A method, system, device and medium for predicting highway pedestrian crossing time

Abstract

The application provides a highway pedestrian crossing time prediction method, system, device and medium, different corresponding relaxation times of pedestrians crossing roads are set based on a kinematic model, preliminary prediction results of pedestrians crossing roads are estimated, real-time information of pedestrians crossing roads is continuously collected, real-time prediction results of pedestrians crossing roads are obtained based on Kalman filter trajectory prediction, a difference between the preliminary prediction results and the real-time prediction results is compared, a combined prediction method is formed, sufficient safety margin can be provided for vehicles receiving active safety warning, the probability of collision between vehicles driving on the highway and operating personnel or illegal crossing personnel is reduced, and road safety is increased; the prediction of pedestrian crossing time and the real-time monitoring of pedestrian movement can be realized, the demand for pedestrian safety, pedestrian movement randomness and vehicle active anti-collision can be met while the prediction accuracy is met; the safety of vehicles and pedestrians on the highway is improved, and new safety protection measures are provided.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of active safety, in particular to a highway pedestrian crossing time prediction method, system, device and medium. BACKGROUND

[0002] The vehicle driving speed is fast in the highway scene, and there are often pedestrians and workers who violate the regulations to cross the guardrail or cross the central separation belt when passing through the village gathering or construction section, which is easy to form a serious vehicle-person conflict. Data shows that vehicle-person collision accidents account for 28% of all highway traffic accidents, with a fatality rate of about 65%. Studies have shown that active safety control of vehicles with the help of V2X technology can effectively reduce the number of vehicle-person collision accidents and the severity, one of the core technologies is pedestrian motion trajectory prediction.

[0003] The current research on vehicle-person active collision avoidance application has the following shortcomings:

[0004] 1. The existing research is a pedestrian motion trajectory prediction method based on conflict points, which determines the vehicle-person collision avoidance strategy by predicting the intersection of the pedestrian motion trajectory and the vehicle driving trajectory to form a conflict point, and its action range is often limited to a certain lane of the road, which is suitable for short-distance vehicle-person active collision avoidance in urban low-speed traffic scenes. For fast traffic scenes on highways, the higher the vehicle driving speed, the farther the safe distance between the vehicle and the pedestrian needs to be maintained. The conflict point-based method increases the risk of vehicle-person collision and makes it difficult to ensure the safety of vehicles and pedestrians.

[0005] 2. The existing research focuses on how to improve the accuracy of pedestrian trajectory prediction, and it is difficult to achieve long-time range trajectory prediction. Although short-time range prediction can meet the accuracy requirements, it still has a high risk of vehicle-person collision in fast traffic scenes, and the real-time active control of the vehicle in the active collision avoidance strategy is required.

[0006] 3. The existing research focuses on short-time range pedestrian motion trajectory prediction, and there is currently no method for predicting the crossing time of pedestrians crossing the road. SUMMARY

[0007] In view of the problems in the prior art, the present application provides a highway pedestrian crossing time prediction method, system, device and medium, which solves the demand for pedestrian safety, pedestrian motion randomness and vehicle active collision avoidance.

[0008] The present application is realized by the following technical solutions:

[0009] A highway pedestrian crossing time prediction method, comprising the following steps:

[0010] The initial information of the pedestrian crossing the road is collected to input a kinematic model, different relaxation times corresponding to different pedestrians crossing the road are set based on the kinematic model, and a preliminary prediction result of the pedestrian crossing the road is predicted;

[0011] Real-time information of the pedestrian crossing the road is continuously collected, and real-time prediction results of the pedestrian crossing the road are obtained based on Kalman filter trajectory prediction, and a difference between the preliminary prediction result and the real-time prediction result is compared;

[0012] Based on the difference between the preliminary prediction result and the real-time prediction result and the relaxation time, the estimated information of the pedestrian crossing the road is continuously reset to obtain a car-person conflict time window.

[0013] Further, the car-person conflict time window is Where t0 is the time when the pedestrian starts to cross the road, t f is the time when the pedestrian ends to cross the road, then the pedestrian crossing time is:

[0014] t T =t f -t0.

