Control method of vehicle, vehicle, and electronic device

Abstract

The application relates to the technical field of vehicle control, in particular to a vehicle control method, a vehicle and an electronic device. The control method comprises the following steps: collecting a steering wheel steering torque signal of a driver within a preset time window; extracting a multi-dimensional feature representing a driver control style based on the steering wheel steering torque signal to obtain a driver behavior feature vector; performing similarity comparison on the driver behavior feature vector and a plurality of pre-stored driver behavior templates of registered drivers respectively, and obtaining corresponding similarity scores; when there is at least one driver behavior template whose similarity score exceeds a first preset threshold and lasts for more than a first preset time length, determining the driver identity corresponding to the driver behavior template as a current driver identity; and controlling at least one vehicle subsystem based on the current driver identity.

Description

Technical Field

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

[0002] In the automotive industry, with the improvement of vehicle intelligence, driver identification and personalized driving adaptation have become important development directions for enhancing the driving experience. Currently, driver identification mainly relies on devices such as physical keys and Bluetooth on mobile phones, or on biometric features such as facial recognition, fingerprints, and seat pressure distribution.

[0003] However, the above-mentioned technical solutions all have inherent defects: identification methods based on physical carriers and static biometrics are easily affected by environmental factors, have insufficient identification stability, and are mostly one-time authentications, which cannot continuously verify the driver's identity during driving and make it difficult to ensure the accurate matching of personalized settings with the actual driver. Summary of the Invention

[0004] This application addresses, to at least some extent, one of the technical problems in the related art.

[0005] In a first aspect, embodiments of this application provide a vehicle control method, the method comprising the following steps: Collect the driver's steering wheel torque signal within a preset time window; Based on the steering wheel torque signal, multi-dimensional features characterizing the driver's driving style are extracted to obtain a driver behavior feature vector; The driver behavior feature vector is compared with the driving behavior templates of multiple pre-stored registered drivers to obtain the corresponding similarity scores. When there is at least one driving behavior template whose similarity score exceeds the first preset threshold and lasts for more than the first preset duration, the driver identity corresponding to the driving behavior template is determined as the current driver identity. Based on the current driver's identity, control at least one vehicle subsystem.

[0006] In this technical solution, by collecting steering wheel torque signals and extracting multi-dimensional features characterizing the driver's driving style, driving behavior is quantified into a computable behavioral feature vector, thus avoiding dependence on external environment or biometric sensors and achieving stable recognition under all-weather conditions. By comparing the real-time behavioral feature vector with multiple pre-stored driving behavior templates and setting similarity thresholds and duration thresholds as judgment conditions, the driver's identity can be continuously verified during driving, effectively preventing misjudgments caused by momentary interference or driver changes, and ensuring accurate matching between personalized settings and the current driver. Furthermore, based on the determined driver identity, at least one vehicle subsystem is controlled, enabling vehicle parameters to adaptively adapt to the driver's driving habits, thereby improving the driving experience and driving safety.

[0007] In some embodiments of this application, it further includes: Collect driver visual feature data. If the driver visual feature data does not match the driver visual feature data of multiple pre-stored registered drivers, then the current driver is determined to be an unregistered driver. In response to the determination that the current driver is an unregistered driver, driving guidance is provided during the preset driving period, and the steering torque signal of the current driver's registered steering wheel is collected; Based on the steering wheel torque signal used for registration, the driver registration behavior feature vector is extracted, and the driver registration behavior feature vector is associated and bound with the driver identity represented by the visual feature data to generate and store the current driver's driving behavior template.

[0008] In this technical solution, visual feature data is introduced as a trigger and identity label for the registration phase, enabling the registration process to be automatically initiated when the vehicle is used for the first time or a new driver is detected. By synchronously collecting steering wheel torque signals and extracting behavioral feature vectors, these vectors are associated with and bound to the driver's identity represented by the visual feature data, generating a unique driving behavior template and thus establishing a high-confidence registration mechanism. This registration mechanism allows subsequent routine recognition to complete identity determination based on the behavioral template without relying on visual features. This solves the template initialization problem when a new driver uses the vehicle for the first time and avoids the interference of repeated registration with the user's driving, thereby improving the system's automation level and user experience.

[0009] In some embodiments of this application, the step of extracting multidimensional features characterizing the driver's driving style based on the steering wheel torque signal to obtain a driver behavior feature vector includes: Based on the steering wheel torque signal, at least one of the following types of features is extracted: time-domain features, frequency-domain features, and driving operation sequence features, to obtain the driver behavior feature vector.

[0010] In the technical solution, this application constructs a driver behavior feature vector by extracting at least one of the time-domain features, frequency-domain features, and driving operation sequence features from the steering wheel torque signal. This can comprehensively characterize the driver's driving style from multiple dimensions such as continuity, nonlinearity, time-series dependence, and dynamic response characteristics. The combination of multi-dimensional features makes the behavior feature vector highly unique and stable, thus accurately distinguishing different drivers in subsequent similarity comparisons. Even in the case of visual sensor failure or harsh environmental conditions, the reliability and robustness of identity recognition can still be guaranteed.

[0011] In some embodiments of this application, the extraction of time-domain and frequency-domain features based on the steering wheel torque signal includes: Under straight-line cruising conditions, the mean, standard deviation, and peak-to-peak value of the steering wheel torque signal are calculated as the time-domain features; under steering event conditions, the rising slope, peak sustain value, and operating impulse of the steering wheel torque signal are calculated as the time-domain features; the operating impulse is used to characterize the cumulative effect of the steering wheel torque intensity and duration during the driver's steering operation. The steering wheel torque signal is subjected to spectral transformation to obtain the energy proportion of a preset frequency band, and / or the approximate entropy of the steering wheel torque signal is calculated as the frequency domain feature.

