Automobile door handle sensing system of intelligent automobile entering device
By introducing a miniature infrared thermal imager and acoustic wave detection at the door handle, and combining temperature gradient and acoustic wave time difference, multimodal feature fusion was achieved, which solved the problem of mis-locking in complex environments of existing systems and improved the reliability and response speed of door unlocking.
Patent Information
- Application Number
- CN202510823662.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-19
- Publication Date
- 2025-10-17
AI Technical Summary
Existing intelligent entry systems rely on sensing methods at the door handles that are susceptible to environmental complexity, leading to frequent false locking incidents. They also fail to effectively distinguish between temperature gradient changes caused by human contact and changes in ambient temperature, and the multiple reflections and attenuation mechanisms of sound waves within the narrow grip groove are not taken into account.
A miniature infrared thermal imager is used to acquire a temperature matrix, calculate the temperature gradient characteristic value and thermal attenuation coefficient, and combine it with the sound wave time difference and sound attenuation coefficient. The comprehensive confidence level is calculated through multimodal feature fusion to determine the credibility of the holding behavior.
Maintaining stable signal output under extreme environments significantly reduces the false unlock rate, improves the reliability and robustness of the system in various climate and contamination scenarios, and shortens the unlock response time.
Smart Images
Figure FT_1 
Figure SMS_5 
Figure SMS_24
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of vehicle door handle sensing, more particularly, it relates to a vehicle door handle sensing system of an intelligent vehicle access device. BACKGROUND
[0002] With the popularity of keyless entry and one-key start function on passenger cars, the vehicle door handle has evolved from a simple mechanical structure to a human-vehicle interaction node integrating multiple sensing devices.
[0003] The existing intelligent access system usually uses capacitive, photoelectric or simple pressure sensors to sense the driver's holding action, but due to the complex environment outside the vehicle, these single sensing methods also respond to rainwater, ice and snow residues, accidental touch or metal foreign objects, resulting in frequent false unlocking events; in addition, the traditional temperature sensing scheme mostly uses thermoelectric elements to measure the absolute temperature of the handle surface, which cannot distinguish between short-term environmental temperature rise and local temperature gradient change caused by human contact, and ultrasonic detection mostly uses a threshold of echo amplitude, without considering the multiple reflections and attenuation mechanism of sound waves in the narrow holding groove, making the system unreliable in high-temperature exposure, low-temperature frosting and complex acoustic background.
[0004] Based on the above scenario, the core problem to be solved by the prior art is: how to establish a multi-modal feature fusion method that can simultaneously quantify the heat field diffusion rate, the sound wave propagation time difference and the attenuation difference of the two, in the limited space of the vehicle door handle, to dynamically evaluate the credibility of the holding behavior, so as to significantly reduce the false triggering rate and ensure the immediate unlocking needs of the legitimate user under different seasons, humidity and noise conditions SUMMARY
[0005] The present application provides a vehicle door handle sensing system of an intelligent vehicle access device, which solves the technical problems raised in the background art.
[0006] The present application provides a vehicle door handle sensing system of an intelligent vehicle access device, which includes: A temperature gradient acquisition module is configured to acquire a temperature matrix of a holding groove area of the vehicle door handle and calculate a temperature gradient characteristic value of the temperature matrix. A thermal attenuation coefficient calculation module is configured to continuously acquire N temperature peak value data of the holding groove area, fit the temperature peak value data to obtain a first fitting function, and calculate a thermal attenuation coefficient of the holding groove area according to a slope of the first fitting function. An acoustic time difference detection module is configured to emit an acoustic signal to the holding groove area and calculate an acoustic time difference between a main echo arrival time and a twice-focused echo time. An acoustic attenuation coefficient calculation module is configured to continuously acquire M echo envelope peak value data of the holding groove area, fit the echo envelope peak value data to obtain a second fitting function, and calculate an acoustic attenuation coefficient of the holding groove area according to a slope of the second fitting function. an attenuation coefficient ratio module configured to calculate an absolute difference between the thermal attenuation coefficient and the acoustic attenuation coefficient; a confidence fusion calculation module configured to perform fusion calculation on the temperature gradient characteristic value, the acoustic travel time and the absolute difference to obtain a comprehensive confidence of the current holding behavior; an unlocking decision execution module configured to calculate a confidence difference between the comprehensive confidence and a preset confidence threshold, and determine whether to unlock the vehicle door according to the confidence difference.
[0007] Preferably, the temperature matrix of the holding groove area of the vehicle door handle is collected, including: a miniature infrared thermal imager is installed at a preset position of the holding groove area; the miniature infrared thermal imager is used to collect thermal radiation information of each position point of the holding groove area to obtain infrared thermal radiation data of the position point; the infrared thermal radiation data is converted into temperature values; the temperature values are arranged and combined according to the spatial distribution of the holding groove area to obtain the temperature matrix.
[0008] Preferably, the temperature gradient characteristic value of the temperature matrix is calculated, including: each temperature data point in the temperature matrix is taken as a center point, and a temperature difference value between the center point and an adjacent temperature data point is calculated; a weighted summation operation is performed on the temperature difference value based on a spatial distance weight of the adjacent temperature data point and the center point to obtain the temperature gradient characteristic value.
