Non-signal-control intersection motor vehicle-pedestrian collision risk multi-fusion early warning method in V2X environment

Abstract

The invention discloses an unsignalized intersection motor vehicle-pedestrian collision risk multi-fusion early warning method in a V2X environment, which focuses on the problem of potential collision early warning between a vehicle and a pedestrian in an unsignalized intersection scene, introduces multi-fusion early warning, enhances the prediction efficiency and robustness of a pedestrian trajectory, and improves the prediction accuracy of the pedestrian trajectory. According to the method, the Monte Carlo algorithm, the Social GAN, the Attention mechanism and the multi-dimensional early warning framework of the elliptical buffer domain are fused, and the detection capability of potential collision is remarkably improved. And aiming at the uncertainty of the pedestrian motion track, a Multi-HeadAttention mechanism is adopted to optimize the Social GAN model, the time sequence prediction capability is enhanced, and the problem that the gradient part disappears is solved, so that the pedestrian track prediction is more accurate. In order to solve the problem of balance between the complexity and the accuracy of an early warning algorithm, the K value of a buffer area is dynamically adjusted, the duplication of algorithm functions is realized, the calculation redundancy is reduced, and the accurate recognition of collision points is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic accidents, and in particular to a multi-fusion early warning method for motor vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment. Background Art

[0002] With the rapid development of modern industry, the number of cars in the world continues to rise, and the frequent occurrence of traffic accidents has had a profound impact on the economy and society, making the research on traffic accident prevention more urgent. Especially at intersections without signal control, due to the obstruction of the driver's vision, the high speed of the car, the frequent crossing of non-motor vehicles and pedestrians, and the uncertainty of their trajectories, traffic accidents occur frequently. In order to solve the fierce conflicts between motor vehicles, non-motor vehicles and pedestrians at intersections without signal control, it is necessary to explore a new method of intersection conflict warning in order to significantly reduce the accident rate and further improve the level of traffic safety.

[0003] Intersection Collision Warning (ICW) is an important part of active automotive safety technology. Early collision warning systems mainly relied on radar and cameras to sense the surrounding environment and issue collision warnings to the driver. However, the limitations of traditional sensors (such as obstacle occlusion) limit the effectiveness and further development of collision warning systems. In recent years, the rise of technologies such as vehicle-to-vehicle communication (V2V) and vehicle-to-infrastructure collaboration (V2I) has enabled traffic participants to achieve all-round and real-time information sharing, becoming a new hotspot in collision risk identification and warning research.

[0004] The collision risk at an unsignalized intersection is essentially a problem of motion conflict between traffic participants. Common conflict risk identification indicators are mainly divided into two categories: those based on time-space conflict indicators and those based on probability conflict indicators. The warning method based on time-space conflict indicators predicts potential collision risks by evaluating the temporal and spatial relationship between vehicles. Among them, the time to collision (TTC) is a commonly used evaluation indicator. For example, Miller et al. proposed an intersection collision detection method based on TTC, and performed risk assessment by calculating the TTC of the vehicle. Lan Liuting et al. used TTC, post encroachment time (PET) and deceleration to safety time (DST) as evaluation indicators for the conflict between right-turning vehicles and pedestrians, and found that TTC and PET were negatively correlated with the severity of the conflict, but positively correlated with DST. Guo Ruijun et al. proposed an improved TTC calculation method, which optimized the application of rectangular and elliptical models in conflict detection based on geometric models, and improved the calculation accuracy and real-time performance. In addition, the safety distance warning method evaluates the risk by analyzing the distance change of vehicle movement. Han Yong et al. proposed a longitudinal and lateral TTC difference model for the collision scenario between cars and electric two-wheelers, optimized the triggering strategy of the AEB system, improved the collision avoidance rate and reduced collision damage. Li Xingjia et al. proposed a positioning fusion algorithm based on unscented Kalman filtering, which improved the positioning accuracy and the safety of the vehicle when positioning fails. Zhao Rui et al. proposed an early warning method based on V2I communication that combines dynamic collision distance detection (DDTC) and circular collision detection (CATC), reducing the missed alarm and false alarm rates.

[0005] Although the warning method based on spatiotemporal conflict indicators has the advantages of small computational complexity and simple operation, it relies on a single indicator (such as TTC or PET) and is difficult to cope with complex traffic scenarios. The warning method based on probabilistic conflict indicators introduces uncertainty and probability models, and evaluates the collision risk of vehicles or pedestrians by quantifying them. For example, Laugier et al. used hidden Markov model (HMM) and Gaussian process to identify vehicle behavior, predict future collision risks, and verify its effectiveness in uncertain environments. Joerer et al. evaluated the collision risk of all possible vehicle trajectories through acceleration probability distribution, further improving the detection accuracy. Berthelot et al. proposed a TTC probability distribution calculation method based on uncertain input, and enhanced the reliability of the system by processing input noise. In response to the risk of pedestrian collision, Peng Liqun et al. proposed a random geometric model of the pedestrian collision area based on the V2P communication technology of the Internet of Vehicles, comprehensively considered factors such as communication delay and positioning error, established a pedestrian-vehicle collision probability model, and verified its accuracy in complex scenarios through simulation experiments. Yang Xu et al. constructed a real-time warning model for human-vehicle collision risk based on the catastrophe theory, and evaluated it by combining BinaryLogistic and the catastrophe series method. Jin Yuanyuan et al. proposed a warning method based on the probability of vehicle collision at intersections, which calculated the collision probability by predicting the future driving behavior of vehicles and verified its effectiveness in simulation.

[0006] In summary, most of the previous studies on intersection collision warning have not fully considered the complexity and variability of pedestrian behavior, and ignored the challenges posed by pedestrian irregular behavior to warning. Therefore, it is necessary to improve the existing technology and provide a technical solution that focuses on the potential collision warning problem between vehicles and pedestrians in unsignalized intersection scenarios. Summary of the invention

[0007] Therefore, based on the above background, the present invention provides a multi-fusion warning method for motor vehicle-pedestrian collision risk at a signal-free intersection in a V2X environment, focusing on the potential collision warning problem between vehicles and pedestrians in a signal-free intersection scenario, and by introducing multi-fusion warning technology, enhancing the prediction efficiency and robustness of pedestrian trajectories, thereby significantly improving the accuracy of detection and warning in a signal-free intersection environment, providing strong support for comprehensively improving road safety.

[0008] The technical solution provided by the present invention is:

[0009] A multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment comprises the following steps:

[0010] 1) Collecting and sharing information about motor vehicles and pedestrians; the information includes movement parameters of motor vehicles and pedestrians, and real-time locations of motor vehicles and pedestrians;

[0011] 2) Input the information of motor vehicles and pedestrians into the vehicle trajectory model and pedestrian trajectory model respectively to calculate and obtain the future movement trajectory of the vehicle and pedestrian;

[0012] 3) Based on the future motion trajectories of vehicles and pedestrians, the elliptical buffer algorithm is used to preliminarily screen pedestrians and identify potential collision targets within the buffer;

[0013] 4) The historical trajectory of the identified potential collision target is input into the noisy pedestrian trajectory model to generate a large number of similar predicted trajectories;

[0014] 5) According to the predicted trajectory generated in step 4), the elliptical buffer algorithm is used to adjust the parameter K of the elliptical buffer to calculate and store the potential collision point, potential collision time and potential collision distance between the vehicle and the potential collision target;

[0015] 6) Using the Monte Carlo algorithm to obtain the expected collision value and collision probability based on the potential collision point, potential collision time and potential collision distance obtained in step 5), and combining the total process time TTM from the driver's awareness of danger to the complete stop of the vehicle, the joint collision probability is calculated based on the joint probability formula;

[0016] 7) Compare the combined collision probability obtained in step 6) with a preset warning threshold. If the combined collision probability exceeds the preset threshold, the collision warning system is immediately triggered to alert the driver.

