Vehicle control method, electronic device, and vehicle

Abstract

This application provides a vehicle control method, electronic device, and vehicle. The method includes: collecting multi-sensor data of the vehicle, including driver state data, vehicle state data, and location information, with the vehicle location information corresponding to the collection time; determining a fatigue score based on the driver's state data; searching a database based on the geographic area code corresponding to the vehicle's location information and the traffic time category corresponding to the collection time to determine the baseline style weights for various historical driving styles of the driver corresponding to the current driving scenario; determining the matching degree between the current driving scenario and the driving scenario represented by the baseline style weights based on the vehicle state data, fatigue score, and baseline style weights; determining the following parameters required by the vehicle's driver assistance system based on the matching degree; and controlling the vehicle's driving using the following parameters. Using the above method, driving safety and comfort can be ensured while taking into account the driver's personalized style.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and in particular to a vehicle control method, electronic device and vehicle. Background Technology

[0002] In the field of intelligent driving, vehicle following control is a crucial aspect of ensuring driving safety and a superior driving experience.

[0003] Car-following algorithms in related technologies are mainly divided into two categories: one is based on fixed rules, which controls following behavior by pre-setting a fixed distance or time interval; the other is a static driving style control algorithm, which determines the driving style by relying on preset parameters or manual selection by the driver. Both types of car-following algorithms generally suffer from problems such as insufficient safety redundancy, rigid control strategies, and poor adaptability, making it difficult to meet the comprehensive requirements of intelligent driving in terms of safety, comfort, and personalization. Summary of the Invention

[0004] This application provides a vehicle control method, electronic device, and vehicle, which can solve technical problems such as insufficient safety redundancy, rigid control strategy, and poor adaptability of vehicle following strategy.

[0005] On one hand, embodiments of this application provide a vehicle control method, the method comprising: collecting multi-sensor data of a vehicle, the multi-sensor data including the driver's state data, the vehicle's state data and location information, the vehicle's location information corresponding to the collection time; determining the driver's fatigue state score based on the driver's state data; searching a database based on the geographic area code corresponding to the vehicle's location information and the traffic time category corresponding to the collection time; determining the baseline style weights of multiple historical driving styles of the driver corresponding to the current driving scenario using a pre-trained classification model based on the vehicle's state data, the fatigue state score and the baseline style weights; adjusting the baseline style weights according to the matching degree to obtain a target style weight; determining the following parameters required by the vehicle's driving assistance system based on the target style weights; and controlling the vehicle's driving using the driving assistance system based on the following parameters.

[0006] In some embodiments of this application, the database stores multiple geographic area codes, each geographic area code corresponding to multiple traffic time periods, and each traffic time period category having a style weight corresponding to the multiple historical driving styles. The step of searching the database based on the geographic area code corresponding to the vehicle's location information and the traffic time period category corresponding to the collection time to determine the baseline style weight of the current driving scenario corresponding to the driver's multiple historical driving styles includes: if the geographic area code corresponding to the traffic time period category is found in the database, determining the baseline style weight of the current driving scenario corresponding to the multiple historical driving styles based on the style weight corresponding to the traffic time period category; if the geographic area code does not correspond to the traffic time period category in the database, determining the baseline style weight of the current driving scenario corresponding to the multiple historical driving styles based on the style weight corresponding to the traffic time period category corresponding to the geographic area code; and if the geographic area code is not found in the database, determining the baseline style weight of the current driving scenario corresponding to the multiple historical driving styles based on multiple preset weights.

[0007] In some embodiments of this application, the method further includes: updating the style weights corresponding to the target style weights in the database using a preset smoothing factor based on the target style weights.

[0008] In some embodiments of this application, the method further includes: obtaining from the database the historical fatigue state score of the driver corresponding to the historical driving scenario and multiple cluster centers for representing the various historical driving styles; the historical fatigue state score corresponds to the historical state data of the vehicle, the historical state data and the historical fatigue state score constitute a historical state feature vector, the multiple cluster centers are determined by clustering multiple historical state feature vectors, each cluster center represents a historical driving style, the historical driving style represented by any cluster center is determined according to the weighted sum of the features of any cluster center, and using the multiple cluster centers, the multiple historical state feature vectors and the historical baseline style weights of the historical driving scenario corresponding to the various historical driving styles, a preset classification network is trained to obtain the classification model.

[0009] In some embodiments of this application, the network layers of the classification network include multiple fully connected layers and a classification output layer; the training method of the classification model includes: concatenating the multiple historical state feature vectors, the multiple cluster centers, and the historical baseline style weights to obtain historical concatenated features; inputting the historical concatenated features into the classification network to obtain the data output by each network layer of the classification network, wherein the data output by the preceding network layer is used as the input data of the following and adjacent network layers; determining the data output by the classification output layer as the historical matching degree between the historical driving scene and the driving scene represented by the historical baseline style weights; calculating the loss value based on the historical matching degree and a preset matching degree label; adjusting the classification network until the loss value is within a preset numerical range; and determining the adjusted classification network as the classification model.

[0010] In some embodiments of this application, the multiple historical driving styles include an aggressive style, a standard style, and a conservative style. The step of adjusting the baseline style weight based on the matching degree to obtain the target style weight includes: if the matching degree is greater than or equal to a first preset threshold, determining the baseline style weight as the target style weight; if the matching degree is less than the first preset threshold but greater than or equal to a second preset threshold, decreasing the baseline style weight corresponding to the aggressive style and increasing the baseline style weight corresponding to the conservative style using a preset first downgrade factor; and updating the standard style based on the adjustment of the baseline style weights corresponding to the aggressive style and the conservative style. The baseline style weight corresponding to the radical style is used to obtain the target style weight. If the matching degree is less than the second preset threshold and greater than or equal to the third preset threshold, the baseline style weight corresponding to the radical style is reduced and the baseline style weight corresponding to the conservative style is increased using a preset second degradation factor. Based on the adjustment of the baseline style weights corresponding to the radical style and the conservative style, the baseline style weight corresponding to the standard style is updated to obtain the target style weight. If the first degradation factor is greater than the second degradation factor, and the matching degree is less than the third preset threshold, the baseline style weight is adjusted to a preset weight to obtain the target style weight.

[0011] In some embodiments of this application, determining the following parameters required by the vehicle's driver assistance system based on the target style weights includes: determining the following distance, lane change delay, and maximum speed of the vehicle based on the target style weights corresponding to the aggressive style and the conservative style; determining the throttle response coefficient of the vehicle based on the target style weights corresponding to the aggressive style, the standard style, and the conservative style; and determining the braking advance time of the vehicle based on the target style weights corresponding to the conservative style.

[0012] In some embodiments of this application, determining the driver's fatigue state score based on the driver's state data includes: acquiring a sequence of facial images of the driver, performing eye movement feature analysis on the driver based on the facial image sequence to obtain the driver's eye movement fatigue score, acquiring multiple head posture data of the driver, determining the driver's head stability score based on the statistical values ​​of the multiple head posture data, acquiring the vehicle's steering wheel pressure signal data, performing wavelet packet energy analysis on the pressure signal data within a preset frequency band to obtain the driver's grip strength abnormality score, and calculating the fatigue state score based on the eye movement fatigue score, the head stability score, and / or the grip strength abnormality score.

[0013] On the other hand, this application provides an electronic device comprising: a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, wherein the computer-readable instructions, when executed by the processor of the electronic device, implement the vehicle control method.

[0014] On the other hand, this application provides a vehicle including the aforementioned electronic device, the electronic device being used to perform the vehicle control method described above.