[0012] In this technical solution, by calculating the mean, standard deviation, and peak-to-peak value of the steering wheel torque signal under straight-line cruising conditions as time-domain features, the driver's hand stability and micro-operation activity during stable driving can be accurately characterized. Under steering event conditions, calculating the rising slope, peak sustain value, and operational impulse of the torque signal as time-domain features effectively reflects the driver's decisiveness, force maintenance ability, and cumulative operational intensity during dynamic operations such as lane changes and turns. Simultaneously, by performing spectral transformation on the torque signal to obtain energy distribution characteristics within a preset frequency band, the frequency domain energy pattern when the driver consciously corrects the steering wheel can be extracted. Furthermore, calculating the signal complexity features further distinguishes between the smoothness and complexity of the driving style. The combination of these time-domain and frequency-domain features quantifies the driver's driving habits from different dimensions, giving the behavioral feature vector stronger individual discriminative power and anti-interference capabilities. This allows for more stable and accurate driver identification in subsequent similarity comparisons, especially suitable for scenarios with poor visual conditions or limited sensors.

[0013] In some embodiments of this application, extracting driving operation sequence features based on the steering wheel torque signal includes: Based on the steering wheel torque signal and steering wheel angle signal, a preset driving operation primitive type is identified. The driving operation primitive type is used to characterize the basic driving operation mode with identifiable start and end boundaries. Key point sequences are extracted from the temporal relationship between the steering torque and steering angle corresponding to the driving operation primitive type to obtain driving operation sequence features.

[0014] In this technical solution, by identifying preset driving operation primitive types, continuous driving behavior can be segmented into operation fragments with clear semantics, thereby focusing on the personalized control mode of the driver when performing specific operations. Furthermore, key point sequences are extracted from the torque-angle temporal relationship corresponding to the identified driving operation primitives to obtain driving operation sequence features. These features can capture the shape differences of the torque-angle trajectory and the distribution patterns of key points when different drivers perform the same operation. Through this method, the driving operation sequence features characterize the driver's control style from a higher-level semantic action dimension, effectively compensating for the shortcomings of time-domain and frequency-domain features in describing complex operational behaviors, significantly improving the distinguishability between different drivers, and thus enhancing the accuracy and robustness of identity recognition.

[0015] In some embodiments of this application, controlling at least one vehicle subsystem based on the current driver identity includes: Based on the current driver's identity, a preset driver association parameter mapping table is queried to generate the corresponding control signal; The control signal is sent to at least one vehicle subsystem to control the corresponding vehicle subsystem.

[0016] In this technical solution, by querying a pre-set driver-related parameter mapping table based on the determined driver identity, the identity recognition result can be quickly mapped to the control parameter set of the corresponding vehicle subsystem, achieving automated decision-making from recognition to adaptation. By sending control signals to at least one vehicle subsystem, the subsystem can collaboratively adjust to a working state that matches the driver's preferences. This approach not only ensures accurate matching between personalized settings and the current driver, avoiding safety risks and decreased experience due to identity mismatch, but also meets the needs of real-time vehicle control due to the pre-stored parameter mapping table and rapid query response, significantly improving driving comfort, control consistency, and overall intelligence level.

[0017] In some embodiments of this application, it further includes: Calculate the mean of a preset sliding window of the driver behavior feature vector; When the mean value of the preset sliding window changes more than a preset change threshold within a preset time window, and no change in the driver's visual feature data is detected, it is determined to be an abnormal driving event. In response to the abnormal driving event, generate a prompt message and / or switch at least one vehicle subsystem to the default safety mode.

[0018] In this technical solution, by calculating the mean of a preset sliding window of the driver's behavioral feature vector, instantaneous noise can be effectively smoothed, reflecting the trend changes in the driver's driving style over a short period of time. The abnormal driving event determination mechanism can accurately identify sudden changes in driving style caused by driver changes, fatigue, or distraction during driving, exhibiting higher robustness and reliability compared to simply relying on instantaneous behavioral feature values ​​or visual sensors. In response to abnormal driving events, the system generates prompts and / or switches at least one vehicle subsystem to a default safety mode, thereby taking timely safety measures when the driver's identity is uncertain or their driving state is abnormal. This avoids risks caused by incorrect personalized parameter adaptation or driver negligence, significantly improving driving safety.

[0019] In some embodiments of this application, it further includes: When the similarity score exceeds the second preset threshold and continues for more than the second preset time, the current driver's identity is determined to be a reliable identification. In response to the determination that the current driver's identity is a reliable identification, the driver behavior template corresponding to the current driver is updated with a preset weight based on the current driver's behavior feature vector; The second preset threshold is greater than the first preset threshold, and the second preset time is longer than the first preset duration.

[0020] In this technical solution, the application sets conditions to trigger high-confidence recognition. Only when the similarity score is significantly higher than the ordinary judgment threshold and remains stable for a relatively long period is the current driver's identity determined to be reliable. In response to this determination, the current driver's corresponding behavior template is updated using the current behavior feature vector with preset weights, achieving a slow and gradual adaptive learning mechanism. This adaptive learning mechanism effectively prevents template distortion caused by single, accidental abnormal driving, momentary interference, or recognition uncertainty. It incorporates only highly reliable and long-term stable control features into the template, thereby tracking the slow changes in driver habits over time while maintaining template stability and accuracy. This avoids template mutations from adversely affecting subsequent recognition, significantly improving the system's long-term reliability and the consistency of user experience.

[0021] Secondly, embodiments of this application provide a vehicle, including a controller, which is used to execute the vehicle control method described in the first aspect above.

[0022] Fourthly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the vehicle control method described in the first aspect above.

[0023] As can be seen from the above technical solutions, additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0024] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of the driver registration process according to an embodiment of this application; Figure 2 This is a schematic flowchart of a vehicle control method according to an embodiment of this application; Figure 3 This is a schematic diagram of the abnormal driving event detection process according to an embodiment of this application; Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application.

