Digital twin loading method, vehicle-mounted device and computer-readable storage medium

Abstract

The application relates to a digital twin loading method, a vehicle-mounted device and a computer readable storage medium. The method comprises the following steps: acquiring a multi-source pre-judgment signal under the condition that a preset condition is met; determining a pre-judgment comprehensive score corresponding to a pre-judgment user based on the multi-source pre-judgment signal, and preloading a first digital twin of the pre-judgment user under the condition that the pre-judgment comprehensive score is greater than or equal to a first preset threshold; determining whether an actual user identity of a current user in the vehicle matches a pre-judgment user identity of the pre-judgment user; and setting the vehicle by using the first digital twin under the condition that the actual user identity matches the pre-judgment user identity. In this way, the digital twin loaded by the vehicle can be ensured to be the digital twin corresponding to the actual user identity, the loading response time of the digital twin is reduced, and the accuracy and efficiency of the digital twin loading are simultaneously considered.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a digital clone loading method, an in-vehicle device, and a computer-readable storage medium. Background Technology

[0002] A digital avatar is a high-fidelity virtual representation of a specific vehicle, driver, or passenger in the physical world within the digital space. It constructs a dynamic digital model that reflects the real-time state, attributes, and preferences of the physical entity by fusing data from multiple sensors. When a driver or passenger enters the vehicle, the corresponding digital avatar can be loaded to automatically restore their personalized settings, such as seat position, air conditioning temperature, music playlist, and navigation preferences, thus automatically achieving personalized settings for different users.

[0003] However, in related technologies, it is usually necessary for the driver or passenger to complete user identification after entering the vehicle before the digital clone can be loaded. Although this can ensure the accuracy of the digital clone loading, there is a large loading delay, and even situations where the function freezes or cannot be used, which seriously affects the user experience. Therefore, how to simultaneously ensure the accuracy and efficiency of digital clone loading has become an urgent technical problem to be solved. Summary of the Invention

[0004] This application provides a digital clone loading method, an in-vehicle device, and a computer-readable storage medium to solve the problem that the prior art cannot simultaneously achieve both accuracy and efficiency in digital clone loading.

[0005] In a first aspect, embodiments of this application provide a method for loading a digital clone, the method comprising: Under the condition of meeting the preset conditions, multi-source prediction signals are acquired, wherein the preset conditions are the trigger conditions for predicting the user of the vehicle, and the multi-source prediction signals refer to the prediction signals acquired from multiple signal sources. Based on the multi-source prediction signal, the prediction comprehensive score corresponding to the prediction user is determined, and if the prediction comprehensive score is greater than or equal to a first preset threshold, the first digital clone of the prediction user is preloaded, wherein the prediction comprehensive score is used to characterize the probability that the prediction user is the actual user who took the bus. Determine whether the actual user identity of the current user in the vehicle matches the predicted user identity; If the actual user identity matches the predicted user identity, the vehicle is configured using the first digital clone.

[0006] Optionally, determining the comprehensive prediction score corresponding to the predicted user based on the multi-source prediction signals includes: Based on the multi-source prediction signals, the normalized value, fusion weight, and basic weight of each key prediction signal in the multi-channel key prediction signals are determined. The multi-source prediction signals include the multi-channel key prediction signals. The normalized value refers to the value obtained after normalizing the original value of each key prediction signal. The fusion weight is used to characterize the importance among the multi-channel key prediction signals. The basic weight is used to characterize the credibility of each key prediction signal in this prediction. Based on the multi-source prediction signal, the user impact factor and environmental impact factor corresponding to this prediction are determined. The user impact factor is used to characterize the probability that each candidate user was predicted in the past. The candidate user refers to all users who have participated in the past prediction. The environmental impact factor is used to characterize the degree of interference of the current environment on this prediction. The comprehensive score of each candidate user is calculated based on the user impact factor, the environmental impact factor, the normalized value, the fusion weight, and the basic weight corresponding to each key prediction signal. The candidate user with the highest overall score is determined as the predicted user, and the overall score of the predicted user is determined as the predicted overall score.

[0007] Optionally, the formula for calculating the overall score of each candidate user is as follows: ; in, Indicates the first The overall score of each candidate user Indicates the first User influence factor of each candidate user This refers to the environmental impact factors. Indicates the first The basic weights corresponding to the key road prediction signals, Indicates the first The fusion weights corresponding to the key prediction signals of the road are Indicates the first The normalized value corresponding to the key road prediction signal.

[0008] Optionally, determining whether the actual user identity of the current user in the vehicle matches the predicted user identity includes: Obtain the first feature set of the current in-vehicle user and the second feature set of the predicted user; Calculate the similarity between the first feature set and the second feature set; If the similarity is greater than or equal to a second preset threshold, it is determined that the actual user identity matches the predicted user identity. If the similarity is less than the second preset threshold, it is determined that the actual user identity does not match the predicted user identity.

[0009] Optionally, after determining whether the actual user identity of the current user in the vehicle matches the predicted user identity, the method further includes: If it is determined that the actual user identity does not match the predicted user identity, the resources occupied by the first digital clone are released, and the second digital clone corresponding to the actual user identity is loaded, so as to use the second digital clone to set up the vehicle.

[0010] Optionally, before preloading the first digital clone of the predicted user, the method further includes: Obtain the valid time information corresponding to the first digital clone in the local preset cache interval, wherein the local preset cache interval caches multiple digital clones and multiple valid time information corresponding to the multiple digital clones, the multiple digital clones and the multiple valid time information correspond one-to-one, and the multiple digital clones include at least the first digital clone; Based on the valid time information corresponding to the first digital clone, determine whether the first digital clone is within its validity period; If it is determined that the first digital clone is within the validity period, the first digital clone is retrieved from the local preset cache range, and the step of preloading the first digital clone corresponding to the predicted user identity is executed.

[0011] Optionally, after determining whether the first digital clone is within its validity period based on the valid time information corresponding to the first digital clone, the method further includes: If it is determined that the first digital clone is not within its validity period, the first digital clone is deleted from the local preset cache range, and a request instruction is sent to the cloud server, wherein the request instruction is used to request the cloud server to regenerate and issue a new first digital clone.

[0012] Optionally, the method further includes: The digital clones cached in the local preset cache range are checked at preset intervals to determine whether each digital clone cached in the local preset cache range is within its validity period. If it is determined that the third digital clone is not within its validity period, the third digital clone is deleted from the local preset cache range, and a request instruction is sent to the cloud server. The request instruction is used to request the cloud server to regenerate and issue a new third digital clone, which is any digital clone within the local preset cache range.

