Traffic conflict behavior recognition method for campus gate road section

Abstract

The invention relates to the technical field of traffic control, and discloses a traffic conflict behavior recognition method for a campus gate road section, which comprises the following steps of: firstly, constructing a traffic behavior and cognitive model suitable for the campus gate road section; and defining key points of conflict behaviors according to the traffic behavior and cognitive model and the campus gate road section traffic conflict bypassing behaviors, identifying the key points of the conflict behaviors and determining change tracks of the conflict behaviors, and finally determining conflict behavior indexes and conflict density indexes on the basis of track data. And judging whether to start traffic safety early warning or not according to the conflict density index. According to the invention, the traffic condition at the campus gate can be better monitored and identified, and the traffic management efficiency is improved.

Description

Technical Field

[0001] The invention relates to the technical field of traffic conflict behavior recognition at a road section at a campus entrance, and in particular to a traffic conflict behavior recognition method for a road section at a campus entrance. Background Art

[0002] In recent years, with the rapid development of the economy, the number of motor vehicles and non-motor vehicles in the city has increased, and traffic congestion and illegal driving have become increasingly serious. In order to ensure smooth traffic and reduce traffic accidents, the traffic in front of the school is very important. At the same time, the traffic management in front of the school is also required to be higher and higher. Because every time when the school starts and ends, there are many parents picking up their children at the school gate, and they park their vehicles at will, which leads to serious traffic congestion and even traffic paralysis. Pedestrians, non-motor vehicles, and motor vehicles are mixed together, and there are many traffic conflicts, which increases more safety hazards and traffic congestion.

[0003] The traditional definition and theoretical system of traffic conflict focuses on motor vehicles, while the traffic participants at the entrance of the campus include more pedestrians and non-motor vehicles. This type of traffic conflict is very different from a simple motor vehicle conflict. Since the basic principles of traffic conflict at the entrance of the campus are consistent with those of motor vehicle conflict, the definition of traffic conflict at the entrance of the campus is still based on the traditional definition of traffic conflict, that is, "if the direction or speed of traffic participants at the entrance of the campus, such as pedestrians, bicycles and electric bicycles, does not change, there is a high probability of collision", which is a prerequisite for the definition of traffic conflict at the entrance of the campus. The difficulty in defining traffic conflict at the entrance of the campus lies in distinguishing between normal interaction and traffic conflict, which is difficult to distinguish by a single indicator or quantitative indicator similar to the definition of motor vehicle traffic conflict. It is necessary to comprehensively consider multiple states such as running status, evasive action, operating space and traffic environment.

[0004] The general identification methods of rapid traffic conflicts involving motor vehicles include two categories: based on spatiotemporal proximity and based on risk avoidance behavior.

[0005] (1) Traffic conflict identification based on spatiotemporal proximity;

[0006] The proximity of traffic participants in time or space can be used as an important basis for traffic conflict judgment. Common time-space proximity conflict measurement indicators are TTC and PET. When traffic participants conflict, if the original movement path and speed remain unchanged, a collision will occur at a certain moment. The time from the beginning of the conflict to the occurrence of the collision is the TTC value. During the conflict, the minimum TTC (TTC min ) value is considered as an indicator of the severity of the conflict. In principle, the TTC minThe smaller the value, the higher the risk of collision and the higher the severity. The TTC index is widely used in motor vehicle rear-end collisions. For vehicles traveling in the same lane, the TTC calculation method is as follows:

[0007]

[0008] In the formula, X j,t is the coordinate of the position of traffic participant j at time t (m); X j+1,t is the coordinates of the position of traffic participant j+1 at time t (m); V j+1,t is the instantaneous speed of traffic participant j+1 at time t (m / s); V j,t is the instantaneous speed of traffic participant j at time t (m / s); L j is the length of the j traffic participant, usually considered in motor vehicle conflicts (m);

[0009] PET is defined as the time difference between traffic participants passing through potential collision points. The PET indicator is simple to calculate. The PET value can be calculated when there is a potential conflict point for traffic participants. The calculation process does not require indicators such as speed and distance. It only needs to record the time when traffic participants arrive at the potential conflict point. However, the premise of PET indicator calculation is that there must be an intersection in the trajectories of traffic participants. Therefore, PET value is often used to calculate traffic conflicts that occur at intersections and other places with intersecting trajectories.

[0010] In addition, time proximity indicators such as time to accident (TA), time exposed collision time (TET), time integrated TTC (TIT), time headway (TH), deceleration rate to avoid the crash (DRAC), and distance proximity indicators such as proportion of stopping distance (PSD) and lateral departure distance (Lateral Distance to Departure) are also used in the identification of various types of traffic conflicts.

[0011] (2) Conflict identification based on risk-avoidance behavior;

[0012] Avoidance behavior refers to the maneuvering behavior of traffic participants to quickly evade conflicts in order to avoid collisions. Rapid avoidance behavior can be detouring, braking, accelerating, or any combination of maneuvers. The advantage of the conflict identification method based on avoidance behavior is that the identification operation is flexible and simple, and it is suitable for situations where quantitative indicators such as speed, distance, and time are not easy to obtain, but the status of traffic participants and traffic status are clear. Its disadvantage is that it is difficult to distinguish between preventive behavior and avoidance behavior, and traffic conflicts in which traffic participants do not have avoidance behavior are easily overlooked. In addition, conflict identification based on avoidance behavior mainly relies on manual judgment, so it is not easy to achieve automated judgment.

[0013] The traffic conflict identification methods involving slow-moving traffic participants such as pedestrians, bicycles and non-motor vehicles mainly include objective indicator conflict identification technology, subjective traffic conflict identification technology and subjective and objective combined traffic conflict identification technology.

[0014] (1) Objective indicator conflict identification technology;

[0015] Temporal proximity indicators such as TTC, PET, and TA are also used in the identification of traffic conflicts involving slow-moving traffic participants. Generally, TTC values ​​less than 1.5s and PET values ​​less than 1s are considered dangerous in urban traffic, but there are often different threshold limits according to different traffic environments and types of traffic participants. Compared with the spatiotemporal proximity indicators suitable for describing conflicts between high-speed traffic participants, conflict indicators based on evasive behavior are more suitable for interactive and intimate slow-moving traffic conflicts. Due to the different traffic characteristics from motor vehicles, the selection of slow-moving conflict indicators based on evasive behavior is also different. The evasive behavior indicators of slow-moving traffic conflicts mainly include spatiotemporal gait parameters (step length and step frequency), acceleration, acceleration change rate, and angular velocity.