[0025] In the above figures: 40. Bus; 41. Processor; 42. Memory; 43. Communication interface. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0027] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0028] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0029] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0030] In the automotive industry, with the increasing intelligence of vehicles, personalized driving experiences have become an important development direction. Different drivers have different preferences and habits regarding vehicle power response, steering feel, chassis suspension, and human-machine interface. Therefore, vehicle systems need to be able to recognize the current driver's identity and automatically adjust relevant subsystem parameters according to their preferences to achieve personalized services tailored to the driver.

[0031] In existing technologies, driver identification mainly relies on physical keys, mobile phone Bluetooth and other devices for identification, or biometric features such as face, fingerprint, and seat pressure distribution for identification.

[0032] However, all of the above technical solutions have inherent defects: Physically based driver identification can only confirm users carrying the device and cannot handle scenarios such as family members sharing keys, which can easily lead to mismatches between personalized settings and real drivers. Driver identification based on static biometrics has a significantly lower recognition rate under conditions such as low light, backlight, and facial occlusion, and it is a one-time authentication that cannot continuously verify the driver's identity during driving. Passive biometric identification, such as driver identification based on seat pressure distribution, is easily affected by occupant clothing and the placement of items, and it is difficult to distinguish drivers of similar body types.

[0033] Therefore, there is an urgent need for a driver identification and adaptive control scheme that can be continuously online and is stable and reliable.

[0034] Based on this, this application proposes a vehicle control method, a vehicle, and electronic equipment. By acquiring steering wheel torque signals and extracting multi-dimensional features characterizing the driver's driving style, a unique behavioral feature vector of the driver is constructed. This vector is then compared with a pre-stored behavioral template to identify the driver's identity in real time, and the vehicle subsystem is adaptively controlled based on the identification results. This application achieves stable and reliable driver identification and personalized adaptive control, solving the technical problems of existing technologies such as susceptibility to environmental interference, poor identification stability, and inability to continuously verify driver identity. This significantly improves driving safety, personalized experience, and the level of vehicle intelligence.

[0035] It should be noted that, for ease of understanding, the key technical concepts involved in this application are briefly explained below.

[0036] EPS (Electric Power Steering) is a power steering system that uses an electric motor to generate auxiliary torque to reduce the driver's steering effort. An electric power steering system typically includes a torque sensor, a vehicle speed sensor, and a control unit, which can adjust the amount of assistance in real time according to the driver's steering intentions and vehicle conditions.

[0037] A Driver Monitoring System (DMS) is a system that monitors the driver's condition in real time using sensors such as cameras. Driver monitoring systems are typically used to detect driver fatigue, distraction, and lack of concentration, and can also identify information such as the driver's facial features, head posture, gaze direction, and seating position.

[0038] Figure 1 This is a flowchart of the driver registration process according to an embodiment of the control method of this application, such as... Figure 1 As shown below, in conjunction with Figure 1 This document details the process of driver registration and driving behavior template generation.

[0039] S61: Collect driver visual feature data. If the driver visual feature data does not match the driver visual feature data of multiple pre-stored registered drivers, then determine that the current driver is an unregistered driver.

[0040] Preferably, the driver's visual feature data is collected by a camera in the driver monitoring system configured in the vehicle, and the visual feature data includes at least one of facial image, head posture, gaze direction and sitting posture model.

[0041] Specifically, driver monitoring systems typically include one or more driver-oriented image acquisition sensors, such as high-resolution infrared cameras or cameras with depth sensing capabilities. During vehicle startup or operation, the cameras continuously acquire images or video streams of the driver at a preset frame rate.

[0042] When acquiring facial images, the system uses face detection algorithms to locate facial regions in the image and extracts key feature points to form a facial feature template. When acquiring head posture, the system uses a 3D posture estimation algorithm to calculate the pitch, yaw, and roll angles of the head relative to the vehicle coordinate system. When acquiring gaze direction, the system uses an eye-tracking model to determine the driver's gaze direction based on the positional relationship between the pupil and corneal reflection centers. When acquiring a seated posture model, the system acquires 3D point cloud data of the driver's torso using a depth camera or structured light sensor, extracting the relative position and angle information of the shoulders, back, and arms. To ensure data quality, the system can automatically adjust the camera's exposure parameters or enable infrared illumination based on ambient lighting conditions, ensuring clear images are acquired even at night or in backlight.

[0043] Furthermore, during the determination process, the similarity between the real-time collected driver visual feature data and the pre-stored driver visual feature data of registered drivers is calculated.

[0044] For facial images, facial feature vectors can be extracted, and cosine similarity or Euclidean distance between them and the feature vectors of each registered driver can be calculated. For head posture, gaze direction, and sitting posture models, corresponding posture differences or model matching scores are calculated respectively. The system can assign preset weights to different categories of visual features and calculate a weighted comprehensive similarity. If the weighted comprehensive similarity is lower than a preset matching threshold, it is preliminarily determined that the current driver may be an unregistered driver.

[0045] In addition, to avoid misjudgment due to momentary interference such as sudden changes in light, drivers briefly turning their heads, or temporary obstruction, the system can set judgment duration conditions. For example, it can require that the weighted comprehensive similarity in multiple consecutive frames be lower than a preset matching threshold before finally confirming that the driver is an unregistered driver.

[0046] Meanwhile, if the system detects a driver whose visual features closely match some features of a registered driver but whose face is obscured by a mask or sunglasses, the system can prompt the driver via the central control screen or voice to "Please remove your mask / sunglasses for identification," or guide the driver to perform auxiliary confirmation actions such as briefly pressing a steering wheel button to confirm whether the driver is a registered driver. If a match still cannot be found after confirmation, the registration process is initiated. Through this mechanism, the system ensures the sensitivity of registration triggering while effectively avoiding erroneous initiation of the registration process due to brief environmental or posture changes, thus improving the robustness of the judgment and the user experience.

[0047] S62: In response to the determination that the current driver is an unregistered driver, driving guidance is provided within a preset driving period, and the steering torque signal of the current driver's registered steering wheel is collected.

