A control method and device of a new energy vehicle braking electric control system

Abstract

The application discloses a control method and device of a new energy automobile brake electric control system, and relates to the technical field of monitoring and analysis, which comprises the following steps: extracting current energy integral data from pedal vibration energy data according to tooth slot torque base frequency data through a band-pass filter; injecting a low-frequency disturbance signal into feedforward gain data and delay compensation phase data, and calculating gradient data of the current energy integral data on the feedforward gain data and the delay compensation phase data; updating the feedforward gain data and the delay compensation phase data based on the gradient data, and feeding back the updated feedforward gain data and the delay compensation phase data to the step of generating hydraulic counteracting torque waveform target data; and cyclically executing extreme value search closed loop until the current energy integral data reaches a minimum value. The application has the effect of improving control efficiency.

Description

Technical Field

[0001] This application relates to the field of monitoring and analysis technology, and in particular to a control method and device for a new energy vehicle braking electronic control system. Background Technology

[0002] With the deep integration of brake-by-wire and electric drive technologies in new energy vehicles, the coordinated control of regenerative braking and hydraulic braking has become crucial for improving vehicle energy efficiency and driving quality. However, under specific operating conditions, due to the physical coupling interference between cross-domain actuators, existing vehicle electronic control strategies often struggle to achieve an effective balance between ensuring continuous energy recovery and maintaining brake pedal comfort. Especially when the vehicle is low on battery power and the driver requests only minimal braking force, the inherent mechanical characteristics of the electric drive system can generate significant negative disturbances to the chassis braking end, resulting in poor pedal tactile feedback to the driver. This has become a technical bottleneck restricting the low-speed driving quality of new energy vehicles.

[0003] Specifically, when a vehicle is in a low-battery state and requests minimal regenerative braking torque, the electromagnetic output of the permanent magnet synchronous motor does very little work. At this time, the inherent periodic cogging torque between the motor's stator and rotor is drastically amplified in the total output torque. This cogging torque pulsation acts as a high-frequency mechanical disturbance wave, directly transmitted to the braking actuator via the drive shaft, thus triggering high-frequency resonant vibration of the brake pedal. Traditional vehicle control strategies typically employ a crude shielding method, directly cutting off the motor's regenerative braking function under such sensitive conditions and relying entirely on mechanical hydraulic braking. This inevitably results in the waste of braking energy, severely weakening the driving range under low battery conditions. Some existing technologies attempt to suppress pulsation by applying reverse damping current only at the motor end, but due to the transmission delay of the vehicle communication bus and the sampling bottleneck of high-frequency signals, not only is accurate real-time force wave cancellation impossible, but secondary vibrations are also easily generated due to phase misalignment. Furthermore, traditional hydraulic braking systems are limited by the large physical inertial response limit of the macroscopic mechanical valve body, and are often considered unable to follow and counteract high-frequency torque fluctuations originating from the motor end, resulting in a deadlock in the engineering implementation of cross-domain collaborative vibration damping. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides a control method and device for a braking electronic control system of a new energy vehicle.

[0005] In a first aspect, this application provides a control method for a braking electronic control system of a new energy vehicle, executed on a braking controller, comprising the following steps: Acquire vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data; When the vehicle speed data is in the sensitive speed range, the battery state of charge data is less than the charge threshold, the brake pedal travel data is greater than zero and the total braking torque request data is less than the torque threshold, and the pedal vibration energy data is greater than the comfort threshold, the joint adjustment vibration damping mode is triggered. In the joint debugging and vibration reduction mode, the tooth cogging feature matrix extracted by the motor controller is received. The tooth cogging feature matrix includes tooth cogging torque fundamental frequency data, harmonic amplitude data and initial phase difference data. Wheel speed fluctuation reference data is extracted, and the wheel speed fluctuation reference data is time-synchronized and phase-locked with the tooth groove feature matrix to reconstruct the end torque pulsation model data. By combining feedforward gain data and delay compensation phase data, the end torque pulsation model data is reverse-calculated to generate target data for hydraulic canceling torque waveform. Extract the anti-jamming high-frequency duty cycle data of the solenoid valve inside the braking system, use the hydraulic cancellation torque waveform target data as the modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generate chatter drive signal data and send it to the brake caliper. Entering the extreme value search closed loop, the extreme value search closed loop includes: The current energy integral data is extracted from the pedal vibration energy data based on the cogging torque fundamental frequency data using a bandpass filter. Inject low-frequency perturbation signals into the feedforward gain data and delay-compensated phase data, and calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data; The feedforward gain data and delay compensation phase data are updated based on gradient data, and the updated feedforward gain data and delay compensation phase data are fed back to the step of generating hydraulic canceling torque waveform target data. The extreme value search loop is executed repeatedly until the current energy integral data reaches a minimum value.

[0006] Secondly, this application provides a control device for a new energy vehicle braking electronic control system, executed on a brake controller, comprising: The data acquisition module is used to acquire vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data. The joint damping mode determination module is used to trigger the joint damping mode when the vehicle speed data is in the sensitive vehicle speed range, the battery state of charge data is less than the charge threshold, the brake pedal travel data is greater than zero and the total braking torque request data is less than the torque threshold, and the pedal vibration energy data is greater than the comfort threshold. The data receiving module is used to receive the tooth cogging feature matrix extracted by the motor controller in the joint debugging and vibration reduction mode. The tooth cogging feature matrix includes tooth cogging torque fundamental frequency data, harmonic amplitude data and initial phase difference data. The reconstruction module is used to extract wheel speed fluctuation reference data, synchronize the wheel speed fluctuation reference data with the tooth groove feature matrix in time, and reconstruct the end torque pulsation model data. The calculation module is used to perform reverse calculation on the end torque pulsation model data by combining feedforward gain data and delay compensation phase data to generate target data of hydraulic canceling torque waveform; The signal generation module is used to extract the anti-jamming high-frequency duty cycle data of the solenoid valve inside the braking system, use the hydraulic cancellation torque waveform target data as the modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generate chatter drive signal data and send it to the brake caliper. The search loop module is used to enter the extreme value search loop, which includes: The current energy integral data is extracted from the pedal vibration energy data based on the cogging torque fundamental frequency data using a bandpass filter. Inject low-frequency perturbation signals into the feedforward gain data and delay-compensated phase data, and calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data; The feedforward gain data and delay compensation phase data are updated based on gradient data, and the updated feedforward gain data and delay compensation phase data are fed back to the step of generating hydraulic canceling torque waveform target data. The extreme value search loop is executed repeatedly until the current energy integral data reaches a minimum value.

[0007] In summary, this application includes at least one of the following beneficial technical effects: 1. This application provides a control method for a new energy vehicle braking electronic control system. Under key operating conditions such as vehicle speed in a sensitive range, battery charge state below a threshold, brake pedal travel, and small total braking torque demand, the method actively triggers a joint damping mode instead of abruptly cutting off regenerative braking. Furthermore, the method extracts the tooth cogging feature matrix through the motor controller and reconstructs the end torque pulsation model by combining wheel speed fluctuation reference data, ultimately generating a hydraulic cancellation torque waveform. This allows the vehicle to maintain regenerative braking participation even when the battery is low and there is a small braking demand, avoiding the waste of braking energy caused by shielding the regenerative function. This effectively extends the vehicle's range when the battery is low and breaks through the technical bottleneck of sacrificing energy efficiency for comfort in traditional strategies. 2. By extracting the high-frequency pulsation sequence from the wheel speed fluctuation reference data and comparing it with the initial phase difference data in the tooth groove feature matrix in real time, synchronous phase compensation data is calculated, and then a real-time phase-locked angle sequence is generated. This enables the accurate reconstruction of the end torque pulsation model data that reflects the actual torque fluctuation at the end of the drive shaft, ensuring that the subsequently generated hydraulic cancellation waveform is strictly matched with the real disturbance in phase. This fundamentally avoids the risk of secondary vibration caused by phase lag or lead, and realizes high-precision coordination of cross-domain actuators. 3. By comparing the current energy integral level with a preset baseline and dynamically calculating the step size adjustment coefficient using a piecewise function: when the error is large, it is linearly positively correlated to quickly approximate the baseline; when the error is small, it decays exponentially to a fixed constant to ensure steady-state accuracy. This effectively overcomes the contradiction between speed and convergence accuracy in the traditional fixed step size gradient descent method. It can respond quickly when the operating conditions change rapidly and avoid parameter oscillation when approaching the optimal state of vibration reduction, thus ensuring the unity of the system's dynamic and static performance. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart of the control method for the braking electronic control system of a new energy vehicle according to an embodiment of this application.

