Data processing method and apparatus, and advanced driving device, storage medium and program product

Abstract

Provided are a data processing method and apparatus, and an advanced driving device, a storage medium and a program product, which relate to the field of advanced driving. The method comprises: acquiring first data of candidate vehicles, wherein the first data is configured to indicate driving behaviors of drivers of the candidate vehicles; on the basis of the first data, determining a target collection vehicle from the candidate vehicles; and acquiring human driving data of the target collection vehicle, wherein the human driving data is data generated during the process of a driver driving a vehicle, and is configured to determine a driving strategy for the vehicle. By means of acquiring the first data for indicating the driving behaviors of the drivers, the driving behaviors of the drivers of the candidate vehicles can be fully understood, thereby effectively identifying the characteristics of the driving behaviors of the drivers of the candidate vehicles. Then, on the basis of the characteristics of the driving behaviors of the drivers, the target collection vehicle is determined from the candidate vehicles, and data collection is targeted at the specific target collection vehicle, thereby improving the efficiency and quality of data collection.

Description

Data processing methods, devices, intelligent driving equipment, storage media and software products Technical Field

[0001] This application relates to the field of intelligent driving, and in particular to data processing methods, devices, intelligent driving equipment, storage media, and program products. Background Technology

[0002] In the field of intelligent driving, human driving data refers to the data generated by drivers while driving a vehicle. Human driving data can be applied to scenarios such as training intelligent driving models.

[0003] Based on current methods for collecting driver and passenger data, the collected data scenarios are complex and the data volume is large, with a large amount of low-quality data. Using this data for model training and other operations is often inefficient and ineffective. Summary of the Invention

[0004] This application provides data processing methods, apparatus, intelligent driving devices, storage media, and program products, with the aim of reducing the proportion of low-quality human-driving data collected.

[0005] In a first aspect, this application provides a data processing method. The steps of the method can be executed by a data processing device, or the method can be executed by a component (such as a chip, chip system, etc.) configured in the data processing device, or it can be implemented by a logic module or software capable of realizing all or part of the functions of the data processing device. This application does not limit the scope of the method.

[0006] As an example and not a limitation, the data processing device can be a device capable of interacting with an advanced driving system, such as a server on the cloud platform of the advanced driving system.

[0007] For example, the method includes: acquiring first data of candidate vehicles, the first data being used to indicate the driving behavior of the driver of the candidate vehicles; determining a target collection vehicle from the candidate vehicles based on the first data; and acquiring driver data of the target collection vehicle, the driver data being data generated by the driver during the driving process, the driver data being used to determine the driving strategy of the vehicle.

[0008] Understandably, the first set of data refers to the historical driver and passenger data generated by the candidate vehicles. This may include, but is not limited to, the candidate vehicles' mileage, vehicle motion status data, or driving behavior data. The volume of this first set of data is relatively small and insufficient for determining the vehicle's driving strategy. The driver and passenger data acquired from the target vehicle differs from this first set of data. This data refers to the driver and passenger data that will be generated on the target vehicle in the future. The volume of this data is larger than that of the first set of data; that is, the volume of this data is sufficient to determine the vehicle's driving strategy.

[0009] Based on the above technical solution, by acquiring the first data indicating the driver's driving behavior, it is possible to fully understand the driving behavior of the candidates' drivers, thereby effectively identifying the driving behavior characteristics of the candidates' drivers. Then, based on the driving behavior characteristics of the drivers, the target collection vehicle is determined from the candidates. By positioning the data collection to a specific target collection vehicle, the efficiency and quality of data collection can be improved, which in turn helps to improve the efficiency of subsequent data utilization and optimize the effect of data processing.

[0010] In conjunction with the first aspect, in some possible implementations, determining the target collection vehicle from the candidate vehicles based on the first data includes: obtaining an evaluation result for the first candidate vehicle based on the data corresponding to the first candidate vehicle in the first data, wherein the data corresponding to the first candidate vehicle is used to indicate the driving behavior of the driver of the first candidate vehicle, the evaluation result of the first candidate vehicle is determined based on the driving behavior of the driver of the first candidate vehicle, and the first candidate vehicle is any vehicle among the candidate vehicles; and determining the target collection vehicle based on the data collection requirements and the evaluation result of each candidate vehicle among the candidate vehicles.

[0011] Based on the above technical solution, the driving behavior of the driver of each candidate vehicle is evaluated based on the data generated by the vehicle itself. The quality of the driver-driver data of the corresponding candidate vehicle can be judged based on the driver's driving behavior evaluation results. The target vehicle for collecting driver-driver data that meets the data collection requirements is selected from the candidate vehicles, which helps to reduce the proportion of low-quality driver-driver data collected and ensure the quality of the collected driver-driver data.

[0012] In conjunction with the first aspect, in some possible implementations, the target acquisition vehicle is determined based on data acquisition requirements and the evaluation results of each candidate vehicle among the candidate vehicles. This includes: determining a target evaluation index based on a screening ratio and the evaluation results of each candidate vehicle among the candidate vehicles, where the screening ratio is the ratio of the target acquisition quantity to the number of candidate vehicles, and the data acquisition requirements include the target acquisition quantity, which is determined based on the amount of data required to determine the vehicle's driving strategy; and determining the target acquisition vehicle based on the target evaluation index, where the target acquisition vehicle is the vehicle among the candidate vehicles that meets the target evaluation index.

[0013] Based on the above technical solution, the number of target data collection vehicles determined in this way may be greater than or equal to the target data collection quantity. If the number of target data collection vehicles exceeds the target data collection quantity, vehicles with the same score as the vehicle corresponding to the screening ratio are also included in the target data collection vehicle range. This avoids wasting data from these vehicles, thus preventing data resource waste.

[0014] In conjunction with the first aspect, in some possible implementations, the target collection vehicle is determined based on the data collection requirements and the evaluation results of each candidate vehicle among the candidate vehicles, including: determining the target collection vehicle based on the target collection quantity and the evaluation results of each candidate vehicle among the candidate vehicles, wherein the number of target collection vehicles is equal to the target collection quantity; wherein the data collection requirements include the target collection quantity, which is determined based on the amount of data required to determine the vehicle's driving strategy.

[0015] Based on the above technical solution, the human and driver data obtained from the target collection vehicle, determined according to the target collection quantity, can meet the data volume required to determine the vehicle's driving strategy.

[0016] In conjunction with the first aspect, in some possible implementations, the evaluation result of the first candidate vehicle is obtained based on the data corresponding to the first candidate vehicle in the first data, including: based on the data corresponding to the first candidate vehicle in the first data, combined with one or more of the following reference indicators: the driver's driving experience, the driver's proficiency in vehicle operation, or driving safety, the evaluation result of the first candidate vehicle is obtained, and the one or more reference indicators are used to evaluate the driving behavior of the driver of the first candidate vehicle.

[0017] Based on the aforementioned technical solution, by combining reference indicators such as the driver's driving experience, proficiency in vehicle operation, and driving safety, the first candidate vehicle can be evaluated, thus more comprehensively reflecting the characteristics of the driver's driving behavior. This multi-dimensional evaluation method helps improve the accuracy and reliability of the evaluation results, thereby providing a more valuable reference for the selection of target data collection vehicles.

[0018] In conjunction with the first aspect, in some possible implementations, each of the one or more reference indicators includes one or more types of events, and each type of event corresponds to a weight; the evaluation result of the first candidate vehicle is determined based on the weight corresponding to each type of event in the one or more types of events, and the number of occurrences of each type of event in the one or more types of events.

[0019] Based on the above technical solution, by assigning weights to events in each reference indicator and determining the evaluation result based on the frequency of occurrence of the events, a detailed analysis of driving behavior can be achieved. This approach quantifies the impact of different events on the overall evaluation result, making the evaluation process more accurate, objective, and reasonable.

[0020] In conjunction with the first aspect, in some possible implementations, the evaluation result is in the form of a score, with the target evaluation index being greater than or equal to the score line, or less than or equal to the score line.

[0021] Based on the above technical solution, the evaluation results are presented in the form of scores, and score thresholds are set for the target evaluation indicators, making the evaluation results more intuitive and easier to understand. This approach facilitates quick determination of whether candidate vehicles meet the expected standards, thereby simplifying the decision-making process and improving evaluation efficiency.

[0022] In conjunction with the first aspect, in some possible implementations, the target vehicle is the candidate vehicle whose score is greater than or equal to the score threshold. The score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories.

[0023] Based on the above technical solution, when the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the frequency of occurrence of each event in the one or more event categories, the higher the score of the first candidate vehicle, the greater the probability that the first candidate vehicle will be selected as the target data collection vehicle. By selecting candidate vehicles with scores greater than or equal to the score threshold as target data collection vehicles, it can be ensured that the selected candidate vehicles meet higher evaluation criteria.

[0024] In conjunction with the first aspect, in some possible implementations, the target vehicle is the candidate vehicle whose score is less than or equal to the score threshold. The score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories.

[0025] Based on the above technical solution, when the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event categories and the frequency of occurrence of each event in the one or more event categories, the lower the score of the first candidate vehicle, the greater the probability that the first candidate vehicle will be selected as the target data collection vehicle. By selecting candidate vehicles with scores less than or equal to the score threshold as target data collection vehicles, it can be ensured that the selected candidate vehicles meet specific evaluation criteria and may be more suitable for specific data collection tasks or conditions.

[0026] In conjunction with the first aspect, in some possible implementations, the method further includes: filtering the driver data, removing data that does not meet quality requirements, and obtaining the target driver data.

[0027] Based on the above technical solution, further screening of the acquired driver and passenger data from the target vehicles can eliminate some data that does not meet quality requirements, thereby further reducing the proportion of low-quality driver and passenger data. This process helps ensure the high reliability and accuracy of subsequent analysis and processing, improving the overall efficiency and effectiveness of data processing.

[0028] In conjunction with the first aspect, in some possible implementations, the data that does not meet quality requirements includes one or more of the following: data that does not meet driving safety requirements; or, data that does not meet driving comfort requirements; or, data that does not meet traffic regulations.

[0029] Based on the above technical solutions, by identifying and eliminating data that does not meet driving safety, driving comfort, or traffic rules, it can be ensured that the remaining data is more in line with actual application needs, which helps to improve the overall quality of the dataset and makes subsequent data analysis and utilization more effective and targeted.

[0030] In conjunction with the first aspect, in some possible implementations, the method further includes: classifying the target driver data to obtain different categories of target driver data.

