A dynamic baseline update and anti-interference method for in-vehicle touch detection

Abstract

The present application relates to the technical field of vehicle-mounted detection, and particularly relates to a dynamic baseline updating and anti-interference method for vehicle-mounted touch detection, comprising the following steps: collecting a capacitive sensor original signal and an environmental temperature and humidity signal, performing state discrimination processing on the original signal and the environmental temperature and humidity signal to obtain a touch-free state signal set; calculating a temperature and humidity change amount, performing adaptive filtering calculation, obtaining dynamic baseline data and forming a dynamic detection baseline; monitoring a change amplitude and a change rate of the dynamic detection baseline, performing abnormal correction on the dynamic detection baseline according to a monitoring result, and obtaining a stabilized detection baseline; calculating a signal deviation, performing noise suppression processing, and obtaining a purified touch detection signal; performing feature parameter quantization extraction, performing hierarchical comparison and verification on the quantized feature parameters and hierarchical thresholds, and obtaining a touch detection result. The present application effectively improves the reliability and accuracy of vehicle-mounted touch detection by using temperature and humidity linkage dynamic baseline adaptive updating.

Description

Technical Field

[0001] This invention relates to the field of vehicle-mounted detection technology, and in particular to a dynamic baseline update and anti-interference method for vehicle-mounted touch detection. Background Technology

[0002] With the development of automotive intelligence, in-vehicle touch interaction has become a core human-machine interaction method in smart cockpits, widely used in central control screens and steering wheel touch areas. Its detection accuracy and anti-interference capabilities directly affect driving safety and user experience. Currently, capacitive sensing technology is the mainstream approach, which identifies touch actions by detecting changes in capacitance signals caused by human touch. Compared to traditional mechanical buttons, it has significant advantages and has become the mainstream solution for in-vehicle interaction.

[0003] However, the unique characteristics of the current in-vehicle environment (such as vibration, temperature changes, and electromagnetic complexity) affect the reliability and stability of capacitive touch detection. Multiple sources of interference exist during vehicle operation, including driving vibrations, low-frequency temperature drift caused by changes in ambient temperature and humidity, high-frequency electromagnetic interference from in-vehicle electronic devices, and foreign objects such as water droplets and sweat on the screen surface. Existing technologies struggle to effectively distinguish these interferences from genuine touch signals, easily leading to misjudgments or missed detections. Traditional solutions often employ fixed baselines or simple adaptive baseline algorithms, unable to perform real-time calibration in response to dynamic changes in temperature and humidity. Due to the lack of effective anomaly detection and correction mechanisms, the baseline is prone to shift, resulting in distorted signal deviation calculations, and in severe cases, even functional failure.

[0004] In summary, existing automotive capacitive touch detection technologies are insufficient to meet the stringent requirements of automotive scenarios. Therefore, developing a detection method capable of dynamic baseline adaptive updates is a pressing technical challenge. Summary of the Invention

[0005] This invention solves the technical problems of baseline deviation, poor interference suppression, and insufficient accuracy in existing vehicle capacitive touch detection by adopting dynamic baseline adaptive update linked to temperature and humidity. It effectively improves the reliability, accuracy, and environmental adaptability of vehicle touch detection, ensuring driving interaction safety and user experience.

[0006] The technical solution proposed in this invention is: a dynamic baseline update and anti-interference method for vehicle-mounted touch detection, the method comprising: The original signal from the capacitive sensor and the ambient temperature and humidity signal are collected, and the original signal and the ambient temperature and humidity signal are processed to obtain a set of no-touch state signals. The temperature and humidity change is calculated based on the ambient temperature and humidity signal. The temperature and humidity change is combined with the temperature and humidity change to perform adaptive filtering calculation on the non-touch state signal set to obtain dynamic baseline data and form a dynamic detection baseline. Monitor the magnitude and rate of change of the dynamic detection baseline, and correct any anomalies in the dynamic detection baseline based on the monitoring results to obtain a stable detection baseline; The signal deviation is calculated based on the stabilized detection baseline and the original signal. The original signal is then subjected to noise suppression processing based on the signal deviation to obtain the purified touch detection signal. The enhanced touch detection signal is subjected to quantification and extraction of feature parameters. The quantified feature parameters are compared and verified with the hierarchical threshold. The touch state is determined based on the hierarchical comparison and verification results to obtain the touch detection results.

[0007] Preferably, the process of acquiring the original signal from the capacitive sensor and the ambient temperature and humidity signal is as follows: The vehicle-mounted capacitive sensor starts signal acquisition at a preset sampling frequency to obtain the original signal from the capacitive sensor corresponding to the touch sensing. Synchronously control the temperature and humidity sensor in the vehicle's touch area to start collecting data and obtain real-time ambient temperature and humidity signals; The acquired raw signals from the capacitive sensor, ambient temperature signal, and ambient humidity signal are synchronously buffered to complete the signal acquisition process and form a raw signal set.

[0008] Preferably, the process for acquiring the set of touchless state signals is as follows: The original signal set is preprocessed by using low-pass filtering to remove high-frequency invalid noise generated by external electromagnetic field coupling and threshold filtering to remove abnormal data, resulting in purified multi-channel synchronization signals. The purified multi-channel synchronization signals are compared with the preset corresponding thresholds to select the timing signals that meet all the conditions. The selected timing signals are then subjected to timing normalization and standardization to obtain the set of touchless state signals.

[0009] Preferably, the process for obtaining the dynamic detection baseline is as follows: Collect ambient temperature and humidity signals corresponding to the non-touch state signal set, calculate the temperature change and humidity change per unit time, and fuse them to obtain the comprehensive temperature and humidity change. By combining the comprehensive temperature and humidity changes, adaptive filtering calculations are performed on the central control channel signal and the steering wheel channel signal, which are concentrated in the non-touch state, respectively, to obtain the effective steady-state characteristic data after filtering. The effective steady-state characteristic data are fitted according to the time series, and the dynamic detection baseline of the central control system and the dynamic detection baseline of the steering wheel are constructed simultaneously.

[0010] Preferably, the process for obtaining the stabilized detection baseline is as follows: Real-time monitoring is performed on the dynamic detection baseline of the central control unit and the dynamic detection baseline of the steering wheel, and the change amplitude and rate of change of the two dynamic detection baselines are collected. For each region, two dynamic detection baselines are set with their own baseline stability judgment thresholds. The monitored change amplitude and change rate are compared with the corresponding thresholds. When the change amplitude or change rate of the dynamic baseline in either region exceeds the corresponding threshold, anomaly correction processing is performed on the baseline of that region. Once the baselines of each region have been corrected to be qualified, the central control stability test baseline and the steering wheel stability test baseline are output synchronously.

