A method for suppressing stray magnetic field of an eddy current type vehicle-mounted angle sensor

Abstract

This invention discloses a method for suppressing stray magnetic fields in an eddy current type vehicle-mounted angle sensor, belonging to the field of sensor technology. It aims to solve the problem of stray magnetic field interference affecting vehicle-mounted angle sensors during use, leading to decreased measurement accuracy. The method includes: setting up a multi-layered magnetic shielding structure to block the intrusion path of external stray magnetic fields; employing a differential symmetric structure in the sensor coil layout to cancel common-mode magnetic field interference; and introducing an adaptive magnetic field compensation algorithm in the signal processing stage to correct measurement deviations caused by stray magnetic fields in real time. Through the synergistic effect of the above technical solutions, this application can effectively suppress the interference of stray magnetic fields on eddy current type angle sensors, significantly improve the measurement accuracy and anti-interference capability of the sensor in complex vehicle-mounted electromagnetic environments, and enhance the stability and reliability of the system.

Description

Technical Field

[0001] This invention belongs to the field of sensor technology, specifically, it relates to a method for suppressing stray magnetic fields in an eddy current type vehicle-mounted angle sensor. Background Technology

[0002] As a core sensing element in modern automotive electronic control systems, sensor technology plays a crucial role in acquiring physical signals in vehicle steering, powertrain, transmission, and chassis stability control systems. Its measurement accuracy and operational reliability directly determine the performance and safety capabilities of the entire vehicle's electronic control system. Onboard angle sensors, as an important branch of the sensor family, are widely used in core application scenarios such as electronic power steering angle monitoring, throttle opening detection, transmission gear identification, and wheel speed and yaw rate measurement.

[0003] Eddy current angle sensors operate based on the principle of electromagnetic induction. They generate an alternating magnetic field by passing an alternating current through an excitation coil. The secondary magnetic field produced by the eddy currents induced in the metal target within this alternating magnetic field is then used to perform non-contact position or angle measurements. This measurement method offers significant advantages such as fast response time, high accuracy, strong environmental adaptability, and no mechanical contact with the target object, demonstrating unique applicability in angle measurement applications within the harsh working environments of automobiles.

[0004] However, existing eddy current-type automotive angle sensors face severe stray magnetic field interference problems in practical applications. Modern automobiles contain numerous electromagnetic interference sources, including drive motors, starters, generators, high-current wiring harnesses, and various permanent magnets. The stray magnetic fields generated by these sources superimpose on the target magnetic field generated by the object being measured, causing deviations and distortions in the magnetic field signal detected by the sensor. This leads to decreased accuracy and output fluctuations in angle measurements. Furthermore, the spatial distribution and intensity of stray magnetic fields in the automotive environment exhibit highly nonlinear characteristics due to dynamic changes in the vehicle's operating state. Traditional hardware and software methods such as position compensation or lookup table correction are insufficient to effectively address this time-varying nonlinear interference coupling, making it difficult for the sensor's measurement accuracy and reliability to meet the increasingly stringent performance requirements of advanced driver assistance and autonomous driving systems. The stray magnetic field interference problem has become a key technical bottleneck restricting the further application of eddy current-type angle sensors in the automotive field, urgently requiring practical and effective methodological solutions. Summary of the Invention

[0005] The purpose of this invention is to provide a method for suppressing stray magnetic fields in an eddy current type vehicle-mounted angle sensor, which mainly solves the problem that vehicle-mounted angle sensors are easily interfered with by stray magnetic fields during use, leading to a decrease in measurement accuracy.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for suppressing stray magnetic fields in an eddy current type vehicle-mounted angle sensor includes the following steps:

[0008] S1. A spatial magnetic field distribution calculation model of typical electromagnetic interference sources inside a car is constructed using the finite element simulation module. Based on the finite element simulation results, a time-domain interference model of stray magnetic fields is constructed using the superposition principle.

[0009] S2, the current waveform of the excitation coil is acquired in real time through the current sensor. The acquired current waveform data is preprocessed by low-pass filtering and then transmitted to the reference input port of the multi-order adaptive notch filter.

