A method, device, medium and program product for contactless entry based on a Bluetooth key

Abstract

The application provides a Bluetooth key-based non-inductive entry method, device, medium and program product, and relates to the technical field of automobile electronics. The device first evaluates the current urgency of the user through a situational stress index, which integrates information in two dimensions of communication abnormality and schedule urgency, and can identify the probability of the user being in a state of anxiety. When the situational stress index exceeds a preset urgent situation threshold, the system automatically enables an effective outlier identification mechanism, which screens candidate outliers that jump upward, retains signal characteristics representing the user's rapid approach, and can more sensitively capture the user's rapid approach trend. Therefore, when the user is in an urgent situation, the technical solution can improve the unlocking response speed and recognition accuracy of the Bluetooth key.

Description

Technical Field

[0001] This application relates to the field of automotive electronics technology, and more particularly to a Bluetooth key-based contactless entry method, device, medium, and program product. Background Technology

[0002] In the modern automotive industry, digital key systems based on Bluetooth Low Energy (BLE) technology, also known as Bluetooth keys, are gradually becoming a mainstream feature that enhances user convenience. This technology allows users to automatically authenticate and unlock their vehicles simply by carrying a personal device such as a smartphone, eliminating the need for traditional physical buttons.

[0003] In related technologies, a series of Received Signal Strength Indication (RSSI) sample values ​​are continuously collected within a preset time window (e.g., several seconds), and smoothing algorithms such as moving average or Kalman filtering are applied to these sample values. This generates a relatively smooth signal strength change curve with a stronger trend. Based on this smoothed curve, when its value stably exceeds a preset unlocking threshold and maintains a certain growth slope, the system determines that the user has a clear intention to approach and authorizes the unlocking operation.

[0004] However, when a user approaches rapidly via a non-standard route in an urgent, running situation, irregular swaying can frequently obstruct the Bluetooth signal, causing RSSI values ​​to oscillate. Within the long sliding window designed to eliminate noise in related technologies, the strict outlier removal mechanism may make it difficult to form an effective recognition curve, reducing the accuracy of recognizing the user's rapid approach and resulting in Bluetooth key response delays. Summary of the Invention

[0005] This application provides a Bluetooth key-based contactless entry method, device, medium, and program product to improve the response speed and recognition accuracy of Bluetooth keys in emergency situations, thereby optimizing the user experience.

[0006] Firstly, this application provides a Bluetooth key-based contactless entry method, which includes: when it is determined that a target mobile device has entered a preset sensing distance range around a vehicle, a situation stress index is obtained by weighting and summing the ratio of the frequency of communication events to a preset normal communication frequency benchmark value based on time decay weight and the presence flag of an emergency category schedule entry in the schedule event metadata; a smooth curve is obtained by performing a moving average processing on a first sample sequence based on a preset smoothing window, wherein the first sample sequence is determined based on multiple received signal strength indication sample values ​​within the preset time window; when it is determined that the situation stress index exceeds a preset emergency situation threshold, the deviation of the first sample sequence from the smooth curve is... Sample values ​​exceeding a preset outlier threshold are marked as candidate outliers, and other sample values ​​are marked as non-outliers. The number of candidate outliers whose received signal strength index (RSI) values ​​are greater than the previous non-outlier is determined, thus obtaining the number of effective outliers. If the ratio of the number of effective outliers to the total number of samples in the first sample sequence is less than a preset outlier tolerance ratio, the effective outliers are combined with the first sample sequence to obtain a second sample sequence. After obtaining the regression slope of the second sample sequence using a linear regression algorithm, if the regression slope is greater than a preset minimum approach speed threshold and is positive, an unlocking command is generated, which controls the door to perform an unlocking operation.

[0007] By employing the above technical solution, the device first assesses the user's current urgency level through a situational stress index. This index integrates information from two dimensions: communication anomaly and schedule urgency, enabling it to identify the probability that the user is in a state of anxiety. When the situational stress index exceeds a preset emergency situation threshold, the system automatically activates an effective outlier identification mechanism. This mechanism filters out candidate outliers with upward signal transitions, preserving signal characteristics representing the user's rapid approach. In traditional solutions, these upward transitions are often considered noise and filtered out, resulting in slow unlocking response. This solution, however, recombines effective outliers with the original sample sequence to form a second sample sequence, and performs linear regression analysis based on this sequence, allowing the system to more accurately capture the trend of the user's rapid approach. Therefore, when the user is in an emergency situation, this technical solution improves the unlocking response speed and recognition accuracy of the Bluetooth key, avoiding the problem of losing proximity signals due to excessive signal smoothing.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the step of weighting and summing the ratio of the communication event frequency to a preset normal communication frequency benchmark value based on time decay weighting with the presence flag of the emergency category schedule entry in the schedule event metadata to obtain the situational stress index specifically includes: after extracting the emergency category schedule entry containing the preset emergency keyword from the schedule event metadata, dividing the preset time window into multiple time sub-intervals according to time order, the multiple time sub-intervals containing corresponding preset weight coefficients; weighting and summing the emergency category schedule entry based on the preset weight coefficients corresponding to the time sub-intervals to obtain a comprehensive schedule stress value; calculating the ratio of the current communication frequency of the target mobile device to the average communication frequency of the corresponding historical time period to obtain the communication anomaly degree; combining the comprehensive schedule stress value and the communication anomaly degree according to a preset ratio to obtain the situational stress index, the preset ratio being obtained based on emergency unlocking events and non-emergency unlocking events marked in the user's historical behavior data through a logistic regression algorithm.

[0009] By employing the aforementioned technical solutions, the time decay weighting mechanism ensures that urgent events closer to the time have a greater impact on the stress index, aligning with the time decay law of human psychological stress. By dividing the preset time window into multiple time sub-intervals and assigning different weight coefficients, the system can differentiate the impact of urgent events at different time intervals on the user's current state. Simultaneously, the calculation of communication anomaly effectively identifies abnormal changes in user behavior by comparing the current communication frequency with the historical average for the same period. This multi-dimensional, adaptive context-aware mechanism enables the system to more accurately identify the user's true emergency state, thereby activating special signal processing logic at appropriate times and improving the overall intelligence level of unlocking judgment.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, the step of determining the number of candidate outliers whose received signal strength indication sample value is greater than that of the previous non-outlier, and obtaining the number of effective outliers, specifically includes: marking sample values ​​in the first sample sequence whose deviation from the smooth curve exceeds a preset outlier threshold as candidate outliers and fitting a local linear trend to multiple preset sliding windows of the first sample sequence to determine the local trend direction of multiple windows; when the received signal strength indication value of the candidate outlier is greater than that of the previous non-outlier, and the change direction of the candidate outlier is consistent with the local trend direction of the window, it is marked as a positive effective outlier.

