CN115273467B - Method for identifying traffic flow interaction event based on multi-region detection data - Google Patents

Method for identifying traffic flow interaction event based on multi-region detection data Download PDF

Info

Publication number
CN115273467B
CN115273467B CN202210830495.5A CN202210830495A CN115273467B CN 115273467 B CN115273467 B CN 115273467B CN 202210830495 A CN202210830495 A CN 202210830495A CN 115273467 B CN115273467 B CN 115273467B
Authority
CN
China
Prior art keywords
lane
event
traffic flow
straight
calculation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202210830495.5A
Other languages
Chinese (zh)
Other versions
CN115273467A (en
Inventor
郝建根
王月
张继锋
江超阳
张俊
沈阳
韩雪
伍丹
王军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing LES Information Technology Co. Ltd
Original Assignee
Nanjing LES Information Technology Co. Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing LES Information Technology Co. Ltd filed Critical Nanjing LES Information Technology Co. Ltd
Priority to CN202210830495.5A priority Critical patent/CN115273467B/en
Publication of CN115273467A publication Critical patent/CN115273467A/en
Application granted granted Critical
Publication of CN115273467B publication Critical patent/CN115273467B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0133—Traffic data processing for classifying traffic situation
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
    • G08G1/0145—Measuring and analyzing of parameters relative to traffic conditions for specific applications for active traffic flow control
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/052—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Abstract

The invention discloses a traffic flow interaction event identification method based on multi-region detection data, which comprises the following steps: dividing multiple detection areas; calibrating lane attributes of a second detection area; detecting data acquisition analysis and calculation window triggering; traffic flow interacts with event recognition. The method is based on multi-region detection, a plurality of detection regions are arranged at different positions of the widened intersection, the lane-level passing number, the number of vehicles in the lane and the change characteristics of traffic demands are further researched through the collection and analysis of multi-region detection data, the traffic flow distribution rules under different traffic states are obtained through space-time correlation analysis, an identification method of traffic flow interaction events in any direction of the widened intersection is established, and effective information is provided for the establishment of a signal control fine microscopic strategy.