[0031] Based on the above technical solutions, classifying target driver data allows for the orderly management and processing of different types of data, which helps to better understand and analyze the data characteristics under different driving scenarios, thereby improving data utilization efficiency and providing more accurate support for applications in specific scenarios.

[0032] In conjunction with the first aspect, in some possible implementations, this category includes one or more of the following: driver's operational behavior, interaction scenarios with other vehicles, interaction scenarios with traffic signs and signals, or scene characteristics of the vehicle's environment.

[0033] Based on the above technical solution, by classifying the target driver data into features such as driver operation behavior, other vehicle interaction scenarios, traffic sign interaction scenarios, or vehicle environment scenarios, the characteristics of each category of data can be analyzed in more detail, which facilitates the subsequent more reasonable and targeted use of the target driver data.

[0034] In conjunction with the first aspect, in some possible implementations, the method further includes: labeling the data in each unit time period of the target driver-passenger data based on one or more of the following: vehicle motion state information, road information, navigation information, speed limit information, or vehicle planning and behavior information.

[0035] Based on the above technical solutions, by annotating the data in each unit of time in the target driver data based on vehicle motion status information, road information, navigation information, speed limit information, or vehicle planning and behavior information, richer contextual information can be provided for the data, which helps to enhance the interpretability and usability of the data and facilitates the more reasonable and targeted use of the target driver data in the future.

[0036] Secondly, this application provides a data processing apparatus, which includes execution steps for performing the first aspect and any possible implementation thereof. The apparatus includes corresponding modules for performing the above-described methods. The modules included in the apparatus can be implemented in software and / or hardware.

[0037] Thirdly, this application provides a data processing apparatus including a processor. The processor is coupled to a memory and can be used to execute a program in the memory to implement the execution steps in the first aspect and any possible implementation thereof.

[0038] Optionally, the data processing device also includes a memory.

[0039] Optionally, the data processing device also includes a communication interface, to which the processor is coupled.

[0040] Fourthly, this application provides a chip system including at least one processor for supporting the implementation of the functions involved in the first aspect and any possible implementation of the first aspect, such as receiving or processing data and / or instruction information involved in the above methods.

[0041] In one possible design, the chip system also includes a memory for storing program instructions and data, which may be located within or outside the processor.

[0042] The chip system can consist of chips or include chips and other discrete components.

[0043] Fifthly, this application provides an intelligent driving device whose driving strategy is determined based on the driver data of the target vehicle obtained by the method in the first aspect and any possible implementation of the first aspect.

[0044] Sixthly, this application provides a computer-readable storage medium storing a program (also referred to as code or instructions) that, when run by a processor, causes the methods in the first aspect and any possible implementation thereof to be executed.

[0045] In a seventh aspect, this application provides a computer program product comprising: a computer program (also referred to as code or instructions) that, when run, causes the methods in the first aspect and any possible implementation thereof to be executed.

[0046] It should be understood that the second to seventh aspects of this application correspond to the technical solutions of the first aspect of this application, and the beneficial effects achieved by each aspect and the corresponding feasible implementation are similar, and will not be repeated here. Attached Figure Description

[0047] Figure 1 is a functional block diagram of an intelligent driving device provided in an embodiment of this application;

[0048] Figure 2 is a schematic diagram of the architecture of the intelligent driving system provided in an embodiment of this application;

[0049] Figure 3 is a schematic flowchart of the data processing method provided in an embodiment of this application;

[0050] Figure 4 is a schematic flowchart of determining the target collection vehicle based on the first data provided in an embodiment of this application;

[0051] Figure 5 is a schematic diagram of a scenario applicable to the method provided in the embodiments of this application;

[0052] Figure 6 is another schematic flowchart of the data processing method provided in the embodiments of this application;

[0053] Figure 7 is a schematic block diagram of the data processing apparatus provided in an embodiment of this application;

[0054] Figure 8 is another schematic block diagram of the data processing apparatus provided in the embodiments of this application. Detailed Implementation

[0055] The technical solutions in this application will now be described with reference to the accompanying drawings.

[0056] First, in this application, the terms “comprising” and “having” and any variations thereof are intended to cover non-exclusive inclusion, for example, an apparatus, system, product or device that includes a series of modules, units or units is not necessarily limited to those modules, units or units that are explicitly listed, but may include other modules, units or units that are not explicitly listed or that are inherent to such apparatus, system, product or device.

[0057] Second, in this application, the words "exemplarily" and "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design that is described as "exemplarily" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of words such as "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0058] Third, in this application, "when," "under the circumstances," "if," and "if" all refer to the device making a corresponding action under certain objective circumstances, and are not time-limited, nor do they require the device to make a judgment action when it is implemented, nor do they mean that there are other limitations.

[0059] Fourth, in this application, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect. For example, "first threshold," "second threshold," and "third threshold" are used to distinguish different thresholds and do not limit their order. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" do not necessarily imply that they are different.

[0060] Fifth, in this application, "preset" can be understood as predefined, defined, pre-defined, stored, pre-stored, pre-negotiated, or pre-configured, etc.

[0061] Sixth, in this application, "at least one (kind, type)" refers to one (kind, type) or multiple (kinds, types). "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone, where A and B can be singular or plural. The character " / " generally indicates an "or" relationship between the preceding and following related objects, but it does not exclude the possibility of indicating an "and" relationship. The specific meaning can be understood in conjunction with the context.

[0062] Seventh, in this application, information C is used to determine information D, including both determining information D based solely on information C and determining it based on information C and other information. Furthermore, information C can also be used to determine information D indirectly, for example, in the case where information D is determined based on information E, and information E is determined based on information C.

[0063] To facilitate understanding of the embodiments of this application, some technical terms or vocabulary involved in this application will be briefly explained below.

[0064] 1. Driver-Driver Data: This refers to data generated by human drivers while driving a vehicle. Driver-Driver Data may include, but is not limited to, the following:

[0065] 1) Vehicle dynamic data (i.e., vehicle motion state data): such as speed, acceleration, steering angle, braking force, etc. This data can help analyze the driver's handling behavior.

[0066] 2) Environmental perception data: such as data from sensors such as cameras, lidar, millimeter-wave radar, etc., used to understand the environment around the vehicle (e.g., may include but is not limited to lane lines, other vehicles, pedestrians, traffic signals, etc.), and / or external condition data (e.g., may include but is not limited to weather, road conditions, traffic conditions, etc., which can affect driving behavior).

[0067] 3) Driving behavior data: such as specific driving operations such as changing lanes, overtaking, stopping, and starting, as well as the timing and conditions of these operations.

[0068] 4) Location and navigation data: such as positioning data (e.g., GPS-based or BeiDou-based) and navigation route information, used to understand the vehicle's route and geographical location.

[0069] 2. Data-driven model: Based on massive amounts of data, the planning control model can learn good driving habits and directly output execution commands to control the vehicle to operate autonomously.

[0070] 3. Features and Labels: Features describe the attributes or variables of the data and serve as inputs for the model to learn and predict. Labels are the outputs or results of each training sample in supervised learning and are the values ​​the model needs to predict. Labels are usually known and are used to guide the model's training process.

[0071] 4. Unlabeled data and labeled data: Unlabeled data refers to data that does not contain labels, and can also be understood as features. Labeled data can include both features and labels, and the training data used to train the model can be labeled data.

[0072] 5. Data labeling: Labeling, also known as tagging, refers to the process of generating corresponding labels for unlabeled data (i.e., features).

[0073] 6. Data Balance (or Weighting): In model training, "data balance" generally refers to a training dataset where the number of data samples for each class is roughly equal. This is an important concept, especially in classification problems, because an imbalanced dataset can lead to a model bias towards the majority class, thus affecting model performance. Data balancing is one of the key steps in improving a model's generalization ability and accuracy, especially when dealing with imbalanced datasets.

[0074] For example, adjusting class weights is a common data balancing technique. Adjusting class weights refers to adjusting the weights of different classes in the loss function during model training so that the model pays more attention to the minority class.

[0075] 7. Vulnerable road users (VRUs): This refers to a group of road users who are more vulnerable to injury than motor vehicles in a traffic environment. VRUs can include pedestrians, cyclists, tricycle riders, e-bike riders, e-tricycle riders, motorcyclists, or other non-motorized vehicle users.

[0076] In the field of intelligent driving, human and driver data are widely used. Taking the training scenario of planning and control models as an example, planning and control are key parts to realize autonomous vehicle navigation and operation, and the training of planning and control models is one of the key links to improve the performance of intelligent driving systems. It usually requires a large amount of human and driver data for training, and the quality of human and driver data determines the accuracy and reliability of planning and control models.

[0077] In the currently known human-driver data collection schemes, the collected human-driver data scenarios are complex and the data volume is large, with a large amount of low-quality data. Using this data for model training and other operations is often inefficient and ineffective.

[0078] To address the aforementioned technical problems, this application proposes a data processing method. By acquiring first data indicating the driver's driving behavior, the driving behavior of candidates can be fully understood, thereby effectively identifying the driving behavior characteristics of the candidates. Based on the driver's driving behavior characteristics, the target collection vehicle is determined from the candidate vehicles. This method positions the data collection to a specific target collection vehicle, improving the efficiency and quality of data collection, which in turn helps to improve the efficiency of subsequent data utilization and optimize the data processing effect.

[0079] Before describing the data processing method provided in the embodiments of this application, the intelligent driving device and intelligent driving system applicable to the method provided in this application will be described below with reference to Figures 1 and 2.

[0080] Figure 1 is a functional block diagram of an intelligent driving device provided in an embodiment of this application.

[0081] As shown in Figure 1, the intelligent driving device 100 may include a perception system 120, a display device 130, and a computing platform 150. The perception system 120 may include several sensors for sensing information about the environment surrounding the intelligent driving device 100. For example, the perception system 120 may include a positioning system, which can be a global navigation satellite system (GNSS), such as the Global Positioning System (GPS) or the BeiDou system. Alternatively, the perception system 120 may also include one or more of the following: an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.

[0082] In the method provided in this application, the sensing system 120 includes several sensors that can be used to sense and acquire first data for instructing the driver’s driving behavior, and can also be used to acquire human-driving data.

[0083] The display device 130 in the cockpit of the intelligent driving equipment 100 can be divided into two categories. For example, the first category can be an in-vehicle display screen; the second category can be a projection display screen, such as a head-up display (HUD).

