Platform privacy data processing method and system based on car using demands of clients

Abstract

The invention relates to a platform privacy data processing method based on a car using demand of a client. The method comprises the following steps: acquiring initial application data of a customer and extracting features to construct a multi-modal demand data set; performing analysis by using a machine self-learning algorithm to generate a preliminary vehicle demand category and a privacy protection strategy; collecting real-time driving behaviors, environment perception and iris tracking data, and constructing an actual driving state data set; analyzing and correcting the preliminary demand category into actual vehicle demand data; and a privacy protection strategy is dynamically adjusted to protect real-time driving state data. The privacy of the customer can be effectively protected, and the platform service quality is improved, so that the service platform use experience of the customer is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data privacy protection, and in particular, to a method and system for processing platform privacy data based on customer vehicle usage requirements. Background Art

[0002] With the development of the sharing economy, travel services such as online car-hailing and car rental platforms have become increasingly popular. While providing convenient travel services, these platforms also collect and process a large amount of customer privacy data, including personal identity information, location data, driving habits, etc. How to effectively protect this privacy data has become an important issue.

[0003] Existing vehicle service platforms only perform simple desensitization processing when collecting data, but this processing method is often not deep enough to prevent data leakage during transmission and processing. In addition, existing methods usually lack the ability to dynamically adjust privacy protection policies and cannot provide targeted privacy protection according to different vehicle usage scenarios and customer requirements, thus reducing the experience of customers when using vehicle service platforms. Therefore, improvement is needed. Summary of the Invention

[0004] In order to improve the experience of customers when using vehicle service platforms, the present application provides a method and system for processing platform privacy data based on customer vehicle usage requirements.

[0005] In a first aspect, the above-mentioned object of the present invention of the present application is achieved through the following technical solutions:

[0006] A method for processing platform privacy data based on customer vehicle usage requirements, the method comprising the steps of:

[0007] Obtain the preliminary application data when a customer submits a vehicle usage requirement, and extract requirement features from the preliminary application data to construct a multi-modal requirement data set of the customer;

[0008] Analyze the multi-modal requirement data set by a pre-set customer requirement analysis model based on a machine learning algorithm to generate a preliminary vehicle usage requirement category of the customer;

[0009] Analyze the preliminary vehicle usage requirement category based on pre-set preliminary privacy protection rules to generate a preliminary privacy protection strategy, and the preliminary privacy protection strategy is used to protect the preliminary application data of the customer;

[0010] Obtain the real-time driving behavior data of the customer, where the real-time driving behavior data includes the frequency of steering wheel rotation and the corresponding rotation angle, and the switching frequency data of the drive pedal and the brake pedal;

[0011] Obtain the environmental perception data of driving, where the environmental perception data includes visual monitoring data and radar data;

[0012] Obtain iris tracking data of the vehicle occupants, and when the in-vehicle camera detects iris features, generate an adversarial noise pattern to cover the sensitive area;

[0013] Associate the real-time driving behavior data, environmental perception data, and iris tracking data based on a pre-set common timeline to construct an actual driving state dataset;

[0014] A pre-set driving state analysis model analyzes the real-time driving state data based on a machine self-learning algorithm to generate driving state correction data, and the driving state correction data is used to correct the preliminary vehicle usage demand category;

[0015] A pre-set actual vehicle usage demand analysis model analyzes the driving state correction data and the preliminary vehicle usage demand category based on a machine self-learning algorithm to generate actual vehicle usage demand data;

[0016] Analyze the actual vehicle usage demand data based on pre-set privacy protection rules to generate a corrected privacy protection strategy, and the corrected privacy protection strategy is used to protect the customer's real-time driving state data.

[0017] By adopting the above technical solutions, by obtaining the customer's preliminary application data and extracting demand characteristics, a multi-modal demand dataset is constructed, and then a machine self-learning algorithm is used to analyze the customer's vehicle usage demand category and generate a preliminary privacy protection strategy. Compared with the prior art, this method not only protects the customer's preliminary application data, but also further enhances the comprehensiveness and dynamics of privacy protection by obtaining the customer's driving behavior data, environmental perception data, and iris tracking data in real time. The prior art often can only perform simple desensitization processing during the data collection stage and is difficult to cope with the risk of data leakage during the transmission and processing process. However, the present invention constructs an actual driving state dataset and uses a driving state analysis model to generate driving state correction data, thereby dynamically adjusting the privacy protection strategy to ensure continuous protection of the customer's privacy throughout the vehicle usage process. In addition, the prior art has deficiencies in dealing with the biometric information of vehicle occupants. The present invention effectively prevents the leakage of sensitive information such as iris data by generating an adversarial noise pattern to cover the sensitive area, further enhancing the strength of privacy protection.

[0018] In a preferred example of the present application, it can be further configured as follows: in the step where a pre-set customer demand analysis model analyzes the multi-modal demand dataset based on a machine self-learning algorithm to generate the customer's preliminary vehicle usage demand category, the following steps are included:

[0019] Obtain each demand characteristic in the multi-modal demand dataset, and the demand characteristics include driving mode parameters, trip purpose, and passenger-carrying status;

[0020] The pre-set customer demand portrait construction model analyzes the multi-modal demand data set based on the machine self-learning algorithm to construct a customer demand portrait;

[0021] Based on the pre-set portrait scoring model, a weighted score is given to the customer demand portrait to calculate the comprehensive demand score of the customer;

[0022] Based on the comprehensive demand score of the customer, a corresponding preliminary vehicle usage demand category is generated, where the preliminary vehicle usage demand category includes tourism travel, business reception, shared car rental, and emergency rescue.

[0023] By adopting the above technical solution, a preliminary vehicle usage demand category is generated based on the comprehensive demand score. The platform can provide services that better meet the actual needs of customers, and at the same time lay a foundation for targeted privacy protection strategies. Through the closed loop of multi-modal feature extraction, portrait construction, dynamic scoring, and intelligent classification, accurate prediction of vehicle usage needs and efficient resource scheduling are realized, thereby improving the customer experience.

[0024] In a preferred example of the present application, it can be further configured as follows: After the step of the pre-set customer demand portrait construction model analyzing the multi-modal demand data set based on the machine self-learning algorithm to construct a customer demand portrait, the following steps are included:

[0025] Construct a vertical federated model across service providers to align the privacy of different user portrait features;

[0026] Specifically, identify the overlapping users in each service provider's data set as the common samples for federated training, and extract the detailed personal data of the overlapping users as the initial common sample data set;

[0027] Establish a logical mapping relationship for the feature fields across service providers, and process the initial common sample data set based on the logical mapping relationship to construct a standard common sample data set;

[0028] Calculate the cross-data features of the data with the mapping relationship through the homomorphic encryption algorithm to generate a joint feature vector;

[0029] Perform user ID preprocessing on the user identifiers of different service providers in the standard common sample data set to unify them into hash values, thereby eliminating format differences.

