A Cover Recognition Method and Detection Device Based on Multi-Frequency Excitation

Abstract

The present invention discloses a covering recognition method and a detection device based on multi-frequency excitation, belonging to the technical fields of electronic technology and sensor technology. The covering recognition method includes: constructing an equivalent circuit of a sensitive capacitor, where the equivalent circuit includes a series-connected equivalent resistor, equivalent inductor, and equivalent capacitor; at different temperatures, exciting the sensitive capacitor with excitation signals of different frequencies, and obtaining the resistance value of the equivalent resistor, the capacitance value of the equivalent capacitor, and the covering category on the sensitive capacitor; taking the same set of temperature values, frequency values, resistance values, and capacitance values as sample data, and taking the corresponding covering category as the true label to construct a sample set; constructing a covering recognition model based on a neural network, and training the covering recognition model through the sample set; realizing covering recognition through the trained covering recognition model; the present invention can accurately and reliably identify the covering.

Description

Technical Field

[0001] The present invention relates to the fields of electronic technology and sensor technology, and particularly to a covering recognition method and a detection device based on multi-frequency excitation. Background Art

[0002] In the field of road surface covering detection, traditional detection methods mainly distinguish different types of road surface coverings, such as ice, water, snow, and dry road surfaces, by the capacitance values reflected by the charge and discharge waveforms of sensitive capacitors. However, although the charge and discharge waveforms of sensitive capacitors are widely used to distinguish different road surface coverings, in some cases, such as when there is similarity between snow and dry road surfaces, there may be a problem of reduced discrimination accuracy. In some cases between ice and snow and dry road surfaces, the similarity of the charge and discharge waveforms of sensitive capacitors may lead to detection errors. Such errors may be caused by certain specific weather or road conditions, making the charge and discharge waveforms unable to clearly distinguish these two road surface coverings. This may lead to the uncertainty and unreliability of road surface covering detection, thus affecting road safety and the reliability of vehicle driving. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a covering recognition method and a detection device based on multi-frequency excitation to improve the accuracy and reliability of road surface covering detection.

[0004] To achieve the above purpose, the present invention is implemented by the following technical solutions:

[0005] In the first aspect, the present invention provides a covering recognition method based on multi-frequency excitation, including:

[0006] Construct an equivalent circuit of a sensitive capacitor, where the equivalent circuit includes a series connection of an equivalent resistance, an equivalent inductance, and an equivalent capacitance;

[0007] At different temperatures, excite the sensitive capacitor with excitation signals of different frequencies, and obtain the resistance value of the equivalent resistance, the capacitance value of the equivalent capacitance, and the covering category on the sensitive capacitor;

[0008] Use the same set of temperature values, frequency values, resistance values, and capacitance values as sample data, and use the corresponding covering category as the true label to construct a sample set;

[0009] Construct a covering recognition model based on a neural network, and train the covering recognition model through the sample set;

[0010] Implement covering recognition through the trained covering recognition model.

[0011] Optionally, the covering recognition model includes a first convolutional layer, a second convolutional layer, a flattening layer, a first fully-connected layer, and a second fully-connected layer connected in sequence;

[0012] The first convolutional layer uses a convolutional kernel of size 2, has 1 input channel, 8 output channels, a stride of 1, and a padding of 0; the first convolutional layer is used to process the input data and extract local features;

[0013] The second convolutional layer uses a convolutional kernel of size 2, has 8 input channels, 16 output channels, a stride of 1, and a padding of 0; the second convolutional layer is used to process the local features and extract global features;

[0014] The flattening layer flattens the multi-dimensional global features into a one-dimensional vector;

[0015] The first fully-connected layer has 16 input features and 16 output features; the first fully-connected layer is used to perform high-level feature representation on the one-dimensional vector;

[0016] The second fully-connected layer has 16 input features and 4 output features; the second fully-connected layer performs classification prediction based on the high-level feature representation;

[0017] After the first convolutional layer, the second convolutional layer, and the first fully-connected layer, ReLU activation functions are respectively provided, and the ReLU activation functions ensure that the features are effectively non-linearized.

[0018] Optionally, during the training process of the covering recognition model, the loss between the prediction result and the true label is calculated through a cross-entropy loss function, and the model parameters are updated through an Adam optimizer.

[0019] In a second aspect, the present invention provides a detection device applicable to the covering recognition method based on multi-frequency excitation as described above. The detection device is configured to, at different temperatures, excite the sensitive capacitor with excitation signals of different frequencies and obtain the resistance value of the equivalent resistance and the capacitance value of the equivalent capacitance; the detection device includes:

[0020] A temperature sensor for collecting temperature values;

[0021] A capacitance excitation circuit connected to one end of the sensitive capacitor for exciting the sensitive capacitor with excitation signals of different frequencies;

[0022] An impedance matching circuit connected to the other end of the sensitive capacitor for matching the impedance of the sensitive capacitor;

[0023] A phase difference detection circuit for obtaining the voltage when the impedance matching circuit reaches a balanced state and the phase difference between the voltage and the voltage and the phase difference between the voltage and the voltage where the voltage is the voltage across the sensitive capacitor, the voltage is the voltage across the impedance matching circuit, and the voltage

[0024] is the voltage across the connection of the sensitive capacitor and the impedance matching circuit; and the phase difference to calculate the resistance value of the equivalent resistance and the capacitance value of the equivalent capacitance.