[0015] In the vehicle control scheme provided in this application embodiment, dual matching based on geographic area coding and traffic time period category can accurately search the database for the baseline style weights of various historical driving styles of the driver corresponding to the current driving scenario, thereby determining the driving scenario represented by the baseline style weights. Since vehicle state data accurately reflects the actual driving condition of the vehicle, and the driver's fatigue score reflects the driver's fatigue level, a pre-trained classification model can accurately determine the matching degree between the current driving scenario jointly represented by the vehicle state data and the driver's fatigue score and the driving scenario represented by the baseline style weights, based on the vehicle state data, the driver's fatigue score, and the baseline style weights. Based on the matching degree, the baseline style weights can be flexibly and adaptively adjusted to ensure that the adjusted target style weights match the current driving scenario while meeting the requirements for safe driving. Based on the target style weights, the required following parameters for the vehicle's driver assistance system are determined, ensuring that the following parameters can balance the driver's personalized style while guaranteeing driving safety and comfort. By using the driver assistance system to control vehicle driving based on the following parameters, adaptive, personalized, safe, and stable following behavior can be achieved, improving the driving experience and driving safety. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and are configured together with the description to explain the principles of this application.

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is an application scenario diagram of a vehicle control method provided in an embodiment of this application.

[0019] Figure 2 This is a flowchart of a vehicle control method provided in an embodiment of this application.

[0020] Figure 3 This is a flowchart illustrating a search and matching method for baseline style weights provided in an embodiment of this application.

[0021] Figure 4 This is a flowchart illustrating a method for determining target style weights provided in an embodiment of this application.

[0022] Figure 5 This is a flowchart of a training method for a classification model provided in an embodiment of this application.

[0023] Figure 6 This is a flowchart illustrating a clustering method provided in an embodiment of this application.

[0024] Figure 7 This is a schematic diagram of a clustering method provided in an embodiment of this application.

[0025] Figure 8 This is a schematic diagram of a clustering method provided in another embodiment.

[0026] Figure 9 This is a flowchart illustrating a classification model training method provided in an embodiment of this application.

[0027] Figure 10 This is a schematic flowchart of a vehicle control method provided in an embodiment of this application.

[0028] Figure 11 This is a schematic flowchart of a vehicle control method provided in another embodiment of this application.

[0029] Figure 12 This is a schematic flowchart of a vehicle control method provided in another embodiment of this application.

[0030] Figure 13This is a schematic flowchart of a vehicle control method provided in another embodiment of this application.

[0031] Figure 14 This is a schematic flowchart of a vehicle control method provided in another embodiment of this application.

[0032] Figure 15 This is a schematic flowchart of a vehicle control method provided in another embodiment of this application.

[0033] Figure 16 This is a schematic flowchart of a vehicle control method provided in another embodiment of this application.

[0034] Figure 17 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0035] It should be noted that in this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and drawings of this application are used to distinguish similar objects, not to describe a specific order or sequence.

[0036] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner. Unless otherwise specified, the following embodiments and features described herein can be combined with each other.

[0037] In the field of intelligent driving, vehicle following control is a crucial aspect of ensuring driving safety and a superior driving experience.

[0038] Car-following algorithms in related technologies are mainly divided into two categories: one is based on fixed rules, which controls following behavior by pre-setting a fixed distance or time interval; the other is a static driving style control algorithm, which determines the driving style by relying on preset parameters or manual selection by the driver. Both types of car-following algorithms generally suffer from problems such as insufficient safety redundancy, rigid control strategies, and poor adaptability, making it difficult to meet the comprehensive requirements of intelligent driving in terms of safety, comfort, and personalization.

[0039] To address the aforementioned technical problems, this application provides a vehicle control method that can ensure both driving safety and comfort while taking into account the driver's individual style.

[0040] like Figure 1 The diagram shown is an application scenario diagram of a vehicle control method provided in an embodiment of this application.

[0041] In this embodiment, the vehicle control method can be applied to electronic device 10. Electronic device 10 can be installed in vehicle 100; for example, electronic device 10 can be an on-board terminal in vehicle 100. Vehicle 100 can be a car, but is not limited to conventional cars, pure electric vehicles, or hybrid vehicles. Furthermore, the vehicle control method provided in this embodiment can also be applied to other types of motor vehicles or non-motor vehicles.

[0042] The vehicle 100 may also include a perception system 11, which is communicatively connected to the electronic device 10. The perception system 11 may include various sensors, such as vision sensors, distance sensors, inertial measurement units, temperature sensors, wheel speed sensors, etc. The vision sensor may be a camera, etc., and the distance sensor may be a lidar, millimeter-wave radar, ultrasonic radar, etc.

[0043] The electronic device 10 and sensor 11 described above are merely examples. In actual applications, the vehicle 100 may include more or fewer components, and this application does not impose any specific limitations on this.

[0044] In some embodiments, electronic device 10 may include network devices and / or user devices. Network devices include, but are not limited to, a single network electronic device, a group of electronic devices comprising multiple network electronic devices, or a cloud based on cloud computing consisting of a large number of hosts or network electronic devices.

[0045] The network where electronic device 10 is located may include, but is not limited to: the Internet, wide area network, metropolitan area network, local area network, and virtual private network (VPN).

[0046] In another embodiment, the electronic device 10 may also be an electronic product that communicates with the vehicle 100. For example, the electronic device 10 may be a personal computer, tablet computer, smartphone, personal digital assistant (PDA), game console, interactive network television (IPTV), smart wearable device, etc.

[0047] like Figure 2The diagram shown is a flowchart of a vehicle control method provided in one embodiment of this application. Depending on different needs, the order of the steps in this flowchart can be adjusted according to actual requirements, and some steps can be omitted. The vehicle control method is applied to electronic devices, such as… Figure 1 and Figure 17 The electronic device 10 shown. The vehicle control method includes the following steps: S11 collects multi-sensor data of the vehicle, including driver status data, vehicle status data and location information, with the vehicle location information corresponding to the collection time.

[0048] In some embodiments of this application, electronic devices can use multiple sensors of the vehicle to detect the road scene in which the vehicle is located, thereby obtaining multi-source perception information.

[0049] The sensors can be multimodal combinations. For example, the sensors may include vision sensors, distance sensors, positioning sensors (such as GNSS receivers), torque sensors, vehicle-mounted network terminals (Telematics Box, T-Box), and controller area network (CAN) buses. The vision sensors may be cameras, webcams, etc., and the distance sensors may be lidar, millimeter-wave radar, ultrasonic radar, etc.

[0050] Table 1 below shows an example of a sensor and collected information provided in one embodiment of this application. Table 1 only shows a portion of the sensors and collected information, and practical applications are not limited to this.

[0051] Table 1 In some embodiments of this application, the driver's state data may include a sequence of facial images of the driver, head posture data, and pressure signal data applied by the driver to the vehicle's steering wheel. For example, the sequence of facial images of the driver can be acquired using the visual sensors shown in Table 1, the driver's head posture data can be acquired using the ToF sensor shown in Table 1, and the pressure signal data can be obtained using a pressure sensor or a CAN bus.

[0052] In some embodiments of this application, the vehicle's location information can be represented by latitude and longitude coordinates in Table 1. The acquisition time corresponding to the vehicle's location information can be a timestamp or a time slot.

[0053] In some embodiments of this application, vehicle state data and driver state data can be collectively referred to as driving behavior data. Vehicle state data may include throttle data, braking data, steering data, following data, lane change data, speed information, etc. Throttle data may include throttle P10, throttle P50, throttle P90, throttle standard deviation, etc.; braking data may include braking P90, braking frequency, etc.; steering data may include steering angular velocity P90, steering angular velocity standard deviation, etc.; following data may include average following headway (THW) and following headway standard deviation; lane change data may include lane change frequency, etc.; and speed information may include average vehicle speed, etc.

[0054] For example, Table 2 below shows an example of vehicle status data provided in one embodiment of this application.

[0055] Table 2 In some embodiments of this application, after obtaining multi-sensor data, the multi-sensor data can be preprocessed, such as through filtering and noise reduction, and anomaly detection. For example, an electronic device can perform Kalman filtering on the vehicle's position to eliminate multipath effects and signal drift.

