An asphalt road mobile load identification method and device, electronic equipment and medium

Abstract

The application discloses a kind of asphalt road mobile load identification method, device, electronic equipment and medium, method includes obtaining load speed, wheel base and asphalt layer temperature;Using the preset strain correction algorithm is corrected to central wavelength data, obtains asphalt road strain parameter;According to vehicle speed and asphalt layer temperature, using the preset numerical simulation method, determine target load size and target load lateral offset;Initial BP neural network model is constructed, and load speed, wheel base, asphalt layer temperature and strain parameter are input into initial BP neural network model, to obtain target BP neural network model;Load speed, wheel base, asphalt layer temperature, strain parameter and target load lateral offset are input into target BP neural network model, to obtain the identification result of asphalt road driving vehicle mobile load.The application solves the technical problem that high-speed mobile load is effectively identified by lacking the non-mapping relationship of mechanical response and load of asphalt road.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of road engineering, and in particular to an asphalt road moving load identification method and device, an electronic device and a storage medium. BACKGROUND

[0002] At present, the highway construction in China develops rapidly, and most of the highways adopt asphalt pavement structure. Many in-service highways in China do not reach the designed service life and end the service prematurely, and have to be maintained on a large scale. The overloading and overspeed of vehicles are one of the important reasons. Therefore, from the perspective of highway management, it is necessary to identify the traffic load information (number of axes, axle distance, load, vehicle speed and driving track, etc.) in the driving process to ensure the normal use of the asphalt road.

[0003] Since the asphalt road belongs to a viscoelastic material, it is difficult to inversely deduce the load size acting on the structure through the structural dynamic characteristics. Therefore, the existing bridge moving load identification technology cannot be applied to the asphalt road. Therefore, it is necessary to find a moving load identification method that can establish a non-mapping relationship between the mechanical response and the load and comprehensively consider various influencing factors. SUMMARY

[0004] The purpose of the present application is to overcome the above technical deficiencies and provide an asphalt road moving load identification method, device, electronic device and storage medium, which solves the technical problem of lacking effective identification of high-speed moving load using the non-mapping relationship between the mechanical response and the load for the asphalt road in the prior art.

[0005] To achieve the above technical purpose, the present application adopts the following technical scheme:

[0006] In a first aspect, the present application provides an asphalt road moving load identification method, comprising:

[0007] obtaining the load speed, axle distance and asphalt layer temperature of a target vehicle;

[0008] correcting the central wavelength data using a preset strain correction algorithm to obtain asphalt road strain parameters;

[0009] determining the target load size and target load lateral offset according to the vehicle speed and asphalt layer temperature using a preset numerical simulation method;

[0010] constructing an initial BP neural network model, inputting the load speed, axle distance, asphalt layer temperature and strain parameters into the initial BP neural network model as input parameters for iterative training to obtain a target BP neural network model;

[0011] The load speed, the wheel base, the asphalt layer temperature, the strain parameter and the load transverse offset are input into the target BP neural network model as input parameters to obtain an identification result of the moving load of the vehicle driving on the asphalt road.

[0012] In some embodiments, the load speed and the wheel base of the target vehicle are acquired based on the fiber grating sensor, and the load speed and the wheel base are acquired based on the fiber grating sensor, and the load speed and the wheel base are acquired based on the fiber grating sensor.

[0013] According to the type of the asphalt road, the positions of the fiber grating sensors are marked and the optical cable is laid;

[0014] The load speed is determined according to the monitoring time interval and the distance between different fiber grating sensors.

[0015] The wheel base is determined according to the signal peak interval obtained by the same fiber grating sensor and the load speed.

[0016] In some embodiments, the target load size and the target load transverse offset are determined by using a preset numerical simulation method, and the target load size and the target load transverse offset are determined by using a preset numerical simulation method.

[0017] A numerical analysis model is established according to the actual structure of the asphalt road.

[0018] Based on the preset dynamic modulus experiment, the viscoelastic parameters are determined according to the viscoelastic characteristics and the correlation between the viscoelastic parameters and the temperature and the vehicle speed.

[0019] The preset load size, the preset load transverse offset, the viscoelastic parameters, the load speed, the asphalt layer temperature and the preset material parameters are input into the numerical analysis model as input parameters, and the load size and the load transverse offset under the strain parameter of the asphalt road are determined as the target load size and the target load transverse offset.

[0020] In some embodiments, before the load speed, the wheel base, the asphalt layer temperature and the strain parameter are input into the initial BP neural network model as input parameters, the input parameters are preprocessed, and the input parameters are preprocessed.

