A method for quantitatively evaluating a game behavior perception utility function of a signal-free pedestrian crossing

Abstract

The application discloses a kind of no signal pedestrian crossing human-vehicle game behavior perception utility function quantitative evaluation method, belong to traffic safety and management field;Including analysis and determination of the game subject at no signal pedestrian crossing and setting the perception utility of game subject;Adopt Monte Carlo simulation and human ear perception similarity respectively quantify the perception utility of each game subject;Determine the acceptable perception level of each game subject;The perception utility function of each game subject is established using logistic regression model;The application can quantitatively analyze the problem that pedestrian crossing and motor vehicle yielding behavior at no signal pedestrian crossing mutually confront, mutually game each other, establish the basis for constructing Jacobian matrix and replicator dynamic equation in human-vehicle game model.

Description

Technical Field

[0001] This invention belongs to the field of traffic safety and management, and specifically relates to a quantitative evaluation method for the perceived utility function of human-vehicle game behavior at unsignalized pedestrian crossings. Background Technology

[0002] Unsignaled pedestrian crossings refer to road sections without traffic signals. They are often located in the middle sections of secondary and secondary roads in urban areas, especially around residential areas, schools, and other facilities, and are an important part of the urban road network. Due to the lack of traffic control and unclear right-of-way, motor vehicles and pedestrians are not completely separated in time or space on these road sections. Pedestrians and motor vehicles rely on their own judgment of the surrounding road conditions, environment, and degree of danger, leading to more frequent interactions and conflicts between motor vehicles and pedestrians / non-motorized vehicles. These crossings are the focus of pedestrian-motor vehicle conflicts and can easily cause traffic accidents in severe cases.

[0003] In the interaction between pedestrians and motor vehicles at unsignalized intersections or crosswalks, drivers may decide to slow down or stop to yield to pedestrians, or not to yield at all. Pedestrians, on the other hand, may choose to wait or cross the street. This interaction is a typical game theory process. In actual driving, both pedestrians and drivers often make decisions based on their observations and judgments of the surrounding environment, pedestrian flow, and traffic conditions. Delay and risk are the main factors influencing the crossing decisions of both vehicles and pedestrians.

[0004] To better explain and predict the interactions and decision-making processes among road users, game theory has been introduced into the modeling of human-vehicle interaction behavior. Analysis of the current state of game theory research reveals two main shortcomings in this method for quantitative evaluation of human-vehicle interaction behavior: (1) The utility functions of strategies in current game models mostly rely on simple data assumptions, lacking support from actual traffic data, and cannot accurately describe the utility values ​​of motor vehicle and pedestrian crossing behaviors. (2) Current human-vehicle game theory models lack attention to perceived risk. Drivers' or pedestrians' perceived risk is a major factor guiding their behavioral decisions. Once the perceived risk exceeds the acceptable level, drivers / pedestrians will accelerate or decelerate accordingly to adjust their perceived risk. Therefore, establishing a risk perception model for drivers and determining acceptable levels is crucial for analyzing the behavior of drivers / pedestrians on unsignalized crosswalks. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the technical solution provided by the present invention is as follows: The purpose of the present invention is to provide a quantitative evaluation method for the perceived utility function of pedestrian-vehicle game behavior at unsignalized pedestrian crossings, which can quantitatively analyze the mutual confrontation and game between pedestrian crossing and motor vehicle yielding behavior at unsignalized pedestrian crossings, and establish the foundation for the construction of payoff function and replication dynamic equation in pedestrian-vehicle game model.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0007] A method for quantitatively evaluating the perceived utility function of pedestrian-vehicle game behavior at unsignalized pedestrian crossings includes the following steps:

[0008] S1: Collect video data of the interaction between pedestrians and vehicles at crosswalks without signals, determine the main players in the interaction between pedestrians and vehicles at crosswalks without signals as pedestrians I1 and vehicles I2, and set the perceived utility of each player to include risk utility and delay utility.

[0009] Let the potential collision damage faced by both pedestrians and motor vehicles simultaneously choosing to cross the crosswalk be the risk utility, and let the risk utility of the pedestrian choosing to cross the street be UR. Cross The risk utility of a motor vehicle choosing to cross a river is denoted as UR. NotYield The delay utility is the time spent waiting for pedestrians and motor vehicles to cross the street at a crosswalk to avoid conflict, denoted as UD for the pedestrian's delay utility. ped The delay utility of motor vehicles is UD veh .

