A highway section vehicle variable speed limit real-time adjustment system and method based on big data

Abstract

The application discloses a highway section vehicle variable speed limit real-time adjustment method and system based on big data, which comprises a face recognition device, a weather detector, a traffic flow detection device, a control center and a variable speed limit display board. According to the driving characteristics of the driver, the current weather quality data and the traffic parameters, a SEM measurement model is constructed, a constraint expression of the latent variable and the manifest variable of the measurement model is established, the driving characteristics, the current weather quality data and the traffic flow parameter data set are normalized and pretreated by using a MinMax scaler, high-dimensional matrix blocks are divided into low-dimensional matrix sets based on a matrix pencil algorithm, and a characteristic matrix is obtained J by constructing the data set J , the characteristic matrix is integrated and solved with the highway section speed limit vehicle speed, and a real-time adjustment vehicle speed is obtained. The application can combine the matrix big data algorithm to realize real-time adjustment control of the fixed speed limit value of the highway section, improve the dynamic ability, realize the variable speed limit in a personalized manner and improve the driving safety.

Description

Technical Field

[0001] This invention belongs to the fields of intelligent road monitoring and control and intelligent algorithm technology, specifically relating to intelligent adjustment systems for speed limits on highways. Background Technology

[0002] In modern transportation systems, highways play a vital role, providing fast and efficient transport routes. However, with increasing traffic density and changing weather conditions, safety risks on highways also rise. To ensure driving safety and optimize traffic flow, it is necessary to adjust fixed speed limits on highway sections in real time.

[0003] Traditional variable speed limit systems on highways are typically based on fixed rules or simple models, failing to consider driver characteristics, current weather quality data, and current traffic parameters. These factors all affect driving safety and efficiency. For example, a driver's habits may be influenced by their driving characteristics, while weather conditions and traffic flow may affect road conditions and relative speeds between vehicles.

[0004] Therefore, a real-time variable speed limit adjustment system for highways is needed that can comprehensively consider these factors. Such a system requires the ability to acquire and process large amounts of data in real time, including driver characteristics, current weather quality data, and current traffic parameters. This data is then used to build a measurement model that describes the factors influencing vehicle speed. Finally, this model is used to adjust the fixed speed limit on highways in real time to ensure driving safety and efficiency. Summary of the Invention

[0005] Purpose of the invention: In view of the problems pointed out in the background art, the present invention proposes a method and system for real-time adjustment of variable speed limits for vehicles on highways based on big data, which realizes real-time adjustment of fixed speed limits for vehicles on highways, improves driving safety, optimizes traffic flow, and promotes information sharing and collaboration.

[0006] Technical solution: This invention discloses a method for real-time adjustment of variable speed limits for vehicles on highways based on big data, comprising the following steps:

[0007] Step 1: Obtain driver's driving characteristics, current weather quality data, and traffic parameter data, and preprocess the data. Driver's driving characteristics include the driver's gender (S1), driving experience (S2), education level (S3), traffic violation penalty records (S4), and sudden illness records (S5). Current weather quality data includes sunny (X1), rainy (X2), foggy (X3), hazy (X4), and sandstorm (X5). Traffic parameter data includes traffic volume (D1), time-averaged speed (D2), and time occupancy rate (D3).

[0008] Step 2: Construct a SEM measurement model, quantify the data in the SEM measurement model, and mathematically express the relationships between latent and manifest variables, and between latent variables in the measurement model.

[0009] Step 3: Fuse driver characteristics, current weather quality data, and current traffic parameter data to construct a feature matrix J = (1 / V m |1 / V c |1 / V k );

[0010] Step 4: Determine the base matrix bundle 1 / V based on the matrix bundle algorithm to analyze the driver's driving characteristics, current weather quality data, and traffic parameter data. m 1 / V c 1 / V k The final characteristic matrix J is determined based on the basic matrix bundle;

[0011] Step 5: Based on the maximum speed limit V of a certain section of the highway. high Minimum speed limit V low The feature matrix J constructed from the dataset is combined with the speed limit of the road segment to obtain the adjusted speed.