[0009] Preferably, N temperature peak data of the holding groove area are continuously collected, the temperature peak data is fitted to obtain a first fitting function, and a thermal attenuation coefficient of the holding groove area is calculated according to a slope of the first fitting function, including: N temperature peak data of the holding groove area are continuously collected at a fixed time interval, wherein the temperature peak data includes a temperature peak value and a collection time point; the temperature peak value and the collection time point are converted into first logarithmic coordinates, and the first logarithmic coordinates are linearly fitted by using a least square method to obtain the first fitting function; a thermal diffusion mapping relationship formula of the slope of the first fitting function and the thermal attenuation coefficient is established based on a thermal diffusion equation in the heat conduction theory; the slope of the second fitting function is substituted into the thermal diffusion mapping relationship formula to obtain the thermal attenuation coefficient.
[0010] Preferably, an acoustic signal is emitted to the holding groove area, and an acoustic travel time between a main echo arrival time and a secondary focus echo time is calculated, including: in response to the current holding behavior, an acoustic signal is emitted to the holding groove area by using an acoustic wave emission unit; Record the arrival time of the main echo from the time the acoustic wave signal is emitted to the time it touches the finger and returns; Record the secondary focusing echo time when the acoustic wave signal is reflected by the groove wall of the grip groove area and then focused again to the groove bottom of the grip groove area; The difference between the main echo arrival time and the secondary focus echo time is calculated to obtain the acoustic wave time difference.
[0011] Preferably, continuously collecting M echo envelope peak data of the grip groove area, fitting the echo envelope peak data to obtain a second fitting function, and calculating the acoustic attenuation coefficient of the grip groove area according to the slope of the second fitting function, including: After transmitting the acoustic wave signal to the grip area, the reflected echo signal is segmented and integrated within a fixed time window to obtain M echo envelope peak data, where the echo envelope peak data includes the echo integral energy value and the marked time point; Converting the integrated energy value and the marked time point into a second logarithmic coordinate, performing a linear fit on the second logarithmic coordinate using the least squares method to obtain a second fitting function; Based on the attenuation model in sound propagation theory, a sound attenuation mapping relationship formula between the slope of the second fitting function and the sound attenuation coefficient is established; Substitute the slope of the second fitting function into the sound attenuation mapping relationship formula to obtain the sound attenuation coefficient.
[0012] Preferably, the temperature gradient characteristic value, the acoustic time difference and the absolute difference are fused and calculated to obtain the comprehensive confidence of the current gripping behavior, including: Normalizing the temperature gradient characteristic value, the acoustic wave time difference, and the absolute difference to obtain the normalized temperature gradient characteristic value, the acoustic wave time difference, and the absolute difference; Perform weighted calculation on the normalized temperature gradient eigenvalue, acoustic time difference and absolute difference to obtain the intermediate fusion amount; The intermediate fusion amount is mapped and the comprehensive confidence is obtained.
[0013] Preferably, determining whether to unlock the door according to the confidence difference includes: Judgment confidence difference Is it greater than 0? like , then unlock the door; like , the doors remain locked.
[0014] Preferably, the reflected echo signal is subjected to segmented integration processing to obtain M echo envelope peak data, including: Divide the fixed time window into M time periods; Perform envelope extraction on the reflected echo signal in each time period to obtain the echo envelope curve of each time period; Integrating the envelope curve to obtain an echo integral energy value of each time period; Taking the middle time of the time period as a marked time point of the echo integral energy value; Combining the echo integral energy value and the marked time point into a wave envelope peak value data.
[0015] The present application has the following advantages: 1、The present application introduces a miniature infrared thermal imager in the grip groove area of the door handle, through temperature gradient sampling, primary-secondary echo time difference detection and other means, the thermal field diffusion rate and the acoustic propagation path can be captured and modeled in real time at the same time, compared with the existing scheme which only relies on a single capacitance or photoelectric sensing, the multi-modal perception architecture can still maintain stable signal output under extreme working conditions such as rain washing, ice and snow covering or strong light directivity, and the false unlocking triggered by single channel distortion is eliminated from the source, and the reliable identification robustness of the whole vehicle in multiple climate and multiple dirt scenes is significantly improved.
[0016] 2、The present application adopts logarithmic linear fitting to extract thermal attenuation coefficient and acoustic attenuation coefficient, and further calculates the absolute difference value of the two as an attenuation heterogeneity index, so that this kind of pseudo grip situation that the temperature rises slowly but the sound energy decays quickly is directly eliminated, compared with the traditional thermal / acoustic amplitude value judgment based on fixed threshold, the present application takes the attenuation dynamic characteristics into the criterion, so that the judgment result has natural immunity to environmental temperature drift, glove heat insulation, gap echo and other complex interference, not only reduces the false positive rate in winter low temperature and summer high temperature, but also avoids additional calibration process, and provides a more relaxed manufacturing tolerance window for batch loading.