[0010] S2: Monte Carlo simulation and human auditory perception similarity are used to quantify the perceived risk of each game player.

[0011] S21: Establish a kinematic model of pedestrians and vehicles during the crossing process at unsignalized pedestrian crossings and derive the criteria for determining pedestrian-vehicle collisions.

[0012] S22: Statistically analyze the distribution of variables involved by game players at unsignalized pedestrian crossings during the crossing process from drone or driving simulation data;

[0013] S23: The collision rate Pcrash during the street crossing process is calculated using Monte Carlo simulation;

[0014] S24: Quantify the perceived risk of game players by drawing on sound measurement and representation methods. P base This serves as a risk benchmark, expressed as the accident rate per million vehicle kilometers.

[0015] S3: Based on the quantification results of perceived utility, determine the acceptable perceived risk level of each game player;

[0016] S31: Obtain the set of perceived risks A of the game players under the yielding condition;

[0017] S32: Obtain the set of perceived risks B of the game players under the condition of not yielding;

[0018] S33: The acceptable perceived risk level of the game players is obtained as A∩B.

[0019] S4: Use logistic regression model to establish the perceived risk utility function and perceived delay utility function of each game player.

[0020] S41: Based on the human-vehicle game scenario and survey data, the model parameters are determined, including: the vehicle speed v at the decision moment. ped Motor vehicle acceleration a veh The distance d between the motor vehicle and the point of conflict veh Pedestrian speed v ped Pedestrian acceleration a ped d, the distance between pedestrians and the conflict point ped The speed difference Δv between motor vehicles and pedestrians, and the squares of each parameter. Δv 2 ;

[0021] S42: Based on the results of the perceived risk quantification, a regression model is constructed and the least squares method is used to determine the weight coefficient and significance of each influencing variable, so as to obtain the perceived risk utility function of each game subject; the standard for significance is that the p value is not less than 0.1.

[0022] S42: The least squares method is used to determine the weight coefficient and significance of each influencing variable, and the delay utility of each game subject is quantified; the standard for significance is that the p-value is not less than 0.1.

[0023] The advantages of this invention are as follows:

[0024] (1) This invention proposes two attributes, “perceived risk utility” and “delay utility”, from the perspectives of safety and efficiency, to evaluate human-vehicle game behavior. It also combines Monte Carlo simulation and human ear perception similarity to quantify the perceived risk utility of the game subjects during the crossing of the street at a pedestrian crossing without signal for the first time.

[0025] (2) This invention proposes for the first time a method for determining the acceptable perceived risk level of game subjects by combining mathematical models.

[0026] (3) Based on measured data, this invention proposes a utility function quantification method for human-vehicle game behavior. Attached Figure Description

[0027] Figure 1 This is a logical structure diagram of the present invention.

[0028] Figure 2This is a schematic diagram of a vehicle-pedestrian collision at a crosswalk without signal control, as described in an embodiment of the present invention.

[0029] Figure 3 This is the distribution of relevant parameters in the embodiments of the present invention.

[0030] Figure 4 This is a distribution map of perceived risks for pedestrians and motor vehicles at an unsignalized crosswalk in an embodiment of the present invention.

[0031] Figure 5 This is a schematic diagram illustrating the acceptable perceived risk level for pedestrians and motor vehicles at an unsignalized crosswalk in an embodiment of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0033] Example 1:

[0034] like Figure 1 As shown, taking the quantification of perceived utility at a pedestrian crossing without a signal as an example, this paper illustrates a method for quantifying and evaluating the perceived utility function of pedestrian and vehicle crossing behavior based on game theory. The main steps are as follows:

[0035] Step S1: Collect video data of the interaction between pedestrians and vehicles at a crosswalk without a signal, determine the main players in the interaction as pedestrians I1 and vehicles I2, and define the perceived utility of each player; the specific steps are as follows:

[0036] (1) Use drones to collect video recordings of pedestrian crossings without signals: manually select the video segments to be captured, that is, segments in the conflict area where pedestrians and motor vehicles may be fighting.

[0037] (2) Determine the game subjects when pedestrians and vehicles are playing against each other at a pedestrian crossing without a signal: pedestrians are game subject I1, and motor vehicles are game subject I2.

