CN113665574B - Prediction of lane changing duration and anthropomorphic trajectory planning method for intelligent vehicles - Google Patents

Abstract

The present application relates to the technical field of intelligent vehicle applications, and in particular to a method for predicting the lane-changing duration and anthropomorphic trajectory planning of an intelligent vehicle, including: extracting the lane-changing trajectory of an excellent driver from natural driving data, and extracting the lane-changing trajectory under the influence of multiple vehicles. Lane duration; obtain the information of the lane-changing vehicle's lap movement, characterize the influence of the lap as a nonlinear mapping of the kinematic parameters of the lap to the average longitudinal acceleration of the lane-changing vehicle, and then obtain the lane-changing duration prediction model under the influence of the lap; When applying the lane-changing duration prediction model to the smart car, based on the expected longitudinal displacement of the lane-changing, the current longitudinal speed and the expected speed of the target lane after the lane-changing, combined with the information of the surrounding vehicles, the prediction model is used to optimize the lane-changing duration, thereby realizing anthropomorphic changing. Road trajectory planning. As a result, the manipulation rules of excellent drivers in the natural driving data are fully excavated, which provides a reference for intelligent vehicles to make scientific and reasonable lane-changing decisions.

Description

Prediction of lane-changing duration and anthropomorphic trajectory planning method for intelligent vehicles

technical field

The present application relates to the technical field of intelligent vehicle applications, and in particular, to a method for predicting the lane-changing duration and anthropomorphic trajectory planning of an intelligent vehicle.

Background technique

In a mixed traffic environment where manual driving and automatic driving are mixed, the lane-changing behavior of autonomous vehicles is affected by the surrounding traffic vehicles, especially the uncertainty of artificially-driven vehicles, which makes the longitudinal motion of the lane-changing trajectory very complex.

In the related art, it is usually assumed that the longitudinal speed of the smart car remains unchanged during the lane-changing process. The lane-changing trajectory obtained by planning the lane-changing trajectory under this ideal operating condition requires more consideration of the lateral motion law. However, this method The lack of analysis of the longitudinal motion of the vehicle during the lane-changing process, and the failure to fully consider the impact of surrounding vehicles on the longitudinal driving behavior of the own vehicle during the lane-changing process, makes the expected performance of the planned trajectory not necessarily able to meet the expected requirements, which may lead to frequent overloading. Deceleration will reduce the comfort of changing lanes, and there may even be safety problems such as excessive lateral acceleration of the vehicle due to insufficient lane-changing time, inducing the risk of instability such as sideslip and roll of the own vehicle, and increased risk of collision with surrounding vehicles. ,waiting to be solved.

SUMMARY OF THE INVENTION

The present application provides a method for predicting the lane-changing duration of an intelligent vehicle and an anthropomorphic trajectory planning method, so as to solve the problem of lane-changing caused by insufficient quantitative assessment of the impact of surrounding vehicles on the longitudinal driving behavior during the lane-changing process of the vehicle in the relevant intelligent vehicle technology. Insufficient rationality, poor interpretability and possible safety risks in the decision-making of time duration, fully excavate the driving and maneuvering rules of excellent drivers in the natural driving data, and provide a reference for intelligent vehicles to make scientific and reasonable lane-changing decisions. The embodiment of the anthropomorphic decision-making concept of "simulating people, surpassing people and serving people" in lane changing decision-making.

The embodiment of the first aspect of the present application provides a method for predicting the lane-changing duration and anthropomorphic trajectory planning of an intelligent vehicle, including the following steps:

Extracting the lane-changing trajectory of the excellent driver from the natural driving data of several drivers, and extracting the lane-changing duration of the excellent driver under the influence of real multi-vehicles from the lane-changing trajectory of the excellent driver;

Acquiring the information of the lane-changing vehicle's movement around the vehicle, and establishing a longitudinal kinematics model of the vehicle under the influence of multiple vehicles according to the vehicle's movement information, and characterizing the influence of the vehicle's movement as the effect of the vehicle's kinematics parameters on changing lanes. The nonlinear mapping of the average longitudinal acceleration of the vehicle, and then the trained lane-changing duration prediction model under the influence of multiple vehicles and the model parameters obtained by identification are obtained;

In the case where the smart car in the network-connected multi-vehicle environment uses the lane-changing duration prediction model to make decisions, when a lane-changing command is issued, based on the expected longitudinal displacement of the lane-changing, the longitudinal speed of the smart car and the target after the lane-changing The expected speed of the lane, combined with the weekly vehicle information of the smart car, use the lane-changing duration prediction model to optimize the lane-changing duration of the smart car, and anthropomorphize the lane-changing duration of the smart car obtained by optimization. trajectory planning to realize the accurate implementation of the lane-changing intention of the intelligent vehicle.

In some examples, the lane-changing trajectory of the excellent driver is extracted from the natural driving data of several drivers, and the lane-changing duration of the excellent driver under the influence of a real multi-vehicle is extracted from the lane-changing trajectory of the excellent driver, include:

calculating the heading angle of the lane-changing trajectory of the excellent driver;

Determine a peak point from the natural driving data based on the heading angle, and search both sides from the peak point to obtain a time point that satisfies a preset condition;

matching the initial interval of the corresponding lane-changing trajectory by the time point;

Calculate the peak-to-peak time difference of the lateral acceleration of the lane-changing trajectory according to the initial interval, and calculate the corresponding time points when the lateral acceleration obtains the maximum value and the minimum value based on the time difference, and obtain the lane-changing trajectory in the real multi-vehicle. A good driver's lane-changing time under the influence.

In some examples, obtaining the trained lane-changing duration prediction model under the influence of multiple vehicles and identifying the obtained model parameters include:

Acceleration correction is performed on the average longitudinal acceleration of the lane-changing vehicle to obtain the corrected average longitudinal acceleration, and the influence of the surrounding car is characterized as a nonlinear mapping of the kinematic parameters of the surrounding vehicle to the average longitudinal acceleration of the lane-changing vehicle ;

According to the vehicle lane-changing longitudinal kinematics model under the influence of the multi-vehicle and the influence of the circling vehicle, which is characterized as a nonlinear mapping of the kinematic parameters of the circling vehicle to the average longitudinal acceleration of the lane-changing vehicle, the multi-vehicle influence is obtained. The lane-changing duration prediction model and the model parameters obtained by identification.

In some examples, the lane change duration prediction model is:

，

,

为自车的换道时长，

为所述自车预期换道纵向位移、

为所述自车平均纵向加速度，

为所述自车换道后在目标车道的预期车速。in,

for the lane changing time of the vehicle,

is the expected lane change longitudinal displacement of the ego vehicle,

is the average longitudinal acceleration of the ego vehicle,

The expected vehicle speed in the target lane after changing lanes for the ego vehicle.