[0012] Furthermore, the SEM measurement model constructed in step 2 is as follows:

[0013] S = e s σ+δ

[0014] X = e x γ+ξ

[0015]

[0016] In the formula: S, X, and D are the output parameters of the external signal source; e s To connect the factor loading matrix of the external signal source output index S to the exogenous latent variables, σ represents the exogenous latent variables of the external signal source output index S, and δ represents the measurement error of the external signal source output index S; e x To connect the factor loading matrix of the external signal source output index X to the exogenous latent variables, γ represents the exogenous latent variables of the external signal source output index X, and ξ represents the measurement error of the external signal source output index X; e d To connect the output index D of the external signal source to the factor loading matrix of the exogenous latent variables, Let δ be the exogenous latent variable of the output index D of the external signal source, μ be the measurement error of the output index D of the external signal source; the measurement error range of δ is [-0.02, +0.03], the measurement error range of ξ is [-0.03, +0.03], and the measurement error range of μ is [-0.04, +0.01].

[0017] For SEM measurement models, data quantification is performed, and the relationships between latent and manifest variables, as well as the relationships between latent variables, are expressed mathematically.

[0018] M = e s1 S1+e s2 S2+e s3 S3+e s4 S4+e s5 S5

[0019] C = e X1 X1+e X2 X2+e X3 X3+e X4 X4+e X4 X5

[0020] K = e D1 D1+e D2 D2+e D3 D3

[0021] Where M represents the driver's driving characteristics latent variable, C represents the current weather quality data characteristics latent variable, and K represents the current traffic parameter data characteristics latent variable; e s1 For the driver's gender factor load, e s2 For the driver's driving experience factor load, e s3 For the driver's education level factor load, e s4 For driver violation penalty record factor load, e s5 Record the driver's sudden illness as a factor load; e X1 For the weather quality sunny factor load, e X2 For weather quality rainfall factor load, e X3 For weather quality fog factor load, e X4 For weather quality haze factor load, e x5 For weather quality dust storm factor load; e D1 For traffic volume factor load, e D2 e is the time-averaged velocity factor load. D3 The factor load is the time occupancy rate.

[0022] Furthermore, when processing the acquired data, the following factor loads are determined:

[0023] S1.1 specifies that when the gender of the driver in the main driver's seat, S1, is male, the factor load e s1 The factor load e is 0.26 when the driver's gender S1 in the driver's seat is female. s1 It is 0.14;

[0024] S1.2 stipulates that the driver's driving experience in the main driver's seat is less than 5 years, subject to factor load e. s2The factor load is 0.19, which stipulates that the driver's driving experience S2 is higher than or equal to 5 years in the driver's seat. s2 It is 0.52;

[0025] S1.3 stipulates that the driver's educational level (S3) is below high school level, with a factor load e. s3 The factor load is 0.09, which specifies that the driver's educational level (S3) in the main driver's seat must be higher than the high school level. s3 It is 0.12;

[0026] S1.4 stipulates that the driver in the main driver's seat is subject to penalties for violations. The S4 factor load is less than 5. s4 The value is 0.46, which indicates that the driver's violation penalty record S4 in the driver's seat is higher than the 5th factor load e. s4 It is 0.05;

[0027] S1.5 stipulates that a driver in the main seat must record a sudden illness. S5 records this information based on the load factor e. s5 The load factor is 0.15, indicating that the driver in the main driver's seat has no record of sudden illness. The S5 factor load is e. s5 It is 0.33;

[0028] S1.6 specifies the weather quality (sunny) factor X1 and load e. X1 It is 0.51;

[0029] S1.7 specifies the weather quality rainfall X2 factor load e X2 It is 0.32;

[0030] S1.8 specifies the weather quality fog X3 factor load e X3 It is 0.11;

[0031] S1.9 specifies the haze X4 factor load e for weather quality. X4 It is 0.05;

[0032] S1.10 specifies the weather quality dust storm X5 factor load e x5 It is 0.01;

[0033] S1.11 specifies that traffic volume D1 = Q / T, and the factor load e of traffic volume D1. D1 =0.34; Q is the total number of vehicles within the observation period and observation distance, and T is the single observation period;

[0034] S1.12 specifies the average speed over a specified time. The factor load e of time-averaged velocity D2 D2 0.52; um is the instantaneous speed of the vehicle in m;

[0035] S1.13 specifies the time occupancy rate Factor loading e of time occupancy rate D3D3 The value is 0.37, tm is the occupancy time of the m-th vehicle, and T is the single observation period.

