Lane change prediction method and device for target vehicle

Abstract

The present invention relates to a lane change prediction method for a target vehicle, the method comprising: receiving a speed and a position of the target vehicle; obtaining a first lane change probability and a second lane change probability of the target vehicle using a first machine learning model and a second machine learning model based on the speed and the position of the target vehicle, respectively; and determining a lane change probability of the target vehicle based on the first lane change probability and the second lane change probability, wherein the first machine learning model and the second machine learning model are pre-trained and different from each other. The present invention also relates to a lane change prediction device for the target vehicle, a computer storage medium, and a vehicle.

Description

Technical Field

[0001] The present invention relates to the field of lane change prediction for a target vehicle, and more specifically, to a lane change prediction method and device for a target vehicle, a computer storage medium, and a vehicle. Background Art

[0002] The dynamics of other vehicles on the road have a crucial impact on road safety. For example, in advanced driver assistance systems, the driving strategies of connected vehicles (ICVs) must consider the movements of other vehicles, primarily their lane changes and lane keeping. Lane changes pose a significant threat to road safety. Therefore, predicting the lane-changing intentions of vehicles on either side of the vehicle can effectively avoid traffic accidents and improve driving safety.

[0003] Currently, existing vehicle lane change intention prediction schemes generally use a single prediction module to predict lane change intention, without considering the potential misjudgment caused by a single model. Therefore, an improved lane change prediction scheme for the target vehicle is desired. Summary of the Invention

[0004] According to one aspect of the present invention, a lane change prediction method for a target vehicle is provided, the method comprising: receiving a speed and a position of the target vehicle; obtaining a first lane change probability and a second lane change probability of the target vehicle respectively based on the speed and the position of the target vehicle using a first machine learning model and a second machine learning model; and determining a lane change possibility of the target vehicle based on the first lane change probability and the second lane change probability, wherein the first machine learning model and the second machine learning model are pre-trained and different from each other.

[0005] As a supplement or replacement for the above solution, in the above method, the speed and position of the target vehicle include: the longitudinal distance d of the target vehicle in the vehicle coordinate system of the host vehicle x ; The lateral distance d of the target vehicle in the lane coordinate system y_ccs ; and the lateral velocity v of the target vehicle in the lane coordinate system y_ccs and longitudinal velocity v x_ccs .

[0006] As a supplement or replacement for the above solution, in the above method, the first machine learning model is at least based on the lateral distance d of the target vehicle in the lane line coordinate system. y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs Train the feature parameters.

[0007] As a supplement or replacement for the above solution, in the above method, the second machine learning model is at least based on the longitudinal distance d of the target vehicle in the vehicle coordinate system of the vehicle. x , the lateral distance d of the target vehicle in the lane coordinate system y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs Train the feature parameters.

[0008] As a supplement or alternative to the above solution, in the above method, the first machine learning model uses a logistic regression method and adopts the following model to calculate the first lane change probability P1:

[0009]

[0010] Among them, fabs(x) is the function of taking the absolute value, flag dir Used to define the direction, cnt dy_ccs represents the number of frames in which the vehicle continuously turns left or right, and c1 to c5 are coefficients obtained after training the first machine learning model.

[0011] As a supplement or alternative to the above solution, in the above method, the second machine learning model uses a support vector machine method and adopts the following model to calculate the second lane change probability P2:

[0012] P2=e1*d x +e2*d y_ccs +e3*V x_ccs +e4*v y_ccs ,

[0013] Among them, e1 to e4 are coefficients obtained after training of the second machine learning model.

[0014] As a supplement or alternative to the above-mentioned scheme, in the above-mentioned method, determining the lane change possibility of the target vehicle based on the first lane change probability and the second lane change probability includes: when the first lane change probability and the second lane change probability both meet the threshold conditions, the target lane change prediction is considered successful and the lane change probability P of the target vehicle is output.

[0015] According to another aspect of the present invention, a lane change prediction device for a target vehicle is provided, the device comprising: a receiving device for receiving the speed and position of the target vehicle; a calculating device for obtaining a first lane change probability and a second lane change probability of the target vehicle respectively based on the speed and position of the target vehicle using a first machine learning model and a second machine learning model; and a determining device for determining the lane change possibility of the target vehicle based on the first lane change probability and the second lane change probability, wherein the first machine learning model and the second machine learning model are pre-trained and different from each other.