[0030] By adopting the above technical solution, without disclosing the original data, the joint analysis and model training of data from multiple service providers are realized. By calculating cross features, the diversity and depth of features are increased, which helps to improve the prediction performance of the model. Homomorphic encryption ensures the confidentiality of data during the calculation process, enhances users' trust in data security, eliminates the differences in data formats of users from different service providers, simplifies the data processing flow, and the data after hash processing is not easily reverse-cracked, further protecting users' privacy. The unified format helps to integrate the data of multiple service providers and improve the efficiency and accuracy of federated learning.

[0031] In a preferred example of the present application, it can be further configured as follows: after the step of preprocessing the user identifiers of different service providers in the standard common sample data set to unify them into hash values, the following steps are included:

[0032] Add Laplace noise to the feature binning boundaries of the standard common sample data set to confuse the statistical distribution;

[0033] Each service provider locally calculates the feature Shapley values of the standard common sample data set, and obtains the global importance ranking corresponding to each data in the standard common sample data through a secure aggregation algorithm, thereby preventing each service provider from directly sharing feature data to prevent reverse inference of user behavior.

[0034] By adopting the above technical solution, each service provider only shares the encrypted Shapley values, the original feature distribution and user behavior patterns are completely hidden, and the global importance ranking guides resource allocation, making the decision interpretable.

[0035] In a preferred example of the present application, it can be further configured as follows: after the step of analyzing the driving state correction data and the preliminary vehicle usage demand categories by a preset actual vehicle usage demand analysis model based on machine self-learning algorithm to generate actual vehicle usage demand data, the following steps are included:

[0036] Compare the actual driving behavior and the predicted value through a federated learning algorithm to generate a driving deviation degree;

[0037] Evaluate the driving deviation degree based on a machine self-learning algorithm to generate a corresponding abnormal behavior index;

[0038] Analyze the environment perception data based on a visual recognition algorithm to dynamically generate an environment threat coefficient;

[0039] A preset driving dynamic risk assessment model comprehensively analyzes the abnormal behavior index and the environment threat coefficient to generate a dynamic risk value of the customer, and the dynamic risk value is used to dynamically evaluate the priority level of privacy protection.

[0040] By adopting the above technical solutions, comprehensively considering driving behaviors and environmental factors, comprehensively evaluating driving risks, automatically enhancing the privacy protection level in high-risk driving situations, reducing the risk of data leakage, and adapting the dynamic weight mechanism to complex scenarios.

[0041] In a preferred example of the present application, it can be further configured as follows: after the step of analyzing the actual vehicle usage demand data based on preset privacy protection rules to generate a revised privacy protection strategy, the following steps are included:

[0042] Obtain service feedback data on customer privacy protection, and associate the actual vehicle usage demand data, the revised privacy protection strategy, and the service feedback data based on the public time axis to construct a privacy service feedback-related data set;

[0043] A preset privacy service adjustment model analyzes the privacy service feedback-related data set based on a machine learning algorithm to generate a privacy leakage-service degradation correlation coefficient, and the privacy leakage-service degradation correlation coefficient is used to adjust the weighted value for calculating the dynamic risk value.

[0044] By adopting the above technical solutions, by analyzing the privacy service feedback-related data set and generating a privacy leakage-service degradation correlation coefficient, the platform can dynamically adjust the priority and intensity of the privacy protection strategy, better balance privacy protection and service quality, and improve customer satisfaction and trust.

[0045] In a preferred example of the present application, it can be further configured as follows: after the step of constructing a vertical federated model across service providers to perform privacy alignment on different user portrait features, the following steps are included:

[0046] Obtain the real-time survival status of the service provider. If the survival status of the service provider has expired, automatically destroy the historical session key and block the decryption ability for the real-time standard common sample data set.

[0047] In a second aspect, the above object of the present invention of the present application is achieved by the following technical solutions:

[0048] A platform privacy data processing device based on customer vehicle usage requirements, the device includes: a multi-modal requirement data set construction unit, configured to obtain preliminary application data when a customer submits a vehicle usage requirement, and perform requirement feature extraction on the preliminary application data to construct a multi-modal requirement data set of the customer;

[0049] A preliminary vehicle usage requirement category generation unit, configured to preset a customer requirement analysis model to analyze the multi-modal requirement data set based on a machine learning algorithm to generate a preliminary vehicle usage requirement category of the customer;

[0050] A preliminary privacy protection policy generation unit, configured to analyze the preliminary vehicle usage demand categories based on preset preliminary privacy protection rules to generate a preliminary privacy protection policy for protecting the customer's preliminary application data;

[0051] A real-time driving behavior data acquisition unit, configured to acquire the customer's real-time driving behavior data;

[0052] An environmental perception data acquisition unit, configured to acquire the environmental perception data of driving;

[0053] An iris tracking data acquisition unit, configured to acquire the iris tracking data of the vehicle occupants, and generate an adversarial noise pattern to cover the sensitive area when the in-vehicle camera detects iris features;

[0054] An actual driving state data set construction unit, configured to associate the real-time driving behavior data, environmental perception data, and iris tracking data based on a preset common time axis to construct an actual driving state data set;

[0055] A driving state correction data generation unit, configured to preset a driving state analysis model to analyze the real-time driving state data based on a machine self-learning algorithm to generate driving state correction data;

[0056] An actual vehicle usage demand data generation unit, configured to preset an actual vehicle usage demand analysis model to analyze the driving state correction data and the preliminary vehicle usage demand categories based on a machine self-learning algorithm to generate actual vehicle usage demand data;

[0057] A corrected privacy protection policy generation unit, configured to analyze the actual vehicle usage demand data based on preset privacy protection rules to generate a corrected privacy protection policy for protecting the customer's real-time driving state data.

[0058] Thirdly, the above object of the present application is achieved by the following technical solutions:

[0059] An electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the above platform privacy data processing method based on customer vehicle usage demands are implemented.

[0060] Fourthly, the above object of the present application is achieved by the following technical solutions:

[0061] A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the steps of the above platform privacy data processing method based on customer vehicle usage demands are implemented.