[0010] S3, the sensor excitation signal generation module uses direct digital frequency synthesis technology to generate a composite excitation signal containing at least two different frequency components, and performs digital sampling based on the composite excitation signal;

[0011] S4, the spectral distribution characteristics of the signal are obtained by performing a fast Fourier transform on the digitized sampled data through a frequency domain analysis algorithm;

[0012] S5, based on a multi-frequency separation algorithm, utilizes the positional difference between the target magnetic field component and the interference magnetic field component in the frequency domain to effectively separate the spectral distribution feature data. The spectral distribution feature data after frequency domain separation is converted back to the time domain by inverse fast Fourier transform.

[0013] S6 uses a multi-stage adaptive notch filter to adaptively filter out power frequency and harmonic magnetic field interference mixed in the signal, and outputs a signal after suppressing power frequency interference.

[0014] S7 utilizes a time-domain infinite impulse response low-pass filter to perform secondary filtering on the output signal of the adaptive notch filter, thereby effectively suppressing residual common-mode interference.

[0015] S8 uses arctangent function calculation to convert the orthogonal magnetic field signal components of the vehicle angle sensor into corresponding angle values.

[0016] Furthermore, the time-domain interference model takes the vehicle powertrain domain, body structure domain, and cockpit electronic equipment domain as the core analysis areas, and uses the spatial coordinates of the drive motor, starter motor, alternator, fuel pump motor, fan motor, and wiper motor as model input parameters. Each interference source is equivalent to a magnetic field dipole array with specific spatial distribution characteristics. The time-domain interference model decomposes the overall interference magnetic field into a static bias component and a dynamic time-varying component. The static bias component is calibrated using the system response test method under Gaussian white noise excitation, and the dynamic time-varying component is mathematically described using an autoregressive moving average model.

[0017] Furthermore, the finite element simulation module adopts an adaptive encryption strategy to control the mesh cell size within 0.1 mm in the near-field region of the interference source, and adopts a gradual coarsening strategy to gradually transition the mesh cell size to 5 mm in the far-field region; after the simulation calculation is completed, it outputs a cloud map of the magnetic field intensity distribution in the three-dimensional space inside the car.

[0018] Furthermore, the calibration process of the static bias component is implemented using a system response test method under Gaussian white noise excitation, including:

[0019] A Gaussian white noise excitation signal with a bandwidth covering DC to 500Hz is applied to the sensor, and the output response of the sensor under the excitation signal is collected. The amplitude and phase characteristics of the static bias component are extracted by statistical averaging. The autoregressive moving average model structure of the dynamic time-varying component is in the form of ARMA, where the autoregressive order is not less than 3.

[0020] Furthermore, in step S3, the digital sampling is achieved by sending the composite excitation signal into the analog front-end preprocessing circuit for processing and then using an analog-to-digital converter. The analog front-end preprocessing circuit includes a first-stage low-noise instrumentation amplifier and a second-stage anti-aliasing low-pass filter.

[0021] Furthermore, the order of the multi-order adaptive notch filter is dynamically determined according to the frequency component complexity of the interference signal: when the number of significant harmonic components contained in the interference signal does not exceed 4, the filter order is set to 4; when the number of harmonic components is in the range of 5 to 6, the filter order is set to 6; when the number of harmonic components reaches 7 or more, the filter order is set to 8; and the notch depth parameter is set to not less than 40dB.

[0022] Furthermore, the time-domain infinite impulse response low-pass filter is implemented using cascaded second-order sections, each consisting of two poles and a finite number of zeros; for the fourth-order filter structure, a configuration of two cascaded second-order sections is adopted; the filter's cutoff frequency parameter is dynamically adjusted within the range of 0.5Hz to 2Hz according to the proportion of high-frequency component energy in the output signal of the vehicle-mounted angle sensor.

