A vehicle-mounted method for detecting pollutants on airport runways based on ultrasonic technology.

Abstract

This invention relates to the field of aviation safety and airport runway inspection, and proposes a vehicle-mounted method for detecting contaminants on airport runways based on ultrasonic technology. The method includes the design of a vehicle-mounted ultrasonic sensor, contaminant type identification, and depth detection technology. The sensor consists of an ultrasonic probe, a lifting motor, a rotating motor, a waterproof cover, a clamp, and an aviation connector. The contaminant type identification steps include: acquiring a contaminant echo signal dataset, extracting waveform feature parameters (such as amplitude, distortion, and signal-to-noise ratio), constructing a CNN deep learning model, training the neural network, and outputting the identification result. The contaminant depth detection technology is based on ultrasonic echo parameter estimation, and the steps include: extracting the echo signal envelope, retrieving feature points, constructing a three-objective function (based on feature point differences and loss function values), using the OMOPSO algorithm to solve for the optimal solution of the model parameters, and calculating the contaminant depth.

Description

Technical Field

[0001] This invention relates to an airport runway ice and snow water information detection technology, including ice and snow water depth detection and type identification, and involves ultrasonic sensors and echo signal processing technology. Background Technology

[0002] In the aviation industry, runway safety is crucial for ensuring the normal takeoff and landing of aircraft. Especially under adverse weather conditions, such as when the runway surface is covered with ice, snow, or water, its anti-skid performance decreases significantly, increasing the risk of tire slippage and aircraft overrunning the runway. This risk not only threatens the lives of passengers and crew but can also have a significant impact on airport operations. Therefore, accurate monitoring of the depth of contaminants on the runway surface has become an important task in airport safety management.

[0003] According to ICAO's Circular 355, contaminants include water, ice, wet snow, dry snow, and slush, requiring precise measurement of runway surface contaminant depth with millimeter-level accuracy. This requirement is based on in-depth research into the relationship between runway surface anti-skid properties and contaminant depth. When the contaminant depth reaches 3 millimeters, the runway's coefficient of friction, whether from dry / wet snow or slush / water, decreases significantly. Furthermore, during high-speed taxiing, fluid contaminants exert hydrodynamic pressure on the aircraft tires, causing them to lose contact with the runway surface. This greatly reduces tire braking effectiveness and increases the risk of the aircraft running off the runway.

[0004] To address this challenge, the development of runway surface contaminant detection technology is essential. Runway surface ice / snow / water detection is mainly divided into two types: contact and non-contact. Contact methods primarily involve embedding sensors on the runway surface to monitor ice / snow / water in real time. Non-contact methods mainly utilize remote sensing sensors to scan, measure distances, and photograph runway contaminants. Common remote sensing sensors include laser, ultrasonic, and infrared sensors, as well as optical / infrared cameras. These remote sensing sensors can be mounted on the runway shoulder (primarily installed at weather stations) or on mobile inspection vehicles. Embedded sensors, as point-based detection methods, have a small detection area, are easily damaged when exposed on the runway surface, have a short lifespan, and are insufficient in terms of accuracy, monitoring cost, and maintenance cost to meet the requirements of airports for runway surface condition assessment. While weather monitoring stations are powerful and provide diverse information, their detection range is limited to the vicinity of their installation location, making it difficult to cover the entire airport area. To monitor multiple runway sections, multiple weather stations are needed; for example, Wellington Airport in New Zealand has installed Remote Weather Stations (RWS) located in the takeoff and landing areas of both runways and at the runway centerline. However, deploying RWS at multiple points significantly increases monitoring costs, and due to height restrictions at airports, the low installation height of remote sensing sensors also reduces the detection range. Vehicle-mounted inspection vehicles, due to their high efficiency and wide detection range, have become the preferred choice for many airports for pavement condition monitoring.

[0005] Currently, vehicle-mounted sensors for detecting ice and snow on runway surfaces mostly employ optical technologies, including laser, spectral, and ultrasonic remote sensing. For example, the MD30 sensor developed by Vaisala in Finland uses laser technology, while the MARWIS sensor developed by Lufft in Germany and the RCM411 sensor developed by Teconer in Finland both use spectral technology. Although these methods can obtain information on the type and depth of pollutants on the pavement, their detection range is limited, and the depth measurement results are inaccurate for transparent media such as slush, making it difficult to achieve millimeter-level accuracy. In addition, ambient light can also interfere with infrared spectral sensors, resulting in poor reliability of the results.

