A motor control method combining inductance and non-inductance for brake-by-wire

Abstract

A motor control method applied to brake-by-wire, which combines the sensing and non-sensing, comprises: obtaining the sensing position of the rotor, and a signal quality factor based on a calibration signal; running an extended Kalman filter (EKF) algorithm to obtain the non-sensing position of the rotor, the speed of the rotor, and the estimated error covariance matrix of the non-sensing position; calculating a residual value based on the sensing position and the non-sensing position; if the signal quality factor is greater than a preset quality threshold value, and the residual value is less than a tolerance threshold value, then determining that the sensing position is reliable; otherwise, determining that the sensing position is unreliable; if the sensing position is reliable, then setting the observation noise covariance matrix as a first value to obtain a fusion position and a fusion speed; if the sensing position is unreliable, then setting the observation noise covariance matrix as a second value which is much larger than the first value to obtain the fusion position and the fusion speed; and generating a control signal for controlling the motor based on the fusion position and the fusion speed. The method improves the accuracy of the motor control of the brake-by-wire.

Description

TECHNICAL FIELD

[0001] The application relates to the fields of motor control and vehicle braking, in particular to a motor control method combining sensing and non-sensing applied to a brake-by-wire system. BACKGROUND

[0002] As a key actuator for realizing advanced driving assistance system (ADAS) and automatic driving functions of modern vehicles, the performance of a brake-by-wire system is directly related to the safety and driving experience of the vehicle. A permanent magnet synchronous motor (PMSM) is an ideal actuator for the brake-by-wire system due to its high power density and high efficiency. The performance of the field-oriented control (FOC) algorithm, which is the core control algorithm of the PMSM, is extremely dependent on the accuracy and reliability of the rotor position information.

[0003] At present, mature solutions in the industry mostly rely on a single position sensor, such as a rotary transformer or an emerging tunnel magnetoresistance (TMR) sensor. As a high-performance magnetic sensor, the TMR sensor has the advantages of high sensitivity, good temperature stability, and wide frequency response, and can directly output absolute position information, simplifying the signal processing process. However, like all physical sensors, the TMR sensor has inherent potential problems such as susceptibility to external stray magnetic field interference and signal harmonic distortion caused by installation eccentricity, and its instantaneous reliability in complex electromagnetic environments and throughout its life cycle still needs further verification.

[0004] Therefore, to solve the reliability bottleneck of the single sensor solution and fully utilize the high-performance advantages of the TMR sensor, the industry urgently needs an innovative motor control strategy that combines sensing and non-sensing technologies and has online diagnosis and seamless fault-tolerant capabilities to improve the robustness and functional safety level of the brake-by-wire system under all operating conditions. SUMMARY

[0005] To solve the above technical problems, the application provides a motor control method combining sensing and non-sensing applied to a brake-by-wire system.

[0006] The purpose of the application is to overcome the reliability risks and performance bottlenecks of existing brake-by-wire systems that rely on a single position sensor, and to provide a redundant position sensing and fault-tolerant control method for a PMSM based on the deep fusion of a TMR sensor and an extended Kalman filter (EKF) non-sensing algorithm. This method not only builds fault redundancy between two position sources, but also realizes performance stacking in daily operation through an adaptive intelligent fusion mechanism, significantly improving the dynamic response, control accuracy, and functional safety level of the system.

[0007] To achieve the above purpose, the technical solution adopted by the application is to build a closed-loop intelligent processing flow, which specifically includes the following steps:

[0008] Firstly, the parallel acquisition and preprocessing of dual-channel information is performed. In the TMR sensor channel, the system reads the absolute position raw signal output by the TMR sensor in real time, and performs online harmonic compensation and temperature drift compensation on the absolute position raw signal to obtain a high-precision calibrated inductive position signal, and a signal quality factor representing the signal-to-noise ratio and stability of the signal is generated. In the non-inductive sampling channel, the system synchronously acquires the phase current and DC bus voltage of the motor.

[0009] Secondly, an extended Kalman filter (EKF) non-inductive observer is run. Based on the acquired current and voltage, the EKF algorithm estimates the non-inductive position and speed of the rotor in real time according to the mathematical model of the motor, and outputs the estimation error variance of the non-inductive position, which directly quantifies the uncertainty of the current estimation result of the EKF algorithm.

