Vehicle lane changing behavior prediction method and device and storage medium

Abstract

The invention relates to a vehicle lane changing behavior prediction method and device and a storage medium. The method comprises the steps of determining at least one target vehicle around a self-vehicle in the driving process of the self-vehicle; for each target vehicle, driving state change characteristics and surrounding environment characteristics of the target vehicle are extracted; and generating a lane changing behavior prediction result of the target vehicle according to the driving state change characteristics and the surrounding environment characteristics. By extracting the driving state change characteristics and the surrounding environment characteristics, the influence factors related to the lane changing behavior of the vehicle can be comprehensively and detailedly obtained, and then the lane changing behavior can be accurately predicted; moreover, the driving state change characteristics and the surrounding environment characteristics can be rapidly collected in real time, so that the lane changing behavior prediction result can be rapidly obtained, the vehicle can conveniently take measures corresponding to the lane changing behaviors of the surrounding vehicles in time, and the safety of automatic driving is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of autonomous driving technology, and in particular, to a method, apparatus, and storage medium for predicting vehicle lane-changing behavior. Background Art

[0002] In related technologies, methods for predicting the lane-changing behavior of surrounding vehicles usually adopt rule-based algorithms, that is, engineers pre-define rules related to lane-changing behavior, and determine the lane-changing intention of the vehicle according to whether the actual behavior of the vehicle meets these rules. However, due to the variability of actual driving scenarios, the above pre-defined rules are difficult to cover the lane-changing behavior of vehicles in detail and comprehensively, resulting in poor accuracy and real-time performance of vehicle lane-changing behavior prediction, so that the autonomous driving system cannot accurately and timely take measures corresponding to the lane-changing behavior, affecting the safety of autonomous driving vehicles. Summary of the Invention

[0003] To overcome the problems existing in related technologies, the present disclosure provides a method, apparatus, and storage medium for predicting vehicle lane-changing behavior.

[0004] According to a first aspect of an embodiment of the present disclosure, there is provided a method for predicting vehicle lane-changing behavior, including:

[0005] During the driving process of the host vehicle, determine at least one target vehicle around the host vehicle;

[0006] For each of the target vehicles, extract the driving state change feature and surrounding environment feature of the target vehicle; according to the driving state change feature and the surrounding environment feature, generate a lane-changing behavior prediction result of the target vehicle.

[0007] Optionally, the extracting the driving state change feature and surrounding environment feature of the target vehicle includes:

[0008] Obtain the historical trajectory of the target vehicle within a preset time period before the current moment and the position information of the target vehicle at the current moment;

[0009] Generate the driving state change feature according to the historical trajectory;

[0010] Generate the surrounding environment feature according to the historical trajectory and the position information.

[0011] Optionally, the generating the driving state change feature according to the historical trajectory includes:

[0012] Determine the historical motion state parameters of the target vehicle according to the historical trajectory;

[0013] Calculate the driving state change feature according to the historical motion state parameters.

[0014] Optionally, generating the surrounding environment features according to the historical trajectory and the position information includes:

[0015] Obtaining the current motion state parameters of the surrounding vehicles of the target vehicle and the lane line information of the lane where the target vehicle is located;

[0016] For each of the surrounding vehicles, determining the relative driving information between the target vehicle and the surrounding vehicle according to the position information and the current motion state parameters of the surrounding vehicle;

[0017] Determining the relative position information between the target vehicle and the lane where it is located according to the position information and the lane line information;

[0018] Determining the displacement information of the target vehicle in the lane where it is located according to the historical trajectory, the position information and the lane line information;

[0019] Wherein, the surrounding environment features include the relative position information, the displacement information, and the relative driving information between the target vehicle and each of the surrounding vehicles.

[0020] Optionally, generating the lane change behavior prediction result of the target vehicle according to the driving state change feature and the surrounding environment feature includes:

[0021] Inputting the driving state change feature and the surrounding environment feature into a pre-trained lane change behavior prediction model to generate the lane change behavior prediction result of the target vehicle.

[0022] Optionally, the lane change behavior prediction model is trained in the following manner:

[0023] Obtaining a training sample set; wherein each training sample in the training sample set includes the driving state change feature and the surrounding environment feature of the sample vehicle, and the lane change behavior annotation category corresponding to the sample vehicle;

[0024] Using the training sample set for model training to obtain the lane change behavior prediction model.