In some examples, the expected lane change longitudinal displacement is determined from a longitudinal position, longitudinal velocity, and longitudinal acceleration state function.

The embodiment of the second aspect of the present application provides an intelligent vehicle lane changing duration prediction and anthropomorphic trajectory planning device, including:

The extraction module is used to extract the lane-changing trajectory of the excellent driver from the natural driving data of several drivers, and extract the lane-changing duration of the excellent driver under the influence of the real multi-vehicle from the lane-changing trajectory of the excellent driver;

The generation module is used to obtain the information about the vehicle's movement around the vehicle, and according to the information about the vehicle's movement, establish a longitudinal kinematics model of the vehicle under the influence of multiple vehicles, and characterize the influence of the vehicle as the movement of the vehicle. The nonlinear mapping of the learning parameters to the average longitudinal acceleration of the lane-changing vehicle, and then the trained lane-changing duration prediction model under the influence of multiple vehicles and the model parameters obtained by identification are obtained; and

The lane-changing trajectory planning module is used for, in the case where the intelligent vehicle in the network-connected multi-vehicle environment uses the lane-changing duration prediction model to make decisions, when a lane-changing command is issued, based on the expected lane-changing longitudinal displacement, the The longitudinal speed of the smart car and the expected speed of the target lane after changing lanes, combined with the weekly traffic information of the smart car, the lane-changing duration prediction model is used to optimize the lane-changing duration of the smart car, and according to the obtained An anthropomorphic trajectory planning is carried out according to the lane-changing duration of the intelligent vehicle, so as to realize the accurate implementation of the lane-changing intention of the intelligent vehicle.

In some examples, the extraction module is specifically used to:

calculating the heading angle of the lane-changing trajectory of the excellent driver;

Determine a peak point from the natural driving data based on the heading angle, and search both sides from the peak point to obtain a time point that satisfies a preset condition;

matching the initial interval of the corresponding lane-changing trajectory by the time point;

Calculate the peak-to-peak time difference of the lateral acceleration of the lane-changing trajectory according to the initial interval, and calculate the corresponding time points when the lateral acceleration obtains the maximum value and the minimum value based on the time difference, and obtain the lane-changing trajectory in the real multi-vehicle. A good driver's lane-changing time under the influence.

In some examples, the generation module is specifically used to:

Acceleration correction is performed on the average longitudinal acceleration of the lane-changing vehicle to obtain the corrected average longitudinal acceleration, and the influence of the surrounding car is characterized as a nonlinear mapping of the kinematic parameters of the surrounding vehicle to the average longitudinal acceleration of the lane-changing vehicle ;

According to the vehicle lane-changing longitudinal kinematics model under the influence of the multi-vehicle and the influence of the circling vehicle, which is characterized as a nonlinear mapping of the kinematic parameters of the circling vehicle to the average longitudinal acceleration of the lane-changing vehicle, the multi-vehicle influence is obtained. The lane-changing duration prediction model and the model parameters obtained by identification.

In some examples, the lane change duration prediction model is:

，

,

为自车的换道时长，

为所述自车预期换道纵向位移、

为所述自车平均纵向加速度，

为所述自车换道后在目标车道的预期车速。in,

for the lane changing time of the vehicle,

is the expected lane change longitudinal displacement of the ego vehicle,

is the average longitudinal acceleration of the ego vehicle,

The expected vehicle speed in the target lane after changing lanes for the ego vehicle.

In some examples, the expected lane change longitudinal displacement is determined from a longitudinal position, longitudinal velocity, and longitudinal acceleration state function.

A third aspect of the present application provides a lane-changing trajectory decision device for an intelligent vehicle, including: a lane-changing intention recognition module and a lane-changing trajectory planning module, wherein the lane-changing duration prediction and The anthropomorphic trajectory planning method serves the lane-changing intention identification module and the lane-changing trajectory planning module.

A fourth aspect of the present application provides a lane-changing trajectory tracking module, on which a computer program is stored, and the program is executed by a processor to implement the lane-changing duration prediction and Anthropomorphic trajectory planning method.

In the embodiment of the present invention, the lane-changing trajectory of the excellent driver can be extracted from a large number of natural driving data of the driver, and the lane-changing duration of the excellent driver under the influence of real multi-vehicles can be extracted from the lane-changing trajectory of the excellent driver, Then, obtain the information of the perimeter vehicle motion of the lane-changing vehicle, and establish a longitudinal kinematic model of the vehicle under the influence of multiple vehicles according to the perimeter vehicle motion information, and characterize the influence of the perimeter vehicle as the average of the kinematic parameters of the perimeter vehicle on the lane-changing vehicle. The nonlinear mapping of longitudinal acceleration can then obtain a lane-changing duration prediction model under the influence of multiple vehicles. In this way, when the lane-changing duration prediction model is applied to a smart car in a networked multi-vehicle environment, the smart car can Longitudinal displacement of lane change, longitudinal speed of smart car and expected speed of target lane after lane change, etc., combined with the traffic information of smart car, the lane-changing duration prediction model is used to optimize the lane-changing duration of smart car, and then, it can be obtained according to the optimization. An anthropomorphic trajectory planning is performed on the lane-changing duration of the intelligent vehicle, so as to realize the accurate implementation of the lane-changing intention of the intelligent vehicle. This solves the problem of insufficient quantification and evaluation of the influence of surrounding vehicles on the longitudinal driving behavior during the lane changing process of the relevant smart car technology, resulting in insufficient rationality, poor interpretability and possible safety of lane change decision-making. To solve the problem of risk, the embodiment of the present invention fully excavates the driving and maneuvering rules of excellent drivers in the natural driving data of a large number of drivers, provides a reference for intelligent vehicles to make scientific and reasonable lane-changing decisions, and makes the lane-changing process of intelligent vehicles more reasonable. , and then improve the safety and reliability of smart cars, which is the embodiment of the anthropomorphic decision-making concept of smart cars "learning, simulating, surpassing and serving people" in lane-changing decision-making.

Additional aspects and advantages of the present application will be set forth, in part, in the following description, and in part will be apparent from the following description, or learned by practice of the present application.

