Intelligent swing arm adjusting and detecting system for automobile driving

Abstract

The invention provides an intelligent swing arm adjusting and detecting system for automobile driving. The intelligent swing arm adjusting and detecting system comprises a multi-source sensing module used for collecting vibration signals, stress distribution and displacement data of a swing arm in real time; the attitude resolving module receives the sensing data and constructs a three-dimensional dynamic swing arm model; the dynamic compensation module calculates attitude deviation based on the model and generates compensation parameters; the execution mechanism module drives the swing arm to adjust according to the compensation parameters; the abnormity diagnosis module analyzes the dynamic response spectrum to determine the fault type; the self-calibration module corrects the drift error of the multi-source sensing data; the control output module integrates the compensation parameter and the diagnosis result to generate a control signal. The stability and reliability of the automobile driving system can be improved, the maintenance cost is reduced, and the breakdown time is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of automotive drive, and more specifically, the present invention relates to an intelligent swing arm adjustment and detection system for automotive drive. Background Art

[0002] In an automotive drive system, the swing arm is a key component, and its performance directly affects the driving stability and controllability of the vehicle. Traditional swing arm adjustment and detection methods mainly rely on manual experience or simple sensor feedback, making it difficult to achieve real-time and precise adjustment of the swing arm attitude and rapid diagnosis of faults. These methods can usually only detect obvious faults, but it is difficult to detect minor deviations and potential faults, resulting in possible instability during vehicle driving and even safety hazards. In addition, traditional systems have poor adaptability under complex working conditions and cannot automatically calibrate sensor data according to environmental changes, affecting the detection accuracy.

[0003] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: it is impossible to achieve dynamic real-time adjustment of the swing arm, the fault diagnosis accuracy is insufficient, and there is a lack of self-calibration function, making it difficult to meet the requirements of modern vehicles for high-performance and high-reliability drive systems. Summary of the Invention

[0004] The present invention provides an intelligent swing arm adjustment and detection system for automotive drive, including:

[0005] A multi-source sensing module configured to collect multi-source sensing data in real time, where the multi-source sensing data includes vibration signals, stress distribution, and displacement data of the swing arm;

[0006] An attitude solution module that receives the multi-source sensing data and constructs a three-dimensional dynamic swing arm model;

[0007] A dynamic compensation module that calculates attitude deviations based on the three-dimensional dynamic swing arm model and generates compensation parameters;

[0008] An actuator module that drives the swing arm to perform dynamic adjustment according to the compensation parameters;

[0009] An anomaly diagnosis module that determines the fault type by analyzing the dynamic response spectrum of the swing arm and generates a diagnosis result;

[0010] A self-calibration module that corrects the drift error of the multi-source sensing data according to environmental parameters;

[0011] A control output module that integrates the compensation parameters and the diagnosis result to generate a control signal.

[0012] Further, the attitude solution module includes:

[0013] The data fusion unit receives vibration signals, stress distributions, and displacement data from the multi-source sensing module and constructs a spatio-temporal state vector:

[0014]

[0015] Among them, is the angular acceleration, ∈ x is the strain value in the x-axis direction, and Δy is the displacement deviation;

[0016] The three-dimensional reconstruction unit establishes a swing arm dynamics model based on the spatio-temporal state vector:

[0017]

[0018] Among them, M is the mass matrix, C is the damping matrix, K is the stiffness matrix, u is the displacement vector, and F(t) is the external excitation vector;

[0019] The deviation analysis unit calculates the residual between the predicted displacement u model and the measured displacement u meas :

[0020] E = ||u model u meas ||2

[0021] Furthermore, the dynamic compensation module performs:

[0022] Construct an adaptive PID controller, and its gain parameter adjustment rule is:

[0023] K p = 0.5 + 0.2E, K i = 0.1e 0.5E , K d = 0.3(1e 2E )

[0024] Among them, K p is the proportional gain coefficient, K i is the integral gain coefficient, K d is the differential gain coefficient, and E is the residual output by the attitude solution module;

[0025] Generate a compensation torque command:

[0026]

[0027] Among them, e is the attitude error vector, and e is equal to the difference between the target displacement u target and the current displacement u current ;

[0028] When the residual E is greater than 0.1, activate the multi-level compensation strategy:

[0029] Superimposed feedforward control term where M is the mass matrix, is the desired angular acceleration;

[0030] Trigger the overload protection mode of the actuator.

[0031] Furthermore, the abnormal diagnosis module executes:

[0032] Extract the time-frequency characteristics of the swing arm vibration signal and calculate the wavelet packet energy entropy H:

[0033]

[0034] where, E j is the wavelet energy of the j-th frequency band, p j is the energy proportion of the j-th frequency band;

[0035] Construct the feature vector V = [H, σ ∈ , f peak T , where σ ∈ is the strain mean square deviation, f peak is the resonance main frequency;

[0036] Adopt an improved SVM classifier for fault diagnosis, and the kernel function is defined as:

[0037]

[0038] where, x i and x j are the input feature vectors, γ is the kernel function width parameter, and d is the polynomial order;

[0039] When the classification confidence is lower than 0.7, activate the redundant sensor cross-validation process.

[0040] Furthermore, the self-calibration module executes:

[0041] Real-time monitor the environmental temperature T env and humidity H, and calculate the drift compensation amount Δ cal :

[0042] Δ cal = 0.3(T env - 40)+0.15(H - 60)

[0043] Correct the sensor output:

[0044] y corrected = y raw ×(1 + Δ cal ·e^t / 200 )

[0045] ​Among them, y raw is the original sensing data, and t is the calibration duration;

[0046] When Δ cal the absolute value is greater than 0.5, start the online calibration program of the sensor array.

