Intelligent vehicle central control methods, devices, equipment, media and software products

By extracting unified features and fusing them at the member level from multi-source perception information within the vehicle, and combining the vehicle's motion state and historical preferences, the interaction methods and control strategies of the central control system are dynamically adjusted. This solves the safety and user experience problems of existing in-vehicle central control systems in complex scenarios, and achieves more efficient driving safety and personalized services.

CN123113597APending Publication Date: 2026-11-17SHENZHEN BLUEWHALE INTEL-CONNECTIVITY S&T CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611472191.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-21
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing in-vehicle central control systems lack unified scene understanding and collaborative decision-making capabilities based on multi-source perception information. This results in the inability to dynamically adjust interaction methods and control strategies when different driving scenarios or changes in occupant status occur, affecting driving safety and user experience.

Method used

By performing unified feature extraction and member-level fusion of multi-source perception information in the vehicle, identifying the current driving scenario in conjunction with the vehicle's motion state, adjusting the response priority of central control operations according to different scenarios, and coordinating the configuration of interactive content and control strategies in conjunction with members' historical preferences and safety rules.

Benefits of technology

It enables interactive operations that avoid safety hazards in complex driving scenarios, improves the matching degree between the central control system's interaction strategy and entertainment content configuration, ensures driving safety and content compliance, and enhances driving safety, interaction efficiency, and user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113597A_ABST
    Figure CN123113597A_ABST
Patent Text Reader

Abstract

This application relates to the field of vehicle infotainment control technology, providing an intelligent in-vehicle central control method, device, equipment, medium, and program product. It obtains a subset of member features by extracting and classifying perceptual data frames. Based on vehicle motion state data and the subset of member features, it determines the current driving scenario type. Then, based on the driving scenario type and the subset of member features, it evaluates the response priority of central control operation commands to generate a control command sequence. Finally, it matches historical preference data from a historical database based on the subset of member features. Based on safety rules and historical preference data, it performs compliance screening and interaction configuration of the control command sequence to generate a set of central control interaction adjustment parameters and entertainment content configuration data. This application improves the matching degree of the central control system's interaction strategy and entertainment content configuration to the current scenario by associating personnel status, vehicle status, and historical preferences in decision-making, thus balancing driving safety, interaction efficiency, and user experience.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to an intelligent vehicle central control method, device, equipment, medium and program product. Background Technology

[0002] With the development of intelligent cockpit and vehicle-to-everything (V2X) technologies, in-vehicle central control systems have evolved from traditional audio-visual playback and navigation control terminals into comprehensive interactive platforms integrating vehicle control, human-machine interaction, entertainment services, and intelligent connectivity. In practical applications, in-vehicle central control systems not only need to respond to the diverse operational needs of drivers and passengers, but also need to dynamically adjust interaction methods, entertainment content, and system functions based on vehicle operating status, passenger composition, driving behavior, and changes in the cabin environment, to balance driving safety, passenger experience, and intelligent service requirements. Simultaneously, with the widespread adoption of smart home devices and vehicle-to-home connectivity applications, in-vehicle central control systems also undertake the important task of collaborative control between vehicles and smart homes. Therefore, they need to possess multimodal perception, scene recognition, and personalized interactive configuration capabilities to adapt to various application scenarios such as family cars, business travel, and shared mobility.

[0003] Existing in-vehicle central control methods typically rely on single interaction methods such as touch, physical buttons, or voice as the primary input means. Upon receiving user input, they directly execute corresponding functions according to preset rules. Some systems can combine limited vehicle operating information, such as vehicle speed and parking status, to perform simple start-stop control for certain functions. Other systems can recommend entertainment resources based on user history or achieve basic vehicle-to-home connectivity control based on vehicle location. However, these methods lack a unified correlation analysis between driver status, passenger type, vehicle motion status, and environmental information. They often employ independent control strategies to complete interactive responses, entertainment recommendations, and function management, making it difficult to dynamically adjust interaction methods and control strategies based on real-time driving scenarios, safety rules, and personalized preferences. Summary of the Invention

[0004] This application provides an intelligent vehicle central control method, device, equipment, medium, and program product to solve the problem of lacking adjustments to vehicle central control devices based on multi-source perception information and understanding of the same scenario.

[0005] In a first aspect, embodiments of this application provide an intelligent vehicle central control method, including: Feature extraction and classification fusion are performed on the collected sensor data frames to obtain the member feature subsets corresponding to each cabin member; The current driving scenario type is determined by performing scenario judgment based on the acquired vehicle motion state data and the member feature subset. Based on the driving scenario type and the member feature subset, the response priority of the received central control operation command is evaluated, and a control command sequence is generated. Based on the subset of member features, corresponding historical preference data is matched from a preset historical database. In order to perform compliance screening and interactive configuration on the control command sequence according to preset security rules and the historical preference data, a set of central control interactive adjustment parameters and corresponding entertainment content configuration data are generated.

[0006] In an optional implementation, before obtaining the perceived data frame, the method further includes: By using a preset sliding time window and time tags, the received raw sensing data is time-synchronized and calibrated to obtain a synchronized sensing data packet. Based on the data type, the synchronous sensing data packets are classified and noise-reducing filtered to obtain a synchronous optimization dataset. The synchronous optimization dataset is time-series encapsulated according to a preset data format to generate perception data frames within each time window.

[0007] In an optional implementation, the step of extracting features and classifying and fusing them from the acquired sensory data frames to obtain a subset of member features corresponding to each cabin member includes: Extract driver state feature set, passenger state feature set, sound source spatial orientation feature set and seat pressure area distribution data from the perception data frame; A time-series cumulative state analysis is performed on the driver state feature set to generate a driver fatigue indicator value. Based on the occupant state feature set, the sound source spatial orientation feature set, and the seat pressure area distribution data, personnel identification is performed to obtain the occupant type identifier corresponding to each seat area; Based on the occupant type identifier and the preset seat area, the driver fatigue indicator value, the occupant state feature set, the sound source spatial orientation feature set, and the seat pressure area distribution data are classified and fused to generate a member feature subset corresponding to each cabin member.

[0008] In an optional implementation, the step of determining the current driving scenario type by performing scenario judgment based on the acquired vehicle motion state data and the subset of member features, and evaluating the response priority of the received central control operation commands based on the driving scenario type and the subset of member features to generate a control command sequence includes: Based on preset speed and driving status rules, the vehicle's current driving status label is matched according to the acquired vehicle motion status data; The current driving scenario type is obtained by matching the driving status label and the occupant type identifier with preset scenarios. Based on the current driving scenario type, the currently active interactive channel type and corresponding response sensitivity threshold are retrieved from the preset priority mapping table. Based on the sensitivity threshold, the received central control operation commands are identified by the source channel and prioritized to generate an initial control command sequence. When the driver fatigue indicator value reaches a preset fatigue threshold, the priority of non-voice commands in the initial control command sequence is downgraded, and the priority of the preset hazard avoidance voice commands identified in the initial control command sequence is upgraded to the highest level, thereby generating a control command sequence.

[0009] In an optional implementation, the step of matching corresponding historical preference data from a preset historical database based on the subset of member features, and then performing compliance screening and interactive configuration on the control command sequence according to preset security rules and the historical preference data to generate a set of central control interaction adjustment parameters and corresponding entertainment content configuration data includes: Based on the crew member type identifier, the historical preference data corresponding to each cabin member is matched from the preset historical database, and a list of prohibited content types and a list of restricted operation types are obtained from the preset security rules based on the crew member type identifier. The historical preference data includes resource retrieval frequency and frequency distribution of historical interaction parameters. The entertainment resource call instructions are removed according to the list of prohibited content types, and the system function configuration instructions are removed according to the list of restricted operation types to obtain a compliant instruction sequence, wherein the control instruction sequence includes the entertainment resource call instructions and the system function configuration instructions; Based on the resource retrieval frequency, the resources to be invoked corresponding to each instruction in the compliant instruction sequence are sorted by priority weighting to generate entertainment content configuration data. Based on the frequency distribution of the historical interaction parameters, decision-making is made on the interaction parameters of each type corresponding to the instructions in the compliant instruction sequence to generate a set of central control interaction adjustment parameters.

