Vehicle lane changing intention recognition method and device, storage medium and electronic equipment

Abstract

The invention discloses a vehicle lane changing intention recognition method and device, a storage medium and electronic equipment, and relates to the technical field of automatic driving automobiles, and the method comprises the steps: obtaining a motion parameter sequence and historical trajectory data of vehicles around a main vehicle, the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes in a preset time window, and the motion parameters at least comprise a transverse speed and a transverse acceleration; inputting the motion parameter sequence into a preset lane changing intention recognition model for lane changing intention recognition to obtain lane changing intentions corresponding to the surrounding vehicles; according to the historical trajectory data, determining a driving mode when the surrounding vehicles change lanes or keep lanes; and according to the lane changing intention and the driving mode, predicting a driving track of the surrounding vehicle. The method can improve the recognition precision of the lane changing intention of the vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous driving vehicles, and in particular to a method, device, storage medium and electronic device for identifying a vehicle's lane-changing intention. Background Art

[0002] In recent years, improper driver operation has become one of the main causes of road accidents around the world, with about 50% of traffic accidents caused by improper operation. It is reported that about 9% of road traffic accidents are caused by lane changing, and about 35% to 50% of road traffic fatalities are directly related to lane changing and overtaking. The research on lane changing intention recognition of autonomous vehicles has always attracted the attention of researchers. By identifying the lane changing intention of autonomous vehicles, the risk of lane changing can be predicted, which can help drivers make lane changing decisions and ensure driving safety.

[0003] At present, when identifying the lane-changing intention of a vehicle, the focus is usually on the lateral displacement of the vehicle. However, the calculation of the lateral displacement of the vehicle has a lag. When the vehicle is driving at high speed, the lateral displacement is difficult to accurately reflect the vehicle's lane-changing intention in advance, resulting in low recognition accuracy of the vehicle's lane-changing intention. Summary of the invention

[0004] The present invention provides a method, device, storage medium and electronic equipment for recognizing a vehicle's lane-changing intention, which are mainly capable of improving the recognition accuracy of the vehicle's lane-changing intention.

[0005] According to a first aspect of an embodiment of the present invention, a method for identifying a vehicle lane change intention is provided, comprising:

[0006] Acquire motion parameter sequences and historical trajectory data of vehicles around the host vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters include at least lateral velocity and lateral acceleration;

[0007] Inputting the motion parameter sequence into a preset lane change intention recognition model to perform lane change intention recognition, and obtaining the lane change intention corresponding to the surrounding vehicles;

[0008] Determining, based on the historical trajectory data, a driving mode of the surrounding vehicles when changing lanes or keeping lanes;

[0009] The driving trajectory of the surrounding vehicles is predicted according to the lane changing intention and the driving mode.

[0010] Optionally, when the preset lane-changing intention recognition model is a Gaussian mixture hidden Markov lane-changing intention recognition model, inputting the motion parameter sequence into the preset lane-changing intention recognition model to perform lane-changing intention recognition to obtain the lane-changing intention corresponding to the surrounding vehicles includes:

[0011] The motion parameter sequence is input into the Gaussian mixture hidden Markov lane change intention recognition model to perform lane change intention recognition, and the lane change intention corresponding to the surrounding vehicles is obtained, wherein the lane change intention includes one of left lane change, right lane change and lane keeping.

[0012] Optionally, determining the driving mode of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data includes:

[0013] Calculating the acceleration of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data;

[0014] The driving mode of the surrounding vehicle is determined according to the acceleration of the surrounding vehicle when changing lanes or keeping lanes, wherein the driving mode includes one of acceleration, deceleration and constant speed.

[0015] Optionally, predicting the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode includes:

[0016] The lane-changing intention and the driving mode are input into a preset trajectory prediction model to perform trajectory prediction, so as to obtain the driving trajectory of the surrounding vehicles within a preset time period in the future.

[0017] Optionally, when the preset trajectory prediction model is a Gaussian mixture hidden Markov trajectory prediction model, the lane change intention and the driving mode are input into the preset trajectory prediction model for trajectory prediction to obtain the driving trajectory of the surrounding vehicles, including:

[0018] The lane change intention and the driving mode are input into the Gaussian mixture hidden Markov trajectory prediction model for trajectory prediction to obtain the driving trajectory of the surrounding vehicles in a future preset time period.

