Method for analyzing vehicle-pedestrian conflict at unsignalized pedestrian crossing under mixed traffic flow

Abstract

The application discloses a kind of mixed traffic flow under the method for analyzing the conflict between man and vehicle of no-signal pedestrian crossing, comprising: obtaining the conflict simulation model parameter corresponding to the no-signal pedestrian crossing of the analyzed pedestrian crossing;Respectively, construct pedestrian crossing behavior model, traditional vehicle behavior model and automatic driving vehicle behavior model, and obtain the conflict simulation model based on pedestrian crossing behavior model, traditional vehicle behavior model and automatic driving vehicle behavior model;The conflict behavior between man and vehicle of no-signal pedestrian crossing is extracted using the conflict simulation model between man and vehicle, and the severity of the extracted conflict behavior between man and vehicle is evaluated.The application is of great significance to guarantee the safety of pedestrians at no-signal crossing after the deployment of autonomous vehicles.

Description

Technical Field

[0001] This invention belongs to the field of road traffic technology, and in particular relates to a method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings in mixed traffic flow, used to evaluate the severity of pedestrian-vehicle conflicts in mixed traffic flow with autonomous vehicles. Background Technology

[0002] With economic and social development, the number of cars on the road is constantly increasing, and road safety is facing severe challenges, especially in areas with complex traffic conditions and unclear right-of-way allocation, such as uncontrolled pedestrian crossings. Due to a lack of necessary protection, pedestrians are a vulnerable group on the road, and interactions between pedestrians and vehicles are often frequent and dangerous. The deployment of autonomous vehicles can reduce the severity of pedestrian-vehicle conflicts by minimizing driver misconduct. However, a transitional period will emerge in the future where autonomous vehicles and conventional vehicles coexist on the road. The heterogeneity of these two types of vehicles in terms of risk perception, driving behavior, and decision-making may further increase the complexity of pedestrian-vehicle conflicts.

[0003] However, the severity of pedestrian-vehicle conflicts under mixed traffic flow remains unknown. Previous studies on pedestrian-vehicle conflicts have focused on pedestrians' crossing behavior in front of regular vehicles, the yielding decisions and behaviors of regular vehicles, and issues such as the identification and severity assessment of pedestrian-vehicle conflicts.

[0004] The invention disclosed in CN114898042A presents a method for predicting the risk of pedestrian-vehicle collisions based on spatiotemporal urgency, belonging to the field of autonomous driving pedestrian-vehicle collision risk prediction. It predicts multimodal pedestrian trajectories based on onboard perspective data; uses a bounding box dynamic detection algorithm to test the intersection of pedestrian trajectories and the vehicle's planned path to identify potential collision trajectories; calculates conflict parameters under potential collision trajectories and without potential collision trajectories; establishes a model to quantify spatiotemporal urgency; outputs the comprehensive pedestrian hazard level; and delineates safe driving zones based on the comprehensive hazard level and the safe distance between pedestrians and vehicles. This invention fully considers all possible pedestrian trajectories, improving upon the limitations of existing pedestrian-vehicle collision risk prediction methods, such as incomplete consideration of factors and subjective evaluation mechanisms, making it more adaptable to autonomous driving and complex traffic environments. The invention disclosed in publication number CN114299607A presents a method for analyzing the risk of pedestrian-vehicle collisions based on autonomous vehicles. First, based on pedestrian crossing characteristics in historical datasets, Gaussian clustering is used to categorize crossing behaviors into different habit types. A joint probability distribution function is then used to help the autonomous vehicle obtain a set of possible future states of the pedestrian based on their current motion state. A graph convolutional neural network (GCN) is then used to consider the dynamic spatiotemporal relationship between pedestrians and vehicles, resulting in a set of possible future trajectories for the pedestrian. Finally, considering the trajectory collision probability and minimum encounter distance, matter-element extension theory is used for feature dimensionality reduction to establish a risk function, enabling real-time assessment of the risk of pedestrian-vehicle collisions. This invention focuses on the abrupt changes in pedestrian motion states, more specifically considering the impact of the spatiotemporal relationship between pedestrians and vehicles on trajectories. The fusion of multiple indicators improves the accuracy and reliability of the collision assessment results, further enhancing the intelligence of autonomous driving and improving ride comfort and safety. However, both patents are specifically designed for autonomous vehicles and involve extremely high computational demands. Existing technical literature contains limited research on assessing the severity of pedestrian-vehicle conflicts in mixed traffic flows. Summary of the Invention

[0005] Technical problem solved: This invention addresses the conflict between pedestrians and vehicles at uncontrolled pedestrian crossings in mixed traffic flow involving autonomous vehicles. It discloses a method for analyzing such conflicts, which is of great significance for ensuring pedestrian traffic safety at uncontrolled pedestrian crossings after the deployment of autonomous vehicles.