Description

Method for identifying traffic flow interaction event based on multi-region detection data
Technical Field
The invention belongs to the technical field of intelligent control of traffic signals, and particularly relates to a traffic flow interaction event identification method based on multi-region detection data.
Background
In the prior art, most of research on judging methods of whether traffic jam events occur is focused on, and research technology for identifying interaction events among different traffic flows in the same direction in the subdivision direction of the traffic jam events is not found. For example: the invention discloses a congestion identification method and a device based on calculation of travel time of vehicle running data, and establishes a congestion road section identification method through threshold judgment. The invention of China patent application No. CN202110850488.7 is a traffic jam state rapid identification prediction method based on electric warning data, which discloses prediction of average travel speed based on an LSTM model, and further judges the jam state by utilizing the ratio of the average travel speed to free flow speed; the technical scheme is a method for judging whether the traffic jam event occurs or not.
In the research direction of multi-region detection data application, the Chinese patent application number is CN201810458053.6, which is a vehicle counting method and system based on multiple detection regions, wherein the states of a first detection region and a second detection region of a target lane in a current frame are acquired by utilizing image frames in a target traffic video, and the vehicle passing counting is performed by analyzing the occupied idle states of the first detection region and the second detection region; the invention has the patent application number of CN201610039673.7, which is a detector layout method based on urban road congestion identification, and reasonably sets two or three groups of detectors based on the congestion characteristics of roads and the effect timeliness of traffic control, so as to accurately obtain the congestion state of the roads and provide effective information for the management and decision of the roads.
According to the method, the existing congestion state recognition technology focuses on the traffic congestion state recognition of road sections or intersections, specific congestion types cannot be determined, and congestion events caused by interaction among traffic flows in the same direction are not beneficial to traffic managers to more pertinently adopt corresponding signal control strategies; although the prior detection technology can realize the acquisition of second-level multi-region detection data, the congestion state is still analyzed by methods of calculating the vehicle passing count, the vehicle passing speed and the like of a transverse section in data application, and the analysis and mining capacity of imported second-level multi-region detection data, especially longitudinal regions, is weaker, and the statistical time granularity is large, the space refinement degree is low, and the event reflecting capacity is weak.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a traffic flow mutual influence event identification method based on multi-region detection data, so as to solve the problems of delay in the discovery time of a congestion event and uncertainty for the congestion type caused by determining whether the congestion occurs or not by a method which is mainly dependent on indexes such as the traffic count, the traffic speed of a transverse section and the like in the prior art; the method is based on multi-region detection, a plurality of detection regions are arranged at different positions of the widened intersection, the lane-level passing number, the number of vehicles in the lane and the change characteristics of traffic demands are further researched through the collection and analysis of multi-region detection data, the traffic flow distribution rules under different traffic states are obtained through space-time correlation analysis, an identification method of traffic flow interaction events in any direction of the widened intersection is established, and effective information is provided for the establishment of a signal control fine microscopic strategy.
In order to achieve the above purpose, the invention adopts the following technical scheme:
the invention discloses a traffic flow mutual influence event identification method based on multi-region detection data, which comprises the following steps:
1) Dividing multiple detection areas: according to the canalization characteristics of the intersection entrance way, a first detection area and a second detection area are respectively arranged for the canal section and the arrival road section, a first detection section is arranged in the first detection area, and a second detection section is arranged in the second detection area;
2) Calibrating lane attribute of a second detection area: determining each lane attribute of the second detection area according to the extension lines of lane boundaries of different lane attributes of the first detection area;
3) Detecting data acquisition analysis and calculation window triggering: each detection area collects transverse detection data and longitudinal detection data, and the number of vehicles passing through the lane, the number of vehicles in the lane and the traffic demand of lane attribute level are calculated once per second; determining whether a traffic flow interaction event calculation window is opened or not according to the lane attribute level passing number change characteristics of the first detection area;
4) Traffic flow interaction event identification: determining judging conditions of traffic flow mutual influence events based on the lane attribute classification condition of the second detection area, and obtaining recognition results of the traffic flow mutual influence events through association calculation of traffic demand change characteristics of the lane attributes of each detection area.
Further, in the step 1), according to the canalization characteristics of the entrance way of the intersection, detection sections and detection areas are arranged at different positions of the entrance way, and the method comprises the following steps:
11 First detection area setting: the first detection area is positioned in the channeling section and comprises a gradual change section, and is used for detecting the number of vehicles in the lane in the first detection area; setting a first detection section at the parking line of the entrance road, and detecting lane-level passing data of the first detection section;