[0016] (2) Subjective traffic conflict identification technology;

[0017] This is because the driving routes of motor vehicles are clear and the driving behaviors are standardized. The traffic conflict process can be fully described by objective indicators such as spatiotemporal proximity index and evasive action index, or even a single indicator. In recent years, technologies such as computer vision and GPS have developed rapidly, and it is not difficult to extract and calculate objective indicators in slow-moving traffic conflicts. Although not limited by technical issues, objective indicators are still not widely used in slow-moving conflict identification. However, in the process of slow-moving traffic conflicts, pedestrians, bicycles, electric bicycles, etc. have high randomness in path selection, poor driving stability, flexible speed changes and detours, and objective indicators can only describe part of the slow-moving traffic conflict. Therefore, some subjective conflict techniques have been proposed, including British Traffic Conflict Techniques (British TCT), French TCT, German TCT, Austrian TCT, and Czech TCT. These subjective conflict techniques use some pre-defined subjective severity rating methods to identify critical events, which are often related to the degree of traffic interaction intimacy and uncontrolled evasive behavior of traffic participants.

[0018] (3) Traffic conflict identification technology combining subjective and objective factors;

[0019] Traffic conflict technology that combines objective indicators and subjective evaluation combines the advantages of the two methods and combines different indicators to better identify slow-moving traffic conflicts. For example, Canadian Traffic Conflict Technology (Canadian TCT) and Dutch Conflict Technology DOCTOR. Canadian TCT combines TTC min It is used in combination with the subjective component of Risk of Collision (ROC) to identify conflicts. ROC is divided into three levels: low, medium, and high, and divided into TTC min <2s, TTC min <1.6s and TTC min <1s three levels, and add them up to estimate the final conflict severity. DOCTOR was developed by SWOV and TNO in the Netherlands. It is a standardized manual observation technology with objective and clear observation units, operated and used by well-trained observers. At present, DOCTOR has been well applied in the field of slow-moving traffic conflicts. van der Horst et al. used DOCTOR to record and analyze the conflict behaviors of cyclists, and used this method to observe serious slow-moving traffic conflicts in Bangladesh, thereby effectively monitoring and evaluating the effectiveness of road safety intervention measures.

[0020] The above analysis shows that the current technology for effectively identifying traffic conflicts between pedestrians, bicycles and electric bicycles in the road section at the school gate is not mature. The traffic density at the school gate is relatively high, and the conflict analysis between two pedestrians can no longer meet the safety needs. Based on the safety domain, a conflict density index is proposed to conduct risk assessment on the detour path, so as to facilitate the early warning and diversion of traffic conditions in the road section at the school gate.

[0021] Therefore, technical personnel in this field need a traffic conflict identification method that can be effectively applied to the road section in front of the school, so as to improve the safety of students leaving school, reduce the traffic pressure on the road section in front of the school during school hours, and alleviate traffic congestion. Summary of the invention

[0022] The purpose of the present invention is to solve the above problems and to design a method for identifying traffic conflict behaviors at the road section at the entrance of a campus.

[0023] To achieve the above-mentioned purpose, the technical solution of the present invention is a method for identifying traffic conflict behaviors at a road section at a campus entrance, the method comprising the following steps:

[0024] Step 1: Conflict identification. Record the conflict variables in the traffic conflict at the entrance of the campus, define the traffic conflict characteristic index to describe and analyze the traffic conflict characteristics at the entrance of the campus, and then grade the severity of the traffic conflict at the entrance of the campus based on the avoidance behavior.

[0025] Step 2: Based on the conflict identification process, the driving behavior cognitive theoretical model constructed by Huguenin is combined with the changes in conflict behaviors and cognition of traffic participants in the road section at the campus gate, and a traffic behavior and cognitive model applicable to the road section at the campus gate is constructed. The key points of conflict behaviors are defined and the conflict stages are divided. The key points of conflict behaviors are defined based on the traffic behavior and cognitive model and the detour behaviors of traffic conflicts in the road section at the campus gate, and then the conflict stages are divided based on the key points of conflict behaviors.

[0026] Step 3: Identify the key points of the conflicting behaviors by analyzing the lateral motion direction, acceleration and deceleration status, and extreme values ​​of trajectory coordinates to accurately identify the key points of the conflicting behaviors, including the starting and winding points, winding back points, dividing points, and the changing points of the starting and winding back, and clarify the changing trajectory of the conflicting behaviors;

[0027] Step 4: Define conflict behavior indicators and calculate and analyze traffic conflict micro-behaviors. Based on trajectory data, define conflict behavior indicators, then calculate and analyze traffic conflict micro-behaviors based on the defined conflict behavior indicators, and determine the conflict density indicators of multiple traffic participants in the road section at the campus gate. Then, determine whether to start traffic warning and traffic diversion procedures based on the conflict density indicators.

[0028] It should be noted that the conflict behavior index includes the 'safety domain', and then the micro-behavior of traffic conflict is calculated and analyzed based on the defined conflict behavior index to comprehensively describe the entire process of traffic conflict, thereby providing a more accurate and comprehensive perspective for evaluating and preventing potential traffic accidents.

[0029] Among them, the "safety domain" is the coordinate system of each trajectory reset, the trajectory starting point is set as the coordinate origin, the virtual trajectory direction is set as the Y-axis, the actual conflict trajectory curve is the actual motion trajectory of the circumnavigator in the coordinate system, the original trajectory is the idealized straight-line distance of the circumnavigator from the starting point to the return point in the negative direction of the Y-axis, and the area enclosed by the actual conflict trajectory curve and the ideal original trajectory straight line is called the safety domain.

