CN114228746B

CN114228746B - Method and device for predicting motion trail of vehicle

Info

Publication number: CN114228746B
Application number: CN202210050559.XA
Authority: CN
Inventors: 颜学术; 李继扬
Original assignee: Beijing Jingwei Hirain Tech Co Ltd
Current assignee: Beijing Jingwei Hirain Tech Co Ltd
Priority date: 2022-01-17
Filing date: 2022-01-17
Publication date: 2024-05-07
Anticipated expiration: 2042-01-17
Also published as: CN114228746A

Abstract

The invention provides a method and a device for predicting a motion trail of a vehicle, which utilize the acquired absolute motion state of a specified target vehicle to determine average kinematic parameters and predict the motion trail to be processed according to the average kinematic parameters; predicting to obtain an initial driving intention based on the motion trail to be processed and the historical motion trail of the appointed target vehicle; determining the type of a lane line of a lane where a specified target vehicle is located and determining obstacle vehicle information; the initial driving intention is filtered by using the road line type and obstacle vehicle information to obtain the final driving intention and determine the transverse target position according to the final driving intention; and predicting the vehicle motion trail of the appointed target vehicle by using the transverse target position and the average kinematics parameter. The method has low requirement on hardware calculation force, and can quickly locate which link is abnormal, so that the abnormality is solved in time to improve the safety.

Description

Method and device for predicting motion trail of vehicle

Technical Field

The invention relates to the technical field of vehicles, in particular to a method and a device for predicting a vehicle motion trail.

Background

During traveling, an intelligent vehicle (i.e., a vehicle having an L2 or higher level automatic driving function) needs to recognize a movement track of a vehicle around the vehicle, for example, a cut-in/cut-out behavior of the vehicle around the vehicle on a lane of the vehicle, in order to ensure traveling safety.

At present, the mode for identifying the motion trail of the vehicle mainly comprises the following steps: and training a large amount of driving data acquired by the acquisition by using a machine learning algorithm to obtain a prediction model for predicting the motion trail. On the one hand, the parameter quantity related to the prediction model is huge, and the calculation platform carrying the prediction model has higher performance requirements, namely, higher hardware calculation power is required for applying the prediction model, and the hardware calculation power of the intelligent vehicle in mass production cannot meet the performance requirements for applying the prediction model; on the other hand, when the calculation result of the prediction model is abnormal, the difficulty of analyzing the cause of the abnormality from the prediction model is high, so that the abnormality cannot be solved in time, and the safety is poor.

Disclosure of Invention

In view of the above, the embodiment of the invention provides a method and a device for predicting a vehicle motion trail, so as to solve the problems of higher requirement on hardware calculation force, poorer safety and the like in the existing mode of predicting the vehicle motion trail.

In order to achieve the above object, the embodiment of the present invention provides the following technical solutions:

An embodiment of the present invention discloses a method for predicting a motion trajectory of a vehicle, which is applied to a device for predicting a motion trajectory of a vehicle mounted in a vehicle, and includes:

Determining a specified target vehicle from the detected target vehicles;

Determining and obtaining an average kinematic parameter by utilizing the absolute motion states of the specified target vehicle acquired in a plurality of history periods, and predicting a to-be-processed motion track of the specified target vehicle based on the average kinematic parameter, wherein the average kinematic parameter is used for describing the average motion state of the specified target vehicle in the plurality of history periods;

Calculating a normalized mean value of motion feature vectors of the specified target vehicle based on the motion trail to be processed and the historical motion trail of the specified target vehicle;

Mapping the normalized mean value into a preset numerical interval to obtain a running intention numerical value, and predicting according to the running intention numerical value to obtain an initial running intention;

Scanning the surrounding environment of the appointed target vehicle to obtain the type of the lane line of the lane where the appointed target vehicle is located and obtain the information of the obstacle vehicle in the lane adjacent to the lane where the appointed target vehicle is located;

Filtering the initial driving intention by using the road line type and the obstacle vehicle information, and predicting to obtain a final driving intention;

determining a lateral target position based on the final travel intent;

And predicting the vehicle motion trail of the appointed target vehicle from the current position to the target position by using the transverse target position and the average kinematic parameter.

Preferably, determining an average kinematic parameter by using the absolute motion states of the specified target vehicle acquired in a plurality of history periods, and predicting a motion track to be processed of the specified target vehicle based on the average kinematic parameter, including:

determining a time interval of a plurality of history periods;

dividing the absolute motion state of the appointed target vehicle acquired in the plurality of history periods by the time interval to obtain an average kinematic parameter;

and processing the average kinematic parameters by using a kinematic model, and predicting the motion trail to be processed of the appointed target vehicle.

Preferably, calculating a normalized mean value of motion feature vectors of the specified target vehicle based on the motion trail to be processed and a prestored historical motion trail of the specified target vehicle includes:

combining the motion trail to be processed and the prestored historical motion trail of the appointed target vehicle to obtain an input motion trail;

Calculating a central line transverse position and a lane line course angle of a lane central line corresponding to the input motion trail by using a lane line expression;

For a vehicle lateral position and a vehicle heading angle of the specified target vehicle in the input motion trajectory, calculating a lateral position deviation between the vehicle lateral position and the center line lateral position, and calculating a heading angle deviation between the vehicle heading angle and the lane heading angle;

Dividing the transverse position deviation by the lane width to normalize the transverse position deviation, averaging all normalized transverse position deviations, and determining a normalized mean value of the transverse position deviation;

Dividing the course angle deviation by the maximum course angle of the appointed target vehicle to normalize the course angle deviation, averaging all normalized course angle deviations, and determining to obtain a normalized mean value of the course angle deviation, wherein the maximum course angle of the appointed target vehicle is calculated based on the speed of the appointed target vehicle.

Preferably, mapping the normalized mean value to a preset value interval to obtain a running intention value, and predicting according to the running intention value to obtain an initial running intention, including:

Mapping the normalized mean value of the transverse position deviation into a preset value interval by using a nonlinear mapping function obtained based on an S-shaped function to obtain a first driving intention value, and mapping the normalized mean value of the course angle deviation into the preset value interval to obtain a second driving intention value;

And under the condition that the first driving intention value and the second driving intention value meet a left lane change rule, predicting the initial driving intention as a left lane change intention, wherein the left lane change rule is as follows: the first driving intention value is in a first left lane change value interval, and the second driving intention value is in a second left lane change value interval;

Under the condition that the first driving intention value and the second driving intention value meet a right lane change rule, predicting the initial driving intention as a right lane change intention, wherein the right lane change rule is as follows: the first driving intention value is in a first right lane change value interval, and the second driving intention value is in a second right lane change value interval;

in a case where the first and second travel intention values do not satisfy the left lane-changing rule, and the first and second travel intention values do not satisfy the right lane-changing rule, the initial travel intention is predicted to be a lane-keeping intention.