[0016] As a supplement or replacement for the above solution, in the above device, the speed and position of the target vehicle include: the longitudinal distance d of the target vehicle in the vehicle coordinate system of the host vehicle x ; The lateral distance d of the target vehicle in the lane coordinate system y_ccs ; and the lateral velocity v of the target vehicle in the lane coordinate system y_ccs and longitudinal velocity v x_ccs .

[0017] As a supplement or replacement for the above solution, the above device also includes: a first training device for training the first machine learning model, wherein the first machine learning model is at least based on the lateral distance d of the target vehicle in the lane line coordinate system. y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs is the characteristic parameter.

[0018] As a supplement or replacement for the above solution, the above device also includes: a second training device for training the second machine learning model, wherein the second machine learning model is at least based on the longitudinal distance d of the target vehicle in the vehicle coordinate system of the vehicle x , the lateral distance d of the target vehicle in the lane coordinate system y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs is the characteristic parameter.

[0019] As a supplement or alternative to the above solution, in the above device, the computing device is configured to calculate the first lane-changing probability P1 by using the following first machine learning model:

[0020]

[0021] Among them, fabs(x) is the function of taking the absolute value, flag dir Used to define the direction, cnt dy_ccsrepresents the number of frames in which the vehicle continuously turns left or right, and c1 to c5 are coefficients obtained after training the first machine learning model.

[0022] As a supplement or alternative to the above solution, in the above device, the computing device is configured to calculate the second lane-changing probability P2 using the following second machine learning model:

[0023] P2=e1*d x +e2*d y_ccs +e3*v x_ccs +e4*V y_ccs ,

[0024] Among them, e1 to e4 are coefficients obtained after training of the second machine learning model.

[0025] As a supplement or alternative to the above solution, in the above device, the determining device is configured to: output the lane changing probability P of the target vehicle when both the first lane changing probability and the second lane changing probability meet a threshold condition.

[0026] According to another aspect of the present invention, a computer storage medium is provided, wherein the medium includes instructions, and the instructions execute the above method when executed.

[0027] According to yet another aspect of the present invention, a vehicle is provided, comprising the device as described above.

[0028] The lane change prediction scheme for the target vehicle in an embodiment of the present invention combines two different models (i.e., a first machine learning model and a second machine learning model) to comprehensively judge the lane change intention of the target vehicle, ensuring the accuracy and stability of the prediction system, thereby improving driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The above and other objects and advantages of the present invention will become more fully apparent from the following detailed description taken in conjunction with the accompanying drawings, wherein the same or similar elements are denoted by the same reference numerals.

[0030] Figure 1 A schematic flow chart of a lane change prediction method for a target vehicle according to one embodiment of the present invention is shown;

[0031] Figure 2 A schematic structural diagram of a lane change prediction device for a target vehicle according to an embodiment of the present invention is shown; and

[0032] Figure 3 A schematic diagram of a scenario for predicting the lane-changing probability of a target vehicle based on the motion state of the target vehicle according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the present invention more clear, the following is a further detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention.

[0034] It should also be noted that, for ease of description, only the part relevant to the present invention, rather than all of the content, is shown in the accompanying drawings. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processing or methods depicted as flow charts. Although flow charts describe various operations (or steps) as sequential processing, many operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of various operations can be rearranged. When its operation is completed, the processing can be terminated, but can also have additional steps not included in the accompanying drawings. The processing can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0035] Although the exemplary embodiments are described as using multiple units to perform exemplary processes, it should be understood that the exemplary processes may also be performed by one or more modules.

[0036] Furthermore, the control logic of the present invention may be embodied as executable program instructions on a computer-readable medium, which is executed by a processor or the like. Examples of computer-readable media include, but are not limited to, ROM, RAM, optical disks, magnetic tapes, floppy disks, flash drives, smart cards, and optical data storage devices. The computer-readable recording medium may also be distributed among computer systems connected to a network, such that the computer-readable medium is stored and executed in a distributed manner, for example, via a vehicle telematics service or a controller area network (CAN).

[0037] It should be understood that the term "vehicle" or other similar terms used herein includes general motor vehicles, such as passenger cars (including sports utility vehicles, buses, trucks, etc.), various commercial vehicles, etc., and includes hybrid vehicles, electric vehicles, etc. A hybrid vehicle is a vehicle that has two or more power sources, such as gasoline-powered and electric vehicles.

