False touch prevention method and system for hidden door handle, storage medium and intelligent automobile
By collecting and processing capacitive sensor signals of hidden door handles, training and detection models to identify false touch events, solving the problem of hidden door handles being accidentally triggered by water conductivity, and improving identification accuracy and user safety experience.
Patent Information
- Application Number
- CN202411693174.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-22
- Publication Date
- 2025-03-18
AI Technical Summary
The hidden door handle is easily triggered by the conductivity of the water during rainy days or when there are water droplets, resulting in unexpected door handle activation, affecting the user experience and posing a safety hazard.
By obtaining the capacitive digital-analog converter mode supported by the capacitive sensor, multiple sets of original touch signals are collected and converted into original touch data through conversion formulas. The annotated training data is obtained based on these data, and the detection model is trained to identify false touch events.
It realizes accurate judgment of touch behavior, improves recognition accuracy and robustness, reduces the occurrence of false triggers, and enhances user experience and security.
Smart Images
Figure CN119646509A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of neural network technology, and in particular to a method, system, storage medium and smart car for preventing accidental touch of a hidden door handle. Background Art
[0002] In modern car design, hidden door handles are becoming increasingly popular in new vehicles due to their aesthetics and reduced wind resistance. This design usually uses capacitive sensing technology to automatically extend or retract the door handle. Specifically, when a person's finger touches the capacitive sensor of the hidden door handle, the sensor detects a change in capacitance. If this change exceeds a preset threshold, the system will assume that someone is trying to use the door handle and trigger the extension or retraction of the door handle accordingly.
[0003] On rainy days or when there are water droplets on the door handle, the conductivity of the water may mistakenly simulate the touch of a finger, causing the door handle to activate unexpectedly. In addition, other objects accidentally touching the capacitive sensor of the door handle may also cause false triggers. These false triggers may not only interfere with the user experience, but also pose a safety hazard. Summary of the invention
[0004] The technical problem to be solved by the embodiments of the present invention is to provide a method, system, storage medium and smart car for preventing accidental touch of hidden door handles, so as to solve the problem of accidental triggering of hidden door handles in the prior art.
[0005] The present invention discloses a method for preventing accidental touch of a hidden door handle, which is applied to the hidden door handle. The hidden door handle includes a capacitive sensor. The method for preventing accidental touch of the hidden door handle includes:
[0006] Acquire a capacitance digital-to-analog converter mode supported by the capacitance sensor, and acquire a conversion formula according to the capacitance digital-to-analog converter mode, wherein the capacitance digital-to-analog converter mode includes a single capacitance digital-to-analog converter mode or a dual capacitance digital-to-analog converter mode;
[0007] When the examiner touches the hidden door handle multiple times, multiple groups of original touch signals are collected, each group of the original touch signals includes the capacitance signal collected by the capacitance sensor during the period from the beginning of the touch to the touching of the hidden door handle each time the examiner touches the hidden door handle;
[0008] Converting the multiple groups of original touch signals into original touch data by using the conversion formula, acquiring labeled training data based on the original touch data, inputting the training data into a detection model for training, and acquiring a pre-trained detection model;
[0009] After the pre-trained detection model is deployed, when a change in the real-time capacitance signal of the capacitance sensor is detected, the changed capacitance signal within a preset time length is collected, the changed capacitance signal is converted into a changed touch signal through the conversion formula, the changed touch signal is input into the pre-trained detection model, the detection result is obtained, and the pop-up and retraction of the hidden door handle are controlled according to the detection result.
[0010] Optionally, the step of acquiring a conversion formula according to the capacitance digital-to-analog converter mode includes:
[0011] In single capacitor digital-to-analog converter mode, the conversion formula is:
[0012]
[0013] In the dual capacitor digital-to-analog converter mode, the conversion formula is:
[0014]
[0015] in,
[0016]
[0017]
[0018] Maxcount = N sub *TxClK Div
[0019] C S =C p +C f
[0020] N sub is the number of sub-conversions, CompClk Div To compensate for the frequency division, C p is the parasitic capacitance, C f is the finger capacitance, C S is the inductive capacitance, C ref is the reference capacitor, TxClk Div is the Tx clock division rate, F Mod is the modulator clock frequency, Maxcount is the maximum number of digital counts converted from the capacitance signal, F Tx is the clock frequency, F comp To compensate for the CDAC clock frequency.