Preferably, the lane line type of the lane where the specified target vehicle is located includes: the lane line type of the left lane line and the lane line type of the right lane line; the track line type is as follows: a lane line that allows lane change or a lane line that does not allow lane change.

Preferably, the initial travel intention is a left lane change intention, a right lane change intention, or a lane keeping intention;

And performing filtering processing on the initial driving intention by using the road line type and the obstacle vehicle information, and predicting to obtain a final driving intention, wherein the filtering processing comprises the following steps:

Determining a lane change condition satisfied by the specified target vehicle by using the lane line type and the obstacle vehicle information, wherein the lane change condition at least comprises: a first condition for indicating that a lane change to the right is permitted, and a second condition for indicating that a lane change to the left is permitted;

In the case where the initial travel intention is a left lane change intention, predicting a final travel intention as a lane keeping intention when the specified target vehicle does not satisfy the second condition; when the specified target vehicle meets the second condition, predicting that the final driving intention is a left lane change intention;

Predicting a final travel intention as a lane keeping intention when the specified target vehicle does not satisfy the first condition in a case where the initial travel intention is a right lane change intention; and when the specified target vehicle meets the first condition, predicting the final driving intention as a right lane change intention.

Preferably, the final driving intention is a left lane change intention, a right lane change intention, or a lane keeping intention;

determining a lateral target position based on the final travel intent, comprising:

When the final driving intention is the left lane change intention, determining the central line transverse position of a lane adjacent to the lane where the appointed target vehicle is positioned as a transverse target position;

When the final driving intention is right lane change intention, determining the central line transverse position of a lane adjacent to the right of the lane where the appointed target vehicle is positioned as a transverse target position;

And determining the transverse position of the central line of the lane where the specified target vehicle is located as the transverse target position under the condition that the final driving intention is lane keeping intention.

Preferably, predicting a vehicle motion trajectory of the specified target vehicle from a current position to a target position using the lateral target position and the average kinematic parameter includes:

Generating a transverse moving track of the appointed target vehicle moving to the transverse target position by using a preset track generation mode;

Generating a longitudinal movement track of the appointed target vehicle according to the average kinematic parameter;

and predicting the vehicle motion trail of the appointed target vehicle from the current position to the target position by combining the transverse motion trail and the longitudinal motion trail.

Preferably, the preset track generation mode is an equal proportion attenuation mode or a polynomial curve fitting mode;

In the case that the track generation mode is an equal proportion attenuation mode, in the generated transverse movement track, a data point at the kth moment is y _k＝αy_k-1,y_k, and the difference value between the transverse position of the appointed target vehicle and the transverse target position at the kth moment is 0< alpha <1;

in the generated longitudinal movement track, the data point at the kth moment is X _k is the longitudinal position of the specified target vehicle aT the kth time, v is the speed, a is the acceleration, T is the length of the fixed period, and v aT the kth time is v _k＝v_k-1 +aT;

And in the predicted motion trail of the vehicle, the trail position of the target appointed vehicle at the kth moment is (x _k,y_k).

A second aspect of an embodiment of the present invention discloses a device for predicting a motion trajectory of a vehicle, the device being mounted in the vehicle, the device comprising:

a first determining unit configured to determine a specified target vehicle from among all the detected target vehicles;

The processing unit is used for determining and obtaining average kinematic parameters by utilizing the absolute motion states of the specified target vehicle acquired in a plurality of history periods, and predicting a to-be-processed motion track of the specified target vehicle based on the average kinematic parameters, wherein the average kinematic parameters are used for describing the average motion states of the specified target vehicle in the plurality of history periods;

a calculating unit, configured to calculate a normalized mean value of motion feature vectors of the specified target vehicle based on the motion trail to be processed and the historical motion trail of the specified target vehicle;

the first prediction unit is used for mapping the normalized mean value into a preset numerical interval to obtain a driving intention value, and predicting to obtain an initial driving intention according to the driving intention value;

The scanning unit is used for scanning the surrounding environment of the appointed target vehicle, obtaining the lane line type of the lane where the appointed target vehicle is located, and obtaining obstacle vehicle information in the lane adjacent to the lane where the appointed target vehicle is located;

the second prediction unit is used for carrying out filtering processing on the initial driving intention by utilizing the road line type and the obstacle vehicle information and predicting to obtain a final driving intention;

a second determination unit configured to determine a lateral target position based on the final travel intention;

And the third prediction unit is used for generating a vehicle motion track of the appointed target vehicle from the current position to the target position by using the transverse target position and the average kinematics parameter.

Based on the method and the device for predicting the motion trail of the vehicle provided by the embodiment of the invention, the method comprises the following steps: determining a designated target vehicle; determining and obtaining average kinematic parameters by utilizing the absolute motion states of the appointed target vehicle acquired in a plurality of history periods, and predicting the motion trail to be processed of the appointed target vehicle based on the average kinematic parameters; calculating a normalized mean value of motion feature vectors of the specified target vehicle based on the motion trail to be processed and the historical motion trail of the specified target vehicle; mapping the normalized mean value into a preset numerical interval to obtain a driving intention numerical value, and predicting according to the driving intention numerical value to obtain an initial driving intention; scanning the surrounding environment of the appointed target vehicle to obtain the type of the lane line of the lane where the appointed target vehicle is located and obtain the information of the obstacle vehicle in the lane adjacent to the lane where the appointed target vehicle is located; the initial driving intention is filtered by using the road line type and obstacle vehicle information, and the final driving intention is predicted; determining a lateral target position based on the final travel intent; and predicting the vehicle motion trail of the appointed target vehicle from the current position to the target position by using the transverse target position and the average kinematics parameter. By adopting the scheme, the historical motion trail and other parameters of the appointed target vehicle are calculated to obtain the running intention, then the motion trail of the vehicle is predicted according to the running intention, less parameters are used for predicting the motion trail of the vehicle, and the requirement on hardware calculation force is low; and the vehicle motion trail is predicted in a layering mode, when the prediction result is abnormal, which link is abnormal can be rapidly positioned, and the abnormality is further solved in time so as to improve the safety.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present invention, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flowchart of a method for predicting a motion trajectory of a vehicle according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of defining an inertial reference coordinate system and a host vehicle coordinate system according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a nonlinear mapping function according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a surrounding environment of a scan specification target vehicle according to an embodiment of the present invention;

FIG. 5 is a flowchart of calculating a normalized mean value according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of predicting a motion trajectory of a vehicle according to an embodiment of the present invention;

Fig. 7 is a block diagram of a device for predicting a motion trajectory of a vehicle according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

In the present disclosure, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

As known from the background art, a machine learning algorithm is generally used to train to obtain a prediction model for predicting a motion track. However, the parameters related to the prediction model are huge, the hardware calculation force requirement on the intelligent vehicle is high, and when the calculation result of the prediction model is abnormal, the difficulty of analyzing the cause of the abnormality from the prediction model is high, so that the abnormality cannot be solved in time, and the safety is poor.