[0017] Further, the motion parameters of the motor vehicle in step S1 include driving speed, driving direction (heading angle), angular velocity (steering angular velocity); the motion parameters of the pedestrian include driving speed, driving direction (heading angle), and steering angle;

[0018] The real-time positions of the motor vehicles and pedestrians include GPS coordinates or roadside relative coordinates.

[0019] Furthermore, the vehicle trajectory model in step S2 is a state vector P constructed based on a two-dimensional plane kinematic model that can update the vehicle's driving direction and position information in real time. V (t), used to predict the future trajectory of the vehicle in the plane;

[0020] The state vector P V (t) See formula (1):

[0021] P V (t)=[x(t),y(t),v x (t),v y (t)] T (1)

[0022] In formula (1), x(t) and y(t) represent the position coordinates of the vehicle on the two-dimensional plane at time t. The formulas of x(t) and y(t) are shown in the following formula (2):

[0023]

[0024] x(t-1) and y(t-1) represent the position of the vehicle at the previous time t-1, and the velocity component of the vehicle at time t is represented by v x (t) and v y (t) indicates that P V Corresponding to the vehicle's speed in the x and y directions respectively.

[0025] Furthermore, the pedestrian trajectory model in step S2 is constructed based on the Social GAN ​​model, by incorporating a multi-head attention mechanism into the generator input part and introducing a diversity loss function.

[0026] Furthermore, the objective function of the adversarial neural network of the pedestrian trajectory model is shown in the following formula (3):

[0027]

[0028] The core of this method is to generate realistic future trajectories through adversarial training between the generator G and the discriminator D. The generator G tries to generate a trajectory G(z) that can "fool" the discriminator D, while the discriminator D tries to distinguish between the real trajectory x and the generated trajectory G(z);

[0029] In the multi-head attention mechanism, the relationship between each time step is calculated as shown in equations (4) to (6):

[0030]

[0031] Z=MultiHead(Q,K,V)(5)

[0032] G(z)=Generator(z,Z)(6)

[0033] Where Q, K, and V are query, key, and value matrices, respectively, T is the transpose, and d k is the dimension of the key; Z is the output of the multi-head self-attention mechanism, G(z) is the trajectory generator, and z represents the latent variable.

[0034] Autonomous attention loss L introduced in pedestrian trajectory model att and diversity loss L var , see the following equations (7) and (8):

[0035]

[0036]

[0037] In the above formula (7), A gen Represents the self-attention matrix generated by the model, A real represents the true self-attention matrix of the model; in formula (8), Y pred (k) represents the kth predicted trajectory. pred (k ' ) represents the k'th predicted trajectory, which is another trajectory generated by the model and is similar to Y pred (k) There are differences between.

[0038] The total loss function of the adversarial neural network is shown in the following formula (9):

[0039] L total =L GAN +λ1L att +λ2L var (9)

[0040] Where λ1 and λ2 are hyperparameters, L GAN is the adversarial loss function.

[0041] Furthermore, in step S4, the prediction trajectory formula generated by introducing noise into the pedestrian trajectory model is shown in the following formula (10):

[0042]

[0043] In the formula, represents the predicted trajectory of the i-th pedestrian of the k-th potential collision target, H k represents the historical trajectory of the input, z is a random noise vector, N(0,σ 2 ) is a Gaussian noise term used to perturb the trajectory at each generation to increase diversity.

[0044] Furthermore, the basic formula of the buffer domain algorithm in steps 3) and 5) is shown in the following formula (11):

[0045]

[0046] In the formula, x c and c are the center points of the ellipse, a and b are the major and minor axes of the ellipse, θ is the rotation angle of the ellipse; x and y are the coordinates of the pedestrian or vehicle target point, respectively;

[0047] The major axis and minor axis of the algorithm are defined as equations (12) to (14):

[0048]

[0049]

[0050] d b =K·v vul ·t rea (14)

[0051] Where L and W are the length and width of the vehicle respectively, d b is the buffer distance used to extend the physical boundary of the vehicle; vul is the speed of the pedestrian, t rea is the driver's reaction time, and K is the adjustment coefficient.

[0052] Furthermore, in step 3), pedestrians are preliminarily screened using the buffer domain algorithm using the following formula (15):

[0053]

[0054] Specifically, the four vertices of the pedestrian rectangle are substituted into the rotation ellipse equation of formula (15) to determine whether it is inside the ellipse;

[0055] In step 5), the potential collision point, potential collision time and potential collision distance between the vehicle and the potential collision target are calculated by the following equations (16) and (17):

[0056] y j =m j x+e j ,j=1,2,3,4(16)

[0057]

[0058] In the above formula, j represents the four sides of the rectangle, where the upper and lower, left and right boundaries are each a straight line; m j is the slope of the jth edge, e j is the intercept of the jth side;

[0059] Specifically, the four vertices and edges of the pedestrian rectangle are substituted into the geometric equation of the rotation ellipse in equation (16) (17).

[0060] Furthermore, the joint probability formula in step 6) is shown in the following formula (24):

[0061] P(t≤TTM,TTC∈R)=P(t≤TTM|TTC∈R)P(TTC∈R)(24)

[0062] The TTC in the above formula represents the time required for a traffic participant to collide, and the joint probability P(t≤TTM, TTC∈R) reflects the situation in which a collision occurs (TTC∈R) and the collision time t is less than TTM in all samples;

[0063] The conditional probability P(t≤TTM / TTC∈R) is used to describe the possibility that the collision time t is less than TTM under the condition that the collision has already occurred (that is, TTC∈R); the collision probability P(TTC∈R) represents the proportion of TTC in all samples that falls into the risk interval R;

[0064] The expected value E(d) of the collision distance and the collision probability P(TTC∈R) calculated by Monte Carlo are shown in the following equations (20) to (22):

[0065] N total =m*h(20)

[0066]

[0067]

[0068] In the above formula (20), h represents the number of potential targets, m represents the number of future trajectories generated by the same potential target; in formula (21), d k represents the collision distance of the kth potential target; N in formula (22) total is the total number of sampling times of the Monte Carlo algorithm, and II is the indicator function used to indicate whether there is a collision;

[0069] In formula (22), is a binary variable representing the collision judgment result, as shown in formulas (18) and (19):

[0070]

[0071]

[0072] In the formula, represents the i-th trajectory of the k-th potential target, m and h are natural integers, and N represents the number of time steps in the trajectory; The elliptical buffer domain formula (16) is used to determine whether the kth potential collision target collides with the vehicle trajectory in the i-th predicted trajectory;

[0073] The calculation of TTM is shown in the following formula (23):

[0074] TTM=t rea +t brak (twenty three)

[0075] In the formula, h represents the number of potential targets, m represents the number of future trajectories generated by the same potential target, and N total is the total number of Monte Carlo sampling, TTM is the driver and vehicle reaction time t rea and braking time t brak The sum of .

[0076] Based on the same inventive concept, the present invention also provides an electronic device, including a memory, a processor, and a program stored in the memory, wherein the processor implements the steps of the above method when executing the program.

