Vehicle control method and vehicle
By generating vehicle control commands based on the target user's perception data, the problem of not considering individual user differences and behavioral intentions in existing technologies is solved, realizing personalized and comfortable intelligent assistance services and improving the user experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BYD CO LTD
- Filing Date
- 2026-02-03
- Publication Date
- 2026-04-21
AI Technical Summary
Existing technologies lack the ability to perceive individual user differences and behavioral intentions in vehicle storage spaces, resulting in an inability to provide personalized and comfortable intelligent assistance services and a poor user experience.
By generating vehicle control commands, based on the target user's perception data, including physiological and behavioral characteristics, the system identifies the user's intention to access items and assists the user in accessing items through vehicle control. It utilizes multimodal perception, skeletal posture analysis, and linkage control mechanisms to adapt to different user needs.
It provides personalized and comfortable intelligent assistance services, enhancing the user's experience of storing and retrieving items in the vehicle's storage space.
Smart Images
Figure CN121893889A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of vehicle control technology, and more specifically, to a vehicle control method and a vehicle. Background Technology
[0002] With the development of vehicle technology, vehicle control is also evolving towards intelligence. In an intelligent control scenario, when a user needs to store or retrieve items in the vehicle's storage space (such as the trunk), the vehicle can proactively control itself accordingly to meet the user's needs. Summary of the Invention
[0003] The purpose of this disclosure is to provide a vehicle control method and vehicle that can provide personalized and comfortable intelligent assistance services for users to access items, thereby enhancing the user experience.
[0004] To achieve the above objectives, in a first aspect, this disclosure provides a vehicle control method, comprising: generating a vehicle control command, the generation of which is based on perception data of a target user, the perception data including the target user's physiological and behavioral characteristics; and controlling the vehicle to assist the target user in storing or retrieving items in the vehicle's storage space.
[0005] Optionally, generating the vehicle control command includes: detecting whether the target user intends to access or retrieve items in the vehicle's storage space, wherein the detection of the intention is based on the target user's travel information and facial posture information; and generating the vehicle control command if the target user is detected to have the intention to access or retrieve items in the vehicle's storage space.
[0006] Optionally, acquiring the physiological characteristics includes: acquiring visual data corresponding to the target user; acquiring tracking and positioning data corresponding to the target user; determining the skeletal point positioning information of the target user, wherein the determination of the skeletal point positioning information includes the visual data; determining the skeletal point pose information of the target user, wherein the determination of the skeletal point pose information includes the skeletal point positioning information and the tracking and positioning data; and determining the physiological characteristics of the target user, wherein the determination of the physiological characteristics includes the skeletal point pose information.
[0007] Optionally, the generation of the vehicle control command may also be based on the difficulty of the target user accessing and retrieving items in the vehicle's storage space. The vehicle control method may further include: recognizing the target user's accessing and retrieving actions, the recognition of which is based on the skeletal point posture information; and determining the difficulty of the target user accessing and retrieving items in the vehicle's storage space, the determination of which is based on the accessing and retrieving actions and the physiological characteristics.
[0008] Optionally, the vehicle control command is generated based on user habit data, and the vehicle control method further includes: obtaining user feedback information corresponding to historical item access operations; generating user habit data, wherein the user habit data is generated based on the user feedback information.
[0009] Optionally, the vehicle control commands include: chassis height control commands and vehicle storage space status control commands, and controlling the vehicle includes: controlling the chassis height of the vehicle and controlling the status of the vehicle storage space.
[0010] Optionally, the chassis height control command includes a target chassis height; the vehicle storage space status control command includes a target status; the vehicle control method further includes: acquiring monitoring information, the monitoring information being used to characterize whether the vehicle chassis height has reached the target chassis height, and / or whether the vehicle storage space status has reached the target status; if the monitoring information characterizes that the vehicle chassis height has not reached the target chassis height, and / or the vehicle storage space status has not reached the target status, regenerating the vehicle control command; and re-controlling the vehicle.
[0011] Optionally, the vehicle control basis further includes risk prediction information for storing and retrieving items in the vehicle's storage space. The risk prediction information is used to characterize whether there is a safety risk in storing and retrieving items in the vehicle's storage space. The vehicle control method further includes: acquiring perception data corresponding to the vehicle's surrounding environment; and determining the risk prediction information, wherein the basis for determining the risk prediction information includes the perception data corresponding to the vehicle's surrounding environment.
