Pressure assignment method and system based on open air spring system

Abstract

The invention provides a pressure assignment method and system based on an open air spring system, and belongs to the technical field of vehicle intelligent chassis control. The method comprises the following steps: acquiring a vehicle state signal in a whole vehicle CAN network and a sensor signal of the open air spring system in real time; performing multi-dimensional feature extraction on the signal, and identifying a vehicle motion state and a state feature parameter through an improved random forest algorithm; dynamically distributing weights of a plurality of filtering modules according to an identification result, and performing weighted filtering on the air spring pressure original signal; a closed-loop feedback mechanism is introduced, the deviation between a filtering pressure value and a theoretical pressure value is calculated in real time, the filtering weight and parameters are dynamically corrected, and finally accurate assignment of the air spring pressure is achieved. According to the method, the problem of pressure assignment deviation caused by inaccurate working condition recognition, single filtering mode and lack of calibration in the prior art is effectively solved, and the control precision of the electric control shock absorber and the operation stability and comfort of the whole vehicle are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of vehicle intelligent chassis control, and particularly relates to a pressure assignment method and system based on an open air spring system. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute prior art.

[0003] In the application scenario of a vehicle open air spring (hereinafter referred to as "open air spring") system, in order to achieve the goal of reducing wind resistance and optimizing handling performance when the vehicle is running at high speed, the system will actively open the electromagnetic valve to directly connect the open air spring with the atmosphere, and release the gas inside the air spring to the atmospheric environment. However, due to the inherent characteristics of the hardware layout structure of the open air spring system, the flow rate of the gas will change significantly with the running state of the vehicle during the dynamic exhaust process, resulting in irregular fluctuations in the internal pressure signal of the air spring, and further causing deviations in the pressure reading and assignment link. This problem has become a core bottleneck affecting the control accuracy of the open air spring system pressure.

[0004] From the perspective of system correlation, the electronically controlled shock absorber needs to calculate the vehicle wheel end spring load based on the air spring pressure, and generate corresponding damping force adjustment instructions according to the skyhook control strategy. If there is a deviation in the assignment of the air spring pressure, it will directly lead to a mismatch between the calculated wheel end spring load and the actual load, so that the electronically controlled shock absorber cannot achieve precise adjustment of the damping force according to the preset target. This control deviation is particularly prominent in vehicle steering, braking, and passing through rough road conditions, and is specifically manifested in problems such as delayed vehicle handling response, unstable vehicle body posture control, and decreased driving comfort, and in severe cases, it can even affect driving safety, which is significantly different from the current high requirements of the whole vehicle on handling stability and comfort.

[0005] Currently, in terms of processing the pressure fluctuations of the open air spring system, the following defects exist: first, the state recognition dimension is single, relying only on the vehicle CAN signal (such as throttle, steering, acceleration), without considering the air spring system itself characteristics (such as capsule skin strain, gas temperature), and cannot accurately identify special conditions such as load mutation; second, the filtering algorithm is fixed, using a single filtering mode to adapt to all conditions, which cannot cope with the frequency and amplitude differences of pressure fluctuations in different conditions, such as high-frequency fluctuations in intense driving and low-frequency fluctuations in emergency braking, which cannot be effectively processed by the same filtering algorithm; finally, there is a lack of feedback calibration mechanism, and the filtering parameters cannot be dynamically adjusted according to the actual assignment deviation once they are set, which can easily lead to a decrease in assignment accuracy due to sensor drift and system aging over a long period of use. These defects make it difficult for the existing pressure assignment strategy to meet the needs of high-end vehicles for handling stability, and innovation is urgently needed. SUMMARY

[0006] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a pressure assignment method and system based on an open-type air spring system, which can effectively solve the problems of inaccurate state recognition, single filtering mode and lack of calibration in the prior art, realize high-precision and high-stability assignment of air spring pressure under all working conditions, and significantly improve the control effect of the electric control shock absorber and the handling stability and comfort of the whole vehicle.