[0021] Optionally, the step of acquiring labeled training data based on the original touch data includes:
[0022] The original touch data is annotated to obtain annotated data, and a portion of the annotated data is selected as the training data, a portion as the verification data, and the remaining portion as the test data.
[0023] Optionally, the step of inputting the training data into the detection model for training includes:
[0024] Centralizing and standardizing all the labeled data to obtain feature data;
[0025] The feature data corresponding to the training data is input into the detection model for training.
[0026] Optionally, the detection model is a back propagation neural network, and the step of inputting the training data into the detection model for training includes:
[0027] A Bayesian regularization method is selected to train the hidden layer of the back propagation neural network.
[0028] Optionally, the step of selecting the Bayesian regularization method to train the hidden layer of the back propagation neural network includes:
[0029] The number of training times is set, and when the hidden layer training of the back-propagation neural network reaches the number of training times, the back-propagation neural network is tested by the test data to prevent the detection model from overfitting.
[0030] Optionally, after the step of selecting the Bayesian regularization method to train the hidden layer of the back propagation neural network, the method further comprises:
[0031] The generalization ability of the detection model is checked using the validation data.
[0032] The present invention also discloses a hidden door handle anti-accidental touch system, which is applied to the hidden door handle. The hidden door handle includes a capacitive sensor. The hidden door handle anti-accidental touch system includes:
[0033] A formula module, used to obtain a capacitance digital-to-analog converter mode supported by the capacitance sensor, and obtain a conversion formula according to the capacitance digital-to-analog converter mode, wherein the capacitance digital-to-analog converter mode includes a single capacitance digital-to-analog converter mode or a dual capacitance digital-to-analog converter mode;
[0034] An acquisition module, used for acquiring multiple groups of original touch signals when the detector touches the hidden door handle multiple times, each group of the original touch signals includes a capacitance signal acquired by the capacitance sensor during a period from the beginning of the touch to the touching of the hidden door handle each time the hidden door handle is touched;
[0035] A training module, used for converting the multiple groups of original touch signals into original touch data through the conversion formula, acquiring labeled training data based on the original touch data, inputting the training data into a detection model for training, and acquiring a pre-trained detection model;
[0036] The detection module is used to deploy the pre-trained detection model, and when a change in the real-time capacitance signal of the capacitance sensor is detected, collect the changed capacitance signal within a preset time length, convert the changed capacitance signal into a changed touch signal through the conversion formula, input the changed touch signal into the pre-trained detection model, obtain the detection result, and control the pop-up and retraction of the hidden door handle according to the detection result.
[0037] The present invention also discloses an intelligent car, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above method.
[0038] The present invention also discloses a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the method described above.
[0039] Compared with the prior art, the beneficial effect of the method for preventing accidental touch of a hidden door handle provided by an embodiment of the present invention is that: when the detector touches the hidden door handle multiple times, multiple groups of original touch signals are collected, and the original touch signals include the capacitance signals collected by the capacitance sensor from the beginning of touching to touching the hidden handle, so as to train the detection model, and then detect whether the currently detected touch event is an accidental touch event based on the pre-trained detection model, and judge the parameter changes in the process from the human hand approaching the door handle to touching the door handle, rather than just relying on a single capacitance threshold for judgment, thereby achieving accurate judgment of touch behavior and improving recognition accuracy and robustness. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] The scheme of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments, in which:
[0041] Figure 1 It is a flow chart of an embodiment of a method for preventing accidental touch of a hidden door handle provided by the present invention;
[0042] Figure 2 It is a schematic diagram of the structure of an embodiment of a back propagation neural network provided by the present invention;
[0043] Figure 3 It is a structural schematic diagram of an embodiment of an anti-mistouch system for a hidden door handle provided by the present invention;
[0044] Figure 4 is a structural schematic diagram of an embodiment of a smart car provided by the present invention;
[0045] Figure 5 It is a schematic structural diagram of an embodiment of the computer-readable storage medium 30 provided by the present invention.
[0046] The reference numerals in the figures are:
[0047] 10. Anti-accidental-touch system for hidden door handles; 11. Formula module; 12. Collection module; 13. Training module; 14. Detection module; 20. Smart car; 21. Processor; 22. Memory; 30. Computer-readable storage medium; 31. Computer program. DETAILED DESCRIPTION
[0048] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. Now, in conjunction with the accompanying drawings, the preferred embodiments of the present invention are described in detail.