[0012] Optionally, generating vehicle control commands includes: at least inputting the perception data corresponding to the target user into a pre-trained decision model to obtain control parameters output by the pre-trained decision model; and converting the control parameters into vehicle control commands.
[0013] In a second aspect, this disclosure provides a vehicle, including: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the executable instructions to implement the vehicle control method as described in the first aspect of this disclosure.
[0014] The above technical solution generates vehicle control commands based on the perceived data of the target user. This vehicle control then assists the target user in storing and retrieving items from the vehicle's storage space. Since the perceived data includes physiological and behavioral characteristics, the intelligent vehicle control is executed based on individual user differences and behavioral intentions. This allows for adaptation to different user needs, providing personalized and comfortable intelligent assistance services for storing and retrieving items, thus enhancing the user experience.
[0015] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description
[0016] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings:
[0017] Figure 1 This is a flowchart illustrating a vehicle control method according to an exemplary embodiment.
[0018] Figure 2 This is a flowchart illustrating an intelligent auxiliary control according to an exemplary embodiment.
[0019] Figure 3 This is a block diagram illustrating a vehicle control device according to an exemplary embodiment.
[0020] Figure 4 This is a functional block diagram of a vehicle according to an exemplary embodiment. Detailed Implementation
[0021] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.
[0022] As described in the background section, in an intelligent control scenario, when a user has a need to store or retrieve items in the vehicle's storage space (such as the trunk), the vehicle can actively control the vehicle accordingly to meet the user's needs for storing or retrieving items.
[0023] In related technologies, intelligent vehicle control relies on the vehicle's internal state (such as whether there are items in the trunk). This control method ignores individual user differences and behavioral intentions. Specifically, users of different heights, body types, or mobility levels (such as the elderly, children, or people with disabilities) have significantly different needs when retrieving and placing luggage. For example, tall users may be able to easily retrieve items without lowering the chassis, while shorter users or those who have difficulty bending over are more sensitive to chassis height and tailgate opening position.
[0024] Therefore, the relevant technologies lack the ability to perceive users' physiological characteristics and postures, making it difficult to adapt to different user needs. Consequently, they cannot provide personalized and comfortable intelligent assistance services for users to access items, resulting in a poor user experience.
[0025] Based on this, this disclosure provides a technical solution that generates vehicle control commands based on perception data of the target user, thereby assisting the target user in storing and retrieving items in the vehicle's storage space through vehicle control. Since the target user's perception data includes physiological and behavioral characteristics, this is equivalent to executing intelligent vehicle control based on individual user differences and behavioral intentions, which can adapt to different user needs. This provides personalized and comfortable intelligent assistance services for users storing and retrieving items, enhancing the user experience.
[0026] Figure 1 This is a flowchart illustrating a vehicle control method according to an exemplary embodiment, such as... Figure 1 As shown, the vehicle control method includes the following steps: Step S11: Generate vehicle control commands. The generation of vehicle control commands is based on the perception data of the target user, which includes the physiological and behavioral characteristics of the target user.
[0027] Step S12: Control the vehicle to assist the target user in storing and retrieving items in the vehicle's storage space.
[0028] Regarding vehicle storage space (vehicle storage space), it can refer to spaces with storage functions such as the vehicle's trunk and front trunk, and there is no limitation here.
[0029] In step S11, it can first detect whether the target user intends to access or retrieve items in the vehicle's storage space. If the target user's intention to access or retrieve items in the vehicle's storage space is detected, then a vehicle control command can be generated.
[0030] In one implementation, whether a user intends to access or retrieve items in the vehicle's storage space can be detected based on information that characterizes the user's intent.
[0031] Therefore, as an optional implementation, step S11 includes: detecting whether the target user has the intention to access or retrieve items in the vehicle's storage space, the detection basis of which includes the target user's travel information and facial posture information; and generating a vehicle control command if the target user is detected to have the intention to access or retrieve items in the vehicle's storage space.
[0032] In one implementation, the target user's travel information may include the user's travel path, travel speed, etc.
[0033] In one implementation, the facial posture information of the target user may include: facial gaze focus direction, facial orientation, etc.
[0034] As an example, situations such as the user's walking path being towards the vehicle's storage space, slowing down their walking speed, focusing their facial gaze towards the vehicle's storage space, or having their face facing the vehicle's storage space can be considered as indicating that the target user intends to access or retrieve items from the vehicle's storage space.