[0048] Preferably, in response to the determination that the current driver is an unregistered driver, the driver registration process is initiated, and a prompt message is output through the in-vehicle human-machine interface to inform the driver that a driving profile is about to be established. For example, the central control screen may display "New driver detected, driving profile being established for you, please continue driving normally for 1-2 minutes," and the voice assistant will simultaneously broadcast the corresponding prompt. The system does not require any additional operation from the driver; the registration process is carried out silently in the background and does not affect normal driving.

[0049] Furthermore, the driving guidance includes real-time monitoring of the current driving conditions. If sufficient effective driving operations, such as straight-line cruising, lane changing, or turning, are not detected within a preset time, guidance information is proactively output to prompt the driver to perform specific operations.

[0050] For example, the system can display messages like "Please safely complete a lane change" or "Please maintain a straight line" on the central control screen, or provide voice prompts like "Please gently turn the steering wheel to change lanes to the left." The guidance information can be intelligently adjusted based on the current vehicle environment, avoiding unnecessary guidance in congested traffic or dangerous scenarios. Drivers can drive normally according to their usual habits, and the system automatically identifies and filters valid signal segments, without needing to strictly follow the guidance sequence.

[0051] Furthermore, to ensure the representativeness and quality of the registered data, the system can set a maximum data collection time. If sufficient valid operating conditions are not collected within the maximum collection time, the system can generate a basic behavior template based on the collected data and gradually improve it in subsequent driving through a self-learning mechanism. Simultaneously, the system can determine whether to prompt the driver to extend the registration period or re-perform certain operations based on the quality score of the collected data.

[0052] In addition to the automatic determination based on DMS visual features mentioned above, the registration process can also be triggered manually or by voice activation. Manual activation means the user actively selects "Add New Driver" or "Create Driving Profile" through the vehicle's user management interface to initiate the registration process. Voice activation means the driver actively initiates the registration process using voice commands.

[0053] S63: Extract the driver's registered behavior feature vector based on the steering wheel torque signal used for registration, and associate and bind the driver's registered behavior feature vector with the driver's identity represented by the visual feature data to generate and store the current driver's driving behavior template.

[0054] Preferably, the driver registration behavior feature vector is obtained by performing multi-dimensional analysis on the steering torque signal of the steering wheel used for registration. This multi-dimensional analysis includes at least one of the following: analysis of the time-domain statistical characteristics, frequency-domain energy distribution characteristics, and operation sequence timing characteristics of the torque signal. The extraction methods, specific calculation formulas, and parameter settings for each dimension of the feature will be described in detail below with reference to specific embodiments.

[0055] Furthermore, when the extracted driver behavior feature vector is associated with the driver identity represented by the visual feature data, the driver registration behavior feature vector can be stored as a standard template for the driver in the local secure storage area, and at least a portion of the visual feature data can be used as the driver's identity index for subsequent fast retrieval, such as facial feature hash value or sitting posture model parameters.

[0056] Furthermore, for the same driver, the stored driving behavior template can be progressively updated through multiple registrations or a self-learning mechanism in subsequent high-confidence recognition stages. This allows the template to track the long-term, slow changes in the driver's operating habits, maintaining recognition accuracy. The specific implementation of the self-learning mechanism update will be further elaborated later.

[0057] Figure 2 This is a flowchart of a vehicle control method according to an embodiment of this application, such as... Figure 2 As shown, the method includes the following steps S1-S5.

[0058] S1: Collect the driver's steering wheel torque signal within a preset time window.

[0059] Preferably, the steering wheel torque signal The torque data is collected by a torque sensor built into the vehicle's electric power steering system. The EPS torque sensor is a high-precision mechanical sensor, typically installed between the steering wheel and steering column. It can measure the torque applied to the steering wheel by the driver in real time and imperceptibly at a sampling rate of at least 1 kHz, while simultaneously outputting corresponding analog or digital signals for the controller to read. Since the EPS system operates continuously after the vehicle is started, this application requires no additional hardware and can obtain raw data reflecting the driver's steering effort in a low-cost and highly reliable manner.

[0060] Furthermore, the length of the preset time window can be set according to different application scenarios. During the daily routine identification of registered drivers, to balance real-time performance and feature stability, a shorter time window is typically used, such as 10 seconds. The system continuously extracts features and matches them with templates using a sliding window approach. During the driver registration phase, to collect sufficiently rich typical working conditions, a longer time window is needed, such as 1-2 minutes, to ensure that the generated behavioral templates are sufficiently representative. Through these settings, this application can adaptively adjust the collection duration, satisfying both the data volume requirements of the registration phase and ensuring rapid response during the routine identification phase.

[0061] S2: Extract multi-dimensional features representing the driver's driving style based on the steering wheel torque signal to obtain the driver's behavior feature vector.

[0062] In some embodiments, at least one of time-domain features, frequency-domain features, and driving operation sequence features is extracted based on the steering wheel torque signal to obtain a driver behavior feature vector.

[0063] Preferably, by extracting time-domain features, the driving style can be characterized from dimensions such as the driver's steady-state maintenance ability, dynamic response speed and operation intensity during straight-line cruising and steering events, which has the advantages of simple calculation and high real-time performance.

[0064] By extracting frequency domain features, individual differences can be reflected from nonlinear dimensions such as the energy distribution and operational complexity of the driver's conscious direction correction, thus compensating for the insensitivity of time domain features to hidden patterns.

[0065] By extracting features from driving operation sequences, the differences in operating habits among different drivers can be further amplified based on the torque-angle trajectory pattern of driving primitives, thereby improving the accuracy and robustness of recognition.

[0066] The three types of features can be used independently or combined with each other, thus flexibly adapting to different computing resources and recognition accuracy requirements, and realizing multi-level representation of driving behavior from macro to micro and from linear to nonlinear.