[0006] Ultrasonic detection of ice, snow, and water is unaffected by transparent media or light, and with appropriate signal processing, it can achieve millimeter-level or even higher accuracy. Furthermore, based on experimentally obtained echo signals of different contaminants, and with distinct waveform characteristics, it holds promise for identifying ice, water, wet snow, dry snow, and slush based on echo signal features. This characteristic makes ultrasonic sensors an ideal choice for measuring contaminant depth on vehicles. Ultrasonic sensors not only meet millimeter-level accuracy requirements but also adapt to various harsh runway conditions, providing strong support for runway surface evaluation, friction / anti-skid performance analysis, and improved management / maintenance efficiency. Therefore, ultrasonic sensors have broad application prospects in the field of runway contaminant depth detection. Summary of the Invention

[0007] To address the aforementioned technical problems, this application aims to design a runway surface contaminant detection technology, including a vehicle-mounted ultrasonic sensor, contaminant type identification technology, and contaminant depth detection technology.

[0008] This application achieves the above-mentioned effects through the following technical solution: a vehicle-mounted airport runway pollutant detection method based on ultrasonic technology, wherein the method is based on the following detection system:

[0009] Design of vehicle-mounted ultrasonic sensor;

[0010] The vehicle-mounted ultrasonic sensor is installed on the vehicle and includes an ultrasonic air probe. The ultrasonic air probe is a single-transmitter, single-receiver air coupling probe. The probe is embedded on a PCB circuit board, and the PCB board is connected to a rotary motor.

[0011] The vehicle-mounted ultrasonic sensor also includes a rotary motor and a lifting motor. The rotary motor controls the rotation direction and speed of the ultrasonic air probe by changing the voltage polarity and magnitude, and the lifting motor realizes the lifting and lowering of the ultrasonic air probe through a lead screw.

[0012] The vehicle-mounted ultrasonic sensor also includes a waterproof cover. The inner wall of the waterproof cover is fixed to the lifting motor, the outside is connected to a fixed rod, an aviation plug is installed on the side, and an opening is made at the bottom for the probe to receive echo signals.

[0013] The vehicle-mounted ultrasonic sensor is installed on the vehicle using a clamp. The clamp consists of a detachable steel connecting rod, a fixing metal plate, and hand-tightening screws. The steel rod can be segmented and disassembled to change its length. The fixing metal plate is used to fix the waterproof cover and the connecting rod.

[0014] The vehicle-mounted ultrasonic sensor is installed on a vehicle and includes a single-transmitter, single-receiver air-coupled ultrasonic air probe. The sensor is equipped with lifting and rotating motors to control the rotating probe's detection height and angle. The lifting motor raises and lowers the probe via a lead screw. The probe is embedded in a PCB circuit board and connected to the rotating motor. A waterproof cover is provided on the probe and motor for waterproofing and dustproofing. An aviation connector is provided on the outside of the waterproof cover for signal transmission.

[0015] The sensor is mounted on the vehicle using a clamp consisting of a detachable steel connecting rod, a fixed metal plate, and hand-tightened screws. The length of the steel rod can be adjusted to accommodate different installation requirements.

[0016] Furthermore, the pollutant detection information includes pollutant type and depth information. The vehicle-mounted ultrasonic sensor first performs type identification and then performs depth calculation. The type identification includes acquiring echo signal samples of various pollutants, extracting echo signal waveform feature parameters, constructing a CNN deep learning model, adjusting the CNN model parameters, and outputting pollutant type identification results.

[0017] The echo signal waveform parameters include the amplitude, distortion level, and signal-to-noise ratio of the echo signal.

[0018] Furthermore, the amplitude of the echo signal is expressed as:

[0019] A = max[S] p (i)]

[0020] Where A represents the amplitude of the echo signal, S p (t) represents the maximum voltage amplitude at the time series points of the obtained echo signal, where i is the time series point of the echo signal;

[0021] The degree of distortion is represented by cosine similarity, which is expressed as:

[0022]

[0023] Among them, S p ,S d S represents the actual ultrasonic echo signal obtained. d This represents the echo signal of the standard mixed exponential model, where N represents the total number of signal sampling points.

[0024] The echo signal of the hybrid exponential model is specifically as follows:

[0025]

[0026] Among them, t i Here, A0 is the amplitude of the signal, m and T are the inherent parameters of the ultrasonic probe, used to characterize the waveform skewness and kurtosis of the signal, τ is the starting point offset, f0 is the center frequency, and θ is the initial phase.

[0027] The signal-to-noise ratio of the echo signal is expressed as:

[0028]

[0029] Among them, P s and P n These represent the power of the signal and the power of the noise, respectively.

[0030] Furthermore, the CNN deep learning model includes convolutional layers, input layers, hidden layers, fully connected layers, and output layers. The convolutional layers are used to extract local features from the echo signal, and the fully connected layers are used to map the extracted features to pollutant categories. The input of the neural network is the waveform parameters of the echo signal, and the output is the type of pollutant.

[0031] Furthermore, the depth detection method includes echo signal envelope extraction, echo signal feature point retrieval, construction of a multi-objective function, obtaining the optimal solution of the multi-objective function using the OMOPSO method, and pollutant depth calculation.

