Passenger car intelligent touch panel control method with adaptive haptic feedback

Abstract

The present application relates to the field of automotive electronics and intelligent cockpit haptic interaction control technology, specifically to a passenger car intelligent touch panel control method with adaptive haptic feedback; in the method, a processor obtains background vibration time series data and touch operation data collected by an inertial measurement unit through a communication bus, performs short-time Fourier transform on the background vibration time series data to generate a background vibration spectrum, maps the touch operation data to a target perception intensity threshold, calculates a perception signal-to-noise ratio residual and determines an optimal injection frequency in combination with a frequency masking coefficient and a haptic sensitivity data table, generates a pulse width modulation control signal with a high-level driving period and a low-level gap period to drive a linear resonant motor, extracts an actual physical vibration frequency based on a back electromotive force collected during the low-level gap period, and performs closed-loop correction according to a frequency deviation value to achieve a low-power, low-strange, stable and perceivable vehicle-mounted haptic feedback.

Description

Technical Field

[0001] This invention relates to the field of automotive electronics and smart cockpit tactile interaction control technology, specifically a control method for a passenger vehicle smart touch panel with adaptive tactile feedback. Background Technology

[0002] As a human-machine interface component that integrates air conditioning adjustment, seat adjustment, sunshade control, and comfort function switching, the intelligent touch panel of passenger vehicles is increasingly important in terms of the application trend of continuous improvement in cabin intelligence and rear-seat interaction experience. Its tactile feedback effect is directly related to the accuracy of passenger operation confirmation and blind operation experience. Therefore, achieving stable and clear tactile feedback control in the dynamic vibration environment of the vehicle is an important foundation for meeting the reliability technical requirements of in-vehicle touch interaction. Existing tactile feedback solutions have many problems. For example, they mostly use open-loop output with fixed vibration frequency and fixed drive intensity, which makes it difficult to adapt to the background vibration changes transmitted to the panel from the chassis, suspension and body structure during vehicle operation. This can easily lead to situations where tactile feedback is masked, feedback intensity is unstable, or drive power is excessively increased to ensure perceptibility, resulting in panel noise, structural resonance and insufficient tactile differentiation of functional areas. In particular, when the actual operating frequency of the linear resonant motor is affected by assembly conditions, temperature rise, aging and finger pressure conditions, existing solutions cannot take into account the human tactile sensitivity characteristics, background vibration interference and the actual mechanical state of the actuator, and cannot effectively guarantee the continuity, accuracy and consistency of tactile feedback under complex in-vehicle conditions. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a control method for a smart touch panel in passenger vehicles with adaptive haptic feedback. Specifically, the technical solution of this invention includes: The method comprises a processor, a communication bus established with the processor, an inertial measurement unit rigidly connected to the back of the touch panel, the touch panel and a linear resonant motor, and a motor drive and sampling circuit connected between the processor and the linear resonant motor, the method including the following steps performed by the processor: The background vibration timing data transmitted from the passenger vehicle chassis to the touch panel and the touch operation data collected by the touch panel are acquired through the communication bus. Short-time Fourier transform is performed on the background vibration time series data to generate a background vibration spectrum, and the touch operation data is mapped to the target perception intensity threshold. Based on the pre-acquired masking effect mapping table, the vibration energy amplitude at each frequency point in the background vibration spectrum is extracted and multiplied by the pre-calibrated frequency masking coefficient extracted from the pre-set masking effect mapping table to obtain the minimum perceptible vibration amount at each frequency point. Obtain a pre-established tactile sensitivity data table, extract the human sensitivity benchmark value of the corresponding frequency point from the pre-established tactile sensitivity data table, calculate the sensing signal-to-noise ratio residual to characterize the relative sensing margin based on the benchmark value and the minimum perceptible vibration amount, and perform peak search processing in the residual to determine the optimal injection frequency; Based on the optimal injection frequency and the target sensing intensity threshold, a pulse width modulation control signal with a high-level driving period and a low-level gap period is generated and sent to the linear resonant motor to drive it to generate tactile feedback vibration. During the low-level gap, the back electromotive force of the linear resonant motor is acquired and the actual physical vibration frequency is extracted. This frequency is then compared with the optimal injection frequency to determine the frequency deviation value. Based on the frequency deviation value, the phase and duty cycle of the pulse width modulation control signal are corrected in a closed loop.

[0004] Optionally, the step of performing short-time Fourier transform processing on the background vibration time-series data to generate a background vibration spectrum includes: Analyze the background vibration time series data to determine the current environmental state label; Extract the basic spectral features that match the current environmental state label from a pre-constructed dataset containing multiple environmental state labels and their corresponding basic spectral features; The background vibration spectrum is generated by performing a short-time Fourier transform on the basic spectral features and the background vibration time series data.

[0005] Optionally, mapping the touch operation data to a target perception intensity threshold includes: Extract the rate of change of pressing area and pressure value from the touch operation data; The rate of change of the pressing area is differentially calculated, that is, the second-order rate of change of the pressing area within adjacent sampling time windows is obtained, and this is used as the contact dynamic feature generated to characterize the clarity of the operation intention; The target perception intensity threshold is generated by multiplying the pre-configured dynamic feature weights and pressure weights by the contact dynamic feature and the pressure value, respectively, and then summing them.

[0006] Optionally, the step of extracting the human sensitivity benchmark value at the corresponding frequency point from a preset tactile sensitivity data table, and calculating the sensing signal-to-noise ratio residual based on the benchmark value and the minimum perceptible vibration amount, includes: Extract the human body sensitivity baseline value at the corresponding frequency point from the tactile sensitivity data table; Calculate the difference between the human body sensitivity benchmark value and the minimum perceptible vibration amount, and generate the perceived signal-to-noise ratio residual at the corresponding frequency point.

[0007] Optionally, the step of performing closed-loop correction processing on the pulse width modulation control signal based on the frequency deviation value is configured as follows: If the frequency deviation value is greater than the preset frequency deviation threshold, the comprehensive period compensation amount is calculated based on the proportional-integral-derivative algorithm using the frequency deviation value. The comprehensive period compensation amount is then multiplied by the preset phase adjustment gain and duty cycle adjustment gain to compensate for the phase and duty cycle of the pulse width modulation control signal, thereby generating an updated pulse width modulation control signal. If the frequency deviation value is less than or equal to the frequency deviation threshold, the current pulse width modulation control signal is maintained.

[0008] Optionally, generating a pulse width modulation control signal with a high-level drive period and a low-level gap period based on the optimal injection frequency and the target sensing intensity threshold includes: The target function identifier is parsed from the touch operation data; Based on a pre-established mapping table containing the correspondence between functional identifiers and harmonic components, the corresponding harmonic components are extracted according to the target functional identifier. A basic waveform is generated using the optimal injection frequency as the fundamental frequency, and the harmonic components are superimposed on the basic waveform to generate the target synthetic waveform. Based on the target synthesized waveform and the target sensing intensity threshold, the pulse width modulation control signal with a high-level drive period and a low-level gap period is generated.