[0006] Technical solution:

[0007] A method for analyzing pedestrian-vehicle conflicts at unsignalized pedestrian crossings under mixed traffic flow, the method comprising the following steps:

[0008] S1. Obtain the simulation model parameters of the pedestrian-vehicle conflict corresponding to the pedestrian crossing without signal control to be analyzed. The simulation model parameters of the pedestrian-vehicle conflict include the basic data required for the construction of the pedestrian-vehicle conflict simulation environment, the basic data required for pedestrian crossing behavior modeling, and the basic data required for conventional vehicle behavior and autonomous vehicle behavior modeling.

[0009] S2, construct pedestrian crossing behavior model, traditional vehicle behavior model and autonomous vehicle behavior model respectively, and construct human-vehicle conflict simulation model based on pedestrian crossing behavior model, traditional vehicle behavior model and autonomous vehicle behavior model; import the human-vehicle conflict simulation model parameters in step S1 into pedestrian crossing behavior model, traditional vehicle behavior model, autonomous vehicle behavior model and human-vehicle conflict simulation model respectively.

[0010] S3 uses a pedestrian-vehicle conflict simulation model to extract pedestrian-vehicle conflict behaviors at uncontrolled pedestrian crossings and evaluates the severity of the extracted pedestrian-vehicle conflict behaviors.

[0011] Furthermore, in step S1, the basic data required for constructing the vehicle-pedestrian conflict simulation environment includes the number of lanes N and the lane width W. L Road segment length L L Uncontrolled pedestrian crossing width W C The length L of the uncontrolled pedestrian crossing C Speed ​​limit V for this road section lim Vehicle arrival rate λ V Pedestrian arrival rate λ P ;

[0012] The basic data required for modeling pedestrian crossing behavior includes pedestrian crossing speed V. P Pedestrian width W P Pedestrian length L P Safety margin γ for pedestrians facing regular vehicles ped，CV Safety margin γ for pedestrians facing autonomous vehicles ped，AV ;

[0013] The basic data required for modeling conventional vehicle behavior and autonomous vehicle behavior includes: vehicle width W. V Vehicle length L V Vehicle speed V V Vehicle desired speed V0, driver reaction time τ CV Expected interval distance s * The deceleration d of a conventional vehicle CV The maximum acceleration 'a' of a conventional vehicle, and the minimum acceptable safe interval 'γ' between a conventional vehicle entering / leaving the conflict zone and a pedestrian leaving / entering the conflict zone. CV The driver's sight distance D for conventional vehicles CVThe penetration rate of autonomous vehicles P AV The deceleration of autonomous vehicles AV Delay time τ of autonomous driving system AV The reference distance s between the autonomous vehicle and the vehicle in front ref Autonomous vehicle detection distance D AV .

[0014] Furthermore, in step S2, the process of constructing the pedestrian crossing behavior model includes the following sub-steps:

[0015] The behavior of pedestrians at uncontrolled crosswalks was modeled using AnyLogic software, resulting in a pedestrian crossing behavior model. Specifically, when facing regular vehicles, pedestrians adjust their behavior according to a safety margin γ. ped，CV When making a pedestrian crossing decision, the safety margin is the marginal safety value that pedestrians maintain to avoid collisions with vehicles; the conflict zone is the overlapping area where the vehicle trajectory and the pedestrian trajectory intersect, when t ped，leave +γ ped，CV ≤t veh，enter When the conflict zone is considered safe, pedestrians move at a speed of V. P Pedestrians will begin crossing the street at a constant speed; otherwise, they will continue to wait. When a pedestrian encounters an autonomous vehicle, if the pedestrian cannot recognize the signal emitted by the vehicle's eHMIs or the eHMIs do not emit a signal, the autonomous vehicle will be treated as a regular vehicle. If the pedestrian recognizes the signal emitted by the autonomous vehicle's eHMIs, a crossing decision will be made based on the signal emitted by the vehicle's eHMIs. When the received signal is "yield," the pedestrian begins to cross at a speed of V. P Pedestrians cross the street at a constant speed; when the signal is "not yielding," pedestrians continue to wait for an opportunity to cross; in the formula, d ped，leave This represents the distance a pedestrian travels from their current location away from the conflict zone. t ped，leaave d is the time required for a pedestrian to leave the conflict zone from their current location. veh，enter t represents the distance the vehicle travels from its current location into the conflict zone. veh，enter The time it takes for the vehicle to travel from its current location into the conflict zone;

[0016] Input the basic data required for pedestrian crossing behavior modeling in step S1 into the pedestrian crossing behavior model.

[0017] Furthermore, in step S2, the construction process of the traditional vehicle behavior model includes the following sub-steps:

[0018] The behavior of conventional vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in a traditional vehicle behavior model. Specifically, at unsignalized pedestrian crossings...