[0025] Optionally, the capacitance excitation circuit includes a DAC module, a high-pass filter circuit, and a first voltage follower connected in sequence. The DAC generates a sine wave signal, which is filtered by the high-pass filter circuit and amplified by the first voltage follower.

[0026] Optionally, the impedance matching circuit includes a plurality of parallel branches, each branch including a series-connected switch and a weighted resistor. By controlling the on / off of the switch on each branch, the total resistance value of the impedance matching circuit is changed where the total resistance value should be less than and as close as possible to which is the impedance of the equivalent capacitance, is the resistance value of the equivalent resistance, and \(i\) is the imaginary unit.

[0027] Optionally, the phase difference detection circuit includes a second voltage follower, three voltage detection circuits, three square wave conversion circuits, and a phase difference detection chip. The input and output terminals of the second voltage follower are respectively connected to both ends of the impedance matching circuit. The input terminals of the three voltage detection circuits are respectively connected to both ends of the sensitive capacitor, both ends of the impedance matching circuit, and both ends of the connection of the sensitive capacitor and the impedance matching circuit for measuring the voltage the voltage and the voltage ; The input terminals of the three square wave conversion circuits are respectively connected to the output terminals of the three voltage detection circuits, and the input terminals of the phase difference detection chip are respectively connected to the output terminals of the three square wave conversion circuits for obtaining the phase difference and the phase difference ;

[0028] ​

[0029]

[0030] Wherein, are respectively the signals of voltage , voltage and voltage . are respectively the signals of phase difference and phase difference .

[0031] Optionally, the voltage detection circuit adopts a differential amplifier circuit, the rotation method circuit adopts a voltage comparator, and the phase difference detection chip adopts a NOR gate logic chip.

[0032] Optionally, the output end of the voltage detection circuit of the voltage is further connected with a full-wave rectification and filtering circuit for sampling the voltage so as to determine the total resistance value .

[0033] Optionally, the calculation formulas for the resistance value of the equivalent resistance and the capacitance value of the equivalent capacitance are:

[0034]

[0035]

[0036]

[0037] Wherein, is the phase difference between the voltage and , is the phase difference between the voltage and , is the impedance of the equivalent capacitance, is the frequency of the excitation signal, is the resistance value of the equivalent resistance, is the capacitance value of the equivalent capacitance, is the total resistance value of the impedance matching circuit.

[0038] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0039] A method and a detection device for covering recognition based on multi-frequency excitation provided by an embodiment of the present invention. The detection device combines multi-frequency sensitive capacitance excitation and impedance matching circuits. By exciting a capacitance to be measured at multiple frequency points, impedance responses at different frequencies are obtained, ensuring that when it is impossible to distinguish at a certain frequency point, the response data at other frequency points can be used for supplementation and confirmation, enabling the system to more comprehensively acquire the characteristics of the capacitance to be measured, thereby improving the recognition accuracy of the covering. Compared with the traditional method of detecting a single data source of charge and discharge waveforms and detecting the peak value and phase difference of sensitive capacitance, the condition of the detection surface can be understood more comprehensively and accurately, and the accuracy of recognizing different coverings is improved. After comprehensively acquiring the characteristics of the capacitance to be measured, through the covering recognition method, these characteristics can be analyzed quickly and in real time, so as to realize rapid response and monitoring of the condition of the detection surface. Description of the Drawings

[0040] Figure 1 is a schematic flowchart of the method for covering recognition based on multi-frequency excitation provided by an embodiment of the present invention;

[0041] Figure 2 is a topological schematic diagram of the equivalent circuit of a sensitive capacitor provided by the prior art;

[0042] Figure 3 is a graph showing the variation of equivalent capacitance and equivalent resistance with frequency at 20°C and -20°C provided by the prior art;

[0043] Figure 4 is a test schematic diagram of equivalent capacitance and equivalent resistance for dry and waterlogged conditions provided by an embodiment of the present invention;

[0044] Figure 5 is a test schematic diagram of equivalent capacitance and equivalent resistance for dry, frozen, and snow-covered conditions provided by an embodiment of the present invention;

[0045] Figure 6 is a schematic structural diagram of a covering recognition model based on a neural network provided by an embodiment of the present invention;

[0046] Figure 7 is a test result graph of a covering recognition model based on a neural network provided by an embodiment of the present invention;

[0047] Figure 8 is a hardware block diagram of the detection device provided by an embodiment of the present invention;