[0077] The above technical solution has the following beneficial effects:

[0078] This paper focuses on the potential collision warning problem between vehicles and pedestrians in the unsignalized intersection scenario, introduces multi-fusion warning, enhances the prediction efficiency and robustness of pedestrian trajectories, and thus significantly improves the accuracy of detection and warning in the unsignalized intersection environment:

[0079] In order to cope with the complexity of the collision scene between motor vehicles and pedestrians, the present invention integrates the Monte Carlo (MC) algorithm, Social GAN, Attention mechanism and a multi-dimensional warning framework of elliptical buffer domain, which can significantly improve the detection capability of potential collisions.

[0080] In order to address the uncertainty of pedestrian motion trajectories, the Multi-HeadAttention mechanism is used to optimize the SocialGAN model, which not only enhances the time series prediction capability, but also effectively solves the problem of partial gradient disappearance, making pedestrian trajectory prediction more accurate.

[0081] In order to solve the balance problem between the complexity and accuracy of the early warning algorithm, the present invention introduces an elliptical buffer domain algorithm, and utilizes the K value of the dynamic adjustment buffer area to achieve the duality of the algorithm function, reduce calculation redundancy, and ensure accurate identification of the collision point.

[0082] It has been verified by simulation that in the scenario of an unsignalized intersection, the accuracy of the present invention in warning of motor vehicle-pedestrian collisions is as high as 94.5%, which is significantly better than the traditional TTC model and braking distance model, and also greatly reduces the missed alarm rate and false alarm rate. It can provide a reliable basis for real-time vehicle warnings and the implementation of avoidance measures, and provide a strong guarantee for pedestrian traffic safety at unsignalized intersections. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Attached Figure 1 This is an example of a non-signaled intersection scenario in an embodiment.

[0084] Attached Figure 2 This is an example of a target recognition scenario at an unsignalized intersection according to an embodiment of the present invention.

[0085] Attached Figure 3 YOLOv5s network architecture of an embodiment.

[0086] Attached Figure 4FIG. 4 is a vehicle motion trajectory diagram of an embodiment.

[0087] Attached Figure 5 This is a graph showing the change in the loss function of the training set of the pedestrian trajectory model of the present invention.

[0088] Attached Figure 6 The network framework for pedestrian trajectory prediction of the present invention.

[0089] Attached Figure 7 This is a flow chart of the multi-fusion early warning method of the present invention.

[0090] Attached Figure 8 It is a schematic diagram of collision detection under the elliptical buffer domain algorithm of the present invention.

[0091] Attached Fig. 9 The Monte Carlo algorithm of the present invention utilizes a pedestrian trajectory model to predict trajectories of potential targets.

[0092] Attached Fig.10 Graph showing the collision probability variation under different thresholds of the embodiment.

[0093] Attached Fig.11 Graph 1 is the collision time distribution of the embodiment.

[0094] Attached Fig.12 Graph 2 is the collision distance distribution of the embodiment.

[0095] Attached Fig.13 It is a simulation accident analysis diagram of T=1s of the embodiment, wherein (a) T=1s real collision scene; T=1s simulation collision scene; (c) T=1s joint collision probability change diagram.

[0096] Attached Fig.14 1. The simulated accident analysis diagram of the embodiment at T=1.45s, wherein (a) T=1.45s real collision scene; (b) T=1.45s simulated collision scene; (c) T=1.45s joint collision probability change diagram.

[0097] Attached Fig.15 The simulated accident analysis diagram of T=4.2s of the embodiment, wherein (a) T=4.2s real collision scene; (b) T=4.2s collision simulation scene without warning; (c) joint collision probability change diagram under T=4.2s warning; (d) T=4.2s collision simulation scene under warning. DETAILED DESCRIPTION

[0098] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0099] The technical solution of the present invention:

[0100] A multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment comprises the following steps:

[0101] 1) Through V2X communication technology, the roadside unit RSU is used to collect and share information about motor vehicles and pedestrians; the information includes the movement parameters of motor vehicles and pedestrians and the real-time locations of motor vehicles and pedestrians;

[0102] Through V2X communication technology, motor vehicles exchange data with RSU through OBU (on-board unit), and RSU aggregates information and shares it with all traffic participants in the area; the information includes the movement parameters of motor vehicles and pedestrians, and the real-time locations of motor vehicles and pedestrians;

[0103] The motion parameters of the motor vehicle in step S1 include the travel speed, travel direction (course angle), and angular velocity (steering angular velocity); the motion parameters of the pedestrian include the travel speed, travel direction (course angle), and steering angle;

[0104] The real-time positions of the motor vehicles and pedestrians include GPS coordinates or roadside relative coordinates.

[0105] 2) Using the vehicle trajectory model and the pedestrian trajectory model to obtain the future motion trajectories of the vehicle and the pedestrian according to the information in step 1);

[0106] The vehicle trajectory model in this step is a state vector P that can update the vehicle's driving direction and position information in real time based on a two-dimensional planar kinematic model. V (t), used to predict the future trajectory of the vehicle in the plane;

[0107] The state vector P V (t) See formula (1):

[0108] P V (t)=[x(t),y(t),v x (t),v y (t)] T (1)

[0109] In formula (1), x(t) and y(t) represent the position coordinates of the vehicle on the two-dimensional plane at time t. The formulas of x(t) and y(t) are shown in the following formula (2):

[0110]

[0111] x(t-1) and y(t-1) represent the position of the vehicle at the previous time t-1, and the velocity component of the vehicle at time t is represented by v x (t) and v y (t) indicates that P V They correspond to the speed of the vehicle in the x and y directions respectively, and △t represents the time interval.

[0112] The pedestrian trajectory model is constructed based on the Social GAN ​​model, which incorporates a multi-head attention mechanism into the input of its generator and introduces a diversity loss function.

[0113] The objective function of the adversarial neural network of the pedestrian trajectory model is shown in the following formula (3):

[0114]

[0115] The core of this method is to generate realistic future trajectories through adversarial training between the generator G and the discriminator D. The generator G tries to generate a trajectory G(z) that can "cheat" the discriminator D, while the discriminator D tries to distinguish between the real trajectory x and the generated trajectory G(z);

[0116] In the multi-head attention mechanism, the relationship between each time step is calculated as shown in equations (4) to (6):

[0117]

[0118] Z=MultiHead(Q,K,V)(5)

[0119] G(z)=Generator(z,Z)(6)

[0120] Where Q, K, and V are query, key, and value matrices, respectively, T is the transpose, and d k is the dimension of the key; Z is the output of the multi-head self-attention mechanism, G(z) is the trajectory generator, and z represents the latent variable.

[0121] Autonomous attention loss L introduced in pedestrian trajectory model att and diversity loss L var , see the following equations (7) and (8):

[0122]

[0123]

[0124] In the above formula (7), A gen Represents the self-attention matrix generated by the model, A real represents the true self-attention matrix of the model; in formula (8), Y pred (k) represents the kth predicted trajectory. pred (k ' ) represents the k'th predicted trajectory, which is another trajectory generated by the model and is similar to Y pred (k) There are differences between.

[0125] The total loss function of the adversarial neural network is shown in the following formula (9):

[0126] L total =L GAN +λ1L att +λ2L var (9)

[0127] Where λ1 and λ2 are hyperparameters, L GAN is the adversarial loss function.