[0036] Furthermore, during the data preprocessing in step 1, the MinMax scaler is used to perform normalization preprocessing on the driving characteristics, current weather quality data, and current traffic parameter dataset.

[0037] Furthermore, in step 4, the basic matrix bundle 1 / V is determined based on the matrix bundle algorithm to identify the driver's driving characteristics, current weather quality data, and traffic parameter data. m 1 / V c 1 / V k The final characteristic matrix J is determined based on the basic matrix bundle. The specific process is as follows:

[0038] S4.1 Based on the matrix bundle algorithm, the SEM measurement model is composed of latent variable components of driving characteristics, latent variable components of weather quality data characteristics, and latent variable components of traffic parameter data characteristics. A Hankel matrix is ​​established based on synchronous data, and the total set of Hankel matrices is decomposed into multiple sub-Hankel matrices. By dividing the Hankel matrix into blocks, dimensionality reduction processing is performed on the matrix type, and overlapping parts of the Hankel matrices are sorted and merged.

[0039] S4.2 performs dimensionality reduction and block partitioning on the high-dimensional matrix data J = (1 / Vm|1 / Vc|1 / Vk), obtaining a latent variable in the J matrix. The element spacing for the latent variable is set to d, and the array is assumed to be uniformly distributed. The input data for the matrix blocks is M = [m1 m2…m ... N ] H Where the input sampled data is the mi matrix block, and W is the covariance matrix of the input sampled data, the covariance matrix is ​​decomposed by dimensionality reduction, and the covariance matrix is ​​defined as the offset constant of the Hankel matrix, resulting in:

[0040]

[0041] S4.3 To find the inverse of the covariance matrix W, we establish an augmented matrix inverse formula O = [o1, o2], which gives us the formula for finding the inverse of the covariance matrix:

[0042]

[0043] in, 1 / V m A base matrix bundle representing driver characteristics;

[0044] S4.4 uses methods S4.2 and S4.3 to solve for the basic matrix bundle of current weather quality and traffic parameter data:

[0045]

[0046]

[0047] in, 1 / V c This is the basic matrix bundle representing the current weather quality;

[0048] 1 / V k This is the basic matrix bundle for current traffic flow parameter data;

[0049] S4.5 constructs the feature matrix J = (1 / Vm|1 / Vc|1 / Vk) based on the dataset obtained from S4.1, S4.3, and S4.4.

[0050]

[0051] Wherein, 1 / V min It is the minimum value of each feature in the data, 1 / V max It is the maximum value of each feature in the data; the feature matrix is ​​constructed from the dataset.

[0052] Furthermore, the speed adjustment (v) is displayed on a variable speed limit sign, allowing drivers to select an appropriate speed.

[0053] This invention also discloses a real-time variable speed limit adjustment system for highway vehicles based on big data, including: a face recognition device, a weather detector, a traffic flow detection device, a control center, and a variable speed limit display board;

[0054] Facial recognition devices are used to acquire drivers' driving characteristics; weather detectors are used to acquire current weather quality data; traffic flow detection devices are used to acquire traffic parameter data.

[0055] The control center is installed in the highway information control center. It uses MinMax scaler, SEM measurement model and matrix bundle algorithm to fit, normalize and reduce the dimensionality of various data parameters to obtain J matrix. Combined with the road section speed limit, the adjustable speed is obtained. The control center is equipped with the steps of the real-time adjustment method for variable speed limit of vehicles on highway sections based on big data as described above.

[0056] Variable speed limit displays are used to show the current variable speed limit value.