[0010] In an optional implementation, the method further includes: Based on the real-time acquired vehicle positioning coordinates and the coordinates of the reference point corresponding to the preset target location, the spatial distance between the vehicle and the target location is calculated. When the spatial distance value is less than the preset trigger threshold, lighting color temperature parameters and air conditioning target temperature parameters are generated based on the current system time and the detected environmental data. Based on the highest priority resource identifier in the entertainment content configuration data, generate a preset set of control parameters for smart home devices; The lighting color temperature parameters, the air conditioning target temperature parameters, and the control parameter set are sent to the corresponding smart home devices via a preset IoT communication link.

[0011] Secondly, embodiments of this application provide an intelligent in-vehicle central control device, comprising: The extraction and fusion module is used to extract features and classify and fuse the collected sensor data frames to obtain the member feature subsets corresponding to each cabin member; The evaluation and sorting module is used to make a scene judgment based on the acquired vehicle motion state data and the member feature subset to determine the current driving scene type, and to evaluate the response priority of the received central control operation commands based on the driving scene type and the member feature subset to generate a control command sequence. The filtering and configuration module is used to match corresponding historical preference data from a preset historical database based on the subset of member features, so as to perform compliance filtering and interactive configuration on the control command sequence according to preset security rules and the historical preference data, and generate a set of central control interactive adjustment parameters and corresponding entertainment content configuration data.

[0012] Thirdly, embodiments of this application provide a computer 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 steps of the above-described intelligent vehicle central control method.

[0013] Fourthly, embodiments of this application provide a readable storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned intelligent vehicle central control method.

[0014] Fifthly, embodiments of this application provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the above-described intelligent vehicle central control method.

[0015] One solution provided by the aforementioned intelligent in-vehicle central control method, device, equipment, medium, and program products first performs unified feature extraction and occupant-level fusion of multi-source perception information within the vehicle to form comprehensive features that accurately characterize the status of occupants in each cabin. Then, it identifies the current driving scenario by combining vehicle motion status and adjusts the response priority of central control operations according to different scenarios. Finally, it coordinates the configuration of interactive content and control strategies based on occupant historical preferences and safety rules. This allows the system to comprehensively consider driving risks, occupant composition, and personalized needs before executing user operations, dynamically optimizing entertainment resource access and central control interaction methods. By associating occupant status, vehicle status, and historical preferences in decision-making, it not only avoids executing potentially unsafe or inappropriate interactive operations in complex driving scenarios but also improves the matching degree of the central control system's interaction strategies and entertainment content configuration to the current scenario while ensuring driving safety and content compliance, thus balancing driving safety, interaction efficiency, and user experience. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the operating environment of the intelligent vehicle central control method in an embodiment of the present invention; Figure 2 This is a flowchart illustrating the intelligent vehicle central control method in an embodiment of the present invention; Figure 3 yes Figure 1 Functional module diagram of the intelligent vehicle central control device integrated with the server; Figure 4 This is a schematic diagram of the structure of the computer device in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. It should also be understood that, as used in this specification and the appended claims, the term "and / or" refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0020] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0022] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0023] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0024] With the development of intelligent cockpit and vehicle-to-everything (V2X) technologies, in-vehicle central control systems have evolved from traditional audio-visual playback and navigation control terminals into comprehensive interactive platforms integrating vehicle control, human-machine interaction, entertainment services, and intelligent connectivity. In practical applications, in-vehicle central control systems not only need to respond to the diverse operational needs of drivers and passengers, but also need to dynamically adjust interaction methods, entertainment content, and system functions based on vehicle operating status, passenger composition, driving behavior, and changes in the cabin environment, to balance driving safety, passenger experience, and intelligent service requirements. Simultaneously, with the widespread adoption of smart home devices and vehicle-to-home connectivity applications, in-vehicle central control systems also undertake the important task of collaborative control between vehicles and smart homes. Therefore, they need to possess multimodal perception, scene recognition, and personalized interactive configuration capabilities to adapt to various application scenarios such as family cars, business travel, and shared mobility.

[0025] Existing in-vehicle central control methods typically rely on single interaction methods such as touch, physical buttons, or voice as the primary input means. Upon receiving user input, they directly execute corresponding functions according to preset rules. Some systems can combine limited vehicle operating information, such as vehicle speed and parking status, to perform simple start-stop control for certain functions. Other systems can recommend entertainment resources based on user history or achieve basic vehicle-to-home connectivity control based on vehicle location. However, these methods lack a unified correlation analysis between driver status, passenger type, vehicle motion status, and environmental information. They often employ independent control strategies to complete interactive responses, entertainment recommendations, and function management, making it difficult to dynamically adjust interaction methods and control strategies based on real-time driving scenarios, safety rules, and personalized preferences.

[0026] The most significant shortcoming of existing in-vehicle central control methods lies in their lack of ability to achieve unified scene understanding and collaborative decision-making based on multi-source perception information. This results in the central control system primarily responding to user operations based on fixed rules or single state information, failing to comprehensively consider driving risks, occupant composition, and user historical preferences, and thus unable to synchronously optimize interaction methods, control commands, and entertainment content. When the vehicle is in different driving scenarios or occupant states change, the system may still execute interactive operations or content calls that are incompatible with the current scenario. This not only easily affects the interaction efficiency and user experience during driving but also makes it difficult to balance driving safety, content compliance, and personalized service needs, failing to achieve intelligent dynamic control of the in-vehicle central control system for complex scenarios.

[0027] To address the aforementioned issues, this application proposes an intelligent in-vehicle central control method, device, equipment, medium, and program product. First, it performs unified feature extraction and occupant-level fusion of multi-source sensory information within the vehicle to form comprehensive features accurately representing the status of each cabin occupant. Then, it identifies the current driving scenario based on the vehicle's motion state and adjusts the response priority of central control operations according to different scenarios. Finally, it collaboratively configures interactive content and control strategies based on occupant historical preferences and safety rules. This allows the system to comprehensively consider driving risks, occupant composition, and personalized needs before executing user operations, dynamically optimizing entertainment resource access and central control interaction methods. By associating occupant status, vehicle status, and historical preferences in decision-making, it not only avoids executing potentially unsafe or inappropriate interactive operations in complex driving scenarios but also improves the matching degree of the central control system's interaction strategies and entertainment content configuration to the current scenario while ensuring driving safety and content compliance, thus balancing driving safety, interaction efficiency, and user experience.

[0028] The intelligent vehicle central control method provided in this embodiment of the invention can be applied to, for example... Figure 1The operating environment shown is a vehicle-to-everything (V2X) environment consisting of a three-layer architecture: the vehicle-side edge computing domain, the cloud service platform, and external collaborative devices. The core computing device of this operating environment is a server integrated into the vehicle's central control unit. This server is equipped with a high-end SOC chip with AI acceleration capabilities, used to execute the intelligent vehicle central control method of this application. The server connects to a group of sensor acquisition devices distributed throughout the cabin via a hybrid network of vehicle Ethernet and CAN bus. This group of sensor acquisition devices specifically includes: an infrared camera mounted above the steering column to capture driver facial image sequences to extract eye opening and closing states and head posture; a ToF sensor embedded in the area in front of the center console to emit and receive reflected light pulses to generate three-dimensional point cloud data of the front cabin space; a microphone array positioned in front of the headliner to collect multi-channel audio signals for sound source localization and voice activity detection; a pressure sensor array positioned under each seat cushion to collect analog quantities of resistance changes caused by occupant weight distribution; and wheel speed pulse signals read through the CAN bus interface to provide a scalar value of vehicle speed. All of the above-mentioned sensing devices are connected to the corresponding data interface of the server via wired connection. Based on the global time reference provided by the high-precision crystal oscillator inside the server, the synchronous binding and noise reduction encapsulation of multi-source raw signals are completed within the sliding time window to generate standardized sensing data frames that can be called by subsequent algorithms.

