Self-adaptive real-time estimation method and system for human driver parameters in mixed traffic flow and storage medium

Abstract

The invention discloses a self-adaptive real-time estimation method and system for human driver parameters in a mixed traffic flow and a storage medium, and relates to a method capable of carrying out self-adaptive real-time estimation on the driver parameters in the mixed traffic flow of a human driving vehicle (HDV) and an automatic driving vehicle (AV). The method adopts a Bayesian estimation method, and comprises the following steps of: firstly, determining prior distribution p (tau) of human driver parameters according to data of an existing vehicle following data following automatic driving vehicle; the likelihood function p (af (t) tau) is expressed after being converted and simplified by a stimulation-reaction car-following model (GM model); the method comprises the following steps: firstly, calculating the posterior distribution of driver parameters according to the posterior distribution of logarithmic normal distribution, and finally estimating the parameters of next unit time according to the parameters estimated in the previous unit time, thereby achieving the effect of self-adaptive real-time estimation.

Description

Technical Field

[0001] The present invention relates to a method, system and storage medium for adaptive real-time estimation of human driver parameters in mixed traffic flow, and is a method capable of adaptively estimating human driver parameters in real-time in mixed traffic flow of human-driven vehicles (HDV) and autonomous vehicles (AV). Background Art

[0002] Human-driven vehicles and autonomous vehicles will coexist for a long time now and in the future. Real-time estimation of human driver parameters in mixed traffic flow is conducive to real-time monitoring of driver behavior and status, taking timely measures to avoid accidents. Through real-time estimation results, intelligent driving assistance systems can adjust according to the driver's current status and behavior to provide more effective assistance. More importantly, based on real-time estimation results, the autonomous driving system can understand the behavior patterns of human drivers in real time, so as to better coordinate with human drivers, improve adaptability and safety in complex environments, and improve the operating efficiency of the road system and driving experience. Therefore, real-time estimation of human driver parameters in mixed traffic flow is of great significance to the development of intelligent transportation systems. Summary of the invention

[0003] The purpose of the present invention is to address the deficiencies of the prior art and propose an adaptive real-time estimation method for human driver parameters in mixed traffic flow. The method is applicable to mixed traffic flow environments, i.e., traffic flows formed by human-driven vehicles and self-driving vehicles.

[0004] The technical solution adopted by the present invention to solve the technical problem is:

[0005] An adaptive real-time estimation method for human driver parameters in mixed traffic flow adopts Bayesian estimation. According to the probability distribution of human driver parameters in the previous unit time, the probability distribution of human driver parameters in the next unit time is obtained by combining with a car-following model, and the Bayesian estimation value of the parameters is calculated, so as to achieve the effect of real-time estimation. The method specifically includes the following steps:

[0006] 1) Based on the existing data set, the cross-correlation time response method is used to obtain the prior distribution of human driver parameters in mixed traffic flow, which can be regarded as a log-normal distribution;

[0007] 2) Calculate the likelihood function of the target parameters based on the car-following model;

[0008] 3) According to the posterior distribution of the lognormal distribution, the posterior distribution of the target parameter is calculated, and then the Bayesian estimate of the parameter is calculated. Finally, based on the parameters estimated in the previous unit time, the parameters for the next unit time are adaptively estimated in real time.

[0009] In the above technical solution, further, the human driver parameter is the driver's reaction time.

[0010] Furthermore, in step 1), the speed, acceleration and the speed of the leading vehicle when the HDV follows the AV are extracted from the existing data set, and the speed difference between the leading and trailing vehicles and the acceleration of the trailing vehicle are selected as two factors. The mutuality between the two factors is calculated according to the cross-correlation time response method, and the parameter value corresponding to the maximum cross-correlation is determined as the target value. The prior distribution of the target parameter is obtained using multiple sets of data.

[0011] Furthermore, in step 2), the GM model in the car-following model is selected as the likelihood function of the Bayesian estimation. The model takes into account the reaction time parameters of the human driver, and also includes the speed of the following vehicle, the speed difference between the front and rear vehicles, and the distance difference between the front and rear vehicles. These data are easy to obtain and can be displayed in real time.