Therefore, the embodiment of the invention provides a method and a device for predicting a vehicle motion trail, which are used for calculating the historical motion trail and other parameters of a specified target vehicle to obtain a running intention, predicting the vehicle motion trail according to the running intention, and predicting the vehicle motion trail with fewer parameters so as to reduce the requirement on hardware calculation force; and the vehicle motion trail is predicted in a layering mode, when the prediction result is abnormal, which link is abnormal can be rapidly positioned, and the abnormality is further solved in time so as to improve the safety.

Referring to fig. 1, a flowchart of a method for predicting a motion trail of a vehicle according to an embodiment of the present invention is shown, where the method is applied to an apparatus for predicting a motion trail of a vehicle mounted in a vehicle, and the method includes:

Step S101: a designated target vehicle is determined from the detected target vehicles.

In addition, a sensing system is installed in a vehicle, and the sensing system may detect each target vehicle and target information of each target vehicle during driving, for example: and detecting information such as the position, speed, acceleration, course angle, category and the like of the target vehicle.

In the specific implementation process of step S101, a specified target vehicle is screened out from all detected target vehicles according to a preset screening rule, where the specified target vehicle is the target vehicle that needs to predict the motion trail of the vehicle.

In some embodiments, from among all detected target vehicles, a target vehicle whose lateral distance from the vehicle is less than a first threshold and whose longitudinal distance from the vehicle is less than a second threshold is determined to be a designated target vehicle. That is, for a certain target vehicle, if the lateral distance of the target vehicle from the vehicle is less than a first threshold value and the longitudinal distance of the target vehicle from the vehicle is less than a second threshold value, the target vehicle may be determined to be a specified target vehicle.

It should be noted that, the above manner of determining the specified target vehicle by the lateral distance and the longitudinal distance from the vehicle is only one example of screening out the specified target vehicle, and the screening rule for screening out the specified target vehicle may be defined according to other requirements, and the specific content of the screening rule is not limited in the embodiment of the present invention.

Step S102: and determining and obtaining average kinematic parameters by utilizing the absolute motion states of the specified target vehicle acquired in a plurality of history periods, and predicting the to-be-processed motion trail of the specified target vehicle based on the average kinematic parameters.

As can be seen from the above, the sensing system can acquire the target information of the target vehicle, specifically, the target information acquired by the sensing system is a relative motion state defined in a coordinate system of the vehicle (the coordinate system is referred to as a host vehicle coordinate system), for example: the speed acquired by the sensing system is the relative speed, and the acceleration acquired by the sensing system is the relative acceleration. In practical application, the acquired relative motion state of the target vehicle needs to be converted into an absolute motion state, and the specific conversion mode is as follows: acquiring chassis information (including information such as vehicle speed, acceleration, yaw rate and the like) of a vehicle (namely, a host vehicle) through a vehicle-mounted bus; according to the obtained chassis information, the relative motion state of the target vehicle under the own vehicle coordinate system is converted into the absolute motion state under the inertial reference coordinate system, and finally the absolute motion states (such as absolute speed, absolute acceleration and the like) of the target vehicle and the vehicle can be obtained.

Note that, the inertial reference coordinate system refers to a geodetic coordinate system in which the current moment coincides with the origin of the own vehicle coordinate system, please refer to fig. 2, fig. 2 is a schematic diagram of defining the inertial reference coordinate system and the own vehicle coordinate system; in fig. 2, the relative motion state of the specified target vehicle in the own vehicle coordinate system is converted to the absolute motion state in the inertial reference coordinate system. Note that in fig. 2, numbers 101 to 104 represent lane lines, number 200 represents a specified target vehicle, and number 300 represents a vehicle (i.e., host vehicle).

It can be understood that in the solution provided in the embodiment of the present invention, the relevant calculation tasks are run according to a certain fixed period, for example, a fixed period of t=20ms; each history period is collected to obtain the absolute motion state of the appointed target vehicle.

The average kinematic parameter is used to describe an average motion state of a specified target vehicle in a plurality of history periods. In the process of implementing step S102, the absolute motion states of the specified target vehicle collected in a plurality of history periods may be used to determine the average kinematic parameters (such as the average speed and the average yaw rate, etc.), so as to effectively suppress the abnormal data caused in the real-time detection process.

Specifically, a time interval of a plurality of history periods is determined, the time interval being a sum of the plurality of history periods, for example: assuming that the number of history periods is N, and the length of each history period is T, the time interval of the N history periods is t=nt; the absolute motion state of the specified target vehicle acquired over a plurality of history periods is divided by the time interval to obtain an average kinematic parameter. For example: dividing the displacement acquired in a plurality of history periods by the time interval to obtain the average speed. Also for example: dividing the angle change acquired in a plurality of history periods by the time interval to obtain the average yaw rate.

After the average kinematic parameters of the specified target vehicle are calculated, the average kinematic parameters are processed by using a kinematic model, and a to-be-processed motion track (also called as a to-be-processed motion track sequence) of the specified target vehicle is predicted, or the to-be-processed motion track of the specified target vehicle is generated.

The kinematic model may be a Constant angular velocity (Constant turn rate and velocity, CTRV) model, a Constant angular velocity (Constant turn RATE AND ACCEL eration, CTRA) model, a Constant velocity model, or the like; specifically, based on the average kinematic parameters of the specified target vehicle, recursion from the kth cycle to the kth+1 cycle is performed using the kinematic model, thereby generating the to-be-processed motion trajectory of the specified target vehicle.

It can be understood that, since the kinematic model can only reflect a motion state for a short time, the time length of the generated motion trajectory to be processed should be less than a certain value (e.g. less than 0.5 s).