[0025] Optionally, the lane change behavior prediction result includes multiple lane change behavior categories and the prediction probability corresponding to each lane change behavior category.

[0026] According to the second aspect of the embodiments of the present disclosure, there is provided a vehicle lane change behavior prediction device, including:

[0027] A target vehicle determination module, configured to determine at least one target vehicle around the own vehicle during the driving process of the own vehicle;

[0028] An extraction module, configured to extract, for each of the target vehicles, the driving state change characteristics and the surrounding environment characteristics of the target vehicle; a prediction module, configured to generate a lane change behavior prediction result of the target vehicle according to the driving state change characteristics and the surrounding environment characteristics.

[0029] According to a third aspect of the embodiments of the present disclosure, there is provided a vehicle lane change behavior prediction device, including:

[0030] A processor;

[0031] A memory for storing executable instructions of the processor;

[0032] Wherein, the processor is configured to:

[0033] During the driving process of the host vehicle, determine at least one target vehicle around the host vehicle;

[0034] For each of the target vehicles, extract the driving state change characteristics and the surrounding environment characteristics of the target vehicle; and generate a lane change behavior prediction result of the target vehicle according to the driving state change characteristics and the surrounding environment characteristics.

[0035] According to a fourth aspect of the embodiments of the present disclosure, there is provided a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the vehicle lane change behavior prediction method provided in the first aspect of the present disclosure are implemented.

[0036] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects: During the driving process of the host vehicle, first determine the target vehicles that need to perform lane change behavior prediction, and then, for each target vehicle, extract the driving state change characteristics used to characterize the driving state change of the target vehicle itself, and the surrounding environment characteristics used to characterize the relationship between the target vehicle and its surrounding vehicles and between the target vehicle and its lane. Finally, based on the driving state change characteristics and the surrounding environment characteristics, obtain the lane change behavior prediction result of the target vehicle. Since the lane change behavior of the target vehicle is related to both its own driving state and the mutual relationship between the target vehicle and its surrounding vehicles and between the target vehicle and its lane, by extracting the driving state change characteristics and the surrounding environment characteristics, it is possible to comprehensively and exhaustively obtain the influencing factors related to the vehicle lane change behavior, and thus accurately predict the lane change behavior; moreover, the above driving state change characteristics and surrounding environment characteristics can be collected quickly and in real time, so that the lane change behavior prediction result can be obtained quickly, which is convenient for the host vehicle to take corresponding measures in a timely manner with respect to the lane change behavior of the surrounding vehicles, and improves the safety of autonomous driving.

[0037] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. Description of the Drawings

[0038] The drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with the present disclosure and, together with the specification, are used to explain the principles of the present disclosure.

[0039] Figure 1 A schematic diagram of a vehicle driving trajectory provided by an embodiment of the present disclosure;

[0040] Figure 2 A flowchart of a method for predicting a vehicle lane change behavior provided by an embodiment of the present disclosure;

[0041] Figure 3 A schematic diagram of area division provided by an embodiment of the present disclosure;

[0042] Figure 4 A block diagram of a device for predicting a vehicle lane change behavior provided by an embodiment of the present disclosure;

[0043] Figure 5 A block diagram of a device for predicting a vehicle lane change behavior according to an embodiment of the present disclosure. Detailed Description of the Embodiments

[0044] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0045] It should be noted that all actions of obtaining signals, information, or data in this application are carried out on the premise of complying with the corresponding data protection regulations and policies of the place where it is located and with the authorization given by the owner of the corresponding device.

[0046] In an autonomous driving system, the ego vehicle (also referred to as the host vehicle) is the vehicle on which the autonomous driving software runs from the first perspective; the target vehicle is the vehicle around the ego vehicle for which a lane change behavior prediction needs to be made. The lane change behavior of a vehicle includes cut in and cut out. Among them, cut in refers to the behavior of the target vehicle squeezing into the driving lane in front of the ego vehicle from the left and right lanes of the ego vehicle, and cut out refers to the behavior of the target vehicle in front of the ego vehicle changing lanes to the left and right sides from the lane where the ego vehicle is located.