[0011] By adopting the above technical solution, the device not only identifies candidate outliers but also further improves the accuracy of effective outlier screening through local trend consistency verification. Local linear trend fitting of multiple preset sliding windows can capture the direction of signal change within a small range, providing contextual information for judging the validity of outliers. When the received signal strength indication value of a candidate outlier is greater than that of the previous non-outlier, it indicates that the point represents an upward jump in the signal; the consistency verification of its change direction with the local trend direction of its window ensures that the jump is a meaningful signal enhancement consistent with the overall approach trend, rather than random noise interference. This dual verification mechanism effectively filters out pseudo-outliers that, although large in amplitude, do not conform to the overall trend, while retaining signal characteristics representing rapid user approach behavior. This provides a data foundation for the subsequent construction of a second sample sequence and regression analysis, ultimately improving the reliability and response speed of the unlocking judgment.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, before the step of performing a moving average processing on the first sample sequence based on a preset smoothing window to obtain a smooth curve, and determining the first sample sequence based on multiple received signal strength indication sample values ​​within a preset time window, the method further includes: after receiving the triaxial accelerometer data sequence and triaxial angular velocity sensor data sequence of the target mobile device within the preset time window from the target mobile device, taking the X-axis as the movement direction, performing a weighted average of the standard deviations of the triaxial accelerometer data sequence on the Y-axis and Z-axis to obtain an effective sway amplitude index of the target mobile device perpendicular to the vehicle's forward direction; calculating the standard deviation of the angular velocity amplitude change of the instantaneous angular velocity values ​​of the triaxial angular velocity sensor on the three axes as the overall attitude change rate of the target mobile device, wherein the angular velocity amplitude is the vector magnitude of the triaxial angular velocity at each time point; generating a real-time signal distortion risk factor based on the effective sway amplitude index and the overall attitude change rate, wherein the real-time signal distortion risk factor is positively correlated with the effective sway amplitude index and positively correlated with the overall attitude change rate; and determining the window length of the preset smoothing window based on the situational stress index and the real-time signal distortion risk factor.

[0013] By adopting the above technical solution, the device acquires the user's motion state information before performing signal smoothing processing, achieving dynamic optimization of signal processing parameters. The effective sway amplitude index, by analyzing the standard deviations of Y-axis and Z-axis acceleration, quantifies the user's motion amplitude perpendicular to the forward direction, reflecting the degree of sway that may obstruct the Bluetooth signal. The overall attitude change rate, by calculating the standard deviations of the three-axis angular velocity amplitude changes, captures the instability of the device's attitude, reflecting the dynamic changes in the signal transmission path. The generation of a real-time signal distortion risk factor fuses these two motion parameters, providing a basis for subsequent signal processing strategies. Based on the situational stress index and this risk factor, the system can intelligently adjust the length of the preset smoothing window, shortening the window to improve response sensitivity when the user's movement is vigorous, and lengthening the window to enhance signal stability when the user's movement is stable. This motion-aware adaptive processing mechanism improves stability under different user behavior patterns.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, the step of determining the window length of the preset smoothing window based on the situational stress index and the real-time signal distortion risk factor specifically includes: when it is determined that the real-time signal distortion risk factor is higher than a preset collaborative threshold, the preset outlier threshold is increased proportionally based on a preset increment, and the window length is shortened proportionally based on a preset reduction.

[0015] By adopting the above technical solution, the device can dynamically adjust the sensitivity parameters of signal processing according to the level of real-time signal distortion risk. When the real-time signal distortion risk factor is higher than the preset collaborative threshold, it indicates that the user's current motion state or environmental conditions may affect the signal quality. At this time, the system adopts a more lenient processing strategy: by proportionally increasing the preset outlier threshold, the system's tolerance for signal fluctuations increases, avoiding misjudging normal signal changes caused by motion interference as invalid noise; at the same time, the window length is proportionally shortened, reducing the degree of smoothing processing, enabling the system to respond more quickly to real signal changes. This collaborative adjustment mechanism prioritizes response speed in high-risk situations, while maintaining sensitivity to valid signals by relaxing the outlier standard. Compared with traditional solutions using fixed parameters, this adaptive adjustment strategy can significantly reduce the misjudgment rate under complex motion and environmental conditions, ensuring that proximity signals are not over-filtered, thereby improving unlocking performance in various dynamic scenarios.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, the step of generating a real-time signal distortion risk factor based on the effective swing amplitude index and the overall attitude change rate, wherein the real-time signal distortion risk factor is positively correlated with the effective swing amplitude index and the overall attitude change rate, specifically includes: comparing the effective swing amplitude index with the historical average swing amplitude to obtain a first comparison value; comparing the overall attitude change rate with the historical average attitude to obtain a second comparison value; obtaining the current geographic location information of the target mobile device, and obtaining a corresponding preset geographic weight from a pre-stored geographic risk database based on the geographic location information, wherein the preset geographic weight is determined based on the historical signal interference or obstruction level of the geographic location area; and weighting and summing the first comparison value and the second comparison value based on the preset geographic weight to obtain the real-time signal distortion risk factor.

[0017] By adopting the above technical solution, the device achieves real-time signal distortion risk assessment based on multi-dimensional environmental perception. The first and second comparison values, by comparing current motion parameters with the user's historical averages, effectively identify the degree of deviation of the user's motion state from their normal state. This personalized benchmark comparison improves the accuracy of anomaly detection. The introduction of preset geographical weights incorporates environmental factors into the risk assessment system. These weights are determined based on the historical signal interference or obstruction levels of the geographical area, considering the differences in the impact of different geographical environments on Bluetooth signal propagation, enabling the risk factor to reflect the potential impact of the environment on signal quality. By weighting and summing the individual's motion anomaly degree with environmental risk, the generated real-time signal distortion risk factor comprehensively considers the dual impact of user behavior and environmental conditions, providing a more comprehensive and accurate decision-making basis for the adaptive adjustment of subsequent signal processing parameters, ultimately improving the system's robustness in complex scenarios.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, the step of combining the effective outlier with the first sample sequence to obtain the second sample sequence specifically includes: rearranging and combining all received signal strength indication sample values ​​in the first sample sequence that are not marked as candidate outliers with all the effective outliers according to the original sampling time order of the first sample sequence to form the second sample sequence.

[0019] By employing the above technical solution, the device ensures that the second sample sequence maintains temporal continuity while optimizing signal quality. This fusion process integrates all original sample values ​​not marked as candidate outliers (representing the stable portion of the signal) with double-verified valid outliers (representing meaningful signal enhancements). This selective retention mechanism avoids the loss of signal features and, by eliminating negative or invalid outliers, reduces noise interference with regression analysis, improving the accuracy and reliability of unlocking judgment.

[0020] In a second aspect, this application provides an apparatus comprising: one or more processors and a memory; the memory being coupled to the one or more processors, the memory being used to store computer program code including computer instructions, the one or more processors invoking the computer instructions to cause the apparatus to perform the method as described in the first aspect and any possible implementation thereof.

[0021] Thirdly, this application provides a computer program product containing instructions that, when run on a device, cause the device to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, this application provides a computer-readable storage medium including instructions that, when executed on a device, cause the device to perform the method described in the first aspect and any possible implementation thereof.

[0023] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0024] 1. By employing signal processing technology that integrates situational stress index and effective outlier identification, the device can intelligently filter and retain upward transition points representing rapid approach based on the user's current urgency level. This effectively solves the technical problem in existing technologies where excessive smoothing leads to misjudging approach signals as noise, resulting in response delays. Consequently, it achieves faster unlocking response and higher recognition accuracy in emergency situations.

[0025] 2. By adopting a technology that integrates user motion perception and adaptive adjustment of signal processing parameters, the device can assess the risk of signal distortion by analyzing the user's real-time motion state (such as swing amplitude and posture changes) and dynamically adjust the length of the smoothing window accordingly. This effectively solves the technical problem in the prior art where fixed parameters cannot adapt to the user's violent movement or complex environment, resulting in poor signal processing performance and improves the robustness of unlocking judgment.