[0013] Secondly, embodiments of this application also provide an in-vehicle device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor, when executing a program stored in memory, implements the digital clone loading method described in the first aspect.

[0014] Thirdly, embodiments of this application also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the digital clone loading method described in the first aspect.

[0015] Compared with the prior art, the technical solution provided in this application has the following advantages: The method provided in this application, under the condition of satisfying preset conditions, acquires multi-source prediction signals, wherein the preset conditions are triggering conditions for predicting the predicted user of the vehicle, and the multi-source prediction signals refer to prediction signals acquired from multiple signal sources; based on the multi-source prediction signals, determines the prediction comprehensive score corresponding to the predicted user, and when the prediction comprehensive score is greater than or equal to a first preset threshold, preloads the first digital clone of the predicted user, wherein the prediction comprehensive score is used to characterize the probability that the predicted user is an actual user in the vehicle; determines whether the actual user identity of the current user in the vehicle matches the prediction user identity of the predicted user; when it is determined that the actual user identity matches the prediction user identity, uses the first digital clone to set up the vehicle. The above method allows for the pre-loading of a first digital clone corresponding to the predicted user identity. After the actual user boards the vehicle, the actual user identity is then matched with the predicted user identity. If the actual user identity matches the predicted user identity, the first digital clone can be used directly to configure the vehicle. This ensures that the digital clone ultimately loaded into the vehicle is the one corresponding to the actual user identity while reducing the loading response time of the digital clone, thus achieving both the accuracy and efficiency of digital clone loading. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0019] Figure 1 A flowchart illustrating a digital clone loading method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a dual-track verification system provided in an embodiment of this application; Figure 3 This is a schematic diagram of a dual-track verification process provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a digital clone loading device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of a vehicle-mounted device provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] The following disclosure provides numerous different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of the invention. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0022] To address the problem that existing technologies cannot simultaneously achieve both accuracy and efficiency in digital clone loading, this application provides a digital clone loading method, an in-vehicle device, and a computer-readable storage medium that can simultaneously achieve both accuracy and efficiency in digital clone loading.

[0023] See Figure 1 , Figure 1 This is a flowchart illustrating a digital clone loading method provided in an embodiment of this application. Figure 1 As shown, the digital clone loading method may include the following steps: Step S102: Under the condition of meeting the preset conditions, acquire multi-source prediction signals, wherein the preset conditions are the trigger conditions for predicting the user of the vehicle, and the multi-source prediction signals refer to the prediction signals acquired from multiple signal sources.

[0024] It should be noted that the digital clone loading method provided in this application embodiment can be applied to any in-vehicle device with sensing, computing and interaction capabilities, such as intelligent cockpit domain controller, in-vehicle infotainment system, digital instrument panel, in-vehicle telematics box (T-Box), in-vehicle driver monitoring system (DMS), etc.

[0025] Specifically, the aforementioned preset conditions refer to the triggering conditions for predicting the user of the vehicle. These conditions can be the detection of a user within a preset distance from the vehicle, and / or the current time reaching the prediction activation time. In other words, when a user is detected within the preset distance from the vehicle, or when the prediction activation time is reached, or when both conditions are met, the prediction of the user will be triggered. At this time, multi-source prediction signals can be acquired to predict the user of the vehicle. The preset distance range can be set according to actual needs, such as 5 meters, 3 meters, etc. The prediction activation time can be learned based on the user's historical vehicle usage data. For example, assuming the user frequently uses the vehicle at 8:00 AM on weekdays, the prediction activation time could be 7:55 AM, etc.

[0026] The aforementioned multi-source prediction signals refer to prediction signals obtained from multiple information sources, which may include, but are not limited to, user-related prediction signals (such as user behavior), environment-related prediction signals (such as illuminance, weather, and pedestrian flow), and vehicle-related prediction signals (such as location signals, device binding signals, reservation order signals, and account association signals).

[0027] Step S104: Based on the multi-source prediction signal, determine the prediction comprehensive score corresponding to the prediction user, and if the prediction comprehensive score is greater than or equal to the first preset threshold, preload the first digital clone of the prediction user, wherein the prediction comprehensive score is used to characterize the probability that the prediction user is the actual user who took the bus.

[0028] Specifically, the first preset threshold can be set according to actual needs, and no specific limitation is made here. The predicted user can be understood as the user most likely to get on the bus based on the multi-source prediction signals.

[0029] The in-vehicle device can determine user impact factors, environmental impact factors, and normalized values, fusion weights, and basic weights corresponding to each key prediction signal based on the acquired multi-source prediction signals. Then, based on this information, it determines the predicted user and their comprehensive prediction score. Next, it compares this comprehensive prediction score with a first preset threshold. If the comprehensive prediction score is greater than or equal to the first preset threshold, it indicates a high probability that the predicted user is actually a user in the vehicle. In this case, the first digital clone of the predicted user can be preloaded, preloading the necessary resources to the in-vehicle device. If the comprehensive prediction score is less than the first preset threshold, it indicates a low probability that the predicted user is actually a user in the vehicle. In this case, the first digital clone of the predicted user is not preloaded.

[0030] It's important to note that this refers to preloading the first digital clone, not formal loading. Preloading simply involves caching the necessary data, configurations, and resources for the first digital clone in the background. This data, configurations, and resources can be destroyed at any time without affecting the main foreground task. Even if the prediction is incorrect, the background can simply release these resources. In other words, formal loading only starts loading resources when needed, which may result in longer user wait times. Preloading, on the other hand, loads resources in advance, and when needed, the pre-prepared resources are instantly switched to the foreground, thus improving response speed and user experience.

[0031] Step S106: Determine whether the actual user identity of the current user in the vehicle matches the predicted user identity.

[0032] Specifically, the in-vehicle device can obtain the actual user identity of the current user in the vehicle (such as facial features, biometric features, driving habit features, etc.) and the predicted user identity, and then calculate the similarity between the two to determine whether the actual user identity of the current user in the vehicle matches the predicted user identity. Of course, the predicted user identity can also be displayed on the vehicle's human-machine interface for the user to select and confirm, in order to determine whether the actual user identity matches the predicted user identity. This application embodiment does not make specific limitations.

[0033] Step S108: If the actual user identity matches the predicted user identity, the vehicle is configured using the first digital clone.

[0034] Specifically, when the actual user identity matches the predicted user identity, the in-vehicle device can use the pre-loaded first digital clone to configure the vehicle without having to reload the first digital clone.