[0019] Optionally, the method comprises:

[0020] Obtaining actual driving trajectories of the surrounding vehicles within the preset time period;

[0021] Determining uncertainty of the predicted driving trajectory according to the predicted driving trajectory and the actual driving trajectory;

[0022] Based on the uncertainty, the prediction accuracy of the preset trajectory prediction model is evaluated.

[0023] Optionally, determining the uncertainty of the predicted driving trajectory according to the predicted driving trajectory and the actual driving trajectory includes:

[0024] Calculating a trajectory covariance matrix according to the predicted driving trajectory and the actual driving trajectory;

[0025] The uncertainty of the predicted driving trajectory is determined according to the trajectory covariance matrix.

[0026] According to a second aspect of an embodiment of the present invention, a vehicle lane change intention recognition device is provided, comprising:

[0027] An acquisition unit, used to acquire a motion parameter sequence and historical trajectory data of vehicles around the host vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters at least include lateral velocity and lateral acceleration;

[0028] an identification unit, configured to input the motion parameter sequence into a preset lane-changing intention identification model to perform lane-changing intention identification, and obtain the lane-changing intention corresponding to the surrounding vehicles;

[0029] a determination unit, configured to determine a driving mode of the surrounding vehicle when changing lanes or keeping lanes according to the historical trajectory data;

[0030] A prediction unit is used to predict the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode.

[0031] According to a third aspect of an embodiment of the present invention, there is provided a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0032] Acquire motion parameter sequences and historical trajectory data of vehicles around the host vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters include at least lateral velocity and lateral acceleration;

[0033] Inputting the motion parameter sequence into a preset lane change intention recognition model to perform lane change intention recognition, and obtaining the lane change intention corresponding to the surrounding vehicles;

[0034] Determining, based on the historical trajectory data, a driving mode of the surrounding vehicles when changing lanes or keeping lanes;

[0035] The driving trajectory of the surrounding vehicles is predicted according to the lane changing intention and the driving mode.

[0036] According to a fourth aspect of an embodiment of the present invention, there is provided an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the following steps are implemented:

[0037] Acquire motion parameter sequences and historical trajectory data of vehicles around the host vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters include at least lateral velocity and lateral acceleration;

[0038] Inputting the motion parameter sequence into a preset lane change intention recognition model to perform lane change intention recognition, and obtaining the lane change intention corresponding to the surrounding vehicles;

[0039] Determining, based on the historical trajectory data, a driving mode of the surrounding vehicles when changing lanes or keeping lanes;

[0040] The driving trajectory of the surrounding vehicles is predicted according to the lane changing intention and the driving mode.

[0041] The present invention provides a vehicle lane change intention recognition method, device, storage medium and electronic device. Compared with the prior art, the present invention can obtain the motion parameter sequence and historical trajectory data of the vehicles around the main vehicle, wherein the motion parameter sequence is composed of the motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters at least include lateral velocity and lateral acceleration. Then, the motion parameter sequence is input into a preset lane change intention recognition model for lane change intention recognition to obtain the lane change intention corresponding to the surrounding vehicles, and then the driving mode of the surrounding vehicles when changing lanes or keeping lanes is determined according to the historical trajectory data. Finally, according to the lane change intention and the driving mode, the driving trajectory of the surrounding vehicles is predicted. It can be seen that when the present invention recognizes the lane change intention of the surrounding vehicles, by introducing lateral velocity and lateral acceleration, the lane change intention of the surrounding vehicles can be accurately identified in the early stage of lane change, thereby improving the recognition accuracy of the lane change intention. In addition, since the present invention takes into account the driving mode of the surrounding vehicles when performing trajectory prediction, the accuracy of the trajectory prediction results can be fully guaranteed when facing complex traffic environments and multi-interaction scenarios.

