Driving behavior intention prediction method, device, storage medium and program product

Abstract

The embodiment of the application discloses a driving behavior intention prediction method and device, a storage medium and a program product, and relates to the technical field of intelligent auxiliary driving. The method comprises the following steps: acquiring a vehicle historical trajectory and a first prediction trajectory corresponding to a future first time period; constructing a reference driving trajectory based on the vehicle historical trajectory and the first prediction trajectory, and extracting a trajectory point set corresponding to the reference driving trajectory; determining trajectory timing features according to the trajectory point position and the corresponding trajectory point timestamp of each trajectory point in the trajectory point set; and inputting the trajectory timing features into a driver intention model to predict the corresponding driving behavior intention. Therefore, the behavior mode of the driver can be comprehensively understood, and the system can more accurately predict the future behavior of the driver.

Description

Technical Field

[0001] This application relates to the field of intelligent assisted driving technology, and in particular to a driving behavior intention prediction device, storage medium and program product. Background Technology

[0002] In the fields of Advanced Driver Assistance Systems (ADAS) and autonomous driving, predicting driving behavior intentions is a key technology for improving road safety and driving efficiency.

[0003] Traditional driving intention prediction methods typically predict the driver's next action based on real-time vehicle operating parameters (such as steering wheel angle, acceleration, and vehicle speed). These methods rely on simplified physical models or rule-based systems to infer short-term driving behavior, such as whether there is an intention to change lanes, accelerate, or stop.

[0004] However, relying solely on real-time vehicle operating parameters makes it difficult for the system to accurately predict driving intentions over a longer period. Therefore, it may not be effective enough in handling complex traffic environments or responding to emergencies, and it cannot provide in-depth insights into driver behavior patterns, resulting in poor adaptability to dynamic environments.

[0005] Currently, the industry has not proposed a better solution to the above problems. Summary of the Invention

[0006] This application provides a driving behavior intention prediction device, equipment, storage medium, and program product to at least solve one of the above-mentioned technical problems.

[0007] In a first aspect, embodiments of this application provide a method for predicting driving behavior intentions, comprising: acquiring a vehicle's historical trajectory and a first predicted trajectory corresponding to a first future time period; constructing a reference driving trajectory based on the vehicle's historical trajectory and the first predicted trajectory, and extracting a set of trajectory points corresponding to the reference driving trajectory; determining trajectory temporal features based on the trajectory point position and corresponding trajectory point timestamp of each trajectory point in the trajectory point set; and inputting the trajectory temporal features into a driver intention model to predict the corresponding driving behavior intentions.

[0008] Secondly, embodiments of this application provide a driving behavior intention prediction device, comprising: a trajectory acquisition unit, configured to acquire a vehicle's historical trajectory and a first predicted trajectory corresponding to a future first time period; a trajectory point set extraction unit, configured to construct a reference driving trajectory based on the vehicle's historical trajectory and the first predicted trajectory, and extract the trajectory point set corresponding to the reference driving trajectory; a trajectory timing determination unit, configured to determine trajectory timing features based on the trajectory point position and corresponding trajectory point timestamp of each trajectory point in the trajectory point set; and a driving intention prediction unit, configured to input the trajectory timing features into a driver intention model to predict the corresponding driving behavior intention.

[0009] Thirdly, embodiments of this application provide a storage medium storing one or more programs including execution instructions, which can be read and executed by electronic devices (including but not limited to computers, servers, or network devices) to perform any of the driving behavior intention prediction methods described above.

[0010] Fourthly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform any of the driving behavior intention prediction methods and systems described above in this application.

[0011] Fifthly, embodiments of this application also provide a computer program product, the computer program product including a computer program stored on a storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to execute any of the above-described driving behavior intention prediction methods and systems.

[0012] Sixthly, embodiments of this application provide a mobile platform on which the computer device described in any of the preceding claims of this application is installed.

