Circulation management method and system for car owner points and rights

Abstract

The present invention relates to a method and system for managing the circulation of car owner points and rights, comprising the following steps: collecting driving behavior data of a target car owner through an intelligent terminal device deployed in a vehicle of the target car owner to obtain behavior collection data; performing data desensitization processing on the behavior collection data to obtain behavior desensitization data; performing behavior pattern recognition on the target car owner based on the behavior desensitization data through a preset recognition algorithm to obtain a driving behavior recognition pattern; calculating the points value of the target car owner based on the driving behavior recognition pattern through a preset point rule in a smart contract to obtain a point value calculation result; calling a point-rights relationship exchange table stored in a database, performing rights and interests matching on the point-rights relationship exchange table based on the point value calculation result to obtain a recommended rights and interests combination plan, thereby solving the technical problem of how to improve driving safety and optimize driving experience in the automotive industry.

Description

Technical Field

[0001] The present invention relates to the technical field of information management, and in particular to a method and system for managing the circulation of vehicle owner points and rights. Background Art

[0002] With the continuous increase in the number of cars, how to improve driving safety and optimize the driving experience has become the focus of the automotive industry. In this context, the circulation management system of car owner points and rights came into being, aiming to encourage good driving habits and correct bad driving behaviors through the data management of car owners' driving behaviors. However, traditional management methods often rely on manual review and simple reward mechanisms, which lack scientificity and real-timeness, and it is difficult to effectively motivate car owners to form good driving habits. In addition, traditional methods are not sufficient to protect personal privacy and cannot meet the high requirements of modern society for personal information security.

[0003] The development of modern technology, especially the application of big data analysis, artificial intelligence and blockchain technology, has provided new ideas for solving the above problems. By collecting driving behavior data through smart terminal devices and processing them with advanced data analysis technology, the quality of driving behavior can be evaluated more accurately. At the same time, the use of smart contracts to automatically execute the points rules not only improves efficiency, but also enhances the transparency and fairness of the system. However, how to ensure the secure transmission and storage of data and prevent the leakage of sensitive information has become one of the technical difficulties that must be overcome in the implementation of such systems.

[0004] In addition, although the original intention of the points and rights transfer management system is good, some challenges may be encountered in actual application. For example, different car owners have different needs for points exchange rights. How to design a point-rights exchange table that is both fair and can meet the needs of most car owners is an issue that needs in-depth research. In addition, due to the complexity and changeability of driving environments and conditions, how to make the system adapt to driving behavior recognition in various situations and ensure the accuracy of points calculation is also a focus of current research. In short, despite many challenges, through continuous technological innovation and improvement, the owner points and rights transfer management system is expected to make important contributions to improving road safety in the future. Summary of the invention

[0005] The main purpose of the present invention is to provide a method, device, equipment and storage medium for the circulation management of car owner points and rights, which solves the technical problem of how to improve driving safety and optimize driving experience in the automotive industry.

[0006] To achieve the above-mentioned purpose, the present invention provides a method for managing the circulation of car owner points and rights, comprising the following steps:

[0007] The driving behavior data of the target vehicle owner is collected by using an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain the behavior collection data;

[0008] Performing data desensitization processing on the behavior collection data to obtain behavior desensitized data;

[0009] Using a preset recognition algorithm, based on the behavior desensitization data, the behavior pattern of the target vehicle owner is recognized to obtain a driving behavior recognition pattern;

[0010] By using the points rules preset in the smart contract, the points value of the target car owner is calculated based on the driving behavior recognition mode to obtain the points value calculation result;

[0011] The point-rights exchange table stored in the database is called, and the point-rights exchange table is matched with the right based on the point value calculation result to obtain a recommended right combination plan.

[0012] Furthermore, the driving behavior data of the target vehicle owner is collected by the intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain the behavior collection data, including:

[0013] The driving behavior data of the target vehicle owner is collected by using an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain the original driving behavior data;

[0014] Preprocessing the original driving behavior data to obtain pure vehicle operation status data;

[0015] Performing multi-dimensional feature extraction on the pure vehicle operation status data to obtain multi-dimensional driving behavior features;

[0016] The multi-dimensional driving behavior characteristics are packaged in the preset data format to obtain behavior collection data.

[0017] Furthermore, the intelligent terminal device is provided with a data desensitization module, and the data desensitization processing is performed on the behavior collection data to obtain behavior desensitized data, including:

[0018] The sensitive attributes of the behavior collection data are disturbed by the data desensitization module to obtain preliminary desensitized data;

[0019] Performing grouping and generalization processing on the preliminary desensitized data to obtain generalized desensitized data;

[0020] Encrypting key fields in the generalized desensitized data to obtain encrypted desensitized data;

[0021] Performing partial data replacement masking on non-critical sensitive information in the generalized desensitized data to obtain partially masked data;

[0022] Combining the encrypted desensitized data and the partially masked data to obtain combined data;

[0023] Performing distributed computing on the combined data to obtain decentralized data;

[0024] The decentralized data is desensitized and verified through a zero-knowledge proof protocol. If the desensitization verification is successful, behavioral desensitized data is obtained.

[0025] Furthermore, the behavior pattern recognition of the target vehicle owner is performed based on the behavior desensitization data by using a preset recognition algorithm to obtain a driving behavior recognition pattern, including:

[0026] By using a preset recognition algorithm, the features in the behavior-massified data that are related to the points-rights exchange table are identified to obtain relevant recognition features; wherein the relevant recognition features include braking frequency features, acceleration features, and speeding frequency features;

[0027] Performing time series segmentation on the relevant identification features to obtain time series segmentation features; wherein the time series segmentation features are time series segmentation features with time tags;

[0028] Performing feature dimensionality reduction on the time series segment features to obtain features after dimensionality reduction;

[0029] Performing feature clustering analysis on the dimension-reduced features to obtain driving behavior pattern labels;

[0030] The driving behavior pattern label is matched with a pattern in a preset driving behavior pattern library to obtain a driving behavior recognition pattern; wherein the driving behavior recognition pattern includes a safe driving mode, an aggressive driving mode and a mixed driving mode.

[0031] Furthermore, the integral value calculation result is obtained by calculating the integral value of the target vehicle owner based on the driving behavior recognition mode according to the integral rules preset in the smart contract, including:

[0032] Reading a preset scoring rule through the smart contract to obtain a reading scoring rule, wherein the reading scoring rule includes a scoring weight and a scoring threshold corresponding to different driving behavior recognition modes;

[0033] Performing integral weight mapping on different driving behavior recognition modes by reading the integral rule to obtain the integral weight of the corresponding driving behavior recognition mode;

[0034] Performing duration statistics on the driving behavior recognition patterns to obtain the duration of each driving behavior recognition pattern;

[0035] Calculating an integral value corresponding to the driving behavior recognition mode based on the integral weight, the integral threshold and the duration;

[0036] Accumulating and / or subtracting the integral values ​​of all the driving behavior recognition modes to obtain a preliminary total integral value;

[0037] The preliminary total integral value is verified for correctness to obtain a verification result. If the verification result is correct, the preliminary total integral value is used as the integral value calculation result.