[0029] At the in-vehicle interaction execution level, the operating environment also includes an in-vehicle terminal execution device group driven by the server. This device group is connected to the server via LVDS or A2B digital audio bus, and specifically includes: a central control display screen, used to receive the central control interaction adjustment parameter set and entertainment content configuration data generated by the server, and to present the compliantly filtered interactive controls and entertainment resource options in a graphical interface; an in-vehicle audio system, used to decode and play music or audiobooks that have been safely filtered according to the audio resource configuration data sent by the server; and a voice interaction module integrated around the central control screen or on the steering wheel. This module collects voice commands through a microphone array, which are then processed by the server to identify the source channel and prioritize them before executing the corresponding voice feedback or function response. Simultaneously, the server establishes a wireless connection with the rear-seat extended screen through its internal Bluetooth communication module, enabling wireless projection of video stream resources from the central control entertainment content configuration data, and supporting independent touch operation by rear passengers. These touch commands are transmitted back to the server via Bluetooth for unified arbitration processing.

[0030] At the vehicle-to-everything (V2X) data interaction level, the server establishes a two-way data link with the external cloud service platform through its integrated 5G+V2X communication module. The cloud service platform includes a cloud server deployed in a remote data center and cloud service instances running on it. The 5G communication module accesses the Internet via a cellular network and exchanges encrypted data with the cloud server to achieve the following functions: First, it uploads anonymized statistical feature data streams (including occupant type aggregation counts after stripping vehicle identification codes, entertainment content click frequency distribution, etc.) after localized differential privacy noise perturbation processing to the cloud server for macro-behavioral trend analysis without involving individual identity tracing; Second, it receives real-time traffic flow, construction area, and weather warnings from roadside units or the cloud via the vehicle-to-everything (V2X) interface as auxiliary input for the server when making scene decisions; Third, when a vehicle enters a preset geofence trigger area, the server packages the generated lighting color temperature parameters, air conditioning target temperature parameters, and highest priority resource identifier into a set of control parameters conforming to the IoT open interconnection standard through the 5G communication module, and forwards them to the home smart gateway in the user's residence via the cloud message queue, thereby driving the smart lighting, air conditioning, and audio systems in the residence to execute preset linkage responses.

[0031] Furthermore, the operating environment also includes user mobile terminal devices as external collaborative nodes. These devices communicate with the vehicle-side server via Bluetooth Low Energy or Wi-Fi Direct for near-field pairing. In car-sharing or temporary vehicle use scenarios, the user authentication module embedded in the mobile terminal transmits the user's identity identifier to the server, triggering the server to retrieve and load the user's historical interaction parameter frequency distribution and resource retrieval frequency data from the locally stored user preference database. This enables cloud loading and cross-vehicle migration of personalized configurations. The locally stored user preference database and residential location reference point data are both stored in the non-volatile memory embedded in the vehicle-side server and isolated and protected by hardware-level encrypted partitions to ensure that matching operations between member feature subsets and historical preference data can still be performed offline.

[0032] In one embodiment, such as Figure 2 As shown, an intelligent vehicle central control method is provided, which is applied to... Figure 1 Taking the runtime environment of [the system] as an example, the following steps are included: Step S1: Extract features and classify and fuse the collected sensory data frames to obtain the member feature subsets corresponding to each cabin member.

[0033] In the operating environment of this embodiment, the data acquisition layer includes various heterogeneous sensors, including an infrared camera mounted above the steering column, a Time of Flight (ToF) sensor embedded in the area in front of the center console, a MEMS microphone ring array arranged in the reading light area at the front of the roof, a wheel speed pulse signal source read through the vehicle controller area network bus interface, a flexible thin-film pressure sensor array arranged under each seat cushion, and an ambient light and humidity sensor arranged in front of the rearview mirror. Each of these sensors operates independently at a different physical sampling frequency. For example, the infrared camera typically outputs a sequence of grayscale facial images at a rate of 30 frames per second, the ToF sensor outputs a three-dimensional point cloud data stream at a rate of 60 frames per second, the microphone array outputs a pulse code modulation format digital audio stream at a sampling rate of 16 kHz, and the controller area network bus data is sent cyclically at millisecond intervals. Because these raw sensing data are independent of each other in terms of time reference and have different data structures, direct cross-sensor joint analysis is not possible. Therefore, the received raw sensing data first needs to be time-synchronized and calibrated.

[0034] Time synchronization calibration is performed using a preset sliding time window based on time stamps. Specifically, the server in the operating environment integrates a high-precision crystal oscillator, which provides the system with a globally unified real-time clock reference in μs. When each frame of raw sensing data from various sensors arrives at the server's data interface, the server immediately appends a time stamp based on this global clock reference to that frame of data. Subsequently, the server uses this time stamp as an index key to assemble all types of raw sensing data falling within the same preset time window into a temporary multi-source data packet. The width of this sliding time window is set to 33ms, which corresponds to the single-frame acquisition cycle of the lowest sampling rate sensor—the infrared camera (1000ms divided by 30 frames / s is approximately 33ms), thus ensuring that at least one complete frame of image data from the infrared camera can be captured within each time window. For cases where a certain data type is missing within a time window due to sensor sampling time deviation, the server uses a zero-order hold method to fill the gap. This involves copying the latest valid data frame received by the sensor within that time window and filling it into the corresponding field of the current data packet. This ensures that the synchronous sensing data packets output by each time window maintain a complete and consistent data structure. By eliminating the time misalignment issues introduced by different sensor startup times, clock drift differences, and inconsistent data transmission delays, data from different physical modes can be correlated and analyzed on the same time reference, avoiding misjudgments in feature extraction due to timing mismatches.

[0035] After acquiring the synchronous sensing data packet, the server performs data classification and noise reduction filtering based on the data type. Since the noise interference mechanisms of sensor signals of different physical types vary, filtering methods matching their respective signal types are required. For facial grayscale image data in the synchronous sensing data packet, the server performs median filtering. This method uses a sliding window of a preset size to traverse each pixel in the image, sorts the grayscale values ​​of that pixel and all its neighboring pixels, and takes the median value of the sorted sequence as the new grayscale value for that pixel. This operation effectively removes isolated polar noise points caused by sensor vibration noise due to vehicle movement or infrared sensor thermal noise—i.e., randomly appearing abnormal bright or dark spots in the image—while preserving the sharpness of facial edge contours to the greatest extent possible, ensuring the accuracy of subsequent eye key point localization. For the 3D point cloud data in the synchronous sensing data packet, the server performs outlier removal processing based on distance statistics. This method calculates the average distance of each 3D point in the point cloud to all its neighboring points in its local neighborhood space, and then calculates the average distance distribution parameters of the entire point cloud. 3D points whose local average distance exceeds a preset multiple standard deviation of the global average distance are identified as isolated drift noise points caused by diffuse reflection of light pulses or ambient light interference and are removed, thereby restoring the smooth geometric structure of the hand and limb surface. For the digital audio stream data in the synchronous sensing data packet, the server performs adaptive spectral subtraction noise reduction processing based on short-time energy spectrum. This method first detects silent segments of non-speech activity in the audio stream, estimates the power spectrum distribution of the current ambient background noise from these silent segments, and then subtracts the estimated power spectrum value of the background noise one by one in each frequency band of the entire audio spectrum, thereby effectively reducing the steady background noise components such as tire noise and air conditioner blower noise, and retaining only the speech components with drastic changes in acoustic characteristics. For the analog resistance data of the seat pressure sensor in the synchronous sensing data packet, the server performs moving weighted average filtering. Different attenuation weight coefficients are assigned to the resistance values ​​of the same sensing point within multiple consecutive sampling periods according to their distance from the current time—the closer the sampling value is to the current time, the greater the weight is given. Then, the weighted values ​​are summed and averaged to obtain the current output value for that point, thus smoothing out instantaneous spikes in pressure values ​​caused by minor vehicle vibrations. Through classification and noise reduction filtering, all types of data are kept stable and reliable within their respective data domains, avoiding false alarms in fatigue detection or incorrect occupant type identification due to noise interference.

[0036] After noise reduction and filtering, the server performs time-series encapsulation of the synchronized optimization dataset according to a preset data format, generating sensing data frames within each time window. The noise-reduced synchronized optimization dataset (e.g., facial image matrix, 3D point cloud coordinate set, audio waveform amplitude sequence, vehicle speed scalar value, pressure distribution 2D array, and ambient illuminance and humidity values) is serialized according to a preset fixed byte length and field arrangement order, forming a standardized sensing data frame with a unified data structure. This sensing data frame unifies multi-source heterogeneous sensor data into input data units with a consistent format and the same time reference.