Description of drawings

The above and/or additional aspects and advantages of the present application will become apparent and readily understood from the following description of embodiments taken in conjunction with the accompanying drawings, wherein:

1 is a flowchart of a method for predicting the lane-changing duration of an intelligent vehicle and an anthropomorphic trajectory planning method provided according to an embodiment of the present application;

2 is a schematic diagram of the variation law of the leftward lane change trajectory, lateral speed and lateral acceleration extracted from the natural driving trajectory set HighD according to an embodiment of the present application;

Fig. 3 is the variation law of the right lane change trajectory, lateral speed and lateral acceleration extracted from the natural driving trajectory set HighD according to an embodiment of the present application;

FIG. 4 is a schematic diagram of a nonlinear mapping relationship of the average equivalent longitudinal acceleration in the process of changing lanes between the surrounding vehicles and the own vehicle according to an embodiment of the present application;

FIG. 5 is an exemplary diagram illustrating the identification of the motion parameters of a weekly vehicle in a networked environment according to an embodiment of the present application;

6 is an example diagram of a statistical analysis result of longitudinal driving behavior of the measured natural driving data set HighD lane change trajectory according to an embodiment of the present application;

7 is a schematic diagram of a histogram and a scatter plot of the comparison between the lane-changing duration of the current vehicle and the lane-changing duration corresponding to the lane-changing trajectory obtained by prediction according to an embodiment of the present application;

8 is a flowchart of a method for predicting the lane-changing duration and anthropomorphic trajectory planning of an intelligent vehicle according to an embodiment of the present application;

FIG. 9 is a schematic block diagram of an apparatus for predicting the lane-changing duration and anthropomorphic trajectory planning of an intelligent vehicle according to an embodiment of the present application.

Detailed ways

The following describes in detail the embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary, and are intended to be used to explain the present application, but should not be construed as a limitation to the present application.

The following describes the method for predicting the lane-changing duration and anthropomorphic trajectory planning of the smart car according to the embodiments of the present application with reference to the accompanying drawings.

Before introducing the lane-changing duration prediction and anthropomorphic trajectory planning method of the smart car according to the embodiment of the present application, the method for extracting the lane-changing trajectory in the related art is briefly introduced.

There are two main methods in related technologies: (1) using a driving simulator to collect lane-changing data for analysis; (2) directly extracting lane-changing trajectories from natural driving data for analysis.

Specifically, the driving simulator data can be directly combined with the driver questionnaire, and the data of different driving lane changes of the same driver in different scenarios can be collected, which is convenient for cluster analysis of driver styles. However, the driving simulator has a limited sample size and There are disadvantages such as differences with the actual driving environment. Therefore, more and more existing lane-changing researches focus on directly extracting and mining the law of lane-changing behavior from high-precision natural driving trajectory data.

However, it is still a big challenge to accurately extract the lane-changing trajectory from the natural driving data. The existing technical solutions for lane-changing trajectory extraction from natural driving data usually give a threshold, and when the vehicle deviates from the center of the lane line, it is considered that the distance is greater than the threshold. The lane change starts, and the lane change is considered to end when the vehicle deviates from the center line of the target lane by less than the threshold, or the lane change trajectory is extracted according to the change process of the lateral speed from zero value to peak value to zero value. When the speed fluctuates, it is easy to cause the extracted lane change trajectory to be incomplete or redundant, which in turn affects the modeling accuracy of the next lane change duration.

Based on the above problems, the present application provides a method for predicting the lane-changing duration and anthropomorphic trajectory planning of an intelligent vehicle. In this method, the lane-changing trajectory of an excellent driver can be extracted from a large number of drivers' natural driving data. , and extract the lane-changing duration of the excellent driver under the influence of the real multi-vehicle from the lane-changing trajectory of the excellent driver, and then obtain the weekly vehicle movement information of the lane-changing vehicle, and establish the vehicle change under the influence of multi-vehicle according to the weekly vehicle movement information. The lane longitudinal kinematics model is used to characterize the influence of the surrounding vehicles as the nonlinear mapping of the kinematic parameters of the surrounding vehicles to the average longitudinal acceleration of the lane-changing vehicles, and then the prediction model of the lane-changing duration under the influence of multiple vehicles can be obtained. After the lane duration prediction model is applied to a smart car in a networked multi-vehicle environment, the smart car can be based on the longitudinal displacement of the expected lane change, the longitudinal speed of the smart car and the expected speed of the target lane after the lane change, etc. information, using the lane-changing duration prediction model to optimize the lane-changing duration of the smart car, and further, anthropomorphic trajectory planning can be carried out according to the lane-changing duration of the smart car obtained by optimization, so as to realize the accurate implementation of the lane-changing intention of the smart car. . This solves the problem of insufficient quantification and evaluation of the influence of surrounding vehicles on the longitudinal driving behavior during the lane changing process of the relevant smart car technology, resulting in insufficient rationality, poor interpretability and possible safety of lane change decision-making. To solve the problem of risk, the embodiment of the present invention fully excavates the driving and maneuvering rules of excellent drivers in the natural driving data of a large number of drivers, provides a reference for intelligent vehicles to make scientific and reasonable lane-changing decisions, and makes the lane-changing process of intelligent vehicles more reasonable. , and then improve the safety and reliability of smart cars, which is the embodiment of the anthropomorphic decision-making concept of smart cars "learning, simulating, surpassing and serving people" in lane-changing decision-making.

Specifically, FIG. 1 is a schematic flowchart of a method for predicting the lane changing duration and anthropomorphic trajectory planning of an intelligent vehicle according to an embodiment of the present application.

As shown in Figure 1, the method for predicting the lane-changing duration and anthropomorphic trajectory planning of an intelligent vehicle includes the following steps:

In step S101, the lane-changing trajectory of the excellent driver is extracted from the natural driving data of several drivers, and the lane-changing duration of the excellent driver under the influence of real multiple vehicles is extracted from the lane-changing trajectory of the excellent driver. That is: extract the lane-changing trajectory of the excellent driver from the massive natural driving data of the driver, and then extract the lane-changing duration of the excellent driver under the influence of real multi-vehicles corresponding to the lane-changing trajectory of the excellent driver.

In a specific example, the lane-changing trajectory of the excellent driver is extracted from the natural driving data of several drivers, and the lane-changing duration of the excellent driver under the influence of real multi-vehicles is extracted from the lane-changing trajectory of the excellent driver, including: Calculate the heading angle of the lane-changing trajectory of the excellent driver; determine a peak point from the natural driving data based on the heading angle, and search from the peak point to both sides to obtain a time point that satisfies the preset condition; The time point matches the initial interval of the corresponding lane-changing trajectory; calculates the peak-to-peak time difference of the lateral acceleration of the lane-changing trajectory according to the initial interval, and calculates the lateral acceleration based on the time difference to obtain a maximum value and a minimum value corresponding to At the time point, the lane-changing duration of the excellent driver under the influence of the real multi-vehicle trajectory of the lane-changing trajectory is obtained.