[0009] Optionally, the step of acquiring the back electromotive force of the linear resonant motor and extracting the actual physical vibration frequency during the low-level gap includes: During the low-level gap, the processor reads the back electromotive force generated by the linear resonant motor through the analog-to-digital converter in the motor drive and sampling circuit; The waveform of the back electromotive force is subjected to zero-crossing detection processing, and the time interval between adjacent zero-crossing points is extracted to calculate the actual physical vibration frequency.

[0010] Compared with the prior art, the present invention has the following beneficial effects: 1. This method acquires background vibration time-series data transmitted from the passenger vehicle chassis to the touch panel via an inertial measurement unit using a processor, and generates a background vibration spectrum by performing a short-time Fourier transform on the background vibration time-series data. This allows the actual vibration state of the touch panel to be used as the basis for tactile feedback control, thereby avoiding the fixed-frequency open-loop output being masked by background vibration under vehicle driving vibration changes. By combining vibration energy amplitude, frequency masking coefficient, and tactile sensitivity data table to calculate the perception signal-to-noise ratio residual and performing peak search to determine the optimal injection frequency, a more suitable feedback frequency can be selected between human sensitivity, background masking effect, and panel structural response. This improves the perceptibility of tactile feedback and reduces the risk of panel noise and structural resonance without simply increasing the driving power.

[0011] 2. This method extracts touch point position, pressure change, contact area change, and trigger duration from touch operation data, and further generates a target perception intensity threshold based on the rate of change of pressing area, contact dynamic characteristics, and pressure value mapping. It can distinguish different operation states such as light touch, heavy press, rapid click, continuous swipe, and passive pressure, so that the tactile feedback intensity is more consistent with the actual operation intention of the occupant. By limiting or reducing feedback output in cases of slight glancing, invalid trigger, or multi-point mixed contact, it can reduce false strong feedback caused by vehicle bumps and improve the accuracy and stability of in-vehicle touch interaction confirmation.

[0012] 3. This method generates a pulse width modulation control signal with a high-level drive period and a low-level gap period based on the optimal injection frequency and the target sensing intensity threshold. During the low-level gap period, the back electromotive force of the linear resonant motor is collected to extract the actual physical vibration frequency. This method can obtain frequency observations directly from the actual mechanical motion state of the motor, thus avoiding inferring the actuator state based solely on the target drive parameters. By comparing the actual physical vibration frequency with the optimal injection frequency, the frequency deviation value is determined. When the frequency deviation value exceeds the frequency deviation threshold, the phase and duty cycle of the pulse width modulation control signal are compensated using a proportional-integral-differential algorithm. This enables the linear resonant motor to maintain an output close to the desired frequency under varying assembly tolerances, temperature rise, aging, and finger pressure conditions, improving the consistency and continuity of tactile feedback.

[0013] 4. This method determines the current environmental state label by analyzing background vibration time-series data, and generates a background vibration spectrum map by combining the basic spectrum features that match the current environmental state label. This can reduce the spectrum jitter caused by the randomness of single-frame vibration data under complex working conditions such as high-speed smooth roads, rough roads, bridge joints, and speed bumps, thereby improving the robustness of optimal injection frequency selection. By parsing the target function identifier from the touch operation data, and superimposing the corresponding harmonic components onto the basic waveform according to the mapping relationship between the function identifier and the harmonic components to generate the target synthetic waveform, it can form differentiated tactile textures for different functions such as air conditioning, seats, and sunshades while maintaining adaptive frequency control, thereby improving the blind operation experience and tactile differentiation of functional areas of the intelligent touch panel in passenger vehicles. Attached Figure Description

[0014] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] like Figure 1 As shown, a control method for a smart touch panel in a passenger vehicle with adaptive haptic feedback is described. The system executing the method includes a processor, a communication bus established with the processor, an inertial measurement unit rigidly connected to the back of the touch panel, the touch panel and a linear resonant motor, and a motor drive and sampling circuit connected between the processor and the linear resonant motor. The method includes the following steps executed by the processor: The background vibration timing data transmitted from the passenger vehicle chassis to the touch panel and the touch operation data collected by the touch panel are acquired through the communication bus. Short-time Fourier transform is performed on the background vibration time series data to generate a background vibration spectrum, and the touch operation data is mapped to the target perception intensity threshold. Based on the pre-acquired masking effect mapping table, the vibration energy amplitude at each frequency point in the background vibration spectrum is extracted and multiplied by the pre-calibrated frequency masking coefficient extracted from the pre-set masking effect mapping table to obtain the minimum perceptible vibration amount at each frequency point. Obtain a pre-established tactile sensitivity data table, extract the human sensitivity benchmark value of the corresponding frequency point from the pre-established tactile sensitivity data table, calculate the sensing signal-to-noise ratio residual to characterize the relative sensing margin based on the benchmark value and the minimum perceptible vibration amount, and perform peak search processing in the residual to determine the optimal injection frequency; Based on the optimal injection frequency and the target sensing intensity threshold, a pulse width modulation control signal with a high-level driving period and a low-level gap period is generated and sent to the linear resonant motor to drive it to generate tactile feedback vibration. During the low-level gap, the back electromotive force of the linear resonant motor is acquired and the actual physical vibration frequency is extracted. This frequency is then compared with the optimal injection frequency to determine the frequency deviation value. Based on the frequency deviation value, the phase and duty cycle of the pulse width modulation control signal are corrected in a closed loop.

