Green wave coordination clearing time calculation method based on sampling trajectory data

Abstract

The invention discloses a green wave coordination clearing time calculation method based on sampling trajectory data, which comprises the following steps of: 1) for a coordination path P (i-j) from an upstream intersection Ii to a downstream intersection Ij, classifying and defining traffic flows fS (i-j) in each flow direction related to the coordination path P (i-j), detecting a parking state S (n, j) and a parking moment TS (n, j) of a sampling vehicle n at the downstream intersection Ij, and calculating the green wave coordination clearing time of the coordination path P (i-j) according to the parking state S (n, j) and the parking moment TS (n, j); the actual passing time TP (n, j) and the ideal passing time TE (n, j) are calculated, and the number of arriving vehicles is calculated in combination with the sampled vehicle tracks; 2) constructing an arrival likelihood function of the traffic flow fS (i-j) in each flow direction, and estimating an arrival flow rate lambda S (i-j) of the traffic flow fS (i-j) in each flow direction; (3) calculating the maximum clearing time # imgabs0 #; (4) determining the passing condition of the target coordination vehicle at the downstream intersection Ij according to the coordination phase green light duration of the upstream intersection and the downstream intersection and the maximum clearing time # imgabs1 #, and calculating the clearing time tC (i-j) of the target coordination vehicle flow fO (i-j) or the target coordination updating vehicle flow fH (i-j). According to the invention, the emptying time of the target coordination vehicle can be effectively calculated.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic signal control, and more specifically, to a method for calculating green wave coordinated clearing time based on sampled trajectory data. Background Art

[0002] At present, the green wave coordination design methods of urban roads can be divided into three types: the first type, the middle type, and the tail type. When the coordination path adopts the first type of green wave coordination design, it is necessary to calculate the queue clearance time for the target coordinated vehicle. However, there are few studies on how to calculate the target coordinated vehicle clearance time by estimating the arrival flow rate of various coordinated vehicle flows based on sampled trajectory data. With the increase of networked vehicles in today's society, trajectory data is more meaningful for research.

[0003] In this regard, the present invention provides a method for calculating the green wave coordinated clearing time based on sampled vehicle trajectory data. Classify and define the traffic flows related to the coordinated path, calculate the actual passing time and ideal passing time of the sampled vehicles at the downstream intersection, calculate the number of arriving vehicles based on the sampled vehicle trajectory, construct an arrival likelihood function, estimate the arrival flow rate of the traffic flows in each direction, calculate the maximum clearing time of all non-target coordinated traffic flows, determine the passing of the target coordinated vehicle at the downstream intersection in combination with the green light duration of the coordinated phase of the upstream and downstream intersections, and calculate the clearing time of the target coordinated vehicle. Summary of the invention

[0004] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide a method for calculating the green wave coordinated clearance time based on sampled trajectory data. Based on the analysis of the green wave coordinated control demand characteristics, the present invention provides the definition and calculation method of the relevant parameters of the sampled trajectory vehicle coordinated control, which provides a research idea for the arrival flow rate estimation.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] The present invention provides a method for calculating green wave coordinated clearing time based on sampled trajectory data, comprising the following steps:

[0007] S1, for the upstream intersection I i To downstream intersection I j The coordination path P (i→j) , classification definition and related flow of vehicles f S(i→j) , detect sampling vehicle n at downstream intersection I j Parking status S (n,j) and parking time T S(n,j) , calculate its actual passing time T P(n,j) and the ideal passing time T E(n,j) , the number of arriving vehicles is calculated by combining the sampled vehicle trajectories;

[0008] S2. Construct the traffic flow f in each direction S(i→j) The arrival likelihood function estimates the traffic flow f in each direction. S(i→j) The arrival flow rate λ S(i→j) ;

[0009] S3. Calculate the maximum clearing time of all non-target coordinated traffic flows

[0010] S4. Coordinate the phase green light duration and maximum clearing time according to the upstream and downstream intersections Determine the target coordinated vehicle at the downstream intersection I j The passing situation of the target coordinated vehicle flow f is calculated O(i→j) Or target coordinated update of traffic flow f H(i→j) The clearing time t C(i→j) .

[0011] As a preferred technical solution, the step S1 is specifically as follows:

[0012] S101, for the upstream intersection I i To downstream intersection I j The coordination path P (i→j) , classification definition and related flow of vehicles f S(i→j) ;

[0013] When the upstream intersection I i To downstream intersection I j Coordination Path P (i→j) The goal of coordinating traffic flow f O(i→j) When designing the green wave coordination of the first type of team, the target coordinated traffic flow f under different circumstances O(i→j) , non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) and f 3(i→j) The definition and selection of are shown in Table 1, where f (i,x→j,y) From the upstream intersection I i The x flow enters the downstream intersection I j The y-flow direction of the vehicle leaving, when x is 1, 2, and 3, it represents the vehicle flow from the upstream intersection I i Turn left, go straight, turn right and go to downstream intersection I j , y takes the values ​​of 1, 2, and 3, which respectively represent that the vehicle flow turns left, goes straight, and turns right at the current entrance to leave the downstream intersection I j , traffic flow (i,x→j,D) The D direction is the target coordinated vehicle flow f O(i→j) At downstream intersection I j The departure direction of the dedicated lane, the traffic flow f (i,x→j,D′) The D′ flow direction is the target coordinated vehicle flow f O(i→j)At downstream intersection I j All departure directions of the shared lane, define a Boolean variable F R(i→j) To determine the upstream intersection I i Define the Boolean variable F to determine whether the right-turn merge flow is controlled. C(i→j) To determine the downstream intersection I j Coordinate whether the entrance lane of the route is a shared lane;

[0014] Table 1 Definitions of target coordinated traffic flow and non-target coordinated traffic flow in different situations

[0015]

[0016] Target coordinated traffic flow O(i→j) The arrival flow rate is denoted as λ O(i→j) , non-target coordinated traffic flow f 1(i→j) and f 2(i→j) The arrival flow rate is denoted as λ 1(i→j) and λ 2(i→j) ; If the upstream intersection I i The right-turn merging traffic flow is not controlled by the signal, so it is defined as the non-target coordinated traffic flow f 3(i→j) The arrival flow rate λ 3(i→j) Perform calculations;

[0017] According to the phase setting of the cycle signal, the traffic flow in the cycle is merged into the downstream intersection I j There are two situations: one is the upstream intersection I i The signal phase is set so that there is no mixing between vehicles entering the downstream intersection I j The second is the upstream intersection I i The signal phase setting allows mixed traffic to enter the downstream intersection I j phenomenon;

[0018] When the upstream intersection I i The signal phase is set so that there is no mixing between vehicles entering the downstream intersection I j When the phenomenon occurs, estimate the target coordinated traffic flow f O(i→j) Coordinate traffic flow with non-targets 1(i→j) 、f 2(i→j) The vehicle arrival situation; when the upstream intersection I i The signal phase setting makes the traffic flow mixed and enter the downstream intersection I j If the traffic flow released in the same period is a non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) or 3(i→j) Two of them are merged and recorded as non-target coordinated merged traffic flow f 4(i→j) , whose arrival flow rate is λ 4(i→j), if the released traffic flow in the same period has both target coordinated traffic flow f O(i→j) There are also other non-target coordinated traffic flows f 1(i→j) 、f 2(i→j) or 3(i→j) , and merge them into the target coordinated update traffic flow f H(i→j) , whose arrival flow rate is λ H(i→j) ;

[0019] According to the above traffic flow division results, the coordinated path P (i→j) The relevant traffic flow in each direction is defined as f S(i→j) , S takes O, H, 1, 2, 3, and 4, representing the target coordinated traffic flow f O(i→j) , target coordinated update of traffic flow f H(i→j) , non-target coordinated traffic flow f 1(i→j) , non-target coordinated traffic flow f 2(i→j) , non-target coordinated traffic flow f 3(i→j) and non-target coordinated merging of traffic flows f 4(i→j) ;