[0047] Figure 1 It is a schematic diagram of a vehicle driving trajectory, as Figure 1 shown, the target vehicle is driving on lane 1, and its driving trajectory is L1 The host vehicle is traveling in lane 2, and its driving trajectory is L 0 L 2 is a possible future driving trajectory of the target vehicle. This trajectory crosses the lane boundary of lane 1 and enters lane 2 where the host vehicle is located. When entering this lane, the target vehicle is in front of the host vehicle. It can be seen that this trajectory indicates that the target vehicle has performed a lane-changing cut-in behavior. If the host vehicle can predict in advance that the target vehicle is about to perform a lane-changing cut-in, it can accurately and timely take countermeasures, such as appropriately decelerating and avoiding, to avoid colliding with the target vehicle and ensure the safety of the host vehicle.

[0048] In related technologies, vehicle lane-changing behavior prediction usually adopts a rule-based algorithm, that is, engineers pre-define rules related to lane-changing behavior and determine the lane-changing intention of the vehicle according to whether the actual behavior of the vehicle meets these rules. This rule-based lane-changing behavior prediction method has good interpretability. However, the pre-defined related rules are limited and cannot comprehensively consider the changes in actual road scenarios and driving scenarios, thus affecting the real-time performance and accuracy of lane-changing behavior prediction.

[0049] In addition, a deep learning model can also be used to predict vehicle lane-changing behavior. This method has a high accuracy rate, but it requires high Graphics Processing Unit (GPU) resources. However, the computing power of the processor in the vehicle is limited. If the deep learning model is deployed in the vehicle, it will increase the overall cost of the vehicle. Moreover, the inference speed of the deep learning model is slow and it is difficult to perform fast real-time prediction.

[0050] Therefore, the present disclosure provides a vehicle lane-changing behavior prediction method, device and storage medium. The following will describe the specific embodiments of the present disclosure in detail with reference to the accompanying drawings.

[0051] Figure 2 is a flowchart of a vehicle lane-changing behavior prediction method shown according to an exemplary embodiment. As Figure 2 shown, the vehicle lane-changing behavior prediction method is applied to the host vehicle and includes the following steps.

[0052] S101: During the driving process of the host vehicle, determine at least one target vehicle around the host vehicle.

[0053] In one embodiment of the present disclosure, all target vehicles around the host vehicle are detected according to sensors provided on the host vehicle, such as cameras, radars, ultrasonic sensors, etc. Herein, the target vehicle refers to a vehicle within a preset distance range of the host vehicle. For example, a vehicle in the same lane as the host vehicle and 10 meters ahead of the host vehicle; a vehicle in the left lane of the host vehicle and 15 meters ahead of the host vehicle, etc. The lane-changing behavior of the target vehicle will affect the driving state of the host vehicle. Therefore, it is necessary to predict its vehicle lane-changing behavior.

[0054] Specifically, the object detection algorithm can be used to analyze the objects detected by the sensors, identify the vehicles within the above-mentioned preset distance range of the host vehicle, and determine them as target vehicles.

[0055] S102: For each target vehicle, extract the driving state change features and surrounding environment features of the target vehicle.

[0056] In one embodiment of the present disclosure, the driving state change features are used to characterize the change situation of different driving states of the target vehicle, and can reflect the influence of the target vehicle's own driving state on its lane-changing behavior. Specifically, the driving state change features may include the change amounts in different dimensions within a preset time period (e.g., 2s) before the current moment of the target vehicle, such as the position change amount, speed change amount, acceleration change amount, driving direction change amount, etc.

[0057] The surrounding environment features are used to characterize the relationship between the target vehicle and its surrounding vehicles, and between the target vehicle and its lane, and can reflect the influence of the surrounding vehicle and lane conditions on the target vehicle's lane-changing behavior. Specifically, the surrounding environment features may include the relative speed, relative distance, relative collision angle, predicted collision time, etc. between the target vehicle and its surrounding vehicles, and may also include the distance of the target vehicle from the lane boundary line, the displacement of the target vehicle within its lane, etc.

[0058] S103: Generate a lane-changing behavior prediction result for the target vehicle according to the driving state change features and surrounding environment features.

[0059] In one embodiment of the present disclosure, according to the driving state change features and surrounding environment features of the above-mentioned target vehicle, it is predicted whether the target vehicle will change lanes, and if the target vehicle will change lanes, its specific lane-changing type (including changing lanes to the left and changing lanes to the right) is determined.