[0037] Subsequently, the server extracts driver state feature sets, occupant state feature sets, sound source spatial orientation feature sets, and seat pressure area distribution data from the perceived data frame. The extraction of the driver state feature set primarily targets the facial image matrix in the perceived data frame. The server uses a pre-trained cascaded regression shape prediction model to locate the coordinates of six key pixels for each of the left and right eyes—including key points on the upper and lower eyelids and inner and outer corners of the eyes. Based on these coordinates, the aspect ratio of the eyes is calculated, defined as the ratio of the average Euclidean distance between the upper and lower key points to the average Euclidean distance between the left and right key points. The extraction of the occupant state feature set primarily targets the three-dimensional point cloud coordinate set in the perceived data frame. The server uses a point cloud segmentation algorithm based on a random forest classifier to separate the point cloud belonging to the human hand region from the background point cloud. Then, using inverse kinematics, the three-dimensional spatial coordinates of the palm and the index fingertip are estimated, and the projection direction and amplitude of the three-dimensional displacement vector between them on the horizontal and vertical planes are calculated, encoding them as gesture motion feature vectors. The extraction of the spatial orientation feature set of the sound source involves analyzing the audio waveform amplitude sequence in the sensing data frame. The server employs a beamforming algorithm based on delay summation to calculate the time delay difference between the microphones in the microphone array receiving the same sound signal, thereby deducing the incident azimuth angle of the sound signal within the passenger compartment and thus determining the seat area where the sound source is located. Seat pressure distribution data is obtained from a two-dimensional array of pressure distribution data in the sensing data frame, recording the pressure values ​​converted from the current resistance values ​​of each point in the pressure sensing array under each seat cushion.

[0038] After extracting the aforementioned feature sets, the server performs a time-series cumulative state analysis on the driver's state feature set. Specifically, the server continuously tracks the calculated eye aspect ratio values ​​within multiple time windows, forming a time-series curve of eye aspect ratio that changes over time. The server detects consecutive periods on this time-series curve where the eye aspect ratio value is consistently below a preset eye-closing threshold. When the cumulative duration of this consecutive period exceeds 2 seconds, the server sets the driver fatigue flag value to a logical true value and records the specific duration of this eye-closing as a continuous quantitative indicator of fatigue level. If no consecutive low-value period exceeding 2 seconds is detected on the eye aspect ratio time-series curve, the fatigue flag value remains a logical false value. Since a single instantaneous decrease in eye aspect ratio value may simply be a normal blink, only through time-series cumulative judgment across multiple frames can normal blinking be distinguished from genuine fatigue-induced eye-closing, thus avoiding frequent false alarms that could interfere with the driver's judgment.

[0039] The server further identifies occupants based on occupant status feature sets, sound source spatial orientation feature sets, and seat pressure area distribution data, obtaining occupant type identifiers corresponding to each seat area. Specifically, the occupant status feature set provides a first type of preliminary judgment on occupant body shape by estimating body shape classification from human volume contours in 3D point clouds; the sound source spatial orientation feature set provides a second type of evidence of the presence of occupants at that location by locating the sound source direction through beamforming; and the seat pressure area distribution data provides a third type of judgment basis by comparing the resultant pressure value of each seat area with preset adult weight pressure benchmark intervals and child weight pressure benchmark intervals. The server assigns different prior confidence weights to these three types of evidence based on the calibration accuracy of their respective source sensors, and then performs a fusion judgment based on Bayesian inference: Assuming the occupant type hypothesis space includes three types—"adult," "child," and "empty seat"—the server pre-defines a prior probability distribution matrix and sets its conditional probability likelihood function based on the calibration accuracy of each of the three types of sensors. Then, using the three types of observational evidence as independent inputs, the server calculates the posterior probability of each occupant type hypothesis given the simultaneous occurrence of these three types of observational evidence using Bayes' theorem. Finally, the occupant type with the highest posterior probability value is taken as the final identification result for that seat area, and the corresponding posterior probability value is output as the confidence weight of that occupant type identifier. Since multiple sensors independently observe the same physical object, any single sensor may produce misjudgments due to occlusion, interference, or viewing angle limitations. Bayesian fusion of multi-source evidence can obtain a comprehensive judgment result with higher reliability than any single sensor.

[0040] Finally, based on the occupant type identifier and preset seat areas, the server classifies and merges the driver fatigue indicator value, occupant state feature set, sound source spatial orientation feature set, and seat pressure area distribution data to generate occupant feature subsets corresponding to each cabin member. Specifically, the server uses preset seat areas—including the driver's seat area, the front passenger seat area, the rear left seat area, and the rear right seat area—as the classification key. The driver fatigue indicator value is assigned to the occupant feature subset corresponding to the driver's seat area. The occupant state feature set is allocated according to the seat area to which its spatial coordinates correspond to in the 3D point cloud. The sound source spatial orientation feature set is allocated according to the seat area corresponding to its located sound azimuth angle. The seat pressure area distribution data is directly allocated according to the physical seat area from which its pressure value originates. For each seat area, the server aggregates and encapsulates all the feature data allocated within that area into a separate subset data structure. This subset includes the occupant type identifier and its confidence weight for that seat area; if the area is the driver's seat, it includes fatigue indicator values ​​and fatigue level quantification; if an occupant is present in the area, it includes gesture motion feature vectors and voiceprint activity feature vectors; and real-time pressure distribution statistics for that area. This reorganizes the previously scattered feature data, organized by sensor type, into centralized feature data organized by individual occupants. This allows for direct indexing of the complete status information of the occupants in a given seat area during scenario decision-making, priority assessment, and safety screening, without requiring distributed queries across multiple data structures.

[0041] Step S2: Based on the acquired vehicle motion state data and the member feature subset, a scenario judgment is made to determine the current driving scenario type. Based on the driving scenario type and the member feature subset, the response priority of the received central control operation command is evaluated, and a control command sequence is generated.

[0042] The vehicle motion status data originates from wheel speed pulse signals periodically sent by the vehicle chassis domain controller via the controller area network (CLAN) bus. These signals are received by the server's CLAN bus interface, and the number of pulses per unit time is counted, combined with tire rolling radius parameters, to convert the signal into a longitudinal vehicle speed value in km / h. This speed value is a direct physical quantity representing the intensity of the vehicle's current motion. However, a single instantaneous speed value cannot fully reflect the vehicle's driving conditions. For example, a vehicle traveling at a constant speed of 80 km / h on a highway, following traffic flow at 60 km / h on an urban expressway, stopping and starting at 10 km / h in congested traffic, and having a zero speed for a prolonged period while parked at a charging station—while these speed values ​​overlap, the functional usability and interactive safety requirements of the in-vehicle central control system are drastically different. Therefore, it is necessary to determine the specific driving state based on the speed value and its change pattern over time. The server uses preset speed and driving state rules to match the acquired vehicle motion state data with the vehicle's current driving state label. Specifically, the server compares the currently received speed value with multiple preset speed state thresholds within a range, and records the speed value's change trend over multiple consecutive time windows. When the speed value is zero and the cumulative duration of this zero-value state within a consecutive time window exceeds 3 minutes, the server labels the driving state at that moment as "parked stationary state." When 30km / h ≥ speed value > 0, it is labeled as "low-speed driving state," corresponding to urban congestion or parking lot entry / exit conditions. When 80km / h ≥ speed value > 30km / h, it is labeled as "medium-speed driving state," corresponding to urban ordinary roads or urban expressways. When the speed value > 80km / h, it is labeled as "high-speed driving state," corresponding to highway cruising conditions. The continuous analog speed value is discretized into a finite set of states with clear physical boundaries and functional semantics, providing discrete input variables that can be directly used for logical operations for subsequent scenario type determination.

