Adaptive vehicle entry method, system, and medium with multi-modal sensor fusion

Abstract

The application relates to a multi-modal sensor fusion adaptive vehicle entry method, system and medium, the method comprising collecting multi-source sensor data; aligning the multi-source sensor data in time and space coordinate systems, and calculating the confidence of each sensor; performing fusion processing on the multi-source sensor data through a Kalman filter to obtain the position, speed and fusion confidence of a mobile terminal relative to a vehicle; performing risk assessment based on the result of the fusion processing to calculate a risk score; selecting an authentication mode through a multi-target utility function according to the device state of the mobile terminal, the fusion confidence and the risk score; and outputting an unlocking control instruction when the unlocking trigger condition is met. The application can realize high-precision and high-safety vehicle entry experience in a complex electromagnetic environment, while significantly reducing system power consumption and improving multi-device compatibility.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and in particular to an adaptive vehicle entry method, system and medium based on multimodal sensor fusion. Background Technology

[0002] With the development of intelligent connected vehicle technology, keyless entry systems have evolved from single-mode to multi-mode. Early solutions were based on Bluetooth 4.0 / 4.2, using RSSI (Signal Strength Index) for near-vehicle unlocking, but suffered from signal interference and low positioning accuracy. To address the shortcomings of single-mode systems, the industry attempted to combine Bluetooth and NFC technologies, with Bluetooth handling long-range detection and NFC for short-range authentication. With the widespread adoption of UWB technology, some high-end models began to introduce a "Bluetooth + UWB" dual-mode solution. UWB achieves centimeter-level positioning through pulse signals and has strong resistance to metal shielding. Meanwhile, biometric technologies (such as fingerprints and facial recognition), while mature in mobile devices, are gradually being introduced into in-vehicle scenarios.

[0003] However, existing technologies still have many shortcomings. Regarding signal stability, Bluetooth solutions suffer severe signal attenuation in strong interference scenarios such as underground parking lots and near high-voltage lines, leading to unlocking delays or failures; the coexistence of multiple devices intensifies channel contention, increasing the false trigger rate; while UWB solutions offer high accuracy, they are costly and difficult to coordinate with Bluetooth timing, making them prone to location jumps. In terms of security, traditional Bluetooth solutions use static keys, making them vulnerable to relay attacks; some manufacturers' NFC solutions lack end-to-end encryption, posing a man-in-the-middle risk; biometric liveness detection is lacking, resulting in a high success rate of forgery attacks. Regarding user experience, existing solutions often require manual operation or active contact with the NFC reader, resulting in a "standing still" phenomenon; poor compatibility with multiple brands of devices leads to high pairing failure rates; and the lack of personalized settings fails to adapt to different user habits. In terms of power management, continuous scanning / connection causes rapid power consumption in mobile phones / vehicle modules, and there is a lack of dynamic power optimization strategies. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide an adaptive vehicle entry method, system and medium based on multimodal sensing fusion. By integrating multiple sensing modes such as Bluetooth, UWB, NFC and biometrics, and combining adaptive signal processing and dynamic power management strategies, it can achieve a high-precision and high-security seamless vehicle entry experience in complex electromagnetic environments, while significantly reducing system power consumption and improving multi-device compatibility.

[0005] To achieve the above objectives, the present invention adopts the following technical solution.

[0006] In a first aspect, the present invention provides an adaptive vehicle entry method based on multimodal sensor fusion, which employs the following technical solution: Multi-source sensor data is collected by a distributed multimodal fusion anchor cluster deployed on the vehicle body. The multimodal fusion anchor cluster includes a Bluetooth transceiver, an ultra-wideband transceiver, and a near-field communication card reader. The multi-source sensor data is aligned in time and space coordinate systems, and the confidence level of each sensor is calculated. The multi-source sensor data is fused using a Kalman filter to obtain the position, speed and fusion confidence of the mobile terminal relative to the vehicle. Based on the results of the fusion process, a risk assessment is performed, and a risk score is calculated. Based on the mobile terminal's device status, the fusion confidence level, and the risk score, an authentication mode is selected using a multi-objective utility function; and When the unlock trigger condition is met, an unlock control command is output.

[0007] Furthermore, in the above method, the multi-source sensing data includes: Bluetooth signal strength data is used to calculate the approximate distance between the mobile terminal and the vehicle; Ultra-wideband ranging data, through two-way time-of-flight measurement, yields the precise distance and angle of arrival between the mobile terminal and the vehicle-mounted anchor point; and Near-field communication trigger signal, used to detect the close proximity of a mobile terminal.

[0008] Furthermore, in the above method, the state vector of the Kalman filter includes the three-dimensional position of the mobile terminal in the vehicle coordinate system, the moving speed, the motion acceleration, and the fusion confidence.

[0009] Furthermore, in the above method, the Kalman filter dynamically adjusts the observation noise covariance matrix according to the real-time confidence of each sensor. When the ultra-wideband signal-to-noise ratio is high and the multipath interference is weak, the observation noise corresponding to the ultra-wideband is reduced. When the Bluetooth signal fluctuates greatly, the observation noise corresponding to the Bluetooth is increased.

[0010] Furthermore, in the above method, the risk score is calculated based on at least one of the following risk characteristics: The distance anomaly risk term characterizes the deviation between Bluetooth distance estimation and UWB ranging results; Signal anomaly risk items characterize sudden changes in signal strength, multipath interference, or abnormal packet loss rate. The risk of abnormal movement is characterized by rapid lateral approach, tailing approach, or simultaneous approach by multiple people. Time and geofencing anomaly risk items; and Equipment integrity anomaly risk item.

[0011] Furthermore, in the above method, the authentication mode includes: Ultra-wideband precision and seamless mode is activated when the mobile terminal's battery level is higher than the first threshold, the ultra-wideband function is normal, and the risk score is lower than the second threshold. Bluetooth-assisted power-saving mode is activated when the mobile terminal's battery level is within a preset range or when the user selects power saving. Near-field communication backup mode is activated when the mobile terminal's battery level falls below the third threshold, wireless communication fails continuously and times out, or near-field communication is directly triggered; and Enhanced security verification mode, enabled when the risk score is above the fourth threshold.

[0012] Furthermore, in the above method, the unlocking trigger condition includes: The fusion confidence level is greater than the confidence threshold; The distance between the mobile terminal and the target car door is less than the unlocking distance threshold; The risk score is less than the risk threshold; and User intent detected.

[0013] Furthermore, the above method also includes: When an authorized user moves away from the vehicle beyond a preset distance for a preset time, and the vehicle is unoccupied and the doors are closed, the vehicle will automatically lock.

[0014] Furthermore, the above method also includes: The sensor scanning cycle is dynamically adjusted based on the fusion confidence level and the risk score. When the fusion confidence level is high and the risk score is low, the scanning cycle is increased. When the user approaches rapidly or the risk score increases, the scanning cycle is shortened.

[0015] Furthermore, the above method also includes: When establishing a secure connection between the mobile terminal and the vehicle, the root key for the session is negotiated via elliptic curve key exchange; and A temporary session key is generated based on the session root key, random number, session counter, and timestamp through a key derivation function.

[0016] Furthermore, in the above method, the multimodal fusion anchor cluster also includes a capacitive touch sensor and a proximity sensor. The capacitive touch sensor is used to detect the user's action of pulling the door handle, and the proximity sensor is used to measure the distance to the object and the approach speed.