[0060] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects: During the driving process of the host vehicle, first determine the target vehicle for which lane change behavior prediction is required, and then, for each target vehicle, extract the driving state change features used to characterize the change in the driving state of the target vehicle itself, and the surrounding environment features used to characterize the relationship between the target vehicle and its surrounding vehicles and between the target vehicle and its lane. Finally, obtain the lane change behavior prediction result of the target vehicle based on the driving state change features and the surrounding environment features. Since the lane change behavior of the target vehicle is related to both its own driving state and the mutual relationship between the target vehicle and its surrounding vehicles and between the target vehicle and its lane, by extracting the driving state change features and the surrounding environment features, it is possible to comprehensively and exhaustively obtain the influencing factors related to the vehicle lane change behavior, and thus accurately predict the lane change behavior; moreover, the above-mentioned driving state change features and surrounding environment features can be collected quickly and in real time, so the lane change behavior prediction result can be obtained quickly, which is convenient for the host vehicle to take corresponding measures in a timely manner for the lane change behavior of the surrounding vehicles, and improve the safety of autonomous driving.

[0061] As an alternative implementation manner, S102 includes: first obtain the historical trajectory of the target vehicle within a preset time period before the current moment and the position information of the target vehicle at the current moment; then, generate the driving state change features of the target vehicle according to the historical trajectory of the target vehicle within a preset time period before the current moment, and generate the surrounding environment features of the target vehicle according to the historical trajectory of the target vehicle within a preset time period before the current moment and the position information of the target vehicle at the current moment.

[0062] In one embodiment, the historical trajectory can be obtained through the data collected by the sensor in real time. For example, the sensor is a camera, and according to the video of the target vehicle collected in the recent 2s, the historical trajectory of the target vehicle can be determined.

[0063] Moreover, according to the positioning information of the host vehicle, search for the target vehicle around the host vehicle that requires lane behavior prediction on the high-precision map. After the target vehicle is searched, the position information of the target vehicle at the current moment can be obtained on the high-precision map. Among them, the high-precision map refers to a high-precision map used in the field of autonomous driving. In addition to including the static data in the conventional map, such as building positions, road attributes, etc., the high-precision map can also include the dynamic data at the current moment, such as the position information, motion state information of moving objects, and the real-time state information of traffic lights, etc.

[0064] As an alternative implementation, a driving state change feature is generated based on the historical trajectory of the target vehicle within a preset duration before the current moment, including: determining the historical motion state parameters of the target vehicle according to the historical trajectory of the target vehicle within a preset duration before the current moment; and calculating the driving state change feature based on the historical motion state parameters.

[0065] In an embodiment of the present disclosure, the historical motion state parameters include information such as the position, speed, acceleration, driving direction, lane where the vehicle is located, and longitudinal and lateral displacements in the Frenet coordinate system (also known as the S-D coordinate system) of the target vehicle within a preset duration before the current moment. The above historical motion state parameters can be extracted from the historical trajectory of the target vehicle, and then, based on these historical motion state parameters, the change amounts of the target vehicle in different driving state dimensions are calculated.

[0066] Exemplarily, sensor data is received at a frequency of 10 hz to obtain the historical trajectory of the target vehicle, and the above historical motion state parameters are extracted from the historical trajectory. Then, the historical motion state parameters are stored in a Least Recently Used with Cache (LRU cache). Finally, the historical motion state parameters within the required time period, such as 20 frames of data within 2 s, are obtained from the cache, and the driving state change feature of the target vehicle is calculated.

[0067] Among them, the calculation method of the driving state change feature is a conventional change amount calculation method. For example, in the speed dimension, the speed change amount of the target vehicle within 2 s before the current moment is the difference between the speed value at the current moment and the speed value 2 s ago.