[0032] Furthermore, the process of extracting the echo signal envelope employs wavelet denoising, Hilbert transform, and cubic spline curve smoothing to filter the echo signal and obtain a smoothed signal, specifically as follows:

[0033] The wavelet denoising process includes using Nakagami wavelet basis functions, selecting wavelet coefficients, decomposition scale and threshold, using soft thresholding to remove signal noise, and finally reconstructing the signal.

[0034] The Hilbert transform is:

[0035] z(i)=H{S p (i)}

[0036] R 0p (i)=|z(i)|

[0037] Among them, H{S p (i)} is the Hilbert operation performed on the echo signal, where z(i) is the complex signal and |z(i)| is the modulus of the complex signal; the cubic spline curve interpolation method is as follows:

[0038]

[0039] Among them, R p (i) is the interpolation curve, which is the smoothed envelope, and s(i) is the cubic spline basis function.

[0040] The basis function s(i) can be expressed as:

[0041] s(i)=ai +b i (xx i )+c i (xx i )2+d i (xx i )

[0042] Wherein, coefficient a i b i c i d i The conditions are obtained by interpolation conditions, curve continuity conditions, and continuity conditions of the first and second derivatives of the curve, respectively.

[0043] Furthermore, the echo signal feature points include:

[0044] The first characteristic point is the peak point of the envelope function of the echo signal, which is represented as:

[0045] Result[i1]suject to R p (i1)=max(R p (i))

[0046] Wherein, Result[i1] represents the retrieval result of the first feature point in the time series;

[0047] The second characteristic point is the peak point of the first derivative of the echo signal envelope function, which is expressed as:

[0048]

[0049] Wherein, Result[i2] represents the retrieval result of the second feature point in the time series. Indicates envelope R p The first derivative of (i).

[0050] Furthermore, the multi-objective function includes:

[0051] The first objective function is to minimize the loss function value of the actual echo signal and the fitted exponential function signal, and its expression is:

[0052]

[0053] Where loss(x) is the loss function with echo signal parameters x = [A, m, T, τ], and S f This is the fitted function;

[0054] The second objective function is to minimize the difference between the first feature points of the actual echo signal and the fitted exponential function signal in the time series. Its expression is:

[0055]

[0056] in, To retrieve the difference between the time series points of the fitted function signal and the actual signal with echo signal parameters x = [A, m, T, τ] and the first feature point, R... f The envelope function of the fitted signal;

[0057] The third objective function, defined as having the smallest difference between the second feature points of the actual echo signal and the fitted exponential function signal on the time series, is expressed as follows:

[0058]

[0059] in, To retrieve the difference between the time series points of the second feature point of the fitted function signal and the actual signal under the echo signal parameters x=[A,m,T,τ];

[0060] The final optimization objective based on the three objective functions is:

[0061] minL(x)=[L(x)1,L(x)2,L(x)3]

[0062]

[0063] Here, L(x)1, L(x)2, and L(x)3 are three objective functions.

[0064] Furthermore, the steps for obtaining the optimal solution of the multi-objective function using the OMOPSO method include: initializing particle position and velocity parameters, updating and mutating particle position and velocity, storing the solution of non-dominated particles, determining the output conditions, and outputting the Pareto optimal solution set.

[0065] The velocity of each particle is expressed as:

[0066]

[0067] Among them, C1 and C2 of OMOPSO are the control... and gbest t The acceleration constants affecting particle velocity, C1 and C2 are random numbers in the range [0,2], ω is a random number, and r1 and r2 are random numbers in the range [0,1] used to increase the randomness of the search process; w t It is the inertial weight of particle i, and controls the trade-off between global history and local history; and Let i be the position and velocity of particle i at time t;

[0068] After obtaining the velocity of each particle, the position of particle i is updated, and the updated particle position is represented as:

[0069]

[0070] The Archive is the solution set of the non-dominated solution of the multi-objective function. The size of the Archive is determined by ∈, and ∈ satisfies the following equation:

[0071]

[0072] The size of Archive is determined by ∈-dominance, where the value of ∈ restricts the size of Archive. The value of ∈ is defined by the user, and the decision vector x1 is ∈-dominance x2.

[0073] Crowding degree is the density of the distribution of solutions to a multi-objective function in the search space, and it is expressed as follows:

[0074]

[0075] Among them, i d This represents the crowding degree of the i-th non-dominated particle in Archive. This represents the j-th objective function value at point i+1. Let represent the j-th objective function value at point i-1, and m be the number of objective functions.

[0076] The beneficial effects of this application are as follows: The vehicle-mounted pollutant information detection method based on ultrasonic technology proposed in this patent can accurately identify pollutants, including water / ice / wet snow / snow slurry / dry snow types, and accurately calculate their depth. At the same time, the vehicle-mounted sensor can dynamically acquire pollutant information on the runway surface, improving detection efficiency. Attached Figure Description

[0077] Figure 1 These are three views of the vehicle-mounted ultrasonic sensor described in this invention.