[0035] In this way, the first digital clone corresponding to the predicted user identity can be preloaded, and after the actual user gets on the vehicle, the actual user identity is matched with the predicted user identity. If the actual user identity matches the predicted user identity, the first digital clone can be used directly to set up the vehicle. Therefore, while ensuring that the digital clone loaded by the vehicle is the digital clone corresponding to the actual user identity, the loading response time of the digital clone is reduced, achieving the goal of simultaneously ensuring the accuracy and efficiency of digital clone loading.

[0036] In an optional embodiment, step S104 above, determining the comprehensive prediction score corresponding to the predicted user based on multi-source prediction signals, includes: Based on multi-source prediction signals, the normalized value, fusion weight, and basic weight of each key prediction signal in the multi-channel key prediction signals are determined. The multi-source prediction signals include multiple key prediction signals. The normalized value refers to the value obtained after normalizing the original value of each key prediction signal. The fusion weight is used to characterize the importance among the multiple key prediction signals. The basic weight is used to characterize the credibility of each key prediction signal in this prediction. Based on multi-source prediction signals, the user impact factor and environmental impact factor corresponding to this prediction are determined. The user impact factor is used to characterize the probability that each candidate user was predicted in the past. Candidate users refer to all users who have participated in the past prediction. The environmental impact factor is used to characterize the degree of interference of the current environment on this prediction. The comprehensive score of each candidate user is calculated based on the user impact factor, environmental impact factor, and the normalized value, fusion weight, and basic weight of each key prediction signal. The candidate user with the highest overall score is identified as the predicted user, and the overall score of the predicted user is determined as the predicted overall score.

[0037] Specifically, the aforementioned user influence factor characterizes the probability of each candidate user being correctly predicted in the past. This user influence factor is determined based on the ratio of each candidate user's historical successful predictions to the total number of historical predictions. This user influence factor can be used to quantify the reliability of each candidate user. Specifically, the user influence factor can be calculated using the following formula: ; in, Indicates the first The number of successful predictions for each candidate user in history. Indicates the first The total number of historical predictions for each candidate user Indicates the first The user influence factor of each candidate user. For candidate users with higher user influence factors, the probability of them being predicted as the predicted user increases; for candidate users with lower user influence factors, the probability of them being predicted as the predicted user decreases.

[0038] The aforementioned environmental impact factor characterizes the degree of interference from the current environment on the current prediction. This environmental impact factor can be dynamically adjusted based on the current environment (such as weather, light intensity, size of obstructions, etc.). It can be used to compensate for deviations in prediction results caused by adverse environmental conditions such as nighttime, rainy days, heavy fog, dense crowds, and obstruction interference. The better the current environment, the closer the environmental impact factor is to 1; the worse the current environment, the smaller the environmental impact factor, thereby lowering the prediction score and preventing misjudgments.

[0039] The normalized values ​​corresponding to the aforementioned key prediction signals refer to the values ​​obtained after normalizing the original values ​​of each key prediction signal. The range of the normalized values ​​for each key prediction signal is [0,1]. The purpose of normalization is to unify the dimensions and eliminate unit differences, so that all key prediction signals can be added and calculated fairly and reasonably in the same mathematical formula. This is a prerequisite for multi-source data fusion operations.

[0040] The fusion weights corresponding to the aforementioned key prediction signals are used to characterize the importance of multiple key prediction signals. More important key prediction signals have higher fusion weights; less important key prediction signals have lower fusion weights. The fusion weights corresponding to each key prediction signal can be directly preset as fixed constants, without dynamic iteration.

[0041] The basic weights corresponding to the above key prediction signals are used to characterize the credibility of each key prediction signal in this prediction.

[0042] The basic weights corresponding to each key prediction signal can be determined by the ratio of the historical hit count of each key prediction signal to the total hit count of all key prediction signals. The sum of the basic weights corresponding to each key prediction signal is 1 to ensure the rationality and comparability of the weighted fusion of key prediction signals. Subsequently, the basic weights are dynamically adjusted based on each prediction result using gain and attenuation step sizes. Specifically, the basic weights corresponding to each key prediction signal can be calculated using the following formula: ; in, Indicates the first The basic weights corresponding to the key prediction signals of the road are Indicates the first The historical hit count of key prediction signals. This represents the total number of hits for all key predictive signals. The value range is 1 to , This represents the total number of critical prediction signals. It should be noted that the historical hit count here refers to the number of times the current state of a particular critical prediction signal is the same as its historical state. For example, assuming the vehicle is currently connected to Bluetooth device A, information on all Bluetooth devices the vehicle has previously connected to can be obtained. If Bluetooth device A has been connected 50 times in the past, then its historical hit count can be determined to be 50.

[0043] For example, suppose the key prediction signals include the distance between the user and the vehicle, the Bluetooth signal strength, the reservation order timeliness, and the historical vehicle frequency. And suppose the historical hit counts of these four key prediction signals are as follows: =40, =30, =15, =35, total hits are 120, then =40 / 120≈0.333, =30 / 120≈0.25, =15 / 120≈0.125, =35 / 120≈0.292. After initially calculating the basic weights corresponding to the key prediction signals, subsequent iterations can be performed based on each matching result: when the predicted user identity matches the actual user identity, then... ′= +ΔW1; If the predicted user identity does not match the actual user identity, then ′= ΔW2, where ΔW1 represents the gain step size and ΔW2 represents the decay step size, for example, ΔW1=0.02, ΔW2=0.03, etc. It is important to note that during the iteration process, the constraint condition must be satisfied: 0 < <1, and each The sum of them equals 1.

[0044] After obtaining the user impact factor, environmental impact factor, and the normalized values, fusion weights, and basic weights corresponding to each key prediction signal, the comprehensive score of each candidate user can be determined based on these factors. The candidate user with the highest comprehensive score is then identified as the predicted user, and the comprehensive score of this predicted user is set as the predicted comprehensive score. This method allows for efficient and accurate determination of the predicted user and its corresponding predicted comprehensive score based on multi-source prediction signals. This facilitates the rapid preloading of the first digital clone when the predicted comprehensive score is high, and the reversion to real-time user identification mode when the predicted comprehensive score is low.

[0045] In one optional embodiment, the formula for calculating the overall score of each candidate user is as follows: ; in, Indicates the first The overall score of each candidate user Indicates the first User influence factor of each candidate user Indicates environmental impact factors, Indicates the first The basic weights corresponding to the key prediction signals of the road are Indicates the first The fusion weights corresponding to the key prediction signals of the road are Indicates the first The normalized value corresponding to the key road prediction signal.