[0049] See also Figure 1 , Figure 1 1 is a flow chart of an embodiment of a method for preventing accidental touch of a hidden door handle provided by the present invention. The method for preventing accidental touch of a hidden door handle comprises the following steps:
[0050] S101: Obtain a capacitance digital-to-analog converter mode supported by the capacitance sensor, and obtain a conversion formula according to the capacitance digital-to-analog converter mode.
[0051] In modern car design, hidden door handles are widely used in new vehicles due to their aesthetics and reduced wind resistance. This design usually uses capacitive sensing technology to automatically extend or retract the door handle. Specifically, when a person's finger touches the capacitive sensor of the hidden door handle, the sensor detects a change in capacitance. If this change exceeds a preset threshold, the system will assume that someone is trying to use the door handle and trigger the extension or retraction of the door handle accordingly.
[0052] CDAC (Capacitive Digital-to-Analog Converter) plays a key role in this process. It converts the capacitance value detected by the capacitive sensor into a digital signal. The CDAC modes supported by the hidden door handle include single CDAC mode and dual CDAC mode. In single CDAC mode, only one CDAC is used to convert digital signals into analog signals. It contains a capacitor array, each of which represents a bit. By switching the charge between these capacitors, digital to analog conversion can be achieved.
[0053] Dual CDAC mode uses two CDACs. In dual CDAC mode, there is usually a primary CDAC and an auxiliary CDAC. The primary CDAC is used to process most of the bit weight, while the auxiliary CDAC is used to process the least significant bit (LSB) part to improve the overall conversion accuracy.
[0054] These two modes use different circuit structures and working principles in the process of capacitance sensor measurement, and the factors that need to be considered are also different. The single CDAC mode has a relatively simple structure, but may not be accurate enough when dealing with small capacitance changes. Although the dual CDAC mode has a complex structure, it can provide higher conversion accuracy and better anti-interference ability, and is especially suitable for application scenarios with high precision requirements.
[0055] In single capacitor digital-to-analog converter mode, the conversion formula is:
[0056]
[0057] In dual capacitor digital-to-analog converter mode, the conversion formula is:
[0058]
[0059] in,
[0060]
[0061]
[0062] Maxcount = N sub *TxClK Div
[0063] C S =C p +C f
[0064] N sub is the number of sub-conversions, CompClk Div To compensate for the frequency division, C p is the parasitic capacitance, C f is the finger capacitance, C S is the inductive capacitance, C ref is the reference capacitor, TxClk Div is the Tx clock division rate, F Mod is the modulator clock frequency, Maxcount is the maximum number of digital counts converted from the capacitance signal, F Tx is the clock frequency, F comp To compensate for the CDAC clock frequency.
[0065] In CDAC, the number of sub-conversions N subRefers to the number of sub-conversion steps required to convert the input capacitance value into a digital signal within a complete conversion cycle. In CDAC, a complete conversion cycle is usually divided into multiple sub-conversion steps, each of which performs certain processing and conversion on the input capacitance to obtain a digital output result.
[0066] The maximum number Maxcount converted from the capacitance signal into a digital count value is used to represent the maximum count range, that is, the digital count value corresponding to the maximum capacitance value that can be represented.
[0067] Sense clock divider SnsClK Div The clock frequency used to control the sensing capacitance measurement affects the accuracy and speed of the measurement. Specifically, the sensing clock divider is usually used to divide the system clock signal to generate the sensing clock signal for measurement. By adjusting the division ratio of the sensing clock divider, the time interval of the sensing capacitance measurement can be controlled, thereby affecting the accuracy and sensitivity of the measurement.
[0068] Compensation capacitor C comp Used to correct errors or deviations that may exist in the measurement process of capacitance sensors. Compensation capacitors are usually used to adjust the sensitivity, stability and accuracy of capacitance sensors to improve the performance of the sensor.
[0069] Compensation divider CompClk Div The role here is to compensate for the capacitor C comp The compensation capacitor C can be corrected by frequency division to adjust the effect of the compensation capacitor, which in turn affects the calculation. comp The error introduced ensures accurate data in dual CDAC mode.
[0070] Reference capacitor C ref Used as a reference value to help measure changes in target capacitance. Specifically, it provides a known fixed value for calibrating and comparing changes in target capacitance. ref By comparing the capacitance of the target object with that of the touch sensor, the change in capacitance of the target object can be accurately measured, thereby realizing the detection and recognition of the external environment or touch input.