[0035] In one implementation, judgment conditions related to travel information and facial posture information can be preset, such as travel information judgment conditions and facial posture judgment conditions. Then, based on the currently acquired travel information and facial posture information, a judgment is made according to the corresponding judgment conditions. If the requirements of the judgment conditions are met, it can be regarded as the target user having the intention to store or retrieve items in the vehicle's storage space.
[0036] In one implementation, the basis for detecting intent may include not only travel information and facial posture information, but also the target user's voice information. This voice information can be detected by a microphone installed in the vehicle's storage space. If the intensity of the target user's voice information is detected to increase from low to high, it indicates that the target user is approaching the vehicle's storage space and can be regarded as having the intent to access or retrieve items in the vehicle's storage space.
[0037] In one implementation, the target user can be a user located near the vehicle's storage space. The target user can be detected first, and then it can be detected whether the target user has the intention to store or retrieve items in the vehicle's storage space.
[0038] In one implementation, target user detection, as well as the detection of target user's movement information and facial posture information, can be achieved through target tracking and trajectory prediction algorithms (models).
[0039] In one implementation, the vehicle's perception layer can provide a data foundation for target tracking and trajectory prediction algorithms (models).
[0040] As an example, vehicles can employ a dual-source heterogeneous sensor fusion strategy. On one hand, a high-definition wide-angle camera positioned in the corresponding location of the vehicle's storage space (e.g., the tailgate) continuously collects visual image streams of the storage space area (e.g., the trunk area) to capture facial features, body movements, and spatial positions of those approaching the user. On the other hand, a millimeter-wave radar integrated in the corresponding location of the vehicle's storage space (e.g., the rear bumper) scans for dynamic targets within a 3-5 meter range behind the vehicle using high-frequency pulses, acquiring their distance, speed, azimuth, and micro-motion characteristics. These two types of data are then fused spatiotemporally through timestamp alignment and spatial coordinate transformation by the vehicle's domain controller, enabling stable identification of moving human targets even in complex environments (such as nighttime, rain, fog, and strong light interference).
[0041] When the fusion algorithm detects a pedestrian moving towards the vehicle's storage space (e.g., towards the tailgate), it initiates a human behavior recognition process. Using target tracking and trajectory prediction models, it determines whether the user has a clear intention to access or retrieve items. For example, in the case of the trunk, if the path points towards the trunk area, the vehicle slows down, and the user's gaze is focused on the tailgate, it is considered an intention to access or retrieve items. If no clear intention to access or retrieve items is detected, the vehicle remains in low-power standby mode to avoid false triggers; once a clear intention to retrieve items is confirmed, vehicle control commands are generated.
[0042] In one implementation, the vehicle control method can be viewed as an intelligent assistance solution for accessing items. Its execution side can be the vehicle's intelligent assistance module. Therefore, in one scenario, after the vehicle is powered on or remote unlocking is triggered, the vehicle control system can initiate a self-test program and activate the intelligent assistance module. The core component of this intelligent assistance module can be a multimodal sensing unit. Upon activation, this multimodal sensing unit immediately enters a standby state, ready to receive real-time sensing data from vehicle cameras and millimeter-wave radar, etc.
[0043] In one implementation, the intention to store or retrieve items in the vehicle's storage space can be, for example, to store, retrieve, or adjust the position of items in the vehicle's storage space, and is not limited thereto.
[0044] In one implementation, the physiological characteristics of the target user, such as height, body proportions (e.g., leg length / torso ratio), and walking stability (e.g., gait symmetry, center of gravity fluctuation), can be obtained through skeletal posture analysis.
[0045] Therefore, as an optional implementation method, the acquisition of physiological characteristics includes: acquiring visual data corresponding to the target user; acquiring tracking and positioning data corresponding to the target user; determining the skeletal point positioning information of the target user, the determination of the skeletal point positioning information including the visual data; determining the skeletal point pose information of the target user, the determination of the skeletal point pose information including the skeletal point positioning information and the tracking and positioning data; and determining the physiological characteristics of the target user, the determination of the physiological characteristics including the skeletal point pose information.
[0046] In one implementation, visual data may involve images, videos, and other data captured by a camera.
[0047] In one implementation, the tracking and positioning data may involve data acquired by millimeter-wave radar, such as depth information, distance information, azimuth information, etc.
[0048] In one implementation, skeletal point positioning information can be determined based on the visual data corresponding to the target user.
[0049] In one implementation, key skeletal points can be located in each image frame of the target user based on a lightweight deep neural network (such as the MobileNetV3-Hourglass structure). The key skeletal points may include: head, shoulder, elbow, wrist, hip, knee, ankle, etc.