[0017] 3、The present application adaptively fuses the temperature gradient characteristics, acoustic time difference and attenuation heterogeneity by weight, constructs a comprehensive confidence, and directly drives the door unlocking logic with the positive and negative relationship between the comprehensive confidence and the threshold, realizes one-key decision chain of high confidence opening immediately and low confidence rejecting, this fusion strategy has explainability and real-time performance: the developer can adjust the weights online according to the whole vehicle test data, to continuously track the change of user grip habit, compared with the scheme which depends on MCU multi-level interrupt or needs cloud confirmation, the present application can complete the judgment locally, and shortens the unlocking response time. BRIEF DESCRIPTION OF DRAWINGS
[0018] Figure 1 A functional module diagram of a door handle sensing system of an automobile intelligent access device is provided for an embodiment of the present application. DETAILED DESCRIPTION
[0019] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and that changes to the functions and arrangements of elements discussed can be made without departing from the scope of the content of this specification. Various examples can omit, substitute, or add various procedures or components as desired. Additionally, features described with respect to some examples can be combined in other examples.
[0020] As shown in Figure 1 A vehicle door handle sensing system of an intelligent vehicle entry device includes: A temperature gradient acquisition module is configured to acquire a temperature matrix of a holding groove region of a vehicle door handle and calculate a temperature gradient characteristic value of the temperature matrix. In an embodiment of the present application, the temperature matrix of the holding groove region of the vehicle door handle includes: A miniature infrared thermal imager is installed at a preset position of the holding groove region. Specifically, the miniature infrared thermal imager is installed at a preset position of the holding groove region, and the miniature infrared thermal imager is installed at an inner side of a top of the holding groove, that is, a position of the holding groove housing corresponding to above a palm of a user when the user holds the holding groove. In a specific implementation, an infrared thermal imager mounting hole is reserved at a stage of injection molding of the top of the holding groove, the size of the hole is adapted to the shape of the infrared thermal imager, the lens of the infrared thermal imager faces the opening direction of the holding groove, and is arranged at an inclination angle of 15°-30° with respect to the bottom surface of the holding groove. This position needs to cover the complete spatial range of the holding groove region, and needs to ensure that the field of view of the thermal imager is adapted to the geometric size of the holding groove, so that there is no blind area for collecting thermal radiation information. The selection of the infrared thermal imager needs to meet the adaptability of the vehicle-mounted environment, and needs to have the ability to interact with the signal of the system main control unit, to ensure that the collected data can be transmitted to the subsequent processing module.
[0021] The thermal radiation information of each position point of the holding groove region is collected by using the miniature infrared thermal imager, to obtain infrared thermal radiation data of the position point. Specifically, the infrared thermal imager converts the infrared radiation energy emitted by different position points of the holding groove region into an electrical signal or a digital signal based on the principle of infrared thermal imaging. During the collection process, the thermal imager continuously works at a preset sampling frequency, to ensure that the dynamic changes of the temperature field of the holding groove region are captured, and the thermal radiation information of each position point corresponds to the radiation characteristics of the surface temperature thereof.
[0022] The infrared thermal radiation data is converted into a temperature value. Specifically, according to the Planck blackbody radiation law, the infrared thermal radiation data collected is processed in combination with the optical parameters of the thermal imager and the response characteristics of the detector, an influence of environmental factors is eliminated by using a calibration algorithm, and the radiation energy data is accurately converted into an absolute temperature value of the corresponding position point.
[0023] The temperature values are arranged and combined according to the spatial distribution of the grip groove area to obtain a temperature matrix.
[0024] Specifically, based on the grip area's three-dimensional geometric model, the area is divided into discrete spatial grid points, each corresponding to a temperature value captured by the infrared thermal imager. The temperature values are then sequentially assigned to the rows and columns of a matrix, following the spatial coordinate order of the grid points, to form a two-dimensional temperature matrix. The row and column indices of this matrix correspond one-to-one with the spatial coordinates of the grip area, fully characterizing the temperature distribution within the grip area and providing a structured data foundation for subsequent temperature gradient eigenvalue calculations.
[0025] In an embodiment of the present invention, calculating the temperature gradient eigenvalue of the temperature matrix includes: Taking each temperature data point in the temperature matrix as the center point, calculate the temperature difference between the center point and the adjacent temperature data points; Specifically, each temperature data point in the temperature matrix is taken as the center point, and all rows and columns of the temperature matrix are traversed. For each center point, the set of its adjacent temperature data points is determined. The selection of adjacent points follows the spatial distribution logic of the temperature matrix, covering the adjacent points of the center point in the horizontal, vertical and diagonal directions. For example, when 8 neighborhoods are selected, 8 adjacent points in the upper and lower, left and right, and four diagonal directions are included; if it is a 4-neighborhood, only the horizontal and vertical directions are included to ensure that the local temperature change trend of the temperature matrix is fully captured; through the absolute difference operation , eliminating the influence of the temperature change direction and retaining only the amplitude information of the temperature difference. The temperature difference reflects the degree of temperature change between the center point and the adjacent points and is the basic parameter for temperature gradient calculation.
[0026] The temperature difference is weighted and summed based on the spatial distance weights between adjacent temperature data points and the center point to obtain the temperature gradient eigenvalue. The calculation formula of the temperature gradient eigenvalue is as follows:
[0027] Where, is the temperature gradient eigenvalue, is the number of rows in the temperature matrix, is the number of columns in the temperature matrix, is the center point, are adjacent temperature data points, is the first Row, No. The temperature value of the temperature data point of the column, is the temperature value of the adjacent temperature data point, It is the center point A collection of adjacent temperature data points, is the spatial distance weight.