[0019] When the visual range of a traditional car driver is D CV A pedestrian appears inside, and when t ped，leave +γ CV ≤t veh，enter or t veh，leave +γ CV ≤t ped，enter At that time, the driver chose to yield, reducing speed d CV Begin deceleration; otherwise, the driver accelerates at time t with an acceleration a. IDM Follow the vehicle in front or drive freely at speed V0; where, d ped，enter t represents the distance a pedestrian travels from their current location into the conflict zone. ped，enter The time it takes for a pedestrian to walk from their current location into the conflict zone. d veh，leave t represents the distance the vehicle has traveled from its current location away from the conflict zone. veh，leave The time taken for the vehicle to leave the conflict zone from its current location; a IDM (t+τ CV ) is t+τ CV The vehicle acceleration at time t, s(t) is the distance between the current vehicle and the vehicle in front at time t, and V is the acceleration of the vehicle at time t. v (t) is the vehicle speed at time t.

[0020] Input the basic data parameters required for conventional vehicle behavior modeling in step S1 into the traditional vehicle behavior model.

[0021] Furthermore, in step S2, the process of constructing the autonomous vehicle behavior model includes the following sub-steps:

[0022] The behavior of autonomous vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in an autonomous vehicle behavior model. Specifically, at unsignalized pedestrian crossings, when the detection range D of the autonomous vehicle... AV When there are no pedestrians, the autonomous vehicle accelerates at time t. Follow the vehicle in front or drive freely at speed V0; For the current vehicle at t+τ AV acceleration at time, k a k v and k d All are model parameters, a n-1 (t) represents the acceleration of the vehicle in front at time t, v n-1 (t) represents the speed of the vehicle in front at time t, v V (t) represents the current speed of the vehicle at time t, s V (t) represents the distance between the current vehicle and the vehicle in front at time t.

[0023] When the detection range of the autonomous vehicle is D AV When a pedestrian appears, yield to the pedestrian and reduce speed d. AV Begin to decelerate;

[0024] Input the basic data parameters required for autonomous vehicle behavior modeling in step S1 into the autonomous vehicle behavior model.

[0025] Furthermore, in step S2, the process of constructing the human-vehicle conflict simulation model includes the following sub-steps:

[0026] Using AnyLogic software, a pedestrian-vehicle conflict simulation model is constructed based on the basic data required for building the pedestrian-vehicle conflict simulation environment in step S1. The uncontrolled pedestrian crossing is located at the midpoint of the road segment, and vehicles start at the beginning of the road with a speed of λ. V The vehicle arrival rate is generated, assuming the probability that a vehicle is an autonomous vehicle is P. AV And randomly assign less than V lim The initial velocity is removed by the system upon reaching the end of the road; pedestrians move at λ on both sides of the uncontrolled crosswalk. P The pedestrian arrival rate is generated, and pedestrians are removed by the system when they reach the other side of the uncontrolled crosswalk.

[0027] Furthermore, in step S3, the process of extracting pedestrian-vehicle conflict behaviors at uncontrolled pedestrian crossings using a pedestrian-vehicle conflict simulation model and evaluating the severity of the extracted pedestrian-vehicle conflict behaviors includes the following sub-steps:

[0028] Safety alternatives are used to identify and assess the severity of pedestrian-vehicle conflicts. The rear intrusion time (PET) is selected as an alternative safety indicator. PET = T2 - T1, where T1 is the time it takes for the vehicle / pedestrian to leave the conflict area, T2 is the time it takes for the pedestrian / vehicle to enter the conflict area, δ1 is the PET threshold when a pedestrian-vehicle conflict occurs, and δ2 is the PET threshold for a serious conflict. If a vehicle and a pedestrian approach each other at an uncontrolled pedestrian crossing, a pedestrian-vehicle conflict is considered to have occurred if PET < δ1, and a serious conflict is considered if PET < δ2.

[0029] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein a computer program is stored therein, which, when executed by a processor, implements the steps in the aforementioned method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings under mixed traffic flow.

[0030] According to another aspect of the present invention, a computer device is provided, the computer device including a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps in the aforementioned method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings under mixed traffic flow.

[0031] Beneficial effects:

[0032] The present invention provides a method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings in mixed traffic flow. By selecting uncontrolled pedestrian crossings as the research object, the method uses AnyLogic software to construct a simulation model to study the severity of pedestrian-vehicle conflicts under mixed traffic flow with autonomous vehicles. This method is of great significance for ensuring pedestrian traffic safety at uncontrolled pedestrian crossings after the deployment of autonomous vehicles. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of a human-vehicle conflict simulation model scenario according to an embodiment of the present invention;

[0034] Figure 2 This is a flowchart illustrating the pedestrian-vehicle conflict analysis method for unsignalized pedestrian crossings under mixed traffic flow, as described in an embodiment of the present invention. Detailed Implementation

[0035] The following embodiments are provided to enable those skilled in the art to more fully understand the present invention, but do not limit the invention in any way.

[0036] This invention provides a method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings in mixed traffic flow, which can accurately assess the severity of pedestrian-vehicle conflicts occurring at uncontrolled pedestrian crossings in mixed traffic flow with autonomous vehicles.