[0017] Furthermore, the above method also includes: When the ultra-wideband signal-to-noise ratio remains below the preset value and the Bluetooth packet loss rate is higher than the preset ratio, switch to sensor fusion enhancement mode to increase the decision weight of proximity sensor and capacitive touch sensor, and prompt the user to use near field communication touch.

[0018] Secondly, the present invention provides an adaptive vehicle entry system based on multimodal sensor fusion, which adopts the following technical solution: A multimodal fusion anchor cluster, distributed across the vehicle body, includes Bluetooth transceivers, ultra-wideband transceivers, and near-field communication card readers for collecting multi-source sensor data. The data alignment module is used to align the multi-source sensor data in terms of time and space coordinates and to calculate the confidence level of each sensor. The fusion processing module is used to fuse the multi-source sensor data through a Kalman filter to obtain the position, speed and fusion confidence of the mobile terminal relative to the vehicle. The risk assessment module is used to perform risk assessment based on the results of the fusion processing and calculate a risk score; The authentication mode selection module is used to select an authentication mode based on the device status of the mobile terminal, the fusion confidence level, and the risk score through a multi-objective utility function; and The unlock control module is used to output unlock control commands when the unlock trigger conditions are met.

[0019] Thirdly, the present invention provides a readable storage medium, which adopts the following technical solution: A readable storage medium storing computer instructions that, when executed by a processor, implement the method as described in any one of the first aspects above.

[0020] In summary, compared with the prior art, the present invention has at least one of the following beneficial technical effects: This invention collects multi-source sensor data, including Bluetooth, UWB, and NFC, through a distributed multimodal fusion anchor cluster. It then utilizes a Kalman filter for spatiotemporal alignment and fusion processing, achieving centimeter-level high-precision positioning in complex electromagnetic environments. This effectively overcomes the problems of signal attenuation and positioning jumps caused by single-sensor modes in highly interference scenarios such as underground parking lots and near high-voltage lines. By dynamically selecting the authentication mode based on a multi-objective utility function of fusion confidence and risk scoring, the system can adaptively switch between UWB precise seamless mode, Bluetooth-assisted energy-saving mode, NFC backup mode, or enhanced security verification mode according to the mobile terminal's battery status, communication quality, and security risk level, significantly reducing system power consumption while ensuring security. Furthermore, a real-time risk assessment mechanism detects potential threats such as abnormal distance, signal, and motion, effectively preventing relay attacks and spoofing attacks, thus providing users with a high-precision, high-security, and low-power seamless vehicle entry experience. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0022] Figure 1 A flowchart of an adaptive vehicle entry method based on multimodal sensor fusion according to an embodiment of the present invention is shown.

[0023] Figure 2 A flowchart of an automatic locking method according to an embodiment of the present invention is shown.

[0024] Figure 3 A flowchart of a method for dynamically adjusting the sensor scanning period according to an embodiment of the present invention is shown.

[0025] Figure 4 A flowchart of a secure key negotiation and derivation method according to an embodiment of the present invention is shown.

[0026] Figure 5 A flowchart of a sensor fusion enhancement mode switching method according to an embodiment of the present invention is shown.

[0027] Figure 6 A structural block diagram of an adaptive vehicle entry system based on multimodal sensor fusion according to an embodiment of the present invention is shown. Detailed Implementation

[0028] 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. Furthermore, it should be understood that the specific embodiments described herein are only for illustration and explanation of this application and are not intended to limit this application.

[0029] It should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments of this application. Furthermore, the descriptions of each embodiment in the following embodiments have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0030] The method steps described in this embodiment of the invention can be executed in the order described in the specific implementation, or the execution order of each step can be adjusted according to actual needs, provided that the technical problem can be solved. These are not listed one by one here.

[0031] The present invention will be further described in detail below with reference to the accompanying drawings.

[0032] Figure 1 A flowchart of an adaptive vehicle entry method 100 based on multimodal sensor fusion according to an embodiment of the present disclosure is shown. Method 100 achieves a highly accurate and secure seamless vehicle entry experience by fusing multiple sensor modalities.

[0033] Reference Figure 1 Method 100 includes step 102, which involves collecting multi-source sensor data through a distributed multi-modal fusion anchor point cluster deployed on the vehicle body. The multi-modal fusion anchor point cluster is distributed across the vehicle's four door B-pillars, rearview mirrors, and trunk. Each anchor point is connected to the central decision-making unit via an in-vehicle Ethernet or CAN FD bus, and sub-microsecond hardware time synchronization is achieved using the IEEE 802.1AS (gPTP) protocol to ensure time consistency of the multi-source sensor data.

[0034] The multimodal fusion anchor cluster includes a Bluetooth transceiver, an ultra-wideband transceiver, and a near-field communication (NFC) reader. The Bluetooth transceiver includes a directional antenna array, supports Bluetooth 5.2 and above, and is used for device discovery, coarse location, and data communication. The ultra-wideband transceiver supports the IEEE 802.15.4z HRP protocol for precise ranging and anti-relay protection. The NFC reader supports Low Power Card Detection (LPCD) mode, with standby power consumption below 5μA, for zero-power backup triggering.

[0035] In some implementations, the multimodal fusion anchor cluster also includes a context-aware sensor group. The context-aware sensor group includes a capacitive touch sensor array integrated into the door handle, ultrasonic or time-of-flight (ToF) proximity sensors arranged around the vehicle body, an in-vehicle driver monitoring camera (DMS), and a vehicle inertial measurement unit (IMU). The capacitive touch sensors detect the user's door handle pulling action; when the waveform of the capacitance value change matches a preset "door handle pulling" pattern, it is determined to be a valid user operation intent. The proximity sensors measure the distance and approach speed of objects; when the standard deviation of multiple consecutive measurements is less than a preset threshold, it is determined to be a high-confidence observation.

[0036] Multi-source sensor data includes Bluetooth signal strength data, ultra-wideband ranging data, and near-field communication trigger signals. Bluetooth signal strength data is used to calculate a coarse distance between the mobile terminal and the vehicle. The Bluetooth transceiver acquires broadcast packets from the mobile terminal, parses the device state metadata within, and reads the signal strength (RSSI). Confidence is calculated based on the stability of the RSSI received from multiple anchor points; a variance less than a preset threshold (e.g., 5 dBm) is considered high confidence. The Bluetooth signal strength data output includes the device ID, the coarse distance range (near / medium / far), and the device state vector.

[0037] Ultra-wideband (UWB) ranging data obtains the precise distance and angle of arrival (AOA) between the mobile terminal and the vehicle-mounted anchor point through two-way time-of-flight (TW-TOF) measurement. The UWB transceiver performs TW-TOF measurements, acquiring precise distance, AOA, and signal-to-noise ratio (SNR). A high confidence level is defined as an SNR greater than a preset threshold (e.g., 25 dB) and weak multipath components. The UWB ranging data outputs precise three-dimensional distance (accuracy ±10 cm) and relative azimuth.

[0038] The near-field communication (NFC) trigger signal is used to detect the proximity of a mobile terminal. The NFC reader monitors changes in radio frequency field strength using LPCD mode. A valid trigger is determined when the field strength change is steep (e.g., greater than 5 A / s) and meets near-field characteristics. The NFC trigger signal outputs a binary trigger signal and a trigger position identifier.

[0039] After acquiring multi-source sensor data, method 100 proceeds to spatiotemporal coordinate alignment and confidence calculation 104. Since the Bluetooth transceiver, ultra-wideband transceiver, near-field communication reader, and context-aware sensor group each have different sampling frequencies and coordinate reference systems, the multi-source sensor data is aligned in time and space coordinate systems, and the confidence of each sensor is calculated to provide a unified data foundation for subsequent fusion processing.