[0047] Furthermore, the overload protection mode performs:

[0048] Calculate the safety threshold τ max :

[0049]

[0050] Among them, τ rated is the rated torque of the actuator;

[0051] When τ c is greater than τ max perform gradient clipping:

[0052]

[0053] Among them, tanh is the hyperbolic tangent function, and τ′ c is the compensated torque after clipping.

[0054] Furthermore, the redundant sensor cross - validation process includes:

[0055] Synchronously collect the data of the three - axis accelerometer and the MEMS gyroscope;

[0056] Calculate the data consistency index R:

[0057]

[0058] Among them, a k is the accelerometer data, and g k is the integrated acceleration of the gyroscope;

[0059] When R is greater than 0.3, determine that the sensor is faulty and switch to the backup channel.

[0060] Furthermore, the stiffness matrix update method includes:

[0061] Dynamically correct the stiffness parameter K based on the strain energy density ij :

[0062]

[0063] Among them, K0 is the initial stiffness, and ε is the strain value;

[0064] When material fatigue is detected, activate the nonlinear stiffness model K nl :

[0065]

[0066] Among them, εpeak,n is the peak strain of the nth high-strain event, and N is the cumulative number of high-strain events.

[0067] Furthermore, the control output module executes:

[0068] Generate a pulse-width modulation signal DPWM to drive the actuator:

[0069] D PWM = 50 + 30 · sigmoid(10(τ c 0.5τ max ))

[0070] Among them, τ max is the maximum safe torque;

[0071] When the detected communication delay exceeds 20 ms, switch to the predictive control mode:

[0072]

[0073] Among them, Δt is the delay time.

[0074] Furthermore, the on-line calibration program includes:

[0075] Apply a known excitation force F cal :

[0076] α newcal = 100sin(2π5t)

[0077] Calculate the sensor response error δ:

[0078]

[0079] Among them, y meas is the measured response data, and y theory is the theoretical response data;

[0080] When δ is greater than 0.1, update the sensor calibration coefficient:

[0081] α new = α old ·(10.2δ)

[0082] Among them, α new is the new calibration coefficient, and α old is the old calibration coefficient.

[0083] The above embodiments of the present invention have at least the following beneficial effects: The intelligent swing arm adjustment and detection system of the present invention can achieve real-time dynamic monitoring and adjustment of the vehicle swing arm. By collecting vibration signals, stress distribution, and displacement data through a multi-source sensing module, and combining with an attitude calculation module to construct a three-dimensional dynamic swing arm model, it can accurately calculate the attitude deviation and generate compensation parameters, thereby driving the actuator for dynamic adjustment. This real-time adjustment ability can improve the stability and handling performance of the vehicle during driving, especially effectively reducing the abnormal vibration and stress accumulation of the swing arm under complex road conditions and extending the service life of the swing arm. At the same time, the abnormal diagnosis module can analyze the dynamic response spectrum of the swing arm, quickly determine the type of fault and generate a diagnosis result, improving the accuracy and efficiency of fault diagnosis.

[0084] In addition, the self-calibration module can correct the drift error of the multi-source sensing data according to environmental parameters, ensure the accuracy of the sensor data, and further improve the reliability and adaptability of the system. The control output module integrates the compensation parameters and the diagnosis result to generate a control signal, realizing precise driving of the actuator. Even under adverse conditions such as communication delay, the system can maintain stable operation through the predictive control mode. The integration of these functions enables the system of the present invention not only to improve the overall performance of the vehicle drive system, but also to reduce maintenance costs and enhance the safety and reliability of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, wherein:

[0086] Figure 1 It is a schematic structural diagram of an intelligent swing arm adjustment and detection system for vehicle drive provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0087] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and then implement the present invention, rather than to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to be able to convey the scope of the present invention fully to those skilled in the art.

[0088] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, equipment, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: completely hardware, completely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0089] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction without any restrictive meaning.

[0090] Reference is now made to Figure 1 , Figure 1 which is a schematic structural diagram of an intelligent swing arm adjustment and detection system for vehicle drive provided by an embodiment of the present invention. As Figure 1 shown, an intelligent swing arm adjustment and detection system 100 for vehicle drive includes:

[0091] A multi-source sensing module 101 configured to collect multi-source sensing data in real time, where the multi-source sensing data includes vibration signals, stress distributions, and displacement data of the swing arm;

[0092] An attitude solution module 102 that receives the multi-source sensing data and constructs a three-dimensional dynamic swing arm model;

[0093] A dynamic compensation module 103 that calculates attitude deviations based on the three-dimensional dynamic swing arm model and generates compensation parameters;

[0094] An actuator module 104 that drives the swing arm to perform dynamic adjustment according to the compensation parameters;

[0095] An anomaly diagnosis module 105 that determines the type of fault by analyzing the dynamic response spectrum of the swing arm and generates a diagnosis result;

[0096] A self-calibration module 106 that corrects the drift error of the multi-source sensing data according to environmental parameters;

[0097] A control output module 107 that integrates the compensation parameters and the diagnosis result to generate a control signal.

[0098] It should be noted that the system includes a multi-source sensing module configured to collect multi-source sensing data in real time, and these data include vibration signals, stress distributions, and displacement data of the swing arm. The multi-source sensing module is the basic data source of the system. Through the collaborative work of multiple sensors, it can comprehensively capture various physical state information of the swing arm during operation. The vibration signal reflects the vibration frequency and amplitude of the swing arm during dynamic operation, the stress distribution reveals the magnitude and direction of the force borne by each part of the swing arm, and the displacement data directly reflects the movement trajectory and position change of the swing arm. The real-time collection of these data provides accurate basic information for subsequent analysis and adjustment.

[0099] Specifically, an accelerometer can be used as the vibration sensor in the multi-source sensing module. It can capture the vibration signals of the swing arm with high precision and high sampling rate. For example, the sampling rate can be set to 1000Hz to ensure that high-frequency vibration characteristics can be captured. A strain gauge can be used as the stress sensor. By attaching it to the key parts of the swing arm, the strain value can be measured in real time, and then the stress distribution can be calculated. A laser displacement sensor or a magnetostrictive displacement sensor can be used as the displacement sensor, with a measurement accuracy reaching the micron level, which can accurately record the displacement changes of the swing arm. The data of these sensors are synchronously collected through a data acquisition card and transmitted to the subsequent modules of the system in the form of digital signals for processing. The frequency range of the vibration signal can cover from 0.1Hz at low frequency to 1000Hz at high frequency. The measurement range of the stress distribution can be set according to the stress limit of the swing arm material, and the measurement range of the displacement data is adjusted according to the actual movement range of the swing arm.

[0100] Preferably, the sensors in the multi-source sensing module can be optimally configured according to different application scenarios. For example, in an automotive drive system with high-precision requirements, redundant sensors can be added to improve the reliability and accuracy of the data. For the vibration sensor, a multi-axis accelerometer can be used, which can not only measure the vibration of the swing arm in a single direction but also capture the vibration in multiple directions, thus more comprehensively reflecting the dynamic behavior of the swing arm. The arrangement position of the stress sensor can be optimized according to the structural characteristics of the swing arm. For example, the sensor density can be increased in the stress concentration area to more accurately monitor the stress distribution. In addition, the sampling frequency of the displacement sensor can be dynamically adjusted according to the movement speed of the swing arm to ensure accurate measurement of displacement changes during high-speed movement.

[0101] In some embodiments, the attitude solution module includes:

[0102] A data fusion unit that receives the vibration signal, stress distribution, and displacement data from the multi-source sensing module and constructs a spatio-temporal state vector:

[0103]

[0104] where, is the angular acceleration, ∈ x is the strain value in the x-axis direction, and Δy is the displacement deviation;

[0105] A three-dimensional reconstruction unit that establishes a swing arm dynamics model based on the spatio-temporal state vector:

[0106]

[0107] where M is the mass matrix, C is the damping matrix, K is the stiffness matrix, u is the displacement vector, and F(t) is the external excitation vector;

[0108] The deviation analysis unit calculates the predicted displacement u of the model model and the measured displacement u meas of the residual:

[0109] E = ||u model u meas ||2

[0110] It should be noted that the attitude solution module is one of the core parts of this system. Its main function is to receive the vibration signals, stress distributions, and displacement data collected by the multi-source sensing module, and construct a three-dimensional dynamic model of the swing arm based on these data. The data fusion unit provides a basis for subsequent model construction by integrating different types of sensing data into a spatio-temporal state vector. The three-dimensional reconstruction unit uses the dynamic equation to establish the dynamic model of the swing arm and further analyze its dynamic behavior. The deviation analysis unit evaluates the accuracy of the model and the actual state of the swing arm by calculating the residual between the predicted displacement of the model and the measured displacement. These steps together ensure that the system can accurately solve the attitude of the swing arm and provide a basis for subsequent dynamic compensation.

[0111] Specifically, the sensing data received by the data fusion unit includes angular acceleration (α¨), x-axis strain value (εx), and displacement deviation (δ). Among them, the angular acceleration is measured by a gyroscope, which reflects the rotational acceleration of the swing arm; the strain value is measured by a strain sensor, which reflects the deformation degree of the swing arm when it is stressed; the displacement deviation is measured by a displacement sensor, which represents the difference between the actual position and the theoretical position of the swing arm. The construction of the spatio-temporal state vector integrates these data into a unified mathematical representation form for subsequent processing. In the three-dimensional reconstruction unit, the establishment of the dynamic model is based on the classical second-order linear system equation, where the mass matrix (M), damping matrix (C), and stiffness matrix (K) are important parameters describing the dynamic characteristics of the swing arm. The mass matrix represents the mass distribution of the swing arm, the damping matrix reflects the energy dissipation characteristics, and the stiffness matrix is related to the elastic characteristics of the swing arm. The external excitation vector (F(t)) represents the external force acting on the swing arm, such as the impact force caused by road unevenness. By setting these parameters, the dynamic response of the swing arm during actual operation can be accurately simulated.

[0112] Preferably, in order to improve the accuracy and reliability of attitude calculation, data filtering techniques, such as Kalman filtering, can be introduced into the data fusion unit to reduce the influence of sensor noise on the spatio-temporal state vector. For the three-dimensional reconstruction unit, the parameters of the mass matrix, damping matrix, and stiffness matrix can be accurately calibrated according to the actual structure and material properties of the swing arm. For example, the mass matrix can be obtained by accurately measuring the mass distribution of the swing arm; the damping matrix can be determined by experimentally testing the damping characteristics of the swing arm; and the stiffness matrix can be calculated based on the material elastic modulus and structural shape of the swing arm. In addition, in order to adapt to complex dynamic environments, an adaptive algorithm can be introduced to dynamically adjust the parameters of these matrices. For example, during the operation of the swing arm, if material fatigue or structural damage is detected, the stiffness matrix can be updated in real time to more accurately reflect the current state of the swing arm.

[0113] In some embodiments, the dynamic compensation module performs:

[0114] Construct an adaptive PID controller, and its gain parameter adjustment rule is:

[0115] K p = 0.5 + 0.2E, K i = 0.1e 0.5E , K d = 0.3(1e 2E )

[0116] Where K p is the proportional gain coefficient, K i is the integral gain coefficient, K d is the derivative gain coefficient, and E is the residual output by the attitude calculation module;

[0117] Generate a compensation torque command:

[0118]

[0119] Where e is the attitude error vector, and e is equal to the difference between the target displacement u target and the current displacement u current ;

[0120] When the residual E is greater than 0.1, activate the multi-level compensation strategy:

[0121] Superimpose the feedforward control term Where M is the mass matrix, is the desired angular acceleration;

[0122] Trigger the overload protection mode of the actuator.

[0123] It should be noted that the dynamic compensation module is a key part of this system for calculating the attitude deviation according to the three-dimensional dynamic swing arm model and generating compensation parameters. Its core function is to construct an adaptive PID controller, dynamically adjust the gain parameters of the controller according to the residuals output by the attitude solution module, and generate a compensation torque command to drive the swing arm for dynamic adjustment. When the residual exceeds the set threshold, the module will activate a multi-level compensation strategy, including superimposing a feedforward control term and triggering the overload protection mode of the actuator, to ensure the stability and safety of the system. The design of this module aims to optimize the operating attitude of the swing arm through precise dynamic compensation and improve the overall performance of the system.

[0124] Specifically, the adjustment rule of the gain parameters of the adaptive PID controller in the dynamic compensation module is dynamically calculated according to the residual E output by the attitude solution module. The proportional gain coefficient (K p ), integral gain coefficient (K i ), and derivative gain coefficient (K d ) are adjusted through the formulas

[0125] K P = 0.5 + 0.2E, K i = 0.1 - 0.5E, K d = 0.3(1 - 2E)

[0126] respectively. The setting of these parameters enables the controller to dynamically optimize the control effect according to the actual deviation. The compensation torque command τ is calculated through the formula τ = K p * e + K i * ∫e + K d * de / dt, where e is the difference between the target displacement and the current displacement, that is, the attitude error vector. When the residual E is greater than 0.1, the system will activate a multi-level compensation strategy, including superimposing a feedforward control term τff = 0.8 * M * α¨des, where M is the mass matrix and α¨des is the desired angular acceleration. At the same time, trigger the overload protection mode of the actuator to prevent the actuator from being damaged due to overload.

[0127] Preferably, the adaptive PID controller in the dynamic compensation module can be further optimized according to the actual application scenario. For example, a fuzzy logic control algorithm can be introduced and combined with PID control to better handle nonlinear and uncertain problems. For the adjustment rule of the gain parameters, it can be fine-tuned according to the actual test data to adapt to different working conditions and swing arm characteristics. In addition, the feedforward control term in the multi-level compensation strategy can be designed in combination with a more complex dynamic model to improve the compensation accuracy. The safety threshold calculation formula in the overload protection mode can be further refined, for example, dynamically adjusted according to the actual working temperature and load conditions of the actuator to improve the reliability and adaptability of the system.

[0128] In some embodiments, the anomaly diagnosis module performs the following:

[0129] Extract the time-frequency features of the swing arm vibration signal and calculate the wavelet packet energy entropy H:

[0130]

[0131] where E j is the wavelet energy of the j-th frequency band, and p j is the energy proportion of the j-th frequency band;

[0132] Construct the feature vector V = [H, σ ∈ , f peak , T where σ ∈ is the mean square deviation of strain, and f peak is the main resonance frequency;

[0133] Adopt an improved SVM classifier for fault diagnosis, and the kernel function is defined as:

[0134]

[0135] where x i and x j are input feature vectors, γ is the kernel function width parameter, and d is the polynomial order;

[0136] When the classification confidence is lower than 0.7, activate the redundant sensor cross-validation process.

[0137] It should be noted that the anomaly diagnosis module is the key part of this system for determining the fault type and generating a diagnosis result by analyzing the dynamic response spectrum of the swing arm. Its core function is to extract the time-frequency features of the swing arm vibration signal, construct a feature vector, and use an improved support vector machine (SVM) classifier for fault diagnosis. When the classification confidence is lower than the set threshold, the module will activate the redundant sensor cross-validation process to ensure the accuracy of the diagnosis result. The design of this module aims to quickly identify potential problems in the operation of the swing arm through accurate fault diagnosis, thereby improving the reliability and maintenance efficiency of the system.

[0138] Specifically, the anomaly diagnosis module first extracts the time-frequency features of the swing arm vibration signal and calculates the wavelet packet energy entropy H. The wavelet packet energy entropy is a feature extraction method based on wavelet transform, which can effectively reflect the time-frequency distribution characteristics of the vibration signal. Among them, the wavelet energy refers to the energy distribution in a specific frequency band, and the energy proportion is the ratio of the energy of this frequency band to the total energy. The feature vector consists of the mean square deviation of strain (σ 2 ), the wavelet packet energy entropy (H), and the main resonance frequency (f peakIt consists of these parameters, which can comprehensively reflect the dynamic characteristics of the swing arm. The improved SVM classifier uses specific kernel functions, such as the combination of the radial basis function (RBF) and the polynomial kernel function. The kernel function width parameter (γ) and the polynomial order (d) can be optimized according to the actual data. When the classification confidence is lower than 0.7, the system will activate the redundant sensor cross-validation process to verify the reliability of the diagnosis result by comparing the data consistency of different sensors.

[0139] Preferably, the calculation of the wavelet packet energy entropy in the abnormal diagnosis module can be optimized by selecting a more suitable wavelet basis function, such as the Daubechies wavelet or the Morlet wavelet, to better adapt to the characteristics of the swing arm vibration signal. For the construction of the feature vector, in addition to the existing parameters, other relevant features, such as the kurtosis or short-time energy of the vibration signal, can be introduced to further improve the accuracy of the fault diagnosis. In the SVM classifier, the selection and parameter optimization of the kernel function can be adjusted by methods such as grid search and cross-validation to ensure the robustness of the classifier under different working conditions.

[0140] Furthermore, the redundant sensor cross-validation process can introduce more types of sensors, such as fiber optic sensors or piezoelectric sensors, to provide more comprehensive data support. When the data consistency index R is greater than 0.3, not only can the sensor failure be determined, but also the system can automatically switch to the backup sensor channel through an intelligent algorithm to further improve the reliability of the system.

[0141] In some embodiments, the self-calibration module performs:

[0142] Monitor the ambient temperature T env and humidity H in real time, and calculate the drift compensation amount Δ cal :

[0143] Δ cal = 0.3(T env - 40)+0.15(H - 60)

[0144] Correct the sensor output:

[0145] y corrected = y raw ×(1 + Δ cal ·e t / 200 )

[0146] where y raw is the original sensing data, and t is the calibration duration;

[0147] When the absolute value of Δ cal is greater than 0.5, start the online calibration program of the sensor array.

[0148] It should be noted that the self - calibration module is a key part of this system for correcting the drift error of multi - source sensing data. Its main function is to monitor the ambient temperature and humidity in real - time, calculate the drift compensation amount based on these environmental parameters, and then correct the sensor output data. The design of this module aims to ensure the accuracy and stability of sensor data through environmental factor compensation, thereby improving the reliability and long - term operating performance of the entire system.

[0149] Specifically, the self - calibration module calculates the drift compensation amount (ΔC) by monitoring the ambient temperature (T) and humidity (H) in real - time. The calculation formula for the drift compensation amount is

[0150] ΔC = 0.3(T - 40)+0.15(H - 60)

[0151] where T and H represent the current ambient temperature and humidity respectively. This formula reflects the influence degree of temperature and humidity changes on the sensor output. The corrected sensor output data is calculated by C corr = C raw ×(1 + ΔC×t / 200), where C raw is the original sensing data and t is the calibration duration. When the absolute value of the drift compensation amount is greater than 0.5, the system will start the online calibration program of the sensor array to further calibrate the sensor. This process ensures that the output data of the sensor is always in a calibrated state under different environmental conditions.

[0152] Preferably, the temperature and humidity monitoring in the self - calibration module can be achieved by high - precision temperature and humidity sensors. For example, a digital temperature and humidity sensor can be used, and its measurement accuracy can reach ±0.5℃ and ±2%RH. For the calculation formula of the drift compensation amount, it can be adjusted according to the actual application scenario. For example, in extreme environmental conditions, the weight coefficients of temperature and humidity can be increased to more accurately reflect the influence of environmental changes on the sensor. In the online calibration program, the applied known excitation force can be a sine - wave signal, such as F cal = 100sin(2π×5t), and the calibration coefficient is updated by calculating the sensor response error. In addition, in order to improve the adaptability of the system, an adaptive algorithm can be introduced to dynamically adjust the calculation formula of the drift compensation amount according to historical data, thereby further optimizing the calibration effect.

[0153] In some embodiments, the overload protection mode performs:

[0154] Calculate the safety threshold τ max :

[0155]

[0156] where τ rated is the rated torque of the actuator;

[0157] When τ c is greater than τ max , perform gradient clipping:

[0158]

[0159] where tanh is the hyperbolic tangent function, and τ′ c is the compensated torque after clipping.

[0160] It should be noted that the overload protection mode is an important functional module in this system to ensure the operation of the actuator within a safe range. Its core function is to calculate the safety threshold and clip the output torque when the output torque of the actuator exceeds this threshold. This process can effectively prevent the actuator from being damaged due to overload and ensure the stability and reliability of the system.

[0161] Specifically, the overload protection mode first calculates the safety threshold τ max , and its formula is τ max = 0.8τ rated ×(1 - 0.05|Δτ| / τ rated ), where τ rated is the rated torque of the actuator, and Δτ is the deviation between the current torque and the rated torque. When the actual output torque τ exceeds the safety threshold τ max , the system will activate the gradient clipping strategy to limit the output torque within a safe range. The compensated torque τ′ after clipping is calculated by the formula τ′ = τ max ×tanh(τ / τ max ), where tanh is the hyperbolic tangent function. This clipping strategy can minimize the impact on the dynamic performance of the system while ensuring the safety of the actuator.

[0162] Preferably, the calculation of the safety threshold in the overload protection mode can be adjusted according to the actual application scenario. For example, in high-load working conditions, the conservativeness of the safety threshold can be appropriately reduced to improve the utilization rate of the system. Specifically, the safety factor can be adjusted from 0.8 to 0.7 to meet higher load requirements.

[0163] Furthermore, the hyperbolic tangent function in the clipping strategy can also be optimized according to the dynamic characteristics of the system. For example, a piecewise linear function or an adaptive clipping algorithm can be introduced to better balance safety and dynamic response. In practical applications, the safety threshold can also be dynamically adjusted according to the working temperature of the actuator by combining the data of the temperature sensor, thereby further improving the adaptability and reliability of the system.

[0164] In some embodiments, the redundant sensor cross-validation process includes:

[0165] Synchronously collect data from a three-axis accelerometer and a MEMS gyroscope;

[0166] Calculate the data consistency index R:

[0167]

[0168] where a k is the accelerometer data, and g k is the integrated acceleration of the gyroscope;

[0169] When R is greater than 0.3, determine that the sensor is faulty and switch to the backup channel.

[0170] It should be noted that the redundant sensor cross-verification process is a key link in this system for improving the reliability of sensor data and the accuracy of fault diagnosis. Its core function is to synchronously collect data from a three-axis accelerometer and a MEMS gyroscope, and calculate the data consistency index R to determine whether the sensor is faulty and switch to the backup channel. This process can quickly switch to the redundant channel when the sensor fails, ensuring the continuous operation of the system and the accuracy of the data.

[0171] Specifically, in the redundant sensor cross-verification process, the three-axis accelerometer is used to measure the acceleration data of the swing arm, while the MEMS gyroscope provides the angular velocity data. Through integral processing, the gyroscope data can be converted into an acceleration form for comparison with the accelerometer data. The calculation formula for the data consistency index R is

[0172] R = (1 / 3)∑|a i - b i | / max(|a i |, |b i |)

[0173] where a i is the accelerometer data, and b i is the integrated acceleration data of the gyroscope. When R is greater than 0.3, the system determines that the sensor data is inconsistent and there may be a fault, thus triggering the sensor switching mechanism. This process ensures that the backup channel can be switched to in a timely manner when the sensor fails, avoiding system failure caused by a single sensor fault.

[0174] Preferably, the threshold of the data consistency index R in the redundant sensor cross - verification process can be adjusted according to the actual application scenario. For example, in the occasion with high - precision requirements, the threshold can be reduced from 0.3 to 0.2 to improve the sensitivity of fault detection. In addition, to further improve the reliability of the system, more types of redundant sensors can be introduced, such as fiber optic sensors or piezoelectric sensors, and a multi - level redundant switching mechanism can be designed. For example, when the main sensor fails, it is first switched to the standby sensor of the same type; if the standby sensor also fails, it is switched to a sensor of a different type but with similar functions. At the same time, a self - diagnosis algorithm can be introduced to verify the sensor data after switching in real - time to ensure its reliability.

[0175] In some embodiments, the stiffness matrix update method includes:

[0176] Dynamically modify the stiffness parameter K based on the strain energy density ij :

[0177]

[0178] where K0 is the initial stiffness and ε is the strain value;

[0179] When material fatigue is detected, activate the non - linear stiffness model K nl :

[0180]

[0181] where εpeak,n is the peak strain of the nth high - strain event and N is the cumulative number of high - strain events.

[0182] It should be noted that the stiffness matrix update method is the key technology in this system for dynamically adjusting the stiffness parameters of the swing arm. Its core function is to dynamically modify the stiffness parameters based on the strain energy density to adapt to possible material fatigue or structural damage of the swing arm during operation. When a high - strain event is detected, the system activates the non - linear stiffness model to further optimize the calculation of the stiffness parameters. This method can ensure that the dynamic model of the swing arm remains accurate and reliable during long - term operation, thereby improving the overall performance of the system.

[0183] Specifically, the stiffness matrix update method dynamically modifies the stiffness parameters through the strain energy density, and the formula is

[0184]

[0185] where K0 is the initial stiffness and ε is the strain value, Denotes the partial derivative of the strain energy density with respect to strain. This formula reflects the influence of the change in strain energy density on stiffness and can dynamically adjust the stiffness parameters to adapt to the actual operating state of the swing arm. When material fatigue is detected, the system activates the nonlinear stiffness model, and the formula is

[0186]

[0187] where is the ratio of the peak strain of the nth high-strain event to the material's ultimate strain, and N is the cumulative number of high-strain events. This nonlinear model can more accurately describe the influence of material fatigue on stiffness, thereby improving the adaptability and reliability of the system.

[0188] Preferably, the initial stiffness K0 in the stiffness matrix update method can be obtained through experimental calibration. For example, by measuring the deformation of the swing arm under different loads through a static loading experiment, and then calculating the initial stiffness value. For the calculation of strain energy density, a more accurate finite element analysis method can be introduced, combined with the actual structure and material properties of the swing arm, to dynamically calculate the strain energy density distribution. In addition, the high-strain event threshold in the nonlinear stiffness model can be adjusted according to the fatigue characteristics of the material. For example, for high-strength materials, the threshold can be appropriately increased to reduce misjudgment. At the same time, machine learning algorithms can be introduced to dynamically adjust the stiffness parameters based on historical data to further improve the adaptive ability of the system.

[0189] In some embodiments, the control output module performs:

[0190] Generate a pulse-width modulation signal DPWM to drive the actuator:

[0191] D PWM = 50 + 30·sigmoid(10(τ c 0.5τ max ))

[0192] where τ max is the maximum safe torque;

[0193] When it is detected that the communication delay exceeds 20 ms, switch to the predictive control mode:

[0194]

[0195] where Δt is the delay time.

[0196] It should be noted that the control output module is a key part of this system for integrating compensation parameters and diagnostic results and generating control signals. Its core function is to drive the actuator by generating a duty cycle pulse width modulation signal (DPWM) to ensure that the dynamic adjustment of the swing arm can be accurately executed. When it is detected that the communication delay exceeds the set threshold, the system will switch to the predictive control mode to maintain the stable operation of the system. The design of this module aims to ensure the efficient and stable operation of the system under different working conditions through flexible control strategies.

[0197] Specifically, the control output module drives the actuator through the duty cycle pulse width modulation signal (DPWM). The generation formula of the DPWM signal is

[0198] DPWM = 50 + 30×tanh(10(τ - 0.5τ max ))

[0199] where τ is the compensation torque, and τ max is the maximum safety torque. This formula realizes the smooth adjustment of the control signal through the hyperbolic tangent function (tanh) to ensure that the output torque of the actuator is within the safe range. When the system detects that the communication delay exceeds 20 ms, it will switch to the predictive control mode. At this time, the calculation formula of the control signal is

[0200] τ pred = τ + 0.5τ dot ×Δt

[0201] where τ dot is the torque change rate, and Δt is the delay time. This mode can predict the future state based on the current state and change rate in the case of communication delay, so as to maintain the dynamic performance of the system.

[0202] Preferably, the generation formula of the DPWM signal in the control output module can be optimized according to the characteristics of the actuator. For example, for a high-precision actuator, the parameters in the formula can be adjusted to improve the control accuracy. Specifically, the proportional coefficient in the formula can be adjusted from 30 to 40 to meet the control requirements of higher precision.

[0203] Furthermore, the delay time Δt in the predictive control mode can be dynamically adjusted by real-time monitoring of the communication state to adapt to different communication environments. For example, when the communication delay time is between 10 - 20 ms, a linear interpolation method can be used to adjust the predictive control parameters. At the same time, an adaptive control algorithm can be introduced to dynamically adjust the control strategy according to the system operation state, further improving the adaptability and stability of the system.

[0204] In some embodiments, the online calibration program includes:

[0205] Applying a known excitation force Fcal :

[0206] F cal = 100sin(2π5t)

[0207] Calculate the sensor response error δ:

[0208]

[0209] where y meas is the measured response data, and y theory is the theoretical response data;

[0210] When δ is greater than 0.1, update the sensor calibration coefficient:

[0211] α new = α old ·(1 + 0.2δ)

[0212] where α new is the new calibration coefficient, and α old is the old calibration coefficient.

[0213] It should be noted that the on-line calibration program in this system is an important part for ensuring the high precision and reliability of the sensor during long-term operation. Its core function is to dynamically update the sensor calibration coefficient by applying a known excitation force and calculating the sensor response error. This process can effectively compensate for the drift or aging phenomenon that may occur during the use of the sensor, thus ensuring the measurement accuracy and stability of the system.

[0214] Specifically, the known excitation force applied during the sensor calibration process is F cal = 100sin(2π × 5t), where the frequency f = 5Hz is selected according to the dynamic response characteristics of the sensor and can cover the main frequency range within its working frequency band. The sensor response error is calculated by the formula

[0215]

[0216] where F measured is the measured response data, and F theoretical is the theoretical response data. When the error δ error is greater than 0.1, the system will update the sensor calibration coefficient, and the update formula is

[0217] C new = C old ×(1 + 0.2 × δ error )

[0218] This process ensures that the sensor can return to a state close to the initial accuracy after each calibration.

[0219] Preferably, the excitation force frequency f in the calibration program can be adjusted according to the actual application scenario of the sensor. For example, under high-precision measurement requirements, the frequency can be increased to 10 Hz to more comprehensively cover the frequency response range of the sensor. In addition, the error threshold δ threshold can be optimized according to the accuracy level of the sensor. For high-precision sensors, the threshold can be reduced from 0.1 to 0.05 to further improve the calibration accuracy.

[0220] Furthermore, the update formula of the calibration coefficient can introduce an adaptive factor to dynamically adjust the update amplitude according to historical error data, thereby improving the adaptability and stability of calibration. In addition, a redundant calibration mechanism can be introduced to further improve the reliability of the calibration result by taking the average value of multiple calibrations.

[0221] The above-mentioned embodiments of the present invention have the following beneficial effects: The intelligent swing arm adjustment and detection system of the present invention can collect the vibration signal, stress distribution and displacement data of the swing arm in real time through the multi-source sensing module, and use the attitude calculation module to construct a three-dimensional dynamic swing arm model, which can accurately calculate the attitude deviation and generate compensation parameters. The actuator module drives the swing arm to perform dynamic adjustment according to the compensation parameters, so as to realize the real-time correction of the swing arm attitude and improve the stability and controllability of vehicle driving. The abnormal diagnosis module can determine the fault type and generate a diagnosis result by analyzing the dynamic response spectrum of the swing arm, quickly identify potential faults, and improve the accuracy and efficiency of fault diagnosis. The self-calibration module corrects the drift error of the multi-source sensing data according to the environmental parameters to ensure the accuracy of the sensor data and enhance the reliability and adaptability of the system. The control output module integrates the compensation parameters and the diagnosis result to generate a control signal to realize the precise drive of the actuator. Even under adverse conditions such as communication delay, the system can maintain stable operation through the predictive control mode. The dynamic compensation module constructs an adaptive PID controller and activates a multi-level compensation strategy when the residual error is large, which can further optimize the adjustment effect and improve the response speed and control accuracy of the system. The overload protection mode can effectively protect the actuator by calculating the safety threshold and performing gradient limiting to avoid damage caused by overload. The redundant sensor cross-verification process can determine the sensor failure and switch to the backup channel when the data consistency index is abnormal to ensure the continuous operation of the system. The stiffness matrix update method can dynamically correct the stiffness parameters based on the strain energy density to better adapt to working condition changes such as material fatigue and improve the long-term stability of the system. The online calibration program can update the calibration coefficient in time by applying a known excitation force and calculating the sensor response error, further improving the accuracy and reliability of the sensor.

[0222] Furthermore, the storage medium of the embodiments of the present application stores program instructions that can implement all of the above methods. Among them, the program instructions can be stored in the above storage medium in the form of a software product, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs, or terminal devices such as computers, servers, mobile phones, and tablets.

[0223] The above description is only some preferred embodiments of the present invention and an explanation of the technical principles applied. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the technical solutions formed by mutually replacing the above features with the (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. An intelligent swing arm adjustment and detection system for vehicle drive, characterized in that, It includes the following modules: A multi-source sensing module configured to collect multi-source sensing data in real time, where the multi-source sensing data includes the vibration signal, stress distribution, and displacement data of the swing arm; An attitude solution module that receives the multi-source sensing data and constructs a three-dimensional dynamic swing arm model; A dynamic compensation module that calculates the attitude deviation based on the three-dimensional dynamic swing arm model and generates compensation parameters; An actuator module that drives the swing arm to perform dynamic adjustment according to the compensation parameters; An abnormal diagnosis module that determines the fault type by analyzing the dynamic response spectrum of the swing arm and generates a diagnosis result; A self-calibration module that corrects the drift error of the multi-source sensing data according to environmental parameters; A control output module that integrates the compensation parameters and the diagnosis result to generate a control signal.

2. The system according to claim 1, wherein The attitude solution module includes: A data fusion unit that receives the vibration signal, stress distribution, and displacement data from the multi-source sensing module and constructs a spatio-temporal state vector: Among them, is the angular acceleration, ∈ x is the strain value in the x-axis direction, and Δy is the displacement deviation; A three-dimensional reconstruction unit that establishes a swing arm dynamics model based on the spatio-temporal state vector, and its calculation formula is: where M is the mass matrix, C is the damping matrix, K is the stiffness matrix, u is the displacement vector, and F(t) is the external excitation vector; Deviation analysis unit, calculating the predicted displacement u of the model model and the measured displacement u meas of the residual E, and its calculation formula is: E = ∥u model u meas ∥2。 3. The system according to claim 1, characterized in that, The dynamic compensation module performs: Construct an adaptive PID controller, and its gain parameter adjustment rule is: K p = 0.5 + 0.2E, K i = 0.1e 0.5E , K d = 0.3(1e 2E ) Among them, K p is the proportional gain coefficient, K i is the integral gain coefficient, K d is the derivative gain coefficient, and E is the residual output by the attitude solution module; Generate a compensation torque command, and its calculation formula is: where, e is the attitude error vector, and e is equal to the target displacement u target minus the current displacement u current ; When the residual E is greater than the preset residual threshold, activate a multi-level compensation strategy, including triggering the overload protection mode of the actuator.

4. The system according to claim 1, characterized in that, The abnormal diagnosis module performs: Extract the time-frequency characteristics of the swing arm vibration signal and calculate the wavelet packet energy entropy H: Among them, E j is the wavelet energy of the j-th frequency band, and p j is the proportion of the energy of the j-th frequency band; Construct the feature vector V = [H, σ ∈ , f peak T , where σ ∈ is the mean square deviation of strain, and f peak is the main resonance frequency;​ Adopt an improved SVM classifier for fault diagnosis, and the kernel function is defined as: where x i and x j are input feature vectors, γ is the kernel function width parameter, and d is the polynomial order; When the classification confidence is lower than the preset confidence threshold, activate the redundant sensor cross-validation process.

5. The system according to claim 1, wherein The self-calibration module performs: Real-time monitor the environmental temperature T env and humidity H, and calculate the drift compensation amount Δ cal : Δ cal = 0.3(T env 40) + 0.15(H60) Correct the sensor output, and its calculation formula is: y corrected = y raw ×(1 + Δ cal ·e t / 200 ) where y raw is the original sensing data, and t is the calibration duration; When Δ cal the absolute value is greater than the preset drift compensation amount threshold, start the online calibration program of the sensor array.

6. The system according to claim 3, wherein The overload protection mode performs: Calculate the safety threshold τ max : Among them, τ rated is the rated torque of the actuator; When τ c is greater than τ max gradient clipping is performed: where tanh is the hyperbolic tangent function, and τ ′ c is the compensated torque after amplitude limiting.

7. The system according to claim 4, wherein The redundant sensor cross-validation process includes: Synchronously collect the data of the triaxial accelerometer and the MEMS gyroscope; Calculate the data consistency index R: where a k is the accelerometer data, and g k is the gyroscope integrated acceleration; When R is greater than the preset consistency index threshold, determine the sensor failure and switch to the backup channel.

8. The system according to claim 2, wherein The update method of the stiffness matrix includes: Dynamic modification of stiffness parameter K based on strain energy density ij : where K0 is the initial stiffness and ε is the strain value; When material fatigue is detected, activate the non-linear stiffness model K nl : where εpeak,n is the peak strain of the nth high-strain event and N is the cumulative number of high-strain events.

9. The system according to claim 1, wherein The control output module performs: Generate a pulse width modulation signal DPWM to drive the actuator: D PWM = 50 + 30 · sigmoid(10(τ c 0.5τ max )) where τ max is the maximum safe torque; When it is detected that the communication delay exceeds 20 ms, switch to the predictive control mode: where Δt is the delay time.

10. The system according to claim 5, characterized in that The online calibration program includes: Apply a known excitation force F cal : F cal = 100sin(2π5t) Calculate the sensor response error δ: where y meas is the measured response data, and y theory is the theoretical response data; When δ is greater than 0.1, update the sensor calibration coefficient: α new = α old ·(10.2δ) Among them, α new is the new calibration coefficient, and α old is the old calibration coefficient.

Images