[0050] In one implementation, lightweight deep neural networks can be deployed on in-vehicle edge computing units such as NPUs (Neural Processing Units) or AI (Artificial Intelligence) acceleration chips to improve the efficiency and accuracy of skeletal point localization.
[0051] In one implementation, the skeletal point positioning information may include the coordinates of each key skeletal point.
[0052] In one implementation, the pose of the skeletal points can be further analyzed by combining skeletal point positioning information and tracking positioning data.
[0053] As an example, based on depth information and skeletal point positioning information in the tracking and positioning data, a user's three-dimensional pose skeleton is reconstructed, which can serve as skeletal point pose information.
[0054] Furthermore, based on the skeletal point pose information, the physiological characteristic parameters of the target user can be calculated. These physiological characteristic parameters can be used for coarse identity classification (such as adult / child / elderly), and thus can serve as the basis for generating vehicle control commands.
[0055] In the embodiments disclosed herein, the reconstruction method of the three-dimensional posture skeleton and the calculation method of physiological characteristic parameters based on the three-dimensional posture skeleton can be referred to the mature technology in the field, and will not be described in detail here.
[0056] In one implementation, the behavioral characteristics of the target user may be related to the target user's actions of accessing or storing items.
[0057] The actions of a target user in storing or retrieving items can be determined based on the posture information of skeletal points.
[0058] As an example, temporal modeling is performed using the motion trajectories of skeletal points in consecutive video frames to identify preparatory actions for storing or retrieving items, such as leaning forward, raising one hand to touch the tailgate handle, or preparing to put a bag in with the other hand. The temporal modeling can employ either an LSTM or a Transformer architecture; neither is specifically chosen here.
[0059] In one implementation, the behavioral characteristics of the target user can be the target user's actions of accessing or storing items, or the motion trajectory of skeletal points in continuous video frames, and are not limited here.
[0060] In one implementation, vehicle control commands are generated based at least on the physiological and behavioral characteristics of the target user. Different physiological and behavioral characteristics correspond to different control parameters.
[0061] Therefore, generating vehicle commands may include: at least inputting the perception data corresponding to the target user into a pre-trained decision model to obtain the control parameters output by the pre-trained decision model; and converting the control parameters into vehicle control commands.
[0062] In this implementation, the perception data corresponding to the target user is used as the basis for the decision of control parameters, and the corresponding control parameters can be obtained. Then, these parameters can be converted into vehicle control commands.
[0063] In one implementation, the pre-trained decision model can be trained based on historical data. This historical data may include historical sensing data and corresponding real control parameters, which can be control parameters that meet user requirements.
[0064] In one implementation, the generation of vehicle control commands may also be based on the difficulty of the target user accessing items in the vehicle's storage space.
[0065] In one implementation, the difficulty of accessing and storing items can be divided into three levels: "low", "medium", and "high".
[0066] In one implementation, the difficulty of accessing or retrieving items can be assessed based on the actions involved in accessing or retrieving items and the user's physiological characteristics.
[0067] Therefore, as an optional implementation, the vehicle control method further includes: identifying the target user's actions of storing and retrieving items, the identification of which is based on skeletal point posture information; and determining the operational difficulty of the target user storing and retrieving items in the vehicle's storage space, the determination of which is based on the actions of storing and retrieving items and physiological characteristics.
[0068] The method for recognizing the action of storing and retrieving items can be referred to the description in the foregoing embodiments, and will not be repeated here.
[0069] In one implementation, the difficulty of accessing or retrieving items can be further determined by combining the actions and physiological characteristics of accessing or retrieving items.
[0070] As an example, a multi-dimensional scoring model can be designed. This model comprehensively considers factors such as the user's height, posture (e.g., obesity or spinal problems causing difficulty bending over), gait stability (e.g., detected limping or swaying), and movement hesitation (e.g., slow initiation of movement, multiple pauses), classifying the difficulty of retrieving items into three levels: "low," "medium," and "high." For example, an elderly person who is 155cm tall and has a slow gait, approaching a vehicle on a rainy day, would be judged as having "high difficulty," regardless of their slow and small movements when retrieving items.
[0071] In one implementation, the generation of vehicle control commands may further include: user habit data, which can characterize the user's preferences in past operations of accessing and storing items.
[0072] Therefore, the vehicle control method may further include: acquiring user feedback information corresponding to historical item access operations; generating user habit data, the user habit data being generated based on user feedback information.