[0037] like Figure 2 As shown in the figure, this embodiment discloses a method for analyzing pedestrian-vehicle conflicts at unsignalized pedestrian crossings under mixed traffic flow, including the following steps:

[0038] (1) Preparation of basic data for the simulation model of human-vehicle conflict

[0039] This includes the basic data required for building the simulation environment, the basic data required for modeling pedestrian crossing behavior, and the basic data required for modeling conventional vehicle behavior and autonomous vehicle behavior. The basic data required for building the simulation environment includes: the number of lanes N, and the lane width W. L Road segment length L L Uncontrolled pedestrian crossing width W C The length L of the uncontrolled pedestrian crossing C Speed ​​limit V for this road section lim Vehicle arrival rate λ V Pedestrian arrival rate λP The basic data required for pedestrian crossing behavior modeling includes: pedestrian crossing speed V. P Pedestrian width W P Pedestrian length L P Safety margin γ for pedestrians facing regular vehicles ped，CV Safety margin γ for pedestrians facing autonomous vehicles ped，AV The basic data required for modeling conventional vehicle behavior and autonomous vehicle behavior includes: vehicle width W. V Vehicle length L V Vehicle speed V V Vehicle desired speed V0, driver reaction time τ CV Expected interval distance s * The deceleration d of a conventional vehicle CV The maximum acceleration 'a' of a conventional vehicle, and the minimum acceptable safe interval 'γ' between a conventional vehicle entering / leaving the conflict zone and a pedestrian leaving / entering the conflict zone. CV The driver's sight distance D for conventional vehicles CV The penetration rate of autonomous vehicles P AV The deceleration of autonomous vehicles AV Delay time τ of autonomous driving system AV The reference distance s between the autonomous vehicle and the vehicle in front ref Autonomous vehicle detection distance D AV .

[0040] (2) Parameter input and simulation model construction, including pedestrian crossing behavior modeling, traditional vehicle behavior modeling, autonomous vehicle behavior modeling, and simulation environment construction.

[0041] ① Pedestrian crossing behavior modeling

[0042] The AnyLogic software was used to model the behavior of pedestrians at uncontrolled crosswalks, and the parameters in step (1) were input. The specific model settings are as follows: When a pedestrian faces a regular vehicle, the pedestrian will act according to the safety margin γ. ped，CV When making a pedestrian crossing decision, the safety margin is the marginal safety value that pedestrians maintain to avoid collisions with vehicles; the conflict zone is the overlapping area where the vehicle trajectory and the pedestrian trajectory intersect, when t ped，leave +γ ped，CV ≤t veh，enter When the conflict zone is considered safe, pedestrians will move at a speed of V. PPedestrians will begin crossing the street at a constant speed; otherwise, they will continue to wait. When faced with an autonomous vehicle, pedestrians will make different crossing decisions based on their familiarity with the vehicle. Autonomous vehicles are equipped with human-machine interfaces (eHMIs) that issue signals to pedestrians waiting to cross the street. When pedestrians are unfamiliar with autonomous vehicles and cannot identify the vehicle type, they will treat it as a regular vehicle. Pedestrians will also make crossing decisions based on safety margins. Therefore, when t... ppd，leave +γ ped，cv ≤t veh，enter When the conflict zone is considered safe, pedestrians will move at a speed of V. P Begin crossing the street at a constant speed; otherwise, continue waiting. In the formula, d ped，leave This represents the distance a pedestrian travels from their current location away from the conflict zone. d veh，enter This refers to the distance the vehicle travels from its current location into the conflict zone. Once pedestrians are familiar with the autonomous vehicle, they will first make a crossing decision based on signals from the vehicle's eHMIs. When the received signal is "yield," the pedestrian begins to cross at speed V. P Pedestrians will cross the street at a constant speed; when the signal is "not yielding," pedestrians will continue to wait for an opportunity to cross; if the eHMI does not issue a signal, pedestrians will cross according to the safety margin γ. ped，AV When making a decision to cross the street, when t ped，leave +γ ped，AV ≤t veh，enter When the conflict zone is considered safe, pedestrians will move at a speed of V. P Begin crossing the street at a constant speed; otherwise, continue waiting.

[0043] ② Traditional vehicle behavior modeling

[0044] The behavior of conventional vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in a traditional vehicle behavior model. Specifically, at unsignalized pedestrian crossings...

[0045] When the visual range of a traditional car driver is D cV A pedestrian appears inside, and when t ped，leave +γ CV ≤t veh，enter or t veh，leave +γ CV ≤t eed，enter At that time, the driver chose to yield, reducing speed d CV Begin deceleration; otherwise, the driver accelerates at time t with an acceleration a. IDM Follow the vehicle in front or drive freely at speed V0; where, d ped，enter t represents the distance a pedestrian travels from their current location into the conflict zone. ped，enterThe time it takes for a pedestrian to walk from their current location into the conflict zone. d veh，leave t represents the distance the vehicle has traveled from its current location away from the conflict zone. veh，leave The time taken for the vehicle to leave the conflict zone from its current location; a IDM (t+τ CV ) is t+τ CV The vehicle acceleration at time t, s(t) is the distance between the current vehicle and the vehicle in front at time t, and V is the acceleration of the vehicle at time t. v (t) is the vehicle speed at time t.