[0068] As an alternative implementation, a surrounding environment feature of the target vehicle is generated based on the historical trajectory of the target vehicle within a preset duration before the current moment and the position information of the target vehicle at the current moment, including: obtaining the current motion state parameters of the surrounding vehicles of the target vehicle and the lane line information of the lane where the target vehicle is located; for each surrounding vehicle, determining the relative driving information between the target vehicle and the surrounding vehicle according to the position information of the target vehicle at the current moment and the current motion state parameters of the surrounding vehicle; determining the relative position information between the target vehicle and its lane according to the position information of the target vehicle at the current moment and the lane line information of the lane where the target vehicle is located; determining the displacement information of the target vehicle within its lane according to the historical trajectory of the target vehicle within a preset duration before the current moment, the position information of the target vehicle at the current moment, and the lane line information of the lane where the target vehicle is located; wherein, the surrounding environment feature includes the relative position information, the displacement information, and the relative driving information between the target vehicle and each surrounding vehicle.

[0069] In one embodiment of the present disclosure, the surrounding vehicles refer to the vehicles around the target vehicle, and the current motion state parameters refer to the position, speed, acceleration, driving direction, lane where the vehicle is located, and transverse and longitudinal displacements in the Frenet coordinate system of the surrounding vehicles at the current moment. Based on the position information of the target vehicle at the current moment, the surrounding vehicles can be searched on the high-precision map, and the current operating state parameters of the surrounding vehicles can be obtained on the high-precision map. At the same time, the lane line information of the lane where the target vehicle is located can also be obtained on the high-precision map. For example, the positions of the lane boundary lines and the position of the lane center line.

[0070] Exemplarily, Figure 3 FIG. is a schematic diagram of area division provided by an embodiment of the present disclosure. As Figure 3 shown, with the driving direction of the target vehicle as the due front, and with the position O of the target vehicle as the center, first determine the detection range of the surrounding vehicles. The detection range includes the lane where the target vehicle is located and the areas in front of and behind the position O of its adjacent lanes on the left and right sides with a preset distance (for example, 60 m). Then, based on the position O and the lane boundary lines of lanes 3 to 5, divide the detection range into 6 areas, namely the left front A, the due front B, the right front C, the left rear D, the due rear E, and the right rear F, and detect whether there are surrounding vehicles in these areas one by one. If there are, extract the current operating state parameters of each surrounding vehicle from the high-precision map; in this example, the surrounding vehicle b is detected in the due front area B 1 , and the surrounding vehicle f is detected in the right rear area F 1 and the surrounding vehicle f 2 .

[0071] The relative driving information is used to describe the relationship between the target vehicle and the surrounding vehicles, and can be calculated from the current motion state parameters of the target vehicle and the current motion state parameters of the surrounding vehicles. Among them, the relative driving information may include the relative distance, relative speed, relative deflection angle, predicted collision time, etc. between the target vehicle and the surrounding vehicles.

[0072] The relative position information is used to describe the position relationship between the target vehicle and the lane where it is located, and can be calculated from the position information of the target vehicle and the lane line information. Among them, the relative position information may include the distances between the target vehicle and the two lane boundary lines of the lane where it is located, the offset of the target vehicle relative to the lane center line of the lane where it is located, etc.

[0073] The displacement information is used to describe the displacement of the target vehicle in its own lane, and can be calculated according to the position information of the target vehicle within a preset time period before the current moment, the position information at the current moment, and the lane line information. Among them, the displacement information may include the transverse and longitudinal displacements of the target vehicle in its own lane.

[0074] Exemplarily, if the target vehicle was driving in a certain lane before a preset duration at the current moment and exited the lane within the preset duration before the current moment, the lateral displacement of the target vehicle in its current lane is the lateral distance between the position of the target vehicle before the preset duration at the current moment and the lane boundary line on the side where it exited; if the target vehicle did not exit the lane within the preset duration before the current moment, the lateral displacement of the target vehicle in its current lane is the lateral distance between the position of the target vehicle at the current moment and the position of the target vehicle before the preset duration at the current moment. If the target vehicle was driving in a certain lane before a preset duration at the current moment and exited the lane within the preset duration before the current moment, the longitudinal displacement of the target vehicle in its current lane is the longitudinal distance between the position of the target vehicle before the preset duration at the current moment and the position when it exited the lane; if the target vehicle did not exit the lane within the preset duration before the current moment, the longitudinal displacement of the target vehicle in its current lane is the longitudinal distance between the position of the target vehicle at the current moment and the position of the target vehicle before the preset duration at the current moment.