[0011] Preferably, the process of obtaining the purified touch detection signal is as follows: Using the central control stability detection baseline and the steering wheel stability detection baseline as reference benchmarks respectively, the signal deviation between the original capacitor signal and the reference benchmark in the corresponding area is calculated independently. Based on the signal deviation values ​​of each region, the noise frequency bands of the corresponding regions are divided. Based on the divided noise frequency bands, the capacitive signals of the non-touch state signal set are subjected to layered noise reduction processing to obtain the noise-reduced signal. The noise-reduced signal is processed by timing smoothing and amplitude normalization to simultaneously obtain the central control air purification touch detection signal and the steering wheel air purification touch detection signal.

[0012] Preferably, the specific process for performing the hierarchical comparison and verification is as follows: Feature extraction was performed on the central control air purification touch detection signal and the steering wheel air purification touch detection signal to obtain the feature parameters of the two signals respectively; The two signals are divided into regions and the hierarchical comparison criteria are set independently. The extracted feature parameters are compared with their respective first-level thresholds to determine their validity and select valid candidate signals. The effective candidate signals are compared with their respective second-level thresholds for stability, and the effective feature signals are output.

[0013] Preferably, the process for obtaining the touch detection result is as follows: The effective feature signals are matched and compared with the preset state calibration rules to obtain preliminary judgment results; By comparing the preliminary judgment result with the core parameters of the effective feature signal in reverse, and combining the preset misjudgment judgment threshold, the misjudgment result is eliminated to obtain the accurate state judgment result. The accurate status determination results are integrated, structured coding and format regularization are performed, and the touch detection results are output.

[0014] The present invention also provides a computer-readable storage medium storing a computer program that is executed by a processor to implement the aforementioned dynamic baseline update and anti-interference method for in-vehicle touch detection.

[0015] The beneficial effects of this invention are: 1. The system collects ambient temperature and humidity signals and calculates the changes in temperature and humidity. This data is then used to adaptively filter the non-touch state signal set to generate a dynamic detection baseline. Simultaneously, the system monitors the amplitude and rate of baseline change and performs anomaly corrections, ultimately obtaining a stable detection baseline. This system dynamically adapts to real-time changes in in-vehicle temperature and humidity, precisely adjusts baseline parameters, and achieves real-time baseline calibration and maintenance. This effectively avoids signal deviation calculation distortion caused by baseline shift, reduces detection misjudgments caused by baseline drift, and significantly improves the basic accuracy of touch detection. It lays a stable and reliable foundation for subsequent signal processing and feature extraction, solving the technical pain points of traditional fixed baselines and simple adaptive baselines that cannot adapt to complex in-vehicle environments and are prone to shifting.

[0016] 2. Based on a stable detection baseline, a continuous time-series signal deviation sequence is constructed and divided into three noise bands (low-frequency, mid-frequency, and high-frequency) through FFT analysis. For each noise band, a dedicated filtering algorithm (trend fitting, moving average, threshold limiting) is used to perform layered and frequency-band noise reduction. This method accurately distinguishes and specifically suppresses noise of different mechanisms, avoiding signal distortion or interference residue caused by traditional single-filter approach. The resulting purified signal waveform is stable and free of noise interference, ensuring that subsequent feature extraction can capture real touch signal changes, effectively improving signal purity. Ultimately, this solves the technical problems of poor targeting and excessive interference residue in traditional noise reduction methods.

[0017] 3. A fixed window is extracted from the purified continuous time-series signal to extract three core feature parameters: steady-state amplitude, rising edge rate, and effective duration. Two-level thresholds are set for different regions, and effective feature signals are filtered through hierarchical comparison logic of validity judgment and stability verification. By capturing the temporal change trend of the signal through a continuous time-series window, the extracted feature parameters are made more representative. The two-level verification logic can effectively eliminate instantaneous interference and spurious signals. The regional adaptation design meets the diverse needs of in-vehicle applications, avoids misjudgment caused by single-frame signal fluctuations, improves the accuracy and reliability of touch status determination, and ensures the safety and convenience of driving interaction. Ultimately, this solves the problem of inaccurate detection caused by traditional single-frame single-point feature extraction and single threshold judgment. Attached Figure Description

[0018] Figure 1 This is a flowchart of a dynamic baseline update and anti-interference method for vehicle-mounted touch detection; Figure 2A flowchart illustrating the temperature and humidity-linked dynamic baseline update process for a dynamic baseline update and anti-interference method used in vehicle-mounted touch detection. Figure 3 A flowchart illustrating the frequency-band hierarchical noise reduction and anti-interference process of a dynamic baseline update and anti-interference method for vehicle-mounted touch detection; Figure 4 This is a flowchart of a two-level hierarchical comparison and verification method for dynamic baseline update and anti-interference in vehicle-mounted touch detection. Detailed Implementation

[0019] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0020] It is understood that the term "a" should be understood as "at least one" or "one or more," that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0021] like Figure 1 As shown, a dynamic baseline update and anti-interference method for vehicle-mounted touch detection is presented, the method comprising: The system collects raw signals from a capacitive sensor and ambient temperature and humidity signals. These signals undergo state discrimination processing to obtain a set of no-touch state signals. Temperature and humidity changes are calculated based on the ambient temperature and humidity signals. Adaptive filtering is then applied to the no-touch state signal set to obtain dynamic baseline data and form a dynamic detection baseline. The amplitude and rate of change of the dynamic detection baseline are monitored. Anomalies are corrected based on the monitoring results to obtain a stable detection baseline. The signal deviation between the stable detection baseline and the raw signal is calculated. Noise suppression is applied to the raw signal based on this deviation to obtain a purified touch detection signal. Feature parameters are quantized and extracted from the purified touch detection signal. These quantized feature parameters are then compared and verified against hierarchical thresholds. The touch state is determined based on the hierarchical comparison and verification results to obtain the touch detection result.

[0022] Furthermore, the process of acquiring the raw signal from the capacitive sensor and the ambient temperature and humidity signals is as follows: The vehicle-mounted capacitive sensor starts signal acquisition at a preset sampling frequency to obtain the original signal from the capacitive sensor corresponding to the touch sensing. Simultaneously, the temperature and humidity sensor in the vehicle-mounted touch area starts acquisition to obtain real-time ambient temperature and humidity signals. The acquired original signals from the capacitive sensor, ambient temperature, and ambient humidity are synchronously buffered to complete the signal acquisition process and form an original signal set.

