnormal sampling points with larger errors, provides strong support for subsequent targeted analysis of vehicle mileage display problems at specific moments, and helps to optimize the instrument display logic from a micro perspective and improve the accuracy and reliability of the overall range display.

[0069] S140, according to the total error, the total error threshold, each segmented error, the segmented error threshold, each single-point error and the single-point error threshold, evaluating the instrument display error of the vehicle.

[0070] Exemplarily, based on the results of the investigation of representative models in the industry, the total error threshold L1 = 5, the segmented error threshold L2 = 7, and the single-point error threshold L3 = 10 are set.

[0071] Exemplarily, when the total error of the vehicle is less than the total error threshold, each segmented error is less than the segmented error threshold, and each single-point error is less than the single-point error threshold, it is determined that the instrument display error of the vehicle meets the threshold requirement. Otherwise, it is determined that the instrument display error of the vehicle does not meet the threshold requirement.

[0072] Based on the above method embodiment, the application provides a test and detection example of the accuracy of the instrument display remaining range of the pure electric vehicle at normal temperature, low temperature and high temperature in combination with a sampling data instance. The detection example is described below in combination with the steps shown in the drawings. Figure 2 The detection example is described below in combination with the steps shown in the drawings.

[0073] An instrument display data acquisition device is installed, wherein a high-strength shock-absorbing video recording device is erected above the main driver instrument display screen and the anti-shake function is turned on. The requirements for the video recording device are: high-definition instrument digital display, high-speed anti-shake, heat dissipation, memory, power supply and the like of the device for long-term use are solved by external power supply. When the test starts, start recording the video, and do not stop in the middle until the test is completed.

[0074] The target test mode is determined according to the test environment temperature and the range length of the vehicle, and the vehicle is tested according to the target test mode. In the process of each test, the data to be recorded include: the instrument display range data, the actual detection hub device range data, and the instrument display battery state of charge data. The second range data includes the hub device range data, wherein in the case of the first shortening method or the second shortening method, the second range data further includes the equivalent range data determined according to the discharge amount of the constant speed section and the average power consumption of the cycle section. The recording requirements for the hub device range data are: recording the cumulative mileage of the vehicle on the hub device, generating a cumulative value every second to obtain a cumulative value sequence of the hub device.

[0075] (1) In the normal temperature test environment, the test mode of the vehicle with a range of ≤8 CLTC-P cycles is the continuous working condition method, i.e. the test is continuously carried out according to the CLTC-P test cycle. The test mode of the vehicle with a range of >8 CLTC-P cycles is the first shortening method, and the normal temperature shortening method is to carry out the test according to the structure of DS1-CSSm-DS2-CSSe, wherein DS1 and DS2 are 2 test cycle sections, each test cycle section includes 2 CLTC-P cycles in sequence, and the cycle sequence numbers are 1~4, i.e. CLTC-P1, CLTC-P2, CLTC-P3, CLTC-P4. CSSm is the first constant speed section, and CSSe is the second constant speed section, and the working condition speed is a constant value of 100 (km) / h. The test starts at the hub under full charge in a 23℃ environment.

[0076] (2) In the low temperature test environment, the test method for the vehicle with the endurance mileage of less than 8 CLTC-P cycles is the continuous working condition method, and the test method for the vehicle with the endurance mileage of more than 8 CLTC-P cycles is the second shortened method. The second shortened method is to test according to the DS-CSS structure, wherein DS is a test cycle section, which is composed of 6 CLTC-P cycles, and the cycle sequence numbers are 1-6, i.e. CLTC-P1, CLTC-P2, CLTC-P3, CLTC-P4, CLTC-P5 and CLTC-P6; and CSS is a constant speed section, and the working condition vehicle speed is a constant value of 100 (km) / h. After being fully charged, the vehicle is immersed in the -7℃ test environment for 12 hours, and then the test is started.

[0077] (3) In the high temperature test environment, the test method for the vehicle with the endurance mileage of less than 8 CLTC-P cycles is the continuous working condition method, and the test method for the vehicle with the endurance mileage of more than 8 CLTC-P cycles is the second shortened method. After being fully charged, the vehicle is immersed in the 30℃ test environment for 0.5 hours, and then the test is started.