[0021] The input parameters are normalized to obtain normalized input parameters.

[0022] According to the normalized input parameters, a preset regularization algorithm is used to correct a standard error function to obtain a target error function.

[0023] In some embodiments, the target BP neural network model is obtained, and the target BP neural network model is obtained.

[0024] A preset error back propagation algorithm is used to train sample data to obtain an optimized neural network model.

[0025] The optimized neural network model is evaluated by using a preset mean absolute error algorithm.

[0026] In some embodiments, the mean absolute error algorithm can be represented by the following formula:

[0027] Wherein, R identified is an identification parameter; R ture is a real parameter, and n is the number of samples.

[0028] In some embodiments, the strain correction algorithm can be represented by the following formula: y=ax+k T ΔT+b,

[0029] Wherein, a and b are fitting parameters, and k T is the temperature sensitivity coefficient of the sensor, x is the offset of the center wavelength, and y is the strain parameter.

[0030] In a second aspect, the present application further provides an asphalt road moving load identification device, comprising:

[0031] An acquisition module is configured to acquire the load speed, wheelbase and asphalt layer temperature of a target vehicle.

[0032] A strain parameter determination module is configured to correct the center wavelength data by using a preset strain correction algorithm to obtain an asphalt road strain parameter.

[0033] A load parameter determination module is configured to determine the target load size and target load lateral offset by using a preset numerical simulation method according to the vehicle speed and asphalt layer temperature.

[0034] A target model determination module is configured to construct an initial BP neural network model, input the load speed, wheelbase, asphalt layer temperature and strain parameter into the initial BP neural network model as input parameters to obtain a target BP neural network model.

[0035] A load result identification module is configured to input the load speed, wheelbase, asphalt layer temperature, strain parameter and load lateral offset into the target BP neural network model as input parameters to obtain the identification result of the moving load of the vehicle driving on the asphalt road.

[0036] In a third aspect, the present application further provides an electronic device, comprising a processor and a memory.

[0037] The memory stores a computer readable program which can be executed by the processor.

[0038] The processor executes the computer readable program to implement the steps of the asphalt road moving load identification method as described above.

[0039] In a fourth aspect, the present application also provides a computer readable storage medium storing one or more programs, which can be executed by one or more processors to implement the steps of the asphalt road moving load identification method as described above.

[0040] Compared with the prior art, the asphalt road moving load identification method, device, electronic equipment and storage medium provided by the present application first acquire the load speed, wheelbase and asphalt layer temperature of the target vehicle, then obtain the longitudinal strain parameters of the asphalt road according to the preset strain correction algorithm, subsequently determine the target load size and target load lateral offset amount by the numerical simulation method, and then train the initial neural network model with the load speed, wheelbase, asphalt layer temperature and strain parameters as input objects to obtain the target BP neural network, and finally identify the moving load of the road driving vehicle according to the target BP neural network model with the load speed, wheelbase, asphalt layer temperature, strain parameters and target load lateral offset amount as input parameters. The present application improves the identification accuracy of the high-speed moving load by the nonlinear mapping characteristics of the BP neural network model and establishes the relationship between the multi-parameters such as load speed, wheelbase, load lateral offset amount, load size, asphalt layer temperature and strain parameters and the vehicle, reduces the risk of abnormal damage of the asphalt road caused by the overload of the moving vehicle, perfects the vehicle information identification method of the asphalt road intelligent sensing system, and provides a basis for the preventive maintenance of the asphalt road structure. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a flowchart of an embodiment of the asphalt road moving load identification method provided by the present application;

[0042] Figure 2 is a flowchart of an embodiment of step S101 in the asphalt road moving load identification method provided by the present application;

[0043] Figure 3 is a structural schematic diagram of an embodiment of the asphalt road in the asphalt road moving load identification method provided by the present application;

[0044] Figure 4 is a flowchart of an embodiment of step S103 in the asphalt road moving load identification method provided by the present application;

[0045] Figure 5 is a schematic diagram of an embodiment of the asphalt road moving load identification device provided by the present application;

[0046] Figure 6 is a running environment schematic diagram of an embodiment of the electronic equipment provided by the present application. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and do not limit the present application.

[0048] The asphalt road moving load identification method, device, equipment or computer readable storage medium related to the present application is used for asphalt road moving load identification method, and the influence of temperature on the asphalt road and the fiber grating sensor is also considered on the basis of the theory of bridge moving load identification, and the vehicle type is identified preferentially, so that the identification accuracy can still maintain a good level under the condition of high speed.

