Collision accident detection method based on traffic flow stability and vehicle running state

Abstract

The application provides a collision accident detection method based on traffic flow stability and vehicle running state, which comprises the following steps: comprehensively considering the road friction coefficient μ t and the accident rate p t of a target road section at a current time t, adopting a pre-constructed road traffic flow stability model to estimate the traffic flow stability l t of the target road section at the current time t, and adopting a vehicle collision accident detection model to detect the risk of a vehicle i having a collision accident at the current time t. The application obtains the road vehicle running state information under adverse weather conditions by using a road along-line meteorological environment detection device and a video recognition device, combines the historical accident data and the road surface state information of the road section, fuses the road traffic and the road surface environment influence, introduces the road section traffic flow stability concept, fuses the vehicle running state information, realizes the detection of the vehicle collision accident, and has the advantage of high accuracy of vehicle collision accident detection.

Description

Technical Field

[0001] This invention relates to the field of traffic safety technology, specifically to a collision accident detection method based on traffic flow stability and vehicle operating status. Background Technology

[0002] With the rapid development of society and the economy, the number of vehicles has been increasing. While the increase in the number of cars has brought convenience to people's lives, it has also brought about traffic safety problems, especially in road sections affected by snow and ice disasters, where vehicle collisions are frequent and cause great harm to people's lives and property. Based on this, a large amount of research has been conducted on vehicle collision accidents in existing technologies, and certain results have been achieved.

[0003] Chinese patent application CN 114582132 B discloses a vehicle collision detection and early warning system and method based on machine vision. The system includes a necessary information acquisition module, a database, a collision accident analysis module, a machine vision perception module, an image processing module, and a vehicle collision early warning module. By employing this system, vehicle collisions are detected and warned through machine vision perception. The system analyzes and processes the images, pre-determining whether target obstacles in the images are occluded. It also proactively interacts with occluded vehicles in invalid areas, maintaining real-time data perception of occluded vehicles. This solves the problem in existing technologies where vehicles are occluded and obstacles cannot be detected or traffic conditions cannot be predicted, thus improving the safety of occluded vehicles.

[0004] The aforementioned invention patents utilize advanced perception and recognition technologies to detect and warn of vehicle collisions from the perspective of the driving vehicle itself. They mainly assess, detect, and warn of vehicle operation status at the micro level, without considering the influence of macro traffic flow on the road vehicle operation status. Therefore, they are insufficient in terms of accuracy and applicability. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a collision accident detection method based on traffic flow stability and vehicle operating status, which can effectively solve the above problems.

[0006] The technical solution adopted in this invention is as follows:

[0007] This invention provides a collision accident detection method based on traffic flow stability and vehicle operating status, comprising the following steps:

[0008] Step S1: For vehicle i entering the target road segment, acquire in real time the target road segment environmental information, target road segment terrain feature information, target road segment traffic feature information, vehicle type of vehicle i, and vehicle i's driving information at the current time t; wherein, the driving information of vehicle i includes the acceleration a of vehicle i along the main line. x,t(i), the lateral acceleration of vehicle i is a y,t (i), the longitudinal distance ΔS between vehicle i and the vehicle in front. x,t (i) and the lateral distance ΔS between vehicle i and the adjacent lanes on the left and right sides. y,t (i);

[0009] Step S2: Using a pre-constructed road segment slipperiness characterization function, and based on the target road segment environmental information, estimate the road segment friction coefficient μ of the target road segment that affects vehicle operation safety at the current time t. t ;

[0010] Step S3: Using a pre-constructed road segment accident rate model, based on the target road segment environmental information, the target road segment terrain feature information, and the target road segment traffic feature information, the accident rate p of the target road segment at the current time t is estimated. t ;

[0011] Step S4, comprehensively consider the road friction coefficient μ of the target road segment at the current time t. t and accident incidence rate p t Using a pre-constructed road traffic flow stability model, the traffic flow stability l of the target road segment at the current time t is estimated. t ;

[0012] Step S5, based on the traffic flow stability of the target road segment at the current time t... t Using pre-constructed vehicle acceleration threshold constraint functions and vehicle distance threshold constraint functions, the vehicle acceleration threshold and vehicle distance threshold for vehicles of the same type as vehicle i at the current time t are estimated; wherein, the vehicle acceleration threshold includes the vehicle acceleration threshold a along the main line. xd,t And the vehicle's lateral acceleration threshold a yd,t The vehicle travel distance threshold includes the longitudinal distance threshold S between the vehicle and the vehicle in front. xd,t And the threshold S of the lateral movement distance of the vehicle deviating from the lane yd,t ;

[0013] Step S6: Using a vehicle collision accident detection model, analyze the acceleration a of vehicle i along the main line. x，t (i) and the vehicle acceleration threshold a along the main line xd，t , vehicle i's lateral acceleration a y,t (i) and the vehicle lateral motion acceleration threshold a yd，t The longitudinal distance ΔS between vehicle i and the vehicle in front x,t (i) and the longitudinal distance threshold S between the vehicle and the vehicle in front. xd，t The lateral distance ΔS between vehicle i and the adjacent lanes on the left and right sides y,t (i) and the vehicle lane departure lateral movement distance threshold Syd,t The size relationship is used to detect the risk of a collision involving vehicle i at the current time t.

[0014] Preferably, the function representing the slipperiness of the road section is:

[0015]

[0016] Where: μ dry RH represents the road friction coefficient of the target road section under dry conditions. t Let T be the ambient humidity of the target road segment at the current time t; t σ represents the ambient temperature of the target road segment at the current time t; β represents the pavement material coefficient of the target road segment; σ represents the ambient temperature of the target road segment at the current time t. t The climate environment type of the target road segment at the current time t is 0, which represents non-severe weather and 1 represents severe weather.

[0017] Preferably, the accident incidence rate model for the road section is as follows:

[0018]

[0019] in: The baseline accident rate coefficient is determined by historical accident data of the target road section; These are the first, second, and third influence weight coefficients, respectively.

[0020] q t v represents the road traffic volume of the target road segment at current time t; t δ represents the average speed of vehicles traveling on the target road segment at the current time t; t γ represents the proportion of large vehicles in the target road segment at time t; γ is the risk weighting coefficient for the impact of the large vehicle proportion, based on δ t Determined; R is the turning radius of the target road segment; θ is the vehicle running gradient of the target road segment; g is the gravity constant; τ is the turning sensitivity coefficient; s t s0 is the visibility of the target road segment at current time t; wt is the ambient light intensity of the target road segment at current time t; w0 is the ambient light risk threshold; f Tr (q t v t δ t f is the influence function of road segment traffic characteristics; ro (R, θ) is the influence function of road segment terrain features; f en (s t w t ) is the function that influences road segment environmental information.

[0021] Preferably, the road traffic flow stability model is as follows:

[0022]

[0023] Where α and ε are the coefficients of influence of accident rate on stability and the coefficient of influence of road friction on stability, respectively.

[0024] Preferably, the vehicle acceleration threshold constraint function is:

[0025]

[0026] Where: d represents the vehicle type of vehicle i, 1 is a passenger car, 2 is a bus, 3 is a light truck, 4 is a truck, and 5 is a freight train; a md k represents the maximum acceleration that a Class D vehicle can achieve. d f(v) is the threshold adjustment coefficient for Class D vehicles; g is the gravity constant; d f(v) represents the anti-skid and rollover constraint function for vehicles of type d on the target road segment based on their historical operating speed during turning. d )∈(0,1).

[0027] Preferably, the vehicle running distance threshold constraint function is:

[0028]

[0029] Where: m d For Class D vehicles, the safety distance adjustment factor is used; v i,t v is the speed of vehicle i at the current time t; f，t t is the speed of vehicle f directly in front of vehicle i at the current time t; t0 is the driver's reaction time;

[0030] D si n is the lateral safety distance between vehicle i and adjacent vehicles in the left and right lanes; d For vehicles classified as category d, the vehicle type adjustment coefficient is used; z i,t This represents the turn signal status of vehicle i at the current time t, where 0 indicates the turn signal is off and 1 indicates it is on.

[0031] Preferably, the vehicle collision accident detection model is as follows:

[0032] Define T c,t (i) represents the collision accident detection result variable of vehicle i at the current time t;

[0033] When ΔS x，t (i)≥S xd，t And a x，t (i)≤a xd，t And ΔS y，t (i)≥S yd，t And a y，t (i)≤a yd，tAt that time, T c,t (i) = 0, indicating that vehicle i will not be involved in a collision at the current time t;

[0034] When it exists Under one of the conditions, T c,t (i) = 1, indicating that vehicle i may be involved in a collision at the current time t;

[0035] When (ΔS) x，t (i) xd，t And a x，t (i)>a xd，t ) or (ΔS y，t (i) yd，t And a y，t (i)>a yd，t When T c,t (i) = 2, indicating that a collision will occur with vehicle i starting from the current time t.

[0036] The collision accident detection method based on traffic flow stability and vehicle operating status provided by this invention has the following advantages:

[0037] This invention utilizes roadside meteorological environment detection and video recognition devices to obtain information on the operating status of vehicles on roads under adverse weather conditions. It combines historical accident data and road surface condition information of the road section, integrates the influence of road traffic and road surface environment, introduces the concept of road section traffic flow stability, and integrates vehicle operating status information to achieve the detection of vehicle collision accidents. It has the advantage of high accuracy in vehicle collision accident detection. Attached Figure Description

[0038] Figure 1 The flowchart of the collision accident detection method based on traffic flow stability and vehicle operating status provided by the present invention. Detailed Implementation

[0039] To make the technical problems solved, the technical solutions, and the beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the invention.

[0040] This invention fully considers the influence of vehicle status, road surface environment, and traffic environment, integrates the influence of road traffic and road surface environment, introduces the concept of road segment traffic flow stability, and integrates vehicle operation status information to realize the detection of vehicle collision accidents and solve the problem of frequent vehicle collision accidents under poor road conditions.

[0041] ​​Specifically, the collision accident detection method based on traffic flow stability and vehicle operation status provided by this invention obtains information on the road vehicle operation status under adverse weather conditions by using roadside meteorological environment detection, video recognition and other devices. It combines historical accident data and road surface condition information of the road section, integrates the influence of road traffic and road surface environment, introduces the concept of road section traffic flow stability, and integrates vehicle operation status information to achieve accurate detection of vehicle collision accidents.

[0042] The main ideas of this invention include:

[0043] Step 1: Introducing the concept of road segment traffic flow stability, we constructed a road segment slipperiness characterization function and a road segment accident rate model, and then constructed a road traffic flow stability model.

[0044] The second step is to integrate vehicle operating status with traffic flow stability, establish a vehicle operating status threshold constraint function that takes into account traffic flow stability, and then form a vehicle collision accident detection method.

[0045] like Figure 1 As shown, this invention provides a collision accident detection method based on traffic flow stability and vehicle operating status, including:

[0046] Steps S1 to S4 are used to construct a road traffic flow stability model based on accident rate and road surface condition: from the perspective of macroscopic road traffic flow operation characteristics, considering the influencing factors such as road surface slipperiness and accident rate under adverse weather conditions, a road traffic flow stability model is constructed.

[0047] Step S1: For vehicle i entering the target road segment, acquire in real time the target road segment environmental information, target road segment terrain feature information, target road segment traffic feature information, vehicle type of vehicle i, and vehicle i's driving information at the current time t; wherein, the driving information of vehicle i includes the acceleration a of vehicle i along the main line. x，t (i), the lateral acceleration of vehicle i is a y，t (i), the longitudinal distance ΔS between vehicle i and the vehicle in front. x，t (i) and the lateral distance ΔS between vehicle i and the adjacent lanes on the left and right sides. y,t (i);

[0048] Step S2: Using a pre-constructed road segment slipperiness characterization function, and based on the target road segment environmental information, estimate the road segment friction coefficient μ of the target road segment that affects vehicle operation safety at the current time t. t ;

[0049] The degree of road surface slippage directly affects vehicle operating safety, causing changes in the road friction coefficient μ. This change in μ is influenced by a combination of factors, including ambient temperature, humidity, and road surface material, exhibiting a non-linear relationship. To best fit the relationship between road surface slippage and the road friction coefficient, this invention constructs a segmented road surface slippage characterization function based on the variation patterns of the road friction coefficient under different influencing conditions:

[0050] The function representing the slipperiness of the road section is:

[0051]

[0052] Where: μ dry The coefficient of friction for the target road section under dry conditions is typically 0.7-0.9; RHt represents the ambient humidity percentage of the target road section at time t; T t σ represents the ambient temperature of the target road section at the current time t (°C); β represents the pavement material coefficient of the target road section, typically 0.02 for asphalt pavement; σ represents the ambient temperature of the target road section at the current time t (°C). t This represents the climate environment type of the target road segment at the current time t. 0 indicates non-severe weather, and 1 indicates severe weather, such as rain, snow, freezing, etc.

[0053] Step S3: Using a pre-constructed road segment accident rate model, based on the target road segment environmental information, the target road segment terrain feature information, and the target road segment traffic feature information, the accident rate p of the target road segment at the current time t is estimated. t ;

[0054] Road traffic stability is closely related to the accident rate of a road segment. The magnitude of the accident rate is directly related to the road alignment, traffic safety facilities, speed limits, and the composition and number of vehicles on the road. Considering that this invention mainly focuses on assessing vehicle operation safety based on vehicle operation status information, the influencing indicators of its accident rate characterization function are selected as follows: vehicle speed, road traffic volume, visibility, light intensity, and historical accidents of the road segment.

[0055] The accident incidence model for the aforementioned road section is as follows:

[0056]

[0057] in: The baseline accident rate coefficient is determined by historical accident data of the target road section; These are the first, second, and third influence weight coefficients, respectively.

[0058] q t v represents the road traffic volume of the target road segment at current time t; t δ represents the average speed of vehicles traveling on the target road segment at the current time t; tγ represents the proportion of large vehicles in the target road segment at time t; γ is the risk weighting coefficient for the impact of the large vehicle proportion, based on δ t Determined; R is the turning radius of the target road segment; θ is the vehicle running gradient of the target road segment; g is the gravity constant, typically 9.8 N / kg; τ is the turning sensitivity coefficient; s t st represents the visibility of the target road segment at current time t; s0 represents the visibility risk threshold; wt represents the visibility risk threshold. t The ambient light intensity of the target road segment at time t is denoted as w0; the ambient light risk threshold is denoted as f. Tr (q t v t δ t f is the influence function of road segment traffic characteristics; ro (R, θ) is the influence function of road segment terrain features; f en (s t w t ) is the function that influences road segment environmental information.

[0059] Step S4, comprehensively consider the road friction coefficient μt and accident occurrence rate p of the target road segment at the current time t. t The stability of traffic flow on the target road segment is described. A pre-constructed road traffic flow stability model is used to estimate the traffic flow stability l of the target road segment at the current time t. t This allows for a macro-level assessment of vehicle operational safety on the target road section.

[0060] The road traffic flow stability model is as follows:

[0061]

[0062] Where α and ε are the coefficients of influence of accident rate on stability and road segment friction coefficient on stability, respectively. Traffic flow stability l t ∈(0,1),l t The larger the road segment, the higher the traffic flow stability.

[0063] Step S5, based on the traffic flow stability of the target road segment at the current time t... t Using pre-constructed vehicle acceleration threshold constraint functions and vehicle distance threshold constraint functions, the vehicle acceleration threshold and vehicle distance threshold for vehicles of the same type as vehicle i at the current time t are estimated; wherein, the vehicle acceleration threshold includes the vehicle acceleration threshold a along the main line. xd，t And the vehicle's lateral acceleration threshold a yd，t The vehicle travel distance threshold includes the longitudinal distance threshold S between the vehicle and the vehicle in front. xd,t And the threshold S of the lateral movement distance of the vehicle deviating from the lane yd,t ;

[0064] The vehicle acceleration threshold constraint function is:

[0065]

[0066] Where: d represents the vehicle type of vehicle i, 1 is a passenger car, 2 is a bus, 3 is a light truck, 4 is a truck, and 5 is a freight train; a md k represents the maximum acceleration that a Class D vehicle can achieve. d f(v) is the threshold adjustment coefficient for Class D vehicles; g is the gravity constant; d f(v) represents the anti-skid and rollover constraint function for vehicles of type d on the target road segment based on their historical operating speed during turning. d )∈(0,1).

[0067] The vehicle travel distance threshold constraint function is:

[0068]

[0069] Where: m d For Class D vehicles, the safety distance adjustment factor is used; v i,t v is the speed of vehicle i at the current time t; f，t D represents the speed of vehicle f directly in front of vehicle i at the current time t; t0 represents the driver's reaction time, typically 2 seconds; si n is the lateral safety distance between vehicle i and adjacent vehicles in the left and right lanes; d For vehicles classified as category d, the vehicle type adjustment coefficient is used; z i,t This represents the turn signal status of vehicle i at the current time t, where 0 indicates the turn signal is off and 1 indicates it is on.

[0070] Specifically, steps S5 to S6 are used for a collision estimation method based on traffic flow stability and vehicle operating status. The operating status of road vehicles has a direct impact on vehicle collision accidents. Integrating numerous vehicle collision accident analyses, it is found that a vehicle's operating status changes before a collision, such as vehicle speed, acceleration, and displacement. Based on this, this invention selects vehicle acceleration and displacement as indicators to determine whether a collision has occurred, and integrates the actual macroscopic traffic flow stability of road sections to perform real-time collision estimation for different vehicle types.

[0071] Step S6: Using a vehicle collision accident detection model, analyze the acceleration a of vehicle i along the main line. x，t (i) and the vehicle acceleration threshold a along the main line xd，t , vehicle i's lateral acceleration a y，t (i) and the vehicle lateral motion acceleration threshold a yd，t The longitudinal distance ΔS between vehicle i and the vehicle in frontx，t (i) and the longitudinal distance threshold S between the vehicle and the vehicle in front. xd，t The lateral distance ΔS between vehicle i and the adjacent lanes on the left and right sides y，t (i) and the vehicle lane departure lateral movement distance threshold S yd，t The size relationship is used to detect the risk of a collision involving vehicle i at the current time t.

[0072] This invention relates to vehicle collision accident detection, which is categorized into rear-end collision accidents and side collision accidents.

[0073] The vehicle collision accident detection model is as follows:

[0074] Define T c,t (i) represents the collision accident detection result variable of vehicle i at the current time t;

[0075] When ΔS x，t (i)≥S xd，t And a x，t (i)≤a xd，t And ΔS y，t (i)≥S yd，t And a y，t (i)≤a yd，t At that time, T c,t (i) = 0, indicating that vehicle i will not be involved in a collision at the current time t;

[0076] When it exists Under one of the conditions, T c,t (i) = 1, indicating that vehicle i may be involved in a collision at the current time t;

[0077] When (ΔS) x，t (i) xd，t And a x，t (i)>a xd，t ) or (ΔS y，t (i) yd，t And a y，t (i)>a yd，t When T c,t (i) = 2, indicating that a collision will occur with vehicle i starting from the current time t.