[0013] The beneficial effects of the embodiments of this application are at least as follows:

[0014] By introducing a comprehensive analysis of vehicle historical trajectories and future predicted trajectories, it is possible to integrate past driving behavior patterns with future driving intentions. Compared with traditional methods that rely solely on real-time vehicle operation parameters, this approach enables a more comprehensive understanding of driver behavior patterns, allowing the system to more accurately predict future driver behavior and further improve road safety and driving efficiency. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating an example of a driving behavior intention prediction method according to an embodiment of this application is shown;

[0017] Figure 2 A flowchart illustrating an example of obtaining a first predicted trajectory according to an embodiment of this application is shown.

[0018] Figure 3 A flowchart is shown as an example of a method for predicting vehicle trajectory based on driving behavior intent according to an embodiment of this application;

[0019] Figure 4 This diagram illustrates an example of the effect of a vehicle trajectory determined according to a conventional trajectory prediction scheme.

[0020] Figure 5 A schematic diagram illustrating the effect of an example of a reference driving trajectory according to an embodiment of this application is shown.

[0021] Figure 6 A schematic diagram illustrating the effect of an example of a second predicted trajectory according to an embodiment of this application is shown.

[0022] Figure 7 A schematic diagram showing the comparison between the second predicted trajectory and the first predicted trajectory according to an embodiment of this application is illustrated.

[0023] Figure 8 A structural block diagram of an example of a driving behavior intention prediction device according to an embodiment of this application is shown;

[0024] Figure 9 This is a schematic diagram of the structure of an embodiment of the electronic device of this application. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0026] It should also be noted that, in this document, the terms "comprising" or "including" include not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0027] The technical solutions in this application, including the collection, storage, use, processing, transmission, provision, and disclosure of users' personal information, comply with relevant laws and regulations and do not violate public order and good morals.

[0028] Figure 1 A flowchart illustrating an example of a driving behavior intention prediction method according to an embodiment of this application is shown.

[0029] Regarding the execution subject of the method in this application embodiment, it can be any controller or processor with computing or processing capabilities. The method for predicting driving behavior intentions executes a method that, compared to the current related technologies that widely use real-time vehicle operation parameters (such as steering wheel angle, acceleration, vehicle speed, etc.) to predict driver intentions, in this application embodiment, by integrating a comprehensive analysis of the vehicle's historical trajectory and future predicted trajectory, the driver's behavior patterns are fully understood, and the driver's behavioral intentions can be predicted more accurately.

[0030] In some examples, it can be integrated into electronic devices or terminals through software, hardware, or a combination of both, and the type of terminal or electronic device can be diverse, such as mobile phones, tablets, or desktop computers. In some scenarios, it can also be set up in the vehicle's infotainment system to provide or enhance intelligent vehicle assistance functions.

[0031] like Figure 1 As shown, in step S110, the vehicle's historical trajectory and the corresponding first predicted trajectory for the first future time period are obtained.

[0032] Here, the vehicle's historical trajectory can be either a historical trajectory from a pre-defined past time period adjacent to the current time, or a historical trajectory from a pre-defined distance adjacent to the current vehicle position, which can be obtained through the vehicle positioning module's recording. Furthermore, the first predicted trajectory can be determined by using various machine learning or physical models to predict the trajectory based on vehicle state parameters (e.g., the current vehicle's body posture, vehicle position, speed, steering wheel angle, etc.), and this is not limited here.

[0033] It should be noted that when using the model to predict the first predicted trajectory in the future, the closer the future time point is to the current time, the more accurate the predicted trajectory point will be; conversely, the farther the future time point is from the current time, the greater the deviation in the predicted trajectory point will be. Therefore, the length of the first preset time period should be determined based on the accuracy of the trajectory prediction results of the above model. For example, the accuracy of the first predicted trajectory should be ensured as much as possible, and it should generally not be too long.