，峰值点可以为朝向角

大于预设朝向角阈值的点，假设朝向角阈值为2，则自然驾驶数据中朝向角

的点均为峰值点，由峰值点向两边搜索，得到满足预设条件的时间点，其中，预设条件可以为

，从而可以由时间点匹配对应的换道轨迹的初始区间

。Specifically, when extracting the lane-changing trajectory of the vehicle from a large amount of natural driving data of drivers, the embodiment of the present application may first calculate the orientation angle of the lane-changing trajectory

, the peak point can be the heading angle

Points larger than the preset heading angle threshold, assuming the heading angle threshold is 2, then the heading angle in the natural driving data

The points are all peak points, search from the peak point to both sides to obtain the time points that meet the preset conditions, where the preset conditions can be

, so that the initial interval of the corresponding lane-changing trajectory can be matched by the time point

.

As a result, the preliminary extraction of the lane-changing trajectory of the vehicle is realized, and the initial interval of the extracted lane-changing trajectory may cause redundant time intervals in the initial interval of the extracted lane-changing trajectory due to the lateral displacement and speed fluctuation of the vehicle. , therefore, in this embodiment of the present application, the lateral acceleration peak-to-peak value can be used to accurately extract the lane change trajectory.

的计算公式如式（1）所示：Further, the peak-to-peak time difference of the lateral acceleration of the lane change trajectory

The calculation formula of is shown in formula (1):

（1）

(1)

为绝对值计算符号，

为侧向加速度取得最大值时对应的时间点，

为侧向加速度取得最小值时对应的时间点，

为换道轨迹对应的换道时长。in,

Calculate the sign for the absolute value,

is the corresponding time point when the lateral acceleration reaches the maximum value,

is the time point corresponding to the minimum value of lateral acceleration,

It is the lane change duration corresponding to the lane change track.

是两倍的侧向加速度峰-峰值时间差

，故换道时间起点

和换道时间终点

分别如式（2）和式（3）所示：It can be seen from this that the lane change duration corresponding to the lane change trajectory

is twice the lateral acceleration peak-to-peak time difference

, so the lane change time starts

and end of lane change time

As shown in formula (2) and formula (3) respectively:

（2）

(2)

（3）

(3)

为取最小值符号，

为取最大值符号。in,

To take the minimum sign,

for the maximum value sign.

和

，侧向加速度峰-峰值对应的时间点即为

和

。由图2和图3可知，该换道轨迹提取方法可以准确而完整的从自然驾驶数据集中提取出换道轨迹，为下一步的换道时长预测建模奠定了可靠的数据基础。For example, as shown in Fig. 2 and Fig. 3, (a) to (c) in Fig. 2 are the leftward lane changing trajectory, lateral speed and lateral acceleration extracted from the natural driving trajectory set HighD, respectively. (a) to (c) in Figure 3 are the changing laws of the right lane-changing trajectory, lateral speed and lateral acceleration extracted from the natural driving trajectory set HighD. In Figure 2 and Figure 3, the lane change start point and lane change end point correspond to the equations (2) and (3) mentioned in

and

, the time point corresponding to the peak-to-peak value of lateral acceleration is

and

. It can be seen from Figures 2 and 3 that the lane-changing trajectory extraction method can accurately and completely extract the lane-changing trajectory from the natural driving data set, which lays a reliable data foundation for the next lane-changing duration prediction modeling.

In step S102 , obtain the information of the cyclic vehicle movement of the lane-changing vehicle, establish a longitudinal kinematics model of the vehicle under the influence of multiple vehicles according to the information of the cyclic vehicle movement, and characterize the influence of the cyclic vehicle as the movement of the surrounding vehicles The nonlinear mapping of the learning parameters to the average longitudinal acceleration of the lane-changing vehicle is obtained, and then the trained lane-changing duration prediction model under the influence of multiple vehicles and the model parameters obtained by identification are obtained. In specific applications, the UAV can be used to obtain the motion information of vehicles around the lane-changing vehicle, for example, the data in the German natural driving data set HighD; the motion information of the vehicles around the lane-changing vehicle can also be obtained by the roadside camera, for example, Get data from the US Natural Driving Dataset NGSIM.

In a specific example, obtaining the trained lane-changing duration prediction model under the influence of multiple vehicles and the model parameters obtained by identification includes: performing acceleration correction on the average longitudinal acceleration of the lane-changing vehicle to obtain the corrected average longitudinal acceleration , the influence of the lap car is represented as the nonlinear mapping of the kinematic parameters of the lap car to the average longitudinal acceleration of the lane-changing vehicle; The influence of is characterized by the nonlinear mapping of the kinematic parameters of the surrounding vehicles to the average longitudinal acceleration of the lane-changing vehicle, and the lane-changing duration prediction model under the influence of the multi-vehicle and the model parameters obtained by identification are obtained.

的影响，等效系数

（即影响值）与车辆周边的预设范围内的其他车辆状态之间的非线性映射关系，可用式（4）表述：It can be understood that under the ideal driver assumption, the driver changes lanes to pursue higher driving speeds, so in the process of changing lanes without considering the influence of the surrounding vehicles, the driver tends to keep a constant speed or accelerate to change lanes. In order to quickly realize lane change and improve driving efficiency, in a complex driving environment, the driver's lane-changing behavior has to consider the influence of the surrounding vehicles, and the influence of the surrounding vehicles on the longitudinal driving behavior of the own vehicle's lane-changing needs to be determined by self-driving. The driver's stress response is generated. Therefore, in the embodiment of the present application, the influence of other vehicles within a preset range around the vehicle may be equivalent to the average longitudinal acceleration of the own vehicle

The effect of the equivalence factor

(that is, the influence value) and the non-linear mapping relationship between other vehicle states within the preset range around the vehicle, which can be expressed by equation (4):

（4）

(4)

、

和

均为回归系数，可根据实际换道轨迹（即步骤S101中提取到的换道轨迹）进行标定：Among them, a feasible scheme of the nonlinear mapping of formula (4) is the regression model scheme of formula (5) and formula (6). In formula (5) and formula (6),

,

and

are regression coefficients, which can be calibrated according to the actual lane-changing trajectory (that is, the lane-changing trajectory extracted in step S101 ):

（5）

(5)

（6）

(6)

的获取方式可以如图4所示，平均纵向加速度

可以使用如式（7）所示的数学公式进行量化的表达：Further, when the embodiment of the present application performs acceleration correction on the average longitudinal acceleration of the own vehicle according to the influence of the surrounding vehicles, the average longitudinal acceleration is

The acquisition method can be shown in Figure 4, the average longitudinal acceleration

It can be expressed quantitatively using the mathematical formula shown in equation (7):

（7）

(7)

为与车辆周边的预设范围内的其他车辆运动状态相关的等效系数，

为理想条件下，只考虑换道效率收益的预期纵向加速度，

的定义可以如式（8）所示：In formula (7),

is the equivalent coefficient related to other vehicle motion states within the preset range around the vehicle,

Under ideal conditions, only considering the expected longitudinal acceleration of lane change efficiency gains,

The definition of can be shown in formula (8):

（8）

(8)

为目标车道的预期行驶速度，在理想驾驶员假设前提下，驾驶员换道是为了追求更高的行驶效率，因此，

，车辆在换道过程应快速的实现纵向行驶速度的提升，而换道过程中，驾驶员又不得不考虑与周车的交互，在与周车的交互博弈过程中，采取的最终实际纵向加速度是与周车博弈的结果，周车的影响主要体现在式（7）所示的影响值

中。In formula (8),

is the expected driving speed of the target lane. Under the assumption of an ideal driver, the driver changes lanes to pursue higher driving efficiency. Therefore,

, the vehicle should quickly increase the longitudinal driving speed during the lane change process, and during the lane change process, the driver has to consider the interaction with the car, and in the interactive game process with the car, the final actual longitudinal acceleration taken is the result of the game with Zhou Che, and the influence of Zhou Che is mainly reflected in the influence value shown in formula (7).

middle.

