Road traffic flow speed prediction method and system considering driving behavior heterogeneity and parameter correlation

Abstract

The present invention discloses a method and system for predicting road traffic flow speed that considers driving behavior heterogeneity and parameter correlation. The method comprises: using the variance of traffic flow speed to represent the heterogeneity of drivers' driving behavior, and using a mathematical statistics model to establish a quantitative relationship between the variance of traffic flow speed and vehicle flow density; using a traffic flow model to establish a quantitative relationship between traffic flow speed, speed variance, and vehicle flow density, and constructing a road traffic flow speed prediction model that considers driving behavior heterogeneity; using a correlated random parameter model to represent the correlation of traffic flow parameters, and constructing a road traffic flow speed prediction model that considers the correlation of traffic flow parameters; and using Bayesian updating to simultaneously calibrate all parameters of a two-stage traffic flow model within the same model framework. After multiple iterative updates, the model fit of the two stages can be optimized. The present invention can more accurately predict road traffic flow speed and improve the accuracy of road traffic status prediction.

Description

Technical Field

[0001] The present invention relates to a method and system for predicting road traffic flow speed considering driving behavior heterogeneity and traffic flow parameter correlation, belonging to the technical field of urban road traffic management and control. Background Art

[0002] Uncertainty in road traffic flow speed manifests itself as discreteness and randomness under conditions of constant traffic density. This is closely related to quantifiable factors such as vehicle type, weather conditions, and individual vehicle speed heterogeneity. It is also influenced by random factors such as driver physiological and psychological characteristics, light intensity, wind speed, and traffic events. Differences in driver perception and feedback of traffic information lead to randomness and differentiated driving behavior, resulting in varying vehicle speeds, which in turn leads to volatility and randomness in traffic flow speed, and thus uncertainty in traffic flow.

[0003] Under the same traffic density, traffic flow speeds are not uniquely determined; they vary within a specific range. This relationship between traffic flow speed and density is called traffic flow speed heterogeneity. Traffic flow speed heterogeneity is an inherent property of traffic flow. Differences in drivers' perception and feedback of traffic information and the road environment in congested conditions lead to differentiated driving behaviors such as stop-and-go and frequent acceleration and deceleration, resulting in differentiated driving speeds for individual vehicles. This, in turn, leads to highly volatile and random traffic flow speeds under the same traffic density. Accurately understanding traffic flow speed heterogeneity is key to achieving accurate prediction of traffic flow speed on urban roads.

[0004] Traffic flow speed prediction models are the foundation of urban road traffic control and an important method for predicting traffic flow status. In the study of traditional short-term traffic flow prediction models, scholars have studied the stability of traffic flow operation status by establishing a deterministic relationship between traffic volume, traffic flow speed, and vehicle flow density, established short-term traffic flow prediction models, and achieved rich research results. However, existing traffic flow speed prediction models mainly focus on the deterministic relationship between traffic flow speed and vehicle flow density, lacking research on the uncertainty relationship between basic elements and key parameters of traffic flow, and lacking research on the correlation between traffic flow parameters. They fail to accurately grasp the operating characteristics and evolution of traffic flow status, affecting the accuracy of traffic flow status prediction and restricting the application of traffic flow speed prediction models in urban road engineering practice. Summary of the Invention

[0005] Purpose of the invention: The purpose of the present invention is to provide a method and system for predicting road traffic flow speed that takes into account the heterogeneity of driving behavior and the correlation of parameters. The method considers the impact of differentiated driving behavior caused by observable and unobservable factors on traffic flow speed, as well as the correlation between traffic flow parameters to construct a traffic flow speed prediction model to improve the accuracy of road traffic flow speed prediction, which will help traffic design and management departments optimize urban road traffic design and transformation plans, formulate reasonable and effective real-time control and induction strategies, and have application value in alleviating urban road traffic congestion.