[0056] S12, Determine the driver's fatigue score based on the driver's status data.

[0057] In some embodiments of this application, the electronic device determines the driver's fatigue state score based on the driver's state data, including: acquiring a sequence of facial images of the driver, and performing eye movement feature analysis on the driver based on the facial image sequence to obtain the driver's eye movement fatigue score; acquiring multiple head posture data of the driver, and determining the driver's head stability score based on the statistical values ​​of the multiple head posture data; acquiring the vehicle's steering wheel pressure signal data, performing wavelet packet energy analysis on the pressure signal data within a preset frequency band to obtain the driver's grip strength abnormality score; and calculating the fatigue state score based on the eye movement fatigue score, head stability score, and / or grip strength abnormality score.

[0058] In some embodiments, electronic devices can utilize eye-tracking fatigue detection algorithms to perform eye-tracking feature analysis on a driver's facial image sequence, thereby obtaining an eye-tracking fatigue score. For example, an eye-tracking fatigue detection algorithm could be the Percentage of Eyelid Closure Over the Pupil Over Time (PERCLOS) algorithm, etc.

[0059] In other embodiments, the electronic device can determine fatigue characteristics such as the duration of eye closure, blinking frequency, and number of yawns of the driver through a sequence of facial images of the driver, and determine an eye movement fatigue score based on the fatigue characteristics.

[0060] In some embodiments, the electronic device can determine the driver's head stability score by statistical values ​​such as the variance or standard deviation of the driver's head posture data over a period of time (e.g., 30 seconds, 60 seconds, etc.).

[0061] In some embodiments, the electronic device can perform wavelet packet energy analysis on the pressure signal data of the vehicle's steering wheel to determine the proportion of wavelet packet energy in a preset frequency band, which is used as the driver's grip strength abnormality score. The preset frequency band can be customized, and this application embodiment does not limit it. For example, the preset frequency band can be from 4Hz to 8Hz.

[0062] In some embodiments, the electronic device may determine the driver’s fatigue score as the sum or weighted sum of eye movement fatigue score, head stability score and / or grip strength abnormality score.

[0063] The weights for eye fatigue score, head stability score, and grip strength abnormality score can be customized.

[0064] In some embodiments, the electronic device can normalize the eye movement fatigue score, head stability score, and / or grip strength abnormality score, and determine the driver's fatigue state score based on the normalized eye movement fatigue score, head stability score, and / or grip strength abnormality score. For example, the fatigue state score can also be expressed as a Driver Monitoring System (DMS) score, and the method for determining the fatigue state score can refer to the following formula (1): DMS_score = 100 - (0.5 × PERCLOS_norm + 0.3 × Head_Var + 0.2 ×Grip_Dev); (1) Among them, DMS_score can represent the fatigue state score, PERCLOS_norm can represent the eye movement fatigue score, Head_Var can represent the head stability score, and Grip_Dev can represent the grip strength abnormality score.

[0065] In some embodiments, the electronic device can acquire multiple fatigue score intervals and the corresponding fatigue state level for each fatigue score interval, so as to determine the driver's fatigue state level based on the fatigue score interval in which the driver's fatigue state score falls. The multiple fatigue score intervals and the corresponding fatigue state level for each fatigue score interval can be customized, and this application embodiment does not impose any limitations on this. For example, Table 3 below shows an example of fatigue score intervals and corresponding fatigue state levels provided in an embodiment of this application.

[0066] Table 3 In this embodiment, the fatigue state score can accurately reflect the driver's fatigue level. For example, a higher fatigue state score indicates a lower level of driver fatigue, while a lower fatigue state score indicates a higher level of driver fatigue.

[0067] S13. Based on the geographic area code corresponding to the vehicle's location information and the traffic time category corresponding to the collection time, search the database to determine the baseline style weights of the current driving scenario corresponding to the driver's various historical driving styles.

[0068] In some embodiments of this application, the electronic device can convert the vehicle's location information into a geographic region code. This geographic region code can be a Hexagonal Hierarchical Spatial Index (H3) geographic code (h3_zone), etc. For ease of distinction, the geographic region code corresponding to the vehicle's location information will be referred to as the target geographic region code below.

[0069] In some embodiments of this application, the database stores multiple geographic area codes, each geographic area code corresponding to multiple traffic time period categories, and each traffic time period category has style weights corresponding to multiple historical driving styles. These multiple historical driving styles can be customized. For example, the multiple historical driving styles may include aggressive style, standard style, and conservative style. In the database, the style weight corresponding to the aggressive style can be expressed as aggressive weight, the style weight corresponding to the standard style can be expressed as standard weight, and the style weight corresponding to the conservative style can be expressed as conservative weight.

[0070] In some embodiments of this application, multiple historical driving styles can be represented in the database using multiple cluster centers, with each cluster center representing a historical driving style. The multiple cluster centers and the historical driving styles represented by each cluster center will be explained in step S14 below. The style weights stored in the database can be preset or obtained by updating preset style weights. For example, updating the style weights in the database can refer to the process in step S15 below of updating the corresponding style weights in the database according to the target style weight.

[0071] In some embodiments of this application, each traffic time period category may have a corresponding time range. The electronic device can determine the corresponding traffic time period category based on the time range in which the data collection time falls. For ease of distinction, the traffic time period category corresponding to the data collection time is defined as the target traffic time period category below. The traffic time period category and its corresponding time range can be customized, and this application does not impose any restrictions on this. For example, Table 4 below shows an example of the time range and corresponding traffic time period category provided in one embodiment of this application.

[0072] Table 4 In other embodiments of this application, in addition to geographic region codes, traffic time categories corresponding to each geographic region code, style weights corresponding to each traffic time category, and the aforementioned cluster centers, the database may also store the driver's historical fatigue state score and user unique identifier (user_id). The historical fatigue state score can be determined based on the driver's historical status data. The method for determining the historical fatigue state score based on historical status data can be found in the description of the fatigue state score determination method in step S12. Using the driver's user unique identifier, the database can be used to query the corresponding driver's geographic region code, traffic time category corresponding to each geographic region code, style weights corresponding to each traffic time category, and multiple cluster centers.

[0073] In some embodiments of this application, the database may include multiple data tables for storing different data. For example, the database may include a main table and auxiliary tables. The main table may be used to store data such as geographic area codes, user unique identifiers, style weights, and historical fatigue state scores, while the auxiliary tables may be used to store data such as the aforementioned cluster centers and the total number of vehicle trips. Table 5 below shows an example of main table data provided in one embodiment of this application. Table 6 below shows an example of auxiliary table data provided in one embodiment of this application. Note that the data in Tables 5 and 6 are only a portion of the data in the database; in practical applications, the database data is not limited to these.

[0074] Table 5 Table 6 In some embodiments of this application, the electronic device searches a database based on the geographic region code corresponding to the vehicle's location information and the traffic time category corresponding to the collection time to determine the baseline style weight of the current driving scenario corresponding to multiple historical driving styles. This includes: if a target traffic time category corresponding to a target geographic region code is found in the database, the baseline style weight of the current driving scenario corresponding to multiple historical driving styles is determined according to the style weight corresponding to the target traffic time category; if a target traffic time category is not found in the database, the baseline style weight of the current driving scenario corresponding to multiple historical driving styles is determined according to the style weight corresponding to the traffic time category corresponding to the geographic region code; if no target geographic region code is found in the database, the baseline style weight of the current driving scenario corresponding to multiple historical driving styles is determined according to multiple preset weights.

[0075] If the target geographic area code is found in the database and corresponds to the target traffic time period category, the electronic device can determine the style weight corresponding to the target traffic time period category as the benchmark style weight for the current driving scenario corresponding to multiple historical driving styles.