[0023] In this embodiment, capacitive sensors covering both the central touchscreen area and the steering wheel touchscreen area are installed. These sensors utilize multi-channel capacitive sensing chips specifically designed for automotive touchscreens, with a 12-bit sampling accuracy and a preset sampling frequency of 10Hz. During signal acquisition, the raw signals from the two capacitive sensors are collected through different channels, ensuring no interference. Furthermore, temperature and humidity sensors are installed in corresponding areas of the central touchscreen area and the steering wheel touchscreen area. The temperature measurement range is set to -40℃ to 85℃, and the humidity measurement range is set to 0%RH to 100%RH. The sampling frequency is consistent with that of the capacitive sensors (10Hz) to ensure signal timing synchronization.

[0024] Preferably, hardware timing triggering is used to acquire the raw signals from the capacitive sensor and the ambient temperature and humidity signals. During acquisition, the FPGA generates a unified trigger signal with a preset trigger interval of 100ms, ensuring synchronization between the capacitive sensor and the temperature and humidity sensor, keeping the time synchronization deviation within ±5ms. Data transmission uses an I2C communication bus with a preset communication rate of 100kHz, transmitting the raw signals from the capacitive sensor (both central control and steering wheel signals), ambient temperature signal, and ambient humidity signal in real time to the buffer module for buffering.

[0025] The detailed implementation process of this step is as follows: After the device is powered on, it first initializes and configures the two capacitive sensors and the temperature and humidity sensor, including the sensing sensitivity, sampling frequency, and ADC sampling accuracy parameters of the capacitive sensors, as well as the measurement range, sampling interval, and communication address parameters of the temperature and humidity sensors. After the configuration is completed, the system triggers the two sensors to start synchronous acquisition at regular intervals and performs continuous sampling according to a fixed sampling period of 10Hz.

[0026] After acquiring the raw signals from the capacitive sensor and the ambient temperature and humidity signals, the format of each round of raw data is validated, and invalid data packets with missing frames or verification errors are removed. A 3σ outlier criterion is used to remove abrupt abnormal data points that exceed the sensor's normal measurement range. The specific process of the outlier criterion is as follows: data exceeding the range from the difference between the data mean and three times the standard deviation to the sum of the data mean and three times the standard deviation are removed. The data mean and standard deviation are statistically derived from the first 100 frames of valid acquired data.

[0027] After verification, the data is timestamped and stored sequentially in the on-chip high-speed cache module according to the acquisition time sequence. Timing sorting and temporary archiving are performed to avoid data disorder, loss, and timing misalignment. The timing sorting and temporary archiving logic is as follows: An independent cache queue is allocated to each sensor data channel. The cache module has a built-in timestamp comparator. When receiving data, it prioritizes verifying the integrity of the timestamps. If there is a time difference caused by I2C bus time-sharing transmission (maximum not exceeding 5ms), linear interpolation is used to complete the timing nodes of the lagging data, ensuring strict alignment of the timestamps of the central control capacitor, steering wheel capacitor, and temperature and humidity data (deviation ≤ ±1ms). After timing alignment, the data is arranged in ascending order along the time axis and temporarily archived. The archiving time is preset to 10s, and the earliest data is automatically overwritten after the timeout. The original signals from the central control capacitor sensor, steering wheel capacitor sensor, ambient temperature signal, and ambient humidity signal are collected and integrated along the time axis to form a multi-channel time-aligned original signal set. The entire process takes ≤100ms, meeting the real-time requirements of vehicle touch detection and steering wheel hands-off monitoring.

[0028] The percentage of valid data in the original signal set is calculated using the following formula: Valid data percentage = (Number of valid data frames in the original signal set ÷ Total number of frames in the original signal set) × 100%. If the percentage of valid data is ≥ 98%, the original signal set is considered complete and qualified; if it is not qualified, the acquisition is retried, with a maximum of 3 retries. If the number of retries exceeds 3, the acquisition is stopped, and a device fault warning is triggered.

[0029] Specifically, the equipment and process details for data acquisition and caching are as follows: The data acquisition device integrates a multi-channel capacitance acquisition unit, a temperature and humidity acquisition unit, and a data cache unit. It has a built-in low-power microprocessor with a main frequency of no less than 1GHz, an operating voltage within the preset range of 3.3-5V, a static power consumption of ≤15mA, a dynamic power consumption of ≤50mA, and is compatible with vehicle ambient temperatures of -40℃ to 85℃ and relative humidity of 0% to 100%, ensuring stable operation in high and low temperature and multi-interference vehicle scenarios.

[0030] Furthermore, the process for acquiring the no-touch state signal set is as follows: The original signal set is preprocessed by using low-pass filtering to remove high-frequency invalid noise generated by external electromagnetic field coupling and threshold filtering to remove abnormal data, resulting in purified multi-channel synchronization signals. The purified multi-channel synchronization signals are compared with the corresponding preset thresholds to select the timing signals that meet all conditions. The selected timing signals are then subjected to timing normalization and standardization to obtain a set of touchless state signals.

[0031] like Figure 2 As shown, specifically, a first-order passive low-pass filter algorithm is used to preprocess the original signal set, with a preset cutoff frequency of 5Hz, to filter out high-frequency invalid noise introduced by vehicle electromagnetic interference, power supply ripple, and vehicle vibration. The filtering process uses a recursive calculation method, and the filtering formula is: ,in, The filter coefficient (can be adjusted to 0.1 when the interference is strong, and to 0.3 when the interference is weak. The variance of the signal before filtering is calculated. If the variance is ≥0.01, it is determined to be strong interference; if the variance is <0.01, it is determined to be weak interference). This is the original data of the current frame. This is the filtering result of the previous frame. This is the filtering result for the current frame. The signal is filtered using the above formula to ensure that the filtered signal is stable and free from significant distortion.

[0032] Based on the rated measurement range of each sensor, the filtered raw signal set is threshold-screened to remove abnormal data such as jumps and sensor false triggers that exceed the range, resulting in purified multi-channel synchronous signals. The threshold range of the raw signal of the capacitive sensor is set to ±10% of its rated range, and the threshold range of the temperature and humidity signal is set to the normal fluctuation range of a typical vehicle environment (temperature -20℃~60℃, humidity 20%RH~80%RH).