[0073] As an example, the system accesses a database of user history behavior stored in the in-vehicle edge computing unit. This database employs a localized encrypted storage mechanism to ensure privacy compliance. The database records the preference data (user feedback information) of each authenticated user (bound via Bluetooth key or facial recognition) in several past item retrieval operations. Examples include: the degree to which the most frequently used vehicle storage spaces are opened (e.g., tailgate opening height), whether the user prefers the vehicle storage spaces (e.g., tailgate) to be fully open or partially open while hovering, their tolerance for chassis descent, and their response latency tolerance.
[0074] By authenticating the target user's identity, identifying the corresponding identity information, and obtaining the feedback information corresponding to that identity information, user habit data can be generated.
[0075] It is understandable that the information recorded in the database may be unorganized and unstatistical. By summarizing, organizing, and statistically analyzing this information, corresponding user habit data can be generated.
[0076] In one implementation, the vehicle control generation command may include not only the various information described in the embodiments of this disclosure, but also more information, such as voice information, command information, etc., which are not limited herein.
[0077] Furthermore, the input to the pre-trained decision model may also include some process data, such as skeleton positioning data and skeleton pose information, which are not limited here.
[0078] Table 1 is an example of input data for a decision model. Referring to Table 1, the input data for a decision model can involve different categories of data, such as physiological characteristics, posture and action information, behavioral and dynamic characteristics, interaction needs, and user preferences. Furthermore, under different categories, multiple different types of data can be involved.
[0079]
[0080] Table 1 In one implementation, the vehicle control commands may include: chassis height control commands and vehicle storage space status control commands.
[0081] This implementation method allows for coordinated control of both the vehicle chassis and the vehicle storage space actuators, thereby enhancing the effectiveness of intelligent assistance.
[0082] Furthermore, controlling the vehicle may include controlling the vehicle's chassis height and controlling the status of the vehicle's storage space.
[0083] It is understandable that sending the chassis height control command to the chassis control actuator can achieve chassis height control, and sending the vehicle storage space status control command to the vehicle storage space control actuator can achieve vehicle storage space status control.
[0084] In one embodiment, the control actuator of the chassis can be an air suspension system, or a control actuator for the vehicle storage space, such as a motor, drive system, etc.
[0085] Taking the trunk as an example of vehicle storage space, for users who find retrieving items difficult, the chassis height will not only be lowered to the corresponding safety limit, but the power tailgate will also open at a greater angle, and its hovering height will be raised by 5-10cm, making the trunk sill closer to the user's waist and minimizing the need to bend over. For taller users, even if retrieving items is "easy," the chassis height may be slightly adjusted or left at the default setting to avoid unnecessary risks to ground clearance caused by lowering the chassis.
[0086] In one embodiment, the air suspension can smoothly lower or raise the vehicle height by adjusting the pressure of each airbag, with a response time controlled within 2 to 3 seconds; and the electric tailgate motor can open to the target angle according to a preset curve and can be precisely hovered at any position.
[0087] In some scenarios, the generated vehicle control commands can be executed directly, while in others, the safety of the surrounding environment needs to be considered, and execution should be postponed.
[0088] Considering the special nature of some scenarios, vehicle control can be based not only on vehicle control commands but also on risk prediction information regarding the storage and retrieval of items in the vehicle's storage space. This risk prediction information is used to characterize whether there is a safety risk in storing or retrieving items in the vehicle's storage space. Combining this risk prediction information with vehicle control commands can improve the safety of storing and retrieving items.
[0089] Therefore, the vehicle control method may further include: acquiring perception data corresponding to the vehicle's surrounding environment; determining risk prediction information, wherein the determination of risk prediction information includes the perception data corresponding to the vehicle's surrounding environment.
[0090] In one implementation, the perception data corresponding to the vehicle's surrounding environment can be environmental perception data within a certain range of the vehicle's storage space, for example, within a fan-shaped area with a radius of 3m centered on the location of the vehicle's storage space.
[0091] In one embodiment, the perception data corresponding to the vehicle's surrounding environment may include obstacle perception data, which can characterize whether obstacles exist in the surrounding environment.
[0092] Based on the perception data corresponding to the vehicle's surrounding environment, it is possible to assess whether there is a safety risk. For example, if an obstacle is approaching from behind and may affect the safety of the current act of accessing or storing items, then it can be determined that there is a safety risk.
[0093] It is understandable that obstacle perception and safety risk assessment are relatively mature technologies in this field, and will not be described in detail here.
