Electric vehicle claim settlement whole-process co-processing system and method based on multi-modal data fusion

Abstract

The invention provides an electric vehicle claim settlement whole-process co-processing system and method based on multi-modal data fusion. The method belongs to the technical field of multi-modal data fusion. Through a multi-modal data fusion technology, the method comprises the following steps: acquiring case data, processing and issuing case information; on the basis of the case information, automatically matching surveyors in an area range to execute a survey task, and generating a survey result; generating a maintenance order based on the survey result and pushing the maintenance order to each franchising maintenance shop in the area; and each franchise maintenance shop offers the price after receiving the maintenance order, and the system determines the claim amount and the maintenance shop to be repaired according to the merchant portrait and the merchant quotation. The system can integrate data from different sources, reduces the interference of a single data source or abnormal data, and improves the accuracy of predicting the survey task load.

Description

Technical Field

[0001] The present invention provides a collaborative disposal system and method for the whole process of electric vehicle claims settlement based on multimodal data fusion, belonging to the technical field of multimodal data fusion. Background Art

[0002] Electric vehicles are currently important means of transportation for the people in China. The number of cases involving electric vehicle claims settlement in traffic surveys has been remaining high. It is particularly important for insurance companies to accurately predict the future survey task volume, reasonably allocate relevant resources, improve the survey efficiency, accurately and quickly generate the claim settlement amount, and help insurance companies reduce losses and burdens. However, there are still many deficiencies in the existing methods for dealing with these problems.

[0003] On the one hand, the existing prediction methods often rely on simple historical data analysis and lack multi-dimensional and comprehensive data fusion and correction mechanisms. This results in the prediction results being often interfered by single data sources or abnormal data, and the accuracy is limited. At the same time, most of the existing prediction models are based on traditional statistical methods or machine learning algorithms, and their prediction ability for the survey task volume data with time-series and non-linear characteristics is limited and difficult to meet the actual business needs.

[0004] On the other hand, in terms of resource allocation, the existing systems often lack intelligent decision support. The scheduling plan of surveyors, the allocation of survey vehicles, and the preparation of survey tools often rely on manual experience or simple rule settings, and it is difficult to achieve the optimal allocation and efficient utilization of resources. Especially in the face of abnormal situations such as emergencies or extreme weather, the emergency response ability and resource allocation flexibility of the existing systems are insufficient, which is likely to affect the smooth progress of the survey work and cause some surveyors to bear too heavy a burden, resulting in low efficiency of the overall disposal process.

[0005] Finally, due to the large number of electric vehicle brands in society, the rapid update and iteration of models, and the messy and non-uniform accessories, it is difficult for insurance companies to determine the claim settlement amount for the repair of electric vehicles during the claim settlement process. In the actual claim settlement process, insurance companies often have disputes with repair shops over the repair amount of the claim settlement. How to reduce losses and burdens has become a problem that needs to be solved. Summary of the Invention

[0006] The present invention provides a collaborative disposal system and method for the whole process of electric vehicle claims settlement based on multimodal data fusion to solve the problems mentioned in the above background art:

[0007] The collaborative disposal method for the whole process of electric vehicle claims settlement based on multimodal data fusion proposed by the present invention includes:

[0008] S1. Obtain case data, process it and publish the case information;

[0009] S2. Automatically match surveyors within the regional scope based on the case information to perform survey tasks and generate survey results;

[0010] S3. Generate a repair order based on the survey results and push it to each franchised repair shop within the region;

[0011] S4. After each franchised repair shop receives the repair order, it makes a quotation, and the system determines the claim settlement amount and the undertaking repair shop according to the merchant portrait and combines the merchant's quotation.

[0012] The full-process collaborative disposal system for electric vehicle claim settlement based on multi-modal data fusion proposed by the present invention includes a memory, a processor, and a computer program stored on the memory and operable on the memory. The processor executes the program to implement the full-process collaborative disposal method for electric vehicle claim settlement based on multi-modal data fusion as described in any one of the above.

[0013] Advantages of the present invention: Through the multi-modal data fusion technology, the system can integrate data from different sources, reduce the interference of single data sources or abnormal data, and improve the accuracy of predicting the survey task volume; through the dynamic matching algorithm based on real-time geographical location and surveyor skill tags, it can automatically match the optimal surveyor to perform survey tasks, greatly reducing manual intervention and improving the execution efficiency and accuracy of survey tasks; according to the task information received by the surveyor through the APP, a reasonable survey route can be planned, reducing delays caused by manual arrangement errors; after generating the repair order, the system uses a multi-dimensional store capacity portrait model to automatically match the most suitable repair shop for repair, optimizing the allocation of franchised repair shop resources, making the repair process more efficient, while reducing competitive errors between merchants, ensuring the balance of repair quality and cost-effectiveness; through the full-process monitoring technology and blockchain technology, the system can record and monitor the entire repair process, improving the traceability and transparency of the whole process; the photos and progress information regularly uploaded by the repair shop are subject to quality inspection to ensure the repair quality, and the system can also issue abnormal warnings in a timely manner to identify potential problems in advance, ensuring the controllability of repair quality and service; the system automatically triggers the claim settlement payment process through smart contracts, reducing manual intervention and improving the claim settlement efficiency; the system combines multi-source data cross-verification technology to ensure the accuracy and consistency of repair records and claim settlements, improving the automation level of claim settlements, reducing errors and claim settlement time, and optimizing the customer experience; through the analysis of historical claim settlement data and prediction models, the system can foresee future survey task volumes, formulate resource allocation strategies in advance, and develop emergency plans for emergencies or extreme weather conditions, enhancing the flexibility and responsiveness of the system, ensuring the efficient use of resources, and improving the robustness and response speed of the overall system. Description of the Drawings

[0014] Figure 1 This is the flowchart of the method steps of the present invention;

[0015] Figure 2 This is the system flowchart of the present invention. Detailed implementation manners

[0016] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for explaining and illustrating the present invention, and are not used to limit the present invention.

[0017] An embodiment of the present invention, as Figure 1 shown, is a collaborative disposal method for the whole process of electric vehicle claims based on multi-modal data fusion. The method includes:

[0018] S1. The case administrator obtains case data, processes it and publishes the case information in the system;

[0019] S2. The system automatically matches the surveyors within the regional scope based on the case information to perform survey tasks and generates survey results;

[0020] S3. The system generates a repair order based on the survey results and pushes it to each franchised repair shop within the region;

[0021] S4. After each franchised repair shop receives the repair order, it makes a quotation, and the system determines the claim amount and the repair shop undertaking the repair based on the merchant portrait and the merchant's quotation.

[0022] The working principle of the above technical solution is as follows: The case administrator obtains relevant information about the involved electric vehicle through the system interface (usually pushed by the insurance company); the case data includes the claim number, license plate of the involved object, liability division, contact information, survey address, etc.; the case administrator conducts a preliminary screening and verification of the obtained case data to ensure the accuracy and integrity of the data; and cleans the data, removes duplicate information, corrects incorrect data, etc.; the processed case data is entered into the system and standardized case information is generated; the system publishes the case information to the internal platform or relevant business processes for subsequent use; the system automatically searches for available surveyors within the regional scope according to conditions such as the geographical location and remarks information in the case information; and conducts intelligent matching considering factors such as the professional skills, experience, and current workload of the surveyors. For example, if the case is very urgent, the surveyor closest to the location can be arranged preferentially; after receiving the survey task assigned by the system, the matched surveyor immediately goes to the appointed location for survey; uses equipment to take pictures of the electric vehicle to record the vehicle damage situation; after the survey is completed, the surveyor enters the survey results into the system, including vehicle damage photos, improvement of case-related information, repair list quotation, etc.; the system reviews and confirms the survey results to ensure their accuracy and reliability. The system automatically generates a repair order according to the survey results, including the electric vehicle brand, model, vehicle damage situation, repair list, etc.; if it is a minor survey such as a slight friction, the loss is small and the liability division is clear, and the owner of the three-wheeled electric vehicle needs to handle it quickly, then, according to the relevant photos and video information from the previous survey, online quick claims settlement is carried out; the system pushes the generated repair order to each affiliated repair shop within the region through the APP for them to view and quote. After receiving the repair order, each affiliated repair shop quotes according to factors such as its own repair ability, parts inventory, and labor cost; among them, the quotation includes detailed items such as parts cost and labor cost; if the electric vehicle owner has a self-selected repair shop (including non-affiliated repair shops), then the self-selected repair shop of the owner is preferably selected for repair at the same price. For example, if the final claim settlement amount determined by the affiliated repair shop is N, and the self-selected repair shop of the owner can accept the final claim settlement amount, then the self-selected repair shop of the customer is preferably selected for repair; the system constructs a merchant portrait based on data such as the historical repair records, customer satisfaction, and repair quality of the affiliated repair shop; the merchant portrait is used to evaluate the reputation and service quality of the affiliated repair shop; the system conducts a comprehensive evaluation based on the merchant portrait and the merchant's quotation, and selects the optimal affiliated repair shop to place an order; after the order is placed, the system notifies the affiliated repair shop and the electric vehicle owner to arrange the repair time and location; at the same time, the system generates a repair contract or service agreement for both parties to confirm and sign.

[0023] The effects of the above technical solutions are as follows: By centrally obtaining and processing case data by case managers and uniformly publishing it in the system, the time and error of information transmission are reduced, the efficiency of case processing is improved, and the customer experience is enhanced; The system automatically matches surveyors within the regional scope to perform survey tasks, avoiding the cumbersome and uncertain manual dispatching, making the survey work faster and more accurate, and improving work efficiency; The system automatically generates repair orders based on the survey results and pushes them to each affiliated repair shop within the region, simplifying the generation and transmission process of repair orders and improving work efficiency; The system intelligently matches the most suitable surveyors and affiliated repair shops according to case information and survey results, achieving optimal allocation of resources; Through the construction of merchant portraits, the system can more comprehensively understand the strength and service quality of affiliated repair shops, provide a strong basis for order placement, further optimize resource allocation, and reduce unnecessary subsequent time expenditure; The system automatically processes case information, survey results, and repair orders, reducing human intervention and enhancing the transparency and fairness of the process; Each affiliated repair shop quotes based on a unified repair order, and the system determines the claim amount and the accepting repair shop in combination with the merchant portrait, ensuring the fairness of quotation and order allocation and improving the order placement success rate; Through fast and accurate case processing and survey services, as well as optimized repair resource allocation, customers can enjoy more efficient and high-quality electric vehicle repair services; Systematic process management reduces human errors and delays, enhancing customers' satisfaction and trust in repair services; The system places orders according to merchant portraits and quotations, providing more business opportunities for affiliated repair shops with strong strength and excellent services, indirectly improving the service quality of each affiliated repair shop and making the competition among affiliated repair shops more benign.