[0067] In some embodiments, extracting time-domain features based on the steering wheel torque signal includes: Under straight-line cruising conditions, the mean, standard deviation, and peak-to-peak value of the steering wheel torque signal are calculated as the time-domain features. Under steering event conditions, the rising slope, peak sustain value, and operating impulse of the steering wheel torque signal are calculated as the time-domain features. The operating impulse is used to characterize the cumulative effect of the steering wheel torque intensity and duration during the driver's steering operation.

[0068] Preferably, the straight-line cruising condition primarily reflects the driver's ability to maintain steady-state control and the level of fine-tuning, while the steering event condition reflects the driver's dynamic response characteristics and the force of their operation. Extracting features from these two conditions separately avoids confusing different types of behavior, thus more accurately portraying the driver's driving style.

[0069] The two operating conditions are distinguished as follows: when the absolute value of the steering wheel angle is less than 5° and the vehicle speed is greater than 30km / h, it is determined to be a straight-line cruising condition; when the rate of change of the steering wheel angle exceeds a preset threshold and lasts for a preset time, combined with changes in vehicle speed and torque, it is determined to be a steering event condition. For example, if the rate of change of the steering wheel angle instantaneously reaches 80° / s, and this high rate of change lasts for 0.2 seconds, while the steering torque rapidly increases from 0.5N·m to 3N·m, the system determines that a steering event condition has begun.

[0070] Preferably, under straight-line cruising conditions, the average value of the steering wheel torque signal is calculated using the following formula:

[0071] in, This represents the number of sampling points for the torque signal during the straight-line cruise segment. For the first The torque value at each sampling point.

[0072] Mean value of steering wheel torque signal It reflects the average hand force a driver uses to maintain their lane, demonstrating hand stability.

[0073] The standard deviation of the steering torque signal under straight-line cruising conditions is calculated as follows:

[0074] It reflects the driver's level of activity in fine-tuning the steering wheel; the higher the value, the higher the level of fine-tuning activity.

[0075] The peak-to-peak value of the steering wheel torque signal is calculated as follows during straight-line cruising:

[0076] in, This represents the i-th steering torque sample value collected during the straight-line cruising segment. This represents the maximum value of the torque within that segment. The minimum value. Peak-to-peak value. It is the difference between the maximum and minimum values, representing the overall fluctuation range of the torque signal.

[0077] Peak-to-peak value represents the maximum fluctuation range of the torque signal, reflecting the driver's correction range in lane keeping.

[0078] Furthermore, under steering event conditions, the formula for calculating the rising edge slope of the steering torque signal is as follows:

[0079] in The torque at the start of the turn. This is the peak torque first reached during the steering process. This represents the time interval from the start to the peak value. It reflects the driver's decisiveness in steering.

[0080] The formula for calculating the peak sustained value of the steering torque signal under steering event conditions is as follows:

[0081] in, This represents the number of sampling points during the stability maintenance phase of a turning event. It reflects the driver's ability to maintain torque in curves.

[0082] The formula for calculating the operating impulse of the steering torque signal under steering event conditions is as follows:

[0083] in For the complete time interval of the turning event; This represents the steering torque value at time t.

[0084] The integral calculated above encloses the area bounded by the torque-time curve over the duration of the steering event. Physically speaking, the operating impulse... It comprehensively reflects the cumulative effect of the intensity and duration of the torque applied by the driver during steering.

[0085] Specifically, if the driver operates decisively and quickly, the peak torque will be high and the duration will be short, and the operating impulse may be moderate or large; if the driver operates slowly and hesitantly, the torque will rise gradually and the duration will be long, and the operating impulse may be large; if the driver operates casually and the torque is small throughout, the operating impulse will be small.

[0086] In some embodiments, extracting frequency domain features based on the steering wheel torque signal includes: The steering wheel torque signal is subjected to spectral transformation to obtain the energy proportion of a preset frequency band, and / or the approximate entropy of the steering wheel torque signal is calculated as the frequency domain feature.

[0087] Preferably, the calculation process for the energy proportion of the preset frequency band is as follows.

[0088] First, for continuous torque signals Perform a Fast Fourier Transform:

[0089] in, This represents the steering torque signal in the time domain. It is a time variable; For frequency variables; The imaginary unit; The transformed frequency domain complex signal has an amplitude that reflects the intensity of the frequency component. This formula converts the torque signal from the time domain to the frequency domain, facilitating the analysis of the energy distribution of different frequency components.

[0090] Based on the continuous torque signal after Fast Fourier Transform The formula for calculating the power spectral density is as follows:

[0091] in, for The model, Power spectral density represents the signal power per unit frequency; the larger the value, the stronger the energy of the corresponding frequency component.

[0092] The preset frequency band can be set to 0.5-2Hz. The 0.5-2Hz frequency band corresponds to the main frequency range of the driver's conscious steering correction. Therefore, this frequency band can effectively extract frequency domain features related to the driver's active driving style.

[0093] The formula for calculating the energy percentage based on the preset frequency band of 0.5-2Hz is as follows:

[0094] in , The numerator is the power integral within the 0.5-2Hz frequency band, and the denominator is the total power across the entire frequency band. The preset frequency band energy ratio reflects the degree of energy concentration for conscious steering corrections by the driver and exhibits individual differences.

[0095] Furthermore, the formula for calculating the approximate entropy of the steering wheel torque signal is as follows:

[0096] in, The embedding dimension (usually 2). For similarity tolerance (usually 0.2 times the standard deviation), The data length is represented by the approximate entropy. The larger the approximate entropy, the more complex and variable the signal; the smaller the entropy value, the more regular and smooth the control style.

[0097] In some embodiments, extracting driving operation sequence features based on steering wheel torque signals includes: Based on the steering wheel torque signal and steering wheel angle signal, a preset driving operation primitive type is identified; the driving operation primitive type is used to characterize the basic driving operation mode with identifiable start and end boundaries; the key point sequence is extracted from the temporal relationship between the steering wheel torque and steering wheel angle corresponding to the driving operation primitive type to obtain the driving operation sequence feature.