[0015] Further, the construction of the kinematic model is based on the data of the pedestrian crossing the road collected in advance, the pedestrians with different crossing times are divided into three categories of aggressive, general and conservative, the overall mean of the crossing time of the three categories of pedestrians is estimated, and the upper limit value of the crossing time corresponding to the different categories of pedestrians is determined based on the Laiyida criterion;

[0016] The estimation of the overall mean of the crossing time of the three categories of pedestrians includes the following steps:

[0017] The overall mean and variance of the i-th category of pedestrians are calculated as μ i and The distribution of the crossing time data of the different categories of pedestrians is tested;

[0018] Based on the data distribution characteristics and the sub-sample mean and variance, the overall mean and variance are estimated to obtain the confidence interval of the overall mean μ i with a confidence level of 1-α;

[0019] The upper limit of the confidence interval is taken as the final μ i estimation value, and then the overall mean of the crossing time is obtained.

[0020] Further, the determination of the upper limit value of the crossing time corresponding to the different categories of pedestrians based on the Laiyida criterion includes the following steps:

[0021] The pedestrian crossing time is (μ i -3σ i , μ i +3σ i) is 0.9974, and the upper limit value of the crossing time corresponding to different categories of pedestrians is determined:

[0022]

[0023] wherein σ i is the relaxation time.

[0024] Further, the process of setting the corresponding relaxation time of different pedestrians crossing the road based on the kinematic model comprises the following steps:

[0025] According to the obtained initial state information of pedestrians crossing the road, the crossing time t T of the pedestrians is predicted by the kinematic model.

[0026] By comparing t T ' with the upper limit value of the crossing time corresponding to different categories of pedestrians, the category of the pedestrian is determined, and the preliminary prediction result and the corresponding relaxation time are determined.

[0027] If t ' is greater than the upper limit value of the crossing time corresponding to the aggressive type of pedestrians, the preliminary prediction result is t ' and the relaxation time σ1.

[0028] If t ' is greater than the upper limit value of the crossing time corresponding to the general type of pedestrians, the preliminary prediction result is t ' and the relaxation time σ2.

[0029] If t ' is greater than the upper limit value of the crossing time corresponding to the conservative type of pedestrians, the preliminary prediction result is t ' and the relaxation time σ3.

[0030] Further, the process of continuously collecting real-time information of pedestrians crossing the road comprises:

[0031] The state vector of the pedestrian at time t-1 is X t-1 ,

[0032] X t-1 = (p, v) T ;

[0033] wherein p is the position of the pedestrian at the current time, and v is the speed of the pedestrian at the current time.

[0034] Further, the process of calculating the real-time prediction result of the pedestrian comprises:

[0035]

[0036]

[0037] wherein X is the state estimation value of the pedestrian at time t; n is the number of lanes, and w​r is the lane width; is the average angle between the walking direction of the pedestrian and the cross section of the road;

[0038] The real-time prediction result is: t-t0+t K ;

[0039] Based on the difference between the preliminary prediction result and the real-time prediction result, it is determined whether the relaxation time needs to be reset.

[0040] If t0+t , the pedestrian crossing time prediction result remains unchanged; if t0+t , the pedestrian crossing time prediction result is reset: and so on, so that the final pedestrian crossing time prediction result is:

[0041]

[0042] Wherein m is the reset number, and the prediction result of the vehicle-pedestrian conflict time window is

[0043] A highway pedestrian crossing time prediction system based on a highway pedestrian crossing time prediction method, comprising:

[0044] A preliminary prediction result module for collecting initial pedestrian crossing road information, inputting a kinematic model, setting corresponding relaxation times for different pedestrians crossing roads based on the kinematic model, and preliminarily predicting the pedestrian crossing road information;

[0045] A prediction result difference module for continuously collecting real-time information of pedestrians crossing roads and obtaining real-time prediction results of pedestrians crossing roads based on Kalman filter trajectory prediction, and comparing the difference between the preliminary prediction result and the real-time prediction result;

[0046] A vehicle-pedestrian conflict time window module for continuously resetting the estimated pedestrian crossing road information based on the difference between the preliminary prediction result and the real-time prediction result, and obtaining the vehicle-pedestrian conflict time window.

[0047] A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of a highway pedestrian crossing time prediction method when executing the computer program.

[0048] A computer readable storage medium storing a computer program, characterized in that the computer program is executed by a processor to implement the steps of a highway pedestrian crossing time prediction method.