[0034] In step S120, a reference driving trajectory is constructed based on the vehicle's historical trajectory and the first predicted trajectory, and the trajectory point set corresponding to the reference driving trajectory is extracted.

[0035] In some implementations, the vehicle's historical trajectory and the first predicted trajectory are stitched together to obtain a reference driving trajectory for the vehicle in the current driving environment. Furthermore, trajectory points can be extracted based on timestamps or distance; for example, multiple trajectory points can be extracted according to preset time intervals or distance intervals to construct a corresponding trajectory point set.

[0036] In step S130, the trajectory temporal characteristics are determined based on the trajectory point position and corresponding trajectory point timestamp of each trajectory point in the trajectory point set.

[0037] In some implementations, the positions of each trajectory point are sequentially connected according to the trajectory point timestamps to obtain the corresponding trajectory temporal features. Furthermore, trajectory features corresponding to each trajectory point (e.g., trajectory point position, lateral offset angle, etc.) can be extracted through trajectory geometric analysis, or vehicle driving state features (e.g., steering wheel angle, vehicle speed, accelerator pedal depth, etc.) can be collected through sensing modules. These features are then integrated into the trajectory temporal features, enriching the feature dimensions of the trajectory temporal features. This is more conducive to learning and predicting driving behavior patterns in the current driving environment. More details will be elaborated below with other examples.

[0038] In step S140, the trajectory temporal features are input into the driver intent model to predict the corresponding driving behavior intent.

[0039] Here, driving behavior intentions can include lateral or longitudinal behavior intentions relative to the lane direction. More specifically, driving behavior intentions can include lateral lane departure probabilities and / or longitudinal speed adjustment probabilities to achieve accurate prediction of the driver's lane-changing, acceleration, or braking behaviors.

[0040] It should be noted that traditional models generally predict driving intentions based on vehicle state parameters, lacking the ability to adapt to dynamic changes in the environment. In this embodiment, by introducing a reference driving trajectory and combining the trajectory position characteristics of the trajectory points in the current driving environment at historical and future times, driving behavior under different traffic conditions can be better analyzed and predicted. This enables advanced driver assistance systems to more effectively handle unexpected events in the current driving environment, such as emergency avoidance or temporary route adjustments.

[0041] More specifically, the driver intent model can employ various non-restricted machine learning models (e.g., deep learning models) or physical models, etc. In some implementations, the driver intent model employs a hidden Markov prediction model, using trajectory data of human drivers under various operating conditions, and predicts the probability of lateral lane departure and / or longitudinal speed adjustment by extracting position, velocity, angular velocity, lateral acceleration, and lane departure speed on the trajectory as feature points.

[0042] Hidden Markov Models (HMMs) are statistical models that assume the state of a system is not directly observable (i.e., "hidden"), but can be indirectly observed through observable random variables. This type of model is well-suited for describing the transitions between different states of a system and inferring state changes through observed behavior. Therefore, based on the lightweight nature of HMMs, they can be deployed locally in the vehicle for rapid real-time computation without interaction with a cloud platform, meeting the high-timeliness requirements for driving intentions under high-speed driving conditions.

[0043] In some implementations, the Baum-Welch algorithm can be used to train a driver intent model. The Baum-Welch algorithm is a special form of expectation-maximization algorithm used to train the parameters of a Hidden Markov Model, such as state transition probabilities and observation probabilities. In practice, the Baum-Welch algorithm optimizes the model parameters through an iterative process to maximize the probability (or likelihood) of a given observation.

[0044] By utilizing driver trajectory data under different conditions (including features such as position, speed, angular velocity, lateral acceleration, and lane departure speed), the model can learn the implicit states behind driving behavior and the transition patterns between states. Thus, after training, the model can predict possible changes in driver behavior, such as the probability of lane departure, based on the input trajectory feature points.