[0024] In one embodiment of the present invention, the S1 includes:

[0025] S11. The case manager logs in to the system through the web terminal to obtain case data related to electric vehicle repair;

[0026] S12. Preprocess the newly created case;

[0027] S13. The case manager classifies and tags the preprocessed case data;

[0028] S14. The case manager publishes the case information in the system.

[0029] The working principle of the above technical solution is as follows: the case administrator uses the designated account and password to log in to the electric vehicle claims management system through the web; the system performs identity authentication to ensure that only authorized case administrators can access the system; the case administrator selects or enters query conditions in the system, such as time range, case number, license plate number, etc., retrieves case data that meets the conditions from the database according to the entered query conditions, and displays it to the case administrator; the case administrator creates a new electric vehicle case involved in the system based on the acquired case data or other information; the content of the newly created case includes but is not limited to the report number (automatically generated by the system), the target license plate (the license plate number of the electric vehicle), the division of responsibilities (the division of the responsible party for the investigation), the contact person (the name of the owner or relevant person in charge), the third party telephone number (the contact number when a third party is involved), the target telephone number (the contact number of the owner), the insurance company contact person (the contact person for handling insurance matters), the insurance company contact number, the damage assessment location (map positioning or manually entered specific address), whether to directly replace the car for compensation (whether the new car is directly compensated due to serious inspection), the allocation of surveyors (system automatic matching or manually designated surveyors), whether to expedite (whether the case needs to be handled as a priority) and remarks (other matters that need to be explained) Item); the system or case manager cleans the data in the newly created case, such as removing extra spaces, correcting typos, unifying data formats, etc.; and verifies the key information in the newly created case according to the built-in data verification rules; for example, check whether the target license plate is complete and meets the format requirements, whether the contact number is valid (for example, whether it is a number and the number of digits is correct), whether the damage assessment location has been mapped or manually entered, etc.; the system feeds back the verification results to the case manager, and the case manager needs to modify or supplement the data that does not meet the requirements; the case manager verifies the data based on the vehicle type (for example, electric two-wheeled vehicles, electric The system classifies the pre-processed case data based on dimensions such as "electric two-wheeler, etc." and "inspection type" (such as collision, spontaneous combustion, flooding, etc.); the system or case administrator labels the classified case data accordingly, such as "electric two-wheeler-collision", etc.; after the case administrator confirms that the pre-processed and classified case data are correct, the case information is published in the system; the system stores the published case information in the database and updates the case status to "pending inspection" or a similar status, indicating that the case is ready for inspection by the surveyor; at the same time, the system sends a notification to the relevant surveyor or franchised repair shop to inform them that there is a new task to be processed.

[0030] The effects of the above technical solutions are as follows: By logging into the system through the web side, case managers can conveniently obtain and create cases of involved electric vehicles, reducing the processing time of paper-based operations and improving the efficiency of case handling. When creating a case, the system provides a detailed case content template (automatically filling in relevant information by pasting the identification board), including key information such as case number, target license plate, and liability division, ensuring the integrity and accuracy of case data, and enhancing the transparency and traceability of case management. Through the built-in data verification rules in the system, it can automatically check the integrity and validity of case data, such as whether the target license plate is filled in completely and whether the contact phone number is valid, avoiding subsequent problems caused by data errors or omissions and ensuring the coherence of the subsequent processing flow. The preprocessed case data is classified and labeled, making the data more consistent and standardized, facilitating subsequent retrieval and analysis. Case managers can reasonably allocate surveyors based on the preprocessed case data to ensure that survey tasks can be completed promptly and accurately, shortening the case handling time and ensuring the efficiency of case handling. Through labeling, the system can quickly identify cases of specific types or under specific conditions, such as vehicle type, survey type, etc., providing strong support for resource allocation and scheduling. The standardization and automation of the case handling process make the claims service more efficient and transparent, enhancing the customer experience and satisfaction. Customers can query the case status in real time through the system or relevant channels to understand the repair progress, enhancing the traceability and credibility of the service. Systematic management provides guarantee for the electric vehicle claims process and a foundation for information management.

[0031] In one embodiment of the present invention, S13 includes:

[0032] The system performs unified standardization and normalization processing on each field in the case; uses natural language processing technology to normalize the terms in the case description, unifying fault problems with different expression methods into standardized descriptions;

[0033] Trains historical case data through machine learning algorithms to generate an automatic classification model; the system automatically performs preliminary classification according to the case content;

[0034] Based on the preliminary classification of the case, automatically generates labels related to the case based on data mining technology; and sets priorities for each label; wherein, the priorities are determined based on physical quantities such as historical processing time, customer importance, and fault urgency;

[0035] By setting label verification rules, the system will automatically detect the compliance of labels, mark and prompt incorrect labels; the system can quickly identify and update relevant labels in real time without reclassifying all case data;

[0036] Introduce tag priority physical quantities (such as priority scores) and case data segmentation physical quantities (such as complexity indices), combine the tag priority scores with the case data complexity indices to generate a comprehensive score; according to the comprehensive score, allocate processing tasks to multiple containers;

[0037] Use a distributed computing framework to segment the case data into multiple batches; the classification and tagging processes are independently executed on multiple nodes; utilize cloud computing resources for elastic expansion and dynamically adjust the computing resource allocation according to the comprehensive score;

[0038] Monitor the processing status and efficiency of each container in real time; according to the feedback loop, continuously optimize the comprehensive scoring system and task allocation strategy.

[0039] The working principle of the above technical solution is as follows: First, the system standardizes and normalizes each field in the case data (such as vehicle type, fault type, customer information, etc.). All case data is converted according to a preset unified format to avoid parsing errors caused by inconsistent data formats. For descriptions containing different terms or different expressions (such as "battery failure" and "accumulator problem"), the system normalizes them through natural language processing technology and converts these expressions into a unified standard description, making the data easier to process. In this way, the system can correct spelling mistakes or ambiguous terms that may appear in the input data; The system uses machine learning algorithms (such as decision trees, random forests) to train historical case data. Based on the trained classification model, the system can automatically classify cases according to their content (such as fault type, repair content, etc.). The classification criteria are dynamically updated. Each time the system receives new case data, it will use the new data to further optimize and update the classification model to ensure classification accuracy. On the basis of case classification, the system uses data mining technology to automatically generate tags related to the case. These tags not only cover the direct information of the case content, but also extract potential associated information, such as the relationship between a certain battery failure and a specific repair tool, or the relationship between some faults and factors such as customer needs and repair cycles. Through these tags, the system can more accurately describe the characteristics of the case and improve the efficiency of subsequent processing; At the same time, the system uses a graph neural network (GNN) to represent the complex relationships between case data by constructing a graph structure. For example, the association between cases can be represented by the connection of nodes and edges in the graph. GNN can learn the characteristics of case data through the graph structure, thereby further optimizing the accuracy of the tags; Once the case is classified and tags are generated, the system will perform a tag compliance check. The system sets strict rules. For example, if a case belongs to the "battery failure" category, it must contain the tag "battery problem". If there is a non-compliant tag situation, the system will automatically detect and prompt an error to ensure the accuracy and standardization of the tags. In this way, users can ensure that the tags of each case meet the predetermined standards. The system dynamically adjusts the priority of tags according to factors such as the importance and frequency of occurrence of the tags. When a case contains multiple tags, the system preferentially displays the tags most relevant to the case content. For example, if a case involves a battery failure and requires a specific tool, the "battery failure" tag may have a higher priority. By using the TF-IDF weight model, the system can calculate the weight of each tag and dynamically adjust the display order of the tags according to the weight. This can avoid the display of redundant tags and improve the processing efficiency of the system; During the tag priority adjustment process, the system will set a priority score for each tag, and this score is determined based on physical quantities such as historical processing time, customer importance, and fault urgency.The historical processing time reflects the average time taken to process cases related to the label. A shorter processing time may indicate that the business corresponding to the label is relatively simple, and the priority can be adjusted appropriately; the customer importance is determined based on factors such as the customer's level and cooperation history. The priority of case labels for important customers will be increased accordingly; the urgency of the fault is comprehensively evaluated based on factors such as the impact of the fault on vehicle use and the urgency of customer feedback. Labels related to urgent faults have a higher priority. The system uses an incremental update mechanism to handle changes in case data. When the data in a case changes (such as a modification of the fault type or an adjustment of the repair service), the system can quickly identify and update the relevant labels without having to reclassify all historical cases. This incremental update ensures the efficiency and real-time nature of the system, enabling the system to respond promptly to new data changes; to efficiently process a large amount of case data, the system uses a distributed computing framework (such as Apache Spark or Flink). During the data partitioning process, the system introduces a physical quantity of label priority (such as a priority score) and a physical quantity for case data partitioning (such as a complexity index); the calculation of the case data complexity index comprehensively considers factors such as the number of labels involved in the case, the diversity of the labels (such as whether it contains multiple different types of labels, such as fault type labels, repair tool labels, etc.), the size of the case data (such as the number of rows and fields in the case record), and the frequency of case data changes (the label update time and quantity statistically based on the incremental update mechanism), etc.; combining the label priority score with the case data complexity index generates a comprehensive score. For example, a weighted summation method can be used, and different weights can be set for the label priority score and the case data complexity index according to business requirements. The calculation formula is: Comprehensive score = Label priority score × Weight 1 + Case data complexity index × Weight 2. Based on the comprehensive score, the system distributes the processing tasks to multiple containers. Batches of case data with a higher comprehensive score (usually indicating greater processing difficulty or higher priority) will be assigned to containers with richer computing resources and stronger performance to ensure processing efficiency and quality; the classification and labeling processes are independently executed on multiple nodes. To handle high-load situations, the system uses containerization technology and cloud computing resources, distributes tasks to multiple containers, and uses cloud computing for elastic scaling, thus ensuring smooth operation even during peak load periods. At the same time, the system uses cloud computing resources for elastic scaling and dynamically adjusts the computing resource allocation according to the comprehensive score. For example, when it is detected that a container is processing case data with a high comprehensive score and heavy load, the system will automatically increase the computing resources of that container, such as increasing the number of CPU cores and memory capacity, to ensure the processing speed; the system monitors the processing status and efficiency of each container in real time, including indicators such as processing time and resource utilization. According to the feedback loop, the comprehensive scoring system and task allocation strategy are continuously optimized.For example, if it is found that the case data in a certain comprehensive score range has problems such as overly long processing time or resource waste in actual processing, the system will re-evaluate the comprehensive score calculation method and task assignment rules for this range, and make adjustments and optimizations to improve the overall performance and processing efficiency of the system; in the entire technical solution, special consideration has been given to the blind spot of combined price calculation in the case data segmentation and task assignment links. When segmenting case data, the complexity of combined price calculation is estimated based on product information, regional information, etc. involved in the case. For example, for case data involving complex situations such as multiple product combinations and special regional price policies, the complexity index will increase accordingly. When assigning tasks, relevant tasks are assigned to appropriate containers according to the priority of combined price calculation and the required computing resources. For example, for batches of cases with high priority and complex combined price calculation, they are assigned to containers with richer computing resources and stronger performance to ensure the accuracy and efficiency of combined price calculation. At the same time, in system monitoring, the execution status of combined price calculation tasks, such as calculation time and calculation result accuracy, is monitored in real time to timely adjust the task assignment strategy and ensure the accuracy and uniformity of combined price calculation. Through these measures, the blind spot problem of combined price calculation is effectively solved, and the accuracy and efficiency of business data processing in franchise repair shops are improved.