[0040] In terms of time alignment, each anchor point uses the IEEE 802.1AS (gPTP) protocol to achieve sub-microsecond hardware time synchronization, unifying the data collected by different sensors to the same time reference. In terms of spatial alignment, the vehicle coordinate system is used as a unified reference system, and the distance, angle and other observations collected by each anchor point are transformed into a three-dimensional spatial position representation in the vehicle coordinate system.

[0041] For Bluetooth signal strength data, a logarithmic distance path loss model is used to estimate the distance between the mobile terminal and the vehicle. The Bluetooth distance estimation formula is: ; in, The distance is estimated based on Bluetooth signal strength. For reference distance (e.g., 1 m). For the signal strength at the reference distance, The current received signal strength, This refers to the path loss index. The values ​​are dynamically assigned based on the vehicle's environment; for example, 2.0–2.3 for open-air environments and 2.5–3.5 for underground parking garages or highly reflective environments. This formula allows RSSI values ​​collected from different anchor points to be converted into multiple candidate distance values ​​for subsequent fusion positioning.

[0042] To reduce the impact of Bluetooth signal fluctuations on unlocking decisions, the Bluetooth signal stability confidence level was further calculated. The Bluetooth signal stability confidence level was calculated using an exponential function based on the RSSI variance: ; in, The variance of RSSI sampled values ​​within a preset time window. This serves as a stability reference variance. When RSSI fluctuations are small, A value close to 1 indicates reliable Bluetooth observations; when RSSI fluctuates significantly, The weight of Bluetooth observations is reduced in subsequent fusion processing to avoid misunderstandings caused by underground parking garages, multipath reflections, or metal vehicle body obstructions.

[0043] For ultra-wideband ranging data, the precise distance between the mobile terminal and the vehicle-mounted anchor point is calculated using two-way time-of-flight (TW-TOF) measurement. The formula for ultra-wideband two-way time-of-flight ranging is: ; in, The speed of electromagnetic wave propagation. The timestamp for when the vehicle sends the ranging request. The timestamp for the mobile terminal receiving the ranging request. The timestamp for sending the response message to the mobile terminal. The timestamp for the response message received by the vehicle is used. Through the above calculations, the distance values ​​of the mobile terminal relative to multiple ultra-wideband anchor points can be obtained, and the user's three-dimensional spatial position relative to the vehicle can be calculated by combining the angle of arrival (AOA). The confidence level of ultra-wideband ranging is determined based on the signal-to-noise ratio (SNR) and the multipath component strength. A high confidence level is defined as an SNR greater than a preset threshold (e.g., 25 dB) and a weak multipath component.

[0044] For near-field communication trigger signals, a valid trigger is determined when the RF field strength change is steep (e.g., greater than 5 A / s) and conforms to near-field characteristics, with the confidence level set to high. For capacitive touch sensors, a valid user operation intent is determined when the waveform of the capacitance value change matches a preset "pull door handle" pattern. For proximity sensors, a high-confidence observation is determined when the standard deviation of multiple consecutive measurements is less than a preset threshold (e.g., 5 cm).

[0045] After calculating the confidence scores of each sensor, method 100 proceeds to Kalman filter fusion processing 106. To unify and fuse Bluetooth signal strength data, ultra-wideband ranging data, near-field communication trigger signals, and observations from the context-aware sensor group, method 100 employs a Kalman filter to fuse the multi-source sensor data, obtaining the mobile terminal's position, speed, and fused confidence score relative to the vehicle. Method 100 uses machine learning algorithms combined with Kalman filtering for real-time fusion of multi-source data and dynamic optimization of communication parameters.

[0046] The state vector of the Kalman filter includes the mobile terminal's three-dimensional position in the vehicle coordinate system, its velocity, acceleration, and fusion confidence. Defined as: ; in, This indicates the three-dimensional position of the mobile terminal in the vehicle coordinate system. Indicates movement speed. Indicates acceleration. This represents the fusion confidence level, with a value ranging from 0 to 1.

[0047] The state prediction process of the Kalman filter uses the following formula: ; ; in, This is the predicted state value. Here is the state transition matrix. To control the input matrix, For exercise input, To predict the state covariance matrix, Let be the state covariance matrix of the previous time step. This is the process noise covariance matrix. Based on typical human motion acceleration (approximately 2 m / s²) 2 )set up.

[0048] The observation update process of the Kalman filter uses the following formula: ; ; ; in, For Kalman gain, For the observation matrix, The observation vector consists of Bluetooth distance, UWB distance, UWB angle, near-field communication trigger state, proximity sensor distance, and touch sensor intent signal. To observe the noise covariance matrix, It is the identity matrix. The updated state estimate, This is the updated state covariance matrix.

[0049] The Kalman filter updates at a frequency of 10 Hz. When a user is detected approaching rapidly (at a speed greater than 1 m / s), the update frequency is increased to 20 Hz to improve response speed and positioning accuracy.

[0050] The Kalman filter dynamically adjusts the observation noise covariance matrix based on the real-time confidence levels of each sensor. When the signal-to-noise ratio of ultra-wideband (UWB) is high and multipath interference is weak, the observation noise corresponding to UWB can be reduced, for example, by setting the UWB observation noise covariance to a certain value. This allows ultra-wideband ranging data to dominate position updates. When Bluetooth signal fluctuations are large, the corresponding Bluetooth observation noise is increased, for example, by setting the Bluetooth observation noise covariance to a certain value. When near-field communication is effectively triggered, the near-field communication triggering result is used as the input for strongly constrained observations, and its observation noise covariance is set to... This is to improve the reliability of near-field authentication.

[0051] The Kalman filter outputs the fused user's precise position, velocity, motion trajectory, and fusion confidence score. The fusion confidence score is calculated based on the confidence scores of each sensor, using the following formula: ; in, For the first Confidence level of sensor-like devices For the corresponding weights, Number of sensor types. Sensor types include Bluetooth transceivers, ultra-wideband transceivers, near-field communication readers, proximity sensors, touch sensors, biometric sensors, and motion sensors. Weights Adjustments can be made dynamically based on the environment. For example, in an underground parking garage environment, the weight of UWB, NFC, and proximity sensors can be increased, while the weight of Bluetooth can be decreased. In scenarios where mobile terminals are low on power, the weight of UWB can be decreased, while the weight of Bluetooth and NFC can be increased.

[0052] After completing the Kalman filter fusion process 106 and obtaining the location, speed, and fusion confidence of the mobile terminal, method 100 proceeds to risk assessment 108. Risk assessment 108 performs dynamic risk assessment based on the results of the fusion process, calculates a risk score, and identifies potential security threats.

[0053] The risk score is calculated based on multiple risk characteristics, including distance anomaly risk items, signal anomaly risk items, motion anomaly risk items, time and geofencing anomaly risk items, and equipment integrity anomaly risk items. The formula for calculating the risk score is: ; in, For distance anomaly risk items, This is a signal anomaly risk item. For the risk item of abnormal movement, For time and geofencing anomaly risk items, This is a risk item related to abnormal equipment integrity. , , , , The weighting coefficients for each risk item satisfy the following conditions: Risk Score The value range is from 0 to 100.