12 Second detection area setting: the second detection area is positioned on the incoming road section and is used for detecting the number of vehicles in a lane of the incoming road section; and a second detection section is arranged at the lower edge of the second detection area and used for detecting lane-level passing data of the second detection section.
Further, the first detection area lane attribute in the step 2) includes, but is not limited to: straight lanes, left turn lanes; the lane attribute of the second detection area includes: straight entrance lane, easy-to-interfere lane, left turn entrance lane.
Further, the lane attribute calibration method of the second detection area in the step 2) specifically includes: marking the lane positioned in the second detection area and positioned on the left straight lane dividing line extension line of the first detection area as an easy-to-disturb lane; the inner lane of the easy-to-disturb lane is marked as a left turn entrance lane, and the outer lane of the easy-to-disturb lane is marked as a straight entrance lane.
In addition, the method for analyzing the event of influence of the right and the left traffic is similar to the method for analyzing the event of the left and the right traffic, and therefore, the description thereof is omitted.
Further, the detection data collection and analysis in the step 3) specifically includes:
each detection area collects transverse detection data and longitudinal detection data and records the transverse detection data and the longitudinal detection data in a passing table PassTable and a queuing table queue table; the PassTable field includes, but is not limited to: transit time, vehicle type, lane number; the QueueTable field includes, but is not limited to: detecting time, the number of vehicles in a lane and the number of tracks; and calculating the number of lane-level passing vehicles, the number of vehicles in the lane and the lane attribute-level traffic demand based on the detection data per second.
Further, the time t is calculated p Lane-level passing number of (a)
Figure BDA0003748089690000031
The calculation formula of (2) is as follows:
Figure BDA0003748089690000032
wherein i is a lane attribute number, i is { t, l, te, le, be }, t, l, te, le, be is a straight lane, a left turn lane, a straight entrance lane, a left turn entrance lane and an easy-to-interfere lane respectively; n is a lane number belonging to the same lane attribute, and when the same lane attribute comprises a plurality of lanes, the lane numbers {1,2, … …, N } of the lanes are marked in sequence from the inner lane to the outer lane;
Figure BDA0003748089690000033
to search for a pass time in the PassTable that corresponds to the lane attribute i, the lane number n, and the pass time is at t p -1,t p ) And calculates the number of the records of the passing of the vehicle.
When the number of passing vehicles of all lanes belonging to the same lane attribute in the first detection area is larger than 0 and the duration time of keeping unchanged exceeds a set reset time threshold, resetting the number of passing vehicles of all lanes under the lane attribute; and when the number of passing vehicles of each lane attribute of the first detection area is 0 and the duration exceeds the set reset time threshold, resetting the number of passing vehicles of all lanes of the second detection area.
Further, the number of vehicles in the lane is collected every second
Figure BDA0003748089690000034
And records, and further calculates traffic demand +.>
Figure BDA0003748089690000035
The calculation method comprises the following steps:
Figure BDA0003748089690000036
wherein, the lane attribute is the number of passing vehicles
Figure BDA0003748089690000037
And the number of vehicles in the lane of the lane attribute level +.>
Figure BDA0003748089690000038
Respectively correspond to the calculation time t p The number of passing vehicles and the number of vehicles in the lane with the lane attribute of i are calculated as follows:
Figure BDA0003748089690000039
Figure BDA00037480896900000310
further, the triggering of the calculation window in the step 3) specifically includes:
determining whether a traffic flow interaction event calculation window is opened or not according to the lane attribute-level passing number change characteristics of the first detection area, and recording that the current calculation window type and the calculation window type to be opened are Cur and Tem respectively, wherein Cur and Tem epsilon { LBTWin, TBLWin, noWin }, LBTWin, TBLWin, noWin are left-turn influence straight-going calculation windows, straight-going influence left-turn calculation windows and closed traffic flow influence calculation windows respectively; stop marks whether a traffic flow influencing event is recognized under the current calculation window, stop=1 indicates that the traffic flow influencing event is recognized, and stop=0 indicates that the traffic flow influencing event is not recognized; when the lane attribute of the first detection area is the number of passing vehicles
Figure BDA0003748089690000041
And->
Figure BDA0003748089690000042
When the calculation window affecting the identification of the straight-going event by left turning is started, the tem=lbtwin; the lane attribute of the first detection area is graded by the number of passing vehicles +.>
Figure BDA0003748089690000043
And->
Figure BDA0003748089690000044
When the left turn event identification calculation window is opened, a calculation window of straight going influencing left turn event identification is opened, and tem=TBLWIN; in other cases, vehicle influence event recognition calculation is not needed, and tem=nowin;
comparing Tem with Cur, if Tem is not equal to Cur and Tem is not equal to NoWin, indicating that a new window type is to be opened, updating the current window type Cur=tem, and recording the starting time t of the calculation window 0 ＝t p ，t p Step 4) is executed for the current calculation time; if tem=cur, tem++nowin and stop=0, indicating that the current calculation window is maintained and no traffic flow influencing event is recognized yet, executing step 4); if tem=cur and stop=1, the traffic flow influence event is identified in the current calculation window, calculation is not needed to be continued, and the detection data collection and analysis step is returned; if tem=noon, it means that the calculation window will be closed, the current window type cur=tem is updated, stop=0 is updated, and the detection data collection and analysis step is returned.