[0128] 3) Based on the future motion trajectories of the vehicle and pedestrians obtained in step 2), the pedestrians are preliminarily screened using the elliptical buffer algorithm to identify potential collision targets within the buffer;

[0129] The basic formula of the buffer domain algorithm (rotated ellipse equation) is shown in the following formula (11):

[0130]

[0131] In the formula, x c and c are the center points of the ellipse, a and b are the major and minor axes of the ellipse, θ is the rotation angle of the ellipse; x and y are the coordinates of the pedestrian or vehicle target point, respectively;

[0132] The major axis and minor axis of the algorithm are defined as equations (12) to (14):

[0133]

[0134]

[0135] d b =K·v vul ·t rea (14)

[0136] Where L and W are the length and width of the vehicle respectively, d b is the buffer distance used to extend the physical boundary of the vehicle;vul is the speed of the pedestrian, t rea is the driver's reaction time, and K is the adjustment coefficient.

[0137]

[0138] Specifically, the four vertices of the pedestrian rectangle are substituted into the rotation ellipse equation of formula (15) to determine whether it is inside the ellipse.

[0139] 4) introducing noise into the pedestrian trajectory model and inputting the historical trajectory of the potential collision target identified in step 3) to generate a large number of similar predicted trajectories;

[0140] In this step, the prediction trajectory formula generated by introducing noise into the pedestrian trajectory model is shown in the following formula (10):

[0141]

[0142] In the formula, represents the predicted trajectory of the i-th pedestrian of the k-th potential collision target, H k represents the historical trajectory of the input, z is a random noise vector, N(0,σ 2 ) is a Gaussian noise term used to perturb the trajectory at each generation to increase diversity.

[0143] 5) According to the predicted trajectory generated in step 4), the elliptical buffer algorithm is used to adjust the parameter K of the elliptical buffer to calculate and store the potential collision point, potential collision time and potential collision distance between the vehicle and the potential collision target;

[0144] In this step, the potential collision point, potential collision time and potential collision distance between the vehicle and the potential collision target are calculated by the following equations (16) and (17):

[0145] y j =m j x+e j ,j=1,2,3,4(16)

[0146]

[0147] In the above formula, j represents the four sides of the rectangle, where the upper and lower, left and right boundaries are each a straight line; m j is the slope of the jth edge, e j is the intercept of the jth side;

[0148] Specifically, the four vertices and edges of the pedestrian rectangle are substituted into the geometric equation of the rotation ellipse in equation (16) (17).

[0149] 6) Using the Monte Carlo algorithm to obtain the expected collision value and collision probability based on the potential collision point, potential collision time and potential collision distance obtained in step 5), and combining the total process time TTM from the driver's awareness of danger to the complete stop of the vehicle, the joint collision probability is calculated based on the joint probability formula;

[0150] The joint probability formula in this step is shown in the following formula (24):

[0151] P(t≤TTM,TTC∈R)=P(t≤TTM|TTC∈R)P(TTC∈R)(24)

[0152] The TTC in the above formula represents the time required for a traffic participant to collide, and the joint probability P(t≤TTM, TTC∈R) reflects the situation in which a collision occurs (TTC∈R) and the collision time t is less than TTM in all samples;

[0153] The conditional probability P(t≤TTM / TTC∈R) is used to describe the possibility that the collision time t is less than TTM under the condition that the collision has already occurred (that is, TTC∈R); the collision probability P(TTC∈R) represents the proportion of TTC in all samples that falls into the risk interval R;

[0154] The expected value E(d) of the collision distance and the collision probability P(TTC∈R) calculated by Monte Carlo are shown in the following equations (20) to (22):

[0155] N total =m*h(20)

[0156]

[0157]

[0158] In the above formula (20), h represents the number of potential targets, m represents the number of future trajectories generated by the same potential target; in formula (21), d k represents the collision distance of the kth potential target; N in formula (22) total is the total number of sampling times of the Monte Carlo algorithm, and II is the indicator function used to indicate whether there is a collision;

[0159] In formula (22), C i k is a binary variable representing the collision judgment result, as shown in formulas (18) and (19):

[0160]

[0161]

[0162] In the formula, represents the i-th trajectory of the k-th potential target, m and h are natural integers, and N represents the number of time steps in the trajectory; The elliptical buffer domain formula (16) is used to determine whether the kth potential collision target collides with the vehicle trajectory in the i-th predicted trajectory;

[0163] The calculation of TTM is shown in the following formula (23):

[0164] TTM=t rea +t brak (twenty three)

[0165] In the formula, h represents the number of potential targets, m represents the number of future trajectories generated by the same potential target, and N total is the total number of Monte Carlo sampling, TTM is the driver and vehicle reaction time t rea and braking time t brak The sum of .

[0166] 7) Compare the combined collision probability obtained in step 6) with a preset warning threshold. If the combined collision probability exceeds the preset threshold, the collision warning system is immediately triggered to alert the driver.

[0167] Example 1: This example further illustrates the present invention.

[0168] (1) Construction of vehicle trajectory model and pedestrian trajectory model

[0169] Firstly, the trajectory models of vehicles and pedestrians are constructed.

[0170] In the unsignalized intersection environment, the present invention uses a kinematic model to construct a vehicle trajectory model in view of the strictness of road traffic rules and the relative predictability of the driver's behavior pattern. In contrast, the movement mode, speed and behavior pattern of pedestrians have greater uncertainty and variability. Therefore, in order to more accurately capture and predict the movement path of pedestrians, the present invention uses an adversarial neural network to construct a pedestrian trajectory model.

[0171] At the same time, in order to build a widely applicable pedestrian trajectory prediction model for unsignalized intersections, this paper uses a drone (DJI Mavic 3) to collect field data at multiple unsignalized intersections in Guilin (see Figure 1 By applying the YOLOv5x algorithm, this embodiment extracts valid trajectory information of 350 pedestrians from the collected video data (see Figure 2 Unsignalized intersection target recognition scenario example), target trajectory information after feature extraction (see Figure 3YOLOv5s network architecture) is used as the training and testing data set for this study. During the data collection process, the data set collects images of pedestrians on the road at intervals of 10 frames, including the time frame number, pedestrian number, and pedestrian X and Y axis coordinate information.

[0172] ① Vehicle trajectory model

[0173] The present invention adopts a simplified two-dimensional plane kinematic model (see Figure 4 ), which is used to describe the motion state of a vehicle at an unsignaled intersection, including characteristics such as displacement, velocity, and acceleration. Based on data sharing of V2X communication technology, the motion parameters of the vehicle (including position, velocity, heading angle, etc.) are extracted in real time through RSU to provide key input for collision warning. Using the vehicle motion parameters, the state vector P is constructed based on the kinematic model. V (t) is used to update the vehicle’s driving direction and position information in real time, thereby predicting the vehicle’s future trajectory in the plane. For details, see Formula 1-2.

[0174] P V (t)=[x(t),y(t),v x (t),v y (t)] T (1)

[0175]

[0176] In the formula, x(t) and y(t) represent the position coordinates of the vehicle on the two-dimensional plane at time t, while x(t-1) and y(t-1) represent the position of the vehicle at the previous time t-1. The velocity component of the vehicle at time t is given by v x (t) and v y (t) indicates that P V Corresponding to the vehicle's speed in the x and y directions respectively.

[0177] ② Pedestrian trajectory model

[0178] This paper is based on the Social GAN ​​model, effectively simulates the social interaction between pedestrians through the social pooling layer, and uses the pedestrian historical trajectory data to predict future trajectories. However, the traditional Social GAN ​​model has limitations in capturing complex temporal dependencies and the gradient vanishing problem.