[0078] Figure 2 This is a three-dimensional diagram of the vehicle-mounted ultrasonic sensor described in this invention;

[0079] Figure 3 This is an ultrasonic echo signal diagram of ultrasonic water / ice / snow mentioned in the embodiments of the present invention;

[0080] Figure 4 This is a flowchart of the CNN-based pollutant type identification method described in this invention;

[0081] Figure 5 This is a flowchart of the depth detection method based on the OMOPSO method described in this invention;

[0082] Figure 6 This is a flowchart illustrating the process of obtaining optimal waveform parameters based on the OMOPSO algorithm described in this invention. Detailed Implementation

[0083] This invention relates to a technology for detecting contaminants on runway surfaces, the specific implementation method of which is as follows:

[0084] like Figure 1 and Figure 2 As shown, this embodiment provides a vehicle-mounted method for detecting pollutants on airport runways based on ultrasonic technology. This method is based on a vehicle-mounted ultrasonic sensor. The invented vehicle-mounted ultrasonic sensor includes an excitation probe 1, a receiving probe 2, a lifting motor 3, a rotary motor 4, a waterproof cover 5, a clamp 6, and an aviation connector 7. The rotary motor is connected to a lead screw 8, and the clamp includes a connecting rod 9 for connecting the sensor and the clamp, and a hand-tightening screw 10 for adjusting the clamp.

[0085] like Figure 1 and Figure 2 As shown, this invention employs a 40kHz single-transmitter, single-receiver air-coupled probe 1, designed with two probes horizontally spaced 15mm apart to ensure the sound wave generating surfaces are on the same horizontal plane. The probe surface is designed to be waterproof to withstand outdoor environments. The maximum input voltage of the probe is 15Vpp, and the operating voltage in this invention is 10Vpp, suitable for a temperature range of -20℃ to 70℃. The probe is fixed to a circuit board, which is treated with epoxy resin for waterproofing and dustproofing, and connected to a rotary motor for precise positioning.

[0086] like Figure 1 and Figure 2 As shown, this invention incorporates a lifting motor 2 and a rotary motor 3 to adjust the detection angle and height of the sensor. The circuit board of the ultrasonic probe is mounted on a rotating plate, which is fixed to the rotating shaft of the motor via screws. The rotation direction and speed of the motor can be controlled by changing the polarity and magnitude of the voltage, with a rotation angle range of -45° to 45°. The rotary motor is mounted on a sliding base, which is mounted on a lead screw connected to the motor's rotating shaft. Rotating the lead screw allows for the probe to be raised or lowered horizontally, with a lifting range of 10cm.

[0087] like Figure 1 and Figure 2 As shown, the waterproof cover 4 of this invention is designed with a height of 15cm, a width of 10cm, and a length of 15cm. The inner wall is fixed to the lifting motor with screws, and the outside is connected to the fixed rod. An opening is provided at the bottom of the waterproof cover for the generation and reception of ultrasonic signals, and an aviation plug is installed on the side for circuit connection and motor control.

[0088] like Figure 1 and Figure 2As shown, the clamp consists of a detachable steel rod, a fixed metal plate, and hand-tightened screws. The ultrasonic testing device is connected to the fixed metal plate via a waterproof cover and is fixed to the first end of the steel rod. The hand-tightened screws are installed at the tail end of the clamp to secure the ultrasonic testing device to the vehicle. The steel rod can be arbitrarily lengthened by segmenting and disassembling to adapt to the installation requirements of different vehicles.

[0089] The vehicle-mounted ultrasonic sensor includes:

[0090] Ultrasonic air probe, position and orientation control motor, waterproof cover, clamp, aviation connector.

[0091] The ultrasonic air probe is specifically a 40kHz single-transmitter, single-receiver air-coupled probe with a maximum input voltage of 15Vpp, a usable input voltage of 10Vpp, and an operating temperature range of -20℃ to 70℃. The horizontal distance between the two probes is 15mm, and the sound wave generating surfaces of both probes are on the same horizontal plane. The probe surface is also waterproof, meaning that water splashes during testing will not affect the probe's measurement operation. The ultrasonic probe is fixed to a circuit board, which is treated with epoxy resin for waterproofing and dustproofing. The circuit board is connected to a rotary motor.

[0092] Specifically, the motor is designed to address the issue that the fixed position of the ultrasonic testing device varies on different testing vehicles during actual testing. When the height and angle of the ultrasonic probe deviate from the preset position, a lifting motor and a rotating motor can be used to control the sensor's detection angle and height. The circuit board of the ultrasonic probe is mounted on a rotating plate, which is connected and fixed to the shaft of the motor by screws. The rotation direction and speed of the motor are controlled by changing the polarity and magnitude of the voltage. Figure 5 As shown, with the horizontal plane as 0°, the probe rotation angle range is -45° to 45°. The rotary motor is mounted on a sliding base, which in turn is mounted on a lead screw. The lead screw is a helical metal rod connected to the motor shaft. The lead screw has threads that mate with a nut. When the motor rotates, the threads guide the nut to move along the axial direction of the lead screw. The probe's horizontal position is raised or lowered by rotating the lead screw. The lead screw is 15cm long, and the maximum probe lifting height is 10cm.