[0040] The effects of the above technical solutions are as follows: Through standardization and normalization processing, the system ensures the consistency and accuracy of case data, reducing errors caused by inconsistent data formats. Using natural language processing technology to correct ambiguous and incorrect terms in the input data makes the description of fault problems clearer and reduces ambiguity. In most existing systems, field standardization and term matching are mostly carried out through simple rules or preset dictionaries, resulting in a significant reduction in the accuracy of data processing; through a classification model automatically generated by machine learning algorithms, the system can classify new cases in real time, reducing the need for manual intervention and improving work efficiency. As new case data continues to flow in, the classification model can be continuously updated dynamically to ensure the accuracy and adaptability of classification, enabling the system to quickly adapt to changing business requirements; automatically generating tags related to cases enables the system to quickly respond to customer needs and improve service quality; through data mining technology, the system mines the associations between case data and generates tags for potential association information, greatly enhancing the richness and practicality of the tags. Many existing systems rely on manually defined rules for case classification and tag generation, lacking the ability of automatic learning and optimization, while this solution can automatically learn from data and discover hidden associations, providing more valuable information for business decisions. Using GNN to optimize tag accuracy improves the accuracy and effectiveness of the tags; the tag compliance check mechanism ensures that the generated tags meet the preset standards, reducing subsequent problems caused by non-compliant tags and enhancing the credibility of the system. The automatic detection and prompt function for incorrect tags helps to correct potential errors in a timely manner and ensure the stability of the system operation; the system's strict control over tag compliance makes the entire data processing process more standardized and reliable; the system dynamically adjusts tag priorities according to the importance and frequency of occurrence of tags to ensure that the most relevant information is presented first, improving the user experience; adopting a distributed computing framework, the system can process case data in parallel on multiple nodes, significantly improving the processing speed, especially performing excellently in the face of large-scale data. And by dividing batches according to tag priorities, data with higher priorities can be processed first, improving the processing efficiency and enabling more flexible response to large-scale case data. By splitting case data into multiple batches and processing them in parallel on different nodes, the system can make full use of computing resources and greatly shorten the data processing time. The elastic expansion ability of cloud computing resources ensures that the system can still operate efficiently under high load, enhancing the reliability of the system; the system can dynamically adjust the allocation of computing resources according to the actual load situation, avoiding resource waste and performance bottlenecks, and ensuring the stable operation of the system under various load conditions; by introducing tag priorities and case data complexity and integrating them into a scoring system to intelligently allocate processing tasks to multiple containers, not only can it ensure that key cases and important tags are processed first, but also improve the overall processing efficiency and load balancing through dynamic adjustment of resource allocation.

[0041] In one embodiment of the present invention, S2 includes:

[0042] S21. The system, based on the location information and urgency level in the case information, combines the multi-dimensional information of the surveyor. The multi-dimensional information of the surveyor includes the real-time location, credit score, and progress of the tasks at hand of the surveyor. Through a dynamic matching algorithm for surveyors based on real-time geographical location and skill tags, the optimal surveyor is automatically matched to perform the survey task;

[0043] S22. The matched surveyor receives the survey task through the APP, views the detailed case information and location information, and plans a reasonable survey route to the survey site;

[0044] S23. The surveyor conducts a detailed survey at the survey site. The detailed survey includes taking photos of the damaged vehicle, recording the survey situation, and collecting relevant evidence, and generating a survey result;

[0045] S24. The surveyor uploads the survey result through the APP, and the system conducts a preliminary review and verification of the survey result.

[0046] The working principle of the above technical solution is as follows: The system extracts location information (such as the damage assessment location) and the degree of urgency from the case information; at the same time, the system obtains multi-dimensional information of all available surveyors, including real-time location, credit score, and the progress of tasks at hand; the system uses a dynamic matching algorithm for surveyors based on real-time geographical location and skill tags, comprehensively considering the distance between the real-time location of the surveyor and the survey site, the credit score of the surveyor (reflecting their historical work performance and reliability), and the progress of the tasks at hand of the surveyor (ensuring that the surveyor is not overloaded); calculates the surveyor who best meets the current case requirements, that is, the surveyor who can reach the survey site the fastest, has a high credit score, and has a light workload; the system automatically matches the surveyor with the case and prepares to send the survey task; the matched surveyor receives the survey task notification through their dedicated APP; the surveyor views the detailed case information and location information through the APP, including the specific address of the survey site, contact information, etc.; the surveyor uses the built-in map function of the APP or a third-party navigation service to plan a reasonable survey route; among them, the route planning considers factors such as traffic conditions, distance, time, etc., to ensure that they can reach the survey site efficiently and accurately; after the surveyor arrives at the survey site, they conduct a detailed survey; the survey content includes taking photos of the damaged vehicle, recording the survey situation, including the time, location, cause, and liability division of the survey; the surveyor generates a detailed survey result report according to the survey situation; the survey results include vehicle information (such as vehicle type, brand, model number, vehicle identification number, vehicle identification number, owner's phone number, repair shop phone number, purchase price range, liability division), vehicle photos (such as front, rear, and side photos of the vehicle, vehicle identification number photos, vehicle identification number photos, license plate photos, damaged vehicle photos), vehicle location (such as positioning and related photos, such as landmark photos near the site), preliminary quotation, whether a replacement vehicle is needed, minor personal injuries and property damages (minor personal injury photos and prices, minor property photos and prices), and the surveyor uploads the survey results to the system through the APP; the system conducts a preliminary review and verification of the uploaded survey results; the review content includes the integrity, accuracy, and compliance of the survey results, etc.; if it is found that there are problems or the requirements are not met in the survey results, the system may require the surveyor to supplement or modify them.

[0047] The effects of the above technical solution are as follows: Through the dynamic matching algorithm of surveyors based on real-time geographical location and skill tags, the system can automatically match the optimal surveyor to perform the survey task, shortening the response time of the surveyor, improving the survey efficiency, and enhancing the flexibility of surveyor management and scheduling; The surveyor receives the task through the APP and views the detailed case information and location information, and can quickly plan a reasonable survey route, further reducing the time to reach the survey site, thereby improving the efficiency of the overall work process; The system comprehensively considers multi-dimensional information such as the real-time location, credit score, and progress of the tasks at hand of the surveyor, ensuring that the matched surveyor not only has the ability to respond quickly, but also has good work performance and reliability, optimizing the balance of resource allocation and tasks, and enhancing the user experience and trust; The surveyor conducts a detailed survey at the survey site, including taking vehicle damage photos that meet the specifications, recording the survey situation, and collecting relevant evidence, generating a detailed survey result, providing an accurate and reliable basis for subsequent repairs and claims, improving the work quality and response speed, and enhancing the system reliability; By automatically matching surveyors through the system, the subjectivity and uncertainty of manual resource allocation are avoided, the optimal allocation of resources is achieved, and fairness is enhanced; The system can dynamically adjust the resource allocation strategy according to the real-time location and task progress of the surveyor to ensure that the survey task can be completed efficiently and accurately; Quick survey response and accurate survey results can improve customer satisfaction and trust; Customers can query the survey progress and results in real time through the system or relevant channels, enhancing the transparency and traceability of the service; The survey results uploaded by the surveyor through the APP contain rich vehicle information, photos, and location data, providing strong support for the system's data management and analysis; The system can mine and analyze these data to understand information such as survey types, vehicle damage conditions, and surveyor work efficiency, providing a data basis for decision-making support.

[0048] In one embodiment of the present invention, the S21 includes:

[0049] S211. First, the system parses the case information, extracts the specific location where the survey occurs and the urgency of the case;

[0050] S212. The system collects the multi-dimensional information of all available surveyors in real time, and dynamically adjusts the screening results based on factors such as traffic conditions and estimated arrival time;

[0051] The multi-dimensional information of the surveyor includes:

[0052] Real-time location: Obtain the current location of the surveyor through GPS or other positioning technologies;

[0053] Credit score: A score calculated based on indicators such as the surveyor's historical work performance, customer feedback, and on-time rate;

[0054] Progress of the task at hand: The status of the task currently being processed by the surveyor, including accepted, in progress, nearly completed, etc.;

[0055] Skill tags: The professional skills and experience areas of the surveyor (such as vehicle type, experience in handling survey types, etc.);

[0056] S213. The system preliminarily filters out all available surveyors within a certain radius based on the survey location information;

[0057] S214. For the preliminarily filtered surveyors, the system further matches them according to the urgency of the case and the skill tags of the surveyors; moreover, during the matching process, the workload balance of the surveyors is further considered through algorithm optimization, and a reasonable allocation is made between the survey tasks and the surveyors;

[0058] S215. The system uses a dynamic matching algorithm based on real-time geographical location and skill tags, comprehensively considers all the above factors, selects the optimal surveyor to execute the survey task; and optimizes the matching strategy in real time according to historical data and real-time situations;

[0059] S216. The system sends a survey task notice to the selected optimal surveyor, including information such as the survey location, urgency, and estimated arrival time; the surveyor receives the notice through the APP and confirms whether to accept the task; if the surveyor is unable to accept the task due to reasons, the system automatically transfers to the next alternative surveyor.