[0062] In summary, the present application includes at least one of the following beneficial technical effects:

[0063] 1. By obtaining the customer's preliminary application data and extracting demand characteristics, a multi-modal demand data set is constructed. Then, a machine self-learning algorithm is used to analyze the customer's vehicle usage demand categories and generate a preliminary privacy protection strategy. Compared with the prior art, this method not only protects the customer's preliminary application data, but also further enhances the comprehensiveness and dynamics of privacy protection by obtaining the customer's driving behavior data, environment perception data, and iris tracking data in real time. The prior art often only performs simple desensitization processing during the data collection stage and is difficult to cope with the risk of data leakage during transmission and processing. However, the present invention constructs an actual driving state data set and uses a driving state analysis model to generate driving state correction data, thereby dynamically adjusting the customer's actual vehicle usage demand and adjusting the corresponding privacy protection strategy to ensure continuous protection of the customer's privacy during the entire vehicle usage process while improving the customer's service platform usage experience. The traditional solution only sets a fixed privacy intensity based on the initial demand, resulting in insufficient protection in high-sensitivity scenarios or excessive interference with the service experience in low-risk scenarios. The multi-source perception real-time correction mechanism of the present solution can trigger dynamic adjustment of the privacy policy, thereby avoiding the situation of insufficient protection in high-sensitivity scenarios or excessive interference with the service experience in low-risk scenarios. In addition, the prior art has deficiencies in processing the biometric information of the occupants in the vehicle. The present invention effectively prevents the leakage of sensitive information such as iris data by generating adversarial noise patterns to cover sensitive areas, further enhancing the intensity of privacy protection;

[0064] 2. Based on the comprehensive demand score, a preliminary vehicle usage demand category is generated. The platform can provide services that more closely meet the actual needs of the customer, and at the same time lay a foundation for targeted privacy protection strategies. Through the closed loop of multi-modal feature extraction, portrait construction, dynamic scoring, and intelligent classification, accurate prediction of vehicle usage demand and efficient resource scheduling are realized, thereby improving the customer's usage experience;

[0065] 3. Without revealing the original data, the joint analysis and model training of the data of multiple service providers are realized. By calculating cross features, the diversity and depth of the features are increased, which helps to improve the prediction performance of the model. Homomorphic encryption ensures the confidentiality of the data during the calculation process, enhances the user's trust in data security, eliminates the data format differences of users of different service providers, simplifies the data processing process, and the data after hash processing is not easily reverse-cracked, further protecting the user's privacy. The unified format helps to integrate the data of multiple service providers and improve the efficiency and accuracy of federated learning; BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 is a flowchart of a platform privacy data processing method based on customer vehicle usage demand in an embodiment of the present application;

[0067] Figure 2 This is a principle block diagram of a platform privacy data processing device based on customers' vehicle usage requirements in an embodiment of the present application;

[0068] Figure 3 This is a schematic diagram of an electronic device in an embodiment of the present application.

[0069] Reference numerals in the drawings:

[0070] 1. Multimodal requirement dataset construction unit; 2. Preliminary vehicle usage requirement category generation unit; 3. Preliminary privacy protection policy generation unit; 4. Real-time driving behavior data acquisition unit; 5. Environment perception data acquisition unit; 6. Iris tracking data acquisition unit; 7. Actual driving state dataset construction unit; 8. Driving state correction data generation unit; 9. Actual vehicle usage requirement data generation unit; 10. Revised privacy protection policy generation unit. Detailed implementation manners

[0071] The present application will be further described in detail below with reference to the accompanying drawings.

[0072] In one embodiment, as Figure 1 shown, the present application discloses a platform privacy data processing method based on customers' vehicle usage requirements, which specifically includes the following steps:

[0073] S01: Obtain the preliminary application data when the customer submits the vehicle usage requirements, and extract the requirement features from the preliminary application data to construct the customer's multimodal requirement dataset;

[0074] Specifically, taking the text requirement as an example, the customer submits the text requirement: "Send three business partners to Pudong Airport at 8 o'clock tomorrow morning, and a luxury vehicle is required";

[0075] Extract the features from the preliminary application data:

[0076] Text semantics: Business reception, luxury vehicle (NLP parsing)

[0077] Structured data: Departure time (08:00), number of passengers (3 people), destination (airport)

[0078] Historical preference: 80% of the user's past orders selected the air conditioner in the silent mode. Construct the customer's multimodal dataset: {Requirement type: business, time: 08:00, number of passengers: 3, vehicle type preference: luxury, environmental requirement: silent}.

[0079] Fuse text, spatio-temporal, and behavioral multi-dimensional data to improve the comprehensiveness of requirement recognition; construct the dataset through non-sensitive features (such as the number of passengers instead of identity information), which conforms to the principle of minimizing collection and protects the customer's identity privacy.

[0080] S02: The pre-set customer demand analysis model analyzes the multi-modal demand data set based on the machine self-learning algorithm to generate the preliminary vehicle usage demand categories of the customer;

[0081] Specifically, in the embodiment of the present application, multi-modal data is input into a pre-trained ResNet-Transformer hybrid model, so as to output preliminary vehicle usage demand categories: business reception category (for example, confidence level 92%), and the labels corresponding to the business reception category include: high privacy sensitivity, and need for comfort guarantee.

[0082] It should be noted that in the embodiment of the present application, the preliminary vehicle usage demand categories include travel, business reception, shared car rental, and emergency rescue. In other embodiments, the preliminary vehicle usage demand categories can be classified according to different classification rules.

[0083] S03: Analyze the preliminary vehicle usage demand categories based on the pre-set preliminary privacy protection rules to generate preliminary privacy protection strategies;

[0084] Specifically, in the embodiment of the present application, the matching rules are: the business reception category corresponds to triggering high-intensity privacy protection; the travel category corresponds to triggering high-intensity privacy protection; shared car rental corresponds to triggering low-intensity privacy protection; emergency rescue corresponds to triggering the lowest-intensity privacy protection (providing vehicle usage data and customer information data according to rescue needs). The preliminary privacy protection strategy is used to protect the preliminary application data of the customer, and the preliminary privacy protection strategy includes data desensitization, data fuzzing, and encrypted storage.

[0085] In the embodiment of the present application, the preliminary privacy protection strategy includes a vehicle position fuzzing radius R = 100m (differential privacy noise injection), MFCC coefficient perturbation of voiceprint commands; default masking of biometric features (iris pixel retention rate ≤ 10%).

[0086] S04: Obtain the real-time driving behavior data of the customer;

[0087] S05: Obtain the environmental perception data of driving;

[0088] S06: Obtain the iris tracking data of the people in the vehicle, and when the in-vehicle camera detects the iris feature, generate an adversarial noise pattern to cover the sensitive area;

[0089] Specifically, for steps S04-S06, the real-time driving behavior data includes the steering wheel rotation frequency and the corresponding rotation angle, and the switching frequency data of the drive pedal and the brake pedal; the environmental perception data includes visual monitoring data and radar data.

[0090] It should be noted that after the in-vehicle camera captures the driver's iris features, the system immediately generates a specific noise pattern to cover the iris area, preventing the illegal collection and utilization of iris data. In the embodiments of the present application, the system calls a pre-trained generative adversarial network (GAN) model to generate adversarial noise patterns in real time to cover the sensitive eye area. The adversarial noise can effectively interfere with illegal iris recognition algorithms, making it impossible to extract real iris features; the noise only covers the iris texture, retaining non-sensitive features such as the eye contour to ensure the accuracy of functions such as fatigue driving monitoring; the noise intensity is adjusted in real time according to the risk level to dynamically adapt to the actual driving situation of the customer.

