Vehicle weighing prediction method and device based on gravitational neural network

Abstract

The invention provides a vehicle weighing prediction method and device based on a gravitational neural network, and belongs to the technical field of parking equipment detection. According to the invention, the weighing sensor and the laser velometer are used to respectively acquire the weighing signal and the driving speed of the vehicle; the weighing signal comprises a first weighing signal and a second weighing signal; determining a combination ratio of the first weighing signal and the second weighing signal for synthesizing the third weighing signal based on the vehicle driving speed; performing preliminary noise reduction on the third weighing signal based on an EEMD wavelet threshold noise reduction algorithm, and then performing multilayer decomposition on the third weighing signal after preliminary noise reduction by using a primary function of wavelet transform to obtain a weighing signal to be measured; and inputting the to-be-measured weighing signal into the gravitational neural network, predicting the actual weight of the vehicle based on the gravitational neural network, and outputting a prediction result. According to the invention, accurate prediction of the vehicle weight can be realized.

Description

Technical Field

[0001] The invention belongs to the field of parking equipment detection, and in particular relates to a vehicle weighing prediction method and device based on a gravitational neural network. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Mechanical multi-story parking equipment has become a relatively common parking method in various scenarios, but existing mechanical multi-story parking equipment is basically not equipped with a vehicle weight detection device. After the vehicle enters the garage, the garage operates directly. If the vehicle weight exceeds the motor torque or load, the equipment will generate a current overload alarm, which will cause greater damage to the equipment. Moreover, the owner of the car has basically left the garage at this time, and the vehicle cannot leave after the equipment alarm, causing congestion in the garage.

[0004] Subsequently, relevant technicians began to add a device for vehicle weighing detection in front of the mechanical three-dimensional parking equipment, but the inventor found that the existing device for vehicle weighing detection still has some technical problems, such as:

[0005] (1) The existing vehicle weighing detection device only detects the weight of the vehicle by installing several weighing sensors at the bottom of the device. However, this method requires the vehicle to be completely stationary, otherwise the detected vehicle weight will have a large error;

[0006] (2) After the existing vehicle weighing detection devices obtain the vehicle weight signal through the weighing sensor, they lack further processing of the signal. This unprocessed signal contains noise signals that will have a significant impact on the vehicle weight detection, resulting in the devices' prediction accuracy for vehicle weight being unsatisfactory. Summary of the invention

[0007] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a vehicle weighing prediction method and device based on a gravitational neural network, aiming to achieve accurate prediction of vehicle weight.

[0008] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0009] A first aspect of the present invention provides a vehicle weight prediction method based on a gravitational neural network.

[0010] A vehicle weighing prediction method based on gravitational neural network, comprising:

[0011] A weighing signal and a driving speed of the vehicle are obtained respectively by using a weighing sensor and a laser speed meter; the weighing signal includes a first weighing signal and a second weighing signal; a combination ratio of the first weighing signal and the second weighing signal for synthesizing a third weighing signal is determined based on the driving speed of the vehicle;

[0012] Preliminary denoising is performed on the third weighing signal based on the EEMD wavelet threshold denoising algorithm, and then, the third weighing signal after preliminary denoising is decomposed by multilayers using the basis function of wavelet transform to obtain the weighing signal to be measured;

[0013] The weighing signal to be measured is input into a gravity neural network, and the actual weight of the vehicle is predicted based on the gravity neural network and a prediction result is output.

[0014] Further, based on the vehicle driving speed, a combination ratio of the first weighing signal and the second weighing signal for synthesizing the third weighing signal is determined. Specifically, a first speed threshold and a second speed threshold are set, and the first speed threshold is greater than the second speed threshold.

[0015] Furthermore, when the vehicle speed is greater than the set first speed threshold, the combined ratios of the first weighing signal and the second weighing signal used to synthesize the third weighing signal are 85% and 15% respectively; when the vehicle speed is between the set first speed threshold and the second speed threshold, the combined ratios of the first weighing signal and the second weighing signal used to synthesize the third weighing signal are 65% and 35% respectively; when the vehicle speed is lower than the set second speed threshold, the combined ratios of the first weighing signal and the second weighing signal used to synthesize the third weighing signal are both 50%.