[0076] If the target geographic area code is not found to correspond to the target traffic time period category in the database, the electronic device can determine the average or weighted average of the style weights corresponding to all traffic time period categories corresponding to the target geographic area code as the baseline style weight for the current driving scenario corresponding to multiple historical driving styles.

[0077] If the target geographic area code is not found in the database, the electronic device can determine multiple preset weights as the baseline style weights for the current driving scenario corresponding to various historical driving styles. These preset weights can be customized, and this embodiment does not impose any restrictions on them. The preset weights can be represented in vector form, referred to as preset weight vectors. For example, assuming that various historical driving styles include aggressive, standard, and conservative styles, the preset weight vectors corresponding to aggressive, standard, and conservative styles can be [0,0,1], where the preset weight corresponding to the aggressive style can be 0, the preset weight corresponding to the standard style can be 0, and the preset weight corresponding to the conservative style can be 1.

[0078] In some embodiments of this application, the sum of the baseline style weights corresponding to the current driving scenario for multiple historical driving styles can be 1.

[0079] In some embodiments of this application, the current driving scenario has a corresponding baseline style weight for each historical driving style. For example, assuming that multiple historical driving styles include aggressive, standard, and conservative styles, the current driving scenario has a baseline style weight (w_agg) for the aggressive style, a baseline style weight (w_std) for the standard style, and a baseline style weight (w_con) for the conservative style. Figure 3 The diagram shown is a flowchart illustrating a search and matching method for baseline style weights provided in an embodiment of this application. Figure 3 The search matching method shown can be referenced from the priority matching methods shown in Table 7 below. Figure 3 The description of the search and matching method for the baseline style weights shown can be found in the description of the method for determining the baseline style weights above.

[0080] Table 7 In this embodiment, by performing dual matching based on geographic region coding and traffic time period category, the baseline style weights of various historical driving styles of the driver corresponding to the current driving scenario can be accurately searched from the database, thereby determining the driving scenario represented by the baseline style weights.

[0081] S14. Using a pre-trained classification model, based on vehicle state data, fatigue state score, and baseline style weights, determine the matching degree between the current driving scenario and the driving scenario represented by the baseline style weights.

[0082] In some embodiments of this application, the network layers of the classification model may include multiple fully connected layers and a classification output layer. The number of fully connected layers can be customized, and this application does not limit this. For example, the number of fully connected layers can be three. The classification output layer can be a classification function, such as sigmoid or softmax.

[0083] In some embodiments of this application, the electronic device can generate a state feature vector X based on the vehicle's state data and the driver's fatigue state score. For example, using the vehicle state data shown in Table 2, the state feature vector X can have 13 dimensions, represented as X = [T10, T50, T90, Tstd, B90, Bfreq, S90, Sstd, THWm, THWstd, LCfreq, Vmean, DMS]. Wherein, T10, T50, T90, Tstd, B90, Bfreq, S90, Sstd, THWm, THWstd, LCfreq, and Vmean are the state data shown in items 1 to 12 of Table 2, and DMS can be the fatigue state score.

[0084] In some embodiments of this application, the database may store multiple cluster centers for drivers. These cluster centers can represent various historical driving styles, with each cluster center representing a specific historical driving style. Each cluster center can be a feature with the same dimension as the state feature vector X. For example, if the state feature vector X has 13 dimensions, each cluster center can be a 13-dimensional feature. Multiple cluster centers for drivers can be queried using the driver's unique identifier.

[0085] In some embodiments of this application, the electronic device can splice the state feature vector X, multiple cluster centers and baseline style weights to obtain spliced ​​features. The electronic device can input the spliced ​​features into a classification model to obtain the data output by each network layer in the classification model. The data output by the preceding network layer is used as the input data of the following and adjacent network layers. The data output by the classification output layer is determined as the matching degree between the current driving scene and the driving scene represented by the baseline style weights.

[0086] For example, suppose there are multiple historical driving styles, including aggressive, standard, and conservative styles, and the state feature vector is a 13-dimensional feature. The driver in the database can include 3 cluster centers, each corresponding to one of the three historical driving styles. Each cluster center has a 13-dimensional feature, and the baseline style weight is a 3-dimensional feature, corresponding to one of the three historical driving styles. Therefore, the concatenated feature can be a 55-dimensional feature, where 55 = 13 + 13 × 3 + 3.

[0087] In this embodiment, since the vehicle's state data can accurately reflect the vehicle's actual driving condition and the driver's fatigue score can reflect the driver's fatigue level, the matching degree between the current driving scenario and the driving scenario represented by the baseline style weight can be accurately determined by using the pre-trained classification model based on the vehicle's state data, the driver's fatigue score, and the baseline style weight.

[0088] S15, adjust the baseline style weights based on the matching degree to obtain the target style weights.

[0089] In some embodiments of this application, the electronic device adjusts the baseline style weight based on the matching degree to obtain the target style weight, including: if the matching degree is greater than or equal to a first preset threshold, determining the baseline style weight as the target style weight; if the matching degree is less than the first preset threshold but greater than or equal to a second preset threshold, using a preset first degrade factor to decrease the baseline style weight corresponding to the aggressive style and increase the baseline style weight corresponding to the conservative style, and updating the baseline style weight corresponding to the standard style based on the adjustment of the baseline style weights corresponding to the aggressive and conservative styles to obtain the target style weight; if the matching degree is less than the second preset threshold but greater than or equal to a third preset threshold, using a preset second degrade factor to decrease the baseline style weight corresponding to the aggressive style and increase the baseline style weight corresponding to the conservative style, and updating the baseline style weight corresponding to the standard style based on the adjustment of the baseline style weights corresponding to the aggressive and conservative styles to obtain the target style weight; if the matching degree is less than the third preset threshold, adjusting the baseline style weight to a preset weight to obtain the target style weight.

[0090] The first, second, and third preset thresholds can all be customized, and this application embodiment does not impose any restrictions on them. For example, the first preset threshold can be 0.7, the second preset threshold can be 0.4, and the third preset threshold can be 0.2.

[0091] For example, assuming the first preset threshold is 0.7, the second preset threshold is 0.4, and the third preset threshold is 0.2, as shown in Table 8 below, these are examples of matching degree levels corresponding to multiple matching degree ranges provided in an embodiment of this application.

[0092] Table 8 In some embodiments, by comparing the matching degree with a first preset threshold, a second preset threshold, and a third preset threshold, the electronic device can determine the degree of matching between the current driving scenario and the driving scenario represented by the baseline style weight. For example, if the matching degree is greater than or equal to the first preset threshold, the electronic device can determine that the current driving scenario matches the driving scenario represented by the baseline style weight; if the matching degree is less than the first preset threshold but greater than or equal to the second preset threshold, the electronic device can determine that the current driving scenario is slightly mismatched with the driving scenario represented by the baseline style weight; if the matching degree is less than the second preset threshold but greater than or equal to the third preset threshold, the electronic device can determine that the current driving scenario is moderately mismatched with the driving scenario represented by the baseline style weight; and if the matching degree is less than the third preset threshold, the electronic device can determine that the current driving scenario is severely mismatched with the driving scenario represented by the baseline style weight.

[0093] The first degradation factor is greater than the second degradation factor. Both the first and second degradation factors can be customized, and this application does not impose any restrictions on them. For example, the first degradation factor can be 0.8, and the second degradation factor can be 0.5.

[0094] In some embodiments, since the electronic device determines the baseline style weight as the target style weight when the matching degree is greater than or equal to the first preset threshold, the degradation factor corresponding to the aggressive style can be considered to be 1 when the matching degree is greater than the first preset threshold. Since the electronic device adjusts the baseline style weight to a preset weight to obtain the target style weight when the matching degree is less than the third preset threshold, the degradation factor corresponding to the aggressive style can be considered to be 0 when the matching degree is less than the third preset threshold.