[0038] Hereinafter, a lane change prediction scheme for a target vehicle according to various exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0039] Figure 1 FIG. 1 shows a flow chart of a lane change prediction method 1000 for a target vehicle according to an embodiment of the present invention. Figure 1 As shown, the lane change prediction method 1000 for a target vehicle includes the following steps:

[0040] In step S110, the speed and position of the target vehicle are received;

[0041] In step S120, based on the speed and position of the target vehicle, a first lane-changing probability and a second lane-changing probability of the target vehicle are obtained using a first machine learning model and a second machine learning model respectively; and

[0042] In step S130, the lane change possibility of the target vehicle is determined based on the first lane change probability and the second lane change probability, wherein the first machine learning model and the second machine learning model are pre-trained and different from each other.

[0043] In the context of the present invention, the term "target vehicle" refers to another vehicle relative to the host vehicle. In one or more embodiments, the target vehicles for which lane change prediction is required are located on the left and right sides in front of the host vehicle.

[0044] Machine learning refers to the process of using algorithms to guide computers to derive appropriate models from known data and then use these models to make decisions about new situations. Therefore, in the context of this invention, the "first machine learning model" and "second machine learning model" refer to different models derived from computers using different algorithms, which can be used to determine (or predict) the lane change probability of a target vehicle.

[0045] Using a single model to predict the target vehicle's lane change can result in significant misjudgment. When the target vehicle is misjudged as changing lanes, the driver's vehicle can accidentally brake on the target vehicle, leading to a rear-end collision or other traffic accident in severe cases. By using both the first and second machine learning models to predict the target vehicle's lane change, the accuracy of the target vehicle's lane change prediction is improved.

[0046] In one embodiment, in the above method 1000, the speed and position of the target vehicle include: the longitudinal distance d of the target vehicle in the vehicle coordinate system of the host vehicle; x ; The lateral distance d of the target vehicle in the lane coordinate system y_ccs ; and the lateral velocity v of the target vehicle in the lane coordinate system y_ccs and longitudinal velocity v x_ccs .

[0047] by Figure 3 For example, it shows a schematic diagram of a scenario 3000 for predicting the lane change probability of a target vehicle based on its motion state. In the scenario 3000, the host vehicle is shown as 310 and the target vehicle is shown as 320. The vehicle coordinate system of the host vehicle is, for example, based on the center of the front bumper of the vehicle as the origin. Figure 3The lane line coordinate system can be a Frenet coordinate system in one embodiment, which uses the lane centerline as a reference line and uses variables s and d to describe the position of the vehicle on the road. The s coordinate represents the distance along the road (also called longitudinal displacement or longitudinal distance) and the d coordinate represents the left and right position on the road (also called lateral displacement or lateral distance). Figure 3 In the example, the projection point of the vehicle 320 on the reference line (i.e., the lane centerline 330) is C. Therefore, the distance between point C and the lateral position B of the target vehicle 320 in the vehicle coordinate system is the lateral distance d of the target vehicle in the lane coordinate system. y_ccs .

[0048] exist Figure 3 In FIG, point A represents the position of the lane centerline at the longitudinal position of the target vehicle 320. Therefore, in one embodiment, the lateral distance d of the target vehicle in the lane line coordinate system is y_ccs =AB*cosθ, where cosθ can be calculated based on the slope of the lane centerline.

[0049] In one embodiment, the first machine learning model is based on at least the lateral distance d of the target vehicle in the lane line coordinate system. y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs For example, the first machine learning model uses a logistic regression method and adopts the following model to calculate the first lane change probability P1:

[0050]

[0051] Among them, fabs(x) is the function of taking the absolute value, flag dir Used to define the direction (for example, left is -1, right is 1), cnt dy_ccs Indicates the number of frames in which the vehicle continuously turns left or right (the number of frames here refers to the number of system operation frames, for example, 50ms / time). When predicting whether the target vehicle turns left, if the target vehicle's d y_ccs If cnt continues to decrease, cnt continues to increase. c1 to c5 are the coefficients obtained after training the first machine learning model. In one embodiment, a big data learning method is used to train a model parameter c1 to c5 based on the speed and position of the target vehicle. Then, when the speed and position of the target vehicle are input, a probability value of the target vehicle changing lanes can be obtained.