[0038] Furthermore, the equity matching of the point value calculation result based on the point-equity exchange table to obtain a recommended equity combination plan includes:

[0039] The historical redemption records and the demand preferences of the target car owner are obtained, and the integral value calculation result is represented in multiple dimensions based on the historical redemption records by a restricted Boltzmann machine in a preset deep belief network to obtain an integral feature representation vector; wherein the row labels in the integral feature representation vector are the weekend usage ratio, the weekday usage ratio and the target car owner's loyalty;

[0040] Constructing a points-equity matrix based on the points-equity exchange table; wherein the rows of the points-equity matrix represent different exchange points, and the columns of the points-equity matrix represent the weekend usage ratio, the weekday usage ratio and the target car owner loyalty;

[0041] Using a cosine similarity algorithm, similarity is calculated between the integral feature representation vector and the vector in the integral-equity matrix to obtain a similarity score vector;

[0042] Performing label mapping on the similarity score vector to obtain a mapped equity label vector;

[0043] Weighting the elements in the mapped equity tag vector based on the demand preference to obtain a mapped equity tag weighted vector;

[0044] Using a conditional random field, the elements in the mapped equity label weighted vector are sorted and labeled in descending order to obtain a sorted and labeled equity label vector;

[0045] When the calculated integral value is greater than a preset redemption threshold, rights matching is performed in the integral-rights relationship redemption table based on the sorted and labeled rights label vector to obtain a recommended rights combination plan; wherein the preset redemption threshold is the sum of the redemption points of each right in the recommended rights combination plan.

[0046] Furthermore, the cosine similarity algorithm is used to calculate the similarity between the integral feature representation vector and the vector in the integral-equity matrix to obtain a similarity score vector, including:

[0047] Using the cosine similarity algorithm, the integral feature representation vector is multiplied by each element in the integral-equity matrix to obtain a preliminary similarity vector; wherein the number of the preliminary similarity vectors is the number of rows of the integral-equity matrix;

[0048] Calculate the modulus of each row in the score-equity matrix to obtain a vector modulus length; wherein the number of the vector modulus lengths is equal to the number of the preliminary similarity vectors;

[0049] Based on the elements in each of the preliminary similarity vectors, normalizing the corresponding vector modulus lengths to obtain a normalized similarity value vector;

[0050] Performing modulus calculation on the integral feature representation vector to obtain the integral representation modulus length;

[0051] Dividing the elements in the normalized similarity value vector by the integral representation modulus length to obtain a preliminary similarity score vector;

[0052] A nonlinear transformation is performed on the similarity score vector to obtain a similarity score vector.

[0053] The present invention also provides a system for the circulation and management of car owner points and rights, including:

[0054] A collection module, used to collect driving behavior data of the target vehicle owner through an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain behavior collection data;

[0055] A desensitization module is used to perform data desensitization processing on the behavior collection data to obtain behavior desensitized data;

[0056] An identification module, used to identify the behavior pattern of the target vehicle owner based on the behavior desensitization data by using a preset identification algorithm to obtain a driving behavior identification pattern;

[0057] A calculation module, used to calculate the integral value of the target vehicle owner based on the driving behavior recognition mode according to the integral rules preset in the smart contract, and obtain the integral value calculation result;

[0058] The matching module is used to call the point-rights relationship exchange table stored in the database, perform rights matching on the point-rights relationship exchange table based on the point value calculation result, and obtain a recommended rights combination plan.

[0059] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0060] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0061] The method for managing the circulation of car owner points and rights provided by the present invention comprises the following steps: collecting driving behavior data of the target car owner through an intelligent terminal device deployed in the vehicle of the target car owner to obtain behavior collection data; performing data desensitization processing on the behavior collection data to obtain behavior desensitization data; performing behavior pattern recognition on the target car owner based on the behavior desensitization data through a preset recognition algorithm to obtain a driving behavior recognition pattern; calculating the point value of the target car owner based on the driving behavior recognition pattern through the point rules preset in the smart contract to obtain a point value calculation result; calling the point-rights relationship exchange table stored in the database, matching the points-rights relationship exchange table based on the point value calculation result to obtain a recommended right combination plan. Through the above-mentioned technical means, the technical problem of how to improve driving safety and optimize driving experience in the automotive industry is solved, and a positive incentive mechanism of points and rights is achieved to encourage car owners to improve their driving behavior, which helps to reduce the incidence of traffic accidents and improve the overall road traffic safety level in the long run. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic diagram of the steps of a method for managing the circulation of car owner points and rights in one embodiment of the present invention;

[0063] Figure 2 It is a structural block diagram of a circulation management system for car owner points and rights in one embodiment of the present invention;

[0064] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0065] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0067] like Figure 1 As shown, Figure 1 It is a schematic diagram of the steps of a method for managing the circulation of car owner points and rights in one embodiment of the present invention;

[0068] In one embodiment of the present invention, a method for managing the circulation of car owner points and rights is provided, comprising the following steps:

[0069] Step S1, collecting driving behavior data of the target vehicle owner through an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain behavior collection data.

[0070] Specifically, the step of "collecting driving behavior data of the target car owner through the intelligent terminal device deployed in the target car owner's vehicle to obtain behavior collection data" mentioned above specifically refers to installing specific intelligent terminal devices inside the target car owner's vehicle. These devices usually include but are not limited to on-board diagnostic system (OBD) interface, GPS positioning module, acceleration sensor and other components. These intelligent terminal devices can monitor and record the vehicle operation status and the owner's driving behavior in real time, such as key parameters such as vehicle speed, acceleration change, braking frequency, driving route, etc. Take a specific scenario as an example. When a car owner drives a vehicle equipped with such intelligent terminal devices, if he frequently accelerates or brakes suddenly during driving, these behaviors will be accurately captured and recorded by the intelligent terminal device to form the original behavior collection data. Subsequently, these data will be uploaded to the cloud server for further processing and analysis. In this way, it can not only help car owners understand their driving habits, but also provide necessary basic data support for subsequent steps such as data desensitization, behavior pattern recognition and integral value calculation. In addition, the design of this smart terminal device also takes into account compatibility and ease of use, ensuring that it can be installed and used on vehicles of different brands and models, thereby achieving wide coverage of a large group of car owners.

[0071] Step S2, performing data desensitization processing on the behavior collection data to obtain behavior desensitized data.

[0072] Specifically, the process of "desensitizing the behavior collection data to obtain desensitized behavior data" mentioned above means that after the intelligent terminal device collects the driving behavior data of the target car owner, in order to protect the personal privacy of the car owner, the system will take a series of technical means to process these original data, remove or replace the information that may directly or indirectly point to the personal identity. For example, in the above-mentioned driving behavior data collection process, the system records the car owner's driving route, speed change and other information. This information itself does not contain the owner's identity information, but if combined with a specific time and place, it is possible to infer the owner's personal information. Therefore, in the data desensitization processing stage, the system will anonymize sensitive fields such as timestamps and geographic locations, such as converting specific time and location information into relative values ​​or fuzzy processing, to ensure that even if the data is illegally obtained, it cannot be traced back to a specific individual. In this way, the value of the data for analysis is retained, and the privacy of the car owner is effectively protected, laying a solid foundation for the subsequent driving behavior pattern recognition and points calculation.

[0073] Step S3, using a preset recognition algorithm, based on the behavior desensitization data, the behavior pattern of the target vehicle owner is recognized to obtain a driving behavior recognition pattern.

[0074] Specifically, the step of "using a preset recognition algorithm to identify the behavior pattern of the target car owner based on the behavior desensitization data to obtain a driving behavior recognition pattern" mentioned above means that after the data desensitization process is completed, the system will use a pre-set algorithm model to analyze the processed behavior data, and then identify the driving behavior pattern of the target car owner. For example, based on previously collected and desensitized data, such as vehicle speed changes, acceleration, braking frequency, driving route and other information, the system can train a model that can accurately judge the type of driving behavior through machine learning algorithms, such as decision trees, random forests or neural networks. In practical applications, suppose that a car owner often accelerates and brakes suddenly during daily driving. After data desensitization, these behaviors will be input into the preset recognition algorithm for analysis. Through the operation of the algorithm, the system can identify that the driving style of the car owner tends to be aggressive, and thus classify it as a specific driving behavior recognition pattern. This pattern not only reflects the actual driving habits of the car owner, but also provides an important basis for the subsequent calculation of the integral value. In addition, through continuous data accumulation and algorithm optimization, the system can continuously improve recognition accuracy and efficiency, better serve the circulation management of car owners' points and rights, and promote the formation of good driving habits.

