Urban road traffic demand prediction method considering road construction influence

Abstract

The invention relates to an urban road traffic demand prediction method considering road construction influence. The method comprises the following steps: acquiring a motor vehicle historical travel OD matrix, a road section traffic volume and travel time in a peak hour in a prediction range; observing the running state of the vehicle during construction in the construction influence area during peak hours, and recording the observed traffic flow and the observed travel time of the vehicle passing through the construction influence area; determining a flow deviation influence factor in combination with the road grade of the construction influence road and the traffic flow; correcting the historical travel OD matrix by adopting a two-stage model according to the observed traffic flow of the vehicle passing through the construction influence area; and according to the observation travel time of the vehicle passing through the construction influence area, the travel time of the road section of the construction influence area is corrected, traffic distribution is performed again, and a traffic demand prediction result is obtained. Compared with the prior art, the urban road network traffic demand can be quickly and accurately predicted under the condition that the traffic capacity of a part of road sections is changed due to road construction.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a method for predicting urban road traffic demand taking into account the impact of road construction. Background Art

[0002] Traffic demand forecasting is one of the core issues in the field of transportation planning and management research. Since the birth of the theory and method of traffic demand forecasting in the 1930s, after years of continuous research and practice, a large number of achievements have been made. With the increasing complexity of practical problems in transportation, traffic volume forecasting has always been a hot issue for theoretical researchers and practitioners in the transportation industry.

[0003] The classic traffic demand forecasting method is the four-stage method, which distributes traffic to each section of the road network through four stages: traffic generation, traffic distribution, traffic mode division, and traffic allocation. However, the classic four-stage method requires a large amount of traffic survey data and has a long implementation cycle. At the same time, the traditional four-stage method uses gravity models and other methods to predict traffic distribution, but this method not only has a high prediction accuracy, but also has relatively complex calculations and extremely high prediction costs. It is difficult to cope with changes in traffic capacity caused by temporary events such as road construction. Affected by urban road construction, the traffic capacity of some sections suddenly changes, which can easily cause traffic jams in severe cases. Therefore, how to quickly and accurately predict the traffic demand of urban road networks in this situation is an urgent problem to be solved. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a method for predicting urban road traffic demand taking into account the impact of road construction. By utilizing multi-source data such as real-time Internet and traffic flow, it is possible to quickly and accurately predict the traffic demand of the urban road network when road construction causes changes in the traffic capacity of some sections.

[0005] The purpose of the present invention can be achieved by the following technical solution: A method for predicting urban road traffic demand considering the impact of road construction, comprising the following steps:

[0006] S1. Obtain the OD (Origin Destination) matrix, road section traffic volume and travel time of the historical motor vehicle trips during the peak hour within the forecast range, where the data in the OD matrix is ​​the traffic flow from one area to another;

[0007] S2. Observe the vehicle operation status during construction in the construction affected area during peak hours, and record the observed traffic flow and observed travel time of vehicles passing through the construction affected area;

[0008] S3. Determine the flow deviation influencing factor based on the road grade and traffic flow of the road affected by the construction;

[0009] S4, based on the observed traffic flow of vehicles passing through the construction impact area in step S2, a two-stage model is used to modify the historical travel OD matrix;

[0010] S5. According to the observed travel time of vehicles passing through the construction affected area in step S2, the travel time of the road section in the construction affected area is corrected, and traffic is redistributed to obtain a traffic demand forecast result.

[0011] Furthermore, the step S1 specifically utilizes mobile phone signaling and Internet data to obtain historical motor vehicle travel OD data, road section traffic volume and travel time during peak hours within the prediction range.

[0012] Furthermore, the step S2 specifically utilizes the front-end acquisition equipment to observe the vehicle operation status during construction in the construction-affected area during peak hours.

[0013] Furthermore, the front-end acquisition equipment includes a camera and a traffic ground sensor coil.

[0014] Furthermore, the observed traffic flow of vehicles passing through the construction affected area in step S2 is specifically the traffic volume passing through the section of the construction area during construction;

[0015] The observed travel time of vehicles passing through the construction affected area is specifically the time t that all vehicles pass through the starting point i of the construction area during the observation period during construction. i and the time t to pass its end point j j The average time difference between .

[0016] Furthermore, the value of the flow deviation impact factor in step S3 is between [0,1]. If the road grade in the construction affected area is higher and the traffic flow is larger, the impact caused by the flow deviation is greater and the flow deviation impact factor is closer to 1, otherwise it is closer to 0.

[0017] Furthermore, the step S4 comprises the following steps:

[0018] S41. Construct a two-stage model, including objective function, traffic assignment model and constraints;

[0019] S42. Use the gradient descent heuristic algorithm to solve the two-stage model and obtain a new travel OD matrix, which is used to correct the historical travel OD matrix.