[0179] Therefore, the present invention incorporates a multi-head attention mechanism into the decoder of the Social GAN ​​model. The multi-head attention mechanism has the function of enhancing the temporal modeling capability of the original model and solving the gradient vanishing problem. It can simultaneously focus on multiple key time steps in the historical trajectory data and dynamically allocate attention weights, thereby accurately extracting information that is crucial to future trajectory prediction.

[0180] The optimized model captures the movement trend and potential intention of pedestrians by inputting the first 8 time steps of historical trajectory data in the generator and combining the multi-head attention mechanism. By processing the output of the attention layer through the LSTM encoder, the model can learn the long-term dependencies of pedestrian historical trajectories and generate trajectory predictions for the next 12 time steps based on this.

[0181] On this basis, the optimized model indirectly combines the social pooling layer and the multi-head attention mechanism through the LSTM encoder. The social pooling layer is responsible for modeling the complex interaction relationship between pedestrians, while the multi-head attention mechanism focuses on capturing the temporal dependencies in the historical trajectory. The LSTM encoder integrates the output features from the two modules to generate a comprehensive latent state representation, which can not only effectively capture the interaction relationship between pedestrians, but also significantly improve the model's processing accuracy of temporal dependencies, thereby achieving better performance in trajectory prediction.

[0182] Formula (3) describes the objective function of its adversarial network, the core of which is to generate realistic future trajectories through adversarial training between the generator G and the discriminator D. The generator G tries to generate a trajectory G(z) that can "fool" the discriminator D, while the discriminator D tries to distinguish between the real trajectory x and the generated trajectory G(z).

[0183]

[0184] In the multi-head attention mechanism, the correlation between each time step is calculated, as shown in Formula 4-6.

[0185]

[0186] Z=MultiHead(Q,K,V)(5)

[0187] G(z)=Generator(z,Z)(6)

[0188] Where Q, K, and V are query, key, and value matrices, respectively, and d k The dimension of the key.

[0189] Z is the context information obtained through the multi-head attention mechanism. z represents the latent variable. The generator G(z) uses this information to generate a prediction trajectory that is more consistent with the temporal correlation. In order to deal with the multimodal problem in trajectory prediction, formula 7-8 introduces the autonomous attention loss L att And the diversity loss function L var , which is calculated as follows:

[0190]

[0191]

[0192] In the above formula (7), A gen Represents the self-attention matrix generated by the model, A real represents the true self-attention matrix of the model; in formula (8), Y pred (k) represents the kth predicted trajectory. pred (k ' ) represents the k'th predicted trajectory, which is another trajectory generated by the model and is similar to Y pred (k) There are differences between.

[0193] This loss function encourages the generation of more diverse trajectories by selecting the one that is farthest from the true trajectory among the multiple candidate trajectories generated. Finally, Formula 9 shows the total loss function L of the adversarial neural network. total , and the training set L total The change curve is shown in Figure 5 .

[0194] L total =L GAN +λ1L att +λ2L var (9)

[0195] Where λ1 and λ2 are hyperparameters that control the importance of each loss term. GAN Supervise the authenticity of the generated trajectory, the self-attention mechanism loss function L att The design of should consider the temporal correlation and spatial dependency between the generated trajectory and the real trajectory, while the diversity loss function L var By encouraging the generation of diverse trajectories, the generation of overly similar trajectories is avoided. At the same time, in order to enable the model to better cope with the interactive scenario of unsignalized intersections, the present invention improves the accuracy and diversity of trajectory prediction by introducing a balance between diversity loss and multimodal trajectories.

[0196] Figure 6The network framework for pedestrian trajectory prediction indicates that the model combines the architecture of Multi-HeadAttention and Social GAN, and is used to generate multiple predicted trajectories of vulnerable traffic groups such as single pedestrians from the speed-trajectory time series dataset. Model optimization The model input data includes the historical trajectory, speed and other dynamic features of the target as the basic information of the model. First, the interactive relationship between the basic information of the target is captured by the Multi-HeadAttention module to generate a context vector; then, the attention mechanism is output through the Add&Norm operation to obtain a normalized context vector to ensure information stability and numerical stability of the model; then, the context vector obtained by the Add&Norm operation is input into the LSTM encoder to extract the historical motion features of the target, and the interactive information of multiple targets is aggregated through the Pooling module to generate a global feature representation; on this basis, the LSTM decoder recursively generates trajectories for multiple time steps in the future according to the above features, and outputs multiple predicted trajectories with diversity for each target.

[0197] In addition, in order to further study the diversity of pedestrian trajectory prediction, by introducing noise in the generation process of the pedestrian trajectory prediction model, multiple similar but not identical prediction trajectories can be generated based on a historical trajectory. The injection of noise increases the randomness and uncertainty of trajectory generation, allowing the model to not only capture the trends and patterns of historical trajectories, but also explore the various possibilities in trajectory evolution. Specifically, by adding noise to the output of the generator, the model can generate multiple similar but not identical prediction trajectories based on the same historical trajectory, see Formula 10. Through this mechanism, the optimization model can provide the MC algorithm (Monte Carlo algorithm) with multiple random samples of pedestrian trajectories, increasing the flexibility and reliability of the algorithm.

[0198]

[0199] In the formula, represents the Monte Carlo random sampling of the i-th pedestrian prediction trajectory of the k-th potential collision target, H k represents the historical trajectory of the input, and z usually represents a random noise vector. 2 ) is a Gaussian noise term used to perturb the trajectory at each generation to increase diversity.

[0200] (2) Collision detection and warning

[0201] Based on the vehicle and pedestrian trajectory model constructed above, the following steps are mainly included to realize the multi-fusion warning in the V2X communication environment: screening of potential collision targets in the elliptical buffer area, joint collision probability assessment combining MC and TTM (Total Time to Maneuver), and collision warning threshold judgment.

[0202] The multi-fusion warning method specifically obtains the future motion trajectory of vehicles and pedestrians through information sharing through the roadside unit (RSU) and combining the vehicle and pedestrian trajectory prediction model; then, the pedestrians are preliminarily screened using the elliptical buffer domain algorithm to identify potential collision targets in the buffer domain; then, after screening out the potential collision targets, a large number of similar predicted trajectories of potential collision targets are generated using the pedestrian trajectory prediction model instead of using the traditional random sampling mechanism; then, by adjusting the parameter K of the elliptical buffer domain, the future collision point and future collision distance (potential collision point, potential collision time and potential collision distance) between the vehicle and the potential collision target are calculated and stored using the geometric equation (reference formula 17); then, the collision data is subjected to the Monte Carlo algorithm to obtain the expected value of the collision probability, collision time and distance, and the joint collision probability is calculated based on the joint probability formula (reference formula 24) in combination with the total process time (TTM) from the driver's awareness of danger to the complete stop of the vehicle. Finally, the calculated joint collision probability is compared with the preset warning threshold. If the joint collision probability exceeds the preset threshold, the collision warning system is immediately triggered to alert the driver so that necessary avoidance measures can be taken. The process of the entire multi-fusion early warning method is shown in Figure 7 .

[0203] ① Ellipse buffer algorithm

[0204] The present invention combines the concept of buffer zone with the geometric collision model, introduces an adjustable parameter K, and dynamically adjusts the major and minor axes of the elliptical buffer domain in combination with the speed of pedestrians, the driver's reaction time and vehicle information. An elliptical buffer domain algorithm is constructed to more efficiently detect and screen potential collision risks.

[0205] The elliptical buffer algorithm introduces a buffer area to expand the collision detection boundary of the vehicle and screen out potential conflict targets. At the same time, the algorithm achieves dual functions based on the adjustment of the K value. One is used for preliminary screening and identification of potential collision targets, and the other is used for accurate determination of the collision point. Figure 8 As shown, the yellow elliptical area represents the target motor vehicle, and the blue rectangle represents pedestrians or other vulnerable traffic groups.