[0046] It should be noted that here This refers to the first The values ​​within the range [0,1] obtained after normalizing the original values ​​of the keyway prediction signal. Specifically, it can be derived from the values ​​of the first... Information such as distance, signal strength, or time difference is obtained from the keyway prediction signal. Then, the parameters are normalized. The processing procedure is as follows: ; in, Indicates the first The normalized value corresponding to the key road prediction signal. Indicates the first The current value of the key prediction signal. Indicates the first The minimum range of the critical prediction signal. Indicates the first The maximum range of the key prediction signal.

[0047] To facilitate understanding, the following explanation will use the distance signal between the user and the vehicle, the Bluetooth signal strength signal, the reservation order validity signal, and the historical vehicle frequency signal as examples.

[0048] =1 indicates the distance signal between the user and the vehicle, with a minimum range of 0m and a maximum range of 10m. The vehicle will trigger sensing when it is within 10 meters. =2 indicates the Bluetooth signal strength, with a minimum range of -90dBm and a maximum range of -30dBm. The larger the value, the stronger the signal. =3 indicates the timeliness signal of the reservation order, with a minimum range of 24 hours and a maximum range of 1 hour. =4 represents the historical vehicle frequency signal, with a minimum range of 0 times / week and a maximum range of 7 times / week.

[0049] In calculation As the user moves, the distance between the user and the vehicle changes in real time, such as 15m→12m→10.2m→9.8m→5m→2m. Each round of sampling refreshes the data. Synchronous real-time recalculation.

[0050] Phase 1: When the distance between the user and the vehicle is greater than 10m (10.1m~15m, not within the perception range), the distance signal is not included in the prediction calculation because it exceeds the distance sensing threshold Td, and a value is directly assigned. =0.

[0051] Phase 2: When the distance between the user and the vehicle is 10m, =1, at which point the distance clue reaches the highest credibility and officially enters the effective perception range. The system begins to integrate information such as Bluetooth signal strength signal, appointment order timeliness signal, and historical vehicle frequency signal to calculate the comprehensive score S.

[0052] Phase 3: The user continues to approach the vehicle, and the distance keeps decreasing. Get smaller Synchronous reduction. 10m is the boundary for predictive activation. When a user just enters the perception range, the weight of distance clues is increased first, and predictive calculation is started as early as possible. When the user continues to get close to the vehicle, the distance is already a certain fact, and it is no longer necessary to increase the score by distance. The identity determination is dominated by high-identity signals such as Bluetooth signal strength, reservation order timeliness, and historical vehicle frequency.

[0053] In calculation At this time, users can be distinguished by the device ID paired with Bluetooth; signal strength is only used for calculation. The signal strength indicates the credibility of the clue. The two work together to make an identity prediction. You cannot identify a person solely by the signal strength.

[0054] The vehicle pre-stores Bluetooth device information for all authorized users: phone identifiers and bound user identity identifiers are all archived. For example, vehicle owner A: Bluetooth device ID=MAC-A, bound identity ID=UID001; family member B: Bluetooth device ID=MAC-B, bound identity ID=UID002, etc. Since passersby do not have bound devices, if Bluetooth is detected but the device ID is not in the authorized list, it is directly determined as an invalid user. Whenever a Bluetooth broadcast packet is detected, the device ID is first extracted and matched against the authorization database to lock the user. Then, the distance of the user from the vehicle and the signal reliability are assessed. The closer the Bluetooth device and the less obstructed it is, the stronger the Bluetooth signal (e.g., -35dBm), and the higher the reliability of matching the currently bound user. Conversely, the farther the Bluetooth device is and the stronger it is (e.g., -80dBm) due to a person or metal obstruction from the vehicle body, the weaker the Bluetooth signal (e.g., -80dBm), and the lower the reliability.

[0055] In calculation In this context, the reservation validity period refers to the remaining time between the current system time and the user's reserved pick-up time. It is a time-based signal from the in-vehicle app / cloud to the vehicle. Users can initiate car reservations in advance via the mobile app (commuting, driving, pick-up / drop-off, long-distance travel). Each reservation is linked to a unique user identifier and the reserved pick-up time. The closer the remaining time is to 1 hour… The closer the number is to 1, the higher the credibility of the appointment lead; the remaining time is approximately 24 hours. Approaching zero, the clues are almost worthless.

[0056] In calculation At that time, it can obtain the user's historical vehicle frequency signal. The higher the historical vehicle frequency → The closer to high; the lower the historical frequency of vehicles → The closer it is to low.

[0057] Scenario 1: Car owner A uses the car normally after get off work (all signals are highly compatible with car owner A, and all roads...) All approach 1) Assumption =2m, =-40dBm, =0.5h, =6 times; So =(2 0) / (10 0) = 0.2; =( 40 ( 90)) / ( 30 ( 90))=50 / 60≈0.833; =(0.5 24) / (1 24)≈0.996; =(6 0) / (7 0)≈0.857; various All high values ​​indicate that all four clues point to car owner A. After merging, the overall score S is much greater than the predicted trigger threshold Ts. The system predicts that the user is car owner A and preloads a digital clone exclusive to car owner A.

[0058] Scenario 2: Strangers pass by the vehicle (all signals) (Approaching 0, no matching users) Assumption =9m, =-85dBm, =20h, =0 times, then =0.9; =0.083; =0.174; =0.

[0059] At this time, apart from distance Apart from being slightly high, all other information is close to 0. Multiple key clues do not match any registered users. The fusion score S is below the trigger threshold. No identity prediction is performed and no clones are preloaded.

[0060] Scenario 3: Car owner B borrows the car temporarily; only partial signal compatibility. High and low differentiation Assume that car owner B has no appointment orders for the day and no frequent driving history, but the Bluetooth device is online and very close. =1.5m, =-35dBm, =18h, =1 time, then =0.15; ≈0.917; ≈0.261; ≈0.143.

[0061] At this point, the normalized value of the Bluetooth signal strength is extremely high, while the normalized values ​​of other signals are extremely low. After weighted fusion, S reaches the trigger threshold, and the system can predict that the user is car owner B and preload car owner B's digital clone.

[0062] Specifically, the overall score This is the core criterion for determining "whether to preload the first digital clone". To ensure that this score is scientific, accurate, and adaptable to complex and ever-changing real-vehicle environments, multiple influencing factors and weights are incorporated into the formula.