[0075] Step S4, calculating the integral value of the target vehicle owner based on the driving behavior recognition pattern through the integral rules preset in the smart contract, and obtaining the integral value calculation result.

[0076] Specifically, the process mentioned above of "calculating the score value of the target car owner based on the driving behavior recognition mode through the preset score rules in the smart contract to obtain the score value calculation result" means that after completing the driving behavior pattern recognition, the system will use the pre-defined score rules in the smart contract to calculate the score value for the different identified driving behavior patterns. Specifically, a smart contract is a computer program that automatically executes the terms of the contract, which can ensure the transparency and consistency of the execution of the score rules. For example, in the driving behavior recognition mode mentioned above, if the system recognizes that a certain car owner's driving style tends to be aggressive, that is, sudden acceleration and sudden braking often occur, then according to the preset score rules in the smart contract, this type of driving behavior may be deducted a certain number of points, and conversely, if it is a smooth driving, it may be rewarded with points. In this way, by automatically executing these score rules through smart contracts, it can be ensured that each driving behavior mode corresponds to a clear score value, and finally form a score value calculation result. This process not only improves the efficiency and accuracy of the points calculation, but also increases the credibility and transparency of the entire points system, so that every car owner can clearly understand the changes in their points and the reasons behind them, thereby motivating car owners to improve their driving behavior and cultivate safer driving habits. At the same time, the application of smart contracts also simplifies the management process, reduces operating costs, and provides strong technical support for the efficient operation of the car owner points and equity transfer management system.

[0077] Step S5, calling the point-rights relationship exchange table stored in the database, performing rights matching on the point-rights relationship exchange table based on the point value calculation result, and obtaining a recommended rights combination plan.

[0078] Specifically, the process of "calling the point-rights exchange table stored in the database, matching the points-rights exchange table with the points-rights exchange table based on the point value calculation result, and obtaining the recommended right combination plan" mentioned above specifically refers to that after completing the point value calculation of the driving behavior, the system will access a pre-built point-rights exchange table, which is stored in the database and records the mapping relationship between different point value intervals and corresponding rights. For example, assuming that the system has calculated that the point value of a certain car owner is 1000 points according to the point rules in the smart contract, the system will query the point-rights exchange table to find the right option corresponding to 1000 points. In the point-rights exchange table, multiple point segments may be set, and each point segment is associated with a set of redeemable rights, such as free car wash service, refueling discount, maintenance discount, etc. Continuing with the above-mentioned car owner as an example, if 1000 points are in the point segment of 800 to 1200 points, the system will select the right combination that best meets the needs of the car owner from all the rights associated with the point segment as the recommended solution. This recommendation may be automatically selected by the system based on the owner's historical preferences, or it may be selected by the owner himself. In this way, not only the points are effectively utilized, but also a variety of service options are provided to the owners, enhancing their satisfaction and loyalty. At the same time, the design of the points-rights exchange table also takes flexibility into account. It can be dynamically updated according to market changes and service provider strategy adjustments to ensure that the recommended rights and interests combination is always close to the actual needs of the owner, thereby promoting the continuous optimization and development of the owner's points and rights and interests circulation management system.

[0079] In a specific embodiment, the driving behavior data of the target vehicle owner is collected by a smart terminal device deployed in the vehicle of the target vehicle owner to obtain the behavior collection data, including:

[0080] The driving behavior data of the target vehicle owner is collected by using an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain the original driving behavior data;

[0081] Preprocessing the original driving behavior data to obtain pure vehicle operation status data;

[0082] Performing multi-dimensional feature extraction on the pure vehicle operation status data to obtain multi-dimensional driving behavior features;

[0083] The multi-dimensional driving behavior characteristics are packaged in the preset data format to obtain behavior collection data.

[0084] Specifically, this series of operations begins with the deployment of smart terminal devices. Taking an actual application scenario as an example, suppose a car owner installs a smart terminal device that integrates multiple sensors such as GPS positioning module, acceleration sensor, gyroscope, etc. in his vehicle. When the vehicle is started, these sensors start to work, monitoring and recording the vehicle's position, speed, acceleration, direction change and other information in real time, which constitutes the original driving behavior data. However, these raw data often contain a lot of noise and unnecessary information, such as slight acceleration changes caused by vehicle vibration. These changes have no substantive significance for understanding driving behavior, but will increase the complexity of subsequent data processing. Therefore, it is necessary to preprocess the original driving behavior data, remove noise, fill missing values, standardize the data format, and finally obtain pure vehicle operation status data. For example, the system may use filtering algorithms to eliminate random fluctuations in sensor readings, or use interpolation methods to fill data gaps caused by signal loss to ensure data integrity and consistency. Next, in order to understand the owner's driving behavior more deeply, the system will perform multi-dimensional feature extraction on the pure vehicle operation status data. This process involves mining key indicators that reflect driving behavior from the data, such as average speed, maximum acceleration, number of emergency brakes, speeding ratio, etc. These features can fully and accurately describe the driving habits of the car owner and provide a basis for subsequent behavioral pattern recognition. For example, the system may calculate that the average speed of the car owner in a trip is 60km / h, the number of emergency brakes is 3 times, and the speeding ratio is 5%. These features together constitute the multi-dimensional driving behavior characteristics. Finally, in order to facilitate data storage, transmission and analysis, the multi-dimensional driving behavior characteristics need to be packaged in a preset data format to form the final behavior collection data. The preset data format usually includes metadata such as data type, unit, and timestamp, which ensures the structuring and standardization of the data and is conducive to subsequent data processing and application. For example, the system may package the above-mentioned average speed, number of emergency brakes, speeding ratio and other features into a JSON format data packet. Each data packet not only contains specific values, but also comes with the corresponding unit and timestamp, which is convenient for the system to quickly parse and use. To sum up, from the collection of raw driving behavior data to the generation of behavior collection data, the whole process involves multiple links such as data collection, preprocessing, feature extraction and data encapsulation. Each link is crucial and together ensures the quality and availability of the data, laying a solid foundation for achieving accurate driving behavior pattern recognition and effective points management.

[0085] In a specific embodiment, the smart terminal device is provided with a data desensitization module, and the data desensitization processing is performed on the behavior collection data to obtain behavior desensitized data, including:

[0086] The sensitive attributes of the behavior collection data are disturbed by the data desensitization module to obtain preliminary desensitized data;

[0087] Performing grouping and generalization processing on the preliminary desensitized data to obtain generalized desensitized data;

[0088] Encrypting the key fields in the generalized desensitized data to obtain encrypted desensitized data;

[0089] Performing partial data replacement masking on non-critical sensitive information in the generalized desensitized data to obtain partially masked data;

[0090] Combining the encrypted desensitized data and the partially masked data to obtain combined data;

[0091] Performing distributed computing on the combined data to obtain decentralized data;

[0092] The decentralized data is desensitized and verified through a zero-knowledge proof protocol. If the desensitization verification is successful, behavioral desensitized data is obtained.