[0026] 3. Due to the adoption of selective signal reconstruction technology based on time continuity, the device can rearrange and combine non-outliers representing stable parts of the signal with effective outliers representing meaningful signal enhancement in the original time order. This effectively solves the technical problem of information loss or data discontinuity caused by simply removing outliers in the existing technology. In this way, it can reduce noise interference while retaining the close signal, providing a data foundation for subsequent regression analysis and improving the accuracy of unlocking judgment. Attached Figure Description

[0027] Figure 1This is a flowchart illustrating a Bluetooth key-based contactless entry method in an embodiment of this application.

[0028] Figure 2 This is another flowchart illustrating a Bluetooth key-based contactless entry method in an embodiment of this application;

[0029] Figure 3 This is a schematic diagram of an exemplary hardware structure of the device in the embodiments of this application. Detailed Implementation

[0030] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0031] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0032] Please see Figure 1 This is a flowchart illustrating a Bluetooth key-based contactless entry method in an embodiment of this application.

[0033] S101. When it is determined that the target mobile device has entered the preset sensing distance range around the vehicle, the ratio of the communication event frequency to the preset normal communication frequency benchmark value is weighted and summed with the existence flag bit of the emergency category schedule entry in the schedule event metadata based on the time decay weight to obtain the situational stress index.

[0034] The target mobile device refers to a smart terminal held by the user that has been pre-paired with the vehicle via Bluetooth, such as a smartphone or smartwatch; the preset sensing distance range refers to the radial distance at which the vehicle's Bluetooth system can stably detect the signal of the target mobile device, for example, 5 to 15 meters; the communication event frequency refers to the number of communication activities that occur on the target mobile device per unit of time, such as making / receiving calls, sending and receiving messages, etc.; the schedule event metadata refers to the structured data obtained from the calendar or schedule application of the target mobile device, which includes information such as event title, time, and participants; the emergency category schedule item refers to a schedule item whose schedule content contains preset keywords such as "emergency", "meeting", "flight", "deadline", etc.; the existence flag is a binary variable, for example, "1" indicates that an emergency schedule exists, and "0" indicates that it does not exist.

[0035] When the device's Bluetooth module periodically scans and detects the signal of a paired target mobile device for the first time, and the signal strength reaches or exceeds the minimum threshold of the corresponding preset sensing distance range, the device requests authorization from the target mobile device via the Bluetooth channel and obtains two types of data: recent communication statistics and schedule event metadata for a specific future time window (such as the next 24 hours). The device first analyzes the schedule event metadata, searching for a preset set of emergency keywords in the schedule title or description using a keyword matching algorithm. Once a match is successful, the device sets the presence flag of the emergency category schedule entry to 1 and assigns it an initial stress value based on its proximity. Subsequently, the device processes the communication statistics, calculates the total number of communication events in the current time period (such as the past 10 minutes), and queries the user's historical behavior database stored locally or in the cloud to obtain a baseline value of the normal communication frequency at similar time points (such as 9 AM on a weekday). The ratio of these two values ​​is then used to obtain a communication anomaly score. Finally, the device uses a preset weighted summation formula to combine the stress value related to the urgency of the schedule with the ratio of communication anomalies. It may apply time decay weights to adjust the influence of recent schedules or communication behaviors, and finally calculates a comprehensive situational stress index.

[0036] It is understood that the acquisition of information such as "schedule event metadata" and "communication event frequency" involved in the embodiments of this application is premised on the user having granted the corresponding system permissions to the application (APP) carrying the method of this application on the target mobile device (such as a smartphone). In a typical implementation, when the user first enables the seamless entry function provided by this application or configures it in the relevant settings interface, the application will issue a clear pop-up prompt to the user through the standard permission management mechanism of the operating system, requesting authorization to access calendar / schedule data and read communication status / statistical information (such as call logs, SMS sending and receiving records, etc.). Only after obtaining the user's explicit consent and active authorization will the device perform subsequent steps such as acquiring data and calculating the situational stress index. When processing all relevant data, this application only accesses the metadata necessary to achieve the purpose of this invention (such as keywords in the schedule title, frequency of communication behavior and timestamps), and does not involve, store or upload any user's communication content, schedule details, contact information or other personal privacy data.

[0037] In some embodiments, the situational stress index can be calculated in several ways: Optionally, the device can extract emergency category schedule entries containing preset emergency keywords (such as "meeting" or "flight") from the target mobile device's schedule event metadata, and then divide a future preset time window (e.g., 24 hours) into multiple time sub-intervals in chronological order, such as 0-1 hour, 1-3 hours, 3-8 hours, etc. Each sub-interval corresponds to a preset weight coefficient that decreases with increasing time distance (e.g., 0.9, 0.6, 0.3). Then, the device performs a weighted summation of all identified emergency category schedule entries according to the preset weight coefficients corresponding to their respective time sub-intervals to obtain a comprehensive schedule stress value. Next, the device calculates the communication frequency of the target mobile device in the current time period (e.g., the past 15 minutes) and retrieves the user's average communication frequency in the same historical period (e.g., Wednesday afternoon) from the historical database, comparing the two to obtain the communication anomaly degree. It is understood that the weight coefficients of each interval can also be quantified based on the relative frequency or probability density of emergency unlocking events occurring within the interval. Finally, the device linearly combines the overall schedule stress value and communication anomaly degree according to a preset ratio to obtain the final situational stress index. This preset ratio is based on model parameters trained using a logistic regression algorithm, using labeled "emergency unlocking events" and "non-emergency unlocking events" from a large amount of user historical behavior data as training samples to ensure optimal weight allocation for each factor. Optionally, the device can also introduce data from other dimensions to calculate the situational stress index. For example, after obtaining explicit authorization from the user, a lightweight NLP analysis engine deployed locally on the target mobile device is activated. This engine performs localized and privacy-preserving analysis of text messages within a preset recent time window (e.g., the past 30 minutes). The analysis process can include two levels: First, through keyword matching, it searches a preset emergency / anxiety dictionary containing words such as "late," "almost there," "traffic jam," "too late," and "trouble," and calculates a basic text stress value based on the number of matched keywords and preset weights; second, through a pre-trained sentiment classification model, it judges the overall sentiment tendency of the message and outputs a quantitative sentiment score (e.g., from -1 representing extreme negative / anxiety to +1 representing positive). The device then merges the base text stress score with the sentiment score to generate a comprehensive "text context factor." Ultimately, this text context factor will be used as a new dimension, weighted and summed together with the aforementioned communication anomaly and comprehensive schedule stress score; its weights can also be calibrated using a machine learning model. Understandably, other methods can also be used to calculate the context stress index, such as using natural language processing technology to analyze the sentiment of recent text messages or instant messaging messages, or including negative or anxious text content as input for stress assessment; this is not limited to these methods.

[0038] S102. Perform a moving average process on the first sample sequence based on a preset smoothing window to obtain a smooth curve.

[0039] Once the device starts continuously receiving Bluetooth broadcast signals from the target mobile device, it records the Received Signal Strength Indication (RSSI) value at a fixed sampling frequency (e.g., once every 100 milliseconds) and stores these values ​​in a buffer in chronological order to form the first sample sequence.