[0017] This embodiment provides a control mechanism for a smart touch panel in a passenger vehicle with adaptive tactile feedback. Specifically, this embodiment uses the smart control panel in the rear center armrest area as the main application scenario. The panel integrates functions such as air conditioning adjustment, seat posture adjustment, sunshade control, and massage mode switching. An inertial measurement unit is rigidly installed on the back of the panel, and a linear resonant motor is installed below the panel. The processor communicates with the above components via the in-vehicle communication bus. During vehicle operation, vibrations generated by the chassis, suspension, and body structure are transmitted to the touch panel along the armrest frame. Passengers' fingers are simultaneously subjected to the combined effects of background vibration and operation feedback vibration. Therefore, the system does not output a fixed vibration immediately after triggering. Instead, it first determines the frequency range of the current background vibration that has a masking effect on human tactile sensation, and then selects a frequency that is more easily perceived and less likely to excite panel structure noise as the feedback injection point. Specifically, the processor acquires background vibration time-series data from the inertial measurement unit. This time-series data reflects the actual mechanical disturbance experienced by the back of the panel, rather than abstract environmental parameters, and therefore can directly characterize the vibration state of the contact interface where the passenger's fingers are located. The processor performs a short-time Fourier transform on this time-series data to obtain a background vibration spectrum organized by time slices. The reason for using the short-time method is that the vehicle road conditions are not stable and unchanging. High-speed asphalt roads, speed bumps, bridge expansion joints, and rough paved roads can cause rapid shifts in the dominant vibration frequency. The system needs to sense the spectrum changes within a preset time window; otherwise, feedback lag is likely to occur. After obtaining the background vibration spectrum, the processor then reads the touch operation data collected by the touch panel. This touch operation data includes touch point position, pressure change, contact area change, and trigger duration. The processor does not directly map the touch operation data to a fixed driving intensity, but first generates a target perception intensity threshold. The mechanism is that different operations correspond to different subjective expectations: for example, a quick tap requires a clear confirmation, a continuous swipe requires a soft and continuous feel, and a deep press requires a more distinct tactile presence. Therefore, the target perception intensity threshold represents the degree of feedback that the passenger should perceive at this time, rather than the amount of current that the motor should output. Furthermore, the processor extracts the vibration energy amplitude of each frequency point from the background vibration spectrum and combines it with a pre-calibrated frequency masking coefficient to obtain the minimum perceptible vibration amount corresponding to each frequency point. Here, the frequency masking coefficient is used to describe the effect of background vibration on the human tactile threshold. The frequency masking coefficient is obtained from a pre-constructed masking effect mapping table, which uses the center frequency and energy amplitude of the background vibration as input indices and records the degradation ratio caused by the sensitivity threshold of adjacent target frequencies. Generally speaking, low-frequency large-displacement vibration of a vehicle will produce a masking effect greater than the preset masking threshold, while mid-to-high frequency vibration may cause additional noise or panel resonance in local structures. Therefore, the system does not simply extract the frequency point where the minimum value of the background vibration energy amplitude is located, but determines the frequency with the highest comprehensive evaluation index within the allowable range of human sensitivity, background masking effect, and panel structure. The processor extracts the human sensitivity baseline value corresponding to each frequency point from the tactile sensitivity data table. This data table can be pre-established based on passenger vehicle tactile interaction calibration experiments, reflecting the typical perception ability of the fingertip at different frequencies. The processor compares it with the aforementioned minimum perceptible vibration amount to obtain the perception signal-to-noise ratio residual. Specifically, the larger the perception signal-to-noise ratio residual, the more likely that the frequency point is to be clearly perceived by the passenger in the current in-vehicle vibration environment without requiring excessive drive power. The processor performs peak search on the residual across the entire frequency band to determine the optimal injection frequency. After determining the optimal injection frequency, the processor generates a pulse width modulation (PWM) control signal with a high-level drive period and a low-level gap period based on this frequency and the target sensing intensity threshold. The high-level drive period is used to inject energy into the linear resonant motor, and the low-level gap period is used to enable the linear resonant motor to output mechanical motion characteristics in a free decay state, so as to collect the back electromotive force (EMF) later. Since the actual mechanical vibration frequency of the linear resonant motor is affected by panel assembly tolerance, temperature rise, aging, passenger finger pressure, and local damping changes of the armrest, the target frequency and the actual frequency are not always consistent. Therefore, the processor collects the back EMF of the motor during the low-level gap period and extracts the actual physical vibration frequency accordingly. It then compares the actual frequency with the optimal injection frequency to obtain the frequency deviation value. If there is a deviation exceeding the preset deviation threshold, the processor performs closed-loop correction on the PWM control signal to make the motor output gradually approach the desired frequency. In abnormal situations, if the inertial measurement unit briefly goes offline, the communication bus experiences momentary congestion, or the background vibration spectrum shows a change in gradient exceeding a preset gradient threshold, the processor can temporarily maintain the optimal injection frequency and pulse width modulation parameters from the previous stable moment and limit the rate of increase of the drive amplitude to avoid abrupt tactile output due to data anomalies. If the touch operation data indicates that the passenger only made a touch with a contact area smaller than a preset area threshold and a duration less than a preset time threshold without forming a valid trigger, the system may not output tactile feedback or only output the lowest level prompt to avoid generating invalid tactile output. If the contact area is greater than or equal to the preset area threshold, or the duration is greater than or equal to the preset time threshold, it is determined to be a valid trigger, and the subsequent mapping steps of the target perception intensity threshold are continued. If the frequency deviation value continues to increase abnormally and cannot converge after multiple closed-loop corrections, it can be determined that the current motor operating point deviates from the resonance zone or the structural damping is abnormal. At this time, the system will fall back to the safe drive mode and only output basic feedback with limited amplitude to prevent panel noise and crosstalk between adjacent functional areas. When the vehicle is cruising at approximately 120 km / h on the highway, the background vibrations experienced by the rear armrest panel are mainly concentrated in frequency bands below the preset frequency threshold. When a passenger taps the area where the air conditioning temperature increases on the rear center armrest screen, the processor receives the touch data and determines that the operation is a clear short-duration tap, with the corresponding target perception intensity threshold exceeding the preset intensity benchmark value. At the same time, the inertial measurement unit shows that the current low-frequency vibration amplitude of the panel is greater than the preset interference threshold. If the traditional fixed vibration frequency is continued, the passenger may have difficulty perceiving it, or the amplitude must be significantly increased to be perceived. To this end, the system avoids the current area of ​​high vibration energy in the spectrum and selects a mid-to-high frequency injection point in the range where the human finger is still in a region where the sensitivity is higher than the preset sensitivity threshold and the panel structure response volatility is lower than the preset volatility threshold. A pulse width modulation control signal is generated to drive the linear resonant motor. During the drive interval, the system reads the back electromotive force and detects that the actual mechanical vibration frequency deviates from the target value due to the continuous pressure of the passenger's finger on the panel. Therefore, the phase and duty cycle are compensated in the drive cycle. The passenger finally obtains a clear and acknowledgment-based tactile feedback without obvious additional noise. Even if the background vibration of the vehicle is strong, there is no need to compensate by increasing the drive power beyond the preset ratio. The purpose of this step is to integrate the dynamic vibration environment inside the vehicle, the human body's tactile perception patterns, and the actual mechanical state of the actuator into the same control closed loop, so as to achieve a vehicle-mounted tactile feedback control effect with low power, low noise, perceptibility, and sustainable stable output. To further clarify, to ensure the uniqueness of the terminology throughout the control link, in the subsequent description of this embodiment, the optimal injection frequency refers to the target operating frequency obtained by searching the peak value of the perceived signal-to-noise ratio residual and used to drive the generation; if the target frequency appears in the text for the sake of narrative fluency, its meaning is the same as the optimal injection frequency and does not refer to any new frequency variable; correspondingly, the actual physical vibration frequency is used to represent the frequency extracted from the actual motion state of the linear resonant motor based on the back electromotive force; if the actual mechanical vibration frequency appears in a local paragraph, its meaning is consistent with the actual physical vibration frequency and is used to compare with the optimal injection frequency to form a frequency deviation value; To further clarify, the linear resonant motor mentioned in this article is sometimes simply referred to as a motor in the implementation details. Both refer to the same actuator object, not different hardware units. The pulse width modulation control signal in each implementation corresponds to the same type of drive signal with a high-level drive period and a low-level gap period. Its name or hardware affiliation is not changed because it carries the basic waveform, harmonic components, or closed-loop compensation. Through the above unified definition, the misunderstanding of multiple meanings or multiple words with the same meaning in different paragraphs can be avoided by the frequency name, actuator name, and drive signal name. In this embodiment, short-time Fourier transform processing is performed on the background vibration time series data to generate a background vibration spectrum, including: Analyze background vibration time series data to determine the current environmental state label; Extract basic spectral features that match the current environmental state label from a pre-constructed dataset containing various environmental state labels and their corresponding basic spectral features; By combining the basic spectral characteristics with the background vibration time series data, a short-time Fourier transform is performed to generate a background vibration spectrum.

