A tidal lane opening double judgment method based on a traffic flow prediction model

Abstract

This invention discloses a dual-determination method for tidal lane opening based on a traffic flow prediction model. It detects the conditions of urban arterial roads in real time and preprocesses the acquired data; it also constructs a quadratic exponential smoothing prediction model to predict the number of vehicles N arriving in subsequent time periods. t+i ; and through N t+i Predicted average vehicle speed v 预 Vehicle density value K 预 Vehicle arrival rate P 预 Road service levels are classified into six levels, from one to six. When the parameters fall within the range of service level one to three, and the vehicle arrival rate P... 预 Following a Poisson distribution, the tidal lane is closed after double determination; when the parameter determination is within the road service level range of four to six, and P 预 Following a binomial distribution, the tidal flow lane is opened after dual determination. This invention improves the time and space resource utilization during the opening process of the tidal flow lane, thereby alleviating traffic congestion and reducing the risk of rear-end collisions.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of tidal lane opening determination in the field of traffic, and particularly relates to a tidal lane opening double determination method based on a traffic flow prediction model. BACKGROUND

[0002] The existing tidal lane opening determination is a single real-time shooting method by means of a camera to obtain the number of vehicles on the road, further determine the saturation of the number of vehicles to the urban trunk road, and determine whether to open or close the tidal lane. This method wastes part of the time in the use of time, and has a safety hazard. After the camera collects information, the computer needs to perform related analysis, and during this period, the front vehicles have already driven to the intersection for queuing, and the tidal lane is still slowly opening. At this time, if the rear vehicles change lanes while the tidal lane is opening, the rear vehicles will generate a certain speed difference with the front vehicles under the condition of acceleration. According to the analysis of the vehicle following characteristics, a high-density vehicle fleet is driving on the road, the distance between the vehicles is small, and the speed of any vehicle in the vehicle fleet is restricted by the speed of the front vehicle in non-free driving. The driver can only adopt the corresponding speed according to the information provided by the front vehicle. Since the speed of the rear vehicle is greater than that of the front vehicle, it is easy to cause a rear-end collision. Therefore, a new prediction method is needed to determine whether to open the tidal lane in advance to improve the time utilization rate and reduce the risk of vehicle rear-end collision during the lane changing process. SUMMARY

[0003] The present application proposes a tidal lane opening double determination method based on a traffic flow prediction model to improve the time utilization rate and space resource utilization rate during the opening process of the tidal lane, thereby relieving traffic congestion and reducing the risk of vehicle rear-end collision.

[0004] Technical scheme: The tidal lane opening double determination method based on the traffic flow prediction model comprises the following steps:

[0005] (1) Real-time detection of the urban trunk road condition, and pre-processing of the obtained data;

[0006] (2) Construction of a quadratic exponential smoothing prediction model to predict the number of subsequent time period prediction vehicles N t+i , and prediction of the average vehicle speed v 预 , the vehicle density value K 预 , and the vehicle arrival rate P 预 in the N t+i prediction interval;

[0007] (3) Division of the road service level, the division of the road service level is one to six levels, when the parameter judgment is in the range of one to three levels of the road service level, and the vehicle arrival rate P 预Subject to Poisson distribution, after double determination, the tidal lane is in closed state; when the parameter determination is in the range of road service level four to six, and P 预 Subject to binomial distribution, after double determination, the tidal lane is in open state.

[0008] Further, the step (1) is implemented as follows:

[0009] The radar is used to detect the urban trunk road condition in real time, including vehicle traffic condition, total number of vehicles, and traffic time size of vehicles, the detected total number of vehicles and traffic time size related data are matrix processed, and are uniformly processed as a transpose matrix.

[0010] Further, the step (2) is implemented as follows:

[0011]

[0012]

[0013] In the formula, S (1) t+1 is the first exponential smoothing value of the t+1 period; S (2) t+1 is the second exponential smoothing value of the t+1 period; S (1) t is the first smoothing value of the t period; S (2) t is the second smoothing value of the t period; N t is the number of records of vehicles in different periods; and α is the smoothing coefficient selected according to different road conditions.