[0016] Furthermore, a first filtering frequency threshold and a second filtering frequency threshold are set; after performing multi-layer decomposition on the third weighing signal after preliminary noise reduction using the basis function of wavelet transform, the third weighing signal that is higher than the set first filtering frequency threshold is directly deleted, the third weighing signal that is between the set first filtering frequency threshold and the second filtering frequency threshold is filtered, and the third weighing signal that is lower than the set second filtering frequency threshold is retained.

[0017] A second aspect of the present invention provides a vehicle weight prediction device based on a gravitational neural network.

[0018] A vehicle weighing prediction device based on a gravitational neural network, comprising:

[0019] A data acquisition unit, the data acquisition unit comprising a weighing sensor and a laser speed meter;

[0020] The signal processing unit is configured to: perform preliminary noise reduction on the third weighing signal based on the EEMD wavelet threshold noise reduction algorithm, and then perform multi-layer decomposition on the third weighing signal after preliminary noise reduction using the basis function of the wavelet transform to obtain the weighing signal to be measured;

[0021] An industrial computer is embedded with a gravity neural network and is configured to: input the weighing signal to be measured into the gravity neural network, predict the actual weight of the vehicle based on the gravity neural network and output the prediction result.

[0022] Furthermore, the weighing prediction device also includes a weighing platform, which is placed underground, that is, the surface of the weighing platform is level with the ground; the weighing platform includes multiple weighing columns, and a weighing sensor is arranged under each weighing column; wherein, the weighing sensors on the weighing columns located at the four corners of the weighing platform are the second weighing sensors, and the weighing sensors on the remaining weighing columns are the first weighing sensors; wherein, the first weighing sensor and the second weighing sensor are used to generate a first weighing signal and a second weighing signal, respectively.

[0023] Furthermore, the weighing sensor on each weighing column is connected to the signal processing unit; at the same time, the signal processing unit is connected to the industrial computer.

[0024] Furthermore, along the driving direction of the vehicle on the weighing platform, a laser speed meter is respectively arranged in front of and behind the weighing platform; the two laser speed meters are used to detect the driving speed of the vehicle when passing the weighing platform.

[0025] The third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in a vehicle weight prediction method based on a gravitational neural network as described in the first aspect of the present invention.

[0026] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in a vehicle weight prediction method based on a gravitational neural network as described in the first aspect of the present invention are implemented.

[0027] One or more of the above technical solutions have the following beneficial effects:

[0028] (1) The present invention utilizes a weighing sensor and a laser speedometer to obtain the weighing signal and driving speed of the vehicle respectively; wherein the weighing signal includes a first weighing signal and a second weighing signal; and the combination ratio of the first weighing signal and the second weighing signal used to synthesize the third weighing signal is determined based on the driving speed of the vehicle. That is, when the driving speed of the vehicle is relatively fast, the weighing signal sent by the weighing sensor closer to the middle position of the weighing platform is used as the main basis for predicting the vehicle weight; when the driving speed of the vehicle is relatively slow, the reference level of the weighing signal sent by the weighing sensors at the four corners of the weighing platform is added. Based on this design, the present invention can realize the weight prediction of a moving vehicle on the basis of ensuring the accuracy of the vehicle weight prediction.

[0029] (2) The present invention performs preliminary noise reduction on the third weighing signal based on the EEMD wavelet threshold noise reduction algorithm, and then uses the basis function of the wavelet transform to perform multi-layer decomposition on the third weighing signal after preliminary noise reduction and obtain the weighing signal to be measured; the weighing signal to be measured is then input into the gravity neural network to predict the actual weight of the vehicle. After filtering the weighing signal, the noise signal that has a more obvious impact on the vehicle weight detection can be filtered out. Therefore, the present invention can achieve accurate prediction of the vehicle weight, and then make a scientific judgment on whether the vehicle can enter the garage.

[0030] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0032] Figure 1 This is a flow chart of a vehicle weight prediction method based on a gravitational neural network in Example 1 of the present invention.

[0033] Figure 2 It is a schematic diagram of the structure of the vehicle weighing prediction device in the second embodiment of the present invention.

[0034] Figure 3 This is a schematic diagram of the position structure of the weighing platform and the laser speed meter in the second embodiment of the present invention.

[0035] Figure 4 It is a structural schematic diagram of the installation position of the weighing sensor in the second embodiment of the present invention. DETAILED DESCRIPTION

[0036] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0037] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0038] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0039] Embodiment 1

[0040] This embodiment discloses a vehicle weight prediction method based on a gravitational neural network.