[0023] Furthermore, in step S8, a piecewise linearized correction table is used to correct the converted angle; the piecewise linearized correction table divides the 360-degree angle range into at least 8 independent calibration intervals, and each interval uses an independent linear correction model for angle correction.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] (1) This invention constructs a time-domain interference model of stray magnetic fields inside a vehicle and combines a multi-frequency separation algorithm with a multi-order adaptive notch filter to achieve adaptive suppression of power frequency and harmonic interference in complex electromagnetic environments on vehicles. The method first establishes a stray magnetic field interference model covering static bias components and dynamic time-varying components based on finite element simulation and superposition principle. Then, it effectively separates the target magnetic field and the interference magnetic field by utilizing the positional difference in the frequency domain. Finally, the multi-order adaptive notch filter dynamically adjusts the filter order and notch depth according to the complexity of the interference frequency components to achieve accurate tracking and filtering of the power frequency fundamental wave and its harmonics, significantly improving the sensor's anti-interference capability in dynamic electromagnetic interference environments.

[0026] (2) This invention generates a composite excitation signal containing at least two different frequency components by employing direct digital frequency synthesis technology, and combines frequency domain analysis algorithms and inverse fast Fourier transform to achieve effective separation of the target magnetic field and the interfering magnetic field, thereby improving the signal-to-noise ratio and measurement sensitivity without increasing hardware costs. The two frequency components of the composite excitation signal are set in the ranges of 5kHz~10kHz and 20kHz~50kHz, respectively, which avoids overlap with the power frequency and its harmonic frequencies, and makes full use of the flat region of the sensor probe response characteristics, effectively reducing the unexpected attenuation of the target signal during signal processing, and ensuring the accuracy and stability of angle measurement.

[0027] (3) This invention achieves multi-stage progressive suppression of residual common-mode interference and high-precision angle calculation by using a cascaded filtering architecture of a time-domain infinite impulse response low-pass filter and an adaptive notch filter, combined with arctangent function operation and piecewise linearization correction. The low-pass filter adopts a Butterworth-type cascaded second-order structure, and the cutoff frequency is dynamically adjusted in the range of 0.5Hz to 2Hz according to the energy ratio of the high-frequency components of the output signal, effectively suppressing residual noise while ensuring dynamic response speed. In the angle calculation stage, a piecewise linearization correction table is used to divide the 360-degree range into multiple independent calibration intervals. Each interval is compensated by an independent linear correction model, which significantly improves the angle measurement resolution and meets the stringent accuracy requirements of vehicle angle sensors in advanced driver assistance and autonomous driving systems. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the overall process of the present invention. Detailed Implementation

[0029] The present invention will be further described below with reference to the accompanying drawings and embodiments. The embodiments of the present invention include, but are not limited to, the following embodiments.

[0030] like Figure 1As shown, this invention discloses a method for suppressing stray magnetic fields in an eddy current type vehicle-mounted angle sensor. This method first constructs a spatial magnetic field distribution calculation model of typical electromagnetic interference sources inside a vehicle using a finite element simulation module. The model takes the vehicle powertrain domain, body structure domain, and cockpit electronic equipment domain as the core analysis areas, using the spatial coordinates of typical interference sources such as drive motors, starter motors, alternators, fuel pump motors, fan motors, and wiper motors as model input parameters. Each interference source is equivalent to a magnetic field dipole array with specific spatial distribution characteristics, and its magnetic moment direction and amplitude are determined based on the operating current waveform characteristics of the interference source. In the finite element mesh generation stage, an adaptive refinement strategy is adopted to control the mesh cell size within 0.1 mm in the near-field region of the interference source to ensure calculation accuracy in areas with drastic magnetic field gradient changes; in the far-field region, a gradual coarsening strategy is adopted to gradually transition the mesh cell size to 5 mm to balance computational efficiency and accuracy requirements. After the simulation calculation is completed, the system outputs a cloud map of the magnetic field intensity distribution in the three-dimensional space inside the car. The spatial resolution of the cloud map is no less than 1 mm, and the magnetic field intensity changes from 0.1 mT to 50 mT are recorded at intervals of 0.01 T.