[0084] Among these, the in-vehicle display screen can be a physical display screen and is an important component of the in-vehicle infotainment system. Multiple displays can be installed in the cabin, such as digital instrument cluster displays and central control screens. In some possible implementations, one or more of the aforementioned in-vehicle displays can be human-machine interfaces (HMIs), for example, the central control screen can be an HMI.

[0085] Additionally, head-up displays (HUDs), also known as head-up display systems, are used to display driving information such as speed and navigation on a display device in front of the driver (e.g., on the windshield). This reduces driver eye movement time, avoids pupil dilation caused by eye shifts, and improves driving safety and comfort. HUDs can include, but are not limited to, combined head-up display (C-HUD) systems, windshield head-up display (W-HUD) systems, and augmented reality head-up display (AR-HUD) systems.

[0086] Some or all of the functions of the intelligent driving device 100 can be controlled by the computing platform 150. The computing platform 150 may include processors 151 to 15n (n is an integer greater than or equal to 1), and the processor may be a circuit with signal processing capabilities.

[0087] In one implementation, the processor can be a circuit capable of reading and executing instructions, such as a central processing unit (CPU), microprocessor, graphics processing unit (GPU) (which can be understood as a type of microprocessor), or digital signal processor (DSP). In another implementation, the processor can implement certain functions through the logical relationships of hardware circuits. These logical relationships are either fixed or reconfigurable. For example, the processor can be a hardware circuit implemented using an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as a field-programmable gate array (FPGA). In reconfigurable hardware circuits, the process of the processor loading a configuration document and configuring the hardware circuit can be understood as the processor loading instructions to implement the functions of some or all of the above units. Furthermore, the processor can also be a hardware circuit designed for artificial intelligence, which can be understood as a type of ASIC, such as a neural network processing unit (NPU), tensor processing unit (TPU), or deep learning processing unit (DPU).

[0088] In addition, the computing platform 150 may also include a memory that can be used to store instructions, and some or all of the processors 151 to 15n can call the instructions in the memory to perform the corresponding functions.

[0089] The computing platform 150 can control the operation of the intelligent driving system, which may include an advanced driving assistance system (ADAS) and an autonomous driving system (ADS). The intelligent driving system can utilize various sensors on the vehicle (including but not limited to: LiDAR, millimeter-wave radar, cameras, ultrasonic sensors, GPS, and inertial measurement units) to acquire information from the vehicle's surroundings, and analyze and process this information to achieve functions such as obstacle perception, target recognition, vehicle localization, path planning, and driver monitoring / alerts, thereby improving the safety, automation, and comfort of driving the vehicle.

[0090] At different levels of autonomous driving (or intelligent driving levels, ranging from L0 to L5, totaling six levels), intelligent driving systems can achieve different levels of automated driving assistance based on artificial intelligence algorithms and information acquired by multiple sensors. These levels of autonomous driving are based on the classification standards of the Society of Automotive Engineers (SAE). Specifically, L0 is no automation; L1 is driver assistance; L2 is partial automation; L3 is conditional automation; L4 is high automation; and L5 is full automation. At levels L1 to L3, the task of monitoring road conditions and reacting is jointly completed by the driver and the system, requiring the driver to take over dynamic driving tasks. Levels L4 and L5 allow the driver to completely transform into a passenger. Currently, the functions that intelligent driving systems can achieve include, but are not limited to: adaptive cruise control, automatic emergency braking, automatic parking, blind spot monitoring, forward cross-traffic alert / braking, rear cross-traffic alert / braking, forward collision warning, lane departure warning, lane keeping assist, rear collision warning, traffic sign recognition, traffic jam assist, and highway assist. It is understandable that the various functions mentioned above can have specific modes at different levels of autonomous driving (L0-L5), and the higher the level of autonomous driving, the more intelligent the corresponding mode.

[0091] For example, in the method provided in this application, the computing platform 150 can perform data preprocessing on the human-driving data on the intelligent driving device 100 obtained through the perception system 120, such as data cleaning or data transformation, and upload the human-driving data to the cloud after preliminary screening. This application does not limit this.

[0092] In addition, it is understood that the driving strategy of the intelligent driving device 100 can be determined based on the driver data of the target vehicle obtained by the data processing method provided in this application.

[0093] It is also understandable that the intelligent driving device 100 can be a device mounted on a vehicle, or it can be the vehicle itself.

[0094] Figure 2 is a schematic diagram of the architecture of the intelligent driving system provided in an embodiment of this application.

[0095] As shown in Figure 2, the system 200 may include a sensing module 210, a human-computer interaction module 220, a display module 230, and a control module 240.

[0096] The perception module 210 may include one or more camera devices or one or more radar sensors from the perception system 120 shown in FIG. 1, for collecting environmental information of the area where the vehicle is located, such as parking line information, obstacle information, etc. The perception module 210 may also process the collected environmental information to build a world model composed of roads, obstacles, etc. for downstream modules (such as human-machine interaction module 220, control module 240). The perception module 210 may send the information it collects and / or determines to the control module 240. For example, in the method provided in this application, the sensors included in the perception module 210 may be used to perceive and acquire first data for instructing the driver's driving behavior, and may also be used to acquire human-driving data.

[0097] The human-machine interaction module 220 may include one or more of the display devices 130 shown in FIG. 1, such as an HMI; the human-machine interaction module 220 may also include a sound-emitting device (such as a speaker, audio system, etc.) and a sound-receiving device (such as a microphone). The display module 230 may include one or more of the display devices 130 shown in FIG. 1, and the display module 230 can display the vehicle interface. The human-machine interaction module 220 can receive user commands (including voice commands, touch screen commands, etc.), and then control the changes of the interface displayed by the display module 230 according to the commands.

[0098] The following will describe in detail, with reference to the accompanying drawings, a data processing method provided in an embodiment of this application.

[0099] Figure 3 is a schematic flowchart of the data processing method provided in the embodiments of this application.

[0100] The method 300 shown in Figure 3 may include steps 310 to 330. The following provides a detailed description of each step in the method 300.

[0101] The steps of this method can be executed by a data processing device, or by a component (such as a chip, chip system, etc.) configured in the data processing device, or by a logic module or software capable of implementing all or part of the functions of the data processing device; this application does not limit this. By way of example and not limitation, the data processing device can be a device capable of interacting with the intelligent driving system, such as a server of the intelligent driving system cloud platform.

[0102] The following describes in detail each step of the method 300, taking the execution of the method 300 by a data processing device as an example.

[0103] In step 310, first data of the candidate vehicle is obtained, which is used to indicate the driving behavior of the driver of the candidate vehicle.

[0104] Candidate vehicles may include commercially produced vehicles equipped with intelligent driving systems (referred to as commercial vehicles). Candidate vehicles may also include vehicles specifically designed for data collection; for ease of description, vehicles specifically designed for data collection will be referred to as dedicated data collection vehicles in the following text.

[0105] The driving behavior of the drivers of the candidate vehicles can be used to screen drivers with specific driving behaviors or specific driving scenarios, good driving habits, or good driving skills. Specific driving behaviors may include, but are not limited to, acceleration, deceleration, lane changing, and overtaking. Similarly, specific driving scenarios may include, but are not limited to, acceleration, deceleration, lane changing, and overtaking scenarios.

[0106] Understandably, the primary data is historical data generated from candidate vehicles.

[0107] For example, the operating data of commercial vehicles equipped with intelligent driving systems can be viewed in the background on the intelligent driving system cloud platform. Therefore, the data processing device can obtain the first data of candidate vehicles from the intelligent driving system cloud platform.

[0108] For example, the first data is the historical driving data generated by the candidate vehicle, which may include, but is not limited to, the candidate vehicle's mileage, vehicle motion status data, or driving behavior data.

[0109] In step 320, the target collection vehicle is determined from the candidate vehicles based on the first data.

[0110] For example, after acquiring the first data of the candidate vehicles, the data processing device can evaluate the driving behavior of the driver of the candidate vehicle based on the first data, and then determine the target collection vehicle from the candidate vehicles based on the evaluation results of the driving behavior of the driver of the candidate vehicle.

[0111] In step 330, the driver data of the target vehicle is acquired.

[0112] The driver-driver data here refers to the data generated by the driver while driving the vehicle, and this data is used to determine the vehicle's driving strategy. It's important to understand that this driver-driver data differs from the first data in step 310 above. The first data is historical driver-driver data generated by the candidate vehicle, and its magnitude is relatively small, insufficient for determining the vehicle's driving strategy. The driver-driver data here refers to future driver-driver data generated on the target vehicle, and its magnitude is greater than the first data; that is, the magnitude of this driver-driver data is sufficient to determine the vehicle's driving strategy.

[0113] For example, after identifying the target vehicles for data collection, the cloud platform of the intelligent driving system can be used to configure the corresponding settings for human-driver data collection on the intelligent driving systems of these target vehicles. This includes setting the upload duration and / or upload format of data from various sensors on the target vehicles (including but not limited to cameras, LiDAR, or millimeter-wave radar), so that during subsequent driving, the sensors on these target vehicles can upload data according to the corresponding settings. Accordingly, the data processing device can then acquire the data collected by the sensors on each target vehicle (i.e., human-driver data).

[0114] As an example, and not a limitation, the duration of data segments uploaded by commercially produced vehicles and dedicated collection vehicles can be set separately for these target data collection vehicles. The duration of data segments uploaded by commercially produced vehicles can differ from that uploaded by dedicated collection vehicles. For example, the duration of data segments uploaded by commercially produced vehicles can be shorter than that uploaded by dedicated collection vehicles. Or, for example, the duration of data segments uploaded by commercially produced vehicles can be in the seconds range, such as 30 seconds, while the duration of data segments uploaded by dedicated collection vehicles can be in the minutes range, such as 5 minutes. Of course, in practical applications, the duration of data segments uploaded by dedicated collection vehicles can also be the same as that uploaded by commercially produced vehicles; for example, both can be in the seconds range.

[0115] As an example and not a limitation, the driver data collected from the vehicles at each target site can be uploaded to the public cloud platform of the intelligent driving system. This application does not limit this.

[0116] Figure 4 is a schematic flowchart of determining the target collection vehicle based on the first data provided in an embodiment of this application.

[0117] In one possible implementation, as shown in Figure 4, step 320 of the above method 300, namely, determining the target collection vehicle from the candidate vehicles based on the first data, includes: step 3201, obtaining the evaluation result of the first candidate vehicle based on the data corresponding to the first candidate vehicle in the first data, wherein the first candidate vehicle is any vehicle among the candidate vehicles; step 3202, determining the target collection vehicle based on the evaluation result of each candidate vehicle among the candidate vehicles.

