A real-time traffic operation safety evaluation method

Abstract

The application relates to a real-time traffic operation safety evaluation method, which comprises the following steps: collecting vehicle trajectory data, generating a vehicle position time sequence for each vehicle in an observation range; determining the speed and direction of each vehicle at the current time t based on the vehicle position time sequence; assuming that each vehicle continues to travel at the speed and direction at the time t for a time length T, and determining that a conflict exists between any two vehicles if the running trajectories of the two vehicles overlap within the time length T; calculating the modulus of the speed vector difference of any two vehicles that produce a conflict at the time t to obtain the size of the conflict between the two vehicles; and adding all the conflicts faced by any vehicle i in the observation range at the time t to obtain the operation safety state value of the vehicle i at the time t. The application initiatively defines and quantifies the conflict between vehicles, and the method can realize real-time and accurate evaluation of the traffic operation safety condition based on the real operation state of the vehicles.

Description

TECHNICAL FIELD

[0001] The present application relates to a real-time traffic operation safety evaluation method, in particular to a method for real-time evaluation of traffic operation safety based on vehicle trajectory data. BACKGROUND

[0002] At present, the methods for evaluating road traffic safety generally include the following categories. The first category is to assess the potential accident risk based on the statistics of the occurred traffic accidents. This method is to make a probabilistic prediction of driving risk according to the history of accident occurrence, which cannot reflect the running state of each vehicle at the micro level and cannot accurately reflect the current traffic safety situation based on the current running state of the vehicle. The second category is to assess the traffic operation safety based on the distribution of the detected vehicle speed. This method classifies the traffic safety level by detecting the speed of the vehicle and counting the number of vehicles and the time distribution in different speed value intervals, which can evaluate the current traffic safety situation to a certain extent, but can only realize the macro evaluation of the road section and cannot accurately reflect the running safety situation of the specific vehicle or the specific road section position. The third category is to assess the traffic safety based on the static traffic conflicts caused by the driving environment, such as the safety risk caused by the change of meteorological conditions or road alignment or the change of traffic operation safety caused by the lane changing of vehicles at the highway exit and entrance, etc. This method also belongs to the macro evaluation of the road section.

[0003] Chinese invention patent ZL202010902854.4 discloses a real-time traffic safety index dynamic comprehensive evaluation system and its construction method. A safety index model is established based on an accident risk prediction model. The data related to the people, vehicles and roads that affect traffic accidents (including historical traffic accident data, road basic attribute data, meteorological data, traffic signal cycle data, traffic flow data, vehicle performance data and driving behavior data) are collected as independent variables, and the expected number of accidents is taken as the dependent variable to establish an accident risk assessment and prediction model, and further to convert the expected number of accidents into a safety index. The core of this method is the accident occurrence probability, which is to evaluate the safety from the perspective of accident occurrence probability, and the space intensity and event granularity are relatively large, such as the probability of accident occurrence in a certain road section within the next half hour, but it cannot evaluate the traffic safety running state of the specific position of the road section.

[0004] Chinese invention patent application 202210032897.0 discloses a multi-dimensional comprehensive traffic safety index calculation method based on the influence of traffic safety. Four aspects of data, including the number of police reports, traffic index, in-transit quantity, and date period, are collected to generate four aspects of safety situation, and the final safety index is obtained by weighted average of the four aspects of safety situation to evaluate the traffic safety situation. This method can only perform macro evaluation and cannot give real-time and accurate evaluation results.

[0005] In general, the existing road traffic safety evaluation methods can only provide macroscopic and qualitative evaluation at the road section level or predict the risk probability of traffic accidents, and cannot realize real-time microscopic and accurate evaluation of specific road sections, lanes and specific vehicles.

[0006] In recent years, with the development of communication technology and intelligent terminal technology, the types of traffic data become more and more rich, the quality of data becomes more and more accurate, and the timeliness of data becomes higher and higher, which provides the possibility for traffic operation safety evaluation methods, but how to use these data has not formed an effective method.

[0007] The full-coverage high-precision vehicle trajectory data can be obtained by a roadside monitoring device, such as a video monitoring device or a radar device. Based on signal and data processing technology, the roadside monitoring device can identify and track multiple vehicle targets within its observation range through the built-in algorithm, and determine the position of each vehicle target in real time, thereby providing high-precision vehicle trajectory data.

[0008] Internet of Vehicles and automatic driving are important development directions of future traffic, and also put forward new requirements for road traffic safety evaluation. Microscopic and accurate traffic operation safety evaluation and real-time monitoring are the inevitable trend of high-quality development of road safety.