[0020] S102, detecting sampling vehicle n at downstream intersection I j Parking status S (n,j) and parking time T S(n,j) , calculate its actual passing time T P(n,j) and the ideal passing time T E(n,j) ; If the sampling vehicle n is at the downstream intersection I j There is a parking state, parking state S (n,j) Take 1, otherwise take 0; if the parking state S of the sampled vehicle n (n,j) is 0, based on the sampling vehicle n at intersection I j The detection time of the trajectory points before and after the stop line and the distance between them and the stop line are calculated by interpolation method. j The actual passing time T of the stop line at the entrance of the coordinated path P(n,j) ; If the parking state S of the sampled vehicle n (n,j) is 1, detecting sample vehicle n at intersection I j The first stop time T of the coordinated path entrance S(n,j) , calculate the ideal number of sampled vehicles n passing through intersection I j The ideal passing time T of the stop line at the entrance of the coordinated path E(n,j) , as shown in formula (1):

[0021]

[0022] Where D (n,j) is the sampling vehicle n at intersection I jThe distance between the coordinated path entrance lane and the stop line when the first stop is made, is the average ideal driving speed of the vehicle on the current road section;

[0023] S103, according to the flow of vehicles f S(i→j) Arrival situation Calculate the number of arriving vehicles; Discretize the continuous arrival time within the signal cycle with a time interval of k seconds. m The number of vehicles that ideally pass the intersection stop line during the time period is counted, and the vehicle arrival ratio P(t m ) is shown in formula (2):

[0024]

[0025] Where N is the total number of vehicles whose trajectories can be collected, m is the time interval number after discrete processing, and F{T E(n,j) =t m} is the ideal passing time T for determining the collection of any sampled vehicle n E(n,j) Is it in t m The indicator variable within a period is calculated as shown in formula (3):

[0026]

[0027] Traverse the sampled trajectory data of the S flow to the vehicle flow, and sequentially include the vehicle trajectory into the data set of different situations U. S takes O, H, 1, 2, 3, and 4, which represent the target coordinated vehicle flow f respectively. O(i→j) , target coordinated update of traffic flow f H(i→j) , non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) 、f 3(i→j) and non-target coordinated merging of traffic flows f 4(i→j) , case U is 1, 2, 3, case 1 corresponds to the first sampled vehicle n in all S-flow traffic within the signal period A The driving trajectory is the parking trajectory. Case 2 corresponds to the front vehicle of two adjacent sampled vehicles in the S flow within the signal period. With the following car B The driving trajectories are all parking trajectories. Case 3 corresponds to the front vehicle of two adjacent sampled vehicles in the S flow within the signal period. With the following car C The driving trajectories are parking trajectories and non-parking trajectories respectively; other sampling trajectory collection situations cannot provide effective information for traffic arrival flow rate estimation, so they are not considered;

[0028] For the sampled trajectory data of the S-flow traffic, all trajectory data that satisfy condition 1 are recorded as set A, and any sampled vehicle n in set A is AThe estimated number of arriving vehicles is recorded as All trajectory data that satisfy condition 2 are recorded as set B, and any adjacent sampled vehicles in set B are With n B The estimated number of arriving vehicles is recorded as All trajectory data that satisfy condition 3 are recorded as set C, and any adjacent sampled vehicles in set C are With n C The estimated number of arriving vehicles is recorded as

[0029] As a preferred technical solution, in step S103, for any sampled vehicle n in set A, A , estimate the coordinated phase red light on time The ideal passing time with the current vehicle trajectory Downstream intersection I j The number of vehicles arriving on the coordinated path corresponding to the traffic flow As shown in formula (4) and formula (5):

[0030]

[0031]

[0032]

[0033] In the formula, [] is the rounding operator, To coordinate the phase green light on time To sample vehicle n A Actual passing time The length of the green light between and are the sampled vehicles n A Pass downstream intersection I j The time corresponding to the coordinated phase of the red and green lights, h s is the saturation headway, Poisson represents the number of arriving vehicles Follows Poisson distribution, λ S(i→j) is the arrival flow rate of traffic to S, The number of vehicles arriving The corresponding time interval.

[0034] As a preferred technical solution, in step S103, for any adjacent sampled vehicles in set B, With n B , estimating the sampled vehicles Ideal passing time With sampled vehicles n B Ideal passing time Number of vehicles arriving between The calculation satisfies formula (7):

[0035]

[0036]

[0037]

[0038] In the formula, For sampling vehicles Actual passing time With sampled vehicles n B Actual passing time The length of the green light between The number of vehicles arriving The corresponding time interval.

[0039] As a preferred technical solution, in step S103, for any adjacent sampled vehicles in set C, With n C , estimating the sampled vehicles Ideal passing time With sampled vehicles n C Actual passing time Number of vehicles arriving between The calculation satisfies formula (10):

[0040]

[0041]

[0042]

[0043] In the formula, For sampling vehicles Actual passing time With sampled vehicles n C Actual passing time The length of the green light between The number of vehicles arriving The corresponding time interval.

[0044] As a preferred technical solution, step S2 is specifically as follows:

[0045] S201, constructing the traffic flow f in each direction S(i→j) The arrival likelihood function of ;

[0046] According to the three sets obtained in step S103, the arrival likelihood function for the S-flow is constructed as shown in formula (13), where h is any adjacent sampled vehicle in set C. With nC The number of vehicles that may exist between the two, the maximum value is the number of vehicles arriving A total of a sampled vehicle trajectories are collected in set A, a total of b pairs of adjacent sampled vehicle trajectories are collected in set B, and a total of c pairs of adjacent sampled vehicle trajectories are collected in set C;

[0047]

[0048] S202: Estimating the traffic flow f in each direction S(i→j) The arrival flow rate λ S(i→j) ; Based on Bayesian inference, take exponential distribution as the arrival flow rate λ S(i→j) The prior distribution of exponential distribution parameter β S(i→j) is a known empirical value, and the arrival flow rate λ S(i→j) is a positive value, then the arrival flow rate λ S(i→j) The prior distribution function and posterior probability expressions of are shown in equations (14) and (15):

[0049]

[0050]

[0051] The maximum a posteriori estimation method is used to estimate the parameters by using the parameter variable value when the posterior probability is the largest as the point estimate of the parameter. The denominator of the posterior probability, P(D), is a constant value and will not affect the maximum value point of the posterior probability function, so it is ignored. The posterior probability function is logarithmically processed and simplified to obtain formula (16):

[0052]

[0053] for The estimated value of is calculated by equation (17), where l represents the number of loop iterations and g represents any adjacent sampled vehicle in set C. With n C The maximum number of vehicles that may exist between

[0054]

[0055] S flows to the traffic flow reaching flow rate λ S(i→j) The average value of The maximum a posteriori estimation method is combined with formula (17) to iterate until convergence, and the result satisfies formula (18):

[0056]

[0057] As a preferred technical solution, the step S202 also includes the following content:

[0058] The convergent average arrival flow rate Multiply it by the vehicle arrival ratio to obtain the vehicle flow f in each direction S(i→j) The arrival flow rate λ at different time periods S(i→j) The expression satisfies formula (19):

[0059]

[0060] As a preferred technical solution, step S3 is specifically as follows:

[0061] S301. Calculate the maximum clearing time of all non-target coordinated vehicle flows The clearing time of all non-target coordinated traffic is defined as t CT ( i→j ), is the maximum clearing time of all non-target coordinated traffic flows, calculated to satisfy equation (20); t S(i→j) Corresponding to the traffic flow f in each direction S(i→j) The time it takes for the queue of vehicles to be cleared, The clearing time t S(i→j) The maximum clearing time to obtain the maximum value, Q S(i→j) Corresponding to the traffic flow f in each direction S(i→j) The number of vehicles in the queue, is the number of queued vehicles Q S(i→j) The maximum number of vehicles in the queue that reaches the maximum value, S (i→j) Coordinate traffic flow for the goal f O(i→j) Or target coordinated update of traffic flow f H(i→j) The Boolean variable F is the sum of the saturated flows of all lanes in the coordinated path. S(i→j) Used to determine whether there is traffic flow f in the traffic flow division result S(i→j) , there is traffic flow f S(i→j) Then F S(i→j) Takes 1, otherwise takes 0, where S takes O, H, 1, 2, 3, 4;

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] In the formula, is the average starting loss time of the queued vehicles, T 1,j With T 2,j For non-target coordinated traffic flow f 1(i→j)With f 2(i→j) The first vehicle ideally passes through downstream intersection I j The ideal time to pass the stop line, T F,j Coordinate traffic flow for the goal f O(i→j) Or target coordinated update of traffic flow f H(i→j) The first vehicle ideally passes through downstream intersection I j The ideal time to pass the stop line, T 4,j For non-target coordinated merging of traffic flows f 4(i→j) The first vehicle ideally passes through downstream intersection I j The ideal time to pass the stop line, T O,j With T E,j Downstream intersection I j The start and end time points of the cycle in which the target coordination phase is located, T D,j For downstream intersection I j The green light end time of the coordination phase.

[0068] As a preferred technical solution, step S4 is specifically as follows:

[0069] S401, based on the green light duration of the upstream and downstream intersections and the maximum clearing time of all non-target coordinated traffic flows Determine the target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) At downstream intersection I j The downstream intersection I j Green light duration t of the coordinated phase G(j→k) Coordinate the maximum clearing time of traffic with all non-targets and upstream intersection I i Green light duration t of the coordinated phase G(i→j) If the sum is compared, Then the target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) If all vehicles in the queue can pass through the downstream intersection in the current cycle and all vehicles in the queue are composed of non-target coordinated vehicles, go to step S402; Then the target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Unable to pass all downstream intersections in the current cycle, the queued vehicles are composed of non-target coordinated traffic flow and target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Composition, go to step S403;

[0070] S402, calculate the non-existent target coordinated vehicle flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Clearing time t in case of parking waitingC(i→j) ; define t C(i→j) From upstream intersection I i To downstream intersection I j The goal of coordinating traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) The time required to clear the queued vehicles when the head-of-line green wave coordination design is adopted; if there is no target coordinated vehicle to stop and wait, then t C(i→j) =t CT ( i→j) ; Define the Boolean variable F AS ( i→j) To determine the traffic flow f S(i→j) Whether the release phase of the target coordination phase is closely followed by the target coordination phase, if so, the Boolean variable F AS(i→j) Take 1, otherwise take 0, where S takes 1, 2, 4; F AS(i→j) Satisfying formula (25):

[0071] F A1(i→j) +F A2(i→j) +F A4(i→j) ≤1 (25)

[0072] If F A1(i→j) +F A2(i→j) +F A4(i→j) =1, then there is a non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) Or non-target coordinated merging of traffic flows f 4(i→j) The release phase follows the target coordination phase, and some non-target coordinated vehicles pass through the downstream intersection I in the current cycle. j ; The green light duration of the traffic flow that is allowed to follow the target coordinated phase to pass through the downstream intersection in the current cycle is defined as t T(i→j) , satisfying formula (26); at this time, the clearing time t C(i→j) Satisfies formula (27);

[0073] t T(i→j) =t G(j→k) -t G(i→j) -t C(i→j) (26)

[0074]

[0075] In the formula, Q T(i→j) is the duration of the green light for the corresponding traffic flow T(i→j) The number of vehicles that can leave the downstream intersection within 100 km is calculated by equation (28). is the corresponding traffic flow f S(i→j) During the green light duration t T(i→j)The average arrival flow rate in the area, S is taken as 1, 2, 4, satisfying formula (29), and the green light duration t T(i→j) Satisfies formula (30);

[0076]

[0077]

[0078]

[0079] Where, T L,j Coordinate traffic flow for the goal f O(i→j) Or target coordinated update of traffic flow f H(i→j) The last vehicle ideally passes downstream intersection I j Ideal passing time of the stop line, calculate the clearing time t C(i→j) Satisfies formula (31):

[0080]

[0081] If F A1(i→j) +F A2(i→j) +F A4(i→j) = 0, then there is no non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) Or non-target coordinated merging of traffic flows f 4(i→j) The release phase is followed by the target coordination phase, and the target coordinated vehicle flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) The clearing time t C(i→j) Equal to the maximum clearing time of all non-target coordinated traffic flows End the optimization process;

[0082] S403, calculate the target coordinated vehicle flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Clearing time t in case of parking waiting C(i→j) ; If there is a target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Stop and wait, then the goal is to coordinate the traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) The clearing time t C(i→j) Satisfy formula (32); Set the coordination phase green light end time T D,j Until the ideal passing time of the last target coordinated vehicle T L,j The ideal release time between W(i→j) Satisfies formula (33);

[0083]

[0084]

[0085]

[0086] In the formula, For traffic flow S(i→j) At the ideal release time t W(i→j) The average arrival flow rate within.

[0087] As a preferred technical solution, the step S403 also includes the following content:

[0088] Calculate the ideal release time t W(i→j) Satisfying formula (35), clearing time t C(i→j) When formula (36) is satisfied, the optimization process ends;

[0089]

[0090] Compared with the prior art, the present invention has the following beneficial effects:

[0091] 1) Based on the analysis of the green wave coordinated control demand characteristics, the present invention provides the definition and calculation method of the relevant parameters of the sampled trajectory vehicle coordinated control, and provides a research idea for the arrival flow rate estimation.

[0092] 2) The present invention can estimate the arrival flow rates of various coordinated vehicle flows based on the sampled trajectory data, and provides a research method for vehicle flow level arrival flow rate estimation for green wave coordination needs.

[0093] 3) The present invention calculates the number of queued vehicles in combination with the arrival flow rates of the traffic flows in each direction, provides a method for calculating the queue clearing time of target coordinated vehicles, and provides a technical basis for the green wave coordination design of the head of the queue. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 It is a flow chart of a method for calculating the coordinated clearing time of a green wave based on sampled trajectory data;

[0095] Figure 2 It is a schematic diagram of the arrival flow rate within the cycle mentioned in the present invention;

[0096] Figure 3 is a schematic diagram of any sampled vehicle in the set A proposed by the present invention;

[0097] Figure 4 is a schematic diagram of any adjacent sampled vehicles in the set B proposed by the present invention;

[0098] Figure 5 is a schematic diagram of any adjacent sampled vehicles in the set C proposed by the present invention;

[0099] Figure 6It is a schematic diagram of the queue situation of the traffic flows in each direction proposed by the present invention;

[0100] Figure 7 It is a schematic diagram of the classification of vehicle arrival conditions proposed by the present invention;

[0101] Figure 8 This is a schematic diagram of the non-target coordination phase closely following the target coordination phase proposed by the present invention;

[0102] Fig. 9 This is a schematic diagram of the non-target coordination phase not closely following the target coordination phase proposed by the present invention;

[0103] Fig.10 This is a schematic diagram of the parking waiting situation of the target coordinated vehicle proposed in the present invention;

[0104] Fig.11 It is a schematic diagram of the arrival flow rate within the cycle mentioned in the present invention;

[0105] Fig.12 This embodiment collects non-target merging vehicles f in a certain period. 4(1→2) The trajectory time diagram of

[0106] Fig.13 is the target coordinated update of vehicle flow f in this embodiment H(1→2) Arrival flow rate diagram;

[0107] Fig.14 is the non-target coordinated merged traffic flow f in this embodiment 4(1→2) Arrival flow rate diagram;

[0108] Fig.15 is the non-target coordinated vehicle flow f in this embodiment 3(1→2) Arrival flow rate diagram. DETAILED DESCRIPTION

[0109] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0110] The intersection of Fenjiang Middle Road and Tongji Road in Chancheng District, Foshan was selected as the upstream intersection I 1 (I i , i=1), the intersection of Tiyu Road and Tongji Road is the downstream intersection I 2 (I j , j = 2), the intersection structure and signal control scheme are as follows Fig.11 As shown, in this embodiment, the straight-moving direction from west to east is taken as the target coordination direction.