[0045] Regarding the implementation details of step S130 above, in some embodiments, at least one trajectory feature group corresponding to a trajectory point is obtained. The trajectory feature group includes at least one of the following: trajectory point position, driving speed, angular velocity, lateral acceleration, and lane departure speed. Then, the trajectory feature groups are combined according to the timestamps corresponding to each trajectory point to determine the trajectory temporal features. Thus, the trajectory temporal features, in addition to containing spatiotemporal feature information of historical and future trajectories, also incorporate vehicle driving state information of each trajectory point, providing multi-dimensional feature information. This allows the driver intent model to more comprehensively understand and learn the driver's driving behavior patterns in the current driving environment.

[0046] This application's embodiments, by integrating feature information from historical and future trajectories to predict driving intentions, demonstrate greater versatility across different vehicle types and driving styles. This not only improves prediction accuracy and foresight but also enhances environmental adaptability and emergency response capabilities, further improving road safety and driving efficiency. Furthermore, the solution's versatility and scalability enable various application scenarios, showcasing broad application prospects.

[0047] More specifically, the driving behavior intention prediction method provided in this application can be used in various business application scenarios, such as accurate trajectory prediction, vehicle collision warning, and smooth driving systems. In one example of this application's embodiments, in a vehicle collision warning system, accurately predicting driving behavior intentions can effectively reduce traffic accidents caused by misunderstandings of driving intentions. In another example of this application's embodiments, based on accurate driving behavior intention prediction, reasonable speed control and lane-changing operations can be achieved, reducing unnecessary braking and acceleration, thereby optimizing fuel consumption and improving driving smoothness.

[0048] Figure 2 A flowchart illustrating an example of obtaining a first predicted trajectory according to an embodiment of this application is shown.

[0049] like Figure 2 As shown, in step S210, at least one real-time vehicle status parameter is acquired.

[0050] Here, the real-time vehicle status parameters can be various operating status parameters of the vehicle at present, such as steering wheel angle, motor speed or engine speed, to match the feature dimensions of the kinematic model.

[0051] In step S220, the real-time state parameters of each vehicle are input into the kinematic model to determine the corresponding kinematic prediction trajectory.

[0052] Here, the kinematic model makes kinematic predictions by combining real-time vehicle state parameters (such as steering wheel angle and speed). In practice, a suitable model can be selected according to requirements, and numerical integration methods can be used to solve the model equations to obtain the kinematically predicted trajectory. Furthermore, the types of kinematic models can be diverse, such as multibody dynamics models, point mass models, etc., and the appropriate type of kinematic model can be switched according to actual business needs or driving environment.

[0053] In step S230, the first predicted trajectory corresponding to the first time period adjacent to the current time is extracted from the kinematic predicted trajectory.

[0054] As described above, the kinematic trajectory predicted by the kinematic model is relatively accurate closer to the current time. For example, if the kinematic model outputs a kinematic trajectory prediction for the next 4 seconds, but only the first 1.5 seconds are relatively accurate, then the length of the first time interval can be selected as 1.5 seconds, and this segment of the trajectory can be extracted as the first predicted trajectory. It should be understood that different kinematic models or vehicle configurations will lead to differences in the accuracy of vehicle trajectory predictions; therefore, the length of the first time interval should be adjusted accordingly to ensure the reliability of the corresponding first predicted trajectory.

[0055] Figure 3 A flowchart illustrating an example of a method for predicting vehicle trajectory based on driving behavior intent according to an embodiment of this application is shown.

[0056] It should be noted that, as described above, traditional trajectory prediction schemes generally rely solely on real-time vehicle operating parameters (such as steering wheel angle, acceleration, and vehicle speed) to predict the driver's next action. However, different drivers may have vastly different operating habits, which will correspondingly lead to personalized trajectory schemes, causing traditional schemes to deviate significantly from the predicted vehicle trajectory.

[0057] Figure 4 This diagram illustrates an example of the effect of a vehicle trajectory determined according to a conventional trajectory prediction scheme.