[0009] Therefore, it is necessary to design a method for real-time quantitative evaluation of traffic operation safety state based on high-precision vehicle trajectory data, which fully excavates the traffic operation safety information contained in the vehicle trajectory data, and scientifically and accurately judges the operation safety state of the vehicle and the traffic operation safety situation of the road section. SUMMARY

[0010] In order to overcome the defects of the prior art, the present application provides a real-time traffic operation safety evaluation method, comprising the following steps:

[0011] Step S01: collecting vehicle trajectory data, and generating a vehicle position time sequence for each vehicle in the observation range:

[0012] (X, Y) 0 = [(x t , y t ), (x t-Δt , y t-Δt ), (x t-2Δt , y t-2Δt ), (x t-3Δt , y t-3Δt ), …]

[0013] Wherein, (x t , y t ) represents the position of the vehicle at time t, (x t-Δt , y t-Δt(x) represents the position of the vehicle at time t-Δt. t-2Δt y t-2Δt (x) represents the position of the vehicle at time t-2Δt. t-3Δt y t-3Δt The position of the vehicle at time t-3Δt is represented by x and y, which represent the horizontal and vertical coordinates of the vehicle's position, respectively. The subscripts t, t-Δt, t-2Δt, and t-3Δt indicate the specific time when the vehicle's position information was collected.

[0014] Step S02: Based on the vehicle position time series, determine the speed and direction of each vehicle at the current time t. The speed of a vehicle at time t is calculated using the following formula:

[0015]

[0016] in, It is the velocity component of the vehicle along the x-axis at time t; It represents the vehicle's velocity component along the y-axis at time t; Δt represents the time interval for collecting vehicle position data; v t Let be the vehicle's velocity scalar at time t;

[0017] Step S03: Assuming each vehicle maintains its speed and direction at time t and continues to travel for a time length T, if the travel trajectories of any two vehicles overlap within the time length T, then it is determined that there is a conflict between the two vehicles.

[0018] Step S04: Calculate the magnitude of the velocity vector difference between any two vehicles that caused the conflict at time t, to obtain the magnitude of the conflict between the two vehicles, i.e.:

[0019]

[0020] in, This indicates the conflict between vehicle i and vehicle j at time t. Let be the velocity vector of vehicle i at time t. Let be the velocity vector of vehicle j at time t. Let be the velocity component of vehicle i along the x-axis at time t. Let be the velocity component of vehicle j along the x-axis at time t. Let be the velocity component of vehicle i along the y-axis at time t. Let be the velocity component of vehicle j along the y-axis at time t;

[0021] Step S05: For any vehicle i within the observation range, sum all the conflicts faced by vehicle i at time t to obtain the operational safety state value of vehicle i at time t, i.e.:

[0022]

[0023] wherein, is the running safety state of the vehicle i at time t, is the conflict between the vehicle i and any other vehicle j within the observation range at time t.

[0024] Preferably, the conflicts between all vehicles within the observation range at time t are added to obtain the running safety state value of the observation range at time t, i.e.,

[0025]

[0026] wherein, E t is the running safety state of the observation range at time t.

[0027] Preferably, the time length T is 0.5-1.5 seconds, more preferably, the time length T is 1 second.

[0028] Preferably, the step S01 further comprises the following steps of correcting the vehicle position time sequence:

[0029] Step S11: based on the vehicle position time sequence before correction, the magnitude of the acceleration of the vehicle at each time is calculated to generate a corresponding acceleration time sequence,

[0030]

[0031] wherein, a t-Δt is the magnitude of the acceleration of the vehicle at time t-Δt, is the magnitude of the acceleration of the vehicle in x direction at time t-Δt, is the magnitude of the acceleration of the vehicle in y direction at time t-Δt, a t-2Δt is the magnitude of the acceleration of the vehicle at time t-2Δt, is the magnitude of the acceleration of the vehicle in x direction at time t-2Δt, is the magnitude of the acceleration of the vehicle in y direction at time t-2Δt, a t-3Δt is the magnitude of the acceleration of the vehicle at time t-3Δt, is the magnitude of the acceleration of the vehicle in x direction at time t-3Δt, is the magnitude of the acceleration of the vehicle in y direction at time t-3Δt, Δt represents the time interval for collecting the vehicle position;

[0032] The magnitude of the acceleration of the vehicle at time t is calculated by the following formula:

[0033]

[0034]

[0035] wherein, a is the magnitude of the acceleration of the vehicle in the x direction at time t, a is the magnitude of the velocity of the vehicle in the x direction at time t, a is the magnitude of the velocity of the vehicle in the x direction at time t-Δt, a is the magnitude of the acceleration of the vehicle in the y direction at time t, a is the magnitude of the velocity of the vehicle in the y direction at time t, a is the magnitude of the velocity of the vehicle in the y direction at time t-Δt, t a is the magnitude of the acceleration of the vehicle at time t;

[0036] Step S12: if the magnitude of the acceleration of the vehicle at any time exceeds the acceleration range [a min , a max ], it is determined that the data of the vehicle at the time is abnormal;

[0037] Step S13: the data of the abnormal time is removed from the acceleration time sequence and the position time sequence of the vehicle, and a new position time sequence (X, Y)' and a new acceleration time sequence A' are generated.