[0078] After each test is completed, each data is sorted and output according to the sampling point requirements. Among them, for the video containing the mileage data displayed by the instrument, the video is processed and data is extracted through the matching software, and the data is extracted at a frequency of once per second in the data extraction process. The finally obtained data sample table is as shown in Table 1: Table 1

[0079] The sampling rule for the cumulative value sequence of the hub device is that the mileage data corresponding to the sampling point of the cycle section is the difference between the cumulative value of the hub device at the current sampling point and the cumulative value of the hub device at the last sampling point, and that after the constant speed section is divided into multiple micro constant speed sections every 10 minutes, the mileage data corresponding to the sampling point is the difference between the cumulative value of the hub device at the current sampling point and the cumulative value of the hub device at the last sampling point. The sampling point of the cycle section is the time point after the end of the cycle section and before the start of the next cycle section.

[0080] In the case that the target test method is the first shortened method or the second shortened method, the equivalent mileage data (which can be referred to as equivalent conversion of the constant speed section) of each micro constant speed section is calculated according to formula (1). For example, in the case that the test method is the first shortened method, the equivalent conversion of the constant speed section is as shown in Table 2: Table 2

[0081] The average power consumption for the CLTC-P3 and CLTC-P4 cycles in Table 2 is 122.7 watt-hours per kilometer. micro-CSS stands for micro-constant speed segment, and the number following it indicates the micro-constant speed segment number.

[0082] For example, when the test method is the second shortening method, the equivalent conversion example of the constant speed section is shown in Table 3: Table 3

[0083] The average power consumption of the CLTC-P5 cycle and the CLTC-P6 cycle in Table 3 is 142 watt-hours per kilometer.

[0084] After each test, the mean absolute error and root mean square error (RMSE) for the corresponding ambient temperature are calculated for the overall error, individual segmented errors, and single-point errors. Each error is compared to the corresponding error threshold to assess the error in the vehicle's instrument display. The calculation method for the overall error, individual segmented errors, and single-point errors is described in the aforementioned formula.

[0085] For example, after a vehicle is tested using the first shortening method at room temperature, the data for the CLTC-P1 cycle segment, the CLTC-P2 cycle segment, and the micro-constant speed segments micro-CSS3 to micro-CSS20, including the hub mileage, equivalent mileage, instrument display mileage, and single-point error, are shown in Table 4: Table 4

[0086] The recorded data for the hub mileage, equivalent mileage, instrument displayed mileage, and single-point error for the CLTC-P3 cycle segment, CLTC-P4 cycle segment, and micro-constant speed segments micro-CSS21 to micro-CSS38 are shown in Table 5: Table 5

[0087] According to the data in Tables 4 and 5, when the first shortening method was used to test the vehicle, the overall mean absolute error was 2.1, the mean absolute error in the high-battery segment was 2.5, the mean absolute error in the medium-battery segment was 1.9, the mean absolute error in the low-battery segment was 3.1, and the maximum single-point error was 5.3. The overall root mean square error was 2.5, the root mean square error in the high-battery segment was 3.2, the root mean square error in the medium-battery segment was 2.2, and the root mean square error in the low-battery segment was 3.8. Single-point root mean square errors were not calculated.

[0088] The vehicle was further tested under the second shortening method at high ambient temperature and the second shortening method at low ambient temperature, and the error results were as follows: During high temperature testing: The overall average absolute error is 6.3, the average absolute error of the high power section is 6.7, the average absolute error of the medium power section is 6.5, the average absolute error of the low power section is 5.1, and the maximum single-point error is 12.4. The overall root mean square error is 6.5, the root mean square error of the high power section is 6.9, the root mean square error of the medium power section is 6.7, and the root mean square error of the low power section is 5.3. The single-point root mean square error is no longer calculated.

[0089] In the low-temperature test: The overall average absolute error is 8.7, the average absolute error of the high power section is 8.7, the average absolute error of the medium power section is 9.6, the average absolute error of the low power section is 5.1, and the maximum single-point error is 13.3. The overall root mean square error is 9.0, the root mean square error of the high power section is 8.8, the root mean square error of the medium power section is 9.8, and the root mean square error of the low power section is 5.3. The single-point root mean square error is no longer calculated.