[0049] The asphalt road moving load identification method provided by the present application, please refer to Figure 1 , comprising:

[0050] S101, acquiring the load speed, the wheelbase and the asphalt layer temperature of the target vehicle;

[0051] S102, correcting the central wavelength data by using a preset strain correction algorithm to obtain the asphalt road strain parameter;

[0052] S103, determining the target load size and the target load lateral offset by using a preset numerical simulation method according to the vehicle speed and the asphalt layer temperature;

[0053] S104, constructing an initial BP neural network model, inputting the load speed, the wheelbase, the asphalt layer temperature and the strain parameter into the initial BP neural network model as input parameters for iterative training to obtain a target BP neural network model;

[0054] S105, inputting the load speed, the wheelbase, the asphalt layer temperature, the strain parameter and the target load lateral offset into the target BP neural network model as input parameters to obtain the identification result of the moving load of the vehicle driving on the asphalt road.

[0055] In the embodiment, firstly, the load speed, the wheel base and the asphalt layer temperature of the target vehicle are acquired, then the strain parameters of the asphalt road in the longitudinal direction are obtained according to a preset strain correction algorithm, subsequently, the target load size and the target load lateral offset are determined through a numerical simulation method, subsequently, the load speed, the wheel base and the asphalt layer temperature and the strain parameters are taken as input objects, the initial neural network model is trained to obtain a target BP neural network, and finally, the load speed, the wheel base, the asphalt layer temperature, the strain parameters and the target load lateral offset are taken as input parameters, and the moving load of the road driving vehicle is identified according to the target BP neural network model; the nonlinear mapping characteristics of the BP neural network model are utilized, and the relationship between the multi-parameters such as the load speed, the wheel base, the load lateral offset, the load size, the asphalt layer temperature and the strain parameters and the vehicle is established, so that the identification precision of the high-speed moving load is improved, the risk of the abnormal damage of the asphalt road caused by the overload of the moving vehicle is reduced, the vehicle information identification method of the asphalt road intelligent sensing system is improved, and a basis for the preventive maintenance of the asphalt road structure is provided.

[0056] It should be noted that the strain parameters of the asphalt road are the load longitudinal strain parameters.

[0057] In some embodiments, referring to Figure 2 , the load speed and the wheel base of the target vehicle are acquired, including:

[0058] S201, according to the type of the asphalt road, marking the position of the fiber grating sensor and laying the optical cable;

[0059] S202, according to the monitoring time interval and the distance between different fiber grating sensors, determining the load speed;

[0060] S203, according to the signal peak interval obtained by the same fiber grating sensor and the load speed, determining the wheel base.

[0061] In the embodiment, different types of asphalt roads are mainly through different vehicles, specifically, the main vehicle types of the asphalt road are monitored according to the traffic volume in previous years, so as to determine the embedding position of the fiber grating sensor.

[0062] In some specific embodiments, the construction site is selected as a test section near a toll station of an in-service highway, the whole section is 1.2 kilometers long, and the middle part includes a toll station exit, a culvert, an uphill and a sharp turn and other special terrains, and the road structure is as follows Figure 3The upper layer (4 cm) is SMA-13, the middle and lower layers (6 cm) are AC-20, the upper base layer (20 cm) is ATB-25, and the lower base layer (40 cm) is a semi-rigid base layer. Two longitudinal FBG sensors are buried along the lane direction. According to the traffic volume data of the test section of the expressway in the past 5 years, combined with the standard model of common trucks, the distance from the center of the vehicle wheel to the edge of the asphalt road surface (about 675 mm) is calculated as the monitoring position, and the monitoring depth is the bottom of each asphalt layer. Temperature sensors are also buried at the same position of each asphalt layer on the shoulder to monitor the real-time temperature. On the one hand, the temperature data will be used as an input parameter, and on the other hand, the temperature compensation will be used to correct the strain data.

[0063] It should be noted that when laying the fiber Bragg grating sensor, the starting end of the FBG sensor is first selected, and the shoulder is cut. The FBG sensor can be pulled from the edge of the asphalt road to the roadside ditch to avoid bending the sensor more than 90°. Then, before the paving process, the specific position of the horizontal and vertical FBG sensors is marked with a marker. In order to improve the survival rate of the sensor and minimize the damage to the road structure caused by the fiber laying, a pre-paved mixture is used to protect and fix the optical cable, and this process is synchronized with the paving process. During the paving process of the asphalt road, the distance between the sensor and the truck or paver needs to be adjusted in real time to avoid direct wheel compaction on the unprotected FBG sensor. Finally, after the compaction process, the FBG sensor is tested for sensitivity and survival rate.