[0038] Preferably, in the new acceleration time sequence A' generated in step S13, the data of the abnormal time is corrected by taking the arithmetic mean of the data of the previous and subsequent times, a corrected acceleration time sequence A" is generated, and a corrected position time sequence (X, Y)" is generated based on the corrected acceleration time sequence A".

[0039] Preferably, in the new acceleration time sequence A' generated in step S13, the data of the abnormal time is corrected by taking the arithmetic mean of the data of the previous and subsequent times, a corrected acceleration time sequence A" is generated, and a corrected position time sequence (X, Y)" is generated based on the corrected acceleration time sequence A".

[0040] Preferably, a min is -8 m / s 2 ~ -3 m / s 2 , a max is 20 m / s 2 ~ 30 m / s 2 ; more preferably, the acceleration range [a min , a max ] is determined according to the type of the vehicle.

[0041] Preferably, the method further comprises:

[0042] Step S06: presenting the running safety state of the observation range in the form of a heat map corresponding to the position of the observation range.

[0043] Preferably, the observation range is a continuous road section.

[0044] The present application defines and quantifies the conflict between vehicles, which provides a basis for scientifically evaluating the traffic operation safety condition. Through the traffic operation safety evaluation method of the present application, the operation safety condition of a road section and a single vehicle can be evaluated in real time and accurately based on the real operation state of the vehicle, so that quantitative evaluation and analysis of the traffic operation safety condition can be realized, thereby providing a scientific and reliable basis for the control of a specific vehicle and the management of traffic flow.

[0045] The application value of the present application lies in improving the traffic management and control level in terms of immediacy (real time) and long-term (long-term). In terms of immediacy, the present technology can accurately and timely understand the traffic operation state from the bottom up based on micro-vehicle trajectory, immediately determine the safety state and resilience of traffic operation, guide managers to comprehensively use control measures, and develop reasonable and efficient traffic management actions, thereby improving the efficiency of traffic management and control. In terms of long-term, the present technology can establish an independent historical database for each observation range, analyze and summarize the traffic operation characteristics in the observation range, support management decision-makers to develop targeted control strategies (such as the planning of emergency sites), optimize the allocation of emergency response resources, reduce the emergency response time, and reasonably arrange control facilities, thereby improving the ability of traffic management and control.

[0046] The method of the present application has the following advantages:

[0047] First, the method has real-time, which can reflect the real-time traffic operation safety condition;

[0048] Second, the evaluation result is in the two-dimensional space of the road surface, realizing lane-level evaluation and breaking the past one-dimensional linear safety evaluation;

[0049] Third, a surrounding vehicle driving vector environment is proposed to determine the operation safety state of the vehicle;

[0050] Fourth, the evaluation logic of the method is from the bottom up, based on a micro evaluation method, and the application scope of the evaluation result is wide, which can realize evaluation of different granularities from micro to macro;

[0051] Fifth, the data source used is available at present and in the future.

[0052] On one hand, the method is a bottom-up discrimination method, which can accurately determine the running safety state of the road, and extends the running safety to the traffic safety in two-dimensional space (two dimensions of the road), breaking the traffic safety in one-dimensional space in the past, so that the traffic running safety state is more flexible in analysis granularity, and can be microscopic or macroscopic; on the other hand, the discrimination of the safety state is not affected by the traffic volume, and is suitable for road sections with free flow and congested traffic flow, thereby providing effective support for fine traffic control; in addition, the method has the ability to continuously expand the data set, and has better scalability.

[0053] Grabbing the real safety running state of the traffic flow, rather than just the safety index, is an important development direction of traffic safety evaluation, and has high data value, and the analysis of the traffic flow safety characteristics of the specific position of the road section based on the accumulated data is an important content of the fine traffic running safety evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0054] Figure 1 A schematic block diagram of a traffic running safety evaluation method according to a preferred embodiment of the application is shown;

[0055] Figure 2 A schematic diagram of vehicle conflicts used to discriminate the traffic running safety state according to a preferred embodiment of the application is shown;

[0056] Figure 3 A schematic diagram of determining the traffic running safety state resilience and influence time according to a preferred embodiment of the application is shown. DETAILED DESCRIPTION

[0057] The overall concept of the application is to define the concept and quantification method of vehicle conflicts, emphasize the instantaneous relationship of vehicles with other vehicles in the movement process, can describe the instantaneous running safety state of a single vehicle movement, and can also be extended to the running safety state of all vehicles in the movement process on a specific road section; based on the real-time driving trajectory data of all vehicles in the observation range, the fixed vector of vehicle movement is analyzed, the instantaneous acceleration, speed and driving direction of each vehicle are determined, the mutual relationship of vehicles in the movement process is judged, the instantaneous running safety state of each vehicle in the observation range is calculated, and finally the instantaneous safety state of the traffic running in the observation range is formed.