[0093] The waterproof cover is specifically designed with the following dimensions: a height of 15cm, a width of 10cm, and a length of 15cm. The inner wall of the waterproof cover is fixed to the lifting motor with screws, and the outer surface of the waterproof cover is connected to a fixing rod with screws. An opening at the bottom of the waterproof cover is used for ultrasonic signal generation and reception, and an aviation connector is installed on the side of the waterproof cover. Sockets are used for connecting the ultrasonic probe circuit, controlling the lifting motor, and controlling the rotation motor. An opening at the bottom of the waterproof cover, measuring 8cm x 8cm, allows the probe to receive the echo signal reflected from the surface of the contaminant.

[0094] The clamp is specifically composed of a detachable steel rod, a fixed metal plate, and hand-tightened screws. The ultrasonic testing device is connected to the fixed metal plate via a waterproof cover and is fixed to the head end of the steel rod with screws. A hand-tightened screw is installed at the tail end of the clamp to secure the ultrasonic testing device to the vehicle. The length of the steel rod can be arbitrarily changed by segmenting and disassembling it.

[0095] Specifically, the aviation plugs are: all three aviation plugs are M12 type aviation plugs, and the aviation plugs are connected to the positive and negative leads of the ultrasonic sensor, the wires of the lifting and rotating motors, inside the waterproof cover.

[0096] The pollutant type identification technology includes: acquiring various pollutant echo signal datasets, extracting waveform feature parameters of the echo signals, constructing a CNN deep learning model, training the neural network, and outputting the pollutant type identification results.

[0097] like Figure 1 and Figure 2 As shown, the three aviation plugs used in this invention are all M12 type. They are connected to the positive and negative leads of the ultrasonic sensor, the wires of the lifting and rotating motors, and are located inside the waterproof cover, ensuring the reliability and stability of the electrical connection.

[0098] The acquisition of the various pollutant echo signal datasets is specifically as follows: Before performing the pollutant type identification task, echo signal datasets containing multiple pollutant types such as water, ice, wet snow, dry snow, and slush are first collected and organized to ensure the model's generalization ability and identification accuracy. Data acquisition involves collecting echo signals from winter snow samples and, in a low-temperature environment laboratory, collecting echo signals of different pollutants using artificial snowmaking methods.

[0099] Extracting the waveform feature parameters of the echo signal is one of the key steps in identifying pollutant types. These echo signal feature parameters include the amplitude, distortion level, and signal-to-noise ratio. The extraction of these feature parameters aids in the subsequent training and identification of deep learning models.

[0100] This invention employs deep learning methods to identify pollutant types based on the echo signal characteristics of different types of pollutants. The identification method is as follows: Figure 4 As shown, it includes the following five steps.

[0101] Step 1: Obtain echo signal samples of various pollutants, including water, ice, wet snow, dry snow, and slush echo signal samples.

[0102] Specifically, the implementation is as follows: Figure 3As shown, echo signals from water, ice, wet snow, dry snow, and slush were collected and processed to ensure the model's generalization ability and recognition accuracy. Data was acquired through echo signal collection from winter snow samples and by collecting echo signals of different pollutants using artificial snowmaking methods in a low-temperature laboratory.

[0103] Step 2: Extract the waveform feature parameters of the echo signal.

[0104] Specifically, the implementation is as follows: Figure 4 As shown in the flowchart, waveform feature parameters of the echo signal are extracted, including the amplitude, distortion level, and signal-to-noise ratio of the echo signal. The extraction of these feature parameters helps in the subsequent training and recognition of deep learning models.

[0105] The amplitude of the echo signal is expressed as:

[0106] A = max[S] p (i)]

[0107] Where A represents the amplitude of the echo signal, S p (t) represents the maximum voltage amplitude at the time series points of the obtained echo signal, where i is the time series point of the echo signal.

[0108] The degree of distortion of the signal shown is represented by cosine similarity, which is expressed as:

[0109]

[0110] Among them, S p ,S d These represent the actual ultrasonic echo signal and the echo signal from the standard mixture exponential model, respectively.

[0111] The echo signal of the hybrid exponential model is specifically as follows:

[0112]

[0113] Among them, t i A0 is the time series point, m and T are the inherent parameters of the ultrasonic probe, used to characterize the waveform skewness and kurtosis of the signal, τ is the starting point offset, f0 is the center frequency, and θ is the initial phase.

[0114] The signal-to-noise ratio of the echo signal is expressed as:

[0115]

[0116] Among them, P s and P n These represent the power of the signal and the power of the noise, respectively.

[0117] Step 3: Build a CNN deep learning model.