[0071] Inductive Capacitor C s is composed of parasitic capacitance C p and finger capacitance C f In capacitive sensors, the sensing capacitor is the capacitive element used to sense external touch.
[0072] Finger capacitance C f When a finger approaches the sensor surface, the capacitance around the finger will affect the capacitance value in the sensor circuit, thereby generating a signal that can be detected. The finger capacitance Cf This can be measured using a capacitive sensor on a hidden door handle.
[0073] Parasitic capacitance C p It is the unintentional capacitance in the circuit of the capacitive sensor due to structure, wiring or other factors.
[0074] Tx clock division frequency TxClk Div Indicates the frequency division ratio of the Tx clock, that is, the frequency division ratio of the Tx clock signal. In the digital-to-analog conversion process, the frequency division ratio of the Tx clock will affect the speed and accuracy of data conversion.
[0075] Modulator clock frequency F Mod Indicates the clock frequency of the modulator. In capacitive sensors, the clock frequency of the modulator is used to control the modulation and demodulation process of the signal. A suitable modulator clock frequency can ensure the stability and accuracy of signal processing.
[0076] Tx clock frequency F Tx Indicates the frequency of the Tx clock. It plays an important role in CADA and affects the speed and accuracy of data conversion. By adjusting the Tx clock frequency F Tx , the speed and accuracy of data conversion can be optimized to meet the requirements of specific applications.
[0077] Compensated CDAC clock frequency F comp It is used to control the process of compensation capacitance. The appropriate compensation CDAC clock frequency can ensure the accuracy and stability of the compensation process, thereby improving the accuracy of data conversion.
[0078] In single CDAC mode, the sensing capacitor C S With reference capacitor C ref Compare and divide the TxClk clock by the Div To determine the digital count value corresponding to the sensing capacitance. Maxcount multiplied by the sensing capacitance C S The total capacitance value without the reference capacitor is given and then divided by the clock division ratio and half of the reference capacitance to get the final raw data value.
[0079] In dual CDAC mode, the formula This is to compensate for the parasitic capacitance C in the circuit. comp This compensation term adjusts the charge of the auxiliary CDAC to offset the error caused by parasitic capacitance, thereby improving the accuracy of the conversion. Div is the compensation frequency, which is related to the modulator clock frequency F Mod and compensated CDAC clock frequency F compIt is used to adjust the operating frequency of the auxiliary CDAC to match the operating status of the main CDAC, ensuring that the two can work together to provide accurate conversion results.
[0080] S102: When the detector touches the hidden door handle multiple times, multiple groups of original touch signals are collected.
[0081] In a specific implementation scenario, multiple testers of different heights, weights or finger features are invited to participate in the experiment to cover different user groups. Each tester is asked to touch or operate the hidden door handle according to a specific process to trigger the capacitive sensor in the hidden handle, so as to obtain the original touch signal through the capacitive sensor. When the tester touches the door handle multiple times, each touch will generate a new set of original touch signals. Each set of original touch signals contains all the changes in capacitance values during the touch process. Specifically, it records the capacitance signal detected by the capacitive sensor from the start of the touch (that is, the finger first approaches the door handle) to the end of the touch (the finger completely leaves the door handle).
[0082] Specifically, during a touch process, as the finger approaches and leaves, the capacitance value around the hidden door handle will change. When the finger approaches the door handle, the capacitance of the finger will affect the electric field distribution around the capacitive sensor, causing the capacitance value to change. This change will be detected and recorded by the capacitive sensor. As the finger continues to touch the door handle, the capacitance value will change accordingly until the touch ends and the finger completely leaves the door handle, and the capacitance value returns to its initial state.
[0083] S103: converting the multiple groups of original touch signals into original touch data through a conversion formula, acquiring labeled training data based on the original touch data, inputting the training data into a detection model for training, and acquiring a pre-trained detection model.
[0084] In a specific implementation scenario, multiple groups of raw touch signals are substituted into the conversion formula to obtain raw touch data, thereby converting the capacitance value detected by the capacitive sensor into a digital signal. The collected raw touch data is preprocessed, including data cleaning, denoising, feature extraction, and other operations to facilitate subsequent model training. For example, a digital filter (such as a low-pass filter, a band-pass filter, etc.) is used to filter out noise within a specific frequency range.