In this way, the trained lane-changing duration prediction model under the influence of multiple vehicles can be obtained according to the vehicle lane-changing longitudinal kinematics model as shown in Equation (9).

（9）

(9)

Optionally, in some embodiments, by solving the one-dimensional quadratic equation shown in equation (9) and excluding negative solutions, the lane-changing duration prediction model can be obtained as:

，

,

为自车的换道时长，

为所述自车预期换道纵向位移、

为所述自车平均纵向加速度，

为所述自车换道后在目标车道的预期车速。in,

for the lane changing time of the vehicle,

is the expected lane change longitudinal displacement of the ego vehicle,

is the average longitudinal acceleration of the ego vehicle,

The expected vehicle speed in the target lane after changing lanes for the ego vehicle.

的数学极限条件下，可求得不考虑纵向加速度时的换道时长

。显然，在

时，式（10）为

型数学极限问题，利用洛必达法则同时求得分子分母对

的导数，如式（11）所示：It should be noted that formula (10) is in

Under the mathematical limit of , the lane change time can be obtained without considering the longitudinal acceleration

. Obviously, in

When, formula (10) is

type mathematical limit problem, using Lhobita's rule to simultaneously find the numerator and denominator pairs

The derivative of , as shown in formula (11):

（11）

(11)

的计算公式（10）中，自车平均纵向加速度

趋向零时的解等于预期换道距离

与换道后目标车道预期车速

的比值，符合匀速运动运动学规律，说明传统的纵向匀速换道时长模型是本申请所提出的换道时长预测模型的特例。From Equation (11), it can be seen that when the lane change time

In the calculation formula (10), the average longitudinal acceleration of the ego vehicle

The solution towards zero is equal to the expected lane change distance

and the expected speed of the target lane after the lane change

The ratio of , conforms to the kinematics law of uniform motion, indicating that the traditional longitudinal uniform lane change duration model is a special case of the lane change duration prediction model proposed in this application.

It should be noted that the nonlinear mapping model of the influence of the surrounding car on the average longitudinal acceleration of the own vehicle can not only be obtained by the above regression modeling method, but also obtained by other methods in other examples, such as: neural network model, support Non-parametric modeling methods such as vector machine models, random forest models, and deep learning models are obtained.

In step S103, in the case where the smart car in the network-connected multi-vehicle environment applies the lane-changing duration prediction model to make a decision, when a lane-changing command is issued, based on the expected longitudinal displacement of the lane-changing, the longitudinal speed of the smart car and the expected speed of the target lane after changing lanes, combined with the weekly traffic information of the smart car, using the lane-changing duration prediction model to optimize the lane-changing duration of the smart car, and based on the optimally obtained changes in the smart car. The anthropomorphic trajectory planning is carried out according to the lane duration, so as to realize the accurate implementation of the lane changing intention of the intelligent vehicle.

That is, after obtaining the lane-changing duration prediction model through steps S101 and S102, the lane-changing duration prediction model can be applied to the intelligent vehicle to perform automatic and intelligent lane-changing control.

It should be understood that in a networked multi-vehicle environment, a smart car can use the in-vehicle perception system and V2X (Vehicle to X, vehicle-to-external information exchange), where X can be a car, road, and cloud, and a smart car can Through the sensors of the perception system (such as lidar, millimeter-wave radar, camera, GPS and inertial navigation, etc.) and V2X communication equipment to obtain information on the movement of vehicles around, in order to assess the risks caused by the movement of surrounding vehicles, so as to provide the replacement for the vehicle. Road decisions provide accurate environmental information.

Therefore, when the lane change intention module and the lane change decision module of the intelligent car issue a lane change command, based on the expected lane change longitudinal displacement, the average longitudinal acceleration of the ego vehicle and the expected speed of the target lane after the lane change, combined with the intelligent vehicle perception system and V2X equipment The obtained weekly vehicle information is used to optimize the lane-changing duration of the smart car by using the lane-changing duration prediction model, and anthropomorphic trajectory planning is performed according to the preferably obtained lane-changing duration of the smart car, so as to realize the described Accurate implementation of lane-changing intent in smart cars.

Wherein, the expected lane change longitudinal displacement is determined according to the longitudinal position, longitudinal velocity and longitudinal acceleration state functions.

为各车的纵向位置、纵向速度和纵向加速度状态函数。Specifically, as shown in FIG. 5 , in the process of changing lanes, the surrounding vehicles that can be considered in the embodiment of the present application are the front and rear vehicles from the current lane and the front and rear vehicles from the target lane, respectively. car, car No. 2 is the car behind the current lane, car No. 3 is the car behind the target lane, car No. 4 is the car in front of the target lane,

It is the state function of longitudinal position, longitudinal velocity and longitudinal acceleration of each vehicle.

This solves the problem of insufficient quantification and evaluation of the influence of surrounding vehicles on the longitudinal driving behavior during the lane changing process of the relevant smart car technology, resulting in insufficient rationality, poor interpretability and possible safety of lane change decision-making. The problem of risk, fully excavate the driving control laws of excellent drivers in the natural driving data, and provide a reference for intelligent vehicles to make scientific and reasonable lane changing decisions. The concept is reflected in the lane change decision.

In order for those skilled in the art to further understand the lane changing duration prediction and anthropomorphic trajectory planning method of the smart car according to the embodiment of the present application, the following takes the natural driving data set HighD as an example for detailed description.

Specifically, there are more than 5,600 lane-changing trajectories in the HighD data set, from which the data of multiple lane-changes are excluded, and the lane-changing trajectories that conform to the environment shown in Figure 5 during the lane-changing process are extracted, and 1,000 lane-changing trajectories are randomly selected. The lane change trajectory is used as the research object.