[0058] Traditional driver trajectory prediction relies solely on the current state of the vehicle, such as using the current steering wheel angle or speed as a control variable, and then projecting the driver's future trajectory forward in fixed steps or for fixed durations. However, this approach suffers from low accuracy in short periods (e.g., within 1.5 seconds) because it depends on the vehicle's motion and cannot anticipate user behavior. Furthermore, as the prediction time increases (e.g., after 1.5 seconds), the accuracy decreases further.

[0059] Currently, as the enabling speed of driver assistance functions gradually increases, the required duration of driver trajectory prediction also increases accordingly. For example, the forward collision warning function needs to issue a warning within a relatively long collision time (e.g., 3.0 seconds) when the vehicle is operating at high speed. To meet this requirement, the driver's trajectory in the next 3.0 seconds needs to be accurately predicted. Inaccurate prediction will lead to false alarms or missed alarms in the forward collision warning function, resulting in poor function performance.

[0060] The following section will elaborate on the process of correcting the driving trajectory based on high-precision driving intention behavior, provided in this embodiment, taking into account the case where the driving behavior intention is a probability of lateral lane departure.

[0061] like Figure 3 As shown, in step S310, the vehicle's historical trajectory and the corresponding first predicted trajectory for the first future time period are obtained.

[0062] In step S320, a reference driving trajectory is constructed based on the vehicle's historical trajectory and the first predicted trajectory, and the trajectory point set corresponding to the reference driving trajectory is extracted.

[0063] Figure 5 A schematic diagram illustrating the effect of an example of a reference driving trajectory according to an embodiment of this application is shown. Figure 5 As shown, the reference driving trajectory of the vehicle's ego includes both the vehicle's historical trajectory and the first predicted trajectory for the future.

[0064] In step S330, the trajectory temporal characteristics are determined based on the trajectory point position and corresponding trajectory point timestamp of each trajectory point in the trajectory point set.

[0065] In step S340, the trajectory temporal features are input into the driver intention model to predict the corresponding lateral lane departure probability.

[0066] Here, the driver intention model adopts a hidden Markov model, which requires the vehicle to generate deviation probabilities in real time during driving.

[0067] More specifically, the formula for calculating the offset probability is as follows:

[0068]

[0069] α 1，i =π i b i (o1), i∈[1,N)

[0070]

[0071] In the formula, P(O|λ) is the offset probability given O and λ, O=(o1,o2,…,ot) is the time series of trajectory feature points, λ={A,B,π} are model parameters, A is the state transition probability matrix, B is the observation probability matrix, π is the initial state probability vector, N is the number of hidden states, and b i (o t Let α be the probability of generating observation ot in the i-th hidden state. j i is the element in the j-th row and i-th column of matrix A, πi is the i-th element of the π vector, and T is the duration of the feature point time series.

[0072] In step S350, the vehicle yaw angle difference is obtained.

[0073] Here, the vehicle yaw angle difference represents the difference between the vehicle's current direction of travel and the heading of the lane centerline, which can be determined through methods such as onboard camera image processing, GPS map matching, LiDAR and environmental modeling.

[0074] In step S360, virtual lane guidance reference elements are generated based on the vehicle's real-time status parameters, vehicle yaw angle difference, and lateral lane deviation probability.

[0075] Here, the virtual lane guidance reference element is the predicted virtual reference position or virtual reference lane line that the vehicle will reach. In some implementations, real-time vehicle state parameters, vehicle yaw angle difference, and lateral lane departure probability are input into the offset calculation model to determine the corresponding lane departure. Then, based on the lane departure, the virtual lane guidance reference element is generated.

