Road actual traffic capacity calculation method under intelligent network connection vehicle formation

Abstract

The invention discloses a road actual traffic capacity calculation method under an intelligent network connection vehicle formation, comprising the following steps: analyzing the intelligent network connection vehicle formation condition running in a road section, and calculating the actual traffic capacity of the road according to historical monitoring data; counting the formation number # imgabs0 # average vehicle speed vi corresponding to the vehicle formation length under different time ti in the line, the vehicle head spacing # imgabs1 # in the formation and the non-formation vehicle head spacing # imgabs2 #, and estimating the vehicle distribution # imgabs5 # of the formation length j under the given time # imgabs4 # according to the historical vehicle formation data set # imgabs3 #; grouping the historical vehicle formation data by using a density clustering OPTICS algorithm; establishing a prediction model for each group of formed clustering data, and predicting vehicle formation length distribution at given time; according to the historical vehicle formation data set # imgabs6 #, estimating the vehicle head distance under the superspeed condition; and based on the average inter-vehicle head # imgabs7 # and the average vehicle speed, calculating the actual road traffic capacity of the vehicle formation under the ultra-high-speed condition.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation, and particularly relates to a method for calculating the actual highway passing capacity under the formation of intelligent connected vehicles. Background Art

[0002] With the deep integration and development of new-generation information technologies such as big data, artificial intelligence, machine vision, blockchain, Beidou, and 5G in the field of highway transportation, digitization, networking, and intelligence are helping the high-quality development of highway transportation. At the same time, highway development will fully apply technologies such as advanced driver assistance, vehicle networking, vehicle-road coordination, and autonomous driving, and use a reliable communication network to greatly improve the formation ability of the highway system, further reduce the headway between vehicles, significantly improve the highway passing capacity, and achieve a doubling of the traffic operation efficiency of the section / road network.

[0003] Therefore, it is urgent to estimate the actual highway passing capacity under the formation of intelligent connected vehicles, so as to explore new laws and characteristics brought by the new traffic flow of the formation operation of intelligent connected vehicles. Summary of the Invention

[0004] Aiming at the deficiencies in the prior art, the present invention provides a method for calculating the actual highway passing capacity under the formation of intelligent connected vehicles.

[0005] The present invention discloses a method for calculating the actual highway passing capacity under the formation of intelligent connected vehicles, including:

[0006] Step 1: Analyze the formation situation of intelligent connected vehicle formations running on the road section, and according to historical monitoring data, count the formation quantities corresponding to the formation lengths of vehicles at different times t i under the line, average vehicle speed v i the headway within the formation, and the headway of non-formation vehicles to obtain the road section vehicle formation data set H = [H 1 , …, H K , where the set of data i i is the data number;

[0007] Step 2: Estimate the vehicle distribution with formation length j at a given time according to the historical vehicle formation data set Estimate the vehicle distribution with formation length j at a given time Use the density clustering OPTICS algorithm to group the historical vehicle formation data, establish a prediction model for each group of clustering data formed, and predict the vehicle formation length distribution at a given time;

[0008] Step 3: Estimate the headway in the case of ultra-high speed according to the historical vehicle formation data set Estimate the headway in the case of ultra-high speed;

[0009] Step 4. Based on the average headway and the average vehicle speed, calculate the actual highway capacity of the vehicle formation under ultra-high speed conditions.

[0010] As a further improvement of the present invention, in Step 1, for any set of data i, it is satisfied that the sum of the vehicles with all formation lengths in the section is equal to the total number N of intelligent connected vehicles in the section i .

