A trunk green wave parking intersection recommendation method

Abstract

The application discloses a trunk green wave parking intersection recommendation method, which overcomes the problem that the existing technology cannot recommend parking intersections for a trunk line that cannot realize bidirectional green waves, and comprises the following steps: calculating the distance between two adjacent intersections on the trunk line to obtain an intersection pair recommendation sequence Z; calculating the average lane number difference between the coordinated direction and the non-coordinated direction of each intersection to obtain an average lane number intersection recommendation sequence A; calculating the average lane flow difference between the coordinated direction and the non-coordinated direction of each intersection to obtain a lane average flow intersection recommendation sequence B; calculating the flow similarity between intersections to obtain an intersection flow trend intersection recommendation sequence; obtaining a final output list from the obtained intersection recommendation sequence, and outputting corresponding recommendation reason labels. The trunk green wave parking intersection recommendation method can recommend a downlink direction parking intersection recommendation list and a recommendation reason for a trunk line that cannot realize bidirectional green waves, and improves the intelligent level of traffic control.

Description

Technical Field

[0001] This invention relates to the field of urban intelligent traffic signal control technology, and in particular to a method for recommending green wave parking intersections on main roads. Background Technology

[0002] With the annual increase in the number of motor vehicles and the expansion of urban road networks, optimizing signal schemes at single intersections is no longer sufficient to meet traffic control needs, leading more and more cities to implement green wave traffic systems. Ideally, an excellent arterial green wave algorithm can enable two-way traffic to pass through intersections sequentially without stopping, greatly improving road efficiency and minimizing delays. However, in practical applications, limitations imposed by road conditions, the number of intersections, intersection characteristics, and phase sequence settings often make it difficult to achieve two-way green waves, inevitably resulting in stopping at some intersections.

[0003] However, existing research on trunk line green wave technology mainly focuses on bidirectional coordination strategies and parameter optimization, with little mention of how to recommend parking point strategies when bidirectional green waves cannot be implemented. For example, a signal coordination control method for long trunk line intersections based on segmented green waves, published by the Chinese Patent Office on December 28, 2021 (publication number CN112509342B), proposes to interrupt the green wave band at appropriate locations and divide the long road segment into several sub-segments for green wave control. However, this method treats the sub-segment division method as a decision variable in the green wave bandwidth optimization model, rather than a solution that can be applied independently of the bandwidth optimization model. Furthermore, it does not distinguish between recommending green waves for the up and down directions separately, nor does it provide labels for the reasons for the recommendations. Summary of the Invention

[0004] The purpose of this invention is to overcome the problems in the existing technology that it cannot recommend parking intersections for trunk lines that cannot implement bidirectional green waves, resulting in a low level of intelligent traffic management. It provides a method for recommending parking intersections for trunk lines with green waves, which can recommend a list of parking intersections in both directions of the trunk line and the reasons for the recommendations, thereby improving the level of intelligent traffic management.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for recommending green wave parking intersections on main roads, comprising the following steps:

[0006] S1: Calculate the distance between two adjacent intersections on the trunk line, sort the intersection pairs according to the distance, and obtain the recommended sequence Z of intersection pairs;

[0007] S2: Calculate the difference in the average number of lanes between the coordinated and non-coordinated directions at each intersection, sort the intersections according to the difference, and obtain the recommended sequence A of intersections with average number of lanes.

[0008] S3: Calculate the difference in average traffic flow between coordinated and non-coordinated lanes at each intersection, sort the intersections according to the difference, and obtain the recommended sequence B of intersections with average lane traffic flow.

[0009] S4: Calculate the traffic flow similarity between intersections based on the average traffic flow of the up and down lanes at each intersection, sort the intersections according to similarity, and obtain the recommended sequence of intersection traffic flow trends.

[0010] S5: Add the intersections that were not included in the final output list from the intersection recommendation sequence obtained from S1-S4 to the final output list to obtain recommended intersections, and mark the recommended intersections with the corresponding recommendation reason labels.