[0048] Figure 9 is a topological schematic diagram of a capacitance excitation circuit provided by an embodiment of the present invention;

[0049] Figure 10 is a topological schematic diagram of an impedance matching circuit provided by an embodiment of the present invention;

[0050] Figure 11 It is the detection principle diagram of the phase difference detection circuit provided by the embodiment of the present invention;

[0051] Figure 12 It is the schematic diagram of vector method analysis provided by the embodiment of the present invention;

[0052] Figure 13 It is the phase difference provided by the embodiment of the present invention and phase difference schematic diagram;

[0053] Figure 14 It is the vector method analysis schematic diagram for negative processing of provided by the embodiment of the present invention;

[0054] Figure 15 It is the topological schematic diagram of the square wave conversion circuit provided by the embodiment of the present invention;

[0055] Figure 16 It is the schematic diagram of the phase difference extraction process provided by the embodiment of the present invention. Specific Embodiments

[0056] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention, and cannot be used to limit the protection scope of the present invention.

[0057] Embodiment 1:

[0058] As Figure 1 shown, the embodiment of the present invention provides a method for identifying a covering based on multi-frequency excitation, including the following steps:

[0059] Step S1, construct an equivalent circuit of the sensitive capacitor.

[0060] For an actual sensitive capacitor, due to the manufacturing process, there are certain parasitic resistors and inductors in the capacitor. The existence of the internal resistor causes energy loss during the charging and discharging processes of the capacitor at the detection frequency, and introduces additional thermal effects. The internal inductor causes an increase in the impedance of the capacitor in the high-frequency range. In high-frequency applications, the equivalent inductor of the capacitor will have a significant impact on the impedance of the capacitor, thus limiting the performance of the capacitor.

[0061] To more accurately describe the actual sensitive capacitor, an equivalent circuit of the sensitive capacitor is constructed. As Figure 2 shown, the equivalent circuit includes a series-connected equivalent resistor , equivalent inductor and equivalent capacitor , and its impedance is:

[0062]

[0063] When the formula satisfies: the capacitor behaves capacitively.

[0064] When the formula satisfies: the capacitor behaves inductively, and at high frequencies, the capacitor exhibits characteristics similar to an inductor.

[0065] When the formula satisfies: at this time, the capacitive reactance vector is equal to the inductive reactance vector, the total impedance of the capacitor is the smallest, and it behaves as a pure resistance characteristic. At this time, the is called the self-resonant frequency of the capacitor.

[0066] In the embodiment of the present invention, the signal frequency is in the low frequency range, and the internal inductance of the capacitor can be ignored. According to the polarization phenomenon of the dielectric, the existing dielectric loss, and the Debye relaxation effect, different detection frequencies and coverings will affect the equivalent capacitance and equivalent resistance of the capacitor, as Figure 3 shown. At 20°C, the equivalent capacitance in the case of water accumulation is larger than that in the dry state, but the equivalent resistance is smaller; at -20°C, the equivalent capacitance and equivalent resistance of ice change greatly with frequency. The equivalent capacitance and equivalent resistance in the dry and snow-covered cases are relatively close and are less affected by frequency.

[0067] Therefore, it is very important to select an appropriate detection frequency. When the detection frequency is fixed, the equivalent capacitance and equivalent resistance of the capacitor only change with different coverings. In the embodiment of the present invention, an excitation signal of 10KHz to 100KHz is selected for multi-frequency excitation for design analysis.

[0068] As Figure 4 、 Figure 5 shown, different road surface coverings are tested under the conditions of -30°C to 60°C and 10KHz to 100KHz.

[0069] According to the average values of the equivalent capacitance value Cs and the equivalent resistance value Rs, the coverings are classified and judged. When the temperature is greater than or equal to 0°C, the average values of the equivalent resistance value Rs and the equivalent capacitance value Cs in the dry and water accumulation cases are as Figure 4 shown. When the temperature is less than 0°C, the average values of the equivalent resistance value Rs and the equivalent capacitance value Cs in the dry, ice, and snow-covered cases are as Figure 5As shown in the figure, the ideal threshold is set according to the average values of Cs and Rs under different covers, and the judgment is made according to the threshold. When the temperature is greater than or equal to 0°C, only two cases of dry and water accumulation are considered, and they can be directly distinguished by the equivalent capacitance value Cs. The equivalent capacitance value Cs when there is water accumulation is generally in the nF level, and the equivalent capacitance value Cs when it is dry is generally in the pF level. The equivalent capacitance value Cs when there is water accumulation is one order of magnitude larger than that when it is dry. Therefore, when the temperature is above zero, the distinction between dry and frozen can be completed by detecting the equivalent capacitance value Cs at any frequency. When the temperature is less than 0°C, three cases of dry, frozen, and snow-covered need to be considered. The equivalent capacitance values Cs in the three cases are relatively close and there are intersections, so it is impossible to directly judge by the Cs value. It can be found from the above tests of the equivalent resistance of dry, snow-covered, and frozen that the equivalent resistance values Rs of the three have distinguishability, but there will be data intersections at multiple temperature points at low frequencies, and there will be data intersections when the temperature is from -10°C to 0°C at medium frequencies, while there are fewer data intersections at high frequencies, and the higher the frequency, the more stable the data. Therefore, it can be known that by combining temperature data, the dry and water accumulation in the above-zero situation can be distinguished by the equivalent capacitance value, and the dry, frozen, and snow-covered in the below-zero situation can be distinguished by the equivalent resistance value.