[0031] Based on finite element simulation results, the system employs the superposition principle to construct a time-domain interference model of the stray magnetic field. This model decomposes the overall interference magnetic field into two components: a static bias component and a dynamic time-varying component. The static bias component mainly consists of the spatial magnetic field concentration effect of the vehicle's ferromagnetic materials and the geomagnetic field component. Its value remains relatively stable at the sensor installation location but exhibits slow drift characteristics, with a drift time constant on the order of several hours. The calibration process for the static bias component is achieved using a system response test method under Gaussian white noise excitation. Specifically, a Gaussian white noise excitation signal with a bandwidth covering DC to 500Hz is applied to the sensor, and the sensor's output response under the excitation signal is acquired. The amplitude and phase characteristics of the static bias component are extracted using a statistical averaging method. The dynamic time-varying component is mathematically described using an autoregressive moving average model with an ARMA(n,n-1) structure, where the autoregressive order n is not less than 3 to ensure that the model can accurately capture the multi-frequency components of the interference magnetic field. The coefficients of the autoregressive moving average model are obtained by fitting historical interference signal data using the least squares method. During the fitting process, a piecewise stationary assumption strategy is adopted, dividing the long-term interference data sequence into several short-term stationary segments for parameter estimation.

[0032] The reference signal acquisition module operates synchronously with the interference model establishment process. This module acquires the current waveform of the excitation coil in real time using a current sensor. The current sensor has a bandwidth of at least 500kHz, a sampling accuracy of at least 12 significant digits, and a sampling rate set at least 20 times the frequency of the excitation signal to ensure complete preservation of waveform details. After low-pass filtering preprocessing, the acquired current waveform data is directly transmitted to the reference input port of the adaptive filter at the original sampling rate, serving as the reference reference signal for subsequent interference suppression processing. Throughout the entire operating cycle, the reference signal acquisition module continuously runs and updates the reference signal buffer in real time. The buffer depth is set to store sampling data for at least 5 complete excitation cycles to meet the data length requirements of subsequent frequency domain analysis algorithms.

[0033] The sensor excitation signal generation module employs direct digital frequency synthesis (DFD) to generate a composite excitation signal containing at least two different frequency components. The mathematical expression of the composite excitation signal is a superposition of single-carrier frequency signals. The frequency f1 of the first sine component is set within the range of 5kHz to 10kHz, and the frequency f2 of the second sine component is set within the range of 20kHz to 50kHz. The frequency settings of the two components must meet the following constraints: the ratio of f1 to f2 should not be an integer to avoid intermodulation interference products during signal processing; both f1 and f2 should be far from the 50Hz power frequency and its harmonic frequencies to prevent unexpected attenuation of the target signal during notch filtering; f1 and f2 should be within the flat region of the sensor probe's response characteristics to ensure consistent measurement sensitivity. The amplitude ratio is dynamically determined based on the spectral energy distribution characteristics of the target magnetic field signal and the interfering magnetic field signal. When the energy proportion of the interfering signal in the frequency band corresponding to f1 is high, the amplitude of the f2 component is appropriately increased to increase the signal-to-noise ratio of the target signal in that frequency band, and vice versa. The amplitude ratio adjustment process is automatically executed once each time the sensor is powered on and initialized, and then triggered as needed during operation based on the signal quality monitoring results.

[0034] In this embodiment, the amplitude ratio of the two frequency components is set to an adjustable parameter between 1:1 and 1:3, with the specific ratio dynamically determined based on the amplitude of the target magnetic field signal. The clock frequency of the direct digital frequency synthesizer is set to at least 40 times the highest output frequency to ensure the frequency resolution and phase noise performance of the synthesized signal. The composite excitation signal is processed by a digitally programmable gain amplifier with a 60dB dynamic range before digital-to-analog conversion to adapt to the needs of different measurement ranges. The signal output stage adopts a differential drive architecture, with the output impedance set to 50 ohms to achieve impedance matching with the sensor coil impedance.

[0035] The composite magnetic field response signal received by the sensor probe is first pre-processed by the analog front-end. The pre-processing circuit includes a first-stage low-noise instrumentation amplifier and a second-stage anti-aliasing low-pass filter. The input noise voltage density of the instrumentation amplifier is no greater than 5nV / sqrt(Hz), and the common-mode rejection ratio is no less than 100dB to ensure effective amplification of weak signals and effective suppression of common-mode interference. The -3dB cutoff frequency of the anti-aliasing low-pass filter is set to 1.2 times the highest excitation frequency f2, and the filter order is no less than 6 to achieve a sufficiently steep transition band attenuation characteristic. The pre-processed analog signal is then digitally sampled by an analog-to-digital converter with a resolution of no less than 16 bits. The sampling rate is set to more than 10 times the highest signal frequency to meet the Nyquist sampling theorem requirements. The digitally sampled data is first stored in a dual-port random access memory (DRAM) to buffer the data between the analog front-end and the digital signal processing unit.