[0046] ③ Autonomous vehicle behavior modeling

[0047] The behavior of autonomous vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in an autonomous vehicle behavior model. Specifically, at unsignalized pedestrian crossings, when the detection range D of the autonomous vehicle... AV When there are no pedestrians, the autonomous vehicle accelerates at time t. Follow the vehicle in front or drive freely at speed V0; For the current vehicle at t+τ AV acceleration at time, k a k v and k d All are model parameters, a n-1 (t) represents the acceleration of the vehicle in front at time t, v n-1 (t) represents the speed of the vehicle in front at time t, v V (t) represents the current speed of the vehicle at time t, s V (t) represents the distance between the current vehicle and the vehicle in front at time t.

[0048] When the detection range of the autonomous vehicle is D AV When a pedestrian appears, yield to the pedestrian and reduce speed d. AV Start decelerating.

[0049] ④ Simulation Environment Construction

[0050] Using AnyLogic software, a simulation model of pedestrian-vehicle conflict is constructed based on the specific data of the simulation model in step (1). The uncontrolled pedestrian crossing is located at the midpoint of the road segment, and vehicles start at the beginning of the road with a speed of λ. V The vehicle arrival rate is generated, and the probability that the vehicle is an autonomous vehicle is P. AV And randomly assign less than V lim The initial velocity is removed by the system upon reaching the end of the road; pedestrians move at λ on both sides of the uncontrolled crosswalk. PThe pedestrian arrival rate is generated, and pedestrians are removed by the system when they reach the other side of the uncontrolled crosswalk.

[0051] (3) Extraction of pedestrian-vehicle conflicts and assessment of the severity of pedestrian-vehicle conflicts

[0052] Safety alternatives are used to identify and assess the severity of pedestrian-vehicle conflicts. The rear intrusion time (PET) is selected as an alternative safety indicator. PET = T2 - T1, where T1 is the time it takes for a vehicle / pedestrian to leave the conflict area, T2 is the time it takes for a pedestrian / vehicle to enter the conflict area, δ1 is the PET threshold when a pedestrian-vehicle conflict occurs, and δ2 is the PET threshold for a severe conflict. When a vehicle and a pedestrian approach each other at an uncontrolled pedestrian crossing, if PET < δ1, it is identified as a pedestrian-vehicle conflict; if PET < δ2, the pedestrian-vehicle conflict is identified as a severe conflict.

[0053] The present invention will be illustrated below with specific embodiments.

[0054] (1) Preparation of basic data for the simulation model of human-vehicle conflict

[0055] This includes the basic data required for building the simulation environment, the basic data required for modeling pedestrian crossing behavior, and the basic data required for modeling conventional vehicle behavior and autonomous vehicle behavior. The basic data required for building the simulation environment includes: the number of lanes N (specifically 4 in this example), and the lane width W. L (Specifically 3.5m in this example), road segment length L L (Specifically 500m in this example), the width W of the uncontrolled pedestrian crossing. C (Specifically 3.5m in this example), length L of uncontrolled pedestrian crossing C (Specifically 15.2m in this example), speed limit value V for the road section. lim (Specifically 40km / h in this example), vehicle arrival rate λ V (Specifically 0.1 pcu / s in this example), pedestrian arrival rate λ P (Specifically 0.05 ped / s in this example).

[0056] The basic data required for pedestrian crossing behavior modeling includes: pedestrian crossing speed V P (Specifically 1 m / s in this example), pedestrian width W P (Specifically 0.7m in this example), pedestrian length L P (Specifically 0.4m in this example), the safety margin γ for pedestrians facing regular vehicles. ped，CV (Specifically 1.5s in this example), the safety margin γ for pedestrians facing autonomous vehicles. ped，AV (Specifically 1s in this example).

[0057] The basic data required for modeling conventional vehicle behavior and autonomous vehicle behavior includes: vehicle width W V (Specifically 2m in this example), vehicle length L V (Specifically 5m in this example), vehicle speed V V (Updated in the simulation model), vehicle desired speed V0 (specifically 40 km / h in this example), driver reaction time τ CV (Specifically 1.5s in this example), the expected interval distance s * (Specifically 8m in this example), the deceleration d of a conventional vehicle CV (Specifically 2.5 m / s in this example) 2 The maximum acceleration 'a' of a conventional vehicle (specifically 3 m / s² in this example) is... 2 The minimum acceptable safe interval γ between regular vehicles entering / leaving the conflict zone and pedestrians leaving / entering the conflict zone. CV (Specifically 1.4s in this example), the driver's sight distance D for a conventional vehicle CV (Specifically 50m in this example), the penetration rate P of autonomous vehicles AV (Specifically 50% in this example), the deceleration d of the autonomous vehicle AV (Specifically 3m / s in this example) 2 ), the delay time τ of the autonomous driving system AV (Specifically 0.1s in this example), the reference distance s between the autonomous vehicle and the vehicle in front. ref (Specifically 2m in this example), the detection distance D of the autonomous vehicle AV (Specifically 90m in this example).