[0063] The in-vehicle device can comprehensively calculate the prediction score based on historical driving habit weights, user behavior confidence coefficients, environmental interference correction coefficients, normalized values ​​of each prediction signal, fusion weighting coefficients of each prediction signal, and basic weights of each prediction signal. If the prediction score is greater than or equal to a first preset threshold, the device can preload the first digital clone corresponding to the predicted user identity.

[0064] In this way, based on user impact factors, environmental impact factors, and the normalized values, fusion weights, and basic weights corresponding to each key prediction signal, the prediction comprehensive score can be calculated efficiently and accurately, which facilitates subsequent accurate determination of whether the first digital clone needs to be preloaded based on the prediction comprehensive score.

[0065] In an optional embodiment, step S106, determining whether the actual user identity of the current user in the vehicle matches the predicted user identity, includes: Obtain the first feature set of the current in-vehicle user and the second feature set of the predicted user; Calculate the similarity between the first feature set and the second feature set; If the similarity is greater than or equal to the second preset threshold, the actual user's identity is determined to match the predicted user's identity. If the similarity is less than the second preset threshold, the actual user's identity is determined to be inconsistent with the predicted user's identity.

[0066] Specifically, the second preset threshold can be set according to actual needs, and no specific limitation is made here.

[0067] When determining whether the actual user identity of the current user in the vehicle matches the predicted user identity, the in-vehicle device can first obtain the first feature set of the current user in the vehicle and the second feature set of the predicted user, and then calculate the similarity between the first feature set and the second feature set. If the similarity is greater than or equal to a second preset threshold, it can be determined that the actual user identity matches the predicted user identity; if the similarity is less than the second preset threshold, it can be determined that the actual user identity does not match the predicted user identity.

[0068] Specifically, the similarity between the first feature set and the second feature set can be calculated using the following formula: ; in, This represents the second feature set (i.e., the set of features used to predict the user's identity). Let ∩ represent the first feature set (i.e., the set of identity features of the current users in the vehicle), and ∩ represent the intersection. ∪ represents the number of identical identity features between the two sets. | represents the total number of all identity features of both parties. | represents the number of elements in the set. This represents the similarity between the first feature set and the second feature set, with a value ranging from 0 to 1.

[0069] For example, the second feature set is: {Car owner A, frequently used seat 1, frequently used time 8 am}, and the first feature set is: {Car owner A, frequently used seat 2, frequently used time 8 am}. Then the intersection of the two is: Car owner A, 8 am (2 items), and the union of the two is: Car owner A, seat 1, seat 2, 8 am (4 items). 2 / 4 = 0.5.

[0070] The similarity between the first feature set and the second feature set is calculated. Then, it can be compared with a second preset threshold, which can be determined jointly by the identity association matching threshold Tu and the fault tolerance offset δ. The identity association matching threshold Tu can be 0.85, meaning that ideally, the identity similarity should reach 85% to be considered a complete match. The fault tolerance offset δ can be 0.10, allowing the similarity to fluctuate downwards by 10% as a tolerance space to adapt to environmental interference and slight feature loss.

[0071] like ≥Tu If δ is true, then the actual user identity matches the predicted user identity. like <Tu If δ, then the actual user identity does not match the predicted user identity.

[0072] In this way, based on the calculated similarity, it is possible to accurately identify whether the actual user identity of the current user in the vehicle matches the predicted user identity, so as to perform secondary verification of the user identity, ensure the accuracy of the user identity, and thus ensure the accuracy of the loaded digital clone.

[0073] In an optional embodiment, after step S106, which determines whether the actual user identity of the current user in the vehicle matches the predicted user identity, the method further includes: If the actual user identity does not match the predicted user identity, the resources occupied by the first digital clone are released, and the second digital clone corresponding to the actual user identity is loaded so that the second digital clone can be used to configure the vehicle.

[0074] In this way, when the actual user identity does not match the predicted user identity, the resources of the first digital clone can be released quickly and the initial loading of the second digital clone can be completed. This enables seamless fault-tolerant switching in the case of prediction errors, avoiding problems such as identity confusion, configuration mismatch, and system lag, and improving the operational stability of the vehicle system.

[0075] In addition, this application can also set a historical data forgetting decay calculation formula, which is triggered after each completion of the complete dual-track verification process of user boarding prediction loading, real-time identity recognition, and matching verification; at the same time, a daily timed periodic trigger for global historical weight decay can be configured, gradually weakening the impact of outdated usage habits on the overall prediction score by progressively decaying the weight of long-standing historical data, so that the overall prediction score continues to use the user's recent behavioral preferences. The historical data forgetting decay calculation formula is as follows: Whis′=Whis λ; Where Whis represents the original weights corresponding to the old historical data, Whis′ represents the new historical weights after decay, and λ represents the forgetting coefficient, which ranges from 0 to 1. The closer to 1, the slower the forgetting. In the vehicle scenario, λ = 0.92. 0.96.

[0076] This application can also set a sliding window sample constraint, using this window to retain only the most recent N boarding behavior records for calculating the weights and the overall prediction score; older records are directly removed from the window and no longer participate in the calculation. This automatically discards outdated usage habits from several years ago, preventing old data from interfering with current predictions, and only learning recent behavior to adapt to changes in user habits. It also controls the amount of computational data, saving onboard computing power and avoiding the need for endless accumulation of historical data. For example, assuming the sliding window sample size N=20 (i.e., only the most recent 20 boarding records are retained), and the current boarding is the t=28th time: t N+1=28 20+1=9, so only the data from the 9th to the 28th times are retained in this window. The old records from the 1st to the 8th times are directly discarded and do not participate in the calculation of weight, similarity, and forgetting decay.

[0077] In an optional embodiment, before step S104 above, preloading and predicting the user's first digital clone, the method further includes: Obtain the valid time information corresponding to the first digital clone in the local preset cache range. The local preset cache range caches multiple digital clones and multiple valid time information corresponding to the multiple digital clones. The multiple digital clones correspond one-to-one with the multiple valid time information. The multiple digital clones include at least the first digital clone. Based on the valid time information corresponding to the first digital clone, determine whether the first digital clone is within its validity period; If it is determined that the first digital clone is within the validity period, the first digital clone is retrieved from the local preset cache range, and the step of preloading the first digital clone corresponding to the user's identity is executed.

[0078] Specifically, the aforementioned valid time information may include the effective time and the expiration time. The aforementioned local preset cache range can cache multiple digital clones and their corresponding valid time information, with each digital clone corresponding to one valid time information.