[0014] The prediction vehicle number equation is:

[0015] N t+i =a t +b t i

[0016] a t =2S (1) t -S (2) t

[0017]

[0018] In the formula, N t+i is the prediction value of the vehicle number of the t+i period; i is the period from the t period; i is an integer, i∈[0,12]; a t , b t are prediction model coefficients.

[0019] Further, the predicted interval average vehicle speed vpre and the predicted vehicle density Kpre in step (2) are calculated as follows:

[0020] The free vehicle speed is selected as v 自由 , and the predicted lane length L 预 = 5v 自 / 60 (the predicted value for a 5-minute period) is calculated. The predicted traffic volume Q 预 , the predicted vehicle flow density K 预 , and the interval vehicle speed v 预 are calculated according to the relevant data:

[0021]

[0022]

[0023]

[0024] In the above equations, N t+i is the predicted number of vehicles in the road section; L is the 5-minute travel length calculated at the free vehicle speed; K 预 is the predicted vehicle flow density; Q 预 is the predicted average vehicle flow; and v 预 is the predicted interval average vehicle speed.

[0025] Further, the predicted vehicle arrival rate Ppre in step (2) is calculated as follows:

[0026] The number of vehicles N t+i predicted by the quadratic smoothing index is (N t+1 , N t+2 , N t+3 , N t+4 ,..., N t+12 ). The relationship with respect to the vehicle flow density K 预 is derived, and it is determined whether the vehicle flow density K 预 meets the condition for Poisson distribution. If the vehicle flow density K 泊松 meets the condition for Poisson distribution, λ 预 is substituted into the Poisson distribution, and the probability of the predicted number of vehicles N t+i for each period is calculated. If the vehicle flow density K 预 does not meet the condition for Poisson distribution, λ t+i is substituted into the binomial distribution, and the probability of the predicted number of vehicles N 预 for each period is calculated.

[0027] The predicted average arrival rate λ 预 is calculated as follows:

[0028]

[0029]

[0030] Poisson distribution:

[0031]

[0032] P = e-λt (N Nt+i ) t t+i λ 预 is the predicted average arrival rate; t is the duration of each counting interval; e is the base of natural logarithm;

[0033] λ 预 Substituting, we get:

[0034] P = e-λt (N Nt+i To be in the predictive confidence interval of Poisson distribution, the range of traffic density and the value of n of Poisson distribution are subject to road conditions and situations;

[0035] Binomial distribution is:

[0036]

[0037] P = (m Nt+i ) t t+i λ 预 is the predicted average arrival rate; t is the duration of each counting interval; m is a positive integer value;

[0038] λ 预 Substituting, we get:

[0039] P = (m Nt+i ) t To be in the predictive confidence interval of binomial distribution, the range of traffic density and the value of n of binomial distribution are subject to road conditions and situations.

[0040] Further, the selection of the smoothing coefficient α is:

[0041] When the road condition is good, the smoothing coefficient α is between 0.25 and 0.35;

[0042] When the road condition is poor, the smoothing coefficient α is between 0.15 and 0.25.

[0043] Further, the division standard of road service level in step (3) is based on different road service levels, specifically referring to the corresponding indicators of different service level judgment parameters (N t+i , v 预 , K 预 ):

[0044] Level one: Nt+i ∈(0,500) vehicles; v 预 ∈(48,56) km / h; K 预 ∈(0,10) vehicles / s / lane;

[0045] Level Two: N t+i ∈(500,1000) vehicles; v 预 ∈(40,58) km / h; K 预 ∈(10,15) vehicles / s / lane;

[0046] Level Three: N t+i ∈(1000,1500) vehicles; v 预 ∈(32,40) km / h; K 预 ∈(20,25) vehicles / s / lane;

[0047] Level Four: N t+i ∈(1500,2000) vehicles; v 预 ∈(24,32) km / h; K 预 ∈(20,25) vehicles / s / lane;

[0048] Level Five: N t+i ∈(2000,2500) vehicles; v 预 ∈(16,24) km / h; K 预 ∈(25,30) vehicles / s / lane;

[0049] Level Six: N t+i ∈(2500,∞) vehicles; v 预 ∈(0,16) km / h; K 预 ∈(30,∞) vehicles / s / lane.