[0091] S07: Correlate the real-time driving behavior data, environmental perception data, and iris tracking data based on a pre-set common timeline to construct an actual driving state data set;

[0092] S08: A pre-set driving state analysis model analyzes the real-time driving state data based on machine self-learning algorithms to generate driving state correction data;

[0093] Among them, the driving state correction data is used to correct the preliminary vehicle usage demand category. Specifically, the driving state analysis model may detect that the driver is fatigued and driving into a school area, and based on historical data and real-time driving behavior (such as frequent direction adjustments and pedal switching), generate correction data to prompt rest, thereby outputting a corrected vehicle usage demand category: the demand category changes from "business" to "high safety priority" (due to fatigue driving + school area).

[0094] S09: A pre-set actual vehicle usage demand analysis model analyzes the driving state correction data and the preliminary vehicle usage demand category based on machine self-learning algorithms to generate actual vehicle usage demand data;

[0095] Specifically, according to the demand type conversion mentioned in S08, the actual vehicle usage demand analysis model inputs: the corrected demand label (high safety priority) + the original business demand, and outputs the corresponding actual vehicle usage demand: {main type: business, sub-strategy: safety enhancement, specific measures: activate assisted driving + increase the biological shielding intensity}.

[0096] S10: Analyze the actual vehicle usage demand data based on pre-set privacy protection rules to generate a corrected privacy protection strategy, and the corrected privacy protection strategy is used to protect the customer's real-time driving state data;

[0097] Specifically, for example: the corrected privacy protection strategy is as follows: the location blur radius increases from 100m to 200m; the iris shielding intensity is adjusted from 10% to 90%; homomorphic encryption is enabled to transmit real-time behavior data.

[0098] What is needed is that in the embodiments of the present application, the vehicle usage data of the customer is combined with the dual destruction conditions of the blockchain smart contract (such as automatically burning the voiceprint data 72 hours after the trip ends and without insurance claims) to achieve full life cycle control.

[0099] In summary, by obtaining the customer's preliminary application data and extracting demand characteristics, a multi-modal demand data set is constructed. Then, a machine self-learning algorithm is used to analyze the customer's vehicle usage demand categories and generate a preliminary privacy protection strategy. Compared with the prior art, this method not only protects the customer's preliminary application data, but also further enhances the comprehensiveness and dynamics of privacy protection by obtaining the customer's driving behavior data, environment perception data, and iris tracking data in real time. The prior art often only performs simple desensitization processing in the data collection stage and is difficult to cope with the risk of data leakage during transmission and processing. However, the present invention constructs an actual driving state data set and uses a driving state analysis model to generate driving state correction data, thereby dynamically adjusting the customer's actual vehicle usage demand and adjusting the corresponding privacy protection strategy to ensure continuous protection of the customer's privacy during the entire vehicle usage process while improving the customer's service platform usage experience. The traditional solution only sets a fixed privacy intensity based on the initial demand, resulting in insufficient protection in high-sensitivity scenarios or excessive interference with the service experience in low-risk scenarios; while the multi-source perception real-time correction mechanism of this solution can trigger dynamic adjustment of the privacy policy, thus avoiding the situation of insufficient protection in high-sensitivity scenarios or excessive interference with the service experience in low-risk scenarios. In addition, the prior art has deficiencies in processing the biometric information of the people in the vehicle. The present invention effectively prevents the leakage of sensitive information such as iris data by generating adversarial noise patterns to cover sensitive areas, further enhancing the intensity of privacy protection.

[0100] In step S02: The pre-set customer demand analysis model analyzes the multi-modal demand data set based on a machine self-learning algorithm to generate the customer's preliminary vehicle usage demand categories. This step includes the following steps:

[0101] S021: Obtain each demand characteristic in the multi-modal demand data set, where the demand characteristics include driving mode parameters, trip purpose, and passenger-carrying status;

[0102] Specifically, for example:

[0103] Driving mode parameters: The user voice inputs "Turn on the sports mode and avoid congested roads", and the system extracts the keywords "sports mode" (power preference) and "avoid congestion" (road condition strategy).

[0104] Trip purpose: The text requirement "Send three customers to Pudong Airport at 8 am tomorrow" → NLP parsing is "Business pick-up and drop-off" (keywords: customers, airport).

[0105] Occupied state: The in-vehicle sensor detects pressure signals from 4 seats + voice command "Adjust the rear air conditioner to 24°C" → Comprehensive judgment shows that the number of passengers (excluding the driver) is 3 people.

[0106] Step S021 fuses multi-source data of voice, text, and sensors to achieve automatic extraction of demand features, improving efficiency compared to traditional form filling; through multi-modal complementary verification (such as cross-verifying the number of passengers through pressure sensors and voice commands), the feature accuracy is improved.

[0107] S022: The pre-set customer demand portrait construction model analyzes the multi-modal demand data set based on machine self-learning algorithms to construct a customer demand portrait;

[0108] Specifically, the input features are: {Driving mode: Sporty; Trip purpose: Business pick-up and drop-off; Number of passengers: 3 people; Historical preference: 80% of orders choose luxury models}.

[0109] Construction of the customer demand portrait construction model: In the embodiment of the present application, the customer demand portrait construction model uses a clustering algorithm (K-means): classifying users as "high-net-worth business group".

[0110] Pre-set label system: Generate portrait labels {Demand intensity: 9.2 / 10, Comfort requirement: High, Privacy sensitivity: High}.

[0111] Step S022 breaks through the limitation of single-dimensional classification (such as only selecting by vehicle type), and depicts user needs through multi-labels three-dimensionally.

[0112] S023: Based on the pre-set portrait scoring model, the customer demand portrait is weighted and scored to calculate the comprehensive demand score of the customer;

[0113] Specifically, the portrait scoring model performs weighted scoring according to different features in the customer demand portrait. For example, the "business travel" feature has a higher weight, and the customer gets a high score due to frequent business travel; the "travel alone" feature has a lower weight, and the corresponding score is lower. Considering all features, the comprehensive demand score of the customer is 75 points (taking a full score of 100 as an example). Through weighted scoring, the platform can accurately quantify customer needs and provide a quantitative basis for subsequent optimization of service and privacy protection strategies.

[0114] S024: Based on the comprehensive demand score of the customer, generate corresponding preliminary vehicle usage demand categories, where the preliminary vehicle usage demand categories include tourist travel, business reception, shared car rental, and emergency rescue;

[0115] Specifically, according to the example in step S023, the comprehensive score of the customer's needs is 75 points. According to the preset scoring rules, 70 - 80 points correspond to the "business reception" category. Therefore, the platform classifies the customer's preliminary car - using needs as "business reception".

[0116] Based on the comprehensive score of the needs to generate the preliminary car - using needs category, the platform can provide services that better meet the actual needs of customers, and at the same time lay a foundation for targeted privacy protection strategies.

[0117] To sum up, through the closed - loop of multi - modal feature extraction, portrait construction, dynamic scoring, and intelligent classification, the accurate prediction of car - using needs and the efficient scheduling of resources are realized, thereby improving the customer's usage experience.