[0030] The cognitive changes of traffic participants in the section of the campus gate in the step 2 refer to the different attitudes, information processing processes and physiological and psychological natural reactions of traffic participants in the section of the campus gate based on the changes in their cognition of traffic conditions and goals. The changes in conflict behaviors of traffic participants in the section of the campus gate refer to the final traffic behaviors formed after the operations taken by traffic participants in the section of the campus gate based on their attitudes and information processing processes and considering the influence of natural reactions.

[0031] The key points of the conflicting behaviors defined in step 3 are as follows:

[0032] ① The detour point is the point where the traffic participants start to detour. Under the traffic conditions where the conflict begins, the action is to detour from normal driving;

[0033] ② The detour change point, which is the point at which the traffic participants believe that the degree of completion of the detour stage has changed;

[0034] ③ The farthest point in the horizontal direction: the point where the traffic participant reaches the farthest distance in the horizontal direction during the detour. The action is manifested as a change in the horizontal detour direction to the left or right.

[0035] ④ The detour change point, which is the point at which the degree of completion of the detour stage considered by traffic participants changes;

[0036] ⑤ Detour point: the point where the traffic participant ends the detour. When the conflict is about to end, the action is manifested as returning to normal driving from detour.

[0037] The conflict stages divided in step 3 include:

[0038] ① The pre-detour stage, from the point where the conflict begins to the point where the detour begins;

[0039] ② Winding stage, from the starting point to the farthest point in the horizontal direction;

[0040] ③ The rewinding stage, from the farthest point horizontally to the rewinding point;

[0041] ④ In the post-detour stage, circle back to the point where the conflict ends.

[0042] The method for identifying the key points of the conflicting behavior in step 4 includes: a method for identifying the starting and winding points, a method for identifying the dividing points, and a method for identifying the starting and winding changing points.

[0043] The conflict behavior indicators defined in step 4 include: lateral decision distance D lateral_decisio , n longitudinal decision distance D long_decision , lateral winding distance D lateral_away , Longitudinal winding distance D long_away , lateral maneuvering distance D lateral_back , Longitudinal maneuvering distance D long_back , lateral safety distance D lateral_safet , y longitudinal safety distance D long_safety , Transverse winding change ratio R lateral_away , longitudinal winding change ratio R long_away , lateral winding change ratio R lateral_back , longitudinal winding change ratio R long_back And security domain S safety zone .

[0044] The process of calculating the traffic conflict micro behavior in step 5 includes:

[0045] Horizontal decision distance D lateral_decision and longitudinal decision distance D long_decision The calculation method is shown in formulas (1) and (2):

[0046] D lateral_decision =|X start_swerve -X start_another | (1)

[0047] D long_decision =|Y start_swerve -Y start_another | (2)

[0048] In the formula, X start_swerve is the X coordinate of the detour starting point of the detour traffic participant (m); start_another is the X coordinate (m) of the other traffic participant when the detour traffic participant is at the detour starting point; Y start_swerve is the Y coordinate of the detour starting point of the detour traffic participant (m); start_another is the Y coordinate (m) of the other traffic participant when the detour conflict participant participates at the detour starting point;

[0049] Transverse winding distance D lateral_away and longitudinal winding distance D long_awayThe calculation formula is shown in formula (3) and (4), the lateral maneuvering distance D lateral_back and longitudinal reversal distance D long_back The calculation method is shown in equations (5) and (6):

[0050] D lateral_away =|X farthest_swerve -X start_swerve | (3)

[0051] D long_away =|Y farthest_swerve -Y start_swerve | (4)

[0052] D lateral_back =|X end_swerve -X farthest_swerve | (5)

[0053] D long_back =|Y end_swerve -Y farthest_swerve | (6)

[0054] Where, X farthest_swerve is the X coordinate of the detour traffic participant demarcation point (m); Y farthest_swerve is the Y coordinate of the detour traffic participant demarcation point (m); end_swerve is the X coordinate of the point where the detour traffic participant returns (m); Y end_swerve is the Y coordinate of the detour participant's return point (m),

[0055] Lateral safety distance D lateral_safety And longitudinal safety distance D long_safety The calculation method is shown in equations (7) and (8):

[0056] D lateral_safety =|X farthest_swerve -X farthest_another | (7)

[0057] D long_safety =|Y farthest_swerve -Y farthest_another | (8)

[0058] Where, X farthest_another is the X coordinate (m) of the other traffic participant when the detour traffic participant is at the demarcation point; Y farthest_another is the Y coordinate of the other traffic participant when the detour traffic participant is at the demarcation point (m);

[0059] Transverse winding change ratio R lateral_away And the longitudinal winding change ratio R long_away The calculation method is as shown in formula (9) and (10), the lateral winding change ratio R lateral_back And the longitudinal winding change ratio Rlong_back The calculation method is as shown in formulas (11) and (12):

[0060]

[0061] In the formula, X change_away is the X coordinate of the detour traffic participant's starting point (m); Y change_away is the Y coordinate of the detour traffic participant's starting point (m); X change_back is the X coordinate of the detour traffic participant's return to the change point (m); Y change_back is the Y coordinate of the detour participant's return to the change point (m),

[0062] Security Domain S safetyzone The calculation method of is as follows:

[0063]

[0064] The conflict density index in step 4 is the ratio of the product of the number of non-motor vehicles and pedestrians in the area and the average area occupied by non-motor vehicles and single pedestrians to the area of ​​the safety zone, as shown in the following formula:

[0065]

[0066] Where N i is the number of individuals of each type; S l is the average area of ​​each type of individuals; ρ confilict is the conflict density index; S safetyzone It is a security domain;

[0067] Among them, ρ confilct It is a real-time changing indicator, which depends on the nearest detour object within the selected boundary. When its value is within the safe range, it can be used as a safe path. The larger the value, the greater the possibility of conflict and the unsafe it is. At this time, a traffic safety warning should be initiated and the staff should be reminded to implement traffic diversion. The smaller the value, the smaller the possibility of conflict and the safer it is. There is no need to issue a traffic safety warning or conduct traffic diversion.