[0058] Based on such concept, the traffic running safety state of the observation range at the current time can be obtained in real time according to the running safety state of all vehicles in the observation range at the current time, and the current decision support for real-time traffic control behavior is provided. In addition, the traffic running safety state data at different times can also be stored to form a traffic running safety state database. Further analysis of these data can also summarize the traffic running rules of the observation range, thereby providing a basis for the formulation of long-term traffic control strategies.

[0059] In some embodiments of the present application, the vehicle trajectory data is obtained by a roadside monitoring device. The roadside monitoring device can identify and track all vehicles entering its observation range, and collect the position information of each vehicle at a frequency of no less than 10 Hz, thereby forming high-precision vehicle trajectory data. The roadside monitoring device includes but is not limited to a video monitoring device or a radar device. Those skilled in the art can understand that high-precision vehicle trajectory data can be provided by detection devices arranged outside the vehicle (such as roadside or other positions on the road) or by devices arranged on the vehicle (such as high-precision GNSS positioning systems). It can be understood that the observation ranges of multiple roadside monitoring devices monitoring continuous road sections can be combined to form a larger observation range, and therefore, in the present application, the observation range refers to a continuous road section and is not limited to the observation range of a single roadside monitoring device.

[0060] Accuracy requirements of vehicle trajectory data: the data collection frequency is no less than 10 Hz, which means that the position information of the vehicle is collected at most every 0.1 seconds. If the data collection frequency is 10 Hz, when the vehicle travels at a speed in the interval [60km / h, 100km / h], the distance between adjacent trajectory points is [1.67m, 2.78m]. Under such data collection accuracy, the vehicle trajectory obtained can have sufficient accuracy.

[0061] The most basic vehicle trajectory data obtained by the roadside monitoring device at least includes vehicle ID, time and position. The vehicle ID is an identification assigned by the roadside monitoring device to each vehicle entering its observation range. The time refers to the specific time when the vehicle position information is collected. The position refers to the position coordinates of the vehicle at the time of collection, for example, the geodetic coordinate system with the roadside monitoring device as the origin can be used as the coordinate system for identifying the position of the vehicle.

[0062] For any vehicle in the observation range, its vehicle trajectory can be represented as the following vehicle position time sequence:

[0063] (X, Y) 0 = [(x t , y t ), (x t-Δt , y t-Δt ), (x t-2Δt , y t-2Δt ), (x t-3Δt , y t-3Δt ), …]

[0064] wherein (x t , y t ) represents the position of the vehicle at time t, (x t-Δt , y t-Δt(x) represents the position of the vehicle at time t-Δt. t-2Δt y t-2Δt (x) represents the position of the vehicle at time t-2Δt. t-3Δt y t-3Δt The value represents the vehicle's position at time t-3Δt, where x and y represent the horizontal and vertical coordinates of the vehicle's position, respectively. The subscripts t, t-Δt, t-2Δt, and t-3Δt indicate the specific time when the vehicle's position information was collected.

[0065] The vehicle's speed and direction can be further calculated based on the vehicle trajectory data. The vehicle's speed at time t can be calculated using the following formula:

[0066]

[0067]

[0068] in, It is the vehicle's velocity along the x-axis at time t, that is, the velocity component along the x-axis; Δt represents the vehicle's velocity along the y-axis at time t, i.e., the velocity component along the y-axis; Δt represents the time interval for collecting vehicle position data; vt is the vehicle's velocity scalar at time t.

[0069] Therefore, the vehicle's running state at time t can be used as the vehicle's position (x). t y t The velocity vector starting from the origin is a fixed vector. This is represented by a fixed velocity vector. This vector contains basic information about the vehicle's motion at time t: position, velocity magnitude, and direction of travel.

[0070] The collected vehicle trajectory data may contain abnormal data points. To avoid the abnormal data interfering with the judgment of the vehicle's motion status, the abnormal data in the vehicle trajectory data can be cleaned or corrected first.

[0071] For example, the vehicle's maximum instantaneous acceleration and maximum braking deceleration can be used as standards for the vehicle's acceleration range [a] min a max This identifies abnormal data in vehicle trajectory data. Depending on vehicle performance, the maximum instantaneous acceleration of a high-performance sports car is 20-30 m / s². 2 The maximum instantaneous acceleration of a typical family car is usually between 5-10 m / s². 2 During emergency braking, the maximum deceleration of a car is typically 7.5-8 m / s². 2 During normal braking, the average deceleration of a car should be 3-4 m / s². 2 Therefore, it can be assumed that the vehicle acceleration is within the range [-8m / s²]. 2 20m / s2 ] is normal, i.e. a min = -8 m / s 2 , a max = 20 m / s 2 ; accelerations outside this range are considered abnormal. In some embodiments of the application, a min = -8 m / s 2 ~ -3 m / s 2 , a max = 20 m / s 2 ~ 30 m / s 2 . In practice, the acceleration range [a min , a max ] can be different for different types of vehicles, i.e. the values of a min and a max may be determined according to the vehicle type.