[0093] Specifically, first, the data desensitization module perturbs the sensitive attributes of the behavior collection data to obtain the preliminary desensitized data. Sensitive attribute perturbation means making minor modifications or adding random noise to the information in the data that may directly or indirectly point to personal identity without affecting the data analysis results. For example, suppose the system records the driving trajectory of a car owner at a specific time and location. The data desensitization module may slightly adjust the specific time and location information, such as moving the exact time point forward or backward by a few minutes, or slightly offsetting the geographical location coordinates by a few meters. In this way, even if the data is leaked, it is impossible to accurately track a specific individual. Next, the preliminary desensitized data is grouped and generalized to obtain the generalized desensitized data. Grouping and generalization means classifying similar data items into the same group and then replacing the specific values with the representative values of the group. For example, the system can group the driving trajectories of car owners by region and represent the driving trajectories of all car owners in the same region with the center point of the region instead of showing the specific driving routes. This not only reduces the detailed information of the data but also further reduces the risk of reverse-deriving personal identity from the data. Then, the keyword fields in the generalized desensitized data are encrypted to obtain the encrypted desensitized data. Data encryption is to convert the data into an unreadable form through cryptographic algorithms, and only those with the correct key can decrypt and restore the data. For example, the system can use the AES (Advanced Encryption Standard) algorithm to encrypt the keyword fields such as the license plate number and mobile phone number of the car owner to ensure that even if the data is intercepted during transmission or storage, it cannot be illegally read. Specifically, suppose the license plate number is "沪A12345" and the mobile phone number is "13812345678". The system will encrypt these keyword fields through the AES algorithm to generate ciphertext. Some data in the generalized desensitized data that is non-critical sensitive information is replaced with masks to obtain the data after partial masking. Partial data replacement masking means replacing some parts of the original data with specific characters or symbols to make it lose its original meaning. For example, the system can replace some characters in the car owner's name with asterisks (*), such as replacing "张三" with "张*". In this way, even if the data is leaked, it is impossible to directly read the complete personal information. Suppose the car owner's name is "张三", and the system can partially mask it as "张*". The encrypted desensitized data and the data after partial masking are combined to obtain the combined data. Data combination means merging the encrypted keyword fields and the non-critical sensitive information after partial masking together to form a new data set. For example, suppose the license plate number "沪A12345" and the mobile phone number "13812345678" have been encrypted, and the name "张三" has been partially masked as "张*". The system will combine these data together to generate the combined data. In this way, even if the data is leaked, it is impossible to directly read the complete personal information, but the usability of the data is still retained. The combined data is subjected to distributed computing to obtain the decentralized data.Distributed computing refers to processing data by dispersing it across multiple nodes, where each node only processes a part of the data, thus avoiding the security risks brought by centralized data storage. For example, the system can split combined data into multiple segments and store them on different servers respectively. Each server is only responsible for processing and storing a part of the data, rather than the complete dataset. In this way, even if a certain node is attacked, it is impossible to obtain the complete data. Finally, the decentralized data is desensitized and verified through the zero-knowledge proof protocol. If the desensitization verification is successful, behaviorally desensitized data is obtained. Zero-knowledge proof is a cryptographic technique that can verify the truth of a statement without revealing any specific information. For example, the system can use the zero-knowledge proof protocol to verify whether the decentralized data has been correctly desensitized without exposing the specific content of the data. If the verification is successful, the system will confirm that the data has achieved the expected desensitization effect and can be safely used for subsequent analysis and processing. Through this series of steps, the system can comprehensively desensitize the behavior collection data, ensuring that the data still has the value of analysis and use while protecting the privacy of car owners. For example, in practical applications, assume that the license plate number of a certain car owner is "沪A12345", the mobile phone number is "13812345678", and the name is "张三". The behaviorally desensitized data generated by the system through the above steps can be used for subsequent driving behavior pattern recognition and point calculation without disclosing the specific identity information of the car owner. In this way, the system not only protects the privacy of car owners but also ensures the security and availability of the data, providing a reliable basis for the transfer management of car owner points and rights and interests.

[0094] In a specific embodiment, encrypting the key fields in the generalized desensitized data to obtain encrypted desensitized data includes:

[0095] Performing standardization processing on the key fields in the generalized desensitized data to obtain standardized field data;

[0096] Performing salt value splicing on the standardized field data to obtain salted field data;

[0097] Converting the salted field data into binary format to obtain binary field data;

[0098] Performing block processing on the binary field data to obtain a data block sequence;

[0099] Using a preset SHA-256 core algorithm to perform block-by-block hash calculation on the data block sequence to obtain an intermediate hash value;

[0100] Converting the intermediate hash value into hexadecimal representation to obtain hash mapping data:

[0101] The key field is encrypted using the X-mapped data to obtain encrypted desensitized data.

[0102] Specifically, first, standardize the keyword fields in the generalized desensitized data to obtain standardized field data. Standardization means unifying the data formats of keyword fields to ensure data consistency and processability. For example, assume a car owner's license plate number is "沪A12345" and mobile phone number is "13812345678". After standardization, the license plate number and mobile phone number will be unified into a specific format, such as the license plate number being unified to "沪A-XXXXX" and the mobile phone number being unified to "+86-XXXXXXXXXXX". Next, perform salt value concatenation on the standardized field data to obtain salted field data. Salt value concatenation is adding a random salt value to the standardized field data to increase data randomness and security. For example, assume the salt value for the license plate number "沪A-12345" is "salt123", then the salted field data is "沪A-12345salt123". The addition of the salt value can prevent the same input data from generating the same hash value, thereby enhancing data security. Then, convert the salted field data into binary format to obtain binary field data. Binary conversion is converting the salted field data from string form to binary form for subsequent hash calculation. For example, the salted field data "沪A-12345salt123" will be converted into binary format, such as "01010101 01010101...". Perform block processing on the binary field data to obtain a sequence of data blocks. Block processing is dividing the binary field data into fixed-size data blocks for block-by-block hash calculation. For example, assume the binary field data is "01010101 01010101...", it can be divided into multiple 512-bit data blocks, such as "01010101 01010101..." and "01010101 01010101...". Use the preset SHA-256 core algorithm to perform block-by-block hash calculation on the sequence of data blocks to obtain intermediate hash values. SHA-256 is a commonly used hash algorithm that can convert input data of any length into a hash value of a fixed length. The system will perform hash calculation on the sequence of data blocks block by block to generate intermediate hash values. For example, perform SHA-256 hash calculation on the first data block "0101010101010101..." to obtain the intermediate hash value "a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6q7r8s9t0u1v2w3x4y5z6". Convert the intermediate hash value into hexadecimal representation to obtain hash mapping data. Hexadecimal representation is for convenient storage and transmission, converting the intermediate hash value into a hexadecimal string.For example, the intermediate hash value "a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6q7r8s9t0u1v2w3x4y5z6" will be converted into the hexadecimal representation "a1b2c3d4e5f6789012345678901234567890123456789012345678901234". Additionally, the standard SHA-256 algorithm outputs a 256-bit (32-byte) hash value, generating a 64-character string in hexadecimal representation. This hash-mapped data is used as a key and further processed through a symmetric encryption algorithm (such as AES-256) or a key derivation function for subsequent encryption operations. The hash-mapped data has irreversibility and high discreteness. Even if the original data is similar, it will generate completely different hash values, thus ensuring the security and uniqueness of data encryption. In practical applications, the system combines this hash-mapped data with preset key materials to form the final encryption key for the protective transformation of key fields, ensuring the confidentiality of data during storage and transmission. Finally, the key fields are encrypted using the hash-mapped data to obtain encrypted and de-identified data. Data encryption combines the hash-mapped data with the key fields to generate the final encrypted and de-identified data. For example, assuming the hash-mapped data is "a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6q7r8s9t0u1v2w3x4y5z6", the system can combine this hash-mapped data with the license plate number "沪A-12345" to generate the encrypted and de-identified data "a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6q7r8s9t0u1v2w3x4y5z6-沪A-12345". Through this series of steps, the system can securely encrypt the key fields in the generalized and de-identified data to generate encrypted and de-identified data. This process not only ensures the privacy and security of the key fields but also retains the integrity and availability of the data, providing a reliable basis for subsequent behavior pattern recognition and point calculation. For example, in practical applications, assuming a car owner's license plate number is "沪A12345" and the mobile phone number is "13812345678", the encrypted and de-identified data generated by the system through the above steps can be used for subsequent point calculation and benefit matching without revealing the specific identity information of the car owner.