[0052] In one embodiment, the second machine learning model is based on at least the longitudinal distance d of the target vehicle in the vehicle coordinate system of the host vehicle. x , the lateral distance d of the target vehicle in the lane coordinate system y_ccs, the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs In one example, the second machine learning model uses a support vector machine method and adopts the following model to calculate the second lane change probability P2:

[0053] P2=e1*d x +e2*d y_ccs +e3*v x_ccs +e4*v y_ccs ,

[0054] Among them, e1 to e4 are coefficients obtained after training of the second machine learning model.

[0055] In one embodiment, step S130 includes: when both the first lane change probability and the second lane change probability meet threshold conditions, the target lane change prediction is deemed successful and the lane change probability P of the target vehicle is output. In an alternative embodiment, step S130 includes: when the first lane change probability P1 meets a first threshold and the second lane change probability P2 meets a second threshold, outputting a final lane change probability P = (P1 + P2) / 2. In yet another alternative embodiment, the final lane change probability P = w1*P1 + w2*P2, where w1 and w2 are weights that can be adjusted as needed.

[0056] Furthermore, those skilled in the art will readily appreciate that the lane change prediction method for a target vehicle provided by one or more of the above-described embodiments of the present invention can be implemented via a computer program. For example, when a computer storage medium (e.g., a USB flash drive) storing the computer program is connected to a computer, the computer program can be run to execute the lane change prediction method for a target vehicle according to one or more embodiments of the present invention.

[0057] refer to Figure 2 , Figure 2 FIG. 2 shows a schematic structural diagram of a lane change prediction device 2000 for a target vehicle according to an embodiment of the present invention. Figure 2 As shown, a lane change prediction device 2000 for a target vehicle includes: a receiving device 210, a computing device 220, and a determining device 230. The receiving device 210 is configured to receive the speed and position of the target vehicle; the computing device 220 is configured to obtain a first lane change probability and a second lane change probability of the target vehicle using a first machine learning model and a second machine learning model based on the speed and position of the target vehicle; and the determining device 230 is configured to determine the lane change probability of the target vehicle based on the first lane change probability and the second lane change probability, wherein the first machine learning model and the second machine learning model are pre-trained and different from each other.

[0058] In the context of the present invention, the term "target vehicle" refers to another vehicle relative to the host vehicle. In one or more embodiments, the target vehicles for which lane change prediction is required are located on the left and right sides in front of the host vehicle.

[0059] Machine learning refers to the process of using algorithms to guide computers to derive appropriate models from known data and then use these models to make decisions about new situations. Therefore, in the context of this invention, the "first machine learning model" and "second machine learning model" refer to different models derived from computers using different algorithms, which can be used to determine (or predict) the lane change probability of a target vehicle.

[0060] Because using a single model to predict a target vehicle's lane change can result in significant misjudgment, a misjudgment of the target vehicle's lane change status can lead to the host vehicle's inadvertent braking of the target vehicle, which can, in severe cases, lead to traffic accidents such as rear-end collisions. The target vehicle lane change prediction device 2000 uses both a first machine learning model and a second machine learning model to predict the target vehicle's lane change, thereby improving the accuracy of the target vehicle's lane change prediction.

[0061] In one embodiment, the speed and position of the target vehicle include: the longitudinal distance d of the target vehicle in the vehicle coordinate system of the host vehicle x ; The lateral distance d of the target vehicle in the lane coordinate system y_ccs ; and the lateral velocity v of the target vehicle in the lane coordinate system y_ccs and longitudinal velocity v x_ccs .

[0062] refer to Figure 3 , which shows a schematic diagram of a scenario 3000 for predicting the lane change probability of a target vehicle based on its motion state. In the scenario 3000, the host vehicle is shown as 310 and the target vehicle is shown as 320. The vehicle coordinate system of the host vehicle is, for example, based on the center of the front bumper of the vehicle as the origin. Figure 3 The lane line coordinate system can be a Frenet coordinate system in one embodiment, which uses the lane centerline as a reference line and uses variables s and d to describe the position of the vehicle on the road. The s coordinate represents the distance along the road (also called longitudinal displacement or longitudinal distance) and the d coordinate represents the left and right position on the road (also called lateral displacement or lateral distance). Figure 3 In the example, the projection point of the vehicle 320 on the reference line (i.e., the lane centerline 330) is C. Therefore, the distance between point C and the lateral position B of the target vehicle 320 in the vehicle coordinate system is the lateral distance d of the target vehicle in the lane coordinate system. y_ccs .