[0057] Beneficial effects:

[0058] 1. This invention improves driving safety: The system reduces the risk of accidents by adjusting the vehicle speed in real time to adapt to current road and weather conditions.

[0059] 2. Optimize traffic flow: By rationally adjusting vehicle speeds, the traffic load on highways can be balanced, congestion can be reduced, and road utilization efficiency can be improved.

[0060] 3. Information sharing and collaboration: Integrating information from different data sources (such as driver characteristics, weather quality, and traffic parameters) and using this information for intelligent decision-making promotes information sharing and collaboration between vehicles.

[0061] 4. Reduce computational complexity: Applying the matrix bundle algorithm to process data decomposes high-dimensional matrices into low-dimensional matrices, reducing computational complexity and enabling the system to perform real-time data processing and decision-making more quickly.

[0062] 5. Personalized speed control: The system takes into account the driving characteristics of individual drivers, and can determine the speed in a more personalized way, in line with the driving habits and reaction patterns of different drivers.

[0063] 6. Handling severe weather: By taking into account current weather quality data, the system can automatically adjust vehicle speed under severe weather conditions to ensure safety.

[0064] 7. Dynamic adjustment capability: Because the system relies on real-time data, it can dynamically adjust the speed limit according to the current traffic conditions, providing more flexible and immediate speed control.

[0065] 8. Improve road management efficiency: The system can provide road managers with real-time data and analysis tools to help them monitor and regulate traffic flow more effectively. Attached Figure Description

[0066] Figure 1 This is a system structure diagram of an example of the present invention;

[0067] Figure 2 This is a schematic diagram of the system control method of the present invention.

[0068] Figure 3 This is a schematic diagram of the SEM structural equation of the present invention;

[0069] Figure 4 This is a schematic diagram of the matrix algorithm structure of the present invention;

[0070] Figure 5 This is a flowchart of the system control method of the present invention. Detailed Implementation

[0071] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0072] This invention discloses a method and system for real-time adjustment of variable speed limits for vehicles on highways based on big data. The system includes: a facial recognition device, a weather detector, a traffic flow detection device, a control center, and a variable speed limit display board. The facial recognition device is used to acquire the driver's driving characteristics, the weather detector is used to acquire current weather quality data, the traffic flow detection device is used to acquire traffic parameter data, and the variable speed limit display board is used to display the current variable speed limit value. The control center is installed in the highway information control center. It uses a MinMax scaler, a SEM measurement model, and a matrix bundle algorithm to fit, normalize, and reduce the dimensionality of various data parameters, obtaining a J-matrix which is combined with the road segment speed limit to obtain the adjustable speed. The control center is equipped with the following method for real-time adjustment of variable speed limits for vehicles on highways based on big data.

[0073] First, data acquisition and preprocessing are performed:

[0074] Step 1: Obtain the driver's driving characteristics through facial recognition equipment, including the driver's gender (S1), driving experience (S2), education level (S3), violation penalty records (S4), and sudden illness records (S5).

[0075] Step 2: Obtain current weather quality data through meteorological detectors, including sunny (X1), rain (X2), fog (X3), haze (X4), and sandstorm (X5).

[0076] Step 3: Obtain traffic parameter data through traffic flow detection equipment, including traffic volume (D1), time average speed (D2), and time occupancy rate (D3).

[0077] S1.1 specifies that the gender of the driver in the main driver's seat (S1) must be male when the factor load e s1 The factor load e is 0.26 when the gender of the driver in the main driver's seat (S1) is female. s1 It is 0.14.

[0078] S1.2 stipulates that the driver's driving experience (S2) in the main driver's seat must be less than 5 years, subject to the load e. s2 The factor load is 0.19, which stipulates that the driver's driving experience (S2) in the main driver's seat must be higher than or equal to 5 years. s2 It is 0.52.

[0079] S1.3 stipulates that the driver's educational level (S3) in the main driver's seat must be below high school level. Factor load e s3 The factor load e is 0.09, which stipulates that the driver's educational level (S3) in the main driver's seat must be higher than that of a high school. s3 It is 0.12.