[0118] After step S022: The pre - set customer needs portrait construction model analyzes the multi - modal needs data set based on the machine self - learning algorithm to construct a customer needs portrait, the following steps are included:

[0119] S0221: Construct a vertical federated model across service providers to perform privacy alignment on different user portrait features;

[0120] For example: Suppose there are multiple travel service providers participating in federated learning. Each service provider has its own user portrait features. For example, travel service provider A may have features such as the customer's taxi - taking frequency and travel distance, while travel service provider B may have features such as the customer's reservation habits and vehicle type preferences. Through vertical federated learning, these different user portrait features are privately aligned, enabling each service provider to build a joint user portrait model without disclosing their respective user data.

[0121] In the embodiment of the present application, there is also a scenario across industry service providers: Automobile company A has user driving behavior data, insurance company B holds user credit records, and bank C has consumption ability data. The three parties establish a federated channel through an encrypted channel and agree on the overlapping user range (such as the user group with car insurance renewal records in the past 3 months).

[0122] Each service provider does not need to share the original data, avoiding the risk of user data leakage. By integrating the features of multiple service providers, the user portrait becomes more comprehensive, improving the prediction ability of the model, promoting cooperation between different service providers, jointly improving service quality, without worrying about data privacy issues, and breaking data islands. Automobile companies can integrate insurance / bank features without exposing any original data; compared with traditional data transactions, each service platform reduces compliance risks.

[0123] It should be noted that when obtaining the real - time survival status of the service provider, if the survival status of the service provider has expired, the historical session key is automatically destroyed, blocking the decryption ability for the real - time standard common sample data set.

[0124] S0222: Identify the overlapping users in the datasets of each service provider as the common samples for federated training, and extract the detailed personal data of the overlapping users as the initial common sample dataset;

[0125] Specifically, through the Private Set Intersection (PSI) technology, based on identifiers such as user ID and mobile phone number, identify the overlapping users who have used these three services simultaneously in the datasets of multiple travel service providers. For example, if user A has registration and usage records in travel service providers A, B, and C, then user A is identified as an overlapping user. Then, extract the detailed personal data of these overlapping users (such as taxi-taking frequency, trip distance, reservation habits, vehicle type preferences, etc.) as the initial common sample dataset.

[0126] This step ensures that the federated training is based on the same user group, improving the consistency and accuracy of model training. The detailed data of the overlapping users provides rich information, which helps to build a high-quality joint model, provides a basis for data collaboration between different service providers, and enhances the feasibility and effectiveness of federated learning.

[0127] S0223: Establish a logical mapping relationship for the feature fields across service providers, and process the initial common sample dataset based on the logical mapping relationship to construct a standard common sample dataset;

[0128] Specifically, in the dataset of travel service provider A, there is a feature of "taxi-taking frequency", while in the dataset of travel service provider B, there is a feature of "number of trips". These two features are essentially the same. Therefore, establish a logical mapping relationship to map "taxi-taking frequency" to "number of trips", and unify the units and definitions. Then, process the initial common sample dataset according to this mapping relationship, so that the corresponding features of all service providers are converted into a unified standard format, thereby constructing a standard common sample dataset.

[0129] Step S0223 has the following technical effects: Feature unification: Eliminates the feature differences between different service providers, improving the comparability of data and the efficiency of model training; Model generalization: The unified feature fields help the model better generalize to the user groups of different service providers; Data integration: Provides a unified data basis for further integrating the data of multiple service providers, enhancing the effect of federated learning.

[0130] S0224: Calculate the cross-data features of the data with the mapping relationship through the homomorphic encryption algorithm to generate a joint feature vector;

[0131] Specifically, for the "number of trips" feature in the standard common sample dataset, assume that the number of trips of user A in the data of travel service provider A is 10 times, and the number of trips of user A in the data of travel service provider B is 8 times. Through a homomorphic encryption algorithm (such as Paillier encryption), these two data are first encrypted, and then their cross-features are calculated, such as calculating their sum, difference, or product, etc. The final encrypted result is a part of the joint feature vector, which can be used for the training of the joint model without decrypting the original data.

[0132] Without revealing the original data, the joint analysis and model training of data from multiple service providers are realized. By calculating cross-features, the diversity and depth of features are increased, which helps to improve the prediction performance of the model. Homomorphic encryption ensures the confidentiality of data during the calculation process and enhances users' trust in data security.

[0133] S0225: Perform user ID preprocessing on the user identifiers of different service providers in the standard common sample dataset to unify them into hash values, thereby eliminating format differences;

[0134] Specifically, in the standard common sample dataset, the user ID of travel service provider A is "123456789", the user ID of travel service provider B is "user_987654321", and the user ID of travel service provider C is "T3_111222333". Through a hash algorithm (such as MD5 or SHA-256), these user IDs in different formats are converted into unified hash values. For example, "123456789" is converted into "abc123def456", "user_987654321" is converted into "def456abc123", and "T3_111222333" is converted into "ghi789jkl345". In this way, all user IDs are unified into hash values of the same length and format, which is convenient for subsequent data processing and model training.

[0135] The format differences of user IDs of different service providers are eliminated, simplifying the data processing flow. The user ID after hash processing is not easily reverse-cracked, further protecting users' privacy. The unified user ID format helps to integrate the data of multiple service providers and improve the efficiency and accuracy of federated learning.

[0136] After step S0225: Perform user ID preprocessing on the user identifiers of different service providers in the standard common sample dataset to unify them into hash values, the following steps are included:

[0137] Add Laplace noise to the feature bin boundaries of the standard common sample dataset to obfuscate the statistical distribution;

[0138] Each service provider locally calculates the Shapley value of the features of the standard common sample dataset, and obtains the global importance ranking corresponding to each data in the standard common sample data through a secure aggregation algorithm, thereby preventing each service provider from directly sharing feature data and preventing the reverse inference of user behavior.

[0139] Based on differential privacy protection, Laplace noise is added to the feature bin boundaries of the standard common sample dataset. Specifically, for example, when a bank and a travel service platform jointly model, it is necessary to bin the user income feature and calculate the boundary sensitivity (the maximum offset of adjacent dataset boundaries); generate Laplace noise, thereby determining the noise value, and adjust the user income feature bin based on the noise value to confuse the statistical distribution. The noise injection makes the bin boundaries randomly shift, blocking the path of reverse inference of individuals through distribution statistics, so that attackers cannot reverse infer individual income through the statistical distribution.

[0140] Each service provider locally calculates the Shapley value of feature i , measuring its marginal contribution to the prediction, where The calculation formula of is:

[0141]

[0142] , N is the complete set of features. Among them, S is the subset of features, f is the model prediction function, and i is the i-th feature in the dataset.

[0143] Based on the secure aggregation algorithm (through Paillier homomorphic encryption or secret sharing technology), the of each service provider is aggregated, and only the global importance ranking is exposed after decryption.