[0006] Technical solution: The above objectives are achieved through the following technical solutions:

[0007] A road traffic flow speed prediction method considering driving behavior heterogeneity and parameter correlation is proposed. A two-stage traffic flow model is used to construct a relationship function between traffic flow speed and traffic density. Bayesian inference is used to estimate parameters of the two-stage traffic flow model within the same model framework. The method includes the following steps:

[0008] The variance of traffic flow speed is used to represent the heterogeneity of drivers’ driving behavior, and the variance of traffic flow speed within a preset time period is established. The relationship function with the traffic density k is defined as g(k|β1,…,β n ), where g(.) is the relationship function, β1,…,β n is a parameter;

[0009] The traffic flow model is used to establish the relationship between the expected value of traffic flow speed μ and density k. The relationship model is μ=f(k|ω1,ω2,λ1,…,λ n )+ε h (k), where ω1 is the free stream velocity u f , ω2 is the congestion density k j Or the optimal density k o ,λ1…,λ n is the characteristic parameter of the traffic flow model, ε h (k) is the error related to density;

[0010] A linear model is used to establish the relationship between the traffic flow speed observation value u and the speed expectation value μ. The relationship model is u=μ+ε, where ε is the random error.

[0011] Correlated random parameter distribution is used to represent the correlation of traffic flow parameters. A road traffic flow speed prediction model considering the correlation of traffic flow parameters is constructed. The two parameters ω1 and ω2 in the traffic flow speed and density relationship model obey the bivariate normal distribution MVN2, which is expressed as follows:

[0012]

[0013] where μω1 ,μ ω2 and is the mean and variance of ω1, ω2, cov(ω1, ω2) is the covariance of ω1, ω2; the expression of the correlation coefficient ρ of ω1, ω2 is:

[0014]

[0015] The normal distribution N is used to represent the error of traffic flow speed, and the error related to traffic density ε h (k) obeys the normal distribution N(0,g(k|β1,…,β n )), the random error ε follows the normal distribution N(0,σ 2 ), where σ 2 is the variance of ε;

[0016] Bayesian updating is used to estimate model parameters;

[0017] The traffic flow speed is predicted using a two-stage traffic flow model, and the road traffic flow speed prediction value considering the heterogeneity of vehicle speed is The expression is:

[0018]

[0019] in Is f in The second-order partial derivative at .

[0020] Furthermore, the Bayesian updating is used to estimate the parameters of the model, including:

[0021] The variance of traffic flow speed follows the normal distribution N, and its expression is:

[0022]

[0023] in and The traffic flow speed variance is The expectation and variance of

[0024] Traffic flow speed follows normal distribution N, which is expressed as:

[0025] u~N(μ,σ 2 )

[0026] where μ and σ 2 are the expectation and variance of the traffic flow speed u, μ=f(k|ω1,ω2,λ1,…,λ n )+ε h (k);

[0027] The Bayesian updating method is used to update the parameters β1,…,β n ,λ1,…,λ n , cov(ω1,ω2), σ 2 Make an estimate.

[0028] Furthermore, when estimating the parameters of the two-stage traffic flow model, the posterior function The expression is:

[0029] in is the probability of u, from the normal distribution N(μ,σ 2 )get, is the probability of ω1,ω2, obtained from the bivariate normal distribution MVN2, is ε h (k), from the normal distribution N(0,g(k|β1,…,β n ))get, yes The probability of get, p(λ1)…p(λ n ), p(σ 2 ) is β1…β n , cov(ω1,ω2),λ1…λ n ,σ 2 The prior distribution of cov(ω1,ω2) estimates the correlation coefficient ρ between ω1 and ω2.

[0030] Furthermore, the prior distribution of the parameters in the two-stage traffic flow model adopts normal distribution N(0,10000), inverse gamma distribution IG(0.01,0.01) and inverse Wishart distribution Its expression is β1~N(0,10000),…,β n ~N(0,10000),λ1~N(0,10000),…,λ n ~N(0,10000), σ 2 ~IG(0.01,0.01),

[0031] Preferably, the g(k|β1,…,β n ) is defined as:

[0032]

[0033] Preferably, the f(k|ω1,ω2,λ1,…,λ n ) is defined as:

[0034]

[0035] where u f and k j are the free flow speed and congestion density parameters of the traffic flow model, C j is the motion wave speed at the congestion density.