[0068] Compared with the prior art, this application has the following advantages:

[0069] 1. The present invention is based on a traffic conflict micro-behavior identification method, which uses a conflict density index to identify the traffic situation at the entrance of the campus, and implements an automatic warning function for traffic conflicts based on the identification situation, prompting staff to guide traffic conflicts and congestion at the entrance of the campus;

[0070] 2. The present invention applies the conflict behavior index to calculate and analyze the micro behavior of traffic conflict, and defines or identifies the "safety domain", so as to comprehensively describe the entire action process of traffic conflict, thereby providing a more in-depth and continuous perspective for traffic safety management, and avoiding the problem that the existing traffic conflict assessment method cannot fully capture the continuity and dynamic changes of the conflict based on safety substitute indicators such as lateral position at key moments, instantaneous speed and comfort zone boundaries. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a flow chart of a method for identifying traffic conflict behaviors at a road section at a campus entrance according to the present invention;

[0072] Figure 2 is the driving behavior and cognitive model diagram of the present invention;

[0073] Figure 3 is a schematic diagram of the key points of the conflict behavior of the present invention;

[0074] Figure 4 is a schematic diagram of the conflict phase involving the detour behavior described in the present invention;

[0075] Figure 5 It is a table of the method for identifying the starting point and the winding point of the present invention;

[0076] Figure 6 It is a partial schematic diagram from the winding starting point to the winding starting change point of the present invention;

[0077] Figure 7 is a schematic diagram of the starting and winding back change points of the present invention;

[0078] Figure 8 is a schematic diagram of the behavioral indicators of the present invention;

[0079] Fig. 9 is a schematic diagram of the security domain of the present invention;

[0080] Fig.10 This is the traffic conflict identification process of the road section at the entrance of the campus described in the present invention;

[0081] Fig.11 is the conflict severity rating table of the present invention;

[0082] Fig.12 is an example diagram of conflicts of different severity levels according to the present invention;

[0083] Fig.13 is the conflict record variable table of the present invention;

[0084] Fig.14 is the traffic conflict characteristic index table of the present invention;

[0085] Fig.15 It is a flow chart of the operation of the campus gate road section conflict technology of the present invention;

[0086] Fig.16 It is a virtual boundary map on both sides of the individual form of the traffic participant described in the present invention. DETAILED DESCRIPTION

[0087] The present invention will be described in detail below in conjunction with the accompanying drawings. Figure 1-16 As shown;

[0088] First, traffic conflicts in the road section in front of the campus gate are identified. In the prior art, there is no accurate and effective identification method for traffic conflicts between pedestrians, bicycles and electric bicycles in the road section in front of the campus. This is the main problem to be solved by this application. In the above-mentioned conflict identification method, DOCTOR combines the advantages of subjective and objective conflict indicators. Its definition, evaluation indicators and operation methods are relatively suitable for slow-moving traffic conflict identification, and have good application verification. However, the identification scope of DOCTOR includes motor vehicle conflicts and non-motor vehicle conflicts, and it is still not detailed and specific enough in the identification of traffic conflicts in the road section in front of the campus. This application is based on DOCTOR and forms a method for identifying traffic conflicts and determining their severity in the road section in front of the campus.

[0089] It should be noted that the identification of traffic conflicts is mainly completed by the definition of conflicts in the road section at the campus gate. Each conflict involves only two traffic participants, and continuous conflicts or multi-person conflicts are broken down into conflicts between multiple two traffic participants. Each traffic conflict is identified by two observers, and undisputed events are coded and recorded. Inconsistent conflict events are discussed and the final judgment results are determined. In order to verify the consistency and reliability of conflict coding, as well as the reliability of observers, random conflict events are selected and re-coded 10 days after the conflict identification is completed. The consistency rate of the two codes is above 85%, which is qualified. The specific process of conflict identification is shown in Fig.10 .

[0090] Then the severity of traffic conflicts in the section in front of the campus gate is graded. The severity of the conflict is mainly determined by the possibility of collision and the severity of the potential consequences of the collision. The conflict situation, such as preventive, controlled, relatively sudden, urgent or almost collision, is a key factor in assessing the severity of the conflict. In addition, unlike the evasive behaviors such as changing lanes and stopping of motor vehicles, the evasive behaviors of slow-moving traffic participants are more likely to be detours, slow down, etc., and the degree of control of the evasive behavior is one of the important indicators to measure the severity. The objective conflict indicators in the severity judgment process use TTC and PET estimates. In addition, due to the slow speed of slow-moving traffic participants, the maneuvering time is not as important as the motor vehicle conflict in judging the severity of traffic conflicts in the section in front of the campus, but whether the operating space is sufficient greatly affects the severity of traffic conflicts in the section in front of the campus. The severity of traffic conflicts in the section in front of the campus gate is also judged by two observers. The specific judgment criteria for each level can be found in. Fig.11 .

[0091] Examples of conflicts at each level of severity are shown in Figure Fig.12 . Fig.12 (a) The conflict shown has ample operating space, a simple traffic environment, and the electric bicycle takes preventive evasive action, operating calmly. The possibility of collision is extremely small, which meets the characteristics of Level 1 conflict. Fig.12 In the conflict shown in (b), the electric bicycle takes a controlled evasive action, the electric bicycle is tilted to a certain extent, and there is ample time and space for the evasive action, which meets the characteristics of a level 2 conflict; Fig.12 In the conflict shown in (c), there is relatively ample time and space for evasion behavior, and the evasion behavior is relatively sudden, which meets the characteristics of level 3 conflict; Fig.12 In the conflict shown in (d), the electric bicycle loses control by pedaling on the ground, and the pedestrian evades in an emergency, leaving limited room for maneuver, which meets the characteristics of a Level 4 conflict. Fig.12 In the conflict shown in (e), the electric bicycle and the pedestrian almost collided, which meets the characteristics of Level 5 conflict.

[0092] Then, traffic conflicts at the entrance of the campus are identified. The conflict coding variables to be recorded during the identification process are shown in Fig.13 As shown in Figure 1, these coded variables constitute a completed conflict event. Fig.14 , including three parts: conflict basic information record, evasion behavior record and conflict severity judgment record. In addition, this paper defines traffic conflict characteristic indicators such as traffic ratio, conflict ratio, evasion ratio, evasion loss ratio and conflict distraction ratio based on the above conflict variables, which can be used to describe and analyze the traffic conflict characteristics of the road section at the entrance of the campus. The definition and calculation method of the indicators can be found in Fig.15 shown.

