Intelligent networked automobile test bed road resistance time sequence data analysis method

Abstract

The invention relates to the technical field of data acquisition, in particular to a road resistance time sequence data analysis method for an intelligent networked automobile test bed. The method comprises the following steps: acquiring an oscillation stationary characteristic value and a response decay time characteristic value of road resistance data; marking a response attenuation time characteristic value of the road resistance data according to a road resistance data difference and an oscillation stationary characteristic value between different road resistance data sequences, and obtaining a target measurement distance between the station test road resistance data in the station test road resistance data sequence and the road test road resistance data in the road test road resistance data sequence; and obtaining a difference evaluation result between the bench test road resistance data sequence and the road test road resistance data sequence according to the target measurement distance. And the accuracy and the credibility of difference or similarity evaluation between the bench test road resistance data and the real road test road resistance data can be improved.

Description

Technical Field

[0001] This invention relates to the field of data acquisition technology, specifically to a method for analyzing road resistance time-series data of an intelligent connected vehicle test bench. Background Technology

[0002] In the R&D and verification process of intelligent connected vehicles, there is a process of analyzing the similarity or difference between road resistance data collected from bench tests and road resistance data collected from real-road tests. This process is mainly for verifying the equivalence and accuracy of testing methods and optimizing vehicle design and performance calibration. Existing technologies typically use traditional dynamic time warping algorithms to calculate the difference or similarity between road resistance data collected from bench tests and real-road tests. Traditional dynamic time warping algorithms generally measure the distance between data points by calculating the absolute value of the difference. However, current advancements in... During bench testing, in the initial stage when the longitudinal driving force or braking force of the vehicle undergoes a step change, a high-frequency damped oscillation caused by additional tire sidewall deformation and rebound will be superimposed on the test road resistance signal. The presence of the high-frequency damped oscillation will cause the road resistance data during the bench test to deviate from the true value. Consequently, the credibility or reliability of the difference or similarity results between the bench test road resistance data and the real road test road resistance data obtained by the traditional dynamic time warping algorithm will be low. Therefore, how to improve the accuracy and credibility of the difference or similarity analysis between the bench test road resistance data and the real road test road resistance data has become an urgent problem to be solved. Summary of the Invention

[0003] To address the aforementioned problems, this invention provides a method for analyzing road resistance time-series data from an intelligent connected vehicle test bench. The specific technical solution adopted is as follows: One embodiment of the present invention provides a method for analyzing road resistance time-series data of an intelligent connected vehicle test bench, comprising the following steps: Acquire the road resistance data sequence of the target vehicle and the power data of each road resistance data in the road resistance data sequence, wherein the road resistance data sequence includes the bench test road resistance data sequence and the road test road resistance data sequence; Based on the difference in power data of all road resistance data within the local time window corresponding to the acquisition time of each road resistance data, the oscillation stability characteristic value of each road resistance data is obtained. The road resistance data in the road resistance data sequence is labeled according to the oscillation stationary characteristic value of the road resistance data to obtain labeled road resistance data. The response decay time characteristic value of the marked road resistance data is obtained based on the nearest extreme value in the power data sequence following the power data of the marked road resistance data. Based on the differences in road resistance data between different road resistance data sequences, the oscillation stationary characteristic value, and the response decay time characteristic value of the marked road resistance data, the target metric distance between the bench test road resistance data in the bench test road resistance data sequence and the road test road resistance data in the road test road resistance data sequence is obtained. The difference assessment result between the bench test road resistance data sequence and the road test road resistance data sequence is obtained based on the target metric distance.

[0004] Beneficial effects: This invention first acquires the road resistance data sequence of the target vehicle and the power data of each road resistance data in the road resistance data sequence; then, based on the difference value of the power data of all road resistance data within the local time window corresponding to the acquisition time of each road resistance data, the oscillation stationary characteristic value of each road resistance data is obtained; subsequently, the road resistance data in the road resistance data sequence is marked according to the oscillation stationary characteristic value of the road resistance data, resulting in marked road resistance data, and the response decay time characteristic value of the marked road resistance data is obtained based on the nearest extreme value after the power data of the marked road resistance data in the power data sequence of the road resistance data sequence; next, based on the road resistance data differences between different road resistance data sequences, the oscillation stationary characteristic value, and the response decay time characteristic value of the marked road resistance data, the target metric distance between the bench test road resistance data in the bench test road resistance data sequence and the road test road resistance data in the road test road resistance data sequence is obtained; finally, the difference evaluation result between the bench test road resistance data sequence and the road test road resistance data sequence is obtained based on the target metric distance. Furthermore, based on the oscillation stationary characteristic value and the response decay time characteristic value, this invention can improve the reliability and accuracy of distance measurement between bench test road resistance data and road test road resistance data in the bench test road resistance data sequence and the road test road resistance data sequence, thereby improving the accuracy and reliability of the assessment of the difference or similarity between bench test road resistance data and real road test road resistance data. Attached Figure Description

[0005] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0006] Figure 1 This is a flowchart of a method for analyzing road resistance timing data on an intelligent connected vehicle test bench according to the present invention. Detailed Implementation

[0007] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the protection scope of the embodiments of the present invention.

[0008] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0009] This embodiment provides a method for analyzing road resistance time-series data on an intelligent connected vehicle test bench, detailed as follows: like Figure 1 As shown, the method for analyzing road resistance time-series data on the intelligent connected vehicle test bench includes the following steps: Step S001: Obtain the road resistance data sequence of the target vehicle and the power data of each road resistance data in the road resistance data sequence. The road resistance data sequence includes the bench test road resistance data sequence and the road test road resistance data sequence.