[0054] Distance anomaly risk item This characterizes the deviation between the Bluetooth estimated distance and the UWB ranging result. A significant difference between the Bluetooth estimated distance and the UWB ranging result indicates a risk of signal spoofing or relay attacks. The formula for calculating the distance anomaly risk term is: ; in, The distance is estimated based on Bluetooth signal strength. For ultra-wideband ranging results, This is the distance difference threshold. In some implementations, the distance difference threshold is... Set to 2 meters, when the difference between the Bluetooth estimated distance and the ultra-wideband ranging result is consistently greater than 2 meters, it is judged as a distance anomaly.

[0055] Signal anomaly risk item This indicates sudden changes in signal strength, multipath interference, or abnormal packet loss rate. When the Bluetooth signal strength changes drastically in a short period of time, or the ultra-wideband signal-to-noise ratio remains below a preset threshold, or the Bluetooth packet loss rate exceeds a preset ratio, the risk of signal anomalies increases accordingly.

[0056] Movement abnormality risk item This characterizes abnormal behaviors such as rapid lateral approach, tailing approach, or simultaneous approach by multiple individuals. The multimodal fusion anchor cluster includes accelerometers, which are used for motion pattern analysis to detect abnormal behaviors such as rapid running, tailing approach, or simultaneous approach by multiple individuals. The formula for calculating the motion anomaly risk term is: ; in, For the user's lateral approach speed, Accelerate the user's movement. The number of users who are simultaneously approaching each other. , , For the corresponding weighting coefficients. In some implementations, when a user rapidly approaches laterally from the non-driving side (lateral speed) If the speed is (m / s) and the driver does not turn to the driver's side for 2 seconds, it is considered an abnormal approach behavior.

[0057] Time and geofencing anomaly risk items This characterizes the risk of triggering unlock requests during time periods outside of a user's historical access habits or in unfamiliar geofenced areas. In some implementations, the time and geofence anomaly risk items are increased when unlock requests occur during non-habitual access periods (e.g., 2 AM to 5 AM) or in unfamiliar geofenced areas.

[0058] Equipment integrity anomaly risk items This indicates the device integrity status of the mobile terminal. When the device integrity verification flag broadcast by the mobile terminal is false, the device status is determined to be abnormal, and the risk of device integrity abnormality increases accordingly.

[0059] In some implementations, an enhanced security authentication process is triggered when the risk score is greater than 60, and a security event is recorded and the current unlock request is rejected when the risk score is greater than 80.

[0060] Based on the risk score obtained from risk assessment 108, method 100 proceeds to authentication mode selection 110. Authentication mode selection 110 selects an authentication mode based on the mobile terminal's device status, fusion confidence, and risk score through a multi-objective utility function.

[0061] The multimodal fusion anchor cluster includes biometric sensors, such as fingerprint sensors, facial recognition sensors, and iris recognition sensors, for user authentication. These biometric sensors, along with Bluetooth transceivers, ultra-wideband transceivers, near-field communication card readers, accelerometers, and proximity sensors, form a six-dimensional sensing array, achieving full-scenario coverage of "long-distance detection – short-distance verification – behavior analysis."

[0062] Method 100 employs a four-level authentication mechanism: Bluetooth RSSI positioning, NFC dynamic key, biometrics, and motion pattern analysis. Bluetooth RSSI positioning is used for long-range detection and coarse positioning. The NFC dynamic key is used for short-range authentication, automatically updating the key with each connection to prevent replay attacks. Biometrics is used for user identification, including liveness detection functions such as fingerprint texture analysis to defend against forgery attacks using fingerprint films or facial photographs. Motion pattern analysis uses an accelerometer to identify abnormal approach behaviors, such as rapid running, tailgating, or multiple people approaching simultaneously, triggering secondary verification or an alarm.

[0063] Method 100 dynamically optimizes the Bluetooth broadcast interval and ultra-wideband channel selection based on environmental conditions. In high-interference scenarios such as underground parking garages, Method 100 automatically adjusts the Bluetooth broadcast interval to reduce channel contention and selects an ultra-wideband channel with a high signal-to-noise ratio to improve ranging accuracy.

[0064] A multi-objective utility function is used to comprehensively evaluate the utility values ​​of various certification modalities. For the first... The utility function for each authentication mode is defined as follows: ; in, For the first The overall utility value of the authentication modes, For safety rating, Rate the user experience. In response to timeliness scoring, For power consumption cost, Assess the current risk level. to These are the weighting coefficients.

[0065] Method 100 selects the authentication mode with the highest utility value as the optimal mode. The formula for selecting the optimal mode is: .

[0066] The selection of the optimal mode simultaneously satisfies the following constraints: ; in, To the maximum allowable response latency, For power consumption budget, This is the risk threshold.

[0067] The authentication modes include Ultra Wideband Precision Seamless Mode, Bluetooth Assisted Power Saving Mode, Near Field Communication Backup Mode, and Enhanced Security Authentication Mode.

[0068] Ultra-wideband Precision Seamless Mode is the default preferred mode. It is activated when the mobile device's battery level is above the first threshold (e.g., 25%), the ultra-wideband function is normal, and the risk score is below the second threshold (e.g., 30). Ultra-wideband Precision Seamless Mode performs ultra-wideband anti-relay ranging and authentication, achieving a seamless unlocking experience with centimeter-level accuracy.

[0069] Bluetooth Assisted Power Saving Mode is activated when the mobile device's battery level is within a preset range (e.g., 10% to 25%) or when the user actively selects power saving. Bluetooth Assisted Power Saving Mode relies on Bluetooth directional positioning and behavioral verification to reduce the frequency of use of the ultra-wideband transceiver, thereby saving power.

[0070] The near-field communication backup mode is enabled under the following conditions: the mobile terminal's battery level is below the third threshold (e.g., 5%), the wireless communication failure timeout (e.g., 10 seconds), or near-field communication is directly triggered. The near-field communication backup mode is authenticated through the near-field communication card reader and inherits full user permissions, providing backup triggering capabilities for scenarios where the mobile terminal has zero battery power.

[0071] The enhanced security verification mode is automatically triggered when the risk score exceeds the fourth threshold (e.g., 60). This mode requires users to enter a PIN code via the in-vehicle touchscreen or undergo multi-factor authentication, such as biometric verification via a driver monitoring camera.

[0072] Based on the optimal authentication mode obtained from authentication mode selection 110, method 100 proceeds to unlock trigger condition judgment 112. Unlock trigger condition judgment 112 comprehensively evaluates the fusion confidence level, the distance between the user and the car door, the risk score, and the user's operational intent to determine whether the unlock conditions are met. The unlock trigger conditions are expressed as follows: ; in, To integrate confidence levels, The confidence threshold. The distance between the mobile terminal and the target car door. To unlock the distance threshold, To score the risk, As a risk threshold, Indicates the user's intention to perform an action.

[0073] The unlocking trigger conditions include the following four items: the fusion confidence score is greater than the confidence threshold; the distance between the mobile terminal and the target door is less than the unlocking distance threshold; the risk score is less than the risk threshold; and user operation intent is detected. The user operation intent is determined by detecting the door handle movement through a capacitive touch sensor or by detecting the user's approach intent through a proximity sensor. When all four conditions are met simultaneously, the result of unlocking trigger condition judgment 112 is yes, and method 100 enters the output unlocking control command 114; when any one of the four conditions is not met, the result of unlocking trigger condition judgment 112 is no, and method 100 enters the maintain lock monitoring 116, maintaining the vehicle locked state and continuing to monitor the user's position and state changes.