、纵向初始速度

和平均等效纵向加速度

的实测结果的统计分布图。由图6可知，纵向位移

服从正态分布，换道过程纵向位移分布在

的区间内。初始纵向速度

呈双峰分布，且集中在

的高速行驶区间内。由自车平均纵向加速度

的分布特性可知，受到周车影响，换道过程可能产生平均加速度为负的场景，而在大部分场景下，驾驶员希望维持平均纵向加速度为零的工况，从而把更多的注意力放在顺利完成换道的横向驾驶行为上。As shown in Fig. 6, Fig. 6(a) to Fig. 6(c) are the longitudinal displacements of these lane changing trajectories respectively

, longitudinal initial velocity

and the mean equivalent longitudinal acceleration

Statistical distribution of the measured results. It can be seen from Figure 6 that the longitudinal displacement

It obeys a normal distribution, and the longitudinal displacement in the lane changing process is distributed in

within the range. initial longitudinal velocity

has a bimodal distribution and is concentrated in

within the high-speed driving area. Average longitudinal acceleration by ego vehicle

It can be seen from the distribution of On the lateral driving behavior that successfully completes the lane change.

Further, as shown in FIG. 7 , FIG. 7( a ) and FIG. 7( b ) are histograms of the comparison between the lane-changing duration of the current vehicle predicted according to the present application and the lane-changing duration corresponding to the initially measured lane-changing trajectory. and scatter plot, it can be seen from the statistical histogram that the lane-changing duration obeys the log-normal distribution, and the log-normal distribution is used to establish the probability density curve for predicting the lane-changing duration and the statistical histogram of the lane-changing duration obtained from the actual measurement. is consistent, indicating that the accuracy of the lane-changing duration of the current vehicle predicted by this application is relatively high. By comparing the scatter plot of the lane-changing duration corresponding to the lane-changing trajectory and the predicted lane-changing duration of the current vehicle, it can be seen that the predicted lane-changing duration is evenly dispersed on both sides of the actual value, indicating that the modeling effect is good.

To sum up, as shown in Figure 8, the application first extracts the complete lane-changing trajectory from the natural driving data, and calculates the lane-changing duration; secondly, the kinematic parameters of the vehicles around the self-vehicle are obtained by using the network connection technology, which is the influence of the surrounding vehicles. Provide data for the quantitative evaluation under Predicted value (the lane-changing duration of the current vehicle obtained using the lane-changing duration prediction model).

According to the method for predicting the lane-changing duration and anthropomorphic trajectory planning of a smart car proposed in the embodiment of the present application, the lane-changing trajectory of an excellent driver can be extracted from a large number of natural driving data of drivers, and the lane-changing trajectory of the excellent driver can be extracted from the lane-changing trajectory of the excellent driver. Extract the lane-changing duration of excellent drivers under the influence of real multi-vehicles, and then obtain the traffic information of the vehicles changing lanes. The influence is represented by the nonlinear mapping of the kinematic parameters of the surrounding vehicles to the average longitudinal acceleration of the lane-changing vehicles, and then the prediction model of the lane-changing duration under the influence of multiple vehicles can be obtained. After the smart car in the environment, the smart car can use the lane-changing duration prediction model to predict the intelligent car according to the expected longitudinal displacement of the lane change, the longitudinal speed of the smart car and the expected speed of the target lane after the lane change, etc. The lane-changing duration of the car is optimized, and further, anthropomorphic trajectory planning can be performed according to the lane-changing duration of the smart car obtained by optimization, so as to realize the accurate implementation of the lane-changing intention of the smart car. This solves the problem of insufficient quantification and evaluation of the influence of surrounding vehicles on the longitudinal driving behavior during the lane changing process of the relevant smart car technology, resulting in insufficient rationality, poor interpretability and possible safety of lane change decision-making. To solve the problem of risk, the embodiment of the present invention fully excavates the driving and maneuvering rules of excellent drivers in the natural driving data of a large number of drivers, provides a reference for intelligent vehicles to make scientific and reasonable lane-changing decisions, and makes the lane-changing process of intelligent vehicles more reasonable. , and then improve the safety and reliability of smart cars, which is the embodiment of the anthropomorphic decision-making concept of smart cars "learning, simulating, surpassing and serving people" in lane-changing decision-making.

Next, the device for predicting the lane changing duration and anthropomorphic trajectory planning of a smart car according to the embodiments of the present application will be described with reference to the accompanying drawings.

FIG. 9 is a schematic block diagram of an apparatus for predicting the lane changing duration and anthropomorphic trajectory planning of an intelligent vehicle according to an embodiment of the present application.

As shown in FIG. 9 , the intelligent vehicle lane change duration prediction and anthropomorphic trajectory planning device 10 includes an extraction module 100 , a generation module 200 and a lane change trajectory planning module 300 .

Wherein, the extraction module 100 is used for extracting the lane-changing trajectory of the excellent driver from the natural driving data of several drivers, and extracting the lane-changing duration of the excellent driver under the influence of the real multi-vehicle from the lane-changing trajectory of the excellent driver;

The generation module 200 is used to obtain the information of the lane-changing vehicle's movement around the vehicle, establish a longitudinal kinematics model of the vehicle lane-changing under the influence of multiple vehicles according to the information on the movement of the vehicle around the vehicle, and characterize the influence of the vehicle as the movement of the vehicle. The nonlinear mapping of the learning parameters to the average longitudinal acceleration of the lane-changing vehicle, and then the trained lane-changing duration prediction model under the influence of multiple vehicles and the model parameters obtained by identification are obtained; and

The lane-changing trajectory planning module 300 is configured to, in the case where the smart car in the network-connected multi-vehicle environment applies the lane-changing duration prediction model to make a decision, when a lane-changing command is issued, based on the expected lane-changing longitudinal displacement, the The longitudinal speed of the smart car and the expected speed of the target lane after changing lanes, combined with the weekly traffic information of the smart car, the lane-changing duration prediction model is used to optimize the lane-changing duration of the smart car, and according to the obtained An anthropomorphic trajectory planning is carried out according to the lane-changing duration of the intelligent vehicle, so as to realize the accurate implementation of the lane-changing intention of the intelligent vehicle.

Optionally, the extraction module 100 is specifically used for:

calculating the heading angle of the lane-changing trajectory of the excellent driver;

Determine a peak point from the natural driving data based on the heading angle, and search both sides from the peak point to obtain a time point that satisfies a preset condition;

matching the initial interval of the corresponding lane-changing trajectory by the time point;

Calculate the peak-to-peak time difference of the lateral acceleration of the lane-changing trajectory according to the initial interval, and calculate the corresponding time points when the lateral acceleration obtains the maximum value and the minimum value based on the time difference, and obtain the lane-changing trajectory in the real multi-vehicle. A good driver's lane-changing time under the influence.