[0118] The data corresponding to the first candidate vehicle is used to indicate the driving behavior of the driver of the first candidate vehicle, and the evaluation result of the first candidate vehicle is determined based on the driving behavior of the driver of the first candidate vehicle.

[0119] Understandably, the first data is a set of data corresponding to each of the candidate vehicles, used to indicate the driving behavior of the driver of that vehicle. Taking the first candidate vehicle as an example, the data processing device can evaluate the driving behavior of the driver of the first candidate vehicle based on the data corresponding to that first candidate vehicle in the first data, and obtain the evaluation result of the first candidate vehicle. Without loss of generality, the data processing device can evaluate the driving behavior of the driver of the corresponding candidate vehicle based on the data corresponding to each candidate vehicle in the first data, and obtain the evaluation result of the corresponding candidate vehicle. In this way, the evaluation result of each of the candidate vehicles is obtained.

[0120] It is also understood that the evaluation results can be for assessing the vehicle profile or for assessing other specific driving scenarios, such as acceleration scenarios, lane changing scenarios, etc. This application does not limit this.

[0121] Furthermore, the data processing device can select the target collection vehicle from the candidate vehicles based on the evaluation results of each candidate vehicle, and then obtain human and driving data from the target collection vehicle.

[0122] Figure 5 is a schematic diagram of a scenario applicable to the method provided in the embodiments of this application.

[0123] As an example and not a limitation, as shown in Figure 5, the data processing method provided in this application embodiment can be applied to the data acquisition and data processing stages before training the planning and control model. Data-driven models require massive amounts of training data to support the training needs of the planning and control model, making the acquisition of massive amounts of human and driver data crucial for data production. Commercially produced vehicles equipped with intelligent driving systems are a suitable data source. These vehicles generate a large amount of human and driver data daily on the roads. Considering that the data provided by commercially produced vehicles far exceeds the needs of model training data, commercial vehicles can be selected.

[0124] As shown in stage 1 of Figure 5, the data processing device can evaluate the driver's driving behavior and use the evaluation results of the driving behavior of drivers of commercially produced vehicles to select potential target vehicles for collecting human-driving data that meet the data collection requirements from candidate vehicles (such as commercially produced vehicle A, commercially produced vehicle B, and commercially produced vehicle C shown in Figure 5).

[0125] As shown in Figure 5, based on the evaluation results of each candidate vehicle, target vehicles A and C are finally selected from candidate vehicles A, B, and C. These are commercial mass-produced vehicles A and C, which correspond to the smiley face icon. This indicates that commercial mass-produced vehicles A and C are potential target vehicles for collecting human-driving data that meet the data collection requirements. On the other hand, commercial mass-produced vehicle B, which corresponds to the sad face icon, is a non-target vehicle for collecting human-driving data that does not meet the data collection requirements.

[0126] Furthermore, as an example rather than a limitation, as shown in Figure 5, a driver's driving behavior can also be specifically represented by a vehicle profile, which can include multiple dimensions of reference indicators. A detailed description of these reference indicators can be found in the relevant descriptions below; for the sake of brevity, it will not be repeated here.

[0127] Once the target vehicle for data collection is identified, driver and passenger data can be obtained from it.

[0128] As shown in Stage 2 of Figure 5, after acquiring the driver and passenger data, the data can be filtered to obtain target driver and passenger data that meets the data collection requirements. This data is then used to train the planning and control model. For example, a filtering algorithm can be used to filter the driver and passenger data to obtain target driver and passenger data that meets the data collection requirements. A detailed description of the filtering algorithm can be found in the relevant description below, which will not be repeated here for the sake of brevity.

[0129] As an example and not a limitation, the driver data collected from the target vehicle can be uploaded to the cloud platform of the intelligent driving system (e.g., a public cloud), and further filtering of the driver data can also be completed on the cloud platform of the intelligent driving system (e.g., a private cloud). This application does not impose any limitations in this regard.

[0130] By evaluating the driving behavior of the driver of each candidate vehicle based on the data generated by the vehicle itself, the quality of the driver-driver data of the corresponding candidate vehicle can be judged based on the evaluation results. Selecting the target vehicle for collecting driver-driver data that meets the data collection requirements from the candidate vehicles helps to reduce the proportion of low-quality driver-driver data collected and ensure the quality of the collected driver-driver data.

[0131] Optionally, the evaluation result of the first candidate vehicle is obtained based on the data corresponding to the first candidate vehicle in the first data, including: based on the data corresponding to the first candidate vehicle in the first data, combined with one or more of the following reference indicators: the driver's driving experience, the driver's proficiency in vehicle operation or driving safety, the evaluation result of the first candidate vehicle is obtained, and the one or more reference indicators are used to evaluate the driving behavior of the driver of the first candidate vehicle.

[0132] For example, a driver's driving experience can be assessed by the vehicle's mileage (i.e., driver-driven mileage). The greater the mileage, the more driving experience the driver has with that vehicle; conversely, the smaller the mileage, the less driving experience the driver has.

[0133] For example, a driver's proficiency in vehicle control can be assessed by the number of times the driver deviates from the lane while driving. The more times the driver deviates from the lane, the less experience they have with that vehicle. Conversely, the fewer times the driver deviates from the lane, the more experience they have with that vehicle.

[0134] For example, driving safety can be assessed by the collision risk that a driver faces while driving the vehicle. The higher the collision risk, the less experience the driver has with the vehicle. Conversely, the lower the collision risk, the more experience the driver has with the vehicle.

[0135] By combining factors such as driver experience, vehicle handling proficiency, and driving safety into the evaluation of candidate vehicles, a more comprehensive reflection of driver behavior can be achieved. This multi-dimensional evaluation approach helps improve the accuracy and reliability of the evaluation results, thus providing a more valuable reference for the selection of target vehicles.

[0136] In one possible implementation, each of the one or more reference metrics includes one or more categories of events.

[0137] For example, the indicator of a driver's proficiency in vehicle control can include, but is not limited to, emergency lane keeping assist (ELKA) events. It is understood that ELKA events can also be referred to as lane departure events.

[0138] For example, driving safety can include, but is not limited to, event data recorder (EDR), collision events, automatic emergency braking (AEB) events, rear autonomous emergency braking (RAEB) events, rear crossing traffic brake (RCTB) events, front crossing traffic brake (FCTB) events, or emergency braking events. It can be understood that the aforementioned EDR, minor collision events, AEB events, RAEB events, RCTB events, FCTB events, or high-threshold emergency braking events can be collectively referred to as collision risk events.

[0139] Optionally, each type of event in the one or more types of events corresponds to a weight; the evaluation result of the first candidate vehicle is determined based on the weight corresponding to each type of event in the one or more types of events, and the number of occurrences of each type of event in the one or more types of events.

[0140] Table 1

[0141] For example, as shown in Table 1 above, each of the candidate vehicles may experience one or more of the aforementioned types of events, in addition to mileage or driver-driven mileage. Each type of event can correspond to a weight; a higher weight indicates a greater safety risk associated with that type of event, meaning such events are less desirable. If, within a unit mileage, events with higher weights occur more frequently, it can be considered that the driver's driving behavior is less in line with the data collection requirements for driver-driven data. In other words, the evaluation results of events other than mileage in the aforementioned one or more types of events are negatively correlated with the evaluation results of the first candidate vehicle.

[0142] Each candidate vehicle receives an evaluation result corresponding to each of the above one or more event categories.

[0143] For example, apart from mileage, the evaluation result for each type of event can be based on the number of times that type of event occurs per unit mileage and the corresponding weight of that type of event.

[0144] For example, the evaluation result for each type of event satisfies: Gs = Cs × Ns, where Gs can represent the score of the s-th type of event, Cs can represent the weight of the s-th type of event among the multiple types of events, and Ns can represent the number of times the s-th type of event occurs per unit mileage. For example, the score of the ELKA event is G2 = 5 × N2; another example is the score of the AEB event, G5 = 15 × N5.

[0145] Understandably, for mileage, the higher the mileage of a vehicle, the more experienced the driver is. In other words, the mileage assessment result of the first candidate vehicle is positively correlated with the overall assessment result of the first candidate vehicle.

[0146] It is also understandable that if the evaluation result of events other than mileage in one or more of the above-mentioned categories of events is negative, the evaluation result of mileage can be positive. Without loss of generality, if the evaluation result of events other than mileage in one or more of the above-mentioned categories of events is positive, the evaluation result of mileage can be negative.

[0147] Taking a negative mileage assessment result as an example, the mileage assessment methods for candidate vehicles can include, but are not limited to, any of the following methods:

[0148] Method 1: Increase the evaluation value by a certain amount every 100 kilometers. For example, increase by -1 for every 100 kilometers. That is, if the first candidate vehicle has driven 210 kilometers, increase by -2 (from...). get).

[0149] Method 2: Different mileage ranges correspond to different evaluation values. For example, the evaluation value for [0 km, 500 km) is -2, the evaluation value for [500 km, 1000 km) is -5, the evaluation value for [1000 km, 1500 km) is -8, the evaluation value for [1500 km, 2000 km) is -10, and so on.

[0150] It is understandable that the evaluation result of the first candidate vehicle can be determined based on the evaluation results corresponding to each type of event that occurs to the first candidate vehicle.

[0151] In other words, after determining the evaluation result of each of the one or more types of events that occurred to the first candidate vehicle, the data processing device can determine the evaluation result of the first candidate vehicle based on the evaluation result of each of the one or more types of events that occurred to the first candidate vehicle.

[0152] By assigning weights to events in each reference indicator and determining the evaluation result based on the frequency of occurrence of each event, a detailed analysis of driving behavior can be achieved. This approach quantifies the impact of different events on the overall evaluation result, making the evaluation process more accurate, objective, and reasonable.

[0153] In one possible implementation, the evaluation result is in the form of a score.

[0154] Understandably, when a driver's driving behavior is reflected through vehicle profiling, the evaluation result can be in the form of a score. The evaluation result can also take other forms, such as a tiered system, like Tier A, Tier B, Tier C, etc., where a higher tier indicates a greater probability of meeting the data collection requirements.

[0155] Optionally, the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories.

[0156] In other words, apart from mileage, the higher the weight and frequency of each event in one or more categories of events that occur to the first candidate vehicle, the lower the score of that first candidate vehicle. Mileage can be set to a negative value.