[0058] (2) Parameter input and simulation model construction, including pedestrian crossing behavior modeling, traditional vehicle behavior modeling, autonomous vehicle behavior modeling, and simulation environment construction.

[0059] ① Pedestrian crossing behavior modeling

[0060] The AnyLogic software was used to model the behavior of pedestrians at uncontrolled crosswalks, and the parameters in step (1) were input. The specific model settings are as follows: When a pedestrian faces a regular vehicle, the pedestrian will act according to the safety margin γ. ped，CV When making a pedestrian crossing decision, the safety margin is the marginal safety value that pedestrians maintain to avoid collisions with vehicles; the conflict zone is the overlapping area where the vehicle trajectory and the pedestrian trajectory intersect, when t ped，leave +γ ped，CV ≤t veh，enter When the conflict zone is considered safe, pedestrians will move at a speed of V. PPedestrians will begin crossing the street at a constant speed; otherwise, they will continue to wait. When faced with an autonomous vehicle, pedestrians will make different crossing decisions based on their familiarity with the vehicle. Autonomous vehicles are equipped with human-machine interfaces (eHMIs) that issue signals to pedestrians waiting to cross the street. When pedestrians are unfamiliar with autonomous vehicles and cannot identify the vehicle type, they will treat it as a regular vehicle. Pedestrians will also make crossing decisions based on safety margins. Therefore, when t... ped，leave +γ ped，CV ≤t veh，enter When the conflict zone is considered safe, pedestrians will move at a speed of V. P Begin crossing the street at a constant speed; otherwise, continue waiting. In the formula, d ped，leave This represents the distance a pedestrian travels from their current location away from the conflict zone. d veh，enter This refers to the distance the vehicle travels from its current location into the conflict zone. Once pedestrians are familiar with the autonomous vehicle, they will first make a crossing decision based on signals from the vehicle's eHMIs. When the received signal is "yield," the pedestrian begins to cross at speed V. P Pedestrians will cross the street at a constant speed; when the signal is "not yielding," pedestrians will continue to wait for an opportunity to cross; if the eHMI does not issue a signal, pedestrians will cross according to the safety margin γ. ped，AV When making a decision to cross the street, when t ped，leave +γ ped，AV ≤t veh，enter When the conflict zone is considered safe, pedestrians will move at a speed of V. P Begin crossing the street at a constant speed; otherwise, continue waiting.

[0061] ② Traditional vehicle behavior modeling

[0062] The behavior of conventional vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in a traditional vehicle behavior model. Specifically, at unsignalized pedestrian crossings...

[0063] When the visual range of a traditional car driver is D CV A pedestrian appears inside, and when t ped，leave +γ CV ≤t veh，enter or t veh，leave +γ CV ≤t ped，enter At that time, the driver chose to yield, reducing speed d CV Begin deceleration; otherwise, the driver accelerates at time t with an acceleration a. IDM Follow the vehicle in front or drive freely at speed V0; where, d ped，enter t represents the distance a pedestrian travels from their current location into the conflict zone. ped，enterThe time it takes for a pedestrian to walk from their current location into the conflict zone. d veh，leave t represents the distance the vehicle has traveled from its current location away from the conflict zone. veh，leave The time taken for the vehicle to leave the conflict zone from its current location; a IDM (t+τ CV ) is t+τ CV The vehicle acceleration at time t, s(t) is the distance between the current vehicle and the vehicle in front at time t, and V is the acceleration of the vehicle at time t. v (t) is the vehicle speed at time t.

[0064] ③ Autonomous vehicle behavior modeling

[0065] The behavior of autonomous vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in an autonomous vehicle behavior model. Specifically, at unsignalized pedestrian crossings, when the detection range D of the autonomous vehicle... AV When there are no pedestrians, the autonomous vehicle accelerates at time t. Follow the vehicle in front or drive freely at speed V0; For the current vehicle at t+τ AV acceleration at time, k a k v and k d All are model parameters, a n-1 (t) represents the acceleration of the vehicle in front at time t, v n-1 (t) represents the speed of the vehicle in front at time t, v V (t) represents the current speed of the vehicle at time t, s V (t) represents the distance between the current vehicle and the vehicle in front at time t.

[0066] When the detection range of the autonomous vehicle is D AV When a pedestrian appears, yield to the pedestrian and reduce speed d. AV Start decelerating.

[0067] ④ Simulation Environment Construction

[0068] Using AnyLogic software, a simulation model of pedestrian-vehicle conflict is constructed based on the specific data of the simulation model in step (1). The uncontrolled pedestrian crossing is located at the midpoint of the road segment, and vehicles start at the beginning of the road with a speed of λ. V The vehicle arrival rate is generated, and the probability that the vehicle is an autonomous vehicle is P. AV And randomly assign less than V lim The initial velocity is removed by the system upon reaching the end of the road; pedestrians move at λ on both sides of the uncontrolled crosswalk. PThe pedestrian arrival rate is generated, and pedestrians are removed by the system when they reach the other side of the uncontrolled crosswalk.