[0041] like Figure 1 As shown, a vehicle weighing prediction method based on gravitational neural network includes:

[0042] Step S1, using a weighing sensor and a laser speedometer to respectively obtain a weighing signal and a driving speed of the vehicle; wherein the weighing signal includes a first weighing signal and a second weighing signal; and determining a combination ratio of the first weighing signal and the second weighing signal for synthesizing a third weighing signal based on the driving speed of the vehicle;

[0043] Step S2, performing preliminary denoising on the third weighing signal based on the EEMD wavelet threshold denoising algorithm, and then performing multi-layer decomposition on the preliminary denoised third weighing signal using the basis function of the wavelet transform to obtain the weighing signal to be measured;

[0044] Step S3: input the weighing signal to be measured into the gravity neural network, predict the actual weight of the vehicle based on the gravity neural network and output the prediction result.

[0045] Based on the above method, the present invention can achieve the weight prediction of a moving vehicle on the basis of ensuring the accuracy of the vehicle weight prediction; at the same time, the noise signal that has a significant impact on the vehicle weight detection can be filtered out through filtering of the weighing signal. Therefore, the present invention can achieve the accurate prediction of the vehicle weight, and then make a scientific judgment on whether the vehicle can enter the garage. In order to facilitate the understanding of the above scheme of this embodiment, the above implementation process is further explained and illustrated, specifically:

[0046] In step S1, the weighing signal and the driving speed of the vehicle are obtained by using a weighing sensor and a laser speed meter respectively, which can be achieved by the following steps:

[0047] Step S1-1, using multiple weighing sensors to collect the weighing signal of the vehicle, there are multiple weighing columns under the weighing platform, and a weighing sensor is arranged under each weighing column. Specifically, there are ten weighing columns under the weighing platform of this embodiment, so ten weighing sensors can be designed. Among them, the weighing sensors located at the four corners of the weighing platform are the first weighing sensors, which are used to send the first weighing signal; the remaining six weighing sensors are the second weighing sensors, which are used to send the second weighing signal.

[0048] The principle of generating the weighing signal is as follows: when the vehicle passes through the weighing platform, the weighing sensor under the weighing platform will be deformed, and after this process, the weighing sensor will convert the received weight signal into an electrical signal; that is to say, the second weighing signal and the third weighing signal are actually electrical signals sent after being processed by the weighing sensor. It is understandable that when the vehicle passes through the weighing platform, the extrusion forces received by each weighing sensor under the weighing platform are actually different, and the existing weighing method is to directly take the average value of the weighing signals collected by these weighing sensors, so the error is relatively large. In order to improve the prediction accuracy, the present invention determines the combination ratio of the first weighing signal and the second weighing signal used to synthesize the third weighing signal based on the vehicle's driving speed, and then uses the third weighing signal for vehicle weighing.

[0049] Step S1-2, using a laser speed meter to obtain the vehicle's speed. Based on the vehicle's travel direction, a laser speed meter is set in front and behind the weighing platform. The laser speed meter emits a laser beam. When a vehicle passes by, the laser beam will irradiate the vehicle; at the same time, the laser speed meter will receive the signal reflected by the laser beam and record the round-trip time of the laser pulse; the vehicle's travel speed can be obtained based on the round-trip time difference obtained by the two laser speed meters.

[0050] Furthermore, a first speed threshold and a second speed threshold are set, wherein the first speed threshold is greater than the second speed threshold. When the vehicle speed is greater than the set first speed threshold, the combination ratios of the first weighing signal and the second weighing signal used to synthesize the third weighing signal are 85% and 15% respectively; when the vehicle speed is between the set first speed threshold and the second speed threshold, the combination ratios of the first weighing signal and the second weighing signal used to synthesize the third weighing signal are 65% and 35% respectively; when the vehicle speed is lower than the set second speed threshold, the combination ratios of the first weighing signal and the second weighing signal used to synthesize the third weighing signal are both 50%. It should be noted that the first speed threshold and the second speed threshold can be set according to actual conditions, and the combination ratios allocated for the above three situations can also be adjusted according to actual needs, and this embodiment does not impose specific restrictions on this.

[0051] Furthermore, the value of the first weighing signal used to synthesize the third weighing signal is the average of the weighing signals sent by the four first weighing sensors; similarly, the value of the second weighing signal used to synthesize the third weighing signal is the average of the weighing signals sent by the other six second weighing sensors.