[0033] Then, the purified multi-channel synchronous signals are filtered according to the corresponding screening thresholds. The preset thresholds are divided into two categories. The first category is capacitance-related thresholds, including the central control capacitor no-touch threshold and the steering wheel capacitor hands-free determination threshold. Continuous steady-state capacitance signals are collected under normal temperature and pressure static interference-free conditions, under the baseline conditions of completely no touch control of the central control unit and naturally untouched hands-free steering wheel. The mean and standard deviation of the signal amplitude are statistically analyzed to calibrate the central control capacitor no-touch threshold range. The threshold range for determining whether the steering wheel capacitor is released from the hand is: , The average signal amplitude of the central control capacitor is calculated by collecting a large number of samples in a completely untouched and idle state. This represents the average signal amplitude of a large number of samples from the steering wheel capacitor when it is naturally unloaded and hands are off the wheel. The standard deviation of the amplitude of the non-touch sample of the central control capacitor. The first category is the standard deviation of the amplitude of the steering wheel capacitor release sample; the second category is the environmental auxiliary threshold, namely the temperature and humidity signal auxiliary threshold. Based on the normal driving and stationary conditions of the vehicle, a fixed normal fluctuation range is defined as temperature -20℃~60℃ and relative humidity 20%RH~80%RH. The interval comparison is performed on the time-aligned multi-channel synchronous signals frame by frame. Only when the three conditions are met simultaneously, such as the amplitude of the central control capacitor signal falling into the central control capacitor no-touch threshold range, the amplitude of the steering wheel capacitor signal falling into the steering wheel capacitor release judgment threshold range, and the real-time ambient temperature and relative humidity falling within the normal range of the temperature and humidity environmental auxiliary threshold, is the frame determined to be a valid reference frame. Then, multiple consecutive valid reference frames are collected according to the time axis to retain the complete time sequence segment. Isolated frames and jump frames that do not meet any of the conditions are removed, and then the time sequence signal that meets all conditions is obtained.

[0034] Finally, time-series signals that meet all conditions are subjected to time-series warping and standardization. Time-series warping and standardization are performed using a combination of time-series interpolation completion, redundant data removal, and amplitude normalization. Linear interpolation is used to complete missing time-series data points. Redundant data removal uses a sliding window deduplication method with a preset window size of 5 frames to remove duplicate data within the window. Amplitude normalization uses a min-max normalization method to map the signal amplitude to the [0,1] interval, eliminating dimensional differences and ensuring consistency in subsequent processing.

[0035] In detail, the specific implementation logic of this step is as follows: The original time-aligned signal set is read from the cache module and sent to the preprocessing unit. A first-order passive low-pass filtering algorithm is started to filter the central control capacitor signal, steering wheel capacitor signal, and temperature and humidity signal frame by frame to remove high-frequency noise. Then, abnormal data points that exceed the normal range are filtered by preset range thresholds to obtain purified multi-channel synchronization signals. Two preset judgment thresholds are retrieved, and the purified multi-channel synchronization signals are compared with the corresponding thresholds frame by frame to perform multi-dimensional joint discrimination. Only time-series signal segments that meet all three conditions are retained. The selected time-series signal segments are time-regularized and standardized to eliminate dimensional differences. Finally, they are integrated to form a time-continuous and feature-stable no-touch state signal set, which is temporarily stored in the feature cache unit to provide clean input for subsequent dynamic baseline construction. The total time for preprocessing and filtering does not exceed 80ms.

[0036] Furthermore, the process for obtaining the dynamic detection baseline is as follows: The ambient temperature and humidity signals corresponding to the no-touch state signal set are collected, and the temperature and humidity changes per unit time are calculated and fused to obtain the comprehensive temperature and humidity change. Based on the comprehensive temperature and humidity change, adaptive filtering calculations are performed on the central control channel signal and steering wheel channel signal in the no-touch state signal set to obtain the effective steady-state feature data after filtering. The effective steady-state feature data are fitted according to the time series and the central control dynamic detection baseline and steering wheel dynamic detection baseline are constructed simultaneously.

[0037] like Figure 3 As shown, the temperature and humidity changes are calculated by the difference between two consecutive frames of temperature and humidity data within a unit of time. The temperature change ΔT = T(n) - T(n-1), and the humidity change ΔH = H(n) - H(n-1), where T(n) and H(n) are the temperature and humidity data of the current frame, respectively, and T(n-1) and H(n-1) are the temperature and humidity data of the previous frame, respectively. The combined temperature and humidity changes are calculated using a weighted fusion method, and the fusion formula is as follows: ,in , Weighting coefficients (preset) =0.6, =0.4), which can be dynamically adjusted according to the characteristics of the vehicle environment, thereby ensuring that the comprehensive change can accurately represent the degree of environmental fluctuation.

[0038] Based on the calculated temperature and humidity changes, a recursive least squares adaptive filtering algorithm is used for filtering. The filtering parameters (iteration step size, convergence threshold) are dynamically adjusted by the comprehensive temperature and humidity changes. When the value is ≥0.5, it is determined that the environmental fluctuation is large. The iteration step size is set to 0.01, the convergence threshold is 1e-6, and the filtering smoothing weight is increased. When the value is less than 0.5, the environment is considered stable. The iteration step size is set to 0.05 and the convergence threshold is 1e-5. The filtering and smoothing weights are reduced to adapt to the dynamic changes in the vehicle environment.

[0039] Then, a linear fitting method is used to fit the effective steady-state feature data. The fitting formula is: y=a×t+b, where t is the time series, y is the baseline amplitude after fitting, and a and b are the fitting coefficients. The least squares method is used to solve the problem (minimize the sum of squares of the fitted value and the actual value). The number of iterations is preset to 50, and the convergence threshold is preset to 1e-5 to ensure that the time series of the fitted baseline is continuous and fits the steady-state features of the no-touch state signal.

[0040] In detail, the specific implementation logic of this step is as follows: Read the set of touchless state signals and corresponding synchronized temperature and humidity time-series data from the buffer, and perform interpolation calculations frame by frame at fixed time intervals (consistent with the sampling frequency, 100ms / frame) to calculate the temperature change per unit time. Humidity change The overall temperature and humidity change was calculated using a weighted fusion formula. This quantitatively characterizes the real-time fluctuations in the in-vehicle environment.