[0011] In S1, intersection pairs with a spacing greater than L need to be selected; in S2, intersections with a difference greater than or equal to 0 need to be filtered; and in S3, intersections with a difference greater than 0 need to be filtered. The method in this application can recommend a list of stopping intersections in both directions for arterial roads that cannot implement bidirectional green waves, along with the reasons for the recommendations. This provides a theoretical basis for the re-division of control sub-zones and the optimization of coordination strategies, using data to drive management and control decisions and improve the intelligence level of traffic management.

[0012] Preferably, step S4 is expressed as follows:

[0013] S4.1: For coordinating the upward direction, obtain the time series W of the average traffic flow at each intersection's approach lanes. m,up The sequence was then normalized. For the normalized time series data, DTW distance was used to calculate pairwise similarity, resulting in the sum of the up-direction similarities at intersection i. The intersections are sorted according to the sum of their similarity scores to obtain a recommended sequence C of intersections with coordinated up-traffic flow trends. The recommendation reason label for the intersections in the sequence is: the coordinated up-traffic flow trend of this intersection is different from that of other intersections.

[0014] S4.2: Calculate the recommended intersection sequence D for coordinating the downstream traffic flow trend using the same method as in S4.1; the recommendation reason label for the intersection in the sequence is: the traffic flow trend of this intersection is different from that of other intersections.

[0015] By acquiring traffic flow data for each lane at each intersection throughout the day at time intervals of t, and averaging the traffic flow of different lanes at the same entrance at the same time interval for each intersection, a time series of the average traffic flow for each entrance lane is obtained. The two directions coordinated by the trunk line are distinguished as uphill and downhill. Uphill and downhill are only used to distinguish directions; which direction is uphill and which is downhill does not affect subsequent judgments.

[0016] Preferably, step S4.1 is further expressed as follows:

[0017] S4.1: Obtain the time series W of the average traffic flow at each intersection's approach lanes. m,up The maximum value w in max,up Using time series W m,up Divide each value in the dataset by the maximum value to obtain the normalized time series W. m ′ ,up ;

[0018] S4.1.2: Let the total number of intersections be m. For each intersection i, calculate the similarity with the other m-1 intersections, and obtain the m-1 similarity values ​​[r]. i,1 r i,2 , ..., r i,m-1 ], and obtain the sum of the similarity in the uphill direction of intersection i.

[0019] S4.1.3: Sort the intersections, placing those with the highest sum of similarity first, to obtain a list of sums of similarity [R]. 1,up R 2,up ... and the intersection recommended sequence coordinates the traffic flow trend C in the upward direction of the intersection.

[0020] Preferably, step S2 is further expressed as:

[0021] Obtain the number of lanes at each intersection entrance and the coordinated direction of the main road, and calculate the average number of entrance lanes n1 in the coordinated direction and the average number of entrance lanes n2 in the non-coordinated direction for each intersection.

[0022] Filter intersections where Δn = n2 - n1 is greater than or equal to 0, and sort the filtered intersections according to the value of Δn, placing the intersections with larger Δn values ​​at the top, to obtain a list of lane number differences [Δn1, Δn2, ...] and a recommended sequence A of intersections with average lane number;

[0023] The recommendation reason label for the intersection in sequence A is: the average number of lanes in the non-coordinated direction is greater than or equal to that in the coordinated direction.

[0024] The main road coordinates directions as east-west and north-south. Coordinated directions usually include two entrances, so n1 is the average number of lanes at the two entrances; uncoordinated directions may include two entrances or one entrance. If there is only one entrance, then n2 is the number of lanes at that entrance; if there are two entrances, then n2 is the average number of lanes at the two entrances.

[0025] Preferably, step S3 is further expressed as:

[0026] Obtain the total daily traffic flow for each lane at each intersection. For each intersection, calculate the average traffic flow q1 for lanes in the coordinating direction and the average traffic flow q2 for lanes in the non-coordinating direction.

[0027] Filter intersections where Δq = q2 - q1 is greater than 0, and sort the filtered intersections according to the value of Δq, placing the ones with larger Δq values ​​first, to obtain a difference list [Δq1, Δq2, ...] and a recommended sequence B of intersections with average lane flow.