[0157] As an example, not a limitation, taking Table 1 as an example, the score of the first candidate vehicle satisfies: G = total score - (G1 + G2 + G3 + ... + Gs) = (G1 + C2 × N2 + C3 × N3 + ... + Cs × Ns), where G represents the score of the first candidate vehicle, G1 can represent the score corresponding to the mileage of the first candidate vehicle, G2 can represent the score of the first candidate vehicle for the second type of event (e.g., the ELKA event), C2 can represent the weight corresponding to the second type of event, N2 represents the number of times the first candidate vehicle experiences the second type of event per unit mileage, G3 can represent the score of the first candidate vehicle for the third type of event (e.g., EDR), C3 can represent the weight corresponding to the third type of event, N3 represents the number of times the first candidate vehicle experiences the third type of event per unit mileage, Gs can represent the score of the first candidate vehicle for the s-th type of event, Cs can represent the weight corresponding to the s-th type of event, and Ns represents the number of times the first candidate vehicle experiences the s-th type of event per unit mileage. Where G1≤0, G2≥0, G3≥0, Gs≥0.

[0158] In another possible implementation, the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event classes and the number of occurrences of each event in the one or more event classes.

[0159] In other words, apart from mileage, the higher the weight and frequency of each type of abnormal driving event in one or more categories that occur to the first candidate vehicle, the higher the score of that first candidate vehicle. Mileage can be set to a negative value.

[0160] As an example, not a limitation, taking Table 1 as an example, the score of the first candidate vehicle satisfies: G = (G1 + G2 + G3 + ... + Gs) = G1 + C2 × N2 + C3 × N3 + ... + Cs × Ns, where G represents the score of the first candidate vehicle, G1 can represent the score corresponding to the mileage of the first candidate vehicle, G2 can represent the score of the first candidate vehicle for the second type of event (e.g., the ELKA event), C2 can represent the weight corresponding to the second type of event, N2 represents the number of times the first candidate vehicle experiences the second type of event per unit mileage, G3 can represent the score of the first candidate vehicle for the third type of event (e.g., EDR), C3 can represent the weight corresponding to the third type of event, N3 represents the number of times the first candidate vehicle experiences the third type of event per unit mileage, Gs can represent the score of the first candidate vehicle for the s-th type of event, Cs can represent the weight corresponding to the s-th type of event, and Ns represents the number of times the first candidate vehicle experiences the s-th type of event per unit mileage. Where G1≤0, G2≥0, G3≥0, Gs≥0.

[0161] Presenting the evaluation results in the form of scores and setting score thresholds for the target evaluation indicators makes the results more intuitive and easier to understand. This approach facilitates quick determination of whether candidate vehicles meet the expected standards, thereby simplifying the decision-making process and improving evaluation efficiency.

[0162] Optionally, the target data collection vehicle is determined based on the evaluation results of each candidate vehicle among the candidate vehicles, including: determining the target data collection vehicle based on the data collection requirements and the evaluation results of each candidate vehicle among the candidate vehicles.

[0163] It's understandable that data collection needs can be specific to a particular scenario or a specific data dimension. For example, for the acceleration strategy of a vehicle's intelligent driving system, data can be specifically collected for acceleration scenarios; similarly, for the lane-changing strategy of a vehicle's intelligent driving system, data can be specifically collected for lane-changing scenarios. It's also understandable that data collection needs can differ for different scenarios.

[0164] Optionally, the data collection requirement includes a target collection quantity, which is determined based on the amount of data required to determine the vehicle's driving strategy.

[0165] Without loss of generality, the number of targets collected can vary depending on the scenario.

[0166] In one possible implementation A, the target collection vehicle is determined based on the data collection requirements and the evaluation results of each candidate vehicle among the candidate vehicles. This includes: determining a target evaluation index based on a screening ratio and the evaluation results of each candidate vehicle among the candidate vehicles, where the screening ratio is the ratio of the target collection quantity to the number of candidate vehicles; and determining the target collection vehicle based on the target evaluation index, where the target collection vehicle is the vehicle among the candidate vehicles that meets the target evaluation index.

[0167] For example, the selection ratio can satisfy: k = n / m, where k represents the selection ratio, m represents the number of candidate vehicles, and n represents the required number of vehicles to be collected (i.e., the target number of vehicles to be collected). The required number of vehicles to be collected is determined based on the amount of data needed to determine the vehicle's driving strategy. For instance, in the training scenario of an intelligent driving planning and control model, this target number of vehicles to be collected is determined based on the amount of data needed to train the intelligent driving planning and control model.

[0168] For example, if m = 10, meaning there are 10 candidate vehicles, and n = 6 is determined based on the amount of data needed to determine the driving strategy of the vehicles, then k = n / m = 6 / 10 = 3 / 5.

[0169] Another example is that the filtering ratio can be preset. For instance, when the number of candidate vehicles is large, a filtering ratio can be directly set, and an appropriate amount of data can be obtained based on this preset filtering ratio. As an example and not a limitation, the preset filtering ratio can be 2 / 3, 3 / 5, 4 / 5, etc., and this application does not limit it.

[0170] In one possible implementation, if the evaluation result is in the form of a score, then the target evaluation index is greater than or equal to the score line, or less than or equal to the score line.

[0171] It is understandable that when the evaluation result is in the form of a score, the target evaluation indicator can be related to a score threshold, such as being greater than or equal to the score threshold, or less than or equal to the score threshold. When the evaluation result is not in the form of a score, the target evaluation indicator can be non-score-dimensional data; for example, if the evaluation result is in a tiered system, the target evaluation indicator can be tier B or above. This application does not impose any limitations on this.

[0172] After obtaining the score of each of the 10 candidate vehicles, the data processing device can sort the 10 candidate vehicles according to their scores.

[0173] Optionally, candidate vehicles can be sorted from highest to lowest score, and a cutoff score can be determined based on the selection ratio.

[0174] For example, if the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the frequency of occurrence of each event category, the data processing device can determine the score of the candidate vehicle ranked 6th (obtained by m×k=10×(3 / 5)) as the score line. In this case, the target vehicle is the candidate vehicle whose score is greater than or equal to the score line. That is, when the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the frequency of occurrence of each event category, the target vehicle is the candidate vehicle whose score is greater than or equal to the score line.

[0175] For example, if the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more categories of events and the frequency of occurrence of each event in the one or more categories of events, the data processing device can determine the score of the candidate vehicle ranked 4th (obtained by m×(1-k)=10×(1-3 / 5)) as the score line. In this case, the target vehicle is the candidate vehicle whose score is less than or equal to the score line. That is, when the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more categories of events and the frequency of occurrence of each event in the one or more categories of events, the target vehicle is the candidate vehicle whose score is less than or equal to the score line.

[0176] Optionally, candidate vehicles can be sorted from low to high based on their scores, and then a cutoff score can be determined according to the selection ratio.

[0177] For example, if the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the frequency of occurrence of each event category, the data processing device can determine the score of the candidate vehicle ranked 4th (obtained from m×(1-k)=10×(1-3 / 5)) as the cutoff score. In this case, the target vehicle is the candidate vehicle whose score is greater than or equal to the cutoff score. That is, when the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the frequency of occurrence of each event category, the target vehicle is the candidate vehicle whose score is greater than or equal to the target cutoff score.

[0178] For example, if the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more categories of events and the frequency of occurrence of each event in the one or more categories of events, the data processing device can determine the score of the candidate vehicle ranked 6th (obtained by m×k=10×(3 / 5)) as the score line. In this case, the target vehicle is the candidate vehicle whose score is less than or equal to the score line. That is, when the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more categories of events and the frequency of occurrence of each event in the one or more categories of events, the target vehicle is the candidate vehicle whose score is less than or equal to the score line.

[0179] Understandably, in real-world applications, the number of target data collection vehicles determined based on method A may be equal to or greater than the target data collection quantity.

[0180] Example 1: If the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the frequency of each event, and the candidate vehicles are sorted from highest to lowest based on their scores (e.g., the scores of 10 candidate vehicles are 100, 85, 85, 80, 80, 80, 75, 70, 70, 65), then the score line is the score of the candidate vehicle ranked 6th (obtained from m×k=10×(3 / 5)), i.e., the score line = 80. Therefore, the target vehicle is the candidate vehicle. If the selected vehicles have a score greater than or equal to 80, that is, the top 6 vehicles, then the number of vehicles to be collected = the target data volume = 6. For example, if the scores are 100, 85, 85, 80, 80, 80, 80, 70, 70, 65, then the score line is the score of the candidate vehicle ranked 6th (obtained by m×k = 10×(3 / 5)), that is, the score line = 80. Then the target vehicles to be collected are the candidates with a score greater than or equal to 80, that is, the top 7 vehicles. In this case, the number of vehicles to be collected = 7 > the target data volume = 6.

[0181] Example 2: When the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more categories of events and the frequency of each event in the one or more categories of events, and the candidate vehicles are sorted from high to low based on their scores, for example, if the scores of 10 candidate vehicles are sorted from high to low as 35, 30, 30, 25, 20, 20, 20, 15, 15, 0, then the score line is the score of the candidate vehicle ranked 4th (obtained by m×k=10×(1-3 / 5)), that is, the target score line = 25. Then the target vehicles to be collected are the vehicles with scores less than or equal to 25 among the candidate vehicles, that is, the vehicles ranked in the last 7 positions. In this case, the number of target vehicles to be collected = 7 > the target data volume = 6.

[0182] Example 3: When the score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the frequency of each event in the one or more event categories, and the candidate vehicles are sorted from low to high based on their scores, for example, if the scores of 10 candidate vehicles are sorted from low to high as 65, 70, 70, 75, 80, 80, 80, 85, 85, 100, then the score line is the score of the candidate vehicle ranked 4th (obtained by m×k=10×(1-3 / 5)), that is, the target score line = 75. Then the target vehicles to be collected are the vehicles with a score greater than or equal to 75 among the candidate vehicles, that is, the vehicles ranked in the last 7 positions. In this case, the number of target vehicles to be collected = 7 > the target data volume = 6.