[0098] Driving operation primitives refer to the basic operational units that constitute complex driving behaviors. They have clear semantics, identifiable start and end boundaries, and relatively fixed pattern characteristics. The driving operation primitives defined in this application include, but are not limited to, typical driving operations such as "low-speed lane changing," "high-speed lane changing," and "ramp turning."

[0099] Specifically, low-speed maneuvering refers to the driver's large-amplitude, low-speed back-and-forth steering wheel operation when parking or making a U-turn in a narrow area. It is usually accompanied by characteristics such as vehicle speed below 5 km / h, steering wheel angle exceeding 90°, and frequent alternation of steering direction.

[0100] High-speed lane changing refers to lane changing operations performed by the driver at high vehicle speeds. It is usually characterized by rapid steering wheel rotation and return to center, accompanied by vehicle speeds exceeding 60 km / h and a turning angle change rate exceeding 80° / s.

[0101] Ramp turns refer to continuous, stable, high-curvature curves that drivers navigate at highway entrances / exits or overpasses. They are typically characterized by a relatively fixed large steering wheel angle, a vehicle speed between 40-80 km / h, and a continuously stable steering torque.

[0102] By identifying these driving operation primitives, the system can segment continuous, unstructured driving behaviors into semantically meaningful operation segments, and then extract more refined torque-angle timing features within each primitive, thereby effectively distinguishing the personalized driving habits of different drivers.

[0103] Preferably, the mathematical expression of the driving operation sequence features is as follows: Define driving operation primitives For each detected primitive, its time interval is recorded. And extract the torque-rotation trajectory:

[0104] Key point sequences are extracted from the trajectory, such as the starting point, peak point, inflection point, and ending point, to form a key point sequence:

[0105] Alternatively, ellipse fitting can be performed to obtain the ellipse parameters: center. Long axis short axis Rotation angle , forming the feature vector:

[0106] The aforementioned ellipse parameters or key point sequence are driving operation sequence features, which can effectively distinguish the personalized operating habits of different drivers when executing the same primitive.

[0107] Preferably, based on the extracted time-domain features, frequency-domain features, and driving operation sequence features, a driver behavior feature vector is constructed. for:

[0108] in, As a scalar feature, or For vector sub-blocks.

[0109] The driver behavior feature vector constructed above integrates features from three dimensions: time domain, frequency domain, and operation sequence. It can comprehensively depict the driver's personalized driving style from multiple perspectives, such as control force, correction frequency, operation complexity, and trajectory shape, significantly improving individual differentiation and recognition robustness. At the same time, each feature has a clear physical meaning, moderate computational complexity, and meets automotive-grade real-time processing requirements, providing a reliable basis for subsequent similarity comparison and identity recognition.

[0110] It should be noted that if no driving operation primitives are detected within the preset time window, such as the driver continuously cruising in a straight line without changing lanes or turning, the driving operation sequence features cannot be extracted. In this case, the system can either set the feature vector components of that part to default values ​​and perform identification based solely on time-domain and frequency-domain features during similarity comparison; or it can skip the matching of driving operation sequence features and calculate the similarity score using only the remaining features.

[0111] S3: Compare the driver's behavior feature vector with the pre-stored driving behavior templates of multiple registered drivers, and obtain the corresponding similarity scores. The driving behavior templates represent the binding relationship between the behavior feature vector and identity information.

[0112] Preferably, the similarity comparison uses cosine similarity calculation. For the current driver's behavioral feature vector... and the pre-stored Mean vector of each driving behavior template The formula for calculating cosine similarity is:

[0113] in Represents the vector dot product. This represents the magnitude of the vector. The cosine similarity value ranges from... The closer the value is to 1, the more consistent the directions of the two vectors are, meaning the current driving style matches the template better.

[0114] Furthermore, when the scale differences among the feature components are large, Mahalanobis distance can be used to measure similarity. The formula for Mahalanobis distance is:

[0115] in and The first The mean vector and covariance matrix of each template. Mahalanobis distance considers the correlation between feature components and can more accurately assess the degree of matching; the smaller the distance, the higher the similarity. The system can convert Mahalanobis distance into a similarity score, for example, through a Gaussian kernel function or by taking the reciprocal normalization.

[0116] In addition, Euclidean distance can be used as a similarity metric to simplify calculations. Regardless of the metric used, the system ultimately outputs the similarity score between the current behavior feature vector and each template for use in subsequent decision-making steps.

[0117] S4: When there is at least one driving behavior template with a similarity score exceeding the first preset threshold and lasting for more than the first preset duration, the driver identity corresponding to the driving behavior template is determined as the current driver identity.

[0118] Preferably, the first preset threshold is set according to the system's requirements for recognition accuracy, for example, it can be set to 0.85 or the corresponding Mahalanobis distance threshold. The first preset duration is a stabilization period set to avoid misjudgment caused by instantaneous noise or accidental fluctuations, for example, it can be set to 10 seconds. During continuous monitoring, the system only confirms a successful match when the similarity score of a certain behavior template continuously exceeds the first preset threshold for the first preset duration; otherwise, monitoring continues or the timer is reset.

[0119] Furthermore, if multiple driving behavior templates simultaneously have similarity scores exceeding a first preset threshold and the duration meets the requirement, the system can determine the driver's identity corresponding to the template with the highest similarity score as the current driver's identity. A secondary arbitration can also be performed using an auxiliary channel to select the identity more consistent with visual characteristics or historical status.