[0042] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 A schematic flow chart of a method for identifying a vehicle lane-changing intention provided by an embodiment of the present invention is shown;

[0045] Figure 2 A schematic diagram of a vehicle lane changing scenario provided by an embodiment of the present invention is shown;

[0046] Figure 3 A schematic diagram of the vehicle lane change intention recognition and trajectory prediction process of the Gaussian mixture hidden Markov model provided in an embodiment of the present invention is shown;

[0047] Figure 4 A schematic diagram of model training provided by an embodiment of the present invention is shown;

[0048] Figure 5 A schematic structural diagram of a vehicle lane change intention recognition device provided by an embodiment of the present invention is shown;

[0049] Figure 6 A schematic diagram of the physical structure of an electronic device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present invention and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0052] The calculation of the vehicle's lateral displacement has a lag. When the vehicle is driving at high speed, it is difficult for the lateral displacement to accurately reflect the vehicle's lane-changing intention in advance, resulting in a low recognition accuracy of the vehicle's lane-changing intention.

[0053] In order to overcome the above defects, an embodiment of the present invention provides a method for identifying a vehicle lane change intention, such as Figure 1 As shown, the method includes:

[0054] Step 101: Obtain motion parameter sequences and historical trajectory data of vehicles around the host vehicle.

[0055] The motion parameter sequence is composed of the motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters at least include lateral velocity and lateral acceleration. In addition, the historical trajectory data includes the longitudinal displacement and lateral displacement of the surrounding vehicles during actual driving.

[0056] The embodiments of the present invention are mainly applicable to the scenarios of identifying lane change intentions and predicting lane change trajectories of surrounding vehicles, such as Figure 2 The execution subject of the embodiment of the present invention is a device or apparatus capable of identifying lane-changing intentions and predicting lane-changing trajectories of surrounding vehicles.

[0057] For the embodiment of the present invention, vehicles traveling on the road within a preset range from the main vehicle can be determined as surrounding vehicles. For example, vehicles within 5 meters from the main vehicle can be determined as surrounding vehicles. In order to identify the lane-changing intention of the vehicles around the main vehicle, it is necessary to obtain the motion parameter sequence of the surrounding vehicles within a preset time window through the vehicle sensor, wherein the size of the preset time window can be set according to actual business needs. In order to improve the recognition accuracy of lane-changing intention, the embodiment of the present invention selects the best preset time window size suitable for the preset lane-changing intention recognition model through experiments. The preset time window includes multiple time steps. For example, each time step is represented as 0.1s. The number of time steps with the best model performance is selected as 10 from 5-60 time steps. It can be seen that the motion parameter sequence of the surrounding vehicles is composed of motion parameters corresponding to 0.1s, 0.2s, ..., 1s.

[0058] Furthermore, considering that the lateral motion characteristics of the vehicle can better reflect the vehicle's lane-changing intention, the embodiment of the present invention divides the motion parameters into three categories. The first category of motion parameters includes: lateral velocity and lateral acceleration; the second category of motion parameters includes: lateral velocity, lateral acceleration and heading angle; the third category of motion parameters includes: lateral velocity, lateral acceleration, longitudinal velocity, longitudinal acceleration and heading angle. It should be noted that the above three categories of motion parameters are all within the protection scope of the embodiment of the present invention. After experimental analysis, the lateral velocity, lateral acceleration and heading angle can be finally selected as a preferred motion parameter.

[0059] Furthermore, while obtaining the motion parameter sequence of the surrounding vehicles, it is also necessary to obtain the historical trajectory data of the surrounding vehicles, that is, the longitudinal displacement and lateral displacement of the surrounding vehicles, so as to predict the trajectory of the surrounding vehicles based on the historical trajectory data.

[0060] It should be noted that the number of vehicles around the main vehicle can be one, two or more. When the number of surrounding vehicles is at least two, it is necessary to obtain the motion parameter sequence and historical trajectory data of each surrounding vehicle, and perform lane change intention recognition and lane change trajectory prediction for each surrounding vehicle.

[0061] Step 102: input the motion parameter sequence into a preset lane change intention recognition model to perform lane change intention recognition, and obtain the lane change intention corresponding to the surrounding vehicles.

[0062] The lane change intention of the surrounding vehicles includes one of left lane change, right lane change and lane keeping. The preset lane change intention recognition model can be a Gaussian mixture hidden Markov lane change intention recognition model, or other models that can recognize time series data.