[0095] In some embodiments, by using the various degradation factors corresponding to the radical style, it is possible to achieve different degrees of degradation of the radical style within different matching ranges. For example, assuming the first preset threshold is 0.7, the second preset threshold is 0.4, and the third preset threshold is 0.2, as shown in Table 9 below, these are examples of multiple matching ranges provided in an embodiment of this application and the degradation factors and degradation levels for each matching range corresponding to the radical style.

[0096] Table 9 In some embodiments, when the matching degree is less than a first preset threshold and greater than or equal to a second preset threshold, the electronic device can determine the target style weight corresponding to the aggressive style by multiplying the first degradation factor by the baseline style weight corresponding to the aggressive style, calculate a preset ratio of the difference between the value 1 and the first degradation factor, determine the target style weight corresponding to the conservative style by summing the preset ratio and the baseline style weight corresponding to the conservative style, and determine the target style weight corresponding to the standard style by adjusting the baseline style weights corresponding to the aggressive style and the conservative style, using the rule that the sum of the target style weights corresponding to the aggressive style, the conservative style and the standard style is 1.

[0097] In some embodiments, when the matching degree is less than a second preset threshold and greater than or equal to a third preset threshold, the electronic device can determine the target style weight corresponding to the aggressive style by multiplying the second degradation factor by the baseline style weight corresponding to the aggressive style, calculate a preset ratio value of the difference between the value 1 and the first degradation factor, sum the preset ratio value with the baseline style weight corresponding to the conservative style, and determine the target style weight corresponding to the standard style by adjusting the baseline style weights corresponding to the aggressive style and the conservative style, using the rule that the sum of the target style weights corresponding to the aggressive style, the conservative style and the standard style is 1.

[0098] The preset ratio can be customized, and this application embodiment does not limit it. For example, the preset ratio can be 0.5. For example, assuming the preset ratio is 0.5, in the case where the matching degree is less than the first preset threshold and greater than or equal to the second preset threshold, and in the case where the matching degree is less than the second preset threshold and greater than or equal to the third preset threshold, the calculation method of the target style weight of each historical driving style corresponding to the current driving scenario can refer to the following formulas (2)-(4): w_agg_final = w_agg_base × degrade_factor; (2) w_con_final = min(1.0, w_con_base + (1 - degrade_factor) × 0.5); (3) w_std_final = 1 - w_agg_final - w_con_final; (4) Among them, w_agg_final can represent the target style weight corresponding to the aggressive style, w_agg_base can represent the base style weight corresponding to the aggressive style, degrade_factor can represent the first degrade factor, w_con_final can represent the target style weight corresponding to the conservative style, w_con_base can represent the base style weight corresponding to the conservative style, and w_std_final can represent the target style weight corresponding to the standard style.

[0099] For example, assuming the baseline style weight for the aggressive style is 0.6, the baseline style weight for the standard style is 0.3, the baseline style weight for the conservative style is 0.1, the first preset threshold is 0.7, the second preset threshold is 0.4, and the third preset threshold is 0.2, if the matching degree is 0.45, since the matching degree 0.45 is less than the first preset threshold 0.7 and greater than the second preset threshold 0.4, the electronic device can determine that the current driving scenario is slightly mismatched with the driving scenario represented by the baseline style weight. In the case where the first degradation factor is 0.8 and the first degradation factor is 0.1, the electronic device can determine the target style weight w_agg_final for the aggressive style is w_agg_final = 0.6 × 0.8 = 0.48, the target style weight w_con_final for the conservative style is w_con_final = 0.1 + (1 - 0.8) × 0.5 = 0.2, and the target style weight w_std_final for the standard style is w_std_final = 1 - 0.48 - 0.2 = 0.32.

[0100] In some embodiments, when the matching degree is less than a third preset threshold, the electronic device can determine the target style weight as a preset weight. For an explanation of the preset weight, refer to the preset weight vector in step S13 above. For example, when the matching degree is less than the third preset threshold, the electronic device can determine the target style weight as a preset weight vector [0,0,1], where the preset weight corresponding to an aggressive style can be 0, the preset weight corresponding to a standard style can be 0, and the preset weight corresponding to a conservative style can be 1.

[0101] For example, such as Figure 4 The diagram shown is a flowchart illustrating a method for determining target style weights according to an embodiment of this application. (The diagram is intended to...) Figure 4 For a detailed explanation of each step, please refer to the description of the method for determining the target style weights above.

[0102] In this embodiment, the baseline style weights can be flexibly and adaptively adjusted based on the matching degree, thereby ensuring that the adjusted target style weights not only match the current driving scenario but also meet the requirements for safe driving.

[0103] In other embodiments of this application, the electronic device can update the style weights and historical fatigue state scores corresponding to the target style weights in the database based on the target style weights and using a preset smoothing factor, thereby updating the database so as to facilitate subsequent training of the classification model.

[0104] For example, electronic devices can use a preset smoothing factor to perform an exponentially weighted moving average (EWMA) calculation on the target style weight and the corresponding style weight in the database, and to perform an exponentially weighted calculation on the driver's fatigue state score and the corresponding fatigue state score in the database, thereby achieving database updates.

[0105] The preset smoothing factor can be customized, and this application embodiment does not impose any restrictions on it. For example, the preset smoothing factor can be 0.3 or 0.4.

[0106] For example, assuming the preset smoothing factor is 0.3, the updated style weights and fatigue state scores can be determined by referring to the following formulas (5)-(8): w_agg_new = 0.3 × w_agg_current + (1-0.3) × w_agg_old; (5) w_std_new = 0.3 × w_std_current + (1-0.3) × w_std_old; (6) w_con_new = 1 - w_agg_new - w_std_new; (7) avg_dms_new = 0.3 × dms_current + (1-0.3) × avg_dms_old; (8) Wherein, w_agg_new represents the updated style weight corresponding to the aggressive style, w_agg_current can represent the target style weight corresponding to the current aggressive style, w_agg_old can represent the style weight corresponding to the aggressive style in the database, w_std_new represents the updated style weight corresponding to the standard style, w_std_current can represent the target style weight corresponding to the current standard style, w_std_old can represent the style weight corresponding to the standard style in the database, w_con_new represents the updated style weight corresponding to the conservative style, avg_dms_new represents the updated fatigue state score, dms_current can represent the current driver's fatigue state score, and avg_dms_old can represent the driver's historical fatigue state score in the database.

[0107] In this embodiment, based on the target style weight, a preset smoothing factor is used to adjust the style weight corresponding to the target style weight and the historical fatigue state score in the database to achieve database updates. This allows the style weight and historical fatigue state score in the database to gradually approach the driver's latest actual state, thereby improving the accuracy and adaptability of subsequent matching.

[0108] In other embodiments of this application, the electronic device can perform attenuation on the style weights and historical fatigue state scores stored in the database according to a preset attenuation period and a preset attenuation factor. The preset attenuation period and preset attenuation factor can be customized, and this application does not limit them. For example, the preset attenuation period can be 30 days, 60 days, etc., and the preset attenuation factor can be 0.6, 0.8, etc. For example, the method for attenuating the style weights and fatigue state scores stored in the database can refer to the following formulas (9)-(12): w_agg = w_agg × β + 0.1 × (1 - β); (9) w_std = w_std × β + 0.1 × (1 - β); (10) w_con = 1 - w_agg - w_std; (11) avg_dms = avg_dms × β; (12) Where w_agg can represent the style weights corresponding to the aggressive style stored in the database, β can represent the preset decay factor, w_std can represent the style weights corresponding to the standard style stored in the database, w_con can represent the style weights corresponding to the conservative style stored in the database, and avg_dms can represent the historical fatigue state scores of the driver stored in the database.

[0109] In this embodiment, the style weights and historical fatigue state scores stored in the database are decayed according to a preset decay period and a preset decay factor. This introduces a time forgetting mechanism, thereby gradually reducing the influence of outdated data in the database. The database content focuses more on reflecting the driver's recent driving style and fatigue state, so as to improve the response speed and adaptability to changes in the driver's current state.