[0183] Example 4: If the score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event categories and the frequency of each event in the one or more event categories, and the candidate vehicles are sorted from low to high based on their scores (e.g., the scores of 10 candidate vehicles are sorted from high to low as 0, 15, 15, 20, 20, 20, 25, 30, 30, 35), then the score line is the score of the candidate vehicle ranked 6th (obtained from m×k=10×(3 / 5)), i.e., the score line = 20. Therefore, the target vehicle is the candidate vehicle with a score less than or equal to... For example, if the scores of 10 candidate vehicles are ranked from highest to lowest as 0, 15, 15, 20, 20, 20, 20, 30, 30, 35, then the score line is the score of the candidate vehicle ranked 6th (obtained by m×k=10×(3 / 5)), which is the target score line = 20. Therefore, the target vehicles to be collected are those with scores less than or equal to 20, which are the vehicles ranked from highest to lowest. In this case, the number of vehicles to be collected = 7 > the target data volume = 6.

[0184] It is also understandable that, based on method A, if the number of target collection vehicles is greater than the target collection quantity, then vehicles that are ranked after the vehicles corresponding to the screening ratio and have the same score as the selected vehicle are also included in the range of target collection vehicles. In this way, the data on these vehicles can be avoided, and thus the waste of resources can be avoided.

[0185] In one possible implementation B, the target collection vehicle is determined based on the data collection requirements and the evaluation results of each candidate vehicle among the candidate vehicles, including: determining the target collection vehicle based on the target collection quantity and the evaluation results of each candidate vehicle among the candidate vehicles, wherein the number of the target collection vehicles is equal to the target collection quantity.

[0186] Understandably, after obtaining the evaluation results (e.g., scores) of each candidate vehicle, these candidate vehicles can be ranked based on the evaluation results of each candidate vehicle.

[0187] Example 5: The score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories. Based on the scores of the candidate vehicles, the candidate vehicles are sorted from high to low. If the target data volume is n, then the target vehicle to be collected is the vehicle ranked in the top n among these candidate vehicles.

[0188] Example 6: The score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories. Based on the scores of the candidate vehicles, the candidate vehicles are sorted from low to high. If the target data volume is n, then the target vehicle to be collected is the vehicle ranked in the last n positions among these candidate vehicles.

[0189] Example 7: The score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories. Based on the scores of the candidate vehicles, the candidate vehicles are sorted from high to low. If the target data volume is n, then the target vehicle to be collected is the vehicle ranked in the last n positions among these candidate vehicles.

[0190] Example 8: The score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories. Based on the scores of the candidate vehicles, the candidate vehicles are sorted from low to high. If the target data volume is n, then the target vehicle to be collected is the vehicle ranked in the top n among these candidate vehicles.

[0191] The driver data obtained from the target vehicle determined by method B is sufficient to meet the data requirements for determining the vehicle's driving strategy.

[0192] In one possible implementation, the method further includes: filtering the driver data and removing data that does not meet quality requirements to obtain the target driver data.

[0193] Understandably, while the driving behavior of the drivers in the target data collection vehicles selected through steps 310 and 320 is generally good and meets the data collection requirements, the driver-driver data generated from these vehicles may still contain some data that does not meet quality standards. Therefore, after obtaining the driver-driver data from the target vehicles, this data can be further filtered to remove data that does not meet quality requirements. This further reduces the proportion of low-quality driver-driver data in the final dataset. This process helps ensure the high reliability and accuracy of the data in subsequent analysis and processing, improving the overall efficiency and effectiveness of data processing.

[0194] Optionally, data that does not meet quality requirements may include one or more of the following: data that does not meet driving safety requirements; or data that does not meet driving comfort requirements; or data that does not meet traffic regulations.

[0195] Understandably, when determining a vehicle's driving strategy, the first priority is to ensure that the vehicle complies with traffic rules, and then driving safety and even driving comfort can also be considered.

[0196] As examples, not limitations, failure to comply with traffic rules may include one or more of the following: improper parking (or illegal parking), illegal reversing, or running a red light; failure to meet driving safety requirements may include one or more of the following: speeding, exceeding lateral acceleration limits, exceeding longitudinal acceleration limits, or exceeding the rate of change of longitudinal acceleration limits; failure to meet driving comfort requirements may include one or more of the following: abrupt changes in driving trajectory, veergence (i.e., not following the planned route provided by navigation), etc.

[0197] In addition, considering the issue of commuting time, driving at a snail's pace when the driver's line of sight is unobstructed (which can be simply referred to as slow driving) can also be classified as data that does not meet quality requirements.

[0198] As an example rather than a limitation, the above-mentioned data that does not meet the quality requirements can be identified through a filtering algorithm.

[0199] For example, an abnormal parking filtering algorithm can be used to filter out data containing abnormal parking behavior. Abnormal parking behavior can include one or more of the following: parking in the middle of the road, parking in areas with no-parking signs, or parking on the sidewalk.

[0200] Data exhibiting abnormal parking behavior can be tagged as abnormal parking, and data with this abnormal parking tag can be identified as data that does not meet quality requirements.

[0201] For example, a violation reversing algorithm can be used to filter out data containing violations of reversing behavior. Violations of reversing behavior can include one or more of the following: reversing on a highway, reversing at an intersection, reversing on a one-way street, reversing on a pedestrian crossing, or reversing without using reversing lights or warning signals or reversing for an extended period of time.

[0202] Data showing illegal reversing behavior can be tagged as illegal reversing, and data with this tag can be identified as data that does not meet quality requirements.

[0203] For example, a speeding filtering algorithm can be used to filter out data that contains speeding behavior.

[0204] Speeding behavior may include one or more of the following: in a highway scenario, the vehicle speed is greater than the first threshold and is higher than the average speed of surrounding vehicles by a third threshold; in an urban scenario, the vehicle speed is greater than the second threshold and is higher than the average speed of surrounding vehicles by a third threshold; or in a non-highway and non-urban scenario, the vehicle speed is higher than the average speed of surrounding vehicles by a third threshold.

[0205] In cases where speeding is determined based on the average speed of surrounding vehicles, in practical applications, the data processing device can filter out all VRUs around the vehicle before determining whether the vehicle speed exceeds a third threshold, such as 20%, 30%, 50%, or 60%.

[0206] Data showing speeding behavior can be tagged with "speeding," and data with this "speeding" tag can be identified as data that does not meet quality requirements.

[0207] For example, in a high-speed scenario, the first threshold is 120 km / h and the third threshold is 20%. If a vehicle's speed is 130 km / h (greater than 120 km / h), and after filtering out all VRUs around the vehicle, the vehicle's speed is 20% higher than the average speed of the surrounding vehicles, then this segment of the vehicle's driver data will be labeled as speeding.

[0208] For example, a lateral acceleration exceeding limit screening algorithm can be used to filter out data exhibiting lateral acceleration exceeding the limit. Lateral acceleration exceeding the limit behavior can include, but is not limited to, behaviors where lateral acceleration exceeds a fourth threshold.

[0209] Data exhibiting excessive lateral acceleration can be tagged with "excessive lateral acceleration," and data with this tag can be identified as data that does not meet quality requirements.

[0210] For example, a longitudinal acceleration exceeding limit screening algorithm can be used to filter out data exhibiting longitudinal acceleration exceeding limits. Longitudinal acceleration exceeding limits can include, but is not limited to, behaviors where lateral acceleration exceeds the fifth threshold.

[0211] Data exhibiting longitudinal acceleration exceeding limits can be tagged with "Longitudinal Acceleration Exceeds Limits," and data with this tag can be identified as data that does not meet quality requirements.

[0212] For example, a filtering algorithm for exceeding the limit of the rate of change of longitudinal acceleration can be used to filter out data that exhibits behavior where the rate of change of longitudinal acceleration exceeds the limit. Such behavior may include, but is not limited to, behaviors where the rate of change of longitudinal acceleration exceeds a sixth threshold.

[0213] Data exhibiting excessive rate of change of longitudinal acceleration can be tagged with "excessive rate of change of longitudinal acceleration," and data with this tag can be identified as data that does not meet quality requirements.

[0214] For example, a trajectory jump filtering algorithm can be used to filter out data exhibiting trajectory jump behavior. Trajectory jump behavior may include, but is not limited to: the difference between the vehicle's actual position and its theoretical position at a certain moment is greater than a seventh threshold, where the theoretical position at that moment is determined based on the vehicle's motion information and vehicle kinematic formula from the previous moment, and the vehicle motion information includes one or more of the following: position, speed, acceleration, or direction of travel.

[0215] Data exhibiting trajectory abrupt changes can be tagged with "trajectory abrupt change," and data with this tag can be identified as data that does not meet quality requirements.

[0216] Considering the issue of commuting efficiency in intelligent driving scenarios, driving data exhibiting slow speeds despite relatively low traffic volume can be identified as failing to meet quality requirements. For example, a slow-driving filtering algorithm can be used to identify data exhibiting slow-driving behavior. Slow-driving behavior can include, but is not limited to: in highway scenarios where traffic volume is less than or equal to the eighth threshold, a speed less than the ninth threshold and lower than the twelfth threshold compared to the average speed of surrounding vehicles; or in urban scenarios where traffic volume is less than or equal to the tenth threshold, a speed less than the eleventh threshold and lower than the twelfth threshold compared to the average speed of surrounding vehicles; or in non-highway and non-urban scenarios, a speed lower than the twelfth threshold compared to the average speed of surrounding vehicles.

[0217] In practical applications, the data processing device can filter out all VRUs around the vehicle before determining whether the vehicle speed is lower than the average speed of the surrounding vehicles by a twelfth threshold, such as 20%, 30%, 50%, or 60%.

[0218] Data exhibiting slow-moving behavior can be labeled as slow-moving data, and data with this label can be identified as data that does not meet quality requirements.

[0219] For example, in a high-speed scenario where the traffic flow is less than or equal to the eighth threshold, the ninth threshold is 80 km / h, and the twelfth threshold is 20%, if a vehicle's speed is 70 km / h (less than 80 km / h), and after filtering out all VRUs around the vehicle, the vehicle's speed is less than 20% of the average speed of the surrounding vehicles, then this segment of the vehicle's driver data will be labeled as "driving at a snail's pace".

[0220] For example, a veer-off driving filtering algorithm can be used to filter out data exhibiting veer-off driving behavior. Veer-off driving behavior can include, but is not limited to, driving outside the planned route provided by the navigation system.

[0221] Data exhibiting yaw behavior can be tagged with "yaw behavior," and data with this tagged label can be identified as data that does not meet quality requirements.

[0222] For example, a red-light violation filtering algorithm can be used to filter out data containing red-light violations.

[0223] Data showing red-light running can be tagged as red-light running, and data with this tag can be identified as data that does not meet quality requirements.

[0224] It is understandable that the methods for filtering the driver data of the target vehicles may include, but are not limited to, the data labeling methods mentioned above.