[0063] For the embodiment of the present invention, the hidden Markov model (HMM) has the characteristic of describing the potential rules of time series data, and the driver's decision-making and driving operation is a dynamic change process of a time series. Therefore, compared with models such as support vector machine (SVM) and neural network, the hidden Markov model (HMM) is more suitable for the driver's lane change intention recognition. At the same time, because the Gaussian mixture model can fit the characteristics of multi-dimensional continuous data to the greatest extent, the embodiment of the present invention selects the Gaussian mixture hidden Markov model as the preset lane change intention recognition model.

[0064] Regarding the process of using a Gaussian mixture hidden Markov model to identify lane changing intentions of surrounding vehicles, step 102 specifically includes: inputting the motion parameter sequence into the Gaussian mixture hidden Markov lane changing intention recognition model to perform lane changing intention recognition, and obtaining the lane changing intentions corresponding to the surrounding vehicles, wherein the lane changing intentions include one of left lane changing, right lane changing and lane keeping.

[0065] Furthermore, before using the Gaussian mixture hidden Markov lane change intention recognition model to recognize the lane change intention of surrounding vehicles, the Gaussian mixture hidden Markov lane change intention recognition model needs to be trained. The embodiment of the present invention mainly considers three aspects when training the Gaussian mixture hidden Markov lane change intention recognition model, namely, the selection of motion parameters, the preset time window size (the number of time steps included), and the Gaussian mixture number.

[0066] Regarding the selection of motion parameters, as mentioned above, the embodiment of the present invention divides the motion parameters into three categories. The above three types of feature parameters are all suitable for the Gaussian mixture hidden Markov lane change intention recognition model. In order to improve the performance of the model, as a preferred method, the embodiment of the present invention selects the second type of motion parameters, namely lateral velocity, lateral acceleration and heading angle. Regarding the preset time window size, assuming that each time step represents 0.1s, the number of time steps with the best model performance is selected from 5-60 time steps through experiments. For example, the number of time steps finally selected is 10, that is, the preset time window contains 10 time steps. Regarding the Gaussian mixture number, considering that the model distribution function has a significant impact on the model performance, the embodiment of the present invention has verified through experiments that the Gaussian mixture number size with the best model performance is selected from Gaussian mixture numbers 1-6. For example, the selected Gaussian mixture number is 5. The training and verification process of the model in the embodiment of the present invention is as follows Figure 3 As shown in , when the training results do not converge, the embodiment of the present invention can increase the number of iterations of model training. The Gaussian mixture hidden Markov lane change intention recognition model established in the embodiment of the present invention has the characteristic of fast convergence speed, which also ensures the precision and accuracy of lane change intention recognition when the vehicle is driving at high speed. Figure 4 As shown, the embodiment of the present invention extracts vehicle data from the NGSIM data set, and uses 70% of the vehicle data as a training set and 30% as a test set.

[0067] Step 103: Determine the driving mode of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data.

[0068] The driving mode includes one of acceleration, deceleration and constant speed.

[0069] For the embodiment of the present invention, in order to predict the trajectory of the surrounding vehicles, it is necessary not only to identify the lane-changing intention of the surrounding vehicles, but also to determine the driving mode of the surrounding vehicles during the lane-changing or lane-keeping process. For this process, step 103 specifically includes: calculating the acceleration of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data; determining the driving mode of the surrounding vehicles according to the acceleration of the surrounding vehicles when changing lanes or keeping lanes, wherein the driving mode includes one of acceleration, deceleration and constant speed.

[0070] Specifically, the historical trajectory data of surrounding vehicles can be used to solve the acceleration during the lane change process, determine the driving mode of the surrounding vehicles when changing lanes, and achieve the above process through high-frequency prediction and low-frequency output.

[0071] Step 104: predicting the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode.

[0072] In the embodiment of the present invention, in order to predict the driving trajectory of the surrounding vehicles, step 104 specifically includes: inputting the lane change intention and the driving mode into a preset trajectory prediction model to perform trajectory prediction, and obtain the driving trajectory of the surrounding vehicles in a future preset time period.

[0073] The preset trajectory prediction model may be a Gaussian mixture hidden Markov trajectory prediction model, or other models capable of trajectory prediction. In addition, the preset time period may be set according to actual business needs, for example, to predict the future t f The x and y values ​​of the lane-changing trajectories of surrounding vehicles within seconds.

[0074] Specifically, the lane change intention and driving mode are input into the Gaussian mixture hidden Markov trajectory prediction model for trajectory prediction, and the driving trajectories of surrounding vehicles in a preset time period in the future are obtained.