[0093] The basic information of the conflict is recorded after the conflict identification is completed, including the conflict type variables and traffic participant characteristic variables. Among them, the conflict angle range of 0°±30° is the same-direction conflict, the conflict angle range of 180°±30° is the opposite-direction conflict, and the conflict in between is defined as the cross conflict. Pedestrian standing conflict refers to the conflict between pedestrians and other traffic participants when they are standing still. The attributes of traffic participants such as age and gender are determined based on observations of appearance, clothing, shape and behavior.

[0094] The avoidance behavior mainly records the avoidance behavior characteristic variables. The avoidance action type and whether it is out of control are obtained through observation. The signs of out-of-control behavior include the rider using the foot to help avoid falling, the bicycle and electric bicycle shaking and difficult to maintain balance, etc. It should be noted that the left and right of the detour direction is defined according to the perspective of the traffic participant, not the observer.

[0095] The judgment of the severity of the conflict is completed according to the above-mentioned traffic conflict severity grading method for the road section in front of the campus, and mainly records the objective conflict indicators, speed estimation and severity judgment results. The TTC estimation value is completed based on the distance estimation and time. The size of the floor tiles, the size of the tree pit and the size of the bus stop sign can provide references for the distance estimation. The PET value is calculated based on the frame difference between the two traffic participants arriving at the potential conflict point. The TTC estimation value and the PET value only record the range of their values ​​(for example, 0-0.5s), and do not need to be estimated to a specific value. The average speed is completed based on a known fixed distance in the road section in front of the campus and the number of frames that the traffic participants pass through this distance and their operation conditions. Since the speed of pedestrians varies in a narrow range and a small amplitude, no speed estimation is performed.

[0096] This application is based on the driving behavior cognitive theory model constructed by Huguenin. Figure 2 . Based on the changes in the cognition of traffic conditions and goals, traffic participants have different attitudes, information processing processes, and physiological and psychological natural reactions. Traffic participants take certain actions based on attitudes and information processing processes, and consider the impact of natural reactions to form the final traffic behavior. Traffic participants' cognition of traffic conditions and goals mainly occurs at the attitude level. In general, traffic behavior is goal-oriented and regulated by attitudes. In complex situations, traffic behavior is mainly controlled by the attitudes of traffic participants. Compared with motor vehicles, traffic participants in the road section in front of the campus have a lower speed. Except for sudden turns and sudden brakes in emergency situations, they generally do not have natural reaction actions that have a greater impact in a short period of time. Therefore, traffic conflict behaviors in the road section in front of the campus are mainly caused by changes in attitudes and corresponding operations after information processing. Correspondingly, if changes are observed in the conflict behaviors of traffic participants in the road section in front of the campus, they often correspond to changes in their behavioral cognition.

[0097] Among all types of evasive behaviors, the average value of behaviors involving detours is as high as 94%, behaviors that only slow down are less than 6%, and backward behaviors rarely occur. In terms of conflict types, opposite-direction conflicts and same-direction conflicts are the main types of conflicts at the entrance of the campus. Therefore, this application focuses on the micro-behaviors involving detours in opposite-direction and same-direction conflicts. This application is based on the driving behavior cognitive theoretical model, and by analyzing the key conflict behaviors during the conflict process, it obtains the cognitive changes and attitudes of traffic participants, and finally constructs a micro-behavior and cognitive model of traffic conflicts at the entrance of the campus.

[0098] The detour behavior in the traffic conflict section at the campus gate is mainly manifested in the change of lateral direction or lateral speed (V x ) changes. Based on this, we can get 5 key points of detour behavior, see the example Figure 3 . Figure 3 V on the left x , and the right side is the corresponding detour behavior key point. The specific behavior key points are: ① detour starting point, the point where the traffic participant starts to detour. Under the traffic conditions where the conflict begins, the action is to detour from normal driving; ② detour change point, the point where the traffic participant believes that the completion degree of the detour stage has changed, and the action is V x Changes in acceleration and deceleration states; ③ The farthest point in the horizontal direction, the point where the traffic participant reaches the farthest distance in the horizontal direction during the detour, and the action is manifested as a change in the lateral detour direction to the left and right; ④ The point of return change, the point where the traffic participant believes that the degree of completion of the detour has changed, and the action is manifested as V x Changes in acceleration and deceleration states; ⑤ The detour point, the point where the traffic participant ends the detour. When the conflict is about to end, the action is manifested as a return to normal driving from detour.

[0099] According to the traffic behavior and cognitive model, the behavior of traffic participants under different traffic conditions is caused by changes in internal cognition. The starting point and the detour point are derived from the judgment of the beginning and end of the conflict. The farthest point in the horizontal direction is the point where the traffic participant makes the detour decision when he believes that the conflict has been resolved to a large extent. The starting point and the detour change point represent the cognitive judgment of the traffic participants on the completion of the starting and detour stages.

[0100] Taking the starting and winding change point as an example, at the beginning of the winding stage, the traffic participant is eager to avoid danger, resolve the conflict, and stay away from the other traffic participant. But when the winding stage is completed to a certain extent, the traffic participant believes that the conflict is no longer dangerous. Although its horizontal movement direction has not changed and it is still circling away from the other traffic participant, the lateral speed has changed from acceleration to deceleration. The focus of the traffic participant's behavior strategy has shifted to the longitudinal forward movement to prepare for the winding stage. Similarly, at the winding change point, the traffic participant believes that he has basically completed the task of winding back. Although the trajectory is still circling back, the focus has shifted to continuing to move forward. The identification of action key points is based on the horizontal component of speed V x and trajectory shape realization, the specific discrimination method will be introduced in the next section.