[0085] In one implementation scenario, the raw touch data is annotated to obtain annotated data, for example, "1" for false touch and "0" for no false touch, for subsequent data processing and analysis. The annotations are divided into three parts: training data (about 90%), verification data (about 5%) and test data (about 5%) for training, verification and evaluation of machine learning models. Training data is a data set used to train machine learning models. Through training data, the model can learn the patterns and characteristics of touch events so that it can accurately identify and predict new touch events. Verification data is used to evaluate the performance of the model during training and help adjust the hyperparameters of the model to improve its performance. Validation data sets are typically used to monitor the generalization ability of the model and avoid overfitting during training. Test data is used to finally evaluate the performance and generalization ability of the trained model. The test data set is usually data that the model has not seen in the training and verification stages to ensure that the model can perform well on new data.
[0086] In another implementation scenario, before the training data is input into the detection model for training, all the labeled data are first centered and then standardized to obtain feature data. Centering refers to subtracting the mean of all data from each data point so that the mean of the data becomes 0. This step helps to eliminate the absolute differences between the data, making the data centered around zero, and making it easier to compare the relative differences between different data points. Through centering, the distribution of the data will be centered around the mean, which helps to reduce the deviation between the data. Standardization refers to dividing the centered data by the standard deviation of all data so that the standard deviation of the data becomes 1. This can unify the scale of the data, ensure that the values between different features or variables have similar ranges, and avoid some features from having too much impact on the model. Standardized data helps to better retain the distribution information between the data, while maintaining the distance relationship between samples, which is conducive to extracting the features of the data. Centering and standardization help retain the distribution and characteristics of the data by adjusting the mean and standard deviation of the data, thereby improving the performance and accuracy of the machine learning model.
[0087] The standardized formula is:
[0088]
[0089] X i is each centralized data, б is the standard deviation of all data, and μ is the mean of all data.
[0090] The feature data corresponding to the training / testing / validation data is input into the detection model for training / testing / validation operations.
[0091] In one implementation scenario, the detection model is a back propagation neural network, which is a feedforward neural network, including an input layer, a hidden layer, and an output layer. The number of nodes in the input layer should match the number of features in the feature data. The number of nodes in the hidden layer can be determined based on the complexity of the problem and requirements. Figure 2 , Figure 2 1 is a schematic diagram of the structure of an embodiment of a back propagation neural network provided by the present invention. Further, a loss function can be set, and the mean square error (MSE) is usually used as the loss function to measure the difference between the network prediction value and the actual value. A suitable optimizer can also be selected, such as gradient descent, Levenberg-Marquardt algorithm, etc., to adjust the network weights to minimize the loss function.
[0092] The Bayesian regularization method is selected to train the hidden layers of the back-propagation neural network. This is a weight decay method that is implemented by adding a regularization term to the loss function. This regularization term is usually a function of the sum of squared weights. By introducing a priori probabilities to constrain the size of the model parameters, it helps to control the growth of network weights and prevent overfitting. The neural network is trained using feature data corresponding to the training data, the output is calculated through forward propagation, and then the weights and biases are adjusted through the back-propagation algorithm to minimize the loss function. In each iteration, the weights are adjusted according to the requirements of Bayesian regularization so that the weights tend to be smaller values.
[0093] Furthermore, the number of training times is set. The number of training times, also known as the number of iterations or epochs, refers to the number of times the entire training data is forward propagated and backward propagated in the neural network. For example, the number of training times is set to 200 times, which means that the entire training data will be completely traversed 200 times. After each 200 training iterations are completed, the model is evaluated using the test data. By testing the detection model at the end of each training cycle, the performance of the detection model on the training data and the test data can be monitored to ensure that the detection model not only performs well on the training data, but also maintains stable performance on unseen data. Specifically, by evaluating the model using the test set at the end of each training, it can be observed whether the model begins to show signs of overfitting, such as the performance on the training data continues to improve, while the performance on the test data begins to decline.
[0094] Train until the detection model converges. Model convergence means that the loss function of the detection model reaches a stable minimum value, or the performance of the detection model no longer improves significantly with the increase in the number of training times. When the model converges, it means that the model has found a good weight and bias setting, which can perform well on the training data and maintain stable performance on the test data.