[0007] To achieve the above object, one or more embodiments of the present application provide the following technical solutions: The present application provides a pressure assignment method based on an open-type air spring system in the first aspect; The present application provides a pressure assignment method based on an open-type air spring system, comprising: collecting vehicle state signals in a CAN network of the whole vehicle and sensor signals of the open-type air spring system in real time; extracting multi-dimensional signal features from the collected vehicle state signals and sensor signals, classifying the extracted multi-dimensional signal features by using an improved random forest algorithm, and outputting the motion state and state feature parameters of the vehicle; dynamically assigning weight coefficients of different filtering modules according to the motion state and state feature parameters, and obtaining filtered air spring pressure values; introducing a pressure assignment deviation closed-loop feedback mechanism, calculating the deviation between the filtered air spring pressure values and the theoretical pressure values corresponding to the actual load of the air spring in real time, and dynamically correcting the weight coefficients and filtering parameters of the filtering modules according to the deviation value; assigning the pressure signal after dynamic weight filtering and closed-loop feedback correction to the air spring to complete precise pressure assignment.

[0008] As a further technical solution, the vehicle state signals include throttle pedal signals, brake pedal signals, steering angle signals, steering angular velocity signals, longitudinal acceleration signals, lateral acceleration signals, and yaw angular velocity signals. The sensor signals include air spring capsule skin strain signals, air spring gas temperature signals, and electromagnetic valve switch state signals.

[0009] As a further technical solution, the multi-dimensional signal feature extraction of the collected vehicle state signals and sensor signals comprises: preprocessing the collected vehicle state signals and sensor signals, wherein the preprocessing includes outlier rejection, signal smoothing and standardization processing; extracting multi-dimensional signal features of the preprocessed vehicle state signals and sensor signals, wherein the multi-dimensional signal features include time domain features, frequency domain features and time domain-frequency domain joint features; As a further technical solution, the extracted multi-dimensional signal features are classified using an improved random forest algorithm, and the motion state and state feature parameters of the vehicle are output, including: A feature matrix is constructed based on the extracted multi-dimensional signal features, and normalization processing is performed; The normalized feature matrix is input into the random forest algorithm to obtain the initial importance of each feature. The initial importance is converted into attention weights based on the multi-dimensional features of the sensor signals. The constructed feature matrix and the attention weights are multiplied element by element to obtain a weighted feature matrix, which strengthens the contribution of the air spring system features in classification; The weighted feature matrix is input into the weighted decision tree, each decision tree outputs the predicted motion state of the vehicle, and the total voting weight of the output different vehicle motion states is counted. The class with the maximum total voting weight is selected as the final recognition result of the vehicle motion state, and the state feature parameters corresponding to the state are output.

[0010] As a further technical solution, the vehicle motion state includes a static state, an aggressive driving state, a bumpy road condition state, an emergency braking condition state, a gentle running state, and a load mutation state; The state feature parameters include fluctuation frequency, fluctuation amplitude, and signal signal-to-noise ratio.

[0011] As a further technical solution, according to the motion state and state feature parameters, the weight coefficients of different filter modules are dynamically allocated to obtain the filtered air spring pressure value, including: A filter library composed of multiple filter modules is constructed, the filter modules at least include mean filter submodules, band-pass filter submodules, high-pass filter submodules, low-pass filter submodules, and Kalman filter submodules. Each filter submodule performs filter processing on the air spring pressure original signal at the same time in parallel, and outputs the corresponding primary filter value respectively; Based on the motion state and state feature parameters, a pre-defined weight allocation rule library is queried to dynamically obtain a set of weight coefficients matched with the current working condition. The set of weight coefficients contains weight coefficients corresponding to each filter module one by one; The primary filter values output by each filter module are weighted and summed with their corresponding weight coefficients to calculate the filtered air spring pressure value.