[0010] Figure 2 This is a schematic diagram of the control device of the new energy vehicle braking electronic control system according to an embodiment of this application. Detailed Implementation

[0011] The following description, in conjunction with the implementation of this invention, is merely an example and illustration of the concept of this invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the inventive concept or exceed the scope defined in these claims, all of which should fall within the protection scope of this invention.

[0012] Application Overview: In existing technologies, the coordinated control of brake-by-wire and electric drive in new energy vehicles largely relies on static mapping and independent execution strategies, making it difficult to balance the continuity of weak energy recovery with the tactile comfort of the brake pedal. Traditional methods, when the vehicle is low on battery and the driver only requests minimal braking force, are susceptible to interference from the physical characteristics of the motor's cogging torque transmitted across domains, leading to high-frequency resonance vibration in the brake pedal. Existing vehicle electronic control systems cannot synchronously and accurately counteract the high-frequency torque wave interference from cross-domain mechanisms. Especially when onboard bus communication delays are aggravated or hydraulic inertia is hindered, open-loop compensation models that simply apply reverse current to the motor end will exhibit severe phase misalignment, thereby triggering more destructive secondary vibrations, making it difficult to meet the demands of high-quality driving at low and medium speeds under harsh operating conditions.

[0013] To address the aforementioned issues, the inventors discovered a nonlinear spatiotemporal misalignment between the stator-rotor pole-slot alternation characteristics of the motor and the hysteresis of the end-effector hydraulic transmission. This misalignment is highly susceptible to environmental temperature drift. Adaptive compensation for cross-domain spatiotemporal errors was achieved by establishing an extreme value search closed loop independent of a precise physical model of the controlled object. During the research, it was found that high-frequency communication delays and elastic expansion of hydraulic lines easily invalidate statically calibrated compensation parameters. However, the high-frequency vibration energy actually accumulated at the human tactile interface can objectively and accurately reflect the final effect of force wave physical interference. Therefore, a method was proposed to dynamically approximate the optimal parameter solution using real-time energy integral gradient feedback. Further experimental verification incorporated the orthogonal partial derivative mapping relationship between the pedal vibration energy and gain and phase into the dynamic adjustment mechanism of the cancellation parameters, forming a cross-domain control system from feature extraction to end-effector micro-execution.

[0014] Specifically, the collaborative vibration damping system first synchronously acquires the cross-domain cogging feature matrix of the drive motor and the instantaneous vibration velocity data at the brake chassis end. By dynamically adjusting the bandpass filter response parameters using the extracted characteristic fundamental frequency, a quantitative index of the current energy integral that can truly characterize the severity of high-frequency pedal jitter is calculated. Simultaneously, orthogonal low-frequency sinusoidal perturbation signals are injected into the feedforward gain and delay compensation phase control link to extract real physical gradient data. When it is detected that the current energy integral has not reached the local minimum stagnation point, the system automatically updates the gain and phase using a dynamic step size mechanism, and reconstructs the target data of the hydraulic cancellation torque waveform using the principle of physical interference and the extracted instantaneous phase and amplitude. During continuous frequency modulation control, the system modulates the reference high-frequency duty cycle of the linear motion anti-jamming module in real time through an amplitude perturbation mechanism based on the macroscopic target waveform output by the optimization, forming a dynamic optimization closed loop for the drive current at the execution end. When it is determined that the energy integral has reached the minimum value and the fluctuation is below the convergence threshold, the injection of perturbation signals is stopped and the current optimal parameters are frozen to continuously output a smooth cancellation command.

[0015] Compared to existing technologies, traditional methods rely on the crude shielding and disconnection function of a single electric drive end or lack fixed anti-phase flow compensation for cross-domain dynamic delay assessment. This can easily lead to wasted braking energy or severe mechanical resonance triggering at low torque output. This solution innovatively integrates the high-frequency electromagnetic microscopic observation characteristics of the motor with the integral characteristics of the actual physical vibration energy at the braking end. An extreme value search optimization model is established to achieve adaptive blocking compensation for cross-domain high-frequency jitter. Unlike existing static open-loop detection models that are prone to failure due to parameter drift, this solution can intelligently switch step sizes and iteratively compensate parameters based on real-time communication delay and hydraulic inertia state. Furthermore, it continuously optimizes and reconstructs the waveform through low-frequency orthogonal micro-perturbation closed-loop feedback, significantly improving the system's vibration damping reliability under complex cross-domain coupling interference conditions.

[0016] Through the above technical solutions, this application effectively overcomes the problems of smooth braking sensation distortion and pedal high-frequency resonance caused by the cross-domain transmission of high-frequency force waves from the motor cogging. While ensuring the continuous recovery of weak regenerative braking energy under extremely low battery conditions, it also improves the accuracy of chassis mechanical reverse interference cancellation. The dynamic orthogonal perturbation gradient extremum search mechanism combines the advantages of high sensitivity in optimizing unknown system characteristics with the stability of the underlying anti-jamming fine-tuning duty cycle. The adaptive optimization and convergence adjustment function of model parameters ensures the long-term vibration damping accuracy of the vehicle under long lifespan and varying pipeline temperature and pressure conditions.

[0017] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0018] Example 1 This application discloses a control method for a braking electronic control system of a new energy vehicle.

[0019] Reference Figure 1 A control method for a braking electronic control system of a new energy vehicle, executed on a braking controller, includes the following steps: Acquire vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data; When the vehicle speed data is in the sensitive speed range, the battery state of charge data is less than the charge threshold, the brake pedal travel data is greater than zero and the total braking torque request data is less than the torque threshold, and the pedal vibration energy data is greater than the comfort threshold, the joint adjustment vibration damping mode is triggered. In the joint debugging and vibration reduction mode, the tooth cogging feature matrix extracted by the motor controller is received. The tooth cogging feature matrix includes tooth cogging torque fundamental frequency data, harmonic amplitude data and initial phase difference data. Wheel speed fluctuation reference data is extracted, and the wheel speed fluctuation reference data is time-synchronized and phase-locked with the tooth groove feature matrix to reconstruct the end torque pulsation model data. By combining feedforward gain data and delay compensation phase data, the end torque pulsation model data is reverse-calculated to generate target data for hydraulic canceling torque waveform. Extract the anti-jamming high-frequency duty cycle data of the solenoid valve inside the braking system, use the hydraulic cancellation torque waveform target data as the modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generate chatter drive signal data and send it to the brake caliper. Entering the extreme value search closed loop, the extreme value search closed loop includes: The current energy integral data is extracted from the pedal vibration energy data based on the cogging torque fundamental frequency data using a bandpass filter. Inject low-frequency perturbation signals into the feedforward gain data and delay-compensated phase data, and calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data; The feedforward gain data and delay compensation phase data are updated based on gradient data, and the updated feedforward gain data and delay compensation phase data are fed back to the step of generating hydraulic canceling torque waveform target data. The extreme value search loop is executed repeatedly until the current energy integral data reaches a minimum value.

[0020] In this embodiment of the invention, acquiring basic vehicle status and vibration data (including vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data) refers to the real-time acquisition of a set of physical quantities of vehicle operating conditions and driver operating intentions through the vehicle bus network and multi-dimensional sensor matrix. Specifically, this can be achieved through wheel speed sensors, BMS battery management system, pedal displacement sensors, and underlying high-frequency vibration accelerometers, which serve as the real-time environmental perception basis and logical trigger input conditions for the system to determine whether to intervene in cross-domain collaborative vibration damping control.