[0049] Compared with the prior art, the present application has the following beneficial technical effects:

[0050] The application provides a high-speed road pedestrian crossing time prediction method, system, device and medium, different pedestrian crossing road corresponding relaxation times are set based on a kinematic model, a preliminary prediction result of pedestrian crossing road is estimated, real-time information of pedestrian crossing road is continuously collected, and a real-time prediction result of pedestrian crossing road is obtained based on Kalman filter trajectory prediction, a difference between the preliminary prediction result and the real-time prediction result is compared, a combined prediction method is formed, sufficient safety margin can be provided for a vehicle receiving active safety warning, a probability of collision between a vehicle driving on a highway and a worker or a person crossing in violation of rules is reduced, and road safety is increased; the pedestrian crossing time prediction and real-time monitoring of pedestrian movement can be realized, pedestrian movement prediction in a long time range can be realized while meeting prediction accuracy, and the needs of pedestrian safety, pedestrian movement randomness and vehicle active anti-collision are met; the safety of vehicles and pedestrians on the highway is improved, and new safety protection measures are provided. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 A high-speed road pedestrian crossing time prediction method flowchart is provided for the application.

[0052] Figure 2 A high-speed road pedestrian crossing time prediction method flowchart is provided for the application. DETAILED DESCRIPTION

[0053] The application will be further described in detail below in combination with specific embodiments, which are an explanation of the application rather than a limitation.

[0054] In order for those skilled in the art to better understand the application, the technical solutions in the embodiments of the application will be described clearly and completely in combination with the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should belong to the protection scope of the application.

[0055] It should be noted that the terms "first", "second", and the like in the description and in the claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0056] The present application provides a highway pedestrian crossing time prediction method, as shown in the following steps: Figure 1 The present application provides a highway pedestrian crossing time prediction method, as shown in the following steps:

[0057] Collecting initial information of pedestrians crossing the road and inputting the kinematic model, setting corresponding relaxation time of pedestrians crossing the road based on the kinematic model, and estimating the preliminary prediction result of pedestrians crossing the road;

[0058] Continuously collecting real-time information of pedestrians crossing the road and obtaining real-time prediction result of pedestrians crossing the road based on Kalman filter trajectory prediction, and comparing the difference between the preliminary prediction result and the real-time prediction result;

[0059] Based on the difference between the preliminary prediction result and the real-time prediction result and the relaxation time, the estimated information of pedestrians crossing the road is constantly reset to obtain the vehicle-pedestrian conflict time window.

[0060] Preferably, the vehicle-pedestrian conflict time window is Wherein, t0 is the time when the pedestrian starts to cross the road, t f is the time when the pedestrian ends to cross the road, then the pedestrian crossing time is:

[0061] t T = t f -t0.

[0062] Specifically, the pedestrian crossing time t T The pedestrian crossing time can be preliminarily determined by the kinematic model, and reset based on the difference between the preliminary prediction result and the real-time prediction result. The present application does not include pedestrian crossing intention prediction, so the vehicle-pedestrian conflict time window prediction is the prediction of the pedestrian crossing time t T ;

[0063] Preferably, the construction of the kinematic model is based on the data of pedestrians crossing the road collected in advance to divide pedestrians with different crossing times into three categories of aggressive, general and conservative, estimate the overall mean of the crossing time of the three categories of pedestrians, and determine the upper limit value of the crossing time corresponding to different categories of pedestrians based on the Wald criterion;

[0064] The estimation of the overall mean of the crossing time of the three categories of pedestrians includes the following steps:

[0065] The overall mean and variance of the i-th category of pedestrians are calculated as μ i and The distribution of the crossing time data of different categories of pedestrians is tested;

[0066] Based on the data distribution characteristics and the sub-sample mean and variance, the overall mean and variance are estimated to obtain the confidence interval of the overall mean μ i with a confidence level of 1-α;

[0067] The upper limit of the confidence interval is taken as the final estimate of μ i , and the overall mean of the crossing time is obtained.

[0068] Further, the determination of the upper limit value of the crossing time corresponding to different categories of pedestrians based on the Wald criterion includes the following steps:

[0069] The probability of the crossing time of pedestrians within the range of (μ i -3σ i , μ i +3σ i ) is 0.9974, and the upper limit value of the crossing time corresponding to different categories of pedestrians is determined as:

[0070]

[0071] Where σ i is the relaxation time.