[0012] As a further technical solution, a pressure assignment deviation closed-loop feedback mechanism is introduced to calculate the deviation between the filtered air spring pressure value and the theoretical pressure value corresponding to the air spring actual load in real time. The weight coefficients and filter parameters of the filter module are dynamically corrected according to the deviation value, including: The load data is obtained in real time by the wheel end spring load sensor, and the theoretical pressure value corresponding to the air spring actual load is generated according to the load data; An instantaneous deviation of the filtered air spring pressure value from the theoretical pressure value is calculated, a two-level deviation threshold system is set based on the instantaneous deviation, and a deviation level is determined according to the interval in which the deviation mean value is located; and the weight coefficient and / or core filter parameter of the filter module are corrected according to different deviation levels.

[0013] The second aspect of the present application provides a pressure assignment system based on an open air spring system.

[0014] A pressure assignment system based on an open air spring system comprises: A signal acquisition module is configured to acquire vehicle state signals in a vehicle CAN network and sensor signals of the open air spring system in real time; A vehicle state recognition module is configured to perform multi-dimensional signal feature extraction on the acquired vehicle state signals and sensor signals, perform classification processing on the extracted multi-dimensional signal features using an improved random forest algorithm, and output a motion state and state characteristic parameters of the vehicle; A weight filter module is configured to dynamically allocate weight coefficients of different filter modules according to the motion state and state characteristic parameters to obtain a filtered air spring pressure value; A closed-loop feedback correction module is configured to introduce a pressure assignment deviation closed-loop feedback mechanism, calculate the deviation of the filtered air spring pressure value from a theoretical pressure value corresponding to an actual load of the air spring in real time, and dynamically correct the weight coefficients and filter parameters of the filter module according to the deviation value; A pressure assignment module is configured to assign the pressure signal corrected by the dynamic weight filtering and closed-loop feedback correction to the air spring to complete pressure assignment.

[0015] The third aspect of the present application provides a computer readable storage medium having a program stored thereon, wherein the program is executed by a processor to implement the steps of the pressure assignment method based on the open air spring system according to the first aspect of the present application.

[0016] The fourth aspect of the present application provides an electronic device comprising a memory, a processor, and a program stored on the memory and executable on the processor, wherein the processor implements the steps of the pressure assignment method based on the open air spring system according to the first aspect of the present application when executing the program.

[0017] The above one or more technical solutions have the following beneficial effects: (1) The present application strengthens the correlation matching of the micro state of the air spring system and the macro working condition of the vehicle by fusing the air spring system exclusive sensor signals and the vehicle CAN signals, and by using the improved random forest algorithm with multi-dimensional feature extraction and attention mechanism. The identification accuracy of special working conditions such as load mutation and asymmetric jolt, which cannot be identified by existing technologies, is improved, the working condition misjudgment rate is effectively reduced, and the mismatch of the filter algorithm caused by working condition misjudgment is avoided.

[0018] (2) The application can adaptively match the optimal filter combination according to the frequency and amplitude characteristics of pressure fluctuations under different working conditions through dynamic weight filtering, and can adapt to high-frequency fluctuations of intense driving, low-frequency fluctuations of emergency braking conditions and wide-frequency impacts of bumpy road conditions under all working conditions, effectively solving the filter failure problem caused by working condition misjudgment and single mode of the prior art.

[0019] (3) The application can monitor the deviation of the filter output value and the theoretical expected value in real time by introducing a pressure assignment deviation closed-loop feedback mechanism, and dynamically corrects the filter parameters and weights accordingly. This mechanism enables the system to have online self-calibration capability, which can effectively compensate for slow errors caused by sensor characteristic drift, system component aging or external environment changes, ensuring the long-term stability of pressure assignment accuracy throughout the vehicle's life cycle. This is a key advantage that existing open source technologies do not have.

[0020] Advantages of additional aspects of the application will be given in part in the following description, some will become apparent from the following description, or will be learned by practicing the application. BRIEF DESCRIPTION OF DRAWINGS

[0021] The accompanying drawings, which form a part of this description, are included to provide further understanding of the application, and are incorporated in and constitute a part of this description. The schematic embodiments of the application and their descriptions are used to explain the application, and do not constitute an improper limitation on the application.

[0022] Figure 1 The method flowchart of the first embodiment.