[0036] A road traffic flow speed prediction system considering driving behavior heterogeneity and parameter correlation includes: a traffic flow model creation module, a parameter estimation module and a traffic flow speed prediction module;

[0037] The traffic flow model creation module is used to use the variance of traffic flow speed to represent the heterogeneity of drivers' driving behavior and establish the variance of traffic flow speed within a preset time period. The relationship function with the traffic density k is defined as g(k|β1,…,β n ), where g(.) is the relationship function, β1,…,β n is a parameter; the traffic flow model is used to establish the relationship between the expected value of traffic flow speed μ and density k, and the relationship model is μ=f(k|ω1,ω2,λ1,…,λ n )+ε h (k), where ω1 is the free stream velocity u f , ω2 is the congestion density k j Or the optimal density k o ,λ1,…,λ n is the characteristic parameter in the traffic flow model, ε h (k) is the error related to density. A linear model is used to establish the relationship between the observed traffic flow speed u and the expected speed μ. The relationship model is u=μ+ε, where ε is the random error. The correlation of traffic flow parameters is represented by the distribution of correlated random parameters. A road traffic flow speed prediction model considering the correlation of traffic flow parameters is constructed. The two parameters ω1 and ω2 in the relationship model obey the bivariate normal distribution MVN2, and the expression is:

[0038]

[0039] in and are the mean and variance of ω1, ω2, and cov(ω1, ω2) is the covariance of ω1, ω2;

[0040] And, the normal distribution N is used to represent the error of traffic flow speed, and the error related to traffic density ε h(k) obeys the normal distribution N(0,g(k|β1,…,β n )); random error ε obeys the normal distribution N(0,σ 2 ), where σ 2 is the variance of ε;

[0041] The parameter estimation module is used to estimate the parameters of the model using Bayesian updating;

[0042] The traffic flow speed prediction module is used to predict the traffic flow speed using a two-stage traffic flow model, taking into account the heterogeneity of vehicle speeds. The expression is:

[0043]

[0044] in Is f in The second-order partial derivative at .

[0045] A computer system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded into the processor, the steps of the method for predicting road traffic flow speed considering the heterogeneity of vehicle speeds are implemented.

[0046] Beneficial effects: The present invention proposes a method for predicting road traffic flow speed that takes into account the heterogeneity of driving behavior and the correlation of parameters. It takes into account the heterogeneity of traffic flow speed caused by the differentiated driving behavior of drivers under the same density and traffic environment, the heterogeneity of traffic flow speed caused by uncertain factors such as traffic accidents, strong winds, and glare effects, and the correlation between traffic flow parameters such as free flow speed, optimal density, and congestion density. Mathematical statistics models and related random parameter models are used to establish a quantitative relationship between traffic flow speed, speed variance and traffic density, and to construct a road traffic flow speed prediction model that takes into account the heterogeneity of driving behavior and the correlation of parameters. Bayesian updating is used to calibrate all parameters of the two-stage traffic flow model within the same model framework. After multiple iterative updates, the model fitting goodness of the two stages can be optimized at the same time. Experiments show that the present invention takes into account the heterogeneity of traffic flow speed and the correlation of traffic flow parameters, and predicts road traffic flow speed more accurately, thereby improving the accuracy of road traffic status prediction, and has application value in alleviating urban road traffic congestion. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the correlation between driving behavior, traffic flow density and individual vehicle speed in the present invention.

[0048] Figure 2 It is a schematic diagram of the model construction framework of the present invention. DETAILED DESCRIPTION

[0049] The present invention is further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, modifications of various equivalent forms of the present invention made by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0050] An embodiment of the present invention discloses a method for predicting road traffic flow speed considering driving behavior heterogeneity and parameter correlation, comprising the following steps:

[0051] In the first step, a traffic flow model is constructed based on traffic flow speed observations, speed variance and traffic density.

[0052] The variance of traffic flow speed is used to represent the heterogeneity of drivers' driving behavior. The variance of traffic flow speed within a preset time period is established using a mathematical statistical model. The relationship function with the traffic density k is defined as g(k|β1,…,β n ), where g(.) is the relationship function, β1,…,β n is a parameter;

[0053] The traffic flow model is used to establish the relationship between the expected value of traffic flow speed μ and density k. The relationship model is μ=f(k|ω1,ω2,λ1,…,λ n )+ε h (k), where ω1 is the free stream velocity u f , ω2 is the congestion density k j Or the optimal density k o ,λ1…,λ n is the characteristic parameter of the traffic flow model, ε h (k) is the error related to density;

[0054] A linear model is used to establish the relationship between the traffic flow speed observation value u and the speed expectation value μ. The relationship model is u=μ+ε, where ε is the random error.

[0055] Correlated random parameter distribution is used to represent the correlation of traffic flow parameters. A road traffic flow speed prediction model considering the correlation of traffic flow parameters is constructed. The two parameters ω1 and ω2 in the traffic flow speed and density relationship model obey the bivariate normal distribution MVN2, which is expressed as follows:

[0056]

[0057] where μ ω1 ,μ ω2 and is the mean and variance of ω1, ω2, cov(ω1, ω2) is the covariance of ω1, ω2; the expression of the correlation coefficient ρ of ω1, ω2 is:

[0058]

[0059] The normal distribution N is used to represent the error of traffic flow speed, and the error related to traffic density ε h (k) obeys the normal distribution N(0,g(k|β1,…,β n )), the random error ε follows the normal distribution N(0,σ 2 ), where σ 2 is the variance of ε;

[0060] Next, we consider the free flow velocity u f , congestion density k j and the motion wave velocity C at the congestion density j Traffic flow parameters are used for example.

[0061] The variance of traffic flow speed within a preset time period t (e.g., 2 minutes) The functional expression of the relationship between θ and traffic density k is shown in formula (1):

[0062]

[0063] The traffic flow model is used to establish the relationship between the expected value of traffic flow speed μ and density k. The relationship function is shown in formula (2):

[0064]

[0065] The linear model is used to establish the relationship between the traffic flow speed observation value u and the speed expectation value μ. The relationship model is shown in formula (3):

[0066]

[0067] Using a correlated random parameter structure, the parameter free flow speed u f and congestion density k j It obeys the bivariate normal distribution MVN2, and its expression is shown in formula (4):

[0068]

[0069] in and is u f ,k j The mean and variance of cov(u f ,k j ) is u f ,k j The covariance of uf ,k j The expression of the correlation coefficient ρ is:

[0070]

[0071] Density-related error ε h (k) obeys the normal distribution N(0,g(k|β1,β2)), and the random error ε obeys the normal distribution N(0,σ 2 ).

[0072] In the second step, the Bayesian updating method is used to estimate the parameters of the two-stage model within the same model framework.

[0073] The traffic flow speed variance obeys the normal distribution N, and its expression is shown in formula (6):

[0074]

[0075] in and The traffic flow speed variance is The expectation and variance of According to Bayesian theory, the posterior function It is expressed as shown in formula (7):

[0076]

[0077] in yes The likelihood, from a normal distribution Obtain, p(β1), p(β2) and They are β1, β2 and The prior distribution of

[0078] Traffic flow speed obeys the normal distribution N, and its expression is shown in formula (8):

[0079] u~N(μ,σ 2 )#(8)

[0080] where μ and σ 2 are the expectation and variance of the mean u of traffic flow speed, Posterior distribution It is expressed as shown in formula (9):

[0081]

[0082] in is the probability of u, from the normal distribution N(μ,σ 2 )get, is u f ,kj The probability of , obtained from the bivariate normal distribution MVN2, is ε h The probability of (k) is obtained from the normal distribution N(0,g(k|β1,β2)), is σ h The probability of 2(k), from the normal distribution Obtain, p(β1), p(β2), p(C j ), p(σ 2 ) is β1,β2, cov(u f ,k j ), C j ,σ 2 The prior distribution of .

[0083] Using normal distribution N(0,10000), inverse gamma distribution IG(0.01,0.01) and inverse Wishart distribution As model parameters β1, β2, C j , σ 2 , cov(u f ,k j ) is a priori function, and its expression distribution is shown in formulas (10)-(13):

[0084] β1~N(0,10000),β2~N(0,10000)#(10)

[0085]

[0086]

[0087]

[0088] The Bayesian updating method is used to update the parameters β1, β2, cov(u f ,k j ), C j ,σ 2 Make an estimate based on cov(u f ,k j ) Estimated value calculation u f ,k j The correlation coefficient ρ.

[0089] The third step is to predict traffic flow speed using a two-stage traffic flow model.

[0090] Prediction of road traffic flow speed considering vehicle speed heterogeneity The expression is shown in formula (14):

[0091]

[0092] in Is f in The second-order partial derivative at .

[0093] Calculation Case

[0094] In order to prove the accuracy of the present invention, traffic flow data of major roads in Hong Kong, China were selected for case analysis. The descriptive characteristics of traffic flow speed, speed variance and traffic density are shown in Table 1.

[0095] Table 1. Descriptive characteristics of traffic data

[0096]

[0097] A two-stage traffic flow model considering driving behavior heterogeneity is constructed, and the Bayesian updating method is used to estimate the parameters of the two-stage traffic flow model considering driving behavior heterogeneity. The model results are shown in Table 2.