[0028] Specifically, the calculation of the spatial distance weight is based on the spatial resolution of the temperature matrix, that is, the actual physical distance of adjacent temperature data points in the holding groove region, the closer the adjacent point to the center point, the greater the weight, and through weighted summation, the temperature gradient characteristic value is more focused on reflecting the significant change of the local temperature field, such as the temperature difference between the finger holding area and the surrounding environment, and the interference of the gentle temperature change is suppressed; the weighted summation result of all center points in the temperature matrix is subjected to global average operation, the local temperature gradient contribution of each center point is averaged according to the total data points of the temperature matrix, and the temperature gradient characteristic value representing the overall temperature change degree of the holding groove region is obtained, and the temperature gradient characteristic value comprehensively reflects the spatial distribution difference of the temperature field of the holding groove region, thereby providing a key thermal field feature for subsequent holding behavior recognition.
[0029] The heat attenuation coefficient calculation module is configured to continuously collect N temperature peak values of the holding groove region, fit the temperature peak value data to obtain a first fitting function, and calculate the heat attenuation coefficient of the holding groove region according to the slope of the first fitting function. In detail, the propagation and attenuation process of the temperature in the holding groove region follows the heat diffusion equation in the heat conduction theory, and the heat diffusion equation describes the law of the heat field changing with time. The N temperature peak values collected continuously record the dynamic attenuation information of the heat field of the holding groove region with time, and the dynamic attenuation information includes the temperature peak value and the corresponding collection time point. After the temperature peak value and the collection time point are converted into logarithmic coordinates, the heat diffusion equation can exhibit linear characteristics in the logarithmic domain. The data in the logarithmic coordinates are linearly fitted by using the least square method, and the first fitting function obtained has a clear mapping relationship between the slope and the heat attenuation coefficient based on the derivation of the heat diffusion equation. Through the mapping relationship, the slope of the first fitting function is substituted into the heat attenuation coefficient, which is accurately extracted from the data reflecting the dynamic attenuation of the heat field, and the heat attenuation coefficient representing the heat energy attenuation rate of the holding groove region with time is accurately extracted, so that the quantitative calculation of the heat attenuation characteristic is realized, and a key heat field dynamic parameter is provided for subsequent holding behavior recognition.
[0030] In the embodiment of the application, the heat attenuation coefficient of the holding groove region is calculated according to the slope of the first fitting function, including: N temperature peak values of the holding groove region are continuously collected at a fixed time interval, wherein the temperature peak value data includes a temperature peak value and a collection time point. Specifically, at a fixed time interval (such as 50 milliseconds, which can be adaptively adjusted according to the thermal response characteristics of the grip groove area), N temperature peak data of the grip groove area are continuously collected. Each temperature peak data contains two parts of information: one is the temperature peak, that is, the maximum temperature value in the temperature matrix of the grip groove area at the time of collection, reflecting the strongest radiation characteristics of the grip groove thermal field at that moment; the other is the collection time point, with the first collection moment as the time origin, and subsequent collection moments are accumulated at time intervals to establish a correlation sequence between temperature peaks and time.
[0031] The temperature peak and the acquisition time point are converted into first logarithmic coordinates, and a linear fitting is performed on the first logarithmic coordinates using the least squares method to obtain a first fitting function; The temperature peak and the corresponding acquisition time point are converted into the first logarithmic coordinates, that is, the common logarithm of the temperature peak is obtained , take the common logarithm of the acquisition time point to get , using the least squares method to transform the logarithmic coordinate data Perform linear fitting and construct the first fitting function , slope Reflects the rate of change of the logarithm of the peak temperature with the logarithm of time, intercept The logarithmic temperature offset at the initial acquisition moment is reflected, and the fitting residual is minimized by the least squares method to ensure that the fitting function accurately represents the attenuation trend of the temperature peak.
[0032] Based on the heat diffusion equation in heat conduction theory, a heat diffusion mapping relationship formula between the slope of the first fitting function and the thermal attenuation coefficient is established. The heat diffusion mapping relationship formula is as follows:
[0033] Where, is the thermal attenuation coefficient, is the slope of the first fitting function; The slope of the second fitting function is substituted into the thermal diffusion mapping relationship formula to obtain the thermal attenuation coefficient.
[0034] Specifically, based on the heat diffusion equation in heat conduction theory, such as the Fourier-Poisson equation that describes the unsteady-state heat conduction of a semi-infinite object, the relationship between the slope of the first fitting function and the thermal attenuation coefficient is derived. The heat diffusion equation describes the spatial distribution and time variation of the temperature field through partial differentials, combined with the geometric boundary conditions of the grip area, and through dimensionless processing and analytical solution, obtains the logarithmic linear relationship between the temperature peak attenuation law and time, and then establishes the heat diffusion mapping relationship formula. The negative sign reflects the physical property that the temperature decreases with time. The coefficient 2 is derived from the analytical solution coefficient of the heat diffusion equation and is related to the thermal diffusivity and geometric dimensions of the grip material.