[0101] According to the definition of key behavior points, the conflict behaviors involving detours at the entrance of the campus can be divided into four stages. Figure 4 : ① Pre-detour stage, from the point where the conflict starts to the detour starting point; ② Detour starting point, from the detour starting point to the farthest point in the horizontal direction; ③ Detour stage, from the farthest point in the horizontal direction to the detour point; ④ Post-detour stage, from the detour point to the point where the conflict ends. The farthest point in the horizontal direction is also the dividing point between the detour starting stage and the detour returning stage (hereinafter referred to as the dividing point).

[0102] Only the avoidance type of "slow down first and then detour" has a pre-detour stage, and only the avoidance type of "detour first and then slow down" has a post-detour stage. Traffic participants often use these two stages to slow down. In addition, some conflicts do not have a detour stage, that is, after reaching the demarcation point, they do not go back to the original trajectory X-axis coordinate position, but move directly along the Y-axis direction. However, all avoidance behaviors involving detours have a detour start stage.

[0103] The behavior key point identification method used in this application includes:

[0104] (1) Method for identifying the starting point and the winding point

[0105] In the conflict detour trajectory at the entrance of the campus, the starting point and the detour point have the same characteristics, which are often the points where the lateral running direction changes or the lateral running direction remains unchanged but the acceleration and deceleration state changes. Based on this analysis, the five types of V corresponding to this behavior key point are summarized. x Feature points and corresponding trajectory shapes, see Figure 5 The identification method of the starting and winding points is to preliminarily determine the locations of the starting and winding points of the traffic participants according to the trajectory shape and running status, and then x Feature points, accurately determine the frame numbers and corresponding spatial coordinates of the starting point and the winding point.

[0106] It should be noted that Type 2 and Type 4 use 0.1m / s as V xThe demarcation value is used to distinguish whether the traffic participants are moving straight or diagonally. x When V is equal to 0, the traffic participants are moving in a straight line. However, this state is too ideal. It is difficult for pedestrians, bicycles and electric bicycles to maintain an absolute vertical position when moving forward. x The conflict process of traffic participants with different values ​​is conducted by two observers in the x The results showed that in V x <0.1m / s, the observer believes that the traffic participants are moving in a straight line. x <0.1m / s, human beings will subjectively believe that they are moving in a straight line, but due to multiple reasons such as balance, they cannot reach V x A completely ideal state equal to 0.

[0107] (2) Method for identifying demarcation points

[0108] The demarcation point is also the point with the longest lateral distance on the detour trajectory. Depending on the detour direction, the X coordinate of the traffic participant's trajectory reaches the maximum or minimum value at this moment, and then decreases or increases. This is the fundamental method for identifying the demarcation point. In addition, the obvious feature of the demarcation point is that the traffic participant changes the detour direction from left to right or from right to left. Correspondingly, at this moment, V x The value of is a 0 value point from negative to positive or from positive to negative, and Figure 5 Medium V x The type 1 of the feature point is the same.

[0109] (3) Method for identifying starting and winding change points

[0110] At the starting and winding changing points, V x The acceleration and deceleration state changes, corresponding to V x is the peak or valley value, and Figure 5 Medium V x The eigenvalues ​​are the same as type 3. It is worth noting that not all trajectories have clear starting and winding change points. Some trajectories are too tortuous and do not have starting and winding change points. In addition, the starting and winding stages are divided into two arcs by the change points, and the centers of the arcs point in opposite directions. Taking the trajectory from the starting point to the starting point as an example, at this time, the center of gravity of the traffic participant is laterally away from the traffic participant, and the lateral speed increases. However, traffic participants often do not accelerate in the longitudinal direction, because this will accelerate their conflicts, which is obviously not good for traffic participants. Therefore, when traffic participants travel an equal distance in the longitudinal direction, that is, Figure 6When b1 is equal to b2, the horizontal travel distance increases, that is, a2 is greater than a1. From this, we can know the changing direction of the arc tangent, and thus determine the orientation of the arc center. The same is true for the three trajectories from the starting point to the dividing point, from the dividing point to the winding change point, and from the winding change point to the winding point. See the specific Figure 7 .

[0111] The definition and calculation method of conflict behavior indicators in this application are as follows:

[0112] The trajectories of traffic participants are output as tuples, including the horizontal and vertical coordinates X of each frame. i ,Y i and instantaneous speed Based on trajectory data, this application defines a series of conflict behavior indicators and introduces their calculation methods. The specific definitions of the indicators are shown in Figure 8 . This application uses a two-sample t-test and analysis of variance (ANOVA) with a significance level of p<0.05 to test the significance of differences in behavioral indicators. The two-sample t-test is used for two independent groups, and ANOVA is used for three or more independent groups. In addition, since Chapter 3 mainly explores the characteristics of conflicts, the location is used as the calculation unit, and the statistical number of each location is used as the calculation sample, so statistical calculations are performed on each location separately. This application uses traffic conflicts as the calculation unit, and the behavioral indicator results of each conflict are used as the calculation sample, so the subsequent analysis summarizes the statistics of the four campus gate sections.

[0113] The decision distance is defined as the distance between the detour participant and the other traffic participant at the detour starting point, that is, at what distance between the detour conflict participant and the other traffic participant does the decision to take detour evasive action begin. lateral_decision and longitudinal decision distance D long_decision The calculation method is shown in equations (1) and (2).

[0114] D lateral_decision =|X start_swerve -X start_another | (1)

[0115] D long_decision =|Y start_swerve -Y start_another | (2)

[0116] In the formula, X start_swerve is the X coordinate of the detour starting point of the detour traffic participant (m); start_another is the X coordinate (m) of the other traffic participant when the detour traffic participant is at the detour starting point; Y start_swerve is the Y coordinate of the detour starting point of the detour traffic participant (m); start_anotherIt is the Y coordinate (m) of the other traffic participant when the detour conflict participant participates at the detour starting point.

[0117] The maneuvering distance is defined as the distance covered by the detour traffic participant during the detour phase and the detour phase. lateral_away and longitudinal winding distance D long_away The calculation formula is shown in formula (3) and (4), the lateral maneuvering distance D lateral_back and longitudinal reversal distance D long_back The calculation method is shown in equations (5) and (6). In addition, the time required for the start-up and winding phases is analyzed and defined as the start-up and winding maneuvering time and the winding maneuvering time, respectively.