[0098] Table 2. Parameter estimation results of random parameter traffic flow model

[0099]

[0100]

[0101] According to the parameter estimation results, the road section traffic flow speed prediction model is obtained as shown in formula (15):

[0102]

[0103] The traffic flow speed is predicted using formula (15), and the results are shown in Table 3.

[0104] Table 3. Traffic flow speed prediction results

[0105]

[0106]

[0107] Based on the same inventive concept, an embodiment of the present invention discloses a road traffic flow speed prediction system that considers driving behavior heterogeneity and parameter correlation, including: a traffic flow model creation module, a parameter estimation module, and a traffic flow speed prediction module; the traffic flow model creation module is used to use the variance of traffic flow speed to represent the heterogeneity of driver driving behavior, and establish the variance of traffic flow speed within a preset time period. The relationship function between traffic flow density k is established; the traffic flow model is used to establish the relationship between the expected value μ of traffic flow speed and density k; the linear model is used to establish the relationship between the observed value u of traffic flow speed and the expected value μ of speed; the correlation of traffic flow parameters is represented by the distribution of related random parameters, and a road traffic flow speed prediction model considering the correlation of traffic flow parameters is constructed; and the normal distribution N is used to represent the error of traffic flow speed, and the error related to traffic density ε h (k) obeys the normal distribution N(0,g(k|β1,…,β n )); random error ε obeys the normal distribution N(0,σ 2 ), where σ 2 is the variance of ε; the parameter estimation module is used to estimate the parameters of the model using Bayesian updating; the traffic flow speed prediction module is used to predict the traffic flow speed using a two-stage traffic flow model. The specific implementation details of each module are consistent with those of the above method embodiment and are not repeated here.

[0108] Based on the same inventive concept, an embodiment of the present invention discloses a computer system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded into the processor, the steps of the method for predicting road traffic flow speed considering the heterogeneity of vehicle driving speeds are implemented.

Claims

1. A method for predicting road traffic flow speed considering driving behavior heterogeneity and parameter correlation, characterized by: The following steps are involved: The variance of traffic flow speed is used to represent the heterogeneity of drivers’ driving behavior, and the variance of traffic flow speed within a preset time period is established. The relationship function with the traffic density k is defined as g(k|β1,…,β n ), where g(.) is the relationship function, β1,…,β n is a parameter; The traffic flow model is used to establish the relationship between the expected value of traffic flow speed μ and density k. The relationship model is μ=f(k|ω1,ω2,λ1,…,λ n )+ε h (k), where ω1 is the free stream velocity u f , ω2 is the congestion density k j Or the optimal density k o ,λ1…,λ n is the characteristic parameter of the traffic flow model, ε h (k) is the error related to density; A linear model is used to establish the relationship between the traffic flow speed observation value u and the speed expectation value μ. The relationship model is u=μ+ε, where ε is the random error. Correlated random parameter distribution is used to represent the correlation of traffic flow parameters. A road traffic flow speed prediction model considering the correlation of traffic flow parameters is constructed. The two parameters ω1 and ω2 in the traffic flow speed and density relationship model obey the bivariate normal distribution MVN2, which is expressed as follows: in and is the mean and variance of ω1, ω2, cov(ω1, ω2) is the covariance of ω1, ω2; the expression of the correlation coefficient ρ of ω1, ω2 is: The normal distribution N is used to represent the error of traffic flow speed, and the error related to traffic density ε h (k) obeys the normal distribution N(0,g(k|β1,…,β n )), the random error ε follows the normal distribution N(0,σ 2 ), where σ 2 is the variance of ε; Bayesian updating is used to estimate model parameters; The traffic flow speed is predicted using a two-stage traffic flow model, and the road traffic flow speed prediction value considering the heterogeneity of vehicle speed is The expression is: in Is f in The second-order partial derivative at .

2. The method for predicting road traffic flow speed considering driving behavior heterogeneity and parameter correlation according to claim 1, characterized in that: The Bayesian updating method is used to estimate the parameters of the model, including: The variance of traffic flow speed follows the normal distribution N, and its expression is: in and The traffic flow speed variance is The expectation and variance of Traffic flow speed follows normal distribution N, which is expressed as: u~N(μ,σ 2 ) where μ and σ 2 are the expectation and variance of the traffic flow speed u, μ=f(k|ω1,ω2,λ1,…,λ n )+ε h (k); The Bayesian updating method is used to update the parameters of the two-stage traffic flow model within the same model framework. Make an estimate.