[0070] Step S2: At different temperatures, the sensitive capacitor is excited by excitation signals of different frequencies, and the resistance value of the equivalent resistance, the capacitance value of the equivalent capacitance, and the cover category on the sensitive capacitor are obtained.

[0071] Through the above analysis, in this embodiment, the cover categories include dry, water accumulation, frozen, and snow-covered.

[0072] Step S3: The same set of temperature values, frequency values, resistance values, and capacitance values are used as sample data, and the corresponding cover category is used as the true label to construct a sample set.

[0073] Step S4: A cover recognition model based on a neural network is constructed, and the cover recognition model is trained through the sample set.

[0074] As Figure 6 shown, specifically in this embodiment, the cover recognition model includes a first convolutional layer, a second convolutional layer, a flattening layer, a first fully connected layer, and a second fully connected layer connected in sequence.

[0075] The first convolutional layer uses a convolutional kernel with a size of 2, the number of input channels is 1, the number of output channels is 8, the stride is 1, and the padding is 0; the number of parameters of this layer is calculated as follows: the convolutional kernel parameters are 2*1*8, and the bias parameters are 8. Selecting a smaller convolutional kernel size helps to capture local features, reduce the amount of calculation, and accelerate the training process.

[0076] The second convolutional layer uses a convolutional kernel of size 2, with 8 input channels, 16 output channels, a stride of 1, and a padding of 0. The number of parameters in this layer is calculated as follows: the convolutional kernel parameters are 2 * 8 * 16, and the bias parameters are 16. Increasing the number of output channels in the second layer helps to extract richer features.

[0077] The flattening layer flattens the multi-dimensional global features into a one-dimensional vector, facilitating subsequent processing by the fully connected layer.

[0078] The first fully connected layer has 16 input features and 16 output features. The number of parameters in this layer is calculated as follows: the weight parameters are 16 * 16, and the bias parameters are 16. The fully connected layer is used to integrate the features extracted by the convolutional layer for high-level feature representation.

[0079] The second fully connected layer has 16 input features and 4 output features. The number of parameters in this layer is calculated as follows: the weight parameters are 16 * 4, and the bias parameters are 4. The second fully connected layer performs classification prediction based on the high-level feature representation and outputs a four-dimensional probability vector. Each value in the vector corresponds to the predicted probability of a corresponding class. The class with the highest predicted probability is selected as the final prediction result according to each type of predicted probability.

[0080] The model designed in the embodiment of the present invention extracts local features of the input data through the convolutional layer, converts them into one-dimensional vectors by the flattening layer, and performs classification prediction through the fully connected layer. The number of parameters of the model is relatively small, making it suitable for processing data classification tasks with four input features.

[0081] ReLU activation functions are also respectively set after the first convolutional layer, the second convolutional layer, and the first fully connected layer. The ReLU activation function is simple and efficient in calculation, has a constant gradient in the positive value interval, helps to alleviate the problem of gradient disappearance, and ensures that the features are effectively non-linearized through the ReLU activation function, thereby enhancing the expression ability of the model.

[0082] The covering recognition model is trained with a sample set, specifically including:

[0083] (1) Constructing sample data: Tests are carried out every 10 °C in the range of -30 °C to 60 °C. The surface of the sensitive capacitor is covered with air, water, ice, and snow, and the temperature, frequency, equivalent series capacitance value Cs, and equivalent series resistance value Rs at different frequencies are tested.

[0084] (2) Data loading and preprocessing: Use DataLoader to load training and test data.

[0085] (3) Model definition: Define the neural network model structure according to actual needs, including the types and parameters of each layer.

[0086] (4) Definition of loss function and optimizer: Select a suitable loss function, such as cross-entropy loss. At the same time, define an optimizer, such as Adam optimizer.

[0087] (5) Training loop: In each training epoch, iterate through the training dataset, perform forward propagation, loss calculation, backward propagation, and parameter update. After each training epoch, output the loss value of the current epoch.

[0088] (6) Prediction: Use the trained model to make predictions on new data. Pass the test data to the trained neural network model to obtain the final prediction results.

[0089] (7) Model evaluation: After training, use the test dataset to evaluate the performance of the model, such as calculating metrics like accuracy.