[0043] After obtaining the driving status tag, the server matches the driving status tag with the occupant type identifier of each seat area obtained in step S1 to obtain the current driving scenario type. The preset scenario matching is achieved through a scenario decision tree that takes driving status labels and occupant type identifiers as inputs and scenario type labels as outputs. The root node of the decision tree uses the driving status label as the first-level splitting condition, roughly dividing the scenario judgment into the parking scenario category and the driving scenario category. Under the parking scenario category, the decision tree further uses whether the occupant type identifier contains children as the second-level splitting condition. When the occupant type identifier of any seat area is detected as a child, the scenario output label is "Parking Child Supervision Scenario", otherwise it is "Parking Adult Entertainment Scenario". Under the driving scenario category, the decision tree uses the specific value of the speed status label as the second-level splitting condition, outputting "High-speed Cruise Scenario", "Medium-speed City Driving Scenario", and "Low-speed Congestion Scenario" respectively. Then, the decision tree uses the driver fatigue identifier value in the member feature subset corresponding to the driver's seat area as the third-level splitting condition. When the fatigue identifier value is a logical true value, the "driver fatigue" suffix is ​​appended to the above driving scenario label. For example, "High-speed Cruise Scenario" with fatigue identifier superimposed is output as "High-speed Cruise Driver Fatigue Scenario". By combining single-dimensional driving status and occupant composition information into composite scenario types with clearly defined interaction strategies, subsequent interaction strategy configurations can respond differently to different scenarios. For example, when the driving status label is "high-speed driving status" and the driver fatigue indicator value is a logical truth value, the scenario matching output "high-speed cruise driver fatigue scenario" directly triggers the highest priority response mechanism for hazard avoidance voice commands. This is fundamentally different from the existing technology's approach of treating all commands equally regardless of driver fatigue.

[0044] After obtaining the current driving scenario type, the server retrieves the currently active interaction channel types and their corresponding response sensitivity thresholds from a preset priority mapping table based on this scenario type. Then, based on these sensitivity thresholds, it identifies the source channel of the received central control operation commands and prioritizes them, generating an initial control command sequence. This priority mapping table is a two-dimensional data structure with scenario type labels as row indices and interaction channel configuration parameters as column contents. The interaction channel types include voice command receiving channels, touchscreen operation channels, gesture recognition interaction channels, and eye-tracking activation channels. Each interaction channel corresponds to an activation status identifier—which takes one of two values: "enabled" or "disabled"—and a response sensitivity threshold value. When the server retrieves the row record corresponding to the current scenario, it first reads the activation status identifiers of each channel, excluding all channels marked "disabled" from subsequent processing, and then reads the response sensitivity thresholds of all channels marked "enabled". For the voice command receiving channel, this sensitivity threshold is reflected as a wake-up confidence threshold, meaning that the command is considered valid only when the confidence score of the voice recognition result exceeds this threshold. For the touchscreen operation channel, this sensitivity threshold is reflected as a touch press duration threshold, meaning that the corresponding operation response is triggered only when the duration of the touch press exceeds this threshold, thus preventing accidental touches during driving. For the gesture recognition interaction channel, this sensitivity threshold is reflected as a hand movement amplitude threshold, meaning that a gesture is recognized as valid only when the displacement amplitude exceeds this threshold, thus filtering out unconscious small-amplitude hand tremors during driving. For the eye-tracking gaze activation channel, this sensitivity threshold is reflected as a gaze persistence time threshold, meaning that the system automatically selects the control and activates the corresponding voice command completion function only when the user's gaze lingers on an interactive control for a continuous duration exceeding this threshold. After determining the activation status and sensitivity threshold of each interaction channel, the server identifies the source channel of each received central control operation command. This is done by detecting the input source identifier field carried by the command—this identifier field is automatically filled in by the underlying driver based on the type of triggering sensor during command generation—to determine whether the command originates from the voice channel, touchscreen channel, gesture channel, or eye-tracking channel. Commands whose source channel is currently marked as "disabled" are discarded by the server. Commands whose source channel is currently marked as "enabled" are included in the initial control command sequence and arranged according to the channel's preset priority in the priority mapping table, with commands from higher-priority channels appearing earlier in the sequence.The goal of this search and sorting is to dynamically set the availability and importance of interaction channels based on the current scenario type. For example, in the "high-speed cruise driver fatigue scenario," the voice command receiving channel is the only channel that remains fully enabled, while the touchscreen swipe operation channel and gesture recognition channel are disabled or desensitized. This is because, under the critical condition of high-speed driving and driver fatigue, any interaction method that requires the driver to take their hands off the steering wheel or their eyes off the road poses a serious safety risk. Only voice interaction allows the driver to keep both hands on the steering wheel to complete all operations. Therefore, in this scenario, the sensitivity threshold of the voice channel is adjusted to the lowest level—the corresponding wake-up confidence threshold is reduced to 0.60—so that the driver can successfully trigger commands with a low volume or a somewhat unclear pronunciation. The touchscreen operation channel, on the other hand, requires a continuous press of more than 0.8 seconds to trigger a single click confirmation, thereby minimizing the probability of accidental touches while retaining the most basic functions.

[0045] After generating the initial control command sequence, the server further determines whether the driver's fatigue indicator value has reached the preset fatigue threshold—the fatigue threshold is the aforementioned condition that the duration of closed eyes exceeds 2 seconds. When the fatigue indicator value is a logical true value, the server performs two priority modulation operations on the initial control command sequence. The first modulation operation involves downgrading the priority of non-voice commands in the initial control command sequence. Specifically, the server iterates through each command in the initial control command sequence, checking its source channel identifier. Commands whose source channel does not belong to the voice command receiving channel are shifted backward, placing them after all voice commands. The second modulation operation prioritizes the preset risk avoidance voice commands identified in the initial control command sequence. Specifically, the server iterates through each voice command in the initial control command sequence, matching its semantic recognition result with a locally stored list of preset risk avoidance command keywords—which includes entries such as "navigate to the nearest service area," "activate rest reminder," and "find nearby parking"—for string matching. When a match is successful, the server extracts the command from its current position and places it at the beginning of the entire initial control command sequence, placing it before all other commands. After this dual modulation of downgrading and prioritization, the server outputs the modulated sequence as the final control command sequence. The order of this sequence directly determines the execution order of multiple concurrent commands. When the system detects that the driver is fatigued, it reduces the incentive for the driver to manually operate the central control screen by suppressing the response priority of non-voice commands. On the other hand, it ensures that the system can respond to the driver's risk avoidance requests as soon as possible in emergency situations by elevating risk avoidance voice commands to absolute priority. Thus, it builds a safety backup mechanism for fatigued driving scenarios at the interaction level.

[0046] Step S3: Match the corresponding historical preference data from the preset historical database according to the member feature subset, and perform compliance screening and interactive configuration on the control command sequence according to the preset security rules and the historical preference data to generate a central control interactive adjustment parameter set and corresponding entertainment content configuration data.

[0047] The historical database is a local storage structure indexed by occupant identification. This database continuously records the interaction behavior logs of occupants in each seating area during each vehicle power-on operation. Log fields include the resource type identifier, resource name, timestamp of retrieval, interaction channel type used when retrieving the resource, and the frequency of modifications to various system function configuration parameters during a single interaction session. Initially empty, the historical database is continuously enriched as the vehicle is used. The total amount of data recorded is limited by the available capacity of the server's non-volatile memory. When the storage space reaches a preset limit, the system deletes the oldest historical records in chronological order to maintain a dynamic balance of storage space. It should be noted that the occupant identification in the historical database is not a personal identification information field, but an anonymized seat identifier generated by combining seat pressure area distribution data and occupant type identifier. This identifier only distinguishes between categorized identities such as "adult in the driver's seat", "adult in the front passenger seat", "child on the left rear seat", and "empty seat on the right rear seat", and does not involve any biometric features or identity information that can be traced back to a specific natural person, thus avoiding the risk of storing personal privacy data at the source.