[0095] Furthermore, after the detection model is trained, the validation data is used to evaluate the generalization ability of the detection model, that is, the performance of the detection model on unseen data. By comparing the performance of the detection model on the validation data, the recognition accuracy of the detection model can be evaluated, that is, the model's ability to classify or predict unknown data.
[0096] In other implementation scenarios, cross-validation can be used. Cross-validation is a technique for evaluating the generalization ability of a model. It divides the labeled data into multiple small subsets, and then uses one of the subsets as validation data and the rest as training data to train and evaluate the detection model. This method ensures that every data point is used for training and validation, thereby improving the reliability of the detection model evaluation. Through cross-validation and parameter tuning, the detection model can be effectively prevented from overfitting the training set data and improving the generalization ability of the detection model on unknown data. Ensure that the detection model performs well on the training data and validation data, and avoid the situation where it performs well on the training data but poorly on the validation data.
[0097] S104: After the pre-trained detection model is deployed, when a change in the real-time capacitance signal of the capacitance sensor is detected, the changed capacitance signal within a preset time period is collected, the changed capacitance signal is converted into a changed touch signal through a conversion formula, the changed touch signal is input into the pre-trained detection model, the detection result is obtained, and the pop-up and retraction of the hidden door handle is controlled according to the detection result.
[0098] In a specific implementation scenario, the trained detection model is converted into a format suitable for running on a microcontroller by compression, quantization or conversion into a format suitable for target hardware, and the converted detection model is deployed to the target microcontroller to perform touch detection in actual applications.
[0099] The capacitive sensor on the hidden door handle is continuously monitored to detect whether there is a change in the capacitive signal. When the capacitive signal is detected, the change in the capacitance value is recorded. The corresponding change capacitance signal is generated. The time length of the change capacitance signal is determined according to the preset time parameter. This preset time can be set according to the needs of the actual application, for example, it can be from a few hundred milliseconds to a few seconds. When the preset time is reached, the storage process ends. This means that the change in capacitance value is stopped and all the capacitance signal data collected during this period is retained, which is the change capacitance signal.
[0100] After substituting the changing capacitance signal into the above conversion formula, it is converted into a changing touch signal, and the changing touch signal is input into the pre-trained detection model. The detection model will analyze the input capacitance signal data and determine whether the touch behavior may be a false trigger by identifying the characteristics of the signal. The detection model will determine whether the touch behavior is a false trigger based on the learned patterns, such as the duration and intensity of the touch, the rate of change of the signal, etc. The detection model will output the recognition result, which is a probability value or a classification label, indicating the possibility that the detected touch behavior is a false trigger.
[0101] Once a non-touch situation is confirmed, the corresponding operation will be performed on the hidden door handle. If the hidden door handle is hidden, it will be popped out so that the user can grab and operate it. If the hidden door handle is deployed, it will be retracted to keep the vehicle's appearance neat. If it is determined to be a false touch, that is, the user accidentally touches it or it is recognized as a touch due to high humidity, rain, etc., no operation will be performed and the hidden door handle will remain in its current state, whether hidden or deployed.
[0102] From the above description, it can be seen that in the present embodiment, when the detector touches the hidden door handle multiple times, multiple groups of original touch signals are collected, and the original touch signals include the capacitance signal collected by the capacitance sensor from the beginning of the touch to the touching of the hidden handle, so as to train the detection model, and then based on the pre-trained detection model, detect whether the currently detected touch event is a false touch event, and judge the parameter changes in the process from the human hand approaching the door handle to touching the door handle, rather than just relying on a single capacitance threshold for judgment, so as to achieve accurate judgment of touch behavior and improve the recognition accuracy and robustness of the model.
[0103] See also Figure 3 , Figure 3 1 is a schematic diagram of a structure of an embodiment of a hidden door handle anti-inadvertent touch system provided by the present invention. The hidden door handle anti-inadvertent touch system 10 includes a formula module 11, a collection module 12, a training module 13, and a detection module 14.