[0041] Combined temperature and humidity changes The weight adjustment coefficients for adaptive filtering of the central control channel and steering wheel channel are mapped respectively, dynamically adjusting the filtering iteration step size and convergence threshold of each of the two signals; the dynamic parameter mapping relationship of the recursive least squares filter is supplemented as follows: The intervals are divided with a increment of 0.1. Linear interpolation is used to map the iteration step size to the convergence threshold, specifically: Step size = 0.05 - 0.08 × Convergence threshold = 1e-5-8e-7× The physical meaning of the iteration step size is the filter update speed. The smaller the step size, the stronger the filter smoothness and the slower the response speed. The physical meaning of the convergence threshold is the filter convergence judgment criterion. The smaller the threshold, the higher the filter accuracy. For each timing signal of the central control and steering wheel in the non-touch state signal set, the recursive least squares adaptive filtering iteration calculation is performed point by point to filter out the slow drift and low-frequency interference deviation of the signal introduced by the ambient temperature and humidity fluctuations, and obtain the effective steady-state characteristic data after two filtering.

[0042] Two effective steady-state feature data are input into the linear fitting module according to the time series. The fitting coefficients a and b are solved by the least squares method to complete the linear fitting process, generating a time-continuous, environmentally adaptive dynamic detection baseline for the central control system and a dynamic detection baseline for the steering wheel. The selection criteria for effective steady-state feature data are to remove fluctuating data points in the filtered signal where the amplitude change between adjacent frames is greater than 0.005, and retain the steady-state data with smooth changes to ensure the stability of the steady-state features. The smoothness of the fitted baseline is checked. The baseline is considered qualified if the amplitude change between adjacent frames is ≤0.01. If it is not qualified, the fitting process is re-executed. Finally, the qualified dual-region dynamic detection baseline is stored in the baseline cache module to realize the synchronous dynamic update of the dual-region baseline with the vehicle environment. The total time for fitting and filtering does not exceed 100ms.

[0043] Furthermore, the process for obtaining the stabilized detection baseline is as follows: The dynamic detection baselines of the central control unit and the steering wheel are monitored in real time, and the change amplitude and rate of change of the two dynamic detection baselines are collected. The baseline stability judgment thresholds are set for the two dynamic detection baselines in different regions. The monitored change amplitude and rate of change are compared with the corresponding thresholds. When the change amplitude or rate of change of the dynamic baseline in any region exceeds the corresponding threshold, the baseline of that region is subjected to abnormal correction processing. After the baselines in each region are corrected to be qualified, the central control unit stabilization detection baseline and the steering wheel stabilization detection baseline are output synchronously.

[0044] The change amplitude is the absolute difference between the dynamic detection baseline amplitudes of two adjacent frames, calculated using the following formula: =|A(n)-A(n-1)|, where A(n) is the baseline amplitude of the current frame and A(n-1) is the baseline amplitude of the previous frame. This represents the absolute difference between the dynamic detection baseline amplitudes of two adjacent frames. The rate of change is the amount of change in the baseline amplitude per unit time, calculated using the following formula: ,in The sampling period is 100ms.

[0045] The baseline stability judgment thresholds set for the above-mentioned regions are as follows: the threshold for the amplitude of the dynamic detection baseline change of the central control system is preset to 0.02, and the threshold for the rate of change is preset to 0.2 / s; the threshold for the amplitude of the dynamic detection baseline change of the steering wheel is preset to 0.015, and the threshold for the rate of change is preset to 0.15 / s. During the judgment process, the two dimensions independently constrain the two baselines to adapt to the differences in signal characteristics between the two regions.

[0046] When any change magnitude or rate of change is detected to exceed the aforementioned threshold, anomaly correction processing is performed on the corresponding dynamic detection baseline. The anomaly correction processing employs a three-level logic: First, baseline freezing is performed, immediately pausing the iterative update of the dynamic baseline in the corresponding region to prevent the anomaly from continuing to spread; next, historical rollback is performed, using the stable baseline from the previous frame in that region as a temporary benchmark to replace the current abnormal baseline; finally, parameter reset is performed, resetting the iterative parameters of the adaptive filtering in that region and restarting the filtering calculation until the baseline in that region returns to the normal fluctuation range (change magnitude ≤ corresponding threshold, change rate ≤ corresponding threshold).

[0047] In detail, the specific implementation logic of this step is as follows: The timing data of the central control dynamic detection baseline and the steering wheel dynamic detection baseline are collected frame by frame, and the amplitude change of the baseline between adjacent frames is calculated in real time. rate of change per unit time ; Retrieve the preset baseline stability judgment threshold for each region, and then monitor the obtained baseline stability threshold. , Each value is compared with its corresponding threshold. If any value in any region exceeds the threshold range, it is determined that the dynamic detection baseline of the corresponding region has experienced abnormal drift or abrupt change.

[0048] For regions identified as abnormal, a three-level anomaly correction logic is immediately executed: the dynamic baseline iteration update process for that region is paused, and its own historical steady-state baseline (the qualified baseline of the previous frame) is invoked for baseline backtracking; the specific implementation of historical backtracking is as follows: if the baseline of the previous frame is in a stable state ( ≤threshold If the threshold is ≤, then the baseline of that frame is directly used as a temporary reference; if the baseline of the previous frame is also abnormal, then trace back sequentially, up to a maximum of 10 frames. If all 10 frames are abnormal, then the preset default baseline (obtained by fitting the initial 1000 frames of no-touch state signals) is used as a temporary reference; the parameter reset rule is: reset the iteration step size and convergence threshold of the adaptive filtering in this region to the default values ​​( Parameters when <0.5: step size 0.05, convergence threshold 1e-5), and at the same time, reset the filter iteration counter, re-filter and calculate the signals of the corresponding channels of the no-touch state signal set, and generate a temporary baseline.

[0049] Real-time monitoring of the magnitude and rate of change of the temporary baseline; if the stability condition is met for three consecutive frames ( ≤threshold If the error is less than or equal to the threshold, the correction is deemed successful, and the adaptive update of the dynamic baseline for that area is restored. If the error is not met, the correction logic is repeated until the error is successful. After the baselines of both the central control and steering wheel areas are successfully corrected, the central control stabilization detection baseline and the steering wheel stabilization detection baseline, which have stable amplitudes, gradual changes, and no abnormal jumps, are output synchronously. These baselines serve as the references for subsequent deviation calculations and noise reduction processing of the two signals. The total time for anomaly monitoring and correction does not exceed 50ms.

[0050] Furthermore, the process of obtaining the purified touch detection signal is as follows: Using the central control stability detection baseline and the steering wheel stability detection baseline as reference benchmarks, the signal deviation between the original capacitor signal and the reference benchmark in the corresponding area is calculated independently. Based on the signal deviation value of each area, the noise frequency band of the corresponding area is divided. Based on the divided noise frequency band, the capacitor signal of the non-touch state signal concentration is subjected to layered noise reduction processing to obtain the noise-reduced signal. The noise-reduced signal is then subjected to time smoothing and amplitude normalization processing to simultaneously obtain the central control purification touch detection signal and the steering wheel purification touch detection signal.