[0048] The server first matches historical preference data for each cabin member from a preset historical database based on the occupant type identifier. This matching operation uses the occupant type identifier as the query key, retrieving all historical interaction records matching that identifier from the historical database. The server then extracts the resource retrieval frequency and historical interaction parameter frequency distribution of all retrieved entries, aggregating them into historical preference data corresponding to that occupant type identifier. The resource retrieval frequency is defined as the ratio of the total number of times various entertainment resources were successfully retrieved and played for a preset duration by this type of occupant within a preset time backtracking window—the window width being set to the most recent 30 calendar days—to the total number of times all types of resources were retrieved within that time window. This ratio is stored as a percentage. The historical interaction parameter frequency distribution is defined as a statistical histogram vector formed by the number of times this type of occupant modified various adjustable parameters in the system function configuration interface within the same time backtracking window—including but not limited to the gain values ​​of each frequency band of the audio equalizer, the percentage value of the display screen brightness, the volume level value of the navigation voice broadcast, and the air conditioning temperature setting value—according to the parameter type dimension. The goal of performing this historical preference data matching is to transform discrete, static occupant type identifiers into statistically significant, dynamically updated preference feature vectors. This allows subsequent personalized configurations to be generated based on the actual behavioral patterns of that occupant type over a past period, rather than a generic user profile template, thereby improving the alignment between recommendation results and the occupant's true intentions. For example, when the occupant type identifier is "adult in the front passenger seat," if the system retrieves data from the historical database showing that this occupant has accessed classical music resources significantly more frequently than pop music resources in the past 30 days, this frequency difference will directly influence the prioritization of subsequent entertainment resources.

[0049] The server synchronously retrieves the list of prohibited content types and the list of restricted operation types from preset security rules based on the occupant type identifier. The security rules are a set of static configuration tables stored locally, pre-configured by the system administrator or vehicle owner through the in-vehicle settings interface. These rules include permitted and prohibited rules differentiated by occupant type. The prohibited content type list defines a set of content type codes that are absolutely inaccessible when an occupant of that type is present for each occupant type. For example, when the occupant type identifier is "child," this list includes entries such as "restricted films," "games / applications containing violent elements," and "adult audiobooks"; when the occupant type identifier is "empty seat," this list is empty. The restricted operation type list defines a set of function operation codes that require additional confirmation or can only be initiated through specific channels when a specific occupant type is present. For example, when the occupant type identifier is "child" and the vehicle is in motion, this list includes entries such as "change navigation destination," "enter the deeper menu of system settings," and "connect external Bluetooth devices." These operations are not absolutely prohibited, but require semantic confirmation via the voice channel before execution. By transforming legal compliance requirements and parental supervision needs into discrete judgment conditions that can be executed by machines, clear criteria for filtering instructions are provided.

[0050] After obtaining the lists of prohibited content types and restricted operation types, the server removes entertainment resource call instructions from the control instruction sequence based on the list of prohibited content types and system function configuration instructions from the control instruction sequence based on the list of restricted operation types, thus obtaining a compliant instruction sequence. Each instruction in the control instruction sequence carries an instruction type field, which indicates whether the instruction belongs to either an "entertainment resource call instruction" or a "system function configuration instruction." For entries carrying the "entertainment resource call instruction" type, the server further reads the content type metadata tag of the target resource pointed to by the instruction. When the value of this metadata tag falls into any of the codes listed in the prohibited content type list, the server completely removes the instruction from the control instruction sequence, preventing it from proceeding to the subsequent sorting and configuration stages. For entries carrying the "system function configuration instruction" type, the server further reads the function operation code requested to be modified by the instruction. When this code falls into any of the codes listed in the restricted operation type list, the server completely removes the instruction from the control instruction sequence, preventing it from proceeding to the subsequent processing stages. After these two rounds of removal, all remaining instructions in the control instruction sequence are compliant instructions that have passed the security screening, and these instructions together constitute the compliant instruction sequence. Before preference matching and personalized sorting, mandatory security filtering is performed to ensure that no instruction that touches the security red line will be executed, regardless of the subsequent preference sorting result. This creates a hard security barrier at the system architecture level that is independent of user preferences.

[0051] Subsequently, the server prioritizes and sorts the resources to be invoked corresponding to each instruction in the compliant instruction sequence based on the resource retrieval frequency in the historical preference data, generating entertainment content configuration data. Specifically, let Ri be the resource to be invoked pointed to by the i-th instruction in the compliant instruction sequence, and Fi be the resource retrieval frequency of this resource in the last 30 days—this frequency value has been retrieved from the historical database in the previous step—then the comprehensive priority score Pi of this resource is calculated by a weighting function. This function multiplies the resource retrieval frequency Fi by a preset weight coefficient Wf, and adds a scene adaptation coefficient Si related to the current driving scenario type. The scene adaptation coefficient Si is determined by the current driving scenario type. For example, when the scenario type is "high-speed cruise driver fatigue scenario", slow-paced music resources are given a higher scene adaptation coefficient, while when the scenario type is "parking child supervision scenario", educational audio resources are given a higher scene adaptation coefficient. After the weighted calculations described above, each instruction in the compliant instruction sequence corresponds to a comprehensive priority score Pi. The server sorts all instructions in the sequence in descending order of Pi value, and extracts the first N instructions from the sorted sequence—where N is preset to 15 by the system—and packages the resource identifier, resource name, resource type, and corresponding Pi value of each extracted instruction into entertainment content configuration data in the sorted order. By combining the occupant's historical behavioral preferences with the real-time semantics of the current scenario, the recommendation results reflect both the user's long-established habitual patterns and the specific needs of the current driving scenario for content types, avoiding the problem of mismatched recommended content with the scene atmosphere caused by simply relying on historical frequency.

[0052] Meanwhile, based on the frequency distribution of historical interaction parameters in historical preference data, the server makes decisions on the interaction parameters of each type corresponding to the instructions in the compliant instruction sequence, generating a set of central control interaction adjustment parameters. Specifically, the server extracts all instruction entries in the compliant instruction sequence that carry the type "system function configuration instruction". For each such instruction, it reads the interaction parameter type code to be modified—this code corresponds to specific parameters such as audio equalizer gain value, display brightness value, navigation voice volume value, or air conditioning temperature value—and then uses this type code as an index to find the value level of this type of parameter that has been most frequently selected by this occupant type in the past 30 days in the historical interaction parameter frequency distribution vector. When a distinct mode exists within the frequency distribution vector, the server uses the specific value corresponding to that mode as the decision output value for that parameter. When the frequency distribution vector exhibits a bimodal or multimodal distribution without a single distinct mode, the server uses the median value of the distribution as the decision output value. When the total number of data samples in the frequency distribution vector is lower than a preset effective statistical sample size threshold—set to 10 times—meaning historical data is insufficient to support reliable decision-making, the server reads the default value preset at the system's factory settings as the decision output value. After completing the decision-making judgment for each interactive parameter involved in all system function configuration instructions in the compliant instruction sequence, the server merges all decision output values ​​according to parameter type codes—for cases where the same instruction carries multiple parameter modification requests, all parameter values ​​are grouped under the same instruction entry—and then all merged parameter-value pairs are encapsulated into a central control interactive adjustment parameter set. The goal of executing this interactive parameter decision-making is to enable the central control system to automatically adjust the various system function parameters to the most commonly used value combination for that occupant type each time it is powered on or after a change of occupants. This avoids the need for occupants to manually adjust the volume, brightness, and equalizer each time they use the vehicle, thereby shifting the interaction burden from the occupant's side to the system side, allowing attention resources during driving to be focused on the driving task itself.

[0053] In an optional implementation, to achieve vehicle-home linkage control, the method further includes: It should be understood that the vehicle's positioning coordinates are provided in real time by a Global Navigation Satellite System (GNSS) receiver integrated on the vehicle's server. This receiver simultaneously receives satellite signals from multiple systems, including GPS, BeiDou, and GLONASS, and uses carrier phase differential positioning technology to improve positioning accuracy to within meters, outputting the current vehicle's geographic coordinates in three dimensions: longitude, latitude, and altitude. The coordinates of the reference point corresponding to the preset target location are stored in the server's non-volatile memory. These reference point coordinates are manually calibrated by the vehicle owner through the map selection interface on the central control display when the system is first used. They are typically calibrated as the latitude and longitude values ​​of the residence, serving as the benchmark reference point for all subsequent spatial distance calculations. After the server retrieves these target location reference point coordinates, it substitutes them with the real-time acquired vehicle positioning coordinates into the geodetic formulas for equal-angle latitude to calculate the geodetic length between them, which characterizes the spatial distance between the vehicle's current location and the preset target location. In this calculation process, the server first converts the vehicle's positioning coordinates and the target location reference point's coordinates from degrees, minutes, and seconds to values ​​expressed in radians. Then, it calculates the longitude difference between the two points, followed by iteratively solving for the normalized latitude. Finally, based on the direct and inverse calculation relationship between the geodetic azimuth and the normalized latitude, it derives the arc length of the geodetic line between the two points. This abstract latitude and longitude coordinate pair is transformed into an intuitive spatial distance value. This value serves as the sole independent variable for triggering all vehicle-home collaborative linkage decisions. When this distance value is large, it indicates that the vehicle is far from the residence, and the system does not need to perform any vehicle-home linkage operation to save computing resources and communication bandwidth. Only when the distance value shrinks to a specific range does the system initiate the subsequent linkage parameter generation process, thereby avoiding frequent occupation of IoT communication link resources when unnecessary.