[0052] In step S2, the third weighing signal is preliminarily denoised based on the EEMD wavelet threshold denoising algorithm, and then the third weighing signal after the preliminarily denoised is multi-layered decomposed using the basis function of the wavelet transform to obtain the weighing signal to be measured; this process can be achieved by the following steps:

[0053] Step S2-1: Preliminary noise reduction is performed on the third weighing signal based on the EEMD wavelet threshold noise reduction algorithm. Specifically:

[0054] Initialize the third weighing signal, set the number of decompositions of the EMD algorithm to m, and the standard deviation of the white noise to α; add a random white noise sequence with a standard deviation of α and zero mean to the original third weighing signal x(t) multiple times, and then perform EMD decomposition on the obtained new sequence to finally obtain the respective residual components and IMF components; perform integrated averaging on the m-times EMD decomposition results to eliminate the influence of white noise, and finally obtain the decomposition result:

[0055]

[0056] Among them, c ij (t) represents the IMF component, c j (t) represents the signal after decomposition by the EMD algorithm. The residual component Res and IMF component after decomposition are preliminarily denoised, and the residual component Res and IMF component after preliminarily denoising are reconstructed. The reconstructed signal is expressed as:

[0057]

[0058] Among them, x ′ (t) represents the reconstructed signal, i.e., the third weighing signal after preliminary noise reduction; c i ′ (t) represents the IMF component after noise reduction, r n ′ (t) represents the residual component after noise reduction.

[0059] Step S2-2: perform multi-layer decomposition on the third weighing signal after preliminary noise reduction using the basis function of wavelet transform and obtain the weighing signal to be measured. Specifically:

[0060] The discrete wavelet transform formula is expressed as:

[0061]

[0062] Where f(t) represents a discrete function, h represents a scaling factor, express The conjugate complex number of , n represents the translation factor, t represents the discrete time parameter; h represents the bandpass filter of a specific frequency band, that is, each h value represents a bandpass filter of a specific frequency band; therefore, by selecting different h values, the corresponding frequency noise and weighing signal can be separated.

[0063] In this embodiment, a first filter frequency threshold and a second filter frequency threshold are set; after the third weighing signal after preliminary noise reduction is decomposed by multi-layer using the basis function of wavelet transform, the third weighing signal higher than the first filter frequency threshold is directly deleted, the third weighing signal between the first filter frequency threshold and the second filter frequency threshold is filtered, and the third weighing signal lower than the second filter frequency threshold is retained. Specifically, the first filter frequency threshold can be set to 1000HZ and the second filter frequency threshold can be set to 250HZ; that is, the signal higher than the first filter frequency threshold is identified as the error of the weighing device itself, and the signal of the filter section can be directly deleted; the signal lower than the third filter frequency threshold is identified as the vibration signal generated by the weighing platform affected by the mass of the car itself when the vehicle passes through the weighing device. The vibration signal has the highest reference significance for calculating the vehicle weight and needs to be fully retained; and the signal based on the first filter frequency threshold and the second filter frequency threshold needs to be filtered.

[0064] Furthermore, this embodiment selects to perform multi-layer decomposition on the third weighing signal after preliminary noise reduction using the DB3 wavelet basis function, and the DB3 wavelet basis function is expressed as:

[0065] [C,L]=wavedec(x4,4,′DB3′);

[0066] Among them, wavedec() represents the wavelet decomposition function, x4 represents the decomposed function (the number of decomposed layers is 4), and DB3 represents the adopted wavelet decomposition function. [C, L] is used to store the approximate and detail coefficients after decomposition, and the detail coefficients are cd4, cd3, cd2, cd1, and ca4; among them, cd4 and cd3 represent the noise of the weighing device itself, cd2 represents the interference generated by the operation of the vehicle engine, and cd1 represents the resonance disturbance noise of the vehicle scale composed of the weighing platform, weighing sensor, and weighing column; ca4 represents the vibration signal generated by the weighing platform when the vehicle passes through the weighing device due to the influence of the vehicle's own mass, that is, the linear change signal really required for the entire filtering process. Therefore, the third weighing signal after preliminary noise reduction is decomposed by multi-layer using the basis function of wavelet transform, and the weighing signal to be measured can be obtained.

[0067] In step S3, the weighing signal to be measured is input into the gravity neural network, and the actual weight of the vehicle is predicted based on the gravity neural network and the prediction result is output.