[0010] Subsequently, a confidence-based dynamic diagnosis of the TMR sensor signal is performed. This step introduces a two-way cross-validation intelligent mechanism: on the one hand, it checks whether the signal quality factor of the TMR is higher than a preset threshold; on the other hand, it calculates the residual error between the calibrated position of the TMR and the non-inductive position of the EKF, and compares the residual error with a dynamic tolerance threshold. One of the key innovations of the present application is that the tolerance threshold is not a fixed value, but is positively correlated with the estimation error variance output by the EKF. This means that when the EKF has a high estimation uncertainty due to the severe system dynamics, the system will intelligently relax the tolerance of the TMR signal; conversely, under stable working conditions, it requires stricter consistency. Only when the signal quality is high and the residual error is within the dynamic threshold, the TMR signal is considered highly reliable.

[0011] Then, adaptive Kalman fusion is performed. This is the core innovation of the present application. The system dynamically adjusts the observation noise covariance matrix of the EKF algorithm according to the diagnosis results. When the TMR signal is reliable, the matrix is set to a small value, so that the EKF fully adopts the absolute position information of the TMR as an observation in the update step, optimally corrects the internal state prediction, and outputs a fused position that combines the absolute accuracy of the TMR and the noise suppression ability of the EKF. When the TMR signal is not reliable, the matrix is set to a maximum value, so that the EKF naturally ignores the unreliable sensor input and degenerates to a pure non-inductive running mode, outputting the non-inductive position as the fused position, achieving seamless fault tolerance.

[0012] Finally, fault-tolerant control based on the fused position is performed. The high-quality fused position signal obtained in the foregoing steps is input into the field-oriented control loop, and through coordinate transformation, current loop regulation and SVPWM modulation, the permanent magnet synchronous motor is finally driven to accurately and stably output the target braking torque. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1A motor control method flow chart of a combined inductive and non-inductive application to line control braking is provided in the embodiments of the present application.

[0014] Figure 2 A TMR original signal calibration and quality evaluation flow chart is provided in the embodiments of the present application.

[0015] Figure 3 A signal reliability dynamic diagnosis flow chart is provided in the embodiments of the present application.

[0016] Figure 4 A dynamic performance enhancement processing flow chart is provided in the embodiments of the present application. DETAILED DESCRIPTION

[0017] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to be limiting on the present application. As used in the specification and the appended claims of the present application, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms used herein are intended to encompass any and all possible combinations of one or more of the listed items.

[0018] Hereinafter, the terms first and second are used only for the purpose of description and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with first and second can explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, the meaning of multiple is two or more, unless otherwise specified.

[0019] The present application provides a motor control method combined inductive and non-inductive application to line control braking, see Figure 1 , including:

[0020] See Figure 2 , first, through the TMR sensor to obtain the absolute position of the rotor real-time signal . The signal usually contains harmonic components and drift caused by sensor installation eccentricity, environmental magnetic field interference and temperature change. In order to improve the reliability of the signal, the system carries out online harmonic compensation and temperature drift compensation, and obtains the calibrated signal, and then obtains the calibrated rotor position , namely inductive position, and the corresponding signal quality factor , which is used to reflect the signal to noise ratio and stability, and its calculation formula is: ;

[0021] Here, signal strength represents the amplitude of the calibrated position signal, and noise level is the fluctuation introduced during measurement by external interference or sensor noise itself. A higher signal quality factor indicates a more stable and reliable TMR sensor output signal. (Calibrated) and This will serve as the core input for subsequent fusion estimation. Simultaneously, the system collects the motor phase current. and DC bus voltage It is used for state estimation of the sensorless EKF observer.

[0022] After data acquisition is complete, the system enters the extended Kalman filter (EKF) sensorless observation phase. Based on the acquired current and voltage data, the EKF algorithm predicts the rotor position using the motor's dynamic mathematical model. and rotational speed The system outputs the position estimation error covariance matrix P. The main diagonal element P(1,1) of the covariance matrix directly reflects the reliability of the rotor position estimation, i.e., the uncertainty of the EKF's prediction of the current state. The EKF state update formula is:

[0023] ,

[0024] ;

[0025] in, This represents the estimated state value. For Kalman gain, For the observed values, For the measurement matrix, To estimate the error covariance matrix, This represents the estimated state value at the previous moment. Indicates and The corresponding estimation error covariance matrix. EKF achieves seamless estimation through two steps: state prediction and observation update. State prediction is based on the PMSM electrical and mechanical equations, while observation update uses real-time acquired current and voltage signals to correct the predicted values, thus achieving optimal state estimation.