[0054] After calculating the spatial distance, the server compares this distance with a preset trigger threshold. This threshold is preset at the factory to 3000 meters. The selection of this value is based on the fact that when the vehicle is 3000 meters from the residence, at a typical urban road average speed of approximately 30 to 40 km / h, the vehicle's arrival time is approximately 5 to 6 minutes. This time window is sufficient for smart home devices to complete the entire startup process from standby to working state. For example, in a high-temperature summer scenario, it typically takes 5 to 10 minutes for the air conditioner in a residence to cool the room from a high temperature to a comfortable temperature after startup. If the trigger threshold is set too low—for example, 500 meters—the air conditioner may not have effectively cooled the room by the time the vehicle arrives, significantly reducing the user experience of vehicle-home integration. If the trigger threshold is set too high—for example, 10000 meters—the air conditioner will be triggered when the vehicle is still far from the residence, meaning the air conditioner will have been running for too long by the time the vehicle actually arrives, resulting in unnecessary energy waste and increasing the risk of discomfort for passengers due to excessively low indoor temperatures.

[0055] When the server determines that the spatial distance is less than the 3000-meter trigger threshold, the system immediately initiates the parameter generation process for vehicle-home collaboration. The first operation of this process is to generate lighting color temperature parameters and air conditioning target temperature parameters based on the current system time and detected environmental data. The environmental data comes from the ambient light and humidity sensor outputs of ambient illuminance values ​​and air humidity percentage values. At the same time, the server obtains real-time meteorological data of the current vehicle's location—including outdoor dry-bulb temperature values—through the vehicle network service interface. When generating lighting color temperature parameters, the server reads the current system time and determines the time period to which that time belongs. When the time falls within the interval between half an hour after sunset and half an hour before sunrise, the system sets the lighting color temperature parameter corresponding to that time to 2700 Kelvin warm light. This color temperature value corresponds to the warm yellow light of traditional incandescent lamps, which helps passengers secrete melatonin in a nighttime environment, facilitating subsequent sleep preparation. When the time falls within the interval between half an hour after sunrise and half an hour before sunset, the system sets the lighting color temperature parameter to 5000 Kelvin cool white light. This color temperature value corresponds to the color temperature of midday sunlight, which helps passengers stay awake and alert in a daytime environment. When generating the target temperature parameters for the air conditioning system, the server first obtains the outdoor dry-bulb temperature value through the vehicle network service. When the outdoor temperature value is higher than 30 degrees Celsius, the system sets the target temperature parameter for the air conditioning system to 24 degrees Celsius. This value is based on the upper limit of the summer comfort temperature of the human thermal comfort model, which can achieve a balance between cooling and energy consumption. When the outdoor temperature value is lower than 10 degrees Celsius, the system sets the target temperature parameter for the air conditioning system to 26 degrees Celsius. This value is based on the lower limit of the winter comfort temperature of the human thermal comfort model. When the outdoor temperature value is between 10 degrees Celsius and 30 degrees Celsius, the system sets the target temperature parameter for the air conditioning system to an energy-saving temperature value with a difference of no more than 8 degrees Celsius from the outdoor temperature—for example, when the outdoor temperature is 22 degrees Celsius, the system sets the target temperature to 26 degrees Celsius—to avoid over-adjustment of the air conditioning system and energy waste. The goal of generating the lighting color temperature and air conditioning target temperature is to pre-set a suitable indoor light and temperature environment based on a comprehensive perception of the time and meteorological dimensions, before the occupants arrive at their residence. This allows the occupants to experience a comfortable microclimate that is significantly different from the current outdoor environment the moment they enter the residence, thus achieving a seamless connection between the vehicle and the home in terms of environmental regulation.

[0056] The server synchronously generates a set of control parameters for preset smart home devices based on the highest priority resource identifier in the entertainment content configuration data. After the priority weighting sort is completed in step S3, the various entertainment resources in the entertainment content configuration data are arranged in descending order according to their comprehensive priority score Pi. The resource identifier carried by the resource ranked first—that is, the highest priority resource—is extracted by the server. This resource identifier is a globally unique string code, which has unique pointing in both the local entertainment resource library and the online streaming service. The server fills this resource identifier as payload data into a specific field of the preset smart home device control parameter set. This control parameter set follows the standard data model structure defined by the Smart Home Interconnection Open Foundation, which includes a device type code field, an operation type field—this field is filled with "AudioPush," i.e., the audio push operation code—and a resource locator field—this field is filled with the aforementioned resource identifier. The goal of generating this set of control parameters is to pre-project the optimal entertainment resources selected by the vehicle after multiple rounds of safety screening and personalized sorting onto the smart audio system in the home before the occupants even step into the house. This ensures that the first audio content the occupants hear upon entering the home is exactly what they have just heard in the car or what the system determines to be their favorite content. This achieves a seamless transition of the entertainment experience from the car audio system to the home audio system, avoiding the cumbersome operation of manually turning on the audio system, searching for resources, and replaying them after entering the home, as is common in traditional scenarios.

[0057] After generating all the aforementioned control parameters, the server sends the lighting color temperature parameters, air conditioning target temperature parameters, and control parameter set to the corresponding smart home devices via a pre-defined IoT communication link. This IoT communication link uses the server's integrated 5G communication module as the sending end, and after accessing the internet via a cellular network, establishes an encrypted transmission channel with a cloud message queue server. The cloud message queue server, based on pre-bound device mapping relationships, forwards the received control parameters to the smart gateway within the residence—this smart gateway acts as the local control hub for all smart devices within the residence. Upon receiving the forwarded lighting color temperature parameters, it converts the parameters into brightness and color temperature adjustment commands conforming to the Zigbee protocol standard and sends them to the smart lighting fixtures. Upon receiving the air conditioning target temperature parameters, it converts the parameters into temperature setting commands conforming to the infrared remote control code standard and sends them to the infrared receiving port of the smart air conditioner. Upon receiving the control parameter set containing resource identifiers, it sends the resource identifiers to the smart speaker system via the home intranet, triggering the smart speaker system to actively retrieve the audio stream corresponding to the resource identifier from the cloud content delivery network and add it to the playback queue. The goal of this transmission operation is to reliably transmit all decision results completed on the vehicle side to the final execution device through the three-level link of vehicle-cloud-home, enabling the entire closed loop of vehicle-home linkage to be physically connected. This upgrades the in-vehicle central control system from a traditional in-vehicle infotainment terminal to a smart life control hub covering all scenarios inside and outside the vehicle. It should be noted that during the entire process of generating and transmitting vehicle-home linkage parameters, the server simultaneously performs localized differential privacy noise perturbation processing on the vehicle positioning coordinates and environmental data used to generate lighting color temperature parameters and air conditioning target temperature parameters, in accordance with the established privacy protection mechanism. The perturbed fields are then separated from the vehicle identification code and encapsulated separately as an anonymized upload data stream before being sent to the cloud. This ensures that sensitive information such as the user's specific location trajectory and personal habits are not exposed while realizing vehicle-home collaboration functions.

