Detection method and system of hidden door handle, intelligent automobile and storage medium
By collecting touch capacitance data from multiple testers and training the detection model, the problem of inaccurate sensing of hidden door handles is solved, personalized touch recognition is achieved, and user experience is improved.
Patent Information
- Application Number
- CN202411698220.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-25
- Publication Date
- 2025-03-14
AI Technical Summary
In the prior art, capacitive sensing is inaccurate, resulting in the problem that hidden door handles are accidentally triggered or difficult to trigger between different individuals.
By collecting touch capacitance data from multiple testers, generating labeled training data, training the detection model, obtaining pre-trained detection model, combining the capacitance digital analog converter mode and the fully connected neural network model, touch detection of the current user is realized.
It realizes learning the differences in touch capacitance data of different body shapes and finger shapes, improves the accuracy and adaptability of personalized touch recognition, and improves the user experience.
Smart Images
Figure CN119616318A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of automobile hidden door handles, in particular to a detection method and system for hidden door handles, a smart automobile and a storage medium. Background Art
[0002] In the field of modern car design, the combination of hidden door handles and capacitive sensing technology has become an innovative trend. This design not only optimizes the appearance of the vehicle, but also improves user experience and safety through intelligent sensing technology. Hidden door handles use capacitive sensing technology to automatically pop out / retract the door handle or open the door with a light touch when the owner or passenger touches it.
[0003] Capacitive sensing technology relies on a preset threshold to determine the user's touch intention to trigger the door handle to pop out or retract. This mechanism may cause false triggering due to differences in capacitance between individuals. Since people of different heights and weights have different finger capacitances, a single threshold setting may not be suitable for everyone, causing some users to have difficulty triggering or be too sensitive. Summary of the invention
[0004] The technical problem to be solved by the embodiments of the present invention is to provide a method for detecting a hidden door handle to solve the problem of inaccurate capacitive sensing in the prior art.
[0005] The present invention discloses a method for detecting a hidden door handle, and the method for detecting a hidden door handle comprises:
[0006] Obtaining a capacitance digital-to-analog converter mode supported by the hidden door handle, and obtaining a 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;
[0007] Collecting touch capacitance data of multiple testers, and substituting the touch capacitance data into the conversion formula to obtain original data;
[0008] Generate labeled training data according to the original data, train the detection model through the training data, and obtain a pre-trained detection model;
[0009] Collect current capacitance data of the current user, substitute the current capacitance data into the conversion formula to obtain current data, input the current data into the pre-trained detection model to obtain detection results, and control the pop-up and closing of the hidden door handle according to the detection results.
[0010] Optionally, after the step of generating raw data according to the number of the capacitive sensors and the touch capacitance data, the method further comprises:
[0011] Obtaining a capacitance reference value when the hidden door handle is not touched, substituting the capacitance reference value into the conversion formula to generate reference data, and obtaining a training difference between the original data and the reference data;
[0012] The step of labeling the original data to generate training data includes:
[0013] The training differences are labeled to obtain labeled data, a portion of the labeled data is selected as the training data, and the remaining portion is used as verification data.
[0014] Optionally, the step of inputting the current data into the pre-trained detection model comprises:
[0015] A difference between the current data and the reference data is obtained as a current difference, and the current difference is input into the pre-trained detection model.
[0016] Optionally, the step of acquiring a conversion formula according to the capacitance digital-to-analog converter mode includes:
[0017] In single capacitor digital-to-analog converter mode, the conversion formula is:
[0018]
[0019] In the dual capacitor digital-to-analog converter mode, the conversion formula is:
[0020]
[0021] in,
[0022] Maxcount = N sub *SnsClK Div
[0023] C s =C p +C f
[0024] N sub is the number of sub-conversions, SnsClK Div is the sensing clock divider, C p is the parasitic capacitance, C f is the finger capacitance, C ref is the reference capacitor, CompClk Div To compensate the divider, C comp For compensation capacitor.
[0025] Optionally, the detection model is a 5-layer fully connected neural network;
[0026] Among them, the input layer has 3 nodes and the hidden layer has 3 layers; the first hidden layer has 16 nodes and is activated by the Relu function; the second hidden layer has 16 nodes and is activated by the Linear function; the third hidden layer has 16 nodes and is activated by the Relu function; the output layer has 5 nodes and is activated by the Sigmoid function.