[0026] See Figure 3 After estimating the sensed and non-sense states, the system performs dynamic validity diagnosis on the TMR signal. Specifically, this involves calculating the residual between the sensed and non-sense locations.

[0027] ;

[0028] Combined with signal quality factor And the main diagonal element P(1,1) of the covariance matrix, to determine the reliability of the TMR sensor output. The determination logic is as follows: when Higher than the preset threshold and When the tolerance threshold positively correlated with P(1,1) is less than the TMR signal is identified as reliable; otherwise, it is determined to be unreliable. The setting of the dynamic tolerance is directly related to the uncertainty of the EKF, that is, the more uncertain the EKF estimates, the greater the tolerated residual. The tolerance threshold The main diagonal element P(1,1) of the EKF estimated error covariance matrix can be adaptively calculated according to: ; wherein, , is an empirical coefficient, The range of is usually set to 0.5-2.0; The value of is generally between 0.1°-1.0°;

[0029] Based on the determination result, a flag signal Flag_TMR_High_Confidence is generated, which is used for mode selection in the subsequent fusion step. At the same time, in the zero speed starting, extremely low speed running or emergency braking working condition, the system generates a performance requirement flag Flag_Perf_Req, which is used to enhance the response weight of the EKF to the inductive signal.

[0030] After completing the dynamic validity diagnosis of the TMR signal, the system enters the adaptive Kalman fusion phase. The core goal of this phase is to dynamically fuse the inductive position information with the EKF inductive estimation to achieve high-precision, low-noise fusion position output and fusion speed . The fusion algorithm adjusts the EKF observation noise covariance matrix to realize mode switching, thereby smoothly transitioning between high confidence and low confidence.

[0031] When Flag_TMR_High_Confidence=1, that is, the TMR signal is determined to be highly reliable, the system enters the high confidence fusion mode. In this mode, the calibrated is input into the EKF update step as an observation. The calculation formula of the Kalman gain is:

[0032] ;

[0033] Here, is the observation noise covariance matrix, whose value is set to the first value , reflecting the high reliability of the TMR measurement. The EKF updates the predicted state based on the TMR observation to obtain the fused state estimation:

[0034] ;

[0035] ​The updated output fusion position Fusion speed In this mode, the system utilizes the absolute position information provided by the TMR sensor to continuously correct the EKF state estimation. At the same time, it leverages the EKF's ability to suppress measurement noise to reduce high-frequency fluctuations and improve the smoothness of position and velocity signals, thereby enhancing the response performance and dynamic bandwidth of the field-oriented control (FOC) loop.

[0036] When Flag_TMR_High_Confidence=0, meaning the TMR signal is deemed untrustworthy, the system enters a low-confidence or fault mode. In this mode, the EKF observer switches to completely undetectable operation, meaning it is no longer used in update steps. Correction is performed. This is achieved by adjusting the observation noise covariance matrix. Set to much greater than The second value This increases the Kalman gain. The update step almost ignores the observations and relies solely on the motor dynamic model for state prediction:

[0037] ;

[0038] This design ensures that the system can seamlessly degrade to a high-performance EKF sensorless mode when the TMR sensor fails or is subjected to strong interference, achieving fault redundancy and functional safety.

[0039] By adjusting the coefficient To improve the system's dynamic response performance while ensuring stability, position and velocity are integrated. After output, it directly enters the FOC fault-tolerant control loop. The FOC algorithm first converts the fused mechanical state into dq-axis currents in the synchronous coordinate system. :

[0040] ; wherein, the desired current command , From the desired torque command And with the motor parameters determined, the two satisfy the torque relationship of PMSM in the dq synchronous coordinate system:

[0041] Where p is the extreme logarithm, It is a permanent magnet flux linkage. , These are the d-axis and q-axis inductances, respectively; preferably, in the base velocity region, i.e., when field weakening control is not activated. Then there is Thus clarifying and The relationship is: Fixed to 0, With proportional change, optionally, in the high-speed field weakening region, to meet the bus voltage, current constraints, make , and amplitude limiting to meet ; wherein is the two-phase measured current, the current control loop according to the expected current instruction generate dq axis voltage instruction , and then through the SVPWM algorithm to generate inverter switching signal, drive PMSM to track torque instruction . In this process, due to the fusion signal noise is low and high dynamic accuracy, current loop control response is fast and stable, while avoiding the control shock caused by TMR transient failure.