[0058] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0059] In one embodiment, an intelligent vehicle central control device is provided, which corresponds one-to-one with the intelligent vehicle central control method described in the above embodiments. For example... Figure 3 As shown, the intelligent vehicle central control device 2 may include: an extraction and fusion module 21, an evaluation and sorting module 22, and a filtering and configuration module 23. Detailed descriptions of each functional module are as follows: The extraction and fusion module 21 is used to extract features and classify and fuse the collected sensor data frames to obtain the member feature subsets corresponding to each cabin member; The evaluation and sorting module 22 is used to make a scene judgment based on the acquired vehicle motion state data and the member feature subset to determine the current driving scene type, and to evaluate the response priority of the received central control operation command based on the driving scene type and the member feature subset to generate a control command sequence. The filtering configuration module 23 is used to match the corresponding historical preference data from the preset historical database according to the subset of member features, so as to perform compliance filtering and interactive configuration on the control command sequence according to the preset security rules and the historical preference data, and generate a set of central control interactive adjustment parameters and corresponding entertainment content configuration data.

[0060] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0061] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0062] This application also provides a computer device, such as... Figure 4 As shown, the computer device includes: at least one processor 32, a memory 31, and a computer program stored in the memory 31 and executable on the at least one processor 32. When the processor 32 executes the computer program, it implements the steps in any of the above method embodiments, or when the processor 32 executes the computer program, it implements the functions of each module / unit in the above device embodiments.

[0063] For example, the computer program may be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 32 to complete this application. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the computer device.

[0064] Those skilled in the art will understand that Figure 4 The computer device described is merely an example and does not constitute a limitation on the computer device. It may include more or fewer components than shown, or combine certain components, or different components. For example, the computer device may also include input / output devices, network access devices, buses, etc.

[0065] The processor 32 mentioned above can be a Central Processing Unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), or field-programmable gate arrays (FPGAs). Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor 32 can be a microprocessor or any conventional processor, etc.

[0066] The memory 31 can be an internal storage unit of the computer device, such as a hard drive or memory. The memory 31 can also be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory 31 can include both internal and external storage units of the computer device.

[0067] This application also provides a readable storage medium storing a computer program that, when executed by a processor, implements the steps described in the various method embodiments above.

[0068] This application provides a computer program product that, when run on an electronic device, enables the electronic device to perform the steps described in the various method embodiments above.

[0069] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of this application can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include at least: any entity or device capable of carrying computer program code to a photographing device / terminal device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium. Examples include USB flash drives, portable hard drives, magnetic disks, or optical disks. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0070] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0071] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0072] In the embodiments provided in this application, it should be understood that the disclosed apparatus / devices and methods can be implemented in other ways. For example, the apparatus / device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0073] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0074] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for intelligent vehicle central control, characterized in that, The method includes: Feature extraction and classification fusion are performed on the collected sensor data frames to obtain the member feature subsets corresponding to each cabin member; The current driving scenario type is determined by performing scenario judgment based on the acquired vehicle motion state data and the member feature subset. Based on the driving scenario type and the member feature subset, the response priority of the received central control operation command is evaluated, and a control command sequence is generated. Based on the subset of member features, corresponding historical preference data is matched from a preset historical database. In order to perform compliance screening and interactive configuration on the control command sequence according to preset security rules and the historical preference data, a set of central control interactive adjustment parameters and corresponding entertainment content configuration data are generated. 2.The intelligent vehicle central control method of claim 1, wherein, Before obtaining the perceived data frame, the method further includes: By using a preset sliding time window and time tags, the received raw sensing data is time-synchronized and calibrated to obtain a synchronized sensing data packet. Based on the data type, the synchronous sensing data packets are classified and noise-reducing filtered to obtain a synchronous optimization dataset. The synchronous optimization dataset is time-series encapsulated according to a preset data format to generate perception data frames within each time window. 3.The intelligent vehicle central control method of claim 1, wherein, The process of extracting features and classifying and fusing them from the collected sensor data frames to obtain the member feature subsets corresponding to each cabin member includes: Extract driver state feature set, passenger state feature set, sound source spatial orientation feature set and seat pressure area distribution data from the perception data frame; A time-series cumulative state analysis is performed on the driver state feature set to generate a driver fatigue indicator value. Based on the occupant state feature set, the sound source spatial orientation feature set, and the seat pressure area distribution data, personnel identification is performed to obtain the occupant type identifier corresponding to each seat area; Based on the occupant type identifier and the preset seat area, the driver fatigue indicator value, the occupant state feature set, the sound source spatial orientation feature set, and the seat pressure area distribution data are classified and fused to generate a member feature subset corresponding to each cabin member. 4.The intelligent vehicle central control method of claim 3, wherein, The step of determining the current driving scenario type by performing scenario judgment based on the acquired vehicle motion state data and the member feature subset, and evaluating the response priority of the received central control operation commands based on the driving scenario type and the member feature subset to generate a control command sequence includes: Based on preset speed and driving status rules, the vehicle's current driving status label is matched according to the acquired vehicle motion status data; The current driving scenario type is obtained by matching the driving status label and the occupant type identifier with preset scenarios. Based on the current driving scenario type, the currently active interactive channel type and corresponding response sensitivity threshold are retrieved from the preset priority mapping table. Based on the sensitivity threshold, the received central control operation commands are identified by the source channel and prioritized to generate an initial control command sequence. When the driver fatigue indicator value reaches a preset fatigue threshold, the priority of non-voice commands in the initial control command sequence is downgraded, and the priority of the preset hazard avoidance voice commands identified in the initial control command sequence is upgraded to the highest level, thereby generating a control command sequence.

5. The intelligent vehicle central control method as described in claim 3, characterized in that, The step of matching corresponding historical preference data from a preset historical database based on the subset of member features, and performing compliance screening and interactive configuration on the control command sequence according to preset security rules and the historical preference data, to generate a set of central control interaction adjustment parameters and corresponding entertainment content configuration data includes: Based on the crew member type identifier, the historical preference data corresponding to each cabin member is matched from the preset historical database, and a list of prohibited content types and a list of restricted operation types are obtained from the preset security rules based on the crew member type identifier. The historical preference data includes resource retrieval frequency and frequency distribution of historical interaction parameters. The entertainment resource call instructions are removed according to the list of prohibited content types, and the system function configuration instructions are removed according to the list of restricted operation types to obtain a compliant instruction sequence, wherein the control instruction sequence includes the entertainment resource call instructions and the system function configuration instructions; Based on the resource retrieval frequency, the resources to be invoked corresponding to each instruction in the compliant instruction sequence are sorted by priority weighting to generate entertainment content configuration data. Based on the frequency distribution of the historical interaction parameters, decision-making is made on the interaction parameters of each type corresponding to the instructions in the compliant instruction sequence to generate a set of central control interaction adjustment parameters. 6.The intelligent vehicle central control method of claim 1, wherein, The method further includes: Based on the real-time acquired vehicle positioning coordinates and the coordinates of the reference point corresponding to the preset target location, the spatial distance between the vehicle and the target location is calculated. When the spatial distance value is less than the preset trigger threshold, lighting color temperature parameters and air conditioning target temperature parameters are generated based on the current system time and the detected environmental data. Based on the highest priority resource identifier in the entertainment content configuration data, generate a preset set of control parameters for smart home devices; The lighting color temperature parameters, the air conditioning target temperature parameters, and the control parameter set are sent to the corresponding smart home devices via a preset IoT communication link.

7. An intelligent vehicle-mounted central control device applied to the intelligent vehicle-mounted central control method of claim 1, characterized in that, The device includes: The extraction and fusion module is used to extract features and classify and fuse the collected sensor data frames to obtain the member feature subsets corresponding to each cabin member; The evaluation and sorting module is used to make a scene judgment based on the acquired vehicle motion state data and the member feature subset to determine the current driving scene type, and to evaluate the response priority of the received central control operation commands based on the driving scene type and the member feature subset to generate a control command sequence. The filtering and configuration module is used to match corresponding historical preference data from a preset historical database based on the subset of member features, so as to perform compliance filtering and interactive configuration on the control command sequence according to preset security rules and the historical preference data, and generate a set of central control interactive adjustment parameters and corresponding entertainment content configuration data.

8. A computer 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 steps of the intelligent vehicle central control method as described in any one of claims 1 to 6.

9. A readable storage medium, the readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent vehicle central control method as described in any one of claims 1 to 6.

10. A computer program product, characterised in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the intelligent vehicle central control method as described in any one of claims 1 to 6.