[0010] During bench testing, the curved contact characteristics between the test bench rollers and tires differ from those of a real flat road. In the initial stage of a step change in the vehicle's longitudinal driving or braking force, a high-frequency damped oscillation caused by additional tire sidewall deformation and rebound is superimposed on the road resistance signal. This high-frequency damped oscillation is added indiscriminately to the road resistance signal reflecting the vehicle's true performance. Consequently, the faster the response of the test vehicle's power or braking system, the more severe the transient impact on the test bench roller and tire system, resulting in a larger amplitude of the generated high-frequency damped oscillation. This causes the dynamic time warping algorithm to generate a large value due to the high-frequency damped oscillation during cost accumulation, leading to a larger final DTW distance. In other words, the presence of high-frequency damped oscillation causes the road resistance data during bench testing to deviate from the true value, resulting in a distance determined based on the absolute value of the difference between road resistance data points. If the accuracy and reliability of the measurement results are low, then the accuracy and reliability of the DTW distance between the bench test road resistance data and the real road test road resistance data determined based on the above distance measurement will also be low. This will lead to low reliability or dependability of the difference or similarity results between the bench test road resistance data and the real road test road resistance data obtained by the traditional dynamic time warping algorithm, affecting the reliability of subsequent vehicle research and development and verification analysis. The DTW distance between sequences can reflect the similarity or difference between sequences. In order to ensure the reliability of subsequent vehicle research and development and verification analysis, this embodiment needs to minimize the negative impact of the deviation of the bench test road resistance data from the true value on the similarity or difference assessment. In other words, it is necessary to improve the accuracy and reliability of the difference or similarity analysis between the bench test road resistance data and the real road test road resistance data, thereby improving or ensuring the reliability of vehicle research and development and verification. Furthermore, for ease of understanding, this embodiment will subsequently describe the process of analyzing the differences or similarities between bench test road resistance data and real road test road resistance data of any vehicle, and will be referred to as the target vehicle. That is, the road resistance data that appears in this embodiment will all belong to the same vehicle.

[0011] In this embodiment, during bench testing and real-road testing of the target vehicle, all bench test road resistance data and all road test road resistance data collected during the bench test and real-road testing of the target vehicle are obtained. The time series of all bench test road resistance data collected during the bench test is recorded as the bench test road resistance data sequence of the target vehicle, and the time series of all road test road resistance data collected during the real-road testing of the target vehicle is recorded as the road test road resistance data sequence of the target vehicle. Both the bench test road resistance data sequence and the road test road resistance data sequence belong to the road resistance data sequence of the target vehicle. Road resistance data refers to the resistance encountered by the vehicle during driving or the equivalent longitudinal force on the vehicle's drive axle that overcomes all driving resistance. That is, road resistance is a resultant force, which includes rolling resistance, wind resistance, etc.

[0012] Since road resistance data cannot be directly measured under normal circumstances, but can be indirectly obtained through other collected data, the specific process for acquiring road resistance data is as follows: First, the vehicle speed and power data of the target vehicle during bench testing and real road testing are collected through onboard sensor networks and data acquisition systems. This yields all bench test speed and power data collected during bench testing, and all road test speed and power data collected during real road testing. The bench test speed and power data are acquired synchronously, and the road test speed and power data are acquired synchronously. That is, at any given moment during bench testing, one bench test power data and one bench test speed data of the target vehicle can be acquired. At any given moment during real road testing, the same data can be acquired. At any given time, the system can acquire road test power data and road test speed data for the target vehicle. The specific time reference and sampling frequency for data acquisition are the same during both bench testing and real-road testing. Since the product of road resistance and vehicle speed equals power, road resistance data can be obtained given the vehicle speed and power. Specifically, for any acquisition time during bench testing of the target vehicle, the ratio of the acquired bench test power data to the acquired bench test speed data is the bench test road resistance data for that acquisition time, which is also the corresponding acquisition time for that road resistance data. Similarly, for any acquisition time during real-road testing of the target vehicle, the ratio of the acquired road test power data to the acquired road test speed data is the road test road resistance data for that acquisition time, which is also the corresponding acquisition time for that road resistance data. In practical applications, the implementer needs to set the data acquisition frequency during bench testing and real-road testing according to actual conditions. For example, in this embodiment, the acquisition frequency can be set to 100 Hz.

[0013] Furthermore, since this embodiment subsequently reduces the negative impact of deviations in test road resistance data from the true value on similarity or difference analysis based on the oscillation variation characteristics and the degree of influence of high-frequency damped oscillations on road resistance data, the oscillation variation characteristics and the degree of influence of high-frequency damped oscillations on road resistance data refer to the response decay time characteristic value and the oscillation stationary characteristic value, respectively. However, since power data reflects various oscillation characteristics more clearly than road resistance data, this embodiment, in order to improve the accuracy and reliability of the analysis of oscillation variation characteristics and the degree of influence of high-frequency damped oscillations on road resistance data, does not directly base its subsequent analysis on... Instead of analyzing and obtaining response decay time and oscillation stability characteristics from road resistance data, this embodiment uses power data for analysis. Therefore, it is necessary to obtain the power data of each road resistance data in the road resistance data sequence of the target vehicle. That is, for any test road resistance data in the test road resistance data sequence of the target vehicle, the test power data collected at the corresponding acquisition time is the power data of the test road resistance data. Similarly, for any road resistance data in the road resistance data sequence of the target vehicle, the road power data collected at the corresponding acquisition time is the power data of the road resistance data.

[0014] Step S002: Based on the difference value of the power data of all road resistance data within the local time window corresponding to the acquisition time of each road resistance data, the oscillation stationary characteristic value of each road resistance data is obtained.