[0060] The working principle of the above technical solution is as follows: First, the system analyzes the case information, extracts the specific location for on-site inspection, including detailed information such as longitude and latitude, address, etc., and at the same time extracts the urgency level of the case, such as normal, urgent, very urgent, etc.; the determination of the urgency level may be based on the specific needs of the customer, such as the customer needs to go to work as soon as possible, is about to be late, etc., which is defined as urgent; the system obtains the current positions of all available surveyors in real time through GPS or other positioning technologies; based on indicators such as the historical work performance, customer feedback, and on-time rate of the surveyors, the system calculates and updates the credit score of each surveyor; the system tracks and records the task status that the surveyor is currently handling, including accepted, in progress, about to be completed, etc.; assigns tags for the professional skills and experience areas to each surveyor, such as vehicle type, experience in handling on-site inspection types, etc.; the system preliminarily screens out all available surveyors within a certain radius according to the on-site inspection location information; this radius may be dynamically adjusted according to factors such as the on-site inspection type, urgency level, etc. For example, if it is a minor on-site inspection and the customer is not in a hurry, the radius may be larger, while if the on-site inspection is more serious and the customer is very anxious, the radius will be correspondingly reduced; for the preliminarily screened surveyors, the system further matches them according to the urgency level of the case and the skill tags of the surveyors; cases with a high urgency level are preferentially matched with surveyors with a high credit score, few tasks at hand, and the experience in handling such on-site inspections; during the matching process, the system further considers the workload balance of the surveyors through algorithm optimization, and the system ensures that the on-site inspection tasks are reasonably distributed among the surveyors, avoiding overloading one surveyor while other surveyors are idle; the system uses a dynamic matching algorithm based on real-time geographical location and skill tags, comprehensively considering all the above factors (location, urgency level, credit score, progress of tasks at hand, skill tags, workload balance, etc.), and selects the optimal surveyor to execute the on-site inspection task; the system optimizes the matching strategy in real time according to historical data and real-time situations, and sends an on-site inspection task notice to the selected optimal surveyor, including information such as the on-site inspection location, urgency level, estimated arrival time, etc.; the surveyor receives the notice through the APP and confirms whether to accept the task; if the surveyor is unable to accept the task due to reasons, the system automatically transfers to the next alternative surveyor to ensure that the on-site inspection task can be completed in a timely and accurate manner.

[0061] The effects of the above technical solution are as follows: By parsing case information, the system quickly extracts key information such as the survey location and urgency, providing a basis for the subsequent matching of surveyors, reducing human errors and subjective biases, and enhancing the intelligence and automation of the system; It collects multi-dimensional information of surveyors in real time, including real-time location, credit score, progress of tasks at hand, and skill tags, enabling the system to comprehensively understand the status and capabilities of surveyors, improving the accuracy of task allocation, increasing work efficiency while reducing task delays; Through a dynamic matching algorithm based on real-time geographical location and skill tags, the system can comprehensively consider various factors, quickly select the optimal surveyor to perform the survey task, improving the efficiency and accuracy of matching, and at the same time enhancing the transparency and trust of the system; The system matches according to the urgency of the survey and the skill tags of the surveyor, ensuring that cases with a high degree of urgency can be processed first, and at the same time matching surveyors with corresponding processing experience, which can efficiently optimize the task processing process, improve the response speed and processing quality, and reduce delays and errors; During the matching process, the system optimizes the workload balance of surveyors through algorithm optimization, avoiding overloading one surveyor while other surveyors are idle, achieving optimal allocation of resources, and at the same time avoiding over-disposing surveyors with mismatched skills or delaying the processing of urgent tasks, thereby reducing the additional costs caused by task delays or incorrect processing; By quickly and accurately matching surveyors, the survey response time is shortened, improving the service experience of customers; The system can give priority to processing urgent tasks according to the urgency of the case, meeting the urgent needs of customers and enhancing customer satisfaction; The system collects multi-dimensional information of surveyors in real time, making the task allocation of surveyors more reasonable and reducing unnecessary waiting and idle time; By receiving survey task notifications through the APP, surveyors can quickly understand the task details and plan a reasonable survey route, improving work efficiency; The system optimizes the matching strategy in real time based on historical data and real-time situations, continuously improving the matching algorithm, and enhancing the adaptability and flexibility of the system; By collecting and analyzing the work data of surveyors and customer feedback, the system can provide support for decision-making, helping enterprises optimize the survey process and improve service quality.

[0062] In one embodiment of the present invention, the S215 includes:

[0063] Construct a dynamic matching model for multi-dimensional information; Based on machine learning algorithms, train historical survey task data to learn the contribution degree of different dimensional information to the success rate of survey tasks; Based on the learning results, construct an evaluation model;

[0064] Collect and process the real-time data of surveyors and the real-time change information of the urgency of cases in real time through data stream processing algorithms;

[0065] Based on the data fusion algorithm, integrate the real-time data and change information collected and processed in real time into a unified surveyor status vector and case requirement vector;

[0066] In the evaluation model, based on the survey location information and the real-time location of the surveyor, conduct a preliminary match to screen out available surveyors within a certain radius;

[0067] For the preliminarily screened surveyors, further conduct in-depth matching according to the urgency of the case, the skill tags of the surveyors, the credit score, and the progress of the tasks at hand;

[0068] Based on the multi-objective optimization algorithm, solve the matching problem;

[0069] According to historical data and real-time situations, adjust and optimize the matching strategy in real time; based on the feedback mechanism, collect the data and results during the execution of the survey task, continuously learn and optimize the evaluation model; and based on online learning, update and adjust the evaluation model in real time;

[0070] Based on the preset strategy adjustment threshold, when the data changes reach a certain degree, trigger the strategy adjustment process.

[0071] The working principle of the above technical solution is as follows: Build a dynamic matching model that integrates multi-dimensional information such as real-time geographical location, skill tags, credit scores, and progress of tasks at hand; Use machine learning algorithms to train historical survey task data to learn the contribution degrees of different dimensions of information (such as geographical location, skills, credit, task progress) to the success rate of survey tasks; Based on these learning results, build an evaluation model that can evaluate the matching degree between surveyors and cases in real time. The evaluation model can comprehensively consider multiple factors and provide an accurate basis for the subsequent matching process; Real-time data of surveyors, including GPS locations, credit score updates, changes in progress of tasks at hand, and skill tag adjustments, as well as information on changes in the real-time urgency of cases, are collected and processed through data stream processing algorithms; These data are obtained in real time through API interfaces or data stream processing platforms (such as Kafka, Flink, etc.) and are processed to ensure the accuracy and timeliness of the data; Based on data fusion algorithms, the real-time data and change information collected and processed in real time are integrated into a unified surveyor status vector and case single demand vector; The surveyor status vector contains information such as the real-time location, skills, credit, and task progress of the surveyor, while the case demand vector contains information such as the urgency of the case and required skills; In the evaluation model, based on the survey location information and the real-time location of the surveyor, a preliminary match is made to screen out available surveyors within a certain radius; Geospatial indexing techniques (such as R-trees or Quad-trees) are used to accelerate the matching process of geographical locations and improve the matching efficiency; The screening radius threshold increases step by step. First, start matching from a smaller radius (such as three kilometers). If there are no suitable surveyors, gradually expand the radius (such as four kilometers, five kilometers) until a match is found; For the preliminarily screened surveyors, further in-depth matching is carried out according to the urgency of the case, the skill tags of the surveyor, the credit score, and the progress of tasks at hand; Based on multi-objective optimization algorithms, reasonable weight coefficients are set to balance the relationship between urgency, skill matching degree, credit score, and progress of tasks at hand, and solve the matching problem; At the same time, a load balancing constraint is introduced to ensure the reasonable distribution of survey tasks among surveyors and avoid overloading some surveyors while some surveyors are idle; According to historical data and real-time situations, the matching strategy is adjusted and optimized in real time; Based on the feedback mechanism, data and results during the execution of survey tasks are collected for the continuous learning and optimization of the evaluation model; Online learning techniques are used to update and adjust the evaluation model in real time to adapt to the changing environment and requirements; Based on a preset strategy adjustment threshold, when the data changes reach a certain degree (such as fluctuating by ±10%), trigger the strategy adjustment process; When the system detects significant changes in the data, it will automatically adjust the matching strategy to ensure the optimization of the matching effect.