[0206] The basic formula of the ellipse buffer algorithm is 11:

[0207]

[0208] In the formula, x c and c are the center points of the ellipse, a and b are the major and minor axes of the ellipse, and θ is the rotation angle of the ellipse.

[0209] The major and minor axes of the algorithm are defined as (12)-(14):

[0210]

[0211]

[0212] d b =K·v vul ·t rea (14)

[0213] Where L and W are the length and width of the vehicle respectively, d b is the buffer distance used to extend the physical boundary of the vehicle. vul is the speed of the pedestrian, t rea is the driver's reaction time, and K is the adjustment coefficient.

[0214] The elliptical buffer algorithm is used to screen potential collision targets:

[0215] Through the above formula, the elliptical buffer algorithm sets a relatively large K value in the initial screening stage to expand the range of the elliptical buffer. This can cover the surrounding traffic participants more widely and ensure that all objects that may pose a threat are identified. Specifically, the algorithm first predicts the future trajectory of vehicles and pedestrians collected by the roadside unit (RSU) and compares it with the expanded elliptical buffer to screen out potential collision targets.

[0216] In the preliminary screening stage, this paper uses the geometric relationship between the ellipse and the rectangle and the inequality condition to quickly screen out the targets that may collide. Specifically, the four vertices of the pedestrian rectangle are substituted into the equation of the rotated ellipse to determine whether it is inside the ellipse, see formula 15.

[0217]

[0218] In the formula, (x c ,y c ) is the coordinate of the center of the ellipse, a and b are the major and minor axes respectively, θ is the rotation angle of the ellipse, (x i ,y i ) are the coordinates of the vertices of the rectangle.

[0219] In the subsequent joint collision probability calculation stage of the elliptical buffer domain algorithm, the K value is reduced to narrow the range of the elliptical buffer domain, and the geometric equation is used to accurately calculate the collision time and collision distance at different time points, providing reliable data support for the subsequent joint collision probability evaluation.

[0220] In the joint collision probability calculation stage, this paper reduces the K value to narrow the range of the elliptical buffer domain, and uses the elliptical buffer domain based on the Monte Carlo algorithm for more accurate collision calculation. On this basis, by substituting the vertices and edges of the rectangle (see Formula 16) into the geometric equation formula of the rotating ellipse (Formula 17), the potential collision point, collision time and collision distance between the vehicle and the target traffic participant can be calculated, see Formulas 13-14.

[0221] y j =m j x+e j ,j=1,2,3,4(16)

[0222]

[0223] In the formula, y j represents the equation of a line, j represents the jth side of a rectangle (the upper and lower, left and right boundaries are each a straight line), m j is the slope of the jth edge, e j is the intercept of the jth side.

[0224] Due to the diversity of traffic participants and the randomness of their behaviors, traditional collision detection methods are difficult to effectively deal with complex interactive scenarios. Therefore, the elliptical buffer domain algorithm constructed by the present invention can improve the early warning capability and robustness in traffic scenarios at unsignalized intersections by expanding the detection boundary and flexibly adjusting the buffer area, especially when dealing with group behaviors of pedestrians, which can improve the adaptability and effectiveness of the early warning algorithm.

[0225] ②Collision probability and warning

[0226] After the potential collision targets are screened out by the elliptical buffer domain algorithm, the scenarios where there is a risk of collision between the vehicle and the pedestrian are identified. In order to accurately quantify the collision time, distance and probability, the Monte Carlo algorithm is then used as the main analysis tool to combine the trajectories of the vehicle and pedestrian trajectory models, and simulate different trajectory scenarios through multiple random sampling to estimate the collision risk between the vehicle and the pedestrian. Specifically, a unique predicted trajectory is obtained from the vehicle trajectory model to represent the future movement path of the vehicle. Multiple possible trajectories are sampled from the pedestrian trajectory prediction model to simulate the uncertain future movement of pedestrians. In each Monte Carlo simulation, a vehicle trajectory and multiple pedestrian trajectories are randomly selected, and the occurrence of collisions is calculated through collision detection, and the frequency of collisions is counted to finally estimate the collision probability (see Formula 22). In addition, in order to improve the accuracy of collision judgment, the present invention introduces the elliptical buffer domain algorithm as a collision judgment criterion, and combines TTM and joint probability analysis to comprehensively evaluate the collision risk of the vehicle.

[0227] Fig. 9The present invention shows that the historical trajectory information of potential collision targets is input into the pedestrian trajectory model, and i pedestrian future trajectory samples are generated for the kth potential target and recorded in (Refer to Formula 18).

[0228] First, in the Monte Carlo random sampling In each time step of each predicted trajectory, the elliptical buffer domain algorithm (see formula 17) is used to detect the relative position relationship between pedestrians and vehicles, so as to make collision judgments. The potential conflict point, collision time and collision distance between the vehicle and the pedestrian are recorded and obtained, and the distribution expectation value E(d) of the collision distance and the collision probability P(TTC∈R) are calculated using the Monte Carlo algorithm. P(TTC∈R) represents the proportion of TTC in all samples that falls into the risk interval R, as shown in the following formula.

[0229]

[0230]

[0231] N total =m*h(20)

[0232]

[0233]

[0234] Where m and h are natural integers, and N represents the number of time steps in the trajectory. is the collision judgment function of the kth potential collision target in the i-th predicted trajectory. k represents the collision distance of the kth potential target, and II represents the indicator function, which is used to indicate whether there is a collision. total is the total number of Monte Carlo samplings.

[0235] Then, after considering the influence of TTM (i.e., the sum of driver reaction time and braking time), the joint collision probability is selected as the final criterion for evaluating collision risk. The joint collision probability is calculated based on the total result of the above trajectory judgment.

[0236] The conditional probability P(t≤TTM / TTC∈R) is used to describe the possibility that the collision time t is less than TTM under the condition that the collision has already occurred (that is, TTC∈R);

[0237] The joint probability P(t≤TTM, TTC∈R) reflects the situation in which a collision occurs (TTC∈R) and the collision time t is less than TTM in all samples, as shown in formula 23-24:

[0238] TTM=trea +t brak (twenty three)

[0239] P(t≤TTM,TTC∈R)=P(t≤TTM|TTC∈R)P(TTC∈R)(24)

[0240] Where TTM is the reaction time between the driver and the vehicle rea and braking time t brak The sum of .

[0241] ②Collision warning

[0242] In this embodiment, the combined collision probability thresholds of 0.25, 0.55 and 0.95 are selected to perform a single-vehicle accident simulation experiment. Fig.10 As can be seen from the figure, when the collision time between the vehicle and the pedestrian is 4.2 seconds, the warning system with a threshold of 0.55 has the best effect. Therefore, this paper determines the optimal joint collision warning threshold to be 0.55. When the collision probability exceeds this value, the system immediately warns the driver to take evasive or braking measures to prevent a collision.

[0243] Embodiment 2: This embodiment performs simulation verification on the present invention.

[0244] For the vehicle collision risk warning at unsignalized intersections and sections, a collision warning algorithm for intersections based on the Internet of Vehicles environment has been constructed in the previous article. In order to verify the reliability of the multi-fusion warning method in unsignalized intersections and sections, this paper designed a simulation experiment based on Matlab and Simulink, and selected several real accident cases in Guilin, Guangxi for verification. The simulation experiment is divided into two parts: case simulation experiment and comparative algorithm simulation experiment. Among them, the case simulation experiment is used to test the effectiveness and practicality of the warning model; the comparative algorithm simulation experiment comprehensively evaluates the reliability of the model from a statistical perspective.