[0069] The simulation model scenario of human-vehicle conflict obtained by combining steps (1) to (2) is as follows: Figure 1 As shown.

[0070] (3) Extraction of pedestrian-vehicle conflicts and assessment of the severity of pedestrian-vehicle conflicts

[0071] Safety alternatives are used to identify and assess the severity of pedestrian-vehicle conflicts. The rear intrusion time (PET) is selected as a substitute safety indicator. PET = T2 - T1, where T1 is the time it takes for the vehicle / pedestrian to leave the conflict area, T2 is the time it takes for the pedestrian / vehicle to enter the conflict area, δ1 is the PET threshold when a pedestrian and vehicle conflict occurs (5s in this example), and δ2 is the PET threshold for a severe conflict (1.5s in this example). When a vehicle and a pedestrian approach each other at an unsignalized crosswalk, if PET < δ1, it is identified as a pedestrian-vehicle conflict; if PET < δ2, the conflict is identified as a severe conflict. An event where a vehicle and a pedestrian approach each other at an unsignalized crosswalk is extracted, PET = T2 - T1 = 3.5s - 1s = 2.5s, indicating that the second pedestrian-vehicle conflict is a non-severe conflict. The other parameters are calculated similarly.

[0072] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for analyzing pedestrian-vehicle conflicts at unsignalized pedestrian crossings under mixed traffic flow, characterized in that, The method for analyzing pedestrian-vehicle conflicts includes the following steps: S1. Obtain the simulation model parameters of the pedestrian-vehicle conflict corresponding to the pedestrian crossing without signal control to be analyzed. The simulation model parameters of the pedestrian-vehicle conflict include the basic data required for the construction of the pedestrian-vehicle conflict simulation environment, the basic data required for pedestrian crossing behavior modeling, and the basic data required for conventional vehicle behavior and autonomous vehicle behavior modeling. S2, construct pedestrian crossing behavior model, traditional vehicle behavior model and autonomous vehicle behavior model respectively, and construct human-vehicle conflict simulation model based on pedestrian crossing behavior model, traditional vehicle behavior model and autonomous vehicle behavior model; import the human-vehicle conflict simulation model parameters in step S1 into pedestrian crossing behavior model, traditional vehicle behavior model, autonomous vehicle behavior model and human-vehicle conflict simulation model respectively. S3 uses a pedestrian-vehicle conflict simulation model to extract pedestrian-vehicle conflict behaviors at uncontrolled pedestrian crossings and evaluates the severity of the extracted pedestrian-vehicle conflict behaviors. Specifically, in step S1, the basic data required for constructing the vehicle-pedestrian conflict simulation environment includes the number of lanes N and the lane width W. L Road segment length L L Uncontrolled pedestrian crossing width W C The length L of the uncontrolled pedestrian crossing C Speed ​​limit V for this road section lim Vehicle arrival rate Pedestrian arrival rate ; The basic data required for modeling pedestrian crossing behavior includes pedestrian crossing speed V. P Pedestrian width W P Pedestrian length L P Safety margin for pedestrians when facing regular vehicles Safety margin for pedestrians when facing autonomous vehicles ; The basic data required for modeling conventional vehicle behavior and autonomous vehicle behavior includes: vehicle width W. V Vehicle length L V Vehicle speed V V Vehicle desired speed V0, driver reaction time Expected interval distance s * The deceleration d of a conventional vehicle CV The maximum acceleration 'a' of a conventional vehicle; the minimum acceptable safe interval between a conventional vehicle entering / leaving the conflict zone and a pedestrian leaving / entering the conflict zone. The driver's sight distance D for conventional vehicles CV The penetration rate of autonomous vehicles P AV The deceleration of autonomous vehicles AV Delay time of autonomous driving system The reference distance s between the autonomous vehicle and the vehicle in front ref Autonomous vehicle detection distance D AV ; Furthermore, in step S2, the process of constructing the pedestrian crossing behavior model includes the following sub-steps: The AnyLogic software was used to model pedestrian behavior at uncontrolled crosswalks, resulting in a pedestrian crossing behavior model. Specifically, when facing regular vehicles, pedestrians adjust their behavior based on safety margins. When making a pedestrian crossing decision, the safety margin is the marginal safety value that pedestrians maintain to avoid collisions with vehicles; the conflict zone is the overlapping area where the vehicle trajectory and the pedestrian trajectory intersect. When the conflict zone is considered safe, pedestrians move at a speed of V. P Begin crossing the street at a constant speed; otherwise, continue waiting. When a pedestrian encounters an autonomous vehicle, if the pedestrian cannot recognize the signal emitted by the vehicle's eHMIs or the eHMIs do not emit a signal, the autonomous vehicle will be treated as a regular vehicle. If the pedestrian recognizes the signal emitted by the autonomous vehicle's eHMIs, a crossing decision will be made based on the signal emitted by the vehicle's eHMIs. When the received signal is "yield," the pedestrian begins crossing at a speed of V. P Pedestrians cross the street at a constant speed; when the signal is "not yielding," pedestrians continue to wait for an opportunity to cross; in the formula, , This represents the distance a pedestrian travels from their current location away from the conflict zone. , This is the time required for a pedestrian to leave the conflict zone from their current location. This represents the distance the vehicle travels from its current location into the conflict zone. The time it takes for the vehicle to travel from its current location into the conflict zone; Input the basic data required for pedestrian crossing behavior modeling in step S1 into the pedestrian crossing behavior model.