[0011] As a further improvement of the present invention, Step 2 specifically includes:

[0012] Step 21. Group the historical vehicle formation data set according to peak periods and off-peak periods;

[0013] Step 22. Normalize the historical vehicle formation data set, and represent it by the ratio of different formation lengths to the total number of intelligent connected vehicles and convert the time t i into the moment t' of a day i ∈[0, 23], then the historical vehicle formation data set changes into the time-series vehicle formation data

[0014] Step 23. For the time-series vehicle formation data, select the Euclidean distance as the clustering similarity metric; for the data numbers i and j of the data Q m and Q b , the minimum average error is expressed as where N 0 is the maximum allowable formation length;

[0015] Step 24. Use the density clustering OPTICS algorithm to cluster the peak periods and off-peak periods respectively. By setting the algorithm parameters, the algorithm parameters are the maximum neighborhood radius ε and the minimum number of samples MinPts included in the neighborhood. Cluster the two historical vehicle formation data sets of the peak periods and off-peak periods respectively to obtain the clustered historical vehicle formation data sets;

[0016] Step 25. For each set of clustered historical vehicle formation data sets, use the Bayesian structural time series model to establish a state observation structure model for the number of vehicles with formation length j;

[0017] Step 26. According to the observed historical vehicle formation data, apply the Kalman filter to update the estimated value of the state; use the nail board method to select variables for the structural model and calculate the regression coefficients;

[0018] Step 27. Then use the Bayesian model averaging method to combine and generate the prediction result at a given time

[0019] As a further improvement of the present invention, in the step 21, the specific grouping method is: compare the total number N of intelligent network-connected vehicles i with the threshold value to determine whether it is in the peak period or the off-peak period; when it is the peak period, otherwise it is the off-peak period.

[0020] As a further improvement of the present invention, in the step 25, the vehicle number state observation structure model

[0021] is as follows:

[0022]

[0023] α t+1 = T t α t + R t η t

[0024]

[0025] η t ~N(0, Q t ) (4)

[0026] where the matrices Z, T, and R are 0-1 matrices, and ε t,j and η t are Gaussian error terms.

[0027] As a further improvement of the present invention, the step 3 specifically includes:

[0028] Step 31: According to the historical headway data under non-ultra-high speed conditions (v i ≤ 120 km / h), use Gaussian Naive Bayes to predict the headway L of intelligent network-connected vehicles in the formation queue and the headway L between non-formation network-connected vehicles 0 and non-formation network-connected vehicles; 3 ;

[0029] Step 32: Calculate the average headway under ultra-high speed conditions and estimate the actual highway traffic capacity; the length of the i-th group of vehicle formations is

[0030] As a further improvement of the present invention, in the step 32, the formula for calculating the average headway is:

[0031]

[0032] where the total number of intelligent network-connected vehicles can be expressed as

[0033] As a further improvement of the present invention, in the step 4, the calculation formula for the actual highway traffic capacity is:

[0034]

[0035] Compared with the prior art, the beneficial effects of the present invention are:

[0036] The present invention analyzes the historical intelligent connected vehicle formation data in a section, and estimates the formation situation of intelligent connected vehicles with different formation lengths based on a given time; at the same time, analyzes the headway between intelligent connected vehicles in the case of ultra-high speed and the headway between non-formation connected vehicles, estimates the actual headway of vehicles; and further estimates the actual traffic capacity of vehicles in the case of the existence of intelligent connected vehicle formations. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 FIG. is a schematic diagram of the formation operation of highway intelligent connected vehicles.

[0038] Figure 2 FIG. is a flowchart of a method for calculating the actual highway traffic capacity under intelligent connected vehicle formations disclosed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0040] The following further describes the present invention in detail with reference to the drawings:

[0041] As Figure 2 shown, the present invention provides a method for calculating the actual highway traffic capacity under intelligent connected vehicle formations. In the case of the operation of intelligent connected vehicles on a highway, by analyzing the historical intelligent connected vehicle formation data in a section, the probability distribution of vehicle formations with different lengths is determined. And estimate the headway between intelligent connected vehicles in the formation queue and the headway between non-formation connected vehicles, and estimate the actual traffic capacity of vehicles in the case of the existence of intelligent connected vehicle formations;

[0042] Specifically, it includes:

[0043] Step 1. Analyze the formation situation of intelligent connected vehicles running on the road section. There are mixed formations and non - formations of intelligent connected vehicles running on the road section, as well as multiple groups of vehicle formations with different formation lengths; as shown in the appendix Figure 1 . Considering environmental factors such as communication delay and control stability, the maximum allowable formation length is N 0 . According to historical monitoring data, count the formation quantity corresponding to the data number i, time t i , vehicle formation length, average vehicle speed v , head - to - head spacing within the formation i and non - formation head - to - head spacing as shown in Table 1. Among them, when the vehicle formation length is 1, it is the number of intelligent connected vehicles not participating in the formation.