[0072] Preferably, the process of setting the corresponding relaxation time of different pedestrians crossing the road based on the kinematic model includes the following steps:

[0073] According to the obtained initial state information of pedestrians crossing the road, the crossing time t T ' of pedestrians is predicted by the kinematic model;

[0074] By comparing t T ' with the upper limit value of the crossing time corresponding to different categories of pedestrians , the category of the pedestrian is determined, and the preliminary prediction result and the corresponding relaxation time are determined;

[0075] If t , the pedestrian is aggressive, and the preliminary prediction result is Relaxation time σ1;

[0076] If the pedestrian is of the general type, and the preliminary prediction result is Relaxation time σ2;

[0077] If the pedestrian is of the conservative type, and the preliminary prediction result is Relaxation time σ3.

[0078] Preferably,

[0079] The process of continuously collecting real-time information of the pedestrian crossing the road is as follows:

[0080] The state vector of the pedestrian at time t-1 is X t-1 ,

[0081] X t-1 = (p, v) T ;

[0082] Wherein, p is the position of the pedestrian at the current time, and v is the speed of the pedestrian at the current time.

[0083] Preferably, the process of calculating the real-time prediction result of the pedestrian is as follows:

[0084]

[0085]

[0086] Wherein, is the estimated value of the state of the pedestrian at time t; n is the number of lanes, w r is the width of the lane; is the average angle between the walking direction of the pedestrian and the cross section of the road;

[0087] The real-time prediction result is: t-t0+t K ;

[0088] Based on the difference between the preliminary prediction result and the real-time prediction result, it is determined whether the relaxation time needs to be reset;

[0089] If the prediction result of the pedestrian crossing time remains unchanged; if the prediction result of the pedestrian crossing time is reset: And so on, the final prediction result of the pedestrian crossing time is:

[0090]

[0091] Wherein, m is the number of resets, and the prediction result of the vehicle-pedestrian conflict time window is

[0092] The application provides a highway pedestrian crossing time prediction system, comprising:

[0093] A preliminary prediction result module is configured to collect initial pedestrian crossing road information, input a kinematic model, set corresponding relaxation times of different pedestrians crossing roads based on the kinematic model, and estimate a preliminary prediction result of the pedestrians crossing the roads.

[0094] A prediction result difference module is configured to continuously collect real-time information of the pedestrians crossing the roads, obtain real-time prediction results of the pedestrians crossing the roads based on Kalman filter trajectory prediction, and compare a difference between the preliminary prediction result and the real-time prediction result.

[0095] A car-person conflict time window module is configured to reset estimated pedestrian crossing road information based on the difference between the preliminary prediction result and the real-time prediction result, and obtain a car-person conflict time window.

[0096] A preferred embodiment provided by the application is:

[0097] The car-person conflict time window is defined, the time when the pedestrian starts to cross the road is set as t0, and the prediction result of the time when the pedestrian ends to cross the road is set as t f , and the prediction result of the car-person conflict time window is The pedestrian crossing time is:

[0098] t T = t f -t0

[0099] Specifically, the pedestrian crossing time t T can be initially determined by the prediction result of the pedestrian crossing time kinematic model, and is reset by the difference between the preliminary prediction result and the real-time prediction result.

[0100] Relaxation time estimation. Taking a certain highway scene as an example, pedestrian crossing data is collected. The relaxation time of the pedestrian crossing time is estimated through the measured data, including the following steps:

[0101] 1) C-class mean fuzzy clustering method is used to divide pedestrians with road crossing time differences into three categories of aggressive, general and conservative, and the overall mean and variance of the ith class of pedestrians are set as μ i and The mean and variance of the sub-sample with a corresponding capacity n i are and

[0102] Specifically, n1, n2 and n3 take values of 56, 88 and 56, respectively; and take values of 6.7, 7.4 and 8.3, respectively; and The values are 0.07, 0.04 and 0.19, respectively;

[0103] 2) The mean crossing time μ of the three types of pedestrians i Estimation is performed, and since the classified data is judged to be normally distributed by normality test, we have:

[0104]

[0105]

[0106] The confidence interval of μ under the confidence level of 1-α is: i

[0107]

[0108] To ensure the safety of pedestrians, the upper limit of the confidence interval is taken as the estimated value of μ: i

[0109]

[0110] Specifically, the confidence level 1-α is 90%, and The value is 1.645; the μ1, μ2 and μ3 are 6.77, 7.47 and 8.43, respectively;