Further, the identifying result in the step 4) includes: left turn affects straight-going events, straight-going affects left-turn events, and no traffic affecting events.
Further, the step 4) specifically includes:
41 Traffic flow interaction event judgment condition setting: determining judging conditions of traffic flow interaction events based on lane attribute distribution conditions of the second detection area; if the straight entrance lane and the left entrance lane exist, judging conditions of the straight event and the straight event are JC1; if the straight entrance lane exists and the left-turn entrance lane does not exist, the judging condition of the left-turn influence straight event is JC1, and the judging condition of the straight influence left-turn event is JC2; if the straight entrance lane does not exist and the left entrance lane exists, the judging condition of the left turning influence straight event is JC2, and the judging condition of the straight turning influence left turning event is JC1; if the straight entrance lane and the left turn entrance lane do not exist, judging conditions of the straight event and the straight event are JC2;
the judgment condition JC1 has the following calculation formula:
Figure BDA0003748089690000045
the EventMark is a parameter for marking the event type of the traffic influencing event, and EventMark epsilon { LBT, TBL, none }, LBT, TBL, none are left-turn influencing straight-going events, straight-going influencing left-turn events and traffic influencing event-free events respectively; t (T) miniGap RisingP is the minimum time interval threshold f1 (t p -T minGap ,t p ) The characterization parameter f1 is in the time segment t p -T miniGap ,t p ]The rising index of f 1 ＝Trend(PassD ie -PassD be ) Ie e { te, le }; trend (f 2) is a parameter which is converted into a parameter which can reflect the change characteristic of f2 by eliminating the random fluctuation of the parameter f2 by a moving average method 2 ＝PassD ie -PassD be ；
The time segment [ t ] p -T miniGap ,t p ]Rise index of internal parameter f1
Figure BDA0003748089690000051
The calculation formula of (2) is as follows:
Figure BDA0003748089690000052
wherein, the liquid crystal display device comprises a liquid crystal display device,
Figure BDA0003748089690000053
the calculation method for the mark parameter f whether the mth second rises compared with the mth-1 second is as follows:
Figure BDA0003748089690000054
wherein, f1 (m) =trend (f 2 (m)), the calculation method of Trend (f 2 (m)) is as follows:
Figure BDA0003748089690000055
the judgment condition JC2 is calculated as follows:
Figure BDA0003748089690000056
wherein Q is min A threshold value for the number of vehicles in the average lane of the time period;
42 Traffic flow interaction event identification): calculating traffic demand TrafficD of each lane attribute in the window according to the traffic flow interaction event judgment conditions and combining the traffic flow interaction event i Number of vehicles in lane QueueVeh interfering with lane be Judging whether a traffic flow influence event occurs or not; if a traffic flow influencing event occurs, namely eventmark=lbt or eventmark=tbl, updating stop=1, and returning to the step 3); if no traffic flow influencing event occurs, i.e. eventmark=none, then directly return to step 3).
The invention has the beneficial effects that:
1) The invention makes up for the technical defect of identifying subdivision directions of congestion events which are mutually influenced by left-right traffic flows in the congestion identification direction in the intelligent traffic field, and provides effective information for signal control refinement strategy formulation;
2) The multi-region data acquisition and analysis method can grasp traffic flow collection and distribution rules of different positions of the intersection inlet more finely;
3) Based on the distribution rule of traffic flows in different traffic states, the invention carries out association analysis on the real-time change characteristics of the traffic flows of different types at different positions of the inlet, establishes a mutual influence identification method of the traffic flows, and improves the timeliness and accuracy of identification.
Drawings
FIG. 1 is a schematic diagram of a traffic situation in which a left turn affects straight traffic;
FIG. 2 is a schematic diagram of a traffic situation in which straight-through affects left-turn traffic;
FIG. 3 is a schematic view of an intersection entrance with a widening section;
FIG. 4 is a schematic diagram of a recognition process of a left turn affecting straight-going event;
FIG. 5 is a schematic diagram of a process for identifying a straight-ahead left turn event;
fig. 6 is a functional block diagram of the method of the present invention.
Detailed Description
The invention will be further described with reference to examples and drawings, to which reference is made, but which are not intended to limit the scope of the invention.
Referring to fig. 6, in the method for identifying traffic flow interaction events based on multi-region detection data according to the present invention, fig. 3 is a schematic view of an intersection entrance with a widening section in an example, and the steps are as follows:
1) Dividing multiple detection areas: according to the canalization characteristics of the intersection entrance way, a first detection area and a second detection area are respectively arranged for the canal section and the arrival road section, a first detection section is arranged in the first detection area, and a second detection section is arranged in the second detection area;
the step 1) of arranging detection sections and detection areas at different positions of the entrance road according to the canalization characteristics of the entrance road of the road junction comprises the following steps:
11 First detection area setting: the first detection area is positioned in the channeling section and comprises a gradual change section, and is used for detecting the number of vehicles in the lane in the first detection area; setting a first detection section at the parking line of the entrance road, and detecting lane-level passing data of the first detection section;
12 Second detection area setting: the second detection area is positioned on the incoming road section and is used for detecting the number of vehicles in a lane of the road section; and a second detection section is arranged at the lower edge of the second detection area and used for detecting lane-level passing data of the second detection section.