[0076] In some implementations, the offset calculation model can employ a mathematical model or a deep learning model to achieve the data conversion from offset probability to offset amount. More specifically, the driver's future offset direction and lane offset amount can be determined by the following formula:

[0077] s=P(O|λ)(f(θ,a θ )+f(γ,aγ)+d)

[0078]

[0079] In the formula, s is the offset, P(O|λ) is the offset probability given O and λ, θ is the steering wheel angle, γ is the vehicle yaw angle difference, and α is the yaw angle difference. θ and α γ is an adjustable parameter (the parameter value can be adjusted according to the actual application and then entered into the above formula for calculation), and d is the distance between the vehicle and the center line of the lane.

[0080] Furthermore, based on the calculated lane offset, the lateral coordinates of the future vehicle position are adjusted, for example, by generating virtual lane guidance reference elements using interpolation algorithms (e.g., spline interpolation).

[0081] In step S370, a planned path from the vehicle's current location to the virtual lane guidance reference element is determined based on the path planning model.

[0082] Specifically, the virtual lane guidance reference element is set as the endpoint of the path planning, and the corresponding planned path is determined by planning through the path planning model. Here, the path planning model can be of various types, such as a gridded path planning model or a sampled base path planning model, etc., and is not limited thereto, to realize path planning from the vehicle's current position to the virtual lane guidance reference element.

[0083] In some examples of embodiments of this application, the path planning model may employ iLQR (iterative Linear Quadratic Regulator). iLQR is an optimization-based path planning algorithm that iteratively linearizes the nonlinear system and solves the linear quadratic regulation problem to gradually approach the global optimum.

[0084] In step S380, the second predicted trajectory is determined based on the planned path.

[0085] In some implementations, the planned path from the current location to the center line of the virtual lane is used as the second predicted trajectory, which is the driver's predicted trajectory. Furthermore, as the vehicle's trajectory is continuously updated, the corresponding second predicted trajectory is also updated in real time.

[0086] Figure 6 A schematic diagram illustrating the effect of an example of a second predicted trajectory according to an embodiment of this application is shown. Figure 6 As shown, Ego represents the vehicle itself, MOT represents the vehicle ahead, and the blue trajectory line represents the second predicted trajectory. Here, since lane information is introduced in the trajectory generation process, when the detection distance of the lane line by the onboard driver assistance system is sufficient (e.g., more than 120 meters), the driver trajectory prediction can be extended to 3.0 seconds to 5.0 seconds based on the lane information of this length, achieving high accuracy for long-term and long-distance prediction.

[0087] In step S390, the first predicted trajectory is updated based on the second predicted trajectory.

[0088] More specifically, the second predicted trajectory is used to replace or calibrate the first predicted trajectory to improve the accuracy of trajectory prediction results by incorporating driver behavioral intent into the vehicle trajectory prediction function. During the calibration of the first predicted trajectory using the second predicted trajectory, a weighted average can be applied. The weights can be gradually adjusted based on the prediction time period. For example, the first predicted trajectory has a higher weight in the short term (e.g., a first preset time period), while the weight of the second predicted trajectory gradually increases over time. By dynamically adjusting the weights by comprehensively considering the accuracy distribution of the motion trajectory, the accuracy of the final output vehicle predicted trajectory can be effectively improved.

[0089] Figure 7 A schematic diagram showing the comparison between the second predicted trajectory and the first predicted trajectory according to an embodiment of this application is illustrated.

[0090] like Figure 7 As shown, ego represents the vehicle, M represents the obstacle target, the red dashed line is the conventional driving trajectory, and the green dashed line is the driving trajectory of this application. The vehicle ego is traveling at a speed of 10 m / s in the right lane. There is a moving target M1 (e.g., an electric bicycle) moving at 3 m / s in front. When the driver notices the slow-moving target in front, he chooses to turn the steering wheel to avoid the target.

[0091] The first predicted trajectory obtained from traditional trajectory prediction shows a tendency to deviate from the current lane laterally. However, due to the small lateral deviation at this point, a forward collision warning is still triggered for the moving target M1. In contrast, in this embodiment, taking into account the driver's driving intentions, the vehicle still travels at 10m / s to the left lane. Before deviating from the driver's predicted trajectory, there is already a significant lateral distance between the vehicle and the moving target M1. While fully considering the driver's intentions, the planned path is long enough to effectively avoid collisions between the vehicle and other obstacles in the environment (such as M2 in the figure), thus preventing the triggering of a forward collision warning and better ensuring safe driving.