[0072] The magnitude of the acceleration of a vehicle at time t is calculated by the following equation:

[0073]

[0074] wherein is the magnitude of the acceleration of the vehicle in the x direction at time t, is the magnitude of the velocity of the vehicle in the x direction at time t, is the magnitude of the velocity of the vehicle in the x direction at time t-Δt, is the magnitude of the acceleration of the vehicle in the y direction at time t, is the magnitude of the velocity of the vehicle in the y direction at time t, is the magnitude of the velocity of the vehicle in the y direction at time t-Δt, and t is the magnitude of the acceleration of the vehicle at time t, and Δt represents the time interval at which the vehicle position is collected.

[0075] For any vehicle within the observation range, based on its time series of vehicle positions, the magnitude of the acceleration of the vehicle at each time can be calculated, thereby generating a corresponding time series of accelerations,

[0076]

[0077] wherein a t-Δt is the magnitude of the acceleration of the vehicle at time t-Δt, is the magnitude of the acceleration of the vehicle in the x direction at time t-Δt, is the magnitude of the acceleration of the vehicle in the y direction at time t-Δt, and t-2Δt is the magnitude of the acceleration of the vehicle at time t-2Δt, is the magnitude of the acceleration of the vehicle in the x direction at time t-2Δt, a is the size of the acceleration of the vehicle in the y direction at time t-2Δt t-3Δt a is the size of the acceleration of the vehicle at time t-3Δt a is the size of the acceleration of the vehicle in the x direction at time t-3Δt a is the size of the acceleration of the vehicle in the y direction at time t-3Δt

[0078] If the size of the acceleration of the vehicle at any time exceeds the acceleration range [a min , a max ], it can be determined that the data of the vehicle at the time is abnormal. The data corresponding to the time is removed from the acceleration time sequence and the position time sequence of the vehicle, thereby generating a new position time sequence (X, Y)' and an acceleration time sequence A'.

[0079] In some embodiments of the present application, the speed and the running direction of the vehicle can be determined based on the new position time sequence (X, Y)', and based on this, whether there is a conflict between the vehicles, the size of the conflict, and thus the running safety state of the vehicle and the observation range can be determined.

[0080] In some embodiments of the present application, the data at the time of data abnormality can be corrected, and a corrected acceleration time sequence A" and a corrected position time sequence (X, Y)" are generated, which are used as the basis for determining the conflict and the running safety state of the vehicle and the observation range.

[0081] In some embodiments, the correction method is to take the arithmetic mean of the data at the time before and after the time of data abnormality in the new acceleration time sequence A', and replace the data at the time of data abnormality with the arithmetic mean, thereby generating the corrected acceleration time sequence A"; and the position data at the time of data abnormality is calculated using the formula

[0082] and , thereby generating the corrected position time sequence (X, Y)". Using the correction method of taking the arithmetic mean of the data at the time before and after the time of data abnormality, compared with simply removing the abnormal data, the missing of part of the data at the time in the acceleration time sequence and the position time sequence can be avoided, and the complexity of subsequent calculation can be effectively reduced; at the same time, the operation load caused by the correction process itself is small, so that the interference caused by the abnormal data can be avoided with small operation resources.

[0083] In other embodiments, the correction method is to correct the data at the time of data abnormality by using a cubic exponential smoothing method, that is, to perform cubic exponential smoothing on the new acceleration time sequence A', to generate a corrected acceleration time sequence A", and then to calculate the position data at the time of data abnormality using the formula

[0084] and The position data at the time of data anomaly is calculated, thereby generating a corrected position time sequence (X, Y)''.

[0085] Generally, the abnormal data can be corrected by using cubic exponential smoothing, thereby effectively reducing the error of subsequent calculation. For example, the acceleration time sequence can be cubic exponential smoothed (taking the x direction as an example) in the following manner:

[0086]

[0087] is the smoothed value of the first smoothing time t, is the smoothed value of the first smoothing time t-1; is the smoothed value of the second smoothing time t, is the smoothed value of the second smoothing time t-1; is the smoothed value of the third smoothing time t, is the smoothed value of the third smoothing time t-1; a is a weighting coefficient, the value range is [0, 1], the smaller the value, the stronger the smoothing effect, and the slower the actual data change is reflected; x t is the actual observation value of time t.

[0088] How to determine the vehicle running safety state and the traffic running safety state based on the vehicle trajectory data is further described below.

[0089] For a specific vehicle, the vehicle running safety state at a certain time is directly related to the traffic running environment in which it is located, which is specifically manifested as whether the vehicle will collide with surrounding vehicles or fixed objects while maintaining the existing motion state. If a collision occurs, it is considered that there is a "conflict". The present application quantifies this "conflict" to represent the vehicle running safety state. Similarly, the traffic running safety state of the observation range is determined by the running safety states of all vehicles in the observation range, which can be represented by the sum of "conflicts" of all vehicles and the distribution of vehicle running safety states.