[0078] The vehicle collision accident detection method of the present invention takes into account the influence of road traffic flow stability on vehicle operating status, and forms a collision accident detection method with macro-micro dual-level constraints.

[0079] ​​This invention utilizes roadside meteorological environment detection and video recognition devices to obtain information on the operating status of vehicles on roads under adverse weather conditions. It combines historical accident data and road surface condition information of the road section, integrates the influence of road traffic and road surface environment, introduces the concept of road section traffic flow stability, and integrates vehicle operating status information to achieve the detection of vehicle collision accidents. It has the advantage of high accuracy in vehicle collision accident detection.

[0080] This invention considers the impact of multiple factors such as road traffic, road conditions, and driving environment on vehicle collision accident detection, constructs a road traffic flow stability model based on accident incidence and road surface conditions, and builds a vehicle collision estimation method based on this model to achieve effective detection of vehicle collisions.

[0081] This invention provides a collision accident detection method based on traffic flow stability and vehicle operating status. It identifies, analyzes, and detects accidents such as scraping and colliding between vehicles or between vehicles and surrounding obstacles caused by adverse environmental factors, so as to achieve early warning and alert for vehicle operation safety.

[0082] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A collision accident detection method based on traffic flow stability and vehicle running state, characterized in that, The method comprises the following steps: Step S1, for the vehicle i driving into the target section, real-time acquisition of the target section environment information, the target section terrain feature information, the target section traffic feature information, the vehicle type of the vehicle i and the driving information of the vehicle i at the current time t; wherein the driving information of the vehicle i includes the running acceleration a x,t (i) of the vehicle i along the main line, the lateral running acceleration a y,t (i) of the vehicle i, the longitudinal distance ΔS x,t (i) between the vehicle i and the front vehicle, and the lateral distance ΔS y,t (i) between the vehicle i and the adjacent lanes on the left and right sides. In step S2, a pre-constructed road section wetness degree representation function is used to estimate the road section friction coefficient μ of the target road section affecting the vehicle running safety according to the target road section environment information at the current time t t ; In step S3, a pre-constructed accident occurrence rate model is used to estimate the accident occurrence rate p of the target road section at the current time t according to the target road section environment information, the target road section terrain feature information, and the target road section traffic feature information. t ; Step S4, comprehensively considering the road friction coefficient μ of the target road section at the current time t t and the accident rate p t , using a pre-constructed road traffic flow stability model to estimate the traffic flow stability l of the target road section at the current time t t ; Step S5, according to the current time t target road section traffic flow stability l t , respectively, using a pre-constructed vehicle running acceleration threshold constraint function and vehicle running distance threshold constraint function, estimated to get the current time t and the same vehicle type as the vehicle running acceleration threshold and vehicle running distance threshold of the vehicle i; wherein the vehicle running acceleration threshold includes the vehicle along the main line running acceleration threshold a xd,t and vehicle lateral motion acceleration threshold a yd,t , the vehicle running distance threshold includes the vehicle and the longitudinal distance threshold S xd,t and vehicle offset lane lateral motion distance threshold S yd,t ; Step S6, using the vehicle collision accident detection model, analyzing the vehicle i along the main line running acceleration a x,t (i) and the vehicle along the main line running acceleration threshold a xd,t , the vehicle i lateral running acceleration a y,t (i) and the vehicle lateral motion acceleration threshold a yd,t , the longitudinal distance ΔS of the vehicle i and the front vehicle x,t (i) and the longitudinal distance threshold S of the vehicle and the front vehicle xd,t , the lateral distance ΔS of the vehicle i and the left and right adjacent lanes y,t (i) and the vehicle lane deviation lateral motion distance threshold S yd,t The size relationship is detected. The risk of vehicle i collision accident at the current time t.