[0050] Further, the double determination implementation process of step (3) is as follows:

[0051] predicted vehicle arrival number N t+i predicted interval average speed v 预 predicted vehicle density value K 预 road service level range judgment:

[0052]

[0053] wherein, logic represents a judgment statement, X1 is defined as road service levels one to three; X2 is defined as road service levels four to six;

[0054] predicted vehicle arrival rate P 预 complies with Poisson distribution or binomial distribution;

[0055]

[0056] Where logic represents the conditional statement; if the conditional statement is 0, it follows a Poisson distribution; if the conditional statement is 1, it follows a binomial distribution.

[0057] When the parameter of logic(1) is within the service level X1 range and the parameter of logic(2) follows a Poisson distribution, the tidal lane is closed; when the parameter of logic(1) is within the service level X2 range and the parameter of logic(2) follows a binomial distribution, the tidal lane is open.

[0058] Beneficial effects: Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention obtains traffic flow data for a certain period in advance to predict the traffic flow for the next period. By analyzing and calculating four judgment parameters, and then performing dual logical judgment, it determines in advance whether to open the tidal lane, thereby alleviating traffic congestion at different times. The present invention can improve the time utilization rate of drivers and the utilization rate of road space resources during the driving process and the opening of the tidal lane, which can alleviate traffic congestion and reduce the risk of rear-end collisions. Attached Figure Description

[0059] Figure 1 This is a flowchart of the present invention;

[0060] Figure 2 This is a flowchart illustrating the detection and judgment process of the dual determination method for the opening of tidal lanes in this invention.

[0061] Figure 3 This invention detects and predicts east-west traffic flow to activate a unidirectional tidal flow lane map. Detailed Implementation

[0062] The present invention will now be described in further detail with reference to the accompanying drawings.

[0063] like Figure 1 As shown, this invention proposes a dual-determination method for the opening of tidal lanes based on a traffic flow prediction model, which specifically includes the following steps:

[0064] Step 1: Use radar to detect the real-time conditions of urban main roads. Urban main road conditions include: vehicle traffic conditions and the total number of vehicles (denoted as N). 实 ), Passage time (denoted as T) 实 ), for the detected T 实 and N 实 The relevant data was matrixed and uniformly processed into a transpose matrix. A fixed-length road of length L was selected at an intersection of urban main roads, and radar detection was used to obtain the time period T taken by vehicles to pass through this fixed-length road of length L. 实 And record the number N of vehicles that pass through. 实Correlation data. Radar records correlation data with 1 hour as a large cycle and 5 minutes as a small cycle. The recorded T 实 and N 实 are matrixed as:

[0065] Time period T 实 Matrix: [t1, t2, t3, t4………t 12 ] T

[0066] Vehicle number N 实 Matrix: [N1, N2, N3, N4……N 12 ] T .

[0067] Step 2: Establish a quadratic exponential smoothing prediction traffic model and predict the vehicle number N t+i arriving in the subsequent time period. Process the predicted N t+i to obtain the judgment parameters: predicted vehicle arrival number N t+i , predicted interval average speed v 预 , predicted vehicle density K 预 , and predicted vehicle arrival rate P 预 .

[0068] Establish a quadratic exponential smoothing prediction traffic model:

[0069]

[0070]

[0071] In the formula, S (1) t+1 is the first exponential smoothing value of the t+1 time period; S (2) t+1 is the quadratic exponential smoothing value of the t+1 time period; S (1) t is the first smoothing value of the t time period; S (2) t is the quadratic smoothing value of the t time period; N t is the recorded number of vehicles in different time periods; α is the smoothing coefficient selected according to different road conditions. When the road condition is good, the smoothing coefficient α is between 0.25 and 0.35; when the road condition is poor, the smoothing coefficient α is between 0.15 and 0.25.

[0072] Prediction vehicle number equation:

[0073] N t+i = a t +b t i

[0074] at = 2S (1) t - S (2) t

[0075]

[0076] wherein N t+i is the predicted value of the number of vehicles in the t+i period; i is the number of periods from the t period; i is an integer, i ∈ [0, 12]; a t , b t are the coefficients of the prediction model.