[0072] The effects of the above technical solutions are as follows: By constructing a dynamic matching model that integrates multi-dimensional information such as real-time geographical location, skill tags, credit scores, and the progress of on-hand tasks, it is possible to more comprehensively evaluate the matching degree between surveyors and cases. Traditional matching algorithms often only consider data from a single dimension (such as distance or time), lacking the integration of multi-dimensional information, resulting in one-sided matching results and being unable to fully reflect the comprehensive requirements of tasks and surveyors. The application of machine learning algorithms enables the model to learn from historical data the contribution degrees of different dimensions of information to the success rate of survey tasks, thereby more accurately evaluating the matching effect, improving the accuracy of matching, enabling survey tasks to be more quickly assigned to the most suitable surveyors, saving the overall processing time, and improving processing efficiency. Existing systems usually adopt rule-based matching methods, relying too much on simple metrics such as distance and time to screen surveyors, ignoring factors such as the abilities of surveyors, historical performance, and the specific requirements of tasks, resulting in the inability to match the best surveyors and potentially affecting the efficiency of subsequent survey tasks. The real-time data collection and processing mechanism ensures that the latest information on the status of surveyors can be obtained and processed in a timely manner, providing a real-time and accurate data basis for the matching process. The matching strategies of traditional systems are often static, do not adjust in real time according to changes during task execution, and do not have the ability of online learning, making it difficult to gradually optimize in long-term use. The data fusion algorithm integrates real-time data into a unified surveyor status vector and case requirement vector, simplifying the complexity of the matching process. The application of geospatial indexing technology accelerates the matching process of geographical locations, further enhancing the survey efficiency. The mechanism of gradually increasing the screening radius threshold enables the system to find suitable surveyors within different ranges, enhancing the flexibility of the system. When there are no suitable surveyors within a small range, the system will automatically expand the search range until a successfully matched surveyor is found, improving the survey efficiency. The application of the multi-objective optimization algorithm enables the system to solve the matching problem based on considering multiple factors (such as urgency, skill matching degree, credit score, and the progress of on-hand tasks). The set weight coefficients balance the relationships between these factors, ensuring the rationality of the matching results. The introduction of load balancing constraints avoids uneven task distribution among surveyors and optimizes resource allocation. Based on historical data and real-time situations, the system can adjust and optimize the matching strategy in real time, enabling the strategy to adapt to changing environments and requirements. The feedback mechanism ensures that the data and results during the execution of survey tasks can be collected in a timely manner and used for the continuous learning and optimization of the model. The application of online learning technology enables the evaluation model to be updated and adjusted in real time, maintaining the accuracy and effectiveness of the model. The preset strategy adjustment threshold ensures that when the data changes reach a certain extent, the system will trigger the strategy adjustment process, avoiding system instability caused by abnormal or excessive data changes. It improves the stability and reliability of the system, enabling the system to operate normally under various circumstances.

[0073] In one embodiment of the present invention, step S3 includes:

[0074] S31. The system automatically generates a repair order according to the survey result;

[0075] S32. The system automatically matches the corresponding affiliated repair shop through a multi-dimensional store capability profiling model by combining the location information and repair item type in the repair case with the multi-dimensional information of each affiliated repair shop in the area;

[0076] S33. The system pushes the repair order to the affiliated repair shop that is matched.

[0077] The working principle of the above technical solution is as follows: After the surveyor completes the on-site survey, the survey result is entered into the system; the system automatically generates a repair order according to the survey result; the repair order details the repair items (such as replacing parts, repairing damaged parts, etc.), required accessories (such as specific part names, models, quantities, etc.), and estimated repair time, etc.; for minor surveys such as minor friction, with little loss and clear liability division, if the owner of the three-wheeled electric vehicle needs to handle it quickly, then, according to the relevant photos and video information of the previous survey, online quick claims are made; the system starts to screen the affiliated repair shops in the area according to the location information (such as the survey location or the customer-specified repair location) and repair item type (such as the brand of the electric vehicle, specific damaged parts or required repair services) in the repair order; the system combines the multi-dimensional information of each affiliated repair shop in the area, including geographical location (distance from the repair location), historical repair success rate (reflecting the technical level and service quality of the affiliated repair shop), stability of the parts supply chain (ensuring that the required parts can be supplied in a timely manner), and user evaluations (reflecting customer satisfaction with the affiliated repair shop), etc.; through the multi-dimensional store capability profiling model, these information are comprehensively analyzed to automatically match the most suitable affiliated repair shop to undertake the repair order; the system pushes the successfully matched repair order to the selected affiliated repair shop; if the owner of the electric vehicle has a self-selected repair shop (including non-affiliated repair shops), then the self-selected repair shop of the owner is preferably selected for repair at the same price. For example, if the final determined claim amount is N, and the self-selected repair shop of the owner can accept the final claim amount, then the customer's self-selected repair shop is preferably selected for repair by the affiliated repair shop; at the same time, the system provides detailed information such as order details (such as repair items, required accessories, estimated repair time, etc.), repair requirements (such as technical standards, operation procedures, etc.), and parts requirements (such as specific part names, models, quantities, and arrival time requirements).

[0078] The effects of the above technical solution are as follows: The system automatically generates a maintenance order based on the survey results, avoiding the cumbersome process of manual entry of order information, greatly shortening the order generation time, and improving work efficiency; through the multi-dimensional store capability portrait model, the corresponding franchised maintenance shops are automatically matched, enabling the maintenance order to be quickly and accurately assigned to the most suitable franchised maintenance shop, further shortening the response time of the maintenance service; the system pushes the maintenance order and detailed information to the franchised maintenance shop, and the franchised maintenance shop can quickly understand the task details and make preparations for maintenance, thus accelerating the maintenance progress; the system combines multi-dimensional information of each franchised maintenance shop in the region for matching, ensuring the reasonable allocation of maintenance resources (such as technology, parts, manpower, etc.); by considering factors such as the geographical location of the franchised maintenance shop, historical maintenance success rate, stability of the parts supply chain, and user evaluations, the system can select the most capable and reputable franchised maintenance shop to undertake the maintenance task, improving the resource utilization efficiency; the automatically generated maintenance order contains information such as detailed maintenance items, required parts, and estimated maintenance time, providing clear maintenance guidance for the franchised maintenance shop and ensuring the accuracy and consistency of the maintenance service; through matching with the multi-dimensional store capability portrait model, the system can select franchised maintenance shops with high technical levels and good service quality, thus enhancing the overall service quality; the fast maintenance response time and high-quality maintenance service can enhance customer satisfaction and trust; the systematic management method makes the maintenance service more transparent and traceable, and customers can understand the maintenance progress and status at any time, improving the transparency and credibility of the service; the system collects and analyzes multi-dimensional information of the franchised maintenance shop, providing data support for the enterprise's decision-making; by analyzing data such as the historical maintenance success rate, stability of the parts supply chain, and user evaluations of the franchised maintenance shop, the enterprise can understand the operation status and service quality of the franchised maintenance shop, providing a decision-making basis for the subsequent selection and management of franchised maintenance shops.

[0079] In an embodiment of the present invention, the S4 includes:

[0080] S41. Determine the claim amount and the undertaking maintenance shop in combination with each quotation information;

[0081] S42. Monitor the maintenance process and control the quality;

[0082] S43. Claim settlement.

[0083] In an embodiment of the present invention, the S41 includes:

[0084] S411. After each franchised maintenance shop receives the maintenance order, it makes a quotation according to the information such as the maintenance items and parts requirements provided by the system;

[0085] S412. During the quotation process, the system adopts a distributed quotation isolation mechanism based on privacy computing to physically isolate and store the quotation data of each affiliated repair shop.

[0086] S413. The system conducts a comprehensive evaluation based on the merchant portrait and the merchant's quotation, and automatically calculates and determines the claim amount and the affiliated repair shop undertaking the repair according to the corresponding algorithm.

[0087] The working principle of the above technical solution is as follows: After each affiliated repair shop receives the repair order pushed by the system, it will make a detailed quotation based on the repair items, required parts, and estimated repair time provided in the order, combined with its own costs, technical level, and market conditions; the quotation content usually includes parts costs, labor costs, and other possible additional costs, etc.; during the quotation process, the system adopts a distributed quotation isolation mechanism based on privacy computing to ensure that the quotation data of each affiliated repair shop is physically isolated and stored; moreover, the quotation data of each affiliated repair shop is stored independently and will not be accessed or tampered with by other affiliated repair shops or unauthorized personnel; the system ensures the security and privacy of the quotation data through encryption algorithms and distributed storage technologies; the system conducts a comprehensive evaluation based on the merchant portrait (including information such as historical repair success rate, user evaluation, and parts supply chain stability) and the merchant's quotation; the merchant portrait reflects the service quality, technical level, and market reputation of the affiliated repair shop and is used to evaluate the capabilities of the affiliated repair shop; the system comprehensively analyzes the merchant portrait and quotation data through algorithms, and automatically calculates and determines the claim amount and the affiliated repair shop undertaking the repair; the optimal quotation and the repair shop undertaking the repair are usually based on a comprehensive consideration of multiple dimensions such as price, service quality, technical level, and parts supply chain stability. The quotations of each repair shop provide a basis for the claim amount of the case, corroborating the rationality, objectivity, and scientificity of the final claim amount.

[0088] The effects of the above technical solutions are as follows: By adopting a distributed quotation isolation mechanism based on privacy computing, the quotation data of each affiliated repair shop is physically isolated and stored, effectively preventing data leakage and abuse, ensuring the security of the business secrets and quotation data of the affiliated repair shops, safeguarding the legitimate rights and interests of the affiliated repair shops, and enhancing the security and compliance of the system; The transparency of the quotation process and the isolation storage mechanism prevent malicious competition and price manipulation, creating a fair and just competition environment for the repair market; Each affiliated repair shop can make independent quotations based on its own costs, technical levels, and market conditions, without being affected by other affiliated repair shops, improving the rationality of the quotations; The information such as repair items and parts requirements provided by the system provides an accurate basis for the affiliated repair shops to make quotations, reducing quotation deviations caused by incomplete or incorrect information; The physical isolation storage of the quotation data ensures the originality and authenticity of the quotations, improving the reliability of the quotations; The system comprehensively evaluates based on the merchant portrait and the merchant's quotation, and can automatically calculate and determine the claim amount and the affiliated repair shop undertaking the repair according to the corresponding algorithm, safeguarding the rights and interests of customers and insurance companies; This evaluation mechanism considers multiple dimensions, such as historical repair success rate, user evaluations, and the stability of the parts supply chain, etc., ensuring that orders can be assigned to the affiliated repair shops with the most capabilities and credibility, improving the order allocation accuracy, enhancing the competitiveness among affiliated repair shops, and ensuring the smoothness of the insurance claim process; Through a fair and transparent quotation mechanism and an optimized order allocation strategy, customers can obtain the most cost-effective repair service plan, reducing repair costs, improving the quality and efficiency of repair services, and enhancing customers' satisfaction and trust in repair services; The information such as quotation data and merchant portraits collected and analyzed by the system provides data support for the enterprise's decision-making, and the quotations of each repair shop provide a basis for the claim amount of the case, corroborating the rationality, objectivity, and scientificity of the final claim amount. During the entire loss assessment process, the roles of survey and loss assessment are completely separated, reflecting the rationality and objectivity of the loss assessment claim amount.