[0118] Specifically, the implementation is as follows: Figure 4 As shown in the flowchart, a convolutional neural network (CNN) is used to process and analyze echo signal data. The CNN model consists of an input layer, hidden layers, fully connected layers, and an output layer. The convolutional layers extract local features from the echo signal, while the fully connected layers map the extracted features to pollutant categories. The neural network input consists of three feature parameters of the ultrasonic echo signal: amplitude A, echo distortion C, and so on. s The signal-to-noise ratio (SNR) is used as the output, and the type of pollutant is also considered. The activation function used for the hidden layer neurons is the sigmoid function. There are three hidden layers, with 20 neurons in each layer.

[0119] Step 4: Adjust CNN model parameters

[0120] Specifically, the implementation is as follows: Figure 4 As shown in the flowchart, the CNN model is trained and optimized using the extracted echo signal feature parameters and known pollutant category labels. During training, the model parameters are continuously adjusted to gradually improve the model's fit to the training data while avoiding overfitting. After training, the model will be able to identify pollutant types from unknown echo signals.

[0121] Step 5: Output Results

[0122] Specifically, the implementation is as follows: Figure 4 As shown in the process, after the model training is completed, the echo signal to be identified is input into the model, and the final identification result is obtained through the forward propagation calculation of the model.

[0123] Since the echo signal waveform obtained by the ultrasonic sensor is a mixed exponential waveform, this invention analyzes the waveform parameters of the echo signal, including x = [A, m, T, τ]. By obtaining the waveform rising edge of the actually obtained echo signal that is closest to the rising edge of the standard mixed exponential model, the waveform parameters are obtained, where τ is the offset of the starting point, which is used to calculate the transit time (ToF) for the final calculation of the contaminant depth.

[0124] The depth detection method described in this invention includes five steps: echo signal envelope extraction, echo signal feature point retrieval, construction of a multi-objective function, obtaining the optimal solution of the multi-objective function using the OMOPSO method, and calculation of pollutant depth.

[0125] Step 1: Extract the envelope of the echo signal

[0126] Specifically, this includes wavelet denoising of the echo signal and extracting the envelope function R using Hilbert transform. 0pAnd by smoothing the envelope function through cubic spline interpolation, the final smoothed envelope function R is obtained. p .

[0127] Specifically, the implementation is as follows: Figure 5 As shown in the flowchart, the wavelet denoising process includes using the Nakagami wavelet basis function, selecting appropriate wavelet coefficients, decomposition scale and threshold, using the soft thresholding method to remove signal noise, and finally reconstructing the signal.

[0128] The Hilbert envelope extraction method is used as follows:

[0129] z(i)=H{S p (i)}

[0130] R 0p (i)=|z(i)|

[0131] Among them, H{S p (i)} is the Hilbert operation performed on the echo signal, z(i) is the complex signal, and |z(i)| is the modulus of the complex signal.

[0132] The cubic spline curve interpolation method is as follows:

[0133]

[0134] Among them, R p (i) is the interpolation curve, which is the smoothed envelope, and s(i) is the cubic spline basis function.

[0135] The basis function s(i) can be expressed as:

[0136] s(i)=a i +b i (xx i )+c i (xx i )2+d i (xx i )

[0137] Wherein, coefficient a i b i c i d i The conditions are obtained by interpolation conditions, curve continuity conditions, and continuity conditions of the first and second derivatives of the curve, respectively.

[0138] Step 2: Retrieval of echo signal feature points.

[0139] Specifically, the implementation is as follows: Figure 5As shown in the flowchart, echo signal feature point detection includes detecting the peak points of the echo signal including the corresponding time series points, and retrieving the peak points of the first derivative of the envelope function to obtain the corresponding time series points.

[0140] The time series point corresponding to the peak point detection of the envelope function is feature point 1, which is represented as follows:

[0141] Result[i1]suject to R p (i1)=max(R p (i))

[0142] Here, Result[i1] represents the retrieval result of feature point 1.

[0143] The first derivative of the envelope function shown is used to obtain the peak point, and the corresponding time series point is identified as feature point 2, which is expressed as follows:

[0144]

[0145] Where Result[i2] represents the retrieval result for feature point 1. Indicates envelope R p The first derivative of (i).

[0146] Step 3: Construct three objective functions

[0147] Specifically, the implementation is as follows: Figure 5 As shown in the flowchart, the three objective functions are to minimize the loss function value of the actual echo signal and the fitted exponential function signal, minimize the time series point difference of feature point 1 of the actual echo signal and the fitted exponential function signal, and minimize the time series point difference of feature point 2 of the actual echo signal and the fitted exponential function signal.

[0148] In objective function 1, the expression that minimizes the loss function value for constructing the actual echo signal and the fitted exponential function signal is:

[0149]

[0150] Where loss(x) is the loss function with echo signal parameters x = [A, m, T, τ], and S f This is the fitted function.

[0151] In objective function 2, the expression that minimizes the time series difference between feature point 1 of the actual echo signal and the fitted exponential function signal is:

[0152]

[0153] in, To retrieve the difference between the time series points of the fitted function signal and the actual signal with echo signal parameters x = [A, m, T, τ], R... f is the envelope function of the fitted function signal.