[0090] As Figure 7 shown, through actual testing, for the classification prediction of temperature, frequency, equivalent series capacitance value, and equivalent series resistance value, the accuracy of drying and water accumulation is relatively high when it is above zero; the accuracy of drying, snow accumulation, and icing is relatively high when it is below zero. It meets the requirements.

[0091] Step S5: Implement the identification of the covering through the trained covering identification model.

[0092] The covering identification method based on multi-frequency excitation provided by the embodiments of the present invention utilizes deeper features in the output data of the sensitive capacitor, and realizes the accurate identification of road surface coverings by establishing a complex pattern recognition and classification model. This method can overcome the difficulty of identifying the similarity of charge and discharge waveforms in traditional methods and improve the accuracy of covering detection. In addition, the intelligent classification prediction algorithm also has the advantage of adapting to different road surface conditions and environmental changes, providing more reliable support for road safety and the stability of vehicle driving.

[0093] Embodiment 2:

[0094] As Figure 8 shown, the embodiments of the present invention provide a detection device applicable to the covering identification method based on multi-frequency excitation as in Embodiment 1. The detection device is configured to, at different temperatures, excite the sensitive capacitor with excitation signals of different frequencies and obtain the resistance value of the equivalent resistance and the capacitance value of the equivalent capacitance; the detection device includes: a temperature sensor, a capacitance excitation circuit, an impedance matching circuit, a phase difference detection circuit, and a microprocessor. The temperature sensor is used to collect the temperature value; the capacitance excitation circuit is connected to one end of the sensitive capacitor and is used to excite the sensitive capacitor with excitation signals of different frequencies; the impedance matching circuit is connected to the other end of the sensitive capacitor and is used to match the impedance of the sensitive capacitor; the phase difference detection circuit is used to obtain the voltage when the impedance matching circuit reaches the balanced state and voltage the phase difference between and voltage between voltage the phase difference , voltage is the voltage across the sensitive capacitor, voltage is the voltage across the impedance matching circuit, voltage is the voltage across the connection of the sensitive capacitor and the impedance matching circuit; the microprocessor is used to calculate the resistance value of the equivalent resistance and the capacitance value of the equivalent capacitance according to the phase difference and the phase difference .

[0095] In the embodiment of the present invention, by exciting the sensitive capacitor to be measured at multiple frequency points, the impedance responses at different frequencies are obtained, ensuring that when it is impossible to distinguish at a certain frequency point, the response data at other frequency points can be used for supplementation and confirmation, enabling the system to more comprehensively obtain the characteristics of the capacitor to be measured, thereby improving the recognition accuracy of the covering. By combining the capacitance excitation circuit with the impedance matching circuit, the impedance response of the sensitive capacitor is changed. By adjusting the impedance of the impedance matching circuit, the impedance matching circuit is brought to a balanced state, thereby reducing the measurement error. Then, through the phase difference detection circuit, the phase difference angle between the signals is measured and calculated to measure and calculate the impedance characteristics of the sensitive capacitor, thereby reflecting the influence of the covering and the detection frequency on the equivalent capacitance and the equivalent resistance.

[0096] As Figure 9 shown, the capacitance excitation circuit includes a DAC module, a high-pass filter circuit, and a first voltage follower U7 connected in sequence. The DAC generates a sine wave signal (10KHz - 100KHz). The sine wave signal is filtered by the high-pass filter circuit and the driving ability is enhanced by the first voltage follower to excite the sensitive capacitor. The high-pass filter circuit consists of a resistor and a capacitor. This circuit has important applications in signal processing, especially in the process of removing low-frequency noise and blocking low-frequency interference, ensuring that the high-frequency components of the signal can be transmitted. The main function of the first voltage follower U7 after the high-pass filter circuit is to provide a high input impedance and a low output impedance, thereby realizing the buffered transmission of the signal. The input signal and the output signal of this circuit are equal, but it can effectively isolate the front-stage signal source from the rear-stage load and prevent the load from affecting the signal source. In this design, a voltage follower is formed by an operational amplifier and resistors and capacitors. The output terminal of the operational amplifier is connected in series to the inverting input terminal through a resistor R17, and a capacitor C21 is connected in parallel. The function of the resistor R17 is to make the positive and negative input biases symmetrical. For an actual operational amplifier, there is a very small input capacitance (distributed parameters) between the input terminal of the operational amplifier and the ground. This input capacitance will form a low-pass filter with the resistor R17 from the output terminal of the operational amplifier to the inverting terminal, affecting the response at high frequencies. The capacitor C21 connected in parallel with the resistor R17 is used to cancel this effect and improve the high-frequency response.

[0097] The impedance matching circuit includes multiple parallel branches, and each branch includes a series-connected switch and a weighted resistor. By controlling the on / off of the switch on each branch, the total resistance value of the impedance matching circuit is changed. , the total resistance value should be less than and as close as possible to , is the impedance of the equivalent capacitor, is the resistance value of the equivalent resistor, is the imaginary unit.