[0042] In the mode switching process, the system through the continuous adjustment of observation noise covariance matrix Smooth transition, avoid the control quantity mutation caused by traditional hard switching. The specific implementation is according to the signal quality factor And the residual Calculate the smoothing factor :

[0043] ,

[0044] ;

[0045] wherein, The lower limit and upper limit of signal quality factor The upper limit of residual tolerance. Through this continuous adjustment, the Kalman gain of EKF in each control cycle can be smoothly changed to realize the disturbance-free switching between sensing and non-sensing.

[0046] The FOC loop of the system generates the control signal of the motor, the mechanical angle For coordinate transformation, fusion speed For current loop speed feedforward compensation, so as to realize accurate torque control. In the emergency stop or emergency braking condition, the controller can quickly increase the bandwidth of the current loop, so that the motor can quickly output braking torque, and at the same time, through the adaptive fusion, the position estimation is continuous and reliable, which prevents braking distortion or braking lag. In the whole implementation process, the hardware advantages of TMR sensor and the dynamic performance of EKF non-sensing algorithm are fully played through adaptive Kalman fusion, which not only guarantees the absolute position accuracy, but also provides non-sensing fault tolerance ability, forming a kind of motor control method with high reliability and high dynamic response.

[0047] The performance of the TMR sensor plays a crucial role in the overall control accuracy of the system, and the quality of its output signal directly affects the reliability of EKF sensorless position estimation and the effectiveness of adaptive Kalman fusion. Therefore, optimizing the sensor's structure, installation, signal acquisition, and compensation strategies is a prerequisite for ensuring the high-performance operation of the entire brake-by-wire system.

[0048] First, the TMR sensor employs a differential output structure. Specifically, two output signals with opposite phases are generated on the same sensor chip and processed by a differential amplifier to suppress common-mode magnetic field interference. The design concept of the differential output structure is based on the common-mode nature of electromagnetic interference (EMI) on the two signals: when external stray magnetic fields or conducted interference act on both output signals simultaneously, the differential amplifier automatically cancels the common-mode components of the two signals when calculating the output voltage, retaining only the differential signal portion. This differential signal is the effective position signal caused by the actual change in magnetic flux. This structure not only effectively reduces the impact of external magnetic interference on measurement accuracy but also enhances the system's stability in the complex electromagnetic environment of automobiles and under conditions of multi-source interference.

[0049] Secondly, the installation position of the TMR sensor has been carefully designed and optimized. During mechanical installation, the distance, coaxiality, and parallelism between the sensor and the permanent magnet have a significant impact on the linearity and harmonic characteristics of the output signal. Precise mechanical installation and positioning ensure that the magnetic flux change caused by magnet rotation is as linear as possible, thereby reducing harmonic distortion caused by installation eccentricity, uneven gaps, or magnetic flux nonlinearity. This optimization ensures that the calibrated position signal... It maintains high linearity across the entire angular range, providing a high-precision sensing input signal for subsequent EKF fusion.

[0050] Regarding temperature compensation, the raw signal output by the TMR sensor... The Kalman filter is susceptible to drift due to changes in ambient temperature. Temperature drift can cause output position deviations, thus affecting the state estimation accuracy of the Kalman filter. To address this issue, the system performs real-time monitoring... Temperature compensation is performed. Two approaches can be used: one is to apply linear or nonlinear correction to the output at different temperatures based on the temperature coefficient curve provided by the sensor manufacturer; the other is to obtain the sensor's temperature response model under actual operating conditions through online calibration. The compensation value is calculated in real time during each sampling period to minimize the impact of temperature on position measurement. The calculation formula is as follows:

[0051] ;

[0052] Where T is the temperature currently measured by the sensor. For the temperature compensation function, the curve obtained from the experiment calibration or the table lookup method can be used. The temperature compensation process can be completed in real time in the microcontroller, ensuring that the system can still output accurate position signals in low or high temperature environments, and ensuring the dynamic performance of the brake-by-wire system under all temperature conditions.