[0023] Figure 2 The system structure diagram of the second embodiment. DETAILED DESCRIPTION

[0024] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the application. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as generally understood by those skilled in the art to which the application belongs.

[0025] It should be noted that the terms used herein are for the purpose of describing specific embodiments only, and are not intended to limit the exemplary embodiments according to the application.

[0026] The embodiments in the application and the features in the embodiments can be combined with each other without conflict.

[0027] Embodiment one The present embodiment discloses a pressure assignment method based on an open air spring system; As shown in Figure 1 A pressure assignment method based on an open air spring system, comprising: Step S101, real-time acquisition of vehicle state signals in the CAN network of the whole vehicle and sensor signals of the open air spring system; Step S102, multi-dimensional signal feature extraction is performed on the acquired vehicle state signals and sensor signals, and the extracted multi-dimensional signal features are classified and processed using an improved random forest algorithm to output the motion state and state characteristic parameters of the vehicle; Step S103, according to the motion state and state characteristic parameters, the weight coefficients of different filtering modules are dynamically allocated to obtain filtered air spring pressure values; Step S104, a pressure assignment deviation closed-loop feedback mechanism is introduced, the deviation between the filtered air spring pressure value and the theoretical pressure value corresponding to the actual load of the air spring is calculated in real time, and the weight coefficients and filtering parameters of the filtering module are dynamically corrected according to the deviation value; Step S105, the pressure signal after dynamic weight filtering and closed-loop feedback correction is assigned to the air spring to complete precise pressure assignment.

[0028] Specifically, the above steps further include the following contents: Step S101, real-time acquisition of vehicle state signals in the CAN network of the whole vehicle and sensor signals of the open air spring system.

[0029] The acquired vehicle state signals include accelerator pedal signals, brake pedal signals, steering angle signals, steering angular velocity signals, longitudinal acceleration signals, lateral acceleration signals, and yaw angular velocity signals.

[0030] The sensor signals of the open air spring system include air spring capsule skin strain signals, air spring gas temperature signals, and electromagnetic valve switch state signals. The air spring capsule skin strain signals are collected by a fiber Bragg strain sensor, which is pasted on the key stress part of the air spring capsule skin, and can reflect the deformation state of the capsule skin in real time, and indirectly reflect the change of the air spring load. The air spring gas temperature signals collect the air spring internal gas temperature by an NTC temperature sensor, and the temperature change can reflect the compression or expansion rate of the gas, which can assist in judging the gas flow state. The electromagnetic valve switch state signals collect the electromagnetic valve opening time and switch frequency by a Hall sensor to judge the stage characteristics of the gas exhaust.

[0031] Step S102, multi-dimensional signal feature extraction is performed on the acquired vehicle state signals and sensor signals, and the extracted multi-dimensional signal features are classified and processed using an improved random forest algorithm to output the motion state and state characteristic parameters of the vehicle.

[0032] Step S1021, the acquired vehicle state signals and sensor signals are preprocessed, including outlier rejection, signal smoothing, and standardization processing.

[0033] For outlier rejection, in this embodiment, 3 The criterion identifies and eliminates outliers. By calculating the mean value of each signal and standard deviation , signal values exceeding the interval are determined as outliers, and the sliding average value of the adjacent 5 valid data of the signal is replaced to ensure signal continuity.

[0034] After eliminating outliers, the signal is smoothed using a sliding average window with a length of Q, as shown in the following formula:

[0035] wherein, is the value of the first sampling point of the original signal; is the value of the first sampling point after smoothing, eliminating high-frequency random noise in the signal; Because there are differences between the dimensions of different signals, Z-score standardization is used to convert all signals to dimensionless data, and the formula is:

[0036] wherein, is the mean value of the smoothed signal, is the standard deviation of the smoothed signal, ensuring that the weight of each signal is balanced during subsequent feature extraction.

[0037] Step S1022, extract the multi-dimensional signal features of the pre-processed vehicle state signals and sensor signals, including time domain features, frequency domain features, and time domain-frequency domain joint features.