[0090] Based on the overall error threshold L1=5, the segmented error threshold L2=7, and the single-point error threshold L3=10, the above vehicle is evaluated, and the results are as follows: The vehicle's instrument display error (or accuracy) meets all threshold requirements at room temperature test environment temperature; at low-temperature test environment temperature and at high-temperature test environment temperature, the vehicle's instrument display error does not meet some threshold requirements.

[0091] The embodiment of the application first accurately determines the target test method according to the test environment temperature and the cruising range length of the vehicle, constructs a comprehensive and unified test framework, and can find an appropriate test method for vehicles with different cruising ranges in different temperature environments. This innovative initiative provides a standardized operation process and data acquisition basis for subsequent error detection, ensures the comparability of test data under different vehicle models and different working conditions from the source, and builds a solid framework for solving technical problems. Specifically, for the test environment temperature at room temperature, vehicles with a cruising range of ≤8 CLTC-P cycles use the continuous working condition method. This test method is highly consistent with the actual driving conditions of the vehicle, and can fully reflect the cruising performance of the vehicle in common driving scenarios such as urban roads. Because the test time of short-range vehicles is relatively short, the continuous working condition method can fully capture the energy consumption and range display characteristics of the vehicle at different driving stages, ensuring the comprehensiveness and accuracy of error detection. For vehicles with a cruising range of >8 CLTC-P cycles, the first shortening method is used, and the test structure includes cycle segments, constant speed segments, cycle segments and constant speed segments in turn. This design takes into account the long cruising range of the vehicle, and simulates the high-speed driving scenario of the vehicle through the setting of the constant speed segment. Such a test structure can effectively shorten the test time while ensuring the representativeness of the test results, balancing the test efficiency and accuracy of long-range vehicles. In low-temperature or high-temperature test environments, vehicles with a cruising range of ≤8 CLTC-P cycles also use the continuous working condition method. This is because, under extreme temperature conditions, the energy consumption and cruising performance of the vehicle will be significantly affected, and the continuous working condition method can fully expose the performance of the vehicle under these special working conditions. For vehicles with a cruising range of >8 CLTC-P cycles, the second shortening method is used, and the test structure is a cycle segment followed by a constant speed segment. This design takes into account the test efficiency of long-range vehicles under extreme temperatures, and simulates the energy consumption of the vehicle during high-speed driving or stable driving through the setting of the constant speed segment, ensuring the reliability of the test results.

[0092] Secondly, when testing the vehicle according to the target test method and collecting the test data of the vehicle, if the target test method is the first shortening method or the second shortening method, the second range data collected not only includes the range data of the hub device, but also innovatively introduces equivalent range data determined according to the discharge amount of the constant speed segment and the average energy consumption of the cycle segment. This improvement effectively solves the problem of large deviation of the total range of the hub caused by high-speed discharge in the shortening method test, making the actual range data more accurate and objective to reflect the actual driving conditions of the vehicle. The introduction of equivalent range data not only enriches the dimension of test data, but also enhances the scientificity and comparability of data, further improves the accuracy of vehicle performance evaluation under different working conditions, provides more reliable data support for subsequent error calculation, and significantly improves the precision and reliability of error detection.

[0093] Finally, by comprehensively analyzing and calculating the collected first mileage data and second mileage data, the overall error, the error of each segment and the error of each single point are determined. The overall error reflects the overall deviation trend, the error of each segment is accurately positioned to the error characteristics of different power segments, and the error of each single point is refined to the display accuracy at each sampling time. This multi-dimensional error quantization method makes the error result more detailed and comparable. At the same time, by comparing with the preset error threshold, the accuracy of the vehicle instrument display error is realized, which provides a strong basis for the accuracy improvement of the pure electric vehicle instrument display mileage and the performance evaluation of different vehicle models. This comprehensive and detailed error evaluation system can not only accurately identify the mileage display error of different vehicle models under various working conditions, but also provide accurate data support for subsequent vehicle improvement and optimization, so that the displayed range error of different vehicle models and different working conditions can be detected, the results can be quantified, and the data can be compared, thereby unifying the evaluation system of the instrument display mileage error.