[0035] In detail, the thermal decay coefficient comprehensively reflects the energy dissipation rate of the hot field of the holding groove in the continuous acquisition period, when the fingers are held, the heat exchange between the holding groove and the fingers changes the size of the thermal decay coefficient, for example, the fingers act as a heat source to supplement The absolute value is reduced to provide a thermal field dynamic feature for subsequent holding behavior recognition.
[0036] An acoustic travel time detection module is configured to emit an acoustic wave signal to the holding groove area, and calculate the acoustic travel time of the main echo arrival time and the secondary focusing echo time. In the embodiments of the present application, the acoustic wave signal is emitted to the holding groove area, and the acoustic travel time of the main echo arrival time and the secondary focusing echo time is calculated, comprising: In response to the current holding behavior, an acoustic wave emission unit is used to emit an acoustic wave signal to the holding groove area. Specifically, when a trigger signal indicating a possible holding behavior is detected, such as when the preliminary features of the temperature gradient characteristic value meet certain conditions, that is, in response to the current potential holding behavior, the acoustic wave emission unit is started to emit an acoustic wave signal to the holding groove area according to preset acoustic wave parameters, such as specific frequency, amplitude, pulse width, etc. These parameters need to be adapted to the spatial size and material characteristics of the holding groove area.
[0037] The main echo arrival time of the acoustic wave signal from emission to touching the fingers and returning is recorded. Specifically, during the propagation of the acoustic wave signal in the holding groove area, if there is a finger holding, the acoustic wave signal will first touch the finger, and the main echo will be formed after being reflected by the finger. The acoustic wave receiving unit captures the main echo signal and accurately records the time interval from the acoustic wave signal emission time to the main echo receiving time. This is the main echo arrival time of the acoustic wave signal from emission to touching the fingers and returning. The secondary focusing echo time of the acoustic wave signal focusing to the groove bottom of the holding groove area again after being reflected by the groove wall is recorded. Specifically, part of the emitted acoustic wave signal will propagate to the groove wall of the holding groove area, and after being reflected by the groove wall, the acoustic wave energy will continue to propagate in the holding groove space. When certain acoustic conditions are met, such as the phase and amplitude of the groove wall reflected acoustic wave meeting the focusing requirements, which are determined by the groove geometry and the acoustic wave frequency, the acoustic wave will be focused to the groove bottom of the holding groove area again. The acoustic wave receiving unit synchronously captures the echo signal generated by the secondary focusing process, and records the time interval from the acoustic wave emission time to the secondary focusing echo receiving time as the secondary focusing echo time.
[0038] The main echo arrival time and the secondary focusing echo time are calculated by difference to obtain the acoustic travel time.
[0039] Specifically, according to the time difference calculation principle, the time difference value of the two is obtained by subtracting the main echo arrival time from the secondary focusing echo time, which is the acoustic time difference, which can reflect the acoustic propagation path and energy distribution changes caused by the finger holding in the holding groove area, and provide acoustic dimension feature data for subsequent comprehensive judgment of the holding behavior.
[0040] The sound attenuation coefficient calculation module is configured to continuously collect M echo envelope peak value data of the holding groove area, fit the echo envelope peak value data to obtain a second fitting function, and calculate the sound attenuation coefficient of the holding groove area according to the slope of the second fitting function. In the embodiment of the application, the M echo envelope peak value data of the holding groove area are continuously collected, the echo envelope peak value data are fitted to obtain a second fitting function, and the sound attenuation coefficient of the holding groove area is calculated according to the slope of the second fitting function, including: The M echo envelope peak value data are obtained by performing segmented integral processing on the reflected echo signal in a fixed time window after the acoustic signal is emitted to the holding groove area, wherein the echo envelope peak value data include an echo integral energy value and a marked time point. The integral energy value and the marked time point are converted into second logarithmic coordinates, and the second logarithmic coordinates are linearly fitted by using the least square method to obtain a second fitting function. Specifically, after the acoustic signal is emitted to the holding groove area, the reflected echo signal is subjected to segmented integral processing in a pre-set fixed time window, and the integral energy value and the corresponding marked time point in each echo envelope peak value data collected are calculated according to the logarithmic transformation rule to obtain and Each group of corresponding is taken as the abscissa, is taken as the ordinate, and the second logarithmic coordinates are constructed to realize the conversion from the original time domain energy-time data to the linearized data in the logarithmic domain, so that the decay law of the echo energy with time presents a linear feature that is easier to fit in the logarithmic coordinates, which is consistent with the mathematical expression form of the attenuation model in the acoustic propagation theory.
[0041] In detail, because the attenuation model in the sound propagation theory has explicitly described the physical law of the energy attenuation of the sound wave with time or distance when the sound wave propagates in the medium, which is consistent with the characteristics of the actual sound wave propagating in the holding groove area, when the peak value data of the collected echo envelope is logarithmically converted and fitted into the second fitting function, the relationship between the sound wave energy attenuation and time presents a linear feature in the logarithmic coordinates, which is consistent with the mathematical expression of the sound attenuation model. Through analyzing the internal relationship between the energy attenuation parameter in the attenuation model and the slope of the fitting straight line in the logarithmic coordinates, and using the quantitative description of the sound attenuation coefficient to the sound wave energy attenuation in the model, the energy attenuation trend reflected by the slope of the fitting function can be corresponded to the physical quantity of the sound attenuation coefficient, so as to establish the mapping relationship between them, and accurately calculate the sound attenuation coefficient through the slope of the fitting function, thereby providing a reliable quantitative basis for judging the holding behavior based on the acoustic signal.