[0074] Method 100 supports three adjustable unlocking distance threshold settings: near, medium, and far, to accommodate different user preferences. In some implementations, the unlocking distance threshold for the near setting is set to 0.5 meters, for the medium setting to 1 meter, and for the far setting to 1.5 meters. Users select the desired unlocking distance setting via the vehicle touchscreen or a mobile application.

[0075] Method 100 includes a user behavior learning function, which uses machine learning algorithms to analyze users' frequently used unlocking distances and time periods to achieve personalized adaptation. The user behavior learning function records users' unlocking behavior data in different scenarios, including unlocking distance, unlocking time, unlocking location, and unlocking frequency, and builds a user behavior model based on historical data. The user behavior learning function dynamically adjusts the unlocking distance threshold and authentication strategy according to the user behavior model; for example, it automatically uses a longer unlocking distance in locations the user habitually visits, and automatically uses a shorter unlocking distance in unfamiliar locations to improve security.

[0076] When the unlocking trigger condition is met, method 100 outputs unlocking control command 114. Unlocking control command 114 is sent to the body actuators, which include the body domain controller, door lock motor, window controller, seat memory function, lighting device, and alarm device. After receiving unlocking control command 114, the body domain controller coordinates the door lock motor, lighting device, and seat memory function to execute the corresponding action sequence.

[0077] The execution sequence of Method 100 includes a pre-wake-up phase, a pre-unlock phase, and an unlocking phase. In the pre-wake-up phase, when the Bluetooth transceiver detects an authorized mobile terminal entering a 30-meter range, it wakes up the relevant processing unit and sensors; the latency of the pre-wake-up phase is less than or equal to 500 milliseconds. In the pre-unlock phase, when the user is located within 1.5 meters of the driver's side door, the fusion confidence is greater than 0.7, and the risk score is at a low-risk level, the welcome function is executed, including folding the rearview mirror, illuminating the welcome lights, and pre-adjusting the seat; the latency of the pre-unlock phase is less than or equal to 200 milliseconds. In the unlocking phase, when the capacitive touch sensor detects the user's action of pulling the door handle, the door is unlocked and the interior lights are illuminated; the latency of the unlocking phase is less than or equal to 100 milliseconds.

[0078] In some implementations, the seat pre-adjustment function retrieves the corresponding seat position parameters from the seat memory function based on the identity of the currently authorized user, and adjusts the seat to the preset position before the user enters the vehicle. Welcome lights include door handle lights, sill welcome lights, and interior ambient lighting; the brightness and color of the welcome lights are adjusted according to user preferences.

[0079] Method 100 achieves automatic vehicle unlocking by determining the unlock trigger condition 112 and outputting an unlock control command 114. To further enhance the user experience and ensure vehicle safety, method 100 also includes an automatic door locking function. Figure 2 A flowchart of an automatic locking method 200 according to an embodiment of the present disclosure is shown. The automatic locking method 200 automatically performs the locking operation after the user leaves the vehicle, preventing the user from forgetting to lock the vehicle due to negligence.

[0080] Reference Figure 2 The automatic door locking method 200 includes monitoring the user's location and vehicle status 202. In monitoring the user's location and vehicle status 202, a multimodal fusion anchor cluster continuously monitors the location of the mobile terminal carried by the authorized user and the current status of the vehicle. The authorized user's location is obtained through fusion positioning of multi-source sensor data collected by a Bluetooth transceiver, an ultra-wideband transceiver, and a near-field communication card reader. The vehicle status includes the door open / closed status, window open / closed status, and engine running status.

[0081] The automatic door locking method 200 includes detecting the distance and duration of movement away from the vehicle 204. In detecting the distance and duration of movement away from the vehicle 204, the distance between the authorized user and the vehicle is calculated based on the fused positioning results, and the duration of the authorized user's movement away from the vehicle is recorded. When the distance between the mobile terminal carried by the authorized user and the vehicle continues to increase and exceeds a preset door locking distance threshold, timing begins and the duration is recorded.

[0082] The automatic door locking method 200 includes detecting occupants and door status 206. In detecting occupants and door status 206, in-vehicle sensors detect the presence of occupants and door status sensors detect whether each door is closed. In-vehicle sensors include a driver monitoring camera, seat pressure sensors, and in-vehicle ultrasonic sensors. Door status sensors detect the open / closed status of each door and output a door closed indicator.

[0083] Based on information obtained from monitoring user location and vehicle status 202, detecting distance and duration of movement away 204, and detecting occupants and door status 206, the automatic locking method 200 proceeds to locking condition judgment 208. Locking condition judgment 208 comprehensively evaluates the authorized user's distance from the vehicle, duration of movement away, occupant status, and door status to determine whether the automatic locking conditions are met. The locking conditions are expressed as follows: ; in, To authorize the distance between the user and the vehicle, The locking distance threshold, To authorize users to remain away from the vehicle for a period of time beyond the lock distance threshold. For duration threshold, For the number of people inside the vehicle, This is a sign indicating that the car door is closed.

[0084] The locking condition judgment 208 includes the following judgment items: whether the authorized user's distance from the vehicle exceeds a preset distance threshold; whether the duration of the authorized user's distance from the vehicle exceeding the preset distance threshold exceeds a preset time threshold; whether the vehicle is unoccupied; and whether the vehicle door is closed. In some embodiments, the preset distance threshold is set to 3 meters, and the preset time threshold is set to 5 seconds. When all authorized devices are more than 3 meters away from the vehicle for 5 seconds, the distance and time conditions are met. In some embodiments, the locking condition is also met when the in-vehicle sensors confirm that the vehicle is unoccupied and the vehicle door remains closed for 30 seconds.

[0085] When the locking condition judgment 208 is true, i.e., the authorized user is more than a preset distance away from the vehicle for a preset time, and the vehicle is unoccupied and the doors are closed, the automatic locking method 200 enters the automatic locking operation 210. The automatic locking operation 210 sends a locking control command to the vehicle body actuator, which sequentially performs the window raising operation, the locking operation, and puts the vehicle into an alert state. The execution delay of the automatic locking operation 210 is less than or equal to 2 seconds. The window raising operation automatically raises all open windows to the fully closed position. The locking operation drives the door lock motors of each door to perform the locking action. The alert state activates the vehicle's anti-theft alarm function, triggering an audible and visual alarm when unauthorized intrusion is detected.

[0086] When the result of the locking condition judgment 208 is negative, the automatic locking method 200 enters the current state maintenance 212. In the current state maintenance 212, the vehicle maintains its current locked or unlocked state and continues to perform the user location and vehicle status monitoring 202, continuously monitoring changes in the authorized user's location and vehicle status.

[0087] In some implementations, when an authorized user leaves the vehicle without locking it and preset reminder conditions are met, the vehicle sends a push notification to the authorized user via a mobile application. The push notification includes a message indicating the vehicle is unlocked, reminding the authorized user that the vehicle is unlocked. Alternatively, the authorized user can remotely lock the vehicle via the mobile application. This remote locking operation sends a locking command to the vehicle via the mobile communication network. Upon receiving the locking command, the vehicle executes a sequence of actions: raising the windows, locking the doors, and entering a warning state.

[0088] To reduce system power consumption while ensuring positioning accuracy and response speed, method 100 dynamically adjusts the sensor scanning cycle based on fusion confidence and risk score. Figure 3 A flowchart of a sensor scanning period dynamic adjustment method 300 according to an embodiment of the present disclosure is shown.