[0075] When performing trajectory prediction, we first determine whether the surrounding vehicles are changing lanes left, right, or keeping lane based on the lane changing intention. Then, we determine the behavior patterns of the surrounding vehicles when changing lanes or keeping lanes based on the driving modes of the surrounding vehicles. Finally, based on the recognition results of the lane changing intention and driving mode, we complete the trajectory prediction through the Gaussian mixture hidden Markov trajectory prediction model.

[0076] Furthermore, when the trajectory prediction is actually performed, the trajectory prediction accuracy of the model can also be evaluated. For this process, the method includes: obtaining the actual driving trajectory of the surrounding vehicles within the preset time period; determining the uncertainty of the predicted driving trajectory based on the predicted driving trajectory and the actual driving trajectory; and evaluating the prediction accuracy of the preset trajectory prediction model based on the uncertainty. Further, determining the uncertainty of the predicted driving trajectory based on the predicted driving trajectory and the actual driving trajectory includes: calculating the trajectory covariance matrix based on the predicted driving trajectory and the actual driving trajectory; and determining the uncertainty of the predicted driving trajectory based on the trajectory covariance matrix.

[0077] For example, the predicted driving trajectories of the surrounding vehicles are (1,2), (3,6), (4,2), (5,2), and X represents the displacement in the x direction and Y represents the displacement in the y direction. We can get X = (1,3,4,5) T , Y=(2,6,2,2) T Since there are only two columns, X and Y, the covariance matrix obtained is a 2*2 matrix. Then, each element of the covariance matrix is ​​calculated separately.

[0078] ∑ 11 =(1-3.25, 3-3.25, 4-3.25, 5-5.25)

[0079] ×(1-3.25, 3-3.25, 4-3.25, 5-5.25) T =8.75

[0080] ∑ 12 =(1-3.25, 3-3.25, 4-3.25, 5-5.25)×(2-3, 6-3, 2-3, 2-3) T =-1

[0081] ∑ 21 =(2-3, 6-3, 2-3, 2-3)×(1-3.25, 3-3.25, 4-3.25, 5-5.25) T =-1

[0082] ∑ 22 =(2-3, 6-3, 2-3, 2-3)×(2-3, 6-3, 2-3, 2-3) T =12

[0083] The final calculated covariance matrix can be expressed as:

[0084]

[0085] In the embodiment of the present invention, the correlation between the predicted driving trajectory and the actual driving trajectory can be represented by a covariance matrix. The larger the covariance value in the matrix, the closer and more consistent the relationship between the two variables. Therefore, the prediction accuracy of the model in actual use can be evaluated by calculating the covariance matrix between the predicted driving trajectory and the actual driving trajectory.

[0086] Furthermore, before actually performing trajectory prediction, the Gaussian mixture hidden Markov trajectory prediction model needs to be trained. Before completing the training and putting it into actual use, the model can also be verified by calculating the covariance matrix. For example, the predicted trajectory 1.5s, 1s and 0s before passing the lane line is compared with the actual trajectory to verify the prediction accuracy of the model. If the prediction accuracy of the model does not meet the requirements, the model continues to be iteratively trained. The Gaussian mixture hidden Markov trajectory prediction model trained in the embodiment of the present invention can predict the trajectory of lane changes of surrounding vehicles at a very sensitive speed, and the predicted vehicle trajectory is highly consistent with the actual trajectory. At the same time, based on the high accuracy of the lane change recognition results and trajectory prediction, the embodiment of the present invention can warn the driver and passengers, and guide the vehicle to plan and control, to ensure the safety of vehicle driving and ride comfort.

[0087] The embodiment of the present invention provides a method for identifying vehicle lane change intentions. When identifying lane change intentions of surrounding vehicles, by introducing lateral velocity and lateral acceleration, the lane change intentions of surrounding vehicles can be accurately identified in the early stages of lane change, thereby improving the recognition accuracy of lane change intentions. In addition, since the embodiment of the present invention takes into account the driving modes of surrounding vehicles when performing trajectory prediction, the accuracy of trajectory prediction results can be fully guaranteed in the face of complex traffic environments and multi-interaction scenarios.