[0094] As shown in Figure 3 The present application provides a pure electric vehicle instrument display mileage error detection system, comprising: A test mode determination module is configured to determine a target test mode according to the test environment temperature and the cruising range length of the vehicle. In a normal temperature test environment, the test mode for a vehicle with a cruising range of less than or equal to 8 CLTC-P cycle ranges is a continuous working condition method, the continuous working condition method is a test mode in which the test is continuously performed according to the CLTC-P cycle, the test mode for a vehicle with a cruising range of more than 8 CLTC-P cycle ranges is a first shortening method, the first shortening method is a test mode in which the test is performed according to a first test structure, the working condition segments included in the first test structure are in turn a cycle segment, a constant speed segment, a cycle segment, and a constant speed segment, the cycle segment of the first test structure includes 2 CLTC-P cycles, in a low temperature test environment or a high temperature test environment, the test mode for a vehicle with a cruising range of less than or equal to 8 CLTC-P cycle ranges is a continuous working condition method, the test mode for a vehicle with a cruising range of more than 8 CLTC-P cycle ranges is a second shortening method, the second shortening method is a test mode in which the test is performed according to a second test structure, the working condition segments included in the second test structure are in turn a cycle segment and a constant speed segment, and the cycle segment of the second test structure includes 6 CLTC-P cycles; A test module is configured to test the vehicle according to the target test mode and collect test data of the vehicle, the test data including first mileage data and second mileage data displayed by the instrument, and the second mileage data including mileage data of a hub device. In the case where the target test mode is the first shortening method or the second shortening method, the second mileage data further includes equivalent mileage data determined according to the discharge capacity of the constant speed segment and the average power consumption of the cycle segment; an error calculation module, configured to determine the full-range error, each segmented error, and each single-point error based on the first mileage data and the second mileage data, wherein the full-range error is statistically calculated based on sampling points within the entire test cycle, the segmented error is calculated based on sampling points in the high-charge segment, the medium-charge segment, or the low-charge segment divided by the battery state of charge interval, and the single-point error is calculated based on a single sampling point; The evaluation module is used to evaluate the vehicle's instrument display error based on the full-range error, the full-range error threshold, each segment error, the segment error threshold, each single-point error and the single-point error threshold.

[0095] In a feasible design, when the target test method is the first shortening method or the second shortening method, the test module collects equivalent mileage data in the following manner: After excluding the parts corresponding to the acceleration process and the vehicle immersion process in the constant speed section, the constant speed section is divided into multiple micro constant speed sections according to the preset time intervals; Determine the discharge amount corresponding to each micro-constant speed segment; Determine the average power consumption of the corresponding cycle segment; Determine the micro constant speed section according to the following formula Equivalent mileage data : ; in, Indicates micro constant speed segment Equivalent mileage data, Indicates micro constant speed segment The discharge capacity, Indicates the number of the micro constant speed segment, Indicates the average power consumption of the cycle segment, where, when the test method is the first shortening method, It represents the average power consumption of the 3rd CLTC-P cycle and the 4th CLTC-P cycle, when the test method is the second shortening method, It represents the average value of the power consumption of the 5th CLTC-P cycle and the 6th CLTC-P cycle.

[0096] In a feasible design, the test module collects vehicle test data in the following ways: Obtain the video of the instrument display screen recorded by the video recording device during the test; Extract the mileage data displayed by each instrument from the video at a frequency of once per second; The mileage data displayed on each instrument is sampled according to a first sampling rule to obtain first mileage data, the first mileage data including mileage data corresponding to each sampling point, the first sampling rule being used to define that mileage data corresponding to a sampling point in a cycle segment is an absolute value of a difference between instrument-displayed mileage data of the current sampling point and instrument-displayed mileage data of a previous sampling point, the sampling point in the cycle segment being a time point after the end of the current cycle segment and before the start of a next cycle segment, and being used to define that, after a constant-speed segment is divided into a plurality of micro constant-speed segments according to a preset time interval, mileage data corresponding to a sampling point in each micro constant-speed segment is an absolute value of a difference between instrument-displayed mileage data of the current sampling point and instrument-displayed mileage data of a previous sampling point.