[0051] The signal deviation value is the difference in amplitude between the original capacitance signal of the corresponding region at the same time and the self-stabilized detection baseline. The calculation formula is as follows: S = |S(n) - B(n)|, where S(n) is the original amplitude of the capacitor signal in the current frame, and B(n) is the amplitude of the stabilized detection baseline in the current frame. Based on the signal deviation values ​​calculated over 100 consecutive frames, three noise frequency bands are defined. For example, in this embodiment, they are divided into low-frequency temperature drift noise (0.1-1Hz), mid-frequency driving vibration interference (1-5Hz), and high-frequency pulse glitches (greater than 5Hz), using a frequency band matching filtering strategy.

[0052] Next, layered noise reduction is performed on the capacitive signals concentrated in the non-touch state, using a frequency-band-specific filtering algorithm. Low-frequency temperature drift noise is filtered using a trend fitting method, which extracts the low-frequency trend component from the deviation value through linear fitting and removes it from the original signal; mid-frequency driving vibration interference is filtered using a moving average method, with a preset moving window size of 5 frames, which suppresses signal fluctuations caused by vibration by averaging the data within the window; high-frequency pulse spikes are filtered using a threshold limiting method, with a preset limiting threshold of 3 times the average deviation value, and pulse spikes exceeding the threshold are directly set to the threshold to remove high-frequency interference.

[0053] Timing smoothing employs a sliding weighted average method with a preset window size of 3 frames and weighting coefficients of [0.2, 0.6, 0.2] to ensure consistent signal timing. Amplitude normalization combines amplitude limiting and normalization. First, the signal amplitude is limited to a reasonable range (0-1), and then normalization is used to eliminate dimensional differences, preserving the true touch / hand-off sensing characteristics.

[0054] In detail, the specific implementation logic of this step is as follows: Using the time-aligned center console and steering wheel stabilization detection baselines as benchmarks, the amplitude difference ΔS between the original signal from the corresponding channel's capacitance sensor and the baseline is calculated point by point, resulting in continuous time-series signal deviation sequences. The amplitude and frequency distributions of the deviation sequences are statistically analyzed for each channel, and the frequency components of the deviation values ​​are analyzed using Fast Fourier Transform (FFT) to classify the noise frequency bands into low-frequency, mid-frequency, and high-frequency categories. The specific FFT analysis parameters are as follows: the window length is set to 1024 points, the number of sampling points is consistent with the window length, the sampling frequency is consistent with the original signal sampling frequency (10Hz), and the frequency resolution is 0.009766Hz. Zero-padding is used to ensure that the frequency coverage range of the FFT analysis is 0-5Hz (Nyquist frequency), ensuring accurate noise frequency band classification.

[0055] Based on the frequency band division results, layered and frequency band-specific noise reduction processing was performed on the original capacitor signals of the central control unit and steering wheel. The specific details are as follows: For low-frequency temperature drift noise, a trend fitting filtering method was used to extract and remove the low-frequency trend component. The trend fitting filtering method is implemented as follows: a sliding window linear fitting was used, with a preset window length of 20 frames (corresponding to 2 seconds) and a polynomial order of 1 (linear fitting). The trend line of the deviation value was fitted window by window, and the amplitude of the trend line was used as the low-frequency temperature drift component. This component was subtracted from the original signal to complete low-frequency noise reduction. The linear fitting formula for discrete points within the window is: ; In the formula: The sampling point number within the window. The slope is the fitted slope. For the fitting intercept, To obtain the low-frequency temperature drift trend component from the fitted signal, the fitted low-frequency trend component is removed from the original capacitance signal. The low-frequency correction signal after noise reduction is: ; For mid-frequency vibration interference, the moving average filtering method is used to suppress fluctuations. The formula is as follows: ; In the formula: For single-sided smooth frame count, To suppress the output signal after mid-frequency vibration interference and smoothly filter out mid-frequency fluctuation interference introduced by continuous vibration of the vehicle body; For high-frequency pulse glitches, a threshold amplitude limiting filter method is used to remove abnormal pulses, with a preset high-frequency pulse discrimination threshold. The discrimination rule is as follows: .

[0056] After performing layered and frequency-band noise reduction, the two denoised signals are subjected to time-series smoothing and amplitude normalization respectively: using a moving weighted average method. Obtaining time-series smoothing The process involves eliminating waveform distortion caused by filtering to ensure signal timing continuity; removing outliers beyond a reasonable range through amplitude limiting; mapping the signal amplitude to the [0,1] interval using the min-max normalization method to eliminate dimensional differences; and finally calculating the normalized amplitude for each sampling moment frame by frame. The specific formula is as follows: .

[0057] Calculate all single frames output sequentially at each of the consecutive sampling times. The signals are sequentially spliced ​​and continuously arranged according to the sampling time sequence to form a complete time-domain waveform sequence, generating a central control purification touch detection signal and a steering wheel purification touch detection signal with stable waveforms and effective suppression of high and low frequency interference noise.

[0058] Furthermore, the specific process for performing hierarchical comparison and verification is as follows: Feature extraction is performed on the central control air purification touch detection signal and the steering wheel air purification touch detection signal to obtain the feature parameters of each signal. The hierarchical comparison standard is set independently for the two signals by region. The extracted feature parameters are compared with their respective first-level thresholds to determine their validity and select valid candidate signals. The valid candidate signals are compared with their respective second-level thresholds to determine their stability and output the valid feature signals.

[0059] like Figure 4 As shown, the feature parameters are extracted in three core quantization features: steady-state amplitude (average amplitude after signal stabilization), rising edge rate of change (rate of amplitude change during the rising phase of the signal), and effective duration (time during which the signal remains within the effective range). Each type of feature parameter is quantized and extracted using time-domain analysis methods, with the extraction accuracy retained to 3 decimal places.

[0060] Hierarchical comparison specifically includes two levels of data comparison, the details of which are as follows: The first level is the signal validity threshold, which is used to distinguish between real sensing signals and random background noise interference. In this embodiment, the central control touch signal validity threshold is set as follows: steady-state amplitude ≥ 0.1, rising edge rate ≥ 0.05 / s, and effective duration ≥ 100ms. The steering wheel release signal validity threshold is set as follows: steady-state amplitude ≤ 0.08, rising edge rate ≤ 0.03 / s, and effective duration ≥ 200ms. Only signals that meet these requirements are valid candidate signals. The second level is the signal stability judgment threshold, which is used to distinguish weak interference jitter from valid signals. The stability threshold for the central control touch signal is set to steady-state amplitude fluctuation ≤ 0.02 and rising edge change rate fluctuation ≤ 0.01 / s. The stability threshold for the steering wheel release signal is set to steady-state amplitude fluctuation ≤ 0.015 and rising edge change rate fluctuation ≤ 0.008 / s. The final filtered signals have good stability.