[0089] In one embodiment of the present invention, the S412 includes:

[0090] Deploy multiple privacy computing nodes on the server side of the system, and each node is responsible for processing the quotation data from different affiliated repair shops;

[0091] Adopt a distributed storage system to store the quotation data from each affiliated repair shop, divide the storage system into multiple independent storage areas, each area corresponding to an affiliated repair shop for physical isolation of data;

[0092] Before data storage, encrypt the quotation data through symmetric encryption or asymmetric encryption algorithms, store the encrypted data in the corresponding storage area, and record the usage of the encryption key as one of the privacy metrics;

[0093] The system assigns the privacy computing tasks of the quotation data to appropriate privacy computing nodes according to the current processing capacity, the load conditions of the privacy computing nodes, and the privacy protection requirements;

[0094] After receiving the task, the privacy computing node processes the quotation data according to predefined algorithms and protocols, and monitors the monitoring metrics during the processing, such as the data access volume and the encryption key usage frequency;

[0095] Based on the data synchronization mechanism, the quotation data of each affiliated repair shop in the system is synchronously transferred to the distributed storage system in real time; during the data synchronization process, an encrypted transmission protocol is used to encrypt the data transmission process;

[0096] Monitor the load conditions and privacy metrics of the privacy computing nodes in real time (such as the encryption key usage frequency and the privacy data access volume), and dynamically adjust the task allocation strategy according to the load conditions and privacy metrics;

[0097] When the load of a certain node is too high or the privacy metrics are abnormal, transfer some tasks to other nodes with lower load or stronger protection capabilities;

[0098] Based on the load warning mechanism, when the overall system load approaches the preset threshold, trigger a warning and take corresponding measures, such as automatically adjusting the task allocation or starting the capacity expansion process;

[0099] Based on the storage resource management mechanism, uniformly manage and schedule the storage resources in the distributed storage system; monitor the usage of storage resources in real time; dynamically adjust the allocation strategy of storage resources according to the usage of storage resources and the privacy protection requirements.

[0100] The working principle of the above technical solution is as follows: The system deploys multiple privacy computing nodes on the server side. These nodes adopt the Trusted Execution Environment (TEE) technology and combine security measures such as Hardware Security Module (HSM) to provide comprehensive security guarantees for data computing. Each node is responsible for processing the quotation data from different affiliated repair shops. Through strict access control and permission management, data isolation and processing are achieved to prevent data leakage and cross-contamination. A distributed storage system is used to store the quotation data from each affiliated repair shop. The storage system is divided into multiple independent storage areas, with each area corresponding to an affiliated repair shop. Through a combination of physical isolation and logical isolation, it is ensured that the data of different affiliated repair shops are independent of each other and do not interfere with each other. Before data storage, the quotation data is encrypted through an encryption algorithm. An appropriate encryption algorithm and key length are selected according to the sensitivity and security requirements of the data. For example, 128 bits are used for particularly sensitive data to ensure data confidentiality. The encrypted data is stored in the corresponding storage area, and only the corresponding affiliated repair shop or a strictly authorized system administrator can access it. At the same time, the usage of the encryption key is recorded, including the key generation time, usage time, usage frequency, etc., as one of the privacy metrics to facilitate the monitoring and management of the key usage. The system allocates the privacy computing tasks of the quotation data to appropriate privacy computing nodes according to the current processing capacity, the load conditions of the privacy computing nodes, and the privacy protection requirements. In the task allocation process, factors such as the computing performance, storage capacity, and network bandwidth of the nodes, as well as the privacy level and processing complexity of the data, are comprehensively considered to ensure that the tasks can be executed efficiently and securely. After receiving the task, the privacy computing node processes the quotation data according to the predefined algorithm and protocol. During the processing, in addition to encryption processing, anonymization technology is also used to process the data to remove personal identity information and sensitive information in the data, further protecting data privacy. At the same time, monitoring metrics during the processing, such as data access volume, encryption key usage frequency, data processing time, etc., are monitored to promptly detect and handle abnormal situations. During the processing, data version control and consistent hashing algorithms are adopted to ensure data consistency between different nodes. A unique version number is assigned to each data object to record the modification history and status of the data, facilitating data traceability and recovery. Through the consistent hashing algorithm, the data is evenly distributed to different nodes. When a node fails or the load changes, the data distribution can be quickly and accurately adjusted to ensure data availability and consistency.Regularly perform consistency checks on the data in the distributed storage system, verify the data using methods such as checksums and hash values, promptly detect and fix data inconsistency issues, and ensure the accuracy and reliability of the data; Based on the data synchronization mechanism, synchronize the quotation data of each affiliated repair shop in the system to the distributed storage system in real time to ensure the timeliness and consistency of the data; During the data synchronization process, use an encrypted transmission protocol (such as TLS / SSL) to ensure the security of the data during transmission and prevent the data from being stolen or tampered with. At the same time, monitor and record the data synchronization process, including information such as synchronization time, synchronized data volume, and synchronization status, for tracking and analyzing the data synchronization situation; Real-time monitor the load situation and privacy metrics of the privacy computing nodes (such as the frequency of encryption key usage and the access volume of private data), and dynamically adjust the task allocation strategy according to the load situation and privacy metrics. When the load of a certain node is too high or the privacy metrics are abnormal, transfer some tasks to other nodes with lower load or stronger protection capabilities to achieve balanced load distribution and optimization of privacy protection. Based on the load warning mechanism, when the overall system load approaches the preset threshold, trigger a warning and take corresponding measures, such as automatically adjusting the task allocation or starting the expansion process. By setting reasonable load thresholds and warning rules, promptly detect the load problems of the system and take effective measures to solve them to ensure the stability and availability of the system; Based on the storage resource management mechanism, uniformly manage and schedule the storage resources in the distributed storage system. Real-time monitor the usage of storage resources, including indicators such as storage space and I / O performance, promptly detect and handle storage resource problems. Dynamically adjust the storage resource allocation strategy according to the usage of storage resources and privacy protection requirements. For example, when the storage space of a certain storage area is insufficient, automatically expand the storage capacity; when the I / O performance of a certain storage node decreases, adjust the data distribution and migrate the data to a node with better performance.

[0101] The effects of the above technical solutions are as follows: By deploying multiple privacy computing nodes on the server side of the system and adopting the Trusted Execution Environment (TEE) technology, the confidentiality and integrity of data during the calculation process are ensured, effectively preventing data leakage and tampering; Selecting a suitable privacy computing protocol, such as multi-party secure computing or federated learning, further enhances the privacy protection ability of the quotation data of franchise repair shops, enabling the data to be fully protected during transmission and calculation. Most existing privacy computing solutions adopt data encryption and privacy protection protocols, but they are often limited to a single link (such as data storage or calculation), rather than throughout the entire data processing chain, resulting in insufficient protection for the entire life cycle of data and increasing the risk of data leakage; Using a distributed storage system to store the quotation data from each franchise repair shop and dividing the storage system into multiple independent storage areas, with each area corresponding to a franchise repair shop, realizes physical isolation of data, effectively preventing cross-interference and leakage of data; Before data storage, the quotation data is encrypted through symmetric encryption or asymmetric encryption algorithms, further improving data security and ensuring that only the corresponding franchise repair shop or authorized system administrator can access the data; The system can dynamically allocate the privacy computing tasks of the quotation data to appropriate privacy computing nodes according to the current processing capacity and the load conditions of the privacy computing nodes, achieving flexible task allocation and efficient processing. Existing privacy computing solutions usually adopt a centralized architecture, with all data processing and storage concentrated on a few nodes, which may lead to performance bottlenecks during the calculation process, further increasing the instability of the system; During the privacy computing process, data version control and consistent hashing algorithms are adopted to ensure the consistency of data between different nodes, effectively preventing the occurrence of data inconsistency problems; Regularly check the consistency of the data in the distributed storage system, promptly detect and repair data inconsistency problems, ensuring the accuracy and reliability of the data; Based on the data synchronization mechanism, the quotation data of each franchise repair shop in the system is synchronously transferred to the distributed storage system in real time, ensuring the timeliness and consistency of the data. Although existing storage systems can provide high availability and scalability, they lack functions in privacy protection, and the guarantee of data security is relatively weak; During the data synchronization process, an encrypted transmission protocol (such as TLS / SSL) is used to ensure the security of the data during transmission, effectively preventing the data from being stolen or tampered with during transmission; By real-time monitoring privacy metrics such as the load of privacy computing nodes, the usage frequency of encryption keys, and the access volume of private data, the task allocation can be dynamically adjusted according to the real-time situation, ensuring that the system resources and privacy protection capabilities are always in the best matching state, and improving the overall operation efficiency and privacy security guarantee level of the system;When the node load is too high or the privacy metric is abnormal, automatically transfer some tasks to nodes with lower load or stronger protection capabilities to achieve balanced utilization of node resources, avoid overloading of individual nodes or failure of privacy protection, and improve the stability and reliability of the system; based on the load warning mechanism, trigger a warning in a timely manner when the overall system load approaches the preset threshold, and automatically take measures such as adjusting task allocation or starting the capacity expansion process to effectively prevent the system from experiencing performance degradation or even crashing due to high load, ensure the continuous and stable operation of the system, and enhance the scalability of the system and its ability to handle high loads.

[0102] In one embodiment of the present invention, the S42 includes:

[0103] S421. During the repair process, the system uses the whole-process monitoring technology based on repair progress prediction to monitor and record the whole process from receiving the order to the end of the order through blockchain technology;

[0104] S422. The franchised repair shop regularly uploads photos and progress information of the repair process during the repair process, and the system conducts quality inspection and verification on the uploaded photos and information;

[0105] S423. If an abnormality or quality problem occurs during the repair process, the system will promptly issue a warning and reminder, and the franchised repair shop will handle and adjust after receiving the warning and reminder.