[0020] Furthermore, the objective function in step S41 is to minimize the modified trip OD matrix d rs Compared with the original travel OD matrix The deviation between the estimates and the traffic flow x ij Observing traffic flow in construction zones The deviation between

[0021] The traffic assignment model is used to indicate that when the travel OD matrix is ​​d, the corresponding road section traffic flow solution x is obtained by solving the traffic assignment model;

[0022] Constraints are used to ensure that the travel volume between regions is non-negative.

[0023] Furthermore, the two-stage model in step S41 is specifically:

[0024]

[0025] Among them, the objective function g(d,x) is about and The least squares form of , vector d, and x, respectively, are used to represent all d rs , and x ij Variable, Z is the traffic zone, d rs is the demand data from cell r to cell s in the modified OD matrix, is the demand data from cell r to cell s in the original OD matrix, A is the set of road sections in the prediction range, is the set of construction sections, x ij is the road flow, t ij is the journey time of the road section, is the observed traffic flow recorded in step S2, and λ is the flow deviation influencing factor.

[0026] Furthermore, the step S5 is specifically based on the Wardrop user equilibrium principle to redistribute traffic.

[0027] Compared with the prior art, the present invention has the following advantages:

[0028] The present invention obtains the historical travel OD matrix of motor vehicles during peak hours within the prediction range, the traffic volume of the road section and the travel time, and then observes the vehicle operation status during construction in the construction-affected area during peak hours, records the observed traffic flow and observed travel time of vehicles passing through the construction-affected area; and combines the road grade and traffic flow of the construction-affected road to determine the flow deviation influencing factor; then, based on the observed traffic flow of vehicles passing through the construction-affected area, a two-stage model is used to correct the historical travel OD matrix; finally, based on the observed travel time of vehicles passing through the construction-affected area, the travel time of the road section in the construction-affected area is corrected, and traffic is redistributed to obtain the traffic demand prediction result. In this way, the multi-source data that can be collected can be used to correct the travel OD matrix, so as to accurately predict the traffic demand within the prediction range in real time, and effectively avoid the occurrence of traffic jams.

[0029] In the present invention, the road grade and traffic flow of the roads affected by the construction are comprehensively considered to determine the flow deviation influencing factor, and the value is selected according to the grade of the road and the traffic saturation: the higher the road grade and the greater the traffic flow in the construction affected area, the greater the impact of the flow deviation, and the closer the flow deviation influencing factor is to 1, and vice versa, the closer it is to 0, thereby fully ensuring the accuracy of the value of the flow deviation influencing factor, which is conducive to improving the reliability of the subsequent construction of the two-stage model.

[0030] The present invention adopts a two-stage model to correct the travel OD matrix. The two-stage model includes: an objective function for minimizing the corrected travel OD matrix d rs Compared with the original travel OD matrix The deviation between the estimates and the traffic flow x ij Observing traffic flow in construction zones The traffic assignment model is used to indicate that when the travel OD matrix is ​​d, the corresponding road section traffic flow solution x is obtained by solving the traffic assignment model; the constraint condition is used to ensure that the travel volume between regions is non-negative. Therefore, by solving the two-stage model, a new travel OD matrix can be obtained, and the optimal solution for traffic flow can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0032] Figure 2 Schematic diagram of the construction impact area in the embodiment;

[0033] Figure 3 Schematic diagram of the process of using the gradient descent heuristic algorithm to solve the two-stage model in the embodiment. DETAILED DESCRIPTION

[0034] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] Example

[0036] like Figure 1 As shown, a method for predicting urban road traffic demand considering the impact of road construction includes the following steps:

[0037] S1. Obtain the historical OD matrix of motor vehicle travel during peak hours within the forecast range, the traffic volume of road sections and the travel time, wherein the data in the OD matrix is ​​the traffic flow from one area to another;

[0038] S2. Observe the vehicle operation status during construction in the construction affected area during peak hours, and record the observed traffic flow and observed travel time of vehicles passing through the construction affected area;

[0039] S3. Determine the flow deviation influencing factor based on the road grade and traffic flow of the road affected by the construction;

[0040] S4, based on the observed traffic flow of vehicles passing through the construction impact area in step S2, a two-stage model is used to modify the historical travel OD matrix;

[0041] S5. According to the observed travel time of vehicles passing through the construction affected area in step S2, the travel time of the road section in the construction affected area is corrected, and traffic is redistributed to obtain a traffic demand forecast result.

[0042] This embodiment applies the above solution, and the main contents are:

[0043] Step 1: Use mobile phone signaling and Internet data to obtain the historical travel OD data of motor vehicles during peak hours within the forecast range (that is, the historical travel OD matrix), road section traffic volume and travel time.