[0094] In one implementation, when the risk prediction information indicates that there is no safety risk in storing or retrieving items in the vehicle's storage space, the vehicle control command can be directly sent to the corresponding actuator to execute the command. Alternatively, if the risk prediction information indicates that there is a safety risk in storing or retrieving items in the vehicle's storage space, the vehicle control command can be withheld until the risk prediction information indicates that there is no safety risk, at which point the vehicle control command can be issued.
[0095] In one implementation, the vehicle's status can be monitored in real time during vehicle control, and corresponding feedback control can be executed based on the real-time monitoring results.
[0096] In one implementation, the chassis height control command includes a target chassis height, and the vehicle storage space status control command includes a target status. The target status may include: target opening height, target response delay duration, etc.
[0097] Therefore, the vehicle control method may further include: acquiring monitoring information, which is used to characterize whether the vehicle chassis height has reached the target chassis height and / or whether the vehicle storage space status has reached the target status; if the monitoring information indicates that the vehicle chassis height has not reached the target chassis height and / or the vehicle storage space status has not reached the target status, regenerating the vehicle control command; and re-controlling the vehicle.
[0098] In this implementation, the vehicle chassis height and the status of the vehicle storage space can be monitored in real time. If the monitoring results indicate that the control results are not up to standard, for example, the vehicle chassis height is not up to standard due to ground slope or the status of the vehicle storage space is not up to standard due to external objects, the vehicle control command can be regenerated.
[0099] In one implementation, in addition to intuitive feedback control of whether the vehicle chassis height and vehicle storage space status meet the standards, the system can also assess whether the user's actions of storing and retrieving items are completed within a set time, and whether the system is properly packaged, etc., without limitation here.
[0100] In one implementation, the target chassis height and the target vehicle storage space status can be fine-tuned in the regenerated vehicle control commands to adapt to the actual scenario.
[0101] Furthermore, when regaining control of the vehicle, the control basis may include newly generated vehicle control commands, as well as real-time risk prediction information, etc. For specific control methods, please refer to the aforementioned embodiments, which will not be repeated here.
[0102] Furthermore, after the user has finished storing or retrieving items, the system can actively control the vehicle's storage space and restore the vehicle chassis to its original state, thereby completing the intelligent assisted operation for storing and retrieving items.
[0103] In this embodiment of the disclosure, after each intelligent assistance operation for a user to access or retrieve items is completed, various data of this intelligent assistance operation can be recorded, and based on this data, decision models and other data can be learned and updated online.
[0104] As an example, data such as raw sensory input, skeletal pose sequences, decision logs, execution trajectories, and subsequent user feedback (e.g., whether the tailgate was manually adjusted) are encrypted and stored in a local database. The edge computing unit periodically initiates an incremental learning process, utilizing online learning algorithms (such as the FedAvg lightweight learning algorithm) to update the decision model based on this data. This allows the intelligent assistance module to gradually adapt to changing user behavior trends; for example, if a user experiences a short-term decline in mobility due to injury, the module can automatically identify this and increase the assistance intensity after several uses. Finally, the entire intelligent assistance module returns to standby mode, awaiting the next trigger, forming a complete operational loop.
[0105] Figure 2 This is a flowchart illustrating an intelligent auxiliary control according to an exemplary embodiment, such as... Figure 2 As shown, the scenario for this intelligent auxiliary control process is the trunk (tailgate) and chassis auxiliary control, that is, the scenario of storing and retrieving items in the trunk. This intelligent auxiliary control process may involve the following steps: After intelligent assistance is activated, it first detects whether a user is approaching. If no user approach is detected, it continues to detect user approach.
[0106] If a user is detected approaching, their dynamic and static characteristics can be analyzed, and various relevant information that can assist in decision-making can be obtained.
[0107] By integrating various information, a corresponding chassis and tailgate linkage control strategy can be generated and further executed.
[0108] During execution, parameters can be dynamically fine-tuned based on feedback data indicating whether the control results are in place, or auxiliary data can be recorded.
[0109] Finally, the decision-making model can be dynamically updated to improve the accuracy of subsequent applications.
[0110] The technical solution of this disclosure improves the intelligent assistance effect of vehicles in the scenario of storing and retrieving items by means of multimodal perception fusion, skeletal posture analysis, linkage control mechanism and localized learning, thereby effectively improving the user experience.