[0089] Reference Figure 3 The scanning period adjustment method 300 includes acquiring the fusion confidence score and risk score 302. In acquiring the fusion confidence score and risk score 302, the fusion confidence score output by the Kalman filter fusion processing 106 is acquired in real time. And the risk score output by Risk Assessment 108 Fusion confidence Risk score reflects the reliability of current multi-source sensor data. It reflects the security risk level of the current scenario.

[0090] After obtaining the fusion confidence score and risk score 302, the scan cycle adjustment method 300 proceeds to the high confidence and low risk judgment 304. The high confidence and low risk judgment 304 assesses whether the fusion confidence score is higher than a preset confidence threshold and whether the risk score is lower than a preset risk threshold. The formula for calculating the scan cycle is: ; in, For the current scan cycle, This is the lower limit of the scan cycle. This is the upper limit of the scan cycle. To integrate confidence levels, Risk scoring is used. When the fusion confidence is high and the risk score is low, the scanning cycle is... Tend to When the fusion confidence is low or the risk score is high, the scanning cycle is... Tend to .

[0091] When the result of judgment 304 (high confidence and low risk) is yes, the scan cycle adjustment method 300 enters the step of increasing the scan cycle 306. Increasing the scan cycle 306 increases the sensor scan cycle to reduce system power consumption. In scenarios with high fusion confidence and low risk score, the user's location has been accurately determined and there is no security threat. Increasing the scan cycle can effectively reduce the sensor's operating frequency, thereby reducing system power consumption.

[0092] If the result of the high confidence and low risk judgment 304 is negative, the scan cycle adjustment method 300 proceeds to the user rapid approach or risk increase judgment 308. The user rapid approach or risk increase judgment 308 further assesses whether the user is rapidly approaching the vehicle or whether the risk score has increased. If the result of the user rapid approach or risk increase judgment 308 is positive, the scan cycle adjustment method 300 proceeds to shorten the scan cycle 310. Shortening the scan cycle 310 shortens the sensor scan cycle to improve response speed and ensure timely response when the user is rapidly approaching or there is a safety risk. If the result of the user rapid approach or risk increase judgment 308 is negative, the scan cycle adjustment method 300 proceeds to maintain the current scan cycle 312, keeping the current scan cycle unchanged.

[0093] In some implementations, the Bluetooth transceiver employs a periodic sleep-wake mechanism, performing a scan operation every 2 seconds. During sleep, the Bluetooth transceiver enters a low-power state, performing device discovery and signal strength acquisition only upon wake-up. Near-field communication (NFC) card readers and ultra-wideband (UWB) transceivers use an on-demand activation method, activating only after the Bluetooth transceiver detects an authorized mobile terminal within a preset range. Through these power management strategies, overall power consumption is reduced by more than 50%. In some implementations, the mobile terminal's standby time exceeds 30 days, with only a 3% increase in daily power consumption.

[0094] To ensure the security of communication data between the mobile terminal and the vehicle, method 100 employs a secure key negotiation and derivation mechanism. Figure 4 A flowchart of a secure key negotiation and derivation method 400 according to an embodiment of the present disclosure is shown. The key negotiation method 400 generates a temporary session key using an elliptic curve cryptography algorithm and automatically updates the key with each connection, achieving dynamic security protection of "one-time key".

[0095] Reference Figure 4The key negotiation method 400 includes establishing a secure connection request 402. The mobile terminal broadcasts its device status, including the hash of its digital certificate, via Bluetooth. After the vehicle verifies the certificate's validity, a secure connection request 402 is initiated between the mobile terminal and the vehicle. In the secure connection request 402, the vehicle and the mobile terminal perform two-way digital certificate verification; that is, the vehicle verifies the mobile terminal's digital certificate, and the mobile terminal verifies the vehicle's digital certificate, to prevent man-in-the-middle attacks.

[0096] Key negotiation method 400 includes an elliptic curve key exchange (ECE) negotiation session root key 404. In the ECE negotiation session root key 404, the mobile terminal and the vehicle negotiate the session root key via an ECE protocol. Elliptic curve key exchange protocols employ either ECDH P-256 or the Chinese national cryptographic algorithm SM2. In some implementations, the SM2 algorithm is used for digital certificate signature verification. Session root key. Updated each time a session is established to prevent replay attacks.

[0097] The key negotiation method 400 includes generating a random number counter timestamp 406. In generating the random number counter timestamp 406, a random number Nonce, a session counter Counter, and a timestamp Timestamp are generated. The random number Nonce is used to increase the randomness of key derivation, the session counter Counter is used to identify the current session sequence number, and the timestamp Timestamp is used to record the key generation time.

[0098] Key negotiation method 400 includes a key derivation function to generate a temporary session key 408. In the key derivation function generating the temporary session key 408, the session root key is used as a basis. A temporary session key is generated using a key derivation function, consisting of a random number (Nonce), a session counter (Counter), and a timestamp (Timestamp). The derivation formula for the temporary session key is: ; in, For temporary session keys, This is the session root key. It is a random number. For session counter, For timestamps, This is a key derivation function based on HMAC. This indicates a data concatenation operation. A new temporary session key is generated for each ultra-wideband ranging or Bluetooth secure connection process, giving the authentication process a one-time key characteristic.

[0099] The key negotiation method 400 includes encrypted communication data 410. In the encrypted communication data 410, a temporary session key is used. The communication data between the mobile terminal and the vehicle is encrypted. In some implementations, the national standard SM4 algorithm is used to encrypt key fields in the ultra-wideband ranging message. During ultra-wideband ranging, the session root key is used in the final message. Derived temporary session key Encrypted random number exchange and time-of-flight verification provide dual anti-relay protection at both the communication and physical layers.

[0100] All key materials are generated and stored by an automotive-grade Hardware Security Module (HSM). User biometric templates are stored only within the vehicle-side HSM, and comparisons are performed within the HSM, with results output as Boolean flags. All operation logs are encrypted using the vehicle's unique key before storage.

[0101] User privacy data includes location data and biometric information. User privacy data is stored locally only and is not uploaded to the cloud, complying with privacy regulations such as GDPR.

[0102] In addition to ensuring the security of communication data through the key negotiation method 400 described above, method 100 also includes a sensor fusion enhancement mode switching mechanism for complex electromagnetic environments. Figure 5 A flowchart of a sensor fusion enhancement mode switching method 500 according to an embodiment of the present disclosure is shown. The fusion enhancement mode switching method 500 is used to adaptively adjust a sensor fusion strategy to maintain the reliability of vehicle access functionality when the quality of ultra-wideband and Bluetooth communication degrades.

[0103] Reference Figure 5 The fusion enhancement mode switching method 500 includes monitoring the UWB signal-to-noise ratio (SNR) and Bluetooth packet loss rate 502. In monitoring the UWB SNR and Bluetooth packet loss rate 502, the multimodal fusion anchor cluster continuously monitors the SNR of the ultra-wideband transceiver and the packet loss rate of the Bluetooth transceiver. The ultra-wideband SNR reflects the quality of the ultra-wideband ranging signal, while the Bluetooth packet loss rate reflects the stability of the Bluetooth communication link. In underground parking garages, near high-voltage lines, or in environments with strong electromagnetic interference, both the ultra-wideband SNR and Bluetooth communication quality can be affected.

[0104] After monitoring the UWB signal-to-noise ratio (SNR) and Bluetooth packet loss rate 502, the fusion enhancement mode switching method 500 enters the enhancement mode trigger condition judgment 504. The enhancement mode trigger condition judgment 504 evaluates whether the UWB SNR is consistently below a preset value and whether the Bluetooth packet loss rate is higher than a preset percentage. In some implementations, when the UWB SNR is consistently below 15 dB and the Bluetooth packet loss rate is higher than 50%, it is determined that the enhancement mode trigger condition is met.