[0044] Step 2: Observe the vehicle operation status during construction in the construction affected area during peak hours, and record the observed traffic flow of vehicles passing through the construction affected area Observation travel time

[0045] like Figure 2 As shown, the red shaded area is the construction area. In step 2:

[0046] ① Traffic flow It is the sum of the traffic volume passing through the section of the construction area during construction;

[0047] ②Travel time That is, the time t when all vehicles pass through the starting point i of the construction area during the observation period during construction i and the time t to pass its end point j jThe average time difference between .

[0048] Step 3: Determine the flow deviation influencing factor λ by combining the road grade and traffic flow of the road affected by the construction;

[0049] The specific implementation of this step is as follows:

[0050] The value of the flow deviation impact factor λ is selected based on the road grade and traffic flow. λ is a flow deviation impact factor between [0,1]. The higher the road grade and the greater the traffic flow in the construction impact area, the greater the impact of the flow deviation. The closer λ is to 1, and vice versa, the closer it is to 0. For example, when the construction road is a traffic trunk road and the traffic flow is saturated, a larger λ value, such as 0.95, is selected; when the construction road is a branch road and the traffic flow is small, a smaller λ value, such as 0.5, can be selected.

[0051] Step 4: Based on the observed traffic before and after the above construction impact, a two-stage model is used to modify the historical travel OD matrix;

[0052] The specific implementation of this step is as follows:

[0053] The following two-stage model is established to correct the original OD matrix. In order to simplify the expression, the vector d is introduced: and x, respectively, are used to represent all d rs , and x ij variable:

[0054]

[0055] Where Z is the traffic zone; d rs is the demand data from cell r to cell s in the modified OD matrix; is the demand data from cell r to cell s in the original OD matrix; A is the set of road sections in the prediction range; is the construction section collection; x ij is the road flow; t ij is the journey time of the road section; is the traffic flow of the observed road section recorded in step 2.

[0056] In this two-stage model, the objective function g(d,x) is and The least squares form aims to minimize the corrected OD matrix d rs With the original OD matrix The deviation between the estimates and the segment flow x ij Observe the flow in the construction area The deviation between .

[0057] Formula (2) involves a traffic assignment model, which means that when the OD matrix is ​​d, the corresponding road section flow solution x is obtained by solving the traffic assignment model. Function X(d) represents the feasible flow of the road section when the OD matrix is ​​d

[0058] gather.

[0059] Since the traffic assignment model has the characteristic of a unique optimal solution for the road section flow, this scheme transforms the objective function g(d,x) into a function form that only depends on the variable d, that is, g(d,x(d)). The function x(d) represents the optimal road section flow solution x when the OD matrix is ​​d.

[0060] For this two-stage model, a heuristic algorithm based on gradient descent is used to solve it. Calculate the objective function g(d,x(d)) for any d rs The partial derivative of :

[0061]

[0062] Integrate the partial derivative of the objective function g with respect to d into a vector form, namely the gradient Since the two-stage model is a minimization problem, once the gradient is found The OD matrix d can be updated along the gradient descent direction indicated by it, so that the objective function g(d,x(d)) decreases.

[0063]

[0064] Here, α represents the step size, which is a positive number used to control the magnitude of the OD matrix update in each iteration. [·] + Indicates that each element of the vector takes a non-negative value operation, which is based on the constraint that the demand in the OD matrix cannot be negative. Based on the above analysis, this solution proposes a corresponding algorithm to solve the two-stage model, such as Figure 3 As shown in the figure, the specific steps for solving the problem are as follows:

[0065] Input: Initial OD matrix Observed road traffic Impact factor λ;

[0066] Step 1 (initialization):

[0067] Iteration starts:

[0068] Step 2: Solve the traffic assignment model based on d, obtain x(d), and calculate the objective function value g old =g(d,x(d));

[0069] Step 3: Calculate the gradient

[0070] Step 4:

[0071] Initialize the step size sequence α∈{1,0.5,0.25,…};

[0072] For each α in the step sequence perform:

[0073] Update d;

[0074] Resolve the traffic assignment to obtain x(d);

[0075] Calculate the new objective function value g new =g(d,x(d));

[0076] If g new <g old but

[0077] Accept the step size α, stop the fourth step, and execute the fifth step;

[0078] otherwise

[0079] Reduce the step size α and continue testing;

[0080] Step 5 (convergence test):

[0081] Observe the changes in the objective function value over multiple consecutive iterations;

[0082] If the decrease in the objective function value gradually decreases and tends to be stable, then

[0083] Terminate the iteration and output d;

[0084] otherwise

[0085] Continue to iterate;

[0086] Output: Corrected OD matrix d.