[0038] (3) The perceived utility of each game player is defined as including risk utility and delay utility: the potential collision loss faced by both pedestrians and motor vehicles when they choose to cross the crosswalk at the same time is defined as risk utility, and the risk utility of the pedestrian choosing to cross the street is denoted as UR. Cross The risk utility of a motor vehicle choosing to cross a river is denoted as UR. NotYieldThe delay utility is the time spent waiting for pedestrians and motor vehicles to cross the street at a crosswalk to avoid conflict, denoted as UD for the pedestrian's delay utility. ped The delay utility of motor vehicles is UD veh .

[0039] Step S2: Quantify the perceived risk of each game player using Monte Carlo simulation and human auditory perception similarity, respectively; the specific steps are as follows:

[0040] (1) Establish a kinematic model of pedestrians and vehicles during the crossing process at an unsignalized pedestrian crossing and derive the criteria for determining pedestrian-vehicle collisions.

[0041] ① Kinematic model of human and vehicle

[0042] like Figure 2 As shown, assuming the conflict point is the origin, the vehicle's direction of travel is the positive x-axis, the pedestrian's direction of travel is perpendicular to the x-axis, and the line connecting the pedestrian's trajectory and the conflict point is the y-axis. Both the pedestrian and the vehicle maintain their current state of motion and do not engage in any game-theoretic activity during the conflict, i.e., their motion states do not change. Therefore, the kinematic equations for the vehicle and the pedestrian are:

[0043]

[0044]

[0045] In the formula, d ped d is the distance of the pedestrian from the conflict point, in meters (m). veh v is the distance between the vehicle and the point of conflict, in meters. ped v is the speed of a pedestrian, measured in m·s⁻¹. veh The speed of a motor vehicle is expressed in m·s⁻¹; t ped The time it takes for a pedestrian to reach the point of conflict at their current speed, measured in seconds (s); t veh The time, measured in seconds, for a motor vehicle traveling at its current speed to reach the point of conflict. veh and a pe d These are the accelerations of motor vehicles and pedestrians, respectively, in m·s⁻².

[0046] ② Criteria for determining pedestrian-vehicle collision

[0047] When the vehicle arrives at the point of conflict, if the pedestrian's distance from the point of conflict is within the interval [-0.5w, 0.5w], then a collision is considered to have occurred. Therefore, the model for determining a pedestrian-vehicle collision is as follows:

[0048] D ped -V ped ·(t r +t veh)∈[-0.5w,0.5w] (3)

[0049] In the formula, w is the width of the vehicle, which is mainly based on passenger cars on urban roads and is taken as 2.0m.

[0050] (2) Statistically analyze the parameter distribution of game players at unsignaled pedestrian crossings during the crossing process from drone or driving simulation data.

[0051] Based on the kinematic equations of motor vehicles and pedestrians and the collision model, the parameters involved in the perceived risk quantification model include the vehicle's speed (v). veh ), distance of vehicle from conflict point (d) veh ), vehicle acceleration (a veh Pedestrian arrival speed (v) ped ), pedestrian distance from conflict point (d) ped ), pedestrian acceleration (a ped ), reaction time (t) r Based on pedestrian-motor vehicle one-to-one game data extracted by drones, the probability distributions of each parameter were obtained. The distributions of each parameter are as follows: Figure 3 As shown.

[0052] like Figure 3 As shown in (a), v is calculated with a step size of 3 m·s⁻¹. veh The probability mass function. Assuming that samples within each interval are equally likely to be drawn, to ensure sampling randomness, a uniform distribution with a step size of 0.5 is added to the sampled values, i.e.:

[0053]

[0054] In the formula, v veh,sample v is the final sampled value of the vehicle's arrival speed. veh,draw The vehicle speed value is extracted from the probability mass function. For a random term following a uniform distribution, Uniform(-1.5, 1.5) represents a uniform distribution with a lower bound of -1.5 and an upper bound of 1.5. Similarly, we obtain d. veh v veh a veh d ped v ped a ped The final sampled values ​​of the 6 parameters.

[0055] For reaction time t r Reaction times were collected using the Silab driving simulator, and the distribution of the reaction time samples is as follows. Figure 3 As shown in (g). The reaction time follows a normal distribution with a mean of 0.5652 s and a variance of 0.1358 s. Generally, t... rTake 0.6s.