[0075] Different from the above displacement information, regardless of whether the target vehicle exits the lane, the lateral and longitudinal displacements of the target vehicle used when obtaining the driving state change characteristics are the lateral distance or longitudinal distance between the position of the target vehicle at the current moment and the position of the target vehicle before the preset duration at the current moment.

[0076] As an optional implementation manner, according to the driving state change characteristics and the surrounding environment characteristics, a lane change behavior prediction result of the target vehicle is generated, including: inputting the driving state change characteristics and the surrounding environment characteristics into a pre-trained lane change behavior prediction model to generate the lane change behavior prediction result of the target vehicle.

[0077] In an embodiment of the present disclosure, the input of the lane change behavior prediction model is the driving state change characteristics and the surrounding environment characteristics of the target vehicle, and the output is the lane change behavior prediction result of the target vehicle. Among them, the above lane change behavior prediction model is a lightweight model, such as an Extreme Gradient Boosting (XGBoost) model, a Light Gradient Boosting Machine (LightGBM) model, etc. Since the lightweight model occupies less GPU resources and has a fast calculation speed, it can perform fast real-time prediction while maintaining accuracy and is suitable for deployment on vehicles.

[0078] As an alternative implementation, the lane change behavior prediction model is trained as follows: Obtain a training sample set; wherein, each training sample in the training sample set includes the driving state change characteristics and surrounding environment characteristics of the sample vehicle, as well as the lane change behavior annotation category corresponding to the sample vehicle; Use the training sample set to train the model to obtain the lane change behavior prediction model.

[0079] Exemplarily, multiple vehicles in the real situation can be used as sample vehicles, and the driving state change characteristics and surrounding environment characteristics of each sample vehicle are extracted. Then, according to the actual lane change behavior of each sample vehicle, the lane change behavior category of the sample vehicle is marked, and the marked label is the lane change behavior annotation category, wherein the lane change behavior annotation category is specifically any one of maintaining the current lane, changing lanes to the left, and changing lanes to the right.

[0080] As an alternative implementation, the lane change behavior prediction result includes multiple lane change behavior categories and the prediction probability corresponding to each lane change behavior category. In this way, the lane change behavior of the predicted target vehicle can be described more comprehensively and accurately. Among them, the multiple lane change behavior categories can include maintaining the current lane, changing lanes to the left, and changing lanes to the right.

[0081] Exemplarily, the lane change behavior prediction result of the target vehicle includes: the prediction probability of the "maintain the current lane" category is 0.2; the prediction probability of the "change lanes to the left" category is 0.7; the prediction probability of the "change lanes to the right" category is 0.1.

[0082] Figure 4 It is a block diagram of a vehicle lane change behavior prediction device shown according to an exemplary embodiment. Refer to Figure 4 , the vehicle lane change behavior prediction device 200 may include a target vehicle determination module 201, an extraction module 202, and a prediction module 203.

[0083] The target vehicle determination module 201 is configured to determine at least one target vehicle around the host vehicle during the driving process of the host vehicle;

[0084] The extraction module 202 is configured to extract the driving state change characteristics and surrounding environment characteristics of each target vehicle;

[0085] The prediction module 203 is configured to generate a lane change behavior prediction result of the target vehicle according to the driving state change characteristics and the surrounding environment characteristics.

[0086] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects: During the driving of the host vehicle, first determine the target vehicle for which the lane change behavior prediction is to be performed, and then, for each target vehicle, extract the driving state change features for characterizing the change of the driving state of the target vehicle itself, and the surrounding environment features for characterizing the relationship between the target vehicle and its surrounding vehicles and between the target vehicle and its lane. Finally, obtain the lane change behavior prediction result of the target vehicle based on the driving state change features and the surrounding environment features. Since the lane change behavior of the target vehicle is related to both its own driving state and the mutual relationship between the target vehicle and its surrounding vehicles and its lane, by extracting the driving state change features and the surrounding environment features, the influencing factors related to the vehicle lane change behavior can be comprehensively and exhaustively obtained, and thus the lane change behavior can be accurately predicted; moreover, the above-mentioned driving state change features and surrounding environment features can be collected quickly and in real time, so the lane change behavior prediction result can be obtained quickly, which is convenient for the host vehicle to take timely measures corresponding to the lane change behavior of the surrounding vehicles and improve the safety of autonomous driving.