[0245] (1) Case simulation analysis

[0246] In order to verify the effectiveness of the early warning system based on the present invention, this embodiment selects a real accident case for simulation analysis. In this case: the initial speed of the vehicle is 52.72km / h, the braking distance is 17.14m, and the average deceleration (MFDD) is 7.35m / s 2 , the driver and the motor vehicle braking reaction time is 1.240s. In the simulation, the vehicle is set to travel at a constant speed of 12.25m / s from the starting coordinate (16.5, -40.0). At the same time, the pedestrian walks at a normal speed from the starting point (10.67, 3.99), and the velocity components at this moment are Vx = 1.45m / s and Vy = 0.29m / s.

[0247] Fig.11 and Fig.12 It shows that when the simulation time is T=0s, the vehicle calculates the distribution of future collision time and future collision distance through the multi-fusion warning method, and the collision distance and collision time show an obvious normal distribution trend. Fig.11 In the data, most of the sampled collision times are close to 4.2 seconds, indicating that the collision times are mainly concentrated in this time period; Fig.12 It shows that the collision frequency between vehicles and pedestrians reaches a peak in the distance range of 63 meters to 64 meters, which indicates that the collision risk in this distance range increases significantly. In summary, the mean of the distribution indicates that the future collision time between vehicles and pedestrians is 4.2 seconds or the future collision distance is 63.5 meters, which also indicates the existence of potential high-risk collision points.

[0248] In the subsequent simulation time, the experiment used the multi-fusion warning method to calculate the joint collision probability between vehicles and pedestrians in real time, and recorded the simulation scenarios and changes in the joint collision probability at key time nodes (T = 1s, T = 1.45s, T = 4.2s), such as Figure 13-15 shown.

[0249] like Fig.13 As shown, at the moment T=1s, the vehicle just starts to detect the pedestrian and the joint collision probability starts to gradually increase. Fig.13 -a and Fig.13 -b shows the real scene and simulation scene at this time. Fig.13 -c shows the change curve of the joint collision probability. At this time, the joint collision probability has not exceeded the warning threshold of 0.55, and the system is in a monitoring state.

[0250] like Fig.14 As shown, at the moment T = 1.45s, Fig.14 -a and 14-b present the real scene and the simulation scene at this time; Fig.14 -c shows the change in the probability of joint collision without warning. As can be seen from the figure, the probability of joint collision quickly climbed to nearly 98%, far exceeding the warning threshold. This shows that if the vehicle does not brake in time, the possibility of collision with pedestrians is very high.

[0251] like Fig.15 As shown in Figure 1, at the moment of T = 4.2s, it means that the vehicle and the pedestrian have collided in the real accident. Fig.14 -a and 14-b. If the early warning is given, Fig.15 -c shows the change of the joint collision probability after the warning. Since the collision reaction time between the driver and the motor vehicle is 1.240s, the joint collision probability rises rapidly and reaches a peak during this time, but drops significantly and returns to zero due to the driver's timely braking. Fig.15-d, the vehicle and pedestrian successfully avoided collision under the warning. The simulation results verified the effectiveness of the warning system in high-risk situations and proved the practicality of the unsignalized intersection collision warning system based on the Internet of Vehicles in practical applications.

[0252] ②Comparative algorithm verification

[0253] Vulnerable traffic participants such as pedestrians often exhibit unstable behavior on the road, especially at traffic intersections, which are extremely unpredictable. Therefore, vehicles need to maintain a safe distance or provide sufficient reaction time to ensure safety. The braking distance safety distance model and TTC model in the traditional algorithm have certain applicability in a variety of traffic scenarios, especially in predicting and avoiding collisions. Therefore, this embodiment selects these two models as comparison algorithms and conducts simulation experiments in combination with the present invention.

[0254] This embodiment selects real traffic accident data of multiple types of unsignalized intersections in Guilin, Guangxi as simulation scenarios, covering intersections such as crosses, T-shaped intersections and road intersections, to comprehensively evaluate the collision warning accuracy of the three models. In addition, the speed range is set to 30-60km / h according to traffic safety regulations and actual accident data of unsignalized intersections on national and provincial roads. In order to ensure the scientificity of the experiment and the rationality of the data, 200 groups of experiments were designed for pedestrians, as shown in Table 1.

[0255] Table 1 Intersection experimental results of different algorithms

[0256] Number of tests Number of warnings False alarm rate False alarm rate Accuracy The present invention 200 150 4.8％ 7.4％ 94.5％ TTC 200 150 16.2％ 23.1％ 82％ Distance Algorithm 200 150 17％ 18.1％ 86.5％

[0257] Table 1 shows the test results of different algorithms when motor vehicles and pedestrians participate in traffic together. It can be seen from Table 1 that the missed alarm rate of the method of the present invention is 4.8%, which is much lower than 16.2% of the TTC model and 17% of the braking distance model; the false alarm rate is 7.4%, which is significantly lower than 23.1% of the TTC model and 18.1% of the braking distance model; the accuracy rate is as high as 94.5%, which is better than 82% of the TTC model and 86.5% of the braking distance model. The results show that the method of the present invention is superior to the TTC model and the model based on braking distance in terms of missed alarm rate, false alarm rate and accuracy, verifying the reliability of the multi-fusion warning method. In addition, the present invention can accurately predict the collision probability between vehicles and pedestrians, and has the advantages of numerical risk reflection, reasonable prejudgment and flexible evaluation.

[0258] The present invention focuses on the conflict between motor vehicles and pedestrians at unsignalized intersections. With the Monte Carlo algorithm as the core, the present invention integrates the SocialGAN model and the pedestrian trajectory prediction model of the Multi-HeadAttention mechanism, and introduces the elliptical buffer domain algorithm. In the traditional algorithm, the TTC model ignores the lateral movement and dynamic changes of pedestrians, which is prone to false alarms and missed alarms in complex scenes; the braking distance model fails to adjust the warning in time and has limited accuracy when dealing with scenes with large speed changes; in contrast, the present invention can mainly benefit from the application of the collision point identification method of the elliptical buffer domain and the V2X technology. The former improves the accuracy of collision prediction, and the latter enhances the ability to obtain information in complex environments. At the same time, the pedestrian trajectory prediction model of the SocialGAN and Multi-HeadAttention mechanism also improves the system's ability to predict the future trajectory of pedestrians, reducing the occurrence of missed alarms and false alarms. Through comparative experiments, combined with real data and traditional algorithms, the superiority of the algorithm in predicting and warning potential collision risks is effectively verified.

[0259] The present invention and its implementation methods are described above, and such description is not restrictive, and the actual methods are not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design implementation methods and embodiments similar to the technical solution without creativity, which should all fall within the protection scope of the present invention.