2) Calibrating lane attribute of a second detection area: determining each lane attribute of the second detection area according to the extension lines of lane boundaries of different lane attributes of the first detection area;
wherein, the first detection area lane attribute in the step 2) includes, but is not limited to: straight lanes, left turn lanes; the lane attribute of the second detection area includes: straight entrance lane, easy-to-interfere lane, left turn entrance lane.
The lane attribute calibration method of the second detection area specifically comprises the following steps: marking the lane positioned in the second detection area and positioned on the left straight lane dividing line extension line of the first detection area as an easy-to-disturb lane; the inner lane of the easy-to-disturb lane is marked as a left turn entrance lane, and the outer lane of the easy-to-disturb lane is marked as a straight entrance lane.
In addition, the method for analyzing the event of influence of the right and the left traffic is similar to the method for analyzing the event of the left and the right traffic, and therefore, the description thereof is omitted.
3) Detecting data acquisition analysis and calculation window triggering: each detection area collects transverse detection data and longitudinal detection data, and the number of vehicles passing through the lane, the number of vehicles in the lane and the traffic demand of lane attribute level are calculated once per second; determining whether a traffic flow interaction event calculation window is opened or not according to the lane attribute level passing number change characteristics of the first detection area;
the detection data acquisition analysis specifically comprises:
each detection area collects transverse detection data and longitudinal detection data and records the transverse detection data and the longitudinal detection data in a passing table PassTable and a queuing table queue table; the PassTable field includes, but is not limited to: transit time, vehicle type, lane number; the QueueTable field includes, but is not limited to: detecting time, the number of vehicles in a lane and the number of tracks; and calculating the number of lane-level passing vehicles, the number of vehicles in the lane and the lane attribute-level traffic demand based on the detection data per second.
Specifically, the time t is calculated p Lane-level passing number of (a)
Figure BDA0003748089690000071
The calculation formula of (2) is as follows:
Figure BDA0003748089690000072
wherein i is a lane attribute number, i is { t, l, te, le, be }, t, l, te, le, be is a straight lane, a left turn lane, a straight entrance lane, a left turn entrance lane and an easy-to-interfere lane respectively; n is a lane number belonging to the same lane attribute, and when the same lane attribute comprises a plurality of lanes, the lane numbers {1,2, … …, N } of the lanes are marked in sequence from the inner lane to the outer lane;
Figure BDA0003748089690000073
to search for a pass time in the PassTable that corresponds to the lane attribute i, the lane number n, and the pass time is at t p -1,t p ) And calculates the number of the records of the passing of the vehicle.
When the number of passing vehicles of all lanes belonging to the same lane attribute in the first detection area is larger than 0 and the duration time of keeping unchanged exceeds a set reset time threshold, resetting the number of passing vehicles of all lanes under the lane attribute; and when the number of passing vehicles of each lane attribute of the first detection area is 0 and the duration exceeds the set reset time threshold, resetting the number of passing vehicles of all lanes of the second detection area.
Collecting number of vehicles in lane per second
Figure BDA0003748089690000074
And records, and further calculates the attribute-level traffic demand of each lane
Figure BDA0003748089690000075
The calculation method comprises the following steps:
Figure BDA0003748089690000076
wherein, the lane attribute is the number of passing vehicles
Figure BDA0003748089690000077
And the number of vehicles in the lane of the lane attribute level +.>
Figure BDA0003748089690000078
Respectively correspond to the calculation time t p The number of passing vehicles and the number of vehicles in the lane with the lane attribute of i are calculated as follows:
Figure BDA0003748089690000079
Figure BDA0003748089690000081
specifically, the computing window trigger specifically includes:
determining whether a traffic flow interaction event calculation window is opened or not according to the lane attribute-level passing number change characteristics of the first detection area, and recording that the current calculation window type and the calculation window type to be opened are Cur and Tem respectively, wherein Cur and Tem epsilon { LBTWin, TBLWin, noWin }, LBTWin, TBLWin, noWin are left-turn influence straight-going calculation windows, straight-going influence left-turn calculation windows and closed traffic flow influence calculation windows respectively; stop marks whether a traffic flow influencing event is recognized under the current calculation window, stop=1 indicates that the traffic flow influencing event is recognized, and stop=0 indicates that the traffic flow influencing event is not recognized; when the lane attribute of the first detection area is the number of passing vehicles
Figure BDA0003748089690000082
And->
Figure BDA0003748089690000083
When the calculation window affecting the identification of the straight-going event by left turning is started, the tem=lbtwin; the lane attribute of the first detection area is graded by the number of passing vehicles +.>
Figure BDA0003748089690000084
And->
Figure BDA0003748089690000085
When the left turn is influenced by opening straight movementA calculation window of event recognition, tem=tblwin; in other cases, vehicle influence event recognition calculation is not needed, and tem=nowin;
comparing Tem with Cur, if Tem is not equal to Cur and Tem is not equal to NoWin, indicating that a new window type is to be opened, updating the current window type Cur=tem, and recording the starting time t of the calculation window 0 ＝t p ，t p Step 4) is executed for the current calculation time; if tem=cur, tem++nowin and stop=0, indicating that the current calculation window is maintained and no traffic flow influencing event is recognized yet, executing step 4); if tem=cur and stop=1, the traffic flow influence event is identified in the current calculation window, calculation is not needed to be continued, and the detection data collection and analysis step is returned; if tem=noon, it means that the calculation window will be closed, the current window type cur=tem is updated, stop=0 is updated, and the detection data collection and analysis step is returned.