[0077] The average speed v 预 of the vehicles in the prediction interval is obtained by judging the parameter: 预 The predicted vehicle density K 自由 (km / h) is selected, and the predicted lane length L 预 = 5v 自 / 60 (km / h) is calculated, and the parameter traffic volume Q 预 , the predicted vehicle flow density K 预 , and the predicted interval speed v 预 are calculated according to the relevant data:

[0078]

[0079]

[0080]

[0081] The predicted vehicle arrival rate P 预 is obtained by judging the parameter: t+i The number of vehicles N t+1 is predicted by the quadratic smoothing index (N t+2 , N t+3 , N t+4 ,..., N t+12 ), and the relationship with the vehicle flow density K 预 is derived, and it is judged whether the K 预 symbol meets the vehicle flow density K 泊松 of the Poisson distribution condition. If it meets the Poisson distribution condition, λ 预 is substituted into the Poisson distribution to calculate the probability of the predicted number of vehicles N t+i in each period; if it does not meet the Poisson distribution condition, λ 预 is substituted into the binomial distribution to calculate the probability of the predicted number of vehicles N t+i in each period:

[0082] The average arrival rate λ 预 is predicted:

[0083]

[0084]

[0085] 1) Poisson distribution:

[0086] λ 预 Substitute:

[0087] 2) Binomial distribution:

[0088]

[0089] λ 预 Substitute:

[0090] The calculated The predicted probability will be subjected to subsequent logical judgment within the confidence interval of Poisson distribution and binomial distribution. The range of application of discrete random distribution of traffic flow density, the value range of m and n is subject to road conditions and circumstances.

[0091] Step 3: Divide the road service level into level one, level two, level three, level four, level five and level six; when the above parameter judgment is in road service level one to three, and P 预 obeys Poisson distribution, the tidal lane is in closed state; when the above parameter judgment is in road service level four to six, and P 预 obeys binomial distribution, the tidal lane is in open state.

[0092] The division standard of road service level is based on different road service levels. Specifically, it refers to the corresponding indicators of different service level judgment parameters (N t+i , vpre, Kpre). For example:

[0093] Service level one: N t+i ∈(0, 500) vehicles; vpre ∈(48, 56) km / h; Kpre ∈(0, 10) vehicles / s / lane;

[0094] Service level two: N t+i ∈(500, 1000) vehicles; vpre ∈(40, 58) km / h; Kpre ∈(10, 15) vehicles / s / lane; Service level three: N t+i ∈(1000, 1500) vehicles; vpre ∈(32, 40) km / h; Kpre ∈(20, 25) vehicles / s / lane; Service level four: N t+i ∈(1500, 2000) vehicles; vpre ∈(24, 32) km / h; Kpre ∈(20, 25) vehicles / s / lane; Service level five: N t+iN < 2000 vehicles; vpre∈ (16, 24) km / h; Kpre∈ (25, 30) vehicles / s / lane; Service Level Six: N t+i N > 2500 vehicles; vpre∈ (0, 16) km / h; Kpre∈ (30, ∞) vehicles / s / lane.

[0095] The judgment process is as follows:

[0096] Parameter N t+i , v 预 , K 预 Road service level range judgment;

[0097]

[0098] Wherein, logic represents the judgment statement, X1 is defined as road service level one to three, and X2 is defined as road service level four to six.

[0099] Parameter P 预 Poisson distribution or binomial distribution;

[0100]

[0101] Wherein, logic represents the judgment statement. If the judgment statement is 0, it is subject to Poisson distribution; if the judgment statement is 1, it is subject to binomial distribution.

[0102] When the parameter of logic (1) is in the service level X1 range, and the parameter of logic (2) is subject to Poisson distribution, the tidal lane is in a closed state; when the parameter of logic (1) is in the service level X2 range, and the parameter of logic (2) is subject to binomial distribution, the tidal lane is in an open state. The relevant X1 parameters N t+i , v 预 , K 预 The range is shown in the following Table 1:

[0103] Table 1: Prediction of relevant parameter range table

[0104]

[0105] Figure 2The detection and judgment flow chart of the double determination method of the tidal lane opening of the application (one-way direction) is as follows: the radar detector counts the number of vehicles on the road to detect and obtain the traffic data in a period; the detected and obtained traffic data is transmitted to the computer for analysis and calculation of the corresponding judgment parameters of the next period; the analyzed and calculated judgment parameters are cyclically input into the computer for corresponding parameter logical judgment; the computer judges that the road service level of the next period is at level 1 to 3 or level 4 to 6; the service level is input into the lane changing robot, and the lane changing robot is operated to open the tidal lane or is in a static state to close the tidal lane.