[0053] At the same time, the system also compensates for periodic errors caused by installation eccentricity, tooth slot effect and non-ideal characteristics of the magnet. The core idea of harmonic compensation is to decompose the known periodic error into a finite Fourier series, and modify the amplitude and phase of each harmonic to eliminate its influence on the output signal. The implementation formula is:

[0054]

[0055] is the amplitude of the th harmonic, is the corresponding phase, is the mechanical angular velocity, is the number of harmonics participating in compensation. By compensating for multiple harmonics one by one, the repetitive periodic error caused by mechanical eccentricity or non-ideal magnetic field can be significantly reduced, further improving the accuracy and smoothness of the calibrated signal. The compensation coefficients and can be obtained through sensor calibration experiments and applied to the signal processing module in the microcontroller in the form of a lookup table or real-time calculation.

[0056] It is worth noting that, in order to ensure the real-time performance of temperature compensation and harmonic compensation, the system uses a pipeline processing structure for : first, temperature compensation is performed to eliminate long-term drift, then harmonic compensation is performed to remove periodic errors, and finally the output is used for EKF fusion. Through this step-by-step processing and priority sorting strategy, the interaction between temperature compensation and harmonic compensation can be effectively avoided, ensuring that the output signal is both accurate and smooth.

[0057] In addition, to further enhance the system's resistance to transient disturbances, a signal filtering unit can be set up in the microcontroller to perform low-pass filtering or Kalman filtering front-end processing on θtmr_comp to suppress the superposition of high-frequency noise. This processing is particularly important in zero-speed starting or emergency braking conditions, as it can prevent EKF state jumps caused by short-term fluctuations in sensor signals, thereby improving the stability and reliability of the control loop.

[0058] During normal operation, the system will adjust the gain of the TMR sensor according to the signal quality factor ​​and the main diagonal element P(1,1) of the rotor position error covariance matrix P estimated by the extended Kalman filter (EKF) to dynamically adjust the tolerance threshold , whose formula is:

[0059] ;

[0060] wherein, , is an empirical coefficient, respectively used to depict the amplification of the covariance to the tolerance threshold and the basic static tolerance. Specifically, is used to map the uncertainty P(1,1) of the position estimation by the EKF to the tolerance range, thereby automatically allowing a larger residual value when the system state prediction uncertainty is large , without prematurely determining the TMR signal as unreliable; provides a minimum tolerance basis to ensure that there is still a certain tolerance space when the EKF estimation is highly certain, to cope with the small noise or transient disturbance existing in the signal.

[0061] the dynamic adjustment logic of the tolerance threshold plays an important role in actual control. Since the estimation uncertainty of the EKF will dynamically change with the motor speed, load change and environmental disturbance, if the tolerance threshold remains fixed, when the EKF uncertainty is high, the residual is easy to exceed the static threshold, causing the system to frequently switch to the non-inductive mode, increasing the instability of the control loop, and even possibly causing torque output fluctuations. By positively correlating with P(1,1), the system can adaptively amplify the tolerance range according to the current state, so that the TMR signal is still reasonably utilized when the EKF prediction uncertainty is high, thereby maintaining the continuity and stability of the fused signal.

[0062] At the same time, the dynamic adjustment of the tolerance threshold also considers the quality factor of the TMR signal. When is high and the EKF uncertainty P(1,1) is low, will approach its minimum value , so that the system strictly utilizes the TMR signal under high-precision conditions to achieve fine correction of the fused position. Conversely, when is low or the EKF uncertainty increases, increases, allowing to fluctuate in a larger range, thereby preventing false judgment of TMR signal failure and avoiding control oscillation caused by frequent switching. This adaptive mechanism is particularly important in zero-speed starting, very low-speed running and load rapid change conditions, as it ensures that the system neither prematurely switches the mode due to small deviations nor loses effective utilization of the TMR signal in critical transient states.