[0079] Before preloading the first digital clone corresponding to the user's identity, the vehicle-mounted device can obtain the validity time information corresponding to the first digital clone in the local preset cache range. Then, based on the validity time information corresponding to the first digital clone, it can determine whether the first digital clone is within the validity period. If the first digital clone is within the validity period, it can obtain the first digital clone from the local preset cache range and perform preloading of the first digital clone corresponding to the user's identity.

[0080] This ensures that the first preloaded digital clone is within its validity period, avoiding the occurrence of expired preloaded digital clones, thereby further improving the accuracy of digital clone loading.

[0081] In an optional embodiment, after determining whether the first digital clone is within its validity period based on the above steps and the valid time information corresponding to the first digital clone, the method further includes: If it is determined that the first digital clone is not within its validity period, the first digital clone is deleted from the local preset cache range, and a request instruction is sent to the cloud server. The request instruction is used to request the cloud server to regenerate and issue a new first digital clone.

[0082] Specifically, if the first digital clone is not within its validity period, the vehicle-mounted device can delete the first digital clone from the local preset cache area and send a request instruction to the cloud server to request the cloud server to regenerate and distribute a new first digital clone to the vehicle-mounted device.

[0083] In this way, after the first digital clone is detected to have expired or become invalid, a request can be made to the cloud server to regenerate and distribute a new first digital clone to the vehicle device, so that the vehicle device can continue to preload the new first digital clone.

[0084] In an optional embodiment, the method further includes: Every preset time interval, each digital clone cached in the local preset cache range is checked to determine whether each digital clone cached in the local preset cache range is within its validity period; If it is determined that the third digital clone is not within its validity period, the third digital clone is deleted from the local preset cache range, and a request instruction is sent to the cloud server. The request instruction is used to request the cloud server to regenerate and issue a new third digital clone, which can be any digital clone within the local preset cache range.

[0085] Specifically, the vehicle-mounted device can also check each digital clone cached in the local preset cache range at preset intervals to determine whether each digital clone cached in the local preset cache range is within the validity period. If it is detected that a digital clone is not within the validity period, the digital clone is deleted from the local preset cache range, and a request instruction is sent to the cloud server to request the cloud server to regenerate and distribute a new digital clone.

[0086] In this way, cache management can be performed on each digital clone cached within the local preset cache range, and the function of automatic expiration can be realized; communication and computing power consumption can be reduced, and storage utilization can be optimized.

[0087] As an optional implementation, the structure of the dual-track verification system provided by this digital clone loading method can be as follows: Figure 2As shown, it can include a dual-track verification master controller, a pre-judgment loading submodule, a real-time identification submodule, a clone switching submodule, and a dynamic update submodule. The dual-track verification master controller is the core scheduling unit of the entire system, responsible for coordinating, initiating, and managing the execution order of all submodules, controlling the logical and orderly operation of the entire process of "pre-judgment loading—real-time identification—clone switching—parameter update," and ensuring the stable operation of the dual-track mechanism. The pre-judgment loading submodule is used to implement the pre-judgment loading function before the user boards the vehicle; it belongs to the pre-judgment track and is used to improve response speed. Its main functions are: 1) Parsing pre-judgment signals: reading and processing multi-source pre-judgment signals such as user reservation information, location, historical habits, and device association to determine the pre-judged user's identity. 2) Cache and pre-load the first digital clone: ​​based on the pre-judged user's identity, pre-loading and caching the corresponding first digital clone so that the user can use it immediately after boarding without waiting. The real-time identification submodule is used to verify the user's true identity after boarding the vehicle; it belongs to the real-time verification track and is used to ensure accurate identification. Its main functions are: 1) Real-time biometric / facial recognition: Collecting users' real biometric features or identity information through facial recognition, fingerprint, account verification, etc. 2) Outputting real identity ID: Outputting the real user ID after recognition, used for matching and verification with the pre-loaded clone. The clone switching submodule is used to perform rapid fault-tolerant correction when the prediction is wrong, which belongs to the dynamic correction link. Its main functions are: 1) Destroying old clone resources: Releasing the system resources occupied by the first digital clone with the prediction error. 2) Initializing the second target clone: ​​Immediately loading the second digital clone that matches the real identity, realizing seamless and rapid switching. The dynamic update submodule is used to form a closed-loop optimization and continuously improve the prediction accuracy. Its main functions are: 1) Matching result weight adjustment: Adjusting the weight of the corresponding signal according to whether the current prediction matches the real identity. 2) Prediction parameter iteration: Updating the prediction parameters to make the next prediction more accurate. Its dual-track verification process is as follows: Figure 3 As shown, the specific steps include the following: Step 302: Collect multi-source prediction signals before the user gets on the vehicle.

[0088] Step 304: Calculate the comprehensive prediction score based on the multi-source prediction signals.

[0089] Step 306: Determine whether the predicted comprehensive score is greater than or equal to the first preset threshold.

[0090] If the predicted overall score is less than the first preset threshold, then proceed to step 308; if the predicted overall score is greater than or equal to the first preset threshold, then proceed to step 310.

[0091] Step 308: Do not preload the first digital clone.

[0092] Step 310: Preload the first digital clone.

[0093] Step 312: After the user gets on the vehicle, identity verification is triggered.

[0094] Step 314: Determine whether the identity matching is successful.

[0095] If the identity is successfully matched, proceed to step 316; if the identity is not matched, proceed to step 318.

[0096] Step 316: Quickly switch to the second digital clone.

[0097] Step 318: Continue using the first digital clone.

[0098] Step 320: Dynamically adjust the weight parameters.

[0099] Compared with existing single serial user identification and digital cloning technologies, this application has the following beneficial technical effects: 1. By adding an independent predictive loading submodule, the first digital clone is cached in advance by parsing multi-source predictive signals before the user gets in the car. This changes the lagging mode of existing technology, which can only identify and load serially after the user gets in the car. It greatly shortens the startup and readiness time of the in-vehicle digital clone and significantly improves the system response efficiency and user experience.

[0100] 2. By setting up a real-time recognition submodule and a pre-judgment loading to form a dual-track verification architecture, the real identity ID is output through real-time biometric / facial recognition and similarity matching verification is performed with the pre-judged identity. This effectively makes up for the shortcomings of existing technologies that lack a real-time verification process and cannot detect prediction errors in a timely manner, thereby improving the accuracy of user identity recognition and clone matching.