[0036] In this embodiment, a frequency domain analysis algorithm is used to perform Fast Fourier Transform (FFT) processing on the digitized sampled data to obtain the spectral distribution characteristics of the signal. The number of computation points for the FFT is set to 2048 to balance the frequency resolution and real-time computation requirements. The frequency resolution is determined by the ratio of the sampling rate to the number of computation points, and under typical sampling rate settings, this resolution is not less than 10Hz. The spectral analysis results include the amplitude and phase information of the target magnetic field component and the interference magnetic field component at different frequency positions. The target magnetic field component exhibits significant peaks at excitation frequencies f1 and f2, and its amplitude has a sine or cosine function relationship with the measured angle; the interference magnetic field component exhibits characteristic spectral peaks at 50Hz and its integer multiples of 50Hz, and also has a noise floor enhancement component distributed in the broadband frequency band. By setting the spectral peak detection threshold and the spectral energy threshold, the system automatically identifies and extracts the characteristic parameters of the target magnetic field component and the interference magnetic field component from the complex spectrum.

[0037] The multi-frequency separation algorithm effectively separates the target magnetic field component from the interference magnetic field component by utilizing their positional differences in the frequency domain. The core processing logic of the algorithm involves preserving the spectral lines at the target frequency position while simultaneously filling or attenuating the spectral lines at the interference frequency position. The frequency-separated spectral data is then converted back to the time domain via an inverse fast Fourier transform, resulting in two independent data streams: one for the target magnetic field signal and one for the interference magnetic field signal. The target magnetic field signal's time-domain waveform is directly transmitted to the subsequent angle calculation module; the interference magnetic field signal's time-domain waveform is transmitted to the desired signal input port and reference signal channel of the adaptive notch filter, forming the data basis for closed-loop interference suppression processing.

[0038] The order of the adaptive notch filter is set between 4 and 8, with the specific order dynamically determined based on the frequency complexity of the interference signal. When the number of significant harmonic components in the interference signal does not exceed four, the filter order is set to 4; when the number of harmonic components is between five and six, the filter order is set to 6; and when the number of harmonic components reaches seven or more, the filter order is set to 8. The adaptive notch filter is implemented using a multi-notch parallel structure, with the center frequency of each notch aligned with the nominal frequency of the 50Hz fundamental frequency and its various harmonics. In the parallel structure, each notch branch operates independently, allowing for independent adjustment of the parameters of a single notch branch when the interference spectrum dynamically changes without redesigning the overall filter structure.

[0039] The precise calibration process for the notch filter center frequency is achieved using the spectral analysis results of the output interference magnetic field signal. The calibration algorithm first searches the spectral data for peak positions with amplitudes exceeding the average noise floor by 20 dB, identifying these peak positions as interference frequency points. For each identified interference frequency point, the system calculates the deviation between this frequency point and the nearest standard power frequency harmonic frequency position, and adds this deviation as a fine-tuning compensation amount to the standard frequency value. The standard power frequency harmonic frequency position is determined by the nominal power grid frequency of 50 Hz and its integer multiples. However, since the power grid frequency fluctuates within a range of ±0.5 Hz in actual operation, the fine-tuning compensation amount for the notch filter center frequency needs to be dynamically tracked and updated within this ±0.5 Hz range. The notch depth parameter is set to no less than 40 dB to ensure sufficient attenuation of the interference signal at the notch filter frequency position.