[0015] This embodiment will next obtain the oscillation stability characteristic value of each road resistance data based on the power data obtained above. The oscillation stability characteristic value can reflect whether the corresponding road resistance data belongs to the high-frequency oscillation region or the signal severely curved region, which is an important parameter for improving the subsequent difference or similarity assessment. Since the method of obtaining the oscillation stability characteristic value of each test road resistance data in the bench test road resistance data sequence is the same as the method of obtaining the oscillation stability characteristic value of each road resistance data in the road test road resistance data sequence, this embodiment will use the process of obtaining the oscillation stability characteristic value of each test road resistance data in any bench test road resistance data sequence as an example for the sake of ease of description. The specific process of obtaining the oscillation stability characteristic value of each test road resistance data in the bench test road resistance data sequence is as follows: For any test road resistance data c in the test road resistance data sequence: First, a time window is constructed centered on the acquisition time corresponding to the test road resistance data c, and denoted as the local time window of the test road resistance data c. The time series of all test road resistance data whose acquisition time is within the local time window of the test road resistance data c and belongs to the test road resistance data sequence is denoted as the local power sequence of the test road resistance data c. Then, the local power sequence of the test road resistance data c is subjected to first-order and second-order differences to obtain first-order difference sequences and second-order difference sequences. The f-th first-order difference value in the first-order difference sequence is the result of subtracting the f-th power data from the (f+1)-th power data in the local power sequence. The g-th second-order difference value in the second-order difference sequence is the result of subtracting the g-th first-order difference value from the (g+1)-th first-order difference value in the local power sequence. The process of first-order and second-order differences is a well-known technique. The method involves obtaining the stationary characteristic value of the test road resistance data c based on the first-order and second-order difference sequences of the local power sequence and the length of the local power sequence. In specific applications, the implementer sets the length of the local time window or the length of the local power sequence according to the typical period of the oscillation signal. Considering that the oscillation frequency introduced by the tire roller system is usually within a specific physical range, this embodiment can set the preset window length to the time period corresponding to the number of data points that can cover approximately one complete oscillation cycle within that frequency range. For example, if the oscillation frequency introduced by the tire roller system is approximately 15 Hz, then at a sampling frequency of 100 Hz, in order to effectively capture the oscillation, the length of the local time window can be set to the length of the time period consisting of 21 consecutive data acquisition times, or the length of the local power sequence can be set to 21. Since the local time window for a certain road resistance data is constructed in this embodiment with the acquisition time of the road resistance data as the center, but it is also required that the road resistance data corresponding to the power data in the local power sequence of the road resistance data be located in the road resistance data sequence where the road resistance data is located, then the length of the local power sequence of the road resistance data may be less than the set length. When this happens, a lengthening method can be used to make the length of the local power sequence of all road resistance data consistent. For example, if the test road resistance data c is the 5th data in the test road resistance data sequence, then the local power sequence of the test road resistance data c is composed of the 1st to the 21st data in the test road resistance data sequence.

[0016] In this embodiment, the specific process of obtaining the oscillating stationary characteristic value of the test road resistance data c based on the first-order and second-order difference sequences of the local power sequence and the length of the local power sequence is as follows: The absolute values ​​of all difference values ​​in the second difference sequence are accumulated, and the accumulated result is recorded as the second-order difference accumulated value. The absolute values ​​of all difference values ​​in the first difference sequence are accumulated, and the accumulated result is recorded as the first-order difference accumulated value. The ratio of the second-order difference accumulated value to the first-order difference accumulated value is calculated and recorded as the difference ratio. The product of the difference ratio and the length of the local power sequence of the test road resistance data c is calculated, and the result of negative correlation mapping of the obtained product is recorded as the oscillating stationary characteristic value of the test road resistance data c. Here, the negative correlation mapping process refers to first mapping the obtained product using the hyperbolic tangent function, and then subtracting the mapping result by a constant 1. The specific calculation expression for the oscillating stationary characteristic value of the test road resistance data c is as follows:

[0017] in, Let M1 be the oscillating stationary eigenvalue of the test road resistance data c, M2 be the second-order difference accumulation value, M1 be the first-order difference accumulation value, and tanh() be the hyperbolic tangent function. The length of the local power sequence of the test road resistance data c can also be the length of the local time window of the test road resistance data c; when When it is larger, that is The smaller the value, the higher the curvature of the local power sequence of the test road resistance data c. This indicates a greater probability that the data acquisition time corresponding to the test road resistance data c is in a high-frequency oscillation phase, or a greater probability that the power data of the test road resistance data c is in a high-frequency oscillation phase. The smaller, that is A larger value indicates a lower degree of curvature in the local power sequence of the bench test road resistance data c, thus reducing the probability that the data acquisition time corresponding to the bench test road resistance data c is in a high-frequency oscillation phase, or that the power data of the bench test road resistance data c is in a high-frequency oscillation phase. Furthermore, for bench test road resistance data, a smaller oscillation stationary characteristic value indicates a higher probability that the corresponding bench test time or the corresponding power data is in a high-frequency oscillation phase, or in other words, the bench test road resistance data is more affected by high-frequency damped oscillations. Similarly, for road test road resistance data, a smaller oscillation stationary characteristic value indicates a higher probability that the corresponding road test time or the corresponding power data is in a high-frequency oscillation phase. However, the high-frequency oscillation phase during road testing is not necessarily high-frequency damped oscillation, but is usually caused by changes in test commands or dynamic changes in the vehicle, such as acceleration or deceleration during the test.

[0018] Furthermore, the first-order differential accumulation value is placed in the denominator as a normalization benchmark. The first-order differential accumulation value represents the total amplitude of power change within a local time window. Dividing by this term allows the final characteristic value to focus on the shape of the signal change rather than its magnitude, eliminating the interference of power amplitude differences under different operating conditions on oscillation identification. The second-order differential accumulation value is a discrete approximation of the signal's local curvature, reflecting the rate of change of the signal direction. In a high-frequency oscillating signal, the power curve rapidly transitions between convex and concave sections, resulting in a large second-order differential value. However, in a unidirectional response signal with a basically unchanged direction, whether it's a rapid step or a slow linear ramp, the direction of the power curve remains essentially unchanged, the power curvature is close to zero, and the second-order differential value is very small. Therefore, the second-order differential accumulation value is key to reflecting the probability that the acquisition time corresponding to the test resistance data c is in the high-frequency oscillation stage. Multiplying by... Its purpose is to enhance the calculated normalized curvature metric, specifically for weak oscillations whose values ​​are very small, by multiplying them by a factor greater than 1. integers This can amplify the difference ratio, allowing the subsequent hyperbolic tangent function to respond more sensitively. It is dimensionless; the hyperbolic tangent function in the above equation can map a large difference ratio to a value approaching... The output of , thus making the final Approaching Conversely, one tends to be The difference ratio will be mapped to approximate The output of makes Approaching This facilitates subsequent efforts to reduce the negative impact of high-frequency damped oscillations on similarity or difference analysis. The participation of the original absolute value of the difference when measuring the final distance between road resistance data is determined, and it is also the criterion for subsequent calculations to determine whether to perform response decay time characteristic value calculation on the road resistance data.