[0040] Specifically, the process begins when the number of samples in the buffer reaches or exceeds the length of a preset smoothing window. The device takes the most recent N sample values ​​(N being the window length), calculates their arithmetic mean, and uses this mean as the latest point on the smoothing curve. When a new RSSI sample value is acquired, the device adds this new sample value to the end of the window, removes the oldest sample value from the front of the window, and then recalculates the arithmetic mean of the N values ​​within the window containing the new sample value to obtain the next point on the smoothing curve. This process continues as new samples arrive, much like a window of length N sliding forward on the first sample sequence, thus generating a dynamically updated smoothing curve.

[0041] In some embodiments, this smoothing process can be implemented in several ways: Optionally, the device can employ a weighted moving average method. In this method, each sample point within the smoothing window is no longer equally weighted. The device pre-sets a set of weight coefficients for the N positions within the window, typically assigning higher weights to sample points closer to the current time point. For example, for a window of length 5, the weights can be set to [0.1, 0.15, 0.2, 0.25, 0.3]. Each time the average is calculated, the device multiplies each sample value within the window by its corresponding weight, then sums these products to obtain the weighted average as a point on the smoothing curve. This approach makes the smoothing result more sensitive to the latest signal changes. Optionally, the device can also employ exponential smoothing. This method does not require maintaining a fixed-length window but instead uses a recursive formula for calculation. The device maintains a smoothed value from the previous time step, and when a new RSSI sample value is received, it calculates the smoothed value for the current time step using the formula: smoothed value_t = α * RSSI_t + (1-α) * smoothed value_{t-1}. Here, α is a smoothing coefficient between 0 and 1. The larger α is, the higher the weight of the new sample, and the faster the smoothing curve responds to changes. Understandably, other methods can also be used to achieve smoothing, such as using a Kalman filter to predict and update the signal state by establishing a state-space model of the signal; this is not limited here.

[0042] S103. When the situational stress index exceeds the preset emergency situation threshold, the sample values ​​in the first sample sequence that deviate from the smooth curve by more than the preset outlier threshold are marked as candidate outliers, and the other sample values ​​are marked as non-outliers.

[0043] The situational stress index calculated by the device in S101 exceeds the preset emergency situation threshold. Understandably, the system can infer that the user may be in an abnormal approach state, such as anxiety or running, and therefore needs to activate a more lenient and special signal processing logic to cope with possible drastic signal fluctuations. Under this premise, the device iterates through each raw RSSI sample value in the first sample sequence. For the i-th sample value RSSI_raw[i] in the sequence, the device finds the smoothed value RSSI_smooth[i] at the corresponding time point on the smoothed curve generated in S102. Then, the device calculates the absolute deviation between the two |RSSI_raw[i] - RSSI_smooth[i]|. If this deviation value is greater than the preset outlier threshold, the device marks the sample point RSSI_raw[i] as a "candidate outlier" in its internal data structure. If the deviation value is less than or equal to the threshold, it is marked as a "non-outlier". This process essentially classifies the first sample sequence, dividing all its points into two sets: "candidate outliers" and "non-outliers".

[0044] Understandably, the preset outlier threshold refers to the threshold for judging whether a sample point is an outlier. Its benchmark value is set from the statistical analysis of the noise distribution of RSSI signals in a large number of static and normal moving scenarios. It is usually taken as 2 to 3 times the noise standard deviation to ensure that it can cover most normal fluctuations.

[0045] In some embodiments, candidate outliers can be identified in several ways: Optionally, the device can calculate the standard deviation of sample points within a smoothing window while calculating the moving average. Then, a preset outlier threshold is defined as a multiple of this local standard deviation, for example, outlier threshold = k * σ_local, where k is a constant (e.g., 2 or 3). It is understood that in regions with stable signals and small local standard deviations, the threshold will adaptively decrease to detect more subtle anomalies; while in regions with drastic signal fluctuations and large local standard deviations, the threshold will adaptively increase to avoid misclassifying normal drastic fluctuations as outliers. Optionally, the device can also employ a percentile-based outlier detection method. The device first sorts the entire first sample sequence, calculates its first quartile (Q1) and third quartile (Q3), and obtains the interquartile range (IQR = Q3 - Q1). Then, the device defines a sample point as a candidate outlier if its value is less than Q1 - 1.5 * IQR or greater than Q3 + 1.5 * IQR. This method makes few assumptions about the shape of the data distribution and is more robust to non-Gaussian noise, especially in emergency situations where the signal distribution may deviate significantly from the normal distribution. Understandably, other methods can also be used to label candidate outliers, such as density-based clustering algorithms (like DBSCAN) to identify outliers in low-density regions; this is not a limitation here.

[0046] In some embodiments, when the device iterates through the i-th sample value RSSI_raw[i] in the first sample sequence, it first calculates the signed deviation D = RSSI_raw[i] - RSSI_smooth[i] between that point and the corresponding smooth curve value RSSI_smooth[i]. After determining that the absolute value of the deviation |D| exceeds a preset outlier threshold, the sign of the deviation D is checked. If D is positive, the device marks the sample point as a "positive candidate outlier" in its internal data structure to characterize a sudden increase in signal strength. If D is negative, it is marked as a "negative candidate outlier" to characterize a sudden decrease in signal strength. In this way, the original set of candidate outliers is subdivided into two subsets.

[0047] S104. Determine the number of candidate outliers whose received signal strength indication sample value is greater than that of the previous non-outlier, and obtain the number of effective outliers.

[0048] After marking all points in the first sample sequence in step S103, the device iterates through all sample points marked as "candidate outliers" in chronological order. For each candidate outlier, the device performs a backtracking search, starting from the position of the candidate outlier and searching backward along the time sequence until the first sample point marked as a "non-outlier" is found; this point is the "previous non-outlier." Then, the device compares the RSSI value of the current candidate outlier with the RSSI value of the found previous non-outlier. If the RSSI value of the candidate outlier is strictly greater than the RSSI value of the previous non-outlier, this indicates that the "abnormal" signal transition is an upward, strengthening transition, consistent with the user's behavior logic of quickly eliminating obstructions or shortening distance. In this case, the device determines this candidate outlier as a "valid outlier" and increments a dedicated counter, "number of valid outliers." If the RSSI value of the candidate outlier is less than or equal to the previous non-outlier, the outlier is ignored, and the counter is not incremented. The device repeats this process for all candidate outliers, and the final counter value is the total number of valid outliers.

[0049] In some embodiments, the determination of valid outliers can be achieved in several ways: Optionally, the device first marks sample values ​​in the first sample sequence whose deviation from the smooth curve exceeds a preset outlier threshold as candidate outliers. Simultaneously, the device fits a local linear trend to multiple preset sliding windows (e.g., window length 5) of the first sample sequence, determining the local trend direction of the multiple windows by calculating the linear regression slope of the data points within each window (positive slope indicates an increasing trend, negative slope indicates a decreasing trend). Only when the received signal strength indication value of a candidate outlier is greater than that of the previous non-outlier, and the direction of change of the candidate outlier itself (i.e., the difference between its value and the previous sample value) is consistent with the local trend direction of its window (e.g., both are positive), is the device marked as a "positive valid outlier" and counted. This dual verification criterion ensures that the accepted outliers are not only a single point of increase but also conform to the signal growth trend within a small surrounding area. Optionally, the device can also incorporate consideration of the signal change rate. When traversing candidate outliers, the device not only compares their values ​​with the previous non-outlier, but also calculates the RSSI difference between the candidate outlier and its immediately preceding time-point sample (regardless of whether it is an outlier), obtaining an instantaneous rate of change. Only when the RSSI value of a candidate outlier is greater than that of the previous non-outlier, and its instantaneous rate of change also exceeds a preset "minimum jump rate threshold," is it counted as a valid outlier. It is understandable that other methods can be used to determine valid outliers; for example, combining the frequency domain characteristics of candidate outliers, only those outliers whose energy is mainly concentrated in a specific frequency range can be retained, while filtering out pseudo-outliers caused by random white noise. This is not limited here.