[0040] The adaptive filtering algorithm uses the normalized minimum mean square error criterion to iteratively update the filter weights. The convergence factor of the algorithm is set in the range of 0.01 to 0.1, and the specific value is determined according to the time-varying characteristics of the interference signal. When the amplitude of the interference signal changes drastically, a larger convergence factor value is used to accelerate the tracking speed; when the amplitude of the interference signal tends to stabilize, a smaller convergence factor value is used to reduce the steady-state error. During the normalization process, the convergence step size is normalized using the input signal power estimate. The power estimate is implemented using the exponentially weighted moving average method, and the estimation time constant is 10 sampling periods. The filter weight update equation adopts the following mathematical form:

[0041]

[0042] in, This represents the filter weight vector at time n. Indicates the convergence factor. Represents the reference input signal vector. Indicates the error signal. This represents the regularization constant that ensures numerical stability. This represents the power estimate of the reference input signal.

[0043] Under dynamic operating conditions where the interference signal amplitude changes by 100%, the adaptive filter maintains a steady-state output within 100ms from the moment the interference amplitude jumps. The steady-state condition is defined as the filter weight update amount being less than a set threshold, set to 0.1% of the initial weight, for 10 consecutive sampling periods. The time-varying characteristic of the notch depth is determined by the convergence direction of the filter weight vector. The notch depth reaches its maximum set value when the weight vector converges to a state matching the frequency components of the interference signal; conversely, the notch depth decreases when the weight vector deviates from the matching state. The convergence direction of the filter weights is driven by an error signal, which is the difference between the desired interference signal and the filter output signal. When the error signal approaches zero, it indicates that the filter has accurately tracked the interference characteristics and completed the adaptive adjustment of the notch parameters.

[0044] A time-domain infinite impulse response (IIR) low-pass filter performs secondary filtering on the output signal of an adaptive notch filter. The filter employs a Butterworth response characteristic to achieve a flat passband amplitude response and reasonable phase shift characteristics. The filter structure is implemented using cascaded second-order sections, each consisting of two poles and a finite number of zeros. The pole positions are determined according to the filter order and cutoff frequency parameters using the Butterworth polynomial coefficient table. For a typical fourth-order filter structure, a configuration of two cascaded second-order sections is used, with the quality factors (Q) of each section set to 0.517 and 1.306, respectively, to achieve the maximum flat amplitude response. The filter's cutoff frequency parameter is dynamically adjusted within the range of 0.5Hz to 2Hz. The adjustment strategy is determined based on the proportion of high-frequency components in the sensor output signal: when the proportion of high-frequency components exceeds 10% of the total energy, the cutoff frequency is increased to accelerate the dynamic response; when the proportion of high-frequency components is less than 5%, the cutoff frequency is decreased to enhance noise suppression.

[0045] The Butterworth low-pass filter exhibits a phase delay of no more than 1 ms within a signal bandwidth of 0 to 500 Hz. This performance is ensured by selecting an appropriate pole distribution structure and implementation during the filter design phase. The direct-type II structure has the smallest inherent delay characteristics compared to other infinite impulse response filter implementations, making it suitable for applications sensitive to phase distortion. Phase equalization technology is also employed during filter design to compensate for the remaining phase nonlinearity within the passband. The group delay characteristics of the compensation network are opposites of the original filter's group delay characteristics, and the overall group delay after cascading remains constant within the signal passband. The phase-equalized filter's group delay jitter is less than 0.1 ms, ensuring that the output signal's time waveform is not distorted.

[0046] This invention employs a differential algorithm to effectively suppress residual common-mode interference. The differential processing architecture comprises two parallel signal acquisition links: a main measurement channel and an auxiliary reference channel. The sensor probe of the main measurement channel is installed in the target magnetic field region, responsible for acquiring the comprehensive signal including both the target magnetic field component and the interfering magnetic field component. The sensor probe of the auxiliary reference channel is installed in the interference field region outside the target magnetic field region, acquiring only the environmental interference magnetic field signal and being insensitive to the target magnetic field. The signal acquisition of the two channels is completely synchronized in the time domain, with synchronization accuracy guaranteed by a shared sampling clock, and the clock synchronization deviation not exceeding one sampling period. The two sampled data are subtracted in the digital domain. In the result, the target magnetic field component is retained because it exists in both channels simultaneously, while the interfering magnetic field component is eliminated in the differential operation because it exists only in the main measurement channel.