Step S103: and calculating a normalized mean value of the motion feature vectors of the specified target vehicle based on the motion trail to be processed and the historical motion trail of the specified target vehicle.

It should be noted that, during the actual running process, the historical motion track of the specified target vehicle in the inertial reference coordinate system in the past period of time is recorded in real time, and the time length of the historical motion track can be customized, for example: the time length of the historical motion trail can be defined as 1 second to 3 seconds according to the general channel change time.

In the specific implementation process of step S103, the normalized mean value of the motion feature vector of the specified target vehicle is calculated using the motion trajectory to be processed and the historical motion trajectory of the specified target vehicle. It is understood that motion feature vectors include, but are not limited to: a lateral position deviation between a vehicle lateral position of the specified target vehicle and a center line lateral position of a lane center line (lane center line of a lane in which the specified target vehicle is located), a heading angle deviation between a vehicle heading angle of the specified target vehicle and a lane line heading angle of the lane center line.

Normalizing the transverse position deviation based on the width of the lane where the appointed target vehicle is located, and calculating the mean value to obtain the normalized mean value of the transverse position deviation; and normalizing the course angle deviation based on the maximum course angle of the appointed target vehicle, and solving the mean value to obtain the normalized mean value of the course angle deviation.

How to calculate the normalized mean value of the lateral position deviation and the normalized mean value of the heading angle deviation is described in detail in fig. 5 of the following embodiment.

Step S104: mapping the normalized mean value to a preset numerical interval to obtain a driving intention numerical value, and predicting according to the driving intention numerical value to obtain an initial driving intention.

In the specific implementation process of step S104, mapping the normalized mean value of the motion characteristic vector onto a preset value interval by using a nonlinear mapping function to obtain a running intention value, and predicting according to the running intention value to obtain an initial running intention; the preset value interval may be a [ -1,1] interval.

In some embodiments, the nonlinear mapping function is determined in advance by using an S-shaped function, specifically, the S-shaped function is scaled and translated to obtain the nonlinear mapping function; the S-type function can be a sigmoid function, a tanh function and the like.

Taking the tanh function as an example, a procedure of how to determine to obtain the nonlinear mapping function is illustrated. the tanh function is as in equation (1). Scaling and translating the tanh function, the resulting nonlinear mapping function is as in equation (2).

Scaling and translating the tanh function, the resulting nonlinear mapping function is as in equation (2).

In the above formula (1) and formula (2), x is an argument, a is a scaling coefficient, and b is a translation coefficient. In practical application, x in the formula (2) is a normalized mean value of the motion feature vector, and f (x) in the formula (2) is a value of the normalized mean value mapped after a preset numerical interval.

In some embodiments, the normalized mean of the lateral position deviation is mapped into a preset value interval to obtain a first driving intention value, and the normalized mean of the heading angle deviation is mapped into the preset value interval to obtain a second driving intention value by using a nonlinear mapping function obtained based on an S-type function.

For example: referring to fig. 3, fig. 3 is a schematic diagram of a nonlinear mapping function, and after mapping the normalized mean value into a preset value interval of [ -1,1], a corresponding driving intention value can be obtained; when the running intention value is in the left lane change value interval (i.e., the value interval indicated by the LLC), the initial running intention is indicated as the left lane change intention; when the running intention value is in the right lane change value interval (namely, the value interval indicated by RLC), the initial running intention is indicated as the right lane change intention; the travel intention value is not in either the left lane change value interval or the right lane change value interval (i.e., in the value interval indicated by LK), indicating that the initial travel intention is to keep the lane. The left lane change value interval in fig. 3 may be-0.6 to-1 and the right lane change value interval may be 0.6 to 1.

After the normalized mean value of the transverse position deviation and the normalized mean value of the course angle deviation are mapped to a preset numerical interval, under the condition that the first running intention numerical value and the second running intention numerical value meet a left lane change rule, predicting the initial running intention as the left lane change intention, wherein the left lane change rule is as follows: the first travel intent value is within a first left lane-change value interval and the second travel intent value is within a second left lane-change value interval.

Under the condition that the first driving intention value and the second driving intention value meet the right lane change rule, predicting the initial driving intention as the right lane change intention, wherein the right lane change rule is as follows: the first travel intent value is within a first right lane-change value interval and the second travel intent value is within a second right lane-change value interval.

In the case where the first travel intention value and the second travel intention value do not satisfy the left lane change rule, and in the case where the first travel intention value and the second travel intention value do not satisfy the right lane change rule, the initial travel intention is predicted to be the lane keeping intention.

Step S105: the surrounding environment of the appointed target vehicle is scanned, the type of the lane line of the lane where the appointed target vehicle is located is obtained, and the obstacle vehicle information in the lane adjacent to the lane where the appointed target vehicle is located is obtained.

In some embodiments, specifying the lane line type of the lane in which the target vehicle is located includes: the lane line type of the left lane line and the lane line type of the right lane line; the track line type is: track lines that allow track changes (e.g., dashed lines) or track lines that do not allow track changes (e.g., solid lines).

In the process of embodying step S105, the surrounding environment of the specified target vehicle is scanned, such as: scanning the type of the lane line of the lane in which the specified target vehicle is located, and scanning obstacle vehicle information in a lane (adjacent left and/or right lanes) adjacent to the lane in which the specified target vehicle is located; obtaining a lane line type of a lane in which the specified target vehicle is located, and obtaining obstacle vehicle information in a lane adjacent to the lane in which the specified target vehicle is located, the obstacle vehicle information indicating at least: whether or not there is an obstacle vehicle in a lane adjacent to the lane in which the specified target vehicle is located.

In some embodiments, the lateral and longitudinal ranges of the surroundings of the specified target vehicle may be defined according to actual conditions; the lateral extent of the surrounding environment is defined as the width of the lane in which the specified target vehicle is located plus the width of the left and right adjacent lanes, for example: the width of the lane where the appointed target vehicle is located is L, and the lanes adjacent to the left and the right of the lane where the appointed target vehicle is located are L, so that the transverse range of the surrounding environment is limited to 3L; the longitudinal extent of the surrounding environment may be adaptively adjusted according to the vehicle speed vx of the given target vehicle, for example: an upper threshold value dis_y_low=k×vx of a longitudinal range of the surrounding environment, a lower threshold value dis_y_high=m×vx of the longitudinal range of the surrounding environment, and the longitudinal range of the surrounding environment is a range formed by dis_y_low and dis_y_high, wherein k and m are adjustment coefficients, and k and m can be adjusted according to a vehicle type and a sensitivity level set by a system.