[0111] like Figure 1 As shown, the present embodiment provides a method for calculating the green wave coordinated clearing time based on sampled trajectory data, comprising the following steps:

[0112] Step S1: For the upstream intersection I1 To downstream intersection I 2 The coordination path P (1→2) , classification definition and related flow of vehicles f S(1→2) , detect sampling vehicle n at downstream intersection I 2 Parking status S (n,2) and parking time T S(n,2) , calculate its actual passing time T P(n,2) and the ideal passing time T E(n,2) , combined with the sampled vehicle trajectories to calculate the number of arriving vehicles, the specific steps are as follows:

[0113] S101, for the upstream intersection I 1 To downstream intersection I 2 The coordination path P (1→2) , classification definition and related flow of vehicles f S(1→2) .

[0114] When the upstream intersection I 1 To downstream intersection I 2 Coordination Path P (1→2) The goal of coordinating traffic flow f O(1→2) When designing the green wave coordination of the first type of team, the target coordinated vehicle flow f in this embodiment is O(1→2) , non-target coordinated traffic flow f 1(1→2) 、f 2(1→2) and f 3(1→2) The definition and selection of are shown in Table 1, where f (1,x→2,y) From the upstream intersection I 1 Enter with x direction and leave downstream intersection I with y direction 2 When x is 1, 2, or 3, it represents the traffic from the upstream intersection I. 1 Turn left, go straight, turn right and go to downstream intersection I 2 , y takes the values ​​of 1, 2, and 3, which respectively represent that the vehicle flow turns left, goes straight, and turns right at the current entrance to leave the downstream intersection I 2 , traffic flow (1,x→2,D) The D direction is the target coordinated vehicle flow f O(1→2) At downstream intersection I 2 The departure direction of the traffic flow f (1,x→2,D′) Where D′ is the target coordinated traffic flow f O(1→2) At downstream intersection I 2 All departure directions of the shared lane, define a Boolean variable F R(1→2) To determine the upstream intersection I 1 Define the Boolean variable F to determine whether the right-turn merge flow is controlled. C(1→2) To determine the downstream intersection I j Coordinate whether the route entrance is a shared lane.

[0115] Table 1 Definition of target coordinated traffic flow and non-target merged traffic flow in this case

[0116]

[0117] Target coordinated traffic flow O(1→2) The arrival flow rate is denoted as λ O(1→2) , non-target coordinated traffic flow f 1(1→2) The arrival flow rate is denoted as λ 1(1→2) In this embodiment, the upstream intersection I 1 The right-turn merging traffic flow is not controlled by the signal, so it is defined as the non-target coordinated traffic flow f 3(1→2) The arrival flow rate λ 3(1→2) Perform calculations.

[0118] According to the phase setting of the periodic signal, the traffic flow in the period can be divided into two types: the upstream intersection I 1 The signal phase is set so that there is no mixing between vehicles entering the downstream intersection I 2 The second is the upstream intersection I 1 The signal phase setting allows mixed traffic to enter the downstream intersection I 2 phenomenon.

[0119] In this embodiment, the upstream intersection I 1 The signal phase setting makes the traffic flow mixed and enter the downstream intersection I 2 Phenomenon, upstream intersection I 1 The traffic released during the left-turn phase release period at the north entrance is non-target coordinated traffic f 1(1→2) and f 3(1→2) , and record their merger as non-target coordinated merged traffic flow f 4(1→2) , whose arrival flow rate is λ 4(1→2) , upstream intersection I 1 During the release period of the through phase of the west entrance, there is both a target coordinated traffic flow f O(1→2) There is also a non-target coordinated traffic flow f 3(1→2) , and merge them into the target coordinated update traffic flow f H(1→2) , whose arrival flow rate is λ H(1→2) ,like Figure 2 As shown in part (a) and part (b) of FIG.

[0120] According to the above traffic flow division results, the coordinated path P (1→2) The relevant traffic flow in each direction is defined as f S(1→2) , S can be O, H, 1, 2, 3, 4, representing the target coordinated traffic flow f O(1→2) , target coordinated update of traffic flow f H(1→2) , non-target coordinated traffic flow f1(1→2) , non-target coordinated traffic flow f 2(1→2) , non-target coordinated traffic flow f 3(1→2) and non-target coordinated merging of traffic flows f 4(1→2) .

[0121] S102, detecting sampling vehicle n at downstream intersection I 2 Parking status S (n,2) and parking time T S(n ,2 ) , calculate its actual passing time T P(n,2) and the ideal passing time T E(n,2) If the sampling vehicle n is at the downstream intersection I 2 There is a parking state, parking state S (n,2) Take 1, otherwise take 0. If the parking state S of the sampled vehicle n (n,2) is 0, based on the sampling vehicle n at intersection I 2 The detection time of the trajectory points before and after the stop line and the distance between them and the stop line are calculated by interpolation method. 2 The actual passing time T of the stop line at the entrance of the coordinated path P(n,2) If the parking state S of the sampled vehicle n (n,2) is 1, detecting sample vehicle n at intersection I 2 The first stop time T of the coordinated path entrance S(n,2) , calculate the ideal number of sampled vehicles n passing through intersection I 2 The ideal passing time T of the stop line at the entrance of the coordinated path E(n,2) , as shown in formula (1).

[0122]

[0123] Where D (n,2) is the sampling vehicle n at intersection I 2 The distance between the coordinated path entrance lane and the stop line when the first stop is made, It is the average ideal driving speed of the vehicle on the current road section.

[0124] Table 2 Non-target merging traffic flow f in a certain period 4(1→2) data

[0125]

[0126] Collect the non-target coordinated merged traffic flow f within a certain period 4(1→2) For example, it passes through the downstream intersection I 2 The parking state quantity S before the stop line (n,1→2) , parking time T S(n,2) , actual passing time T P(n,2)and the ideal passing time T E(n,2) As shown in Table 2.

[0127] S103, according to the flow of vehicles f S(1→2) Arrival situation Calculate the number of arriving vehicles. Discretize the continuous arrival time within the signal cycle with k = 1 second as the time interval. m The number of vehicles that ideally pass the intersection stop line during the time period is counted, and the vehicle arrival ratio P(t m ) as shown in formula (2).

[0128]

[0129] Where N is the total number of vehicles whose trajectories can be collected, m is the time interval number after discrete processing, and F{T E(n,2) =t m} is the ideal passing time T for determining the collection of any sampled vehicle n E(n,2) Is it in t m The indicator variable within a period is calculated as shown in formula (3).

[0130]

[0131] Traverse the sampled trajectory data of the S flow to the vehicle flow, and sequentially include the vehicle trajectory into the data set of different situations U. S can take O, H, 1, 2, 3, and 4, which represent the target coordinated vehicle flow f respectively. O(1→2) , target coordinated update of traffic flow f H(1→2) , non-target coordinated traffic flow f 1(1→2) 、f 2(1→2) 、f 3(1→2) and non-target coordinated merging of traffic flows f 4(1→2) , case U can be 1, 2, 3, case 1 corresponds to the first sampled vehicle n in all S-flow traffic within the signal period A The driving trajectory is the parking trajectory. Case 2 corresponds to two adjacent sampled vehicles in the S flow within the signal period. (front car) and n B (The following vehicle) has a parking trajectory. Case 3 corresponds to two adjacent sampled vehicles in the S-flow traffic flow within the signal period. (front car) and n C The driving trajectories of the following vehicles are respectively parking trajectories and non-parking trajectories. Other sampling trajectory collection situations cannot provide effective information for traffic arrival flow rate estimation, so they are not considered.