[0087] Optionally, the extraction module 202 includes:

[0088] The first acquisition sub-module is configured to acquire the historical trajectory of the target vehicle within a preset time period before the current moment and the position information of the target vehicle at the current moment;

[0089] The first generation sub-module is configured to generate the driving state change features according to the historical trajectory;

[0090] The second generation sub-module is configured to generate the surrounding environment features according to the historical trajectory and the position information.

[0091] Optionally, the first generation sub-module includes:

[0092] The first determination sub-module is configured to determine the historical motion state parameters of the target vehicle according to the historical trajectory;

[0093] The calculation sub-module is configured to calculate the driving state change features according to the historical motion state parameters.

[0094] Optionally, the second generation sub-module includes:

[0095] The second acquisition sub-module is configured to acquire the current motion state parameters of the surrounding vehicles of the target vehicle and the lane line information of the lane where the target vehicle is located;

[0096] A second determination sub-module, configured to determine, for each of the surrounding vehicles, relative driving information between the target vehicle and the surrounding vehicle according to the position information and current motion state parameters of the surrounding vehicle;

[0097] A third determination sub-module, configured to determine relative position information between the target vehicle and its lane according to the position information and the lane line information;

[0098] A fourth determination sub-module, configured to determine displacement information of the target vehicle within its lane according to the historical trajectory, the position information, and the lane line information;

[0099] Wherein, the surrounding environment features include the relative position information, the displacement information, and relative driving information between the target vehicle and each of the surrounding vehicles.

[0100] Optionally, the prediction module 203 is configured to input the driving state change features and the surrounding environment features into a pre-trained lane-changing behavior prediction model to generate the lane-changing behavior prediction result of the target vehicle.

[0101] Optionally, the lane-changing behavior prediction model is trained by a model training device, wherein the model training device includes:

[0102] An acquisition module, configured to acquire a training sample set; wherein each training sample in the training sample set includes driving state change features and surrounding environment features of a sample vehicle, and a lane-changing behavior annotation category corresponding to the sample vehicle;

[0103] A training module, configured to perform model training using the training sample set to obtain the lane-changing behavior prediction model.

[0104] Optionally, the lane-changing behavior prediction result includes multiple lane-changing behavior categories and a prediction probability corresponding to each lane-changing behavior category.

[0105] Regarding the vehicle lane-changing behavior prediction device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0106] The present disclosure also provides a computer-readable storage medium, on which computer program instructions are stored, and when the program instructions are executed by a processor, the steps of the vehicle lane-changing behavior prediction method provided by the present disclosure are implemented.

[0107] Figure 5is a block diagram of a vehicle 600 shown according to an exemplary embodiment. For example, the vehicle 600 can be a hybrid vehicle, or a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicles. The vehicle 600 can be an autonomous vehicle.

[0108] Referring to Figure 5 , the vehicle 600 can include various subsystems. For example, the infotainment system 610, the perception system 620, the decision control system 630, the drive system 640, and the computing platform 650. Among them, the vehicle 600 can also include more or fewer subsystems, and each subsystem can include multiple components. In addition, each subsystem and each component of the vehicle 600 can be interconnected by wired or wireless means.

[0109] In some embodiments, the infotainment system 610 can include a communication system, an entertainment system, and a navigation system, etc.

[0110] The perception system 620 can include several sensors for sensing information about the environment around the vehicle 600. For example, the perception system 620 can include a global positioning system (the global positioning system can be a GPS system, or a Beidou system, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter wave radar, ultrasonic radar, and a camera device.

[0111] The decision control system 630 can include a computing system, a vehicle controller, a steering system, an accelerator, and a braking system.

[0112] The drive system 640 can include components that provide motive power for the vehicle 600. In one embodiment, the drive system 640 can include an engine, an energy source, a powertrain, and wheels. The engine can be one or a combination of an internal combustion engine, an electric motor, and an air compression engine. The engine can convert the energy provided by the energy source into mechanical energy.

[0113] Some or all of the functions of the vehicle 600 are controlled by the computing platform 650. The computing platform 650 can include at least one processor 651 and a memory 652. The processor 651 can execute instructions 653 stored in the memory 652.