[0111] 3) Determine the upper limit value of the crossing time of pedestrians of different categories, and based on the Laplace criterion, the probability that the pedestrian crossing time distribution is within the interval (μ i -3σ i , μ i +3σ i ) is 0.9974, so the upper limit value of the crossing time of pedestrians of different categories is:

[0112]

[0113] Among them, At the same time, σ i is determined as the relaxation time when monitoring the motion trajectory of the corresponding category of pedestrians, and is estimated by the sub-sample variance of the crossing data of each category of pedestrians, which are 0.25, 0.20 and 0.44, respectively. Substituting formula (10) gives and are 7.52, 8.08 and 9.73, respectively.

[0114] Prediction of pedestrian crossing time;

[0115] The kinematic model is used to predict the pedestrian crossing time, and the pedestrian crossing time prediction result can be obtained in time when the pedestrian starts crossing, including the following steps:

[0116] 1) Calculate the pedestrian crossing time t according to the kinematic model​​T

[0117]

[0118]

[0119] where v0 is the initial speed of the pedestrian crossing, is the lateral speed component of v0, n is the number of lanes, w r is the lane width, and θ is the angle between the direction of the pedestrian and the road cross section;

[0120] The instantaneous speed and direction of the pedestrian at 0.2 seconds after the pedestrian starts crossing are selected to calculate the crossing time t T of the pedestrian; T

[0121] 2) The category of the pedestrian is determined by comparing t T with the upper limit value of the crossing time corresponding to the different categories of pedestrians obtained in step 1, so as to determine the preliminary crossing time prediction result and the corresponding relaxation time;

[0122] Specifically, if the pedestrian is aggressive, and the preliminary prediction result is t with a relaxation time σ1;

[0123] If the pedestrian is general, and the preliminary prediction result is t with a relaxation time σ2;

[0124] If the pedestrian is conservative, and the preliminary prediction result is t with a relaxation time σ3;

[0125] Step 4: Real-time monitoring model of pedestrian motion trajectory;

[0126] The Kalman filter trajectory prediction algorithm is used to monitor the pedestrian trajectory in real time, and the core of the algorithm is to correct the estimated value at the previous time according to the latest measurement value, which can realize real-time monitoring of the pedestrian trajectory.

[0127] The pedestrian crossing behavior is monitored in real time from 1 second after the pedestrian starts crossing with a monitoring time interval of 0.2 seconds;

[0128] The motion state equation and the observation equation of the pedestrian are:

[0129] X t = FX t-1 + W t ;

[0130] Z t = HX​​t +V t ;

[0131] where Z t is the observation vector, F and H are the state transition matrix and observation matrix respectively, W t and V t are the process noise and observation noise respectively, Q and R are their covariance matrices respectively;

[0132] Specifically, the state prediction equation of the pedestrian and the transmission process of the model uncertainty are as follows:

[0133]

[0134]

[0135] where, is the state estimation value at t-1 time, is the state estimation value at t time, is the state estimation value update value at t time; P t-1 is the noise covariance matrix at t-1 time, P t - is the noise covariance matrix estimation value based on t-1 time;

[0136] The calculation process of the Kalman gain and the pedestrian state update equation are as follows:

[0137] K t = P t - H T (HP t - H T +R) -1 ;

[0138]

[0139] where K t is the Kalman gain;

[0140] Further, the updated noise covariance matrix is as follows:

[0141]

[0142] The constant velocity model is used to determine the state transition matrix H =

[10] ; W t and V t are zero-mean Gaussian white noise processes; R = 0.04.

[0143] The fusion analysis determines the final pedestrian crossing time prediction result by fusing the pedestrian crossing time prediction result based on the kinematic model and the real-time monitoring result of the pedestrian trajectory prediction algorithm of Kalman filter, and includes the following two parts:

[0144] 1) Calculate the real-time remaining time of the pedestrian crossing the road, calculate the remaining time of the pedestrian crossing the road at time t through the real-time pedestrian motion state prediction result of Kalman filter t K , and obtain the pedestrian crossing time prediction result under real-time monitoring:

[0145]

[0146]

[0147] Wherein, is the estimated value of the pedestrian state at time t; is the average angle between the walking direction of the pedestrian and the cross section of the road, which is estimated from sample data Take 0.99; then the pedestrian crossing time prediction result under real-time monitoring by Kalman filter prediction algorithm is t-t0+t K ;