[0080] S1.4 stipulates that the driver in the main driver's seat has fewer than 5 violations of driving regulations (S4) as a factor load e. s4The value is 0.46, which stipulates that the driver's violation penalty record (S4) in the main driver's seat is higher than the 5th factor load e. s4 It is 0.05.

[0081] S1.5 stipulates that a record of a sudden illness of the driver in the driver's seat (S5) must be kept on file due to load e. s5 The factor load is 0.15, which specifies that the driver in the main driver's seat has no record of sudden illness (S5). s5 It is 0.33.

[0082] S2.1 specifies the weather quality (X1) factor load e. X1 It is 0.51.

[0083] S2.2 specifies the weather quality rainfall (X2) factor load e X2 It is 0.32.

[0084] S2.3 specifies the weather quality fog (X3) factor load e X3 It is 0.11.

[0085] S2.4 specifies the haze (X4) factor load e for weather quality. X4 It is 0.05.

[0086] S2.5 specifies the weather quality dust storm (X5) factor load e x5 It is 0.01.

[0087] S3.1 specifies that traffic volume (D1) = Q / T, and the factor load e of traffic volume (D1) is... D1 The value is 0.34. Where Q represents the total number of vehicles within the observation period and observation distance, and T represents a single observation period.

[0088] S3.2 Average speed over specified time Factor load e of time-averaged velocity (D2) D2 The value is 0.52. Where Q represents the number of vehicles within the observation period and observation distance, and um represents the instantaneous speed of vehicle m.

[0089] S3.3 specifies the time occupancy rate Factor loading e of time occupancy (D3) D3 The value is 0.37. Where Q is the number of vehicles within the observation period and observation distance, tm is the occupancy time of the m-th vehicle, and T is the single observation period.

[0090] Secondly, after data acquisition, a SEM measurement model is constructed. The data in the SEM measurement model is then quantified, and the latent and manifest variables, as well as the relationships between latent variables, are mathematically expressed, as follows:

[0091] Step 1: Constructing the SEM measurement model:

[0092] S = e s σ+δ

[0093] X = e x γ+ξ

[0094]

[0095] In the formula: S, X, and D are the output parameters of the external signal source; e s To connect the factor loading matrix of the external signal source output index S to the exogenous latent variables, σ represents the exogenous latent variables of the external signal source output index S, and δ represents the measurement error of the external signal source output index S; e x To connect the factor loading matrix of the external signal source output index X to the exogenous latent variables, γ represents the exogenous latent variables of the external signal source output index X, and ξ represents the measurement error of the external signal source output index X; e d To connect the output index D of the external signal source to the factor loading matrix of the exogenous latent variables, Let δ be the exogenous latent variable of the output index D of the external signal source, μ be the measurement error of the output index D of the external signal source; the measurement error range of δ is [-0.02,+0.03], the measurement error range of ξ is [-0.03,+0.03], and the measurement error range of μ is [-0.04,+0.01].

[0096] Step 2: Quantify the data for the SEM measurement model, and mathematically express the relationships between latent and manifest variables, and between latent variables, as follows:

[0097] M = e s1 S1+e s2 S2+e s3 S3+e s4 S4+e s5 S5

[0098] C = e X1 X1+e X2 X2+e X3 X3+e X4 X4+e X4 X5

[0099] K = e D1 D1+e D2 D2+e D3 D3

[0100] Where M represents the driver's driving characteristics latent variable, C represents the current weather quality data characteristics latent variable, and K represents the current traffic parameter data characteristics latent variable; e s1 For the driver's gender factor load, e s2 For the driver's driving experience factor load, e s3For the driver's education level factor load, e s4 For driver violation penalty record factor load, e s5 Record the driver's sudden illness as a factor load; e X1 For the weather quality sunny factor load, e X2 For weather quality rainfall factor load, e X3 For weather quality fog factor load, e X4 For weather quality haze factor load, e x5 For weather quality dust storm factor load; e D1 For traffic volume factor load, e D2 e is the time-averaged velocity factor load. D3 The time occupancy factor load. See the schematic diagram of the SEM structural equations. Figure 3 .