Step S106: and carrying out filtering processing on the initial driving intention by using the road line type and the obstacle vehicle information, and predicting to obtain the final driving intention.

In the specific implementation step S106, the lane line type of the lane where the specified target vehicle is located is obtained by scanning, and after the obstacle vehicle information is obtained by scanning, the lane line type and the obstacle vehicle information are utilized to determine the lane change condition satisfied by the specified target vehicle, where the lane change condition at least includes: a first condition indicating that a lane change to the right is allowed, and a second condition indicating that a lane change to the left is allowed. The first condition may specifically be: the lane line type of the lane line on the right side of the lane where the specified target vehicle is located is a broken line (i.e. lane line which can be changed), and no obstacle vehicle exists in the lane adjacent to the right of the lane where the specified target vehicle is located; the second condition may specifically be: the lane line type of the lane line on the left side of the lane in which the specified target vehicle is located is a broken line, and there is no obstacle vehicle in the lane adjacent to the left of the lane in which the specified target vehicle is located.

In the case where the initial travel intention is the left lane change intention, when the specified target vehicle does not satisfy the second condition (i.e., cannot change lanes to the left), the final travel intention is predicted to be the lane keeping intention; when the specified target vehicle satisfies the second condition (i.e., can change lane to the left), the final travel intention is predicted as the lane change intention to the left.

When the first condition is not satisfied by the specified target vehicle (i.e., lane change to the right is not possible) in the case where the initial travel intention is the right lane change intention, the final travel intention is predicted to be the lane keeping intention; if the specified target vehicle satisfies the first condition (i.e., lane change to the right), the final travel intention is predicted as the lane change intention to the right.

For example: referring to fig. 4, fig. 4 is a schematic diagram illustrating a surrounding environment of a scan-designated target vehicle, wherein a number 201 and a number 202 represent lane lines, a number 400 represents an obstacle vehicle, and a number 500 represents the designated target vehicle.

As shown in fig. 4, the lane line type of the lane line on the right side of the lane in which the specified target vehicle is located is a solid line, and there is an obstacle vehicle in the lane adjacent to the left of the lane in which the specified target vehicle is located; in the case where the initial travel intention is the left lane change intention, since there is an obstacle vehicle in the lane adjacent to the left of the lane in which the specified target vehicle is located (i.e., lane change to the left is impossible), the final travel intention of the specified target vehicle is predicted to be the lane keeping intention (i.e., in-lane keeping). Meanwhile, as for the scanned surrounding environment mentioned in the above step S105, as can be seen from fig. 4, the longitudinal range of the surrounding environment (the environment scanning area in fig. 4) is a range composed of dis_y_low and dis_y_high.

Step S107: a lateral target position is determined based on the final travel intent.

As can be seen from the above step S106, the final driving intention is a left lane change intention, a right lane change intention, or a lane keeping intention; in the process of concretely implementing step S107, based on the final travel intention, a lateral target position is determined in the lane line Frenet coordinate system, the lateral target position being: the position to be reached by the lateral movement of the target vehicle is specified.

In some embodiments, in the case where the final travel intention is a left lane change intention, determining a center line lateral position of a lane adjacent to the lane where the specified target vehicle is located to the left as a lateral target position; when the final driving intention is the right lane change intention, determining the central line transverse position of a lane adjacent to the right of the lane where the specified target vehicle is located as a transverse target position; in the case where the final travel intention is to keep the lane intention, the center line lateral position of the lane in which the specified target vehicle is located is determined as the lateral target position.

Step S108: and predicting the vehicle motion trail of the appointed target vehicle from the current position to the target position by using the transverse target position and the average kinematics parameter.

In the specific implementation of step S108, a lateral movement track for moving the specified target vehicle to the lateral target position is generated by using a preset track generation method, where the preset track generation method includes, but is not limited to: an equal-proportion attenuation mode, a polynomial curve fitting mode and the like; in the case of proportional damping, the data point at the kth time in the transverse movement track is as in formula (3).

y_k＝αy_k-1 (3)

In the formula (3), y _k is the difference between the lateral position of the specified target vehicle and the lateral target position at the kth time, 0< α <1, and the smaller α is, the faster the y _k decays, that is, the faster the lateral position of the specified target vehicle changes, and α can specifically take a value in the range of 0.90 to 0.99.

It is understood that the acceleration of the specified target vehicle in the longitudinal direction is set to be constant, and a longitudinal movement locus of the specified target vehicle in which the data point at the kth time is as in formula (4) is generated based on the average kinematic parameter determined in step S102 described above.

In the formula (4), x _k is the longitudinal position of the specified target vehicle at the kth time, v is the speed, a is the acceleration, and T is the length of the fixed period. The calculation of v at the kth time is as in equation (5).

v_k＝v_k-1+aT (5)

And (3) through (5) are utilized, the transverse movement track and the longitudinal movement track are combined, the vehicle movement track from the current position to the target position of the appointed target vehicle is predicted, and the track position of the appointed target vehicle in the vehicle movement track at the kth moment is marked as (x _k,y_k).

In combination with the above-mentioned contents of step S101 to step S108, the method for predicting the motion trail of the vehicle according to the embodiment of the present invention predicts the motion trail of the vehicle of the specified target vehicle in two layers. The first layer is a driving intention prediction layer, and the final driving intention of the specified target vehicle is judged by continuously observing the motion state of the specified target vehicle acquired in a history period, calculating the motion characteristic of the specified target vehicle relative to the lane line and considering the dynamic information (whether an obstacle vehicle exists) and the static information (the lane line type of the lane line) of the surrounding environment of the specified target vehicle. The second layer is a vehicle motion trail prediction layer for predicting the vehicle motion trail of the appointed target vehicle in a certain future time period by appointing the final driving intention of the target vehicle.

In the embodiment of the invention, the running intention is obtained by calculating the historical motion trail and other parameters of the appointed target vehicle, and then the motion trail of the vehicle is predicted according to the running intention, so that less parameters are used for predicting the motion trail of the vehicle, and the requirement on hardware calculation force is lower; and the vehicle motion trail is predicted in a layering mode, when the prediction result is abnormal, which link is abnormal can be rapidly positioned, and the abnormality is further solved in time so as to improve the safety.

The process of calculating the normalized mean value of the lateral position deviation and the normalized mean value of the heading angle deviation referred to in step S103 in the above embodiment of the present invention is shown in fig. 5, which is a flowchart for calculating the normalized mean value according to the embodiment of the present invention, and includes the following steps:

Step S501: and combining the motion trail to be processed and a prestored historical motion trail of the appointed target vehicle to obtain an input motion trail.