Optionally, the generating module 200 is specifically used for:

Acceleration correction is performed on the average longitudinal acceleration of the lane-changing vehicle to obtain the corrected average longitudinal acceleration, and the influence of the surrounding car is characterized as a nonlinear mapping of the kinematic parameters of the surrounding vehicle to the average longitudinal acceleration of the lane-changing vehicle ;

According to the vehicle lane-changing longitudinal kinematics model under the influence of the multi-vehicle and the influence of the circling vehicle, which is characterized as a nonlinear mapping of the kinematic parameters of the circling vehicle to the average longitudinal acceleration of the lane-changing vehicle, the multi-vehicle influence is obtained. The lane-changing duration prediction model and the model parameters obtained by identification.

Optionally, the lane change duration prediction model is:

，

,

为自车的换道时长，

为所述自车预期换道纵向位移、

为所述自车平均纵向加速度，

为所述自车换道后在目标车道的预期车速。in,

for the lane changing time of the vehicle,

is the expected lane change longitudinal displacement of the ego vehicle,

is the average longitudinal acceleration of the ego vehicle,

The expected vehicle speed in the target lane after changing lanes for the ego vehicle.

Optionally, the expected lane change longitudinal displacement is determined from a longitudinal position, longitudinal velocity and longitudinal acceleration state function.

It should be noted that the foregoing explanations of the embodiment of the method for predicting the lane-changing duration and anthropomorphic trajectory planning of a smart car are also applicable to the device for predicting the lane-changing duration and anthropomorphic trajectory planning of a smart car in this embodiment, which will not be repeated here.

According to the device for predicting the lane-changing duration and anthropomorphic trajectory planning of a smart car proposed in the embodiment of the present application, the lane-changing trajectory of an excellent driver can be extracted from a large number of natural driving data of drivers, and the lane-changing trajectory of the excellent driver can be extracted from the lane-changing trajectory of the excellent driver. Extract the lane-changing duration of excellent drivers under the influence of real multi-vehicles, and then obtain the traffic information of the vehicles changing lanes. The influence is represented by the nonlinear mapping of the kinematic parameters of the surrounding vehicles to the average longitudinal acceleration of the lane-changing vehicles, and then the prediction model of the lane-changing duration under the influence of multiple vehicles can be obtained. After the smart car in the environment, the smart car can use the lane-changing duration prediction model to predict the intelligent car according to the expected longitudinal displacement of the lane change, the longitudinal speed of the smart car and the expected speed of the target lane after the lane change, etc. The lane-changing duration of the car is optimized, and further, anthropomorphic trajectory planning can be performed according to the lane-changing duration of the smart car obtained by optimization, so as to realize the accurate implementation of the lane-changing intention of the smart car. This solves the problem of insufficient quantification and evaluation of the influence of surrounding vehicles on the longitudinal driving behavior during the lane changing process of the relevant smart car technology, resulting in insufficient rationality, poor interpretability and possible safety of lane change decision-making. To solve the problem of risk, the embodiment of the present invention fully excavates the driving and maneuvering rules of excellent drivers in the natural driving data of a large number of drivers, provides a reference for intelligent vehicles to make scientific and reasonable lane-changing decisions, and makes the lane-changing process of intelligent vehicles more reasonable. , and then improve the safety and reliability of smart cars, which is the embodiment of the anthropomorphic decision-making concept of smart cars "learning, simulating, surpassing and serving people" in lane-changing decision-making.

In addition, an embodiment of the present application provides an intelligent vehicle lane-changing trajectory decision-making device, including: a lane-changing intention recognition module and a lane-changing trajectory planning module, wherein the lane-changing duration prediction and anthropomorphic method of the intelligent vehicle described in any one of the above embodiments The integrated trajectory planning method serves the lane change intention identification module and the lane change trajectory planning module.

In addition, an embodiment of the fourth aspect of the present application provides a lane-changing trajectory tracking module, on which a computer program is stored, characterized in that the program is executed by a processor, so as to realize the above-mentioned intelligent vehicle lane-changing duration prediction and anthropomorphism method for trajectory planning.

In the description of this specification, description with reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples", etc., mean specific features described in connection with the embodiment or example , structure, material or feature is included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials or characteristics described may be combined in any suitable manner in any one or N of the embodiments or examples. Furthermore, those skilled in the art may combine and combine the different embodiments or examples described in this specification, as well as the features of the different embodiments or examples, without conflicting each other.

In addition, the terms "first" and "second" are only used for descriptive purposes, and should not be construed as indicating or implying relative importance or implying the number of indicated technical features. Thus, a feature delimited with "first", "second" may expressly or implicitly include at least one of that feature. In the description of the present application, "N" means at least two, such as two, three, etc., unless otherwise expressly and specifically defined.

Any process or method description in the flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or N more executable instructions for implementing custom logical functions or steps of the process , and the scope of the preferred embodiments of the present application includes alternative implementations in which the functions may be performed out of the order shown or discussed, including performing the functions substantially concurrently or in the reverse order depending upon the functions involved, which should It is understood by those skilled in the art to which the embodiments of the present application belong.

The logic and/or steps represented in flowcharts or otherwise described herein, for example, may be considered an ordered listing of executable instructions for implementing the logical functions, may be embodied in any computer-readable medium, For use with, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other system that can fetch instructions from and execute instructions from an instruction execution system, apparatus, or apparatus) or equipment. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport the program for use by or in connection with an instruction execution system, apparatus, or apparatus. More specific examples (non-exhaustive list) of computer readable media include the following: electrical connections (electronic devices) with one or N wires, portable computer disk cartridges (magnetic devices), random access memory (RAM), Read Only Memory (ROM), Erasable Editable Read Only Memory (EPROM or Flash Memory), Fiber Optic Devices, and Portable Compact Disc Read Only Memory (CDROM). In addition, the computer readable medium may even be paper or other suitable medium on which the program may be printed, as the paper or other medium may be optically scanned, for example, followed by editing, interpretation, or other suitable medium as necessary process to obtain the program electronically and then store it in computer memory.

It should be understood that various parts of this application may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the N steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented by any one of the following techniques known in the art, or a combination thereof: discrete with logic gates for implementing logic functions on data signals Logic circuits, ASICs with suitable combinational logic gates, Programmable Gate Arrays (PGA), Field Programmable Gate Arrays (FPGA), etc.

Those of ordinary skill in the art can understand that all or part of the steps carried by the methods of the above embodiments can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, and the program is stored in a computer-readable storage medium. When executed, one or a combination of the steps of the method embodiment is included.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing module, or each unit may exist physically alone, or two or more units may be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware, and can also be implemented in the form of software function modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.

The above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, and the like. Although the embodiments of the present application have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limitations to the present application. Embodiments are subject to variations, modifications, substitutions and variations.