[0092] It should be understood that the above combination Figure 7 The description of the second predicted trajectory is for illustrative purposes only. Different drivers, due to their varying driving habits, may exhibit different driving intentions even in the same road conditions. For example, some drivers may prefer to drive in the lane adjacent to an obstacle to facilitate a quicker return to their current lane, while others may prefer to change lanes to a more distant lane to minimize the risk of collision. Therefore, the second predicted trajectory generated by the system will be adjusted accordingly for drivers with different operating habits to meet their individual driving preferences and achieve better adaptability and flexibility.

[0093] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0094] Figure 8 A structural block diagram of an example of a driving behavior intention prediction device according to an embodiment of this application is shown.

[0095] like Figure 8 As shown, the driving behavior intention prediction device 800 includes a trajectory acquisition unit 810, a trajectory point set extraction unit 820, a trajectory timing determination unit 830, and a driving intention prediction unit 840.

[0096] The trajectory acquisition unit 810 is used to acquire the vehicle's historical trajectory and the first predicted trajectory for the corresponding first time period in the future.

[0097] The trajectory point set extraction unit 820 is used to construct a reference driving trajectory based on the vehicle's historical trajectory and the first predicted trajectory, and to extract the trajectory point set corresponding to the reference driving trajectory.

[0098] The trajectory timing determination unit 830 is used to determine the trajectory timing characteristics based on the trajectory point position and the corresponding trajectory point timestamp of each trajectory point in the trajectory point set.

[0099] The driving intention prediction unit 840 is used to input the trajectory temporal features into the driver intention model to predict the corresponding driving behavior intention.

[0100] In some embodiments, this application provides a non-volatile computer-readable storage medium storing one or more programs including execution instructions, which can be read and executed by electronic devices (including but not limited to computers, servers, or network devices) to perform any of the driving behavior intention prediction methods described above.

[0101] In some embodiments, this application also provides a computer program product, the computer program product including a computer program stored on a non-volatile computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform any of the above-described driving behavior intention prediction methods.

[0102] In some embodiments, this application also provides an electronic device including: at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform a driving behavior intention prediction method.

[0103] The driving behavior intention prediction device described in the above embodiments of this application can be used to execute the driving behavior intention prediction method described in the present application, and accordingly achieve the technical effects achieved by the driving behavior intention prediction method described in the above embodiments of this application, which will not be elaborated here. In the embodiments of this application, the relevant functional modules can be implemented by a hardware processor.

[0104] Figure 9 This is a schematic diagram of the hardware structure of an electronic device for performing a driving behavior intention prediction method according to another embodiment of this application, as shown below. Figure 9 As shown, the device includes:

[0105] One or more processors 910 and memory 920, Figure 9 Take the 910 processor as an example.

[0106] The device for performing the driving behavior intention prediction method may further include: an input device 930 and an output device 940.

[0107] The processor 910, memory 920, input device 930, and output device 940 can be connected via a bus or other means. Figure 9 Taking the example of a connection between China and Israel via a bus.

[0108] The memory 920, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the driving behavior intention prediction method in the embodiments of this application. The processor 910 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions, and modules stored in the memory 920, thereby implementing the driving behavior intention prediction method in the above-described method embodiments.

[0109] The memory 920 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the driving behavior intention prediction device. Furthermore, the memory 920 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 920 may optionally include memory remotely located relative to the processor 910, and this remote memory may be connected to the driving behavior intention prediction device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0110] The input device 930 can receive input digital or character information and generate signals related to user settings and function control of the driving behavior intention prediction device. The output device 940 may include a display device such as a display screen.