In the specific implementation process of step S501, the historical motion trail of the specified target vehicle and the predicted motion trail to be processed are combined to obtain the input motion trail.

Step S502: and calculating the central line transverse position and the lane line course angle of the central line of the lane corresponding to the input motion trail by using the lane line expression.

It should be noted that, the lane center line corresponding to the input motion track is the lane center line of the lane where the specified target vehicle is located; the lane line expression includes: a left lane line expression and a right lane line expression.

In some embodiments, the left lane line expression is as in equation (6) and the right lane line expression is as in equation (7).

y_l＝c_0l+c_1lx+c_2lx²+c_3lx³ (6)

y_r＝c_0r+c_1rx+c_2rx²+c_3rx³ (7)

In the formula (6), x is the longitudinal position of the specified target vehicle in the own vehicle coordinate system, y _l is the lateral position of the left lane line of the specified target vehicle in the own vehicle coordinate system, and c _0l、c_1l、c_2l and c _3l are the 3 rd order polynomial coefficients of the left lane line of the specified target vehicle. In the formula (7), x is the longitudinal position of the specified target vehicle in the own vehicle coordinate system, y _r is the lateral position of the right lane line of the specified target vehicle in the own vehicle coordinate system, and c _0r、c_1r、c_2r and c _3r are the 3 rd order polynomial coefficients of the right lane line of the specified target vehicle.

Combining the formula (6) and the formula (7), and designating the central line transverse position of the central line of the lane where the target vehicle is located as y _c as the formula (8); the lane line heading angle θ _c of the lane center line is as in formula (9).

y_c＝(y_l+y_r)/2 (8)

θ_c≈c₁+2c₂x+3c₃x² (9)

In the formula (9) ,c₁＝(c_1l+c_1r)/2,c₂＝(c_2l+c_2r)/2,c₃＝(c_3l+c_3r)/2.

Step S503: for a vehicle lateral position and a vehicle heading angle of a specified target vehicle in an input motion trajectory, calculating a lateral position deviation between the vehicle lateral position and a center line lateral position, and calculating a heading angle deviation between the vehicle heading angle and a lane heading angle.

The input motion track includes a plurality of track points, each track point corresponding to a lateral position of the vehicle and a heading angle of the vehicle where the specified target vehicle exists.

For a vehicle transverse position and a vehicle course angle of a specified target vehicle corresponding to a certain track point, calculating a transverse position deviation dy (namely a transverse position deviation vector) between the vehicle transverse position and a central line transverse position through a formula (10), and calculating a course angle deviation dθ (namely a course angle deviation vector) between the vehicle course angle and a track line course angle through a formula (11).

dy＝y_t-y_c (10)

dθ＝θ_t-θ_c (11)

In formula (10), y _t is the vehicle lateral position corresponding to a certain track point, and y _c is the center line lateral position. In the formula (11), θ _t is a vehicle heading angle corresponding to a certain track point, and θ _c is a track heading angle.

Through the formula (10) and the formula (11), the transverse position deviation between the transverse position of the vehicle corresponding to each track point in the input motion track and the transverse position of the central line can be calculated, and the course angle deviation between the course angle of the vehicle corresponding to each track point in the input motion track and the course line course angle can be calculated; i.e. a plurality of lateral position deviations and a plurality of heading angle deviations are calculated.

Step S504: dividing the transverse position deviation by the lane width to normalize the transverse position deviation, averaging all normalized transverse position deviations, and determining to obtain a normalized mean value of the transverse position deviation.

In the specific implementation process of step S504, the lateral position deviation is normalized according to the lane width of the lane where the specified target vehicle is located, which is obtained by real-time detection, to obtain a normalized value of the lateral position deviation. Specifically, the lateral position deviation is divided by the lane width to normalize the lateral position deviation, and the normalized value of the lateral position deviation is dy_norm=dy/L, dy is the lateral position deviation, and L is the lane width.

From the above, it can be seen that, each track point in the input motion track is calculated to obtain a corresponding lateral position deviation, and after normalizing the lateral position deviation, a normalized value of the lateral position deviation corresponding to each track point can be obtained, so that a normalized value of a plurality of lateral position deviations can be obtained.

And averaging all normalized transverse position deviations, or averaging all normalized transverse position deviations, and determining to obtain the normalized average value of the transverse position deviations.

Step S505: dividing the course angle deviation by the maximum course angle of the appointed target vehicle to normalize the course angle deviation, averaging all normalized course angle deviations, and determining to obtain a normalized mean value of the course angle deviation.

The maximum heading angle of the specified target vehicle is calculated based on the vehicle speed of the specified target vehicle. Specifically, the maximum lateral acceleration of the specified target vehicle is set, and the maximum heading angle is calculated in real time according to the speed of the specified target vehicle.

In the specific implementation process of step S505, the course angle deviation is normalized according to the maximum course angle of the specified target vehicle obtained by real-time calculation, so as to obtain the normalized value of the course angle deviation. Specifically, the heading angle deviation is divided by the maximum heading angle of the specified target vehicle to normalize the heading angle deviation, so as to obtain a normalized value dθ_norm=dθ/θ of the heading angle deviation, dθ is the heading angle deviation, and θ is the maximum heading angle.

From the above, it can be seen that each track point in the input motion track is calculated to obtain a corresponding course angle deviation, and after normalizing the course angle deviation, a normalized value of the course angle deviation corresponding to each track point can be obtained, and then a normalized value of a plurality of course angle deviations can be obtained.

And averaging all the normalized course angle deviations, or averaging all the normalized course angle deviations, and determining to obtain the normalized mean value of the course angle deviations.

In the embodiment of the invention, the motion trail to be processed and the historical motion trail are combined to obtain the input motion trail. And calculating the normalized mean value of the transverse position deviation and the course angle deviation according to the lane line expression and the input motion trail. Determining the driving intention by using the calculated normalized mean value, further predicting the vehicle motion trail based on the driving intention and the information of the surrounding environment, wherein the predicted vehicle motion trail uses fewer parameters and has lower requirement on hardware calculation force; and the vehicle motion trail is predicted in a layering mode, when the prediction result is abnormal, which link is abnormal can be rapidly positioned, and the abnormality is further solved in time so as to improve the safety.

To better explain the contents of the method for predicting the movement track of the vehicle in the above embodiments, a schematic diagram of predicting the movement track of the vehicle is illustrated in fig. 6.