Claims (9)

1. An intelligent automobile lane change duration prediction and anthropomorphic track planning method is characterized by comprising the following steps:

extracting a lane changing track of an excellent driver from natural driving data of a plurality of drivers, and extracting lane changing time length of the excellent driver under the influence of real multiple vehicles from the lane changing track of the excellent driver;

acquiring the vehicle-to-vehicle motion information of the lane-changing vehicle, establishing a vehicle lane-changing longitudinal kinematics model under the influence of multiple vehicles according to the vehicle-to-vehicle motion information, representing the influence of the vehicles as the nonlinear mapping of the vehicle-to-vehicle kinematics parameters to the average longitudinal acceleration of the lane-changing vehicle, and further acquiring the trained influence of the multiple vehiclesThe lane change duration prediction model and the model parameters obtained by identification are as follows:

wherein

in order to prolong the lane change time of the self vehicle,

the lane change longitudinal displacement is expected for the self-vehicle,

is the average longitudinal acceleration of the own vehicle,

the expected speed of the vehicle in the target lane after the vehicle is changed;

under the condition that an intelligent automobile in a network multi-automobile environment uses the lane change duration prediction model to make a decision, when a lane change instruction is given, the lane change duration of the intelligent automobile is optimized by using the lane change duration prediction model based on the expected lane change longitudinal displacement, the longitudinal speed of the intelligent automobile and the expected speed of a target lane after lane change in combination with the week information of the intelligent automobile, and the lane change duration of the intelligent automobile is subjected to anthropomorphic track planning according to the lane change duration of the intelligent automobile obtained through optimization, so that the lane change intention of the intelligent automobile is accurately implemented.

2. The method according to claim 1, wherein the extracting of the lane change trajectory of the excellent driver from the natural driving data of several drivers and the extracting of the excellent driver lane change duration under the influence of real multi-vehicle from the excellent driver lane change trajectory comprises:

calculating an orientation angle of a lane change trajectory of the excellent driver;

determining a peak point from the natural driving data based on the orientation angle, and searching from the peak point to two sides to obtain a time point meeting a preset condition;

matching the initial interval of the corresponding lane changing track by the time point;

and calculating the peak-peak time difference of the lateral acceleration of the track changing track according to the initial interval, and calculating corresponding time points when the lateral acceleration obtains the maximum value and the minimum value based on the time difference to obtain the excellent track changing time of the driver under the influence of real multiple vehicles.

3. The method of claim 1, wherein the obtaining a trained lane change duration prediction model under the influence of multiple vehicles and identifying obtained model parameters comprises:

performing acceleration correction on the average longitudinal acceleration of the lane changing vehicle to obtain the corrected average longitudinal acceleration, and representing the influence of the week vehicle as nonlinear mapping of the kinetic parameters of the week vehicle on the average longitudinal acceleration of the lane changing vehicle;

and according to the longitudinal vehicle lane change kinematics model under the influence of the multiple vehicles and the influence characterization of the vehicles in the week, obtaining a lane change duration prediction model under the influence of the multiple vehicles and model parameters obtained by identification by nonlinear mapping of the vehicle lane change kinematics parameters to the average longitudinal acceleration of the lane change vehicle.

4. The method of claim 1, wherein the expected lane-change longitudinal displacement is determined as a function of longitudinal position, longitudinal velocity, and longitudinal acceleration state.

5. The utility model provides an intelligent automobile lane change duration prediction and anthropomorphic track planning device which characterized in that includes:

the extraction module is used for extracting a lane changing track of an excellent driver from natural driving data of a plurality of drivers and extracting the lane changing time length of the excellent driver under the influence of real multiple vehicles from the lane changing track of the excellent driver;

generatingThe module is used for acquiring the vehicle-to-vehicle motion information of a lane-changing vehicle, establishing a vehicle lane-changing longitudinal kinematic model under the influence of multiple vehicles according to the vehicle-to-vehicle motion information, representing the influence of the vehicles as the nonlinear mapping of the vehicle-to-vehicle kinematic parameters to the average longitudinal acceleration of the lane-changing vehicle, and further acquiring a trained lane-changing duration prediction model under the influence of the multiple vehicles and model parameters obtained by identification, wherein the lane-changing duration prediction model is as follows:

wherein

in order to prolong the lane change time of the self vehicle,

the lane change longitudinal displacement is expected for the self-vehicle,

is the average longitudinal acceleration of the own vehicle,

the expected speed of the vehicle in the target lane after the vehicle is changed; and

and the lane change track planning module is used for optimizing the lane change time length of the intelligent automobile by using the lane change time length prediction model according to the lane change time length prediction model and the acquired lane change time length of the intelligent automobile according to the optimization when a lane change instruction is issued under the condition that the intelligent automobile applies the lane change time length prediction model to make a decision, and the lane change time length prediction model is combined with the week automobile information of the intelligent automobile, so that the accurate implementation of the lane change intention of the intelligent automobile is realized.

6. The apparatus according to claim 5, wherein the extraction module is specifically configured to:

calculating an orientation angle of a lane change trajectory of the excellent driver;

determining a peak point from the natural driving data based on the orientation angle, and searching from the peak point to two sides to obtain a time point meeting a preset condition;

matching the initial interval of the corresponding lane changing track by the time point;

and calculating the peak-peak time difference of the lateral acceleration of the track changing track according to the initial interval, and calculating corresponding time points when the lateral acceleration obtains the maximum value and the minimum value based on the time difference to obtain the excellent track changing time of the driver under the influence of real multiple vehicles.

7. The apparatus of claim 5, wherein the generating module is specifically configured to:

performing acceleration correction on the average longitudinal acceleration of the lane changing vehicle to obtain the corrected average longitudinal acceleration, and representing the influence of the week vehicle as nonlinear mapping of the kinetic parameters of the week vehicle on the average longitudinal acceleration of the lane changing vehicle;

and according to the longitudinal vehicle lane change kinematics model under the influence of the multiple vehicles and the influence characterization of the vehicles in the week, obtaining a lane change duration prediction model under the influence of the multiple vehicles and model parameters obtained by identification by nonlinear mapping of the vehicle lane change kinematics parameters to the average longitudinal acceleration of the lane change vehicle.

8. An intelligent automobile track change decision-making device is characterized by comprising: the lane change intention identification module and the lane change trajectory planning module, wherein the intelligent automobile lane change duration prediction and anthropomorphic trajectory planning method according to any one of claims 1-4 serves the lane change intention identification module and the lane change trajectory planning module.

9. A lane change trajectory tracking module, on which a computer program is stored, wherein the program is executed by a processor to implement the intelligent automobile lane change duration prediction and anthropomorphic trajectory planning method according to any one of claims 1-4.

Images