[0064] Further, the laid fiber Bragg grating sensor is used to measure the load speed and wheelbase of the target vehicle.

[0065] In some embodiments, referring to Figure 4 , the preset numerical simulation method is used to determine the target load size and target load lateral offset, including:

[0066] S401, a numerical analysis model is established according to the actual structure of the asphalt road;

[0067] S402, based on the preset dynamic modulus experiment, the viscoelastic parameters are determined according to the correlation between the viscoelastic properties and the temperature and speed.

[0068] S403, the preset load size, the preset load lateral offset, the viscoelastic parameters, the load speed, the asphalt layer temperature and the preset material parameters are used as inputs of the numerical analysis model, and the load size and load lateral offset under the strain parameters of the asphalt road are determined as the target load size and target load lateral offset.

[0069] In the embodiment, by constructing a numerical model, taking the strain parameters of the asphalt road as the target, inputting the viscoelastic parameters, load speed, asphalt layer temperature and preset material parameters, and setting the load size and load lateral offset, the load size and load lateral offset in the current state are determined as the target load size and target load lateral offset by continuously adjusting the load size and load lateral offset to make the output structure of the numerical model consistent with the strain parameters of the asphalt road.

[0070] It should be noted that the material characteristics of the numerical model are shown in the following table:

[0071]

[0072] Further, the viscoelastic properties of the material are characterized by defining the shear modulus in the form of Prony series, which is specifically shown in the following formula:

[0073]

[0074]

[0075] In the formula, g(t) is obtained by normalizing the shear modulus G(t), N is the number of Prony series, g i is the parameter of Prony series, τ i is the delay time. The shear modulus can be obtained by converting the relaxation modulus E(t) by formula (2), and μ is the Poisson's ratio of the material.

[0076] The uniaxial compression dynamic modulus at different temperatures and frequencies is obtained by dynamic modulus test (T0738-2011), and then the isotheral line of the viscoelastic parameters and loading frequency measured at different temperatures is translated along the specified temperature by using the time-temperature equivalence principle, and the master curve of the dynamic modulus is drawn, which is shown in the following formula:

[0077]

[0078]

[0079] In the formula, C1 is a fitting parameter, dimensionless; C2 is a fitting parameter, ℃; T0 is the reference temperature, 20℃. δ is the minimum value of the dynamic modulus, MPa; α is the difference between the maximum and minimum values of the dynamic modulus, MPa; λ, β and γ are the shape parameters of the dynamic modulus master curve, which determine the correlation between the viscoelastic parameters and temperature and vehicle speed, thereby determining the viscoelastic parameters.

[0080] In some embodiments, before the load speed, wheelbase, asphalt layer temperature and strain parameters are input into the initial BP neural network model, the input parameters are preprocessed, including:

[0081] normalizing the input parameters to obtain normalized input parameters;

[0082] correcting a standard error function using a preset regularization algorithm according to the normalized input parameters to obtain a target error function.

[0083] In this embodiment, in order to avoid paralysis of the network in the saturation region of the activation function due to singular sample data during training, the sample needs to be normalized first, and the sample data is mapped to the range of 0-1, as shown in formula 5:

[0084]

[0085] In the formula, x i* is the i-th normalized data; x i is the i-th original data; x max is the maximum value of the original data; and x min is the minimum value of the original data.

[0086] Further, in order to improve the generalization ability of the network, the standard error function of the network is corrected using regularization, as shown in formula 6:

[0087]

[0088] In the formula, γ is a proportionality coefficient, and takes a value (0, 1); is the error sum of squares; n is the number of weights and threshold values, and w is the weight vector.

[0089] In some embodiments, the target BP neural network model is obtained, including:

[0090] training the sample data using a preset error back propagation algorithm to obtain an optimized neural network model;

[0091] evaluating the optimized neural network model using a preset mean absolute error algorithm.

[0092] In this embodiment, the process of training the sample data using the stochastic gradient descent of the error back propagation algorithm is as follows:

[0093] (1) Forward calculation of the net input and activity value of each hidden layer;

[0094] (2) Back propagation calculation of the error term of each layer;

[0095] (3) Calculation of the partial derivative of each layer parameter as shown in formulas 7 and 8, and updating of the weight and bias term until the error rate of the neural network on the validation set is not decreasing.