[0144] Each service provider locally calculates the Shapley value of the feature. Suppose the Shapley value of a certain feature "travel frequency" is 0.3, indicating that the average contribution of this feature to the model prediction is 30%. Through the secure aggregation algorithm, the Shapley values of all service providers are aggregated to obtain the global importance ranking. For example, "travel frequency" ranks first and "vehicle type preference" ranks third.

[0145] Zero exposure of effect data: Each party only shares the encrypted Shapley value (scalar), and the original feature distribution and user behavior pattern are completely hidden. Decision interpretability: The global importance ranking guides resource allocation.

[0146] After step S09: The pre-set actual vehicle usage demand analysis model analyzes the driving state correction data and the preliminary vehicle usage demand categories based on the machine self-learning algorithm to generate actual vehicle usage demand data, the following steps are included:

[0147] S091: Compare the actual driving behavior with the predicted values through the federated learning algorithm to generate the driving deviation degree;

[0148] Specifically, each vehicle-mounted terminal calculates the difference between the actual driving behavior (such as steering wheel angle, acceleration) and the output value of the prediction model locally, uploads the gradient through Paillier homomorphic encryption, and the central server aggregates the encrypted gradients to update the global model and outputs the driving deviation degree (such as trajectory deviation rate, speed fluctuation variance).

[0149] S092: Evaluate the driving deviation degree based on the machine self-learning algorithm to generate the corresponding abnormal behavior index;

[0150] Specifically, in the embodiment of the present application, based on the Isolation Forest and LOF (Local Outlier Factor) algorithms, analyze the spatio-temporal distribution characteristics of the deviation degree; output the abnormal behavior index (0-1.0 in the present application), and the higher the value, the greater the behavior risk.

[0151] Example: Scenario: Nighttime highway section, the driver has not rested for 2 consecutive hours, and the frequency of fine-tuning the steering wheel has increased suddenly. Evaluation process:

[0152] Input: Steering wheel fine-tuning frequency (0.8Hz → 2.3Hz), blink interval (3s → 1.2s);

[0153] Isolation Forest identification: The fine-tuning frequency is located in the abnormal branch depth = 3 (normal branch depth > 8) → Index = 0.86; Beneficial effects:

[0154] Early warning: Trigger a fatigue reminder when the index is greater than the preset threshold (earlier than the traditional DMS system);

[0155] Anti-interference: Strong robustness to sensor noise.

[0156] S093: Analyze the environmental perception data based on the visual recognition algorithm to dynamically generate the environmental threat coefficient;

[0157] Specifically, based on YOLOv5, dynamically detect road targets (pedestrians, vehicles), and combine with the spatio-temporal convolutional network (STCNN) to predict the trajectory conflict probability;

[0158] Threat coefficient = number of targets × motion entropy × reciprocal of collision time. Enhance image recognition through GAN, and the threat coefficient error is reduced.

[0159] S094: The pre-set driving dynamic risk assessment model comprehensively analyzes the abnormal behavior index and the environmental threat coefficient to generate the dynamic risk value of the customer, and the dynamic risk value is used to dynamically evaluate the priority level of privacy protection;

[0160] Specifically, for example: The driving dynamic risk assessment model combines the abnormal behavior index of 85 and the environmental threat coefficient of 0.8, and calculates the dynamic risk value of the customer as 0.75 (full score 1.0) through a preset formula. According to the risk value, the system raises the privacy protection priority level, restricts unnecessary data sharing, and gives priority to protecting the customer's privacy in high-risk situations.

[0161] To sum up, by comprehensively considering driving behavior and environmental factors, comprehensively evaluating driving risks, automatically raising the privacy protection level in high-risk driving situations, reducing the risk of data leakage, and the dynamic weight mechanism adapts to complex scenarios.

[0162] After step S10: Analyzing the actual vehicle usage demand data based on pre-set privacy protection rules to generate a revised privacy protection strategy, the following steps are included:

[0163] S101: Obtain service feedback data on customer privacy protection, and associate the actual vehicle usage demand data, the revised privacy protection strategy, and the service feedback data based on the public timeline to construct a privacy service feedback-related data set;

[0164] Specifically, after using the platform service, the customer provides service feedback data on privacy protection through a questionnaire survey or an evaluation system. For example, after using the platform service, customer A gives a score of 4 (full score 5) to the platform's privacy protection measures and points out that he is concerned about how the platform processes his driving behavior data. The platform associates customer A's actual vehicle usage demand data (such as business trips), the revised privacy protection strategy (such as encrypting driving behavior data), and the service feedback data through the public timeline to construct a privacy service feedback-related data set.

[0165] By associating the customer's privacy protection service feedback with the actual vehicle usage demand data and the revised privacy protection strategy, the platform can better understand the customer's satisfaction and concerns about privacy protection measures, providing data support for subsequent optimization of the privacy protection strategy.

[0166] S102: The pre-set privacy service adjustment model analyzes the privacy service feedback-related data set based on machine learning algorithms to generate a privacy leakage-service degradation correlation coefficient, and the privacy leakage-service degradation correlation coefficient is used to adjust the weighted value of the dynamic risk value calculation;

[0167] Specifically, the privacy service adjustment model of the platform uses machine learning algorithms (such as logistic regression or neural networks) to analyze the dataset related to privacy service feedback. For example, through analysis, the privacy service adjustment model finds that when the customer's satisfaction with privacy protection measures is less than 3 points, the risk of privacy leakage increases by 20%. Based on this, the model generates a privacy leakage-service degradation correlation coefficient of 0.2, indicating a medium-strength negative correlation between the privacy leakage risk and service satisfaction. This coefficient will be used to adjust the weighted value of the dynamic risk value calculation to reflect the impact of privacy protection measures on the overall risk.

[0168] By analyzing the dataset related to privacy service feedback and generating a privacy leakage-service degradation correlation coefficient, the platform can dynamically adjust the priority and intensity of privacy protection policies, better balance privacy protection and service quality, and improve customer satisfaction and trust.