As shown in FIG. 4, the attribute level passing number change characteristics of each lane in the first detection area are analyzed, and the passage veh is performed at the time of 9:00:02 l =0 and PassVeh t >0, thus opening a calculation window of left turn affecting straight-going event recognition at the moment 9:00:02, cur=lbtwin, and recording the start moment t of the calculation window 0 =9:00:02, step 4) is performed; maintaining at LBTWin, cur=lbtwin and stop=0, steps 4) from 9:00:03 to 9:00:15; 9:00:16 to 9:00:21 are maintained at LBTWin, cur=lbtwin and stop=1, and the detection data acquisition analysis step is returned; and closing a calculation window of which left turn affects straight movement, wherein Cur=NoWin, and returning to a detection data acquisition and analysis step.
As shown in FIG. 5, the attribute level passing number change characteristics of each lane in the first detection area are analyzed, and the passage veh is performed at the time of 18:00:23 t =0 and PassVeh l >0, thus opening a calculation window that straight-going affects left turn event recognition at 18:00:23, cur=tblwin, and recording the start time t of the calculation window 0 =18:00:23, step 4) is performed; 18:00:24 to 18:00:38 are maintained at TBLWin, cur=tblwin and stop=0, step 4) is performed; 18:00:39 to 18:00:43 are maintained at TBLWin, cur=tblwin and stop=1, and the detection data collection and analysis step is returned; closing straight line affects left turn 18:00:44And calculating a window, cur=NoWin, and returning to the detection data acquisition and analysis step.
4) Traffic flow interaction event identification: determining judging conditions of traffic flow interaction events based on the lane attribute classification conditions of the second detection areas, and obtaining recognition results of the traffic flow interaction events through association calculation of traffic demand change characteristics of the lane attributes of each detection area; referring to fig. 1, fig. 2;
the identification result comprises: left turn affects straight-going events, straight-going affects left-turn events, and no traffic affecting events.
Specifically, 41) traffic flow interaction event judgment condition setting: determining judging conditions of traffic flow interaction events based on lane attribute distribution conditions of the second detection area; if the straight entrance lane and the left entrance lane exist, judging conditions of the straight event and the straight event are JC1; if the straight entrance lane exists and the left-turn entrance lane does not exist, the judging condition of the left-turn influence straight event is JC1, and the judging condition of the straight influence left-turn event is JC2; if the straight entrance lane does not exist and the left entrance lane exists, the judging condition of the left turning influence straight event is JC2, and the judging condition of the straight turning influence left turning event is JC1; if the straight entrance lane and the left turn entrance lane do not exist, judging conditions of the straight event and the straight event are JC2;
the judgment condition JC1 has the following calculation formula:
Figure BDA0003748089690000091
the EventMark is a parameter for marking the event type of the traffic influencing event, and EventMark epsilon { LBT, TBL, none }, LBT, TBL, none are left-turn influencing straight-going events, straight-going influencing left-turn events and traffic influencing event-free events respectively; t (T) miniGap RisingP is the minimum time interval threshold f1 (t p -T minGap ,t p ) The characterization parameter f1 is in the time segment t p -T miniGap ,t p ]The rising index of f 1 ＝Trend(PassD ie -PassD be ) Ie e { te, le }; trend (f 2) is a parameter which is converted into a parameter which can reflect the change characteristic of f2 by eliminating the random fluctuation of the parameter f2 by a moving average method 2 ＝PassD ie -PassD be ；
The time segment [ t ] p -T miniGap ,t p ]Rise index of internal parameter f1
Figure BDA0003748089690000092
The calculation formula of (2) is as follows:
Figure BDA0003748089690000093
wherein, the liquid crystal display device comprises a liquid crystal display device,
Figure BDA0003748089690000094
the calculation method for the mark parameter f whether the mth second rises compared with the mth-1 second is as follows:
Figure BDA0003748089690000095
wherein, f1 (m) =trend (f 2 (m)), the calculation method of Trend (f 2 (m)) is as follows:
Figure BDA0003748089690000101
the judgment condition JC2 is calculated as follows:
Figure BDA0003748089690000102
wherein Q is min A threshold value for the number of vehicles in the average lane of the time period;
42 Traffic flow interaction event identification): calculating traffic demand TrafficD of each lane attribute in the window according to the traffic flow interaction event judgment conditions and combining the traffic flow interaction event i In-lane interfering with a laneVehicle number QueueVeh be Judging whether a traffic flow influence event occurs or not; if a traffic flow influencing event occurs, namely eventmark=lbt or eventmark=tbl, updating stop=1, and returning to the step 3); if no traffic flow influencing event occurs, i.e. eventmark=none, then directly return to step 3).
In time period [9:00:02,9:00:14]Binding PassD per second i Obtaining EventMark=none according to the formula (5), and returning to the step (3); calculating the time t p At 9:00:15, according to equation (5), the result eventmark=lbt is output, stop=1 is updated, and step 3 is returned.
In time period [18:00:23,18:00:37]EventMark=none output according to formula (9), return to step 3); calculating the time t p At 18:00:38, according to equation (9), the result eventmark=lbt is output, stop=1 is updated, and the process returns to step 3).
The present invention has been described in terms of the preferred embodiments thereof, and it should be understood by those skilled in the art that various modifications can be made without departing from the principles of the invention, and such modifications should also be considered as being within the scope of the invention.