[0018] This embodiment provides an enhanced generation mechanism for background vibration spectrum maps. Specifically, in the aforementioned basic scheme, if the spectrum map is directly generated based solely on the background vibration time-series data collected within a preset time period, the single-frame data may be random when the vehicle is in complex conditions such as rapid road surface changes, turning and merging, short-term crossing of bridge gaps, or transient changes in suspension attitude. This can easily lead to abrupt changes in spectrum recognition features. Therefore, this embodiment introduces an environmental state label before spectrum generation to combine real-time data with preset vehicle vibration prior features, thereby improving the stability and interpretability of the spectrum map. Specifically, the processor first performs state analysis on the background vibration time series data to determine the current environmental state label. This label can be used to characterize the typical operating scenarios of the vehicle, such as idling, low-speed urban roads, smooth high-speed roads, rough high-speed roads, passing speed bumps, and passing through continuous potholes. The environmental state label represents the classification of the current vibration source and propagation mechanism, rather than a single description of the driving mode. Taking a smooth high-speed road as an example, the background vibration mostly comes from tire rolling and vehicle aerodynamic excitation, and its spectral structure is relatively continuous with low frequencies dominating. When passing through speed bumps, the vibration is more likely to manifest as pulse impacts and multimodal short-term excitations. Pre-classifying these states helps to reduce misjudgments when processing transient data. After the label is determined, the processor extracts the basic spectral features that match it from the pre-built dataset. These basic spectral features can be understood as a typical profile of the vibration spectrum of the handrail area under the corresponding operating scenario. The processor uses this profile as a priori reference for real-time analysis and then participates in short-time Fourier transform processing together with the currently collected background vibration time series data. The significance of doing this is that the real-time data retains the authenticity of the scene, while the basic spectral features provide a structured background understanding. The background vibration spectrum map formed by the combination of the two can reflect the current instantaneous changes without completely deviating from the overall vibration law of the vehicle under this working condition due to isolated impacts. Furthermore, to avoid directly splicing the frequency domain prior with the original time domain signal in a way that results in inconsistent physical quantities, the preferred approach is to constrain or weight the short-time Fourier transform results, rather than directly superimposing the basic spectral features onto the background vibration time series data itself. Specifically, the processor first performs a short-time Fourier transform on the background vibration time-series data within the current window to obtain the real-time spectrum; then, using the basic spectral features as frequency band weights, frequency band confidence intervals, or spectral smoothing references, it enhances, suppresses, or smooths the real-time spectrum within the corresponding frequency range, thereby outputting the background vibration spectrum; if the basic spectral features are stored in the form of several main frequency bands, bandwidth ranges, and energy distribution ratios, the processor can prioritize maintaining the instantaneous peak positions in the real-time spectrum and only perform amplitude limiting smoothing on the scattered spikes in non-dominant frequency bands to ensure that the spectrum is still based on the real-time sampling results. Furthermore, when the environmental state label changes, the processor prefers not to immediately and completely replace the basic spectral features, but to perform a short-time transition fusion on the basic spectral features corresponding to the previous label and the current label, so as to reduce the spectral abrupt change at the state boundary. This will not destroy the true expression of the current vibration window by the short-time Fourier transform, and will also avoid the subsequent optimal injection frequency from jumping beyond the preset jump threshold between adjacent windows due to label jitter. In abnormal situations, if the processor cannot reliably determine the environmental state label, such as when the data segment is shorter than the preset analysis time threshold, the sensor is saturated by a momentary impact, or the overlap of different state features is higher than the preset overlap threshold, the label can be set to an unconfirmed state. In this case, only real-time background vibration time series data is used to generate a spectrum diagram, or the basic spectrum features corresponding to the most recently confirmed state are called for conservative fusion. If the environmental state cannot be stably identified in multiple consecutive time windows, the system can reduce the switching sensitivity of the optimal injection frequency to prioritize ensuring continuous and stable tactile feedback output rather than frequent jumps. In the aforementioned intelligent control panel application scenario, a vehicle carrying rear passengers is traveling on an urban expressway and enters an elevated section with obvious bridge joints. The inertial measurement unit first detects relatively stable continuous low-frequency vibrations, which the processor identifies as a high-speed, smooth road surface. When the vehicle enters the bridge joint area, periodic impact peaks appear in the time-series data, but the overall vehicle speed and vibration distribution are still close to high-speed conditions. At this time, the processor will not misjudge the system as an extremely bumpy road surface due to short-term impacts. Instead, it retrieves the basic spectral characteristics corresponding to the high-speed road surface as a reference and generates a background vibration spectrum map together with the real-time impact data. In this way, when passengers operate the sunshade control area, the system can still output stable and consistent tactile feedback without abnormal frequency jumps caused by one or two bridge joint impacts. The purpose of this step is to improve the robustness of the background vibration spectrum under complex in-vehicle conditions by introducing environmental state labels and basic spectral features, thereby achieving a more stable basis for spectral hole search. In this embodiment, mapping touch operation data to a target perception intensity threshold includes: Extract the rate of change of pressed area and pressure value from the touch operation data; The rate of change of the pressing area is calculated by differentiation, that is, the second-order rate of change of the pressing area within adjacent sampling time windows is obtained, and this is used as the contact dynamic feature generated to characterize the clarity of the operation intention. The target perception intensity threshold is generated by multiplying the pre-configured dynamic feature weights and pressure weights by the contact dynamic feature and pressure value, respectively, and then summing them.