[0120] By setting a first preset threshold and a first preset duration as dual conditions, this application effectively avoids misjudgments caused by instantaneous noise, brief driver distraction, or occasional control fluctuations. Driver identity is only confirmed when the similarity score consistently and stably exceeds the threshold for a preset time, thus significantly improving the robustness and reliability of the identification. Furthermore, when multiple templates simultaneously meet the conditions, the probability of identity confusion is further reduced by selecting the highest similarity score or combining it with an auxiliary channel for secondary arbitration. This determination mechanism ensures both the timeliness of identification and the high confidence of the determination results, providing an accurate identity basis for subsequent personalized control and safety decisions, while avoiding the poor user experience caused by frequent identity switching.

[0121] S5: Based on the current driver's identity, control at least one vehicle subsystem.

[0122] In some embodiments, based on the determined driver identity, a preset driver association parameter mapping table is queried to generate a corresponding control signal; Control signals are sent to at least one vehicle subsystem to control the corresponding vehicle subsystem.

[0123] Preferably, Table 1 is a driver-related parameter mapping table. As shown in Table 1, this mapping table associates the driver's identity with adjustable parameters of multiple vehicle subsystems.

[0124]

[0125] Table 1 Driver-related parameter mapping table In some embodiments, such as Figure 3 As shown, the vehicle control method of this application also includes an abnormal driving event detection step S7, which includes the following steps S71-S73.

[0126] S71: Calculate the mean of the preset sliding window of the driver behavior feature vector.

[0127] Preferably, the calculation method for the mean of the preset sliding window is to set the length of the sliding window as follows: For example, if we take 20 seconds, we can extract a behavioral feature vector within each window. The default sliding window mean is the arithmetic mean of the feature vectors of the most recent N windows:

[0128] Where c is the current window number. A sliding window averaging can smooth out transient noise and reflect short-term trends in driver handling style.

[0129] S72: When the change in the mean of the preset sliding window exceeds the preset change threshold within the preset time window, and no change in the driver's visual feature data is detected, it is determined to be an abnormal driving event.

[0130] Preferably, the magnitude of the change is measured using Euclidean distance.

[0131] in Set a preset time window (e.g., 20 seconds). This is the short-term average value at previous times. The preset change threshold can be set according to the statistical distribution of characteristic changes in historical data, for example, three times the standard deviation.

[0132] Meanwhile, the driver monitoring system continuously monitors the driver's visual characteristics, such as facial features, head posture, and sitting posture. If no facial changes or significant adjustments to sitting posture are detected, it indicates that the feature transition is not caused by a change of driver.

[0133] When both of the above conditions are met, it is determined to be an abnormal driving event, which may correspond to fatigue driving, a sudden change in driving style due to distraction, or temporary operation by an unregistered driver.

[0134] Furthermore, to avoid misjudgments caused by short-term fluctuations, the system can require the change amplitude to exceed a threshold and continue for more than a preset verification time before finally confirming the abnormal event, such as 5 seconds.

[0135] Through the aforementioned abnormal driving event detection mechanism, this application can identify sudden changes in driving style caused by driver changes, fatigue, or distraction in real time. Upon confirming the abnormality, it promptly issues a warning and switches to a safe mode, effectively avoiding driving risks caused by mismatched personalized parameters or poor driver condition, and significantly improving driving safety. Simultaneously, by combining driver change detection with exclusion of normal driver change scenarios, false alarms are reduced, improving the system's intelligence and user experience.

[0136] S73: In response to an abnormal driving event, generate a prompt message and / or switch at least one vehicle subsystem to the default safety mode.

[0137] Preferably, the prompt information can be output through text display on the instrument panel or voice broadcast to remind the driver to pay attention to the current status.

[0138] If the abnormal state persists for more than a preset duration, such as 30 seconds, the system will switch at least one vehicle subsystem to the default safety mode. The default safety mode uses conservative parameter settings to avoid safety risks caused by identity mismatch or abnormal driver status.

[0139] Furthermore, the system simultaneously records anomaly event logs, including the occurrence time, abnormal feature values, and current identified identity, for subsequent analysis and system optimization. If the DMS detects facial changes or the driver confirms their own driving via a button during an anomaly, the system can exit the anomaly mode and resume normal personalized adaptation. Additionally, if the mean feature value stabilizes automatically after a short period of anomaly, the system can automatically de-anomaly. Through these mechanisms, this application can promptly issue warnings and take safety measures when the driver's state changes abruptly, significantly improving driving safety.

[0140] In some embodiments, the vehicle control method of this application further includes a self-learning mechanism, which includes: When the similarity score exceeds the second preset threshold and continues to exceed the second preset time, the current driver's identity is determined to be a reliable identification; in response to the determination of high confidence identification, the driver behavior template corresponding to the current driver is updated with a preset weight based on the current driver's behavior feature vector; wherein the second preset threshold is greater than the first preset threshold, and the second preset time is longer than the first preset time.

[0141] Preferably, the second preset threshold can be, for example, 0.95, and the second preset time can be, for example, 30 seconds. The preset weight is a minimal weight, for example, a value of... This ensures that template updates are slow and gradual, preventing template distortion due to single anomalies or accidental fluctuations.

[0142] The update method uses an exponentially weighted moving average for the current driver's behavioral feature vector. and the mean vector of its existing template Covariance Matrix The updated formula is as follows:

[0143]

[0144] Furthermore, the self-learning mechanism is triggered only when the following conditions are met simultaneously: the system has identified the current driver with a high degree of confidence, no driver switching event or abnormal driving event has been detected, and the vehicle is in a stable driving state.

[0145] This mechanism allows driver behavior templates to track long-term, gradual changes in driver operating habits while maintaining template stability and recognition accuracy. Users can also manually reset the template or restore factory settings at any time via the vehicle's infotainment system menu. This self-learning process is completed locally on the vehicle, without transmitting raw data to the cloud, ensuring user privacy and security.

[0146] It should be noted that the steps shown in the above process or in the flowchart of the accompanying figures can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0147] This application also provides a vehicle, including a controller, which is used to execute the control method described in the above embodiments.