[0106] According to Figure 3 The detection and prediction of east-west direction (two-way) traffic opening of the single direction tidal lane are shown as follows:

[0107] (1) The east-west direction radar detector counts the number of vehicles on the road to detect and obtain the east-west direction traffic data in a period;

[0108] (2) The detected and obtained traffic data is transmitted to the computer for analysis and calculation of the corresponding judgment parameters of the next period; the specific judgment parameters are as follows:

[0109] (2.1) A quadratic exponential smoothing prediction traffic model is established to predict the number of arriving vehicles, denoted as N t+i . The selection of the smoothing coefficient a should meet the road conditions. The quadratic exponential smoothing prediction traffic model is as follows:

[0110] First-order exponential smoothing:

[0111] Second-order exponential smoothing:

[0112] The relevant prediction equation is N t+i = a t +b t i

[0113] (2.2) N t+i is obtained, and the following four parameters are calculated according to the free vehicle speed: traffic volume Q 预 , predicted traffic density K 预 , predicted interval speed v 预 , and predicted vehicle arrival rate P 预 . The calculation formula is as follows:

[0114]

[0115] Poisson distribution:

[0116] Substituting λ 预 , we get:

[0117] Binomial distribution:

[0118] Lambda 预 Substitute:

[0119] (3) The four judgment parameters calculated by analysis are cycled into the computer, and the corresponding parameter logic judgment is performed. The judgment statement logic is as described in the claim.

[0120] Wherein when the parameter of logic (1) is in the service level X1 range, and the parameter of logic (2) is subject to Poisson distribution, the tidal lane is in a closed state; when the parameter of logic (1) is in the service level X2 range, and the parameter of logic (2) is subject to binomial distribution, the tidal lane is in an open state.

[0121] (4) The service levels X1 and X2 input by the logic language logic judgment are input into the lane changing robot, and the lane changing robot runs to open the tidal lane or is in a static state to close the tidal lane.

[0122] (5) Due to the uneven nature of the road direction, as shown in Figure 3 When the east direction tidal lane is opened, the west direction tidal lane is in a closed state.

[0123] The above embodiments are only for illustrating the technical concept and characteristics of the present application, and the purpose is to enable those skilled in the art to understand the content of the present application and implement it, and cannot limit the protection scope of the present application. Any equivalent transformation or modification according to the spirit and essence of the present application should be covered within the protection scope of the present application.