[0101] Finally, the driver characteristics, current weather quality data, and current traffic parameter data are fused to construct the feature matrix J = (1 / V) m |1 / V c |1 / V k Based on the matrix bundle algorithm, a base matrix bundle 1 / V is used to determine the driver's driving characteristics, current weather quality data, and traffic parameter data. m 1 / V c 1 / V k The final feature matrix J is determined based on the basic matrix bundle. See the schematic diagram of the matrix algorithm structure. Figure 4 .

[0102] Step 1: Use the MinMax scaler to normalize and preprocess the driving characteristics, current weather quality data, and current traffic parameter datasets to ensure the accuracy and completeness of the data. Construct a feature matrix J based on the processed dataset, and fuse the driver characteristics, current weather quality data, and current traffic parameter data to obtain J = (1 / Vm|1 / Vc|1 / Vk).

[0103] Step Two:

[0104] S4.1 Based on the matrix bundle algorithm, the SEM measurement model is composed of latent variable components of driving characteristics, latent variable components of weather quality data characteristics, and latent variable components of traffic parameter data characteristics. A Hankel matrix is ​​established based on synchronous data, and the total set of Hankel matrices is decomposed into multiple sub-Hankel matrices. By dividing the Hankel matrix into blocks, dimensionality reduction processing is performed on the matrix type, and overlapping parts of the Hankel matrices are sorted and merged.

[0105] S4.2 performs dimensionality reduction and block partitioning on the high-dimensional matrix data J = (1 / Vm|1 / Vc|1 / Vk), obtaining a latent variable in the J matrix. The element spacing for the latent variable is set to d, and the array is assumed to be uniformly distributed. The input data for the matrix blocks is M = [m1 m2…m ... N ] H Where the input sampled data is the mi matrix block, and W is the covariance matrix of the input sampled data, the covariance matrix is ​​decomposed by dimensionality reduction, and the covariance matrix is ​​defined as the offset constant of the Hankel matrix, resulting in:

[0106]

[0107] S4.3 To find the inverse of the covariance matrix W, we establish an augmented matrix inverse formula O = [o1, o2], which gives us the formula for finding the inverse of the covariance matrix:

[0108]

[0109] in, 1 / V m The basic matrix bundle for driver characteristics.

[0110] S4.4 uses the methods described in S4.2 and S4.3 above to solve for the basic matrix bundle of the current weather quality and traffic parameter data.

[0111]

[0112] in, 1 / V c This is the basic matrix bundle representing the current weather quality;

[0113] 1 / V k This is the base matrix bundle for current traffic flow parameter data.

[0114] S4.5, based on S4.1, S4.3, and S4.4, constructs the feature matrix J = (1 / Vm|1 / Vc|1 / Vk) from the dataset.

[0115]

[0116] Wherein, 1 / V min It is the minimum value of each feature in the data, 1 / V max It is the maximum value of each feature in the data.

[0117] Obtain the dataset and construct the feature matrix

[0118] Based on the obtained feature matrix J, the maximum speed limit on a certain section of the highway is V. high The minimum speed limit is V.low The feature matrix J constructed from the dataset is combined with the speed limit of the road segment to obtain the adjusted speed.

[0119] The speed adjustment (v) is displayed on the variable speed limit sign so that drivers can select the appropriate speed.