[0019] This embodiment provides a refined mapping mechanism for the target perception intensity threshold. Specifically, in the aforementioned scheme, if the feedback strength is determined solely based on whether a touch event is triggered, then light touch, heavy press, slow contact, and rapid click will trigger similar tactile outputs, which weakens the effectiveness of tactile feedback as confirmation information in human-computer interaction. Especially in the moving environment inside a vehicle, the passenger's hand support posture is not fixed, and relying solely on static pressure values ​​can easily confuse passive pressure caused by vehicle bumps with active confirmation pressing by the passenger. Therefore, this embodiment introduces the rate of change of pressing area and its dynamic change characteristics to distinguish between the initiative and decisiveness of the contact process. Specifically, the processor extracts the rate of change of the pressing area and the pressure value from the touch operation data. The rate of change of the pressing area reflects the speed at which the fingertip spreads contact with the touch surface. Under normal circumstances, an active click will result in a rapid increase in the contact area in a short period of time, while involuntary contact or panel contact caused by vehicle bumps is more likely to show a slow, drifting change. The processor further analyzes the trend of the rate of change of the pressing area to generate contact dynamic features. These dynamic features characterize the transient rate of change in the contact establishment process and the clarity of the operation intention. The system then fuses the contact dynamic features and the pressure value according to preset weights to obtain the target perception intensity threshold. The physical significance of this fusion mechanism is that pressure values ​​can better reflect whether the contact has reached an effective triggering state, while the dynamic characteristics of the contact can better reflect whether the action has obvious human instruction attributes. After the two are combined, the system can effectively distinguish between high-dynamic light clicks and low-dynamic gravity pressure. Although the former has pressure less than the preset pressure judgment threshold, it requires clear confirmation of tactile sensation. Although the latter has pressure greater than the preset pressure judgment threshold, it may not necessarily trigger the same strong feedback, so as to avoid misleading strong vibrations when the vehicle is bumpy. Under an anomaly handling mechanism, if the touch panel detects multiple simultaneous contacts, or if there are mixed sliding and pressing actions at the same time, the processor can prioritize the determination based on the continuity and pressure stability of the main contact point, and appropriately reduce the weight of the contact dynamic features to avoid the target perception intensity threshold being artificially high due to multiple contacts; if there is a water film, gloves or contaminants on the sensor surface causing unstable contact area recognition, the system can increase the priority of the pressure value in the fusion process, or if necessary, fall back to a conservative medium intensity feedback level. In the same rear armrest scenario, a passenger is using the front passenger seat adjustment panel to adjust the seat back angle. During the first operation, the passenger briefly taps the reclining area at a speed greater than the preset speed threshold. The touch panel detects a rapid increase in contact area and a clear increase in pressure. Based on this, the system generates a higher target perception intensity threshold and outputs clearer confirmation feedback. During the second operation, the vehicle is passing through a continuous, uneven road surface. The passenger's palm edge is slightly pressed near the screen. Although the local pressure reading increases, the contact area changes gradually and lacks obvious dynamic changes. Therefore, the system does not provide the same level of strong feedback, thus avoiding misidentifying unintentional pressure as a function trigger. The purpose of this step is to upgrade the touch intent from a single static contact amount to a comprehensive recognition of the active operation, so as to achieve tactile intensity control that is more in line with the real interaction habits in the car. In this embodiment, the human body sensitivity benchmark value at the corresponding frequency point is extracted from a preset tactile sensitivity data table, and the perceived signal-to-noise ratio residual is calculated based on the benchmark value and the minimum perceptible vibration amount, including: Extract the baseline human sensitivity values ​​at corresponding frequency points from the tactile sensitivity data table; The difference between the human body sensitivity benchmark and the minimum perceptible vibration is calculated, and the perceptible signal-to-noise ratio residual at the corresponding frequency point is generated.

[0020] This embodiment provides a mechanism for constructing the signal-to-noise ratio residual. Specifically, in the aforementioned scheme, if the system only observes the intensity of the background vibration spectrum, it usually tends to select the frequency point with the lowest background energy for feedback output. However, the lowest background energy is not equivalent to the highest human perception. The sensitivity of fingers to vibrations at different frequencies varies significantly. Even if the background interference is small in some frequency bands, the human perception efficiency is low. While in some frequency bands, although the background vibration energy is not at its minimum, the vibration amount only needs to be slightly greater than the minimum perceptible vibration amount to produce a significant tactile response. Therefore, this embodiment introduces a human sensitivity benchmark value to unify the level of environmental quietness and the level of human perception into the same evaluation metric. Specifically, the processor extracts the human sensitivity benchmark value corresponding to each candidate frequency point from a preset tactile sensitivity data table. This data table can be derived from human tactile experiments calibrated under typical usage postures of in-vehicle touch panels, and is particularly suitable for the perception characteristics when the fingertips touch a rigid panel in the rear armrest operation scenario. The processor compares this benchmark value with the minimum perceptible vibration amount obtained under background vibration conditions to form a perception signal-to-noise ratio residual. Specifically, when the human sensitivity at a specific frequency point is high and the background masking effect is lower than the preset masking influence threshold, the perception signal-to-noise ratio residual is characterized by a larger relative margin, indicating that the system can obtain a more significant subjective tactile response with lower driving energy. Furthermore, to avoid confusion regarding the dimensionality of high sensitivity and low threshold, the human body sensitivity benchmark value is preferably represented by a threshold type with the same dimension as the minimum perceptible vibration amount or one-to-one conversion. In other words, the tactile sensitivity data table can record the benchmark perceptible vibration threshold of the human body at each frequency point, or the equivalent threshold value converted from the sensitivity index. In this way, the difference calculation corresponds to the relative margin between the minimum perceptible vibration amount required to be achieved under the current background masking and the benchmark perceptible level of the human body under ideal reference conditions, and the two have a consistent comparison direction. Furthermore, in the specific implementation, the more favorable the difference result, the more suitable the frequency point is as a feedback injection point in the current environment; if the benchmark value used is defined as the benchmark perceptible threshold, the processor can internally assign a higher priority to the smaller threshold before performing difference evaluation; if the benchmark value used is defined as the equivalent sensitivity score, it can first be converted into an equivalent threshold according to a predetermined conversion rule before performing the difference calculation with the minimum perceptible vibration; through the above unification, the difference generation method in the embodiment remains unchanged, while the physical meaning of the residual is clearly defined as a perceptible margin index that can be used for horizontal comparison at the same frequency point; The key to this residual construction method lies in avoiding simplifying the spectrum selection problem to a simple structural vibration problem. The core purpose of in-vehicle haptic feedback is not simply to drive the panel to produce mechanical displacement, but to adapt the vibration output to the human tactile perception model to provide clear interactive feedback. Therefore, the perceived signal-to-noise ratio residual actually constitutes a cross-domain evaluation quantity: one end connects the physical vibration environment of the vehicle and the panel, and the other end connects to the physiological perception law of the human finger. Only by incorporating both of these parts can the selection of the optimal injection frequency have practical interactive significance. Under an anomaly handling mechanism, if the calibration data for certain frequency points in the tactile sensitivity data table is missing, the processor can use the smooth interpolation results of neighboring frequency points, or directly mark the frequency point as low priority to avoid rashly selecting it when the information is insufficient; if the background vibration amplitude is greater than the preset environmental interference threshold, causing the minimum perceptible vibration amount of most frequency points to generally increase, the processor can prioritize the frequency point with the relatively optimal residual, and simultaneously reduce the crispness requirement of the target tactile sensation, instead outputting a more stable confirmation vibration to ensure the consistency of the interaction; In the scenario described above where the vehicle enters a rough asphalt road surface, multiple low-frequency points in the panel's background spectrum exhibit high vibration energy, while a few higher-frequency points have low background vibration energy. If the system only considers the background energy, it might select a frequency with vibration energy below a preset noise threshold but not sensitive to human fingers, resulting in the motor working but not being noticeably perceived by the passenger. This embodiment, by consulting the human sensitivity benchmark, determines that another candidate frequency point, although having slightly higher background energy, falls within the range where fingers are more sensitive, ultimately resulting in a better residual signal-to-noise ratio. Therefore, the system selects this frequency point to inject tactile feedback, allowing the passenger to clearly confirm that the operation has taken effect when adjusting the massage mode. The purpose of this step is to establish a frequency evaluation standard that takes into account both background masking and human sensitivity, so as to achieve the optimal frequency selection that is closer to the real subjective touch. To further clarify, in order to ensure the uniqueness of terminology, the human sensitivity benchmark value in this embodiment refers to the reference value that corresponds one-to-one with each candidate frequency point in the tactile sensitivity data table. Before entering the difference calculation, the reference value is organized into an equivalent threshold characterization that can be directly compared with the minimum perceptible vibration amount. That is, if the following paragraphs describe it as reflecting the level of human sensitivity in a subjective sense, it is an explanation of the source and significance of the reference value, and does not mean that it is separated from the threshold form and used as a scoring variable of another dimension in the calculation link. To further clarify, throughout the text, the perceived signal-to-noise ratio residual refers to the comparison result of the same type of frequency points, that is, the result quantity generated by the difference between the human sensitivity benchmark value of the corresponding frequency point and the minimum perceptible vibration amount of the corresponding frequency point; if for the sake of explanation, there are expressions such as "more favorable result is better than margin" or "better residual", these are all textual explanations of the engineering significance of the difference result being applicable to frequency optimization, and do not introduce new evaluation parameters or new residual definitions; In this embodiment, the pulse width modulation control signal is subjected to closed-loop correction processing based on the frequency deviation value, configured as follows: If the frequency deviation value is greater than the preset frequency deviation threshold, the comprehensive period compensation amount is calculated based on the proportional-integral-differential algorithm using the frequency deviation value. The comprehensive period compensation amount is then multiplied by the preset phase adjustment gain and duty cycle adjustment gain to compensate for the phase and duty cycle of the pulse width modulation control signal, thereby generating an updated pulse width modulation control signal. If the frequency deviation value is less than or equal to the frequency deviation threshold, the current pulse width modulation control signal is maintained.