[0118] D lateral_away =|X farthest_swerve -X start_swerve | (3)

[0119] D long_away =|Y farthest_swerve -Y start_swerve | (4)

[0120] D lateral_back =|X end_swerve -X farthest_swerve | (5)

[0121] D long_back =|Y end_swerve -Y farthest_swerve | (6)

[0122] In the formula, X farthest_swerve is the X coordinate of the detour traffic participant demarcation point (m); Y farthest_swerve is the Y coordinate of the detour traffic participant demarcation point (m); end_swerve is the X coordinate of the point where the detour traffic participant returns (m); Y end_swerve is the Y coordinate (m) of the point where the detour traffic participant returns.

[0123] The safety distance is defined as the distance between the detour participant and the other traffic participant at the demarcation point. At this point, the detour participant believes that the distance between him and the other traffic participant is safe and starts to detour. The lateral safety distance D lateral_safety And longitudinal safety distance D long_safety The calculation method is shown in equations (7) and (8).

[0124] D lateral_safety =|X farthest_swerve -X farthest_another | (7)

[0125] D long_safety =|Y farthest_swerve -Y farthest_another| (8)

[0126] In the formula, X farthest_another is the X coordinate (m) of the other traffic participant when the detour traffic participant is at the demarcation point; Y farthest_another It is the Y coordinate (m) of the other traffic participant when the detour traffic participant is at the dividing point.

[0127] The starting and winding change ratio is defined as the ratio of the distance from the starting and winding point to the starting and winding change point to the entire starting and winding maneuvering distance, and the winding change ratio is defined as the ratio of the distance from the dividing point to the winding change point to the entire winding maneuvering distance. The concept of change ratio describes the degree to which the cognitive state of traffic participants changes when they complete the starting and winding phases or the winding phases. lateral_away And the longitudinal winding change ratio R long_away The calculation method is shown in equations (9) and (10), and the lateral winding change ratio R lateral_back And the longitudinal winding change ratio R long_back The calculation method is shown in equations (11) and (12).

[0128]

[0129]

[0130] In the formula, X change_away is the X coordinate of the detour traffic participant's starting point (m); Y change_away is the Y coordinate of the detour traffic participant's starting point (m); X change_back is the X coordinate of the detour traffic participant's return to the change point (m); Y change_back is the Y coordinate (m) of the detour traffic participant's return to the change point.

[0131] In existing research, a series of safety alternative measures such as the lateral position, instantaneous speed and comfort zone boundary when a traffic participant exceeds another traffic participant are often used to evaluate the entire conflict process. For example, the comfort zone boundary describes the minimum lateral distance when a traffic participant exceeds another traffic participant during the conflict. However, most of these indicators describe the entire conflict process through a key moment in the conflict, and the description of the conflict is not complete. This application proposes a new indicator, the safety domain, which can be used to describe the entire conflict process. The safety domain is defined as the area enclosed by the actual conflict trajectory and the virtual original trajectory if no conflict occurs, that is, the additional road area occupied by the traffic participant due to the conflict. For ease of calculation, the coordinate system of each trajectory is reset, the trajectory starting point is set to the coordinate origin, and the virtual trajectory direction is set to the Y axis. It should be clear that this coordinate system is only applicable to the calculation of the safety domain indicator, and does not affect the actual coordinates of the trajectory and other indicator calculations. The safety domain S for the actual conflict trajectory and the virtual original trajectory with or without intersection safetyzoneSee Fig. 9 , the calculation method is shown in formula (13).

[0132]

[0133] It should be noted that the traffic density at the school gate is relatively high, and simple conflict analysis between two routes can no longer meet the complex traffic warning needs at the school gate. Therefore, based on the safety domain, this application proposes a conflict density index to conduct risk assessment on detour routes, so as to facilitate the path selection and risk warning of traffic participants.

[0134] First, when selecting the detour object, two virtual boundaries are established that are tangent to the individual shapes of traffic participants and parallel to the direction of travel. The individual with the shortest distance from the traffic participants within the boundary range is the detour object, such as Fig.16 As shown;

[0135] Then, a risk assessment index is constructed. When the detourer selects a detour object, a corresponding safety zone appears. The product of the number of non-motor vehicles and pedestrians in the zone and the average area occupied by non-motor vehicles and a single pedestrian is the ratio of the accumulated total area to the safety zone area, which is the conflict density index.

[0136] Therefore, the conflict density index in step 4 is determined by the ratio of the product of the number of non-motor vehicles and pedestrians in the area and the average area occupied by non-motor vehicles and single pedestrians to the area of ​​the safety zone, as shown in the following formula:

[0137]

[0138] Where N i is the number of individuals of each type; S l is the average area of ​​each type of individuals; ρ conflict is the conflict density index; S safety zone It is a security domain;

[0139] Among them, ρ conflict When the value of exceeds the safety threshold, it means that the possibility of conflict increases, the traffic environment at the entrance of the campus is unsafe, and it is necessary to start a traffic safety warning and implement traffic diversion; ρ conflict When the value is lower than the safety threshold, it means that the possibility of conflict is reduced, the traffic environment at the entrance of the campus is safe, and there is no need to activate the traffic warning.

[0140] The above technical solutions only reflect the preferred technical solutions of the technical solutions of the present invention. Some changes that may be made to certain parts thereof by technicians in this technical field all reflect the principles of the present invention and fall within the protection scope of the present invention.

Claims

1. A method for identifying traffic conflict behaviors at the entrance of a campus, characterized in that: The method comprises the following steps: Step 1: Conflict identification. Record the conflict variables in the traffic conflict at the entrance of the campus, define the traffic conflict characteristic index to describe and analyze the traffic conflict characteristics at the entrance of the campus, and then grade the severity of the traffic conflict at the entrance of the campus based on the avoidance behavior. Step 2: Based on the conflict identification process, a traffic behavior and cognitive model suitable for the road section at the campus gate is constructed to define the key points of conflict behavior and divide the conflict stages; Step 3: Identify the key points of the conflicting behaviors by analyzing the lateral movement direction, acceleration and deceleration status, and extreme values ​​of trajectory coordinates of traffic participants to accurately identify the key points of the conflicting behaviors; Step 4: Define the conflict behavior indicators of each traffic participant and calculate and analyze them. Based on the trajectory data, define the conflict behavior indicators, then calculate and analyze the micro-behavior of traffic conflicts based on the defined conflict behavior indicators, and determine the conflict density indicators of multiple traffic participants in the road section in front of the campus gate. Then, determine whether to initiate a traffic safety warning based on the conflict density indicators.