[0154] In objective function 3, the expression that minimizes the time series difference between feature point 1 of the actual echo signal and the fitted exponential function signal is:

[0155]

[0156] in, The feature points of the fitted function signal and the actual signal under the echo signal parameters x=[A,m,T,τ] are 2.

[0157] time series

[0158] Point-based retrieval of differences.

[0159] The final optimization objective based on the three objective functions is:

[0160] minL(x)=[L(x)1,L(x)2,L(x)3]

[0161]

[0162] Here, L(x)1, L(x)2, and L(x)3 are three objective functions.

[0163] Step 4: Obtain the optimal solution for the multi-objective function

[0164] Specifically, the implementation is as follows: Figure 5 and Figure 6 As shown in the flowchart, the key steps in obtaining the optimal solution of a multi-objective function using the OMOPSO method include determining particle velocity, archive size, and crowding.

[0165] The speed of each example is expressed as:

[0166]

[0167] In OMOPSO, C1 and C2 are not fixed values, ω is randomly selected from [0.1, 0.5], c1 and c2 are randomly selected from [1.5, 2], r1 and r2 are random numbers uniformly distributed between 0 and 1, and C1 and C2 are control values. and gbest t The acceleration constant that affects particle velocity. Furthermore, w t It is the inertial weight of particle i, and controls the trade-off between global history and local history. and Let be the position and velocity of particle i at time t.

[0168] After obtaining the velocity of each particle, the position of particle i is updated, and the updated position can be represented as:

[0169]

[0170] The size of the Archive is described as follows:

[0171]

[0172] The size of Archive is determined by ∈-dominance, where the value of ∈ restricts the size of Archive. The value of ∈ is generally defined by the user. If ∈ satisfies the following formula, then the decision vector x1 can be regarded as ∈-dominance x2.

[0173] Crowding density was represented using NSGA-II as follows:

[0174]

[0175] Among them, i d This represents the crowding degree of the i-th non-dominated particle in Archive. This represents the j-th objective function value at point i+1. Let m represent the j-th objective function value at point u-1, where m is the number of objective functions (with a value of 3).

[0176] Step 5: Calculate the depth of contaminants

[0177] Specifically, the implementation is as follows: Figure 5 The flowchart shown in the figure illustrates that the pollutant depth is calculated using the Time-of-Flight (ToF) method, where ToF is the time delay in x = [A, m, T, τ]. The specific method is as follows:

[0178]

[0179] Where ToF1 and ToF2 are the transit times of ultrasound waves under the main road surface and with pollutants, respectively, h is the depth of the pollutants, and v is the distance between the ultrasonic waves and ... u This represents the speed at which ultrasound travels through the air.