[0132] For the sampled trajectory data of the S-flow traffic, all trajectory data that satisfy condition 1 are recorded as set A, and any sampled vehicle n in set A is A The estimated number of arriving vehicles is recorded as All trajectory data that satisfy condition 2 are recorded as set B, and any adjacent sampled vehicles in set B are With n B The estimated number of arriving vehicles is recorded as All trajectory data that satisfy condition 3 are recorded as set C, and any adjacent sampled vehicles in set C are With n C The estimated number of arriving vehicles is recorded as

[0133] a) For any sampled vehicle n in set A A , estimate the coordinated phase red light on time The ideal passing time with the current vehicle trajectory Downstream intersection I 2 The number of vehicles arriving on the coordinated path corresponding to the traffic flow As shown in formula (4) and formula (5).

[0134]

[0135]

[0136]

[0137] In the formula, [] is the rounding operator, To coordinate the phase green light on time To sample vehicle n A Actual passing time The length of the green light between and are the sampled vehicles n A Pass downstream intersection I 2 The time corresponding to the coordinated phase of the red and green lights, h s =2(s) is the saturation headway, Poisson is the number of arriving vehicles Follows Poisson distribution, λ S(1→2) is the arrival flow rate of traffic to S, The number of vehicles arriving The corresponding time interval, such as Figure 3 shown.

[0138] b) For any adjacent sampled vehicles in set B With n B ,like Figure 4 As shown, the estimated sample vehicles Ideal passing time With sampled vehicles n B Ideal passing time Number of vehicles arriving between The calculation satisfies formula (7).

[0139]

[0140]

[0141]

[0142] In the formula, For sampling vehicles Actual passing time With sampled vehicles n B Actual passing time The length of the green light between The number of vehicles arriving The corresponding time interval.

[0143] c) For any adjacent sampled vehicles in set C With n C ,like Figure 5 As shown, the estimated sample vehicles Ideal passing time With sampled vehicles n C Actual passing time Number of vehicles arriving between The calculation satisfies formula (10).

[0144]

[0145]

[0146]

[0147] In the formula, For sampling vehicles Actual passing time With sampled vehicles n C Actual passing time The green light duration between S3 ( nC,1→2) The number of vehicles arriving The corresponding time interval.

[0148] Collect the non-target coordinated merged traffic flow f within a certain period 4(1→2) For example, the trajectory time diagram is as follows Fig.12 As shown, the valid trajectory data in set A is shown in Table 3, and the trajectory data in set B is shown in Table 4.

[0149] Table 3 Valid trajectory data in set A

[0150]

[0151]

[0152] Table 4 Trajectory data in set B

[0153]

[0154] Step S2: construct the traffic flow f in each direction S(1→2) The arrival likelihood function estimates the traffic flow f in each direction. S(1→2) The arrival flow rate λ S(1→2) , the specific steps are as follows:

[0155] S201, constructing the traffic flow f in each direction S(1→2) The arrival likelihood function of .

[0156] According to the three sets obtained in step S103, the arrival likelihood function for the S-flow is constructed as shown in formula (13), where h is any adjacent sampled vehicle in set C. With n C The number of vehicles that may exist between the two, the maximum value is the number of vehicles arriving A total of a sampled vehicle trajectories are collected in set A, a total of b pairs of adjacent sampled vehicle trajectories are collected in set B, and a total of c pairs of adjacent sampled vehicle trajectories are collected in set C.

[0157]

[0158] S202: Estimating the traffic flow f in each direction S(1→2) The arrival flow rate λ S(1→2) Based on Bayesian inference, we take the exponential distribution as the arrival flow rate λ S(1→2) The prior distribution of exponential distribution parameter β S(1→2) is a known empirical value, and the arrival flow rate λ S(1→2) is a positive value, then the arrival flow rate λ S(1→2) The prior distribution function and posterior probability expressions of are shown in equations (14) and (15).

[0159]

[0160]

[0161] The maximum a posteriori estimation method is used to estimate the parameters, using the parameter variable value when the posterior probability is the maximum as the point estimate of the parameter. The denominator of the posterior probability, P(D), is a constant value and will not affect the maximum value point of the posterior probability function, so it can be ignored. The posterior probability function is logarithmized and simplified to obtain equation (16).

[0162]

[0163] for The estimated value of is calculated by equation (17), where l represents the number of loop iterations and g represents any adjacent sampled vehicle in set C. With n C The maximum number of vehicles that may exist between

[0164]

[0165] S flows to the traffic flow reaching flow rate λ S(1→2) The average value of The maximum a posteriori estimation method is combined with equation (17) to iterate until convergence, and the result satisfies equation (18).

[0166]

[0167] Furthermore, the converged average arrival flow rate By multiplying it by the vehicle arrival ratio, we can get the vehicle flow f in each direction S(1→2) The arrival flow rate λ at different time periods S(1→2) The expression satisfies formula (19).

[0168]

[0169] In this embodiment, the target coordinated update vehicle flow f is calculated. H(1→2) Arrival flow rate, non-target coordinated merging flow f 4(1→2) Arrival flow rate, non-target coordinated traffic flow f 3(1→2) Arrival flow rate Fig.13 , Fig.14 , Fig.15 shown.

[0170] Step S3: Calculate the maximum clearing time of all non-target coordinated traffic flows Specifically:

[0171] S301. Calculate the maximum clearing time of all non-target coordinated vehicle flows The clearing time of all non-target coordinated traffic is defined as t CT ( 1→2) , is the maximum clearing time of non-target coordinated traffic flow, and the calculation satisfies equation (20). S(1→2) Corresponding to the traffic flow f in each direction S(1→2) The time it takes for the queue of vehicles to be cleared, The clearing time t S(1→2) The maximum clearing time to obtain the maximum value, Q S(1→2) Corresponding to the traffic flow f in each direction S(1→2) The number of vehicles in the queue, is the number of queued vehicles QS(1→2) The maximum number of vehicles in the queue that reaches the maximum value, S (1→2) Coordinate traffic flow for the goal f O(1→2) Or target coordinated update of traffic flow f H(1→2) The Boolean variable F is the sum of the saturated flows of all lanes in the coordinated path. S(1→2) Used to determine whether there is traffic flow f in the traffic flow division result S(1→2) , there is traffic flow f S(1→2) Then F S(1→2) Take 1, otherwise take 0, where S can take O, H, 1, 2, 3, 4. In this embodiment, F O(1→2) =0, F H(1→2) =1, F 1(1→2) =0, F 2(1→2) =0, F 3(1→2) =1, F 4(1→2) =1.

[0172]

[0173]

[0174]

[0175]

[0176]

[0177] In the formula, is the average start-up loss time of the queued vehicles, such as Figure 6 As shown, T 1,2 With T 2,2 For non-target coordinated traffic flow f 1(1→2) With f 2(1→2) The first vehicle ideally passes through downstream intersection I 2 The ideal time to pass the stop line, T F,2 Coordinate traffic flow for the goal f O(1→2) Or target coordinated update of traffic flow f H(1→2) The first vehicle ideally passes through downstream intersection I 2 The ideal time to pass the stop line, T 4,2 For non-target coordinated merging of traffic flows f 4(1→2) The first vehicle ideally passes through downstream intersection I 2 The ideal time to pass the stop line, T O,2 With T E,2 Downstream intersection I 2 The start and end time points of the cycle in which the target coordination phase is located, T D,2 For downstream intersection I 2 The green light end time of the coordination phase.

[0178] According to this embodiment, after calculation

[0179] Step S4: Coordinate the green light duration and maximum clearing time of the phase according to the upstream and downstream intersections Determine the target coordinated vehicle at the downstream intersection I j The passing situation of the target coordinated vehicle flow f is calculated O(1→2) Or target coordinated update of traffic flow f H(1→2) The clearing time t C(1→2) , specifically:

[0180] S401, based on the green light duration of the upstream and downstream intersections and the maximum clearing time of all non-target coordinated traffic flows Determine the target coordinated traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) At downstream intersection I 2 The downstream cross

[0181] Oral I 2 Green light duration t of the coordinated phase G(2→3) Coordinate the maximum clearing time of traffic with all non-targets and upstream intersection I 1 Green light duration t of the coordinated phase G(1→2) If the sum is compared, Then the target coordinated traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) All vehicles in the queue are able to pass through the downstream intersection in the current cycle, and all vehicles in the queue are composed of non-target coordinated vehicles. Then the target coordinated traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) It is not possible to pass all downstream intersections in the current cycle, such as Figure 7 As shown, the queued vehicles consist of non-target coordinated traffic flow and target coordinated traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) composition.

[0182] After calculation, in this embodiment, the green light duration t of the upstream intersection coordination phase G(1→2) =32s, downstream intersection coordination phase green light duration t G(2→3) =45s, so Target coordinated update of vehicle flow f H(1→2) If all vehicles can pass through the downstream intersection in the current cycle and all queued vehicles are composed of non-target coordinated vehicles, go to step S402.

[0183] S402, calculate the non-existent target coordinated vehicle flow f O(1→2)Or target coordinated update of traffic flow f H(1→2) Clearing time t in case of parking waiting C(1→2) . Define t C(1→2) From upstream intersection I 1 To downstream intersection I 2 The goal of coordinating traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) The time required to clear the queued vehicles when the head-of-line green wave coordination design is adopted. If there is no target coordinated vehicle waiting when the vehicle arrives, then t C(1→2) =t CT(1→2) . Define a Boolean variable F AS(1→2) To determine the traffic flow f S(1→2) Whether the release phase of the target coordination phase is closely followed by the target coordination phase, if so, the Boolean variable F AS(i→j) Takes 1, otherwise takes 0, where S can take 1, 2, or 4. F AS(1→2) Satisfies formula (25).

[0184] F A1(1→2) +F A2(1→2) +F A4(1→2) ≤1 (25)

[0185] If F A1(1→2) +F A2(1→2) +F A4(1→2) =1, then there is a non-target coordinated traffic flow f 1(1→2) 、f 2(1→2) Or non-target coordinated merging of traffic flows f 4(1→2) The release phase follows the target coordination phase, and some non-target coordinated vehicles can pass through the downstream intersection I in the current cycle. 2 In this embodiment, F A1(i→j) =0, F A2(i→j) =0, F A4(i→j) =1, non-target coordinated merging of traffic flow f 4(1→2) The release phase follows the target coordination phase. The green light duration of the traffic flow that is allowed to pass through the downstream intersection in the current cycle is defined as t T(1→2) , satisfying equation (26). At this time, the clearing time t C(1→2) Satisfies formula (27).

[0186] t T(1→2) =t G(2→3) -t G(1→2) -t C(1→2) (26)

[0187]

[0188] In the formula, Q T(1→2) is the duration of the green light for the corresponding traffic flow T(1→2)The number of vehicles that can leave the downstream intersection within 100 km is calculated by equation (28). is the corresponding traffic flow f S(1→2) During the green light duration t T(1→2) The average arrival flow rate in the area, S can be 1, 2, 4, satisfying equation (29), and the green light duration t T(1→2) Satisfies formula (30).

[0189]

[0190]

[0191]

[0192] Where, T L,2 Coordinate traffic flow for the goal f O(1→2) Or target coordinated update of traffic flow f H(1→2) The last vehicle ideally passes downstream intersection I 2 The ideal passing time of the stop line, further, the clearing time t can be calculated C(1→2) Satisfying formula (31), such as Figure 8 shown.

[0193]

[0194] If F A1(1→2) +F A2(1→2) +F A4(1→2) = 0, then there is no non-target coordinated traffic flow f 1(1→2) 、f 2(1→2) Or non-target coordinated merging of traffic flows f 4(1→2) The release phase of is followed by the target coordination phase, such as Fig. 9 As shown, the target coordinated traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) The clearing time t C(1→2) Equal to the maximum clearing time of all non-target coordinated traffic flows

[0195] After calculation, the green light duration is t T(1→2) =5.5437s, clearing time t C(i→j) =7.4562s, i.e., the target coordinated update of vehicle flow f H(1→2) The clearing time when the team-head green wave coordinated design method is adopted is 7.4562 seconds, and the optimization process ends.

[0196] S403, calculate the target coordinated vehicle flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) Clearing time t in case of parking waiting C(1→2) .like Fig.10As shown, if there is a target coordinated traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) Stop and wait, then the goal is to coordinate the traffic flow f O(1→2) Or target coordinated update of traffic flow f H(1→2) The clearing time t C(1→2) Satisfy equation (32). Set the coordination phase green light end time T D,2 Until the ideal passing time of the last target coordinated vehicle T L,2 The ideal release time between W(1→2) Satisfies formula (33).

[0197]

[0198]

[0199]

[0200] In the formula, For traffic flow S(1→2) At the ideal release time t W(1→2) The average arrival flow rate within . Further, calculate the ideal release time t W(1→2) Satisfying formula (35), clearing time t C(1→2) When equation (36) is satisfied, the optimization process ends.

[0201]

[0202]

[0203] Since there is no target coordinated vehicle flow f in this embodiment O(1→2) Or target coordinated update of traffic flow f H(1→2) In the case of parking and waiting, there is no need to calculate this step.

[0204] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for calculating green wave coordinated clearing time based on sampled trajectory data, characterized in that: The following steps are involved: S1, for the upstream intersection I i To downstream intersection I j The coordination path P (i→j) , classification definition and its related flow direction traffic f S(i→j) , detect sampling vehicle n at downstream intersection I j Parking status S (n,j) and parking time T S(n,j) , calculate its actual passing time T P(n,j) and the ideal passing time T E(n,j) , the number of arriving vehicles is calculated by combining the sampled vehicle trajectories; S2. Construct the traffic flow f in each direction S(i→j) The arrival likelihood function estimates the traffic flow f in each direction. S(i→j) The arrival flow rate λ S(i→j) ; S3. Calculate the maximum clearing time of all non-target coordinated traffic flows S4. Coordinate the phase green light duration and maximum clearing time according to the upstream and downstream intersections Determine the target coordinated vehicle at the downstream intersection I j The passing situation of the target coordinated vehicle flow f is calculated O(i→j) Or target coordinated update of traffic flow f H(i→j) The clearing time t C(i→j) .

2. A method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 1, characterized in that: The step S1 is specifically as follows: S101, for the upstream intersection I i To downstream intersection I j The coordination path P (i→j) , classification definition and its related flow direction traffic f S(i→j) ; When the upstream intersection I i To downstream intersection I j Coordination Path P (i→j) The goal of coordinating traffic flow f O(i→j) When designing the green wave coordination of the first type of team, the target coordinated traffic flow f under different circumstances O(i→j) , non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) and f 3(i→j) The definition and selection of are shown in Table 1, where f (i,x→j,y) From the upstream intersection I i The x flow enters the downstream intersection I j The y-flow direction of the vehicle leaving, when x is 1, 2, and 3, it represents the vehicle flow from the upstream intersection I i Turn left, go straight, turn right and go to downstream intersection I j , y takes the values ​​of 1, 2, and 3, which respectively represent that the vehicle flow turns left, goes straight, and turns right at the current entrance to leave the downstream intersection I j , traffic flow (i,x→j,D) The D direction is the target coordinated vehicle flow f O(i→j) At downstream intersection I j The departure direction of the dedicated lane, the traffic flow (i,x→j,D′) The D′ flow direction is the target coordinated vehicle flow f O(i→j) At downstream intersection I j All departure directions of the shared lane, define a Boolean variable F R(i→j) To determine the upstream intersection I i Define the Boolean variable F to determine whether the right-turn merge flow is controlled. C(i→j) To determine the downstream intersection I j Coordinate whether the entrance lane of the route is a shared lane; Table 1 Definitions of target coordinated traffic flow and non-target coordinated traffic flow in different situations Target coordinated traffic flow O(i→j) The arrival flow rate is denoted as λ O(i→j) , non-target coordinated traffic flow f 1(i→j) and f 2(i→j) The arrival flow rate is denoted as λ 1(i→j) and λ 2(i→j) ; If the upstream intersection I i The right-turn merging traffic flow is not controlled by the signal, so it is defined as the non-target coordinated traffic flow f 3(i→j) The arrival flow rate λ 3(i→j) Perform calculations; According to the phase setting of the cycle signal, the traffic flow in the cycle is merged into the downstream intersection I j There are two types of situations: one is the upstream intersection I i The signal phase is set so that there is no mixing between vehicles entering the downstream intersection I j The second is the upstream intersection I i The signal phase setting allows mixed traffic to enter the downstream intersection I j phenomenon; When the upstream intersection I i The signal phase is set so that there is no mixing between vehicles entering the downstream intersection I j When the phenomenon occurs, estimate the target coordinated traffic flow f O(i→j) Coordinate traffic flow with non-targets 1(i→j) 、f 2(i→j) The vehicle arrival situation; when the upstream intersection I i The signal phase setting makes the traffic flow mixed and enter the downstream intersection I j If the traffic flow released in the same period is a non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) or 3(i→j) Two of them are merged and recorded as non-target coordinated merged traffic flow f 4(i→j) , whose arrival flow rate is λ 4(i→j) , if the released traffic flow in the same period has both target coordinated traffic flow f O(i→j) There are also other non-target coordinated traffic flows f 1(i→j) 、f 2(i→j) or 3(i→j) , and merge them into the target coordinated update traffic flow f H(i→j) , whose arrival flow rate is λ H(i→j) ; According to the above traffic flow division results, the coordinated path P (i→j) The relevant traffic flow in each direction is defined as f S(i→j) , S takes O, H, 1, 2, 3, and 4, representing the target coordinated traffic flow f O(i→j) , target coordinated update of traffic flow f H(i→j) , non-target coordinated traffic flow f 1(i→j) , non-target coordinated traffic flow f 2(i→j) , non-target coordinated traffic flow f 3(i→j) and non-target coordinated merging of traffic flows f 4(i→j) ; S102, detecting sampling vehicle n at downstream intersection I j The parking state S (n,j) and parking time T S(n,j) , calculate its actual passing time T P(n,j) and the ideal passing time T E(n,j) ; If the sampling vehicle n is at the downstream intersection I j There is a parking state, parking state S (n,j) Take 1, otherwise take 0; if the parking state S of the sampled vehicle n (n,j) is 0, based on the sampling vehicle n at intersection I j The detection time of the trajectory points before and after the stop line and the distance between them and the stop line are calculated by interpolation method. j The actual passing time T of the stop line at the entrance of the coordinated path P(n,j) ; If the parking state S of the sampled vehicle n (n,j) is 1, detecting sample vehicle n at intersection I j The first stop time T of the coordinated path entrance S(n,j) , calculate the ideal number of sampled vehicles n passing through intersection I j The ideal passing time T of the stop line at the entrance of the coordinated path E(n,j) , as shown in formula (1): Where D (n,j) is the sampling vehicle n at intersection I j The distance between the coordinated path entrance lane and the stop line when the first stop is made, is the average ideal driving speed of the vehicle on the current road section; S103, according to the flow of vehicles f S(i→j) Arrival situation Calculate the number of arriving vehicles; Discretize the continuous arrival time within the signal cycle with a time interval of k seconds. m The number of vehicles that ideally pass the intersection stop line during the time period is counted, and the vehicle arrival ratio P(t m ) is shown in formula (2): Where N is the total number of vehicles whose trajectories can be collected, m is the time interval number after discrete processing, and F{T E(n,j) =t m } is the ideal passing time T for determining the collection of any sampled vehicle n E(n,j) Is it in t m The indicator variable within a period is calculated as shown in formula (3): Traverse the sampled trajectory data of the S flow to the vehicle flow, and sequentially include the vehicle trajectory into the data set of different situations U. S takes O, H, 1, 2, 3, and 4, which represent the target coordinated vehicle flow f respectively. O(i→j) , target coordinated update of traffic flow f H(i→j) , non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) 、f 3(i→j) and non-target coordinated merging of traffic flows f 4(i→j) , case U is 1, 2, 3, case 1 corresponds to the first sampled vehicle n in all S-flow traffic within the signal period A The driving trajectory is the parking trajectory. Case 2 corresponds to the front vehicle of two adjacent sampled vehicles in the S flow within the signal period. With the following car B The driving trajectories are all parking trajectories. Case 3 corresponds to the front vehicle of two adjacent sampled vehicles in the S flow within the signal period. With the following car C The driving trajectories are parking trajectories and non-parking trajectories respectively; other sampling trajectory collection situations cannot provide effective information for traffic arrival flow rate estimation, so they are not considered; For the sampled trajectory data of the S-flow traffic, all trajectory data that satisfy condition 1 are recorded as set A, and any sampled vehicle n in set A is A The estimated number of arriving vehicles is recorded as All trajectory data that satisfy condition 2 are recorded as set B, and any adjacent sampled vehicles in set B are With n B The estimated number of arriving vehicles is recorded as All trajectory data that satisfy condition 3 are recorded as set C, and any adjacent sampled vehicles in set C are With n C The estimated number of arriving vehicles is recorded as 3. A method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 2, characterized in that: In step S103, for any sampled vehicle n in set A, A , estimate the coordination phase red light on time The ideal passing time with the current vehicle trajectory Downstream intersection I j The number of vehicles arriving on the coordinated path corresponding to the traffic flow As shown in formula (4) and formula (5): In the formula, [] is the rounding operator, To coordinate the phase green light on time To sample vehicle n A Actual passing time The length of the green light between and are the sampled vehicles n A Pass downstream intersection I j The time corresponding to the coordinated phase of the red and green lights, h s is the saturation headway, Poisson represents the number of arriving vehicles Follows Poisson distribution, λ S(i→j) is the arrival flow rate of traffic to S, The number of vehicles arriving The corresponding time interval.

4. A method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 2, characterized in that: In step S103, for any adjacent sampled vehicles in set B, With n B , estimating the sampled vehicles Ideal passing time With sampled vehicles n B Ideal passing time Number of vehicles arriving between The calculation satisfies formula (7): In the formula, For sampling vehicles Actual passing time With sampled vehicles n B Actual passing time The length of the green light between The number of vehicles arriving The corresponding time interval.

5. The method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 2 is characterized in that: In step S103, for any adjacent sampled vehicles in set C, With n C , estimating the sampled vehicles Ideal passing time With sampled vehicles n C Actual passing time Number of vehicles arriving between The calculation satisfies formula (10): In the formula, For sampling vehicles Actual passing time With sampled vehicles n C Actual passing time The length of the green light between The number of vehicles arriving The corresponding time interval.

6. The method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 1, characterized in that: The step S2 is specifically as follows: S201, constructing the traffic flow f in each direction S(i→j) The arrival likelihood function of ; According to the three sets obtained in step S103, the arrival likelihood function for the S-flow is constructed as shown in formula (13), where h is any adjacent sampled vehicle in set C. With n C The number of vehicles that may exist between the two, the maximum value is the number of vehicles arriving A total of a sampled vehicle trajectories are collected in set A, a total of b pairs of adjacent sampled vehicle trajectories are collected in set B, and a total of c pairs of adjacent sampled vehicle trajectories are collected in set C; S202: Estimating the traffic flow f in each direction S(i→j) The arrival flow rate λ S(i→j) ; Based on Bayesian inference, take exponential distribution as the arrival flow rate λ S(i→j) The prior distribution of exponential distribution parameter β S(i→j) is a known empirical value, and the arrival flow rate λ S(i→j) is a positive value, then the arrival flow rate λ S(i→j) The prior distribution function and posterior probability expressions of are shown in equations (14) and (15): The maximum a posteriori estimation method is used to estimate the parameters by using the parameter variable value when the posterior probability is the largest as the point estimate of the parameter. The denominator of the posterior probability, P(D), is a constant value and will not affect the maximum value point of the posterior probability function, so it is ignored. The posterior probability function is logarithmically processed and simplified to obtain formula (16): for The estimated value of is calculated by equation (17), where l represents the number of loop iterations and g represents any adjacent sampled vehicle in set C. With n C The maximum number of vehicles that may exist between S flows to the traffic flow reaching flow rate λ S(i→j) The average value of The maximum a posteriori estimation method is combined with formula (17) to iterate until convergence, and the result satisfies formula (18):

7. A method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 6, characterized in that: The step S202 also includes the following contents: The convergent average arrival flow rate Multiply it by the vehicle arrival ratio to obtain the vehicle flow f in each direction S(i→j) The arrival flow rate λ at different time periods S(i→j) The expression satisfies formula (19):

8. The method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 1, characterized in that: The step S3 is specifically as follows: S301. Calculate the maximum clearing time of all non-target coordinated vehicle flows The clearing time of all non-target coordinated traffic is defined as t CT(i→j) , is the maximum clearing time of all non-target coordinated traffic flows, calculated to satisfy equation (20); t S(i→j) Corresponding to the traffic flow f in each direction S(i→j) The time it takes for the queue of vehicles to be cleared, The clearing time t S(i→j) The maximum clearing time to obtain the maximum value, Q S(i→j) Corresponding to the traffic flow f in each direction S(i→j) The number of vehicles in the queue, is the number of queued vehicles Q S(i→j) The maximum number of vehicles in the queue that reaches the maximum value, S (i→j) Coordinate traffic flow for the goal f O(i→j) Or target coordinated update of traffic flow f H(i→j) The Boolean variable F is the sum of the saturated flows of all lanes in the coordinated path. S(i→j) Used to determine whether there is traffic flow f in the traffic flow division result S(i→j) , there is traffic flow f S(i→j) Then F S(i→j) Takes 1, otherwise takes 0, where S takes O, H, 1, 2, 3, 4; In the formula, is the average starting loss time of the queued vehicles, T 1,j With T 2,j For non-target coordinated traffic flow f 1(i→j) With f 2(i→j) The first vehicle ideally passes downstream intersection I j The ideal time to pass the stop line, T F,j Coordinate traffic flow for the goal f O(i→j) Or target coordinated update of traffic flow f H(i→j) The first vehicle ideally passes through downstream intersection I j The ideal time to pass the stop line, T 4,j For non-target coordinated merging of traffic flows f 4(i→j) The first vehicle ideally passes downstream intersection I j The ideal time to pass the stop line, T O,j With T E,j Downstream intersection I j The start and end time points of the cycle in which the target coordination phase is located, T D,j For downstream intersection I j The green light end time of the coordination phase.

9. The method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 1, characterized in that: The step S4 is specifically as follows: S401, based on the green light duration of the upstream and downstream intersections and the maximum clearing time of all non-target coordinated traffic flows Determine the target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) At downstream intersection I j The downstream intersection I j Green light duration t of the coordinated phase G(j→k) Coordinate the maximum clearing time of traffic with all non-targets and upstream intersection I i Green light duration t of the coordinated phase G(i→j) If the sum is compared, Then the target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) If all vehicles in the queue can pass through the downstream intersection in the current cycle and all vehicles in the queue are composed of non-target coordinated vehicles, go to step S402; Then the target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Unable to pass all downstream intersections in the current cycle, the queued vehicles are composed of non-target coordinated traffic flow and target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Composition, go to step S403; S402, calculate the non-existent target coordinated vehicle flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Clearing time t in case of parking waiting C(i→j) ; define t C(i→j) From upstream intersection I i To downstream intersection I j The goal of coordinating traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) The time required to clear the queued vehicles when the head-of-line green wave coordination design is adopted; If the vehicle arrives and there is no target coordinated vehicle to stop and wait, then t C(i→j) =t CT(i→j) ; Define the Boolean variable F AS(i→j) To determine the traffic flow f S(i→j) Whether the release phase of the target coordination phase is closely followed by the target coordination phase, if so, the Boolean variable F AS(i→j) Take 1, otherwise take 0, where S takes 1, 2, 4; F AS(i→j) Satisfying formula (25): F A1(i→j) +F A2(i→j) +F A4(i→j) ≤1 (25) If F A1(i→j) +F A2(i→j) +F A4(i→j) =1, then there is a non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) Or non-target coordinated merging of traffic flows f 4(i→j) The release phase follows the target coordination phase, and some non-target coordinated vehicles pass through the downstream intersection I in the current cycle. j ; The green light duration of the traffic flow that is allowed to follow the target coordinated phase to pass through the downstream intersection in the current cycle is defined as t T(i→j) , satisfying formula (26); at this time, the clearing time t C(i→j) Satisfies formula (27); t T(i→j) =t G(j→k) -t G(i→j) -t C(i→j) (26) In the formula, Q T(i→j) is the duration t of the green light for the corresponding traffic flow T(i→j) The number of vehicles that can leave the downstream intersection within 100 km is calculated by equation (28). is the corresponding traffic flow f S(i→j) During the green light duration t T(i→j) The average arrival flow rate in the area, S is taken as 1, 2, 4, satisfying formula (29), and the green light duration t T(i→j) Satisfies formula (30); Where, T L,j Coordinate traffic flow for the goal f O(i→j) Or target coordinated update of traffic flow f H(i→j) The last vehicle ideally passes downstream intersection I j Ideal passing time of the stop line, calculate the clearing time t C(i→j) Satisfies formula (31): If F A1(i→j) +F A2(i→j) +F A4(i→j) = 0, then there is no non-target coordinated traffic flow f 1(i→j) 、f 2(i→j) Or non-target coordinated merging of traffic flows f 4(i→j) The release phase is followed by the target coordination phase, and the target coordinated vehicle flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) The clearing time t C(i→j) Equal to the maximum clearing time of all non-target coordinated traffic flows End the optimization process; S403, calculate the target coordinated vehicle flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Clearing time t in case of parking waiting C(i→j) ; If there is a target coordinated traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) Stop and wait, then the goal is to coordinate the traffic flow f O(i→j) Or target coordinated update of traffic flow f H(i→j) The clearing time t C(i→j) Satisfy formula (32); Set the coordination phase green light end time T D,j Until the ideal passing time of the last target coordinated vehicle T L,j The ideal release time between W(i→j) Satisfies formula (33); In the formula, For traffic flow S(i→j) At the ideal release time t W(i→j) The average arrival flow rate within.

10. A method for calculating green wave coordinated clearing time based on sampled trajectory data according to claim 9, characterized in that: The step S403 also includes the following contents: Calculate the ideal release time t W(i→j) Satisfying formula (35), clearing time t C(i→j) When formula (36) is satisfied, the optimization process ends;

Images