Claims

1. A method for double determination of tidal lane opening based on a traffic volume prediction model, characterized in that, The method comprises the following steps: (1) Real-time detection of urban trunk road conditions, and preprocessing of the obtained data; (2) Construct a quadratic exponential smoothing prediction model to predict the number of vehicle arrivals N in the subsequent time period t+i ; and predict the average vehicle speed v t+i , the vehicle density value K 预 , and the vehicle arrival rate P 预 in the prediction interval 预 ; (3) dividing the road service level, the division of the road service level is one to six levels, when the parameter judgment is in the range of one to three of the road service level, and the vehicle arrival rate P 预 obeys the Poisson distribution, after double judgment, the tidal lane is in the closed state; when the parameter judgment is in the range of four to six of the road service level, and P 预 obeys the binomial distribution, after double judgment, the tidal lane is in the open state; The secondary exponential smoothing prediction model of step (2) is: In the formula, S (1) t+1 S is the first exponential smoothing value of the t+1 period; S (2) t+1 S is the second exponential smoothing value of the t+1 period; S (1) t S is the first smoothing value of the t period; S (2) t S is the second smoothing value of the t period; N t S is the number of records of the vehicle in different periods; and α is the smoothing coefficient selected according to different road conditions. Prediction vehicle number equation: N t+i = a t + b t i a t = 2S (1) t -S (2) t In the formula, N t+i is the vehicle number prediction value of the t+i period; i is the period number from the t period; i is an integer, i∈[0, 12]; a t , b t is a prediction model coefficient; The predicted interval average vehicle speed v 预 The predicted vehicle density K 预 The implementation process is as follows: The selected free-flow speed is v 自由 , the predicted lane length L 预 = 5v 自 / 60, the predicted value for a 5-minute period, the traffic volume Q 预 , the flow density K 预 , and the section speed v 预 are calculated according to the relevant data where N t+i is the predicted number of vehicles for the t+i period; L is the 5-minute travel length calculated at free speed; K 预 is the predicted traffic density; Q 预 is the predicted average traffic volume; v 预 is the predicted average vehicle speed for the interval The vehicle arrival rate P described in step (2) 预 The prediction is implemented as follows: N t+i N t+1 N t+2 N t+3 N t+4 N t+12 N 预 N 预 N 泊松 N 预 N t+i N 预 N t+i ​ Predicted average arrival rate λ 预 : Poisson distribution: where P Nt+i is the probability of predicting the arrival of N t+i vehicles within a counting interval t; λ 预 is the predicted average arrival rate; and t is the duration of each counting interval. e is the base of natural logarithm; λ 预 Substituting, we get: P Nt+i In order to be in the prediction confidence interval of Poisson distribution, the range of traffic density and the value of n of Poisson distribution should be based on road conditions and situations. Binomial distribution: where P Nt+i is the probability of predicting the arrival of N t+i vehicles within a counting interval t; λ 预 is the predicted average arrival rate; and t is the duration of each counting interval. m is a positive integer value; λ 预 Substituting gives: P Nt+i The range of traffic density and the value of n of binomial distribution should be determined according to the road conditions and situations in order to be within the prediction confidence interval of binomial distribution. 2.The method of claim 1, wherein, The implementation process of step (1) is as follows: Real-time detection of urban trunk road conditions by radar, including vehicle traffic conditions, total number of vehicles, and vehicle travel time, matrix processing of the detected total number of vehicles and travel time related data, and unified processing into a transpose matrix. 3.The method of claim 1, wherein, The selection of the smoothing coefficient a is: When the road condition is good, the smoothing coefficient a is between 0.25 and 0.35; When the road condition is poor, the smoothing coefficient a is between 0.15 and 0.

25. 4.The method of claim 1, wherein, The classification criteria of the road service level in step (3) are based on different road service levels, specifically referring to the judgment parameters (N t+i 、 预 、 预 ) corresponding to the indexes: Class one: N t+i ∈(0,500) vehicles; v 预 ∈(48,56) km / h; K 预 ∈(0,10) vehicles / s / lane; Class two: N t+i ∈(500,1000) vehicles; v 预 ∈(40,58) km / h; K 预 ∈(10,15) vehicles / s / lane; Class three: N t+i ∈(1000,1500) vehicles; v 预 ∈(32,40) km / h; K 预 ∈(20,25) vehicles / s / lane; Class four: N t+i ∈(1500,2000) vehicles; v 预 ∈(24,32) km / h; K 预 ∈(20,25) vehicles / s / lane; Level five: N t+i ∈(2000,2500) vehicles; v 预 ∈(16,24) km / h; K 预 ∈(25,30) vehicles / s / lane; Class six: N t+i ∈(2500,∞) vehicles; v 预 ∈(0,16) km / h; K 预 ∈(30,∞) vehicles / s / lane. 5.The method of claim 1, wherein, The implementation process of step (3) is as follows: Predicted number of vehicles N t+i , Predicted average speed v 预 , Predicted density value K 预 Road service level range determination: Wherein, logic represents a judgment statement, X1 is defined as road service level one to three, and X2 is defined as road service level four to six. Predicted vehicle arrival rate P 预 Poisson or binomial distribution Wherein, logic represents a judgment statement; if the judgment statement is 0, it obeys Poisson distribution; if the judgment statement is 1, it obeys binomial distribution. When the parameters of logic(1) are in the service level X1 range, and the parameters of logic(2) obey Poisson distribution, the tidal lane is in the closed state; when the parameters of logic(1) are in the service level X2 range, and the parameters of logic(2) obey binomial distribution, the tidal lane is in the open state.

Images