2. The method for detecting a collision accident based on traffic flow stability and vehicle operating state according to claim 1, wherein, The road section wetness degree representation function is: wherein: μ dry is the road friction coefficient of the target section under dry conditions; RH t is the environmental humidity of the target section at the current time t; T t is the environmental temperature of the target section at the current time t; β is the road surface material coefficient of the target section; σ t is the climate environment type of the target section at the current time t, 0 represents non-adverse weather, and 1 represents adverse weather. 3.The collision accident detection method based on traffic flow stability and vehicle running state according to claim 1, wherein, The road section accident occurrence rate model is: Wherein: is a reference accident rate coefficient, determined by historical accident data of the target section; are respectively a first influence weight coefficient, a second influence weight coefficient and a third influence weight coefficient. q t is the road traffic volume of the target road section at the current time t; v t is the average speed of vehicles running on the target road section at the current time t; δ t is the proportion of large vehicles in traffic on the target road section at the current time t; γ is the risk weight coefficient of the proportion of large vehicles, determined according to δ t ; R is the linear turning radius of the target road section; θ is the running slope of the target road section; g is the gravitational constant; τ is the turning sensitivity coefficient; s t is the visibility of the target road section at the current time t; s0 is the critical threshold of the visibility risk; w t is the ambient light intensity of the target road section at the current time t; w0 is the critical threshold of the ambient light risk; f Tr (q t , v t , δ t ) is the influence function of the traffic characteristics of the road section; f ro (R, θ) is the influence function of the terrain characteristics of the road section; f en (s t , w t ) is the influence function of the environmental information of the road section.

4. The method for detecting a collision accident based on traffic flow stability and vehicle operating state according to claim 1, wherein, The road traffic flow stability model is: Wherein: α, ε are respectively the influence coefficient of accident occurrence rate on stability, and the influence coefficient of road section friction coefficient on stability.

5. The method for detecting a collision accident based on traffic flow stability and vehicle operating state according to claim 1, wherein, The vehicle operation acceleration threshold constraint function is: wherein: d is the vehicle type of vehicle i, 1 is a small passenger car, 2 is a large passenger car, 3 is a small truck, 4 is a large truck, and 5 is a truck train; a md is the maximum acceleration that can be reached by the vehicle of type d running; k d is the threshold adjustment coefficient of the vehicle of type d; g is the gravitational constant; f(v d ) is the turning driving anti-skid and rollover constraint function of the historical running speed of the vehicle of type d on the target section, f(v d ) ∈ (0, 1).

6. The method for detecting a collision accident based on traffic flow stability and vehicle operating state according to claim 1, wherein, The vehicle operation distance threshold constraint function is: Wherein: m d is the safety distance adjustment coefficient of the d-type vehicle; v i,t is the running speed of the vehicle i at the current time t; v f,t is the running speed of the vehicle f in front of the vehicle i at the current time t; t0is the driver reaction time D si is the lateral safety distance of vehicle i and the vehicles adjacent to the left and right lanes; n d is the vehicle type adjustment coefficient of the d-type vehicle; z i,t is the state of the turn signal of vehicle i at the current time t, 0 is not on, 1 is on.

7. The method for detecting a collision accident based on traffic flow stability and vehicle operating state according to claim 1, wherein, The vehicle collision accident detection model is: Definition T c,t (i) denotes the collision accident detection result variable of the vehicle i at the current time instant t; when ΔS x,t (i) ≥ S xd,t and a x,t (i) ≤ a xd,t and ΔS y,t (i) ≥ S yd,t and a y,t (i) ≤ a yd,t then, T c,t (i) = 0, means that the vehicle i will not have a collision accident at the current time t; When one of the conditions exists T c,t (i) = 1, indicating that vehicle i can have a collision accident at the current time t; when (ΔS x,t (i)<S xd,t and a x,t (i)>a xd,t ) or (ΔS y,t (i)<S yd,t and a y,t (i)>a yd,t ), T c,t (i) = 2, indicating that vehicle i will have a collision accident from the current time t.

Images