[0019] Furthermore, the method for calculating and obtaining the oscillating stationary characteristic value of the road test road resistance data is the same as the method for calculating and obtaining the oscillating stationary characteristic value of the bench test road resistance data. Therefore, this embodiment will not describe in detail the process of obtaining the oscillating stationary characteristic value of the road test road resistance data.

[0020] Therefore, this embodiment can obtain the oscillation stationary characteristic values ​​of each road resistance data in the bench test road resistance data sequence and the road test road resistance data sequence through the above process.

[0021] Step S003: Mark the road resistance data in the road resistance data sequence according to the oscillation stationary characteristic value of the road resistance data to obtain marked road resistance data; obtain the response decay time characteristic value of the marked road resistance data according to the nearest extreme value in the power data sequence of the road resistance data sequence that is after the power data of the marked road resistance data.

[0022] Because the test road resistance data deviates from the true value when the corresponding test time or power data is in the high-frequency damped oscillation stage, the test road resistance data is composed of the true road resistance and the high-frequency damped oscillation. Therefore, the confidence or reliability of the distance measurement obtained solely based on the absolute value of the difference between the instantaneous test road resistance data and the instantaneous road resistance data is low. Consequently, the accuracy and reliability of the difference or similarity results between the test road resistance data sequence and the road resistance data are also low. Furthermore, since the energy decay rate or change rate of different oscillations during the test process varies, and the change in oscillation energy can reflect the true state of the vehicle during the test, this embodiment... To improve the accuracy and reliability of the difference or similarity assessment results, other similarity metrics will be introduced, namely, the response decay time feature value, which characterizes energy attenuation or change. However, not all road resistance data has the necessity and conditions for calculating the response decay time feature value. For example, the response decay time feature value cannot be calculated for data in the non-oscillating phase or in the stationary phase. Therefore, this embodiment only calculates the response decay time feature value for data in the oscillating phase. That is, before calculating the response decay time feature value, it is necessary to obtain road resistance data that can be used for the calculation, namely, marked road resistance data. The specific process for obtaining marked road resistance data is as follows: For any test road resistance data c in the test road resistance data sequence, the time series constructed from the oscillating stationary eigenvalues ​​of all road resistance data in the test road resistance data sequence is denoted as the oscillating stationary eigenvalue sequence of the test road resistance data sequence. The h-th oscillating stationary eigenvalue in the oscillating stationary eigenvalue sequence of the test road resistance data sequence is the oscillating stationary eigenvalue of the h-th test road resistance data in the test road resistance data sequence. All maxima in the oscillating stationary eigenvalue sequence of the test road resistance data sequence are obtained, and it is determined whether the test road resistance data c belongs to a maximum value in the oscillating stationary eigenvalue sequence of the test road resistance data sequence. If so, then... This indicates that the time corresponding to the test road resistance data c is in the oscillation phase, and the test road resistance data c meets the conditions for calculating the response decay time characteristic value. Therefore, it is necessary to calculate the response decay time characteristic value of the test road resistance data c. Thus, the test road resistance data c needs to be marked and recorded as marked road resistance data. Conversely, if the test road resistance data c is not a maximum value in the oscillating stationary characteristic value sequence of the test road resistance data sequence, it indicates that the time corresponding to the test road resistance data c is not in the oscillation phase, and the test road resistance data c does not meet the conditions for calculating the response decay time characteristic value; therefore, it does not need to be marked. Furthermore, the process of marking the road resistance data in the road resistance data sequence to obtain marked road resistance data is the same as the process of marking the road resistance data in the test road resistance data sequence to obtain marked road resistance data. Therefore, this embodiment will not describe the process of marking the road resistance data in the road resistance data sequence to obtain marked road resistance data.

[0023] Therefore, this embodiment obtains marked road resistance data through the above process. After obtaining the marked road resistance data, the response decay time characteristic value of the corresponding marked road resistance data is obtained based on the nearest extreme value in the power data sequence of the road resistance data sequence that is located after the power data of the marked road resistance data. The response decay time characteristic value of the marked road resistance data can mainly reflect the change of power, that is, energy, at the power data of the corresponding marked road resistance data, and can also reflect the time or speed of recovery to stability after being subjected to external impact. The larger the response decay time characteristic value, the slower and more persistent the change. Furthermore, the difference in response decay time characteristic values ​​between data is subsequently incorporated into the distance measurement, which can avoid the influence of the instantaneous test road resistance deviating from the true value on the similarity or difference analysis.