[0101] 3. By configuring the clone switching submodule, the old clone resources can be quickly destroyed and the second target clone can be initialized and loaded when identity matching fails. This enables seamless fault-tolerant switching in the event of a predicted error, avoiding problems such as identity confusion, configuration mismatch, and system lag, and improving the stability of the vehicle system.

[0102] 4. By adding a dynamic update submodule, relying on the matching result weight adjustment and prediction parameter iteration mechanism, combined with the gain step size for successful matching and the attenuation step size for failed matching, the basic weight of each signal source is adaptively optimized; at the same time, a sliding window sample constraint and historical data forgetting attenuation mechanism are introduced to retain only the recent valid vehicle entry records for calculation and gradually weaken the weight of long-standing vehicle usage habits.

[0103] 5. By coordinating user feature classes, environmental perception classes, signal source weight classes, judgment logic control classes, and self-learning iteration classes, and combining environmental interference correction coefficients, multi-signal fusion weighting coefficients, and signal normalization processing, it can adaptively adapt to complex vehicle scenarios such as nighttime, rainy days, dense crowds, and building obstructions, reducing the impact of environmental interference on prediction results and improving the system's adaptability and robustness under different working conditions.

[0104] 6. A joint judgment mechanism of multi-source signal weighted fusion, Jaccard identity similarity matching, and threshold plus fault tolerance offset is adopted to quantify the identity matching standard and avoid the problem of easy misjudgment by a single judgment rule; at the same time, the timing of prediction triggering is precisely controlled by distance sensing threshold and time window parameters to reduce invalid preloading process and save computing resources and operating power of the vehicle system.

[0105] 7. Normalized statistics and mean calculation are used for the weights of each signal source, user habit weights, and behavioral confidence. The parameter values ​​are standardized, quantifiable, and iterative. Compared with the existing empirical fixed parameter mode, it has stronger adaptive and scalable capabilities, which facilitates subsequent function iterations and adaptation to multi-user and multi-vehicle scenarios.

[0106] 8. By binding valid time information to each digital clone and performing time verification, the problem of historical digital clones being in effect for a long time due to the lack of defined validity period in existing technologies is effectively solved. This makes the digital clone highly matched with the user's real-time car use scenario, avoids the erroneous reuse of digital clones that are not in the corresponding time period (such as the user's prediction on Monday still being effective on Tuesday), greatly improves the accuracy and rationality of digital clones, and enhances the user's car use experience.

[0107] 9. Within the validity period of the digital clone, the local cache can be read directly without repeatedly requesting the cloud server, which greatly reduces the communication frequency between the vehicle and the cloud, reduces data transmission loss, reduces the computing power consumption of the vehicle, and improves the response speed of the digital clone's preloading. At the same time, expired cache can be automatically cleaned up to avoid invalid cache occupying vehicle storage resources and optimize storage resource utilization.

[0108] See Figure 4 , Figure 4 This is a schematic diagram of the structure of a digital clone loading device provided in an embodiment of this application, as shown below. Figure 4 As shown, the digital clone loading device 400 includes: The first acquisition module 402 is used to acquire multi-source prediction signals when preset conditions are met. The preset conditions are the trigger conditions for predicting the user of the vehicle. The multi-source prediction signals refer to the prediction signals acquired from multiple signal sources. The first determining module 404 is used to determine the comprehensive prediction score corresponding to the predicted user based on the multi-source prediction signal, and preload the first digital clone of the predicted user when the comprehensive prediction score is greater than or equal to the first preset threshold. The comprehensive prediction score is used to characterize the probability that the predicted user is the actual user who took the bus. The second determining module 406 is used to determine whether the actual user identity of the current user in the vehicle matches the predicted user identity. Setting module 408 is used to set up the vehicle using the first digital clone when the actual user identity is determined to match the predicted user identity.

[0109] Furthermore, the first determining module 404 includes: The first determination submodule is used to determine the normalized value, fusion weight, and basic weight of each key prediction signal in the multi-source prediction signals based on the multi-source prediction signals. The multi-source prediction signals include multiple key prediction signals. The normalized value refers to the value obtained after normalizing the original value of each key prediction signal. The fusion weight is used to characterize the importance between the multiple key prediction signals. The basic weight is used to characterize the credibility of each key prediction signal in this prediction. The second determination submodule is used to determine the user impact factor and environmental impact factor corresponding to the current prediction based on multi-source prediction signals. The user impact factor is used to characterize the probability that each candidate user was predicted in the past. Candidate users refer to all users who have participated in the past prediction. The environmental impact factor is used to characterize the degree of interference of the current environment on the current prediction. The first calculation submodule is used to calculate the comprehensive score of each candidate user based on the user impact factor, environmental impact factor, and the normalized value, fusion weight and basic weight corresponding to each key prediction signal. The third determination submodule is used to determine the candidate user with the highest comprehensive score as the predicted user, and to determine the comprehensive score of the predicted user as the predicted comprehensive score.

[0110] Furthermore, the formula for calculating the overall score of each candidate user is as follows: ; in, Indicates the first The overall score of each candidate user Indicates the first User influence factor of each candidate user Indicates environmental impact factors, Indicates the first The basic weights corresponding to the key prediction signals of the road are Indicates the first The fusion weights corresponding to the key prediction signals of the road are Indicates the first The normalized value corresponding to the key road prediction signal.

[0111] Furthermore, the second determining module 406 includes: The acquisition submodule is used to acquire the first feature set of the current in-vehicle user and the second feature set of the predicted user. The second calculation submodule is used to calculate the similarity between the first feature set and the second feature set; The fourth determination submodule is used to determine whether the actual user identity matches the predicted user identity when the similarity is greater than or equal to the second preset threshold. The fifth determination submodule is used to determine whether the actual user identity does not match the predicted user identity when the similarity is less than the second preset threshold.

[0112] Furthermore, the digital clone loading device 400 also includes: The release module is used to release the resources occupied by the first digital clone and load the second digital clone corresponding to the actual user identity when the actual user identity does not match the predicted user identity, so as to use the second digital clone to set up the vehicle.

[0113] Furthermore, the digital clone loading device 400 also includes: The second acquisition module is used to acquire the valid time information corresponding to the first digital clone in the local preset cache interval. The local preset cache interval caches multiple digital clones and multiple valid time information corresponding to the multiple digital clones. The multiple digital clones correspond one-to-one with the multiple valid time information. The multiple digital clones include at least the first digital clone. The third determining module is used to determine whether the first digital clone is within its validity period based on the valid time information corresponding to the first digital clone. The third acquisition module is used to acquire the first digital clone from the local preset cache range when it is determined that the first digital clone is within the validity period, and to perform the step of preloading the first digital clone corresponding to the user's identity.