[0103] In a specific embodiment, the behavior pattern of the target car owner is recognized based on the behavior de-identified data through a preset recognition algorithm to obtain a driving behavior recognition pattern, including:

[0104] Through a preset recognition algorithm, the features related to the integral-benefit relationship conversion table in the behavior de-identified data are recognized to obtain relevant recognition features; wherein, the relevant recognition features include braking frequency features, acceleration features, and speeding frequency features;

[0105] Performing time series segmentation on the relevant identification features to obtain time series segmentation features; wherein the time series segmentation features are time series segmentation features with time tags;

[0106] Performing feature dimensionality reduction on the time series segment features to obtain features after dimensionality reduction;

[0107] Performing feature clustering analysis on the dimension-reduced features to obtain driving behavior pattern labels;

[0108] The driving behavior pattern label is matched with a pattern in a preset driving behavior pattern library to obtain a driving behavior recognition pattern; wherein the driving behavior recognition pattern includes a safe driving mode, an aggressive driving mode and a mixed driving mode.

[0109] Specifically, first, through the preset recognition algorithm, the system will identify the features related to the points-rights relationship exchange table in the behavior desensitized data, and obtain the relevant recognition features. It should be noted that these relevant recognition features are data that have not been partially masked; these relevant recognition features mainly include braking frequency features, acceleration features and speeding frequency features. For example, suppose that the system has collected and desensitized the driving data of a certain car owner, which includes the time of each braking, the change in acceleration and the number of speeding. The recognition algorithm will extract key features such as braking frequency, acceleration change and speeding frequency from these data as the basis for subsequent analysis. Next, these relevant recognition features are time-series segmented to obtain time-series segmentation features. Time-series segmentation refers to dividing continuous driving data into multiple time periods according to time intervals, and the data in each time period is used as an independent sample. For example, the system can divide the driving data of a day into one time period per hour, and the features such as braking frequency, acceleration change and speeding number in each time period will be recorded with a specific time label. Doing so not only helps to retain the continuity in time, but also facilitates subsequent analysis and processing. Then, the time series segment features are reduced in dimension to obtain the reduced features. Feature reduction is to reduce the dimension of data, remove redundant information, and improve computational efficiency. Common methods include principal component analysis (PCA) and linear discriminant analysis (LDA). For example, the system can use the PCA algorithm to reduce the dimension of features such as braking frequency, acceleration change, and speeding times in each time period, extract the most important features, and form reduced-dimensional features. The reduced-dimensional features can still retain the main information of the original data, but the computational complexity is greatly reduced. Next, feature clustering analysis is performed on the reduced-dimensional features to obtain the driving behavior pattern label. Feature clustering analysis is to classify data points with similar features into the same category. Commonly used clustering algorithms include K-means and DBSCAN. For example, the system can use the K-means algorithm to perform cluster analysis on the reduced-dimensional features and divide the owner's driving behavior into several categories, each category corresponding to a driving behavior pattern. The system generates a driving behavior pattern label for each category, such as "safe driving mode" and "aggressive driving mode". Finally, the driving behavior pattern label is matched with the pattern in the preset driving behavior pattern library to obtain the driving behavior recognition pattern. The preset driving behavior pattern library contains a variety of typical driving behavior patterns, such as safe driving mode, aggressive driving mode and mixed driving mode. The system will compare the generated driving behavior pattern label with the pattern in the pattern library to find the best matching pattern. For example, if the system concludes through cluster analysis that a car owner's driving behavior pattern label is "aggressive driving mode", the corresponding "aggressive driving mode" will be found in the pattern library and used as the final driving behavior recognition pattern.In summary, through this series of steps, the system can extract key features from behavioral desensitization data, perform time series segmentation, feature dimension reduction, feature clustering analysis and pattern matching, and finally generate accurate driving behavior recognition patterns. This process not only ensures the security and privacy of data, but also provides a reliable basis for subsequent point value calculation and rights matching, which helps to motivate car owners to form good driving habits and improve the overall road traffic safety level.

[0110] In a specific embodiment, the integral value calculation result is obtained by calculating the integral value of the target vehicle owner based on the driving behavior recognition mode according to the integral rules preset in the smart contract, including:

[0111] Reading a preset scoring rule through the smart contract to obtain a reading scoring rule, wherein the reading scoring rule includes a scoring weight and a scoring threshold corresponding to different driving behavior recognition modes;

[0112] Performing integral weight mapping on different driving behavior recognition modes by reading the integral rule to obtain the integral weight of the corresponding driving behavior recognition mode;

[0113] Performing duration statistics on the driving behavior recognition patterns to obtain the duration of each driving behavior recognition pattern;

[0114] Calculating an integral value corresponding to the driving behavior recognition mode based on the integral weight, the integral threshold and the duration;

[0115] Accumulating and / or subtracting the integral values ​​of all the driving behavior recognition modes to obtain a preliminary total integral value;

[0116] The preliminary total integral value is verified for correctness to obtain a verification result. If the verification result is correct, the preliminary total integral value is used as the integral value calculation result.

[0117] Specifically, first, the preset points rules are read through the smart contract to obtain the read points rules. These read points rules include the points weights and points thresholds corresponding to different driving behavior recognition modes. For example, suppose three driving behavior recognition modes are preset in the smart contract: safe driving mode, aggressive driving mode and mixed driving mode. Each mode has a corresponding points weight and points threshold. The points weight of the safe driving mode may be high, while the points weight of the aggressive driving mode is low or even negative, and the points weight of the mixed driving mode is between the two. The points threshold is the critical value that triggers the change of points. For example, the points threshold of the aggressive driving mode may be set to deduct a certain number of points each time the number of sudden braking or acceleration exceeds a certain number. Next, the points weights of different driving behavior recognition modes are mapped by reading the points rules to obtain the points weights of the corresponding driving behavior recognition modes. The system will find the preset points rules according to the identified driving behavior mode and determine the points weight of each mode. For example, if the system recognizes that a car owner's driving behavior is mainly in safe driving mode, the system will map the points weight of the safe driving mode (for example, +10 points / hour) to the points calculation of the car owner. Then, the duration of the driving behavior recognition mode is counted to obtain the duration of each driving behavior recognition mode. The system records the total time the owner drives in different driving behavior modes. For example, suppose a car owner drives for 4 hours in a day, of which 2 hours are in safe driving mode, 1 hour is in aggressive driving mode, and 1 hour is in mixed driving mode. The system counts the duration of these three modes respectively. Based on the corresponding integral weight, integral threshold and duration, the integral value of the corresponding driving behavior recognition mode is calculated. The system calculates the integral value of each mode based on the integral weight and duration of each mode. For example, for the safe driving mode, the integral value = 2 hours × 10 points / hour = 20 points; for the aggressive driving mode, assuming the integral weight is -5 points / hour, the integral value = 1 hour × (-5 points / hour) = -5 points; for the mixed driving mode, assuming the integral weight is 5 points / hour, the integral value = 1 hour × 5 points / hour = 5 points. The integral values ​​of all driving behavior recognition modes are accumulated and / or subtracted to obtain the preliminary total integral value. The system will add or subtract the points in all modes to get the preliminary total points. For example, the preliminary total points of the above car owner = 20 points + (-5 points) + 5 points = 20 points. The preliminary total points are verified for correctness to get a verification result. The system will use a series of verification mechanisms to ensure the correctness and fairness of the points calculation. For example, the system may check for abnormal data or logical errors to ensure that the calculation of each point value is based on accurate data and rules. If the verification result is correct, the system will use the preliminary total points as the final point value calculation result.In summary, through this series of steps, the system can read the preset points rules from the smart contract, perform point weight mapping for different driving behavior recognition modes, count the duration, calculate the points, and perform accumulation and / or subtraction calculations, and finally generate accurate points calculation results. This process not only ensures the transparency and fairness of the points calculation, but also provides a reliable foundation for subsequent rights and interests matching and incentive mechanisms, which helps to promote car owners to form good driving habits and improve the overall road traffic safety level.