[0111] The one or more modules are stored in the memory 920, and when executed by the one or more processors 910, they execute the driving behavior intention prediction method in any of the above method embodiments.

[0112] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0113] The electronic devices in this application embodiments exist in various forms, including but not limited to:

[0114] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and primarily aim to provide voice and data communication. These terminals include: smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones, etc.

[0115] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, possessing computing and processing capabilities, and generally also have mobile internet access features. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0116] (3) Portable entertainment devices: These devices can display and play multimedia content. This category includes audio and video players (such as iPods), handheld game consoles, e-book readers, as well as smart toys and portable car navigation devices.

[0117] (4) Server: A device that provides computing services. The components of a server include a processor, hard disk, memory, system bus, etc. Servers are similar to general computer architectures, but because they need to provide highly reliable services, they have higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0118] (5) Other electronic devices with data interaction functions.

[0119] In some embodiments, this application also provides a mobile platform on which the computer device described in any embodiment of this application is installed. The mobile platform includes, but is not limited to, vehicles, tracked robots, bipedal robots, quadrupedal robots, etc., wherein the vehicle can be a passenger car, pickup truck, truck, etc. It should be noted that the above are merely examples, and this application does not limit the specific form of the mobile platform.

[0120] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0121] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for predicting driving behavior intention, comprising: Obtain the vehicle's historical trajectory and the first predicted trajectory for the corresponding first time period in the future; A reference driving trajectory is constructed based on the vehicle's historical trajectory and the first predicted trajectory, and the trajectory point set corresponding to the reference driving trajectory is extracted. Based on the trajectory point position and corresponding trajectory point timestamp of each trajectory point in the trajectory point set, the trajectory temporal characteristics are determined; The trajectory temporal features are input into the driver intent model to predict the corresponding driving behavior intent; The step of determining the trajectory temporal characteristics based on the trajectory point position and corresponding trajectory point timestamp of each trajectory point in the trajectory point set includes: Obtain the trajectory feature group corresponding to each trajectory point; the trajectory feature group includes at least one of the following: trajectory point position, driving speed, angular velocity, lateral acceleration, and lane departure speed; The trajectory temporal characteristics are determined by combining the timestamps corresponding to each trajectory point into the respective trajectory feature groups. Wherein, the driving behavior intention includes a lateral lane departure probability, and after inputting the trajectory temporal features into the driver intention model to predict the corresponding driving behavior intention, the method includes: Obtain the vehicle's yaw angle difference; Virtual lane guidance reference elements are generated based on the vehicle's real-time status parameters, the vehicle's yaw angle difference, and the lateral lane deviation probability.

2. The method according to claim 1, characterized in that, The acquisition of the vehicle's historical trajectory and the corresponding first predicted trajectory for the first future time period includes: Obtain at least one real-time vehicle status parameter; The real-time state parameters of each vehicle are input into the kinematic model to determine the corresponding kinematic prediction trajectory. From the kinematic predicted trajectory, extract the first predicted trajectory corresponding to the first time period adjacent to the current time.

3. The method according to claim 1, characterized in that, The step of generating virtual lane guidance reference elements based on real-time vehicle status parameters, vehicle yaw angle difference, and lateral lane departure probability includes: The real-time vehicle status parameters, the vehicle yaw angle difference, and the lateral lane departure probability are input into the lane departure calculation model to determine the corresponding lane departure. Based on the lane offset, a virtual lane guidance reference element is generated.

4. The method according to claim 3, characterized in that, After generating virtual lane guidance reference elements based on real-time vehicle status parameters, vehicle yaw angle difference, and lateral lane departure probability, the method further includes: The planned path from the vehicle's current position to the virtual lane guidance reference element is determined based on the path planning model; Based on the planned path, determine the second predicted trajectory; The first predicted trajectory is updated based on the second predicted trajectory.

5. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-4.

6. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-4.

7. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-4.

Images