[0063] exist Figure 3In FIG, point A represents the position of the lane centerline at the longitudinal position of the target vehicle 320. Therefore, in one embodiment, the lateral distance d of the target vehicle in the lane line coordinate system is y_ccs =AB*cosθ, where cosθ can be calculated based on the slope of the lane centerline.

[0064] although Figure 2 Not shown, in one embodiment, the device 2000 may further include: a first training device for training the first machine learning model, wherein the first machine learning model is at least based on the lateral distance d of the target vehicle in the lane line coordinate system. y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs In one embodiment, the device 2000 may further include: a second training device for training the second machine learning model, wherein the second machine learning model is based on at least the longitudinal distance d of the target vehicle in the vehicle coordinate system of the vehicle. x , the lateral distance d of the target vehicle in the lane coordinate system y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs is the characteristic parameter.

[0065] Specifically, in one embodiment, the computing device 220 is configured to calculate the first lane-changing probability P1 using the following first machine learning model:

[0066]

[0067] Among them, fabs(x) is the function of taking the absolute value, flag dir Used to define the direction (for example, left is -1, right is 1), cnt dy_ccs Indicates the number of frames in which the vehicle continuously turns left or right (the number of frames here refers to the number of system operation frames, for example, 50ms / time). When predicting whether the target vehicle turns left, if the target vehicle's d y_ccs If cnt continues to decrease, cnt continues to increase. c1 to c5 are the coefficients obtained after training the first machine learning model. In one embodiment, a big data learning method is used to train a model parameter c1 to c5 based on the speed and position of the target vehicle. Then, when the speed and position of the target vehicle are input, a probability value of the target vehicle changing lanes can be obtained.

[0068] In one embodiment, the computing device 220 is configured to calculate the second lane-changing probability P2 using the following second machine learning model:

[0069] P2=e1*d x+e2*d y_ccs +e3*V x_ccs +e4*v y_ccs ,

[0070] Among them, e1 to e4 are coefficients obtained after training of the second machine learning model.

[0071] In one embodiment, the determining device 230 is configured to output the lane change probability P of the target vehicle when both the first lane change probability and the second lane change probability meet threshold conditions. In an alternative embodiment, the determining device 230 is configured to output the final lane change probability P = (P1 + P2) / 2 when the first lane change probability P1 meets a first threshold and the second lane change probability P2 meets a second threshold. In yet another alternative embodiment, the determining device 230 is configured to calculate the final lane change probability P according to the following formula: P = w1*P1+w2*P2, where w1 and w2 are weights (which can be adjusted as needed).

[0072] In addition, the lane change prediction device 2000 for a target vehicle according to one or more embodiments of the present invention may be included in an advanced driving assistance system (ADAS). In one embodiment, the advanced driving assistance system (ADAS) may be installed in a vehicle.

[0073] Advanced Driver Assistance Systems, also known as ADAS, utilize various sensors installed on the vehicle (e.g., millimeter-wave radar, lidar, monocular and binocular cameras, and satellite navigation) to sense the surrounding environment at all times while the car is in motion, collect data, and identify, detect, and track static and dynamic objects. Combined with navigation map data, these systems perform systematic calculations and analysis, allowing the driver to proactively detect potential dangers and effectively improving driving comfort and safety. In one embodiment, the advanced driver assistance system includes a navigation and real-time traffic system TMC, an electronic police system ISA (Intelligent speed adaptation or intelligent speed advice), a vehicle network (Vehicular communication systems), an adaptive cruise control ACC), a lane departure warning system LDWS (Lane Departure Warning System), a lane keeping system (Lane keep assistance), a collision avoidance system or pre-crash system (Collision Avoidance System or Precrash System), a night vision system (Night Vision System), an adaptive light control (Adaptive light control), a pedestrian protection system (Pedestrian protection system), an automatic parking system (Automatic parking), traffic sign recognition (Traffic sign recognition), blind spot detection (Blind spot detection), driver drowsiness detection (Driver drowsiness detection), a hill descent control system (Hill descent control) and an electric vehicle warning sounds system.

[0074] In summary, the lane change prediction scheme for the target vehicle in an embodiment of the present invention combines two different models (i.e., the first machine learning model and the second machine learning model) to comprehensively judge the lane changing intention of the target vehicle, ensuring the accuracy and stability of the prediction system, thereby improving driving safety.