[0104] The formula module 11 is used to obtain the capacitance digital-to-analog converter mode supported by the capacitance sensor, and obtain the conversion formula according to the capacitance digital-to-analog converter mode. The capacitance digital-to-analog converter mode includes a single capacitance digital-to-analog converter mode or a dual capacitance digital-to-analog converter mode; the acquisition module 12 is used to collect multiple groups of original touch signals when the detector touches the hidden door handle multiple times, and each group of original touch signals includes the capacitance signal collected by the capacitance sensor from the beginning of the touch to the touching of the hidden door handle each time the hidden door handle is touched; the training module 13 is used to convert the multiple groups of original touch signals into original touch data through the conversion formula, obtain labeled training data based on the original touch data, input the training data into the detection model for training, and obtain a pre-trained detection model; the detection module 14 is used to deploy the pre-trained detection model, and when the real-time capacitance signal of the capacitance sensor is detected to change, collect the changed capacitance signal within a preset time length, convert the changed capacitance signal into a changed touch signal through the conversion formula, input the changed touch signal into the pre-trained detection model, obtain the detection result, and control the pop-up and retraction of the hidden door handle according to the detection result.
[0105] In single capacitor digital-to-analog converter mode, the conversion formula is:
[0106]
[0107] In dual capacitor digital-to-analog converter mode, the conversion formula is:
[0108]
[0109] in,
[0110]
[0111]
[0112] Maxcount = N sub *TxClK Div
[0113]
[0114] N sub is the number of sub-conversions, CompClk Div To compensate for the frequency division, C p is the parasitic capacitance, C f is the finger capacitance, C S is the inductive capacitance, C ref is the reference capacitor, TxClk Div is the Tx clock division rate, F Mod is the modulator clock frequency, Maxcount is the maximum number of digital counts converted from the capacitance signal, FTx is the clock frequency, F comp To compensate for the CDAC clock frequency.
[0115] The training module 13 is further used to annotate the original touch data to obtain annotated data, select a part of the annotated data as training data, a part as verification data, and the remaining part as test data.
[0116] The training module 13 is also used to perform centralization and standardization operations on all labeled data to obtain feature data; and input the feature data corresponding to the training data into the detection model for training.
[0117] The detection model is a back propagation neural network, and the training module 13 is also used to select a Bayesian regularization method to train the hidden layer of the back propagation neural network.
[0118] The training module 13 is also used to set the number of training times. When the hidden layer training of the back propagation neural network reaches the number of training times, the back propagation neural network is tested through test data to prevent the detection model from overfitting.
[0119] The training module 13 is also used to check the generalization ability of the detection model using the verification data.
[0120] From the above description, it can be seen that in the present embodiment, when the detector touches the hidden door handle multiple times, multiple groups of original touch signals are collected, and the original touch signals include the capacitance signal collected by the capacitance sensor from the beginning of the touch to the touching of the hidden handle, so as to train the detection model, and then based on the pre-trained detection model, detect whether the currently detected touch event is a false touch event, and judge the parameter changes in the process from the human hand approaching the door handle to touching the door handle, rather than just relying on a single capacitance threshold for judgment, so as to achieve accurate judgment of touch behavior and improve recognition accuracy and robustness.
[0121] See also Figure 4 , Figure 4 2 is a schematic diagram of the structure of an embodiment of the smart car provided by the present invention. The smart car 20 includes a processor 21 and a memory 22. The processor 21 is coupled to the memory 22. The memory 222 stores a computer program, and the processor 1 executes the computer program when working to implement the above method. The detailed steps can be referred to above and will not be repeated here.
[0122] See also Figure 5 , Figure 53 is a schematic diagram of the structure of an embodiment of the computer-readable storage medium 30 provided by the present invention. The computer-readable storage medium 30 stores at least one computer program 31, which is used to be executed by the processor to implement the above method. The detailed steps can be referred to above and will not be repeated here. In one embodiment, the computer-readable storage medium can be a storage chip, a hard disk, or a mobile hard disk or a USB flash drive, a CD, or other readable and writable storage tools in the terminal, or a server, etc.
[0123] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0124] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0125] It should be understood that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. For those skilled in the art, the technical solutions described in the above embodiments can be modified, or some of the technical features therein can be replaced by equivalents; and all these modifications and replacements should fall within the scope of protection of the claims attached to the present invention.