[0087] Step 5: Based on the observed travel time The travel time of the road section in the construction-affected area is corrected, and traffic is redistributed (in this embodiment, traffic is redistributed according to the Wardrop user equilibrium principle), and finally a traffic demand prediction result is obtained.

[0088] In summary, when this scheme is applied in practice, when predicting road traffic demand, the multi-source data that can be collected can be directly used to correct the travel distribution OD matrix, so as to achieve accurate prediction of traffic demand within the prediction range. At the same time, the execution process of the method of this scheme is relatively simple, and the required data and calculation amount are small, so it can quickly predict traffic demand and obtain real-time prediction results.

Claims

1. A method for predicting urban road traffic demand considering the impact of road construction, characterized in that: The following steps are involved: S1. Obtain the historical OD matrix of motor vehicle travel during peak hours within the forecast range, the traffic volume of road sections and the travel time, wherein the data in the OD matrix is ​​the traffic flow from one area to another; S2. Observe the vehicle operation status during construction in the construction affected area during peak hours, and record the observed traffic flow and observed travel time of vehicles passing through the construction affected area; S3. Determine the flow deviation influencing factor based on the road grade and traffic flow of the road affected by the construction; S4, based on the observed traffic flow of vehicles passing through the construction impact area in step S2, a two-stage model is used to modify the historical travel OD matrix; S5. According to the observed travel time of vehicles passing through the construction affected area in step S2, the travel time of the road section in the construction affected area is corrected, and traffic is redistributed to obtain a traffic demand forecast result.

2. The method for predicting urban road traffic demand considering the impact of road construction according to claim 1 is characterized in that: The step S1 specifically utilizes mobile phone signaling and Internet data to obtain historical motor vehicle travel OD data, road section traffic volume and travel time during peak hours within the forecast range.

3. The method for predicting urban road traffic demand considering the impact of road construction according to claim 1 is characterized in that: The step S2 specifically utilizes the front-end acquisition equipment to observe the running status of vehicles during construction in the construction-affected area during peak hours.

4. The method for predicting urban road traffic demand considering the impact of road construction according to claim 3 is characterized in that: The front-end acquisition equipment includes a camera and a traffic ground sensor coil.

5. The method for predicting urban road traffic demand considering the impact of road construction according to claim 1 is characterized in that: The observed traffic flow of vehicles passing through the construction affected area in step S2 is specifically the traffic volume passing through the section of the construction area during construction; The observed travel time of vehicles passing through the construction affected area is specifically the time t that all vehicles pass through the starting point i of the construction area during the observation period during construction. i and the time t to pass its end point j j The average time difference between .

6. The method for predicting urban road traffic demand considering the impact of road construction according to claim 1 is characterized in that: The value of the flow deviation impact factor in step S3 is between [0,1]. If the road grade in the construction affected area is higher and the traffic flow is larger, the impact caused by the flow deviation will be greater and the flow deviation impact factor will be closer to 1, otherwise it will be closer to 0.

7. The method for predicting urban road traffic demand considering the impact of road construction according to claim 1 is characterized in that: The step S4 comprises the following steps: S41. Construct a two-stage model, including objective function, traffic assignment model and constraints; S42. Use the gradient descent heuristic algorithm to solve the two-stage model and obtain a new travel OD matrix, which is used to correct the historical travel OD matrix.

8. The method for predicting urban road traffic demand considering the impact of road construction according to claim 7 is characterized in that: The objective function in step S41 is to minimize the modified travel OD matrix d rs Compared with the original travel OD matrix The deviation between the estimates and the traffic flow x ij Observing traffic flow in construction zones The deviation between The traffic assignment model is used to indicate that when the travel OD matrix is ​​d, the corresponding road section traffic flow solution x is obtained by solving the traffic assignment model; Constraints are used to ensure that the travel volume between regions is non-negative.

9. The method for predicting urban road traffic demand considering the impact of road construction according to claim 8 is characterized in that: The two-stage model in step S41 is specifically: Among them, the objective function g(d,x) is about and The least squares form of , vector d, and x, respectively, are used to represent all d rs , and x ij Variable, Z is the traffic zone, d rs is the demand data from cell r to cell s in the modified OD matrix, is the demand data from cell r to cell s in the original OD matrix, A is the set of road sections in the prediction range, is the set of construction sections, x ij is the road flow, t ij is the journey time of the road section, is the observed traffic flow recorded in step S2, and λ is the flow deviation influencing factor.

10. A method for predicting urban road traffic demand considering the impact of road construction according to any one of claims 1 to 9, characterized in that: The step S5 is specifically based on the Wardrop user equilibrium principle to redistribute traffic.

Images