[0097] In a feasible design, the test module implements extraction of the instrument-displayed mileage data from the video at a frequency of once per second in the following manner: The video is collected at a frequency of 5 frames per second to form a sequence of image data; For a mileage number region in each image in the sequence of image data, independent recognition is performed by using a pre-trained number recognition model to output 5 groups of results containing specific numbers and corresponding confidence levels per second; Based on a highest-confidence filtering mechanism, the highest-confidence number is selected from the 5 groups of results corresponding to each second as the instrument-displayed mileage data corresponding to the second.

[0098] In a feasible design, the test module implements recognition of the mileage number region in each image in the following manner: If the current image is an initial image in the sequence of image data, four vertex coordinates of the mileage number region in the initial image are determined by using a target detection algorithm, and offsets of the four vertex coordinates of the mileage number region in the initial image relative to a preset picture boundary are recorded as four reference parameters; If the current image is not the initial image in the sequence of image data, four vertex coordinates of the mileage number region in the current image are determined by calculating a picture translation amount based on a phase correlation tracking technology, and similarities between offsets of each vertex coordinate of the current image relative to the preset picture boundary and corresponding reference parameters of the initial image are calculated, if all the similarities are higher than a similarity threshold, recognition of the mileage number region of the next image is continued until recognition of the mileage number region of each image in the sequence of image data is completed, if there is a similarity lower than the similarity threshold, the four vertex coordinates of the mileage number region in the current image are determined again based on the phase correlation tracking technology, and the similarities between the offsets of each vertex coordinate of the current image relative to the preset picture boundary and the corresponding reference parameters of the initial image are calculated again until all the similarities of the current image are higher than the similarity threshold.

[0099] In a feasible design, the test module implements collection of test data of the vehicle in the following manner: The cumulative mileage of the vehicle is recorded in real time by the rotating hub device, and a cumulative value is generated every second to obtain a cumulative value sequence; The accumulated value sequence is sampled according to the second sampling rule to obtain the mileage data of the hub device. The second sampling rule is used to stipulate that the mileage data corresponding to the sampling point of the cycle segment is the difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point. The sampling point of the cycle segment is the moment after the end of the current cycle segment and before the beginning of the next cycle segment, and after the constant speed segment is divided into multiple micro-constant speed segments according to the preset time interval, the moment of the end point of each micro-constant speed segment is used as the sampling point, and the difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point is used as the mileage data corresponding to the current sampling point.

[0100] In a feasible design, the full-distance error includes the full-distance mean absolute error and the full-distance root mean square error. The error calculation module determines the full-distance error based on the first mileage data and the second mileage data in the following manner: Obtaining the instrument display mileage data of each valid sampling point except the zero value point and the initial point in the first mileage data; Obtaining measured mileage data for each valid sampling point in the second mileage data, wherein when the target test method is the continuous operating method, the measured mileage data is mileage data of the hub device; when the target test method is the first shortened method or the second shortened method, the measured mileage data is equivalent mileage data; The overall mean absolute error is determined according to the following formula: ; in, represents the overall mean absolute error, Indicates the number of valid sampling points, Indicates the number of valid sampling points in the full test cycle, Indicates valid sampling points The instrument displays mileage data. Indicates valid sampling points Measured mileage data; The root mean square error of the entire process is determined according to the following formula: ; in, It represents the root mean square error of the whole process.

[0101] In a feasible design, the segmentation error includes the segmentation mean absolute error and the segmentation root mean square error. The error calculation module determines the segmentation error of the high-battery segment based on the first mileage data and the second mileage data in the following manner: The segmented mean absolute error of the high power segment is determined according to the following formula: ; wherein, denotes the segment average absolute error of the high-charge segment, denotes the number of the first valid sampling point of the high-charge segment, denotes the number of the last valid sampling point of the high-charge segment; The segment root mean square error of the high-charge segment is determined according to the following formula: ; wherein, denotes the segment root mean square error of the high-charge segment.