[0169] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not imply the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0170] In one embodiment, a platform privacy data processing device based on customer vehicle usage requirements is provided. The platform privacy data processing device based on customer vehicle usage requirements corresponds one-to-one with the platform privacy data processing method based on customer vehicle usage requirements in the above embodiments. As Figure 2 shown, the platform privacy data processing device based on customer vehicle usage requirements includes:

[0171] A multimodal demand dataset construction unit 1, configured to obtain the preliminary application data when the customer submits vehicle usage requirements, and perform demand feature extraction on the preliminary application data to construct a multimodal demand dataset of the customer;

[0172] A preliminary vehicle usage requirement category generation unit 2, configured to preset a customer demand analysis model to analyze the multimodal demand dataset based on a machine learning algorithm to generate a preliminary vehicle usage requirement category of the customer;

[0173] A preliminary privacy protection policy generation unit 3, configured to analyze the preliminary vehicle usage requirement category based on a preset preliminary privacy protection rule to generate a preliminary privacy protection policy, where the preliminary privacy protection policy is used to protect the preliminary application data of the customer;

[0174] A real-time driving behavior data acquisition unit 4, configured to acquire the real-time driving behavior data of the customer;

[0175] An environment perception data acquisition unit 5, configured to acquire the environment perception data of driving;

[0176] An iris tracking data acquisition unit 6, configured to acquire iris tracking data of in-vehicle personnel, and generate an adversarial noise pattern to cover a sensitive area when the in-vehicle camera detects an iris feature;

[0177] An actual driving state dataset construction unit 7, configured to associate the real-time driving behavior data, environment perception data, and iris tracking data based on a preset common timeline to construct an actual driving state dataset;

[0178] A driving state correction data generation unit 8, configured to preset a driving state analysis model to analyze the real-time driving state data based on a machine self-learning algorithm to generate driving state correction data;

[0179] An actual vehicle usage demand data generation unit 9, configured to preset an actual vehicle usage demand analysis model to analyze the driving state correction data and the preliminary vehicle usage demand categories based on a machine self-learning algorithm to generate actual vehicle usage demand data;

[0180] A corrected privacy protection policy generation unit 10, configured to analyze the actual vehicle usage demand data based on a preset privacy protection rule to generate a corrected privacy protection policy, and the corrected privacy protection policy is used to protect the real-time driving state data of the customer.

[0181] For the specific limitations of the platform privacy data processing device based on customer vehicle usage requirements, reference can be made to the limitations of the platform privacy data processing method based on customer vehicle usage requirements in the above text, which will not be elaborated here. Each module in the above platform privacy data processing device based on customer vehicle usage requirements can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in or independent of the processor in the electronic device in the form of hardware, or stored in the memory of the electronic device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to the above modules.

[0182] In one embodiment, an electronic device is provided. The electronic device can be a server, and its internal structure diagram can be as Figure 3 shown. The electronic device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the electronic device is used to store the database. The network interface of the electronic device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a platform privacy data processing method based on customer vehicle usage requirements.

[0183] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented:

[0184] Obtain the preliminary application data when the customer submits the vehicle usage requirement, and extract the requirement features from the preliminary application data to construct a multi-modal requirement data set of the customer;

[0185] The pre-set customer requirement analysis model analyzes the multi-modal requirement data set based on the machine self-learning algorithm to generate the preliminary vehicle usage requirement categories of the customer;

[0186] Analyze the preliminary vehicle usage requirement categories based on the pre-set preliminary privacy protection rules to generate a preliminary privacy protection strategy, which is used to protect the customer's preliminary application data;

[0187] Obtain the real-time driving behavior data of the customer, where the real-time driving behavior data includes the steering wheel rotation frequency and the corresponding rotation angle, and the switching frequency data of the drive pedal and the brake pedal;

[0188] Obtain the environmental perception data of the driving, where the environmental perception data includes visual monitoring data and radar data;

[0189] Obtain the iris tracking data of the people in the vehicle, and when the in-vehicle camera detects the iris feature, generate an adversarial noise pattern to cover the sensitive area;

[0190] Associate the real-time driving behavior data, environmental perception data, and iris tracking data based on the pre-set common time axis to construct an actual driving state data set;

[0191] The pre-set driving state analysis model analyzes the real-time driving state data based on the machine self-learning algorithm to generate driving state correction data, which is used to correct the preliminary vehicle usage requirement categories;

[0192] The pre-set actual vehicle usage requirement analysis model analyzes the driving state correction data and the preliminary vehicle usage requirement categories based on the machine self-learning algorithm to generate actual vehicle usage requirement data;

[0193] Analyze the actual vehicle usage requirement data based on the pre-set privacy protection rules to generate a corrected privacy protection strategy, which is used to protect the customer's real-time driving state data.

[0194] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0195] Obtain the preliminary application data when the customer submits the vehicle usage requirements, and extract the requirement characteristics from the preliminary application data to construct a multi-modal requirement dataset of the customer;

[0196] The pre-set customer requirement analysis model analyzes the multi-modal requirement dataset based on the machine self-learning algorithm to generate the preliminary vehicle usage requirement categories of the customer;

[0197] Analyze the preliminary vehicle usage requirement categories based on the pre-set preliminary privacy protection rules to generate a preliminary privacy protection strategy, which is used to protect the customer's preliminary application data;

[0198] Obtain the customer's real-time driving behavior data, where the real-time driving behavior data includes the steering wheel rotation frequency and the corresponding rotation angle, and the switching frequency data of the drive pedal and the brake pedal;

[0199] Obtain the environmental perception data of driving, where the environmental perception data includes visual monitoring data and radar data;

[0200] Obtain the iris tracking data of the vehicle occupants, and when the in-vehicle camera detects the iris feature, generate an adversarial noise pattern to cover the sensitive area;

[0201] Associate the real-time driving behavior data, environmental perception data, and iris tracking data based on the pre-set common timeline to construct an actual driving state dataset;

[0202] The pre-set driving state analysis model analyzes the real-time driving state data based on the machine self-learning algorithm to generate driving state correction data, which is used to correct the preliminary vehicle usage requirement categories;

[0203] The pre-set actual vehicle usage requirement analysis model analyzes the driving state correction data and the preliminary vehicle usage requirement categories based on the machine self-learning algorithm to generate actual vehicle usage requirement data;

[0204] Analyze the actual vehicle usage requirement data based on the pre-set privacy protection rules to generate a corrected privacy protection strategy, which is used to protect the customer's real-time driving state data.

[0205] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0206] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0207] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for processing platform privacy data based on customers' vehicle usage requirements, characterized in that, The method includes the steps of: obtaining preliminary application data when a customer submits a vehicle usage demand, and extracting demand characteristics from the preliminary application data to construct a multi-modal demand dataset of the customer; A pre-set customer demand analysis model analyzes the multi-modal demand dataset based on a machine self-learning algorithm to generate a preliminary vehicle usage demand category of the customer; Analyze the preliminary vehicle usage demand category based on pre-set preliminary privacy protection rules to generate a preliminary privacy protection strategy, which is used to protect the customer's preliminary application data; Obtain the customer's real-time driving behavior data, where the real-time driving behavior data includes the steering wheel rotation frequency and the corresponding rotation angle, and the switching frequency data of the drive pedal and the brake pedal; Obtain the environmental perception data of driving, where the environmental perception data includes visual monitoring data and radar data; Obtain the iris tracking data of the people in the vehicle, and when the in-vehicle camera detects the iris feature, generate an adversarial noise pattern to cover the sensitive area; Associate the real-time driving behavior data, environmental perception data, and iris tracking data based on a pre-set common timeline to construct an actual driving state dataset; A pre-set driving state analysis model analyzes the real-time driving state data based on a machine self-learning algorithm to generate driving state correction data, which is used to correct the preliminary vehicle usage demand category; A pre-set actual vehicle usage demand analysis model analyzes the driving state correction data and the preliminary vehicle usage demand category based on a machine self-learning algorithm to generate actual vehicle usage demand data; Analyze the actual vehicle usage demand data based on pre-set privacy protection rules to generate a corrected privacy protection strategy, which is used to protect the customer's real-time driving state data.