[0050] In some embodiments, before traversing candidate outliers, the device first backtracks through the time series, initializing the most recent non-outlier as the "current valid reference point." When evaluating the first candidate outlier, its RSSI value is compared with the RSSI value of this "current valid reference point." If the candidate outlier's RSSI value is larger, it is determined to be a valid outlier, and the "current valid reference point" is immediately updated to this newly determined valid outlier. Subsequently, when evaluating the next consecutive candidate outlier, the device uses this newly updated, more recent "current valid reference point" for comparison. This process is repeated cyclically, ensuring that for a continuously ascending series of outliers, each point is compared with its immediately preceding valid point, thus forming a ladder-like chain of validity judgments.

[0051] S105. If the ratio of the number of valid outliers to the total number of samples in the first sample sequence is less than the preset outlier tolerance ratio, the valid outliers are combined with the first sample sequence to obtain the second sample sequence.

[0052] After calculating the total number of valid outliers in step S104, the device calculates the ratio of the number of valid outliers to the total number of samples in the first sample sequence. Then, the device compares this ratio to a preset outlier tolerance ratio. If the calculated ratio is less than the preset tolerance ratio, it means that although some outliers exist, their proportion in the entire sample sequence is still within an acceptable range, and the overall signal structure has not been completely destroyed. In this case, the device continues with the combination operation. The combination operation refers to the device creating a new sequence, the second sample sequence, which contains all the "non-outliers" from the first sample sequence, as well as all the sample points determined to be "valid outliers." Sample points marked as "candidate outliers" but not determined to be "valid" (usually a sudden drop in signal strength) are discarded. If the calculated ratio is greater than or equal to the tolerance ratio, it indicates that there are too many signal anomalies, the current data reliability is low, and the device may terminate the current unlocking judgment process and wait for the next set of more stable signal data.

[0053] In some embodiments, the generation of the second sample sequence can be implemented in several ways: Optionally, this step specifically involves the device collecting all received signal strength indication sample values ​​(i.e., all non-outliers) from the first sample sequence that are not marked as candidate outliers, along with all valid outliers determined in S104, into a temporary set. Then, the device sorts all points in this temporary set in ascending order according to the original sampling timestamp of each sample point in the first sample sequence. Through this sorting operation, all the retained points are rearranged back to their original temporal order, thereby forming a temporally continuous second sample sequence that has eliminated negative outliers. This sequence retains both the stationary portion of the signal and incorporates upward jumps representing rapid approach. Optionally, the device can use a "repair" rather than "rejection" approach. The device creates a second sample sequence of the same length as the first sample sequence. First, the values ​​of all non-outliers and valid outliers are directly copied to their corresponding positions in the second sample sequence. Then, for those outliers marked as candidate but not valid (i.e., rejected points), the device does not leave them blank but fills them with a substitute value. This substitute value can be the value of the smoothed curve at that point in time, or the value of its nearest non-outlier in time. In this way, the generated second sample sequence maintains the same data point density and temporal continuity as the original sequence, avoiding potential problems caused by missing data points in subsequent regression analysis. Understandably, other methods can also be used to combine the second sample sequences; for example, effective outliers can be weighted before combination, with their weights determined based on their deviation from the smoothed curve. Effective outliers with greater deviations may be assigned higher weights in subsequent regressions; this is not limited here.

[0054] S106. After obtaining the regression slope of the second sample sequence based on the linear regression algorithm, and if the regression slope is greater than the preset minimum approach speed threshold and is positive, an unlocking command is generated.

[0055] After successfully generating the second sample sequence, the device treats the data points in the second sample sequence as a set of points in a two-dimensional coordinate system, where the X-axis represents the sampling time and the Y-axis represents the RSSI value. The device applies a linear regression algorithm (such as the least squares method) to this set of points to calculate a straight line equation y=mx+c that best fits the distribution trend of these data points, where m is the regression slope. This slope m precisely quantifies the average rate of change of RSSI signal strength within the analyzed time window. Subsequently, the device performs a dual conditional judgment on this calculated regression slope m: First, it checks if m is positive. Only a positive slope indicates that the signal strength is increasing over time, meaning the target mobile device is approaching the vehicle. A negative slope or zero slope indicates moving away or the distance remains unchanged, and unlocking cannot be triggered. Second, based on determining that m is positive, it further checks whether m is greater than a preset minimum approach speed threshold. Understandably, the minimum approach speed threshold indicates that the user's approach intention is clear and sufficiently rapid, excluding ambiguous scenarios such as slow signal drift caused by environmental changes or the user lingering near the vehicle. The device will determine that the user intends to unlock the door quickly only when both of these conditions are met, and will immediately generate an unlocking command that conforms to the vehicle safety protocol. This command will be sent to the door locking module via the internal bus (such as the CAN bus) to complete the seamless entry.

[0056] In some embodiments, after applying a linear regression algorithm to the second sample sequence to obtain the regression slope m, the device also simultaneously calculates the goodness of fit (Coefficient of Determination, R² value) of this regression. The R² value ranges from 0 to 1; the closer it is to 1, the higher the explanatory power of the regression line on the data, and the stronger the linear relationship. The device expands the original unlock judgment condition into a triple AND gate: first, it determines that the regression slope m is positive; second, it determines that m is greater than a preset minimum approach speed threshold; finally, it determines that the calculated R² value is greater than a preset "minimum goodness of fit threshold" (e.g., 0.75). Only when these three conditions are met simultaneously can the device determine that the user has a stable and clear approach intention and generate an unlock command. If the slope is high but the R² value is low, it indicates that the data points are scattered and the linear trend is unreliable. The device will abandon this unlock judgment and wait to collect the next set of data.

[0057] It is understandable that if the situational stress index does not exceed the preset emergency situation threshold, the special outlier marking logic for the emergency situation in step S103 will not be triggered.

[0058] In some embodiments, if the situational stress index does not reach the emergency threshold, step S103 may be omitted. In this case, after step S102, a more conventional signal processing step can be performed directly (e.g., trend judgment based directly on the smoothed curve or more stringent outlier removal). Then, steps S104 and S105 can be skipped, and the regression analysis in step S106 can be performed directly. However, in this case, the input sequence for the regression analysis will be a conventionally processed sequence, rather than a second sample sequence containing valid outliers. No limitation is imposed here.

[0059] Understandably, by preserving these effective outliers that jump upwards, this scheme enables the reconstructed signal sequence (i.e., the second sample sequence) to exhibit a steeper regression slope in linear regression analysis. This means that, at the same user approach speed, the signal strength growth trend calculated by this scheme is more significant, and it can reach and exceed the preset minimum approach speed threshold earlier.