[0042] Based on the attenuation model in the sound propagation theory, the slope of the second fitting function and the sound attenuation mapping relationship formula of the sound attenuation coefficient are established, wherein the sound attenuation mapping relationship formula is as follows:
[0043] In the formula, a is the sound attenuation coefficient, is the sound attenuation coefficient, is the slope of the second fitting function; The slope of the second fitting function is substituted into the sound attenuation mapping relationship formula to obtain the sound attenuation coefficient.
[0044] Specifically, first, according to the sound propagation theory, it is clear that when the sound wave propagates in the holding groove area, its energy will attenuate according to a certain law with the propagation distance, time and other factors. The attenuation law is described by the sound propagation attenuation model. The model reflects the internal relationship between the energy attenuation degree and the propagation process parameters. By analyzing the correlation between the energy attenuation characteristics in the sound propagation attenuation model and the slope of the second fitting function, and through theoretical derivation of the model and adaptation verification with the actual collected data, the energy attenuation rate characteristics corresponding to the slope of the second fitting function are sorted out, and the corresponding relationship between the slope and the sound attenuation coefficient is established, that is, the sound attenuation mapping relationship formula, so that the sound attenuation coefficient reflecting the speed of the sound wave energy attenuation in the holding groove area can be calculated according to the slope of the fitting function combined with the law of the sound propagation attenuation model, thereby providing a key parameter basis for subsequently judging the holding behavior by using the acoustic characteristics.
[0045] In the embodiments of the present application, the reflected echo signal is subjected to segmented integral processing to obtain M echo envelope peak value data, including: The fixed time window is equally divided into M time periods; The envelope of the reflected echo signal in each time period is extracted to obtain the echo envelope curve of each time period; Integrate the envelope curve to obtain the echo integral energy value of each time period; Take the middle time of the time period as the marked time point of the echo integral energy value; Combine the echo integral energy value and the marked time point into the wave envelope peak value data.
[0046] Specifically, according to the propagation characteristics of the reflected echo signal and the analysis requirements, a fixed time window covering the complete time length of the sound wave emission and reflection and propagation in the holding groove area is determined, and the window is then equally divided into M continuous and non-overlapping time periods to ensure that the energy changes of the echo signal at different propagation stages are captured in detail. Then, for the reflected echo signal in each time period, an envelope extraction algorithm is used to remove the high-frequency carrier component and retain the envelope curve reflecting the energy change trend, presenting the energy fluctuation characteristics of the echo signal corresponding to the time period. Then, the envelope curve of each time period is integrated to calculate the area enclosed by the curve and the time axis, and the echo integral energy value is obtained to quantify the total echo energy of the corresponding time period. Then, the middle time of each time period is selected as the marked time point of the echo integral energy value to balance the energy distribution in the time period and enhance the physical meaning and comparability of the data. Finally, the echo integral energy value of each time period is associated and combined with its marked time point to form M sets of echo envelope peak values.
[0047] The attenuation coefficient comparison module is used to calculate the absolute difference between the thermal attenuation coefficient and the acoustic attenuation coefficient; In detail, first, the thermal attenuation coefficient and the acoustic attenuation coefficient of the holding groove area are obtained through the aforementioned steps, the thermal attenuation coefficient is derived from the temperature peak value data fitting and the thermal diffusion mapping relationship, reflecting the attenuation characteristics of the holding groove area thermal field energy over time; the acoustic attenuation coefficient is obtained based on the echo envelope peak value data fitting and the sound propagation attenuation model mapping, reflecting the attenuation law of the holding groove area sound field energy. Then, according to the calculation logic of the absolute difference in mathematics, the thermal attenuation coefficient and the acoustic attenuation coefficient are substituted into the difference operation, i.e. the absolute value of the difference between the two values, to obtain the absolute difference between the thermal attenuation coefficient and the acoustic attenuation coefficient. This difference can comprehensively measure the difference degree of the attenuation characteristics of the thermal and acoustic physical fields in the holding groove area, and provide a key differentiated feature parameter for subsequent fusion calculation of the comprehensive confidence of the holding behavior.
[0048] The confidence fusion calculation module is used to fuse and calculate the temperature gradient characteristic value, the acoustic time difference and the absolute difference to obtain the comprehensive confidence of the current holding behavior; In the embodiments of the present application, the temperature gradient characteristic value, the acoustic time difference and the absolute difference are fused and calculated to obtain the comprehensive confidence of the current holding behavior, including: The temperature gradient characteristic value, the acoustic time difference and the absolute difference are normalized to obtain the normalized temperature gradient characteristic value, the acoustic time difference and the absolute difference; In detail, since the temperature gradient characteristic value, the acoustic travel time and the absolute difference value are different in numerical range and dimension, in order to ensure that each characteristic can participate fairly in the fusion calculation, the three types of characteristic data need to be normalized. The normalization algorithm is used to map the data to a unified numerical interval, eliminate the influence caused by the original numerical range and dimension difference, and obtain the normalized temperature gradient characteristic value, the acoustic travel time and the absolute difference value which can be directly fused and operated.