[0088] Further, as Figure 1 In a specific implementation, an embodiment of the present invention provides a vehicle lane change intention recognition device, such as Figure 5 As shown, the device includes: an acquisition unit 21, an identification unit 22, a determination unit 23 and a prediction unit 24.

[0089] The acquisition unit 21 can be used to acquire the motion parameter sequence and historical trajectory data of the vehicles around the main vehicle, wherein the motion parameter sequence is composed of the motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters include at least lateral velocity and lateral acceleration.

[0090] The recognition unit 22 may be used to input the motion parameter sequence into a preset lane-changing intention recognition model to perform lane-changing intention recognition, and obtain the lane-changing intention corresponding to the surrounding vehicles.

[0091] The determination unit 23 may be configured to determine the driving mode of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data.

[0092] The prediction unit 24 may be configured to predict the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode.

[0093] In a specific application scenario, when the preset lane change intention recognition model is a Gaussian mixture hidden Markov lane change intention recognition model, the recognition unit 22 can be specifically used to input the motion parameter sequence into the Gaussian mixture hidden Markov lane change intention recognition model to perform lane change intention recognition, and obtain the lane change intention corresponding to the surrounding vehicles, wherein the lane change intention includes one of left lane change, right lane change and lane keeping.

[0094] In a specific application scenario, the determination unit 23 can be specifically used to calculate the acceleration of the surrounding vehicles when changing lanes or keeping lanes based on the historical trajectory data; and determine the driving mode of the surrounding vehicles based on the acceleration of the surrounding vehicles when changing lanes or keeping lanes, wherein the driving mode includes one of acceleration, deceleration and constant speed.

[0095] In a specific application scenario, the prediction unit 24 can be specifically used to input the lane change intention and the driving mode into a preset trajectory prediction model to perform trajectory prediction, so as to obtain the driving trajectory of the surrounding vehicles within a preset time period in the future.

[0096] Furthermore, when the preset trajectory prediction model is a Gaussian mixture hidden Markov trajectory prediction model, the prediction unit 24 can also be specifically used to input the lane change intention and the driving mode into the Gaussian mixture hidden Markov trajectory prediction model for trajectory prediction, so as to obtain the driving trajectory of the surrounding vehicles in a preset time period in the future.

[0097] In a specific application scenario, the device further includes: an evaluation unit.

[0098] The acquisition unit 21 may also be used to acquire the actual driving tracks of the surrounding vehicles within the preset time period.

[0099] The determination unit 23 may also be configured to determine the uncertainty of the predicted driving trajectory according to the predicted driving trajectory and the actual driving trajectory.

[0100] The evaluation unit may be used to evaluate the prediction accuracy of the preset trajectory prediction model.

[0101] In a specific application scenario, the determination unit 23 may also be specifically used to calculate a trajectory covariance matrix according to the predicted driving trajectory and the actual driving trajectory; and determine the uncertainty of the predicted driving trajectory according to the trajectory covariance matrix.

[0102] It should be noted that for other corresponding descriptions of the functional modules involved in the vehicle lane change intention recognition device provided in the embodiment of the present invention, reference can be made to Figure 1 The corresponding description of the method shown will not be repeated here.

[0103] Based on the above Figure 1 The method shown, accordingly, an embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the following steps are implemented: obtaining a motion parameter sequence and historical trajectory data of vehicles surrounding the main vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters include at least lateral velocity and lateral acceleration; inputting the motion parameter sequence into a preset lane change intention recognition model to perform lane change intention recognition, and obtaining the lane change intention corresponding to the surrounding vehicles; determining the driving mode of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data; and predicting the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode.

[0104] Based on the above Figure 1 The method shown and Figure 5 The embodiment of the device shown in the figure, the embodiment of the present invention also provides a physical structure diagram of an electronic device, such as Figure 6 As shown, the electronic device includes: a processor 31, a memory 32, and a computer program stored in the memory 32 and executable on the processor, wherein the memory 32 and the processor 31 are both arranged on a bus 33, and the processor 31 implements the following steps when executing the program: obtaining a motion parameter sequence and historical trajectory data of vehicles surrounding the main vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters include at least lateral velocity and lateral acceleration; inputting the motion parameter sequence into a preset lane change intention recognition model for lane change intention recognition, and obtaining the lane change intention corresponding to the surrounding vehicles; determining the driving mode of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data; and predicting the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode.