[0038] The extracted time domain features include mean, variance, peak value and change rate. The mean value is used to represent the smoothness of the vehicle state; the variance can reflect the dispersion degree of the signal, and is used to reflect the pressure fluctuation amplitude, such as the significantly larger spring capsule skin strain signal variance in intense driving than in smooth running state; the peak value is used to capture the extreme fluctuation of the signal, such as the larger longitudinal acceleration signal peak value in the case of emergency braking; the change rate is obtained by calculating the difference between the adjacent sampling point signal values, and is used to identify the mutation characteristics of the signal, such as the significantly increased air temperature signal change rate of the spring when the load suddenly changes.

[0039] ​​​For signals with dynamic fluctuation characteristics, such as longitudinal, lateral acceleration, yaw rate, etc., the time domain signal is converted into frequency domain signal by fast Fourier transform, and the frequency domain features reflecting the fluctuation frequency characteristics are extracted. Among them, the frequency domain features include spectral peak, center frequency and frequency band energy ratio. The spectral peak is used to determine the dominant fluctuation frequency of the signal, such as the spectral peak of the wheel end acceleration signal under the bumping road condition, which is concentrated in the frequency of 15-35Hz; the center frequency is used to distinguish the frequency characteristics of different working conditions, such as the center frequency of the longitudinal acceleration signal under the emergency braking condition, which is lower than 2Hz, and the center frequency under the intense driving condition, which is higher than 5Hz; the frequency band energy ratio is used to accurately match the working condition, such as the frequency band energy ratio of 15-35Hz under the bumping road condition, which is more than 60%.

[0040] For the scene with high signal complexity, such as air spring capsule skin strain and wheel end acceleration signal, which needs to reflect the time domain mutation and frequency domain distribution at the same time, wavelet transform is used to realize time-frequency domain joint analysis, and wavelet entropy and short-time Fourier transform energy distribution are extracted. The wavelet entropy is used to distinguish the complexity of the signal, such as the wavelet entropy of the air spring capsule skin strain signal under the load mutation, which is significantly higher than that under the stable working condition; the short-time Fourier transform energy distribution is used to capture the dynamic changes of the signal in time and frequency dimensions, such as when the vehicle switches from smooth running to intense driving, the short-time Fourier transform energy distribution shifts from low frequency interval to high frequency interval.

[0041] Step S1023, constructing a feature matrix based on the extracted multi-dimensional signal features , wherein N is the number of training samples, each sample corresponds to signal features under a continuous working condition, M is the total number of features, including whole vehicle CAN signal derived features and air spring exclusive signal derived features. At the same time, a label matrix is constructed , which labels the corresponding vehicle motion state of each sample, the vehicle motion state includes static state, intense driving state, bumping road condition state, emergency braking condition state, smooth running state and load mutation state; the feature matrix is normalized by Min-Max normalization to map the feature values to the interval [0, 1] and eliminate the interference of different feature value range differences on the algorithm.

[0042] Step S1024, inputting the normalized feature matrix into the traditional random forest, and calculating the initial importance of each feature by Gini coefficient gain method, as follows:

[0043] , wherein, is the initial importance of each feature, . is the number of decision trees in the random forest; is the number of samples of the th decision tree; is the number of samples of the impurity of the parent node of the decision tree; the number of child nodes; the number of the first child node of the decision tree; the number of the first child node of the decision tree; the number of the first child node of the decision tree; the number of the first child node of the decision tree; the number of the first child node of the decision tree; the number of the first child node of the decision tree.

[0044] For the air spring exclusive signal derived features such as the variance of the capsule skin strain, the rate of change of the gas temperature, and the wavelet entropy of the strain signal, an enhancement coefficient is set, and the initial importance is converted into attention weight through a Softmax function, as shown in the following formula:

[0045] wherein, is the attention weight; is the initial importance of the air spring exclusive signal derived feature; is the enhancement coefficient. is the type coefficient of the first feature.

[0046] The constructed feature matrix is multiplied element by element with the attention weight to obtain a weighted feature matrix, as shown below:

[0047] wherein, is the weighted feature matrix.