[0049] The normalized temperature gradient characteristic value, the acoustic travel time and the absolute difference value are weighted and calculated to obtain an intermediate fusion quantity, wherein the calculation formula of the intermediate fusion quantity is as follows:
[0050] In the formula, is the normalized temperature gradient characteristic value, is the normalized acoustic travel time, is the normalized absolute difference value, is a temperature weight coefficient, is an acoustic travel time weight coefficient, is an attenuation difference penalty coefficient. In detail, according to the physical meaning and contribution degree of the temperature gradient characteristic, the acoustic travel time and the absolute difference value in representing the holding behavior, corresponding weight coefficients are respectively given to the three types of normalized characteristics. The normalized temperature gradient characteristic value, the acoustic travel time and the absolute difference value are linearly combined by weighted calculation to obtain the intermediate fusion quantity. The intermediate fusion quantity comprehensively combines the information of the three types of characteristics in the holding behavior judgment, and highlights the effect of different characteristics on the holding behavior confidence judgment. The determination of the temperature weight coefficient, the acoustic travel time weight coefficient and the attenuation difference penalty coefficient relies on multiple sets of holding and non-holding measured data. The control variable method is used to do comparative experiments, the influence of different coefficients on the recognition accuracy and the misjudgment rate is observed, a multi-objective optimization model is constructed based on the theory of thermal and acoustic characteristics, the optimal recognition effect is taken as the target, the iterative algorithm is used to solve the coefficient combination, and the coefficients are adaptively adjusted according to the feedback of different scenes. The whole vehicle level experiment is verified to ensure that the coefficients adapt to the actual situation and reliably distinguish the holding behavior.
[0051] The intermediate fusion quantity is mapped and operated to obtain a comprehensive confidence, wherein the calculation formula of the comprehensive confidence is as follows:
[0052] In the formula, is the comprehensive confidence, is a slope adjustment coefficient, is the intermediate fusion quantity, is a threshold shift parameter, is an exponential function.
[0053] In detail, to transform the intermediate fusion amount S into a result conforming to the definition of confidence (0 to 1 interval, the greater the value, the more reliable the holding behavior), a mapping operation based on an exponential function is introduced, and the formula The intermediate fusion amount is nonlinearly transformed, the output comprehensive confidence reflects the possibility of the current holding behavior, and an intuitive and quantitative basis for subsequent unlocking decision is provided.
[0054] The unlocking decision execution module is configured to calculate a confidence difference value of the comprehensive confidence and a preset confidence threshold, and determine whether to unlock the vehicle door according to the confidence difference value.
[0055] In the embodiment of the present application, whether to unlock the vehicle door according to the confidence difference value includes: determining whether the confidence difference value is greater than 0: if , the vehicle door is unlocked; if , the vehicle door remains in a locked state.
[0056] The difference value between the comprehensive confidence and the preset confidence threshold is calculated, and is denoted as the confidence difference value The difference value reflects the deviation of the comprehensive confidence of the current holding behavior from the system preset judgment standard, and then, the confidence difference value is logically judged, and the core is to compare its size relationship with 0. When the confidence difference value is greater than or equal to 0, it means that the comprehensive confidence of the current holding behavior reaches or exceeds the system preset unlocking judgment standard, indicating that this holding behavior is probably a valid operation of a legal user. At this time, the unlocking decision execution module triggers the vehicle door unlocking process, sends an unlocking instruction to the vehicle door lock control mechanism, and completes the vehicle door unlocking action. If is less than 0, it means that the comprehensive confidence of the current holding behavior does not reach the system preset unlocking requirement, and there is a possibility of illegal operation or invalid holding. The unlocking decision execution module controls the vehicle door to remain in a locked state and refuses to execute the unlocking operation, so as to protect the safety of the vehicle. Through simple and clear logical judgment, precise control of the vehicle door unlocking permission is realized.
[0057] The above describes the embodiment of the present application, but the present application is not limited to the above specific embodiments. The above specific embodiments are only illustrative and not limiting. Those skilled in the art can make many forms under the inspiration of the present application, which are all within the protection scope of the present application.
Claims
1. A door handle sensing system for a car intelligent entry device, characterized in that: include: The temperature gradient acquisition module is used to collect the temperature matrix of the grip groove area on the door handle and calculate the temperature gradient eigenvalue of the temperature matrix; a thermal attenuation coefficient calculation module, configured to continuously collect N temperature peak data of the grip groove area, fit the temperature peak data to obtain a first fitting function, and calculate the thermal attenuation coefficient of the grip groove area according to the slope of the first fitting function; The acoustic wave time difference detection module is used to transmit acoustic wave signals to the grip area and calculate the acoustic wave time difference between the main echo arrival time and the secondary focus echo time; an acoustic attenuation coefficient calculation module, configured to continuously collect M echo envelope peak data of the grip groove area, fit the echo envelope peak data to obtain a second fitting function, and calculate the acoustic attenuation coefficient of the grip groove area according to the slope of the second fitting function; Attenuation coefficient comparison module, used to calculate the absolute difference between thermal attenuation coefficient and acoustic attenuation coefficient; The confidence fusion calculation module is used to fuse the temperature gradient characteristic value, the acoustic wave time difference and the absolute difference to obtain the comprehensive confidence of the current gripping behavior; The unlocking decision execution module is used to calculate the confidence difference between the comprehensive confidence and the preset confidence threshold, and determine whether to unlock the door based on the confidence difference.