[0021] Among them, the comprehensive judgment conditions for triggering the joint debugging vibration reduction mode (i.e., vehicle speed is in the sensitive range, battery power is less than the threshold, pedal travel is greater than zero, torque request is less than the threshold and vibration energy is greater than the threshold) refer to the pre-calibrated high-frequency cogging resonance boundary scenario of the motor that is easy to induce and can be perceived by the human body. Specifically, logical rules can be established through bench calibration experiments or subjective evaluation data statistical mapping to accurately define the appropriate timing for cross-domain interference intervention between the chassis and electric drive, avoiding ineffective control intervention and wasted computing power under non-resonance conditions. The cogging feature matrix refers to the high-frequency electromagnetic excitation source feature set containing cogging torque fundamental frequency data, harmonic amplitude data, and initial phase difference data. Specifically, this can be achieved through the extraction of fast Fourier transform and high-frequency injection observation model based on the rotor position and stator current of the drive motor at the bottom layer of the motor controller. This is used to break down information silos and provide a precise dimensionality-reduced mathematical description of the physical source of cross-domain interference force waves for the chassis braking end. Time synchronization phase-locked loop and end torque pulsation model data reconstruction refers to the process of spatiotemporal alignment and mapping of the wheel speed fluctuation reference data extracted from the chassis with the tooth groove feature matrix received across the domain. This can be achieved through digital bandpass filtering, Hilbert transform to extract instantaneous phase, and kinematic space delay compensation algorithm. This is used to accurately restore the true mechanical force wave shape transmitted across the domain to the physical space of the brake caliper.

[0022] Among them, the hydraulic cancellation torque waveform target data refers to the target control command generated by combining the optimized feedforward gain data and delay compensation phase data to perform anti-phase superposition calculation on the end torque pulsation model. Specifically, it can be calculated through the mechanical waveform cancellation interference equation and the hydraulic volume expansion attenuation model. It is used to generate a destructive interference command source with the same amplitude and opposite phase as the interference force wave at the hydraulic actuator. The perturbation modulation and chatter drive signal data generation refers to the process of mapping the macroscopic hydraulic cancellation target data to the microscopic electrical control execution layer. Specifically, it can be achieved by using the cancellation target as a modulation wave to perform amplitude modulation and duty cycle reconstruction on the high-frequency duty cycle of the anti-jamming solenoid valve inside the braking system. It is used to output precise high-frequency cancellation hydraulic pressure through the microscopic chatter of the solenoid valve coil without affecting the macroscopic braking safety pressure build-up.

[0023] Among them, the extreme value search closed loop refers to an adaptive iterative feedback mechanism that dynamically finds the best vibration compensation parameters by introducing exploration signals. Specifically, it can be achieved by extracting the current energy integral data of a specific frequency band through bandpass filtering, injecting orthogonal low-frequency sinusoidal perturbations, calculating the energy gradient using a decoupled correlation integral algorithm, and dynamically updating the feedforward gain and phase. This is used to completely overcome the distortion of static compensation parameters caused by cross-domain communication delay drift and hydraulic pipeline temperature drift. The current energy integral data reaching the minimum value means that the high-frequency jitter energy at the pedal end is canceled to the lowest level of convergence stagnation state by the reverse physical interference of the actuator. Specifically, it can be achieved by calculating the average energy fluctuation over multiple consecutive iteration cycles and determining when the absolute values ​​of the gradients are both less than the tolerance threshold. This is used to terminate the injection of perturbation signals and freeze the output of the optimal vibration damping parameters for a long time.

[0024] The working process and principle of this application are as follows: First, vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data are acquired. Then, when the vehicle speed data is within the sensitive speed range, the battery state of charge data is less than the charge threshold, the brake pedal travel data is greater than zero, and the total braking torque request data is less than the torque threshold, while the pedal vibration energy data is greater than the comfort threshold, a joint adjustment vibration reduction mode is triggered. Next, in the joint adjustment vibration reduction mode, the tooth cogging feature matrix extracted by the motor controller is received. The tooth cogging feature matrix includes tooth cogging torque fundamental frequency data, harmonic amplitude data, and initial phase difference data. Subsequently, wheel speed fluctuation reference data is extracted, and the wheel speed fluctuation reference data is time-synchronized and phase-locked with the tooth cogging feature matrix to reconstruct the end torque pulsation model data. Based on this pulsation model, combined with feedforward gain data and delay compensation phase data, it is used for reverse calculation to generate hydraulic cancellation torque waveform target data. Then, the anti-jamming high-frequency duty cycle data of the solenoid valve inside the braking system is extracted, and the hydraulic... The target data of the anti-jamming torque waveform is used as a modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generating flutter drive signal data and sending it to the brake caliper. Then, it enters the extreme value search closed loop. In this closed loop, the current energy integral data is first extracted from the pedal vibration energy data based on the toothed torque fundamental frequency data through a bandpass filter. Low-frequency perturbation signals are injected into the feedforward gain data and delay compensation phase data. The gradient data of the current energy integral data with respect to the feedforward gain data and delay compensation phase data is calculated. Then, the feedforward gain data and delay compensation phase data are updated based on the gradient data. The updated feedforward gain data and delay compensation phase data are fed back to the step of generating hydraulic anti-jamming torque waveform target data. Finally, the above extreme value search closed loop is executed repeatedly until the current energy integral data reaches a minimum value. In this way, dynamic compensation and adaptive parameter tuning of cross-domain high-frequency resonance jitter interference error are achieved, ensuring that the vehicle maintains excellent pedal tactile smoothness and continuous braking energy recovery under harsh operating conditions.

[0025] Furthermore, the wheel speed fluctuation reference data and the tooth groove feature matrix are synchronized and phase-locked in time to reconstruct the end torque pulsation model data, including: Based on the tooth cogging torque fundamental frequency data, the wheel speed fluctuation reference data is filtered and separated to extract the high-frequency wheel speed pulsation sequence data with the same frequency as the tooth cogging torque fundamental frequency data; Real-time dynamic phase data of high-frequency wheel speed pulsation sequence data is extracted, and the difference between the real-time dynamic phase data and the initial phase difference data is compared to calculate the synchronous phase compensation data. A reference phase sequence data is generated based on the cogging torque fundamental frequency data. The reference phase sequence data is added to the synchronous phase compensation data to generate real-time phase-locked angle sequence data. Harmonic amplitude data is used as the amplitude reference, and real-time phase-locked angle sequence data is used as the periodically changing independent variable to construct a fluctuation mapping relationship. Based on the fluctuation mapping relationship, the end torque pulsation model data is output.

[0026] In this embodiment of the invention, the raw angular velocity signal from the wheel speed sensor is first extracted as reference data for wheel speed fluctuations. A digital bandpass filter is then introduced, using the fundamental frequency data of the tooth cogging torque in the tooth cogging feature matrix as the center frequency for filtering and separation. Considering the torsional stiffness characteristics of the vehicle driveshaft, the bandpass filter is constructed based on a standard second-order infinite impulse response model. After its transfer function is converted into a discrete-time difference equation, it filters out the low-frequency components caused by macroscopic vehicle speed changes and the broadband noise caused by random road surface excitation, thereby extracting high-frequency pulsation sequence data of wheel speed with a frequency equivalent to the fundamental frequency data of the tooth cogging torque. The quality factor Q of the bandpass filter is a preset filter sharpness parameter, which is set to an adjustable range of 5 to 10. To avoid signal loss due to small changes in speed caused by an excessively narrow passband, and to ensure sufficient suppression of mechanical noise interference from non-tooth cogging frequencies, a value of 8 is preferred.

[0027] Next, the high-frequency pulsation sequence data of the wheel speed is analyzed, and the real-time dynamic phase data of the sequence is extracted using the Hilbert transform algorithm. This algorithm, by converting the pulsating signal in the real domain to an analytic signal in the complex domain, overcomes the computational deficiency of traditional zero-crossing detection methods, which are easily affected by environmental noise and cause phase jumps in small-amplitude high-frequency signals. It can continuously and smoothly acquire the instantaneous phase at any given moment. After calculating the real-time dynamic phase data, it is compared with the initial phase difference data contained in the tooth cogging feature matrix. The calculation formula is as follows: in, For synchronous phase compensation data, it represents the spatial hysteresis angle of the transmission chain at the brake disc relative to the force wave source at the motor end; This refers to real-time dynamic phase data obtained through Hilbert transform; The received initial phase difference data reflects the initial offset angle of the electromagnetic excitation source on the rotor side of the motor.