[0098] As Figure 10 shown, specifically in this embodiment, the impedance matching circuit mainly consists of four analog switch chips (U9, U12, U15, U18). Each chip contains two independent and selectable single-pole double-throw switches, with ultra-low capacitance and charge injection characteristics, which are particularly suitable for data acquisition and sample-and-hold applications, and can provide an ideal solution in occasions with high requirements for low interference and fast stability. Each switch can conduct equally in both directions in the on state, and its input signal range extends to the power supply voltage. The present invention designs 8-way weighted resistors, so 4 analog switch chips are used. For measuring impedance elements of different sizes, switches are used to switch different ranges.

[0099] The phase difference detection circuit includes a second voltage follower, three voltage detection circuits, three square-wave conversion circuits, and a phase difference detection chip; the input end and the output end of the second voltage follower are respectively connected to both ends of the impedance matching circuit, and the input ends of the three voltage detection circuits are respectively connected to both ends of the sensitive capacitor, both ends of the impedance matching circuit, and both ends after the sensitive capacitor and the impedance matching circuit are connected, for measuring voltage , voltage and voltage ; the input ends of the three square-wave conversion circuits are respectively connected to the output ends of the three voltage detection circuits, and the input ends of the phase difference detection chip are respectively connected to the output ends of the three square-wave conversion circuits, for obtaining the phase difference and the phase difference .

[0100] Specifically in this embodiment, the voltage detection circuit uses a differential amplifier circuit. The input ends of the three voltage detection circuits are respectively connected to both ends of the sensitive capacitor, both ends of the impedance matching circuit, and both ends after the sensitive capacitor and the impedance matching circuit are connected to measure voltage , voltage and voltage , as Figure 11 shown, in the figure, A1 corresponds to the second voltage follower, and A2, A3, and A4 respectively correspond to the three voltage detection circuits.

[0101] Voltage ,Voltage and voltage And the input and output voltage signals of the total circuit and , the three voltage signals are analyzed using the vector method in circuit analysis. The vector method analysis diagram is as follows Figure 12 As shown, take a reference direction, which is the direction of current, and the relationships between several voltage signals are:

[0102]

[0103]

[0104]

[0105] No matter how Cs and Rs change, Should be based on The radius of the circle is , and its modulus should be a constant. for and The phase difference angle between for and The phase difference between and The schematic diagram is as follows Figure 13 As shown, if these two angles are measured, the equivalent series resistance Rs and equivalent series capacitance Cs of the sensitive capacitor element can be calculated by combining the tangent angle formula with the total parallel resistance R of the half-bridge circuit and the circuit excitation frequency f. The calculation formula is as follows:

[0106]

[0107]

[0108]

[0109] In the formula, is voltage and The phase difference between is voltage and The phase difference between is the impedance of the equivalent capacitor, is the frequency of the excitation signal, is the resistance value of the equivalent resistor, is the capacitance value of the equivalent capacitor, is the total resistance of the impedance matching circuit.

[0110] However, a problem will occur in the practical application of the above method, that is, when Rs and are too small, it will be very difficult to measure . Therefore, negative processing can be performed on through reverse difference amplification, that is . As shown in Figure 14 which is the schematic diagram of vector method analysis after negative processing , at this time, even if Rs and become smaller, the measured and will still be very large.

[0111] The total parallel resistance value R of the impedance matching circuit needs to meet certain requirements, which are divided into 3 cases:

[0112] 1. If , waveform distortion will occur;

[0113] 2. If , the measurement accuracy is the best, and

[0114] 3. If is too small, the measurement accuracy will deteriorate, resulting in too small an included angle between

[0115] Therefore, should be selected to be less than and as close as possible to , but it cannot be judged by whether is greater than and close to , because when is very small, it is very difficult to accurately judge. It should be judged by measuring whether is less than and close to . should be as large as possible without distortion, and should be gradually reduced during measurement, so that it is better when is just less than

[0116] By measuring , that is , the output terminal of the voltage detection circuit of voltage is also connected to a full-wave rectification and filtering circuit, which is used to sample voltage , and then determine the total resistance value An operational amplifier can be used to perform full-wave rectification on an AC signal, convert the input bipolar AC signal into a unipolar signal, and also amplify the signal. Although the output voltage of the rectifier circuit is in a single direction, it contains a large AC component. Therefore, in the present invention, a capacitor is connected in parallel at the output end of the rectifier circuit to form a capacitor filter circuit. The larger the capacitance of the filter capacitor, the smaller the low-frequency cut-off frequency, and the better the filtering effect.

[0117] The value of R should satisfy in the circuit design Figure 10 , a total of 8 paths, and the R value range of 256 combinations should be as large as possible greater than the impedance range of the sensitive capacitor at all selected frequencies. In each test, an algorithm is used to find the R value that is most suitable for the current frequency and the current covering material.