[0225] It is also understandable that for a given data segment, if any of the above behaviors are present, the data will be determined as not meeting the quality requirements and will therefore not be selected as the target driver data.

[0226] In one possible implementation, these various filtering algorithms can run on the private cloud platform of the intelligent driving system. That is, after the target vehicle uploads driver data to the public cloud platform of the intelligent driving system, the server of the public cloud platform can transfer this driver data to the private cloud platform. The private cloud platform server then filters this driver data based on the aforementioned filtering algorithms to obtain target driver data that meets quality requirements.

[0227] By identifying and removing data that does not meet driving safety, driving comfort, or traffic rules, we can ensure that the remaining data better meets the needs of actual applications, which helps to improve the overall quality of the dataset and makes subsequent data analysis and utilization more effective and targeted.

[0228] Optionally, the method may include: classifying the target driver data to obtain different categories of target driver data.

[0229] After obtaining the target driver data that meets the quality requirements, the target driver data can be classified to obtain different types of target driver data.

[0230] Classifying target driver and passenger data allows for the orderly management and processing of different data types, which helps to better understand and analyze data characteristics under different driving scenarios, thereby improving data utilization efficiency and providing more accurate support for applications in specific scenarios.

[0231] In one possible implementation, the target driver data categories may include one or more of the following: driver's operational behavior, interaction scenarios with other vehicles, interaction scenarios with traffic signs and signals, or scene characteristics of the vehicle's environment.

[0232] Categories can be considered as segment-level classifications or segment-level data annotations of target driver and vehicle data. Each category can also include one or more subcategories. For example, driver actions may include, but are not limited to: efficient lane changing, turning at intersections, etc.; other vehicle interaction scenarios may include, but are not limited to: lane changing, traffic jams, etc.; traffic signal interaction scenarios may include, but are not limited to: traffic light interaction. Scene characteristics of the vehicle's environment may include, but are not limited to: intersections, ramps, or gaps, etc.

[0233] Understandably, a piece of data may be divided into multiple different categories.

[0234] By classifying target driver data into features such as driver behavior, interaction scenarios with other vehicles, traffic signal interaction scenarios, or the environment in which the vehicle is located, we can analyze the characteristics of each category of data in more detail, which will facilitate the more reasonable and targeted use of the target driver data in the future.

[0235] Optionally, the method further includes: labeling the data in each unit time period of the target driver data based on one or more of the following: vehicle motion status, road information, navigation information, speed limit information, or vehicle planning and behavior information.

[0236] It is understandable that, based on one or more of the following: vehicle motion status information, road information, navigation information, speed limit information, or vehicle planning and behavior information, data annotation is performed on each unit of time in the target driver data. This can be understood as unit-time level data annotation or frame-level data annotation.

[0237] The vehicle's motion status information may include one or more of the following: vehicle speed, lateral acceleration, longitudinal acceleration, or driving direction. Road information may include, but is not limited to, road curvature or road type. Navigation information may include, but is not limited to, driving direction or drivable candidate roads. Speed ​​limit information may include, but is not limited to, maximum speed limit or minimum speed limit. Vehicle planning and behavior information may include, but is not limited to, lateral movements (e.g., lateral obstacle avoidance behavior, lateral speed, or lateral acceleration) and longitudinal movements (e.g., longitudinal speed, or longitudinal acceleration).

[0238] By annotating the data within each unit of time in the target driver-passenger data based on vehicle motion status information, road information, navigation information, speed limit information, or vehicle planning and behavior information, richer contextual information can be provided to the data. This helps to enhance the interpretability and usability of the data, making it easier to use the target driver-passenger data more reasonably and in a more targeted manner in the future.

[0239] As an example, and not a limitation, after annotating the target driver and passenger data at the fragment and / or frame levels, the annotation results can be organized into a JavaScript object notation (JSON) format and stored in a data table along with the corresponding data fragments.

[0240] For example, in the training scenario of an intelligent driving planning and control model, a data balancing step can be performed subsequently. This involves reading the labeled results from the corresponding data table and balancing the JSON files of the labeled results according to the requirements of the planning and control model. In short, it involves frame extraction processing based on the file storage format of the planning and control model and the output JSON files of the labeled data, setting the data balance, and then providing it to the planning and control model as input data for training. As shown in Figure 5, after training the planning and control model, the vehicle profile can be optimized and updated in reverse based on the model's performance. That is, by combining the performance of the planning and control model, it can be determined in which aspects the planning and control model can be further improved, and then the reference indicators used to determine the vehicle profile can be optimized and updated.

[0241] To facilitate understanding of the data processing method provided in this application, the data processing method provided in this application will be explained again below with reference to Figure 6.

[0242] Figure 6 is another schematic flowchart of the data processing method provided in the embodiments of this application.

[0243] As shown in Figure 6, the data processing method provided in this application may include two major steps: screening the target vehicles for collecting human and driver data and screening high-quality human and driver data.

[0244] In Phase 1, the data processing device can evaluate candidate vehicles (which may include commercially produced vehicles and / or dedicated data collection vehicles) based on data used to indicate driver behavior, and determine the target data collection vehicle based on the evaluation results of each candidate vehicle. A detailed description can be found in the relevant description in Method 300 above; for brevity, it will not be repeated here.

[0245] After the target vehicle is identified, the data processing device can send collection (information) to the target vehicle. Correspondingly, the target vehicle collects and uploads the driver and passenger data, that is, the data processing device obtains the driver and passenger data from the target vehicle.

[0246] In stage 2, after acquiring the driver and passenger data of the target vehicle, the data processing device can filter the acquired data, that is, remove data that does not meet the quality requirements (i.e., low-quality data) to obtain the target driver and passenger data (i.e., high-quality or excellent data). A detailed description can be found in the relevant description in method 300 above; for the sake of brevity, it will not be repeated here.

[0247] It is understandable that before filtering the acquired driver data, data management, data mining, or data transformation can be performed on the driver data, that is, data preprocessing can be performed on the driver data. For example, it may include, but is not limited to, aligning the data acquired by different sensors in time, for example, by frame extraction or frame interpolation. This application does not limit this.

[0248] In addition, the data processing device can also perform data annotation on the target driver and passenger data, such as including fragment-level data annotation and / or frame-level data annotation. A detailed description can be found in the relevant description in method 300 above, and for the sake of brevity, it will not be repeated here.

[0249] Based on the above technical solution, firstly, by acquiring the first data indicating the driver's driving behavior, a comprehensive understanding of the driving behavior of candidates can be achieved, thereby effectively identifying the characteristics of the driving behavior of candidates. Then, based on these characteristics, the target vehicle for data collection is determined from the candidate vehicles. This targeted data collection improves the efficiency and quality of data acquisition, thus enhancing the efficiency of subsequent data utilization and optimizing data processing. Secondly, by evaluating the driver's driving behavior based on the first data, and utilizing the evaluation results of the candidates' drivers' driving behavior, the target vehicle for data collection that potentially meets quality requirements is selected. This helps reduce the proportion of low-quality driver data collected, ensuring the quality of the collected data and improving the efficiency and quality of data collection, which in turn aids in determining vehicle driving strategies. Furthermore, filtering the acquired driver data and removing data that does not meet quality requirements further reduces the proportion of low-quality driver data to be used. Finally, labeling the target driver data facilitates more rational and targeted use of this data in the future.

[0250] The data processing method provided in this application has been described in detail above with reference to the accompanying drawings. The data processing apparatus provided in the embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0251] It should be understood that the data processing apparatus shown in Figures 7 and 8 can be used to implement the functions of the data processing apparatus in the above method embodiments, and thus can also achieve the beneficial effects of the above method embodiments. In the embodiments of this application, the apparatus can be the apparatus in any of the method embodiments shown in Figure 3, or it can be a component (such as a chip, chip system, processor, etc.) configured in the data processing apparatus, or it can be a logic module or software capable of implementing some or all of the functions of the data processing apparatus.

[0252] Figure 7 is a schematic block diagram of the data processing apparatus provided in an embodiment of this application.

[0253] As shown in Figure 7, the data processing device 700 includes an acquisition module 710 and a processing module 720. This data processing device 700 can be used to implement the functions of the data processing device in any of the method embodiments shown in Figure 3.

[0254] For example, when the data processing device 700 is used to implement the function of the data processing device in the method embodiment shown in FIG3, the acquisition module 710 can be used to acquire first data of candidate vehicles, which is used to indicate the driving behavior of the driver of the candidate vehicle; the processing module 720 can be used to determine the target collection vehicle from the candidate vehicles based on the first data; the acquisition module 710 can also be used to acquire the driver data of the target collection vehicle, which is the data generated by the driver during the driving process, and the driver data is used to determine the driving strategy of the vehicle.

[0255] Optionally, the processing module 720 can be specifically used to obtain the evaluation result of the first candidate vehicle based on the data corresponding to the first candidate vehicle in the first data. The data corresponding to the first candidate vehicle is used to indicate the driving behavior of the driver of the first candidate vehicle. The evaluation result of the first candidate vehicle is determined based on the driving behavior of the driver of the first candidate vehicle. The first candidate vehicle is any vehicle among the candidate vehicles. Based on the evaluation result of each candidate vehicle among the candidate vehicles, the target collection vehicle is determined.

[0256] Optionally, the processing module 720 can be specifically used to determine a target evaluation index based on a screening ratio and the evaluation results of each candidate vehicle among the candidate vehicles. The screening ratio is the ratio of the target number of data to the number of candidate vehicles. The data collection requirement includes the target number of data to be collected, which is determined based on the amount of data required to determine the driving strategy of the vehicle. Based on the target evaluation index, the target data collection vehicle is determined, which is the vehicle among the candidate vehicles that meets the target evaluation index.

[0257] Optionally, the processing module 720 can be specifically used to determine the target collection vehicle based on the target collection quantity and the evaluation result of each candidate vehicle among the candidate vehicles, wherein the number of target collection vehicles is equal to the target collection quantity; wherein the data collection requirement includes the target collection quantity, which is determined based on the amount of data required to determine the vehicle's driving strategy.

[0258] Optionally, the processing module 720 can be specifically used to obtain an evaluation result of the first candidate vehicle based on the data corresponding to the first candidate vehicle in the first data, combined with one or more of the following reference indicators: the driver's driving experience, the driver's proficiency in vehicle operation, or driving safety. The one or more reference indicators are used to evaluate the driving behavior of the driver of the first candidate vehicle.