[0063] In addition, in actual implementation, the value of the experience coefficient , may be optimized through simulation and real vehicle experiments. The range of the experience coefficient is usually set to 0.5-2.0, which is adjusted according to the sensitivity of the prediction error of the EKF to the tolerance amplification; The value of the experience coefficient is generally between 0.1°-1.0°, so as to provide a small signal tolerance space and ensure full use of the TMR signal under high precision. Through experimental verification, this dynamic tolerance strategy can realize smooth mode switching under different working conditions, improve the robustness of the system to transient disturbances and external interference, avoid torque fluctuation and control delay caused by misjudgment, and further improve the safety and response performance of the brake-by-wire system. In summary, The dynamic adjustment of the experience coefficient is not only a key parameter for judging the reliability of the TMR signal, but also an important guarantee for smooth switching and fault-tolerant control of the adaptive Kalman fusion algorithm.

[0064] In the PMSM control system, zero-speed starting and extremely low-speed running belong to the working conditions in which the blind area of the sensorless algorithm is most likely to occur. At this time, due to the close-to-zero back electromotive force of the motor, the traditional sensorless EKF cannot reliably estimate the rotor position and speed, which may cause unstable or delayed torque output during the starting stage. To solve this problem, the performance enhancement flag Flag_Perf_Req is introduced in the adaptive Kalman fusion module, which is used to enhance the response ability of the EKF to the inductive signal under specific working conditions, so as to improve the dynamic performance and starting reliability of the system. See Figure 4 The setting conditions of the performance enhancement flag Flag_Perf_Req include: zero-speed starting condition, the actual speed of the motor , wherein is a threshold value less than the lowest measurable speed of the system. Extremely low-speed running, Emergency braking condition, braking request signal is activated, and the expected output torque is greater than the safety braking threshold Tq_emergency. When any of the above conditions is met, the Flag_Perf_Req is set to 1, indicating that the system enters the performance enhancement mode. In the performance enhancement mode, the amplitude of the Kalman gain matrix is adjusted to improve the response weight of the EKF to the TMR inductive position signal, so as to realize fast and reliable position signal updating. The enhancement strategy formula is as follows:

[0065] ;

[0066] , wherein is the Kalman gain calculated by the standard EKF update step, is the gain adjustment coefficient. A value greater than 1 indicates an amplified level of confidence in the perceived signal, thereby accelerating state convergence during low-speed or startup phases. The value of should be adjusted while ensuring system stability and noise suppression. A range of 1.2 ≤ 0.5 is generally recommended. ≤3, the specific selection depends on the system motor parameters, sampling period, controller bandwidth and noise level, smaller values ​​are preferred. For example, values ​​of 1.2-1.5 are suitable for light-load, low-noise operating conditions, and can moderately enhance dynamic response without introducing excessive noise. Larger values... For example, 2-3 is used for zero-speed start-up or rapid acceleration scenarios, prioritizing the rapid locking of the rotor position and accepting some of the influence of high-frequency noise.

[0067] By enhancing EKF has a larger correction step for the sensed position signal in each sampling period, making the fused position... It rapidly approximates the actual rotor position, achieving rapid locking at zero speed and stable control at extremely low speeds. Increasing the gain may amplify transient noise from the sensor, therefore, it is dynamically adjusted. And combined with signal quality factor This ensures that gain amplification only applies to high-confidence signals, thus balancing response and noise suppression. The system can quickly lock the rotor position during zero-speed startup, avoiding torque fluctuations caused by initial state uncertainties in sensorless algorithms. Under extremely low speed and emergency braking conditions, the fused position signal... and fusion speed The response speed to braking commands is significantly improved.