[0027] Optionally, the Relu function formula is:
[0028] Relu(x)=max(0,x)
[0029] The Linear function formula is:
[0030] Linear(x)=XA T +b
[0031] The Sigmoid function formula is:
[0032]
[0033] Optionally, the loss function of the fully connected neural network is:
[0034]
[0035] The present invention also discloses a detection system for a hidden door handle, the detection system for a hidden door handle comprising:
[0036] A formula module, used to obtain a capacitance digital-to-analog converter mode supported by the hidden door handle, and obtain a 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;
[0037] An acquisition module, used for acquiring touch capacitance data of multiple testers, and substituting the touch capacitance data into the conversion formula to obtain original data;
[0038] A training module, used to generate labeled training data according to the original data, train the detection model through the training data, and obtain a pre-trained detection model;
[0039] The detection module is used to collect the current capacitance data of the current user, substitute the current capacitance data into the conversion formula to obtain the current data, input the current data into the pre-trained detection model, obtain the detection result, and control the pop-up and closing of the hidden door handle according to the detection result.
[0040] 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.
[0041] 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.
[0042] Compared with the prior art, the beneficial effect of the hidden door handle detection method provided by the embodiment of the present invention is that by collecting touch capacitance data of multiple testers for training the detection model, and then detecting whether a touch operation is currently occurring based on the pre-trained detection model, it is possible to learn the differences in touch capacitance data caused by factors such as different body postures and finger shapes, thereby realizing personalized touch recognition. Compared with the fixed threshold detection method, it has higher adaptability, personalized recognition capability, and data-driven accuracy, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] The scheme of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments, in which:
[0044] Figure 1 It is a flow chart of an embodiment of a method for detecting a hidden door handle provided by the present invention;
[0045] Figure 2 It is a structural schematic diagram of the detection model provided by the present invention;
[0046] Figure 3 It is a structural schematic diagram of an embodiment of a detection system for a hidden door handle provided by the present invention;
[0047] Figure 4 It is a structural schematic diagram of an embodiment of a smart car provided by the present invention;
[0048] Figure 5 It is a structural schematic diagram of an embodiment of a computer-readable storage medium provided by the present invention.
[0049] The reference numerals in the figures are:
[0050] 10. Hidden door handle detection system; 11. Formula module; 12. Acquisition module; 13. Training module; 14. Detection module; 20. Smart car; 21. Processor; 22. Memory; 30. Computer-readable storage medium; 31. Computer program. DETAILED DESCRIPTION
[0051] 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.
[0052] See also Figure 1 , Figure 1 1 is a flow chart of an embodiment of a method for detecting a hidden door handle provided by the present invention. The method for detecting a hidden door handle comprises the following steps:
[0053] S101: Obtain a capacitance digital-to-analog converter mode supported by the hidden door handle, and obtain a conversion formula according to the capacitance digital-to-analog converter mode.
[0054] In a specific implementation scenario, a capacitive digital-to-analog converter mode supported by the hidden door handle is obtained, and the capacitive digital-to-analog converter is used to convert the capacitance value of the touch capacitance data into a digital signal. The capacitive digital-to-analog converter modes supported by the hidden door handle include a single CDAC (Capacitive Digital-to-Analog Converter) mode or a dual CDAC mode. In the single CDAC mode, only one capacitive digital-to-analog converter is used to convert the digital signal into an analog signal. It contains a capacitor array, each of which represents a bit. By switching charges between these capacitors, digital-to-analog conversion can be achieved. The dual CDAC mode uses two capacitive digital-to-analog converters. In the dual CDAC mode, there is usually a main CDAC and an auxiliary CDAC. The main CDAC is used to process most of the bit weights, while the auxiliary CDAC is used to process the least significant bit (LSB) part to improve the overall conversion accuracy. They use different circuit structures and working principles in the measurement process of the capacitive sensor, and the factors that need to be considered are also different.
[0055] Specifically, 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] Maxcount = N sub *SnsClK Div
[0060] C s =C p +C f
[0061] Nsub is the number of sub-conversions, SnsClK Div is the sensing clock divider, C p is the parasitic capacitance, C f is the finger capacitance, C ref is the reference capacitor, CompClk Div To compensate the divider, C comp For compensation capacitor.
[0062] In CDAC, the number of sub-conversions N sub Refers 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.
[0063] 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.
[0064] Parasitic capacitance C p It is the unintentional capacitance in the circuit of the capacitive sensor due to structure, wiring or other factors.
[0065] 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.