Referring to fig. 6, chassis information of a vehicle (i.e., a host vehicle) is acquired through a vehicle-mounted bus 601, information such as a vehicle speed, an acceleration, a yaw rate, and the like of the host vehicle is acquired, and the acquired information is sent to a lane change intention calculation module 603.

The lane line information is detected by a perception system 602 and sent to a vehicle motion trail generation module 604 and a lane change intention calculation module 603; the sensing system 602 screens out the appointed target vehicle from the detected target vehicles, sends the current environmental information of the appointed target vehicle to the lane change intention calculating module 603, and sends the historical information such as the historical position and the historical speed of the appointed target vehicle to the vehicle motion track generating module 604 and the lane change intention calculating module 603, wherein the historical information such as the historical position and the historical speed is acquired in a historical period.

The lane change intention calculation module 603 calculates the lane change intention (i.e., the final travel intention mentioned above) of the specified target vehicle based on the received information and sends it to the vehicle motion trajectory generation module 604; the vehicle motion trajectory generation module 604 predicts a vehicle motion trajectory of the specified target vehicle based on the received information.

Details of the data processing related to fig. 6 can be seen from the content shown in fig. 1 to 5 in the above embodiment of the present invention, and will not be described herein.

Corresponding to the method for predicting the motion trail of the vehicle provided by the embodiment of the invention, referring to fig. 7, the embodiment of the invention also provides a structural block diagram of a device for predicting the motion trail of the vehicle, the device is mounted on the vehicle, and the device comprises: a first determination unit 701, a processing unit 702, a calculation unit 703, a first prediction unit 704, a scanning unit 705, a second prediction unit 706, a second determination unit 707, and a third prediction unit 708;

A first determination unit 701 for determining a specified target vehicle from among the detected target vehicles.

In a specific implementation, the first determining unit 701 is specifically configured to: from all the detected target vehicles, a target vehicle whose lateral distance from the vehicle is less than a first threshold value and whose longitudinal distance from the vehicle is less than a second threshold value is determined as a specified target vehicle.

The processing unit 702 is configured to determine an average kinematic parameter using the absolute motion states of the specified target vehicle acquired in the plurality of history periods, and predict a motion trajectory to be processed of the specified target vehicle based on the average kinematic parameter.

The average kinematic parameter is used to describe an average motion state of a specified target vehicle in a plurality of history periods.

In a specific implementation, the processing unit 702 is specifically configured to: determining a time interval of a plurality of history periods; dividing the absolute motion state of the appointed target vehicle acquired in a plurality of history periods by the time interval to obtain an average kinematic parameter; and processing the average kinematic parameters by using the kinematic model, and predicting the motion trail to be processed of the appointed target vehicle.

A calculating unit 703 for calculating a normalized mean value of the motion feature vectors of the specified target vehicle based on the motion trajectories to be processed and the historical motion trajectories of the specified target vehicle.

The first prediction unit 704 is configured to map the normalized mean value to a preset value interval to obtain a driving intention value, and predict an initial driving intention according to the driving intention value.

The scanning unit 705 is configured to scan the surrounding environment of the specified target vehicle, obtain the lane type of the lane where the specified target vehicle is located, and obtain the obstacle vehicle information in the lane adjacent to the lane where the specified target vehicle is located.

In some embodiments, specifying the lane line type of the lane in which the target vehicle is located includes: the lane line type of the left lane line and the lane line type of the right lane line; the track line type is: a lane line that allows lane change or a lane line that does not allow lane change.

The second prediction unit 706 is configured to perform filtering processing on the initial travel intention by using the lane type and the obstacle vehicle information, and predict a final travel intention.

In a specific implementation, the initial travel intention is a left lane change intention, a right lane change intention, or a lane keeping intention; the second prediction unit 706 specifically is configured to: and determining lane changing conditions met by the appointed target vehicle by using the lane line type and the obstacle vehicle information, wherein the lane changing conditions at least comprise: a first condition for indicating that a lane change to the right is permitted, and a second condition for indicating that a lane change to the left is permitted;

in the case where the initial travel intention is the left lane change intention, predicting the final travel intention as the lane keeping intention when the specified target vehicle does not satisfy the second condition; when the specified target vehicle meets the second condition, predicting the final driving intention as a left lane change intention;

In the case where the initial travel intention is the right lane change intention, predicting the final travel intention as the lane keeping intention when the specified target vehicle does not satisfy the first condition; when the specified target vehicle satisfies the first condition, the final travel intention is predicted as the right lane change intention.

A second determination unit 707 for determining a lateral target position based on the final travel intention.

In a specific implementation, the final driving intention is a left lane change intention, a right lane change intention or a lane keeping intention; the second determining unit 707 specifically is configured to: when the final driving intention is the left lane change intention, determining the central line transverse position of a lane adjacent to the left of the lane where the specified target vehicle is positioned as a transverse target position; when the final driving intention is the right lane change intention, determining the central line transverse position of a lane adjacent to the right of the lane where the specified target vehicle is located as a transverse target position; in the case where the final travel intention is to keep the lane intention, the center line lateral position of the lane in which the specified target vehicle is located is determined as the lateral target position.

The third prediction unit 708 is configured to generate a vehicle motion trajectory specifying the target vehicle from the current position to the target position, using the lateral target position and the average kinematic parameter.

In a specific implementation, the third prediction unit 708 is specifically configured to: generating a transverse moving track for the appointed target vehicle to move to a transverse target position by using a preset track generation mode; generating a longitudinal movement track of the appointed target vehicle according to the average kinematic parameters; and predicting the vehicle motion trail of the appointed target vehicle from the current position to the target position by combining the transverse motion trail and the longitudinal motion trail.

In some embodiments, the preset track generation mode is an equal-proportion attenuation mode or a polynomial curve fitting mode; in the case where the trajectory generation mode is the equal-proportion attenuation mode, the content of the data points in the generated lateral movement trajectory is referred to the above formula (3); the content of the data points in the generated longitudinal movement track is shown in the above formula (4) and formula (5); and (3) combining the formula (3) to the formula (5), wherein the track position of the target appointed vehicle at the kth moment is (x _k,y_k) in the predicted vehicle motion track.

Preferably, in connection with fig. 7, the computing unit 703 comprises: the execution principle of each subunit is as follows:

and the combining subunit is used for combining the motion trail to be processed and the prestored historical motion trail of the appointed target vehicle to obtain the input motion trail.