Claims (7)

1. The method for identifying the traffic flow interaction event based on the multi-region detection data is characterized by comprising the following steps:
1) Dividing multiple detection areas: according to the canalization characteristics of the intersection entrance way, a first detection area and a second detection area are respectively arranged for the canal section and the arrival road section, a first detection section is arranged in the first detection area, and a second detection section is arranged in the second detection area;
2) Calibrating lane attribute of a second detection area: determining each lane attribute of the second detection area according to the extension lines of lane boundaries of different lane attributes of the first detection area;
3) Detecting data acquisition analysis and calculation window triggering: each detection area collects transverse detection data and longitudinal detection data, and the number of vehicles passing through the lane, the number of vehicles in the lane and the traffic demand of lane attribute level are calculated once per second; determining whether a traffic flow interaction event calculation window is opened or not according to the lane attribute level passing number change characteristics of the first detection area;
4) Traffic flow interaction event identification: determining judging conditions of traffic flow interaction events based on the lane attribute classification conditions of the second detection areas, and obtaining recognition results of the traffic flow interaction events through association calculation of traffic demand change characteristics of the lane attributes of each detection area;
the identification result in the step 4) comprises the following steps: left turn affects straight-going events, straight-going affects left-going events and no traffic flow affects events;
the step 4) specifically comprises the following steps:
41 Traffic flow interaction event judgment condition setting: determining judging conditions of traffic flow interaction events based on lane attribute distribution conditions of the second detection area; if the straight entrance lane and the left entrance lane exist, judging conditions of the straight event and the straight event are JC1; if the straight entrance lane exists and the left-turn entrance lane does not exist, the judging condition of the left-turn influence straight event is JC1, and the judging condition of the straight influence left-turn event is JC2; if the straight entrance lane does not exist and the left entrance lane exists, the judging condition of the left turning influence straight event is JC2, and the judging condition of the straight turning influence left turning event is JC1; if the straight entrance lane and the left turn entrance lane do not exist, judging conditions of the straight event and the straight event are JC2;
the judgment condition JC1 has the following calculation formula:
Figure FDA0004250375710000011
the EventMark is a parameter for marking the event type of the traffic influencing event, and EventMark epsilon { LBT, TBL, none }, LBT, TBL, none are left-turn influencing straight-going events, straight-going influencing left-turn events and traffic influencing event-free events respectively; t (T) miniGap RisingP is the minimum time interval threshold f1 (t p -T minGap ，t p ) The characterization parameter f1 is in the time segment t p -T miniGap ，t p ]The rising index of f 1 ＝Trend(PassD ie -PassD be ) Ie e { te, le }; trend (f 2) is a parameter which is converted into a parameter which can reflect the change characteristic of f2 by eliminating the random fluctuation of the parameter f2 by a moving average method 2 ＝PassD ie -PassD be ；
The time segment [ t ] p -T miniGap ，t p ]Rise index of internal parameter f1
Figure FDA0004250375710000021
The calculation formula of (2) is as follows:
Figure FDA0004250375710000022
wherein, the liquid crystal display device comprises a liquid crystal display device,
Figure FDA0004250375710000023
the calculation method for the mark parameter f whether the mth second rises compared with the mth-1 second is as follows:
Figure FDA0004250375710000024
wherein, f1 (m) =trend (f 2 (m)), the calculation method of Trend (f 2 (m)) is as follows:
Figure FDA0004250375710000025
the judgment condition JC2 is calculated as follows:
Figure FDA0004250375710000026
wherein Q is min A threshold value for the number of vehicles in the average lane of the time period;
42 Traffic flow interaction event identification): according to the vehicle flow phaseTraffic demand TrafficD of each lane attribute in window is calculated by combining traffic flow interaction event judgment conditions and traffic flow interaction event judgment conditions i Number of vehicles in lane QueueVeh interfering with lane be Judging whether a traffic flow influence event occurs or not; if a traffic flow influencing event occurs, namely eventmark=lbt or eventmark=tbl, updating stop=1, and returning to the step 3); if no traffic flow influencing event occurs, i.e. eventmark=none, then directly return to step 3).
2. The method for identifying traffic interaction events based on multi-zone detection data according to claim 1, wherein the first detection zone lane attribute in step 2) includes, but is not limited to: straight lanes, left turn lanes; the lane attribute of the second detection area includes: straight entrance lane, easy-to-interfere lane, left turn entrance lane.
3. The method for identifying traffic flow interaction event based on multi-region detection data according to claim 2, wherein the lane attribute calibration method for the second detection region in step 2) specifically comprises: marking the lane positioned in the second detection area and positioned on the left straight lane dividing line extension line of the first detection area as an easy-to-disturb lane; the inner lane of the easy-to-disturb lane is marked as a left turn entrance lane, and the outer lane of the easy-to-disturb lane is marked as a straight entrance lane.
4. The method for identifying traffic flow interaction events based on multi-zone detection data according to claim 1, wherein the detection data collection analysis in step 3) specifically comprises:
each detection area collects transverse detection data and longitudinal detection data and records the transverse detection data and the longitudinal detection data in a passing table PassTable and a queuing table queue table; the PassTable field includes, but is not limited to: transit time, vehicle type, lane number; the QueueTable field includes, but is not limited to: detecting time, the number of vehicles in a lane and the number of tracks; and calculating the number of lane-level passing vehicles, the number of vehicles in the lane and the lane attribute-level traffic demand based on the detection data per second.
5. The method for identifying traffic flow interaction event based on multi-zone detection data according to claim 4, wherein the time t is calculated p Lane-level passing number of (a)
Figure FDA0004250375710000031
The calculation formula of (2) is as follows:
Figure FDA0004250375710000032
wherein i is a lane attribute number, i is { t, l, te, le, be }, t, l, te, le, be is a straight lane, a left turn lane, a straight entrance lane, a left turn entrance lane and an easy-to-interfere lane respectively; n is a lane number belonging to the same lane attribute, and when the same lane attribute comprises a plurality of lanes, the lane numbers {1,2, & gt, N };
Figure FDA0004250375710000033
to search for a pass time in the PassTable that corresponds to the lane attribute i, the lane number n, and the pass time is at t p -1，t p ) And calculates the number of the records of the passing of the vehicle.
6. The method for identifying traffic flow interaction event based on multi-zone detection data according to claim 4, wherein the number of vehicles in a lane per second is collected
Figure FDA0004250375710000034
And records, and further calculates traffic demand of each lane attribute
Figure FDA0004250375710000035
The calculation method comprises the following steps:
Figure FDA0004250375710000036
wherein, the lane attribute is the number of passing vehicles
Figure FDA0004250375710000037
And the number of vehicles in the lane of the lane attribute level +.>
Figure FDA0004250375710000038
Respectively correspond to the calculation time t p The number of passing vehicles and the number of vehicles in the lane with the lane attribute of i are calculated as follows:
Figure FDA0004250375710000039
Figure FDA00042503757100000310
7. the method for identifying traffic flow interaction event based on multi-zone detection data according to claim 4, wherein the calculating window triggering in step 3) specifically comprises:
determining whether a traffic flow interaction event calculation window is opened or not according to the lane attribute-level passing number change characteristics of the first detection area, and recording that the current calculation window type and the calculation window type to be opened are Cur and Tem respectively, wherein Cur and Tem epsilon { LBTWin, TBLWin, noWin }, LBTWin, TBLWin, noWin are left-turn influence straight-going calculation windows, straight-going influence left-turn calculation windows and closed traffic flow influence calculation windows respectively; stop marks whether a traffic flow influencing event is recognized under the current calculation window, stop=1 indicates that the traffic flow influencing event is recognized, and stop=0 indicates that the traffic flow influencing event is not recognized; when the lane attribute of the first detection area is the number of passing vehicles
Figure FDA0004250375710000041
And->
Figure FDA0004250375710000042
When the calculation window affecting the identification of the straight-going event by left turning is started, the tem=lbtwin; the lane attribute of the first detection area is graded by the number of passing vehicles +.>
Figure FDA0004250375710000043
And->
Figure FDA0004250375710000044
When the left turn event identification calculation window is opened, a calculation window of straight going influencing left turn event identification is opened, and tem=TBLWIN; in other cases, vehicle influence event recognition calculation is not needed, and tem=nowin;
comparing Tem with Cur, if Tem is not equal to Cur and Tem is not equal to NoWin, indicating that a new window type is to be opened, updating the current window type Cur=tem, and recording the starting time t of the calculation window 0 ＝t p ，t p Step 4) is executed for the current calculation time; if tem=cur, tem++nowin and stop=0, indicating that the current calculation window is maintained and no traffic flow influencing event is recognized yet, executing step 4); if tem=cur and stop=1, the traffic flow influence event is identified in the current calculation window, calculation is not needed to be continued, and the detection data collection and analysis step is returned; if tem=noon, it means that the calculation window will be closed, the current window type cur=tem is updated, stop=0 is updated, and the detection data collection and analysis step is returned.
CN202210830495.5A 2022-07-15 2022-07-15 Method for identifying traffic flow interaction event based on multi-region detection data Active CN115273467B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210830495.5A CN115273467B (en) 2022-07-15 2022-07-15 Method for identifying traffic flow interaction event based on multi-region detection data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210830495.5A CN115273467B (en) 2022-07-15 2022-07-15 Method for identifying traffic flow interaction event based on multi-region detection data