[0106] The working principle of the above technical solution is as follows: The system uses advanced algorithms and models, combines historical maintenance data with factors such as the complexity of the current maintenance project and the supply situation of required parts, to predict the maintenance progress. The franchised maintenance shops can reasonably arrange the maintenance plan according to the prediction results; from receiving the maintenance order to the end of the maintenance process until the order is completed, the system uses blockchain technology for full-process monitoring and recording to ensure the authenticity and traceability of the maintenance process data; the system records the key information during the maintenance process, such as maintenance steps, parts replacement situation, maintenance time nodes, etc., on the blockchain in real time, providing a basis for subsequent maintenance quality assessment and also helping the enterprise to quickly find out the truth in case of disputes; the franchised maintenance shops need to regularly upload photos and progress information of the maintenance process during the maintenance. These photos and information can intuitively reflect the actual situation of the maintenance, such as the maintenance site, replaced parts, maintenance progress, etc.; the system conducts quality inspection and verification on the uploaded photos and information, including checking the clarity of the photos and verifying the integrity of the information, to ensure the authenticity, accuracy, and reliability of the uploaded data; if the system detects quality problems in the uploaded data, such as blurred photos or missing information, it will promptly notify the franchised maintenance shop to re-upload or supplement the information; the system uses preset algorithms and models to monitor the key indicators during the maintenance process in real time. If it is found that the maintenance progress lags behind, the parts replacement is abnormal (for example, non-genuine parts are replaced), the maintenance quality does not meet the standards (for example, there are safety hazards in the vehicle after maintenance), etc., the system will promptly issue a warning and reminder; the system sends the warning and reminder information to the franchised maintenance shop through methods such as message push and email notification. After receiving the warning and reminder, the franchised maintenance shop needs to immediately check and adjust the maintenance process to ensure that the maintenance quality and progress meet the requirements; after dealing with the abnormality, the franchised maintenance shop needs to feedback the processing result to the system. The system will verify the feedback result to ensure that the abnormal problem is properly solved.

[0107] The effects of the above technical solution are as follows: Through the whole-process monitoring technology based on maintenance progress prediction and combined with blockchain technology for full-process monitoring and recording, it ensures that every link of the maintenance process from receiving an order to the end of the order is accurately and tamper-proofly recorded, ensuring the transparency and traceability of the process, and enhancing the trust of all parties in the maintenance process; The franchised maintenance shop regularly uploads photos and progress information of the maintenance process during the maintenance process, and the system conducts quality inspection and verification on these photos and information to ensure the authenticity and accuracy of the uploaded data. The enterprise can timely understand the maintenance progress and actual situation, providing reliable data support for subsequent decision-making; The system can real-time monitor key indicators during the maintenance process. Once anomalies or quality problems are found, such as lagging maintenance progress, abnormal parts replacement, unqualified maintenance quality, etc., it will promptly issue warnings and reminders, enabling the franchised maintenance shop to quickly respond and handle these anomalies or quality problems, ensuring the smooth progress of the maintenance process, improving the maintenance quality and customer satisfaction; Through the combination of the whole-process monitoring technology and blockchain technology, the enterprise can achieve refined management of the maintenance process, optimize the maintenance process, and improve the maintenance efficiency; At the same time, the warning and reminder functions provided by the system also help the enterprise to timely discover and solve potential problems, reducing risks and costs during the maintenance process; The transparent maintenance process, accurate maintenance progress information, and timely handling of anomalies all enhance the trust and satisfaction of customers with the maintenance service; Customers can understand the maintenance progress at any time, have a clearer understanding of the maintenance process, and thus improve their evaluation of the maintenance service.

[0108] In one embodiment of the present invention, the S43 includes:

[0109] S431. After the maintenance is completed, the franchised maintenance shop uploads the maintenance information, and the system audits and verifies the uploaded information;

[0110] S432. If the audit is passed, the system will automatically trigger the automatic claim trigger mechanism based on the smart contract, and automatically make claim payments according to the preset claim rules and conditions;

[0111] S433. During the claim process, the system conducts correlation analysis on the maintenance records, parts codes, and payment vouchers through multi-source data cross-verification technology;

[0112] S434. The system records and statistics the data of the whole process, and generates a situation statistical report on dates, regions, case volumes, etc.;

[0113] S435. Use the prediction model to predict the situation in the next period of time and make preparations in advance;

[0114] S436. After the claim payment is completed, the system will automatically update the case status and notify the vehicle owner to pick up the vehicle or conduct subsequent processing;

[0115] S437. Meanwhile, the system will conduct regular reviews and summaries, and visually display the situation statistics; for example, the case volume for brands, dates, regions, and weather conditions, etc. will be displayed in a visual way such as a line chart.

[0116] The working principle of the above technical solution is as follows: After the repair is completed, the franchised repair shop needs to upload the repair information, including the photos after repair and a detailed repair list; after the system receives this information, it will automatically conduct reviews and verifications. The review content includes the clarity of the photos, the integrity of the repair list, the rationality of the use of parts, etc.; if the information meets the preset standards, the review passes; otherwise, the system will notify the franchised repair shop to supplement or modify; then review again, and after the review passes, the system will automatically trigger the automatic claim settlement trigger mechanism based on the smart contract; the smart contract is a set of pre-set rules and conditions, and when these conditions are met, the contract will be automatically executed; the system automatically conducts claim settlement payments according to the preset claim settlement rules and conditions, such as repair costs, parts prices, insurance terms, etc.; during the claim settlement process, the system conducts correlation analysis on repair records, parts codes, payment vouchers, etc. through multi-source data cross-verification technology; by comparing data from different sources, the system can verify the authenticity and accuracy of the repair information; if data inconsistencies or anomalies are found, the system will conduct further investigations or notify the franchised repair shop to give explanations; the system records and statistics the data of the whole process, including various links such as repair information, claim settlement payment, and data verification; based on these data, the system can generate situation statistics reports on dates, regions, case volume, etc.; these reports provide data support for the enterprise's decision-making, such as resource allocation, business order planning, etc.; the system uses prediction models, such as time series analysis, machine learning algorithms, etc., to predict the situation in a future period; for example, predict the future survey task volume according to historical data and weather conditions; according to the prediction results, the enterprise can make preparations in advance, such as reasonably allocating relevant resources, adjusting work plans, etc.; after completing the claim settlement payment, the system will automatically update the status, such as changing from "under repair" to "completed"; meanwhile, the system will notify the vehicle owner to pick up the vehicle or conduct subsequent processing, such as providing pick-up location, time, etc. information; the system will conduct regular reviews and summaries, evaluate and analyze various links such as repair services and claim settlement processes; meanwhile, the system will visually display the situation statistics, such as displaying the case volume for brands, dates, regions, and weather conditions, etc. through line charts, bar charts, etc.

[0117] The effects of the above technical solutions are as follows: After the repair is completed, the franchised repair shop uploads the repair information, and the system audits and verifies the uploaded information to ensure the accuracy and integrity of the repair information, improving work efficiency, reducing human errors, and ensuring the smooth progress of the claims settlement process; if the audit is passed, the system will automatically trigger the claims automatic trigger mechanism based on smart contracts, and automatically make claims payments according to the preset claims rules and conditions; this automated payment mechanism simplifies the claims process, speeds up the payment speed, and improves customer satisfaction; during the claims settlement process, the system uses multi-source data cross-verification technology to conduct a correlation analysis of repair records, parts codes, and payment vouchers to ensure the authenticity and legality of the claims settlement process; this cross-verification mechanism effectively prevents fraud and protects the interests of insurance companies and customers; the system records and statistics the data of the entire process and generates a situation statistical report on dates, regions, case volumes, etc.; this data recording and statistics mechanism provides valuable data support for insurance companies, helping insurance companies with business analysis and decision-making; through a prediction model, the situation in the next period of time is predicted, and a pre-plan is made in advance, such as predicting the future survey task volume based on historical data and weather conditions, and reasonably allocating relevant resources; this prediction and pre-plan mechanism improves the response ability of insurance companies and ensures the continuity and stability of business; after the claims payment is completed, the system will automatically update the case status and notify the vehicle owner to pick up the vehicle or perform subsequent processing; the timely update and notification mechanism improves the customer experience and reduces the customer waiting time; the system will conduct regular reviews and summaries, and visually display the situation statistics, such as displaying the case volume and weather conditions for brands, dates, regions, etc. through visual methods such as line charts; the review summary and visual display mechanism helps insurance companies discover problems and trends in their business, providing strong support for future business development and decision-making.

[0118] In one embodiment of the present invention, the S435 includes:

[0119] The system first collects and collates past claims data, including multi-dimensional data such as claims case volume, claims amount, survey task volume, weather conditions, and holiday impacts; cleans and preprocesses the collected data;

[0120] Based on historical data, extracts feature variables related to future survey task volume, constructs a feature engineering; trains a machine learning model using historical data;

[0121] Fuses information from different data sources, such as weather forecast data, holiday arrangement data, historical claims data, etc., and calibrates the fused data;

[0122] Use the trained model to predict the survey task volume for a period of time in the future, generate prediction results, and evaluate the prediction results by calculating indicators such as prediction error and accuracy to verify the prediction performance of the model;

[0123] According to the prediction results, formulate a resource allocation strategy, such as adjusting the scheduling plan of survey personnel, deploying survey vehicles, and preparing necessary survey tools and equipment; based on the differences in survey task volumes in different regions and different time periods, optimize the allocation and efficient utilization of resources;

[0124] For possible abnormal situations, formulate an emergency plan, clarify the emergency response process and division of responsibilities; and establish an emergency response mechanism;

[0125] Continuously monitor the prediction model and resource allocation strategy, adjust and optimize in a timely manner according to the actual implementation situation, and regularly evaluate the accuracy of the prediction model and the effectiveness of the resource allocation strategy.