[0068] The gravitational neural network is a network that optimizes the BP neural network based on the gravitational search algorithm GSA. With the support of the gravitational search algorithm, the initial weights and thresholds in the BP neural network will be further optimized, so that the processing speed and calculation accuracy of the BP neural network for dynamic weighing data can be greatly improved. Among them, the principle of the gravitational search algorithm is: in the process of algorithm-based search, all particles will be affected by gravity and move relative to each other. The movement speed of particles with larger mass in the particle population is slower than that of particles with smaller mass, and the inertial mass and gravity of the particles are determined by the fitness value; therefore, the GSA algorithm regards each particle as a solution to the problem, and the population composed of all particles will form a solution vector.

[0069] The GSA algorithm continuously adjusts and iterates the weight threshold of the BP neural network, and uses the weight threshold of the BP neural network as the position information of the individual particles to search for the optimal solution in the search space. In the specific implementation process, the actual weight of the vehicle can be predicted based on the gravitational neural network through the following process:

[0070] Step S3-1, construct a gravitational search algorithm model.

[0071] First, construct the initial particle population of the gravitational search algorithm, namely:

[0072]

[0073] in, represents the position of the i-th particle in the d-dimensional space, and YN represents the total number of particles in the particle population. In order to obtain the gravitational force (the gravitational force on each particle), the mass of the particle is solved based on the following formula:

[0074]

[0075] Among them, Z i (t) represents the mass of the ith particle, fit i (t) represents the fitness value of the ith particle at the tth iteration, and worst(t) represents the worst fitness value of the particle at the tth moment. According to the law of gravity, the gravitational force that the ith particle can receive from the jth particle can be expressed as:

[0076]

[0077] Among them, M i(t) represents the inertial mass of the i-th particle at the t-th moment and the inertial mass of the force-applying particle, M j (t) represents the inertial mass of the jth particle at the tth moment and the inertial mass of the force particle, R ij (t) represents the Euclidean distance between the i-th particle and the j-th particle at time t. G(t) represents the gravitational constant at time t, G0 represents the initial value of the gravitational constant, t max represents the maximum number of iterations, and α represents the attenuation factor. Furthermore, the gravitational force on the particle is expressed as: Among them, r j Represents a random number in the interval [0,1].

[0078] Then, the acceleration of the i-th particle at the t-th time in the d-dimensional space is defined as:

[0079]

[0080] Under the action of gravity, the speed and position of each particle will change in each iterative optimization process. Specifically, the update process of the speed and position of the i-th particle in the d-dimensional space is as follows:

[0081] v i (t+1)=gen i *v i (t)+A i (t);

[0082] x i (t+1)=x i (t)+v i (t+1);

[0083] Among them, v i (t+1) represents the acceleration of the ith particle at the t+1th time (i.e., the t+1th iteration), x i (t+1) represents the position of the i-th particle at the t+1th iteration, A i (t) represents the acceleration of the i-th particle at the t-th time, gen i Represents a random number in the interval [0,1].

[0084] Step S3-2: In this embodiment, the BP neural network includes an input layer, a hidden layer and an output layer; wherein the number of nodes in the hidden layer is calculated based on the following formula:

[0085]

[0086] Among them, yc represents the number of nodes in the hidden layer, sr represents the number of nodes in the input layer, sc represents the number of nodes in the output layer, and c represents a constant in the interval [1,10]. After repeated verification, it is shown that the model effect is best when the number of hidden layer nodes is 6.

[0087] Step S3-3, parameter initialization. Set the number of individual particles to N = 30 groups, and the maximum number of iterations to t max =600, the target error is 0.002, the moving range of the particle speed is [-0.5, 0.5], and the moving range of the particle position is [-1, 1]; the initial gravitational constant G0 = 10, the attenuation factor α = 0.5, and the initial velocity of the population is randomly initialized.

[0088] The error function f(w) of the BP neural network is used as the fitness value function, and the weight threshold in the BP neural network is selected as the position information of the individual particle, which is recorded as: X = [x1, x2, ..., x YN ]. The fitness value function f(w) is expressed as:

[0089]

[0090] Where u represents the number of output nodes, q s represents the expected output, p s Represents the predicted output.

[0091] Step S3-3, calculate the fitness function value of each particle in the population and select the optimal fitness value; calculate the mass of each particle and update G and worst. Then, calculate the gravity and acceleration of each particle, and optimize and update the speed and position of the particle according to the speed and position update formula of the gravity search algorithm. If the current fitness value is greater than the previous optimal fitness value, update the position of the particle, otherwise do not update. Repeat this iterative process until the maximum number of iterations is reached.