[0048] In step S1025, the weighted feature matrix is input into a weighted decision tree, and the feature weights of the weighted feature matrix are used to feature selection is performed to ensure that the air spring exclusive features with high weights are more likely to participate in node splitting; and a weighted impurity is used to evaluate the splitting effect, and the formula is as follows:

[0049] wherein, is the weighted impurity, is the vehicle motion state category, 1 represents a static state, 2 represents an intense driving state, 3 represents a bumpy road condition state, 4 represents an emergency braking condition state, 5 represents a gentle running state, and 6 represents a load mutation state; is the node sample, and the optimal splitting threshold is determined by minimizing the weighted impurity. Each decision tree is grown until the number of node samples is less than a preset threshold or the node impurity is less than 0.1, and no pruning operation is performed to retain the distinguishing ability of the decision tree to the feature details.

[0050] Then, the output of each decision tree is the predicted motion state of the vehicle ​, calculate the voting weight of each decision tree For:

[0051] wherein, is the number of split features of the th decision tree; is the split feature set of the tree, reflecting the decision-making credibility of the decision tree based on high-weight features.

[0052] For each of the six vehicle states, the total voting weight obtained is counted, The class with the maximum total voting weight is selected as the final recognition result, and the fluctuation frequency, fluctuation amplitude and signal-to-noise ratio and other characteristic parameters corresponding to the state are output, providing the basis for subsequent dynamic weight filtering.

[0053] Step S103, according to the motion state and state characteristic parameters, dynamically assigning weight coefficients of different filtering modules to obtain filtered air spring pressure values.

[0054] A filter library composed of multiple filtering modules is constructed, the filtering modules at least including mean filtering sub-module, band-pass filtering sub-module, high-pass filtering sub-module, low-pass filtering sub-module, Kalman filtering sub-module; each filtering sub-module filters the air spring pressure original signal at the same time in parallel, and outputs the corresponding primary filtered value respectively; Based on the motion state and state characteristic parameters, a pre-defined weight allocation rule library is queried to dynamically obtain a set of weight coefficients matched with the current working condition; the set of weight coefficients contains weight coefficients corresponding to each filtering module one by one; The primary filtered values output by each filtering module are weighted and summed to calculate the filtered air spring pressure value, as follows:

[0055] wherein, is the filtered air spring pressure value; , , , , are weight coefficients respectively; , , , , are the primary filtered values output by each filtering sub-module.

[0056] Step S104, introduce pressure assignment deviation closed loop feedback mechanism, real-time acquisition of load data through wheel end spring on load sensor, combined with air spring effective area parameter, calculate the theoretical pressure value through mechanical equilibrium equation, as follows:

[0057] Wherein, is the theoretical pressure value; is the load data; is the air spring effective area.

[0058] The instantaneous deviation of the filtered air spring pressure value and the theoretical pressure value is as follows:

[0059] Wherein, e is the instantaneous deviation.

[0060] According to the instantaneous deviation, two levels of deviation threshold system are set, specifically, a slight deviation threshold and a significant deviation threshold are set. According to the threshold range of the obtained instantaneous deviation, the deviation level is judged. Wherein, the first level is slight deviation correction, only adjusting the weight coefficient of the filter module; the second level is significant deviation correction, synchronously correcting the weight coefficient and the core filter parameter.

[0061] Step S105, assign the pressure signal after dynamic weight filtering and closed loop feedback correction to the air spring, complete the pressure precise assignment.