[0102] In a feasible design, the error calculation module determines each single-point error according to the first mileage data and the second mileage data in the following manner: The absolute value of the difference between the instrument display mileage of each valid sampling point in the first mileage data and the measured mileage data of the corresponding valid sampling point in the second mileage data is determined as the single-point error of the corresponding valid sampling point.

[0103] Other embodiments and effects of the above device are described in the method for detecting instrument display mileage error of a pure electric vehicle, and are not repeated here.

[0104] The basic principles of the present application are described above in combination with specific embodiments, but it should be pointed out that the advantages, advantages, effects and the like mentioned in the present application are only examples and not limitations, and these advantages, advantages, effects and the like cannot be considered as the must-have of each embodiment of the present application. In addition, the above specific details are only for the purpose of example and for the purpose of understanding, and are not limited to the above specific details. The above details do not limit the present application to the above specific details.

[0105] It should be understood that although each step in the flowchart of the accompanying drawings is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps has no strict sequence limitation, and they can be executed in other order. Moreover, at least part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order is not necessarily sequential, but can be alternately executed with other steps or sub-steps or stages of other steps.

[0106] The block diagrams of the devices, apparatuses, equipment, systems referred to in this application are only illustrative examples, and are not intended to require or imply that the connection, arrangement, configuration must be as shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, equipment, systems can be connected, arranged, configured in any manner. Words such as "include", "contain", "have", and the like are open-ended words, mean "including but not limited to", and can be used interchangeably with each other. The words "or" and "and" used herein mean the word "and / or", and can be used interchangeably with each other, unless the context clearly indicates otherwise. The word "such as" used herein means the phrase "such as but not limited to", and can be used interchangeably with each other.

[0107] It is also necessary to point out that in the devices, apparatuses and methods of the present application, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalents of the present application.

[0108] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0109] The above description has been given for the purpose of illustration and description. Furthermore, this description does not intend to limit the embodiments of the present application to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for detecting mileage error displayed on an electric vehicle instrument, characterized in that: include: The target test mode is determined according to the test environment temperature and the driving range of the vehicle. Among them, under the normal test environment temperature, the test mode used for vehicles with a driving range of ≤8 CLTC-P cycle mileage is the continuous working condition method, which is a test mode for continuously testing according to the CLTC-P cycle. The test mode used for vehicles with a driving range of >8 CLTC-P cycle mileage is the first shortened method, which is a test mode for testing according to the first test structure. The working condition sections of the first test structure are the cycle section, the constant speed section, and the cycle section. , constant speed section, the cyclic section of the first test structure includes 2 CLTC-P cycles, at a low test ambient temperature or a high test ambient temperature, the test method used for vehicles with a driving range ≤8 CLTC-P cycle mileage is the continuous operating method, and the test method used for vehicles with a driving range >8 CLTC-P cycle mileage is the second shortened method, which is a test method conducted according to the second test structure. The operating sections included in the second test structure are a cyclic section and a constant speed section in sequence, and the cyclic section of the second test structure includes 6 CLTC-P cycles; Testing the vehicle according to the target test method and collecting vehicle test data, the test data including first mileage data and second mileage data displayed on an instrument panel, the second mileage data including mileage data of a rotating hub device, wherein, when the target test method is the first shortened method or the second shortened method, the second mileage data also includes equivalent mileage data determined based on the discharge amount in the constant speed section and the average power consumption in the cycle section; Determining a full-range error, each segmented error, and each single-point error based on the first mileage data and the second mileage data, wherein the full-range error is statistically calculated based on sampling points within a full test cycle, the segmented error is calculated based on sampling points of a high-power segment, a medium-power segment, or a low-power segment divided by the battery state of charge interval, and the single-point error is calculated based on a single sampling point; The vehicle's instrument display error is evaluated based on the full-range error, full-range error threshold, each segment error, segment error threshold, each single-point error and single-point error threshold.