[0118] In a specific embodiment, the equity matching of the integral value calculation result based on the integral-equity exchange table to obtain a recommended equity combination scheme includes:

[0119] The historical redemption records and the demand preferences of the target car owner are obtained, and the integral value calculation result is represented in multiple dimensions based on the historical redemption records by a restricted Boltzmann machine in a preset deep belief network to obtain an integral feature representation vector; wherein the row labels in the integral feature representation vector are the weekend usage ratio, the weekday usage ratio and the target car owner's loyalty;

[0120] Constructing a points-equity matrix based on the points-equity exchange table; wherein the rows of the points-equity matrix represent different exchange points, and the columns of the points-equity matrix represent the weekend usage ratio, the weekday usage ratio and the target car owner loyalty;

[0121] Using a cosine similarity algorithm, similarity is calculated between the integral feature representation vector and the vector in the integral-equity matrix to obtain a similarity score vector;

[0122] Performing label mapping on the similarity score vector to obtain a mapped equity label vector;

[0123] Weighting the elements in the mapped equity tag vector based on the demand preference to obtain a mapped equity tag weighted vector;

[0124] Using a conditional random field, the elements in the mapped equity label weighted vector are sorted and labeled in descending order to obtain a sorted and labeled equity label vector;

[0125] When the calculated integral value is greater than a preset redemption threshold, rights matching is performed in the integral-rights relationship redemption table based on the sorted and labeled rights label vector to obtain a recommended rights combination plan; wherein the preset redemption threshold is the sum of the redemption points of each right in the recommended rights combination plan.

[0126] Specifically, first, the system will obtain the historical redemption records and demand preferences of the target car owner. This information is crucial for recommending a suitable combination of benefits. For example, if a car owner has frequently redeemed car wash services in the past few months, but rarely redeemed gas discounts, the system will record these historical redemption records. At the same time, the system will also understand the demand preferences of car owners through questionnaires, user feedback, etc. For example, car owners may prefer to choose maintenance services rather than car wash services. Next, through the restricted Boltzmann machine in the preset deep belief network, the integral value calculation results are multi-dimensionally represented based on the historical redemption records to obtain the integral feature representation vector. The restricted Boltzmann machine is a neural network used for unsupervised learning that can extract useful features from a large amount of historical data. In this process, the system converts the car owner's historical redemption records into a vector, in which the row labels include the weekend usage ratio, the weekday usage ratio, and the target car owner's loyalty. For example, if a car owner uses his car 30% of the time on weekends and 70% on weekdays, and has a high loyalty (e.g., he has not redeemed benefits of other brands in the past year), the system will generate a score feature representation vector such as [0.3, 0.7, 0.9]. Then, a score-benefit matrix is ​​constructed based on the score-benefit relationship exchange table. The rows of the score-benefit matrix represent different redemption points, and the columns represent the weekend usage ratio, weekday usage ratio, and target car owner loyalty. For example, suppose the score-benefit relationship exchange table contains three benefits: car wash service (requires 100 points), gas discount (requires 200 points), and repair and maintenance service (requires 300 points). The system will generate a vector for each benefit, such as the vector of the car wash service may be [0.4, 0.6, 0.8], which represents the usage ratio of the car wash service on weekends and weekdays and the loyalty of the car owner. The cosine similarity algorithm is used to calculate the similarity between the score feature representation vector and the vector in the score-benefit matrix to obtain a similarity score vector. The cosine similarity algorithm is used to measure the similarity between two vectors. The system calculates the cosine similarity between the integral feature representation vector and each vector in the integral-equity matrix to generate a similarity score vector. For example, assuming that the integral feature representation vector is [0.3, 0.7, 0.9], the system calculates its cosine similarity with the car wash service vector [0.4, 0.6,0.8] and obtains a score, such as 0.95. Label mapping is performed on the similarity score vector to obtain a mapped equity label vector. Label mapping refers to mapping each score in the similarity score vector to the corresponding equity label. For example, the system maps the similarity score of 0.95 for car wash service to the label "car wash service" to generate a mapped equity label vector, such as [car wash service, gas discount, repair and maintenance service]. The elements in the mapped equity label vector are weighted based on demand preferences to obtain a mapped equity label weighted vector.The system weights each element in the mapped equity label vector according to the owner's demand preferences. For example, if the system learns that the owner prefers maintenance services, the system will increase the weight of maintenance services and generate a mapped equity label weight vector, such as [0.95, 0.85, 1.2]. Using conditional random fields, the elements in the mapped equity label weight vector are sorted and labeled from large to small to obtain a sorted and labeled equity label vector. Conditional random fields are a machine learning model for sequence labeling that can sort equity labels according to weighted similarity scores. The system sorts the elements in the mapped equity label weight vector from large to small to generate a sorted and labeled equity label vector, such as [maintenance service, car wash service, gas discount]. When the calculated integral value is greater than the preset redemption threshold, the system matches the equity in the integral-equity relationship redemption table based on the sorted and labeled equity label vector to obtain a recommended equity combination plan. The preset redemption threshold is the sum of the redemption points of each equity in the recommended equity combination plan. For example, assuming that the calculated score of a car owner is 350 points and the preset redemption threshold is 300 points, the system will label the equity label vector according to the sorting, and recommend that the car owner give priority to maintenance services (300 points), and the remaining 50 points can choose car wash services (100 points). Finally, the system will generate a recommended equity combination plan, such as "maintenance service + car wash service". In summary, through this series of steps, the system can comprehensively consider the car owner's historical redemption records, demand preferences and score values, and recommend the most suitable equity combination plan for him. This process not only improves the accuracy and personalization of recommendations, but also enhances the satisfaction and loyalty of car owners, and promotes a virtuous cycle of the points and equity circulation management system.

[0127] In a specific embodiment, the conditional random field is used to sort and label the elements in the mapped equity label weighted vector from large to small to obtain a sorted and labeled equity label vector, including:

[0128] Decomposing the elements in the weighted vector of the mapped equity tags into equity type features to obtain an equity feature state sequence; wherein the equity feature state sequence includes a refueling discount state, a parking discount state, and a repair and maintenance discount state;

[0129] Constructing a transfer probability matrix for the equity feature state sequence using a preset conditional probability algorithm to obtain an equity state transfer matrix;

[0130] Performing potential function calculation on the equity characteristic state sequence based on the equity state transfer matrix to obtain an equity state potential energy vector;

[0131] Perform conditional probability reasoning on the equity state potential energy vector to obtain an equity annotation probability distribution; wherein the equity annotation probability distribution includes a high-frequency equity probability, a medium-frequency equity probability, and a low-frequency equity probability;

[0132] Using a conditional random field, based on the equity labeling probability distribution, the elements in the mapped equity label weighted vector are sorted and labeled in descending order to obtain a sorted and labeled equity label vector.