[0021] This embodiment provides a closed-loop correction mechanism for pulse width modulation control signals. Specifically, in the aforementioned scheme, although the processor has selected the optimal injection frequency based on background vibration and human perception model, whether the linear resonant motor can ultimately operate stably near that frequency is also affected by the state of the mechanical system itself. The different positions of passengers' fingers on the panel, the different pressure applied, the differences in the assembly of the internal fasteners of the armrest, changes in ambient temperature, and the aging of components will all change the equivalent damping and equivalent mass of the system, thereby causing the actual mechanical vibration frequency to deviate from the target injection frequency. Without closed-loop correction, tactile feedback will exhibit response lag, energy dissipation, or even induce structural abnormal noise as the usage state changes. Specifically, the processor compares the actual physical vibration frequency with the target frequency to obtain the frequency deviation value. When the deviation exceeds a preset threshold, it indicates that the current drive parameters can no longer effectively lock the optimal operating point of the motor and panel assembly. At this time, the processor compensates for the phase and duty cycle of the pulse width modulation control signal using a proportional-integral-derivative method. Phase compensation is mainly used to adjust the timing of energy injection, making the drive more in line with the natural motion state of mechanical vibration. Duty cycle compensation is used to fine-tune the magnitude and duration of injected energy in order to restore tactile clarity without significantly increasing noise and structural crosstalk. When the frequency deviation value does not exceed the threshold, the processor keeps the current drive signal unchanged to avoid over-adjustment that could cause new fluctuations. The engineering significance of this threshold-based closed-loop strategy is that the in-vehicle haptic system does not need to perform high-frequency compensation for minor disturbances. Minor vibration disturbances already exist inside the vehicle. If the closed loop is too sensitive, it can easily cause the haptic output to oscillate at high frequencies, manifested as frequency shifts and sudden changes in subjective tactile intensity. By setting a frequency offset threshold, the closed-loop compensation can be applied to systematic deviations that exceed the preset frequency offset threshold, while minor high-frequency disturbances within the tolerance range are not adjusted. Under an abnormal handling mechanism, if the frequency deviation value exceeds the limit for several consecutive cycles but cannot converge after compensation, the processor can determine that the current operating condition has exceeded the normal resonance control range. For example, the panel is pressed against by a heavy object for a long time, the motor is loosely installed, or the structural components are abnormally rubbing. At this time, the system can enter the restricted output mode, prioritizing the absence of abnormal noise and accidental touch, while temporarily reducing the clarity of touch. If the noise amplitude of the back EMF sampling signal itself is greater than the preset noise tolerance threshold, causing the frequency deviation judgment to be unstable, the processor can temporarily freeze the phase compensation, only retaining the duty cycle adjustment with the adjustment rate lower than the preset rate threshold, and re-enable the complete closed loop after the sampling returns to normal. In the aforementioned high-speed driving scenario, when rear passengers adjust the air conditioning fan speed, their fingers continuously press on the armrest screen, thus altering the local mechanical boundary conditions of the panel. During the low-level interval, the system detects that the actual physical vibration frequency begins to deviate from the originally set injection frequency, and the deviation exceeds the allowable range. The processor then fine-tunes the drive phase of the next cycle and appropriately changes the duty cycle to rematch the motor's energy input with the current mechanical state. After several corrections, the actual vibration stabilizes again, and the click sensation perceived by the passenger remains clear without any residual vibration. If the deviation cannot converge during this process, the system automatically lowers the output limit to prevent acoustic noise caused by structural resonance from appearing on the panel. The purpose of this step is to transform the tactile feedback from an open-loop set value to an adaptive output oriented towards the actual mechanical state, thereby achieving a consistent tactile feel under different assembly states and different passenger contact states. To further explain, in this embodiment, the target injection frequency, the originally set injection frequency, and the target frequency are all synonymous descriptions of the optimal injection frequency in the embodiment, all referring to the same frequency object used as the driving reference before the closed-loop correction; when the processor performs the closed-loop correction, it does not redefine a new independent target quantity, but adjusts the phase and duty cycle of the pulse width modulation control signal around the optimal injection frequency, so that the actual physical vibration frequency of the linear resonant motor converges to it. To further explain, maintaining the current pulse width modulation control signal means that when the frequency deviation value is less than or equal to the frequency deviation threshold, the phase parameters and duty cycle parameters that are already in effect in the current cycle are maintained without triggering new PID compensation calculations. If the system also has basic waveform parameters, harmonic parameters, or drive envelope parameters that are predetermined based on the target perception intensity threshold or target function identifier, these parameters will still be maintained normally according to their respective upstream logic, and their names or control levels will not be changed due to the closed-loop correction determination in this embodiment. Through the above limitations, the object boundaries between the closed-loop correction link and the aforementioned frequency selection link and waveform synthesis link can be kept consistent, avoiding the situation where the same driving quantity is repeatedly named or mixed in different segments. In this embodiment, based on the optimal injection frequency and the target sensing intensity threshold, a pulse width modulation control signal with a high-level drive period and a low-level gap period is generated, including: Extract the target function identifier from the touch operation data; Based on a pre-established mapping table containing the correspondence between functional identifiers and harmonic components, the corresponding harmonic components are extracted according to the target functional identifier. A basic waveform is generated using the optimal injection frequency as the fundamental frequency, and harmonic components are superimposed on the basic waveform to generate the target synthetic waveform. Based on the target synthesized waveform and the target sensing intensity threshold, a pulse width modulation control signal with a high-level drive period and a low-level gap period is generated.