[0047] The amplitude consistency deviation between the main channel and the reference channel is controlled within 1% to ensure the effectiveness of differential operations. Amplitude consistency calibration is completed during system initialization. The calibration process uses a standard magnetic field source to apply excitation signals of the same amplitude to both channels, acquires the output responses of both channels, and calculates the amplitude scaling factor. This scaling factor is stored in the system parameter configuration area and is used to adjust the amplitude matching of the data to be processed before each differential operation. The common-mode interference rejection ratio (CMIR) in the signal after differential operation is no less than 60 dB. This metric is obtained by calculating the ratio of common-mode gain to differential-mode gain under ideal conditions and converting it to logarithmic coordinates. Any small amount of residual interference components in the differential operation results are further filtered out by a subsequent narrowband low-pass filter. The cutoff frequency of this low-pass filter is set within the same dynamic adjustment range as the time-domain infinite impulse response filter.

[0048] The angle calculation algorithm uses the arctangent function to convert the orthogonal magnetic field signal components of the sensor into corresponding angle values. The orthogonal signal output by the sensor typically contains two orthogonal paths: a cosine component and a sine component, with equal amplitudes and a 90-degree phase difference. Under ideal conditions, the angle value at any given time is related to the orthogonal components by the arctangent function. The mathematical expression for angle calculation is:

[0049]

[0050] in, This represents the calculated target angle value. This represents the sampled voltage value of the sinusoidal channel. This represents the sampled voltage value of the cosine channel. The angle value range generated by the above arctangent operation is from -90 degrees to +90 degrees. When the sign combination of the sine and cosine channels indicates that the angle value is in the second or third quadrant, quadrant correction processing is required on the original output of the arctangent function to obtain the correct angle value from 0 to 360 degrees.

[0051] This embodiment employs a piecewise linearization calibration table for angle correction after conversion. The table divides the 360-degree angle range into at least eight independent calibration intervals, each with an angle span of 45 degrees. The boundary positions between adjacent intervals are determined based on the distribution characteristics of zero-crossing and extreme points of the orthogonal signals, typically set at discontinuities or locations with significant nonlinearity. Each interval uses an independent linear correction model, which includes two parameters: a slope coefficient and an offset coefficient. The slope coefficient compensates for the amplitude gain error of the signal within the interval, while the offset coefficient compensates for the DC bias error. The specific form of the linear correction model is that the corrected angle value equals the original angle value multiplied by the slope coefficient plus the offset coefficient. Typical values ​​for the slope coefficient range from 0.98 to 1.02, and typical values ​​for the offset coefficient range from -0.5 degrees to +0.5 degrees. The application of piecewise linearization calibration technology enables an angle measurement resolution of 0.01 degrees, meeting the accuracy requirements of vehicle-mounted angle sensors.

[0052] The above embodiments are merely one of the preferred embodiments of the present invention and should not be used to limit the scope of protection of the present invention. Any modifications or refinements made to the main design concept and spirit of the present invention that are not of substantial significance, but solve the same technical problem as the present invention, should be included within the scope of protection of the present invention.

Claims

1. An eddy current vehicle-mounted angle sensor stray magnetic field suppression method, characterized by, Includes the following steps: S1. A spatial magnetic field distribution calculation model of typical electromagnetic interference sources inside a car is constructed using the finite element simulation module. Based on the finite element simulation results, a time-domain interference model of stray magnetic fields is constructed using the superposition principle. S2, the current waveform of the excitation coil is acquired in real time through the current sensor. The acquired current waveform data is preprocessed by low-pass filtering and then transmitted to the reference input port of the multi-order adaptive notch filter. S3, the sensor excitation signal generation module uses direct digital frequency synthesis technology to generate a composite excitation signal containing at least two different frequency components, and performs digital sampling based on the composite excitation signal; S4, the spectral distribution characteristics of the signal are obtained by performing a fast Fourier transform on the digitized sampled data through a frequency domain analysis algorithm; S5, based on a multi-frequency separation algorithm, utilizes the positional difference between the target magnetic field component and the interference magnetic field component in the frequency domain to effectively separate the spectral distribution feature data. The spectral distribution feature data after frequency domain separation is converted back to the time domain by inverse fast Fourier transform. S6 uses a multi-stage adaptive notch filter to adaptively filter out power frequency and harmonic magnetic field interference mixed in the signal, and outputs a signal after suppressing power frequency interference. S7 utilizes a time-domain infinite impulse response low-pass filter to perform secondary filtering on the output signal of the adaptive notch filter, thereby effectively suppressing residual common-mode interference. S8 uses arctangent function calculation to convert the orthogonal magnetic field signal components of the vehicle angle sensor into corresponding angle values.