[0075] Although the above description only describes some embodiments of the present invention, it should be understood by those skilled in the art that the present invention may be implemented in many other forms without departing from its spirit and scope. Therefore, the examples and embodiments shown are to be considered illustrative rather than restrictive, and the present invention may encompass various modifications and substitutions without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A lane change prediction method for a target vehicle, characterized in that: The method comprises: receiving the speed and position of the target vehicle; Based on the speed and position of the target vehicle, using a first machine learning model and a second machine learning model to respectively obtain a first lane change probability and a second lane change probability of the target vehicle; and Determine the lane change possibility of the target vehicle based on the first lane change probability and the second lane change probability, wherein the first machine learning model and the second machine learning model are pre-trained and different from each other, wherein the first machine learning model is based on at least the lateral distance d of the target vehicle in the lane line coordinate system. y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs The second machine learning model is trained with at least the longitudinal distance d of the target vehicle in the vehicle coordinate system of the vehicle x , the lateral distance d of the target vehicle in the lane coordinate system y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs Train feature parameters; The first machine learning model uses a logistic regression method and adopts the following model to calculate the first lane change probability P1: Among them, fabs(x) is the function of taking the absolute value, flag dir Used to define the direction, cnt dy_ccs represents the number of frames in which the vehicle continuously turns left or right, and c1 to c5 are coefficients obtained after training the first machine learning model; The second machine learning model uses a support vector machine method and adopts the following model to calculate the second lane change probability P2: P2=e1*d x +e2*d y_ccs +e3*v x_ccs +e4*v y_ccs , Among them, e1 to e4 are coefficients obtained after training of the second machine learning model.

2. The method according to claim 1, wherein The speed and position of the target vehicle include: The longitudinal distance d of the target vehicle in the vehicle coordinate system of the vehicle x ; The lateral distance d of the target vehicle in the lane line coordinate system y_ccs ;as well as The lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs .

3. The method according to claim 1, wherein Determining the lane change possibility of the target vehicle according to the first lane change probability and the second lane change probability includes: When both the first lane-changing probability and the second lane-changing probability meet the threshold conditions, the target lane-changing prediction is deemed successful and the lane-changing probability P of the target vehicle is output.

4. A lane change prediction device for a target vehicle, characterized in that: The device comprises: a receiving device for receiving the speed and position of the target vehicle; a computing device for obtaining, based on the speed and position of the target vehicle, a first lane-changing probability and a second lane-changing probability of the target vehicle using a first machine learning model and a second machine learning model, respectively; and a determining device for determining a lane change possibility of the target vehicle based on the first lane change probability and the second lane change probability, wherein the first machine learning model and the second machine learning model are pre-trained and different from each other; Also included is: a first training device for training the first machine learning model, wherein the first machine learning model is at least based on the lateral distance d of the target vehicle in the lane line coordinate system. y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs is a characteristic parameter; a second training device for training the second machine learning model, wherein the second machine learning model is at least based on the longitudinal distance d of the target vehicle in the vehicle coordinate system of the vehicle x , the lateral distance d of the target vehicle in the lane coordinate system y_ccs , the lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs is the characteristic parameter; The computing device is configured to calculate the first lane-changing probability P1 using the following first machine learning model: Among them, fabs(x) is the function of taking the absolute value, flag dir Used to define the direction, cnt dy_ccs represents the number of frames in which the vehicle continuously turns left or right, and c1 to c5 are coefficients obtained after training the first machine learning model; The computing device is configured to calculate the second lane-changing probability P2 using the following second machine learning model: P2=e1*d x +e2*d y_ccs +e3*v x_ccs +e4*v y_ccs , Among them, e1 to e4 are coefficients obtained after training of the second machine learning model.

5. The apparatus of claim 4, wherein: The speed and position of the target vehicle include: The longitudinal distance d of the target vehicle in the vehicle coordinate system of the vehicle x ; The lateral distance d of the target vehicle in the lane line coordinate system y_ccs ;as well as The lateral velocity v of the target vehicle in the lane line coordinate system y_ccs and longitudinal velocity v x_ccs .

6. The apparatus of claim 4, wherein: The determining device is configured to output a lane-changing probability P of the target vehicle when both the first lane-changing probability and the second lane-changing probability meet a threshold condition.

7. A computer storage medium, characterized in that The medium includes instructions that, when executed, perform the method of any one of claims 1 to 3.

8. A vehicle comprising the apparatus of any one of claims 4 to 6.

Images