[0118] The amplification factor of the differential amplifier circuit is designed to be 1 times, and metal film resistors are selected for the resistors. After passing through the differential amplifier circuit, three signals are obtained 、 、 . In order to obtain the phase difference between them, the phase conversion circuit uses a voltage comparator, as shown in Figure 15 and Figure 16 . Using the voltage comparators U10, U13, and U19, 、 、 are respectively converted into square waves, and then input to the phase difference detection chip, which is a NOR gate logic chip with 4 two-input NOR gates. and and have a relationship of two NOR operations. and and also have a relationship of two NOR operations. Finally, the signals of and are output through the logic chip. These two square wave signals are input into the single-chip microcomputer, and the single-chip microcomputer timer is used to detect the high-level duration to measure the duty cycle.

[0119] A square wave is a periodic signal composed of alternating high and low levels. Its characteristic is that it is convenient for phase difference comparison. After the signal passes through the voltage comparator, according to the setting of the threshold, it will become a square wave signal bounded by the threshold. When the voltage of the input signal is higher than the threshold, the comparator outputs a high level; otherwise, when it is lower than the threshold, it outputs a low level. By comparing the phases of the converted square wave signals, the phase difference between the signals can be measured and calculated. Logic symbols are assigned to the square waves converted from the three groups of signals. , , are respectively assigned the logic symbols A, B, and C. The phase difference angles of the two phases and The corresponding signals are respectively assigned to Y1 and Y2. The three groups of square waves are respectively compared and analyzed to establish a truth table. When A = 1 and B = 0, Y1 = 1; in other cases, Y1 = 0. Similarly, when C = 1 and B = 0, Y2 = 1; in other cases, Y2 = 0.

[0120]

[0121]

[0122] In the formula, are respectively the voltages , voltage and voltage signals, are respectively the phase differences and phase difference signals.

[0123] Specifically, in this embodiment, the temperature sensor uses a PT100 platinum resistance. The measurement range of the PT100 platinum resistance is usually between -200°C and 850°C, which can cover a wide temperature range. Platinum metal has high chemical stability, and the temperature measurement of the PT100 platinum resistance has good long-term stability. The resistance of the PT100 changes approximately linearly with temperature, so it is easy to establish an accurate corresponding relationship with temperature. The temperature is obtained by measuring the resistance value of the PT100 platinum resistance.

[0124] The detection device also includes a power supply circuit, which is composed of two parts: an analog circuit and a digital circuit. Since there are high requirements for the measurement accuracy, the power supply circuit part needs to be reliable and stable. By isolating the analog power supply and the digital power supply, the electromagnetic coupling between them can be reduced, and the possibility of mutual interference can be lowered. Since there may be stray interference in the ground return current, the digital ground and the analog ground need to be separated and finally connected to the common ground of the power input to reduce the adverse impact of the ground return current on the analog circuit.

[0125] The detection device provided by the embodiment of the present invention obtains impedance responses at different frequencies, ensuring that when it is impossible to distinguish at a certain frequency point, the response data at other frequency points can be used for supplementation and confirmation, enabling the system to more comprehensively obtain the characteristics of the capacitance to be measured, thereby improving the recognition accuracy of the covering. At different frequency points, the equivalent series capacitance value Cs of the sensitive capacitance element changes, resulting in the change of the impedance value Xc of the equivalent series capacitance, and then the impedance value R of the best-matched half-bridge circuit also changes. Under certain coverings, the equivalent series capacitance value Cs and the equivalent series resistance value Rs of the sensitive capacitance element will decay, while under certain other coverings, they hardly change. By changing different excitation frequencies, it is judged whether the equivalent series capacitance value Cs and the equivalent series resistance value Rs of the sensitive capacitance element will decay or not, thereby increasing the criterion to ensure that when it is impossible to distinguish at a certain frequency point, the response data at other frequency points can be used for supplementation and confirmation, enabling the system to more comprehensively obtain the characteristics of the capacitance to be measured, thereby improving the recognition accuracy of the covering.

[0126] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can be implemented in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can be implemented in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0127] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one or more of the processes Figure 1 or multiple processes and / or blocks

[0128] These computer program instructions can also be stored in a computer-readable memory capable of guiding the computer or other programmable data processing devices to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in Figure 1 one or more of the processes Figure 1The functions specified in one or more boxes.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide for implementing the steps of the functions specified in Figure 1 One process or more processes and / or boxes Figure 1 The steps of the functions specified in one box or more boxes.