[0105] When the result of the enhancement mode trigger condition judgment 504 is yes, the fusion enhancement mode switching method 500 enters the sensor fusion enhancement mode 506. In the sensor fusion enhancement mode 506, method 100 switches to the sensor fusion enhancement mode, increasing the decision weights of the proximity sensor and the capacitive touch sensor. The sensor fusion enhancement mode compensates for the loss of positioning accuracy caused by the degradation of ultra-wideband and Bluetooth signal quality by increasing the weights of the proximity sensor and the capacitive touch sensor in the fusion confidence calculation. In the sensor fusion enhancement mode, the proximity sensor is used to measure the distance between the user and the vehicle, and the capacitive touch sensor is used to detect the user's door handle pulling action. Together, they provide auxiliary judgment basis for the user's approach intention and location.

[0106] After switching to sensor fusion enhancement mode 506, fusion enhancement mode switching method 500 proceeds to prompting near-field communication touch 508. In prompting near-field communication touch 508, method 100 prompts the user to use near-field communication touch to access the card reader. Prompt methods include sending a push notification via a mobile terminal application, indicating the near-field communication card reader's location by flashing vehicle lights, or playing a voice prompt via the vehicle's speakers. Following the prompts, the user brings their mobile terminal close to the near-field communication card reader to complete authentication and unlocking operations via near-field communication.

[0107] When the result of the enhanced mode trigger condition judgment 504 is negative, the fusion enhanced mode switching method 500 enters the current fusion mode maintenance 510. In the current fusion mode maintenance 510, method 100 keeps the current sensor fusion strategy unchanged and continues to perform UWB signal-to-noise ratio and Bluetooth packet loss rate monitoring 502 to continuously monitor changes in communication quality.

[0108] Method 100 also includes a sensor failure degradation strategy. When the capacitive touch sensor fails, the unlocking condition is downgraded to a high-confidence positioning combined with the user remaining stationary for more than 2 seconds. In this degradation mode, when the fusion confidence output by the Kalman filter fusion processing 106 is higher than a preset threshold, and the user is detected to have remained stationary near the target door for more than 2 seconds, the unlocking trigger condition is determined to be met. This degradation strategy ensures that the vehicle entry function can still operate normally in the event of a capacitive touch sensor failure.

[0109] Method 100 also includes a backup mechanism for scenarios where the mobile terminal has zero battery power. When the mobile terminal is completely out of power, the near-field communication (NFC) reader detects the proximity of the mobile terminal using Low Power Card Detection (LPCD) mode. The mobile terminal's NFC chip has a pre-installed offline authorization credential encrypted with a Hardware Security Module (HSM) shared key. The NFC reader reads and verifies this offline authorization credential; upon successful verification, it inherits the corresponding user permissions and performs the unlocking operation. This backup mechanism provides ultimate backup capability for scenarios where the mobile terminal has zero battery power.

[0110] Method 100 also includes an arbitration strategy for scenarios where multiple authorized users approach the vehicle simultaneously. When multiple authorized users approach the vehicle at the same time, Method 100 determines the response priority based on an arbitration strategy of spatial proximity, temporal priority, and primary user priority. Spatial proximity means that the authorized user closest to the vehicle has higher priority. Temporal priority means that the authorized user who enters the detection range first has higher priority. Primary user priority means that the vehicle owner or the preset primary user has the highest priority. Method 100 responds to the device with the highest priority and sends a status notification to other user terminals, informing them that the vehicle is currently responding to another authorized user.

[0111] Each step of the above method 100 is executed by the multimodal sensor fusion adaptive vehicle entry system 600. Figure 6 A structural block diagram of an adaptive vehicle access system 600 with multimodal sensor fusion according to an embodiment of the present disclosure is shown.

[0112] Reference Figure 6 The vehicle entry system 600 includes a multimodal fusion anchor cluster 602, a data alignment module 610, a fusion processing module 612, a risk assessment module 614, an authentication mode selection module 616, and an unlock control module 618.

[0113] A multimodal fusion anchor cluster 602 is distributed across the vehicle body to collect multi-source sensor data. In some embodiments, the multimodal fusion anchor cluster 602 is distributed across the B-pillars of the four doors, rearview mirrors, and the trunk. The multimodal fusion anchor cluster 602 includes a Bluetooth transceiver 604, an ultra-wideband transceiver 606, and a near-field communication (NFC) reader 608. The Bluetooth transceiver 604 is used for device discovery, coarse positioning, and data communication. The ultra-wideband transceiver 606 is used for precise ranging and anti-relay protection. The NFC reader 608 is used for zero-battery backup triggering. The multi-source sensor data collected by the multimodal fusion anchor cluster 602 includes Bluetooth signal strength data, ultra-wideband ranging data, and NFC trigger signals. Specific implementation details are described in step 102 of method 100 above and will not be repeated here.

[0114] In some implementations, the multimodal fusion anchor cluster 602 also includes distance sensors, including millimeter-wave radar and lidar, for precise distance measurement. The millimeter-wave radar and lidar, together with the Bluetooth transceiver 604, the ultra-wideband transceiver 606, and the near-field communication card reader 608, constitute a multimodal sensing array to achieve multi-dimensional user location awareness.

[0115] The data alignment module 610 is coupled to the multimodal fusion anchor cluster 602 and is used to align the time and space coordinates of multi-source sensor data and calculate the confidence level of each sensor. The specific implementation of the data alignment module 610 is described in the description of spatiotemporal coordinate alignment and confidence level calculation 104 in the above method 100, and will not be repeated here.

[0116] The fusion processing module 612 is coupled to the data alignment module 610 and is used to perform fusion processing on multi-source sensor data through a Kalman filter to obtain the position, speed, and fusion confidence of the mobile terminal relative to the vehicle. The specific implementation of the fusion processing module 612 is described in the Kalman filter fusion processing 106 of method 100 above, and will not be repeated here.

[0117] The risk assessment module 614 is coupled to the fusion processing module 612 and is used to perform risk assessment and calculate risk scores based on the results of the fusion processing. The specific implementation of the risk assessment module 614 is described in the risk assessment 108 section of method 100 above, and will not be repeated here.

[0118] The authentication mode selection module 616 is coupled to the risk assessment module 614 and is used to select an authentication mode based on the mobile terminal's device status, fusion confidence, and risk score through a multi-objective utility function. The specific implementation of the authentication mode selection module 616 is described in the authentication mode selection 110 section of method 100 above, and will not be repeated here. In some implementations, the authentication mode selection module 616 includes a decision evaluator, which evaluates the mode utility every 200 milliseconds. The formula for calculating the mode utility is: ; in, The utility value for the ultra-wideband precision and seamless mode. For safety rating, Rate the user experience. This refers to power consumption cost. When the mobile terminal's battery level drops to 20%, the power consumption cost coefficient in the utility function increases, and the system automatically switches to Bluetooth-assisted power-saving mode.

[0119] The unlock control module 618 is coupled to the authentication mode selection module 616 and is used to output an unlock control command when the unlock trigger condition is met. The specific implementation of the unlock control module 618 refers to the description of unlock trigger condition judgment 112 and output unlock control command 114 in the above method 100, and will not be repeated here.