[0022] This embodiment provides a pulse width modulation signal generation mechanism with functional differentiation capability. Specifically, in the aforementioned scheme, the system can adaptively select a better injection frequency based on the background vibration environment and adjust the feedback perception intensity according to the strength of the operation. However, if all functional areas rely on only the same tactile style, even if the passenger can feel that it has been triggered, it is still difficult to distinguish the specific function category currently triggered without looking at the screen. For the rear control panel scenario, blind operation capability is an important part of in-vehicle interaction design. Therefore, this embodiment further introduces target function identification and harmonic component mapping to form tactile codes with differentiated textures within the same optimal injection frequency framework. Specifically, the processor first parses the target function identifier from the touch operation data. This identifier is used to characterize the function category of the current trigger area, such as air conditioning temperature adjustment, seat massage mode, seat forward and backward sliding, sunshade opening and closing, etc. The processor extracts the harmonic components corresponding to the target function based on a pre-established mapping table of function identifiers and harmonic components. The harmonic components here are not for pursuing complex waveforms themselves, but to change the spectral composition of the tactile waveform. Near the same fundamental frequency, different harmonic combinations will make the tactile sensation present subjective differences such as high sharpness, low frequency high amplitude, long oscillation time or low delay, thereby carrying additional information. The processor generates a basic waveform based on the optimal injection frequency, and then superimposes the corresponding harmonic components to obtain the target synthetic waveform. Subsequently, the processor combines the target perception intensity threshold to convert the synthetic waveform into a pulse width modulation control signal with a high-level drive period and a low-level gap period. On the one hand, the basic waveform ensures that the feedback falls within the frequency range that is easily perceived in the current environment; on the other hand, the harmonic components give different functions distinguishable tactile characteristics. In this way, the system does not deviate from the adaptive frequency selection logic, and can also endow the tactile feedback function with semantics. In abnormal situations, if the harmonic component corresponding to a certain functional area is likely to resonate with the panel structure under the current background vibration environment, the processor can reduce or temporarily shield the harmonic component, retaining only the basic waveform and a small number of low-risk harmonics to prevent local abnormal noise; if the target function identifier resolution fails, for example, if the touch boundary area covers two functional areas at the same time or the interface is being refreshed, the system can call the default haptic template, prioritize ensuring confirmation feedback, and restore functional coding after the interface state stabilizes. On the rear armrest screen of the same vehicle, the passenger tapped the area for increasing the air conditioning temperature and the area for switching the seat massage mode. Both operations occurred when the vehicle was traveling at medium to high speeds, with strong low-frequency vibrations in the background. The system selected similar optimal injection frequencies for both operations to ensure perceptibility. However, during the waveform synthesis phase, the air conditioning function corresponded to a combination of harmonics with a smooth waveform envelope and a single high-frequency component, presenting a short, confirming touch. The massage mode corresponded to a combination of harmonics with a dominant low-frequency component and a wide frequency spectrum, presenting a gentle touch with a long start-up period. Even without looking down at the screen, the passenger could perceive the two different feedback textures with their fingers, thus helping to determine the function category. The purpose of this step is to add functional differentiation capability on the basis of adaptive frequency control, so as to achieve the blind operation tactile coding effect of the vehicle touch panel in complex driving environment; In this embodiment, the back electromotive force of the linear resonant motor is acquired and the actual physical vibration frequency is extracted during the low-level gap period, including: During the low-level gap, the processor reads the back electromotive force generated by the linear resonant motor through the analog-to-digital converter in the motor drive and sampling circuit. The waveform of the back electromotive force is subjected to zero-crossing detection processing, and the time interval between adjacent zero-crossing points is extracted to calculate the actual physical vibration frequency.

[0023] This embodiment provides a mechanism for extracting the actual physical vibration frequency based on back electromotive force. Specifically, in the aforementioned scheme, closed-loop correction relies on the perception of the actual mechanical state of the linear resonant motor. If the vibration state is inferred solely based on the target drive parameters output to the linear resonant motor, the system only obtains the expected drive characteristics and cannot characterize the actual physical mechanical vibration state. Considering the installation structure of the vehicle touch panel and the constant changes in finger contact state, it is necessary to introduce a detection quantity that can directly reflect the motor's motion state. This embodiment uses the method of reading the back electromotive force during the low-level gap to obtain this information. Specifically, when the pulse width modulation control signal enters the low-level gap period, the drive circuit no longer actively injects energy into the linear resonant motor. The motor's mover continues to move under the action of inertia and elastic restoring force. At this time, a back electromotive force related to mechanical motion is generated in the coil. Since this signal originates from the actual physical motion process of the actuator, it can reflect whether the current vibration frequency has deviated from the target operating point. The processor reads the back electromotive force waveform during the low-level gap period and performs zero-crossing detection on it. Zero-crossing detection is to identify the moment when the waveform crosses the reference baseline from positive to negative or from negative to positive. The time interval between adjacent zero-crossing points corresponds to the actual vibration period, so the actual physical vibration frequency can be extracted accordingly. Furthermore, to avoid ambiguity in period conversion, adjacent zero-crossing points preferably refer to adjacent alternating positive and negative zero-crossing events in the waveform; under this definition, the time interval between adjacent zero-crossing points corresponds to half a mechanical vibration cycle, and the processor converts it to a complete cycle when calculating the actual physical vibration frequency; if the implementation uses zero-crossing points in the same direction as the criterion, such as two consecutive zero-crossing moments from negative to positive, then the time interval can directly correspond to one mechanical vibration cycle; both of the above implementations belong to the method of extracting the actual physical vibration frequency based on zero-crossing detection, only the period conversion rules are different; The reason for using zero-crossing detection instead of complex waveform reconstruction is that the vehicle control environment emphasizes real-time performance and anti-interference capabilities; the zero-crossing moment is relatively easy to extract from the noisy background and has less dependence on the absolute amplitude, making it suitable for online frequency tracking of small actuators such as linear resonant motors; especially in the rear armrest screen scenario, the processor often has to undertake tasks such as touch parsing, interface refresh coordination and bus communication management, so a simple and robust frequency extraction method is more conducive to vehicle-level implementation. Furthermore, to suppress false zero crossings caused by local distortion of the back EMF waveform, the processor can perform amplitude limiting, bandpass filtering, or moving average processing on the sampled waveform before zero crossing detection, and add a minimum interval threshold and polarity continuity check after zero crossing detection; only when the candidate zero crossing point meets the expected time interval range and has basic consistency with the preceding and following periods is it used for frequency calculation; this does not change the main zero crossing detection process in the embodiment, and can improve the reliability of frequency estimation under vehicle impact and electromagnetic interference environments. Under an anomaly handling mechanism, if the low-level gap is too short, resulting in insufficient back EMF sampling window, the processor can extend the gap in one or more subsequent drive cycles to prioritize frequency extraction quality. If the back EMF waveform exhibits false zero-crossing points due to electromagnetic noise, power supply ripple, or structural impact interference, the processor can filter out unreliable zero-crossing events by combining minimum time interval limits and adjacent cycle consistency checks. If back EMF characteristics cannot be reliably obtained within several consecutive cycles, the system can temporarily stop closed-loop frequency correction and retain only conservative fixed drive to avoid larger deviations caused by erroneous feedback. During the operation of the sunshade control area by the rear passengers, the vehicle happened to pass through a section of bridge deck with dense joints, and the panel was subjected to a short-term impact. The back EMF waveform read by the processor during a low-level interval showed local distortion, and the first zero-crossing recognition was unstable. The system did not immediately adjust the drive based on this single abnormal result, but instead combined the back EMF waveforms of the subsequent two cycles for consistency verification. After confirming the true zero-crossing interval, a stable actual physical vibration frequency was obtained again. Thereafter, the closed-loop correction continued to take effect, keeping the tactile feedback of the sunshade button continuous and preventing significant drift due to the instantaneous impact. The purpose of this step is to provide frequency observations directly derived from the mechanical motion of the motor for closed-loop control, thereby enabling real-time tracking of the actuator's operating state and providing a reliable basis for subsequent drive corrections.