[0105] When identifying the lane-changing intention of surrounding vehicles, the embodiment of the present invention can accurately identify the lane-changing intention of surrounding vehicles in the early stage of lane changing by introducing lateral velocity and lateral acceleration, thereby improving the recognition accuracy of lane-changing intention. In addition, since the embodiment of the present invention takes the driving mode of surrounding vehicles into consideration when predicting trajectories, it can fully guarantee the accuracy of trajectory prediction results in complex traffic environments and multi-interaction scenarios.

[0106] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0107] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying a vehicle's lane-changing intention, characterized in that: include: Acquire motion parameter sequences and historical trajectory data of vehicles around the host vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters include at least lateral velocity and lateral acceleration; Inputting the motion parameter sequence into a preset lane-changing intention recognition model to perform lane-changing intention recognition, and obtaining the lane-changing intention corresponding to the surrounding vehicles; Determining, based on the historical trajectory data, a driving mode of the surrounding vehicles when changing lanes or keeping lanes; The driving trajectory of the surrounding vehicles is predicted according to the lane changing intention and the driving mode.

2. The method according to claim 1, characterized in that When the preset lane-changing intention recognition model is a Gaussian mixture hidden Markov lane-changing intention recognition model, inputting the motion parameter sequence into the preset lane-changing intention recognition model to perform lane-changing intention recognition to obtain the lane-changing intention corresponding to the surrounding vehicles includes: The motion parameter sequence is input into the Gaussian mixture hidden Markov lane change intention recognition model to perform lane change intention recognition, and the lane change intention corresponding to the surrounding vehicles is obtained, wherein the lane change intention includes one of left lane change, right lane change and lane keeping.

3. The method according to claim 1, characterized in that Determining the driving mode of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data includes: Calculating the acceleration of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data; The driving mode of the surrounding vehicle is determined according to the acceleration of the surrounding vehicle when changing lanes or keeping lanes, wherein the driving mode includes one of acceleration, deceleration and constant speed.

4. The method according to claim 1, characterized in that: Predicting the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode, including: The lane-changing intention and the driving mode are input into a preset trajectory prediction model to perform trajectory prediction, so as to obtain the driving trajectory of the surrounding vehicles within a preset time period in the future.

5. The method according to claim 4, characterized in that When the preset trajectory prediction model is a Gaussian mixture hidden Markov trajectory prediction model, the lane change intention and the driving mode are input into the preset trajectory prediction model for trajectory prediction to obtain the driving trajectory of the surrounding vehicles, including: The lane change intention and the driving mode are input into the Gaussian mixture hidden Markov trajectory prediction model for trajectory prediction to obtain the driving trajectory of the surrounding vehicles in a future preset time period.

6. The method according to claim 4 or 5, characterized in that: The method comprises: Obtaining actual driving trajectories of the surrounding vehicles within the preset time period; Determining uncertainty of the predicted driving trajectory according to the predicted driving trajectory and the actual driving trajectory; Based on the uncertainty, the prediction accuracy of the preset trajectory prediction model is evaluated.

7. The method according to claim 6, characterized in that Determining uncertainty of the predicted driving trajectory according to the predicted driving trajectory and the actual driving trajectory includes: Calculating a trajectory covariance matrix according to the predicted driving trajectory and the actual driving trajectory; The uncertainty of the predicted driving trajectory is determined according to the trajectory covariance matrix.

8. A vehicle lane change intention recognition device, characterized in that: include: An acquisition unit, used to acquire a motion parameter sequence and historical trajectory data of vehicles around the host vehicle, wherein the motion parameter sequence is composed of motion parameters of the surrounding vehicles at different time nodes within a preset time window, and the motion parameters at least include lateral velocity and lateral acceleration; an identification unit, configured to input the motion parameter sequence into a preset lane-changing intention identification model to perform lane-changing intention identification, and obtain the lane-changing intention corresponding to the surrounding vehicles; a determination unit, configured to determine a driving mode of the surrounding vehicles when changing lanes or keeping lanes according to the historical trajectory data; A prediction unit is used to predict the driving trajectory of the surrounding vehicles according to the lane change intention and the driving mode.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images