[0068] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0069] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0070] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0071] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited thereto, any modification or replacement within the technical range disclosed by the present application can be easily thought by any person skilled in the art, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0072] Finally: the above merely provides the preferred embodiments of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A motor control method combining sensory and sensorless techniques applied to brake-by-wire, characterized in that, include: The raw signal output by the TMR sensor is subjected to online harmonic compensation and temperature drift compensation to obtain a calibration signal. The sensed position of the rotor and the signal quality factor are obtained based on the calibration signal. Based on the acquired phase current and DC bus voltage, the Extended Kalman Filter (EKF) algorithm is run to obtain the sensorless position of the rotor, the rotor speed, and the estimation error covariance matrix of the sensorless position; the residual value is calculated based on the sensorless position and the sensorless position. If the signal quality factor is greater than the preset quality threshold and the residual value is less than the tolerance threshold, then the sensing location is determined to be reliable. Otherwise, the perceived location is deemed unreliable; among them, the value of the tolerance threshold is positively correlated with the estimation error covariance matrix; If the sensed location is reliable, the observation noise covariance matrix is ​​set to the first value, thereby increasing the Kalman gain value, so that the EKF algorithm can perform fusion based on the sensed location in the update step to obtain the fused location and fusion speed. If the sensed location is unreliable, the observation noise covariance matrix is ​​set to a second value that is much larger than the first value, so that the Kalman gain value approaches zero, and the EKF algorithm ignores the sensed location for fusion in the update step to obtain the fused location and fusion speed. Based on the fusion position and fusion speed, control signals are generated for controlling the motor; The weight of the sensed location in the update step can be controlled by setting the value of the observation noise covariance matrix of the EKF algorithm. When the sensed location is reliable, the observation noise covariance matrix is ​​set to the first value, so that the sensed location participates in the state update of EKF as an observation, and the fused location is obtained. When the perceived location is unreliable, the observation noise covariance matrix is ​​set to a second value that is much larger than the first value, so that the EKF update step ignores the perceived location and the unperceived location is used as the fused location.

2. The motor control method combining sensory and sensorless braking applied to line-of-wire braking according to claim 1, characterized in that, include: According to the formula Determine the tolerance threshold ,in, The variance of the estimation error for the non-sensory position is given. and These are preset coefficients.

3. The motor control method combining sensory and sensorless braking applied to line-of-wire braking according to claim 1, characterized in that, include: Read the sensed position output from the TMR sensor; The state vector of the EKF algorithm is initialized using the sensed position, where the rotor position state in the EKF algorithm is assigned the sensed position, and the rotor speed state in the EKF algorithm is assigned zero.

4. The motor control method combining sensory and sensorless braking applied to line-of-wire braking according to claim 1, characterized in that, include: Based on the determination of whether the sensed location is reliable or not, the observation noise covariance matrix in the extended Kalman filter algorithm is smoothly changed. The numerical values ​​are used to achieve a seamless transition; A smoothing factor is calculated based on the signal quality factor and the residual value. And according to the formula To dynamically set the observation noise covariance matrix; where, This represents the matrix value corresponding to the perceived location being reliable. Let be the matrix value corresponding to the unreliable perceived location, and .

5. A motor control method combining sensory and sensorless braking applied to line-of-wire braking according to claim 1, characterized in that, include: Based on the pre-calibrated harmonic error spectrum, the original signal is corrected using a Fourier series compensation model; The harmonic compensation model is ,in, , For the first Compensation coefficient for subharmonics This is for the compensation order.

6. A motor control method combining sensory and sensorless braking applied to line-of-wire braking according to claim 1, characterized in that, include: When the motor is in any of the following operating conditions: zero-speed start-up, extremely low-speed operation, or emergency braking, a performance requirement flag is generated. When the performance requirement flag is valid and the sensed location is determined to be trustworthy, increase the Kalman gain in the EKF algorithm corresponding to the location state.

7. A motor control method combining sensory and sensorless braking applied to line-of-wire braking according to claim 1, characterized in that, include: The TMR sensor adopts a differential output structure, and the differential signal output by the TMR sensor is acquired and converted through the differential input interface or differential amplifier circuit in the motor controller. The installation location of the TMR sensor is optimized based on finite element analysis to maximize the linearity of the sensing magnetic flux change period, thereby reducing harmonic distortion of the original signal from a mechanical structure perspective.

8. A motor control method combining sensory and sensorless braking applied to line-of-wire braking according to claim 1, characterized in that, include: Using the fusion position, the sampled current is transformed to a synchronous rotating coordinate system through coordinate transformation to obtain the direct-axis current and the quadrature-axis current; Closed-loop regulation of direct-axis and quadrature-axis currents generates corresponding voltage commands. Based on voltage commands and fused position, a space vector pulse width modulation algorithm is used to generate the PWM signal that drives the inverter.

Images