[0148] 2) Reset the pedestrian crossing time prediction result based on the relaxation time, compare the real-time monitoring result of the pedestrian crossing time prediction result and the preliminary crossing time prediction result in step two, and determine whether the relaxation time needs to be reset;

[0149] If , the pedestrian crossing time prediction result remains unchanged;

[0150] If , the pedestrian crossing time prediction result is reset: And so on, the final pedestrian crossing time prediction result is:

[0151]

[0152] Wherein m is the reset number, and the final prediction result of the vehicle-pedestrian conflict time window is

[0153] In still another embodiment of the present application, a computer device is provided, which comprises a processor and a memory, the memory being configured to store a computer program, the computer program comprising program instructions, and the processor being configured to execute the program instructions stored in the computer storage medium. The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, and are particularly suitable for loading and executing one or more instructions in the computer storage medium to implement a corresponding method flow or a corresponding function; the processor in the embodiments of the present application can be used for the operation of the highway pedestrian crossing time prediction method.

[0154] In still another embodiment of the present application, the present application further provides a storage medium, specifically a computer readable storage medium (Memory), which is a memory device in the computer device, and is configured to store programs and data. It can be understood that the computer readable storage medium herein can include the built-in storage medium in the computer device, and of course can also include the expansion storage medium supported by the computer device. The computer readable storage medium provides a storage space, which stores the operating system of the terminal. Moreover, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM memory, or a non-volatile memory such as at least one disk memory. One or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the highway pedestrian crossing time prediction method in the above embodiments.

[0155] Those skilled in the art will appreciate that embodiments of the application can be devised for a method, a system, or a computer program product. Accordingly, the present application can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0156] The present application is described in reference to the flowchart and / or block diagrams of the method, apparatus (system) and computer program product according to embodiments of the present application. It will be understood that each block of the flowchart and / or block diagrams, and combinations of blocks in the flowchart and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0157] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0158] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks. Figure 1 one or more functions specified in the flowchart and / or block diagram block or blocks.

[0159] Finally, it should be noted that the above-described embodiments are merely intended for describing and illustrating, not limiting, the technical solutions of the present application; even though the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for forecasting freeway pedestrian crossing time, characterized by, The method comprises the following steps: The initial information of the pedestrian crossing the road is collected to input a kinematic model, and corresponding relaxation time of different pedestrians crossing the road is set based on the kinematic model to preliminarily predict the crossing time of the pedestrian; The process of setting the corresponding relaxation time of different pedestrians crossing the road based on the kinematic model comprises the following steps: According to the acquired state information of the pedestrian initially crossing the road, the pedestrian crossing time t is predicted through a kinematics model T '; By comparing t T Upper limit values of crossing time corresponding to different categories of pedestrians Determine the category to which the pedestrian belongs, determine the preliminary prediction result and the corresponding relaxation time If then the pedestrian is aggressive, and the preliminary prediction is the relaxation time σ1; If then the pedestrian is of the general type and the preliminary prediction result is the relaxation time σ2; If then the pedestrian is conservative, and the preliminary prediction is the relaxation time σ3; The construction of the kinematic model divides pedestrians with different crossing times into three categories of aggressive, general and conservative based on the data of pedestrians crossing the road collected in advance, estimates the overall mean of the crossing time of the three categories of pedestrians, and determines the upper limit value of the crossing time corresponding to different categories of pedestrians based on the Laplace criterion; The estimation of the overall mean of the crossing time of the three categories of pedestrians comprises the following steps: The total mean and variance of the ith class of pedestrians are μ i and Distribution test is performed on the crossing time data of different classes of pedestrians; Based on the data distribution characteristics and the sub-sample mean and variance, the population mean and variance are estimated, and the confidence interval of the population mean μ i with confidence 1-α is obtained. Taking the upper limit of the confidence interval as the final μ i The estimated value, and then the overall mean of the crossing time; The determination of the upper limit value of the crossing time corresponding to different categories of pedestrians based on the Laplace criterion comprises the following steps: The probability of pedestrian crossing time in (μ i - 3σ i , μ i + 3σ i ) range is 0.9974, and the upper limit value of crossing time corresponding to different categories of pedestrians is determined: where σ i is the relaxation time; Real-time information of the pedestrian crossing the road is continuously collected, and real-time prediction results of the pedestrian crossing the road are obtained based on Kalman filtering trajectory prediction, and the difference between the preliminary prediction results and the real-time prediction results is compared; Based on the difference between the preliminary prediction results and the real-time prediction results and the relaxation time, the estimated information of the pedestrian crossing the road is continuously reset to obtain a vehicle-pedestrian conflict time window.