[0148] A controller is typically an integrated electronic control module containing a microprocessor, memory, input / output interfaces, and other components. It can receive data from various sensors and systems in real time and make corresponding control decisions based on preset algorithms and logic. Controllers are generally installed in the vehicle's cockpit or engine compartment to ensure a relatively safe location that facilitates electrical connections with other vehicle systems.

[0149] The vehicle provided in this embodiment is used to execute the control method described above, and therefore can achieve the same effect as the implementation method described above.

[0150] The beneficial effects of the above embodiments can be referred to the beneficial effects of the corresponding methods provided above, and will not be repeated here.

[0151] In addition, combined Figure 2 The vehicle control method described in this application embodiment can be implemented by an electronic device. Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application.

[0152] The electronic device may include a processor 41 and a memory 42 storing computer program instructions.

[0153] Specifically, the processor 41 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0154] The memory 42 may include a large-capacity memory for data or instructions. For example, and not limitingly, the memory 42 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 42 may include removable or non-removable (or fixed) media. Where appropriate, the memory 42 may be internal or external to a data processing device. In a particular embodiment, the memory 42 is non-volatile memory. In a particular embodiment, the memory 42 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), Extended Data Out Dynamic Random-Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0155] The memory 42 can be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 41.

[0156] The processor 41 implements any of the vehicle control methods described in the above embodiments by reading and executing computer program instructions stored in the memory 42.

[0157] In some embodiments, the electronic device may further include a communication interface 43 and a bus 40. For example, Figure 4 As shown, the processor 41, memory 42, and communication interface 43 are connected through bus 40 and complete communication with each other.

[0158] The communication interface 43 is used to enable communication between the various modules, units, and / or devices in the embodiments of this application. The communication interface 43 can also enable data communication with other components such as external devices, image / data acquisition devices, databases, external storage, and image / data processing workstations.

[0159] The electronic device can execute the vehicle control method in the embodiments of this application based on the acquired computer program instructions.

[0160] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0161] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for controlling a vehicle, characterized in that, The method includes: Collect the driver's steering wheel torque signal within a preset time window; Based on the steering wheel torque signal, multi-dimensional features characterizing the driver's driving style are extracted to obtain a driver behavior feature vector. The driver behavior feature vector is compared with the driving behavior templates of multiple pre-stored registered drivers to obtain the corresponding similarity scores. When there is at least one driving behavior template whose similarity score exceeds the first preset threshold and lasts for more than the first preset duration, the driver identity corresponding to the driving behavior template is determined as the current driver identity. Based on the current driver's identity, control at least one vehicle subsystem.

2. The vehicle control method according to claim 1, characterized in that, Also includes: Collect driver visual feature data. If the driver visual feature data does not match the driver visual feature data of multiple pre-stored registered drivers, then the current driver is determined to be an unregistered driver. In response to the current driver being an unregistered driver, driving guidance is provided during the preset driving period, and the steering torque signal of the steering wheel used for registration of the current driver is collected; Based on the steering torque signal of the registered steering wheel, the driver's registration behavior feature vector is extracted, and the driver's registration behavior feature vector is associated and bound with the driver's identity represented by the visual feature data to generate and store the current driver's driving behavior template.

3. The vehicle control method according to claim 1, characterized in that, The step of extracting multidimensional features characterizing the driver's driving style based on the steering wheel torque signal to obtain a driver behavior feature vector includes: Based on the steering wheel torque signal, at least one of the following types of features is extracted: time-domain features, frequency-domain features, and driving operation sequence features, to obtain the driver behavior feature vector.

4. The vehicle control method according to claim 3, characterized in that, The extraction of time-domain and frequency-domain features based on the steering wheel torque signal includes: Under straight-line cruising conditions, the mean, standard deviation, and peak-to-peak value of the steering wheel torque signal are calculated as the time-domain features; under steering event conditions, the rising slope, peak sustain value, and operating impulse of the steering wheel torque signal are calculated as the time-domain features; the operating impulse is used to characterize the cumulative effect of the steering wheel torque intensity and duration during the driver's steering operation. The steering wheel torque signal is subjected to spectral transformation to obtain the energy proportion of a preset frequency band, and / or the approximate entropy of the steering wheel torque signal is calculated as the frequency domain feature.

5. The vehicle control method according to claim 3, characterized in that, Extracting driving operation sequence features based on the steering wheel torque signal includes: Based on the steering wheel torque signal and steering wheel angle signal, a preset driving operation primitive type is identified. The driving operation primitive type is used to characterize the basic driving operation mode with identifiable start and end boundaries. Key point sequences are extracted from the temporal relationship between the steering torque and steering angle corresponding to the driving operation primitive type to obtain driving operation sequence features.

6. The vehicle control method according to claim 1, characterized in that, The control of at least one vehicle subsystem based on the current driver's identity includes: Based on the current driver's identity, a preset driver association parameter mapping table is queried to generate the corresponding control signal; The control signal is sent to at least one vehicle subsystem to control the corresponding vehicle subsystem.

7. The vehicle control method according to claim 1, characterized in that, Also includes: Calculate the mean of a preset sliding window of the driver behavior feature vector; When the mean value of the preset sliding window changes more than a preset change threshold within a preset time window, and no change in the driver's visual feature data is detected, it is determined to be an abnormal driving event. In response to the abnormal driving event, generate a prompt message and / or switch at least one vehicle subsystem to the default safety mode.

8. The vehicle control method according to any one of claims 1-7, characterized in that, Also includes: When the similarity score exceeds the second preset threshold and continues for more than the second preset duration, the current driver's identity is determined to be a reliable identification. In response to the current driver's identity being reliably identified, the driver behavior template corresponding to the current driver is updated with a preset weight based on the current driver's behavior feature vector; Wherein, the second preset threshold is greater than the first preset threshold, and the second preset duration is longer than the first preset duration.

9. A vehicle, characterized in that, Includes a controller for performing the control method as described in any one of claims 1 to 8.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the control method as described in any one of claims 1 to 8.

Images