[0012] Furthermore, in step 3), the prior distribution and likelihood function of the target parameter are substituted into the posterior distribution formula of the lognormal distribution to obtain the posterior distribution of the target parameter. Then, according to the formula Adaptive Bayesian estimates of human driver parameters are computed.

[0013] An adaptive real-time estimation system for human driver parameters in mixed traffic flow, used to implement any of the methods described above.

[0014] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements any of the methods described above.

[0015] An electronic device, comprising:

[0016] one or more processors;

[0017] A memory for storing one or more programs;

[0018] When the one or more programs are executed by the one or more processors, the one or more processors implement any of the above methods.

[0019] The beneficial effects of the present invention are:

[0020] The method of the present invention can provide real-time estimation of the parameters of surrounding human-driven vehicles for autonomous driving vehicles, which can be used for subsequent important tasks such as trajectory planning and AEB (automatic emergency braking). Existing methods mainly use calibrated parameters of surrounding human-driven vehicles, lack a real-time adaptive adjustment mechanism, and cannot cope with complex traffic flow environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the process of the present invention.

[0022] Figure 2 , Figure 3 The existing research results on the changes of driver parameters over time show that they are all volatile. DETAILED DESCRIPTION

[0023] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0024] According to a specific example of the present invention, an adaptive real-time estimation method of human driver parameters in mixed traffic flow of the present invention is based on the Bayesian estimation method, according to the probability distribution of the human driver parameters in the previous unit time, combined with the following model to calculate the probability distribution of the human driver parameters in the next unit time, and calculate the Bayesian estimation value of the parameter. Through the above method flow, the effect of real-time estimation of human driver parameters such as human driver reaction time is achieved, so that the estimation information is sent to each autonomous driving vehicle to achieve more intelligent and effective assistance; Figure 1 As shown, the method specifically comprises the following steps:

[0025] 1) Based on the existing data set, the cross-correlation time response method is used to obtain the prior distribution of human driver parameters in mixed traffic flow, which can be regarded as a log-normal distribution;

[0026] Extract the speed v when HDV follows AV in the existing data set f (t), acceleration a f (t) and the speed v of the preceding vehicle l (t), select the speed difference Δv of the front and rear vehicles l-f (t) and the acceleration of the rear vehicle, the cross-correlation between the two factors is calculated according to the cross-correlation time response method, where the cross-covariance between the speed difference between the front and rear vehicles and the acceleration of the rear vehicle is as follows:

[0027]

[0028] in, They are Δv l-f (t), a f (t), τ is the time delay imposed on the following vehicle, i.e., the reaction time of HDV. The two vehicles following each other are a pair, and N is the number of pairs.

[0029] Compute the cross-correlation between two factors:

[0030]

[0031] in, is Δvl-f The standard deviation of (t), for a f (t) standard deviation;

[0032] The parameter value corresponding to the maximum cross-correlation is determined as the target value, and the prior distribution of the target parameter is obtained using multiple sets of data; the prior distribution can be written as a log-normal distribution with parameters μ and σ, that is:

[0033]

[0034] 2) Calculate the likelihood function of the target parameters based on the stimulus-response (GM) model in the car-following model;

[0035] The GM model in the car-following model is selected as the likelihood function of Bayesian estimation. The model takes into account the reaction time parameters of human drivers, as well as the speed of the following vehicle, the speed difference between the front and rear vehicles, and the distance difference between the front and rear vehicles. These data are easy to obtain and can be displayed in real time. The GM model in the car-following model is shown in the following formula, which indicates that the acceleration of the following vehicle with reaction time is proportional to the speed difference between the front and rear vehicles, the speed of the following vehicle with reaction time, and inversely proportional to the distance difference between the front and rear vehicles. β and δ represent exponents, which are set according to the requirements. Different values ​​can convert the model into Greenshields, Grenberg, and Underwood models. The dimension of α(β, δ) depends on β and δ.