2. The method for analyzing pedestrian-vehicle conflicts at unsignalized pedestrian crossings under mixed traffic flow as described in claim 1, characterized in that, In step S2, the construction process of the traditional vehicle behavior model includes the following sub-steps: The behavior of conventional vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in a traditional vehicle behavior model. Specifically, at unsignalized pedestrian crossings... When the visual range of a traditional car driver is D CV Pedestrians appeared inside, and when or At that time, the driver chose to yield, reducing speed d CV Begin deceleration; otherwise, the driver accelerates at time t with an acceleration a. IDM Follow the vehicle in front or drive freely at speed V0; where, , This represents the distance a pedestrian would travel from their current location into the conflict zone. The time it takes for a pedestrian to walk from their current location into the conflict zone. , This represents the distance the vehicle has traveled from its current location away from the conflict zone. The time it takes for the vehicle to travel from its current location away from the conflict zone; yes The vehicle acceleration at time t, and s(t) is the distance between the current vehicle and the vehicle in front at time t. It is the vehicle speed at time t. ; Input the basic data parameters required for conventional vehicle behavior modeling in step S1 into the traditional vehicle behavior model.

3. The method for analyzing pedestrian-vehicle conflicts at unsignalized pedestrian crossings under mixed traffic flow as described in claim 1, characterized in that, In step S2, the process of constructing the autonomous vehicle behavior model includes the following sub-steps: The behavior of autonomous vehicles at unsignalized pedestrian crossings was modeled using AnyLogic software, resulting in an autonomous vehicle behavior model. Specifically, at unsignalized pedestrian crossings, when the detection range D of the autonomous vehicle... AV When there are no pedestrians, the autonomous vehicle accelerates at time t. Follow the vehicle in front or drive freely at speed V0; For the current vehicle acceleration at time, k a , k v and k d All are model parameters, a n-1 (t) represents the acceleration of the vehicle in front at time t, v n-1 (t) represents the speed of the vehicle in front at time t. (t) represents the speed of the vehicle at time t. (t) represents the distance between the current vehicle and the vehicle in front at time t. When the detection range of the autonomous vehicle is D AV When a pedestrian appears, yield to the pedestrian by reducing speed d. AV Begin to decelerate; Input the basic data parameters required for autonomous vehicle behavior modeling in step S1 into the autonomous vehicle behavior model.

4. The method for analyzing pedestrian-vehicle conflicts at unsignalized pedestrian crossings under mixed traffic flow as described in claim 1, characterized in that, In step S2, the process of constructing the human-vehicle conflict simulation model includes the following sub-steps: Using AnyLogic software, a pedestrian-vehicle conflict simulation model is constructed based on the basic data required for building the pedestrian-vehicle conflict simulation environment in step S1. The uncontrolled pedestrian crossing is located at the midpoint of the road segment, and vehicles start at the beginning of the road... The vehicle arrival rate is generated, assuming the probability that a vehicle is an autonomous vehicle is P. AV And randomly assign less than V lim The initial speed is removed by the system upon reaching the end of the road; pedestrians on both sides of the uncontrolled crosswalk... The pedestrian arrival rate is generated, and pedestrians are removed by the system when they reach the other side of the uncontrolled crosswalk.

5. The method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings under mixed traffic flow as described in claim 1, characterized in that, Step S3 involves using a pedestrian-vehicle conflict simulation model to extract pedestrian-vehicle conflict behaviors at uncontrolled pedestrian crossings, and evaluating the severity of the extracted conflicts. This process includes the following sub-steps: Safety alternatives are used to identify and assess the severity of pedestrian-vehicle conflicts. The rear intrusion time (PET) is selected as an alternative safety indicator, where PET = T2 - T1, where T1 is the time it takes for the vehicle / pedestrian to leave the conflict area, and T2 is the time it takes for the pedestrian / vehicle to enter the conflict area. The PET threshold for pedestrian-vehicle conflicts. The PET threshold for severe conflict; if a vehicle and a pedestrian approach each other at an unsignalized crosswalk, when PET < When PET < If so, the conflict between people and vehicles is judged as a serious conflict.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps in the method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings under mixed traffic flow as described in any one of claims 1 to 5.

7. A computer device, characterized in that, The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the method for analyzing pedestrian-vehicle conflicts at uncontrolled pedestrian crossings under mixed traffic flow as described in any one of claims 1 to 5.

Images