2. The method of claim 1, wherein, The vehicle-person conflict time window is Wherein, t0 is the time when the pedestrian starts to cross the road, t f is the time when the pedestrian ends to cross the road, then, the pedestrian crossing time is: t T = t f - t0.

3. The method of claim 1, wherein, The process of continuously collecting real-time information of the pedestrian crossing the road is as follows: The state vector of the pedestrian at time t-1 is X t-1 , X t-1 = (p, v) T ; Wherein, p is the position of the pedestrian at the current time, and v is the speed of the pedestrian at the current time.

4. The method of claim 3, wherein, The process of calculating the real-time prediction results of the pedestrian is as follows: wherein, is the pedestrian state estimation value at time t; n is the number of lanes, w r is the lane width; is the average angle between the pedestrian walking direction and the road cross section; Real-time prediction result is: t-t0+t K ; Based on the difference between the preliminary prediction results and the real-time prediction results, it is determined whether the relaxation time needs to be reset; If The pedestrian crossing time prediction result Remains unchanged; if The pedestrian crossing time prediction result is reset: And so on, the final pedestrian crossing time prediction result is: ; where m is the number of resets, and the prediction of the car-person conflict time window is obtained as 5. A high-speed road pedestrian crossing time prediction system characterized by, The method comprises the following steps: The initial information of the pedestrian crossing the road is collected to input a kinematic model, and corresponding relaxation time of different pedestrians crossing the road is set based on the kinematic model to preliminarily predict the crossing time of the pedestrian; The process of setting the corresponding relaxation time of different pedestrians crossing the road based on the kinematic model comprises the following steps: According to the acquired state information of the pedestrian initially crossing the road, the pedestrian crossing time t is predicted through a kinematics model T '; By comparing t T Upper limit values of crossing time corresponding to different categories of pedestrians Determine the category to which the pedestrian belongs, determine the preliminary prediction result and the corresponding relaxation time If then the pedestrian is aggressive, and the preliminary prediction is the relaxation time σ1; If then the pedestrian is of the general type and the preliminary prediction result is the relaxation time σ2; If then the pedestrian is conservative, and the preliminary prediction is the relaxation time σ3; The construction of the kinematic model divides pedestrians with different crossing times into three categories of aggressive, general and conservative based on the data of pedestrians crossing the road collected in advance, estimates the overall mean of the crossing time of the three categories of pedestrians, and determines the upper limit value of the crossing time corresponding to different categories of pedestrians based on the Laplace criterion; The estimation of the overall mean of the crossing time of the three categories of pedestrians comprises the following steps: The total mean and variance of the ith class of pedestrians are μ i and Distribution test is performed on the crossing time data of different classes of pedestrians; Based on the data distribution characteristics and the sub-sample mean and variance, the population mean and variance are estimated, and the confidence interval of the population mean μ i with confidence 1-α is obtained. Taking the upper limit of the confidence interval as the final μ i estimate, and thereby obtain the overall mean of the crossing time; The determination of the upper limit value of the crossing time corresponding to different categories of pedestrians based on the Laplace criterion comprises the following steps: The probability of pedestrian crossing time in (μ i - 3σ i , μ i + 3σ i ) range is 0.9974, and the upper limit value of crossing time corresponding to different categories of pedestrians is determined: where σ i is the relaxation time; A prediction result difference module is configured to continuously collect real-time information of the pedestrian crossing the road, and obtain real-time prediction results of the pedestrian crossing the road based on Kalman filtering trajectory prediction, and compare the difference between the preliminary prediction results and the real-time prediction results; A vehicle-pedestrian conflict time window module is configured to reset the estimated information of the pedestrian crossing the road based on the difference between the preliminary prediction results and the real-time prediction results, and obtain a vehicle-pedestrian conflict time window.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the method for predicting the crossing time of a pedestrian on a highway according to any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 6. The computer program is executed by the processor to implement the steps of the method for predicting the crossing time of a pedestrian on a highway according to any one of claims 1 to 4.

Images