[0110] S16, based on the target style weights, determine the following parameters required by the vehicle's driver assistance system.

[0111] In some embodiments of this application, the following parameters can be one or more, such as following distance, lane change delay, maximum vehicle speed, throttle response, and braking advance. These following parameters can also be described as following control parameters. The driver assistance system can be an Advanced Driver Assistance System (ADAS), which can include one or more of the following systems: Adaptive Cruise Control (ACC), lane keeping assist system, and automatic lane change system.

[0112] In some embodiments of this application, the electronic device determines the following parameters required by the vehicle's driving assistance system based on the target style weights, including: determining the following distance, lane change delay, and maximum speed of the vehicle based on the target style weights corresponding to the aggressive style and the conservative style; determining the throttle response coefficient of the vehicle based on the target style weights corresponding to the aggressive style, the standard style, and the conservative style; and determining the braking advance time of the vehicle based on the target style weights corresponding to the conservative style.

[0113] For example, the method for determining the above-mentioned following parameters can refer to the calculation formulas shown in Table 10 below: Table 10 In Table 10, , , , This indicates the default parameter. For example, = 2.0 seconds (base following distance) =0.8 seconds (baseline lane change delay) = 120 km / h (base maximum speed) = 0.3 seconds (baseline braking advance).

[0114] In this embodiment, since the target style weight matches the current driving scenario and meets the requirements of safe driving, the following parameters required by the vehicle's driving assistance system are determined according to the target style weight, so that the following parameters can ensure driving safety and comfort while taking into account the driver's personalized style.

[0115] S17 utilizes a driver assistance system to control vehicle movement based on following parameters.

[0116] In some embodiments of this application, the electronic device can control the vehicle's driving based on following parameters, using the vehicle's adaptive cruise control system, lane keeping system, automatic lane changing system, etc., and provide the driver with various driving information through the vehicle's human-machine interface.

[0117] For example, electronic devices can provide different prompts to the driver on the human-computer interaction interface based on the different matching degrees between the current driving scenario and the driving scenario represented by the baseline style weight. For example, Table 11 below shows examples of prompts corresponding to different matching degree levels provided in an embodiment of this application.

[0118] Table 11 In this embodiment, since the following parameters can ensure driving safety and comfort while taking into account the driver's personalized style, the driving assistance system can control the vehicle's driving based on the following parameters to achieve adaptive, personalized, safe and stable following behavior, thereby improving the driving experience and driving safety.

[0119] In the vehicle control scheme provided in this application embodiment, dual matching based on geographic area coding and traffic time period category enables accurate retrieval of baseline style weights for various historical driving styles of the driver corresponding to the current driving scenario from the database, thereby determining the driving scenario represented by the baseline style weights. Since vehicle state data accurately reflects the actual driving condition of the vehicle, and driver fatigue score reflects the driver's fatigue level, a pre-trained classification model can accurately determine the matching degree between the current driving scenario jointly represented by the vehicle state data and driver fatigue score and the driving scenario represented by the baseline style weights, based on the vehicle state data, driver fatigue score, and baseline style weights. Based on the matching degree, the baseline style weights can be flexibly and adaptively adjusted to ensure that the adjusted target style weights match the current driving scenario while meeting the requirements for safe driving. Based on the target style weights, the following parameters required by the vehicle's driver assistance system are determined, ensuring that the following parameters can balance the driver's personalized style while guaranteeing driving safety and comfort. By using the driver assistance system to control vehicle driving based on the following parameters, adaptive, personalized, safe, and stable following behavior can be achieved, improving the driving experience and driving safety.

[0120] In some embodiments of this application, the electronic device can obtain the driver's historical fatigue state score and multiple cluster centers representing various historical driving styles from a database, and train a preset classification network to obtain a classification model. For example... Figure 5 The diagram shown is a flowchart of a training method for a classification model provided in an embodiment of this application, including the following steps: S21. Obtain the driver's historical fatigue state score for the corresponding historical driving scenario and multiple cluster centers to represent various historical driving styles from the database.

[0121] In some embodiments of this application, the historical fatigue state score corresponds to the historical state data of the vehicle, and the historical state data and the historical fatigue state score constitute a historical state feature vector.

[0122] For an explanation of the historical state data, please refer to the description of the vehicle state data in step S11 above. For an explanation of the historical fatigue state score, please refer to the description in step S12 above. For an explanation of the historical state feature vector, please refer to the explanation of the state feature vector X in step S14.

[0123] In some embodiments of this application, the electronic device can use a clustering algorithm to cluster multiple historical state feature vectors to obtain the aforementioned multiple cluster centers.

[0124] The electronic device can flexibly select a clustering algorithm. For example, the clustering algorithm can be one or more, such as Density-Based Spatial Clustering (DBSCAN) or k-means clustering. For example, the electronic device can use the DBSCAN algorithm to cluster multiple historical state feature vectors to obtain the above-mentioned multiple cluster centers, where the neighborhood radius of the cluster (DBSCAN eps) can be 150 meters.

[0125] In some embodiments of this application, the electronic device can acquire multiple driving segments (sensor data) of the vehicle through sensors. Each driving segment (e.g., a 10-second window) is converted into a 13-dimensional feature vector. Each 13-dimensional historical state feature vector includes the vehicle's historical state data and the driver's historical fatigue state score. Continuing with the 12 vehicle state data shown in Table 2, the dimension of the historical state feature vector can be 13-dimensional, represented as X = [T10, T50, T90, Tstd, B90, Bfreq, S90, Sstd, THWm, THWstd, LCfreq, Vmean, DMS].

[0126] For example, assuming there are N driving segments, the electronic device can input an input matrix consisting of N historical state feature vectors corresponding to the N driving segments into a clustering algorithm to obtain multiple cluster centers, each cluster center representing a historical driving style. For example, the input matrix... Shape: N × 13. Assume that there are multiple historical driving styles, including aggressive, standard, and conservative styles. The number of cluster centers can be 3, corresponding to the aggressive, standard, and conservative styles, respectively.

[0127] For example, multiple cluster centers can be represented as: Each cluster center is also a 13-dimensional vector, which can represent the key features of the corresponding historical driving style.

[0128] Cluster centers are also called clusters. As shown in Table 12 below, these are examples of the feature values ​​of multiple clusters provided in an embodiment of this application.

[0129] Table 12 In some embodiments of this application, an electronic device can determine the historical driving style represented by any cluster center based on the weighted summation of the features of any cluster center.

[0130] The electronic device can perform a weighted summation of some or all features of any cluster center, and this application embodiment does not limit this. For example, if each cluster center has 13-dimensional features, the electronic device can perform a weighted summation of the 3rd dimension T90, the 7th dimension S90, the 11th dimension LCfreq, the 5th dimension B90, the 9th dimension THWm, and the 13th dimension DMS of any cluster center to obtain the weighted summation score Agg_Score of any cluster center, and determine the historical driving style represented by each cluster center based on the ranking of the weighted summation scores of multiple cluster centers.

[0131] The weights used in the weighting can be customized, and this application does not impose any restrictions on them. For example, the weight corresponding to the 3rd dimension T90 can be 0.30, the weight corresponding to the 7th dimension S90 can be 0.20, the weight corresponding to the 11th dimension LCfreq can be 0.15, the weight corresponding to the 5th dimension B90 can be 0.10, the weight corresponding to the 9th dimension THWm can be -0.10, and the weight corresponding to the 13th dimension DMS can be -0.15.

[0132] For example, taking the three clusters shown in the table, the weighted sum score Agg_Score for clusters 0, 1, and 2 can be calculated using the following formula: Cluster 0: Agg_Score = 0.30×0.85 + 0.20×0.45 + 0.15×0.80 + 0.10×0.75 -0.10×0.24 - 0.15×0.85 = 0.255 + 0.09 + 0.12 + 0.075 - 0.024 - 0.1275 =0.3885.