[0062] Example two The embodiment discloses a pressure assignment system based on an open air spring system; As Figure 2 shown, a pressure assignment system based on an open air spring system, comprising: A signal acquisition module 201 is configured to: real-time acquisition of vehicle state signals and sensor signals of the open air spring system in the vehicle CAN network; A vehicle state recognition module 202 is configured to: multi-dimensional signal feature extraction is performed on the collected vehicle state signals and sensor signals, and the extracted multi-dimensional signal features are classified and processed by using an improved random forest algorithm, and the motion state and state characteristic parameters of the vehicle are output; A weight filtering module 203 is configured to: according to the motion state and state characteristic parameters, dynamically allocate the weight coefficients of different filter modules, and obtain the filtered air spring pressure value; A closed loop feedback correction module 204 is configured to: introduce pressure assignment deviation closed loop feedback mechanism, real-time calculation of the deviation between the filtered air spring pressure value and the theoretical pressure value corresponding to the actual load of the air spring, dynamic correction of the weight coefficient and filter parameter of the filter module according to the deviation value; The pressure assignment module 205 is configured to assign the pressure signal filtered by the dynamic weight and corrected by the closed loop feedback to the air spring, thereby completing pressure assignment.

[0063] Embodiment three An object of this embodiment is to provide a computer-readable storage medium.

[0064] A computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the pressure assignment method based on an open air spring system according to embodiment 1.

[0065] Embodiment four An object of this embodiment is to provide an electronic device.

[0066] An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor implements the steps of the pressure assignment method based on an open air spring system according to embodiment 1 when executing the program.

[0067] The steps and methods involved in the above embodiments two, three and four correspond to embodiment one, and the specific embodiments can be seen from the relevant description of embodiment one. The term "computer-readable storage medium" should be understood to include a single medium or multiple media of one or more instruction sets; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and causing the processor to perform any of the methods of the present application.

[0068] Those skilled in the art should understand that the above-mentioned modules or steps of the present application can be realized by a general computer device, alternatively, they can be realized by program codes executable by a computing device, so that they can be stored in a storage device and executed by a computing device, or they can be made into individual integrated circuit modules, or a plurality of modules or steps can be made into a single integrated circuit module. The present application is not limited to any specific combination of hardware and software.

[0069] The above describes the specific embodiments of the present application in conjunction with the accompanying drawings, but is not a limitation on the scope of protection of the present application. Those skilled in the art should understand that various modifications or variations made by those skilled in the art on the basis of the technical solutions of the present application without creative labor are still within the scope of protection of the present application.

Claims

1. A method for assigning pressure values ​​based on an open spring system, characterized in that, include: Real-time acquisition of vehicle status signals from the vehicle's CAN network and sensor signals from the open air spring system; Multi-dimensional signal features are extracted from the collected vehicle state signals and sensor signals. The extracted multi-dimensional signal features are then classified using an improved random forest algorithm to output the vehicle's motion state and state feature parameters. Based on the motion state and state characteristic parameters, the weight coefficients of different filtering modules are dynamically allocated to obtain the filtered air spring pressure value. A closed-loop feedback mechanism for pressure assignment deviation is introduced to calculate the deviation between the filtered air spring pressure value and the theoretical pressure value corresponding to the actual load of the air spring in real time, and dynamically adjust the weight coefficient and filtering parameters of the filtering module based on the deviation value. The pressure signal, after dynamic weight filtering and closed-loop feedback correction, is assigned to the air spring to complete the precise pressure assignment.

2. The pressure assignment method based on an open spring system as described in claim 1, characterized in that, The vehicle status signals include: accelerator pedal signal, brake pedal signal, steering angle signal, steering angular velocity signal, longitudinal acceleration signal, lateral acceleration signal, and yaw rate signal; The sensor signals include: air spring bladder strain signal, air spring gas temperature signal, and solenoid valve switch status signal.

3. The pressure assignment method based on an open spring system as described in claim 1, characterized in that, The multi-dimensional signal feature extraction of the collected vehicle status signals and sensor signals includes: The collected vehicle status signals and sensor signals are preprocessed, including outlier removal, signal smoothing, and standardization. Multi-dimensional signal features are extracted from the preprocessed vehicle state signal and sensor signal. These multi-dimensional signal features include time-domain features, frequency-domain features, and time-frequency domain joint features.