[0044] Table 1

[0045]

[0046] For any set of data i, the sum of vehicles with all formation lengths in the road section is equal to the total number of intelligent connected vehicles N i , that is:

[0047]

[0048] Therefore, for K sets of historical monitoring data, the vehicle formation data set H of the road section can be obtained as H = [H 1 ,..., H K . Among them, the set of data i Among them, the average vehicle speed

[0049] Step 2. Estimate the vehicle distribution with formation length j at a given time according to the historical vehicle formation data set . Use the density - based clustering OPTICS algorithm to group the historical vehicle formation data, and establish a prediction model for each formed cluster data set to predict the vehicle formation length distribution at a given time;

[0050]

[0050] Specifically include:

[0051] Step 21. Group the historical vehicle formation data set according to peak hours and off - peak hours; among them, compare the total number of intelligent connected vehicles N i with the threshold to determine whether it is in peak hours or off - peak hours; when it is peak hours, otherwise it is off - peak hours;

[0052] Step 22: Normalize the historical vehicle formation data set and represent it by the ratio of different formation lengths to the total number of intelligent connected vehicles And convert the time t i (in hours) into the moment t' of the day i ∈[0, 23], then the historical vehicle formation data set changes to the time-series vehicle formation data

[0053] Step 23: For the time-series vehicle formation data, select the Euclidean distance as the clustering similarity metric; for the data Qi m and Qj b , the minimum average error can be expressed as

[0054] Step 24: Considering the scale of the actual data and the uncertainty of the actual number of clustering clusters, use the density clustering OPTICS algorithm to cluster the peak period and the off-peak period respectively. By setting the algorithm parameters, the algorithm parameters are the maximum neighborhood radius ε and the minimum number of samples MinPts in the neighborhood. Cluster the two historical vehicle formation data sets of the peak period and the off-peak period respectively to obtain the clustered historical vehicle formation data set

[0055] Step 25: For each clustered historical vehicle formation data set, use the Bayesian structural time series model to establish a vehicle number state observation structure model for the formation length j; the vehicle number state observation structure model is specifically as follows: where the matrices Z, T, R are 0-1 matrices, ε t,j and η t are Gaussian error terms

[0056]

[0057] α t+1 = T t α t + R t η t

[0058]

[0059] η t ~N(0, Q t ) (8)

[0060] Step 26: According to the observed historical vehicle formation data, apply the Kalman filter to update the estimated value of the state; use the spike-and-slab method to perform variable selection on the structural model and calculate the regression coefficients

[0061] Step 27: Then use the Bayesian model averaging method to combine and generate at a given time The predicted results below

[0062] Step 3. According to the historical vehicle formation data set Estimate the headway at ultra-high speed (i.e., the average vehicle speed );

[0063] Specifically, it includes:

[0064] Step 31. According to the historical headway data under non-ultra-high speed (v i ≤120 km / h), use Gaussian Naive Bayes to predict the headway L between the connected vehicles in the formation queue and the headway L between non-formation connected vehicles 0 ; among them, non-formation vehicles include the leading vehicle of the vehicle formation queue and the connected vehicle in front, two connected vehicles on the road that are not in the formation queue, etc. 3

[0065] Step 32. Calculate the average headway under ultra-high speed conditions, and estimate the actual highway capacity; the length of the i-th group of vehicle formations is Then the average headway The calculation formula is:

[0066]

[0067] Among them, the total number of connected vehicles Can be expressed as

[0068] Step 4. Based on the average headway And the average vehicle speed, calculate the actual highway capacity of the vehicle formation under ultra-high speed conditions; among them, the calculation formula of the actual highway capacity is:

[0069]

[0070] At the same time, the correction coefficient f of the formation length is the ratio of the actual capacity C to the reference capacity C p , and the calculation is as follows:

[0071]

[0072] In practical applications, different average vehicle speeds can be selected to calculate the correction coefficient, and the actual capacity can be calculated conveniently through a list.