[0024] Furthermore, since the method for obtaining the response decay time characteristic value of all marked road resistance data in this embodiment is the same, for ease of understanding, this embodiment will subsequently describe the process of obtaining the response decay time characteristic value of any marked road resistance data A in the test road resistance data sequence as an example. The specific process for obtaining the response decay time characteristic value of marked road resistance data A is as follows: First, let the power data of the marked road resistance data A be denoted as power data a, and let the power data sequence of the bench test road resistance data sequence be denoted as sequence B. The power data of the kth bench test road resistance data in the bench test road resistance data sequence is the kth power data in the power data sequence of the bench test road resistance data sequence. Among all power data that are located after power data a and belong to sequence B, find the extreme value that is closest to power data a at that position, and denot it as the starting power data corresponding to the marked road resistance data A. Extreme values ​​include maximum and minimum values. For example, if power data a is in sequence B... If the first extreme value appearing after the z-th power data is the (z+3)-th power data in sequence B, then the (z+3)-th power data is the starting power data corresponding to the marked road resistance data A. The acquisition time of the road resistance data corresponding to the starting power data of marked road resistance data A is recorded as the starting time of marked road resistance data A. For example, if a certain power data is the v-th power data in sequence B, then the acquisition time of the road resistance data corresponding to this power data is the time when the v-th test road resistance data in the test road resistance data sequence is acquired. A preset local time period following the start time of the marked road resistance data A is taken and recorded as the preset local time period corresponding to the marked road resistance data A. The arithmetic mean of all test road resistance data in the test road resistance data sequence whose acquisition time falls within the preset local time period is recorded as the local steady-state power corresponding to the marked road resistance data A. Then, based on the local steady-state power corresponding to the marked road resistance data A, the time period to be analyzed corresponding to the marked road resistance data A and the initial change amplitude of the start time corresponding to the marked road resistance data A are obtained. In addition, in specific applications, the implementer needs to set the length of the preset local time period according to the actual situation such as the acquisition frequency. For example, in this embodiment, the length of the preset local time period can be set to 1 second. Moreover, if there is no extreme value closest to power data a among all power data that are located after power data a and belong to sequence B, the response decay time characteristic value of the marked road resistance data A is not calculated. Alternatively, if there is no extreme value closest to power data a among all power data that are located after power data a and belong to sequence B, the extreme value closest to power data a is selected in the entire sequence B as the starting power data corresponding to the marked road resistance data A.

[0025] The specific process for obtaining the analysis time period corresponding to the marked path resistance data A and the initial change amplitude of the starting time corresponding to the marked path resistance data A is as follows: The absolute value of the difference between the starting power data corresponding to the marked path resistance data A and the local steady-state power corresponding to the marked path resistance data A is recorded as the initial change amplitude of the starting time corresponding to the marked path resistance data A; the sequence consisting of all power data in sequence B following the starting power data corresponding to the marked path resistance data A is recorded as the analysis sequence of the marked path resistance data A. For example, if the z1th power data in sequence B is the starting power data corresponding to the marked path resistance data A, then the power data from the z1+1th power data in sequence B to the starting power data in sequence B... The last data point in the sequence constitutes the analysis sequence of the marked road resistance data A. The variation amplitude of each power data point in the analysis sequence is obtained, and the variation amplitude of any power data point in the analysis sequence is the absolute value of the difference between that power data point and the local steady-state power corresponding to the marked road resistance data A. Based on the difference between the variation amplitude of each power data point in the analysis sequence and the initial variation amplitude at the starting time corresponding to the marked road resistance data A, the variation amplitude difference value of each power data point in the analysis sequence is obtained, and the variation amplitude difference value of any power data point in the analysis sequence is the difference between the variation amplitude of that power data point and the initial variation amplitude at the starting time corresponding to the marked road resistance data A. The absolute value of the difference between the initial change amplitudes is calculated; the ratio of the difference in change amplitude of each power data in the sequence to be analyzed to the initial change amplitude at the starting time corresponding to the marked road resistance data A is obtained and recorded as the magnitude change value of each power data in the sequence to be analyzed. The magnitude change value of any power data in the sequence to be analyzed is the ratio of the difference in change amplitude of that power data to the initial change amplitude at the starting time corresponding to the marked road resistance data A; it is determined whether the magnitude change value of the u-th power data in the sequence to be analyzed for the marked road resistance data A is greater than a preset change threshold. If it is not greater, the magnitude change value of the (u+1)-th power data in the sequence to be analyzed is then determined. The system checks whether the value is greater than a preset change threshold, and so on, until the magnitude change of the corresponding power data exceeds the preset change threshold. The system then stops judging when the magnitude change of the power data exceeds the preset change threshold. The acquisition time of the road resistance data corresponding to the power data is recorded as the end time of the marked road resistance data A. The time period from the start time of the marked road resistance data A to the end time of the marked road resistance data A is taken as the time period to be analyzed corresponding to the marked road resistance data A. For example, if the first data in the sequence to be analyzed to have a magnitude change value greater than the preset change threshold is the (u+4)th power data in the sequence to be analyzed, then the acquisition time of the road resistance data corresponding to the (u+4)th power data is the end time of the marked road resistance data A.In addition, in specific applications, implementers need to set a preset change threshold according to the actual situation. For example, the preset change threshold can be set to five percent, that is, the time when the power data corresponding to the first change amplitude in the analysis sequence of the marked road resistance data A is collected when the difference between the first change amplitude and the initial change amplitude exceeds five percent of the initial change amplitude is the end time of the marked road resistance data A.

[0026] Then, based on the power data of the test road resistance data collected during the analysis period corresponding to the marked road resistance data A, the local steady-state power corresponding to the marked road resistance data A, and the initial change amplitude at the starting time corresponding to the marked road resistance data A, the response decay time characteristic value of the marked road resistance data A is obtained; and the specific expression for the response decay time characteristic value of the marked road resistance data A is as follows:

[0027] in, To label the response decay time characteristic value of the road resistance data A, To mark the start time of the time period to be analyzed corresponding to the road resistance data A, To mark the end time of the time period to be analyzed corresponding to the road resistance data A, This refers to the power data of the test road resistance data collected at time t within the analysis time period corresponding to road resistance data A. To label the local steady-state power corresponding to the road resistance data A, exp() is an exponential function with a base of constant e, and τ is the unknown quantity to be optimized. To mark the acquisition time t in the time period to be analyzed corresponding to the road resistance data A, and The time interval between To mark the initial change magnitude of the road resistance data A at the starting time, argmin represents the set of independent variable values ​​that minimize the objective function. When the objective function involves integration (such as functionals in variational calculus), argmin is used to find the function or parameter that maximizes the integral expression. Therefore, in this embodiment... This optimization process aims to attenuate the power deviation amplitude. Fit to exponential decay function The purpose of the above is not to assert that the decay process strictly follows an exponential law, but to use this function as an analytical template to extract features from the actual decay process. In other words, it is used to analyze and obtain the characteristic time of the system's recovery to stability after being impacted. This embodiment treats the brief transient response process as the behavioral manifestation of a micro-dynamic system. It can reflect the speed at which the system returns to stability from an excited state or the speed at which the power recovers from an oscillating state to a stable state. The changes in oscillation and other factors are directly related to the vehicle's state. Therefore, the calculated τ value can reflect the vehicle's state. Thus, based on the response decay time characteristic value, it can be used to measure the difference or distance between data. The larger the value, the slower the energy change or decay at the time of data collection corresponding to the marked road resistance data A. The smaller the value, the faster the energy change or decay occurs at the time of data collection corresponding to the marked road resistance data A.