[0120] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent transformations or modifications made in accordance with the spirit and essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for real-time adjustment of variable speed limits for vehicles on highways based on big data, characterized in that, Includes the following steps: Step 1: Obtain driver's driving characteristics, current weather quality data, and traffic parameter data, and preprocess the data. Driver's driving characteristics include the driver's gender (S1), driving experience (S2), education level (S3), traffic violation penalty records (S4), and sudden illness records (S5). Current weather quality data includes sunny (X1), rainy (X2), foggy (X3), hazy (X4), and sandstorm (X5). Traffic parameter data includes traffic volume (D1), time-averaged speed (D2), and time occupancy rate (D3). Step 2: Construct a SEM measurement model, quantify the data in the SEM measurement model, and establish constraint expressions for the latent and manifest variables of the measurement model; The specific SEM measurement model is as follows: S=e s s+d X=e x c+x In the formula: S, X, and D are the output parameters of the external signal source; e s To connect the factor loading matrix of the external signal source output index S to the exogenous latent variables, σ represents the exogenous latent variables of the external signal source output index S, and δ represents the measurement error of the external signal source output index S; e x To connect the factor loading matrix of the external signal source output index X to the exogenous latent variables, γ represents the exogenous latent variables of the external signal source output index X, and ξ represents the measurement error of the external signal source output index X; e d To connect the output index D of the external signal source to the factor loading matrix of the exogenous latent variables, Let δ be the exogenous latent variable of the output index D of the external signal source, μ be the measurement error of the output index D of the external signal source; the measurement error range of δ is [-0.02, +0.03], the measurement error range of ξ is [-0.03, +0.03], and the measurement error range of μ is [-0.04, +0.01]. For SEM measurement models, data quantification is performed, and the relationships between latent and manifest variables, as well as between latent variables, are expressed mathematically. M=e s1 S1+e s2 S2+e s3 S3+e s4 S4+e s5 S5 C=e X1 X1+e X2 X2+e X3 X3+e X4 X4+e X4 X5 K=e D1 D1+e D2 D2+e D3 D3 Where M represents the driver's driving characteristics latent variable, C represents the current weather quality data characteristics latent variable, and K represents the current traffic parameter data characteristics latent variable; e s1 For the driver's gender factor load, e s2 For the driver's driving experience factor load, e s3 For the driver's education level factor load, e s4 For driver violation penalty record factor load, e s5 Record the driver's sudden illness as a factor load; e X1 For the weather quality sunny factor load, e X2 For weather quality rainfall factor load, e X3 For weather quality fog factor load, e X4 For weather quality haze factor load, e x5 For weather quality dust storm factor load; e D1 For traffic volume factor load, e D2 e is the time-averaged velocity factor load. D3 The factor load is the time occupancy rate. Step 3: Fuse driver characteristics, current weather quality data, and current traffic parameter data to construct a feature matrix J = (1 / V m |1 / V c |1 / V k ); Step 4: Based on the matrix bundle algorithm, the high-dimensional matrix block is split into a low-dimensional matrix set to determine the basic matrix bundle 1 / V of the driver's driving characteristics, current weather quality data, and traffic parameter data. m 1 / V c 1 / V k The final characteristic matrix J is determined based on the basic matrix bundle; S4.1 Based on the matrix bundle algorithm, the SEM measurement model is composed of latent variable components of driving characteristics, latent variable components of weather quality data characteristics, and latent variable components of traffic parameter data characteristics. A Hankel matrix is ​​established based on synchronous data, and the total set of Hankel matrices is decomposed into multiple sub-Hankel matrices. By dividing the Hankel matrix into blocks, dimensionality reduction processing is performed on the matrix type, and overlapping parts of the Hankel matrices are sorted and merged. S4.2 For high-dimensional matrix data J = (1 / V m |1 / V c |1 / V k Dimensionality reduction and block partitioning are performed to obtain a latent variable in the J matrix. The element spacing of the latent variable is set to d, and the array is assumed to be uniformly distributed. The input data for the matrix block is M = [m1m2…m N ] H , where m i The input sampled data for the matrix block, where W is the covariance matrix of the input sampled data, is decomposed into a dimensionless covariance matrix and defined as the offset constant of the Hankel matrix, resulting in: S4.3 To find the inverse of the covariance matrix W, we establish an augmented matrix inverse formula O = [o1, o2], which gives us the formula for finding the inverse of the covariance matrix: in, 1 / V m A base matrix bundle for driver characteristics; S4.4 uses methods S4.2 and S4.3 to solve for the basic matrix bundle of current weather quality and traffic parameter data: in, 1 / V c This is the basic matrix bundle representing the current weather quality; 1 / V k This is the basic matrix bundle for current traffic flow parameter data; S4.5, based on S4.1, S4.3, and S4.4, constructs the feature matrix J = (1 / V) from the dataset. m |1 / V c |1 / V k ), Wherein, 1 / V min It is the minimum value of each feature in the data, 1 / V max It is the maximum value of each feature in the data; the feature matrix is ​​constructed from the dataset. Step 5: Based on the maximum speed limit V of a certain section of the highway. high Minimum speed limit V low The feature matrix J constructed from the dataset is combined with the speed limit of the road segment to obtain the adjusted speed.