[0259] Optionally, each of the one or more reference indicators includes one or more types of events, and each type of event corresponds to a weight; the evaluation result of the first candidate vehicle is determined based on the weight corresponding to each type of event in the one or more types of events, and the number of occurrences of each type of event in the one or more types of events.

[0260] Optionally, the evaluation result is in the form of a score, and the target evaluation indicator is greater than or equal to the score line, or less than or equal to the score line.

[0261] Optionally, the target vehicle is a vehicle among the candidate vehicles whose score is greater than or equal to the score threshold. The score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories.

[0262] Optionally, the target vehicle is a vehicle among the candidate vehicles whose score is less than or equal to the score line. The score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories.

[0263] Optionally, the processing module 720 can also be used to filter the driver data, remove data that does not meet the quality requirements, and obtain the target driver data.

[0264] Optionally, the data that does not meet quality requirements may include one or more of the following: data that does not meet driving safety requirements; or, data that does not meet driving comfort requirements; or, data that does not meet traffic regulations.

[0265] Optionally, the processing module 720 can also be used to classify the target driver data to obtain different categories of target driver data.

[0266] Optionally, the category includes one or more of the following: driver's operating behavior, interaction scenarios with other vehicles, interaction scenarios with traffic signs and signals, or scene characteristics of the vehicle's environment.

[0267] Optionally, the processing module 720 can also be used to annotate the data in each unit of time in the target driver data based on one or more of the following: vehicle motion status information, road information, navigation information, speed limit information, or vehicle planning and behavior information.

[0268] For a more detailed description of each of the above modules, please refer directly to the relevant description in the method embodiment shown in Figure 3, which will not be repeated here.

[0269] It should be understood that the module division in the embodiments of this application is illustrative and only represents a logical functional division. In actual implementation, there may be other division methods. Furthermore, the functional modules in the various embodiments of this application can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0270] Figure 8 is another schematic block diagram of the data processing apparatus provided in the embodiments of this application.

[0271] The data processing device 800 can be a chip system, or it can be a device configured with a chip system to implement the methods described in the above method embodiments. In the embodiments of this application, the chip system can be composed of chips, or it can include chips and other discrete devices.

[0272] As shown in FIG8, the data processing device 800 may include a processor 810, which can be used to execute computer programs or instructions in memory to implement the steps performed by the relevant device in any embodiment of the method embodiment shown in FIG3.

[0273] Optionally, the data processing device 800 further includes a communication interface 820. The communication interface 820 can be used to communicate with other devices via a transmission medium, thereby enabling the data processing device 800 to communicate with other devices. The communication interface 820 can be, for example, a transceiver, interface, bus, circuit, or a device capable of transmitting and receiving functions. The processor 810 can use the communication interface 820 to input and output data and to implement the method described in any embodiment corresponding to FIG3. Specifically, the data processing device 800 can be used to implement the functions of the data processing device in the above method embodiments.

[0274] Optionally, the data processing apparatus 800 further includes at least one memory 830 for storing program instructions and / or data. The memory 830 is coupled to the processor 810. The coupling in this embodiment is an indirect coupling or communication connection between devices, units, or modules, and can be electrical, mechanical, or other forms, used for information exchange between devices, units, or modules. The processor 810 may operate in conjunction with the memory 830. The processor 810 may execute program instructions stored in the memory 830.

[0275] In this application, the memory 830 can be integrated into the processor 810, or the processor 810 and the memory 830 can be set up separately; this application does not limit this.

[0276] It should be understood that the coupling in the embodiments of this application is an indirect coupling or communication connection between devices, units, or modules, which can be electrical, mechanical, or other forms, used for information interaction between devices, units, or modules. The processor 810 may operate in conjunction with the memory 830. The specific connection medium between the processor 810, communication interface 820, and memory 830 is not limited in the embodiments of this application. In Figure 8, the processor 810, communication interface 820, and memory 830 are connected via a bus 840. The bus 840 is represented by a thick line in Figure 8. The connection methods between other components are only illustrative and not intended to be limiting. The bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used in Figure 8, but this does not indicate that there is only one bus or one type of bus.

[0277] This application also provides a computer program product, which includes a computer program (also referred to as code or instructions) that, when run, can implement the steps performed by the relevant devices in any of the embodiments shown in Figures 3 to 6.

[0278] This application also provides a computer-readable storage medium storing a computer program (also referred to as code or instructions). When the computer program is run, it can implement the steps performed by the relevant devices in any of the embodiments shown in FIG3 to FIG6.

[0279] This application provides a chip system including at least one processor for supporting the implementation of the steps performed by the relevant devices in any of the embodiments shown in FIG3 to FIG6.

[0280] In one possible design, the chip system also includes a memory for storing program instructions and data, which may be located within or outside the processor.

[0281] The chip system can consist of chips or include chips and other discrete components.

[0282] It should be understood that the processor in the embodiments of this application can be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method embodiments can be completed by the integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0283] It should also be understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), EEPROM, or flash memory. Volatile memory can be RAM, which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0284] The terms "unit," "module," etc., used in this specification can be used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. In the embodiments of this application, "unit" and "module" have the same meaning and can be used interchangeably.

[0285] Those skilled in the art will recognize that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application. In the several embodiments provided in this application, it should be understood that the disclosed apparatus, devices, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0286] 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0287] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0288] In the above embodiments, the functions of each functional unit can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions (programs). When the computer program instructions (programs) are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs, DVDs), or semiconductor media (e.g., solid-state disks, SSDs), etc.

[0289] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the technology, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0290] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A data processing method, characterized in that, The method includes: Acquire first data of the candidate vehicle, the first data being used to indicate the driving behavior of the driver of the candidate vehicle; Based on the first data, the target collection vehicle is determined from the candidate vehicles; The driver data of the target vehicle is acquired. The driver data is the data generated by the driver while driving the vehicle. The driver data is used to determine the vehicle's driving strategy.

2. The method according to claim 1, characterized in that, The step of determining the target collection vehicle from the candidate vehicles based on the first data includes: Based on the data corresponding to the first candidate vehicle in the first data, the evaluation result of the first candidate vehicle is obtained. The data corresponding to the first candidate vehicle is used to indicate the driving behavior of the driver of the first candidate vehicle. The evaluation result of the first candidate vehicle is determined based on the driving behavior of the driver of the first candidate vehicle. The first candidate vehicle is any vehicle among the candidate vehicles. Based on the data collection requirements and the evaluation results of each candidate vehicle among the candidate vehicles, the target collection vehicle is determined.

3. The method according to claim 2, characterized in that, The process of determining the target vehicle for data collection based on data acquisition requirements and the evaluation results of each candidate vehicle includes: Based on the screening ratio and the evaluation results of each candidate vehicle among the candidate vehicles, a target evaluation index is determined. The screening ratio is the ratio of the target number of data to the number of candidate vehicles. The data collection requirements include the target number of data to be collected, which is determined based on the amount of data required to determine the driving strategy of the vehicle. Based on the target evaluation index, the target data collection vehicle is determined, and the target data collection vehicle is the vehicle among the candidate vehicles that meets the target evaluation index.

4. The method according to claim 2, characterized in that, The process of determining the target vehicle for data collection based on data acquisition requirements and the evaluation results of each candidate vehicle includes: Based on the target number of vehicles to be collected and the evaluation results of each candidate vehicle among the candidate vehicles, the target vehicles to be collected are determined, and the number of target vehicles to be collected is equal to the target number of vehicles to be collected. The data collection requirements include the target collection quantity, which is determined based on the amount of data required to determine the vehicle's driving strategy.

5. The method according to any one of claims 2 to 4, characterized in that, The step of obtaining the evaluation result of the first candidate vehicle based on the data corresponding to the first candidate vehicle in the first data includes: Based on the data corresponding to the first candidate vehicle in the first data, and combined with one or more of the following reference indicators: driver's driving experience, driver's proficiency in vehicle operation or driving safety, the evaluation result of the first candidate vehicle is obtained. The one or more reference indicators are used to evaluate the driving behavior of the driver of the first candidate vehicle.

6. The method according to claim 5, characterized in that, Each of the one or more reference indicators includes one or more types of events, and each type of event corresponds to a weight; the evaluation result of the first candidate vehicle is determined based on the weight corresponding to each type of event in the one or more types of events, and the number of occurrences of each type of event in the one or more types of events.

7. The method according to claim 6, characterized in that, The evaluation results are in the form of scores, and the target evaluation index is greater than or equal to the score line, or less than or equal to the score line.

8. The method according to claim 7, characterized in that, The target vehicle is the candidate vehicle whose score is greater than or equal to the score threshold. The score of the first candidate vehicle is negatively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories.

9. The method according to claim 8, characterized in that, The target vehicle is the candidate vehicle whose score is less than or equal to the score line. The score of the first candidate vehicle is positively correlated with the weight of each event in the one or more event categories and the number of occurrences of each event in the one or more event categories.

10. The method according to any one of claims 1 to 9, characterized in that, The method further includes: The driver data is filtered to remove data that does not meet the quality requirements, thus obtaining the target driver data.

11. The method according to claim 10, characterized in that, The data that does not meet the quality requirements includes one or more of the following: Data that does not meet driving safety requirements; or, Data that does not meet driving comfort requirements; or, Data that does not comply with traffic rules.

12. The method according to claim 10 or 11, characterized in that, The method further includes: The target driver data is classified to obtain different categories of target driver data.

13. The method according to claim 12, characterized in that, The category includes one or more of the following: driver's operating behavior, interaction scenarios with other vehicles, interaction scenarios with traffic signs and signals, or scene characteristics of the vehicle's environment.

14. The method according to claim 12 or 13, characterized in that, The method further includes: Based on one or more of the following: vehicle motion status information, road information, navigation information, speed limit information, or vehicle planning and behavior information, data is labeled for each unit of time in the target driver data.

15. A data processing apparatus, characterized in that, The apparatus includes a module for performing the method as described in any one of claims 1 to 14.

16. A data processing apparatus, characterized in that, Including processor and memory, among which, The memory is used to store programs; The processor is used to invoke the program so that the device performs the method as described in any one of claims 1 to 14.

17. An intelligent driving device, characterized in that, The driving strategy of the intelligent driving device is determined based on the human driving data of the target vehicle obtained by the data processing method of any one of claims 1 to 14.

18. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed, the method as described in any one of claims 1 to 14 is performed.

19. A computer program product, characterized in that, Includes a computer program, which, when run, causes the method as described in any one of claims 1 to 14 to be performed.

Images