[0111] Figure 3 This is a block diagram illustrating a vehicle control device 300 according to an exemplary embodiment, such as... Figure 3 As shown, the vehicle control device 300 includes: The generation module 301 is used to generate vehicle control commands. The generation of the vehicle control commands is based on the perception data of the target user, which includes the physiological and behavioral characteristics of the target user.
[0112] The control module 302 is used to control the vehicle to assist the target user in storing and retrieving items in the vehicle's storage space.
[0113] Optionally, the generation module 301 is further configured to: detect whether the target user has the intention to access or retrieve items in the vehicle's storage space, the detection basis of which includes the target user's travel information and facial posture information; and generate the vehicle control command if the target user is detected to have the intention to access or retrieve items in the vehicle's storage space.
[0114] Optionally, the device further includes: a physiological feature acquisition module, configured to: acquire visual data corresponding to the target user; acquire tracking and positioning data corresponding to the target user; determine skeletal point positioning information of the target user, wherein the determination of the skeletal point positioning information includes the visual data; determine skeletal point pose information of the target user, wherein the determination of the skeletal point pose information includes the skeletal point positioning information and the tracking and positioning data; and determine physiological features of the target user, wherein the determination of the physiological features includes the skeletal point pose information.
[0115] Optionally, the device further includes: an operation difficulty determination module, used to: identify the target user's actions of storing and retrieving items, the identification basis of the actions of storing and retrieving items including the skeletal point posture information; and determine the operation difficulty of the target user storing and retrieving items in the vehicle storage space, the determination basis of the operation difficulty including the actions of storing and retrieving items and the physiological characteristics.
[0116] Optionally, the generation module 301 is further configured to: obtain user feedback information corresponding to historical item access operations; generate user habit data, wherein the generation of the user habit data includes the user feedback information.
[0117] Optionally, the control module 302 is also used to: control the chassis height of the vehicle and control the state of the vehicle's storage space.
[0118] Optionally, the generation module 301 is further configured to: determine whether the chassis height and / or the vehicle storage space status have reached the target status; and, if the monitoring information indicates that the vehicle chassis height has not reached the target chassis height and / or the vehicle storage space status has not reached the target status, regenerate the vehicle control command; the control module 302 is further configured to: re-control the vehicle.
[0119] Optionally, the device further includes a risk prediction module, used to: acquire perception data corresponding to the vehicle's surrounding environment; and determine the risk prediction information, wherein the determination of the risk prediction information is based on the perception data corresponding to the vehicle's surrounding environment.
[0120] Optionally, the generation module 301 is further configured to: input at least the perception data corresponding to the target user into the pre-trained decision model to obtain the control parameters output by the pre-trained decision model; and convert the control parameters into vehicle control commands.
[0121] Figure 4 This is a functional block diagram of a vehicle 400 according to an exemplary embodiment. The vehicle 400 may include various subsystems, such as an infotainment system 410, a perception system 420, a decision control system 430, a drive system 440, and a computing platform 450. The vehicle 400 may also include more or fewer subsystems, and each subsystem may include multiple components. Furthermore, each subsystem and each component of the vehicle 400 can be interconnected via wired or wireless means.
[0122] In some embodiments, the infotainment system 410 may include a communication system, an entertainment system, and a navigation system, etc.
[0123] The perception system 420 may include several sensors for sensing information about the environment surrounding the vehicle 400. For example, the perception system 420 may include a global positioning system (which may be GPS, BeiDou, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.
[0124] The decision control system 430 may include a computing system, a vehicle controller, a steering system, a throttle, and a braking system.
[0125] The drive system 440 may include components that provide powered motion to the vehicle 400. In one embodiment, the drive system 440 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of internal combustion engines, electric motors, and compressed air engines. The engine is capable of converting energy provided by the energy source into mechanical energy.
[0126] Some or all of the functions of the vehicle 400 are controlled by a computing platform 450. The computing platform 450 may include at least one processor 451 and a memory 452, the processor 451 being able to execute instructions 453 stored in the memory 452.
[0127] Processor 451 can be any conventional processor, such as a commercially available CPU. Processors may also include graphics processing units (GPUs), field-programmable gate arrays (FPGAs), systems-on-chips (SoCs), application-specific integrated circuits (ASICs), or combinations thereof.
[0128] The memory 452 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.
[0129] In addition to instruction 453, memory 452 can also store data, such as road maps, route information, vehicle position, direction, speed, and other data. The data stored in memory 452 can be used by computing platform 450.
[0130] In this embodiment of the disclosure, the processor 451 may execute instructions 453 to complete all or part of the steps of the vehicle control method described above.