[0024] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A control method for a smart touch panel in a passenger vehicle with adaptive haptic feedback, characterized in that, The system executing the method includes a processor, a communication bus established with the processor, an inertial measurement unit rigidly connected to the back of a touch panel, a touch panel and a linear resonant motor, and a motor drive and sampling circuit connected between the processor and the linear resonant motor. The method includes the following steps executed by the processor: The background vibration timing data transmitted from the passenger vehicle chassis to the touch panel and the touch operation data collected by the touch panel are acquired through the communication bus. Short-time Fourier transform is performed on the background vibration time series data to generate a background vibration spectrum, and the touch operation data is mapped to the target perception intensity threshold. Based on the pre-acquired masking effect mapping table, the vibration energy amplitude at each frequency point in the background vibration spectrum is extracted and multiplied by the pre-calibrated frequency masking coefficient extracted from the pre-set masking effect mapping table to obtain the minimum perceptible vibration amount at each frequency point. Obtain a pre-established tactile sensitivity data table, extract the human sensitivity benchmark value of the corresponding frequency point from the pre-established tactile sensitivity data table, calculate the sensing signal-to-noise ratio residual to characterize the relative margin of perception based on the benchmark value and the minimum perceptible vibration amount, and perform peak search processing in the residual to determine the optimal injection frequency; Based on the optimal injection frequency and the target sensing intensity threshold, a pulse width modulation control signal with a high-level driving period and a low-level gap period is generated and sent to the linear resonant motor to drive it to generate tactile feedback vibration. During the low-level gap, the back electromotive force of the linear resonant motor is collected and the actual physical vibration frequency is extracted. The frequency deviation value is determined by comparing it with the optimal injection frequency. Based on the frequency deviation value, the phase and duty cycle of the pulse width modulation control signal are corrected in a closed loop. The step of generating a background vibration spectrum by performing short-time Fourier transform processing on the background vibration time-series data includes: Analyze the background vibration time series data to determine the current environmental state label; Extract the basic spectral features that match the current environmental state label from a pre-constructed dataset containing multiple environmental state labels and their corresponding basic spectral features; The background vibration spectrum is generated by performing a short-time Fourier transform on the basic spectral features and the background vibration time series data. The step of mapping touch operation data to a target perception intensity threshold includes: Extract the rate of change of pressing area and pressure value from the touch operation data; The rate of change of the pressing area is differentially calculated, that is, the second-order rate of change of the pressing area within adjacent sampling time windows is obtained, and this is used as the contact dynamic feature generated to characterize the clarity of the operation intention; The target perception intensity threshold is generated by multiplying the pre-configured dynamic feature weights and pressure weights by the contact dynamic feature and the pressure value, respectively, and then summing them. The step of generating a pulse width modulation control signal with a high-level drive period and a low-level gap period based on the optimal injection frequency and the target sensing intensity threshold includes: The target function identifier is parsed from the touch operation data; Based on a pre-established mapping table containing the correspondence between functional identifiers and harmonic components, the corresponding harmonic components are extracted according to the target functional identifier. A basic waveform is generated using the optimal injection frequency as the fundamental frequency, and the harmonic components are superimposed on the basic waveform to generate the target synthetic waveform. Based on the target synthesized waveform and the target sensing intensity threshold, the pulse width modulation control signal with a high-level drive period and a low-level gap period is generated.

2. The control method for a smart touch panel of a passenger vehicle with adaptive haptic feedback according to claim 1, characterized in that, The step of extracting the human body sensitivity benchmark value at the corresponding frequency point from the preset tactile sensitivity data table, and calculating the sensing signal-to-noise ratio residual based on the benchmark value and the minimum perceptible vibration amount, includes: Extract the human body sensitivity baseline value at the corresponding frequency point from the tactile sensitivity data table; Calculate the difference between the human body sensitivity benchmark value and the minimum perceptible vibration amount, and generate the perceived signal-to-noise ratio residual at the corresponding frequency point.

3. The control method for a smart touch panel of a passenger vehicle with adaptive haptic feedback according to claim 1, characterized in that, The closed-loop correction processing of the pulse width modulation control signal based on the frequency deviation value is configured as follows: If the frequency deviation value is greater than the preset frequency deviation threshold, the comprehensive period compensation amount is calculated based on the proportional-integral-differential algorithm using the frequency deviation value. The comprehensive period compensation amount is then multiplied by the preset phase adjustment gain and duty cycle adjustment gain to compensate for the phase and duty cycle of the pulse width modulation control signal, thereby generating an updated pulse width modulation control signal. If the frequency deviation value is less than or equal to the frequency deviation threshold, the current pulse width modulation control signal is maintained.

4. The control method for a smart touch panel of a passenger vehicle with adaptive haptic feedback according to claim 1, characterized in that, The process of acquiring the back electromotive force of the linear resonant motor and extracting the actual physical vibration frequency during the low-level gap includes: During the low-level gap, the processor reads the back electromotive force generated by the linear resonant motor through the analog-to-digital converter in the motor drive and sampling circuit; The waveform of the back electromotive force is subjected to zero-crossing detection processing, and the time interval between adjacent zero-crossing points is extracted to calculate the actual physical vibration frequency.

Images