2. The method for suppressing stray magnetic fields of an eddy current type vehicle-mounted angle sensor according to claim 1, characterized in that, The time-domain interference model takes the vehicle powertrain domain, body structure domain, and cockpit electronic equipment domain as the core analysis areas, and uses the spatial position coordinates of the drive motor, starter motor, alternator, fuel pump motor, fan motor, and wiper motor as the model input parameters. Each interference source is equivalent to a magnetic field dipole array with specific spatial distribution characteristics. The time-domain interference model decomposes the overall interference magnetic field into a static bias component and a dynamic time-varying component. The static bias component is calibrated using the system response test method under Gaussian white noise excitation, while the dynamic time-varying component is mathematically described using an autoregressive moving average model.

3. The method for suppressing stray magnetic fields of an eddy current type vehicle-mounted angle sensor according to claim 2, characterized in that, The finite element simulation module adopts an adaptive encryption strategy to control the mesh cell size within 0.1 mm in the near-field region of the interference source, and a gradual coarsening strategy to gradually transition the mesh cell size to 5 mm in the far-field region; after the simulation calculation is completed, it outputs a cloud map of the magnetic field intensity distribution in the three-dimensional space inside the car.

4. The method for suppressing stray magnetic fields of an eddy current type vehicle-mounted angle sensor according to claim 3, characterized in that, The calibration process of the static bias component is implemented using a system response test method under Gaussian white noise excitation, including: A Gaussian white noise excitation signal with a bandwidth covering DC to 500Hz is applied to the sensor, and the output response of the sensor under the excitation signal is collected. The amplitude and phase characteristics of the static bias component are extracted by statistical averaging. The autoregressive moving average model structure of the dynamic time-varying component is in the form of ARMA, where the autoregressive order is not less than 3.

5. The method for suppressing stray magnetic fields of an eddy current type vehicle-mounted angle sensor according to claim 4, characterized in that, In step S3, the digital sampling is achieved by sending the composite excitation signal into the analog front-end preprocessing circuit for processing and then using an analog-to-digital converter. The analog front-end preprocessing circuit includes a first-stage low-noise instrumentation amplifier and a second-stage anti-aliasing low-pass filter.

6. The method for suppressing stray magnetic fields of an eddy current type vehicle-mounted angle sensor according to claim 5, characterized in that, The order of the multi-order adaptive notch filter is dynamically determined based on the frequency component complexity of the interference signal: when the number of significant harmonic components in the interference signal does not exceed 4, the filter order is set to 4; when the number of harmonic components is in the range of 5 to 6, the filter order is set to 6; when the number of harmonic components reaches 7 or more, the filter order is set to 8; and the notch depth parameter is set to not less than 40dB.

7. The method for suppressing stray magnetic fields of an eddy current type vehicle-mounted angle sensor according to claim 6, characterized in that, The time-domain infinite impulse response low-pass filter is implemented using cascaded second-order sections, each consisting of two poles and a finite number of zeros. For the fourth-order filter structure, a configuration of two cascaded second-order sections is adopted. The cutoff frequency parameter of the filter is dynamically adjusted within the range of 0.5Hz to 2Hz according to the proportion of high-frequency component energy in the output signal of the vehicle angle sensor.

8. The method for suppressing stray magnetic fields of an eddy current type vehicle-mounted angle sensor according to claim 7, characterized in that, In step S8, a piecewise linearized correction table is used to correct the converted angle; the piecewise linearized correction table divides the 360-degree angle range into at least 8 independent calibration intervals, and each interval uses an independent linear correction model for angle correction.

Images