[0120] The vehicle entry system 600 also includes a central decision unit, which runs on the vehicle infotainment system (IVI) or a standalone electronic control unit (ECU). The central decision unit operates middleware based on an extension of the AUTOSAR Adaptive platform, responsible for aggregating all sensor data, executing fusion algorithms, performing risk assessments, and making final control decisions. A data alignment module 610, a fusion processing module 612, a risk assessment module 614, an authentication mode selection module 616, and an unlock control module 618 are integrated into the central decision unit.

[0121] The Vehicle Access System 600 supports OTA (Over-The-Air) upgrades to continuously optimize signal processing algorithms and power management strategies, adapting to future technological advancements such as the high bandwidth transmission of Bluetooth 6.0 and further developments in ultra-wideband technology. The Vehicle Access System 600 supports protocol standardization, such as the ICCOA protocol, enabling cross-brand device compatibility between automakers and mobile phone manufacturers. The Vehicle Access System 600 also supports scenario-based functions, including a pet mode and the ability to keep the power on even when temporarily removed from the vehicle.

[0122] In terms of performance, the Vehicle Entry System 600 achieves an average seamless unlocking time of 128 milliseconds in typical scenarios, with a standby vehicle current of less than 50μA. The Vehicle Entry System 600 achieves 100% defense against relay attacks under laboratory conditions. The unlocking success rate of the Vehicle Entry System 600 reaches 99.5%, with a 40% improvement in success rates in complex environments such as underground parking garages. The response time of the Vehicle Entry System 600 is reduced to less than 0.8 seconds, and the risk of unauthorized unlocking is reduced to below 0.005%.

[0123] This invention also includes a readable storage medium.

[0124] A computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method described in any of the above embodiments. The computer-readable storage medium may include any entity or device capable of carrying a computer program, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc. The computer program includes computer program code. The computer program code may be in the form of source code, object code, an executable file, or some intermediate form, etc. The computer-readable storage medium may include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), and a software distribution medium, etc.

[0125] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0126] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a system including a processing module or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0127] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention 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 the present invention.

Claims

1. An adaptive vehicle entry method based on multimodal sensor fusion, characterized in that, include: Multi-source sensor data is collected by a distributed multimodal fusion anchor cluster deployed on the vehicle body. The multimodal fusion anchor cluster includes a Bluetooth transceiver, an ultra-wideband transceiver, and a near-field communication card reader. The multi-source sensor data is aligned in time and space coordinate systems, and the confidence level of each sensor is calculated. The multi-source sensor data is fused using a Kalman filter to obtain the position, speed and fusion confidence of the mobile terminal relative to the vehicle. Based on the results of the fusion process, a risk assessment is performed, and a risk score is calculated. Based on the device status of the mobile terminal, the fusion confidence level, and the risk score, an authentication mode is selected through a multi-objective utility function. as well as When the unlock trigger condition is met, an unlock control command is output.

2. The method according to claim 1, characterized in that, The multi-source sensing data includes: Bluetooth signal strength data is used to calculate the approximate distance between the mobile terminal and the vehicle; Ultra-wideband ranging data, through two-way time-of-flight measurement, yields the precise distance and angle of arrival between the mobile terminal and the vehicle-mounted anchor point; and Near-field communication trigger signal, used to detect the close proximity of a mobile terminal.

3. The method according to claim 1, characterized in that, The state vector of the Kalman filter includes the three-dimensional position of the mobile terminal in the vehicle coordinate system, the moving speed, the motion acceleration, and the fusion confidence.

4. The method according to claim 1, characterized in that, The Kalman filter dynamically adjusts the observation noise covariance matrix based on the real-time confidence level of each sensor. When the ultra-wideband signal-to-noise ratio is high and multipath interference is weak, the observation noise corresponding to the ultra-wideband signal is reduced. When the Bluetooth signal fluctuates greatly, the observation noise corresponding to the Bluetooth signal is increased.

5. The method according to claim 1, characterized in that, The risk score is calculated based on at least one of the following risk characteristics: The distance anomaly risk term characterizes the deviation between Bluetooth distance estimation and UWB ranging results; Signal anomaly risk items characterize sudden changes in signal strength, multipath interference, or abnormal packet loss rate. The risk of abnormal movement is characterized by rapid lateral approach, tailing approach, or simultaneous approach by multiple people. Time and geofencing anomaly risk items; as well as Equipment integrity anomaly risk item.

6. The method according to claim 1, characterized in that, The authentication modes include: Ultra-wideband precision and seamless mode is activated when the mobile terminal's battery level is higher than the first threshold, the ultra-wideband function is normal, and the risk score is lower than the second threshold. Bluetooth-assisted power-saving mode is activated when the mobile terminal's battery level is within a preset range or when the user selects power saving. Near-field communication backup mode is activated when the mobile terminal's battery level falls below the third threshold, wireless communication fails continuously and times out, or near-field communication is directly triggered; and Enhanced security verification mode, enabled when the risk score is above the fourth threshold.

7. The method according to claim 1, characterized in that, The unlocking trigger conditions include: The fusion confidence level is greater than the confidence threshold; The distance between the mobile terminal and the target car door is less than the unlocking distance threshold; The risk score is less than the risk threshold; and User intent detected.

8. The method according to claim 1, characterized in that, Also includes: When an authorized user moves away from the vehicle beyond a preset distance for a preset time, and the vehicle is unoccupied and the doors are closed, the vehicle will automatically lock.

9. The method according to claim 1, characterized in that, Also includes: The sensor scanning cycle is dynamically adjusted based on the fusion confidence level and the risk score. When the fusion confidence level is high and the risk score is low, the scanning cycle is increased. When the user approaches rapidly or the risk score increases, the scanning cycle is shortened.

10. The method according to claim 1, characterized in that, Also includes: When establishing a secure connection between the mobile terminal and the vehicle, the root key of the negotiation session is negotiated through elliptic curve key exchange. as well as A temporary session key is generated based on the session root key, random number, session counter, and timestamp through a key derivation function.

11. The method according to claim 1, characterized in that, The multimodal fusion anchor cluster also includes a capacitive touch sensor and a proximity sensor. The capacitive touch sensor is used to detect the user's action of pulling the door handle, and the proximity sensor is used to measure the distance to the object and the approach speed.

12. The method according to claim 1, characterized in that, Also includes: When the ultra-wideband signal-to-noise ratio remains below the preset value and the Bluetooth packet loss rate is higher than the preset ratio, switch to sensor fusion enhancement mode to increase the decision weight of proximity sensor and capacitive touch sensor, and prompt the user to use near field communication touch.

13. An adaptive vehicle entry system based on multimodal sensor fusion, characterized in that, include: A multimodal fusion anchor cluster, distributed across the vehicle body, includes Bluetooth transceivers, ultra-wideband transceivers, and near-field communication card readers for collecting multi-source sensor data. The data alignment module is used to align the multi-source sensor data in terms of time and space coordinates and to calculate the confidence level of each sensor. The fusion processing module is used to fuse the multi-source sensor data through a Kalman filter to obtain the position, speed and fusion confidence of the mobile terminal relative to the vehicle. The risk assessment module is used to perform risk assessment based on the results of the fusion processing and calculate a risk score; The authentication mode selection module is used to select an authentication mode based on the device status of the mobile terminal, the fusion confidence level, and the risk score through a multi-objective utility function. as well as The unlock control module is used to output unlock control commands when the unlock trigger conditions are met.

14. A readable storage medium, characterized in that, The readable storage medium stores computer instructions that, when executed by a processor, implement the method as described in any one of claims 1-12.

Images