[0066] Compensation divider CompClk Div The role here is to compensate for the capacitor C comp The frequency division process is performed to adjust the influence of the compensation capacitor, which in turn affects the calculation. The compensation capacitor C can be corrected by the frequency division operation of the compensation divider. comp The error introduced ensures accurate data in dual CDAC mode.
[0067] 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.
[0068] 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 C f The capacitance data can be obtained by measuring the capacitance sensor of the hidden door handle, and can be any one of touch capacitance data, current capacitance data, and capacitance reference value.
[0069] In single CDAC mode, only one capacitance digital-to-analog converter (CDAC) is used to make the measurement. The formula for single CDAC mode takes into account the total capacitance, C s With the induction clock divider SnsClK Div and reference capacitor C ref In dual CDAC mode, two CDACs are usually used, one for measuring the total capacitance C s , and the other is used for compensation capacitor C comp , it is necessary to consider the difference between the total capacitance and the compensation capacitance, as well as the effect of the compensation capacitance on the total capacitance.
[0070] S102: collecting touch capacitance data of multiple testers, and substituting the touch capacitance data into a conversion formula to obtain original data.
[0071] 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 and record the touch capacitance data. The touch capacitance data is substituted into the conversion formula to obtain the raw data, thereby converting the capacitance value of the touch capacitance data into a digital signal.
[0072] S103: Generate labeled training data according to the original data, train the detection model through the training data, and obtain a pre-trained detection model.
[0073] In a specific embodiment, the collected raw 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. When generating training data using labeled data, it is usually necessary to convert the data into a format acceptable to the model, such as a feature vector.
[0074] In other implementation scenarios, when the hidden door handle is started or no touch is detected for a long time, the capacitance value of the capacitive sensor is recorded as the capacitance reference value. Substitute the capacitance reference value into the conversion formula to obtain the reference data. For each new raw data, calculate the training between it and the baseline data. This training difference reflects the change between the capacitance value of the capacitive sensor in the hidden door handle when the tester touches the hidden door handle and the capacitance value when there is no touch in the environment, which can help eliminate the influence of environmental factors on the capacitance value.
[0075] The training difference is labeled to obtain labeled data, with "1" being used to indicate a touch event and "0" being used to indicate a non-touch event, providing accurate labels for the training model so that the model can learn how to identify touch events. All labeled data is divided into a training set and a test set. For example, 95% of the data can be selected as a training set for training the model, and the remaining 5% of the data can be used as a test set to evaluate the performance and generalization ability of the model.
[0076] Furthermore, in order to help the detection model better understand the characteristics of the data and improve the performance and generalization ability of the model, all training data are normalized before training the detection model so that all data are distributed in the range of 0 to 1. This can eliminate the dimensional influence between different features and make the model training more stable and converge faster.
[0077] The minimum-maximum normalization method can be used to linearly scale the data to a specific range and find the maximum value X in the training data. max and minimum value X min , for each training data X i , and obtain its corresponding normalized value according to the following formula:
[0078]
[0079] The normalized training data is input into the detection model for training. Specifically, the training rounds and data batches can be set first. The training rounds are the number of times the neural network iterates learning on the entire training set, and the data batches are the number of samples used for each model update. A larger batch size can speed up training but may take up more memory. In general, the choice of batch size depends on the size of the dataset and the computing resources.
[0080] The training data is then input into the detection model for training, and the training results are evaluated with the validation data. Multiple rounds of training are performed based on the training data, and the weights and biases of the monitoring model are continuously adjusted through the back propagation algorithm so that the model can better fit the data set. During the training process, the parameters and status of the detection model are saved regularly to prevent accidental data loss. Saving the detection model can be achieved by saving the model's weights, structure, and optimizer status to a file.
[0081] The trained detection model is converted into a format suitable for running on the microcontroller by compression, quantization or conversion into a format suitable for the target hardware, and the converted detection model is deployed to the target microcontroller for touch detection in actual applications.
[0082] S104: Collect the current capacitance data of the current user, substitute the current capacitance data into the conversion formula to obtain the current data, input the current data into the pre-trained detection model, obtain the detection result, and control the pop-up and closing of the hidden door handle according to the detection result.
[0083] In a specific implementation scenario, the current capacitance data is continuously detected by the capacitance sensor. When the current user touches the hidden door handle, the current capacitance data of the current user is collected, and the current capacitance data is substituted into the corresponding conversion formula to obtain the current data. The difference between the current data and the reference data is used as the current difference, and the current difference is input into the pre-trained detection model for calculation to obtain the detection result output by the model. The detection result is a probability distribution, which represents the possibility of each category (touch or no touch), and the category with the largest probability in the probability distribution is selected as the prediction result.