[0073] ​The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for calculating the actual highway capacity under intelligent networked vehicle platooning, characterized in that: include: Step 1: Analyze the intelligent connected vehicle fleet running on the road section, and count the number of vehicles on the route at different times t according to historical monitoring data. i The number of vehicles in the formation corresponding to the length of the vehicle formation under Average speedv i , the distance between the fronts of the vehicles in the formation and non-platooning headway Get the road section vehicle formation data set H = [H1,…,H K ], where the set of data i i is the data number; Step 2: Based on the historical vehicle formation data set Estimated given time Vehicle distribution under formation length j The density clustering OPTICS algorithm is used to group the historical vehicle formation data, and a prediction model is established for each group of clustered data to predict the distribution of vehicle formation length at a given time. Step 3: Based on the historical vehicle formation data set Estimate headway distance in ultra-high-speed situations; Step 4: Based on average headroom and average vehicle speed, calculate the actual highway capacity of vehicle platoons in super-high-speed conditions.

2. The method for calculating the actual highway capacity under the intelligent networked vehicle formation according to claim 1, characterized in that: In step 1, for any set of data i, the sum of all vehicles that meet the formation length in the section is equal to the total number of intelligent connected vehicles N in the section i .

3. The method for calculating the actual highway capacity under the intelligent networked vehicle platoon as claimed in claim 1, characterized in that: The step 2 specifically includes: Step 21, grouping the historical vehicle formation data set according to peak period and off-peak period; Step 22: Normalize the historical vehicle formation data set and represent it by the ratio of different formation lengths to the total number of intelligent connected vehicles. And the time t i Converted to time of day t′ i ∈[0,23], the historical vehicle formation data set changes to the time series vehicle formation data Step 23: For the time series vehicle formation data, select the Euclidean distance as the cluster similarity metric; for data number i and j data Q m and Q b , the minimum average error is expressed as Among them, N0 is the maximum allowed formation length; Step 24, using the density clustering OPTICS algorithm to cluster the peak period and the flat peak period respectively, by setting the algorithm parameters, the algorithm parameters are the maximum radius ε of the neighborhood and the minimum number of samples MinPts in the neighborhood, respectively clustering the two historical vehicle formation data sets of the peak period and the flat peak period to obtain the clustered historical vehicle formation data set; Step 25: for each set of clustered historical vehicle formation data, a vehicle number state observation structure model for formation length j is established using a Bayesian structural time series model; Step 26: Based on the observed historical vehicle formation data, apply the Kalman filter to update the estimated value of the state; use the nail board method to select variables for the structural model and calculate the regression coefficient; Step 27, then use the Bayesian model averaging method to generate a given time The prediction results of 4. The method for calculating the actual highway capacity under the intelligent networked vehicle formation as claimed in claim 3, characterized in that: In step 21, the specific grouping method is: compare the total number of intelligent networked vehicles N i With threshold to determine whether it is during peak or off-peak period; It is the peak period, otherwise it is the off-peak period.

5. The method for calculating the actual highway capacity under the intelligent networked vehicle platoon as claimed in claim 3, characterized in that: In step 25, the vehicle number state observation structure model is as follows: Among them, the matrices Z, T, and R are 0-1 matrices, and ε t,j and η t is the Gaussian error term.

6. The method for calculating the actual highway capacity under the intelligent networked vehicle platoon as claimed in claim 1, characterized in that: The step 3 specifically includes: Step 31: According to the non-ultra-high speed case (v i ≤120km / h) historical headway data Gaussian Naive Bayes is used to predict the headway distance L0 of intelligent connected vehicles in a platoon and the headway distance L3 between non-platoon connected vehicles; Step 32: Calculate super high speed The average headway distance under the condition is used to estimate the actual traffic capacity of the highway; the length of the i-th vehicle formation is 7. The method for calculating the actual highway capacity under the intelligent networked vehicle platoon as claimed in claim 6, characterized in that: In step 32, the average headroom The calculation formula is: Among them, the total number of intelligent connected vehicles It can be expressed as 8. The method for calculating the actual highway capacity under the intelligent networked vehicle platoon as claimed in claim 7, characterized in that: In step 4, the calculation formula of the actual highway capacity is:

Images