[0061] In detail, the specific implementation logic of this step is as follows: Temporal feature parameters were quantized and extracted from the central control panel purification touch detection signal and the steering wheel purification touch detection signal. Using sliding window analysis (window size 5 frames), three core feature parameters were calculated for each signal: steady-state amplitude, rise time rate, and effective duration. The specific values ​​for each parameter were labeled. The rise time rate was calculated as follows: the interval where the signal amplitude rises from below half the hysteresis threshold to above the actual touch threshold was defined as the signal rise phase. The start time t1, end time t2, and corresponding amplitudes S1 and S2 of this interval were recorded. The rise time rate was calculated as (S2-S1) / (t2-t1). If the rise phase duration was less than 10ms, it was considered an invalid rise time, and the rise time rate was recorded as 0. The statistical logic for the effective duration was as follows: the signal amplitude must be continuously maintained within the corresponding threshold range, allowing for brief fluctuations of no more than 10ms (below / above the threshold), with the cumulative fluctuation duration not exceeding 10% of the total duration; otherwise, the duration was considered invalid.

[0062] The two-level comparison standard preset by region is retrieved. First, the feature parameters extracted by the two channels are compared with their own first-level effective admission threshold. Random invalid interference signals with too low amplitude, too weak change rate, or insufficient duration are eliminated, and effective candidate signals with touch sensing and off-hand sensing features are retained respectively.

[0063] The candidate signals from the central control and steering wheel, which are first-level admitted, are further compared with their own second-level stability thresholds to verify the stability and continuity of each channel signal, and to eliminate false touch and false release signals with short-term jitter and instantaneous jumps. The signals that pass the two-level comparison are then checked for integrity to ensure that the feature parameters are not missing and the fluctuations are within the allowable range. Finally, the identified valid touch feature signals of the central control and valid release feature signals of the steering wheel are output, providing an accurate basis for subsequent state determination. The total time for hierarchical comparison and verification does not exceed 50ms.

[0064] Furthermore, the process of obtaining the touch detection results is as follows: The effective feature signals are matched and compared with the preset state calibration rules to obtain a preliminary judgment result; the preliminary judgment result is compared with the core parameters of the effective feature signals in reverse, and the false judgment results are eliminated by combining the preset false judgment threshold to obtain the accurate state judgment result; the accurate state judgment result is integrated, structured encoding and format regularization are performed, and the touch detection result is output.

[0065] The preset state calibration rules are divided into regions as follows: The central control touch state calibration rule is based on the characteristic parameter range corresponding to different touch strengths and durations, distinguishing between three states: light touch (steady-state amplitude 0.1-0.3, effective duration 100-500ms), long press (steady-state amplitude ≥0.3, effective duration ≥500ms), and false trigger interference (characteristic parameters exceed the access threshold but do not meet the stability threshold). The steering wheel off-hand state calibration rule is based on the characteristic parameter ranges of the grip and idle states, distinguishing between three states: normal grip (steady-state amplitude ≥0.1, effective duration ≥200ms), completely off-hand (steady-state amplitude ≤0.08, effective duration ≥200ms), and slight false trigger (characteristic parameter fluctuations exceed the stability threshold).

[0066] The reverse comparison adopts the feature parameter backtracking verification method, which compares the feature parameter range corresponding to the preliminary judgment result with the actual core parameters of the effective feature signal, calculates the deviation value (the absolute difference between the actual parameter and the mean of the range), and sets the misjudgment judgment threshold to 0.02. If the deviation value is > 0.02, it is judged as a misjudgment result and is removed; if the deviation value is ≤ 0.02, the judgment result is confirmed to be valid and an accurate state judgment result is obtained.

[0067] The structured coding adopts JSON format, and the standardized detection results contain 6 types of core information: central control touch status (valid / invalid, touch type), steering wheel off-hand status (holding / releasing), trigger time, feature parameter value, judgment confidence, and anomaly mark; after format standardization, it is output according to the preset communication protocol to adapt to the interface requirements of the vehicle touch main controller and the vehicle safety control unit.

[0068] In detail, the specific implementation logic of this step is as follows: The system retrieves the effective feature signals from the central control unit and steering wheel from the hierarchical comparison and verification output, loads the preset dual-region state calibration rules, and clarifies the feature parameter interval boundaries corresponding to various states. The interval boundaries of the state calibration rules are based on the following: by collecting 1000 sets of experimental samples with different touch strengths (light touch, long press) and grip states (grip, release), the distribution intervals of feature parameters under various states are statistically analyzed, and a 95% confidence interval is taken as the boundary. For example, the boundary of the steady-state amplitude of light touch is determined by: statistically analyzing the minimum value of steady-state amplitude of 0.1 and the maximum value of 0.3 under the light touch state, ensuring that the amplitude of more than 95% of the light touch samples falls within this interval. The core parameters (steady-state amplitude, rising edge change rate, effective duration) of the two effective feature signals are matched one by one with the corresponding calibration rule intervals to complete the automatic calibration of touch attribution and release state, and obtain the preliminary judgment results of the central control touch type and steering wheel grip / release state.

[0069] The validity of the preliminary judgment results is verified by comparing the mean value of the feature parameter interval corresponding to the preliminary judgment results with the actual core parameters of the valid feature signals in reverse and calculating the deviation value. Combined with the preset misjudgment judgment threshold (0.02), the misjudgment results with a deviation value > 0.02 are eliminated, and the valid judgment results with a deviation value ≤ 0.02 are retained to obtain the accurate state judgment information of the dual region.

[0070] The calibrated dual-region status information is structured and encoded, and various core information is integrated using JSON format. The confidence level is then labeled (confidence level = 1 - deviation value / threshold, range 0-1). The calculation formula for confidence level is refined as follows: the threshold is uniformly adopted from the corresponding characteristic parameter threshold in the second-level stable threshold (e.g., steady-state amplitude threshold 0.02, rising edge rate threshold 0.01 / s). If multiple characteristic parameters exist, the average confidence level of each parameter is taken as the final confidence level. The encoded information is formatted, data fields are standardized, and data formats are normalized to ensure communication compatibility with the vehicle's main control unit. Finally, standardized touch detection results are uniformly output according to the preset communication protocol, achieving accurate anti-interference recognition of the vehicle's central control touch and reliable synchronous monitoring of the steering wheel hands-off state. The total time for result generation and formatting does not exceed 50ms.