[0090] First, based on the current speed fixed vector of each vehicle, if the vehicle trajectories overlap in space within the effective reaction time of the driver, it is considered that a conflict occurs between the vehicles; the magnitude of the conflict is determined by the speed of the vehicles and the angle of collision. The effective reaction time of the driver is the time used by the driver to accurately perceive the driving environment and make the correct operation. When the driver faces a sudden situation, the effective reaction time is the most direct indicator related to driving safety. Generally, the effective reaction time of the driver is between 0.3-1 seconds. In order to make the conflict determination consistent with the driving behavior of most drivers, in the preferred embodiment of the present application, the maximum value of the effective reaction time of the driver is used as the time threshold in the conflict definition. That is, within 1 second (the maximum effective reaction time of the driver) after t time, if the trajectories of the two vehicles overlap in space, there is a conflict between the two vehicles at t time.

[0091] As shown in FIG. 3, for example, on a one-way three-lane road, the vehicle 3 driving in the third lane changes the driving direction at t time to change lanes, which will cause a conflict with the vehicle 2 normally driving in the second lane within 1 second in the future; the vehicle 1 and the vehicle 2 driving in the second lane will cause a conflict within 1 second in the future due to the large speed difference at t time. That is, if each vehicle maintains the motion state at t time, the vehicle 2 will encounter two conflicts within 1 second in the future. Figure 2

[0092] The severity of the conflict or the magnitude of the conflict is directly related to the speed of the vehicle, so the modulus of the speed vector difference of the two vehicles causing the conflict can be used as the magnitude of the conflict, that is:

[0093]

[0094] wherein, represents the conflict between the vehicle i and the vehicle j at t time, is the speed fixed vector of the vehicle i at t time, is the speed fixed vector of the vehicle j at t time, is the speed component of the vehicle i along the x-axis at t time, is the speed component of the vehicle j along the x-axis at t time, is the speed component of the vehicle i along the y-axis at t time, is the speed component of the vehicle j along the y-axis at t time.

[0095] For a specific vehicle, the running safety state of the vehicle is the sum of the conflicts faced by the vehicle. Therefore, the running safety state of any vehicle i in the observation range at t time can be represented by calculating the sum of the conflicts between the vehicle i and all other vehicles j in the observation range at t time, that is:

[0096] ​

[0097] wherein, is the running safety state of vehicle i at time t, is the conflict between vehicle i and any other vehicle j within the observation range at time t.

[0098] Further, the running safety state of the observation range at time t is the sum of the conflicts of all vehicles within the observation range at time t, i.e.,

[0099]

[0100] wherein, E t is the running safety state of the observation range at time t.

[0101] In the above manner, the running safety state of a single vehicle and the traffic running safety state of the observation range can be quantified as a specific conflict value according to the motion of the vehicle on the road section; the unit thereof is the same as that of the speed, i.e., km / h.

[0102] In the present application, the conflict has a clear physical meaning: the size of the velocity vector difference of the vehicles that produce conflicts (or are likely to collide) under the current traffic environment. The physical meaning can express the collision intensity to some extent, and presents the size of the danger of the vehicle. At the same time, under this definition, the running safety states of vehicles can be directly compared with each other without the need for standardization.

[0103] Since the running safety states of vehicles or observation ranges can be directly compared in size, they can be directly presented in the form of a heat map, i.e., different colors are used to represent different values or ranges of values of the running safety states of vehicles or observation ranges. For example, based on the corresponding relationship of positions, the running safety state heat map can be directly superimposed on the two-dimensional plan view of the road to form a traffic running safety state distribution map of the observation area.

[0104] With the movement of the vehicle, the traffic operation safety state and the observation range operation safety state also change, and these changes can be reflected in the traffic operation safety state distribution map in real time. In the preferred embodiment of the application, the operation safety state is divided into different levels according to the value, and different colors are respectively corresponded, for example, the operation safety state value less than 20 km / h is represented by light green; between 20 km / h and 40 km / h, represented by yellow; between 40 km / h and 60 km / h, represented by orange; between 60 km / h and 80 km / h, represented by red; more than 80 km / h, represented by dark brown. In this way, when the operation safety state value changes in a small range, the corresponding level will not change, thereby shielding the fluctuations of the operation safety state in a small range and avoiding the traffic operation safety state distribution map from changing too frequently. At the same time, one or more safety thresholds can be set in advance, and when the vehicle operation safety state or the observation range operation safety state exceeds a certain safety threshold, the corresponding position or area is presented in a flashing or other easily noticeable way, thereby realizing the alarm function for a specific traffic operation safety state.

[0105] By the above-mentioned method of determining the conflict quantity and calculating the conflict value size, the total amount of vehicle conflicts in the observation range at the current time can be used as an indicator of the safety state of traffic operation, and the traffic operation safety state of the observation range can be output in real time. Therefore, the manager can make an immediate and effective response, thereby improving the efficiency of traffic control.