[0056] (3) The collision probability during the crossing process was calculated using Monte Carlo simulation.

[0057] The input parameters are randomly sampled and analyzed based on the simulation parameter distribution function. A Monte Carlo simulation model is used to simulate the collision probability (also known as objective operational risk) of drivers and pedestrians crossing the street while maintaining their current motion state. The number of simulation samplings is set to 106. The ratio of the number of samples that meet the collision conditions in the simulation to the total number of simulations is the collision probability Pcrash of pedestrian-vehicle interaction behavior.

[0058] (4) Quantify the perceived risk of game players by drawing on sound measurement and representation methods.

[0059] To accurately measure the perceptual ambiguity of risk, based on the calculation of objective operational risk probability, and drawing on the correlation between sound measurement and representation methods (physically measurable sound pressure and relative quantities that conform to human auditory perception), a perceptual risk measurement method conforming to human physiological perception characteristics is proposed by introducing the definition of decibels (the logarithm of the ratio of the measured value to the baseline value), as shown in Equation (5). That is, the logarithm of the ratio of objective operational risk to baseline operational risk, in decibels (dB).

[0060]

[0061] In the formula, PRI is the perceived risk index of human-vehicle interaction behavior, in dB; P crash The collision probability of this interaction is calculated using the Monte Carlo simulation method based on the pedestrian and vehicle operating states; P base This serves as a risk benchmark, expressed as the accident rate per million vehicle kilometers. According to relevant data, the United States requires an accident rate of no more than 0.5 accidents per million kilometers, while the European Union sets a standard of no more than 1.0 traffic accidents per million kilometers. The average of these two values ​​(0.75 * 10⁻⁶) is taken as the basic risk value.

[0062] Taking a pedestrian crossing sample with a speed of 4.12 km / h, an acceleration of 0.0001 m / s⁻², a distance of 5.78 m from the conflict point, a vehicle speed of 23.16 km / h, an acceleration of -0.72 m / s⁻², and a distance of 33.95 m from the conflict point as an example, and with 106 simulations, the collision probability is 0.216, meaning the perceived risk of this interaction is 5.46 dB.

[0063] Step S3: Based on the quantification results of perceived utility, determine the acceptable perceived risk level for each game player. The specific steps are as follows:

[0064] (1) Obtain the set of perceived risks A of the game players under the yield condition.

[0065] The perceived risk distribution of motor vehicles under the yield condition (i.e., pedestrians cross the conflict point first) can be obtained based on the constructed perceived risk measurement model. Then, the perceived risk set A1 of motor vehicles under the yield condition can be determined based on the minimum and maximum values ​​of the distribution.

[0066] The perceived risk distribution of pedestrians under the yield condition (i.e., the motor vehicle passes the conflict point first) can be obtained based on the constructed perceived risk measurement model. Then, the perceived risk set A2 of pedestrians under the yield condition can be determined based on the minimum and maximum values ​​of the distribution.

[0067] (2) Obtain the set of perceived risks B of the game players under the condition of not yielding.

[0068] The perceived risk distribution of motor vehicles under the non-yielding condition (i.e., the motor vehicle passes the conflict point first) can be obtained based on the constructed perceived risk measurement model. Then, the perceived risk set B1 of motor vehicles under the non-yielding condition can be determined based on the minimum and maximum values ​​of the distribution.

[0069] The perceived risk distribution of pedestrians under the non-yield condition (i.e., pedestrians first cross the conflict point) can be obtained based on the constructed perceived risk measurement model. Then, the perceived risk set B2 of pedestrians under the non-yield condition can be determined based on the minimum and maximum values ​​of this distribution.

[0070] (3) The acceptable perceived risk level of the game players is obtained as A∩B.

[0071] The intersection of the perceived risk sets of a motor vehicle under the two conditions of yielding and not yielding, A1∩B1=C1[a,b], represents the driver's acceptable perceived risk level. When the perceived risk is within the range of [a,b], a game will occur between the motor vehicle and the pedestrian, meaning the driver will make a decision to yield or preemptively strike before reaching the conflict point. When PRI≥b, the driver loses the game with the pedestrian and chooses to slow down or stop and wait. When PRI≤b, the driver has the upper hand in the game and chooses to accelerate or maintain a constant speed to cross the crosswalk before the pedestrian reaches the conflict point.