The first calculating subunit is used for calculating the central line transverse position and the lane line course angle of the lane central line corresponding to the input motion trail by using the lane line expression;

And the second calculating subunit is used for calculating the transverse position deviation between the transverse position of the vehicle and the transverse position of the central line and calculating the course angle deviation between the course angle of the vehicle and the course line course angle for the designated transverse position of the vehicle and the course angle of the vehicle in the input motion track.

The first processing subunit is used for dividing the transverse position deviation by the lane width to normalize the transverse position deviation, averaging all normalized transverse position deviations and determining to obtain a normalized mean value of the transverse position deviation.

The second processing subunit is used for dividing the course angle deviation by the maximum course angle of the appointed target vehicle to normalize the course angle deviation, averaging all normalized course angle deviations, determining to obtain a normalized mean value of the course angle deviation, and calculating the maximum course angle of the appointed target vehicle based on the speed of the appointed target vehicle.

Accordingly, the first prediction unit 704 is specifically configured to: mapping the normalized mean value of the transverse position deviation into a preset numerical interval to obtain a first driving intention value by using a nonlinear mapping function obtained based on the S-shaped function, and mapping the normalized mean value of the course angle deviation into the preset numerical interval to obtain a second driving intention value;

Under the condition that the first driving intention value and the second driving intention value meet the left lane change rule, predicting the initial driving intention as the left lane change intention, wherein the left lane change rule is as follows: the first driving intention value is in a first left lane change value interval, and the second driving intention value is in a second left lane change value interval;

Under the condition that the first driving intention value and the second driving intention value meet the right lane change rule, predicting the initial driving intention as the right lane change intention, wherein the right lane change rule is as follows: the first driving intention value is in a first right lane change value interval, and the second driving intention value is in a second right lane change value interval;

In summary, the embodiment of the invention provides a method and a device for predicting a vehicle motion track, which calculate a historical motion track and other parameters of a specified target vehicle to obtain a driving intention, predict the vehicle motion track according to the driving intention, and have fewer parameters used for predicting the vehicle motion track and lower requirements on hardware calculation force; and the vehicle motion trail is predicted in a layering mode, when the prediction result is abnormal, which link is abnormal can be rapidly positioned, and the abnormality is further solved in time so as to improve the safety.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for a system or system embodiment, since it is substantially similar to a method embodiment, the description is relatively simple, with reference to the description of the method embodiment being made in part. The systems and system embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of functionality in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of predicting a vehicle motion trajectory, the method being applied to a device for predicting a vehicle motion trajectory mounted in a vehicle, the method comprising:

Determining a specified target vehicle from the detected target vehicles;

Determining and obtaining an average kinematic parameter by utilizing the absolute motion states of the specified target vehicle acquired in a plurality of history periods, and predicting a to-be-processed motion track of the specified target vehicle based on the average kinematic parameter, wherein the average kinematic parameter is used for describing the average motion state of the specified target vehicle in the plurality of history periods; the motion trail to be processed is a motion trail recursively obtained based on a kinematic model in a short time length;

determining lane changing conditions which are satisfied by the appointed target vehicle for predicting the initial driving intention by utilizing the lane type and the obstacle vehicle information, and predicting to obtain a final driving intention according to the lane changing conditions;

determining a lateral target position based on the final travel intent;

2. The method of claim 1, wherein determining an average kinematic parameter using the absolute motion states of the specified target vehicle acquired in a plurality of history periods, and predicting a trajectory of the specified target vehicle to be processed based on the average kinematic parameter, comprises:

determining a time interval of a plurality of history periods;

And processing the average kinematic parameters by using the kinematic model, and predicting the motion trail to be processed of the appointed target vehicle.

3. The method of claim 1, wherein the calculating a normalized mean of motion feature vectors of the specified target vehicle based on the pending motion profile and the historical motion profile of the specified target vehicle comprises:

combining the motion trail to be processed and the historical motion trail of the appointed target vehicle to obtain an input motion trail;

4. The method according to claim 3, wherein mapping the normalized mean value into a preset value interval to obtain a driving intention value, and predicting from the driving intention value to obtain an initial driving intention, comprises:

5. The method of claim 1, wherein the specifying a lane line type of a lane in which the target vehicle is located comprises: the lane line type of the left lane line and the lane line type of the right lane line; the track line type is as follows: a lane line that allows lane change or a lane line that does not allow lane change.

6. The method of claim 1, wherein the initial travel intent is a left lane change intent, a right lane change intent, or a lane keeping intent;

Determining a lane change condition satisfied by the specified target vehicle for predicting the initial driving intention by using the lane type and the obstacle vehicle information, and predicting to obtain a final driving intention according to the lane change condition, wherein the lane change condition comprises:

7. The method of claim 1, wherein the final travel intent is a left lane change intent, a right lane change intent, or a lane keeping intent;

Determining a central line transverse position of a lane adjacent to the lane where the specified target vehicle is located to the left as a transverse target position when the final driving intention is a left lane change intention;

determining a central line transverse position of a lane adjacent to the right of the lane where the specified target vehicle is located as a transverse target position when the final driving intention is a right lane change intention;

and determining the transverse position of the central line of the lane where the specified target vehicle is located as a transverse target position under the condition that the final driving intention is lane keeping intention.

8. The method according to claim 1, wherein the preset trajectory generation mode is an equal-proportion decay mode or a polynomial curve fitting mode;

And in the predicted vehicle motion trail, the trail position of the appointed target vehicle at the kth moment is (x _k,y_k).

9. An apparatus for predicting a motion trajectory of a vehicle, the apparatus being mounted in the vehicle, the apparatus comprising:

a first determination unit configured to determine a specified target vehicle from among the detected target vehicles;

The processing unit is used for determining and obtaining average kinematic parameters by utilizing the absolute motion states of the specified target vehicle acquired in a plurality of history periods, and predicting a to-be-processed motion track of the specified target vehicle based on the average kinematic parameters, wherein the average kinematic parameters are used for describing the average motion states of the specified target vehicle in the plurality of history periods; the motion trail to be processed is a motion trail recursively obtained based on a kinematic model in a short time length;

A second prediction unit, configured to determine a lane change condition satisfied by the specified target vehicle for which the initial travel intention is predicted, using the lane type and the obstacle vehicle information, and predict a final travel intention according to the lane change condition;

A third prediction unit, configured to generate a lateral movement track of the specified target vehicle moving to the lateral target position by using a preset track generation manner; generating a longitudinal movement track of the appointed target vehicle according to the average kinematic parameter; and predicting the vehicle motion trail of the appointed target vehicle from the current position to the target position by combining the transverse motion trail and the longitudinal motion trail.