[0028] In addition, in the above formula This is a theoretical attenuation model. This attenuation function is used to approximate the actual observed signal attenuation process, and to determine the optimal model parameter τ that best represents the real situation. The goal is not to assert that the actual physical process strictly follows an exponential law, but rather to abstract a complex attenuation process into a characteristic parameter that represents its overall attenuation trend. In the above formula... For the actual observed object, it represents the actual power at acquisition time t. Deviation from its local steady-state power The magnitude of the deviation of the power data of the road resistance data collected at time t during the analysis period from its local steady-state power. The sequence of deviation amplitudes formed by the amplitudes of the above equations records how the system gradually returns to stability from its peak state after being subjected to an impact; in the above equation... The initial magnitude of attenuation at the location corresponding to the acquisition time of the marked road resistance data A. The exponential decay function describes the decay process over time. The square of the difference in the above equation represents the square of the difference between the actual observed value and the theoretical model value. This is the idea behind the least squares method: a small difference is calculated as a small penalty, while a large difference is subject to an exponentially increasing penalty. This represents the integral over the time period to be analyzed corresponding to the marked road resistance data A, starting from the extreme value moment. By the end of the event The sum of the squared differences is the integral, which represents the overall fitting error of the theoretical model to the actual process for a given value of τ. for .

[0029] Therefore, this embodiment can obtain the response decay time characteristic value of the marked road resistance data through the above process.

[0030] Step S004: Based on the differences in road resistance data between different road resistance data sequences, the oscillation stationary characteristic value, and the response decay time characteristic value of the marked road resistance data, obtain the target metric distance between the bench test road resistance data in the bench test road resistance data sequence and the road test road resistance data in the road test road resistance data sequence; obtain the difference evaluation result between the bench test road resistance data sequence and the road test road resistance data sequence based on the target metric distance.

[0031] After obtaining the response decay time characteristic value, the target metric distance between the bench test road resistance data and the road test road resistance data in the bench test road resistance data sequence and the road test road resistance data sequence is obtained based on the differences in road resistance data between different road resistance data sequences, the oscillation stationarity characteristic value, and the response decay time characteristic value of the marked road resistance data. The specific acquisition process is as follows: For the i-th bench test road resistance data and the j-th road test road resistance data in the bench test road resistance data sequence: determine whether the i-th bench test road resistance data and the j-th road test road resistance data both belong to the marked road resistance data. If they do not both belong to the marked road resistance data, then the absolute value of the difference between the i-th bench test road resistance data and the j-th road resistance data is directly used as the target metric distance between the i-th bench test road resistance data and the j-th road resistance data. If they both belong to the marked road resistance data, then the target metric distance between the i-th bench test road resistance data and the j-th road resistance data is obtained based on the numerical difference and response decay time characteristic value difference between the i-th bench test road resistance data and the j-th road resistance data, the mean of the road test road resistance data sequence, and the oscillation stationary characteristic value of the i-th bench test road resistance data.

[0032] In this embodiment, the specific process of obtaining the target metric distance between the i-th bench test road resistance data and the j-th road test road resistance data, based on the numerical difference and response decay time characteristic value difference between the i-th bench test road resistance data and the j-th road test road resistance data, the mean of the road test road resistance data sequence, and the oscillation stationary characteristic value of the i-th bench test road resistance data, is as follows: Based on the numerical difference and response decay time characteristic value difference between the i-th and j-th road test road resistance data, the mean of the road test road resistance data sequence, and the oscillation stationary characteristic value of the i-th road test road resistance data, a first distance characterization value and a second distance characterization value are determined between the i-th and j-th road test road resistance data. The sum of the first and second distance characterization values ​​is used as the target metric distance between the i-th and j-th road test road resistance data. The specific process for obtaining the first and second distance characterization values ​​between the i-th and j-th road test road resistance data is as follows: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] The result of multiplying the absolute value of the difference between the road resistance data by the oscillation stationary characteristic value of the i-th bench test road resistance data is denoted as the first distance characterization value between the i-th bench test road resistance data and the j-th road test road resistance data. The normalized result of the absolute value of the difference between the response decay time characteristic value of the i-th bench test road resistance data and the response decay time characteristic value of the j-th road test road resistance data is denoted as the time characteristic difference value. The mean of the road test road resistance data sequence is denoted as the road test road resistance mean. The product of the complement of the oscillation stationary characteristic value of the i-th bench test road resistance data, the time characteristic difference value, and the road test road resistance mean is used as the second distance characterization value between the i-th bench test road resistance data and the j-th road test road resistance data.

[0033] Furthermore, based on the numerical difference and response decay time characteristic value difference between the i-th bench test road resistance data and the j-th road test road resistance data, the mean of the road test road resistance data sequence, and the oscillation stationary characteristic value of the i-th bench test road resistance data, the specific expression for the target metric distance between the i-th bench test road resistance data and the j-th road test road resistance data is obtained as follows:

[0034] in, Let be the target metric distance between the i-th bench test road resistance data and the j-th road test road resistance data. Let be the oscillating stationary characteristic value of the i-th test road resistance data. For the i-th test circuit resistance data, For the j-th road test road resistance data, Let j be the response decay time characteristic value of the j-th road test road resistance data. Let be the response decay time characteristic value of the i-th test road resistance data. and The comparison is for the purpose of Normalization is performed to eliminate dimensions. Let be the mean of the road test road resistance data sequence; and since the smaller the oscillation stationary characteristic value of the i-th test road resistance data, the greater the probability that the i-th test road resistance data is affected by high-frequency damped oscillations, then the less reference should be made to the difference in instantaneous values, that is, the less reference should be made to the difference in instantaneous values. In other words The smaller, The more dominant the target, the more accurate the distance measurement result. When it is larger, The more dominant the target, the more accurate the distance measurement result; The smaller the value, the more similar the dynamic characteristics of the test vehicle are at the time corresponding to the i-th bench test road resistance data point and the time corresponding to the j-th road test road resistance data point. For example, the vehicle speed may be more similar. The dynamic characteristics of the vehicle are directly related to the road resistance data, so similar dynamic characteristics can indicate similar road resistance. This serves as the evaluation benchmark, and the magnitude and scale of the entire cost calculation should be determined by this benchmark data to ensure the objectivity and consistency of the evaluation. That is, multiplying by [the benchmark data] in the above formula... Its function is to achieve dimensional unification and to give the cost term physical meaning. The temporal differences in response characteristics can be transformed into an equivalent force cost, ensuring that the physical meaning of each term in the final cost function is consistent. In the DTW algorithm, the cost typically refers to the distance between corresponding points in two time series. Furthermore, The larger and When it is larger, The larger, The larger the value, the greater the difference between the i-th bench test road resistance data and the j-th road test road resistance data.

[0035] After obtaining the target metric distance between the bench test road obstacle data and the road test road obstacle data in the bench test road obstacle data sequence, matrix construction and path backtracking are performed based on the obtained target metric distance to calculate the DTW distance between the bench test road obstacle data sequence and the road test road obstacle data sequence. The DTW distance between the bench test road obstacle data sequence and the road test road obstacle data sequence is the difference assessment result between the bench test road obstacle data sequence and the road test road obstacle data sequence. The larger the DTW distance between the bench test road obstacle data sequence and the road test road obstacle data sequence, the less similar the bench test road obstacle data sequence and the road test road obstacle data sequence are. In this embodiment, only the method of obtaining the metric distance is changed when obtaining the DTW distance between the bench test road obstacle data sequence and the road test road obstacle data sequence. The matrix construction, path backtracking and other processes after obtaining the metric distance are consistent with the traditional dynamic time warping, so they will not be described in detail.

[0036] Thus, this embodiment completes the analysis of the differences or similarities between bench test road resistance data sequences and road test road resistance data sequences, and outputs the analysis results to the user terminal or test report generation system. Furthermore, by referencing less the differences in instantaneous road resistance and more the differences in response decay time characteristic values ​​when bench test road resistance data is more likely to be affected by high-frequency damped oscillations, this embodiment can maximize the accuracy and reliability of the corresponding distance measurement, thereby improving the accuracy and reliability of the analysis of the differences or similarities between bench test road resistance data and real road test road resistance data. In addition, the results of the analysis of the differences or similarities between bench test road resistance data and real road test road resistance data can be used in various engineering application scenarios, such as quantitatively evaluating and calibrating the simulation accuracy of vehicle test benches, comparing the real impact of different control strategies or hardware modifications on the transient response performance of vehicles, and automatically filtering out abnormal test results from a large amount of test data that differ from the benchmark data by a preset threshold, thereby achieving automated monitoring and quality control of the testing process.

[0037] In summary, this embodiment first acquires the road resistance data sequence of the target vehicle and the power data of each road resistance data in the road resistance data sequence; then, based on the difference value of the power data of all road resistance data within the local time window corresponding to the acquisition time of each road resistance data, the oscillation stationary characteristic value of each road resistance data is obtained; next, the road resistance data in the road resistance data sequence is marked according to the oscillation stationary characteristic value of the road resistance data to obtain marked road resistance data, and the response decay time characteristic value of the marked road resistance data is obtained according to the nearest extreme value after the power data of the marked road resistance data in the power data sequence of the road resistance data sequence; then, based on the road resistance data differences between different road resistance data sequences, the oscillation stationary characteristic value, and the response decay time characteristic value of the marked road resistance data, the target metric distance between the bench test road resistance data in the bench test road resistance data sequence and the road test road resistance data in the road test road resistance data sequence is obtained; finally, the difference evaluation result between the bench test road resistance data sequence and the road test road resistance data sequence is obtained based on the target metric distance. Furthermore, based on the oscillation stationary characteristic value and the response decay time characteristic value, this embodiment can improve the reliability and accuracy of distance measurement between bench test road resistance data and road test road resistance data in the bench test road resistance data sequence and the road test road resistance data sequence, thereby improving the accuracy and reliability of the assessment of the difference or similarity between bench test road resistance data and real road test road resistance data.

[0038] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for analyzing road resistance time-series data on an intelligent connected vehicle test bench, characterized in that, The method includes the following steps: Acquire the road resistance data sequence of the target vehicle and the power data of each road resistance data in the road resistance data sequence, wherein the road resistance data sequence includes the bench test road resistance data sequence and the road test road resistance data sequence; Based on the difference in power data of all road resistance data within the local time window corresponding to the acquisition time of each road resistance data, the oscillation stability characteristic value of each road resistance data is obtained. The road resistance data in the road resistance data sequence is labeled according to the oscillation stationary characteristic value of the road resistance data to obtain labeled road resistance data. The response decay time characteristic value of the marked road resistance data is obtained based on the nearest extreme value in the power data sequence following the power data of the marked road resistance data. Based on the differences in road resistance data between different road resistance data sequences, the oscillation stationary characteristic value, and the response decay time characteristic value of the marked road resistance data, the target metric distance between the bench test road resistance data in the bench test road resistance data sequence and the road test road resistance data in the road test road resistance data sequence is obtained. The difference assessment result between the bench test road resistance data sequence and the road test road resistance data sequence is obtained based on the target metric distance.

2. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 1, characterized in that, The method for obtaining the stationary oscillation eigenvalues ​​includes: For any road resistance data in the road resistance data sequence, the time series consisting of the power data of all road resistance data within the local time window corresponding to the acquisition time of the road resistance data is denoted as the local power sequence of the road resistance data. Based on the results of performing first-order and second-order differences on the local power sequence, the oscillation stationary characteristic value of the road resistance data is obtained.

3. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 2, characterized in that, A method for obtaining the oscillatory stationary characteristic value of the road resistance data based on the results of first-order and second-order differencing of the local power sequence includes: The results of performing first-order and second-order differences on the local power sequence are denoted as the first difference sequence and the second difference sequence, respectively. The ratio of the sum of the absolute values ​​of all difference values ​​in the second difference sequence to the sum of the absolute values ​​of all difference values ​​in the first difference sequence is denoted as the difference ratio. The result of negatively mapping the product of the difference ratio and the length of the local power sequence is denoted as the oscillating stationary characteristic value of the road resistance data.

4. The method for analyzing road resistance time series data of an intelligent connected vehicle test bench as described in claim 1, characterized in that, The method for obtaining the marked road resistance data includes: For any road resistance data in the road resistance data sequence, if the road resistance data is a maximum value in the oscillating stationary eigenvalue sequence of the road resistance data sequence, then the road resistance data is marked and denoted as marked road resistance data. The oscillating stationary eigenvalue sequence of the road resistance data sequence is a time series constructed from the oscillating stationary eigenvalues ​​of all road resistance data in the road resistance data sequence.

5. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 1, characterized in that, The method for obtaining the response decay time feature value of the marked road resistance data includes: For any marked road resistance data A in the road resistance data sequence: The power data of the marked road resistance data A is denoted as power data a, and the power data sequence of the road resistance data sequence is denoted as sequence B. The extreme value in sequence B that is located after power data a and is closest to power data a in position is denoted as the starting power data. The acquisition time of the road resistance data corresponding to the starting power data is denoted as the starting time of the marked road resistance data A. The average power data of all road resistance data in the road resistance data sequence whose acquisition time is within a preset local time period after the starting time is denoted as the local steady-state power. Sequence B is a time series constructed from the power data of all road resistance data in the road resistance data sequence. Based on the local steady-state power, the time period to be analyzed corresponding to the marked road resistance data A and the initial change amplitude at the starting time are obtained; Based on the power data, local steady-state power, and initial change amplitude of the road resistance data in the time period to be analyzed, the response decay time characteristic value of the marked road resistance data A is obtained.

6. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 5, characterized in that, The method for obtaining the time period to be analyzed corresponding to the marked road resistance data A includes: The sequence consisting of all power data following the initial power data in sequence B is denoted as the sequence to be analyzed. It is determined whether the ratio of the difference in the variation amplitude of the u-th power data in the sequence to be analyzed to the initial variation amplitude is greater than a preset variation threshold. If not, the determination continues to be made to determine whether the ratio of the difference in the variation amplitude of the (u+1)-th power data in the sequence to be analyzed to the initial variation amplitude is greater than the preset variation threshold. This process continues until the ratio of the difference in the variation amplitude of the corresponding power data to the initial variation amplitude is greater than the preset variation threshold. The acquisition time of the road resistance data corresponding to the power data is recorded as the end time of the marked road resistance data A. The time period from the start time to the end time is recorded as the time period to be analyzed corresponding to the marked road resistance data A.

7. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 6, characterized in that, The initial change amplitude is the absolute value of the difference between the initial power data and the local steady-state power. The difference in the change amplitude of any power data in the sequence to be analyzed is the absolute value of the difference between the change amplitude of the power data and the initial change amplitude. The change amplitude of any power data in the sequence to be analyzed is the absolute value of the difference between the power data and the local steady-state power.

8. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 7, characterized in that, The specific calculation expression for obtaining the response decay time characteristic value of the marked road resistance data A is as follows: ； in, To label the response decay time characteristic value of the road resistance data A, This refers to the start time of the time period to be analyzed corresponding to the marked road resistance data A. This refers to the end time of the time period to be analyzed corresponding to the marked road resistance data A. The power data refers to the power data of the bench test road resistance data collected at time t within the time period to be analyzed. Let be the local steady-state power, exp() be an exponential function with base e, and τ be the unknown quantity to be optimized. The acquisition time t and The time interval between The initial change range is given.

9. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 1, characterized in that, The method for obtaining the target metric distance includes: For the i-th bench test road resistance data and the j-th road test road resistance data in the bench test road resistance data sequence: Determine whether the i-th bench test road resistance data and the j-th road test road resistance data both belong to the marked road resistance data. If they do not both belong to the marked road resistance data, the absolute value of the difference between the i-th bench test road resistance data and the j-th road test road resistance data is directly used as the target metric distance between the i-th bench test road resistance data and the j-th road test road resistance data. If they both belong to the marked road resistance data, the first distance characterization value and the second distance characterization value between the i-th bench test road resistance data and the j-th road test road resistance data are obtained based on the numerical difference and response decay time characteristic value difference between the i-th bench test road resistance data and the j-th road test road resistance data, the mean of the road test road resistance data sequence, and the oscillation stationary characteristic value of the i-th bench test road resistance data. The sum of the first distance characterization value and the second distance characterization value is used as the target metric distance between the i-th bench test road resistance data and the j-th road test road resistance data.

10. The method for analyzing road resistance time-series data of an intelligent connected vehicle test bench as described in claim 9, characterized in that, The method for obtaining the first distance representation value and the second distance representation value includes: The result of multiplying the absolute value of the difference between the i-th test road resistance data and the j-th road resistance data by the oscillation stationary characteristic value of the i-th test road resistance data is denoted as the first distance characterization value between the i-th test road resistance data and the j-th road resistance data. The normalized result of the absolute value of the difference between the response decay time feature value of the i-th bench test road resistance data and the response decay time feature value of the j-th road test road resistance data is denoted as the time feature difference value. The mean of the road test road resistance data sequence is denoted as the road test road resistance mean. The product of the complement of the oscillation stationary feature value of the i-th bench test road resistance data, the time feature difference value, and the road test road resistance mean is used as the second distance characterization value between the i-th bench test road resistance data and the j-th road test road resistance data.

Images