[0060] In this embodiment, a signal processing mechanism that integrates situational stress index and effective outlier identification is adopted. Therefore, after sensing the user's emergency state, the signal sequence can be intelligently retained and used to reconstruct the signal sequence by the upward jump point (effective outlier) representing the rapid approach. This effectively solves the problem of response delay caused by the over-smoothing of the approach signal being misjudged as noise in the prior art, thereby achieving faster unlocking response and higher recognition accuracy in emergency situations.

[0061] In the above embodiment, the device can improve unlocking response speed by capturing signal transitions representing approach intentions when a user is in an emergency situation, through a situational stress index and effective outlier identification mechanism. However, in practical applications, when the above method is implemented, the distortion of the Bluetooth signal increases when the user approaches the vehicle in a complex electromagnetic environment, resulting in limitations in signal processing effectiveness. This method can improve the accuracy of quickly identifying the user's approach intentions under complex motion and environmental conditions by dynamically adjusting the signal processing parameters and adjusting the sensitivity of the signal processing.

[0062] Please see Figure 2 This is another flowchart illustrating a Bluetooth key-based contactless entry method in an embodiment of this application.

[0063] S201. When it is determined that the target mobile device has entered the preset sensing distance range around the vehicle, the situational stress index is obtained by weighting and summing the ratio of the frequency of communication events to the preset benchmark value of normal communication frequency based on the time decay weight and the existence flag of the emergency category schedule entry in the schedule event metadata.

[0064] Step S201 and Figure 1Step S101 in the illustrated embodiment is similar and will not be repeated here.

[0065] S202. After receiving the triaxial acceleration sensor data sequence and triaxial angular velocity sensor data sequence of the target mobile device within a preset time window, the effective swing amplitude index of the target mobile device in the direction perpendicular to the vehicle's forward movement is obtained by weighted averaging of the standard deviations of the triaxial acceleration sensor data sequence on the Y and Z axes with the X-axis as the direction of movement.

[0066] Among them, the triaxial accelerometer data sequence represents the set of acceleration measurements of the target mobile device over time in the three orthogonal directions of X, Y, and Z; the triaxial angular velocity sensor data sequence represents the set of rotational velocity measurements of the target mobile device over time in the three orthogonal axes of X, Y, and Z.

[0067] After determining that the target mobile device has entered a preset sensing distance range, the device continuously receives data sequences collected by the target mobile device's built-in triaxial accelerometer and triaxial angular velocity sensor within a preset time window (e.g., the most recent 5 seconds) via Bluetooth Low Energy (BLE) or other wireless communication methods. Upon receiving the triaxial accelerometer data sequence, the device first aligns this data with the vehicle's coordinate system or performs relative motion analysis to determine which axis represents the vehicle's forward direction (X-axis). Then, the device focuses on the acceleration data along the Y and Z axes perpendicular to the vehicle's forward direction. For the Y-axis acceleration data sequence, the device calculates its standard deviation to quantify the device's sway in the left-right direction. Similarly, for the Z-axis acceleration data sequence, the device calculates its standard deviation to quantify the device's sway in the up-down direction. Finally, the device weights these two standard deviations according to preset weights (e.g., the relative importance of Y-axis and Z-axis sway to the signal can be determined based on experience or training data) to obtain an index of the effective sway amplitude of the target mobile device perpendicular to the vehicle's forward direction. The higher this indicator, the more violently the equipment shakes when it approaches a vehicle.

[0068] In some embodiments, the effective swing amplitude index can be calculated in several ways: Optionally, the device can first preprocess the received triaxial accelerometer data sequence, for example, by applying a low-pass filter to remove high-frequency noise and performing gravity component compensation to eliminate the influence of Earth's gravity on acceleration measurements during static or slow motion. Then, the device calculates the variance of the Y-axis and Z-axis acceleration data sequences within a preset time window, where the variance is the square of the standard deviation, which more directly reflects the dispersion of the data. Next, the device weights and sums the Y-axis and Z-axis variances according to preset weights (e.g., Y-axis weight 0.6, Z-axis weight 0.4, because lateral sway may have a greater impact on signal obstruction), and finally takes the square root of the weighted sum to obtain the effective swing amplitude index. Optionally, the device can also use frequency domain analysis. The device performs a Fast Fourier Transform (FFT) on the Y-axis and Z-axis acceleration data sequences to obtain their spectra. Then, the device identifies and extracts energy components related to the frequency range of human walking or running (e.g., 0.5Hz to 3Hz) from the spectra. The device performs a weighted summation of the energy within these specific frequency ranges to determine the effective swing amplitude. It is understood that other methods can also be used to calculate the effective swing amplitude, such as using wavelet transform to analyze swing characteristics at different scales; this is not a limitation here.

[0069] S203. Calculate the standard deviation of the angular velocity amplitude change of the instantaneous angular velocity values ​​of the three-axis angular velocity sensor as the overall attitude change rate of the target mobile device.

[0070] The instantaneous angular velocity value refers to the rotational speed of the target mobile device around its X, Y, and Z axes at a certain moment; the angular velocity amplitude refers to the magnitude of the three-axis angular velocity vector at each time point, indicating the overall speed of the device's rotation.

[0071] After receiving the triaxial angular velocity sensor data sequence from the target mobile device, the device iterates through each time point in the sequence. For each time point, the device acquires the instantaneous angular velocity values ​​(ωx, ωy, ωz) on the X, Y, and Z axes. Then, the device calculates the vector magnitude of these three instantaneous angular velocity values, i.e., angular velocity amplitude = sqrt(ωx^2 + ωy^2 + ωz^2). This angular velocity amplitude represents the overall rotational speed of the device at that instant. The device constructs a new sequence from all the angular velocity amplitudes calculated within a preset time window. Finally, the device calculates the standard deviation of this angular velocity amplitude sequence. This standard deviation is the overall attitude change rate of the target mobile device. Understandably, a high overall attitude change rate indicates that the target mobile device's attitude (such as tilting or rotating a phone) is changing frequently and drastically as it approaches a vehicle. This is typically associated with the unstable state of a user holding the device while running, searching for items, or performing other complex actions.

[0072] In some embodiments, the overall attitude change rate can be calculated in several ways: Optionally, the device can first preprocess the received triaxial angular velocity sensor data sequence, for example, by applying a medium-range filter to smooth the data and remove spike noise. Then, the device calculates the angular velocity amplitude at each time point. Next, instead of directly calculating the standard deviation of these amplitudes, the device calculates the absolute value of the difference between the angular velocity amplitudes of two adjacent time points, obtaining a sequence of "instantaneous changes". Finally, the device calculates the average or median of this sequence of instantaneous changes as the overall attitude change rate. This method focuses more on the frequency and amplitude of attitude changes rather than the absolute rotational speed. Optionally, the device can also combine attitude estimation algorithms. The device can use accelerometer and gyroscope data (through algorithms such as Kalman filtering or complementary filtering) to estimate the attitude (e.g., Euler angles or quaternions) of the target mobile device in real time. Then, the device calculates the rate of change or standard deviation of these attitude parameters (e.g., pitch, roll, yaw) over time within a preset time window. For example, the device can calculate the root mean square (RMS) values ​​of the rates of change of pitch and roll angles as the overall attitude change rate. This method directly quantifies the device's directional changes in space, providing more intuitive information on attitude stability. It is understandable that other methods can be used to calculate the overall attitude change rate, such as analyzing the autocorrelation function of angular velocity data to assess its periodicity and randomness; this is not limited to these methods here.

