Characteristic variable selection method in highway safety intelligent monitoring system

Abstract

The invention discloses a characteristic variable selection method in a highway safety intelligent monitoring system, and the method comprises the steps: obtaining the vibration information of a highway through a three-axis acceleration sensor, and extracting the time domain characteristics and frequency domain characteristics in the vibration information of the highway; extracting time-frequency domain features in the expressway vibration information by using an empirical mode decomposition method, and constructing a feature variable pool based on the extracted time-domain features, frequency-domain features and time-frequency domain features; performing optimization selection on the characteristic variable pool by adopting a pelican optimization algorithm to form a characteristic variable selection method in the highway safety intelligent monitoring system; according to the pelican algorithm, redundant features are effectively reduced, the training efficiency and prediction precision of the expressway state monitoring system are improved, the expressway state monitoring system has higher real-time performance and adaptability, and the response ability and accuracy to complex road conditions are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of highway monitoring, and in particular to a method for selecting characteristic variables in a highway safety intelligent monitoring system. Background Art

[0002] With the continuous expansion of the transportation network and the increasing demand for the use of highways, the intelligent monitoring system for highway safety plays a vital role in highway management and maintenance. At present, common monitoring systems rely on sensors, cameras and other equipment to collect data such as road vibration and traffic density, and realize real-time monitoring of road conditions and early warning of safety hazards through feature extraction and analysis of these data. Its traditional monitoring methods mainly focus on extracting statistical features from the time domain or frequency domain to assist in identifying the road surface state. However, due to the complex environment of highways, simple time domain and frequency domain feature extraction methods often cannot effectively deal with the nonlinear and non-stationary characteristics of road vibration signals, making key information easily ignored, thus affecting the accuracy and comprehensiveness of monitoring.

[0003] In order to overcome the limitations of traditional methods, multidimensional feature extraction methods that combine time domain, frequency domain and time-frequency domain analysis have gradually emerged in recent years. This type of method captures the dynamic changes of vibration signals at different times and frequencies by introducing time-frequency domain features, which improves the recognition accuracy of road conditions to a certain extent. However, with the increase in the number and dimension of features, the existing feature selection methods mostly rely on manual or experience, and it is difficult to efficiently screen out key variables in the high-dimensional feature space, resulting in a large number of redundant feature variables, which not only increases the system's computational burden, but also affects the reliability and real-time performance of the monitoring model. Therefore, how to automatically screen out the optimal feature combination from massive features to ensure that the highway monitoring system has sufficient accuracy and can efficiently cope with complex road conditions has become one of the main challenges facing the current intelligent monitoring system. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed. Therefore, the present invention provides a method for selecting characteristic variables in a highway safety intelligent monitoring system to solve the problems raised in the background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for selecting characteristic variables in a highway safety intelligent monitoring system, comprising:

[0007] Acquire highway vibration information through a triaxial acceleration sensor, and extract time domain features and frequency domain features from the highway vibration information;

[0008] The empirical mode decomposition method is used to extract the time-frequency domain features in the highway vibration information, and a feature variable pool is constructed based on the extracted time-domain features, frequency-domain features, and time-frequency-domain features.

[0009] The Pelican optimization algorithm is used to optimize the feature variable pool to form a feature variable selection method in the highway safety intelligent monitoring system.

[0010] As a preferred solution of the characteristic variable selection method in the highway safety intelligent monitoring system of the present invention, the time domain characteristics include:

[0011] Dimensional characteristics and dimensionless characteristics;

[0012] The dimensioned features include maximum value, minimum value, peak difference, mean, variance, standard deviation, root mean square, kurtosis and skewness;

[0013] The dimensionless characteristics include peak factor, pulse factor, waveform factor, margin factor and kurtosis factor.

[0014] As a preferred solution of the characteristic variable selection method in the highway safety intelligent monitoring system of the present invention, the empirical mode decomposition method includes:

[0015] A road condition discrimination factor is established, and the number of decomposition layers in the empirical mode decomposition method is set according to the current road conditions on the highway.

[0016] As a preferred solution of the characteristic variable selection method in the highway safety intelligent monitoring system of the present invention, the empirical mode decomposition method is used to extract the time-frequency domain features in the highway vibration information, including:

[0017] The vibration signals in the x-axis, y-axis and z-axis directions are detected respectively by three-axis acceleration sensors. The empirical mode decomposition is used to decompose each single-axis vibration signal into multiple frequency bands. The time-frequency domain features are extracted in each frequency band to obtain the time-frequency domain features of the highway vibration information.