Claims

1. A multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment, characterized in that: The steps include: 1) Collecting and sharing information about motor vehicles and pedestrians; the information includes movement parameters of motor vehicles and pedestrians, and real-time locations of motor vehicles and pedestrians; 2) Input the information of motor vehicles and pedestrians into the vehicle trajectory model and pedestrian trajectory model respectively to calculate and obtain the future movement trajectory of the vehicle and pedestrian; 3) Based on the future motion trajectories of vehicles and pedestrians, the elliptical buffer algorithm is used to preliminarily screen pedestrians and identify potential collision targets within the buffer; 4) The historical trajectory of the identified potential collision target is input into the noisy pedestrian trajectory model to generate a large number of similar predicted trajectories; 5) According to the predicted trajectory generated in step 4), the elliptical buffer algorithm is used to adjust the parameter K of the elliptical buffer to calculate and store the potential collision point, potential collision time and potential collision distance between the vehicle and the potential collision target; 6) Using the Monte Carlo algorithm to obtain the expected collision value and collision probability based on the potential collision point, potential collision time and potential collision distance obtained in step 5), and combining the total process time TTM from the driver's awareness of danger to the complete stop of the vehicle, the joint collision probability is calculated based on the joint probability formula; 7) Compare the combined collision probability obtained in step 6) with a preset warning threshold. If the combined collision probability exceeds the preset threshold, the collision warning system is immediately triggered to alert the driver.

2. According to claim 1, a multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment is characterized in that: The motion parameters of the motor vehicle in step S1 include driving speed, driving direction, and angular velocity; the motion parameters of the pedestrian include driving speed, driving direction, and steering angle; The real-time positions of the motor vehicles and pedestrians include GPS coordinates or roadside relative coordinates.

3. According to claim 1, a multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment is characterized in that: The vehicle trajectory model in step S2 is a state vector P constructed based on a two-dimensional plane kinematic model that can update the vehicle's driving direction and position information in real time. V (t), used to predict the future trajectory of the vehicle in the plane; The state vector P V (t) See formula (1): P V (t)=[x(t),y(t),v x (t),v y (t)] T (1) In formula (1), x(t) and y(t) represent the position coordinates of the vehicle on the two-dimensional plane at time t. The formulas of x(t) and y(t) are shown in the following formula (2): x(t-1) and y(t-1) represent the position of the vehicle at the previous time t-1, and the velocity component of the vehicle at time t is represented by v x (t) and v y (t) indicates that P V Corresponding to the vehicle's speed in the x and y directions respectively.

4. According to claim 1, a multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment is characterized in that: The pedestrian trajectory model in step S2 is based on the Social GAN ​​model, which incorporates a multi-head attention mechanism into the generator input and introduces a diversity loss function.

5. According to claim 4, a multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment is characterized in that: The objective function of the adversarial neural network of the pedestrian trajectory model is shown in the following formula (3): The core is to generate realistic future trajectories through adversarial training between the generator G and the discriminator D. The generator G tries to generate a trajectory G(z) that can "cheat" the discriminator D, while the discriminator D tries to distinguish between the real trajectory x and the generated trajectory G(z); In the multi-head attention mechanism, the relationship between each time step is calculated as shown in equations (4) to (6): Z=MultiHead(Q,K,V) (5) G(z)=Generator(z,Z) (6) Where Q, K, and V are query, key, and value matrices, respectively, T is the transpose, and d k is the dimension of the key; Z is the output of the multi-head self-attention mechanism, G(z) is the trajectory generator, and z represents the latent variable. Autonomous attention loss L introduced in pedestrian trajectory model att and diversity loss L var , see the following equations (7) and (8): In the above formula (7), A gen Represents the self-attention matrix generated by the model, A real represents the true self-attention matrix of the model; in formula (8), Y pred (k) represents the kth predicted trajectory. pred (k ' ) represents the k'th predicted trajectory, which is another trajectory generated by the model and is similar to Y pred (k) There are differences between. The total loss function of the adversarial neural network is shown in the following formula (9): L total =L GAN +λ1L att +λ2L var (9) Where λ1 and λ2 are hyperparameters, L GAN is the adversarial loss function.

6. The multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment according to claim 5 is characterized in that: In step S4, the prediction trajectory formula generated by introducing noise into the pedestrian trajectory model is shown in the following formula (10): T i k =G(H k ,z)+N(0,σ 2 ) (10) Where, T i k represents the predicted trajectory of the i-th pedestrian of the k-th potential collision target, H k represents the historical trajectory of the input, z is a random noise vector, N(0,σ 2 ) is a Gaussian noise term used to perturb the trajectory at each generation to increase diversity.

7. The multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment according to claim 1 is characterized in that: The basic formula of the buffer domain algorithm in steps 3) and 5) is shown in the following formula (11): In the formula, x c and c are the center points of the ellipse, a and b are the major and minor axes of the ellipse, θ is the rotation angle of the ellipse; x and y are the coordinates of the pedestrian or vehicle target point, respectively; The major axis and minor axis of the algorithm are defined as equations (12) to (14): d b =K·v vul ·t rea (14) Where L and W are the length and width of the vehicle respectively, d b The buffer distance used to extend the physical boundaries of the vehicle; v vul is the speed of the pedestrian, t rea is the driver's reaction time, and K is the adjustment coefficient.

8. The multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment according to claim 7 is characterized in that: in, In step 3), pedestrians are preliminarily screened using the buffer domain algorithm using the following formula (15): Specifically, the four vertices of the pedestrian rectangle are substituted into the rotation ellipse equation of formula (15) to determine whether it is inside the ellipse; In step 5), the potential collision point, potential collision time and potential collision distance between the vehicle and the potential collision target are calculated by the following equations (16) and (17): y j =m j x+e j ,j=1,2,3,4 (16) In the above formula, j represents the four sides of the rectangle, where the upper and lower, left and right boundaries are each a straight line; m j is the slope of the jth edge, e j is the intercept of the jth side; Specifically, the four vertices and edges of the pedestrian rectangle are substituted into the geometric equation of the rotation ellipse in equation (16) (17).

9. The multi-fusion warning method for vehicle-pedestrian collision risk at a non-signaled intersection in a V2X environment according to claim 1 is characterized in that: The joint probability formula in step 6) is shown in the following formula (24): P(t≤TTM,TTC∈R)=P(t≤TTM|TTC∈R)P(TTC∈R) (24) The TTC in the above formula represents the time required for a traffic participant to collide, and the joint probability P(t≤TTM, TTC∈R) reflects the situation in which a collision occurs (TTC∈R) and the collision time t is less than TTM in all samples; The conditional probability P(t≤TTM / TTC∈R) is used to describe the possibility that the collision time t is less than TTM under the condition that the collision has already occurred (ie, TTC∈R); The collision probability P(TTC∈R) represents the proportion of TTCs in all samples that fall into the risk interval R; The expected value E(d) of the collision distance and the collision probability P(TTC∈R) calculated by Monte Carlo are shown in the following equations (20) to (22): N total =m*h (20) In the above formula (20), h represents the number of potential targets, m represents the number of future trajectories generated by the same potential target; in formula (21), d k represents the collision distance of the kth potential target; N in formula (22) total is the total number of sampling times of the Monte Carlo algorithm, and II is the indicator function used to indicate whether there is a collision; In formula (22), C i k is a binary variable representing the collision judgment result, as shown in formulas (18) and (19): Where, T i k represents the ith trajectory of the kth potential target, m and h are natural integers, and N represents the number of time steps in the trajectory; C i k The elliptical buffer domain formula (16) is used to determine whether the kth potential collision target collides with the vehicle trajectory in the i-th predicted trajectory; The calculation of TTM is shown in the following formula (23): TTM=t rea +t brak (23) In the formula, h represents the number of potential targets, m represents the number of future trajectories generated by the same potential target, and N total is the total number of Monte Carlo sampling, TTM is the driver and vehicle reaction time t rea and braking time t brak The sum of .

10. An electronic device comprising a memory, a processor, and a program stored in the memory, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images