[0126] The working principle of the above technical solution is as follows: The system first collects and collates past claims data, including multi-dimensional information such as the number of claims cases, claim amounts, survey task volumes, weather conditions, and holiday impacts; collects and collates the prices of various electric vehicle components and accessories listed in the quotation links of each repair shop in each case to form a reference database of different accessory price data for different brands of electric vehicles. The data may come from internal systems, external data sources (such as weather forecast services), or manual input; the system conducts preliminary collation and cleaning of this data to remove outliers, missing values, or incorrect data; performs preprocessing on the data, such as data normalization, standardization, discretization, etc.; based on historical data, extracts feature variables related to future survey task volumes, such as historical case volumes in the same period, weather types, holiday distributions, seasonal factors, etc.; constructs feature engineering to enhance the expression ability of these features for survey task volume prediction through data transformation, combination, etc.; uses the processed historical data to train a machine learning model so that the model can learn the relationship between the survey task volume and the feature variables; integrates information from different data sources, such as weather forecast data, holiday arrangement data, historical claims data, etc.; corrects the integrated data to ensure the accuracy and consistency of the data; the correction may include steps such as data alignment, error correction, and filling of missing values; uses the trained model to predict the survey task volume for a future period of time to generate prediction results; evaluates the prediction results by calculating indicators such as prediction error and accuracy to verify the prediction performance of the model; if the prediction performance does not meet the requirements, adjusts the model parameters or retrains the model; based on the prediction results, formulates a resource allocation strategy, such as adjusting the scheduling plan of survey personnel, allocating survey vehicles, and preparing necessary survey tools and equipment; based on the differences in survey task volumes in different regions and different time periods, conducts optimal allocation and efficient utilization of resources; the resource allocation strategy may include dynamic adjustment, reservation of backup resources, and priority allocation of key resources; formulates an emergency plan for possible abnormal situations (such as extreme weather, emergencies, etc.); clarifies the emergency response process and division of responsibilities to ensure rapid response and effective handling in case of abnormal situations; establishes an emergency response mechanism, including a warning system, emergency teams, and material reserves; continuously monitors the prediction model and resource allocation strategy and adjusts and optimizes them in a timely manner according to the actual implementation situation; regularly evaluates the accuracy of the prediction model and the effectiveness of the resource allocation strategy to ensure that they can adapt to the changing environment and requirements; the optimization may include model parameter adjustment, feature engineering improvement, resource allocation strategy adjustment, etc.

[0127] The effects of the above technical solutions are as follows: The system collects and organizes past claim settlement data, including multi-dimensional information, providing a solid data foundation for decision-making, optimizing processes and improving efficiency; by collecting and organizing the prices of various electric vehicle components and accessories listed in the quotation links of each repair shop in each case, a price data reference library for different accessories of different brands of electric vehicles is formed, and the machine learning model is trained using the database, and finally a reference value (AI quotation) of the claim settlement amount for the electric vehicle loss involved in each case is generated. Data cleaning and preprocessing ensure the accuracy and reliability of the data, providing a high-quality data source for subsequent analysis and prediction; extracting feature variables related to future survey task volumes and constructing a feature engineering, through data transformation, combination, etc., enhancing the expression ability of features for survey task volume prediction, improving the accuracy and stability of the prediction model, making the prediction results more reliable, enhancing the robustness and training efficiency of the model; using historical data to train the machine learning model, enabling the model to learn the rules and patterns in the historical data, improving the accuracy of prediction and enhancing the generalization ability of the model; the trained model can accurately predict the survey task volume for a period of time in the future, providing a scientific basis for resource allocation; integrating information from different data sources, such as weather forecast data, holiday arrangement data, historical claim settlement data, etc., improving the comprehensiveness and accuracy of prediction; calibrating the integrated data to further reduce data errors and improve the reliability of prediction results; according to the prediction results, formulating resource allocation strategies, such as adjusting the scheduling plan of survey personnel, allocating survey vehicles, etc., realizing the optimal allocation and efficient utilization of resources, improving the efficiency and quality of survey work, and reducing resource waste; formulating emergency plans and emergency response mechanisms for possible abnormal situations, such as extreme weather, emergencies, etc., ensuring that in the event of abnormal situations, a rapid response and effective disposal can be made, guaranteeing the smooth progress of survey work, and reducing potential risks and losses; continuously monitoring the prediction model and resource allocation strategies, and adjusting and optimizing them in a timely manner according to the actual implementation situation; regularly evaluating the accuracy of the prediction model and the effectiveness of the resource allocation strategy to ensure the continuous improvement and perfection of the system.

[0128] An embodiment of the present invention, as Figure 2 shown, is an electric vehicle claim settlement full-process collaborative disposal system based on multi-modal data fusion, including a memory, a processor, and a computer program stored on the memory and executable on the memory. The processor executes the program to implement the electric vehicle claim settlement full-process collaborative disposal method based on multi-modal data fusion as described in any one of the above.

[0129] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A collaborative disposal method for the whole process of electric vehicle claims settlement based on multimodal data fusion, characterized in that The method includes: S1. Obtain case data, process it, and publish case information; S2. Automatically match surveyors within the regional scope based on the case information to perform survey tasks and generate survey results; S3. Generate a repair order based on the survey results and push it to each franchised repair shop within the region; S4. After each franchised repair shop receives the repair order, it makes a quotation, and the system determines the claim amount and the repairing repair shop according to the merchant portrait and the merchant's quotation.

2. The method for collaborative disposal of the whole process of electric vehicle claims settlement based on multi-modal data fusion according to claim 1, wherein The S1 includes: S11. Obtain case information related to the involved electric vehicle and create a new case; S12. Preprocess the newly created case; S13. Classify and label the preprocessed case data; S14. Publish the case information in the system.

3. The method for collaborative disposal of the whole process of electric vehicle claim settlement based on multi-modal data fusion according to claim 1, wherein The S2 includes: S21. The system, according to the location information and urgency in the case information, combines the multi-dimensional information of the surveyors, and through the surveyor dynamic matching algorithm based on real-time geographical location and skill tags, automatically matches the optimal surveyor to perform the survey task; S22. The matched surveyor receives the survey task through the APP, views the detailed case information and location information, and plans the survey route to go to the survey site; S23. The surveyor conducts a detailed survey at the survey site and generates survey results; S24. The surveyor uploads the survey results through the APP, and the system conducts a preliminary review and verification of the survey results.

4. The method for collaborative disposal of the whole process of electric vehicle claim settlement based on multimodal data fusion according to claim 1, wherein The S3 includes: S31. The system automatically generates a repair order according to the survey results; S32. The system, according to the location information and repair project type in the repair order, combines the multi-dimensional information of each franchised repair shop within the region, and through the multi-dimensional store ability portrait model, automatically matches the corresponding franchised repair shop; S33. The system pushes the repair order to the matched franchised repair shop.

5. The method for collaborative disposal of the entire process of electric vehicle claim settlement based on multi-modal data fusion according to claim 1, characterized in that The S4 includes: S41. Determine the claim amount and the repairing repair shop by combining each quotation information; S42. Monitor the repair process and control the quality; S43. Claim settlement and settlement.

6. The method for collaborative disposal of the entire process of electric vehicle claims settlement based on multi-modal data fusion according to claim 5, characterized in that, The S41 includes: S411. After each franchised repair shop receives the repair order, it makes a quotation according to the information provided by the system; S412. During the quotation process, the system adopts a distributed quotation isolation mechanism based on privacy computing to physically isolate and store the quotation data of each franchised repair shop; S413. The system conducts a comprehensive evaluation according to the merchant portrait and the merchant's quotation, and automatically calculates and determines the claim amount and the repairing repair shop for the franchised repair shop according to the corresponding algorithm.

7. The method for collaborative disposal of the whole process of electric vehicle claim settlement based on multi-modal data fusion according to claim 6, characterized in that, The S412 includes: Deploy multiple privacy computing nodes on the server side of the system, and each node is responsible for processing the quotation data from different franchised repair shops; Adopt a distributed storage system to store the quotation data from each franchised repair shop, divide the storage system into multiple independent storage areas, each area corresponds to a franchised repair shop, and conduct physical isolation of the data; Before data storage, encrypt the quotation data through symmetric encryption or asymmetric encryption algorithms, store the encrypted data in the corresponding storage area, and record the usage of the encryption key as one of the privacy indicators; The system allocates the privacy computing tasks of the quotation data to appropriate privacy computing nodes according to the current processing capacity, the load conditions of the privacy computing nodes, and the privacy protection requirements; After receiving the tasks, the privacy computing nodes process the quotation data according to predefined algorithms and protocols, and monitor the monitoring metrics during the processing, such as the data access volume and the encryption key usage frequency; Based on the data synchronization mechanism, the quotation data of each affiliated repair shop in the system is synchronously transferred to the distributed storage system in real time; during the data synchronization process, an encrypted transmission protocol is used to encrypt the data transmission process; Monitor the load conditions and privacy metrics of the privacy computing nodes in real time, and dynamically adjust the task allocation strategy according to the load conditions and privacy metrics; When the load of a certain node is too high or the privacy metrics are abnormal, transfer some tasks to other nodes with lower load or stronger protection capabilities; Based on the load warning mechanism, when the overall load of the system approaches the preset threshold, trigger a warning and take corresponding measures, such as automatically adjusting the task allocation or starting the expansion process; Based on the storage resource management mechanism, uniformly manage and schedule the storage resources in the distributed storage system; monitor the usage of the storage resources in real time; dynamically adjust the allocation strategy of the storage resources according to the usage of the storage resources and the privacy protection requirements.

8. The method for collaborative disposal of the whole process of electric vehicle claim settlement based on multi-modal data fusion according to claim 5, wherein The S42 includes: S421. During the repair process, the system monitors the whole process through the whole-process monitoring technology based on the repair progress prediction and records it through the blockchain technology; S422. The repairing repair shop regularly uploads the repair process photos and progress information during the repair process, and the system conducts quality inspection and verification on the uploaded photos and information; S423. If any abnormality or quality problem occurs during the repair process, send out a warning and reminder, and the repairing repair shop processes and adjusts after receiving the warning and reminder.

9. The method for collaborative disposal of the entire process of electric vehicle claims settlement based on multi-modal data fusion according to claim 5, wherein, The S43 includes: S431. After the repair is completed, the repairing shop uploads the repair information, and the system audits and verifies the uploaded information; S432. If the audit is passed, the system will automatically trigger the automatic claim settlement trigger mechanism based on the smart contract, and automatically conduct claim settlement payment according to the preset claim settlement rules and conditions; S433. During the claim settlement process, the system conducts correlation analysis on the repair records, parts codes, and payment vouchers through the multi-source data cross-verification technology; S434. The system records and statistics the data of the whole process and generates a situation statistical report; S435. Predict the situation in the next period of time through the prediction model and make a plan in advance; S436. After completing the claim settlement payment, the system will automatically update the case status and notify the vehicle owner to pick up the vehicle or conduct subsequent processing; S437. At the same time, the system will conduct regular review and summary, and visually display the situation statistics.

10. An electric vehicle claim settlement full-process collaborative disposal system based on multi-modal data fusion, characterized in that, It includes a memory, a processor, and a computer program stored on the memory and executable on the memory. The processor executes the program to implement the full-process collaborative disposal method for electric vehicle claim settlement based on multi-modal data fusion as described in any one of claims 1-9.

Images