[0018] As a preferred solution of the characteristic variable selection method in the highway safety intelligent monitoring system of the present invention, a characteristic variable pool is constructed based on the extracted time domain features, frequency domain features and time-frequency domain features, including:

[0019] Based on the obtained time domain features, frequency domain features and time-frequency domain features, the time domain features, frequency domain features and time-frequency domain features of different types of vehicles and different types of highway conditions are considered to construct a feature variable pool.

[0020] As a preferred solution of the characteristic variable selection method in the highway safety intelligent monitoring system of the present invention, the characteristic variable pool is optimized and selected by using the Pelican optimization algorithm to form a characteristic variable selection method in the highway safety intelligent monitoring system, including:

[0021] In the initial stage, the position of each pelican in the pelican group is initialized by the upper and lower limits of the pelican search space and a random number;

[0022] In the exploration phase, by comparing the fitness function value of each pelican in the initialized pelican group at the tth iteration and the t+1th iteration, it is determined whether to update the position of each pelican;

[0023] During the development phase, the position of the pelican at t+1 after the update was further adjusted;

[0024] In the convergence stage, the optimal pelican position vector is determined, and the optimal pelican position vector is the vibration information characteristic variable.

[0025] As a preferred solution of the characteristic variable selection method in the highway safety intelligent monitoring system of the present invention, the determination of whether to update the position of each pelican includes:

[0026] At the t+1th iteration, if the fitness function value of the position of the i-th pelican is greater than or equal to the fitness function value of the position of the i-th pelican at the t-th iteration, the position of the i-th pelican is updated;

[0027] At the t+1th iteration, if the fitness function value of the position of the i-th pelican is less than the fitness function value of the position of the i-th pelican at the t-th iteration, the position of the i-th pelican is updated according to the position at the t-th iteration.

[0028] As a preferred solution of the characteristic variable selection method in the highway safety intelligent monitoring system of the present invention, the fitness function value includes:

[0029] Highway state prediction and prediction probability, the highway state prediction is obtained by a feature variable pool, and the prediction probability is obtained by a highway fault model trained by a convolutional neural network.

[0030] Compared with the prior art, the invention has the following beneficial effects:

[0031] 1. The present invention can effectively extract the dynamic change characteristics of vibration signals at different times and frequencies, thereby accurately depicting the multi-level state information of highways; while retaining the basic characteristics of time domain and frequency domain, the vibration signal is decomposed into multiple intrinsic mode components (IMFs) through empirical mode decomposition (EMD), and characteristic information in different frequency bands is extracted, ensuring the comprehensiveness and integrity of vibration characteristics;

[0032] 2. The Pelican optimization algorithm is used to optimize and screen the feature variable pool, automatically screening out redundant and noisy feature variables; this avoids the traditional feature selection method that relies on manual or experience, effectively reduces the dimension of the feature space, and improves the training efficiency and prediction accuracy of the highway status monitoring model, thereby achieving accurate and efficient monitoring results;

[0033] 3. The three-axis acceleration sensor can simultaneously obtain vibration signals in the x, y, and z directions, which can better cope with the complex working conditions and changeable traffic environment on the highway; through the screening and optimization of characteristic variables, the sensitivity and response speed to changes in road conditions are improved, so that the highway safety intelligent monitoring system can maintain stable and reliable monitoring performance under complex road conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0035] Figure 1 It is an overall flow chart of a characteristic variable selection method in a highway safety intelligent monitoring system according to an embodiment of the present invention;

[0036] Figure 2 A flowchart of the Pelican optimization algorithm of a characteristic variable selection method in a highway safety intelligent monitoring system according to an embodiment of the present invention;

[0037] Figure 3 This is a comparison diagram of the training losses of optimized and non-optimized feature models of a feature variable selection method in a highway safety intelligent monitoring system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0039] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0040] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0041] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0042] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0043] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0044] Example 1

[0045] Reference Figure 1 and Figure 2, which is the first embodiment of the present invention, and provides a method for selecting characteristic variables in a highway safety intelligent monitoring system, comprising:

[0046] S1. Obtain highway vibration information through a triaxial acceleration sensor, and extract time domain features and frequency domain features from the highway vibration information;

[0047] Furthermore, the time domain features are divided into dimensional features and dimensionless features;

[0048] Furthermore, dimensional features include maximum value, minimum value, peak difference, mean, variance, standard deviation, root mean square, kurtosis and skewness;

[0049] Specifically, the maximum value X max It is expressed as:

[0050] X max =max{x i} (1)

[0051] Among them, x i is the i-th timing information x detected by the triaxial acceleration sensor;

[0052] Specifically, the minimum value X min It is expressed as:

[0053] X min =min{x i} (2)

[0054] Specifically, the peak difference Vpp is expressed as:

[0055] Vpp=X max -X min =max{x i}-min{x i} (3)

[0056] Specifically, the mean It is expressed as:

[0057]

[0058] Specifically, the variance σ is expressed as:

[0059]

[0060] Among them, N represents a total of N-dimensional feature variables;

[0061] Specifically, the standard deviation s is expressed as:

[0062]

[0063] Specifically, the root mean square Xrms It is expressed as:

[0064]

[0065] Specifically, the kurtosis k is expressed as:

[0066]

[0067] Specifically, the skewness α is expressed as:

[0068]

[0069] Furthermore, dimensionless characteristics include,peak factor, impulse factor, shape factor, margin factor, and kurtosis factor;

[0070] Specifically, the peak factor C is expressed as:

[0071]

[0072] Specifically, the pulse factor I is expressed as:

[0073]

[0074] Specifically, the waveform factor S is expressed as:

[0075]

[0076] Specifically, the margin factor C e It is expressed as:

[0077]

[0078] Specifically, the kurtosis factor C q It is expressed as:

[0079]

[0080] Furthermore, the frequency domain features are divided into mean square frequency, frequency variance, root mean square frequency, center of gravity frequency, frequency standard deviation, etc.;

[0081] S2. Using the empirical mode decomposition method to extract the time-frequency domain features in the highway vibration information, and constructing a feature variable pool based on the extracted time-domain features, frequency-domain features, and time-frequency domain features;

[0082] Furthermore, the vibration signals in the x-axis, y-axis and z-axis directions are detected by a three-axis acceleration sensor, and each single-axis vibration signal is decomposed into multiple frequency bands by using empirical mode decomposition, and the time-frequency domain features are extracted in each frequency band to obtain the time-frequency domain features in the highway vibration information;

[0083] Furthermore, a road condition discrimination factor is established, and the number of decomposition levels in the empirical mode decomposition method is set according to the current road conditions on the highway;

[0084] Exemplarily, each uniaxial vibration signal obtains 4 frequency bands through empirical mode decomposition, and each frequency band contains more than 18 characteristic variables, including frequency domain features and time-frequency domain features; the frequency domain and time-frequency domain characteristic variables extracted from the triaxial vibration signal (x, y, z) exceed 216 in total, in addition to 14 time domain characteristic variables extracted on each axis; therefore, all time domain, frequency domain and time-frequency domain characteristic variables extracted from the triaxial vibration signal are combined, totaling more than 230 characteristic variables;

[0085] It should be noted that if the road conditions on the highway are very complex, the larger the number of decomposition layers in the empirical mode decomposition method, the larger the number of characteristic variables obtained;

[0086] Preferably, in the solution of the present invention, the number of decomposition levels in the empirical mode decomposition method is set to 3, and the penalty factor is set to 400 to represent the road conditions on the highway under normal conditions;

[0087] Specifically, the time-frequency domain feature x(t) is expressed as:

[0088]

[0089] Among them, c i (t) is the eigenmode component, r n (t) is the residual;

[0090] Specifically, the time-frequency domain features also include: envelope entropy, fuzzy entropy, power spectrum entropy, energy entropy, spread entropy, permutation entropy, approximate entropy, singular spectrum entropy and sample entropy;

[0091] Furthermore, based on the obtained time domain features, frequency domain features and time-frequency domain features, the time domain features, frequency domain features and time-frequency domain features under different types of vehicles and different types of highway conditions are considered to construct a feature variable pool;

[0092] Specifically, the status of the expressway can be divided into: normal, vehicle-to-vehicle collision, vehicle-to-guardrail collision, vehicle parking failure, road construction, road collapse, road icing, road congestion, road crosswind weather, etc.

[0093] It should be noted that in actual application scenarios, there are many types of vehicles on the highway, and the characteristic variables corresponding to different types of vehicles and different types of highway states are not the same, resulting in a large characteristic variable pool of vibration information. Therefore, the Pelican optimization algorithm is needed to reduce the characteristic variables from high-dimensional space to low-dimensional space;

[0094] S3. Use the Pelican optimization algorithm to optimize the feature variable pool and form a feature variable selection method in the highway safety intelligent monitoring system;

[0095] Furthermore, in the initial stage, the position of each pelican in the pelican group is initialized by the upper and lower limits of the pelican search space and a random number;

[0096] Specifically, the initial stage is expressed as:

[0097]