2. The method according to claim 1, characterized in that When the target test method is the first shortening method or the second shortening method, the step of collecting equivalent mileage data includes: After excluding the parts corresponding to the acceleration process and the vehicle immersion process in the constant speed section, the constant speed section is divided into multiple micro constant speed sections according to the preset time intervals; Determine the discharge amount corresponding to each micro-constant speed segment; Determine the average power consumption of the corresponding cycle segment; Determine the micro constant speed section according to the following formula Equivalent mileage data : ； in, Indicates micro constant speed segment Equivalent mileage data, Indicates micro constant speed segment The discharge capacity, Indicates the number of the micro constant speed segment, Indicates the average power consumption of the cycle segment, where, when the test method is the first shortening method, It represents the average power consumption of the 3rd CLTC-P cycle and the 4th CLTC-P cycle, when the test method is the second shortening method, It represents the average value of the power consumption of the 5th CLTC-P cycle and the 6th CLTC-P cycle.

3. The method according to claim 1 or 2, characterized in that The collecting of vehicle test data includes: Obtain the video of the instrument display screen recorded by the video recording device during the test; Extracting mileage data displayed by each instrument from the video at a frequency of once per second; The mileage data displayed on each instrument is sampled according to a first sampling rule to obtain first mileage data. The first mileage data includes mileage data corresponding to each sampling point. The first sampling rule is used to stipulate that the mileage data corresponding to the sampling point of the cycle segment is the absolute value of the difference between the mileage data displayed on the instrument at the current sampling point and the mileage data displayed on the instrument at the previous sampling point. The sampling point of the cycle segment is the time after the end of the current cycle segment and before the beginning of the next cycle segment. It is also stipulated that after the constant speed segment is divided into multiple micro-constant speed segments according to a preset time interval, the time of the endpoint of each micro-constant speed segment is used as the sampling point, and the absolute value of the difference between the mileage data displayed on the instrument at the current sampling point and the mileage data displayed on the instrument at the previous sampling point is used as the mileage data corresponding to the current sampling point.

4. The method according to claim 3, characterized in that The extracting mileage data displayed by each instrument from the video at a frequency of once per second includes: The video is captured at a frequency of 5 frames per second to form an image data sequence; The mileage number area in each frame of the image data sequence is independently identified using a pre-trained number recognition model, and five sets of results containing specific numbers and corresponding confidence levels are output per second; Based on the highest confidence screening mechanism, the number with the highest confidence is selected from the five groups of results corresponding to each second as the mileage data displayed on the instrument corresponding to that second.

5. The method according to claim 4, characterized in that The recognition process of the mileage digit area in each frame image includes the following steps: If the current frame image is the initial frame in the image data sequence, determine the coordinates of the four vertices of the mileage number area in the initial frame using a target detection algorithm, and record the offsets of the four vertices of the mileage number area in the initial frame relative to the preset picture boundary as four reference parameters; If the current frame image is not the initial frame in the image data sequence, the four vertex coordinates of the mileage digital area of ​​the current frame are determined by calculating the screen translation based on the phase correlation tracking technology, and the similarity between the offset of each vertex coordinate of the current frame relative to the preset screen boundary and the corresponding reference parameter of the initial frame is calculated. If all similarities are higher than the similarity threshold, the mileage digital area of ​​the next frame is continuously identified until the mileage digital area of ​​each frame image in the image data sequence is identified. If there is a similarity lower than the similarity threshold, the four vertex coordinates of the mileage digital area of ​​the current frame are re-determined based on the phase correlation tracking technology, and the similarity between the offset of each vertex coordinate of the current frame relative to the preset screen boundary and the corresponding reference parameter of the initial frame is calculated again until all similarities of the current frame are higher than the similarity threshold.

6. The method according to claim 1 or 2, characterized in that The collecting of vehicle test data includes: The cumulative mileage of the vehicle is recorded in real time by the rotating hub device, and a cumulative value is generated every second to obtain a cumulative value sequence; The accumulated value sequence is sampled according to the second sampling rule to obtain the mileage data of the hub device. The second sampling rule is used to stipulate that the mileage data corresponding to the sampling point of the cycle segment is the difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point. The sampling point of the cycle segment is the moment after the end of the current cycle segment and before the beginning of the next cycle segment. It is also stipulated that the constant speed segment is divided into multiple micro-constant speed segments according to the preset time interval, and the moment of the end point of each micro-constant speed segment is used as the sampling point. The difference between the accumulated value of the hub device at the current sampling point and the accumulated value of the hub device at the previous sampling point is used as the mileage data corresponding to the current sampling point.