[0130] The above is only the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A detection device applicable to a covering recognition method, characterized in that, The detection device is configured to, at different temperatures, excite the sensitive capacitor with excitation signals of different frequencies, and obtain the resistance value of the equivalent resistance and the capacitance value of the equivalent capacitance of the sensitive capacitor; The detection device includes: a temperature sensor for collecting temperature values; a capacitance excitation circuit connected to one end of the sensitive capacitor for exciting the sensitive capacitor with excitation signals of different frequencies; An impedance matching circuit, connected to the other end of the sensitive capacitor, is used to match the impedance of the sensitive capacitor; the impedance matching circuit includes a plurality of parallel branches, and each branch includes a switch and a weighting resistor connected in series. By controlling the on / off of the switch on each branch, the total resistance value of the impedance matching circuit is changed. , the total resistance value should be less than and as close as possible to , which is the impedance of the equivalent capacitance, where is the resistance value of the equivalent resistance, and is the imaginary unit; A phase difference detection circuit, configured to obtain voltages when the impedance matching circuit reaches a balanced state and voltage the phase difference between 、voltage and voltage the phase difference between wherein the voltage is the voltage across the sensitive capacitor, and the voltage is the voltage across the impedance matching circuit, and the voltage is the voltage across the connection of the sensitive capacitor and the impedance matching circuit; A microprocessor for calculating a resistance value of an equivalent resistance and a capacitance value of an equivalent capacitance according to the phase difference and the phase difference ; The cover identification method includes: constructing an equivalent circuit of the sensitive capacitor, the equivalent circuit including a series-connected equivalent resistance, equivalent inductance, and equivalent capacitance; at different temperatures, exciting the sensitive capacitor with excitation signals of different frequencies, and obtaining the resistance value of the equivalent resistance, the capacitance value of the equivalent capacitance, and the cover category on the sensitive capacitor; using the temperature value, frequency value, resistance value, and capacitance value of the same group as sample data, and using the corresponding cover category as the true label to construct a sample set; constructing a cover identification model based on a neural network, and training the cover identification model through the sample set; realizing cover identification through the trained cover identification model.

2. The detection device according to claim 1, wherein The capacitance excitation circuit includes a DAC module, a high-pass filter circuit, and a first voltage follower connected in sequence. The DAC generates a sine wave signal, and the sine wave signal is filtered by the high-pass filter circuit and enhanced by the first voltage follower.

3. The detection device according to claim 1, characterized in that The phase difference detection circuit includes a second voltage follower, three voltage detection circuits, three square wave conversion circuits, and a phase difference detection chip; the input terminal and the output terminal of the second voltage follower are respectively connected to both ends of the impedance matching circuit, and the input terminals of the three voltage detection circuits are respectively connected to both ends of the sensitive capacitor, both ends of the impedance matching circuit, and both ends after the connection of the sensitive capacitor and the impedance matching circuit for measuring voltages , voltage , and voltage ; the input terminals of the three square wave conversion circuits are respectively connected to the output terminals of the three voltage detection circuits, and the input terminals of the phase difference detection chip are respectively connected to the output terminals of the three square wave conversion circuits for obtaining phase differences and phase difference ; ； ； Wherein, are respectively the signals of voltage , voltage and voltage . are respectively the signals of phase difference and phase difference .

4. The detection device according to claim 3, wherein, The voltage detection circuit uses a differential amplifier circuit, the conversion method circuit uses a voltage comparator, and the phase difference detection chip uses a NOR logic chip.

5. The detection device according to claim 3, wherein The voltage The output terminal of the voltage detection circuit is also connected to a full-wave rectification and filtering circuit for sampling the voltage and further determining the total resistance value .

6. The detection device according to claim 1, wherein The calculation formulas for the resistance value of the equivalent resistance and the capacitance value of the equivalent capacitance are: ； ； ； wherein, is the voltage and is the phase difference therebetween, is the voltage and is the phase difference therebetween, is the impedance of the equivalent capacitance, is the frequency of the excitation signal, is the resistance value of the equivalent resistance, is the capacitance value of the equivalent capacitance, is the total resistance value of the impedance matching circuit.

7. The detection device according to claim 1, characterized in that The cover identification model includes a first convolutional layer, a second convolutional layer, a flattening layer, a first fully connected layer, and a second fully connected layer connected in sequence; The first convolutional layer uses a convolutional kernel of size 2, has an input channel number of 1, an output channel number of 8, a stride of 1, and a padding of 0; the first convolutional layer is used to process the input data and extract local features; The second convolutional layer uses a convolutional kernel of size 2, has an input channel number of 8, an output channel number of 16, a stride of 1, and a padding of 0; the second convolutional layer is used to process the local features and extract global features; The flattening layer flattens the multi-dimensional global features into a one-dimensional vector; The first fully connected layer has an input feature number of 16 and an output feature number of 16; the first fully connected layer is used to perform high-level feature representation on the one-dimensional vector; The second fully connected layer has an input feature number of 16 and an output feature number of 4; The second fully connected layer performs classification prediction based on the high-level feature representation; After the first convolutional layer, the second convolutional layer, and the first fully connected layer, ReLU activation functions are respectively provided, and the ReLU activation functions are used to ensure that the features are effectively non-linearized.

8. The detection device according to claim 1, wherein During the training process of the cover identification model, the loss between the prediction result and the true label is calculated through a cross-entropy loss function, and the model parameters are updated through an Adam optimizer.

Images