[0133] Cluster 1: Agg_Score = 0.30×0.55 + 0.20×0.25 + 0.15×0.40 + 0.10×0.50 -0.10×0.50 - 0.15×0.80= 0.165 + 0.05 + 0.06 + 0.05 - 0.05 - 0.12= 0.155.

[0134] Cluster 2: Agg_Score = 0.30×0.30 + 0.20×0.12 + 0.15×0.15 + 0.10×0.30 -0.10×0.70 - 0.15×0.55 = 0.09 + 0.024 + 0.0225 + 0.03 - 0.07 - 0.0825=0.014.

[0135] For example, the electronic device determines the historical driving style represented by each cluster center based on the ranking of the weighted summation scores of multiple cluster centers. This includes: determining that the cluster center corresponding to the highest weighted summation score represents an aggressive style; determining that the cluster center corresponding to the lowest weighted summation score represents a conservative style; and determining that the cluster center corresponding to the weighted summation score between the highest and lowest weighted summation scores represents a standard style. For example, taking into account the weighted summation scores of 0.3885, 0.155, and 0.014 corresponding to clusters 0 to 2, the historical driving styles represented by clusters 0 to 2 can be referenced to the style labels shown in Table 13 below.

[0136] Table 13 The method of determining the historical driving style represented by each cluster center based on the weighted sum of scores from multiple cluster centers is merely an example. In practical applications, electronic devices can also map the weighted sum of scores from multiple cluster centers to corresponding historical driving styles based on preset style mapping rules.

[0137] like Figure 6 The diagram shown is a flowchart illustrating a clustering method provided in an embodiment of this application. Figure 7 The diagram shown is a schematic representation of a clustering method provided in an embodiment of this application. Figure 8 The diagram shown is a schematic representation of a clustering method provided in another embodiment. For Figures 6 to 8 For a detailed explanation of each step in the process, please refer to the description of clustering above.

[0138] S22, using multiple cluster centers, multiple historical state feature vectors, and historical baseline style weights corresponding to various historical driving styles for historical driving scenarios, a pre-set classification network is trained to obtain a classification model.

[0139] In some embodiments of this application, the classification network includes multiple fully connected layers and a classification output layer. The classification output layer can be a softmax layer or a sigmoid layer; this application does not limit this. In addition to fully connected layers and a classification output layer, the classification network may also include other network layers, such as ReLU layers, regularized Dropout layers, etc.

[0140] In some embodiments of this application, the electronic device uses multiple cluster centers, multiple historical state feature vectors, and historical baseline style weights corresponding to various historical driving styles of the driver's historical driving scenarios to train a preset classification network to obtain a classification model. This includes: concatenating multiple historical state feature vectors, multiple cluster centers, and historical baseline style weights to obtain historical concatenated features; inputting the historical concatenated features into the classification network to obtain the data output by each network layer of the classification network, wherein the data output by the preceding network layer is used as the input data of the following and adjacent network layers; determining the data output by the classification output layer as the historical matching degree of the driving scenario represented by the historical driving scenario and the historical baseline style weights; calculating the loss value based on the historical matching degree and the preset matching degree label; adjusting the classification network until the loss value is within a preset numerical range; and determining the adjusted classification network as the classification model.

[0141] The loss value can be cross-entropy loss, mean squared error, root mean square error, etc. The preset numerical range can be customized, and this embodiment does not impose any restrictions on it. For example, the preset numerical range can be the interval 0-1.

[0142] The parameters used to train the classification network, such as the initial learning rate, batch size, optimizer, number of training epochs, and Dropout, can be customized, and this application embodiment does not impose any restrictions on them. For example, the initial learning rate can be 0.001, the optimizer can be Adam, the batch size can be set to 128, the number of training epochs can be 50, and Dropout can be set to 0.3.

[0143] For example, such as Figure 9 The diagram shown is a flowchart illustrating a classification model training method provided in an embodiment of this application. (The diagram is intended for...) Figure 9 For a detailed explanation of each step, please refer to the description of the training method for the classification model above.

[0144] In this embodiment, since each cluster center represents a historical driving style and includes the key features of the corresponding historical driving style, a preset classification network is trained using multiple cluster centers, multiple historical state feature vectors, and historical baseline style weights corresponding to the driver's historical driving scenarios for multiple historical driving styles to obtain a classification model. This allows the classification model to learn the mapping relationship between driving behavior features and driving style categories, thereby ensuring that the classification model can accurately predict the matching degree between driving scenarios.

[0145] The vehicle control method provided in this application embodiment may include five steps: data acquisition and clustering classification, database construction and mapping, model consistency training, scene matching, and real-time monitoring and degradation. Table 14 below provides illustrative examples of each step in the vehicle control method provided in one embodiment of this application.

[0146] Table 14 Following steps one through five as shown in Table 14, as follows Figure 10 The diagram shown is a schematic flowchart of a vehicle control method provided in an embodiment of this application. Figure 11 The diagram shown is a schematic flowchart of a vehicle control method provided in another embodiment of this application. Figure 12 The diagram shown is a flowchart illustrating a vehicle control method according to another embodiment of this application. Figure 13 The diagram shown is a flowchart illustrating a vehicle control method according to another embodiment of this application. Figure 14 The diagram shown is a schematic representation of a vehicle control method provided in an embodiment of this application. Figure 15 The diagram shown is a schematic flowchart of a vehicle control method provided in an embodiment of this application. Figure 16 The diagram shown is a flowchart illustrating a vehicle control method according to another embodiment of this application. Figure 15 The diagram shown is a flowchart illustrating a vehicle control method according to another embodiment of this application. Figure 16 The diagram shown is a schematic flowchart of a vehicle control method provided in another embodiment of this application. (About...) Figures 8 to 16 For further explanation, please refer to the explanations of steps S11 to S17 and steps S21 to S22.

[0147] like Figure 17 The diagram shown is a structural schematic of an electronic device provided in an embodiment of this application. The electronic device 10 can be an electronic device, computer, mobile phone, tablet computer, laptop computer, server, etc. This application embodiment does not impose any restrictions on the specific type of the electronic device 10.

[0148] exist Figure 17 The electronic device 10 may include a communication module 101, a memory 102, a processor 103, an input / output (I / O) interface 104, and a bus 105. The processor 103 is coupled to the communication module 101, the memory 102, and the input / output interface 104 via the bus 105.

[0149] Communication module 101 may include a wired communication module and / or a wireless communication module. The wired communication module may provide one or more wired communication solutions such as Universal Serial Bus (USB) and Controller Area Network (CAN). The wireless communication module may provide one or more wireless communication solutions such as Wireless Fidelity (Wi-Fi), Bluetooth (BT), mobile communication networks, frequency modulation (FM), near field communication (NFC), and infrared (IR).

[0150] Memory 102 may include one or more random access memory (RAM) and one or more non-volatile memory (NVM). The RAM can be directly read and written by the processor 103, and can be used to store executable programs (e.g., machine instructions) of other running programs, as well as user and application data. The RAM may include 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), etc.

[0151] Non-volatile memory can also store executable programs and user and application data, and can be pre-loaded into random access memory for direct reading and writing by the processor 103. Non-volatile memory can include disk storage devices and flash memory.

[0152] The memory 102 is used to store one or more computer programs. The one or more computer programs are configured to be executed by the processor 103. The one or more computer programs include multiple instructions that, when executed by the processor 103, can implement a vehicle control method executed on the electronic device 10.

[0153] In other embodiments, such as Figure 17 The electronic device 10 shown also includes an external memory interface for connecting to an external memory to expand the storage capacity of the electronic device 10.