Publications (2)

Publication Number Publication Date
CN115273467A CN115273467A (en) 2022-11-01
CN115273467B true CN115273467B (en) 2023-07-04

Family

ID=83764486

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210830495.5A Active CN115273467B (en) 2022-07-15 2022-07-15 Method for identifying traffic flow interaction event based on multi-region detection data

Country Status (1)

Country Link
CN (1) CN115273467B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116403419B (en) * 2023-06-07 2023-08-25 贵州鹰驾交通科技有限公司 Traffic light control method based on vehicle-road cooperation

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103093623B (en) * 2013-01-09 2014-10-29 东南大学 Prediction method of urban road signalized intersection direct-left conflict number
GB2510206A (en) * 2013-04-23 2014-07-30 Richard Charles David Andrew Traffic lights which prevent conflict between left-turning motor vehicles and bicycles
CN105390004B (en) * 2015-10-15 2017-08-01 大连理工大学 A kind of paired intersection time-space distribution distribution method of the short track influence of left-hand rotation
CN105405301B (en) * 2015-11-23 2017-02-22 山东理工大学 Right-turn signal induction control method for eliminating straight-right-turn convergence conflict
KR101722168B1 (en) * 2016-08-09 2017-04-11 한국정보기술 주식회사 System and method for guding detour of vehicle waiting for turn left using cctv
CN206097478U (en) * 2016-10-18 2017-04-12 长安大学 Non -motor signal controlling system of turning left based on infrared induction
CN106504527B (en) * 2016-10-19 2018-12-28 东南大学 A kind of signalized intersections directly left conflict and its impact analysis method
CN107016851B (en) * 2017-05-24 2019-06-28 大连理工大学 The method that a kind of quantitative analysis city built environment influences road journey time
CN107798868B (en) * 2017-10-17 2019-02-01 大连理工大学 A kind of signalized intersections u-turn opening bit selecting method that traffic flow composition influences
CN113428081B (en) * 2020-03-19 2023-08-29 荷兰移动驱动器公司 Traffic safety control method, vehicle-mounted device and readable storage medium
CN111445699B (en) * 2020-04-13 2021-10-26 黑龙江工程学院 Intersection traffic conflict discrimination method based on real-time vehicle track
CN111968370B (en) * 2020-08-18 2022-02-22 南京慧尔视智能科技有限公司 Intelligent variable lane sensing system and method for microwave radar
CN113963554A (en) * 2021-12-02 2022-01-21 吉林大学 Method for calculating left-turn lane traffic capacity under left-turn permission phase
CN114677838B (en) * 2022-05-27 2022-08-09 南京莱斯信息技术股份有限公司 Lapping phase starting discrimination method based on full time

Also Published As

Publication number Publication date
CN115273467A (en) 2022-11-01

Similar Documents

Publication Publication Date Title
CN111508094B (en) Highway congestion finding method based on ETC portal frame and gate traffic data
CN106355903B (en) Multilane traffic volume detection method based on video analysis
CN105913661B (en) A kind of express highway section traffic state judging method based on charge data
CN110717433A (en) Deep learning-based traffic violation analysis method and device
CN115273467B (en) Method for identifying traffic flow interaction event based on multi-region detection data
CN109191830A (en) A kind of congestion in road detection method based on video image processing
CN113870564B (en) Traffic jam classification method and system for closed road section, electronic device and storage medium
CN109754598B (en) Congestion grouping identification method and system
US20220383738A1 (en) Method for short-term traffic risk prediction of road sections using roadside observation data
CN108848460A (en) People's vehicle correlating method based on RFID and GPS data
CN111243338A (en) Vehicle acceleration-based collision risk evaluation method
CN109087534A (en) A kind of traffic conflict detection method based on vehicle driving trace
CN113380036B (en) Queuing length calculation method based on electronic police data
CN115691120A (en) Congestion identification method and system based on highway running water data
CN116153078A (en) Road safety assessment method and device based on millimeter wave radar and storage medium
CN114973659A (en) Method, device and system for detecting indirect event of expressway
CN116631186A (en) Expressway traffic accident risk assessment method and system based on dangerous driving event data
CN114267173B (en) Multisource data fusion method, device and equipment for space-time characteristics of expressway
CN111666859A (en) Dangerous driving behavior identification method
Ansariyar et al. Investigating the Vehicle-Bicyclists Conflicts using LIDAR sensor technology at signalized intersections
CN117037109A (en) Traffic target detection method based on improved YOLOv5 model
CN114241756B (en) Method and system for dynamically using hard road shoulder during construction of expressway
CN116257797A (en) Single trip track identification method of motor vehicle based on Gaussian mixture model
CN113128847A (en) Entrance ramp real-time risk early warning system and method based on laser radar
CN107038870A (en) A kind of vehicle peccancy detection method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant