Bluetooth digital key authentication method and system based on multi-anchor fusion

Abstract

The present application relates to the technical field of identity authentication, in particular to a Bluetooth digital key authentication method and system based on multi-anchor fusion, comprising the following steps: a vehicle is provided with multiple anchor nodes to receive continuous wave signals and signal response packets transmitted by a Bluetooth digital key, a transmission timing phase state matrix is generated, the internal frequency components of the transmission timing phase state matrix are divided into trunk translation frequency parameters and arm periodic swing frequency parameters, and a Doppler frequency shift feature group is generated.In the present application, abnormal features can be extracted from time distribution deviation, and unlicensed intermediate person hijacking state features are generated accordingly, and vehicle anti-theft stop lock instructions are further established, so that the identity authentication process has three-layer effects of pre-screening, identification and blocking after approaching, which not only enhances the identification accuracy of legal users, but also enhances the inhibition capability of relay hijacking, disguised approach and asynchronous forwarding.

Description

Technical Field

[0001] This invention relates to the field of identity verification technology, and in particular to a Bluetooth digital key authentication method and system based on multi-anchor point fusion. Background Technology

[0002] Existing identity verification technologies, in practice, focus primarily on confirming the authenticity of the identity subject, verifying the legitimacy of the access recipient, and ensuring the trustworthiness of the interaction process. Authentication is largely based on credential matching, session establishment, and compliance with access rules. While these technologies can determine whether a request has apparent legitimacy, they struggle to characterize whether the request initiation process aligns with genuine proximity behavior. In near-field wireless scenarios, even after legitimate digital credentials are forwarded, transferred, or proxied, the apparent identity information may remain consistent. Judging solely by the validity of the credentials and the success of the interaction often fails to distinguish between a genuine holder approaching a vehicle on foot and a remote device extending the signal via an intermediate link. Abnormal behavior can easily infiltrate normal sessions. Therefore, improvements are needed. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the existing technology and to propose a Bluetooth digital key authentication method and system based on multi-anchor point fusion.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a Bluetooth digital key authentication method based on multi-anchor point fusion, comprising the following steps: The vehicle is equipped with multiple anchor nodes to receive the continuous wave signal and signal response packet transmitted by the Bluetooth digital key, generate a transmission timing phase state matrix, divide the frequency components inside the transmission timing phase state matrix into torso translation frequency parameters and arm periodic swing frequency parameters, and generate Doppler frequency shift feature groups. The trunk translation frequency parameters within the Doppler frequency shift feature group are converted into radial movement velocity values, the arm periodic swing frequency parameters within the Doppler frequency shift feature group are converted into human walking step frequency values, gait motion fusion indexes are calculated, preset anti-unauthorized movement interval parameters are retrieved, and unauthorized device approach judgment marks are generated by comparison. When the unauthorized device proximity determination flag shows a state of not exceeding the outer boundary, the arrival timestamp parameters of the corresponding packets of the received signal response packets from the aforementioned multiple anchor nodes are called to calculate the discrete amount of the arrival interval of the response packets. Based on the discrete amount of the arrival interval of the response packets, the link deterministic timing jitter score is calculated and generated. The anti-relay hijacking hardware tolerance baseline value is retrieved. When the link deterministic timing jitter is greater than the anti-relay hijacking hardware tolerance baseline value, an unauthorized man-in-the-middle hijacking state feature is generated. Based on the unauthorized man-in-the-middle hijacking state feature, a vehicle anti-theft stop and lock command is established.

[0005] Preferably, the step of obtaining the Doppler frequency shift feature set is as follows: Obtain the signal response packets corresponding to the continuous wave signals received by multiple anchor nodes respectively, extract the packet arrival timestamp and instantaneous phase sequence value attached to each signal response packet, classify the packet arrival timestamp and instantaneous phase sequence value according to the node number of the anchor node, and then arrange the corresponding instantaneous phase sequence values ​​one by one according to the order of packet arrival timestamps to form a time-series spliced ​​phase sequence group. According to the time-series splicing phase sequence group, read the packet arrival timestamps arranged sequentially under each node number, read the instantaneous phase sequence value corresponding to each packet arrival timestamp, perform arrangement splicing and recombination according to the packet arrival timestamp as the vertical time sequence position and the instantaneous phase sequence value as the horizontal phase sampling position, write the arrangement results of the anchor nodes into the same structure in a unified order, and generate the transmission time-series phase state matrix. The frequency components that change continuously within the transmission timing phase state matrix are read line by line. The frequency segments corresponding to the continuous displacement of the human torso are screened out and marked as torso translation frequency parameters. The frequency segments corresponding to the reciprocating swing of the arm are screened out and marked as arm periodic swing frequency parameters. Then, compression mapping, parameter alignment and scale unification dimensionality reduction calibration operations are performed on the torso translation frequency parameters and the arm periodic swing frequency parameters to obtain the Doppler frequency shift feature group.

[0006] Preferably, the steps for obtaining the gait motion fusion index are as follows: Based on the Doppler frequency shift characteristic group, the frequency values ​​of the torso translation frequency parameter at continuous sampling positions are read and the radial movement speed value is calculated item by item in combination with the carrier wavelength. The absolute value of the difference between the radial movement speed values ​​corresponding to adjacent sampling positions within the observation segment is calculated and the average value is calculated. Then, the frequency values ​​of the arm period swing frequency parameter at continuous sampling positions are read and the human walking step frequency value is calculated according to the number of swing cycles per unit time. The absolute value of the difference between the human walking step frequency values ​​corresponding to adjacent sampling positions within the observation segment is calculated and the average value is calculated, forming a set of radial movement speed value and human walking step frequency value fluctuation parameters. Based on the radial movement speed value and human walking cadence value fluctuation parameter group, the average radial movement speed, the average human walking cadence, the average absolute value of the radial movement speed difference, and the average absolute value of the human walking cadence difference within the same observation segment are extracted to calculate the gait motion fusion index.

[0007] Preferably, the step of obtaining the unauthorized device proximity determination identifier is as follows: The lower boundary value and the outer boundary value of the preset anti-unauthorized movement interval parameters are retrieved, and the gait motion fusion index is compared with the lower boundary value and the outer boundary value of the interval one by one. When the gait motion fusion index is between the lower boundary value and the outer boundary value of the interval, it is recorded as not exceeding the outer boundary. When the gait motion fusion index is greater than the outer boundary value of the interval or less than the lower boundary value of the interval, it is recorded as exceeding the outer boundary, thus obtaining the unauthorized device approach determination mark.

[0008] Preferably, the step of obtaining the discrete amount of the arrival interval of the response packet is as follows: If the unauthorized device proximity determination flag shows a state of not exceeding the outer boundary, then multiple anchor nodes are called to receive the signal response packet corresponding to the packet arrival timestamp parameter. All packet arrival timestamp parameters are rearranged according to the time sequence of the packet arrival timestamp parameters. The time difference between two adjacent packet arrival timestamp parameters is calculated item by item and formed into a continuous record. Then, all time differences are written into a unified sequence structure according to the observation segment order to obtain the response packet arrival time interval sequence. Based on the response packet arrival time interval sequence, the deviation between the response packet arrival time interval and the average response packet arrival time interval is calculated item by item. All deviations are squared and summed. Then, the deviations are divided equally according to the number of response packet arrival time intervals and square rooted. The resulting value is used as the standard deviation of the response packet arrival time interval sequence to obtain the response packet arrival time interval discreteness.

[0009] Preferably, the step of obtaining the link deterministic timing jitter score is as follows: Based on the discrete amount of the response packet arrival interval, the drift rate parameter of the internal hardware crystal oscillator of the Bluetooth digital key is extracted and the average value of the response packet arrival time interval is read to calculate the link deterministic timing jitter score.

[0010] Preferably, the step of obtaining the unauthorized man-in-the-middle hijacking state characteristics is as follows: Retrieve the anti-relay hijacking hardware tolerance baseline value, read the current value of the link deterministic timing jitter score, compare the link deterministic timing jitter score with the anti-relay hijacking hardware tolerance baseline value, write the comparison result that the link deterministic timing jitter score is greater than the anti-relay hijacking hardware tolerance baseline value into the anomaly judgment field, and write the judgment conclusion that there is an asynchronous forwarding anomaly of the device into the status record area to obtain the device asynchronous forwarding anomaly judgment result; Based on the device asynchronous forwarding anomaly determination result, the determination conclusion of the device asynchronous forwarding anomaly in the status record area is extracted. The preset hijacking marking rules are called to write the anomaly source category, anomaly timing category, and anomaly forwarding category item by item. The determination conclusion of the device asynchronous forwarding anomaly is mapped to the unauthorized man-in-the-middle hijacking mark content. Then the unauthorized man-in-the-middle hijacking mark content is written into the security status field to generate the unauthorized man-in-the-middle hijacking status feature.

[0011] Preferably, the step of obtaining the vehicle anti-theft locking command is as follows: Based on the unauthorized man-in-the-middle hijacking state characteristics, the command sending interface of the vehicle engine control unit is invoked to write the control code content of the blocking command to the vehicle engine control unit, the prohibition state of cutting off the vehicle start permission is written to the vehicle start permission register area, and the closure state of terminating the peripheral signal interaction is written to the peripheral signal interaction control area. Then, the blocking command, prohibition state and closure state are associated and solidified to establish the vehicle anti-theft stopping and locking command.

[0012] The present invention also provides a system comprising: The multi-anchor point signal acquisition and Doppler feature extraction module is used to set multiple anchor point nodes in the vehicle to receive the continuous wave signal and signal response packet transmitted by the Bluetooth digital key, generate the transmission time-phase state matrix, divide the frequency components inside the transmission time-phase state matrix into torso translation frequency parameters and arm periodic swing frequency parameters, and generate Doppler frequency shift feature groups. The gait motion fusion determination module is used to convert the trunk translation frequency parameters in the Doppler frequency shift feature group into radial movement speed values, convert the arm periodic swing frequency parameters in the Doppler frequency shift feature group into human walking step frequency values, calculate gait motion fusion index, retrieve preset anti-unauthorized movement interval parameters, and compare and generate unauthorized device approach determination identifiers. The link timing jitter detection module is used to call the arrival timestamp parameters of the corresponding packets of the received signal response packets of the aforementioned multiple anchor nodes when the unauthorized device approach determination flag shows a state of not exceeding the outer limit, calculate the discrete amount of the arrival interval of the response packets, and generate the link deterministic timing jitter score based on the discrete amount of the arrival interval of the response packets. The relay hijacking detection and vehicle locking module is used to retrieve the anti-relay hijacking hardware tolerance baseline value. When the link deterministic timing jitter is greater than the anti-relay hijacking hardware tolerance baseline value, an unauthorized man-in-the-middle hijacking state feature is generated. Based on the unauthorized man-in-the-middle hijacking state feature, a vehicle anti-theft stop and lock command is established.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, multiple anchor nodes synchronously receive continuous wave signals and signal response packets around the Bluetooth digital key proximity authentication link. The packet arrival timestamps and instantaneous phase sequence values ​​are reconstructed into a transmission timing phase state matrix. Then, torso translation frequency parameters and arm periodic swing frequency parameters are extracted from the frequency components within the matrix to form a Doppler frequency shift feature group. The authentication basis is expanded from a single identity credential to a joint determination of spatial location, human movement trajectory, limb swing rhythm, and link timing fluctuations. This results in finer granularity of identity verification and a more complete characterization of proximity behavior. Radial movement velocity values ​​and human walking gait frequency values ​​are further fused into a gait fusion index, which is compared with preset anti-unauthorized movement interval parameters. This allows for early identification of device activity trajectories that do not conform to real-world usage habits during the proximity phase, reducing the exploitable timeframe after unauthorized devices approach the vehicle. The response packet arrival timestamp parameter continues to participate in the calculation of the response packet arrival interval discreteness and the deterministic timing jitter of the link. The authentication perspective has been successfully extended from the surface communication to the underlying link rhythm consistency and hardware clock stability. When encountering scenarios such as asynchronous forwarding, remote handling, and link grafting, abnormal features can be extracted from the time distribution deviation, and unauthorized man-in-the-middle hijacking state features can be generated accordingly. This further establishes vehicle anti-theft stopping and locking commands, making the identity verification process have a three-layer effect of pre-approach screening, approach identification, and post-approach blocking. This not only enhances the accuracy of legitimate user identification but also enhances the ability to suppress relay hijacking, spoofed approach, and asynchronous forwarding. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the steps of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0016] Please see Figure 1 This invention provides a technical solution for a Bluetooth digital key authentication method based on multi-anchor point fusion, comprising the following steps: The vehicle is equipped with multiple anchor nodes to receive the continuous wave signal and signal response packet transmitted by the Bluetooth digital key, generate a transmission timing phase state matrix, divide the frequency components inside the transmission timing phase state matrix into torso translation frequency parameters and arm periodic swing frequency parameters, and generate Doppler frequency shift feature groups. The trunk translation frequency parameters within the Doppler frequency shift feature group are converted into radial movement velocity values, the arm periodic swing frequency parameters within the Doppler frequency shift feature group are converted into human walking step frequency values, gait motion fusion indexes are calculated, preset anti-unauthorized movement interval parameters are retrieved, and unauthorized device approach judgment marks are generated by comparison. When the unauthorized device approaches and the determination flag shows that it has not exceeded the outer boundary, the arrival timestamp parameters of the corresponding packets of the signal response packets received by the aforementioned multiple anchor nodes are called to calculate the discrete amount of the arrival interval of the response packets. Based on the discrete amount of the arrival interval of the response packets, the link deterministic timing jitter score is calculated and generated. The baseline value of the anti-relay hijacking hardware tolerance is retrieved. When the deterministic timing jitter of the link is greater than the baseline value of the anti-relay hijacking hardware tolerance, an unauthorized man-in-the-middle hijacking state feature is generated. Based on the unauthorized man-in-the-middle hijacking state feature, a vehicle anti-theft stop and lock command is established.

[0017] The steps for obtaining the Doppler frequency shift feature set are as follows: Obtain the signal response packets corresponding to the continuous wave signals received by multiple anchor nodes respectively, extract the packet arrival timestamp and instantaneous phase sequence value attached to each signal response packet, classify the packet arrival timestamp and instantaneous phase sequence value according to the node number of the anchor node, and then arrange the corresponding instantaneous phase sequence values ​​one by one according to the order of packet arrival timestamps to form a time-series spliced ​​phase sequence group. Based on the timing splicing phase sequence group, read the packet arrival timestamps arranged sequentially under each node number, read the instantaneous phase sequence value corresponding to each packet arrival timestamp, perform arrangement splicing and recombination according to the packet arrival timestamp as the vertical timing position and the instantaneous phase sequence value as the horizontal phase sampling position, write the arrangement results of the anchor nodes into the same structure in a unified order, and generate the transmission timing phase state matrix. The frequency components that change continuously within the transmission timing phase state matrix are read line by line. The frequency segments corresponding to the continuous displacement of the human torso are screened out and marked as torso translation frequency parameters. The frequency segments corresponding to the reciprocating swing of the arm are screened out and marked as arm periodic swing frequency parameters. Then, compression mapping, parameter alignment and scale unification dimensionality reduction calibration operations are performed on the torso translation frequency parameters and the arm periodic swing frequency parameters to obtain the Doppler frequency shift feature group.

[0018] Specifically, the process involves acquiring the signal response packets corresponding to the continuous wave signals received by multiple anchor nodes, parsing the physical layer data structure of each response packet, reading the packet arrival timestamp recorded in the data header, extracting the in-phase and quadrature signal components contained in the load area, calculating the arctangent function value of the ratio of the quadrature signal component to the in-phase signal component to obtain the corresponding instantaneous phase sequence value, extracting the phase transition points in the instantaneous phase sequence value, and performing phase expansion processing by adding or subtracting twice the value of pi to the current phase if the absolute value of the difference between adjacent phases is greater than pi, obtaining a continuous true phase sequence. The process also involves reading the hardware identifier code corresponding to each signal response packet, converting the hardware identifier code into a numerical sequence as the node number (e.g., assigning the left door anchor point to number 1, the right door anchor point to number 2, and the trunk anchor point to number 3), and placing the packet arrival timestamps and instantaneous phase sequence values ​​with the same node number into the same temporary storage area. In each temporary storage area, the packet arrival timestamp is retrieved and compared with a preset time validity lower limit. The method for setting the time validity lower limit is to read the minimum circuit response time of the hardware receiver processing a single data packet and add a hardware clock drift margin. For example, if the minimum circuit response time is 1.5 milliseconds and the clock drift margin is set to 0.1 milliseconds, then the time validity lower limit is set to 1.6 milliseconds. The time difference between two adjacent packet arrival timestamps is calculated. If the time difference is less than 1.6 milliseconds, it is determined to be an invalid retransmission packet caused by signal multipath reflection. The timestamp and the instantaneous phase sequence value are simultaneously removed. After cleaning, the remaining packet arrival timestamps are arranged in ascending order of time. The instantaneous phase sequence values ​​corresponding to the sorted packet arrival timestamps are extracted and concatenated item by item. The packet arrival timestamps under each node number are arranged and combined with the corresponding phase data streams and packaged and bound to form a time-series concatenated phase sequence group.

[0019] Based on the aforementioned generated time-series spliced ​​phase sequence group, the node numbers contained within are read, and the packet arrival timestamps arranged sequentially under each node number are retrieved, along with the continuous real phase sequence corresponding to each packet arrival timestamp. The global minimum time value among all packet arrival timestamps is identified as the starting time reference point, and the global maximum time value is identified as the ending time reference point. The observation interval between the starting and ending time reference points is divided according to the preset sampling step size parameter. The specific method for setting the sampling step size parameter is to obtain the broadcast time interval of the continuous wave signal transmitted by the Bluetooth digital key. For example, if the broadcast time interval is 20 milliseconds according to the configuration parameters, half of the broadcast time interval is taken as the sampling step size, i.e., the sampling step size parameter is set to 10 milliseconds. A globally reference time series with equal intervals is generated at fixed intervals of 10 milliseconds. For each node number, the recorded packet arrival timestamps are... For time nodes registered with the global reference time series, if there is no corresponding packet arrival timestamp in the global reference time series, the instantaneous phase sequence values ​​corresponding to the arrival timestamps of the two adjacent real packets before and after the time node are extracted. The supplementary phase value at the time node is calculated using linear interpolation. The instantaneous phase sequence value of all nodes is then filled in. The time nodes in the global reference time series are used as the vertical time sequence arrangement positions, and the node numbers and the corresponding filled instantaneous phase sequence values ​​are used as the horizontal phase sampling positions. A two-dimensional arrangement, splicing and recombination is performed. According to the unified order of the node numbers from smallest to largest, for example, the phase sequence of node number 1 is placed in column 1, the phase sequence of node number 2 is placed in column 2, and the phase sequence of node number 3 is placed in column 3. The arrangement sequences of all anchor nodes are written into the same two-dimensional data structure in this order to generate the transmission time sequence phase state matrix.

[0020] The frequency components that continuously change within the previously generated transmission timing phase state matrix are read line by line. Fourier transforms are performed on the frequency data sequences at different time series, converting them to the frequency domain. The center frequency value of the corresponding spectrum in the frequency domain is compared item by item with the preset frequency band boundaries. The method for setting the frequency band boundaries is to obtain the typical radial velocity of different parts of the human body during movement and combine it with the Doppler frequency shift calculation principle for conversion. For example, if the low-frequency Doppler range corresponding to the radial velocity range of the human torso during normal walking is 5 Hz to 15 Hz, then 5 Hz to 15 Hz is set as the torso translation frequency boundary. If the high-frequency Doppler range corresponding to the highest superimposed radial velocity during arm swing is 15 Hz to 30 Hz, then 15 Hz to 30 Hz is set as the arm periodic swing frequency boundary. Therefore, the frequency band boundaries are set within the 5 Hz to 15 Hz range. The frequency bands between 15 Hz and 30 Hz are extracted and marked as trunk translation frequency parameters. The frequency bands between 15 Hz and 30 Hz are extracted and marked as arm periodic swing frequency parameters. The total numerical sequence of the trunk translation frequency parameters and arm periodic swing frequency parameters are extracted and marked. The total numerical sequence of the two types of parameters is divided by the maximum absolute value in their respective measurement intervals to perform a unified scale calculation. The parameter values ​​are all limited to the interval between -1 and 1. The processed data is truncated into data segments of the same length according to the set fixed number of sampling points to complete the parameter alignment operation. A covariance matrix is ​​constructed on the aligned high-dimensional frequency data sequence. The principal component eigenvectors with the first 3 eigenvalues ​​in the covariance matrix are extracted. The original frequency sequence is projected into the low-dimensional space formed by the first 3 principal component eigenvectors to complete the dimension reduction calibration operation of compressing and mapping the original frequency components, and the Doppler frequency shift feature group is obtained.

[0021] The steps for obtaining gait fusion metrics are as follows: Based on the Doppler frequency shift characteristic group, the frequency values ​​of the torso translation frequency parameter at continuous sampling positions are read and the radial movement velocity value is calculated item by item in combination with the carrier wavelength. The absolute value of the difference between the radial movement velocity values ​​corresponding to adjacent sampling positions within the observation segment is calculated and the average value is calculated. Then, the frequency values ​​of the arm period swing frequency parameter at continuous sampling positions are read and the human walking step frequency value is calculated according to the number of swing cycles per unit time. The absolute value of the difference between the human walking step frequency values ​​corresponding to adjacent sampling positions within the observation segment is calculated and the average value is calculated, forming a set of radial movement velocity value and human walking step frequency value fluctuation parameters. Based on the radial movement speed and human walking cadence fluctuation parameter set, the average radial movement speed, average human walking cadence, average absolute value of radial movement speed difference, and average absolute value of human walking cadence difference within the same observation segment are extracted. The gait fusion index is then calculated using the following formula: ; in, As an indicator of gait fusion, It is the average of all radial velocity values ​​within the same observation segment. This represents the average of all human walking cadence values ​​within the same observation segment. It is the average of the absolute values ​​of the radial movement velocity differences between adjacent sampling positions within the same observation segment. It is the average of the absolute values ​​of the differences in human walking gait frequency values ​​between adjacent sampling locations within the same observation segment.

[0022] Specifically, based on the aforementioned Doppler frequency shift feature set, the frequency values ​​of the torso translation frequency parameter at continuous sampling positions are read. The center transmission frequency of the Bluetooth wireless communication signal is retrieved; for example, if the center frequency is 2.4 GHz, the corresponding carrier wavelength is calculated to be approximately 0.125 meters using the constant of the speed of light. The extracted torso translation frequency parameter value is multiplied by the carrier wavelength and divided by 2, and the radial velocity value of the target body at this time is calculated item by item. For example, if the torso frequency at a certain sampling point is 24 Hz, the radial velocity value is calculated to be 1.5 meters per second. All sampling positions within the current observation segment are traversed, and the difference between the radial velocity values ​​at two adjacent sampling positions is calculated. The absolute values ​​of these differences are processed, and then all absolute values ​​of the differences are added together and divided by the total number of differences to calculate the average difference. At the same time, all radial velocity values ​​within the observation segment are added together and divided by the total number to obtain the average radial velocity. Then, the specific values ​​are read... The arm swing frequency parameter within the sampling group is calculated at each continuous sampling location. Since a complete arm swing cycle corresponds to two steps during normal walking, the arm swing frequency parameter is multiplied by 2 to convert it into the human walking step frequency value per unit time. For example, if the arm swing frequency at a sampling point is 0.9 Hz, the converted human walking step frequency value is 1.8 steps per second. All human walking step frequency values ​​within the current observation segment are iterated through, and the difference between the step frequency values ​​at two adjacent sampling locations is calculated and the absolute value is taken. The absolute values ​​of all step frequency differences are added together and divided by the total number of differences to calculate the average step frequency difference. The sum of all step frequency values ​​is then divided by the total number to calculate the average human walking step frequency. The calculated average radial movement speed, average absolute value of radial movement speed difference, average human walking step frequency, and average absolute value of step frequency difference are all packaged and stored in the same memory block to form the radial movement speed value and human walking step frequency value fluctuation parameter group.

[0023] In the gait motion fusion index calculation formula, the product of radial velocity and stride frequency is used to reflect the overall macroscopic motion energy characteristics of the target. At the same time, a comprehensive fluctuation penalty term composed of velocity variation coefficient and stride frequency variation coefficient is introduced. The negative exponential function of the natural constant base is used to map the comprehensive fluctuation. When the fluctuation of the target motion is greater, the penalty term is smaller. This weakens the index weight of abnormal human gait, such as the translation of wheeled robots or the waving of robotic arms, which are low-fluctuation or irregular high-fluctuation prosthetic fake signals, and enhances the distinguishability of real human motion characteristics. The steps for obtaining the values ​​are as follows: extract the raw average value of all radial movement velocity values ​​within the observation segment from the aforementioned parameter set, and then perform maximum-minimum normalization processing. The calculation formula is as follows: ,in For the extracted raw average velocity, and These are the preset lower and upper limits of the human walking speed range, respectively. The benchmark values ​​are set based on the slowest and fastest normal walking speeds statistically analyzed from real pedestrian daily tests. For example, setting a lower limit... The maximum speed is 0.5 meters per second. The original average value was extracted at 2.5 meters per second. The value is 1.5 meters per second. Substituting this into the normalization formula, we get... ; The steps to obtain the values ​​are as follows: extract the raw average of all human walking step frequencies within the same observation segment, and then perform maximum-minimum normalization to eliminate dimensions. The calculation formula is: ,in This is the original average step frequency. and These are the lower and upper limits of the cadence range, set based on the extreme values ​​at both ends of human cadence physiological test data. For example, setting the lower limit... 1.0 steps per second, maximum At 3.0 steps per second, the extracted raw average value The value is 1.8 steps per second. Substituting this into the calculation formula yields... ; The steps for obtaining the value are as follows: First, extract the raw average of the absolute values ​​of the radial movement velocity differences between adjacent sampling positions within the observation segment. To maintain the same dimensional system to support the accurate calculation of the subsequent dimensionless coefficient of variation, scaling and normalization are performed using the same velocity range. The calculation formula is as follows: ,in The average of the absolute values ​​of the extracted original velocity differences, for example, the extracted... It is 0.2 meters per second, here The numerical value reflects the magnitude of the change in velocity over time. Substituting it into the normalization formula yields... ; The acquisition steps are as follows: extract the original average value of the absolute value of the difference between human walking gait frequency values ​​at adjacent sampling locations within the same observation segment, and perform scaling and normalization processing using the same gait frequency range. The calculation formula is: ,in The average of the absolute values ​​of the extracted raw step frequency differences, for example, the extracted... The step rate is 0.1 steps per second. By using the same boundary values, the difference in fluctuations is also mapped to the dimensionless interval. Substituting this into the normalization formula yields... ; Calculations based on parameters: Substituting the aforementioned normalization parameters, the coefficient of variation of the radial movement velocity is calculated as follows: ; The coefficient of variation of human walking cadence is calculated as follows: ; Calculate the sum of squares of the coefficients of variation: ; Calculate the square root of the sum of squares: ; Calculate the overall volatility penalty term: ; Calculate the normalized macroscopic kinetic energy characteristics: ; Calculate the final gait motion fusion index: ; The results show that even after completely eliminating the dimensional differences, the extracted normalized moving target features still have high stability and frequency and speed matching characteristics that conform to the biological movement laws of the human body. This proves that the value of the index is within a reasonable range, and the signal is very likely to come from a real human body carrying a legitimate Bluetooth digital key approaching smoothly. This value will serve as the core comparative reference value for subsequent accurate determination of the legitimacy of the device's approach.

[0024] The steps for obtaining the unauthorized device proximity detection identifier are as follows: The lower boundary value and the outer boundary value of the preset anti-unauthorized movement interval parameters are retrieved. The gait motion fusion index is compared with the lower boundary value and the outer boundary value of the interval one by one. When the gait motion fusion index is between the lower boundary value and the outer boundary value of the interval, the state of not exceeding the outer boundary is recorded. When the gait motion fusion index is greater than the outer boundary value of the interval or less than the lower boundary value of the interval, the state of exceeding the outer boundary is recorded, and the unauthorized device approach judgment mark is obtained.

[0025] Specifically, the lower bound and outer boundary values ​​of the preset anti-unauthorized movement interval parameters are retrieved. The method for setting these values ​​is as follows: Doppler radar monitoring data of 50 adults of different body types approaching a vehicle while walking normally are collected at the test site. These sample data are then substituted into the aforementioned fusion index calculation model with normalization processing. The distribution of fusion indices for all legitimate real human subjects is statistically analyzed, and the dimensionless numerical range corresponding to the 95% confidence interval is identified. For example, if the statistical analysis shows that the lowest lower bound of the normalized fusion index for a 95% real human subject approaching the vehicle is 0.10 and the highest upper bound is 0.30, then the lower bound is set to 0.10, and the outer boundary is set to 0.30. The gait fusion index obtained from the aforementioned calculation is extracted and compared with the lower bound of 0.10 and the outer boundary of 0.30, respectively. A comparison of magnitudes is performed. When the gait motion fusion index value is within the closed interval of 0.10 to 0.30, for example, the previously calculated value of 0.158 is within this interval, the target's movement pattern is determined to conform to the biomechanical characteristics of a real human body, and the current state is recorded in the status field as "not exceeding the outer boundary". When the gait motion fusion index value is greater than 0.30, it indicates that the target's movement frequency or speed has increased abnormally and exceeds the range of normal human gait characteristics. Or when the gait motion fusion index value is less than 0.10, it indicates that the target has a stiff translational trajectory or extremely violent fluctuations. In both of these abnormal situations, it is determined to be recorded as "exceeding the outer boundary". Based on the previously written data of "not exceeding the outer boundary" or "exceeding the outer boundary", the status code is assigned and converted to obtain the unauthorized device approach determination mark used to characterize the legality of the current approach device.

[0026] The steps for obtaining the discrete value of the arrival interval of the response packet are as follows: If the unauthorized device approach determination flag shows a state of not exceeding the outer boundary, then multiple anchor nodes are called to receive the signal response packet corresponding to the packet arrival timestamp parameter. All packet arrival timestamp parameters are rearranged according to the time sequence of the packet arrival timestamp parameters. The time difference between two adjacent packet arrival timestamp parameters is calculated item by item and formed into a continuous record. Then, all time differences are written into a unified sequence structure according to the observation segment order to obtain the response packet arrival time interval sequence. Based on the response packet arrival time interval sequence, the deviation between the response packet arrival time interval and the average response packet arrival time interval is calculated for each item. All deviations are squared and summed. Then, the deviations are divided equally according to the number of response packet arrival time intervals and square root operations are performed. The resulting value is used as the standard deviation of the response packet arrival time interval sequence to obtain the response packet arrival time interval discreteness.

[0027] Specifically, if the aforementioned unauthorized device proximity determination flag shows a state where it has not exceeded the outer boundary, then after reading the flag determination value in the status field to confirm that it is within the safe range, the underlying timing feature analysis process is triggered. This process extracts the timestamp field attached to the end of the data frame of the physical layer data packet, strips the payload data by parsing the protocol frame header, locates the hardware counter snapshot value at the time of reception, and uniformly extracts and collects the timestamp parameters of all anchor nodes into a temporary one-dimensional array structure. The original node numbering classification mode is abandoned, and a quicksort algorithm is used to rearrange all packet arrival timestamp parameters in the array according to their numerical order from smallest to largest. For example, if 100 data packet timestamp snapshots are collected, the rearranged sequence forms an increasing time series from timestamp value 1000.1 milliseconds to 1020.5 milliseconds. The sorted timestamp array is then traversed, starting from the second timestamp, item by item... Extract the timestamp value of the current location and subtract the timestamp value of the previous adjacent location to calculate the time difference between the arrival timestamps of two adjacent packets. For example, extract the second timestamp of 1000.3 milliseconds and subtract the first timestamp of 1000.1 milliseconds to calculate the first time interval as 0.2 milliseconds. Then extract the third timestamp of 1000.8 milliseconds and subtract the second timestamp of 1000.3 milliseconds to calculate the second time interval as 0.5 milliseconds. In this process, even if there are large differences due to suspected channel packet loss, they are completely retained to truly reflect the link jitter. The calculated time differences of 0.2 milliseconds and 0.5 milliseconds are formed into a series of continuous records. According to the time flow of the original observation segment, all the record values ​​containing 99 time differences are sequentially appended to a unified floating-point sequence structure to obtain the response packet arrival time interval sequence.

[0028] Based on the previously generated sequence of response packet arrival time intervals, all time difference data stored in the sequence structure is read, and the total number of time difference data in the sequence is counted. For example, if a total of 99 time interval values ​​are found, these 99 time interval values ​​are summed in a floating-point arithmetic unit. For example, the sum of all interval values ​​is 1980.0 milliseconds. The total sum is divided by the total number of time difference data to calculate the average response packet arrival time interval. For example, dividing 1980.0 by 99 yields an average of 20.0 milliseconds. Then, starting from the beginning of the sequence, each value is re-traversed, and the absolute deviation between the specific response packet arrival time interval and the previously calculated average response packet arrival time interval is calculated. For example, if the first time interval is 20.4 milliseconds, subtracting the average of 20.0 milliseconds yields a first deviation of +0.4 milliseconds; if the second time interval is 19.7 milliseconds, subtracting the average of 20.0 milliseconds yields a second deviation of -0.3 milliseconds. To eliminate the masking effect of positive and negative deviations canceling each other out during subsequent summation, each calculated deviation is squared and raised to the power of the result. For example, squaring a positive 0.4 milliseconds yields 0.16 square milliseconds, and squaring a negative 0.3 milliseconds yields 0.09 square milliseconds. Then, a cumulative summation operation is performed on all the squared positive values. For example, adding all 99 squared values, including 0.16 and 0.09, yields a total sum of squares of 0.891 square milliseconds. The numerical values ​​are evenly distributed according to the total number of response packet arrival time interval data. That is, the total summation value is divided by the total number of 99. For example, dividing 0.891 by 99 yields a variance value of 0.009 square milliseconds. Finally, the underlying mathematical library function is called to perform a square root operation on the variance value. For example, taking the square root of 0.009 square milliseconds yields an approximate value of 0.0948 milliseconds. The obtained square root value is assigned to the standard deviation variable of the aforementioned response packet arrival time interval sequence to obtain the discrete value of the response packet arrival interval.

[0029] The steps for obtaining the deterministic timing jitter score of the link are as follows: Based on the discreteness of the response packet arrival interval, the drift rate parameter of the internal hardware crystal oscillator of the Bluetooth digital key is extracted, and the average value of the response packet arrival interval is read. The deterministic timing jitter of the link is calculated using the following formula: ; in, For deterministic timing jitter in links, This is the discreteness of the response packet arrival interval, i.e., the standard deviation of the sequence of response packet arrival time intervals. The average value of the sequence of time intervals between response packet arrivals. The drift rate parameter of the internal hardware crystal oscillator clock of the Bluetooth digital key. This indicates the proportion of time-scale correction that the hardware crystal oscillator clock drift rate produces on the average time interval.

[0030] Specifically, in the deterministic timing jitter score calculation formula, the crystal oscillator drift rate at the hardware level is introduced as a correction factor to transform the measured absolute time interval discrete quantity into a dimensionless relative timing jitter score. This eliminates the inherent error caused by the inaccuracy of the physical hardware clock of the communication equipment itself. At the same time, it ensures that the numerator standard deviation and the denominator average time interval are in the same dimension for ratio calculation, so that the jitter score calculated at the end can purely reflect the additional time jitter caused by the tampering of the man-in-the-middle hijacking device or asynchronous forwarding, thus improving the accuracy of the anti-relay attack judgment. The acquisition steps are as follows: First, read the response packet arrival interval discrete value from the memory block generated in the preceding steps. This value represents the standard deviation of the response packet arrival time interval sequence, reflecting the degree of fluctuation in data packet arrival times. It is obtained by retrieving the result calculated using statistical standard deviation methods. To ensure absolute consistency of magnitude when substituted into the formula, the raw tick count value of the hardware clock is intercepted during the reading of the underlying registers, multiplied by the microsecond duration corresponding to a single tick, and then divided by one thousand to convert it to standard millisecond units. For example, the standard deviation value obtained after performing the sum of squares and square root operations is read, and after unit verification and alignment, it is explicitly acquired. The specific value is 0.0948 milliseconds; The acquisition steps are as follows: synchronously read the average value of the response packet arrival time interval sequence from the aforementioned time difference statistics block. This value reflects the macroscopic average communication period of Bluetooth communication within the observation segment. The acquisition method is to retrieve the calculation result from the previous step, which sums all time intervals and divides by the total number. For example, read the average value data obtained in the previous step and confirm that it is measured in milliseconds. The specific value is 20.0 milliseconds, which is used as the denominator base to measure the relative jitter amplitude; The steps for obtaining this information are as follows: First, read the electronic tag model identifier of the Bluetooth digital key physical layer communication chip. Then, match the corresponding hardware crystal oscillator clock drift rate parameter in the factory hardware feature database. This parameter characterizes the inherent frequency offset ratio of the crystal oscillator due to changes in ambient temperature and manufacturing defects. Its value should be a dimensionless proportionality constant. For example, if the table shows that the crystal oscillator accuracy class of the currently used chip is ±30 parts per million, extract the maximum positive drift amount and convert it to standard decimal form. The specific conversion formula is to divide the read parts per million value by a one-million constant. For example, dividing 30 by 1,000,000 yields 0.00003. This process clearly obtains the desired result. The specific value is 0.00003; Calculations based on parameters: The corrected time interval denominator baseline value is calculated as follows: .

[0031] Deterministic timing jitter in link calculations is divided into: .

[0032] The results indicate that the timing jitter in the current communication link mainly originates from normal slight fluctuations in the radio channel and inherent hardware errors. It has not yet reached the level of large-scale asymmetric jitter on the millisecond scale introduced by malicious asynchronous relay forwarding. This proves that the current communication packet timing is consistent and highly deterministic, and is within the normal safety tolerance range. This jitter score will serve as a key indicator for subsequent comparison of the baseline value of the anti-relay hijacking hardware tolerance.

[0033] The steps to obtain the unauthorized man-in-the-middle hijacking status characteristics are as follows: Retrieve the anti-relay hijacking hardware tolerance baseline value, read the current value of the link deterministic timing jitter, compare the link deterministic timing jitter with the anti-relay hijacking hardware tolerance baseline value, write the comparison result where the link deterministic timing jitter is greater than the anti-relay hijacking hardware tolerance baseline value into the anomaly judgment field, and write the judgment conclusion that there is an asynchronous forwarding anomaly of the device into the status record area to obtain the device asynchronous forwarding anomaly judgment result; Based on the device asynchronous forwarding anomaly determination result, the determination conclusion of the device asynchronous forwarding anomaly in the status record area is extracted. The preset hijacking marking rules are called to write the anomaly source category, anomaly timing category, and anomaly forwarding category item by item. The determination conclusion of the device asynchronous forwarding anomaly is mapped to the unauthorized man-in-the-middle hijacking mark content. Then the unauthorized man-in-the-middle hijacking mark content is written into the security status field to generate the unauthorized man-in-the-middle hijacking status feature.

[0034] Specifically, the baseline value for anti-relay hijacking hardware tolerance is retrieved. The method for setting this baseline value is as follows: In an open area with a clean electromagnetic environment, a legitimate Bluetooth digital key configured at the factory is used to conduct one thousand normal short-range communication interactions with the vehicle test node. The deterministic timing jitter scores generated in each interaction are collected. These one thousand normal timing jitter scores are then distributed normally, and the mean and standard deviation are calculated. The mean plus three times the standard deviation is used as the upper limit of the normal communication jitter. For example, if the mean is 0.005 and the standard deviation is 0.001, adding three times 0.001 to 0.005 gives a baseline value of 0.008. This 0.008 is set as the baseline value for anti-relay hijacking hardware tolerance. The tolerance baseline value is obtained by reading the current value of the link deterministic timing jitter score calculated above. For example, in this test, the jitter score of the target device is read as 0.015. The jitter score of 0.015 is compared with the anti-relay hijacking hardware tolerance baseline value of 0.008. Since 0.015 is greater than the upper limit of the baseline of 0.008, it is determined that the current signal is subject to illegal delay interference. The comparison result that the timing jitter score is greater than the anti-relay hijacking hardware tolerance baseline value is encoded and written into the abnormal judgment field of the vehicle memory, triggering the abnormal alarm processing logic. The clear judgment conclusion that there is an abnormal situation of device asynchronous forwarding is written into the status record area of ​​the global variable in the form of a Boolean true value, thus obtaining the device asynchronous forwarding abnormal judgment result.

[0035] Based on the aforementioned generated device asynchronous forwarding anomaly determination result, the global variable memory block is accessed and the true value of the Boolean value stored in the status record area is extracted to confirm the determination conclusion of the device asynchronous forwarding anomaly. A preset hijacking marking rule pre-configured within the vehicle security gateway is invoked. This marking rule contains a hexadecimal fault code mapping table corresponding to different anomaly characteristics. Based on the numerical deviation magnitude generated by the aforementioned comparison process, the anomaly source category, anomaly timing category, and anomaly forwarding category are matched and written item by item. For example, if the jitter delay is determined to be a micro-timeline misalignment of more than 10 milliseconds, and the anomaly source category is matched in the rule table as an external hacker fake base station, the source fault code hexadecimal 0x is written. A1 matches the abnormal timing category as millisecond-level asynchronous delay, and writes the timing fault code hexadecimal 0xB2. Matches the abnormal forwarding category as bidirectional link asymmetric forwarding, and writes the forwarding fault code hexadecimal 0xC3. These three fault codes are concatenated in sequence to map the judgment of the existence of device asynchronous forwarding abnormality to the unauthorized man-in-the-middle hijacking mark content. For example, the complete hexadecimal hijacking mark content 0xA1B2C3 is obtained by concatenating it. This string of hexadecimal hijacking mark content is transmitted to the underlying protection controller through bus communication and is forcibly written into the security status field with read-only lock protection for solidification and storage, generating the unauthorized man-in-the-middle hijacking status feature.

[0036] The steps to obtain the vehicle anti-theft locking command are as follows: Based on the characteristics of unauthorized man-in-the-middle hijacking, the command sending interface of the vehicle engine control unit is invoked to write the control code content of the blocking command to the vehicle engine control unit, write the prohibition state of cutting off the vehicle start permission to the vehicle start permission register area, write the closure state of terminating the peripheral signal interaction to the peripheral signal interaction control area, and then associate and solidify the blocking command, prohibition state and closure state to establish the vehicle anti-theft stopping and locking command.

[0037] Specifically, based on the aforementioned unauthorized man-in-the-middle hijacking state characteristics, the complete hexadecimal hijacking flag content is read from the security status field. The corresponding target node is located via the vehicle's internal controller area network bus. The emergency safety command issuance interface reserved by the vehicle engine control unit is invoked to generate a string of control instructions with the highest system priority. Control codes for forced shutdown and ignition blocking commands are written to the execution memory area of ​​the vehicle engine control unit, for example, writing the highest priority hexadecimal blocking instruction 0xFF00. The vehicle start permission register is located, and the original start permission flag bit in the register is overwritten from a high level 1 in the normal state. Write a low level 0 to the register area to write a prohibited state that cuts off the vehicle start permission. Address and locate the peripheral signal interaction control area of ​​the antenna control module responsible for external communication. Pull all the power enable pins of Bluetooth communication and near-field communication RF antennas low to cut off the power supply. Write a closed state that terminates all peripheral signal interaction to the peripheral signal interaction control area. Logically bind the three execution results: the previously written hexadecimal blocking command 0xFF00, the register flag bit 0 indicating the prohibited state, and the antenna power-off closed state. Store the bound results in the anti-theft record area of ​​the non-volatile memory to complete the association and solidify operation, and establish the vehicle anti-theft stop and lock command.

[0038] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A Bluetooth digital key authentication method based on multi-anchor point fusion, characterized in that, Includes the following steps: The vehicle is equipped with multiple anchor nodes to receive the continuous wave signal and signal response packet transmitted by the Bluetooth digital key, generate a transmission timing phase state matrix, divide the frequency components inside the transmission timing phase state matrix into torso translation frequency parameters and arm periodic swing frequency parameters, and generate Doppler frequency shift feature groups. The trunk translation frequency parameters within the Doppler frequency shift feature group are converted into radial movement velocity values, the arm periodic swing frequency parameters within the Doppler frequency shift feature group are converted into human walking step frequency values, gait motion fusion indexes are calculated, preset anti-unauthorized movement interval parameters are retrieved, and unauthorized device approach judgment marks are generated by comparison. When the unauthorized device proximity determination flag shows a state of not exceeding the outer boundary, the arrival timestamp parameters of the corresponding packets of the received signal response packets from the aforementioned multiple anchor nodes are called to calculate the discrete amount of the arrival interval of the response packets. Based on the discrete amount of the arrival interval of the response packets, the link deterministic timing jitter score is calculated and generated. The anti-relay hijacking hardware tolerance baseline value is retrieved. When the link deterministic timing jitter is greater than the anti-relay hijacking hardware tolerance baseline value, an unauthorized man-in-the-middle hijacking state feature is generated. Based on the unauthorized man-in-the-middle hijacking state feature, a vehicle anti-theft stop and lock command is established.

2. The Bluetooth digital key authentication method based on multi-anchor point fusion according to claim 1, characterized in that, The steps for obtaining the Doppler frequency shift feature set are as follows: Obtain the signal response packets corresponding to the continuous wave signals received by multiple anchor nodes respectively, extract the packet arrival timestamp and instantaneous phase sequence value attached to each signal response packet, classify the packet arrival timestamp and instantaneous phase sequence value according to the node number of the anchor node, and then arrange the corresponding instantaneous phase sequence values ​​one by one according to the order of packet arrival timestamps to form a time-series spliced ​​phase sequence group. According to the time-series splicing phase sequence group, read the packet arrival timestamps arranged sequentially under each node number, read the instantaneous phase sequence value corresponding to each packet arrival timestamp, perform arrangement splicing and recombination according to the packet arrival timestamp as the vertical time sequence position and the instantaneous phase sequence value as the horizontal phase sampling position, write the arrangement results of the anchor nodes into the same structure in a unified order, and generate the transmission time-series phase state matrix. The frequency components that change continuously within the transmission timing phase state matrix are read line by line. The frequency segments corresponding to the continuous displacement of the human torso are screened out and marked as torso translation frequency parameters. The frequency segments corresponding to the reciprocating swing of the arm are screened out and marked as arm periodic swing frequency parameters. Then, compression mapping, parameter alignment and scale unification dimensionality reduction calibration operations are performed on the torso translation frequency parameters and the arm periodic swing frequency parameters to obtain the Doppler frequency shift feature group.

3. The Bluetooth digital key authentication method based on multi-anchor point fusion according to claim 1, characterized in that, The steps for obtaining the gait motion fusion index are as follows: Based on the Doppler frequency shift characteristic group, the frequency values ​​of the torso translation frequency parameter at continuous sampling positions are read and the radial movement speed value is calculated item by item in combination with the carrier wavelength. The absolute value of the difference between the radial movement speed values ​​corresponding to adjacent sampling positions within the observation segment is calculated and the average value is calculated. Then, the frequency values ​​of the arm period swing frequency parameter at continuous sampling positions are read and the human walking step frequency value is calculated according to the number of swing cycles per unit time. The absolute value of the difference between the human walking step frequency values ​​corresponding to adjacent sampling positions within the observation segment is calculated and the average value is calculated, forming a set of radial movement speed value and human walking step frequency value fluctuation parameters. Based on the radial movement speed value and human walking cadence value fluctuation parameter group, the average radial movement speed, the average human walking cadence, the average absolute value of the radial movement speed difference, and the average absolute value of the human walking cadence difference within the same observation segment are extracted to calculate the gait motion fusion index.

4. The Bluetooth digital key authentication method based on multi-anchor point fusion according to claim 1, characterized in that, The steps for obtaining the unauthorized device proximity determination identifier are as follows: The lower boundary value and the outer boundary value of the preset anti-unauthorized movement interval parameters are retrieved, and the gait motion fusion index is compared with the lower boundary value and the outer boundary value of the interval one by one. When the gait motion fusion index is between the lower boundary value and the outer boundary value of the interval, it is recorded as not exceeding the outer boundary. When the gait motion fusion index is greater than the outer boundary value of the interval or less than the lower boundary value of the interval, it is recorded as exceeding the outer boundary, thus obtaining the unauthorized device approach determination mark.

5. The Bluetooth digital key authentication method based on multi-anchor point fusion according to claim 1, characterized in that, The steps for obtaining the discrete value of the arrival interval of the response packet are as follows: If the unauthorized device proximity determination flag shows a state of not exceeding the outer boundary, then multiple anchor nodes are called to receive the signal response packet corresponding to the packet arrival timestamp parameter. All packet arrival timestamp parameters are rearranged according to the time sequence of the packet arrival timestamp parameters. The time difference between two adjacent packet arrival timestamp parameters is calculated item by item and formed into a continuous record. Then, all time differences are written into a unified sequence structure according to the observation segment order to obtain the response packet arrival time interval sequence. Based on the response packet arrival time interval sequence, the deviation between the response packet arrival time interval and the average response packet arrival time interval is calculated item by item. All deviations are squared and summed. Then, the deviations are divided equally according to the number of response packet arrival time intervals and square rooted. The resulting value is used as the standard deviation of the response packet arrival time interval sequence to obtain the response packet arrival time interval discreteness.

6. The Bluetooth digital key authentication method based on multi-anchor point fusion according to claim 1, characterized in that, The steps for obtaining the link deterministic timing jitter score are as follows: Based on the discrete amount of the response packet arrival interval, the drift rate parameter of the internal hardware crystal oscillator of the Bluetooth digital key is extracted and the average value of the response packet arrival time interval is read to calculate the link deterministic timing jitter score.

7. The Bluetooth digital key authentication method based on multi-anchor point fusion according to claim 1, characterized in that, The steps for obtaining the unauthorized man-in-the-middle hijacking status characteristics are as follows: Retrieve the anti-relay hijacking hardware tolerance baseline value, read the current value of the link deterministic timing jitter score, compare the link deterministic timing jitter score with the anti-relay hijacking hardware tolerance baseline value, write the comparison result that the link deterministic timing jitter score is greater than the anti-relay hijacking hardware tolerance baseline value into the anomaly judgment field, and write the judgment conclusion that there is an asynchronous forwarding anomaly of the device into the status record area to obtain the device asynchronous forwarding anomaly judgment result; Based on the device asynchronous forwarding anomaly determination result, the determination conclusion of the device asynchronous forwarding anomaly in the status record area is extracted. The preset hijacking marking rules are called to write the anomaly source category, anomaly timing category, and anomaly forwarding category item by item. The determination conclusion of the device asynchronous forwarding anomaly is mapped to the unauthorized man-in-the-middle hijacking mark content. Then the unauthorized man-in-the-middle hijacking mark content is written into the security status field to generate the unauthorized man-in-the-middle hijacking status feature.

8. The Bluetooth digital key authentication method based on multi-anchor point fusion according to claim 1, characterized in that, The steps for obtaining the vehicle anti-theft stop and lock command are as follows: Based on the unauthorized man-in-the-middle hijacking state characteristics, the command sending interface of the vehicle engine control unit is invoked to write the control code content of the blocking command to the vehicle engine control unit, the prohibition state of cutting off the vehicle start permission is written to the vehicle start permission register area, and the closure state of terminating the peripheral signal interaction is written to the peripheral signal interaction control area. Then, the blocking command, prohibition state and closure state are associated and solidified to establish the vehicle anti-theft stopping and locking command.

9. The system of the Bluetooth digital key authentication method based on multi-anchor point fusion according to any one of claims 1-8, characterized in that, include: The multi-anchor point signal acquisition and Doppler feature extraction module is used to set multiple anchor point nodes in the vehicle to receive the continuous wave signal and signal response packet transmitted by the Bluetooth digital key, generate the transmission time-phase state matrix, divide the frequency components inside the transmission time-phase state matrix into torso translation frequency parameters and arm periodic swing frequency parameters, and generate Doppler frequency shift feature groups. The gait motion fusion determination module is used to convert the trunk translation frequency parameters in the Doppler frequency shift feature group into radial movement speed values, convert the arm periodic swing frequency parameters in the Doppler frequency shift feature group into human walking step frequency values, calculate gait motion fusion index, retrieve preset anti-unauthorized movement interval parameters, and compare and generate unauthorized device approach determination identifiers. The link timing jitter detection module is used to call the arrival timestamp parameters of the corresponding packets of the received signal response packets of the aforementioned multiple anchor nodes when the unauthorized device approach determination flag shows a state of not exceeding the outer limit, calculate the discrete amount of the arrival interval of the response packets, and generate the link deterministic timing jitter score based on the discrete amount of the arrival interval of the response packets. The relay hijacking detection and vehicle locking module is used to retrieve the anti-relay hijacking hardware tolerance baseline value. When the link deterministic timing jitter is greater than the anti-relay hijacking hardware tolerance baseline value, an unauthorized man-in-the-middle hijacking state feature is generated. Based on the unauthorized man-in-the-middle hijacking state feature, a vehicle anti-theft stop and lock command is established.

Images