[0096]

[0097]

[0098] In the formula, is a loss function; W l is the weight matrix of the lth layer; α l-1 is the activity value of the l-1th layer; δ l is the error function of the lth layer.

[0099] Further, in order to evaluate the accuracy of the BP neural network identification, the difference between the identification parameters and the true parameters is measured, and the mean absolute error MAE (Mean Absolute Error) is used, as shown in formula 9:

[0100]

[0101] In the formula, R identified is the identification parameter; R ture is the true parameter.

[0102] In the training process, the parameters of each neural network are adjusted, and finally the parameters are selected to make the MAE reach the minimum, and the BP neural network parameter settings are shown in the following table:

[0103]

[0104] In some embodiments, the strain correction algorithm can be represented by the following formula: y=ax+k T ΔT+b,

[0105] Where a and b are fitting parameters, k T is the temperature sensitivity coefficient of the sensor, x is the offset of the center wavelength, and y is the strain parameter.

[0106] In this embodiment, the FGB sensor is implanted in the rotating compaction test piece, and uniaxial compression test is carried out. The real-time strain of the test piece is measured by the resistance strain sensor, and the measurement results of the FBG sensor are analyzed by regression analysis to obtain the average strain correction curve. During the cooperative deformation test, the influence of temperature on the FBG sensor is considered, the test results are modified by temperature compensation, and finally the average strain correction curve y=1.094x+0.0370ΔT+23.87 is drawn through multiple test results. The real-time strain and temperature data collected by the FBG sensor in the later stage will be calculated through the correction curve. Due to the difference between different vehicle types, the interval between the data peaks is different (the wheelbase is different), so the whole vehicle data needs to be split into single axle data, and the time interval also needs to be converted into time increment steps, so that the measured strain time history data can be matched with the single axle strain time history data of numerical simulation.

[0107] Based on the asphalt road moving load identification method, the asphalt road moving load identification device 500 is correspondingly provided, please refer to Figure 5 The asphalt road moving load identification device 500 comprises an acquisition module 510, a strain parameter determination module 520, a load parameter determination module 530, a target model determination module 540 and a load result identification module 550.

[0108] The acquisition module 510 acquires the load speed, the wheelbase and the asphalt layer temperature of the target vehicle.

[0109] The strain parameter determination module 520 is configured to correct the central wavelength data by using a preset strain correction algorithm to obtain the strain parameter of the asphalt road.

[0110] The load parameter determination module 530 is configured to determine the target load size and the target load lateral offset by using a preset numerical simulation method according to the vehicle speed and the asphalt layer temperature.

[0111] The target model determination module 540 is configured to construct an initial BP neural network model, input the load speed, the wheelbase, the asphalt layer temperature and the strain parameter into the initial BP neural network model as input parameters to obtain a target BP neural network model.

[0112] The load result identification module 550 is configured to input the load speed, the wheelbase, the asphalt layer temperature, the strain parameter and the load lateral offset into the target BP neural network model as input parameters to obtain the identification result of the moving load of the vehicle driving on the asphalt road.

[0113] As shown in Figure 6 Based on the asphalt road moving load identification method, the present application further correspondingly provides an electronic device, which can be a mobile terminal, a desktop computer, a notebook, a palm computer and a server and the like computing device. The electronic device comprises a processor 610, a memory 620 and a display 630. Figure 6 Only part of the components of the electronic device are shown, but it should be understood that all the shown components are not required to be implemented, and more or less components can be alternatively implemented.

[0114] The memory 620 can be an internal storage unit of the electronic device in some embodiments, such as a hard disk or a memory of the electronic device. The memory 620 can also be an external storage device of the electronic device in other embodiments, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, and the like equipped on the electronic device. Further, the memory 620 can include both an internal storage unit and an external storage device of the electronic device. The memory 620 is used to store application software and various data installed on the electronic device, such as program codes installed on the electronic device. The memory 620 can also be used to temporarily store data that has been output or will be output. In an embodiment, the memory 620 stores an asphalt road mobile load identification program 640, which can be executed by the processor 610 to implement the asphalt road mobile load identification method of the embodiments of the present application.

[0115] The processor 610 can be a central processing unit (CPU), a microprocessor, or other data processing chip in some embodiments, and is used to run program codes or process data stored in the memory 620, such as to execute the asphalt road mobile load identification method.