[0114] The processor 651 can be any conventional processor, such as a commercially available CPU. The processor may also include, for example, a Graphic Process Unit (GPU), a Field Programmable Gate Array (FPGA), a System on Chip (SOC), an Application Specific Integrated Circuit (ASIC), or a combination thereof.

[0115] The memory 652 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, magnetic disks, or optical disks.

[0116] In addition to the instructions 653, the memory 652 can also store data, such as road maps, route information, data on the position, direction, speed, etc. of the vehicle. The data stored in the memory 652 can be used by the computing platform 650.

[0117] In the embodiments of the present disclosure, the processor 651 can execute the instructions 653 to complete all or part of the steps of the above-described vehicle lane-changing behavior prediction method.

[0118] In another exemplary embodiment, a computer program product is also provided. The computer program product includes a computer program that can be executed by a programmable device. The computer program has a code portion for executing the above-described vehicle lane-changing behavior prediction method when executed by the programmable device.

[0119] Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only to be considered as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0120] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A method for predicting vehicle lane-changing behavior, characterized in that, it includes: During the driving process of the host vehicle, determine at least one target vehicle around the host vehicle; For each of the target vehicles, extract the driving state change characteristics and surrounding environment characteristics of the target vehicle; according to the driving state change characteristics and the surrounding environment characteristics, generate the lane-changing behavior prediction result of the target vehicle.

2. The method according to claim 1, characterized in that, The extracting the driving state change characteristics and surrounding environment characteristics of the target vehicle includes: Obtain the historical trajectory of the target vehicle within a preset time period before the current moment and the position information of the target vehicle at the current moment; Generate the driving state change characteristics according to the historical trajectory; Generate the surrounding environment characteristics according to the historical trajectory and the position information.

3. The method according to claim 2, characterized in that, The generating the driving state change characteristics according to the historical trajectory includes: Determine the historical motion state parameters of the target vehicle according to the historical trajectory; Calculate the driving state change characteristics according to the historical motion state parameters.

4. The method according to claim 2, characterized in that, The generating the surrounding environment characteristics according to the historical trajectory and the position information includes: Obtain the current motion state parameters of the surrounding vehicles of the target vehicle and the lane line information of the lane where the target vehicle is located; For each of the surrounding vehicles, determine the relative driving information between the target vehicle and the surrounding vehicle according to the position information and the current motion state parameters of the surrounding vehicle; Determine the relative position information between the target vehicle and its lane according to the position information and the lane line information; Determine the displacement information of the target vehicle within its lane according to the historical trajectory, the position information and the lane line information; Wherein, the surrounding environment characteristics include the relative position information, the displacement information, and the relative driving information between the target vehicle and each of the surrounding vehicles.

5. The method according to claim 1, characterized in that, The generating the lane-changing behavior prediction result of the target vehicle according to the driving state change characteristics and the surrounding environment characteristics includes: Input the driving state change characteristics and the surrounding environment characteristics into a pre-trained lane-changing behavior prediction model to generate the lane-changing behavior prediction result of the target vehicle.

6. The method according to claim 5, characterized in that, The lane-changing behavior prediction model is trained in the following manner: Obtain a training sample set; wherein, each training sample in the training sample set includes the driving state change characteristics and surrounding environment characteristics of the sample vehicle, and the lane-changing behavior annotation category corresponding to the sample vehicle; Use the training sample set for model training to obtain the lane-changing behavior prediction model.

7. The method according to any one of claims 1-6, characterized in that, The lane-changing behavior prediction result includes multiple lane-changing behavior categories and the prediction probability corresponding to each lane-changing behavior category.

8. A vehicle lane-changing behavior prediction device, It is characterized in that It includes: A target vehicle determination module configured to determine at least one target vehicle around the host vehicle during the driving of the host vehicle; An extraction module configured to extract the driving state change characteristics and surrounding environment characteristics of each target vehicle; A prediction module configured to generate a lane change behavior prediction result of the target vehicle according to the driving state change characteristics and the surrounding environment characteristics.

9. A vehicle lane change behavior prediction device It is characterized in that It includes: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to: Determine at least one target vehicle around the host vehicle during the driving of the host vehicle; Extract the driving state change characteristics and surrounding environment characteristics of each target vehicle; generate a lane change behavior prediction result of the target vehicle according to the driving state change characteristics and the surrounding environment characteristics.

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

Images