Pavement unevenness joint estimation method based on baseline correction
By combining the least squares method and the EEMD algorithm for baseline correction, the problem of baseline drift in pavement unevenness estimation is solved, and higher estimation accuracy and stability are achieved, which improves the effect of pavement unevenness recognition.
Patent Information
- Application Number
- CN202411324334.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-23
- Publication Date
- 2024-12-10
AI Technical Summary
When estimating road surface unevenness using the state parameters of suspension response, baseline drift is prone to occur, resulting in loss of authenticity of the estimation results.
Using the combined estimation method of road surface unevenness based on baseline correction, combined with the least squares algorithm and EEMD algorithm, preliminary estimation of road surface unevenness, trend term extraction and multiple iteration correction are performed through the unsprung acceleration signal, to remove baseline drift and improve estimation accuracy. .
It effectively reduces baseline drift phenomenon, improves the accuracy and stability of road surface unevenness recognition, and enhances the performance of signal processing and model prediction.
Smart Images
Figure CN119106568A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of automobile driving, and in particular relates to a road surface roughness joint estimation method based on baseline correction. Background Art
[0002] As an important part of transportation infrastructure, the quality of road surface conditions is directly related to vehicle driving safety, comfort and road service life. Road surface roughness, that is, the change in height along the length of the road relative to the reference plane, is a key indicator for measuring road surface conditions. Its evaluation and control has always been a hot issue in road engineering and vehicle dynamics research. It is not only related to the dynamic response of the vehicle and the comfort of passengers, but also closely related to the fatigue life of vehicle components, fuel consumption and road maintenance costs. Especially when driving at high speeds, road surface roughness may cause significant changes in the vertical load of the vehicle, and even cause the tire to detach from the road in extreme cases, thereby affecting the generation of driving force or braking force of the vehicle. This puts higher requirements on the real-time monitoring and evaluation of road surface roughness. Real-time adjustment of vehicle speed and suspension parameters to adapt to road conditions has become an important means to improve vehicle driving safety and comfort. Therefore, the development of effective road surface roughness identification and evaluation methods is of great significance for intelligent vehicle systems and road maintenance management.
[0003] However, when estimating road roughness using the state parameters of the suspension response, baseline drift is inevitable. That is, the acceleration signal includes the so-called baseline drift (also called trend term) due to instrument instability, noise, actual initial acceleration, etc., that is, small steps or distortions on the motion reference level. This will distort the estimated results or even completely lose their authenticity. Summary of the invention
[0004] The purpose of the embodiments of the present invention is to provide a road roughness joint estimation method based on baseline correction, aiming to solve the problems raised in the above background technology.
[0005] The embodiment of the present invention is implemented as follows: a method for joint estimation of road roughness based on baseline correction includes the following specific steps:
[0006] Step 1: Estimate the initial road roughness based on the unsprung acceleration signal and the transfer function between the unsprung acceleration and the road excitation;
[0007] Step 2: Calculate the trend item in the initial road surface roughness according to the least squares algorithm, remove the trend item, and preliminarily correct the baseline;
[0008] Step 3: According to the EEMD (Ensemble Empirical Mode Decomposition) algorithm, the trend term of the road surface roughness is extracted again and the baseline is corrected to obtain the final estimated road surface roughness.
[0009] A further technical solution is that in step 1, a road surface estimator is established using the transfer function of a quarter vehicle model, in which the vehicle body mass is simplified to m s , the wheel mass is simplified to m u , the suspension spring stiffness is expressed in k s The suspension damping coefficient is represented by c, and the tire stiffness is represented by k t express;
[0010] According to Newton's second law, the following dynamic differential equation is established:
[0011]
[0012] Among them, z s and z u Represent the vertical displacement of the vehicle body and wheels respectively, and are their time derivatives, z r is the time-varying road roughness input;
[0013] By performing Laplace transformation on equations (1) and (2) and eliminating z s , the transfer function between the unsprung acceleration and the road excitation is obtained:
[0014]
[0015] Based on the above formula, a road roughness observation system is established which only observes the unsprung acceleration.
[0016] A further technical solution is that in step 2, the road roughness baseline is first fitted using the least squares method, and a polynomial function z(t) is fitted using the least squares principle, as shown below:
[0017] z(t)=a2t 2 +a1t+a0 (4)
[0018] Then, the initially calculated road roughness z r Subtract the polynomial function to get the corrected road roughness z rc1 , the formula is as follows:
[0019] z rc1 =z r -z(t) (5).