[0116] The display 630 can be an LED display, a liquid crystal display, a touch liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, and the like in some embodiments. The display 630 is used to display information of the asphalt road mobile load identification device and to display a visualized user interface. The components 610-630 of the electronic device communicate with each other through a system bus.

[0117] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware (such as a processor, a controller, and the like) to complete, and the program can be stored in a computer-readable storage medium. The program can include the processes of the above-mentioned embodiments when executed. The storage medium can be a memory, a disk, an optical disk, and the like.

[0118] The specific embodiments of the application described above do not constitute a limitation on the scope of protection of the application. Any various other corresponding changes and modifications made according to the technical concept of the application should be included in the scope of protection of the claims of the application.

Claims

1. A method for identifying a moving load on an asphalt road, characterized by, The method comprises the following steps: obtaining the load speed, wheelbase and asphalt layer temperature of a target vehicle; correcting the central wavelength data by using a preset strain correction algorithm to obtain asphalt road strain parameters; determining the target load size and target load lateral offset according to the speed and asphalt layer temperature by using a preset numerical simulation method; constructing an initial BP neural network model, inputting the load speed, wheelbase, asphalt layer temperature and strain parameters into the initial BP neural network model for iterative training to obtain a target BP neural network model; inputting the load speed, wheelbase, asphalt layer temperature, strain parameters and target load lateral offset into the target BP neural network model to obtain the identification result of the moving load of the vehicle on the asphalt road; obtaining the load speed and wheelbase of a target vehicle, comprising: marking the positions of fiber grating sensors and laying optical cables according to the type of asphalt road; determining the load speed according to the monitoring time interval and distance between different fiber grating sensors; determining the wheelbase according to the signal peak interval obtained by the same fiber grating sensor and the load speed; The method comprises the following steps: establishing a numerical analysis model according to the actual structure of the asphalt road; determining the viscoelastic parameters based on the preset dynamic modulus experiment according to the correlation between the viscoelastic characteristics and the temperature, speed and loading frequency; inputting the preset load size, preset load lateral offset, viscoelastic parameters, load speed, asphalt layer temperature and preset material parameters into the numerical analysis model as input, and determining the load size and load lateral offset under the asphalt road strain parameters as the target load size and target load lateral offset by taking the asphalt road strain parameters as the fitting target; The strain correction algorithm can be represented by the following equation: , where a and b are fitting parameters are fitting parameters, is the temperature sensitivity coefficient of the sensor, x is the offset of the center wavelength, and y is the strain parameter.

2. The asphalt road moving load identification method according to claim 1, characterized by, The method further comprises the following steps before inputting the load speed, wheelbase, asphalt layer temperature and strain parameters into the initial BP neural network model as input parameters: normalizing the input parameters to obtain normalized input parameters; correcting the standard error function by using a preset regularization algorithm according to the normalized input parameters to obtain a target error function.

3. The asphalt road moving load identification method according to claim 2, characterized by, The method comprises the following steps: training sample data by using a preset error back propagation algorithm to obtain an optimized neural network model; evaluating the optimized neural network model by using a preset mean absolute error algorithm.

4. The asphalt road moving load identification method according to claim 3, characterized by, The average absolute error algorithm can be represented by the following equation: , wherein, is the true parameter; is the true parameter, n is the number of samples.

5. An asphalt road mobile load identification device for implementing the asphalt road mobile load identification method according to any one of claims 1 to 4, characterized by, The method comprises the following steps: an obtaining module for obtaining the load speed, wheelbase and asphalt layer temperature of a target vehicle; a strain parameter determination module for correcting the central wavelength data by using a preset strain correction algorithm to obtain asphalt road strain parameters; a load parameter determination module for determining the target load size and target load lateral offset according to the speed and asphalt layer temperature by using a preset numerical simulation method; A target model determining module is configured to construct an initial BP neural network model, input the load speed, the wheelbase, the asphalt layer temperature and the strain parameter into the initial BP neural network model as input parameters for iterative training, and obtain a target BP neural network model; A load result identifying module is configured to input the load speed, the wheelbase, the asphalt layer temperature, the strain parameter and the load lateral offset into the target BP neural network model as input parameters, and obtain an identification result of the moving load of the vehicle driving on the asphalt road.

6. An electronic device, characterized in that: Comprise: A processor and a memory; The memory has stored thereon a computer readable program which can be executed by the processor; The processor executes the computer readable program to implement the steps in the asphalt road moving load identification method according to any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer readable storage medium stores one or more programs which can be executed by one or more processors to implement the steps in the asphalt road moving load identification method according to any one of claims 1-4.

Images