[0028] Subsequently, using the time axis as the integration reference, a reference phase sequence data is generated based on the cogging torque fundamental frequency data. This reference phase sequence data is then added to the synchronous phase compensation data, calculated using the following formula: in, The generated real-time phase-locked angle sequence data represents the relative mechanical phase angle that is fully aligned to the physical space of the brake caliper at the current moment; This is the fundamental frequency data for tooth cogging torque.

[0029] Finally, a nonlinear physical space mapping model is established, using the harmonic amplitude data in the tooth cogging feature matrix as the amplitude reference, and the real-time phase-locked angle sequence data as the periodically changing independent variable to construct the wave mapping relationship. The calculation formula is as follows: in, The final output is the end torque pulsation model data, which characterizes the instantaneous disturbance torque transmitted to the brake disc; The harmonic amplitude data represents the magnitude of the extreme values ​​of cogging torque fluctuations evaluated and issued by the motor controller; a sine function is used. Establishing a mapping relationship allows for the perfect reconstruction of the physical nature of the spatial periodic force waves generated when the rotor poles and stator slots of a permanent magnet synchronous motor are staggered and aligned. Through this mapping, the reference torque for subsequent feedforward cancellation can be accurately calculated.

[0030] Furthermore, the target data of the hydraulic cancellation torque waveform is used as a modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generating chatter drive signal data and sending it to the brake caliper, including: Extract the baseline duty cycle value and carrier frequency data of the anti-jamming high-frequency duty cycle data; The target data of the hydraulic canceling torque waveform is normalized to generate fluctuation envelope data. Dynamic duty cycle bias data is generated by superimposing and multiplying the baseline duty cycle value using the fluctuation envelope data. The dynamic duty cycle offset data and carrier frequency data are fused and reconstructed to generate flutter drive signal data. The flutter drive signal data is converted into a drive current with a corresponding duty cycle and sent to the linear solenoid valve in the brake caliper.

[0031] In this embodiment of the invention, for the conversion and modulation of the underlying signal at the execution end in the joint debugging and vibration damping mode, the reference duty cycle value and carrier frequency data of the anti-jamming high-frequency duty cycle data inside the underlying brake-by-wire system are first extracted. The carrier frequency data typically originates from the clock division setting of the underlying hardware microcontroller. Next, in order to losslessly map the target data of the hydraulic cancellation torque waveform in the macroscopic mechanical dimension to the microscopic electronic control domain, and to avoid direct drive generating large hydraulic shocks that could affect braking safety, the system normalizes the target data of the hydraulic cancellation torque waveform, generates fluctuation envelope data, and performs a multiplication operation with the reference duty cycle value to generate dynamic duty cycle offset data. This fusion calculation is based on a variation of the amplitude modulation formula. in, For the generated dynamic duty cycle offset data; The extracted baseline duty cycle value; The target data for the hydraulic canceling torque waveform input in the preceding steps; The equivalent reference torque constant of the solenoid valve is extracted and pre-calibrated; This is the preset modulation scaling factor. This modulation scaling factor is mainly used to constrain the envelope depth of high-frequency fluctuations. It is set to an adjustable range of 0.05 to 0.15. In order to avoid excessively large perturbation amplitude causing macroscopic displacement of the main valve core and thus interfering with normal brake pressure build-up, and at the same time to ensure that the small chatter energy is sufficient to penetrate the bulk modulus of the hydraulic oil and be transmitted to the brake caliper, the preferred value is 0.10.

[0032] Subsequently, the generated dynamic duty cycle bias data and carrier frequency data are input to a hardware timer module for fusion and reconstruction, generating a flutter drive signal data composed of a pulse width modulation waveform with a real-time dynamic change in duty cycle. Finally, this flutter drive signal data is converted into a drive current corresponding to the duty cycle and sent to the linear solenoid valve in the brake caliper. To accurately quantify the underlying electrical output, the conversion process uses a drive current calculation formula derived from a variation of the high-frequency AC impedance characteristics: in, This is the drive current data output to the solenoid valve coil; The power supply voltage for the vehicle power system is collected in real time; The DC equivalent resistance parameter of the linear solenoid valve coil; This is a correction coefficient for the high-frequency electromagnetic conversion efficiency of the coil. Traditional duty cycle equivalent current formulas typically only involve dividing the voltage by the resistance and then multiplying by the duty cycle. However, this invention introduces a different formula based on this. The necessity and advantage of this variation are that it can accurately compensate for the non-negligible core eddy current loss and high-frequency hysteresis loss of the solenoid valve under high-frequency carrier flutter conditions, making the calculated micro-current command more closely approximate the actual physical response boundary of the actuator, and effectively preventing the tiny canceling force wave signal from being swallowed up by mechanical dead zone and heat loss.

[0033] Furthermore, low-frequency perturbation signals are injected into the feedforward gain data and delay-compensated phase data to calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data, including: Generate sinusoidal disturbance wave data with a frequency lower than the fundamental frequency of the cogging torque; The sinusoidal perturbation wave data is superimposed onto the feedforward gain data according to the first proportional coefficient, and simultaneously superimposed onto the delay compensation phase data according to the second proportional coefficient to form the parametric perturbation excitation data. The current energy integral data within the time window under the action of the collected parameter perturbation excitation data; The correlation integrals of the current energy integral data and the sinusoidal disturbance wave data within the time window are calculated to separate the first correlation integral value and the second correlation integral value. The first correlation integral value is used as the gradient data of the current energy integral data with respect to the feedforward gain data, and the second correlation integral value is used as the gradient data of the current energy integral data with respect to the delay compensation phase data.

[0034] In this embodiment of the invention, for the dynamic optimization process of unknown system characteristics in the extreme value search closed loop, the system first uses the built-in digital oscillator to generate sinusoidal disturbance wave data with a frequency much lower than the fundamental frequency data of the tooth cogging torque, which serves as the excitation source for exploring the optimal canceling torque parameters. The disturbance angular frequency of the sinusoidal disturbance wave data is set to an adjustable range of 10% to 20% of the fundamental frequency of the tooth cogging torque. In order to avoid the aliasing and coupling of high-frequency micro-perturbations with the inherent mechanical delay characteristics of the brake hydraulic system, and to ensure that the parameter optimization speed can keep up with the changes in actual vehicle speed and road conditions, the disturbance angular frequency is preferably 20 radians per second.

[0035] Next, in order to achieve decoupling optimization of multiple control parameters, the system uses the principle of quadrature modulation to directly superimpose the generated sinusoidal disturbance wave data onto the feedforward gain data according to the first proportional coefficient as a sinusoidal reference wave. At the same time, after phase shifting the sinusoidal disturbance wave data by half of pi, it is superimposed onto the delay compensation phase data as mutually orthogonal cosine waves according to the second proportional coefficient, thereby combining to form parameter perturbation excitation data containing two independent orthogonal excitation components.