2. The method for real-time adjustment of variable speed limits for vehicles on highways based on big data, as described in claim 1, is characterized in that... When processing the acquired data, the following factor loads are determined: S1.1 specifies that when the gender of the driver in the main driver's seat, S1, is male, the factor load e s1 The factor load e is 0.26 when the driver's gender S1 in the driver's seat is female. s1 It is 0.14; S1.2 stipulates that the driver's driving experience (S2) is less than 5 years, subject to factor load e. s2 The factor load is 0.19, which stipulates that the driver's driving experience S2 is higher than or equal to 5 years in the driver's seat. s2 It is 0.52; S1.3 stipulates that the driver's educational level (S3) is below high school level, with a factor load e. s3 The factor load is 0.09, which specifies that the driver's educational level (S3) in the main driver's seat must be higher than the high school level. s3 It is 0.12; S1.4 stipulates that the driver in the main driver's seat is subject to penalties for violations. The S4 factor load is less than 5. s4 The value is 0.46, which indicates that the driver's violation penalty record S4 in the driver's seat is higher than the 5th factor load e. s4 It is 0.05; S1.5 stipulates that a driver in the main seat must record a sudden illness. S5 records this information based on the load factor e. s5 The load factor is 0.15, indicating that the driver in the main driver's seat has no record of sudden illness. The S5 factor load is e. s5 It is 0.33; S1.6 specifies the weather quality (sunny) factor (X1) and load e. X1 It is 0.51; S1.7 specifies the weather quality rainfall X2 factor load e X2 It is 0.32; S1.8 specifies the weather quality fog X3 factor load e X3 It is 0.11; S1.9 specifies the haze X4 factor load e for weather quality. X4 It is 0.05; S1.10 specifies the weather quality dust storm X5 factor load e x5 It is 0.01; S1.11 specifies that traffic volume D1 = Q / T, and the factor load e of traffic volume D1. D1 =0.34; Q is the total number of vehicles within the observation period and observation distance, and T is the single observation period; S1.12 specifies the average speed over a specified time. The factor load e of time-averaged velocity D2 D2 0.52; um is the instantaneous speed of the vehicle in m; S1.13 specifies the time occupancy rate Factor loading e of time occupancy rate D3 D3 The value is 0.37, tm is the occupancy time of the m-th vehicle, and T is the single observation period.

3. The method for real-time adjustment of variable speed limits for vehicles on highways based on big data, as described in claim 1, is characterized in that... In step 1, when preprocessing the data, the MinMax scaler is used to normalize the driving characteristics, current weather quality data, and current traffic parameter dataset.

4. A method for real-time adjustment of variable speed limits for vehicles on highways based on big data, as described in any one of claims 1 to 3, characterized in that, The speed adjustment (v) is displayed on the variable speed limit sign so that drivers can select the appropriate speed.

5. A real-time variable speed limit adjustment system for vehicles on highways based on big data, characterized in that, include: Facial recognition equipment, weather detectors, traffic flow detection equipment, control centers, and variable speed limit displays; Facial recognition devices are used to acquire drivers' driving characteristics; weather detectors are used to acquire current weather quality data; traffic flow detection devices are used to acquire traffic parameter data. The control center is installed in the highway information control center. It uses MinMax scaler, SEM measurement model and matrix bundle algorithm to fit, normalize and reduce the dimensionality of various data parameters to obtain J matrix. Combined with the road section speed limit, the speed limit can be adjusted. The control center is equipped with the steps of the real-time adjustment method for variable speed limit of vehicles on highway sections based on big data as described in any one of claims 1 to 4 above. Variable speed limit displays are used to show the current variable speed limit value.

Images