[0114] Furthermore, the digital clone loading device 400 also includes: The first deletion module is used to delete the first digital clone from the local preset cache range when it is determined that the first digital clone is not within the validity period, and send a request instruction to the cloud server. The request instruction is used to request the cloud server to regenerate and issue a new first digital clone.

[0115] Furthermore, the digital clone loading device 400 also includes: The detection module is used to detect each digital clone cached in the local preset cache range at preset intervals to determine whether each digital clone cached in the local preset cache range is within its validity period. The second deletion module is used to delete the third digital clone from the local preset cache range when it is determined that the third digital clone is not within the validity period, and send a request instruction to the cloud server. The request instruction is used to request the cloud server to regenerate and issue a new third digital clone, which can be any digital clone within the local preset cache range.

[0116] It should be noted that the digital clone loading device 400 can implement the steps of the digital clone loading method provided in any of the aforementioned method embodiments and achieve the same technical effect, which will not be elaborated here.

[0117] like Figure 5 As shown in the illustration, this application also provides an in-vehicle device, including a processor 511, a communication interface 512, a memory 513, and a communication bus 514, wherein the processor 511, the communication interface 512, and the memory 513 communicate with each other via the communication bus 514. Memory 513 is used to store computer programs; In one embodiment of this application, when the processor 511 executes the program stored in the memory 513, it implements the digital clone loading method provided in any of the foregoing method embodiments.

[0118] This application also provides a vehicle that includes an on-board device. The on-board device can implement the steps of the digital clone loading method provided in any of the foregoing method embodiments and achieve the same technical effect, which will not be described in detail here.

[0119] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the digital clone loading method provided in any of the foregoing method embodiments.

[0120] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0121] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0122] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0123] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for loading digital clones, characterized in that, The method includes: Under the condition of meeting the preset conditions, multi-source prediction signals are acquired, wherein the preset conditions are the trigger conditions for predicting the user of the vehicle, and the multi-source prediction signals refer to the prediction signals acquired from multiple signal sources. Based on the multi-source prediction signal, the prediction comprehensive score corresponding to the prediction user is determined, and if the prediction comprehensive score is greater than or equal to a first preset threshold, the first digital clone of the prediction user is preloaded, wherein the prediction comprehensive score is used to characterize the probability that the prediction user is the actual user who took the bus. Determine whether the actual user identity of the current user in the vehicle matches the predicted user identity; If the actual user identity matches the predicted user identity, the vehicle is configured using the first digital clone.

2. The method according to claim 1, characterized in that, The step of determining the comprehensive prediction score corresponding to the predicted user based on the multi-source prediction signals includes: Based on the multi-source prediction signals, the normalized value, fusion weight, and basic weight of each key prediction signal in the multi-channel key prediction signals are determined. The multi-source prediction signals include the multi-channel key prediction signals. The normalized value refers to the value obtained after normalizing the original value of each key prediction signal. The fusion weight is used to characterize the importance among the multi-channel key prediction signals. The basic weight is used to characterize the credibility of each key prediction signal in this prediction. Based on the multi-source prediction signal, the user impact factor and environmental impact factor corresponding to this prediction are determined. The user impact factor is used to characterize the probability that each candidate user was predicted in the past. The candidate user refers to all users who have participated in the past prediction. The environmental impact factor is used to characterize the degree of interference of the current environment on this prediction. The comprehensive score of each candidate user is calculated based on the user impact factor, the environmental impact factor, the normalized value, the fusion weight, and the basic weight corresponding to each key prediction signal. The candidate user with the highest overall score is determined as the predicted user, and the overall score of the predicted user is determined as the predicted overall score.

3. The method according to claim 2, characterized in that, The formula for calculating the overall score of each candidate user is as follows: ； in, Indicates the first The overall score of each candidate user Indicates the first User influence factor of each candidate user This refers to the environmental impact factors. Indicates the first The basic weights corresponding to the key road prediction signals, Indicates the first The fusion weights corresponding to the key prediction signals of the road are Indicates the first The normalized value corresponding to the key road prediction signal.

4. The method according to claim 1, characterized in that, Determining whether the actual user identity of the current user in the vehicle matches the predicted user identity includes: Obtain the first feature set of the current in-vehicle user and the second feature set of the predicted user; Calculate the similarity between the first feature set and the second feature set; If the similarity is greater than or equal to a second preset threshold, it is determined that the actual user identity matches the predicted user identity. If the similarity is less than the second preset threshold, it is determined that the actual user identity does not match the predicted user identity.

5. The method according to claim 4, characterized in that, After determining whether the actual user identity of the current user in the vehicle matches the predicted user identity, the method further includes: If it is determined that the actual user identity does not match the predicted user identity, the resources occupied by the first digital clone are released, and the second digital clone corresponding to the actual user identity is loaded, so as to use the second digital clone to set up the vehicle.

6. The method according to claim 1, characterized in that, Before preloading the first digital clone of the predicted user, the method further includes: Obtain the valid time information corresponding to the first digital clone in the local preset cache interval, wherein the local preset cache interval caches multiple digital clones and multiple valid time information corresponding to the multiple digital clones, the multiple digital clones and the multiple valid time information correspond one-to-one, and the multiple digital clones include at least the first digital clone; Based on the valid time information corresponding to the first digital clone, determine whether the first digital clone is within its validity period; If it is determined that the first digital clone is within the validity period, the first digital clone is retrieved from the local preset cache range, and the step of preloading the first digital clone corresponding to the predicted user identity is executed.

7. The method according to claim 6, characterized in that, After determining whether the first digital clone is within its validity period based on the valid time information corresponding to the first digital clone, the method further includes: If it is determined that the first digital clone is not within its validity period, the first digital clone is deleted from the local preset cache range, and a request instruction is sent to the cloud server, wherein the request instruction is used to request the cloud server to regenerate and issue a new first digital clone.

8. The method according to claim 6, characterized in that, The method further includes: The digital clones cached in the local preset cache range are checked at preset intervals to determine whether each digital clone cached in the local preset cache range is within its validity period. If it is determined that the third digital clone is not within its validity period, the third digital clone is deleted from the local preset cache range, and a request instruction is sent to the cloud server. The request instruction is used to request the cloud server to regenerate and issue a new third digital clone, which is any digital clone within the local preset cache range.

9. A vehicle-mounted device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the digital clone loading method according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the digital clone loading method according to any one of claims 1-8.

Images