2. The method for processing platform privacy data based on customer vehicle usage requirements according to claim 1, characterized in that In the step where a pre-set customer demand analysis model analyzes the multi-modal demand dataset based on a machine self-learning algorithm to generate a preliminary vehicle usage demand category of the customer, the steps include as follows: Obtain each demand characteristic in the multi-modal demand dataset, where the demand characteristic includes a driving mode parameter, a trip purpose, and a passenger-carrying status; A pre-set customer demand portrait construction model analyzes the multi-modal demand dataset based on a machine self-learning algorithm to construct a customer demand portrait; Perform a weighted score on the customer demand portrait based on a pre-set portrait scoring model to calculate the comprehensive demand score of the customer; Generate a corresponding preliminary vehicle usage demand category based on the comprehensive demand score of the customer, where the preliminary vehicle usage demand category includes tourism travel, business reception, shared car rental, and emergency rescue.

3. A method for processing platform privacy data based on customer vehicle usage requirements according to claim 2, characterized in that, After the step where a pre-set customer demand portrait construction model analyzes the multi-modal demand dataset based on a machine self-learning algorithm to construct a customer demand portrait, the steps include as follows: Construct a vertical federated model across service providers to align the privacy of different user portrait features; Specifically, identify the overlapping users in each service provider's dataset as the common samples for federated training, and extract the detailed personal data of the overlapping users as the initial common sample dataset; Establish a logical mapping relationship for feature fields across service providers, and process the initial common sample data set based on the logical mapping relationship to construct a standard common sample data set; Calculate the cross-data features of the data with the mapping relationship through the homomorphic encryption algorithm to generate a joint feature vector; Perform user ID preprocessing on the user identifiers of different service providers in the standard common sample data set to unify them into hash values, thereby eliminating format differences.

4. A method for processing platform privacy data based on customer vehicle usage requirements according to claim 3, characterized in that, After the step of performing user ID preprocessing on the user identifiers of different service providers in the standard common sample data set to unify them into hash values, the following steps are included: Add Laplace noise to the feature bin boundaries of the standard common sample data set to obfuscate the statistical distribution; Each service provider locally calculates the feature Shapley values of the standard common sample data set, and obtains the global importance ranking corresponding to each data in the standard common sample data through the secure aggregation algorithm, thereby preventing each service provider from directly sharing feature data to prevent reverse inference of user behavior.

5. A method for processing platform privacy data based on customer vehicle usage requirements according to claim 1, characterized in that, After the step of analyzing the driving state correction data and the preliminary vehicle usage demand categories based on a machine learning algorithm in a pre-set actual vehicle usage demand analysis model to generate actual vehicle usage demand data, the following steps are included: Compare the actual driving behavior and the predicted value through the federated learning algorithm to generate a driving deviation degree; Evaluate the driving deviation degree based on a machine learning algorithm to generate a corresponding abnormal behavior index; Analyze the environment perception data based on a visual recognition algorithm to dynamically generate an environment threat coefficient; A pre-set driving dynamic risk assessment model comprehensively analyzes the abnormal behavior index and the environment threat coefficient to generate a dynamic risk value of the customer, and the dynamic risk value is used to dynamically evaluate the priority level of privacy protection.

6. A method for processing platform privacy data based on customer vehicle usage requirements according to claim 5, characterized in that, After the step of analyzing the actual vehicle usage demand data based on a pre-set privacy protection rule to generate a corrected privacy protection policy, the following steps are included: Obtain the service feedback data of customer privacy protection, and associate the actual vehicle usage demand data, the corrected privacy protection policy and the service feedback data based on the public time axis to construct a privacy service feedback related data set; A pre-set privacy service adjustment model analyzes the privacy service feedback related data set based on a machine learning algorithm to generate a privacy leakage-service degradation related coefficient, and the privacy leakage-service degradation related coefficient is used to adjust the weighted value of the dynamic risk value calculation.

7. A method for processing platform privacy data based on customer vehicle usage requirements according to claim 3, characterized in that, After the step of constructing a vertical federated model across service providers and aligning the privacy of different user portrait features, the following steps are included: Obtain the real-time survival status of the service provider. If the survival status of the service provider has expired, automatically destroy the historical session key and block the decryption ability of the real-time standard common sample data set.

8. A platform privacy data processing device based on customer vehicle usage requirements, applied to the method for processing platform privacy data based on customer vehicle usage requirements according to any one of claims 1 to 7, characterized in that, The device includes: A multi-modal demand data set construction unit (1) for obtaining the preliminary application data when the customer submits a vehicle usage demand, and extracting demand features from the preliminary application data to construct a multi-modal demand data set of the customer; The preliminary vehicle usage requirement category generation unit (2) is configured to preset a customer requirement analysis model to analyze the multi-modal requirement data set based on a machine self-learning algorithm to generate the preliminary vehicle usage requirement category of the customer; The preliminary privacy protection policy generation unit (3) is configured to analyze the preliminary vehicle usage requirement category based on a preset preliminary privacy protection rule to generate a preliminary privacy protection policy, and the preliminary privacy protection policy is used to protect the preliminary application data of the customer; The real-time driving behavior data acquisition unit (4) is configured to acquire the real-time driving behavior data of the customer; The environment perception data acquisition unit (5) is configured to acquire the environment perception data of driving; The iris tracking data acquisition unit (6) is configured to acquire the iris tracking data of the vehicle occupants, and generate an adversarial noise pattern to cover the sensitive area when the in-vehicle camera detects the iris feature; The actual driving state data set construction unit (7) is configured to associate the real-time driving behavior data, the environment perception data, and the iris tracking data based on a preset common time axis to construct an actual driving state data set; The driving state correction data generation unit (8) is configured to preset a driving state analysis model to analyze the real-time driving state data based on a machine self-learning algorithm to generate driving state correction data; The actual vehicle usage requirement data generation unit (9) is configured to preset an actual vehicle usage requirement analysis model to analyze the driving state correction data and the preliminary vehicle usage requirement category based on a machine self-learning algorithm to generate actual vehicle usage requirement data; The corrected privacy protection policy generation unit (10) is configured to analyze the actual vehicle usage requirement data based on a preset privacy protection rule to generate a corrected privacy protection policy, and the corrected privacy protection policy is used to protect the real-time driving state data of the customer.

9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, the steps of a platform privacy data processing method based on customer vehicle usage requirements as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, the steps of a platform privacy data processing method based on customer vehicle usage requirements as described in any one of claims 1 to 7 are implemented.

Images