Claims
1. A method for preventing accidental touch of a hidden door handle, characterized in that: Applied to a hidden door handle, the hidden door handle includes a capacitive sensor, and the hidden door handle anti-mistouch method includes: Acquire a capacitance digital-to-analog converter mode supported by the capacitance sensor, and acquire a conversion formula according to the capacitance digital-to-analog converter mode, wherein the capacitance digital-to-analog converter mode includes a single capacitance digital-to-analog converter mode or a dual capacitance digital-to-analog converter mode; When the examiner touches the hidden door handle multiple times, multiple groups of original touch signals are collected, each group of the original touch signals includes the capacitance signal collected by the capacitance sensor during the period from the beginning of the touch to the touching of the hidden door handle each time the examiner touches the hidden door handle; Converting the multiple groups of original touch signals into original touch data by using the conversion formula, acquiring labeled training data based on the original touch data, inputting the training data into a detection model for training, and acquiring a pre-trained detection model; After the pre-trained detection model is deployed, when a change in the real-time capacitance signal of the capacitance sensor is detected, the changed capacitance signal within a preset time length is collected, the changed capacitance signal is converted into a changed touch signal through the conversion formula, the changed touch signal is input into the pre-trained detection model, the detection result is obtained, and the pop-up and retraction of the hidden door handle are controlled according to the detection result.
2. The method for preventing accidental touch of a hidden door handle according to claim 1, characterized in that: The step of acquiring a conversion formula according to the capacitance digital-to-analog converter mode comprises: In single capacitor digital-to-analog converter mode, the conversion formula is: In the dual capacitor digital-to-analog converter mode, the conversion formula is: in, Maxcount=N sub *TxClK Div C S =C p +C f N sub is the number of sub-conversions, CompClk Div To compensate for the frequency division, C p is the parasitic capacitance, C f is the finger capacitance, C S is the inductive capacitance, C ref is the reference capacitor, TxClk Div is the Tx clock division rate, F Mod is the modulator clock frequency, Maxcount is the maximum number of digital count values converted from the capacitance signal, F Tx is the clock frequency, F comp To compensate for the CDAC clock frequency.
3. The method for preventing accidental touch of a hidden door handle according to claim 1, characterized in that: The step of acquiring labeled training data based on the original touch data comprises: The original touch data is annotated to obtain annotated data, and a portion of the annotated data is selected as the training data, a portion as the verification data, and the remaining portion as the test data.
4. The method for preventing accidental touch of a hidden door handle according to claim 3, characterized in that: The step of inputting the training data into the detection model for training comprises: Centralizing and standardizing all the labeled data to obtain feature data; The feature data corresponding to the training data is input into the detection model for training.
5. The method for preventing accidental touch of a hidden door handle according to claim 3, characterized in that: The detection model is a back propagation neural network, and the step of inputting the training data into the detection model for training comprises: A Bayesian regularization method is selected to train the hidden layer of the detection model.
6. The method for preventing accidental touch of a hidden door handle according to claim 5, characterized in that: The step of selecting the Bayesian regularization method to train the hidden layer of the back propagation neural network includes: The number of training times is set, and when the hidden layer training of the detection model reaches the number of training times, the detection model is tested by the test data to prevent the detection model from overfitting.
7. The method for preventing accidental touch of a hidden door handle according to claim 6, characterized in that: After the step of selecting the Bayesian regularization method to train the hidden layer of the detection model, the method further comprises: The generalization ability of the detection model is checked using the validation data.
8. A hidden door handle anti-accidental touch system, characterized in that: Applied to a hidden door handle, the hidden door handle includes a capacitive sensor, and the anti-mistouch system of the hidden door handle includes: A formula module, used to obtain a capacitance digital-to-analog converter mode supported by the capacitance sensor, and obtain a conversion formula according to the capacitance digital-to-analog converter mode, wherein the capacitance digital-to-analog converter mode includes a single capacitance digital-to-analog converter mode or a dual capacitance digital-to-analog converter mode; An acquisition module, used for acquiring multiple groups of original touch signals when the detector touches the hidden door handle multiple times, each group of the original touch signals includes a capacitance signal acquired by the capacitance sensor during a period from the beginning of the touch to the touching of the hidden door handle each time the hidden door handle is touched; A training module, used for converting the multiple groups of original touch signals into original touch data through the conversion formula, acquiring labeled training data based on the original touch data, inputting the training data into a detection model for training, and acquiring a pre-trained detection model; The detection module is used to deploy the pre-trained detection model, and when a change in the real-time capacitance signal of the capacitance sensor is detected, collect the changed capacitance signal within a preset time length, convert the changed capacitance signal into a changed touch signal through the conversion formula, input the changed touch signal into the pre-trained detection model, obtain the detection result, and control the pop-up and retraction of the hidden door handle according to the detection result.
9. A smart car, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 7.