[0084] If the neural network model predicts "touch" and detects that the hidden door handle is in the pop-up state (open state), the operation of closing the door handle is performed. If the neural network model predicts "touch" but detects that the hidden door handle is in the closed state, the operation of popping out the door handle is performed. If the neural network model predicts "not touched", no operation is performed and the current state of the door handle remains unchanged.
[0085] From the above description, it can be seen that in this embodiment, by collecting touch capacitance data of multiple testers for training the detection model, and then detecting whether a touch operation is currently occurring based on the pre-trained detection model, it is possible to learn the differences in touch capacitance data caused by factors such as different body postures and finger shapes, thereby achieving personalized touch recognition. Compared with the fixed threshold detection method, it has higher adaptability, personalized recognition capability, and data-driven accuracy, thereby improving the user experience.
[0086] Please refer to Figure 2 , Figure 2 It is a structural schematic diagram of the detection model provided by the present invention. The detection model is a 5-layer fully connected neural network, including an input layer, three hidden layers and an output layer. The input layer has 3 nodes for receiving data as the input of the neural network, and the first hidden layer has 16 nodes, which are activated by the ReLU (Rectified Linear Unit) function. The function of the ReLU function is to change all negative values to 0 and keep the positive values unchanged. The second hidden layer has 16 nodes and is activated by the Linear function. The Linear function is a linear function used to perform linear transformation on data. The third hidden layer has 16 nodes, which are activated again by the ReLU function. The output layer has 5 nodes and is activated by the Sigmoid function. The Sigmoid function compresses the output value to between 0 and 1, and is used to output probability values in binary classification problems.
[0087] The Relu function formula is:
[0088] Relu(x)=max(0,x)
[0089] For input x, if x is greater than 0, then output x, otherwise output 0.
[0090] The linear function formula is:
[0091] Linear(x)=XA T +b
[0092] This is a linear function where X is the input, A is the weight matrix, and b is the bias term.
[0093] The Sigmoid function formula is:
[0094]
[0095] The input value is mapped to between 0 and 1 and used in the output layer to make probability predictions for binary classification problems.
[0096] The loss function of a fully connected neural network is:
[0097]
[0098] The loss function uses the cross entropy loss function to measure the difference between the model output and the actual label, y i is the actual label, x i is the predicted value of the model output, and n is the number of samples. The cross entropy loss function is commonly used in binary classification problems.
[0099] Through the nonlinear feature extraction of multiple hidden layers, the fully connected neural network can better capture the complex patterns and associations in the touch signal and improve the accuracy of detection. This application uses a detection model for detection, which can learn and adapt to the touch data of different testers, realize personalized touch recognition, thereby improving the flexibility and generalization ability of detection, and learning the characteristics of touch signals based on a large amount of data, thereby improving the accuracy and stability of detection.
[0100] See also Figure 3 , Figure 3 1 is a schematic diagram of a structure of an embodiment of a detection system for a hidden door handle provided by the present invention. The detection system 10 for a hidden door handle comprises: a formula module 11, a collection module 12, a training module 13, and a detection module 14.
[0101] The formula module 11 is used to obtain the capacitance digital-to-analog converter mode supported by the hidden door handle, 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 touch capacitance data of multiple testers, and substitute the touch capacitance data into the conversion formula to obtain the original data; the training module 13 is used to annotate the original data to generate training data, train the detection model through the training data, and obtain a pre-trained detection model; the detection module 14 is used to collect the current capacitance data of the current user, substitute the current capacitance data into the conversion formula to obtain the current data, input the current data into the pre-trained detection model, obtain the detection result, and control the pop-up and closing of the hidden door handle according to the detection result.
[0102] The acquisition module 12 is also used to obtain a capacitance reference value when the hidden door handle is not touched, substitute the capacitance reference value into a conversion formula to generate reference data, and obtain a training difference between the original data and the reference data.
[0103] The training module 13 is also used to label the training differences to obtain labeled data, select a part of the labeled data as training data, and use the remaining part as verification data.
[0104] The detection module 14 is also used to obtain the difference between the current data and the reference data as the current difference, and input the current difference into the pre-trained detection model.
[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] Maxcount = N sub *SnsClK Div
[0111] C s =C p +C f
[0112] N sub is the sub-conversion number, SnsClK Div is the sensing clock divider, C p is the parasitic capacitance, C f is the finger capacitance, C ref is the reference capacitor, CompClk Div To compensate the divider, C comp For compensation capacitor.