[0072] The intersection of the pedestrian's perceived risk sets under the two conditions of yielding and not yielding, A2∩B2=C2[m,n], represents the pedestrian's acceptable perceived risk level. When the perceived risk is within the interval [m,n], the pedestrian will engage in a game with the vehicle, deciding whether to slow down and wait or accelerate before reaching the conflict point. When PRI≥m, the pedestrian loses the game and tends to yield and wait to cross the street. When PRI≤b, the pedestrian has the upper hand in the game and chooses to accelerate or cross at a constant speed.

[0073] This embodiment collected 1199 pedestrian-motor vehicle one-to-one conflicts, and the perceived risk distribution was obtained based on the perceived risk measurement model as follows: Figure 4 As shown. By Figure 4(a) It can be known that the perceived risk range when the motor vehicle yields is [1.13dB, 5.85dB]. Figure 4 (b) The perceived risk range for a motor vehicle when it does not yield is [-2.52dB, 2.48dB]. Therefore, the intersection of the perceived risk ranges for motor vehicles under the two conditions of yielding and not yielding when pedestrians cross the street is [1.13dB, 2.48dB], that is, the acceptable perceived risk level for motor vehicles is [1.13dB, 2.48dB]. Figure 5 As shown in (a).

[0074] Depend on Figure 4 (c) It can be seen that the perceived risk range when the pedestrian yields is [0dB, 6.13dB]. Figure 4 (d) The perceived risk range for the pedestrian when not yielding is [-1.92dB, 1.85dB]. Therefore, the intersection of the perceived risk ranges for the pedestrian under the two conditions of yielding and not yielding is [0dB, 1.85dB], that is, the acceptable perceived risk level for the pedestrian is [0dB, 1.85dB]. Figure 5 As shown in (b).

[0075] Step S4: Use a logistic regression model to establish the perceived risk utility function and perceived delay utility function for each game player. The specific steps are as follows:

[0076] (1) Determine model parameters by combining human-vehicle game scenario and survey data.

[0077] Fourteen potential factors that may affect perceived risk were selected as model parameters, including the vehicle speed v at the decision time. ped Motor vehicle acceleration a veh The distance d between the motor vehicle and the point of conflict veh Pedestrian speed v ped Pedestrian acceleration a ped d, the distance between pedestrians and the conflict point ped The speed difference Δv between motor vehicles and pedestrians, and the squares of each parameter. Δv 2 .

[0078] (2) Construct the risk perception utility function of each game player

[0079] First, a regression model was constructed using the quantified risk perception results of motor vehicles and pedestrians as dependent variables and 14 parameters. Then, the least squares method was used to determine the weight coefficient and significance of each influencing variable. Finally, a significance level of 0.1 was selected, and stepwise regression was used to determine significant variables, thus obtaining the risk perception utility function of each game subject.

[0080] Table 1. Parameter estimation results of the risk utility model.

[0081]

[0082]

[0083] Table 1 shows that, at a 90% confidence level, vehicle speed, pedestrian distance from the conflict point, vehicle acceleration, and vehicle speed significantly affect the pedestrian's perceived risk; while vehicle distance from the conflict point and the speed difference between the pedestrian and the vehicle significantly affect the vehicle's perceived risk. Based on this, the risk utility function for the vehicle's choice to cross the street can be expressed as follows:

[0084] UR NotYield =b1Δv 2 +b2d veh = -0.0178Δv 2 +0.0620d veh

[0085] The risk utility function of the pedestrian choosing to cross the street can be expressed as:

[0086]

[0087] (3) Construct the perceived delay utility function for each game player

[0088] Pedestrian and motor vehicle delays refer to the waiting time for pedestrians and motor vehicles to cross the street at the crosswalk to avoid conflict by yielding to each other; that is, the difference between the actual travel time and the driver's theoretical crossing time. In actual decision-making, if the driver (pedestrian) does not need to slow down, the driver (pedestrian) delay value UD = 0; if the driver (pedestrian) needs to slow down, then the delay UD > 0.