[0133] Specifically, in the process of managing the transfer of car owner points, the system will first receive the weighted vector of the mapped rights and interests label, which contains the initial weight information of various rights and interests. When the system decomposes the rights and interests type characteristics of this vector, it will classify and organize different types of rights and interests according to their attribute characteristics to form a rights and interests characteristic state sequence. For example, when the system receives a vector containing multiple rights and interests labels, it will decompose it into different state sequences such as refueling preferential state (such as discount rights for 92 / 95 / 98 gasoline), parking preferential state (such as mall parking, roadside parking exemption rights) and maintenance preferential state (such as four-wheel alignment, oil change preferential rights). After obtaining the rights and interests characteristic state sequence, the system constructs the rights and interests state transfer matrix through the preset conditional probability algorithm. This matrix describes the conversion probability relationship between different rights and interests states. For example, by analyzing historical data, the system finds that when the car owner uses the refueling preferential rights and interests, the probability of switching to the parking preferential rights and interests is 0.6, the probability of switching to the maintenance preferential rights and interests is 0.3, and the probability of continuing to use the refueling preferential rights and interests is 0.1. These probability data constitute an important part of the rights and interests state transfer matrix. After the equity state transfer matrix is ​​constructed, the system will calculate the potential function of the equity feature state sequence based on this matrix to obtain the equity state potential energy vector. This calculation process takes into account the strength of the association between different equity states. For example, the system will calculate the potential energy value of using the shopping mall parking discount after using the refueling discount during the weekend, or the potential energy value of using the refueling discount after using the maintenance discount on weekdays. These potential energy values ​​reflect the applicability of different equity combinations in different scenarios. Next, the system performs conditional probability reasoning on the equity state potential energy vector to generate an equity labeling probability distribution. In this distribution, the high-frequency equity probability may represent a refueling discount equity used 2-3 times a week (with a probability value of 0.5), the medium-frequency equity probability may correspond to a parking discount equity used 3-4 times a month (with a probability value of 0.3), and the low-frequency equity probability may be a maintenance discount equity used 1-2 times a quarter (with a probability value of 0.2). Finally, the system uses conditional random fields to sort and label the elements in the weighted vector of mapped equity labels from large to small based on the equity labeling probability distribution. Specifically, assuming that the car owner currently has 5,000 points available, the system will sort all redeemable benefit combinations according to the probability distribution calculated previously. For example, the benefit combination of "20 yuan off for purchases of 100 yuan or more of No. 92 gasoline + 2 hours of free parking at the mall" may be marked as the highest priority because it meets both high-frequency usage scenarios (refueling) and medium-frequency usage scenarios (parking), while a combination such as "four-wheel alignment discount + oil change discount" may be marked as a lower priority because it belongs to a low-frequency usage scenario. In this way, the system can provide car owners with benefit redemption suggestions that best suit their usage habits and needs, thereby optimizing the circulation effect of the entire points benefit.

[0134] In a specific embodiment, the cosine similarity algorithm is used to calculate the similarity between the integral feature representation vector and the vector in the integral-equity matrix to obtain a similarity score vector, including:

[0135] Using the cosine similarity algorithm, the integral feature representation vector is multiplied by each element in the integral-equity matrix to obtain a preliminary similarity vector; wherein the number of the preliminary similarity vectors is the number of rows of the integral-equity matrix;

[0136] Calculate the modulus of each row in the score-equity matrix to obtain a vector modulus length; wherein the number of the vector modulus lengths is equal to the number of the preliminary similarity vectors;

[0137] Based on the elements in each of the preliminary similarity vectors, normalizing the corresponding vector modulus lengths to obtain a normalized similarity value vector;

[0138] Performing modulus calculation on the integral feature representation vector to obtain the integral representation modulus length;

[0139] Dividing the elements in the normalized similarity value vector by the integral representation modulus length to obtain a preliminary similarity score vector;

[0140] A nonlinear transformation is performed on the similarity score vector to obtain a similarity score vector.

[0141] Specifically, first, the cosine similarity algorithm is used to perform a dot product operation on the integral feature representation vector and each element in the integral-equity matrix to obtain a preliminary similarity vector. The cosine similarity algorithm measures the similarity of two vectors by calculating the cosine value of the angle between them. Specifically, the system will perform a dot product operation on the integral feature representation vector and each row vector in the integral-equity matrix to generate a preliminary similarity vector. For example, assuming that the integral feature representation vector is [0.3, 0.7, 0.9], and there are three vectors in the integral-equity matrix, namely [0.4, 0.6, 0.8], [0.5, 0.5, 0.5] and [0.2, 0.8, 0.9], the system will calculate the dot products of these three vectors with the integral feature representation vector respectively to obtain a preliminary similarity vector , that is, [0.98, 0.75, 1.21]. The modulus of each row vector in the integral-equity matrix. For example, for the vector [0.4, 0.6, 0.8], its modulus is Similarly, the modulus lengths of the other two vectors are calculated as and Based on the elements in each preliminary similarity vector, the corresponding vector modulus is normalized to obtain the normalized similarity value vector. The normalization is to eliminate the influence of vector length and make the similarity calculation more accurate. The system will divide each element in the preliminary similarity vector by the corresponding vector modulus. For example, for the preliminary similarity vector [0.98, 0.75, 1.21], the corresponding vector moduli are [1.08, 0.87, 1.22], and the normalized similarity value vector is [0.98 / 1.08, 0.75 / 0.87, 1.21 / 1.22], that is, [0.91, 0.86, 0.99]. The integral feature representation vector is modulo calculated to obtain the integral representation modulus length. The system will calculate the modulus of the integral feature representation vector. For example, for the integral feature representation vector [0.3, 0.7, 0.9], its modulus is Divide the elements in the normalized similarity value vector by the integral representation modulus length to obtain a preliminary similarity score vector. The system divides each element in the normalized similarity value vector by the modulus length of the integral feature representation vector. For example, for the normalized similarity value vector [0.91, 0.86, 0.99], the integral representation modulus length is 1.18, and the preliminary similarity score vector is [0.91 / 1.18, 0.86 / 1.18, 0.99 / 1.18], that is, [0.77, 0.73, 0.84]. Perform a nonlinear transformation on the similarity score vector to obtain the final similarity score vector. The nonlinear transformation is to further adjust the similarity score to make it more in line with the needs of practical applications. Common nonlinear transformation methods include sigmoid function, ReLU function, etc. For example, the system can use the sigmoid function to transform the preliminary similarity score vector to obtain the final similarity score vector. The formula of the sigmoid function is Assuming the initial similarity score vector is [0.77, 0.73, 0.84], after the sigmoid function transformation, the final similarity score vector is , that is, [0.68, 0.67, 0.70]. In summary, through this series of steps, the system can use the cosine similarity algorithm to calculate the similarity between the integral feature representation vector and the vector in the integral-equity matrix, and finally generate a similarity score vector. This process not only ensures the accuracy and reliability of the similarity calculation, but also provides a scientific basis for subsequent equity matching recommendations, which helps to improve the satisfaction and loyalty of car owners and promote the healthy development of the points and equity circulation management system.

[0142] The above describes the method for managing the transfer of car owner points and rights in the embodiment of the present invention. The following describes the management system for the transfer of car owner points and rights in the embodiment of the present invention. Figure 2In one embodiment of the present invention, a management system for the circulation of car owner points and rights includes:

[0143] The collection module 21 is used to collect driving behavior data of the target vehicle owner through an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain behavior collection data;

[0144] A desensitization module 22 is used to perform data desensitization processing on the behavior collection data to obtain behavior desensitized data;

[0145] The recognition module 23 is used to perform behavior pattern recognition on the target vehicle owner based on the behavior desensitization data by using a preset recognition algorithm to obtain a driving behavior recognition pattern;

[0146] A calculation module 24 is used to calculate the integral value of the target vehicle owner based on the driving behavior recognition mode according to the integral rules preset in the smart contract to obtain the integral value calculation result;

[0147] The matching module 25 is used to call the point-rights relationship exchange table stored in the database, perform rights matching on the point-rights relationship exchange table based on the point value calculation result, and obtain a recommended rights combination plan.