[0098] Where M represents a total of M pelicans, i represents the i-th pelican, i = 1, 2, ... M; n represents the n-th characteristic variable, n = 1, 2, ... N; P i represents the position of the i-th pelican, P i =low+rand×(up-low), where low represents the lower limit of the Pelican search space, up represents the upper limit of the Pelican search space, and rand represents a random number between 0 and 1;

[0099] Furthermore, in the exploration phase, by comparing the fitness function value of each pelican in the initialized pelican group at the tth iteration and the t+1th iteration, it is determined whether to update the position of each pelican;

[0100] Specifically, the position of the i-th pelican at the t-th iteration is expressed as:

[0101]

[0102] Where t represents the tth iteration, t+1 represents the t+1th iteration, S represents the location of the prey, and λ randomly selects one from 1 and 2; F(P s ) is the target fitness, F(P i t ) represents the fitness of the i-th pelican;

[0103] Specifically, updating the position of each pelican is expressed as:

[0104]

[0105] Furthermore, the rules for determining whether to update the position of each pelican are as follows:

[0106] Rule 1: At the t+1th iteration, if the fitness function value of the position of the i-th pelican is greater than or equal to the fitness function value of the position of the i-th pelican at the tth iteration, the position of the i-th pelican is updated;

[0107] Rule 2: At the t+1th iteration, if the fitness function value of the position of the i-th pelican is less than the fitness function value of the position of the i-th pelican at the t-th iteration, then the position of the i-th pelican is updated according to the position at the t-th iteration;

[0108] Furthermore, during the development phase, the position of the pelican after the update at the t+1th time is further adjusted;

[0109] Specifically, the position of the i-th pelican at the t+1th time is expressed as:

[0110]

[0111] Where T represents the total number of iterations, and γ represents a constant between 0 and 1;

[0112] It should be noted that, in the development phase, the position of the i-th pelican is also updated using formula (17);

[0113] Furthermore, in the convergence stage, the optimal pelican position vector is determined, and the optimal pelican position vector is the vibration information characteristic variable;

[0114] It should be noted that by using the position of the i-th pelican updated in the development phase, formula (17) is obtained again to determine the optimal pelican position vector;

[0115] Furthermore, highway status prediction and prediction probability,the highway status prediction is obtained by the feature variable pool, and the prediction probability is obtained,through the highway fault model trained by convolutional neural network;

[0116] Specifically, the optimization process of the overall Pelican algorithm and the convolutional neural network training process are referenced Figure 2 ;

[0117] It should be noted that the pelican position needs to be updated in both the exploration and development stages, but their functions are different: the exploration stage is to search the global extensively to find the possible optimal area; the development stage is to conduct a local fine search to optimize the potential optimal solution that has been found; therefore, the update in the exploration stage is to ensure that possible optimal solution areas are not missed, while the update in the development stage is to accurately locate the optimal solution in the area; the two-stage update strategy balances global search and local optimization, which not only increases the probability of finding the global optimal solution, but also avoids the algorithm from falling into the local optimal solution, thereby effectively reducing the dimension of the feature space.

[0118] Example 2

[0119] Reference Figure 3, which is the second embodiment of the present invention, and which provides a method for selecting characteristic variables in a highway safety intelligent monitoring system, including: further illustrating the beneficial effects of the scheme of the present invention by means of simulation experiments;

[0120] In order to solve the problem of high-dimensional feature variable redundancy, the feature variable pool is screened by the Pelican optimization algorithm; the Pelican optimization algorithm is divided into four stages: the initialization stage generates the initial solution for the Pelican group, the exploration stage searches for potential optimal solution areas in the feature space in a large range, the development stage focuses on optimizing local optimal solutions, and finally determines the best feature variable combination in the convergence stage; after multiple iterations of the algorithm, the feature variables are reduced to a low-dimensional space through the optimization strategy, and finally the best feature combination with the number of features ranging from 30 to 50 is screened out; the processed low-dimensional feature combination is input into the highway state recognition model, and the convolutional neural network (CNN) is used for model training; by comparing and analyzing with the traditional full feature variable pool model that has not been optimized, the model training time, recognition accuracy and resource consumption are measured respectively, as shown in Table 1;

[0121] Table 1

[0122]

[0123]

[0124] From the data in Table 1, it can be seen that in the process of model training, the optimized feature model shows a significant reduction in training time, with an average training time reduction of about 35%, and a significant reduction in resource consumption. In addition, from the perspective of recognition accuracy, the average accuracy of the optimized model in the recognition of different working conditions has increased by more than 5%, especially in the complex "vehicle collision" and "road construction" working conditions, the recognition rates after feature optimization have increased by 8% and 7% respectively, effectively coping with the variability of complex road conditions.