[0106] In addition, the observed vehicle trajectory data can be structured and processed to form a standardized data set. For example, the smallest data unit of the data set can include: vehicle ID, time t, horizontal coordinate of vehicle position, vertical coordinate of vehicle position, lateral speed, longitudinal speed, lateral acceleration, longitudinal acceleration, conflict quantity and operation safety state. Such a data unit reflects the vehicle operation state of a specific vehicle at a specific time. With the growth of observation time, the accumulated vehicle trajectory data also gradually increases, and the data set continuously increases. Through such a data set, not only can the operation safety state of the vehicle at each time be reproduced, but also the traffic operation safety state of the entire observation range or road section can be reproduced, thereby supporting the analysis and summary of the characteristics and laws of traffic operation, and effectively supporting the improvement of traffic control efficiency.

[0107] (1) According to the historical data set of the observation area, the traffic flow is disturbed by such a scale at the current time, and can recover in a short time and will not appear a state of too high conflict level in the recovery process, which will not cause long-term congestion or safety hazards to the traffic, that is, the traffic system can be relatively stable self-recovery in a period of time (such as short-term slow-down of road section caused by vehicle emergency braking). For this kind of problem, the manager can not take management action, but monitor and let the traffic system self-recover;

[0108] (2) The traffic running state can recover in a short time, but there will be a state of too high conflict level in the recovery process, that is, the traffic system can recover in a short time, but the safety risk is high during the recovery process (such as vehicle breakdown in the middle of the road). For this kind of problem, the manager should first solve the problem of reducing and controlling the conflict level, such as information reminder on variable information board, dynamic adjustment of road speed limit, arrangement of personnel and emergency vehicles for on-site command and rescue, and other measures;

[0109] (3) The traffic running state cannot recover in a short time, and the number of affected vehicles exceeds the acceptable range, that is, the traffic system has low resilience and needs a long time to recover (such as traffic congestion caused by multiple vehicle collisions under heavy traffic flow). For this kind of problem, the management should control the traffic from three aspects: one is the removal of the disturbance source, the second is the control of traffic flow, and the third is the traffic diversion. Comprehensive use of traffic control measures can reduce the influence time and the number of affected vehicles.

[0110] Through the vehicle trajectory data set, not only can the traffic running safety state change with time be reproduced, but also the spatial distribution of the traffic running safety state can be obtained.

[0111] For a specific road section or observation range, the trajectory data of all vehicles passing through in a certain period of time forms a two-dimensional matrix corresponding to the road plane space where the specific road section or observation range is located. By superimposing the running safety state values of different vehicles at the same matrix position, a traffic running safety state spatial distribution matrix (referred to as matrix B) of the specific road section or observation range in the time period can be formed.

[0112] On the basis of matrix B, the traffic running safety state at any position in the road section or observation range can be obtained, which is convenient for summarizing the characteristics of the traffic running safety state of the road section or observation range, such as conflict high-occurrence points and areas, conflict high-occurrence time, and different levels of conflict characteristics. According to the conflict characteristics shown by the data, the manager can plan the emergency response, optimize the resource layout and quantity allocation from the perspective of minimizing the total time of personnel and emergency rescue equipment arriving at the scene after a specific disturbance event occurs, and shorten the emergency response time after the specific disturbance event occurs, so as to improve the response ability of the manager, that is, improve the traffic control ability.

[0113] It can be seen that, by the present application, the traffic control level can be improved from both short-term and long-term aspects. In the short-term aspect, instant information output is provided to managers to provide instant and effective data to support the decision of current traffic control behavior, and the efficiency is improved in the reflection time of the event and the control of the event. In the long-term aspect, through the continuous accumulation of observation area data set, the traffic operation characteristics gradually become clear and obvious, where the events occur frequently, which type of event occurs frequently, and where the type of event occurs frequently, etc. gradually highlight. According to the characteristics, the layout and quantity of emergency resources are optimized from the perspective of reducing emergency response time.

[0114] Those skilled in the art should recognize that the embodiments described herein are exemplary and non-limiting, and the present application is not limited to the specific embodiments described above.

Claims

1. A method for real-time traffic operational safety evaluation, comprising the following steps: Step S01: collecting vehicle trajectory data, and generating a vehicle position time series for each vehicle in the observation range; (X,Y) 0 = [(x t ,y t ), (x t-Δt ,y t-Δt ), (x t-2Δt ,y t-2Δt ), (x t-3Δt ,y t-3Δt ),...] wherein (x t ,y t ) represents the position of the vehicle at time t, (x t-Δt ,y t-Δt ) represents the position of the vehicle at time t-Δt, (x t-2Δt ,y t-2Δt ) represents the position of the vehicle at time t-2Δt, (x t-3Δt ,y t-3Δt ) represents the position of the vehicle at time t-3Δt, x and y are used to represent the horizontal coordinate and the vertical coordinate of the vehicle position respectively, and the subscripts t, t-Δt, t-2Δt, t-3Δt represent the specific time at which the vehicle position information is collected; Step S02: determining the speed and direction of each vehicle at the current time t based on the vehicle position time series, and calculating the speed of the vehicle at time t by the following formula: wherein, is the velocity component of the vehicle along the x-axis at time t; is the velocity component of the vehicle along the y-axis at time t; Δt represents the time interval at which the vehicle position is collected; t is the velocity scalar of the vehicle at time t; Step S03: assuming that each vehicle continues to travel at the speed and direction at time t for a time length T, and determining that there is a conflict between any two vehicles if their trajectories overlap within the time length T; Step S04: calculating the modulus of the speed vector difference between any two vehicles that have a conflict at time t, to obtain the size of the conflict between the two vehicles, i.e. wherein, denotes a conflict between vehicle i and vehicle j at time t, is the velocity vector of vehicle i at time t, is the velocity vector of vehicle j at time t, is the velocity component of vehicle i along the x-axis at time t, is the velocity component of vehicle j along the x-axis at time t, is the velocity component of vehicle i along the y-axis at time t, is the velocity component of vehicle j along the y-axis at time t; Step S05: for any vehicle i in the observation range, adding all the conflicts faced by vehicle i at time t to obtain the operational safety state value of vehicle i at time t, i.e. wherein, is the running safety state of vehicle i at time t, is the conflict between vehicle i and any other vehicle j within the observation range at time t.

2. The method of claim 1, further comprising: Step S06: adding all the conflicts between all the vehicles in the observation range at time t to obtain the operational safety state value of the observation range at time t, i.e. E t is the observed running safety state at time t.

3. The method of claim 1 or 2, wherein, The time length T is 0.5-1.5 seconds.

4. The method of claim 3, wherein, The time length T is 1 second.

5. The method of claim 1 or 2, wherein, Step S01 further comprises the following steps of modifying the vehicle position time series: Step S11: calculating the magnitude of the acceleration of the vehicle at each time based on the vehicle position time series before modification, to generate a corresponding acceleration time series, wherein a t-Δt is the magnitude of the acceleration of the vehicle at the time t-Δt, is the magnitude of the acceleration of the vehicle in the x direction at the time t-Δt, is the magnitude of the acceleration of the vehicle in the y direction at the time t-Δt, a t-2Δt is the magnitude of the acceleration of the vehicle at the time t-2Δt, is the magnitude of the acceleration of the vehicle in the x direction at the time t-2Δt, is the magnitude of the acceleration of the vehicle in the y direction at the time t-2Δt, a t-3Δt is the magnitude of the acceleration of the vehicle at the time t-3Δt, is the magnitude of the acceleration of the vehicle in the x direction at the time t-3Δt, is the magnitude of the acceleration of the vehicle in the y direction at the time t-3Δt, Δt represents the time interval for collecting the position of the vehicle; The magnitude of the acceleration of the vehicle at time t is calculated by the following formula: wherein is the magnitude of the acceleration of the vehicle in the x-direction at time t, is the velocity component of the vehicle along the x-axis at time t, is the velocity component of the vehicle along the x-axis at time t - Δt, is the magnitude of the acceleration of the vehicle in the y-direction at time t, is the velocity component of the vehicle along the y-axis at time t, is the velocity component of the vehicle along the y-axis at time t - Δt, a t is the magnitude of the acceleration of the vehicle at time t; Step S12: If the acceleration magnitude of the vehicle at any time exceeds the acceleration range [a min ,a max ], then it is determined that the data of the vehicle at that time is abnormal. Step S13: removing the data at the data abnormal time from the acceleration time series and the position time series of the vehicle, to generate a new position time series (X, Y)' and a new acceleration time series A'.

6. The method of claim 5, wherein, In the new acceleration time series A' generated in step S13, the data at the data abnormal time is modified by taking the arithmetic average of the data at the previous and subsequent times, to generate a modified acceleration time series A'', and a modified position time series (X, Y)'' is generated based on the modified acceleration time series A''.

7. The method of claim 5, wherein, In the new acceleration time series A' generated in step S13, the data at the data abnormal time is modified by taking the three-exponential smoothing method, to generate a modified acceleration time series A'', and a modified position time series (X, Y)'' is generated based on the modified acceleration time series A''.

8. The method of claim 5, wherein, a min -8 m / s 2 ~ -3 m / s 2 , a max 20 m / s 2 ~ 30 m / s 2 .

9. The method of claim 5, wherein, The acceleration range [a min ,a max ] is determined depending on the vehicle type. 10.The method of claim 1 or 2, further comprising: Step S06: presenting the operational safety state of the observation range in the form of a heat map corresponding to the position of the observation range.

Images