2. The method for identifying traffic conflict behaviors at the entrance of a campus according to claim 1, characterized in that: The key points of the conflicting behaviors defined in step 2 are as follows: ① The detour starting point is the point where the traffic participants start to detour. Under the traffic conditions where the conflict begins, the action is to detour from normal driving; ② The detour change point, which is the point at which the traffic participants believe that the degree of completion of the detour stage has changed; ③ The farthest point in the horizontal direction: the point where the traffic participant reaches the farthest distance in the horizontal direction during the detour. The action is manifested as a change in the horizontal detour direction to the left or right. ④ The detour change point, which is the point at which the degree of completion of the detour stage considered by traffic participants changes; ⑤ Detour point: the point where the traffic participant ends the detour. When the conflict is about to end, the action is manifested as returning to normal driving from detour.

3. The method for identifying traffic conflict behaviors at the entrance of a campus according to claim 2, characterized in that: The conflict stages divided in step 2 include: ① The pre-detour stage, from the point where the conflict begins to the point where the detour begins; ② Winding stage, from the starting point to the farthest point in the horizontal direction; ③ The rewinding stage, from the farthest point horizontally to the rewinding point; ④ In the post-detour stage, circle back to the point where the conflict ends.

4. The method for identifying traffic conflict behaviors at the entrance of a campus according to claim 3, characterized in that: The method for identifying the key points of the conflicting behavior in step three includes: a method for identifying the starting and winding points, a method for identifying the dividing points, and a method for identifying the starting and winding changing points.

5. The method for identifying traffic conflict behaviors at the entrance of a campus according to claim 1, characterized in that: The conflict behavior indicators defined in step 4 include: lateral decision distance D lateral_decisio , n longitudinal decision distance D long_decision , lateral winding distance D lateral_away , Longitudinal winding distance D long_away , lateral maneuvering distance D lateral_back , Longitudinal maneuvering distance D long_back , lateral safety distance D lateral_safety , longitudinal safety distance D long_safety , Transverse winding change ratio R lateral_away , longitudinal winding change ratio R long_away , lateral winding change ratio R lateral_back , longitudinal winding change ratio R long_back And security domain S safetyzone .

6. The method for identifying traffic conflict behaviors at the entrance of a campus according to claim 5, characterized in that: The process of calculating the traffic conflict micro-behavior in step 4 includes: Horizontal decision distance D lateral_decision and longitudinal decision distance D long_decision The calculation method is shown in formulas (1) and (2): D lateral_decision =|X start_swerve -X start_another | (1) D long_decision =|Y start_swerve -Y start_another | (2) Where, X start_swerve is the X coordinate of the detour starting point of the detour traffic participant (m); start_another is the X coordinate (m) of the other traffic participant when the detour traffic participant is at the detour starting point; Y start_swerve is the Y coordinate of the detour starting point of the detour traffic participant (m); start_another is the Y coordinate (m) of the other traffic participant when the detour conflict participant participates at the detour starting point; Transverse winding distance D lateral_away and longitudinal winding distance D long_away The calculation formula is shown in formula (3) and (4), the lateral maneuvering distance D lateral_back and longitudinal reversal distance D long_back The calculation method is shown in equations (5) and (6): D lateral_away =|X farthest_swerve -X start_swerve | (3) D long_away =|Y farthest_swerve -Y start_swerve | (4) D lateral_back =|X end_swerve -X farthest_swerve | (5) D long_back =|Y end_swerve -Y farthest_swerve | (6) Where, X farthest_swerve is the X coordinate of the detour traffic participant demarcation point (m); Y farthest_swerve is the Y coordinate of the detour traffic participant demarcation point (m); end_swerve is the X coordinate of the point where the detour traffic participant returns (m); Y end_swerve is the Y coordinate of the detour participant's return point (m), Lateral safety distance D lateral_safety And longitudinal safety distance D long_safety The calculation method is shown in equations (7) and (8): D lateral_safety =|X farthest_swerve -X farthest_another | (7) D long_safety =|Y farthest_swerve -Y farthest_another | (8) Where, X farthest_another is the X coordinate (m) of the other traffic participant when the detour traffic participant is at the demarcation point; Y farthest_another is the Y coordinate of the other traffic participant when the detour traffic participant is at the demarcation point (m); Transverse winding change ratio R lateral_away And the longitudinal winding change ratio R long_away The calculation method is as shown in formula (9) and (10), the lateral winding change ratio R lateral_back And the longitudinal winding change ratio R long_back The calculation method is as shown in formulas (11) and (12): Where, X change_awa is the X coordinate of the detour traffic participant's starting point (m); Y change_away is the Y coordinate of the detour traffic participant's starting point (m); X change_back is the X coordinate of the detour traffic participant's return to the change point (m); Y change_back is the Y coordinate of the detour participant's return to the change point (m), Security Zone S safetyzon The calculation method of is as follows:

7. The method for identifying traffic conflict behaviors at the entrance of a campus according to claim 6, characterized in that: The conflict density index in step 4 is the ratio of the product of the number of non-motor vehicles and pedestrians in the area and the average area occupied by non-motor vehicles and single pedestrians to the area of ​​the safety zone, as shown in the following formula: Where N i is the number of individuals of each type; S l is the average area of ​​each type of individuals; ρ conflict is the conflict density index; S safetyzone It is a security domain; Among them, ρ conflict When the value of exceeds the safety threshold, it means that the possibility of conflict increases, the traffic environment at the entrance of the campus is unsafe, and it is necessary to start a traffic safety warning and implement traffic diversion; ρ conflict When the value is lower than the safety threshold, it means that the possibility of conflict is reduced, the traffic environment at the entrance of the campus is safe, and there is no need to initiate a traffic safety warning.

Images