[0036] Subsequently, the system monitors the system response in real time and collects the current energy integral data within a specific time window under the continuous action of the parameter perturbation excitation data. To ensure the mathematical completeness of the periodic signal integration operation, the time window is preset to an integer multiple of the aforementioned sinusoidal perturbation period, preferably a single complete period. After acquiring the current energy integral data within the time window, to eliminate the interference of random road vibration noise and steady-state DC bias on the gradient direction, the system performs a correlation integration operation based on mean-removing logic on the current energy integral data and the corresponding sinusoidal perturbation wave data, thereby accurately separating the first and second correlation integral values. This core operation is solved based on the decoupled correlation integral variant formula: in, The first relevant integral value separated will be directly used as the gradient data output of the current energy integral data to the feedforward gain data; The separated second correlation integral value will be output as the gradient data of the current energy integral data to the delay-compensated phase data; The collected real-time energy integral data characterizes the high-frequency oscillation energy level of the pedal at different times; The energy moving average baseline data within the time window; The preferred perturbation angular frequency; At the current sampling time, The time variable is the integral time. The set time window length; is the first proportionality coefficient, and is the dimensionless scaling constant that limits gain jumps; The second proportionality coefficient represents the maximum amplitude of the phase perturbation. Traditional extremum search algorithms typically rely on an additional cascaded high-pass digital filter to block the DC steady-state component of the objective function, but this inevitably introduces additional phase delay and severely slows down the system convergence speed. The variant formula designed in this embodiment directly introduces a sliding window mean difference term inside the integrator kernel. Within a limited time window, DC filtering and quadrature demodulation separation are completed simultaneously in one go. At the same time, the pre-designed normalization adjustment coefficient ensures that the dimension of the output gradient is strictly equivalent to the true dimension of the partial derivative of the cost function with respect to the tuned parameter, thus completely eliminating the dimension distortion problem that may exist in conventional algorithms.

[0037] Furthermore, the feedforward gain data and delay compensation phase data are updated based on the gradient data, including: Extract the absolute value level of the current energy integral data; The absolute numerical level is compared with a preset energy baseline to calculate the error level data; The error level data is input as an independent variable into a piecewise function to calculate the dynamic step size adjustment coefficient. When the error level data is in the first interval, the dynamic step size adjustment coefficient is linearly positively correlated with the error level data. When the error level data enters the second interval with lower values, the dynamic step size adjustment coefficient decays exponentially to a fixed constant. Multiply the gradient data by the dynamic step size adjustment coefficient to generate incremental data for parameter updates; The parameter update increment data is added to the feedforward gain data and delay compensation phase data of the previous control cycle respectively to complete the closed-loop parameter update.

[0038] In this embodiment of the invention, for the parameter adaptive optimization stage at the end of the extreme value search closed loop, the system first extracts the absolute value level of the current energy integral data as an objective basis for evaluating the severity of the current pedal vibration. This absolute value level is then compared with a preset energy baseline by subtraction to calculate the error level data purely caused by the cogging effect. The preset energy baseline represents the inherent, unavoidable mechanical background noise of the vehicle chassis and hydraulic lines. This baseline value is set to an adjustable range of 0.01 joules to 0.10 joules. To avoid the algorithm misjudging normal minor vibrations of the road surface as residual cogging force waves that need to be eliminated, thus causing invalid optimization oscillations, the preset energy baseline is preferably 0.05 joules.

[0039] Next, to balance global search speed and local convergence stability, the system inputs the calculated error level data as an independent variable into a carefully designed piecewise function to calculate the dynamic step size adjustment coefficient. The variant calculation formula of this piecewise function is as follows: in, The calculated dynamic step size adjustment coefficient; The error level data is calculated from the preceding steps; The energy boundary threshold for dividing the optimization interval is set to an adjustable range of 0.5 joules to 2.0 joules, preferably 1.0 joules; The reference energy constant is used for unifying dimensions; the default value here is 1.0 joule to achieve dimensionless error level data. The first interval (i.e.) The linear proportional gain in the high vibration region; The second interval (i.e.) The fixed constant lower limit (in the low-vibration region). Traditional extreme value search often uses a fixed optimization step size, which easily leads to extremely slow convergence in the early stage and violent oscillations near the minimum point in the later stage. The variant piecewise function constructed in this embodiment, when the error level data is in the first interval, the step size coefficient is linearly positively correlated with the error level data, driving the system to approach the optimal solution quickly and in large steps; when the error level data enters the second interval with lower values, the step size coefficient strictly decays exponentially and smoothly to a fixed constant according to the natural constant. This ensures stability when finally reaching the minimum point, while retaining the ability to fine-tune slowly changing characteristics such as hydraulic oil temperature drift.

[0040] Subsequently, the system multiplies the gradient data obtained in the previous step with the dynamic step size adjustment coefficient, and combines this with the physical dimension conversion rate to generate parameter update incremental data. The incremental decomposition formula is as follows: as well as ;in, The update increment of the feedforward gain; The update increment for delay compensation phase; and These are the first and second relevant integral values ​​input in the preceding steps, respectively; The baseline learning rate is the gain dimension, measured in reciprocals of joules. The baseline learning rate is defined as the phase dimension, measured in square radians per joule. A negative sign indicates that the algorithm performs minimum gradient descent optimization in the opposite direction of the gradient. Finally, the calculated parameter update increments are added to the feedforward gain data and delay compensation phase data from the previous control cycle, respectively, thus completing the closed-loop parameter update.

[0041] Furthermore, before receiving the tooth cogging feature matrix extracted by the motor controller, the following data processing steps are performed by the motor controller: Real-time acquisition of rotor electrical angle data and stator three-phase current data of the drive motor; The rotor electrical angle data and stator three-phase current data are input into the high-frequency injection observation model to calculate the instantaneous electromagnetic torque sequence data; Fast Fourier Transform is performed on the instantaneous electromagnetic torque sequence data to extract the dominant harmonic frequency in the spectrum as the fundamental frequency data of the cogging torque, and the amplitude corresponding to the dominant harmonic frequency is extracted as the harmonic amplitude data. The phase corresponding to the dominant harmonic frequency is aligned and compared with the mechanical zero position of the rotor electrical angle data to calculate the initial phase difference data. The tooth cogging torque fundamental frequency data, harmonic amplitude data, and initial phase difference data are packaged into a tooth cogging feature matrix and sent through the vehicle communication network.

[0042] In this embodiment of the invention, to ensure that the chassis braking end can accurately obtain the physical source parameters required for cross-domain collaborative vibration damping, before the brake controller receives the tooth cogging feature matrix extracted by the motor controller, the system first performs the high-frequency electromagnetic feature analysis and extraction steps based on the underlying computing logic of the motor controller. This process first collects the rotor electrical angle data and stator three-phase current data of the drive motor in real time through the resolver sensor and current transformer. The acquisition operation follows the preset underlying communication sampling frequency, which is set to an adjustable range of 10 kHz to 20 kHz. To avoid the Nyquist aliasing phenomenon from masking the weak high-frequency tooth cogging force wave pulsation, and at the same time taking into account the computing power load limit of the microcontroller, 10 kHz is preferred.

[0043] Next, the system transforms the collected stator three-phase current data into AC and DC axis current data through coordinate transformation, and inputs it synchronously with the rotor electrical angle data into the high-frequency injection observation model. Since traditional electromagnetic torque observation models rely solely on constant permanent magnet calibration flux, they cannot accurately capture the microscopic high-frequency magnetoresistive fluctuations caused by the stator-rotor relative positional overlap. Therefore, this embodiment uses a variant formula for instantaneous torque that incorporates high-frequency current perturbation response to calculate the instantaneous electromagnetic torque sequence data. The variant calculation formula is as follows: in, The instantaneous electromagnetic torque sequence data obtained from the calculation; This refers to the pole pair number parameter of the drive motor; For pre-calibrated permanent magnet flux linkage data; as well as These are the quadrature-axis current data and the direct-axis current data obtained after coordinate transformation, respectively. and The data for calibrated direct-axis inductance and quadrature-axis inductance; The preset high-frequency coupling amplification factor is set to an adjustable range of 0.5 to 1.5 to enhance the signal-to-noise ratio of the high-frequency characteristics, preferably set to 1.0; This refers to the high-frequency response current data separated from the quadrature-axis current by an internal digital bandpass filter. This is the rated reference current data for the motor. The necessity and core advantage of this variant formula lies in the fact that by multiplying and coupling the dimensionless high-frequency current response perturbation term with the macroscopic fundamental torque, the observability of high-frequency reluctance torque such as cogging torque in the total torque is greatly amplified, preventing it from being completely submerged by the macroscopic driving torque.

[0044] Subsequently, the system performs Fast Fourier Transform (FFT) processing on the instantaneous electromagnetic torque sequence data within the set analysis time window, extracting the dominant harmonic frequency from the spectrum as the fundamental frequency data of the cogging torque, and extracting the peak value corresponding to the dominant harmonic frequency as the harmonic amplitude data. Simultaneously, it extracts the complex phase angle of this frequency component at the start of the analysis time window as the initial phase data of the spectrum. Considering that the sampling delay of the underlying analog-to-digital converter of the motor controller and the time consumed in pre-communication calculations will introduce a significant spatial phase shift to the high-frequency signal, in order to accurately map the phase extracted from the frequency domain back to the actual physical mechanical space, the system aligns and compares the phase corresponding to the dominant harmonic frequency with the mechanical zero position of the rotor electrical angle data, and calculates the initial phase difference data using a phase transformation formula that includes a dynamic delay compensation mechanism. in, The initial phase difference data obtained from the calculation; Initial phase spectral data extracted using Fast Fourier Transform; This refers to the inherent number of cogging force wave cycles corresponding to one revolution of the motor. To analyze the rotor electrical angle data at the start of the time window; This represents the current mechanical angular velocity data of the motor; This is the pre-calibrated controller hardware comprehensive delay time data. Compared to the traditional simple angle difference, this variant formula introduces a dynamic product compensation term of speed and delay time, which can completely eliminate the spatial angle slip caused by time difference under speed fluctuation conditions.

[0045] Finally, the system packages the tooth cogging torque fundamental frequency data, harmonic amplitude data, and initial phase difference data into a tooth cogging feature matrix containing dimensionality reduction feature information, and sends it to the brake controller through the vehicle communication network to complete the simplified cross-domain data transmission.

[0046] Furthermore, based on the fundamental frequency data of the tooth cogging torque, the current energy integral data is extracted from the pedal vibration energy data using a bandpass filter, including: The dynamic passband range is determined by using the cogging torque fundamental frequency data as the center frequency and combining it with the preset bandwidth offset. Adjust the frequency response parameters of the bandpass filter to match the dynamic passband range; Input the pedal vibration energy data into the bandpass filter after adjusting the frequency response parameters to filter out vibration sequence data in a specific frequency band; The vibration sequence data in a specific frequency band is summed by squaring to generate the current energy integral data.

[0047] In this embodiment of the invention, for the quantitative evaluation stage of the target cost function in the extreme value search closed loop, the system needs to extract the current energy integral data from the pedal vibration energy data based on the tooth cogging torque fundamental frequency data using a bandpass filter. This process first uses the tooth cogging torque fundamental frequency data in the received tooth cogging feature matrix as the center frequency, and combines it with a preset bandwidth offset to determine the dynamic passband interval. The preset bandwidth offset represents the target frequency drift margin that the control system is allowed to capture. This preset bandwidth offset is set to an adjustable range of 2 Hz to 5 Hz. To avoid the high-frequency jitter signal instantly leaving the monitoring and filtering range when the motor speed fluctuates slightly due to an excessively narrow passband, and to ensure effective filtering of non-co-frequency wideband background noise such as random road surface excitations, a bandwidth offset of 3 Hz is preferred. Based on the determined upper and lower limits of the dynamic passband interval, the system recalculates the quality factor of the digital filter and adjusts the frequency response parameters of the discrete-time second-order Butterworth bandpass filter accordingly, ensuring that its amplitude-frequency characteristics perfectly match the dynamic passband interval.

[0048] Next, the system inputs the pedal vibration energy data (specifically, the instantaneous vibration velocity sequence data of the pedal surface in this embodiment) collected in real time by the underlying vibration sensor into the bandpass filter after adjusting the frequency response parameters. Through differential iterative calculation, it accurately filters and outputs vibration sequence data of a specific frequency band containing only the characteristics of tooth cogging force wave transmission.

[0049] Subsequently, in order to accurately convert the aforementioned instantaneous velocity sequence into a macroscopic energy index that can characterize the driver's actual tactile discomfort, the system performs a sum-of-squares operation on the vibration sequence data in a specific frequency band, thereby generating the current energy integral data. This core operation is based on a variation of the work done by damping dissipation at the human body interface in mechanical engineering, and its variation calculation formula is as follows: in, The final calculated current energy integral data characterizes the high-frequency jitter energy that is transmitted and accumulated on the pedal interface within the evaluation time window; The equivalent damping dissipation coefficient data of the pedal structure and the driver's foot contact interface were extracted for pre-calibration. This is the index of discrete sampling points for the underlying modular sampling. To evaluate the total number of sampling points included within the time window; To filter vibration sequence data of a specific frequency band from the output of a pre-bandpass filter; This refers to the sampling time interval data of the underlying hardware system. Traditional control strategies often simply accumulate the absolute values ​​of acceleration or velocity signals when evaluating vibration, lacking the true meaning of physical work done and failing to objectively measure the physiological discomfort felt by the human body. However, this embodiment introduces the multiplication of the equivalent damping dissipation coefficient with the time discrete integral of the square of the velocity in a specific frequency band, directly and accurately restoring the real mechanical energy transformed into human discomfort in the physical dissipation work domain, greatly enhancing the perception resolution of the subsequent extreme value search algorithm for subtle jitter changes.

[0050] Furthermore, the steps to determine if the current energy integral data has reached a minimum include: Within multiple consecutive iterations of the extreme value search closed loop, store the sequence of changes in the current energy integral data; The average absolute value of the change difference sequence is calculated to obtain the mean energy fluctuation data. When the mean value of energy fluctuation is less than the convergence threshold and the absolute value of gradient data is less than the gradient tolerance threshold, the current energy integral data is determined to have reached a minimum value. Once the minimum value is reached, the injection of low-frequency disturbance signals into the feedforward gain data and delay compensation phase data is stopped, and the current feedforward gain data and delay compensation phase data are frozen.

[0051] In this embodiment of the invention, for the convergence determination and parameter freezing stage of the extreme value search closed loop, in order to accurately identify whether the system has truly reached the optimal parameter solution that minimizes the high-frequency jitter of the pedal, the system first constructs a sliding data buffer to store the change difference sequence of the current energy integral data within multiple consecutive iteration cycles of the extreme value search closed loop. The historical evaluation window length is set to a preset number of cycles, which is set to an adjustable range of 5 to 15 iteration cycles. To avoid misjudgments caused by accidental sensor sampling noise due to an excessively short data window, and to ensure that the closed loop strategy does not experience convergence determination lag due to an excessively long window, the preset number of cycles is preferably 10 iteration cycles.

[0052] Next, the system extracts the energy changes before and after each discrete period within the aforementioned window, and performs an average calculation on the absolute values ​​of the change difference sequence based on exponential decay weighting to obtain the average energy fluctuation data. Traditional simple arithmetic averages are easily affected by historical large fluctuations, slowing down the judgment process. Therefore, this embodiment uses a modified moving average formula that incorporates a time penalty factor for solving the problem: in, The calculated average energy fluctuation data characterizes the overall stability of the recent energy curve; The index for the current latest iteration cycle; This is the periodic offset index for historical backtracking, with values ​​ranging from 0 to... ; The optimal historical evaluation window length is set to 10. The current energy integral data is collected and stored for the corresponding period; The preset time decay penalty factor is used to assign a higher judgment weight to data that is closer to the current time. It is set to an adjustable range of 0.1 to 0.5, preferably 0.2.

[0053] Subsequently, the system compares the calculated average energy fluctuation data with the convergence threshold, which represents the energy noise floor resulting from the limiting resolution of the chassis hydraulic micro-adjustment, and is preferably set to 0.005 joules. Relying solely on stable energy fluctuations may still lead to a plateau in the algorithm; therefore, the system must simultaneously perform double verification on the pre-calculated gradient data. Specifically, it checks whether the absolute value of the first relevant integral value (gain gradient) is less than the first gradient tolerance threshold, and whether the absolute value of the second relevant integral value (phase gradient) is less than the second gradient tolerance threshold. When the absolute values ​​of the average energy fluctuation data and both gradient values ​​simultaneously meet the strict condition of being less than the corresponding thresholds, the system confirms from both physical and mathematical dimensions that the optimization function has reached the true local minimum stationary point, and determines that the current energy integral data has reached a minimum. Once the minimum value is reached, the system immediately sends a blocking command to the internal oscillator, forcing the first and second proportional coefficients of the aforementioned generated parameter perturbation excitation data to zero, thereby stopping the injection of low-frequency perturbation signals into the feedforward gain data and delay compensation phase data, and freezing the feedforward gain data and delay compensation phase data obtained at the current moment as the optimal cancellation parameters for long-term output.

[0054] Furthermore, by combining feedforward gain data and delay-compensated phase data, the end-torque pulsation model data is inversely calculated to generate target data for the hydraulic cancelling torque waveform, including: Extract instantaneous phase data from the end torque pulsation model data; The instantaneous phase data and the delayed compensation phase data are vector superimposed and subtracted to generate the anti-phase compensation angle data. Generate inverse amplitude coefficient data based on feedforward gain data; By combining the reverse amplitude coefficient data and the reverse phase compensation angle data, the trigonometric function equation is reconstructed to generate the target data of the hydraulic canceling torque waveform.

[0055] In this embodiment of the invention, for the force wave cancellation execution stage after the extreme value search closed loop converges, in order to accurately generate a physical reaction force at the chassis brake caliper that can destructively interfere with the motor tooth groove pulsation, the system must combine the feedforward gain data and delay compensation phase data frozen in the optimization to perform reverse calculation on the end torque pulsation model data, thereby generating the target data of the hydraulic cancellation torque waveform. The system first analyzes the end torque pulsation model data that maintains continuous synchronous operation, extracts the instantaneous phase data used to characterize the periodic spatial position of the force wave at the current moment, and simultaneously extracts the basic wave harmonic amplitude data inherent in the model.

[0056] Next, to accurately align the microsecond-level physical spatiotemporal misalignment generated by the cross-domain transmission process and the hydraulic pressure-building process, the system performs vector superposition and subtraction operations on the extracted instantaneous phase data and the delay-compensated phase data obtained from the extreme value search iteration to generate anti-phase compensation angle data. Subsequently, the system generates inverse amplitude coefficient data based on the frozen feedforward gain data, and fuses this inverse amplitude coefficient data with the anti-phase compensation angle data to reconstruct a trigonometric function equation containing completely destructive interference characteristics. This reconstruction process is based on a variation of the mechanical waveform destructive superposition principle, and its variation calculation formula is as follows: in, To calculate the target data of the generated hydraulic canceling torque waveform, the high-frequency reverse canceling torque required to be output by the chassis braking system is characterized; Instantaneous phase data extracted from end-torque pulsation model data; The delay-compensated phase data is the output of the closed-loop freeze for extreme value search; the difference between the two constitutes the... This refers to the generated anti-phase compensation angle data; The feedforward gain data of the closed-loop freeze output is used for extreme value search; This is the fundamental wave harmonic amplitude data extracted synchronously from the pulsation model; The preset hydraulic volume expansion attenuation coefficient; This represents an artificially introduced physical phase-reversal operator. The combined terms inside the square brackets represent the generated inverse amplitude coefficient data. Traditional force wave cancellation equations typically only use simple gain multiplication, neglecting the energy dissipation inevitably caused by the elastic expansion of the hose wall when high-frequency perturbations propagate in a slender brake hose, resulting in a weaker cancellation amplitude reaching the caliper. The variant formula introduced in this embodiment adds a preset hydraulic volume expansion attenuation coefficient. The coefficient is set to an adjustable range of 1.05 to 1.20. To avoid overcompensation leading to caliper pressure build-up overshoot, and to ensure that sufficient cancellation energy is maintained after the compensating hydraulic pressure penetrates the elastic deformation barrier of the oil pipe, the coefficient is preferably 1.10.

[0057] Example 2 This application also discloses a control device for a new energy vehicle braking electronic control system.

[0058] Reference Figure 2 A control device for a new energy vehicle braking electronic control system, executed on the brake controller, comprising: The data acquisition module is used to acquire vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data. The joint damping mode determination module is used to trigger the joint damping mode when the vehicle speed data is in the sensitive vehicle speed range, the battery state of charge data is less than the charge threshold, the brake pedal travel data is greater than zero and the total braking torque request data is less than the torque threshold, and the pedal vibration energy data is greater than the comfort threshold. The data receiving module is used to receive the tooth cogging feature matrix extracted by the motor controller in the joint debugging and vibration reduction mode. The tooth cogging feature matrix includes tooth cogging torque fundamental frequency data, harmonic amplitude data and initial phase difference data. The reconstruction module is used to extract wheel speed fluctuation reference data, synchronize the wheel speed fluctuation reference data with the tooth groove feature matrix in time, and reconstruct the end torque pulsation model data. The calculation module is used to perform reverse calculation on the end torque pulsation model data by combining feedforward gain data and delay compensation phase data to generate target data of hydraulic canceling torque waveform; The signal generation module is used to extract the anti-jamming high-frequency duty cycle data of the solenoid valve inside the braking system, use the hydraulic cancellation torque waveform target data as the modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generate chatter drive signal data and send it to the brake caliper. The search loop module is used to enter the extreme value search loop, which includes: The current energy integral data is extracted from the pedal vibration energy data based on the cogging torque fundamental frequency data using a bandpass filter. Inject low-frequency perturbation signals into the feedforward gain data and delay-compensated phase data, and calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data; The feedforward gain data and delay compensation phase data are updated based on gradient data, and the updated feedforward gain data and delay compensation phase data are fed back to the step of generating hydraulic canceling torque waveform target data. The extreme value search loop is executed repeatedly until the current energy integral data reaches a minimum value.

[0059] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention, they should all fall within the protection scope of the present invention.

[0060] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0061] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A control method for a braking electronic control system of a new energy vehicle, characterized in that, The action performed on the brake controller includes the following steps: Acquire vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data; When the vehicle speed data is in the sensitive speed range, the battery state of charge data is less than the charge threshold, the brake pedal travel data is greater than zero and the total braking torque request data is less than the torque threshold, and the pedal vibration energy data is greater than the comfort threshold, the joint adjustment vibration damping mode is triggered. In the joint debugging and vibration reduction mode, the tooth cogging feature matrix extracted by the motor controller is received. The tooth cogging feature matrix includes tooth cogging torque fundamental frequency data, harmonic amplitude data and initial phase difference data. Wheel speed fluctuation reference data is extracted, and the wheel speed fluctuation reference data is time-synchronized and phase-locked with the tooth groove feature matrix to reconstruct the end torque pulsation model data. By combining feedforward gain data and delay compensation phase data, the end torque pulsation model data is reverse-calculated to generate target data for hydraulic canceling torque waveform. Extract the anti-jamming high-frequency duty cycle data of the solenoid valve inside the braking system, use the hydraulic cancellation torque waveform target data as the modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generate chatter drive signal data and send it to the brake caliper. Entering the extreme value search closed loop, the extreme value search closed loop includes: The current energy integral data is extracted from the pedal vibration energy data based on the cogging torque fundamental frequency data using a bandpass filter. Inject low-frequency perturbation signals into the feedforward gain data and delay-compensated phase data, and calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data; The feedforward gain data and delay compensation phase data are updated based on gradient data, and the updated feedforward gain data and delay compensation phase data are fed back to the step of generating hydraulic canceling torque waveform target data. The extreme value search loop is executed repeatedly until the current energy integral data reaches a minimum value.

2. The control method for a new energy vehicle braking electronic control system according to claim 1, characterized in that, By performing time-synchronized phase-locking with the wheel speed fluctuation reference data and the tooth groove feature matrix, the end torque pulsation model data is reconstructed, including: Based on the tooth cogging torque fundamental frequency data, the wheel speed fluctuation reference data is filtered and separated to extract the high-frequency wheel speed pulsation sequence data with the same frequency as the tooth cogging torque fundamental frequency data; Real-time dynamic phase data of high-frequency wheel speed pulsation sequence data is extracted, and the difference between the real-time dynamic phase data and the initial phase difference data is compared to calculate the synchronous phase compensation data. A reference phase sequence data is generated based on the cogging torque fundamental frequency data. The reference phase sequence data is added to the synchronous phase compensation data to generate real-time phase-locked angle sequence data. Harmonic amplitude data is used as the amplitude reference, and real-time phase-locked angle sequence data is used as the periodically changing independent variable to construct a fluctuation mapping relationship. Based on the fluctuation mapping relationship, the end torque pulsation model data is output.