[0028] The recommendation reason label for the intersection in sequence B is: the average flow rate of the lanes in the non-coordinated direction is greater than that in the coordinated direction.

[0029] Preferably, step S5 is further expressed as:

[0030] S5.1: Determine if the intersection recommendation sequence Z is empty. If it is empty, no processing is required; if it is not empty, add it to the final recommendation list E. up E down ;

[0031] S5.2: Determine if the average number of lanes per intersection recommendation sequence A is empty. If it is empty, no processing is required; if it is not empty, set the flag. i,up At the intersection where 0 = 0, add it to E. up And modify the flag i,up =1, set the flag i,down At the intersection where 0 = 0, add it to E. down And modify the flag i,down =1;

[0032] S5.3: Determine if the lane average traffic flow intersection recommendation sequence B is empty. If it is empty, no processing is required; if it is not empty, set the flag. i,up At the intersection where 0 = 0, add it to E. up And modify the flag i,up =1, set the flag i,down At the intersection where 0 = 0, add it to E. down And modify the flag i,down =1;

[0033] S5.4: Coordinate the traffic flow trend at the intersection in the upbound direction and the recommended sequence C in the flag. i,up Add the intersection with a value of 0 to E. up The flag in the intersection recommendation sequence D will coordinate the traffic flow trend at the intersection in the downlink direction. i,down At the intersection where 0 = 0, add it to E. down ;

[0034] S5.5: Output a recommended list of parking intersections in the uphill direction. up Recommended list of parking intersections in the downhill direction (E) down Two recommendation lists, along with corresponding intersection recommendation reason tags.

[0035] Initialize the flag for each intersection i,up=0, flag i,down =0, used to indicate whether the intersection has been included in the final output list. i,up =0 or flag i,down =0 indicates that it was not included in the final output list, flag i,up =1 or flag i,down =1 indicates that it has been included in the final output list.

[0036] Preferably, step S5.1 is further expressed as follows:

[0037] For intersection Z, select one intersection to add to the final recommendation list: Recommendation list E for parking intersections in the uphill direction. up Choose to travel in the upstream direction at the downstream intersection Z. 11 Recommended list E for parking intersections in the downhill direction down Choose to travel in the downstream direction at intersection Z. 12 For those who have already joined E up The crossroads flag i,up =1, for those already added to E down The crossroads flag i,down =1.

[0038] Since Z represents a pair of intersections, only one intersection can be selected when adding it to the final recommendation list.

[0039] Preferably, step S1 is further expressed as:

[0040] Obtain the distance between two adjacent intersections on the main road and treat two adjacent intersections as a single intersection pair;

[0041] Filter intersection pairs with an intersection spacing greater than L, sort them according to the spacing, and place the intersection pairs with larger spacing at the front to obtain the intersection pair recommendation sequence Z;

[0042] The reason for recommending intersections in sequence Z is: intersection spacing is too large.

[0043] L is a configuration parameter, typically 800 meters.

[0044] Therefore, the present invention has the following beneficial effects: it can recommend a list of parking intersections in both directions of the trunk line and the reasons for the recommendations for trunk lines that cannot achieve bidirectional green wave, providing a theoretical basis for the re-division of control sub-zones and the optimization of coordination strategies, using data to drive management and control decisions, and improving the level of intelligence in traffic management. Attached Figure Description

[0045] Figure 1 This is a flowchart of the steps of the method of the present invention. Detailed Implementation

[0046] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0047] like Figure 1 The illustrated embodiment shows a method for recommending green wave parking intersections on main roads. The operation flow is as follows: Step 1: Calculate the distance between adjacent intersections on the main road, sort the intersection pairs according to the distance, and obtain the intersection pair recommendation sequence Z; Step 2: Calculate the difference in the average number of lanes between the coordinated and non-coordinated directions at each intersection, sort the intersections according to the difference, and obtain the average number of lanes intersection recommendation sequence A; Step 3: Calculate the difference in the average traffic flow between the coordinated and non-coordinated directions at each intersection, sort the intersections according to the difference, and obtain the average traffic flow intersection recommendation sequence B; Step 4: Calculate the traffic flow similarity between intersections based on the average traffic flow of the up and down lanes at each intersection, sort the intersections according to the similarity, and obtain the intersection traffic flow trend intersection recommendation sequence; Step 5: Obtain recommended intersections from the intersection recommendation sequences obtained in Steps 1 to 4 that are not included in the final output list, and label the recommended intersections with corresponding recommendation reason tags.

[0048] The technical solution of this application will be further illustrated below with specific examples.

[0049] Step 1: Calculate the distance between two adjacent intersections on the main road, select intersection pairs with a distance greater than L, sort the intersection pairs according to the distance, and obtain the recommended sequence of intersection pairs Z = [Z1, Z2, ...].

[0050] Based on the distance between two adjacent intersections on the main road, obtain the intersection pair recommendation sequence Z = [Z1, Z2, ...], and label the corresponding recommendation reason tags:

[0051] Obtain the distance between two adjacent intersections on the main road, and treat the two adjacent intersections as a pair; filter the intersection pairs with a distance greater than L, and sort them according to the distance, placing the intersection pairs with larger distances at the front, to obtain the recommended sequence of intersection pairs Z = [Z1, Z2, ...].

[0052] The recommendation reason label for intersections in the recommended sequence Z is: the intersection spacing is too large.

[0053] Where L is a configuration parameter, and in this embodiment, L = 800m by default.

[0054] Specifically:

[0055] In this embodiment, the main road has 8 intersections, each marked with an intersection ID. The distance between two intersections is shown in the table below:

[0056]

[0057]

[0058] Based on the criteria, the recommended sequence of intersection pairs is obtained: Z = [300044-300045].

[0059] Step 2: Calculate the difference between the average number of lanes in the coordinated direction and the non-coordinated direction at each intersection, filter the intersections with a difference greater than or equal to 0, sort the intersections according to the difference, and obtain the recommended sequence A of intersections with the average number of lanes.

[0060] Based on the number of lanes at each intersection entrance and the coordinated direction of the main road, obtain the average number of lanes recommended sequence A = [A1, A2, ...], and label the corresponding recommendation reasons:

[0061] Obtain the number of lanes at each intersection entrance and the coordinated direction of the trunk line; calculate the average number of entrance lanes n1 in the coordinated direction and the average number of entrance lanes n2 in the non-coordinated direction for each intersection; calculate Δn = n2 - n1, filter intersections with Δn greater than or equal to 0, and sort the filtered intersections according to the value of Δn, placing the intersections with larger Δn at the front, to obtain a list of lane number difference [Δn1, Δn2, ...] and a recommended sequence of intersections with average lane number A = [A1, A2, ...].

[0062] The intersection recommendation reason label in the average number of lanes per intersection recommendation sequence A is: the average number of lanes in the non-coordinated direction is greater than or equal to that in the coordinated direction.

[0063] The main road coordination directions are east-west and north-south. Coordinated directions typically include two entrances, so n1 is the average number of lanes at the two entrances. Uncoordinated directions may include two entrances or one entrance. If there is only one entrance, then n2 is the number of lanes at that entrance; if there are two entrances, then n2 is the average number of lanes at the two entrances.

[0064] In this embodiment, the calculated lane number difference for each intersection is listed in the table below:

[0065]

[0066] The recommended sequence of intersections with average number of lanes is obtained as A = [300048].

[0067] Step 3: Calculate the difference in average traffic flow between the coordinated and non-coordinated directions at each intersection, remove intersections with a difference greater than 0, sort the intersections by the difference, and obtain the recommended sequence B of intersections with average lane traffic flow.

[0068] Based on the total daily traffic flow of lanes in both coordinated and non-coordinated directions at each intersection, obtain the recommended intersection sequence B = [B1, B2, ...] for average lane traffic flow, and label the corresponding recommendation reasons:

[0069] Obtain the total daily traffic flow for each lane at each intersection. For each intersection, calculate the average traffic flow q1 for lanes in the coordinating direction and the average traffic flow q2 for lanes in the non-coordinating direction. Calculate Δq = q2 - q1, filter intersections with Δq greater than 0, and sort the filtered intersections according to the value of Δq, placing those with larger Δq values ​​at the top, to obtain the difference list [Δq1, Δq2, ...] and the recommended sequence of intersections with average lane traffic flow B = [B1, B2, ...].

[0070] The intersection recommendation reason label in the intersection recommendation sequence B with average lane flow is: the average lane flow in the non-coordinated direction is greater than that in the coordinated direction.

[0071] The table below shows the differences in the average lane flow rates calculated for each intersection:

[0072]

[0073] The recommended intersection sequence for average lane flow is obtained as B = [300049, 300048, 300050].

[0074] Step 4: Calculate the traffic flow similarity between intersections based on the average traffic flow of the up and down lanes at each intersection, sort the intersections according to the similarity, and obtain the recommended sequence of intersection traffic flow trends.

[0075] Based on the traffic flow data of each lane at each intersection throughout the day at time intervals of t, obtain the intersection traffic flow trend recommendation sequence and label it with the corresponding recommendation reason tags:

[0076] In this embodiment, the traffic flow data of each lane at each intersection is obtained at 5-minute intervals throughout the day. The average traffic flow of different lanes at the same entrance at each intersection at the same time interval is taken to obtain the time series of the average traffic flow of each entrance lane.

[0077] In this embodiment, the two directions of trunk line coordination are distinguished as the up direction and the down direction. The terms "up" and "down" are only used to distinguish the directions, and which direction is up and which is down does not affect the subsequent judgment.

[0078] (1) Coordinate the traffic flow trend at the intersection in the uphill direction and recommend the intersection sequence C.

[0079] For coordinating the upward direction, the time series W of the average inlet lane flow rate at each intersection is taken. m,up =[w 1,up w 2,up w 3,up ,……]; Obtain the time series W m,up The maximum value w in max,up Using time series W m,up Each value in the range is divided by the maximum value w. max,upObtain the normalized time series W m ′ ,up =[w1 ′ ,up w2 ′ ,up w3 ′ ,up ...].

[0080] For the normalized time-series data of the average northbound lane flow at each intersection, DTW distance was used to calculate pairwise similarity. The calculation method is as follows: assuming a total of m intersections, for each intersection i, the similarity is calculated with the other m-1 intersections, obtaining m-1 similarity values ​​[r]. i,1 r i,2 , ..., r i,m-1 ].

[0081] Let the sum of the similarities in the uphill direction of intersection i be . Sort the intersections according to their total similarity score, placing those with higher total similarity scores first, to obtain a list of total similarity scores [R]. 1,up R 2,up ... and the recommended sequence of intersection traffic flow trend in the upbound direction C = [C1, C2, ...].

[0082] The intersection recommendation reason label in the intersection recommendation sequence C, which coordinates the traffic flow trend of the intersection in the upward direction, is: the traffic flow trend of this intersection in the coordinated direction is different from that of other intersections.

[0083] The table below shows the total similarity of the northbound directions at each intersection:

[0084]

[0085]

[0086] The recommended intersection sequence for coordinating the traffic flow trend in the upward direction is obtained as C = [300050, 300051, 300049, 300046, 300048, 300044, 300045, 300047].

[0087] (2) Coordinate the traffic flow trend at intersections in the downhill direction and recommend the intersection sequence D.

[0088] For the coordinated downlink direction, the same method as for the uplink direction is used to calculate and obtain a list of summed similarities [R]. 1,down R 2,down ... and the recommended sequence of intersection traffic flow trend in the downhill direction, D = [D1, D2, ...].

[0089] The intersection recommendation reason label in the down-direction intersection traffic flow trend recommendation sequence D is: the traffic flow trend of this intersection is different from that of other intersections.

[0090] The table below shows the total similarity of the down-direction intersections:

[0091] Intersection Number Downlink Similarity Sum 300044 11.56 300045 8.64 300046 9.01 300047 9.78 300048 9.05 300049 10.33 300050 12.56 300051 9.69

[0092] The recommended intersection sequence for coordinating the downstream traffic flow trend is obtained as D = [300050, 300044, 300049, 300047, 300051, 300048, 300046, 300045].

[0093] Step 5: Extract recommended intersections from the intersection recommendation sequences obtained in Steps 1 to 4 that were not included in the final output list, and label the recommended intersections with the corresponding recommendation reason tags.

[0094] The final conclusion is output by combining the calculation results from steps one through four: Let E be the final recommended list of parking intersections in the uphill direction. up The recommended list for parking intersections in the downhill direction is E. down Initialize the flag for each intersection. i,up =0, flag i,down =0, used to indicate whether the intersection has been included in the final output list.

[0095] (1) Intersection pair recommendation sequence Z.

[0096] Determine if the recommended sequence Z = [Z1, Z2, ...] for each intersection is empty. If it is empty, no further processing is needed; otherwise, add it to the final recommendation list E. up E down .

[0097] Since Z represents a pair of intersections, only one intersection can be selected when adding it to the final recommendation list. The selection method is to check the upstream recommendation list E. up Choose to travel in the upstream direction at the downstream intersection Z. 11 For the downstream recommendation list E down Choose to travel in the downstream direction at intersection Z. 12 For those who have already joined E up The crossroads flag i,up =1, for those already added to E down The crossroads flag i,down =1.

[0098] That is, based on the recommended sequence Z = [300044-300045] for the intersection pair, E is obtained. up =[300045], E down=[300044].

[0099] (2) Recommended sequence A for intersections with average number of lanes.

[0100] Check if the recommended sequence A = [A1, A2, ...] for intersections with an average number of lanes is empty. If it is empty, no processing is needed; otherwise, check the flag corresponding to each intersection in turn. i,up and flag i,down The value of the flag i,up At the intersection where 0 = 0, add it to E. up E up =[Z 11 Z 12 [A1, A2, ...], and modify the flag. i,up =1; Set flag i,down At the intersection where 0 = 0, add it to E. down E down =[Z 12 Z 22 ...A1, A2, ...], and modify the flag. i,down =1.

[0101] That is, based on (1), according to the recommended sequence A = [300048] of the intersection with the average number of lanes, E is obtained. up =[300045, 300048], E down =[300044, 300048].

[0102] (3) Recommended intersection sequence B for average lane flow.

[0103] Determine if the intersection recommendation sequence B = [B1, B2, ...] based on average lane flow is empty. If it is empty, no processing is needed; otherwise, check the corresponding flags for each intersection. i,up and flag i,down The value of the flag i,up At the intersection where 0 = 0, add it to E. up E up =[Z 11 Z 12 ..., A1, A2, ..., B1, B2, ...], and modify the flag. i,up =1; Set flag i,down At the intersection where 0 = 0, add it to E. down E down =[Z 12 Z 22 ...A1, A2, ..., B1, B2, ...], and modify the flag. i,down =1.

[0104] That is, based on (2), according to the intersection recommended sequence B = [300049, 300048, 300050] based on the average lane flow rate, E is obtained. up =[300045, 300048, 300049, 300050], E down =[300044, 300048, 300049, 300050,].

[0105] (4) Intersection traffic flow trend and recommended sequence of intersections.

[0106] The flag in the recommended sequence C = [C1, C2, ...] will coordinate the traffic flow trend at the intersection in the upbound direction. i,up Add the intersection with a value of 0 to E. up E up ==[Z 11 Z 12 ..., A1, A2, ..., B1, B2, ..., C1, C2, ...]; will coordinate the downstream intersection traffic flow trend intersection recommendation sequence D = [D1, D2, ...] flags i,down At the intersection where 0 = 0, add it to E. down E down =[Z 12 Z 22 ,...A1,A2,...,B1,B2,...,D1,D2,...].

[0107] Based on (3), according to the recommended intersection sequence C = [300050, 300051, 300049, 300046, 300048, 300044, 300045, 300047] and the recommended intersection sequence D = [300050, 300044, 300049, 300047, 300051, 300048, 300046, 300045], E is obtained. up =[300045, 300048, 300049, 300050, 300051, 300046, 300044, 300047], E down =[300044, 300048, 300049, 300050, 300047, 300051, 300046, 300045], which is the final output list of recommended intersections.

[0108] (5) Final output of the recommended list of intersections.

[0109] E obtained in (4) up and E down This is the final output list of recommended intersections.

[0110] The final output is a list of recommended parking intersections in the uphill direction, E. up The corresponding intersection recommendation reason tags are shown in the table below:

[0111] Intersection Number Recommendation Reason Label 300045 Intersection Distance Too Large 300048 Average Lane Count of Non-Coordinated Direction is Greater Than or Equal to Coordinated Direction 300049 Average Lane Flow of Non-Coordinated Direction is Greater Than Coordinated Direction 300050 Average Lane Flow of Non-Coordinated Direction is Greater Than Coordinated Direction 300051 Flow Trend of Coordinated Direction at This Intersection is Different Than Other Intersections 300046 Flow Trend of Coordinated Direction at This Intersection is Different Than Other Intersections 300044 Flow Trend of Coordinated Direction at This Intersection is Different Than Other Intersections 300047 Flow Trend of Coordinated Direction at This Intersection is Different Than Other Intersections

[0112] The final output is a list of recommended parking intersections in the downhill direction, E. down The corresponding intersection recommendation reason tags are shown in the table below:

[0113]

[0114]

[0115] The method in this embodiment can generate a list of recommended parking intersections in both directions of the trunk line that cannot achieve bidirectional green wave, along with the reasons for the recommendations. This provides a theoretical basis for the re-division of control sub-zones and the optimization of coordination strategies, using data to drive management and control decisions and improve the level of intelligence in traffic management.

[0116] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Other variations and modifications are possible without departing from the technical solutions described in the claims.

Claims

1. A trunk green wave parking intersection recommendation method, characterized in that, The method comprises the following steps: S1: calculate the distance between two adjacent intersections on the trunk, sort the intersection pairs according to the distance, and obtain an intersection pair recommendation sequence Z; S2: calculate the average lane number difference between the coordinated direction and the non-coordinated direction of each intersection, sort the intersections according to the difference, and obtain an average lane number intersection recommendation sequence A; S3: calculate the lane average flow difference between the coordinated direction and the non-coordinated direction of each intersection, sort the intersections according to the difference, and obtain a lane average flow intersection recommendation sequence B; S4: calculate the flow similarity between intersections according to the lane average flow of the upstream and downstream directions of each intersection, sort the intersections according to the similarity, and obtain an intersection flow trend intersection recommendation sequence; S5: add the intersections in the intersection recommendation sequence obtained in S1-S4 that are not in the final output list to the final output list to obtain recommended intersections, and mark the recommended intersections with corresponding recommended reason labels; For the intersection pair in the intersection pair recommendation sequence Z, select one of the intersections to join the final recommendation list: for the upstream recommendation list, select the intersection that travels in the downstream direction in the upstream direction, and for the downstream recommendation list, select the intersection that travels in the downstream direction in the downstream direction.

2. The method of claim 1, wherein, The step S4 is further represented as: S4.1: For coordinated uplink, obtain the time series W of the average value of each intersection entrance lane flow m,up And normalize the sequence, for the normalized time series data, use DTW distance to calculate the similarity between each other, and get the uplink similarity sum of intersection i as Sort the intersections according to the size of the similarity sum, and get the coordinated uplink intersection flow trend intersection recommendation sequence C; The intersection recommendation reason label in the sequence is: the coordinated direction flow trend of this intersection is different from that of other intersections; S4.2: calculate the coordinated downstream direction intersection flow trend intersection recommendation sequence D using the same method as in S4.1; the recommended reason label in the sequence is that the coordinated direction flow trend of the intersection is different from that of other intersections.

3. The method of claim 2, wherein, The step S2 is further represented as: S4.1: Obtain the time series W of average values of the flow of each approach lane of the intersection m,up The maximum value w in W max,up Divide each value in the time series W m，up by the maximum value w max，up to obtain the normalized time series W' m，up ; S4.1.2: Set the total number of intersections as m, for each intersection i, respectively calculate the similarity with other m-1 intersections, obtain m-1 similarity values [r i，1 , r i，2 , …, r i，m-1 ], and get the uplink direction similarity sum of intersection i S4.1.3: Rank the intersections, put the one with larger similarity sum in front, get the similarity sum list [R 1,up , R 2,up , …] and the intersection recommendation sequence coordination uplink direction intersection flow trend C.

4. The method of claim 1, wherein, Obtain the lane number of each entrance of the intersection and the coordinated direction of the trunk, calculate the average number of entrance lanes n1 in the coordinated direction and the average number of entrance lanes n2 in the non-coordinated direction of each intersection; Screen the intersections with Δn=n2-n1 greater than or equal to 0, sort the screened intersections according to the value of Δn, and place the intersections with larger Δn in front to obtain the lane number difference list [Δn1, Δn2, …] and the average lane number intersection recommendation sequence A; The recommended reason label in the sequence A is that the average lane number in the non-coordinated direction is greater than or equal to that in the coordinated direction. The step S3 is further represented as:

5. The method of claim 1 or 2 or 4, wherein, Obtain the total flow of each lane of each intersection throughout the day, for each intersection, calculate the average value q1 of the lane flow in the coordinated direction and the average value q2 of the lane flow in the non-coordinated direction; Screen the intersections with Δq=q2-q1 greater than 0, sort the screened intersections according to the value of Δq, and place the intersections with larger Δq in front to obtain the difference list [Δq1, Δq2, …] and the lane average flow intersection recommendation sequence B; the recommended reason label in the sequence B is that the lane average flow in the non-coordinated direction is greater than that in the coordinated direction. The step S5 is further represented as:

6. The method of claim 1 or 2 or 4, wherein, The step S1 is further represented as: S5.1: judge whether the intersection recommendation sequence Z is empty, if empty, no need to process; if not empty, add to the final recommendation list E up , E down ; S5.2: judge whether the recommended sequence A of average lane number is empty, if empty, no need to process; if not empty, add the intersection with flag i,up = 0 to E up , and modify flag i,up = 1; add the intersection with flag i,down = 0 to E down , and modify flag i,down = 1; S5.3: judge whether the recommended sequence B of the average flow of the lane is empty, if empty, no need to process; if not empty, add the intersection with flag i,up = 0 to E up , and modify flag i,up = 1, add the intersection with flag i，down = 0 to E down , and modify flag i，down = 1; S5.4: Add to E the intersections with flag = 0 in the intersection recommended sequence C coordinating the uplink direction intersection traffic trend i，up S5.5: Add to E the intersections with flag = 0 in the intersection recommended sequence D coordinating the downlink direction intersection traffic trend up S5.6: Add to E the intersections with flag = 0 in the intersection recommended sequence E coordinating the intersection traffic trend i，down S5.7: Add to E the intersections with flag = 0 in the intersection recommended sequence F coordinating the intersection traffic trend down ; S5.5: Output the uplink intersection recommendation list E up , the downlink intersection recommendation list E down two recommendation lists, and the corresponding intersection recommendation reason labels.

7. The method of claim 6, wherein, In step S5.1, it is further expressed as: for those already added E up The crossroads flag i，up =1, for those already added to E down The crossroads flag i，down =1.

8. The method of claim 1 or 2 or 4, wherein, Obtain the distance between two adjacent intersections on the trunk, and take the two adjacent intersections as an intersection pair; Screen the intersection pairs with a distance greater than L, sort them according to the distance, and place the intersection pairs with a larger distance in front to obtain the intersection pair recommendation sequence Z; The recommended reason label in the sequence Z is that the distance between the intersections is too large. ​

Images