[0148] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0149] Reference Figure 3 The present invention also provides a computer device in an embodiment, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer 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 computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0150] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0151] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0152] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and 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-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0153] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0154] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for managing the circulation of car owner points and rights, characterized in that: The following steps are involved: The driving behavior data of the target vehicle owner is collected by using an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain the behavior collection data; Performing data desensitization processing on the behavior collection data to obtain behavior desensitized data; Using a preset recognition algorithm, based on the behavior desensitization data, the behavior pattern of the target vehicle owner is recognized to obtain a driving behavior recognition pattern; By using the points rules preset in the smart contract, the points value of the target car owner is calculated based on the driving behavior recognition mode to obtain the points value calculation result; Calling the point-equity exchange table stored in the database, performing equity matching on the point-equity exchange table based on the point value calculation result, and obtaining a recommended equity combination plan; The historical redemption records and the demand preferences of the target car owner are obtained, and the integral value calculation result is represented in multiple dimensions based on the historical redemption records by a restricted Boltzmann machine in a preset deep belief network to obtain an integral feature representation vector; wherein the row labels in the integral feature representation vector are the weekend usage ratio, the weekday usage ratio and the target car owner's loyalty; Constructing a points-equity matrix based on the points-equity exchange table; wherein the rows of the points-equity matrix represent different exchange points, and the columns of the points-equity matrix represent the weekend usage ratio, the weekday usage ratio and the target car owner loyalty; Using a cosine similarity algorithm, similarity is calculated between the integral feature representation vector and the vector in the integral-equity matrix to obtain a similarity score vector; Performing label mapping on the similarity score vector to obtain a mapped equity label vector; Weighting the elements in the mapped equity tag vector based on the demand preference to obtain a mapped equity tag weighted vector; Using a conditional random field, the elements in the mapped equity label weighted vector are sorted and labeled in descending order to obtain a sorted and labeled equity label vector; When the calculated integral value is greater than a preset redemption threshold, rights matching is performed in the integral-rights relationship redemption table based on the sorted and labeled rights label vector to obtain a recommended rights combination plan; wherein the preset redemption threshold is the sum of the redemption points of each right in the recommended rights combination plan.

2. The method for managing the circulation of vehicle owner points and rights according to claim 1, characterized in that: The driving behavior data of the target vehicle owner is collected by using the intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain the behavior collection data, including: The driving behavior data of the target vehicle owner is collected by using an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain the original driving behavior data; Preprocessing the original driving behavior data to obtain pure vehicle operation status data; Performing multi-dimensional feature extraction on the pure vehicle operation status data to obtain multi-dimensional driving behavior features; The multi-dimensional driving behavior characteristics are packaged in a preset data format to obtain behavior collection data.

3. The method for managing the circulation of vehicle owner points and rights according to claim 1, characterized in that: The smart terminal device is provided with a data desensitization module, and the data desensitization processing is performed on the behavior collection data to obtain behavior desensitized data, including: The sensitive attributes of the behavior collection data are disturbed by the data desensitization module to obtain preliminary desensitized data; Performing grouping and generalization processing on the preliminary desensitized data to obtain generalized desensitized data; Encrypting key fields in the generalized desensitized data to obtain encrypted desensitized data; Performing partial data replacement masking on non-critical sensitive information in the generalized desensitized data to obtain partially masked data; Combining the encrypted desensitized data and the partially masked data to obtain combined data; Performing distributed computing on the combined data to obtain decentralized data; The decentralized data is desensitized and verified through a zero-knowledge proof protocol. If the desensitization verification is successful, behavioral desensitized data is obtained.

4. The method for managing the circulation of car owner points and rights according to claim 1, characterized in that: The method of performing behavior pattern recognition on the target vehicle owner based on the behavior desensitization data by using a preset recognition algorithm to obtain a driving behavior recognition pattern includes: By using a preset recognition algorithm, the features in the behavior-massified data that are related to the points-rights exchange table are identified to obtain relevant recognition features; wherein the relevant recognition features include braking frequency features, acceleration features, and speeding frequency features; Performing time series segmentation on the relevant identification features to obtain time series segmentation features; wherein the time series segmentation features are time series segmentation features with time tags; Performing feature dimensionality reduction on the time series segment features to obtain features after dimensionality reduction; Performing feature clustering analysis on the dimension-reduced features to obtain driving behavior pattern labels; The driving behavior pattern label is matched with a pattern in a preset driving behavior pattern library to obtain a driving behavior recognition pattern; wherein the driving behavior recognition pattern includes a safe driving mode, an aggressive driving mode and a mixed driving mode.

5. The method for managing the circulation of vehicle owner points and rights according to claim 1, characterized in that: The point value calculation result obtained by calculating the point value of the target vehicle owner based on the driving behavior recognition mode according to the point value rules preset in the smart contract includes: Reading a preset scoring rule through the smart contract to obtain a reading scoring rule, wherein the reading scoring rule includes a scoring weight and a scoring threshold corresponding to different driving behavior recognition modes; Performing integral weight mapping on different driving behavior recognition modes by reading the integral rule to obtain the integral weight of the corresponding driving behavior recognition mode; Performing duration statistics on the driving behavior recognition patterns to obtain the duration of each driving behavior recognition pattern; Calculating an integral value corresponding to the driving behavior recognition mode based on the integral weight, the integral threshold and the duration; Accumulating and / or subtracting the integral values ​​of all the driving behavior recognition modes to obtain a preliminary total integral value; The preliminary total integral value is verified for correctness to obtain a verification result. If the verification result is correct, the preliminary total integral value is used as the integral value calculation result.

6. The method for managing the circulation of vehicle owner points and rights according to claim 1, characterized in that: The method of using the cosine similarity algorithm to calculate the similarity between the integral feature representation vector and the vector in the integral-equity matrix to obtain a similarity score vector includes: Using the cosine similarity algorithm, the integral feature representation vector is multiplied by each element in the integral-equity matrix to obtain a preliminary similarity vector; wherein the number of the preliminary similarity vectors is the number of rows of the integral-equity matrix; Calculate the modulus of each row in the score-equity matrix to obtain a vector modulus length; wherein the number of the vector modulus lengths is equal to the number of the preliminary similarity vectors; Normalizing the corresponding vector modulus based on the elements in each of the preliminary similarity vectors to obtain a normalized similarity value vector; Performing modulus calculation on the integral feature representation vector to obtain the integral representation modulus length; Dividing the elements in the normalized similarity value vector by the integral representation modulus length to obtain a preliminary similarity score vector; A nonlinear transformation is performed on the similarity score vector to obtain a similarity score vector.

7. A circulation management system for car owner points and rights, characterized in that: include: A collection module, used to collect driving behavior data of the target vehicle owner through an intelligent terminal device deployed in the vehicle of the target vehicle owner to obtain behavior collection data; A desensitization module, used for performing data desensitization processing on the behavior collection data to obtain behavior desensitized data; An identification module, used to identify the behavior pattern of the target vehicle owner based on the behavior desensitization data through a preset identification algorithm to obtain a driving behavior identification pattern; A calculation module, used to calculate the integral value of the target vehicle owner based on the driving behavior recognition mode according to the integral rules preset in the smart contract, and obtain the integral value calculation result; A matching module, used to call the point-equity relationship exchange table stored in the database, perform equity matching on the point-equity relationship exchange table based on the point value calculation result, and obtain a recommended equity combination plan; The historical redemption records and the demand preferences of the target car owner are obtained, and the integral value calculation result is represented in multiple dimensions based on the historical redemption records by a restricted Boltzmann machine in a preset deep belief network to obtain an integral feature representation vector; wherein the row labels in the integral feature representation vector are the weekend usage ratio, the weekday usage ratio and the target car owner's loyalty; Constructing a points-equity matrix based on the points-equity exchange table; wherein the rows of the points-equity matrix represent different exchange points, and the columns of the points-equity matrix represent the weekend usage ratio, the weekday usage ratio and the target car owner loyalty; Using a cosine similarity algorithm, similarity is calculated between the integral feature representation vector and the vector in the integral-equity matrix to obtain a similarity score vector; Performing label mapping on the similarity score vector to obtain a mapped equity label vector; Weighting the elements in the mapped equity tag vector based on the demand preference to obtain a mapped equity tag weighted vector; Using a conditional random field, the elements in the mapped equity label weighted vector are sorted and labeled in descending order to obtain a sorted and labeled equity label vector; When the calculated integral value is greater than a preset redemption threshold, rights matching is performed in the integral-rights relationship redemption table based on the sorted and labeled rights label vector to obtain a recommended rights combination plan; wherein the preset redemption threshold is the sum of the redemption points of each right in the recommended rights combination plan.

8. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images