4. The pressure assignment method based on an open spring system as described in claim 1, characterized in that, The extracted multi-dimensional signal features are classified using an improved random forest algorithm, outputting the vehicle's motion state and state feature parameters, including: A feature matrix is ​​constructed based on the extracted multi-dimensional signal features and then normalized; a label matrix is ​​constructed based on the vehicle's motion state. The normalized feature matrix is ​​input into the random forest algorithm to obtain the initial importance of each feature; the initial importance is transformed into attention weights based on the multi-dimensional features of the sensor signals; the constructed feature matrix is ​​multiplied element by element with the attention weights to obtain a weighted feature matrix, which enhances the contribution of the features of the air spring system to classification; The weighted feature matrix is ​​input into a weighted decision tree. Each decision tree outputs the predicted motion state of the vehicle. The total voting weight of the different vehicle motion states is calculated. The category with the largest total voting weight is selected as the final recognition result of the vehicle motion state. At the same time, the state feature parameters corresponding to the state are output.

5. The pressure assignment method based on an open spring system as described in claim 4, characterized in that, The vehicle motion states include stationary state, aggressive driving state, bumpy road condition state, emergency braking state, smooth running state, and sudden load change state. The state characteristic parameters include fluctuation frequency, fluctuation amplitude, and signal-to-noise ratio.

6. The pressure assignment method based on an open spring system as described in claim 1, characterized in that, Based on the motion state and state characteristic parameters, the weight coefficients of different filtering modules are dynamically allocated to obtain the filtered air spring pressure value, including: A filter library consisting of multiple filter modules is constructed. The filter modules include at least a mean filter submodule, a bandpass filter submodule, a high-pass filter submodule, a low-pass filter submodule, and a Kalman filter submodule. Each of the filter submodules performs parallel filtering on the original air spring pressure signal at the same time and outputs the corresponding primary filter value. Based on the motion state and state characteristic parameters, a predefined weight allocation rule library is queried to dynamically obtain a set of weight coefficients that match the current working condition; the set of weight coefficients contains weight coefficients that correspond one-to-one with each filtering module. The primary filter values ​​output by each filter module are weighted and summed with their corresponding weight coefficients to calculate the filtered air spring pressure value.

7. The pressure assignment method based on an open spring system as described in claim 1, characterized in that, A closed-loop feedback mechanism for pressure assignment deviation is introduced to calculate the deviation between the filtered air spring pressure value and the theoretical pressure value corresponding to the actual load of the air spring in real time. The weighting coefficients and filtering parameters of the filtering module are dynamically adjusted based on the deviation value, including: Load data is acquired in real time by a load sensor on the wheel end spring, and the theoretical pressure value corresponding to the actual load of the air spring is generated based on the load data. Calculate the instantaneous deviation between the filtered air spring pressure value and the theoretical pressure value. Based on the instantaneous deviation, set a two-level deviation threshold system and determine the deviation level according to the interval in which the mean deviation value is located. Adjust the weight coefficients and / or core filtering parameters of the filtering module according to different deviation levels.

8. A pressure assignment system based on an open spring system, characterized in that: include: The signal acquisition module is configured to: acquire vehicle status signals and sensor signals from the open air spring system in the vehicle's CAN network in real time; The vehicle state recognition module is configured to: extract multi-dimensional signal features from the collected vehicle state signals and sensor signals, classify the extracted multi-dimensional signal features using an improved random forest algorithm, and output the vehicle's motion state and state feature parameters. The weighted filtering module is configured to dynamically allocate weight coefficients of different filtering modules according to the motion state and state characteristic parameters to obtain the filtered air spring pressure value. The closed-loop feedback correction module is configured to: introduce a pressure assignment deviation closed-loop feedback mechanism, calculate in real time the deviation between the filtered air spring pressure value and the theoretical pressure value corresponding to the actual load of the air spring, and dynamically correct the weight coefficient and filtering parameters of the filtering module based on the deviation value. The pressure assignment module is configured to assign the pressure signal, after dynamic weight filtering and closed-loop feedback correction, to the air spring to complete the pressure assignment.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the pressure assignment method based on an open spring system as described in any one of claims 1-7.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the pressure assignment method based on an open spring system as described in any one of claims 1-7.

Images