2. The door handle sensing system of the automobile intelligent entry device according to claim 1, characterized in that: Collect the temperature matrix of the grip area on the door handle, including: Install a miniature infrared thermal imager at a preset position in the grip area; Use a miniature infrared thermal imager to collect thermal radiation information from various points in the grip area to obtain infrared thermal radiation data of the points; Convert infrared thermal radiation data into temperature values; The temperature values are arranged and combined according to the spatial distribution of the grip groove area to obtain a temperature matrix.
3. The door handle sensing system of the automobile intelligent entry device according to claim 1, characterized in that: Calculate the temperature gradient eigenvalues of the temperature matrix, including: Taking each temperature data point in the temperature matrix as the center point, calculate the temperature difference between the center point and the adjacent temperature data points; The temperature gradient characteristic value is obtained by performing a weighted summation operation on the temperature difference based on the spatial distance weights between adjacent temperature data points and the center point.
4. The door handle sensing system of the automobile intelligent entry device according to claim 1, characterized in that: Continuously collect N temperature peak data of the grip groove area, fit the temperature peak data to obtain a first fitting function, and calculate the thermal attenuation coefficient of the grip groove area according to the slope of the first fitting function, including: Continuously collecting N temperature peak data of the grip groove area at fixed time intervals, wherein the temperature peak data includes the temperature peak and the collection time point; The temperature peak and the acquisition time point are converted into first logarithmic coordinates, and a linear fitting is performed on the first logarithmic coordinates using the least squares method to obtain a first fitting function; Based on the heat diffusion equation in heat conduction theory, a heat diffusion mapping relationship formula between the slope of the first fitting function and the thermal attenuation coefficient is established; The slope of the second fitting function is substituted into the thermal diffusion mapping relationship formula to obtain the thermal attenuation coefficient.
5. The door handle sensing system of the automobile intelligent entry device according to claim 1, characterized in that: Transmit an acoustic wave signal to the grip area and calculate the acoustic wave time difference between the main echo arrival time and the secondary focus echo time, including: In response to the current gripping behavior, the acoustic wave transmitting unit transmits an acoustic wave signal to the gripping groove area; Record the arrival time of the main echo from the time the acoustic wave signal is emitted to the time it touches the finger and returns; Record the secondary focusing echo time when the acoustic wave signal is reflected by the groove wall of the grip groove area and then focused again to the groove bottom of the grip groove area; The difference between the main echo arrival time and the secondary focus echo time is calculated to obtain the acoustic wave time difference.
6. The door handle sensing system of the automobile intelligent entry device according to claim 1, characterized in that: Continuously collect M echo envelope peak data of the grip groove area, fit the echo envelope peak data to obtain a second fitting function, and calculate the acoustic attenuation coefficient of the grip groove area according to the slope of the second fitting function, including: After transmitting the acoustic wave signal to the grip area, the reflected echo signal is segmented and integrated within a fixed time window to obtain M echo envelope peak data, where the echo envelope peak data includes the echo integral energy value and the marked time point; Converting the integrated energy value and the marked time point into a second logarithmic coordinate, performing a linear fit on the second logarithmic coordinate using the least squares method to obtain a second fitting function; Based on the attenuation model in sound propagation theory, a sound attenuation mapping relationship formula between the slope of the second fitting function and the sound attenuation coefficient is established; Substitute the slope of the second fitting function into the sound attenuation mapping relationship formula to obtain the sound attenuation coefficient.
7. The door handle sensing system of a car intelligent entry device according to claim 1, characterized in that: The temperature gradient eigenvalue, acoustic time difference, and absolute difference are fused and calculated to obtain the comprehensive confidence of the current gripping behavior, including: Normalizing the temperature gradient characteristic value, the acoustic wave time difference, and the absolute difference to obtain the normalized temperature gradient characteristic value, the acoustic wave time difference, and the absolute difference; Perform weighted calculation on the normalized temperature gradient eigenvalue, acoustic time difference and absolute difference to obtain the intermediate fusion amount; The intermediate fusion amount is mapped and the comprehensive confidence is obtained.
8. The door handle sensing system of a car intelligent entry device according to claim 1, characterized in that: Determine whether to unlock the door based on the confidence difference, including: Judgment confidence difference Is it greater than 0? like ,, then unlock the door; like , the doors remain locked.
9. The door handle sensing system of the automobile intelligent entry device according to claim 6, characterized in that: Perform segmented integration processing on the reflected echo signal to obtain M echo envelope peak data, including: Divide the fixed time window into M time periods; Perform envelope extraction on the reflected echo signal in each time period to obtain the echo envelope curve of each time period; Perform integration operation on the envelope curve to obtain the echo integral energy value of each time period; The middle moment of the time period is used as the marking time point of the echo integral energy value; Combine the echo integral energy value and the marked time point into the wave envelope peak data.