[0092] Specifically, before inputting the parameters after the iteration into the BP neural network, this embodiment sets the number of nodes in the input layer of the BP neural network to 2, namely the vehicle speed and dynamic weighing signal; the static vehicle weight is used as the output layer node, so the hidden layer of the BP neural network is set to a single layer of 4 nodes. At the same time, the trainlm function is selected as the learning function, the tangent function tansing with the minimum error is selected as the transfer function from the input layer to the hidden layer, and the linear function purelin is selected as the transfer function from the hidden layer to the output layer.

[0093] Therefore, when the maximum number of iterations is reached, the particle position corresponding to the latest optimal fitness value (corresponding to the dynamic weighing signal) is used as the optimal individual value, and the dynamic weighing signal and the detected vehicle speed are used as input parameters of the BP neural network (the dynamic weighing signal and the vehicle speed correspond to a node in the input layer respectively) and input into the network model. The static weight of the vehicle, that is, the predicted value of the vehicle weighing, can be output by the output node of the output layer of the BP neural network.

[0094] In the specific implementation process, when the actual load of the parking equipment (the predicted value of the vehicle weighing) exceeds 95% of the rated load, the overload limiter in the garage will send out an alarm signal; when the actual load of the parking equipment (the predicted value of the vehicle weighing) exceeds the rated load by 100% to 110%, the overload limiter will take effect, and the lifting power supply of the parking equipment in the garage should be automatically cut off. Based on the above method, the accurate prediction of the vehicle weight can be achieved, and then a scientific judgment can be made on whether the vehicle can enter the garage.

[0095] Embodiment 2

[0096] This embodiment discloses a vehicle weight prediction device based on a gravitational neural network.

[0097] A vehicle weighing prediction device based on a gravitational neural network, comprising:

[0098] A data acquisition unit, the data acquisition unit comprising a weighing sensor and a laser speed meter;

[0099] The signal processing unit is configured to: perform preliminary noise reduction on the third weighing signal based on the EEMD wavelet threshold noise reduction algorithm, and then perform multi-layer decomposition on the third weighing signal after preliminary noise reduction using the basis function of the wavelet transform to obtain the weighing signal to be measured;

[0100] An industrial computer is embedded with a gravity neural network and is configured to: input the weighing signal to be measured into the gravity neural network, predict the actual weight of the vehicle based on the gravity neural network and output the prediction result.

[0101] like Figure 3 As shown, the vehicle weighing prediction device also includes a weighing platform 10, which is placed underground, that is, the surface of the weighing platform is level with the ground; the weighing platform 10 includes a plurality of weighing columns 7, and a weighing sensor 8 is arranged under each weighing column 7.

[0102] Furthermore, along the driving direction of the vehicle on the weighing platform, a laser speed meter 9 is respectively arranged in front and behind the weighing platform; the two laser speed meters 9 are used to detect the driving speed of the vehicle when passing the weighing platform.

[0103] like Figure 4 As shown, the weighing sensors located on the weighing columns at the four corners of the weighing platform 10 are the second weighing sensors 8, and the weighing sensors on the remaining weighing columns are the first weighing sensors 11; wherein, the first weighing sensor 11 and the second weighing sensor 8 are used to generate a first weighing signal and a second weighing signal, respectively.

[0104] The weighing sensor on each weighing column is connected to the signal processing unit (i.e. Figure 2 At the same time, the signal processing circuit is connected to an industrial computer. Specifically, Figure 2 The four weighing sensors A, B, C, and D shown in the figure are used to represent the second weighing sensors located on the weighing columns at the four corners of the weighing platform 10, and 1 to 6 represent the six first weighing sensors located on the weighing platform 10; it is not difficult to see that the four second weighing sensors and the six first weighing sensors are all connected to the signal processing point unit.

[0105] Embodiment 3

[0106] The purpose of this embodiment is to provide a computer-readable storage medium.

[0107] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in a vehicle weight prediction method based on a gravity neural network as described in the first embodiment of the present disclosure.

[0108] Embodiment 4

[0109] The purpose of this embodiment is to provide an electronic device.

[0110] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, the steps in a vehicle weight prediction method based on a gravity neural network as described in the first embodiment of the present disclosure are implemented.