[0036]

[0037] Transform and simplify the formula as follows:

[0038]

[0039] Among them, let:

[0040] (x l (t-τ)-x f (t-τ) β =(x l-f (t-τ) β

[0041] v l (t-τ)-v f (t-τ)=v l-f (t-τ)

[0042] Calculate the given sample set t = (t1, t2, t3, ..., t n ) is the likelihood function p(a f (t)|τ):

[0043]

[0044] 3) According to the posterior distribution of the lognormal distribution, the posterior distribution of the target parameter is calculated, and then the Bayesian estimate of the parameter is calculated. Finally, based on the parameters estimated in the previous unit time, the parameters for the next unit time are adaptively estimated in real time.

[0045] Substitute the prior distribution and likelihood function of the above target parameters into the posterior distribution formula of the lognormal distribution to obtain the posterior distribution of the target parameters:

[0046]

[0047] Then according to the formula The adaptive Bayesian estimates of the human driver parameters are calculated:

[0048]

[0049] Figure 2 , Figure 3 According to the research on the change of driver parameters over time at home and abroad, it can be seen that the driver's parameters are volatile. The method proposed in the present invention can adaptively estimate its parameters based on real-time data, thereby improving the operating performance and safety level of the autonomous driving vehicle.

[0050] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0051] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0052] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0053] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0054] The specific implementation methods described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific implementation methods described or replace them with similar methods, but they will not deviate from the spirit of the present invention or exceed the scope defined by the attached claims.

Claims

1. An adaptive real-time estimation method for human driver parameters in mixed traffic flow, characterized in that: This method uses Bayesian estimation to calculate the probability distribution of human driver parameters in the next unit time based on the probability distribution of human driver parameters in the previous unit time in combination with the following model, and then calculates the Bayesian estimation value of the parameters to achieve adaptive real-time estimation.

2. The method for adaptive real-time estimation of human driver parameters in mixed traffic flow according to claim 1, characterized in that: The specific steps include: 1) Based on the existing data set, the cross-correlation time response method is used to obtain the prior distribution of human driver parameters, i.e., target parameters, in mixed traffic flow, which is considered to be a log-normal distribution; 2) Calculate the likelihood function of the target parameters based on the car-following model; 3) According to the posterior distribution of the lognormal distribution, the adaptive posterior distribution of the target parameter is calculated, and then the Bayesian estimate of the target parameter is calculated. Finally, based on the parameters estimated in the previous unit time, the parameters of the next unit time are adaptively estimated in real time.

3. The adaptive real-time estimation method for human driver parameters in mixed traffic flow according to claim 2, characterized in that: The human driver parameter is the driver reaction time.

4. The method for adaptive real-time estimation of human driver parameters in mixed traffic flow according to claim 3, characterized in that: In step 1), the speed, acceleration and speed of the HDV following the AV as well as the speed of the leading vehicle are extracted from the existing data set, the speed difference between the leading and trailing vehicles and the acceleration of the trailing vehicle are selected as two factors, the mutuality between the two factors is calculated according to the cross-correlation time response method, the parameter value corresponding to the maximum cross-correlation is determined as the target value, and the prior distribution of the target parameter is obtained using multiple sets of data.

5. The method for adaptive real-time estimation of human driver parameters in mixed traffic flow according to claim 3, characterized in that: In step 2), the GM model in the car-following model is used as the likelihood function of the Bayesian estimate.

6. The method for adaptive real-time estimation of human driver parameters in mixed traffic flow according to claim 3, characterized in that: In step 3), the prior distribution of the target parameter and the likelihood function of the target parameter are substituted into the posterior distribution formula of the lognormal distribution to obtain the posterior distribution of the target parameter, and the Bayesian estimation value of the human driver parameter is calculated based on the posterior distribution.

7. An adaptive real-time estimation system for human driver parameters in mixed traffic flow, characterized in that: Used to implement the method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. An electronic device, characterized in that: The device comprises: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

Images