[0113] Among them, the detection model is a 5-layer fully connected neural network; the input layer has 3 nodes and the hidden layer has 3 layers; the first hidden layer has 16 nodes and is activated by the Relu function; the second hidden layer has 16 nodes and is activated by the Linear function; the third hidden layer has 16 nodes and is activated by the Relu function; the output layer has 5 nodes and is activated by the Sigmoid function.
[0114] Among them, the Relu function formula is:
[0115] Relu(x)=max(0,x)
[0116] The linear function formula is:
[0117] Linear(x)=XA T +b
[0118] The Sigmoid function formula is:
[0119]
[0120] Among them, the loss function of the fully connected neural network is:
[0121]
[0122] From the above description, it can be seen that in this embodiment, by collecting touch capacitance data of multiple testers for training the detection model, and then detecting whether a touch operation is currently occurring based on the pre-trained detection model, it is possible to learn the differences in touch capacitance data caused by factors such as different body postures and finger shapes, thereby achieving personalized touch recognition. Compared with the fixed threshold detection method, it has higher adaptability, personalized recognition capability, and data-driven accuracy, thereby improving the user experience.
[0123] 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.
[0124] See also Figure 5 , Figure 5 3 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.
[0125] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods 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 this 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).
[0126] 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.
[0127] 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 detecting a hidden door handle, characterized in that: The detection method of the hidden door handle comprises: Obtaining a capacitance digital-to-analog converter mode supported by the hidden door handle, and obtaining a 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; Collecting touch capacitance data of multiple testers, and substituting the touch capacitance data into the conversion formula to obtain original data; Generate labeled training data according to the original data, train the detection model through the training data, and obtain a pre-trained detection model; Collect current capacitance data of the current user, substitute the current capacitance data into the conversion formula to obtain current data, input the current data into the pre-trained detection model to obtain detection results, and control the pop-up and closing of the hidden door handle according to the detection results.
2. The method for detecting a hidden door handle according to claim 1, characterized in that: The step of generating labeled training data according to the original data comprises: Obtaining a capacitance reference value when the hidden door handle is not touched, substituting the capacitance reference value into the conversion formula to generate reference data, and obtaining a training difference between the original data and the reference data; The training differences are labeled to obtain labeled data, a portion of the labeled data is selected as the training data, and the remaining portion is used as verification data.
3. The method for detecting a hidden door handle according to claim 2, characterized in that: The step of inputting the current data into the pre-trained detection model comprises: A difference between the current data and the reference data is obtained as a current difference, and the current difference is input into the pre-trained detection model.
4. The method for detecting a hidden door handle according to any one of claims 1 to 3, 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 *SnsClK Div C s =C p +C f N sub is the sub-conversion number, SnsClK Div is the sensing clock divider, C p is the parasitic capacitance, C f is the finger capacitance, C ref is the reference capacitor, CompClk Div To compensate the divider, C comp For compensation capacitor.
5. The method for detecting a hidden door handle according to claim 1, characterized in that: The detection model is a 5-layer fully connected neural network; Among them, the input layer has 3 nodes and the hidden layer has 3 layers; the first hidden layer has 16 nodes and is activated by the Relu function; the second hidden layer has 16 nodes and is activated by the Linear function; the third hidden layer has 16 nodes and is activated by the Relu function; the output layer has 5 nodes and is activated by the Sigmoid function.
6. The method for detecting a hidden door handle according to claim 5, characterized in that: The Relu function formula is: Relu(x)=max(0,x) The Linear function formula is: Linear(x)=XA T +b The Sigmoid function formula is:
7. The method for detecting a hidden door handle according to claim 5, characterized in that: The loss function of the fully connected neural network is:
8. A detection system for a hidden door handle, characterized in that: The detection system of the hidden door handle comprises: A formula module, used to obtain a capacitance digital-to-analog converter mode supported by the hidden door handle, and obtain a 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; An acquisition module, used for acquiring touch capacitance data of multiple testers, and substituting the touch capacitance data into the conversion formula to obtain original data; A training module, used to generate labeled training data according to the original data, train the detection model through the training data, and obtain a pre-trained detection model; The detection module is used to collect the current capacitance data of the current user, substitute the current capacitance data into the conversion formula to obtain the current data, input the current data into the pre-trained detection model, obtain the detection result, and control the pop-up and closing 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.