3. The control method for a new energy vehicle braking electronic control system according to claim 1, characterized in that, The target data of the hydraulic cancellation torque waveform is used as a modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generating chatter drive signal data and sending it to the brake caliper, including: Extract the baseline duty cycle value and carrier frequency data of the anti-jamming high-frequency duty cycle data; The target data of the hydraulic canceling torque waveform is normalized to generate fluctuation envelope data. Dynamic duty cycle bias data is generated by superimposing and multiplying the baseline duty cycle value using the fluctuation envelope data. The dynamic duty cycle offset data and carrier frequency data are fused and reconstructed to generate flutter drive signal data. The flutter drive signal data is converted into a drive current with a corresponding duty cycle and sent to the linear solenoid valve in the brake caliper.

4. The control method for a new energy vehicle braking electronic control system according to claim 1, characterized in that, Inject low-frequency perturbation signals into the feedforward gain data and delay-compensated phase data, and calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data, including: Generate sinusoidal disturbance wave data with a frequency lower than the fundamental frequency of the cogging torque; The sinusoidal perturbation wave data is superimposed onto the feedforward gain data according to the first proportional coefficient, and simultaneously superimposed onto the delay compensation phase data according to the second proportional coefficient to form the parametric perturbation excitation data. The current energy integral data within the time window under the action of the collected parameter perturbation excitation data; The correlation integrals of the current energy integral data and the sinusoidal disturbance wave data within the time window are calculated to separate the first correlation integral value and the second correlation integral value. The first correlation integral value is used as the gradient data of the current energy integral data with respect to the feedforward gain data, and the second correlation integral value is used as the gradient data of the current energy integral data with respect to the delay compensation phase data.

5. The control method for a new energy vehicle braking electronic control system according to claim 4, characterized in that, Updating feedforward gain data and delay-compensated phase data based on gradient data includes: Extract the absolute value level of the current energy integral data; The absolute numerical level is compared with a preset energy baseline to calculate the error level data; The error level data is input as an independent variable into a piecewise function to calculate the dynamic step size adjustment coefficient. When the error level data is in the first interval, the dynamic step size adjustment coefficient is linearly positively correlated with the error level data. When the error level data enters the second interval with lower values, the dynamic step size adjustment coefficient decays exponentially to a fixed constant. Multiply the gradient data by the dynamic step size adjustment coefficient to generate incremental data for parameter updates; The parameter update increment data is added to the feedforward gain data and delay compensation phase data of the previous control cycle respectively to complete the closed-loop parameter update.

6. The control method for a new energy vehicle braking electronic control system according to claim 1, characterized in that, Before receiving the tooth cogging feature matrix extracted by the motor controller, the following data processing steps are performed by the motor controller: Real-time acquisition of rotor electrical angle data and stator three-phase current data of the drive motor; The rotor electrical angle data and stator three-phase current data are input into the high-frequency injection observation model to calculate the instantaneous electromagnetic torque sequence data; Fast Fourier Transform is performed on the instantaneous electromagnetic torque sequence data to extract the dominant harmonic frequency in the spectrum as the fundamental frequency data of the cogging torque, and the amplitude corresponding to the dominant harmonic frequency is extracted as the harmonic amplitude data. The phase corresponding to the dominant harmonic frequency is aligned and compared with the mechanical zero position of the rotor electrical angle data to calculate the initial phase difference data. The tooth cogging torque fundamental frequency data, harmonic amplitude data, and initial phase difference data are packaged into a tooth cogging feature matrix and sent through the vehicle communication network.

7. The control method for a new energy vehicle braking electronic control system according to claim 1, characterized in that, The current energy integral data is extracted from the pedal vibration energy data based on the cogging torque fundamental frequency data using a bandpass filter, including: The dynamic passband range is determined by using the cogging torque fundamental frequency data as the center frequency and combining it with the preset bandwidth offset. Adjust the frequency response parameters of the bandpass filter to match the dynamic passband range; Input the pedal vibration energy data into the bandpass filter after adjusting the frequency response parameters to filter out vibration sequence data in a specific frequency band; The vibration sequence data in a specific frequency band is summed by squaring to generate the current energy integral data.

8. The control method for a new energy vehicle braking electronic control system according to claim 1, characterized in that, The steps to determine if the current energy integral data has reached a minimum include: Within multiple consecutive iterations of the extreme value search closed loop, store the sequence of changes in the current energy integral data; The average absolute value of the change difference sequence is calculated to obtain the mean energy fluctuation data. When the mean value of energy fluctuation is less than the convergence threshold and the absolute value of gradient data is less than the gradient tolerance threshold, the current energy integral data is determined to have reached a minimum value. Once the minimum value is reached, the injection of low-frequency disturbance signals into the feedforward gain data and delay compensation phase data is stopped, and the current feedforward gain data and delay compensation phase data are frozen.

9. The control method for a new energy vehicle braking electronic control system according to claim 1, characterized in that, By combining feedforward gain data and delay-compensated phase data, the terminal torque pulsation model data is inversely calculated to generate target data for the hydraulic cancelling torque waveform, including: Extract instantaneous phase data from the end torque pulsation model data; The instantaneous phase data and the delayed compensation phase data are vector superimposed and subtracted to generate the anti-phase compensation angle data. Generate inverse amplitude coefficient data based on feedforward gain data; By combining the reverse amplitude coefficient data and the reverse phase compensation angle data, the trigonometric function equation is reconstructed to generate the target data of the hydraulic canceling torque waveform.

10. A control device for a new energy vehicle braking electronic control system, applied to the control method of the new energy vehicle braking electronic control system according to any one of claims 1-9, characterized in that, Executed in the brake controller, including: The data acquisition module is used to acquire vehicle speed data, battery state of charge data, brake pedal travel data, total braking torque request data, and pedal vibration energy data. The joint damping mode determination module is used to trigger the joint damping mode when the vehicle speed data is in the sensitive vehicle speed range, the battery state of charge data is less than the charge threshold, the brake pedal travel data is greater than zero and the total braking torque request data is less than the torque threshold, and the pedal vibration energy data is greater than the comfort threshold. The data receiving module is used to receive the tooth cogging feature matrix extracted by the motor controller in the joint debugging and vibration reduction mode. The tooth cogging feature matrix includes tooth cogging torque fundamental frequency data, harmonic amplitude data and initial phase difference data. The reconstruction module is used to extract wheel speed fluctuation reference data, synchronize the wheel speed fluctuation reference data with the tooth groove feature matrix in time, and reconstruct the end torque pulsation model data. The calculation module is used to perform reverse calculation on the end torque pulsation model data by combining feedforward gain data and delay compensation phase data to generate target data of hydraulic canceling torque waveform; The signal generation module is used to extract the anti-jamming high-frequency duty cycle data of the solenoid valve inside the braking system, use the hydraulic cancellation torque waveform target data as the modulation wave to perform perturbation modulation on the anti-jamming high-frequency duty cycle data, generate chatter drive signal data and send it to the brake caliper. The search loop module is used to enter the extreme value search loop, which includes: The current energy integral data is extracted from the pedal vibration energy data based on the cogging torque fundamental frequency data using a bandpass filter. Inject low-frequency perturbation signals into the feedforward gain data and delay-compensated phase data, and calculate the gradient data of the current energy integral data with respect to the feedforward gain data and delay-compensated phase data; The feedforward gain data and delay compensation phase data are updated based on gradient data, and the updated feedforward gain data and delay compensation phase data are fed back to the step of generating hydraulic canceling torque waveform target data. The extreme value search loop is executed repeatedly until the current energy integral data reaches a minimum value.

Images