Claims

1. A vehicle-mounted method for detecting pollutants on airport runways based on ultrasonic technology, characterized in that, The method is based on the following detection system: Vehicle-mounted ultrasonic sensor; The vehicle-mounted ultrasonic sensor is installed on the vehicle and includes an ultrasonic air probe. The ultrasonic air probe is a single-transmitter, single-receiver air coupling probe. The probe is embedded on a PCB circuit board, and the PCB board is connected to a rotary motor. The vehicle-mounted ultrasonic sensor is installed on a vehicle and includes a single-transmitter, single-receiver air-coupled ultrasonic air probe. The sensor is equipped with lifting and rotating motors to control the rotating probe's detection height and angle. The lifting motor raises and lowers the probe via a lead screw. The probe is embedded in a PCB circuit board and connected to the rotating motor. A waterproof cover is provided on the probe and motor for waterproofing and dustproofing. An aviation connector is provided on the outside of the waterproof cover for signal transmission. The sensor is mounted on the vehicle using a clamp consisting of a detachable steel connecting rod, a fixed metal plate, and hand-tightened screws. The length of the steel rod can be adjusted to accommodate different installation requirements. The pollutant detection information includes pollutant type and depth information. The vehicle-mounted ultrasonic sensor first performs type identification and then performs depth calculation. The type identification includes acquiring echo signal samples of various pollutants, extracting echo signal waveform feature parameters, constructing a CNN deep learning model, adjusting the CNN model parameters, and outputting pollutant type identification results. The echo signal waveform parameters include the amplitude, distortion level, and signal-to-noise ratio of the echo signal; The amplitude of the echo signal is expressed as: A=max[S p (i)] Where A represents the amplitude of the echo signal, S p (t) represents the maximum voltage amplitude at the time series points of the obtained echo signal, where i is the time series point of the echo signal; The degree of distortion is represented by cosine similarity, which is expressed as: Among them, S p ,S d S represents the actual ultrasonic echo signal obtained. d This represents the echo signal of the standard mixed exponential model, where N represents the total number of signal sampling points; The echo signal of the hybrid exponential model is specifically as follows: Among them, t i Here, A0 is the amplitude of the signal, m and T are the inherent parameters of the ultrasonic probe, used to characterize the waveform skewness and kurtosis of the signal, τ is the starting point offset, f0 is the center frequency, and θ is the initial phase. The signal-to-noise ratio of the echo signal is expressed as: Among them, P s and P n These represent the power of the signal and the power of the noise, respectively. The depth detection method includes echo signal envelope extraction, echo signal feature point retrieval, construction of a multi-objective function, obtaining the optimal solution of the multi-objective function and calculating the pollutant depth using the OMOPSO method; The echo signal feature points include: The first characteristic point is the peak point of the envelope function of the echo signal, which is represented as: Result[i1]suject to R p (i1)=max(R p (i)) Wherein, Result[i1] represents the retrieval result of the first feature point in the time series; The second characteristic point is the peak point of the first derivative of the echo signal envelope function, which is expressed as: Wherein, Result[i2] represents the retrieval result of the second feature point in the time series. Indicates envelope R p The first derivative of (i); The multi-objective function includes: The first objective function is to minimize the loss function value of the actual echo signal and the fitted exponential function signal, and its expression is: Where loss(x) is the loss function with echo signal parameters x = [A, m, T, τ], and S f This is the fitted function; The second objective function is to minimize the difference between the first feature points of the actual echo signal and the fitted exponential function signal in the time series. Its expression is: in, To retrieve the difference between the time series points of the fitted function signal and the actual signal with echo signal parameters x = [A, m, T, τ] and the first feature point, R... f The envelope function of the fitted signal; The third objective function, defined as having the smallest difference between the second feature points of the actual echo signal and the fitted exponential function signal on the time series, is expressed as follows: in, To retrieve the difference between the time series points of the second feature point of the fitted function signal and the actual signal under the echo signal parameters x=[A,m,T,τ]; The final optimization objective based on the three objective functions is: minL(x)=[L(x)1,L(x)2,L(x)3] Where L(x)1, L(x)2, and L(x)3 are three objective functions; The steps for obtaining the optimal solution of a multi-objective function using the OMOPSO method include: initializing particle position and velocity parameters, updating and mutating particle position and velocity, storing solutions for non-dominated particles, determining output conditions, and outputting the Pareto optimal solution set. The velocity of each particle is expressed as: Among them, C1 and C2 of OMOPSO are the control... and gbest t The acceleration constants affecting particle velocity, C1 and C2 are random numbers in the range [0,2], ω is a random number, and r1 and r2 are random numbers in the range [0,1] used to increase the randomness of the search process; w t It is the inertial weight of particle i, and controls the trade-off between global history and local history; and Let i be the position and velocity of particle i at time t; After obtaining the velocity of each particle, the position of particle i is updated, and the updated particle position is represented as: The Archive is the solution set of the non-dominated solutions of the multi-objective function. The size of the Archive is determined by ∈, and ∈ satisfies the following equation: The size of Archive is determined by ∈-dominance, where the value of ∈ restricts the size of Archive. The value of ∈ is defined by the user, and the decision vector x1 is ∈-dominancex2. Crowding degree is the density of the distribution of solutions to a multi-objective function in the search space, and it is expressed as follows: Among them, i d This represents the crowding degree of the i-th non-dominated particle in Archive. This represents the j-th objective function value at point i+1. Let represent the j-th objective function value at point i-1, and m be the number of objective functions.

2. The vehicle-mounted airport runway pollutant detection method based on ultrasonic technology according to claim 1, characterized in that, The CNN deep learning model includes convolutional layers, input layers, hidden layers, fully connected layers, and output layers. The convolutional layers are used to extract local features from the echo signal, and the fully connected layers are used to map the extracted features to pollutant categories. The input of the neural network is the waveform parameters of the echo signal, and the output is the type of pollutant.

3. The vehicle-mounted airport runway pollutant detection method based on ultrasonic technology according to claim 1, characterized in that, The process of extracting the envelope of the echo signal employs wavelet denoising, Hilbert transform, and cubic spline curve smoothing to filter the echo signal and obtain a smoothed signal. Specifically: The wavelet denoising process includes using Nakagami wavelet basis functions, selecting wavelet coefficients, decomposition scale and threshold, using soft thresholding to remove signal noise, and finally reconstructing the signal. The Hilbert transform is: z(i)=H{S p (i)} R 0p (i)=|z(i)| Among them, H{S p (i)} is the Hilbert operation performed on the echo signal, z(i) is the complex signal, and |z(i)| is the modulus of the complex signal; The cubic spline curve interpolation method is as follows: Among them, R p (i) is the interpolation curve, which is the smoothed envelope, and s(i) is the cubic spline basis function; The basis function s(i) can be expressed as: s(i)=a i +b i (x-x i )+c i (x-x i )2+d i (x-x i ) Wherein, coefficient a i b i c i d i The conditions are obtained by interpolation conditions, curve continuity conditions, and continuity conditions of the first and second derivatives of the curve, respectively.

Images