[0111] The steps involved in the apparatuses of the above embodiments 2, 3 and 4 correspond to the method embodiment 1, and the specific implementation methods can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0112] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0113] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A vehicle weight prediction method based on gravitational neural network, characterized in that: include: A weighing sensor and a laser speed meter are used to obtain a weighing signal and a driving speed of the vehicle respectively; the weighing signal includes a first weighing signal and a second weighing signal; determining a combination ratio of the first weighing signal and the second weighing signal for synthesizing a third weighing signal based on the vehicle travel speed; Preliminary denoising is performed on the third weighing signal based on the EEMD wavelet threshold denoising algorithm, and then, the third weighing signal after preliminary denoising is decomposed by multilayers using the basis function of wavelet transform to obtain the weighing signal to be measured; The weighing signal to be measured is input into a gravity neural network, and the actual weight of the vehicle is predicted based on the gravity neural network and a prediction result is output.

2. A vehicle weight prediction method based on gravitational neural network as claimed in claim 1, characterized in that: The combination ratio of the first weighing signal and the second weighing signal for synthesizing the third weighing signal is determined based on the vehicle speed. Specifically, a first speed threshold and a second speed threshold are set, and the first speed threshold is greater than the second speed threshold.

3. A vehicle weight prediction method based on gravitational neural network as claimed in claim 2, characterized in that: When the vehicle speed is greater than the set first speed threshold, the combined ratio of the first weighing signal and the second weighing signal used to synthesize the third weighing signal is 85% and 15% respectively; when the vehicle speed is between the set first speed threshold and the second speed threshold, the combined ratio of the first weighing signal and the second weighing signal used to synthesize the third weighing signal is 65% and 35% respectively; when the vehicle speed is lower than the set second speed threshold, the combined ratio of the first weighing signal and the second weighing signal used to synthesize the third weighing signal is 50%.

4. A vehicle weight prediction method based on gravitational neural network as claimed in claim 1, characterized in that: A first filter frequency threshold and a second filter frequency threshold are set; after performing multi-layer decomposition on the third weighing signal after preliminary noise reduction using the basis function of wavelet transform, the third weighing signal that is higher than the set first filter frequency threshold is directly deleted, the third weighing signal that is between the set first filter frequency threshold and the second filter frequency threshold is filtered, and the third weighing signal that is lower than the set second filter frequency threshold is retained.

5. A vehicle weight prediction device based on gravitational neural network, characterized in that: include: A data acquisition unit, the data acquisition unit comprising a weighing sensor and a laser speed meter; The signal processing unit is configured to: perform preliminary noise reduction on the third weighing signal based on the EEMD wavelet threshold noise reduction algorithm, and then perform multi-layer decomposition on the third weighing signal after preliminary noise reduction using the basis function of the wavelet transform to obtain the weighing signal to be measured; An industrial computer is embedded with a gravity neural network and is configured to: input the weighing signal to be measured into the gravity neural network, predict the actual weight of the vehicle based on the gravity neural network and output the prediction result.

6. A vehicle weight prediction device based on gravitational neural network as claimed in claim 5, characterized in that: The weighing prediction device also includes a weighing platform, which is placed underground, that is, the surface of the weighing platform is level with the ground; The weighing platform includes multiple weighing columns, and a weighing sensor is arranged under each weighing column; wherein the weighing sensors located on the weighing columns at the four corners of the weighing platform are the second weighing sensors, and the weighing sensors on the remaining weighing columns are the first weighing sensors; wherein the first weighing sensor and the second weighing sensor are used to generate a first weighing signal and a second weighing signal, respectively.

7. A vehicle weight prediction device based on a gravitational neural network as described in any one of claims 5 to 6, characterized in that: The weighing sensor on each weighing column is connected to the signal processing unit; at the same time, the signal processing unit is connected to the industrial computer.

8. The vehicle weight prediction device based on gravitational neural network as claimed in claim 5, characterized in that: Along the driving direction of the vehicle on the weighing platform, a laser speed meter is respectively arranged in front of and behind the weighing platform; the two laser speed meters are used to detect the driving speed of the vehicle when passing the weighing platform.

9. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, the steps in a vehicle weight prediction method based on a gravity neural network as described in any one of claims 1 to 4 are implemented.

10. An electronic device comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the vehicle weight prediction method based on gravity neural network as described in any one of claims 1-4 are implemented.

Images