[0125] Combination Figure 2 , we can see that the loss value of the optimized model (solid line) decreases significantly faster than that of the unoptimized model (dashed line) in the early and middle stages of training, and finally reaches a lower loss value; this verifies the beneficial effect of feature optimization, that is, by reducing redundant feature variables, the convergence efficiency of the model is improved and the training complexity is reduced;

[0126] Compared with the traditional full-feature input, the feature variable screening scheme after using the Pelican optimization algorithm not only reduces the interference of redundant features, but also makes the highway status monitoring model more accurate and efficient; the optimized low-dimensional feature set effectively reduces the computational complexity while retaining the key features, indicating that the scheme of the present invention is innovative and has significant advantages in improving the real-time and adaptability of highway safety monitoring systems.

[0127] Those skilled in the art will appreciate that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program codes. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0128] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0129] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0131] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0132] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for selecting characteristic variables in a highway safety intelligent monitoring system, characterized in that: include: Acquire highway vibration information through a triaxial acceleration sensor, and extract time domain features and frequency domain features from the highway vibration information; The empirical mode decomposition method is used to extract the time-frequency domain features in the highway vibration information, and a feature variable pool is constructed based on the extracted time-domain features, frequency-domain features, and time-frequency-domain features. The Pelican optimization algorithm is used to optimize the feature variable pool to form a feature variable selection method in the highway safety intelligent monitoring system.

2. The characteristic variable selection method in the highway safety intelligent monitoring system according to claim 1 is characterized in that: The time domain features include: Dimensional characteristics and dimensionless characteristics; The dimensioned features include maximum value, minimum value, peak difference, mean, variance, standard deviation, root mean square, kurtosis and skewness; The dimensionless characteristics include peak factor, pulse factor, waveform factor, margin factor and kurtosis factor.

3. The characteristic variable selection method in the highway safety intelligent monitoring system according to claim 1 is characterized in that: The empirical mode decomposition method comprises: A road condition discrimination factor is established, and the number of decomposition layers in the empirical mode decomposition method is set according to the current road conditions on the highway.

4. The characteristic variable selection method in the highway safety intelligent monitoring system according to claim 3 is characterized in that: The empirical mode decomposition method is used to extract the time-frequency domain features of highway vibration information, including: The vibration signals in the x-axis, y-axis and z-axis directions are detected respectively by three-axis acceleration sensors. The empirical mode decomposition is used to decompose each single-axis vibration signal into multiple frequency bands. The time-frequency domain features are extracted in each frequency band to obtain the time-frequency domain features of the highway vibration information.

5. The characteristic variable selection method in the highway safety intelligent monitoring system according to claim 2 or 4 is characterized in that: Based on the extracted time domain features, frequency domain features, and time-frequency domain features, a feature variable pool is constructed, including: Based on the obtained time domain features, frequency domain features and time-frequency domain features, the time domain features, frequency domain features and time-frequency domain features under different types of vehicles and different types of highway conditions are considered to construct a feature variable pool.

6. The characteristic variable selection method in the highway safety intelligent monitoring system according to claim 5 is characterized in that: The Pelican optimization algorithm is used to optimize the characteristic variable pool to form a characteristic variable selection method in the highway safety intelligent monitoring system, including: In the initial stage, the position of each pelican in the pelican group is initialized by the upper and lower limits of the pelican search space and a random number; In the exploration phase, by comparing the fitness function value of each pelican in the initialized pelican group at the tth iteration and the t+1th iteration, it is determined whether to update the position of each pelican; During the development phase, the position of the pelican at t+1 after the update was further adjusted; In the convergence stage, the optimal pelican position vector is determined, and the optimal pelican position vector is the vibration information characteristic variable.

7. The characteristic variable selection method in the highway safety intelligent monitoring system according to claim 6 is characterized in that: The determining whether to update the position of each pelican includes: At the t+1th iteration, if the fitness function value of the position of the i-th pelican is greater than or equal to the fitness function value of the position of the i-th pelican at the t-th iteration, the position of the i-th pelican is updated; At the t+1th iteration, if the fitness function value of the position of the i-th pelican is less than the fitness function value of the position of the i-th pelican at the t-th iteration, the position of the i-th pelican is updated according to the position at the t-th iteration.

8. The characteristic variable selection method in the highway safety intelligent monitoring system according to claim 6 is characterized in that: The fitness function value includes: Highway state prediction and prediction probability, the highway state prediction is obtained by a feature variable pool, and the prediction probability is obtained by a highway fault model trained by a convolutional neural network.

Images