[0089] First, based on the delay values ​​from the on-site survey of the road section, a regression model was constructed with 14 parameters. Then, the least squares method was used to determine the weight coefficients and significance of each influencing variable. Finally, a significance level of 0.1 was selected, and stepwise regression was used to determine the significant variables, thus obtaining the delay utility of each game player.

[0090] Table 2. Parameter estimation results for the delay utility model.

[0091]

[0092] Based on this, the delay utility function of motor vehicles can be expressed as:

[0093]

[0094] The pedestrian's delay utility function can be expressed as:

[0095] UD ped =f1v veh+f2d ped +f3d veh =-0.4651v veh -0.7965d ped +0.1170d veh

[0096] Finally, it should be noted that the specific embodiments described above are merely illustrative of the present invention, enabling those skilled in the art to understand and apply it, and are not intended to limit the invention. Any modifications or improvements made by those skilled in the art to the above embodiments within the spirit and principles of the present invention, based on the technical essence of the present invention, should be within the protection scope of the present invention.

Claims

1. A method for quantitatively evaluating the perceived utility function of pedestrian-vehicle game behavior at unsignalized pedestrian crossings, characterized in that, Includes the following steps: S1. Collect video data of the interaction between pedestrians and vehicles at crosswalks without signals, determine the main players in the interaction between pedestrians and vehicles at crosswalks without signals as pedestrians I1 and vehicles I2, and set the perceived utility of each player. S2. Monte Carlo simulation and human auditory perception similarity are used to quantify the perceived risk of each game player. The Monte Carlo simulation and human auditory perception similarity are used to quantify the perceived risk of each game player, and the specific methods are as follows: S21. Establish a kinematic model of pedestrians and vehicles during the crossing process at unsignalized pedestrian crossings and derive the criteria for determining pedestrian-vehicle collisions. S22. Statistically analyze the distribution of variables involved by game players at unsignalized pedestrian crossings during the crossing process from drone or driving simulation data; S23. The collision rate Pcrash during the street crossing process is calculated using Monte Carlo simulation. S24. Quantify the perceived risk of game players by drawing on sound measurement and representation methods: ;in, This serves as a risk benchmark, expressed as the accident rate per million vehicle kilometers. S3. Based on the quantification results of perceived utility, determine the acceptable perceived risk level of each game player; S4. Use logistic regression model to establish the perceived risk utility function and perceived delay utility function of each game player; The specific method for establishing the perceived risk utility function and perceived delay utility function for each game player is as follows: S41. The model parameters are determined by combining the human-vehicle game scenario and survey data, including: the speed of the vehicle at the decision moment. Motor vehicle acceleration Distance of motor vehicles from the point of conflict Pedestrian speed Pedestrian acceleration Distance between pedestrians and conflict points Speed ​​difference between motor vehicles and pedestrians and the squares of each parameter , , , , , , ; S42. Based on the results of the perceived risk quantification, a regression model is constructed and the least squares method is used to determine the weight coefficient and significance of each influencing variable, so as to obtain the perceived risk utility function of each game subject. The standard for significance is that the P value is not less than 0.

1. S43. The delay utility of each game player is quantified using the least squares method. The significance criterion is that the P-value is not less than 0.

1.

2. The method for quantitatively evaluating the perceived utility function of pedestrian-vehicle game behavior at a signalless pedestrian crossing as described in claim 1, characterized in that, In step S1, the perceived utility of each game player includes risk utility and delay utility; the potential collision loss faced by both pedestrians and motor vehicles simultaneously choosing to cross the crosswalk is considered risk utility, and the risk utility of the pedestrian choosing to cross the street is denoted as... The risk utility of a motor vehicle choosing to cross a river is denoted as ; The waiting time for pedestrians and vehicles to cross the street at a crosswalk in order to avoid conflict is called the delay utility. Let the delay utility of the pedestrian be denoted as . The delay utility of motor vehicles is .

3. The method for quantitatively evaluating the perceived utility function of pedestrian-vehicle game behavior at a signalless pedestrian crossing as described in claim 1, characterized in that, In step S3, based on the quantification results of perceived utility, the acceptable perceived risk level of each game player is determined. The specific method is as follows: S31. Obtain the set of perceived risks A of the game players under the condition of yielding; S32. Obtain the set of perceived risks B of the game players under the condition of not yielding; S33, The acceptable perceived risk level of the game players is obtained as follows: .

Images