[0073] S204. Based on the effective swing amplitude index and the overall attitude change rate, generate a real-time signal distortion risk factor.

[0074] After the device has calculated the effective sway amplitude index (from S202) and overall attitude change rate (from S203) of the target mobile device perpendicular to the vehicle's direction of travel, it uses this information to assess the potential distortion risk of the current Bluetooth signal. First, the device compares the currently calculated effective sway amplitude index with the pre-stored historical average sway value of the user, obtaining a first comparison value, which reflects the degree of abnormality of the current sway relative to the user's normal state. Next, the device compares the currently calculated overall attitude change rate with the pre-stored historical average attitude value of the user, obtaining a second comparison value, which reflects the degree of abnormality of the current attitude change relative to the user's normal state. Simultaneously, the device acquires the current geographic location information of the target mobile device (e.g., via GPS or Wi-Fi positioning) and, based on this geographic location information, retrieves the corresponding preset geographic weights from a pre-stored geographic risk database. This geographic risk database stores historical signal interference or obstruction data for different geographic areas; for example, areas with tall buildings or strong electromagnetic interference have a higher risk of signal distortion, and correspondingly higher geographic weights. Finally, the device weights and sums the first and second comparison values ​​based on the acquired preset geographic weights to obtain the final real-time signal distortion risk factor. The higher this factor, the greater the likelihood that the current signal is distorted due to motion and environmental factors.

[0075] In some embodiments, the generation of real-time signal distortion risk factors can be achieved in several ways: Optionally, the device can first compare the effective sway amplitude index with the historical sway average to obtain a first comparison value, for example, by calculating the ratio or difference between the current index and the historical average. Similarly, the device compares the overall attitude change rate with the historical attitude average to obtain a second comparison value. Next, the device obtains the current geographic location information of the target mobile device and obtains the corresponding preset geographic weight from a pre-stored geographic risk database based on the geographic location information. The preset geographic weight is determined based on the historical signal interference or obstruction level of the geographic location area. Finally, the device performs a weighted summation of the first comparison value and the second comparison value according to a preset linear combination formula, combined with the preset geographic weight, to obtain the real-time signal distortion risk factor. Optionally, the device can also use a machine learning model to generate the real-time signal distortion risk factor. The device uses the effective sway amplitude index, the overall attitude change rate, and geographic location information (or its encoded features, such as region ID, building density, etc.) as input features and inputs them into a pre-trained regression model (e.g., support vector regression, random forest regression, or neural network). This model directly outputs a quantified real-time signal distortion risk factor by learning the complex nonlinear relationships between a large amount of historical data (including user motion data, geographic location, and actual signal quality degradation). This method can more comprehensively consider the interactions between multiple factors and provide a more accurate risk assessment. Understandably, other methods can also be used to generate the real-time signal distortion risk factor, such as combining environmental noise sensor data or weather information to further refine the risk factor; this is not limited here.

[0076] S205. Based on the situational stress index and the real-time signal distortion risk factor, determine the window length of the preset smoothing window.

[0077] After the device has calculated the situational stress index (from S201) and the real-time signal distortion risk factor (from S204), it uses these two key indicators to dynamically adjust the parameters of subsequent signal smoothing processing, especially the window length of the preset smoothing window. The device first determines whether the real-time signal distortion risk factor is higher than a preset collaborative threshold. If the risk factor is higher than this threshold, it indicates that the current signal is likely to be affected by motion and environmental interference. To avoid over-smoothing leading to signal feature loss, the device takes the following measures: First, it proportionally increases the preset outlier threshold used for outlier detection in subsequent steps based on a preset increment. This means the device will have a higher tolerance for signal fluctuations, allowing larger deviations to be considered outliers. Second, it proportionally shortens the window length of the preset smoothing window based on a preset decrement. Shortening the window length aims to reduce the degree of smoothing, allowing the smoothing curve to respond more quickly to real changes in the signal, thus enabling more timely detection of rising signal strength trends when a user approaches rapidly. If the real-time signal distortion risk factor is below the collaborative threshold, the device may maintain the default window length and outlier threshold, or make fine adjustments based on the situational stress index. For example, when the stress index is high but the risk factor is low, the window may still be shortened appropriately to improve the response speed.

[0078] In some embodiments, the window length of the preset smoothing window can be determined in several ways: Optionally, the device can first, if it is determined that the real-time signal distortion risk factor is higher than a preset cooperative threshold, proportionally increase the preset outlier threshold based on a preset increment (e.g., a percentage or a fixed value), for example, new outlier threshold = original outlier threshold * (1 + increment ratio). Simultaneously, the window length is proportionally shortened based on a preset decrement (e.g., a percentage or a fixed value), for example, new window length = original window length * (1 - decrement ratio). If the real-time signal distortion risk factor is lower than the cooperative threshold, the device adjusts the window length according to the situational stress index: when the situational stress index is high, the window length is moderately shortened; when the situational stress index is low, the window length remains at the default value or is moderately extended. This hierarchical adjustment strategy ensures the rationality of the smoothing parameters under different situations. Optionally, the device can also use a multidimensional lookup table or a rule-based expert system to determine the window length. The input to the lookup table is the situational stress index and the quantified level of the real-time signal distortion risk factor (e.g., low, medium, high), and the output is the corresponding window length and outlier threshold. For example, when both the "situational stress index" and the "real-time signal distortion risk factor" are high, the lookup table might return a very short window length and a large outlier threshold; when both are low, it would return a longer window length and a smaller outlier threshold. This approach allows the device to make flexible and non-linear parameter adjustments based on empirical knowledge or preset rules. It is understood that other methods can also be used to determine the window length of the preset smoothing window, such as using a fuzzy logic controller, taking the situational stress index and the real-time signal distortion risk factor as fuzzy inputs, and obtaining the optimal window length and outlier threshold through fuzzy inference; this is not limited here.

[0079] It is understandable that step S205 achieves dynamic adaptive adjustment of signal smoothing parameters by integrating the situational stress index and the real-time signal distortion risk factor. In some embodiments, if the situational stress index does not exceed the preset emergency situation threshold, the operations of expanding the preset outlier threshold and shortening the window length in step S205 may not be performed. Instead, default parameters or parameters determined based on other factors may be used. In this way, step S206 can be executed directly after step S204 is completed. This is not limited here.

[0080] S206. Perform a moving average process on the first sample sequence based on a preset smoothing window to obtain a smooth curve.

[0081] Steps S206~S210 and Figure 1 Steps S102 to S106 in the illustrated embodiment are similar and will not be repeated here.

[0082] S207. When the situational stress index exceeds the preset emergency situation threshold, the sample values ​​in the first sample sequence that deviate from the smooth curve by more than the preset outlier threshold are marked as candidate outliers, and the other sample values ​​are marked as non-outliers.

[0083] S208. Determine the number of candidate outliers whose received signal strength indication sample value is greater than that of the previous non-outlier, and obtain the number of effective outliers.