[0020] A further technical solution is that in step 3, first, in n trials, a new time series is constructed by adding a white noise time series to the original signal time series:
[0021] z n (t) = Z re1 +u n (t) (6)
[0022] Where, for n=1,…,N, N is the number of integrated trials;
[0023] Secondly, based on the EEMD algorithm, the time series z contaminated by noise n (t) is decomposed into a series of intrinsic mode functions and the sum of residuals, namely:
[0024]
[0025] Where m is the number of iterations and n is the number of integrations;
[0026] Then the previous two steps are repeated N times. In each iterative test, a different white noise time series u(t) is added to the original signal time series. Finally, the eigenmode function obtained by averaging the N tests is integrated to obtain the final eigenmode function, which is the road roughness z obtained by the EEMD algorithm. rc2 .
[0027] The embodiment of the present invention provides a road surface roughness joint estimation method based on baseline correction, which adopts the least squares method combined with the EEMD method to perform baseline correction of road surface roughness. Through this combination, EEMD can solve the disadvantage that the least squares method needs to predict the trend item type in advance, and improve the adaptability of the model, while the least squares method can effectively reduce the endpoint effect problem in EEMD and enhance the accuracy of trend extraction. The combined use of the two methods can provide higher performance and accuracy in signal processing and model prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 It is a schematic diagram of a quarter vehicle model;
[0029] Figure 2 is the road roughness obtained without baseline correction;
[0030] Figure 3 is the road roughness after least squares correction;
[0031] Figure 4 is the road roughness corrected by the least square method combined with the EEMD algorithm;
[0032] Figure 5A flowchart of a road roughness joint estimation method based on baseline correction is provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0033] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0034] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.
[0035] like Figure 5 As shown, a road roughness joint estimation method based on baseline correction is provided in one embodiment of the present invention, comprising the following specific steps:
[0036] Step 1: Estimate the initial road roughness based on the unsprung acceleration signal using the transfer function between the unsprung acceleration and the road excitation.
[0037] The road surface estimator is established using the transfer function of the quarter vehicle model, which embodies the most basic characteristics of the vehicle vertical model and can accurately evaluate the impact of road surface roughness on vehicle stability.
[0038] The quarter vehicle vertical model is a simplified vehicle dynamics model, which is mainly used to analyze the dynamic response of the vehicle suspension system in the vertical direction. This model usually includes the following key components and connections:
[0039] (1) Wheel mass (unsprung mass): represents the equivalent mass of the wheel and tire, excluding the elastic and damping elements of the suspension system.
[0040] (2) Body mass (sprung mass): represents the equivalent mass of the vehicle body and is connected to the wheel mass through the suspension system.
[0041] (3) Tire spring: simulates the elastic properties of the tire, usually with linear or nonlinear stiffness characteristics.
[0042] (4) Suspension spring: simulates the stiffness of the elastic element of the suspension system (such as coil spring).
[0043] (5) Shock absorber: A damping element that simulates the suspension system (such as a hydraulic shock absorber), which controls the movement of the suspension system and attenuates vibrations.
[0044] (6) Road excitation: It simulates the effect of road roughness on the suspension system during vehicle driving, usually represented by a band-limited white noise model.
[0045] The model is Figure 1As shown, Sprung Mass is the vehicle body mass, simplified to m s , Unsprung Mass is the wheel mass, simplified to m u , k s is the suspension spring stiffness, c is the suspension damping coefficient, k t is the tire stiffness.
[0046] According to Newton's second law, the following dynamic differential equation is established:
[0047]
[0048] Among them, z s and z u Represent the vertical displacement of the vehicle body and wheels respectively, and are their time derivatives, z r is the time-varying road roughness input.
[0049] By performing Laplace transformation on equations (1) and (2) and eliminating z s , and obtain the transfer function between the unsprung acceleration and the road excitation.
[0050]
[0051] Based on the above formula, a road roughness observation system can be established that only uses unsprung acceleration as the observation object.
[0052] Step 2: Calculate the trend term in the initial road roughness based on the least squares algorithm, remove the trend term, and preliminarily correct the baseline.
[0053] When only step 1 is performed, severe baseline drift will occur (see Figure 2 ).