7. The method according to claim 1 or 2, characterized in that The full-distance error includes a full-distance mean absolute error and a full-distance root mean square error. Determining the full-distance error based on the first mileage data and the second mileage data includes: Obtaining instrument displayed mileage data of each valid sampling point except the zero value point and the initial point in the first mileage data; Obtaining measured mileage data of each valid sampling point in the second mileage data, wherein when the target test method is the continuous working condition method, the measured mileage data is mileage data of the hub device; when the target test method is the first shortened method or the second shortened method, the measured mileage data is equivalent mileage data; The overall mean absolute error is determined according to the following formula: ； in, represents the overall mean absolute error, Indicates the number of valid sampling points, Indicates the number of valid sampling points in the full test cycle, Indicates valid sampling points The instrument displays mileage data. Indicates valid sampling points Measured mileage data; The root mean square error of the entire process is determined according to the following formula: ； in, It represents the root mean square error of the whole process.

8. The method according to claim 3, characterized in that The segmentation error includes the segmentation mean absolute error and the segmentation root mean square error. Determining the segmentation error of the high battery segment based on the first mileage data and the second mileage data includes the following steps: The segmented mean absolute error of the high power segment is determined according to the following formula: ； in, Represents the segmented mean absolute error of the high power segment, Indicates the number of the first valid sampling point in the high power segment. Indicates the number of the last valid sampling point in the high power segment; The segmented RMS error of the high power segment is determined according to the following formula; ； in, Indicates the segmented RMS error of the high battery segment.

9. The method according to claim 7, characterized in that Determining each single-point error based on the first mileage data and the second mileage data includes the following steps: The absolute value of the difference between the instrument displayed mileage of each valid sampling point in the first mileage data and the measured mileage data of the corresponding valid sampling point in the second mileage data is determined as the single point error of the corresponding valid sampling point.

10. A pure electric vehicle instrument display mileage error detection system, characterized in that: include: The test mode determination module is used to determine the target test mode according to the test environment temperature and the driving range of the vehicle. Among them, under the normal test environment temperature, the test mode used by the vehicle with a driving range of ≤8 CLTC-P cycle mileage is the continuous working condition method, which is a test mode for continuously testing according to the CLTC-P cycle. The test mode used by the vehicle with a driving range of >8 CLTC-P cycle mileage is the first shortened method, which is a test mode for testing according to the first test structure. The working condition sections of the first test structure are cycle section, constant speed section, and continuous speed section. The first test structure comprises a cycle section, a constant speed section, and a cycle section. The cycle section of the first test structure includes two CLTC-P cycles. Under a low test ambient temperature or a high test ambient temperature, the test method used for vehicles with a driving range of ≤8 CLTC-P cycle mileage is the continuous operating condition method. The test method used for vehicles with a driving range of >8 CLTC-P cycle mileage is the second shortened method. The second shortened method is a test method conducted according to the second test structure. The operating condition sections included in the second test structure are a cycle section and a constant speed section in sequence. The cycle section of the second test structure includes six CLTC-P cycles. a testing module, configured to test the vehicle according to the target test method and collect test data of the vehicle, the test data including first mileage data and second mileage data displayed on an instrument panel, the second mileage data including mileage data of a rotating hub device, wherein, when the target test method is the first shortened method or the second shortened method, the second mileage data also includes equivalent mileage data determined based on a discharge amount in a constant speed section and an average power consumption in a cycle section; an error calculation module, configured to determine, based on the first mileage data and the second mileage data, a full-range error, segmented errors, and single-point errors, wherein the full-range error is statistically calculated based on sampling points within a full test cycle, the segmented error is calculated based on sampling points in a high-battery segment, a medium-battery segment, or a low-battery segment divided by the battery state of charge interval, and the single-point error is calculated based on a single sampling point; The evaluation module is used to evaluate the vehicle's instrument display error based on the full-range error, the full-range error threshold, each segment error, the segment error threshold, each single-point error and the single-point error threshold.

Images