[0084] S209. If the ratio of the number of valid outliers to the total number of samples in the first sample sequence is less than the preset outlier tolerance ratio, the valid outliers are combined with the first sample sequence to obtain the second sample sequence.

[0085] S210. After obtaining the regression slope of the second sample sequence based on the linear regression algorithm, and if the regression slope is greater than the preset minimum approach speed threshold and is positive, an unlocking command is generated.

[0086] In this embodiment, by employing triaxial accelerometer and triaxial angular velocity sensor data based on the target mobile device, combined with geographic location information, a real-time signal distortion risk factor is generated. Therefore, the device can assess in real-time the degree of interference that the user's motion state and environment may cause to the Bluetooth signal. Furthermore, by dynamically adjusting the window length and outlier threshold of the preset smoothing window based on the situational stress index and the real-time signal distortion risk factor, the device can adaptively adjust the sensitivity of signal smoothing and outlier identification according to the user's current urgency and signal distortion risk. This effectively solves the problem in existing technologies where fixed parameters lead to sluggish response in emergency situations or misjudgments when signal distortion is high. Thus, a balance between response speed and recognition accuracy of the Bluetooth key is achieved in different situations, improving the user experience.

[0087] The exemplary device 300 provided in the embodiments of this application is described below. Figure 3 This is an exemplary hardware structure diagram of the device 300 provided in the embodiments of this application.

[0088] In some embodiments, the device 300 is a computer device or includes a computer device. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements the methods in the embodiments of this application.

[0089] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0090] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0091] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0092] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0093] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for contactless entry based on a Bluetooth key, characterized in that, The method includes: When the Bluetooth key enters the vehicle within a preset sensing distance, the situational stress index is obtained by weighting and summing the ratio of the frequency of communication events to the preset normal communication frequency benchmark value based on the time decay weight and the presence flag of the emergency category schedule entry in the schedule event metadata. The first sample sequence is processed by a moving average based on a preset smoothing window to obtain a smooth curve. The first sample sequence is determined based on multiple received signal strength indication sample values ​​within a preset time window. If the situational stress index is determined to exceed a preset emergency situation threshold, the sample values ​​in the first sample sequence that deviate from the smooth curve by more than a preset outlier threshold are marked as candidate outliers, and the other sample values ​​besides the candidate outliers are marked as non-outliers. The number of candidate outliers whose received signal strength indication sample value is greater than that of the previous non-outlier is determined, and the number of effective outliers is obtained. If the ratio of the number of effective outliers to the total number of samples in the first sample sequence is less than a preset outlier tolerance ratio, the effective outliers are combined with the first sample sequence to obtain a second sample sequence. After obtaining the regression slope of the second sample sequence based on the linear regression algorithm, if the regression slope is determined to be greater than the preset minimum approach speed threshold and is positive, an unlocking command is generated, and the unlocking command controls the car door to perform an unlocking operation.

2. The method according to claim 1, characterized in that, The step of weighting and summing the ratio of the communication event frequency to a preset benchmark value of normal communication frequency based on time decay weighting, and the presence flag of the emergency category schedule entry in the schedule event metadata to obtain the situational stress index specifically includes: After extracting the emergency category schedule entries containing preset emergency keywords from the schedule event metadata, the preset time window is divided into multiple time sub-intervals according to time order, and the multiple time sub-intervals contain corresponding preset weight coefficients; The comprehensive schedule pressure value is obtained by weighting and summing the emergency category schedule items based on the preset weight coefficients corresponding to the time sub-intervals; The communication anomaly degree is obtained by calculating the ratio of the current communication frequency of the Bluetooth key to the average communication frequency of the corresponding historical period. The overall schedule stress value and the communication anomaly degree are combined according to a preset ratio to obtain the situational stress index. The preset ratio is obtained by using a logistic regression algorithm based on emergency unlocking events and non-emergency unlocking events marked in the user's historical behavior data.

3. The method according to claim 1, characterized in that, The step of determining the number of candidate outliers whose received signal strength indication sample value is greater than that of the previous non-outlier, and obtaining the number of effective outliers, specifically includes: The sample values ​​in the first sample sequence whose deviation from the smooth curve exceeds a preset outlier threshold are marked as candidate outliers, and the local linear trend is fitted to multiple preset sliding windows of the first sample sequence to determine the local trend direction of multiple windows. When the received signal strength index sample value of a candidate outlier is greater than that of the previous non-outlier, and the direction of change of the candidate outlier is consistent with the local trend direction of the window, the corresponding candidate outlier is marked as a positive valid outlier.

4. The method according to claim 1, characterized in that, Before the step of performing a moving average process on the first sample sequence based on a preset smoothing window to obtain a smooth curve, and determining the first sample sequence based on multiple received signal strength indication sample values ​​within a preset time window, the method further includes: After receiving the triaxial accelerometer data sequence and triaxial angular velocity sensor data sequence of the Bluetooth key within the preset time window, the standard deviation of the triaxial accelerometer data sequence on the Y and Z axes is weighted and averaged with the X-axis as the direction of movement to obtain the effective swing amplitude index of the Bluetooth key in the direction perpendicular to the vehicle's forward movement. The standard deviation of the instantaneous angular velocity amplitude change of the triaxial angular velocity sensor on the three axes is calculated as the overall attitude change rate of the Bluetooth key, where the angular velocity amplitude is the vector magnitude of the triaxial angular velocity at each time point; Based on the effective swing amplitude index and the overall attitude change rate, a real-time signal distortion risk factor is generated. The real-time signal distortion risk factor is positively correlated with the effective swing amplitude index and the overall attitude change rate. The window length of the preset smoothing window is determined based on the situational stress index and the real-time signal distortion risk factor.

5. The method according to claim 4, characterized in that, The step of determining the window length of the preset smoothing window based on the situational stress index and the real-time signal distortion risk factor specifically includes: If the real-time signal distortion risk factor is determined to be higher than the preset collaborative threshold, the preset outlier threshold is increased proportionally based on the preset increment, and the window length is shortened proportionally based on the preset reduction.

6. The method according to claim 4, characterized in that, The step of generating a real-time signal distortion risk factor based on the effective swing amplitude index and the overall attitude change rate, wherein the real-time signal distortion risk factor is positively correlated with the effective swing amplitude index and the overall attitude change rate, specifically includes: The effective swing amplitude index is compared with the historical average swing value to obtain a first comparison value; The overall attitude change rate is compared with the historical attitude average to obtain a second comparison value; The current geographic location information of the Bluetooth key is obtained, and a corresponding preset geographic weight is obtained from a pre-stored geographic risk database based on the geographic location information. The preset geographic weight is determined based on the historical signal interference or obstruction level of the geographic location area. The first comparison value and the second comparison value are weighted and summed based on the preset geographical weight to obtain the real-time signal distortion risk factor.

7. The method according to claim 1, characterized in that, The step of combining the effective outliers with the first sample sequence to obtain the second sample sequence specifically includes: The received signal strength indication sample values ​​of all unmarked candidate outliers in the first sample sequence are rearranged and combined with all the valid outliers according to the original sampling time order of the first sample sequence to form the second sample sequence.

8. A device, characterized in that, The device includes: a memory and one or more processors; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code including computer instructions, and the one or more processors invoking the computer instructions to cause the device to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on the device, the device causes the device to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the device, the device causes the device to perform the method as described in any one of claims 1-7.

Images