[0154] Processor 103 may include one or more processing units, such as application processors (APs), modem processors, graphics processing units (GPUs), image signal processors (ISPs), controllers, video codecs, digital signal processors (DSPs), and / or neural network processing units (NPUs). These different processing units may be independent devices or integrated into one or more processors.

[0155] The processor 103 provides computing and control capabilities. For example, the processor 103 is used to execute computer programs stored in the memory 102 to implement the vehicle control method described above.

[0156] The input / output interface 104 is used to provide a channel for user input or output. For example, the input / output interface 104 can be used to connect various input / output devices, such as a mouse, keyboard, touch device, display screen, etc., so that users can enter information or visualize information.

[0157] Bus 105 is used at least to provide a channel for communication between communication modules 101, memory 102, processor 103, and input / output interface 104 in electronic device 10.

[0158] It is understood that the structures illustrated in the embodiments of this application do not constitute a specific limitation on the electronic device 10. In other embodiments of this application, the electronic device 10 may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0159] This application also provides a computer-readable storage medium storing a computer program, which includes program instructions. When the program instructions are executed, the method implemented can refer to the methods in the above embodiments of this application.

[0160] The computer-readable storage medium can be the internal memory of the electronic device described in the above embodiments, such as the hard disk or memory of the electronic device. Alternatively, the computer-readable storage medium can be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., installed on the electronic device.

[0161] In some embodiments, a computer-readable storage medium may include a stored program area and a stored data area, wherein the stored program area may store an operating system, an application program required for at least one function, etc.; and the stored data area may store data created based on the use of the electronic device, etc.

[0162] This application provides a computer program product including computer instructions stored in a computer-readable storage medium. An electronic device's processor reads and executes the computer instructions from the computer-readable storage medium, causing the electronic device to perform the vehicle control method described in this application.

[0163] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0164] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0165] Furthermore, the functional modules 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. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0166] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of this application is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within this application. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0167] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices described in this application may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0168] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this application without departing from the spirit and scope of the technical solutions of this application.

Claims

1. A vehicle control method characterized by, The method includes: Collect multi-sensor data of the vehicle, including the driver's status data, the vehicle's status data and location information, with the vehicle's location information corresponding to the collection time; Based on the driver's status data, determine the driver's fatigue status score; Based on the geographic area code corresponding to the vehicle's location information and the traffic time category corresponding to the collection time, the database is searched to determine the baseline style weights of the current driving scenario corresponding to the driver's various historical driving styles. Using a pre-trained classification model based on the vehicle's state data, the fatigue state score, and the baseline style weights, the matching degree between the current driving scenario and the driving scenario represented by the baseline style weights is determined. The baseline style weight is adjusted based on the matching degree to obtain the target style weight, and the following parameters required by the vehicle's driving assistance system are determined based on the target style weight. The driving assistance system controls the vehicle's movement based on the following parameters.

2. The vehicle control method according to claim 1, characterized in that, The database stores multiple geographic region codes, each corresponding to multiple traffic time period categories, and each traffic time period category has a style weight corresponding to the various historical driving styles. The process of searching the database based on the geographic region code corresponding to the vehicle's location information and the traffic time period category corresponding to the collection time includes: If the geographical region code corresponding to the traffic time period category is found in the database, the baseline style weight of the current driving scenario corresponding to the multiple historical driving styles is determined according to the style weight corresponding to the traffic time period category. If the geographic region code is not found to correspond to the traffic time period category in the database, the baseline style weight of the current driving scenario corresponding to the multiple historical driving styles is determined according to the style weight of the traffic time period category corresponding to the geographic region code. If the geographic region code is not found in the database, the baseline style weight corresponding to the current driving scenario and the various historical driving styles is determined according to multiple preset weights.

3. The vehicle control method according to claim 2, characterized in that, The method further includes: Based on the target style weights, the style weights corresponding to the target style weights are updated in the database using a preset smoothing factor.

4. The vehicle control method according to claim 1, characterized in that, The method further includes: From the database, the historical fatigue state score of the driver corresponding to the historical driving scenario and multiple cluster centers used to represent the various historical driving styles are obtained; the historical fatigue state score corresponds to the historical state data of the vehicle, and the historical state data and the historical fatigue state score constitute a historical state feature vector. The multiple cluster centers are determined by clustering multiple historical state feature vectors. Each cluster center represents a historical driving style, and the historical driving style represented by any cluster center is determined based on the weighted sum of the features of any cluster center. The classification model is obtained by training a preset classification network using the multiple cluster centers, the multiple historical state feature vectors, and the historical baseline style weights corresponding to the multiple historical driving styles for the historical driving scenarios.

5. The vehicle control method according to claim 4, characterized in that, The classification network comprises multiple fully connected layers and a classification output layer; the training method for the classification model includes: By concatenating the multiple historical state feature vectors, the multiple cluster centers, and the historical baseline style weights, historical concatenated features are obtained. The historical splicing features are input into the classification network to obtain the data output by each network layer of the classification network, wherein the data output by the preceding network layer is used as the input data for the following and adjacent network layers. The data output by the classification output layer is determined as the historical matching degree between the historical driving scenario and the driving scenario represented by the historical baseline style weight; Calculate the loss value based on the historical matching degree and the preset matching degree label; The classification network is adjusted until the loss value is within a preset range, and the adjusted classification network is then determined as the classification model.

6. The vehicle control method according to claim 1, characterized in that, The various historical driving styles include aggressive, standard, and conservative styles. The step of adjusting the baseline style weight based on the matching degree to obtain the target style weight includes: If the matching degree is greater than or equal to the first preset threshold, the benchmark style weight is determined as the target style weight; If the matching degree is less than the first preset threshold and greater than or equal to the second preset threshold, the baseline style weight corresponding to the aggressive style is reduced and the baseline style weight corresponding to the conservative style is increased by using the preset first degradation factor. Based on the adjustment of the baseline style weights corresponding to the aggressive style and the conservative style, the baseline style weight corresponding to the standard style is updated to obtain the target style weight. If the matching degree is less than the second preset threshold and greater than or equal to the third preset threshold, the baseline style weight corresponding to the aggressive style is reduced and the baseline style weight corresponding to the conservative style is increased using the preset second degradation factor. Based on the adjustment of the baseline style weights corresponding to the aggressive style and the conservative style, the baseline style weight corresponding to the standard style is updated to obtain the target style weight. The first degradation factor is greater than the second degradation factor. If the matching degree is less than the third preset threshold, the baseline style weight is adjusted to the preset weight to obtain the target style weight.

7. The vehicle control method according to claim 6, characterized in that, The step of determining the following parameters required by the vehicle's driver assistance system based on the target style weight includes: Based on the target style weights corresponding to the aggressive style and the conservative style, the following distance, lane change delay and maximum speed of the vehicle are determined. The throttle response coefficient of the vehicle is determined based on the target style weights corresponding to the aggressive style, the standard style, and the conservative style. The braking advance time of the vehicle is determined based on the target style weight corresponding to the conservative style.

8. The vehicle control method according to any one of claims 1 to 7, characterized in that, Determining the driver's fatigue score based on the driver's status data includes: The driver's facial image sequence is acquired, and eye movement feature analysis is performed on the driver based on the facial image sequence to obtain the driver's eye movement fatigue score; Acquire multiple head posture data of the driver, and determine the driver's head stability score based on the statistical values ​​of the multiple head posture data; The pressure signal data of the vehicle's steering wheel is acquired, and wavelet packet energy analysis is performed on the pressure signal data within a preset frequency band to obtain the driver's grip strength abnormality score. The fatigue state score is calculated based on the eye movement fatigue score, the head stability score, and / or the grip strength abnormality score.

9. An electronic device, characterized in that, The electronic device includes: processor; The memory; and the application program, wherein the application program is stored in the memory and configured to be executed by the processor to implement the vehicle control method as described in any one of claims 1 to 8.

10. A vehicle, characterized in that, The vehicle includes the electronic equipment as described in claim 9.

Images