[0071] The processes described above with reference to the flowcharts in the embodiments disclosed in this invention can be implemented as computer software programs. The embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), it performs the functions defined in the methods of this application. It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wire segments, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless segments, wire segments, optical fibers, RF, etc., or any suitable combination thereof.

[0072] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0073] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functions and structural principles of the present invention have been shown and explained in the embodiments. Without departing from the stated principles, the implementation of the present invention may have any variations or modifications.

Claims

1. A dynamic baseline update and anti-interference method for vehicle-mounted touch detection, characterized in that, The method includes: The original signal from the capacitive sensor and the ambient temperature and humidity signal are collected, and the original signal and the ambient temperature and humidity signal are processed to obtain a set of no-touch state signals. The temperature and humidity change is calculated based on the ambient temperature and humidity signal. The temperature and humidity change is combined with the temperature and humidity change to perform adaptive filtering calculation on the non-touch state signal set to obtain dynamic baseline data and form a dynamic detection baseline. Monitor the magnitude and rate of change of the dynamic detection baseline, and correct any anomalies in the dynamic detection baseline based on the monitoring results to obtain a stable detection baseline; The signal deviation is calculated based on the stabilized detection baseline and the original signal. The original signal is then subjected to noise suppression processing based on the signal deviation to obtain the purified touch detection signal. The purified touch detection signal is subjected to quantization and extraction of feature parameters. The quantized feature parameters are compared and verified with the hierarchical threshold. The touch state is determined based on the hierarchical comparison and verification results to obtain the touch detection results.

2. The dynamic baseline update and anti-interference method for vehicle-mounted touch detection according to claim 1, characterized in that, The process of acquiring the original signal from the capacitive sensor and the ambient temperature and humidity signals is as follows: The vehicle-mounted capacitive sensor starts signal acquisition at a preset sampling frequency to obtain the original signal from the capacitive sensor corresponding to the touch sensing. Synchronously control the temperature and humidity sensor in the vehicle's touch area to start collecting data and obtain real-time ambient temperature and humidity signals; The acquired raw signals from the capacitive sensor, ambient temperature signal, and ambient humidity signal are synchronously buffered to complete the signal acquisition process and form a raw signal set.

3. The dynamic baseline update and anti-interference method for vehicle-mounted touch detection according to claim 2, characterized in that, The process of acquiring the no-touch state signal set is as follows: The original signal set is preprocessed by using low-pass filtering to remove high-frequency invalid noise generated by external electromagnetic field coupling and threshold filtering to remove abnormal data, resulting in purified multi-channel synchronization signals. The purified multi-channel synchronization signals are compared with the preset corresponding thresholds to select the timing signals that meet all conditions. The selected timing signals are then subjected to timing normalization and standardization to obtain a set of touchless state signals.

4. The dynamic baseline update and anti-interference method for vehicle-mounted touch detection according to claim 3, characterized in that, The process of obtaining the dynamic detection baseline is as follows: Collect ambient temperature and humidity signals corresponding to the non-touch state signal set, calculate the temperature change and humidity change per unit time, and fuse them to obtain the comprehensive temperature and humidity change. By combining the comprehensive temperature and humidity changes, adaptive filtering calculations are performed on the central control channel signal and steering wheel channel signal, which are concentrated in the non-touch state, respectively, to obtain the effective steady-state characteristic data after filtering. The effective steady-state characteristic data are fitted according to the time series, and the dynamic detection baseline of the central control system and the dynamic detection baseline of the steering wheel are constructed simultaneously.

5. The dynamic baseline update and anti-interference method for vehicle-mounted touch detection according to claim 4, characterized in that, The process of obtaining the stabilized detection baseline is as follows: Real-time monitoring is performed on the dynamic detection baseline of the central control unit and the dynamic detection baseline of the steering wheel, and the change amplitude and rate of change of the two dynamic detection baselines are collected. The two dynamic detection baselines are divided into regions, and their respective baseline stability judgment thresholds are set. The monitored change amplitude and change rate are compared with the corresponding thresholds. When the change amplitude or change rate of the dynamic baseline in either region exceeds the corresponding threshold, anomaly correction processing is performed on the baseline of that region. Once the baselines of each region have been corrected to be qualified, the central control stability test baseline and the steering wheel stability test baseline are output synchronously.

6. The dynamic baseline update and anti-interference method for vehicle-mounted touch detection according to claim 5, characterized in that, The process of obtaining the purified touch detection signal is as follows: Using the central control stability detection baseline and the steering wheel stability detection baseline as reference benchmarks respectively, the signal deviation between the original capacitor signal and the reference benchmark in the corresponding area is calculated independently. Based on the signal deviation values ​​of each region, the noise frequency bands of the corresponding regions are divided. Based on the divided noise frequency bands, the capacitive signals of the non-touch state signal set are subjected to layered noise reduction processing to obtain the noise-reduced signal. The noise-reduced signal is processed by timing smoothing and amplitude normalization to simultaneously obtain the central control air purification touch detection signal and the steering wheel air purification touch detection signal.

7. The dynamic baseline update and anti-interference method for vehicle-mounted touch detection according to claim 6, characterized in that, The specific process for performing the hierarchical comparison and verification is as follows: Feature extraction was performed on the central control air purification touch detection signal and the steering wheel air purification touch detection signal to obtain the feature parameters of the two signals respectively; The two signals are divided into regions and the hierarchical comparison criteria are set independently. The extracted feature parameters are compared with their respective first-level thresholds to determine their validity and select valid candidate signals. The effective candidate signals are compared with their respective second-level thresholds for stability, and the effective feature signals are output.

8. The dynamic baseline update and anti-interference method for vehicle-mounted touch detection according to claim 7, characterized in that, The process of obtaining the touch detection result is as follows: The effective feature signals are matched and compared with the preset state calibration rules to obtain preliminary judgment results; By comparing the preliminary judgment result with the core parameters of the effective feature signal in reverse, and combining the preset misjudgment judgment threshold, the misjudgment result is eliminated to obtain the accurate state judgment result. The accurate status determination results are integrated, structured coding and format regularization are performed, and the touch detection results are output.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is executed by a processor to implement a dynamic baseline update and anti-interference method for vehicle-mounted touch detection as described in any one of claims 1-8.

Images