3. The method for predicting road traffic flow speed considering driving behavior heterogeneity and parameter correlation according to claim 2, characterized in that: When estimating the parameters of the two-stage traffic flow model, the posterior function The expression is: in is the probability of u, from the normal distribution N(μ,σ 2 )get, is the probability of ω1,ω2, obtained from the bivariate normal distribution MVN2, is ε h (k), from the normal distribution N(0,g(k|β1,…,β n ))get, yes The probability of get, yes σ 2 The prior distribution of The estimated values ​​are used to calculate the correlation coefficient ρ of ω1 and ω2.

4. The method for predicting road traffic flow speed considering driving behavior heterogeneity and parameter correlation according to claim 2, characterized in that: The prior distribution of the parameters in the two-stage traffic flow model adopts normal distribution N(0,10000), inverse gamma distribution IG(0.01,0.01), and inverse Wishart distribution Its expression is β1~N(0,10000),…,β n ~N(0,10000),λ1~N(0,10000),…,λ n ~N(0,10000), σ 2 ~IG(0.01,0.01), 5. The method for predicting road traffic flow speed considering driving behavior heterogeneity and parameter correlation according to claim 1 is characterized by: The g(k|β1,…,β n ) is defined as:

6. The method for predicting road traffic flow speed considering vehicle speed heterogeneity according to claim 1, characterized in that: The f(k|ω1,ω2,λ1,…,λ n ) is defined as: where u f and k j are the free flow speed and congestion density parameters of the traffic flow model, C j is the speed of the motion wave at the congestion density.

7. A road traffic flow speed prediction system considering driving behavior heterogeneity and parameter correlation, characterized in that: include: Traffic flow model creation module, parameter estimation module and traffic flow speed prediction module; The traffic flow model creation module is used to use the variance of traffic flow speed to represent the heterogeneity of drivers' driving behavior and establish the variance of traffic flow speed within a preset time period. The relationship function with the traffic density k is defined as g(k|β1,…,β n ), where g(.) is the relationship function, β1,…,β n is a parameter; the traffic flow model is used to establish the relationship between the expected value of traffic flow speed μ and density k, and the relationship model is μ=f(k|ω1,ω2,λ1,…,λ n )+ε h (k), where ω1 is the free stream velocity u f , ω2 is the congestion density k j Or the optimal density k o ,λ1,…,λ n is the characteristic parameter in the traffic flow model, ε h (k) is the error related to density. A linear model is used to establish the relationship between the observed traffic flow speed u and the expected speed μ. The relationship model is u=μ+ε, where ε is the random error. The correlation of traffic flow parameters is represented by the distribution of correlated random parameters. A road traffic flow speed prediction model considering the correlation of traffic flow parameters is constructed. The two parameters ω1 and ω2 in the relationship model obey the bivariate normal distribution MVN2, and the expression is: in and are the mean and variance of ω1, ω2, and cov(ω1, ω2) is the covariance of ω1, ω2; And, the normal distribution N is used to represent the error of traffic flow speed, and the error related to traffic density ε h (k) obeys the normal distribution N(0,g(k|β1,…,β n )); random error ε obeys the normal distribution N(0,σ 2 ), where σ 2 is the variance of ε; The parameter estimation module is used to estimate the parameters of the model using Bayesian updating; The traffic flow speed prediction module is used to predict the traffic flow speed using a two-stage traffic flow model, taking into account the heterogeneity of vehicle speeds. The expression is: in Is f in The second-order partial derivative at .

8. The road traffic flow speed prediction system considering driving behavior heterogeneity and parameter correlation according to claim 7 is characterized by: The g(k|β1,…,β n ) is defined as:

9. The road traffic flow speed prediction system considering vehicle speed heterogeneity according to claim 7, characterized in that: The f(k|ω1,ω2,λ1,…,λ n ) is defined as: where u f and k j are the free flow speed and congestion density parameters of the traffic flow model, C j is the speed of the motion wave at the congestion density.

10. A computer system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the computer program is loaded into a processor, the steps of the method for predicting road traffic flow speed considering the heterogeneity of vehicle driving speeds according to any one of claims 1 to 6 are implemented.

Images