[0131] In another exemplary embodiment, a controller is also provided, which may be part of the aforementioned vehicle. The controller may be an integrated circuit (IC) or a chip, wherein the integrated circuit may be a single IC or a collection of multiple ICs; the chip may include, but is not limited to, the following types: GPU (Graphics Processing Unit), CPU (Central Processing Unit), FPGA (Field Programmable Gate Array), DSP (Digital Signal Processor), ASIC (Application Specific Integrated Circuit), SOC (System on Chip), etc. The aforementioned integrated circuit or chip may be used to execute executable instructions (or code) to implement the aforementioned vehicle control method. The executable instructions may be stored in the integrated circuit or chip, or obtained from other devices or equipment; for example, the integrated circuit or chip may include a processor, memory, and an interface for communicating with other devices. The executable instructions can be stored in the memory, and when the executable instructions are executed by the processor, the above-mentioned control method for rail transit vehicles can be implemented; or, the integrated circuit or chip can receive the executable instructions through the interface and transmit them to the processor for execution to implement the above-mentioned vehicle control method.
[0132] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.
[0133] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.
[0134] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.
Claims
1. A vehicle control method, characterized in that, include: The vehicle control command is generated based on the perception data of the target user, which includes the physiological and behavioral characteristics of the target user. The vehicle is controlled to assist the target user in accessing items in the vehicle's storage space.
2. The vehicle control method according to claim 1, characterized in that, The generation of vehicle control commands includes: The system detects whether the target user intends to access or retrieve items in the vehicle's storage space. The detection of this intention is based on the target user's movement information and facial posture information. The vehicle control command is generated when the system detects that the target user intends to access or retrieve items in the vehicle's storage space.
3. The vehicle control method according to claim 1, characterized in that, The acquisition of the physiological characteristics includes: Obtain the visual data corresponding to the target user; Obtain the tracking and location data corresponding to the target user; Determine the skeletal point positioning information of the target user, wherein the determination of the skeletal point positioning information includes the visual data; Determine the skeletal point pose information of the target user, wherein the determination of the skeletal point pose information is based on the skeletal point positioning information and the tracking positioning data; The physiological characteristics of the target user are determined, and the determination of the physiological characteristics is based on the skeletal point pose information.
4. The vehicle control method according to claim 3, characterized in that, The generation of the vehicle control command also includes the difficulty of the target user accessing items in the vehicle's storage space, and the vehicle control method further includes: The system identifies the target user's actions of accessing and storing items, and the identification of these actions is based on the skeletal point pose information. The difficulty of the target user accessing and retrieving items in the vehicle's storage space is determined, and the determination of the difficulty is based on the actions of accessing and retrieving items and the physiological characteristics.
5. The vehicle control method according to claim 1, characterized in that, The generation of the vehicle control commands also includes: user habit data, and the vehicle control method further includes: Get user feedback information corresponding to historical item access operations; User habit data is generated, and the generation of user habit data includes the user feedback information.
6. The vehicle control method according to claim 1, characterized in that, The vehicle control commands include: chassis height control commands and vehicle storage space status control commands; controlling the vehicle includes: The vehicle's chassis height and the state of its storage space are controlled.
7. The vehicle control method according to claim 6, characterized in that, The chassis height control command includes a target chassis height; the vehicle storage space status control command includes a target status. The vehicle control method further includes: Acquire monitoring information, which is used to characterize whether the vehicle chassis height has reached the target chassis height, and / or whether the vehicle storage space status has reached the target status; If the monitoring information indicates that the vehicle chassis height has not reached the target chassis height, and / or the vehicle storage space status has not reached the target status, the vehicle control command will be regenerated. Take control of the vehicle again.
8. The vehicle control method according to claim 1, characterized in that, The vehicle control method also includes risk prediction information for storing and retrieving items in the vehicle's storage space. This risk prediction information is used to characterize whether there is a safety risk associated with storing and retrieving items in the vehicle's storage space. Acquire perception data corresponding to the environment surrounding the vehicle; The risk prediction information is determined based on the perception data corresponding to the vehicle's surrounding environment.
9. The vehicle control method according to any one of claims 1 to 8, characterized in that, The generation of vehicle control commands includes: At least the perception data corresponding to the target user is input into the pre-trained decision model to obtain the control parameters output by the pre-trained decision model; The control parameters are converted into vehicle control commands.
10. A vehicle, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to execute the executable instructions to implement the vehicle control method as described in any one of claims 1 to 9.