[0054] Therefore, the least squares method is first used to fit the road roughness baseline. The polynomial function z(t) is fitted by the least squares principle. Following the principle that the sample points are evenly distributed on both sides of the mean line and the polynomial order is as low as possible, after multiple comparisons, it is believed that the second-order fitting effect is better.
[0055] z(t)=a2t 2 +a1t+a0 (4)
[0056] Then, the initially calculated road roughness z r Subtract the polynomial function to get the corrected road roughness z rc1 (See Figure 3 ). The formula is as follows:
[0057] z rc1 =zr -z(t) (5)
[0058] Step 3: According to the EEMD (Ensemble Empirical Mode Decomposition) algorithm, the trend term of the road surface roughness is extracted again and the baseline is corrected to obtain the final estimated road surface roughness.
[0059] As can be seen from the figure, although the road roughness after correction is greatly improved compared with that before treatment, there is still a certain gap in accuracy. This is because the trend item type is not predicted in advance and the polynomial fit is not ideal, so the EEMD method is used to further process it.
[0060] First, in n trials, a new time series is constructed by adding the white noise time series to the original signal time series:
[0061] z n (t) = z re1 +u n (t) (6)
[0062] Where, for n=1,…,N, N is the number of integrated trials.
[0063] Secondly, based on the EEMD algorithm, the time series z contaminated by noise n (t) is decomposed into a series of intrinsic mode functions and the sum of residuals, namely:
[0064]
[0065] Where m is the number of iterations and n is the number of integrations.
[0066] Then the previous two steps are repeated N times. In each iterative test, a different white noise time series u(t) is added to the original signal time series. Finally, the eigenmode function obtained by averaging the N tests is integrated to obtain the final eigenmode function, which is the road roughness z obtained by the EEMD algorithm. rc2 (See Figure 4 ).
[0067] Depend on Figure 4 It can be seen that the baseline drift phenomenon in the initially obtained road roughness can be effectively removed by combining the least squares method with EEMD processing, and the road roughness recognition effect is better.
[0068] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A road roughness joint estimation method based on baseline correction, characterized in that: The specific steps include: Step 1: Estimate the initial road roughness based on the unsprung acceleration signal and the transfer function between the unsprung acceleration and the road excitation; Step 2: Calculate the trend item in the initial road surface roughness according to the least squares algorithm, remove the trend item, and preliminarily correct the baseline; Step 3: According to the EEMD algorithm, the trend item of the road surface roughness is extracted again and the baseline is corrected to obtain the final estimated road surface roughness.
2. The method for joint estimation of road roughness based on baseline correction according to claim 1, characterized in that: In step 1, a road surface estimator is established using the transfer function of a quarter vehicle model, in which the vehicle body mass is simplified to m s , the wheel mass is simplified to m u , the suspension spring stiffness is expressed in k s The suspension damping coefficient is represented by c, and the tire stiffness is represented by k t express; According to Newton's second law, the following dynamic differential equation is established: Among them, z s and z u Represent the vertical displacement of the vehicle body and wheels respectively, and are their time derivatives, z r is the time-varying road roughness input; By performing Laplace transformation on equations (1) and (2) and eliminating z s , the transfer function between the unsprung acceleration and the road excitation is obtained: Based on the above formula, a road roughness observation system is established which only observes the unsprung acceleration.
3. The method for joint estimation of road roughness based on baseline correction according to claim 2, characterized in that: In step 2, the road roughness baseline is first fitted using the least squares method, and a polynomial function z(t) is fitted using the least squares principle, as shown below: z(t)=a2t 2 +a1t+a0 (4) Then, the initially calculated road roughness z r Subtract the polynomial function to get the corrected road roughness z rc1 , the formula is as follows: z rc1 =z r -z(t) (5)。 4. The method for joint estimation of road roughness based on baseline correction according to claim 3, characterized in that: In step 3, first, in n trials, a new time series is constructed by adding a white noise time series to the original signal time series: z n (t)=z rc1 +u n (t) (6) Where, for n=1,…,N, N is the number of integrated trials; Secondly, based on the EEMD algorithm, the time series z contaminated by noise n (t) is decomposed into a series of intrinsic mode functions and the sum of residuals, namely: Where m is the number of iterations and n is the number of integrations; Then the previous two steps are repeated N times. In each iterative test, a different white noise time series u(t) is added to the original signal time series. Finally, the eigenmode function obtained by averaging the N tests is integrated to obtain the final eigenmode function, which is the road roughness z obtained by the EEMD algorithm. rc2 .